Social Work Research

Bright, Charlotte, Lyn

2020 Social Work Research

doi: 10.1093/swr/svaa012

I write these words in early June in the northeastern United States. Incongruously, the sun is shining, birds are chirping, and flowers bloom. This apparent paradise is at odds with the two pandemics gripping this nation and the world at large: coronavirus disease (COVID-19) and racism. Recent weeks and months have seen a string of horrifying events, one after another. The brutal police violence that took the lives of Breonna Taylor and George Floyd and the vigilante murder of Ahmaud Arbery may seem to have little to do with the novel coronavirus that has killed over 400,000 people as of this writing. However, these traumatic, disastrous outcomes are linked by the overarching and systemic racism that affects Black Americans and people of color worldwide. Other scholars and leaders have characterized racism as a pandemic, sooner and more eloquently than I am doing in these pages. In a recent statement, Dr. Sandra Shullman, president of the American Psychological Association (APA), has identified systemic racism as a pandemic that manifests in pervasive trauma, leading to physical and mental health consequences inequitably borne by African Americans (APA, 2020). Last year, the American Academy of Pediatrics made a clear connection between racism and health conditions among children and adolescents (Trent, Dooley, & Dougé, 2019). The American Medical Association (Ehrenfeld & Harris, 2020) has focused attention on the negative public health outcomes of police violence, which affects Black individuals both with and without direct experience of physically or verbally violent acts committed by police. Communities are angry. Protests of police violence have spread far beyond the borders of the United States to such places as Japan and the United Kingdom. Worldwide, there is some solidarity in the frustration and betrayal we feel as citizens when those charged to protect us engage in oppression and violence. The news about police joining protesters in marching, kneeling, and chanting in places like Flint, Michigan, and Camden, New Jersey, underscores the shared humanity in demanding a more equal society. These moments of shared purpose are fleeting, however, especially when paired with instances of police using tear gas, flash grenades, rubber bullets, and batons against protesters. Moreover, demands for justice in individual cases equate to holding perpetrators accountable—through firings, criminal prosecution, and convictions—which are a bitter recourse after deaths that should have been prevented. This type of justice may be both appropriate and necessary, but it represents dissatisfying individual solutions to systemic problems. True justice for George Floyd, Breonna Taylor, and so many others would mean that their deaths never happened. In the earliest days of widespread COVID-19 discussion, I heard many comments about how “We’re all in the same boat.” Although vulnerability to the virus is associated with both age and pre-existing conditions, it is not inherently more dangerous to any particular racial or ethnic group than to others … until we factor in the many ways in which systemic racism perpetuates health disparities as well as disparate working and living conditions. A recent study captured this disparity perfectly: The preponderance of Black Americans in occupations, environments, and situations that increase exposure to the novel coronavirus is not accidental but grounded in the historical and modern-day structural violence of racism. Racism is a form of structural violence because it produces socially unjust conditions that predispose Black communities to disability and death—a reality that is both normalized and reproduced within the practices and policies of enduring public and private institutions. (Poteat, Millett, Nelson, & Beyrer, 2020, p. 6) The pernicious, persistent health disparities that racism and discrimination perpetuate make comments about everyone being in the same boat disingenuous. These comments willfully ignore the very different circumstances people face, not just in the United States but also elsewhere. Most nations probably have their own examples of systemic vulnerability. In Sweden, for example, immigrants from Somalia, Syria, and Iraq comprise a disproportionate number of hospitalizations for COVID-19 (Rothschild, 2020). India and Sri Lanka have seen an uptick in discrimination and violence toward Muslims, who are at times accused of bringing or spreading the virus (Amarasuriya, 2020; Frayer, 2020). In the United Kingdom, a recent analysis found that Black, Asian, and other minority groups have death rates from COVID-19 two to three times higher than their White counterparts (Aldridge et al., 2020). The unequal burden of COVID-19 falls on people of color, immigrants, and religious minorities across the globe. As some astute observers have said, “We may all be weathering the same storm, but we are not all in the same boat.” Systemic racism in a society may be analogous to a systemic illness in the body; it permeates and sickens, and treating the symptoms offers temporary relief at best. Structural solutions are required to resolve this pandemic. My aspiration is that we as scholars will focus our skills and values on a research agenda that centers human rights and social justice, that calls out racism by its name, that uplifts the voices of scholars of color and communities of color while reinforcing the responsibility of majority populations to dismantle White supremacy, and that draws connections between micro and macro effects in terms of experiences, services, and policy. In my last editorial, written as COVID-19 was just beginning to be recognized as a crisis in the United States, I identified a preliminary research agenda for social work scholarship in the era of COVID-19. What was unclear to me at that time was how disproportionate the impact of the virus would be on communities of color. Social work scholars have an ethical obligation to confront systemic racism and discrimination, and to identify and work to resolve health disparities. This is true at all times, but has a particular urgency today. Social problems that were intractable to begin with have become even more concerning in the context of isolation, unemployment, and lack of resources that accompany COVID-19. In the United States, African Americans and other communities of color experience a disproportionate share of these problems. Unemployment is higher among Black versus White men and women in the United States, and Latina women have the highest unemployment rates of all, using data from April 2020. COVID-19-related employment losses exacerbate existing disparities in the experience of poverty (Gould & Wilson, 2020). Social distancing may promote public health, but it disrupts collective coping methods, which are particularly relevant for African Americans (Sharpe, 2015). In these and other ways, the newer pandemic, COVID-19, feeds into intractable problems caused by the ongoing pandemic of racial inequality. Both minority and majority populations have experienced particular risk from COVID-19 and the public health response. Food banks and school lunch programs are overburdened due to exploding demand. Health care is frequently tied to employment in the United States, making access to quality care out of reach for some at exactly the time they may have the greatest need. From school-age children whose educators can no longer monitor their well-being to older adults with chronic health conditions who struggle to access needed care when providers are overwhelmed with COVID-19 patients, the primary concerns of social work research participants have shifted dramatically. Research needs to be nimble enough to capture this moment in time, and to focus attention on the broader context of individual and group behavior. Fortunately, social work scholarship mirrors social work practice in its long history of recognizing the nuances of person-in-environment, and of “beginning where the client (or participant) is.” Social work scholars are better positioned than most to pivot to the unique circumstances in which we and our populations of interest find ourselves. In the age of COVID-19 and the ongoing pandemic of racism, social work scholars are necessary to push back against health disparities, inequitable allocation of resources, unjust policies, and the White supremacist system in which all nations that have been colonized by Europeans inhabit. Social work scholarship can investigate the effectiveness of policy solutions to confront the economic, social, and political realities of COVID-19 and systemic racism. Amid calls for dramatic reforms to policing, social work research can fill a needed gap in knowledge about which of a number of possible reforms actually reduce violence (Engel, McManus, & Isaza, 2020) and participate in academic partnerships with police to facilitate valid data collection and analysis (Engel & Whalen, 2010). Beyond policing reform, social work scholars can and have worked to document the effects of economic and social policy innovations (universal basic income, ban-the-box initiatives, child savings accounts, supports to employers to retain workers) with the potential to increase social stability in times of economic and health crises. The charge that we must dismantle racism, that we should resolve health disparities, and that we need to focus on COVID-19 recovery is daunting. This work requires our focus, energy, creativity, and collaborative efforts. Our own feelings of grief, anger, and helplessness should fuel our desire to identify and disseminate effective strategies and solutions. Social work scholars can bring to bear their skills in community-focused research design and methods, capacity to network with other scholars and community members, knowledge in organizing and social action, and ethics. The impact of these two overlapping pandemics will be long-lasting, and social work research will not only document their effects, but also champion effective solutions. The deaths of Breonna Taylor, George Floyd, and thousands of others are senseless in that they are indefensible. But they are not meaningless. As demonstrations and protests around the globe show, these unjust killings serve as painful reminders of our current reality, even as they call us to action. For social work scholars, part of this call is to be antiracist in our research as well as in our teaching and civic lives. I eagerly anticipate a new wave of scholarship directly focused on ameliorating injustice, dismantling White supremacy, and ushering in a more equitable future. I extend my gratitude to those already doing this work and welcome those who are new to it. Your efforts are very needed. At this time, social work researchers face difficulties gaining access to research materials, meeting with research teams, and connecting with research participants. Social work research thrives on community engagement, and despite the technological advances that keep us connected, opportunities for meeting children and teachers in schools, observing skill-based training programs, and attending grassroots organizing events are sorely missed. Recent antiracism demonstrations have filled this gap in some ways by creating venues for public protest. Alongside these in-person protests, virtual public scholarship has blossomed. I have been astounded at the high-quality webinars, Instagram and Facebook Live events, and media statements social work researchers have contributed. Personally, I have learned a great deal about such topics as antiracist scholarship, behavioral health access in the context of COVID-19, social workers as volunteers, suicide risk and resources, and using new measures and tools to record behavior change. When we think about dissemination and impact, we will do well to remember these shining examples of accessible but rigorous scholarly input. I challenge myself and each one of us to continue the outreach enabled by technology and to participate as both producers and consumers of this kind of knowledge. This is one way we can use our skills to fight the pandemic of racism. I attended a meeting recently in which a common question was “When can we expect to return to normal?” I waited, with baited breath, to learn the date I would experience the magic of this return. To no reader’s surprise, the response was … No one really knows. Despite our collective eagerness to resume scholarship, teaching, and practice as it was before March, it is unclear what the course of COVID-19 in our collective and connected society will be. Will we see peaks and valleys; sporadic and localized spikes; or some other pattern of illness, death, and recovery? Models and forecasts change as our knowledge grows and the assumptions underpinning them evolve. Despite some nations’ standing as “ahead” of others in terms of duration of COVID-19 infections, different communities and countries appear to have diverse trajectories of virus-related illness and death. We have much to learn. I am not the first to observe that “back to normal” implies that life will return to a system characterized by oppressive and racist practices and social, economic, health, and political inequalities. As in the case of public scholarship, it makes sense to consider how we can redesign “normal” based on what we have learned and accomplished. The endeavor of research itself is fraught with inequalities that promote and reward Eurocentric scholarship at the expense of researchers and communities of color (Thorp, 2020). If “back to normal” means White supremacy, if it means disproportionate disease and poverty burden on people of color, and if it means researchers ignore the ethical and social dimensions of their study, we should seek a new normal. Like many of you, I commit to working for a society in which normality means true equality—of access to justice, safety, and health. I look forward to working with you in solidarity toward our shared goals. References Aldridge R. W. , Lewer D. , Katikireddi S. V. , Mathur R. , Pathak N. , Burns R. , et al. ( 2020 ). Black, Asian and minority ethnic groups in England are at increased risk of death from COVID-19: Indirect standardisation of NHS mortality data . Wellcome Open Research, 5 ( 88) , Article 88. Google Scholar Crossref Search ADS WorldCat Amarasuriya H. ( 2020 , May 30). Sri Lanka’s COVID-19 response is proof that demonisation of minorities is normalised. Wire. Retrieved from https://thewire.in/south-asia/sri-lanka-covid-19-demonisation-minorities American Psychological Association. ( 2020 , May 29). “We are living in a racism pandemic,” says APA president [Press Release]. Retrieved from https://www.apa.org/news/press/releases/2020/05/racism-pandemic?fbclid=IwAR38_9yur82elNvLShHRZFrPlXKNhSzYKw-NWaO7YVk6pkOOJHguL9hAilI Ehrenfeld J. M. , Harris P. A. ( 2020 , May 29). Police brutality must stop. Retrieved from https://www.ama-assn.org/about/leadership/police-brutality-must-stop Engel R. S. , McManus H. D. , Isaza G. T. ( 2020 ). Moving beyond “best practice”: Experiences in police reform and a call for evidence to reduce officer-involved shootings . ANNALS of the American Academy of Political and Social Science, 687 ( 1 ), 146 – 165 . Google Scholar Crossref Search ADS WorldCat Engel R. S. , Whalen J. L. ( 2010 ). Police–academic partnerships: Ending the dialogue of the deaf, the Cincinnati experience . Police Practice & Research: An International Journal, 11 ( 2 ), 105 – 116 . Google Scholar Crossref Search ADS WorldCat Frayer L. ( 2020 , April 23). Blamed for coronavirus outbreak, Muslims in India come under attack. NPR. Retrieved from https://www.npr.org/2020/04/23/839980029/blamed-for-coronavirus-outbreak-muslims-in-india-come-under-attack Gould E. , Wilson V. ( 2020 , June 1). Black workers face two of the most lethal preexisting conditions for coronavirus: Racism and economic inequality. Economic Policy Institute. Retrieved from https://www.epi.org/publication/black-workers-covid/ Poteat T. , Millett G. , Nelson L. E. , Beyrer C. ( 2020 ). Understanding COVID-19 risks and vulnerabilities among Black communities in America: The lethal force of syndemics . Annals of Epidemiology, 47 , 1 – 3 . Google Scholar Crossref Search ADS WorldCat Rothschild N. ( 2020 , April 21). The hidden flaw in Sweden’s anti-lockdown strategy. Foreign Policy. Retrieved from https://foreignpolicy.com/2020/04/21/sweden-coronavirus-anti-lockdown-immigrants/ Sharpe T. L. ( 2015 ). Understanding the sociocultural context of coping for African American family members of homicide victims: A conceptual model. Trauma, Violence, & Abuse, 16 ( 1 ), 48 – 59 . Google Scholar Crossref Search ADS WorldCat Thorp H. H. ( 2020 , June 8). Time to look in the mirror [Editorial]. Science. doi:10.1126/science.abd1896 Trent M. , Dooley D. G. , Dougé J. ( 2019 ). The impact of racism on child and adolescent health . Pediatrics, 144 ( 2 ), e20191765 . Google Scholar Crossref Search ADS WorldCat © 2020 National Association of Social Workers This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Reed, Lauren, A;McCullough Cosgrove,, Jenny;Sharkey, Jill, D;Felix,, Erika

2020 Social Work Research

doi: 10.1093/swr/svaa011

Abstract Digital dating abuse (DDA), which is the use of social media and mobile phones to abuse a dating partner, is a common and harmful form of dating violence among youths. To date, this issue has not been studied among Latinx youths. The current study examined DDA among a sample of 70 Latinx youths with dating experience, using survey data on participants’ experiences with traditional forms of offline dating violence, DDA victimization and perpetration, healthy relationship knowledge, and self-reported conflict resolution skills. Results showed that Latinx youths experienced DDA and that there was a strong link between DDA and offline forms of dating abuse. Most participants reported high levels of healthy relationship knowledge and conflict resolution skills, but results indicated a link between DDA experience and fewer positive conflict resolution behaviors. The study has implications for the assessment and prevention of DDA among diverse populations of youths, supports the incorporation of conflict resolution skills in dating violence prevention efforts, and calls for future research on the cultural context of DDA among Latinx youths. Teenage dating abuse (TDA) poses a significant public health risk to youths in the United States. TDA involves a repeated pattern of behaviors to exert power and control over a dating partner, and comes in many forms including psychological, physical, and sexual abuse (Olsen, Vivolo-Kantor, & Kann, 2017). A recent meta-analysis found that among teenage romantic partners, physical dating violence, such as hitting and kicking, occurs in 20% of teenage relationships and forced sexual activity was reported in about 10% of relationships (Wincentak, Connolly, & Card, 2017). The use of digital media technology in teenage interpersonal interactions is ubiquitous in the United States, with 95% of teenagers reporting that they have access to a smartphone (Anderson & Jiang, 2018). As digital media technologies, including social media and mobile phone use, have become an increasingly important means of communication for youths in dating relationships, these media have also come under investigation as a potential context and tool for dating violence (Henry & Powell, 2018). When social media and mobile phones are used in a pattern of behavior to harass, pressure, coerce, or threaten a current or former dating partner, it is called digital dating abuse (DDA) (Guerra-Reyes et al., 2017; Reed, Tolman, & Ward, 2017). Youths who experience TDA in their face-to-face relationships are also likely to experience DDA (Kernsmith, Victor, & Smith-Darden, 2018; Temple et al., 2016). As research on DDA emerges, most data come from majority-White youths and young adult populations. The current study sought to focus on understanding DDA among Latinx youths (Latinx is a gender-inclusive term for people of Latin American origin or descent), an underrepresented and underserved youth population in the United States. Given that the dating experiences of Latinx youths may vary when compared with their White peers, it is likely that their digital dating experiences may vary too (Vagi, O’Malley Olsen, Basile, & Vivolo-Kantor, 2015). Despite implications for how social workers, educators, and mental health professionals understand and support Latinx youths’ healthy dating relationships, to our knowledge, studies have yet to explicitly examine DDA among this population. Therefore, the current study explored the frequency and experience of DDA among a sample of Latinx youths, including their knowledge of relationship skills and conflict resolution as a possible target for DDA prevention. DDA among Teenagers DDA is a common and harmful form of TDA. DDA behaviors include monitoring or controlling behaviors, directly aggressive behaviors, and sexual coercion or unwanted sexual behaviors (Reed et al., 2017). Specific behaviors include pressuring a partner for sexual photos (sexting), sending or posting mean or hurtful public messages about your partner, or looking at a partner’s private digital information without their permission. Recent surveys of high school–age youths found that a quarter reported experiencing DDA (Korchmaros, Ybarra, Langhinrichsen-Rohling, Boyd, & Lenhart, 2013; Zweig, Dank, Yahner, & Lachman, 2013). The potential consequences of experiencing these behaviors include declines in physical and mental health and well-being (Zweig, Lachman, Yahner, & Dank, 2014). Therefore, researchers and practitioners should focus on digital contexts to understand the full scope of psychological dating abuse. Previous research suggests that girls and boys experience DDA at about equal rates, but that the experience and context of these DDA experiences differ (Reed et al., 2017; Yahner, Dank, Zweig, & Lachman, 2015; Zweig et al., 2014). For example, girls are more likely to experience pressure to engage in sexting and are also more likely to be upset by DDA experiences than boys (Kernsmith et al., 2018; Reed et al., 2017). Thus, we examined gender differences in DDA among Latinx youths. Dating, Dating Violence, and Relationship Communication among Latinx Youths Latinx youths experience specific and unique familial expectations around dating. Previous research has shown that Latinx families attribute familial honor to a daughter’s sexual reputation (Rueda, Nagoshi, & Williams, 2014). In a qualitative study of 75 youths, interviewed in groups separated by gender and ethnicity, Latinx participants reported that they were forbidden to date and kept their romantic relationships a secret from parents (Rueda et al., 2014). In another study, Latinx female youths were more likely than youths from other ethnic backgrounds to avoid any conversations with their parents about healthy relationships (McCullough Cosgrove, LeCroy, Fordney, & Voelkel, 2018). Given the possible greater taboo of adolescent romantic relationships (especially for girls) and lack of communication about romantic relationships with their parents, Latinx youths may be less likely to discuss unhealthy relationship behaviors or dating abuse with their parents. Therefore, the current study sought to examine Latinx adolescents’ knowledge about healthy relationship skills and conflict resolution, and whether this knowledge is associated with experiences of DDA. Research on dating communication and dating violence among Latinx youths remains an understudied topic. Although no studies have focused specifically on DDA among Latinx youths, some larger studies included a significant minority of Latinx youths in their sample. For example, a study of 1,008 youths with 36.5% identifying as Latinx found that 13% of all participants reported experiencing digital monitoring behaviors and digital pressure to talk about sex from their partner in the past three months (Dick et al., 2014). A more recent study in which the participants were 37% Latinx youths found that 51% of all those surveyed reported experiencing DDA and 32% reported DDA perpetration (Cutbush, Williams, Miller, Gibbs, & Clinton-Sherrod, 2018). Although these studies did not disaggregate their findings by racial or ethnic group, results suggest that DDA is an issue for Latinx youths. The few studies that focus on offline forms of TDA among Latinx adolescents find that they experience TDA at rates similar to those of their White peers (Sabina, Cuevas, & Cotignola-Pickens, 2016) with sexual TDA being experienced at significantly higher rates (Kann et al., 2014). However, other research found that physical TDA was higher for Latinx students than non-Latinx White students (Wechsler, 2012). Approximately 14% to 19.5% of Latinx youths reported experiencing offline dating abuse behaviors (Kann et al., 2014; Sabina et al., 2016). Latinx youths have widespread access to digital media technology; a recent Pew Research study found that 95% of Latinx youths report having access to a smartphone and 82% have access to a computer at home (Anderson & Jiang, 2018). Given differences in the experiences of TDA among Latinx youths, combined with the widespread use and increasing social importance of digital communication, it is imperative to examine DDA among Latinx youths. Current Study The current study is the first, to our knowledge, to explore the experience of  DDA among a sample of Latinx youths. This study explored the frequency of three types of DDA: digital direct aggression, digital monitoring and control, and digital sexual abuse behaviors. We also explored participants’ knowledge of healthy relationship and conflict resolution skills. Our four research questions and hypotheses are (1) For Latinx youths, how common are the three forms of DDA? Are there gender differences in DDA victimization and perpetration? We expected that rates of reported DDA would be similar among our sample of  Latinx youths to those in other cultural groups, as studies including a significant portion of Latinx youths found that DDA was common. Most DDA research finds equal rates of DDA frequency among girls and boys, therefore, we expected this to be the same for Latinx youths. (2) What is the association between online and offline dating abuse? We hypothesized that like other cultural groups, Latinx youths would report strong positive associations between online and offline dating abuse. (3) How much dating relationship knowledge and conflict resolution skills do Latinx youths report? (4) Are relationship knowledge and self-reported conflict resolution skills related to their DDA experience? The third and fourth research questions are largely exploratory. The goal of this study was to explore DDA experiences within a sample of Latinx youths and examine their relationship knowledge and conflict resolution skills as potential areas of future social work intervention to prevent DDA. Method Participants High school students participated in a quasi-experimental study assessing the effectiveness of a school-based TDA prevention program on the Central Coast of California. A total of 174 students were recruited to participate in the program, and 143 of these students completed baseline surveys. Among those, 70 students identified as Latinx or Hispanic and reported dating experience. This sample of 70 students was used for all analyses in the current study. From the sample of 70 Latinx students with dating experience, most participants identified as girls (73.1%), one student identified as genderqueer or transgender, and participants ranged in age from 14 to 18 years (M = 15.65). Students were in grades 9 (35.7%), 10 (28.6%), 11 (15.7%), or 12 (20%). Most students (65.2%) participated in a free or reduced-price lunch program and most (78.6%) reported dating or hooking up only with the opposite sex. Procedure The study was conducted in three large public high schools during the 2015–2016 school year and was approved by a university institutional review board for ethical research with human subjects. These schools agreed to take part in a TDA prevention program, and participants were recruited into the program and this study through advertisements in the school, a school assembly, and referrals by teachers and school counselors. Parent or guardian consent was obtained to participate in the study. Participation in the TDA prevention program was voluntary, and parents could opt out of the research study and still request to have their child participate in the program. Few parents (n = 8) opted out of the research study. Participation was voluntary and confidential. Students were asked to complete the online survey using school computers or school-provided iPads, under supervision of the research team during one 50-minute class period. Students were instructed that they could withdraw from the study at any time and could skip any question that they did not feel comfortable answering. Measures Dating Experience All items in this section were created for or modified from our previous research (Reed et al., 2017) for use in this study. Participants were asked, “Are you currently in a dating relationship?” with yes or no as response options. If the response was affirmative, they were asked, “How long have you been in this relationship?’ with response options ranging from 1 = less than a month to 5 = more than a year. Participants were also asked, “How old is your current or most recent dating partner?” with response options “The same age I am,” “One year younger than I am,” “One year older than I am,” “More than one year older than I am,” and “More than one year younger than I am.” Participants also reported the number of dating partners in the past year, with possible responses from 1 to 6+. Finally, participants were asked about their gender identification with options including girl/woman, boy/man, and transgender/genderqueer. Participants were then asked, “What is the gender of the people you typically date or hook up with?” with responses including only girls/women, mostly girls/women, both girls/women and boys/men, mostly boys/men, and only boys/men. These items were used to determine whether participants had same-sex dating or sexual experience. Healthy Relationship Knowledge The six-item Perceived Relationship Knowledge Scale (Bradford, Wade Stewart, Higginbotham, & Skogrand, 2015) assessed healthy relationship knowledge. Example items included, “My knowledge of how to listen effectively to a partner” and “My awareness of the importance of spending time together,” with the following response options provided: 1 = was/is poor, 2 = was/is fair, 3 = was/is good, and 4 = was/is excellent (α = .85 for the current sample). Items were modified to fit an adolescent sample, in that the word “spouse” was removed from some items that indicated “spouse/partner.” Conflict Resolution Skills A modified and expanded version of the Healthy Conflict Resolution in Peer and Dating Relationships Scale was used (Ball et al., 2012). This measure assesses frequency of 10 healthy conflict resolution behaviors when having an argument with a close friend or dating partner. We added 14 unhealthy conflict resolution behaviors to assess the frequency of both healthy and unhealthy conflict resolution behaviors when having an argument with a dating partner. For the adapted 24-item measure, participants were given the prompt: “In your current/most recent relationship, when you have a fight, how often do you do the following?” with items such as “Put off talking until we both calm down” and “Listen to their side of the story” from the original measure and, “Blame the other person for the problem” and “Yell at them,” which were added. Responses ranged from 1 = never to 5 = always (α = .86 for the current sample). Dating Violence Victimization and Perpetration Experience with physical, sexual, and psychological TDA was assessed using measures from Foshee et al. (1998) that were designed for a school-based TDA prevention program evaluation. Participants were given the prompt: “How often has your current or most recent dating partner done the following things to you?” and then were asked about psychological abuse, physical abuse, and sexual abuse. All Cronbach’s alphas shown are for the current sample. Psychological abuse victimization was measured using 13 items, such as “Said things to hurt my feelings on purpose” and “Threatened to hurt me” (α = .95). Parallel items were used to measure perpetration (α = .64). Physical abuse victimization was measured using 14 items. Participants were given a similar prompt but told to only include it when the dating partner “did it to you first” and not count self-defense, with items including “Slapped me” and “Tried to choke me” (α = .89). Parallel items were used to measure perpetration (α = .75). Sexual abuse victimization was measured using two items from the Foshee et al. (1998) measure and one item from Zweig et al. (2014). Participants answered if their partner forced them into sexual activities or pressured them to have sex when they did not want to (α = .86). For all subscales, response options ranged from 0 = never to 3 = very often. Parallel items measured perpetration, but no students reported perpetrating sexual abuse. DDA Victimization and Perpetration DDA was assessed with a 36-item measure (Reed et al., 2017; Reed, Ward, Tolman, Lippman, & Seabrook, 2018) of victimization and perpetration of three types of DDA behaviors: (1) digital direct aggression, (2) digital monitoring and control, and (3) digital sexual abuse. For victimization, participants were given the following prompt: “How often has your current or most recent dating partner done the following things to you using the Internet or a cell phone?” Response options ranged from 0 = never to 3 = very often. All Cronbach’s alphas shown are for the current sample. The Digital Direct Aggression Victimization subscale (eight items, α = .79) measured digital behaviors meant to hurt, humiliate, or threaten a dating partner using social media or a mobile phone (for example, “Sent me a threatening message” and “Spread rumors about me”). Parallel items measured perpetration (α = .34). One item from this scale was removed from reliability analysis (“Threatened to harm my partner physically”) because it was not reported by any of the participants. The Digital Monitoring and Control Victimization subscale (six items, α = .91) involved using social media or mobile phones to keep track, intrude on the privacy, and control the activities and relationships of a dating partner (for example, “Monitored who I talk to and am friends with” and “Looked at my private information [text messages, e-mails, etc.] to check up on me without permission.”) Parallel items measured perpetration (α = .74). The Digital Sexual Abuse Victimization subscale (four items, α = .82) included pressuring a dating partner for online or offline sexual behavior and engaging in unwanted distribution of sexual images (for example, “Sent me a sexual or naked photo that I did not ask for” and “Pressured me to sext”). Parallel items measured perpetration (α = .76). Results Students were selected because they had dating experience, and 34.3% of students were currently in a dating relationship at the time of survey, with most in this relationship for less than a year (83.3%). About half of participants (53.2%) reported having one dating partner in the last calendar year, 30.6% reported two partners, 8.1% reported three partners, 3.2% reported five partners, and 4.8% (n = 3) reported having six or more dating partners. Dating Violence and DDA Experience among Latinx Youths Analyses showed that all three types of offline TDA and all three types of DDA were common among this sample of  Latinx youths in their current or most recent relationship as compared with rates in other cultural groups. Table 1 displays the prevalence and mean frequency of reported TDA and DDA experiences for all subscales. T tests revealed no significant gender differences for both TDA and DDA subscales, but there were some differences by individual item. To correct for multiple tests, a p value of .01 was selected to minimize Type II error. Table 2 provides the frequency of each type of DDA behavior by gender. The most commonly reported types of DDA victimization behaviors among this sample of Latinx youths were pressure to sext (27.2%), pressure to have sex or do other sexual activities (27.0%), and most of the digital monitoring and control behaviors (between 19.6% and 33.7%). There were gender differences in some individual DDA victimization behaviors that met or approached our adjusted statistical significance. Girls were more likely to report that their partner sent a mean or hurtful private message, t(63.97) = 2.64, p = .010 and that their partner looked at their private digital information to check up on them without permission, t(62.07) = 2.81, p = .007. There was only one significant gender difference in the perpetration of individual DDA behaviors. Girls were found to be more likely than boys to report pressuring their partner to sext, t(48) = 2.86, p = .006, as no boys reported pressuring their partner to sext. This indicates that although there are no gender differences overall in the frequency reported of each DDA subscale, several individual behaviors occur more often among Latinx girls. Table 1: Rates and Frequency of Teenage Dating Abuse and Digital Dating Abuse among Latinx Youths in Current or Most Recent Relationship (N = 72) Abuse Type . Percent of Sample Experienced . Frequency M (SD) . Frequency Range  (0–3) . Dating violence experiences  Sexual abuse   Perpetration 0.0 — —   Victimization 28.6 0.25 (0.60) 0–3  Physical abuse   Perpetration 24.6 0.05 (0.12) 0–0.64   Victimization 40.0 0.14 (0.29) 0–1.64  Psychological abuse   Perpetration 44.9 0.11 (0.16) 0–0.62   Victimization 59.4 0.40 (0.63) 0–2.77 Digital dating abuse experiences  Digital sexual abuse   Perpetration 18.6 0.10 (0.30) 0–1.75   Victimization 37.1 0.37 (0.65) 0–3  Digital direct aggression   Perpetration 36.2 0.10 (0.16) 0–0.5   Victimization 44.3 0.23 (0.40) 0–2.13  Digital monitoring/control   Perpetration 53.6 0.28 (0.41) 0–2.00   Victimization 57.1 0.58 (0.77) 0–2.67 Abuse Type . Percent of Sample Experienced . Frequency M (SD) . Frequency Range  (0–3) . Dating violence experiences  Sexual abuse   Perpetration 0.0 — —   Victimization 28.6 0.25 (0.60) 0–3  Physical abuse   Perpetration 24.6 0.05 (0.12) 0–0.64   Victimization 40.0 0.14 (0.29) 0–1.64  Psychological abuse   Perpetration 44.9 0.11 (0.16) 0–0.62   Victimization 59.4 0.40 (0.63) 0–2.77 Digital dating abuse experiences  Digital sexual abuse   Perpetration 18.6 0.10 (0.30) 0–1.75   Victimization 37.1 0.37 (0.65) 0–3  Digital direct aggression   Perpetration 36.2 0.10 (0.16) 0–0.5   Victimization 44.3 0.23 (0.40) 0–2.13  Digital monitoring/control   Perpetration 53.6 0.28 (0.41) 0–2.00   Victimization 57.1 0.58 (0.77) 0–2.67 Note: For the range of mean frequency, 0 = never happened and 3 = very often. Open in new tab Table 1: Rates and Frequency of Teenage Dating Abuse and Digital Dating Abuse among Latinx Youths in Current or Most Recent Relationship (N = 72) Abuse Type . Percent of Sample Experienced . Frequency M (SD) . Frequency Range  (0–3) . Dating violence experiences  Sexual abuse   Perpetration 0.0 — —   Victimization 28.6 0.25 (0.60) 0–3  Physical abuse   Perpetration 24.6 0.05 (0.12) 0–0.64   Victimization 40.0 0.14 (0.29) 0–1.64  Psychological abuse   Perpetration 44.9 0.11 (0.16) 0–0.62   Victimization 59.4 0.40 (0.63) 0–2.77 Digital dating abuse experiences  Digital sexual abuse   Perpetration 18.6 0.10 (0.30) 0–1.75   Victimization 37.1 0.37 (0.65) 0–3  Digital direct aggression   Perpetration 36.2 0.10 (0.16) 0–0.5   Victimization 44.3 0.23 (0.40) 0–2.13  Digital monitoring/control   Perpetration 53.6 0.28 (0.41) 0–2.00   Victimization 57.1 0.58 (0.77) 0–2.67 Abuse Type . Percent of Sample Experienced . Frequency M (SD) . Frequency Range  (0–3) . Dating violence experiences  Sexual abuse   Perpetration 0.0 — —   Victimization 28.6 0.25 (0.60) 0–3  Physical abuse   Perpetration 24.6 0.05 (0.12) 0–0.64   Victimization 40.0 0.14 (0.29) 0–1.64  Psychological abuse   Perpetration 44.9 0.11 (0.16) 0–0.62   Victimization 59.4 0.40 (0.63) 0–2.77 Digital dating abuse experiences  Digital sexual abuse   Perpetration 18.6 0.10 (0.30) 0–1.75   Victimization 37.1 0.37 (0.65) 0–3  Digital direct aggression   Perpetration 36.2 0.10 (0.16) 0–0.5   Victimization 44.3 0.23 (0.40) 0–2.13  Digital monitoring/control   Perpetration 53.6 0.28 (0.41) 0–2.00   Victimization 57.1 0.58 (0.77) 0–2.67 Note: For the range of mean frequency, 0 = never happened and 3 = very often. Open in new tab Table 2: Percentage of Digital Dating Abuse (DDA) Behavior Reporting by Subscale and Gender . . Girls (n = 49) . Boys (n = 17) . DDA Subscale . Total Victimization . Victimization . Perpetration . Victimization . Perpetration . Digital sexual abuse  Pressured to sext 27.2 30.6 16.3** 17.6 0.0  Sent a sexual/naked photo that the partner did not want/ask for 18.5 16.7 10.4 5.9 5.9  Sent a sexual or naked photo/video to others without permission 5.4 4.1 2.0 5.9 6.2  Pressured to have sex or do other sexual activities 27.0 33.3 4.2 11.8 0.0  Total 37.1 36.7 20.4 29.4 11.8 Digital direct aggression  Shared an embarrassing photo or video with others without permission 20.7 14.3 16.7 23.5 29.4  Sent a mean or hurtful private message 24.2 28.6** 18.7 11.8 17.6  Posted a mean or hurtful public message 8.7 10.2 8.3 0.0 0.0  Spread a rumor 19.8 16.3 4.2 18.7 5.9  Sent a threatening message 12.1 16.3 2.1 6.2 0.0  Threatened to physically harm 10.1 14.9** 0.0 0.0 0.0  Used cell phone or online account to pretend to be me/my partner 8.8 8.2 0.0 0.0 5.9  Used information from a social networking site to tease or put down 6.6 6.1 4.2 0.0 5.9  Total 44.3 44.9 36.7 35.3 41.2 Digital monitoring and control  Pressured to respond quickly to calls, texts, or other messages 33.7 36.7 25.0 29.4 23.5  Monitored whereabouts and activities 33.3 33.3 29.8 31.2 23.5  Sent so many messages that I/my partner felt uncomfortable 32.6 36.7 6.2 29.4 0.0  Pressured for passwords to access cell phone or online accounts 20.9 22.4 8.3 13.3 6.2  Looked at private information to check up on me/my partner without permission 19.6 24.5** 12.5 6.2 11.8  Monitored who I/my partner talks to/is friends with 41.3 44.9 29.2 47.1 41.2  Total 57.1 57.1 53.1 58.8 58.8 . . Girls (n = 49) . Boys (n = 17) . DDA Subscale . Total Victimization . Victimization . Perpetration . Victimization . Perpetration . Digital sexual abuse  Pressured to sext 27.2 30.6 16.3** 17.6 0.0  Sent a sexual/naked photo that the partner did not want/ask for 18.5 16.7 10.4 5.9 5.9  Sent a sexual or naked photo/video to others without permission 5.4 4.1 2.0 5.9 6.2  Pressured to have sex or do other sexual activities 27.0 33.3 4.2 11.8 0.0  Total 37.1 36.7 20.4 29.4 11.8 Digital direct aggression  Shared an embarrassing photo or video with others without permission 20.7 14.3 16.7 23.5 29.4  Sent a mean or hurtful private message 24.2 28.6** 18.7 11.8 17.6  Posted a mean or hurtful public message 8.7 10.2 8.3 0.0 0.0  Spread a rumor 19.8 16.3 4.2 18.7 5.9  Sent a threatening message 12.1 16.3 2.1 6.2 0.0  Threatened to physically harm 10.1 14.9** 0.0 0.0 0.0  Used cell phone or online account to pretend to be me/my partner 8.8 8.2 0.0 0.0 5.9  Used information from a social networking site to tease or put down 6.6 6.1 4.2 0.0 5.9  Total 44.3 44.9 36.7 35.3 41.2 Digital monitoring and control  Pressured to respond quickly to calls, texts, or other messages 33.7 36.7 25.0 29.4 23.5  Monitored whereabouts and activities 33.3 33.3 29.8 31.2 23.5  Sent so many messages that I/my partner felt uncomfortable 32.6 36.7 6.2 29.4 0.0  Pressured for passwords to access cell phone or online accounts 20.9 22.4 8.3 13.3 6.2  Looked at private information to check up on me/my partner without permission 19.6 24.5** 12.5 6.2 11.8  Monitored who I/my partner talks to/is friends with 41.3 44.9 29.2 47.1 41.2  Total 57.1 57.1 53.1 58.8 58.8 Note: Figures shown in bold represent significant gender differences, with greater mean frequency bolded. ** p < .01. Open in new tab Table 2: Percentage of Digital Dating Abuse (DDA) Behavior Reporting by Subscale and Gender . . Girls (n = 49) . Boys (n = 17) . DDA Subscale . Total Victimization . Victimization . Perpetration . Victimization . Perpetration . Digital sexual abuse  Pressured to sext 27.2 30.6 16.3** 17.6 0.0  Sent a sexual/naked photo that the partner did not want/ask for 18.5 16.7 10.4 5.9 5.9  Sent a sexual or naked photo/video to others without permission 5.4 4.1 2.0 5.9 6.2  Pressured to have sex or do other sexual activities 27.0 33.3 4.2 11.8 0.0  Total 37.1 36.7 20.4 29.4 11.8 Digital direct aggression  Shared an embarrassing photo or video with others without permission 20.7 14.3 16.7 23.5 29.4  Sent a mean or hurtful private message 24.2 28.6** 18.7 11.8 17.6  Posted a mean or hurtful public message 8.7 10.2 8.3 0.0 0.0  Spread a rumor 19.8 16.3 4.2 18.7 5.9  Sent a threatening message 12.1 16.3 2.1 6.2 0.0  Threatened to physically harm 10.1 14.9** 0.0 0.0 0.0  Used cell phone or online account to pretend to be me/my partner 8.8 8.2 0.0 0.0 5.9  Used information from a social networking site to tease or put down 6.6 6.1 4.2 0.0 5.9  Total 44.3 44.9 36.7 35.3 41.2 Digital monitoring and control  Pressured to respond quickly to calls, texts, or other messages 33.7 36.7 25.0 29.4 23.5  Monitored whereabouts and activities 33.3 33.3 29.8 31.2 23.5  Sent so many messages that I/my partner felt uncomfortable 32.6 36.7 6.2 29.4 0.0  Pressured for passwords to access cell phone or online accounts 20.9 22.4 8.3 13.3 6.2  Looked at private information to check up on me/my partner without permission 19.6 24.5** 12.5 6.2 11.8  Monitored who I/my partner talks to/is friends with 41.3 44.9 29.2 47.1 41.2  Total 57.1 57.1 53.1 58.8 58.8 . . Girls (n = 49) . Boys (n = 17) . DDA Subscale . Total Victimization . Victimization . Perpetration . Victimization . Perpetration . Digital sexual abuse  Pressured to sext 27.2 30.6 16.3** 17.6 0.0  Sent a sexual/naked photo that the partner did not want/ask for 18.5 16.7 10.4 5.9 5.9  Sent a sexual or naked photo/video to others without permission 5.4 4.1 2.0 5.9 6.2  Pressured to have sex or do other sexual activities 27.0 33.3 4.2 11.8 0.0  Total 37.1 36.7 20.4 29.4 11.8 Digital direct aggression  Shared an embarrassing photo or video with others without permission 20.7 14.3 16.7 23.5 29.4  Sent a mean or hurtful private message 24.2 28.6** 18.7 11.8 17.6  Posted a mean or hurtful public message 8.7 10.2 8.3 0.0 0.0  Spread a rumor 19.8 16.3 4.2 18.7 5.9  Sent a threatening message 12.1 16.3 2.1 6.2 0.0  Threatened to physically harm 10.1 14.9** 0.0 0.0 0.0  Used cell phone or online account to pretend to be me/my partner 8.8 8.2 0.0 0.0 5.9  Used information from a social networking site to tease or put down 6.6 6.1 4.2 0.0 5.9  Total 44.3 44.9 36.7 35.3 41.2 Digital monitoring and control  Pressured to respond quickly to calls, texts, or other messages 33.7 36.7 25.0 29.4 23.5  Monitored whereabouts and activities 33.3 33.3 29.8 31.2 23.5  Sent so many messages that I/my partner felt uncomfortable 32.6 36.7 6.2 29.4 0.0  Pressured for passwords to access cell phone or online accounts 20.9 22.4 8.3 13.3 6.2  Looked at private information to check up on me/my partner without permission 19.6 24.5** 12.5 6.2 11.8  Monitored who I/my partner talks to/is friends with 41.3 44.9 29.2 47.1 41.2  Total 57.1 57.1 53.1 58.8 58.8 Note: Figures shown in bold represent significant gender differences, with greater mean frequency bolded. ** p < .01. Open in new tab Association between Online and Offline Dating Violence Zero-order correlation analyses (see Table 3) examined whether the frequency of online DDA experience was associated with offline TDA experience among this sample of Latinx youths. To correct for multiple tests, we restricted the p value to .01 to minimize Type II error. As hypothesized, there was a strong positive correlation overall between the three DDA subscales and offline forms of dating violence, with the exception of sexual abuse perpetration. Digital sexual abuse perpetration and digital direct aggression perpetration showed little to no association with offline forms of  TDA among these Latinx youths. Table 3: Zero-Order Correlations between Digital Dating Abuse Experiences, Offline Teenage Dating Abuse Experiences, Healthy Relationships Knowledge, and Conflict Resolution Skills . Digital Sexual Abuse . Digital Direct Aggression . Digital Monitoring/Control . Subscale . Victimization . Perpetration . Victimization . Perpetration . Victimization . Perpetration . Psychological victimization .638*** .243 .860*** .510*** .807*** .567*** Psychological perpetration .343** .049 .329** .643*** .389** .268 Physical victimization .354** .019 .692*** .179 .588*** .495*** Physical perpetration .184 −.017 .562*** .181 .417*** .368** Sexual victimization .584*** .068 .822*** .261 .660*** .351** Sexual perpetration Relationship knowledge –.084 .039 –.211 –.272 –.137 .055 Conflict resolution skills –.177 –.026 –.355** –.260 –.366** –.059 . Digital Sexual Abuse . Digital Direct Aggression . Digital Monitoring/Control . Subscale . Victimization . Perpetration . Victimization . Perpetration . Victimization . Perpetration . Psychological victimization .638*** .243 .860*** .510*** .807*** .567*** Psychological perpetration .343** .049 .329** .643*** .389** .268 Physical victimization .354** .019 .692*** .179 .588*** .495*** Physical perpetration .184 −.017 .562*** .181 .417*** .368** Sexual victimization .584*** .068 .822*** .261 .660*** .351** Sexual perpetration Relationship knowledge –.084 .039 –.211 –.272 –.137 .055 Conflict resolution skills –.177 –.026 –.355** –.260 –.366** –.059 Note: No participants reported sexual abuse perpetration. ** p < .01. ***p < .000. Open in new tab Table 3: Zero-Order Correlations between Digital Dating Abuse Experiences, Offline Teenage Dating Abuse Experiences, Healthy Relationships Knowledge, and Conflict Resolution Skills . Digital Sexual Abuse . Digital Direct Aggression . Digital Monitoring/Control . Subscale . Victimization . Perpetration . Victimization . Perpetration . Victimization . Perpetration . Psychological victimization .638*** .243 .860*** .510*** .807*** .567*** Psychological perpetration .343** .049 .329** .643*** .389** .268 Physical victimization .354** .019 .692*** .179 .588*** .495*** Physical perpetration .184 −.017 .562*** .181 .417*** .368** Sexual victimization .584*** .068 .822*** .261 .660*** .351** Sexual perpetration Relationship knowledge –.084 .039 –.211 –.272 –.137 .055 Conflict resolution skills –.177 –.026 –.355** –.260 –.366** –.059 . Digital Sexual Abuse . Digital Direct Aggression . Digital Monitoring/Control . Subscale . Victimization . Perpetration . Victimization . Perpetration . Victimization . Perpetration . Psychological victimization .638*** .243 .860*** .510*** .807*** .567*** Psychological perpetration .343** .049 .329** .643*** .389** .268 Physical victimization .354** .019 .692*** .179 .588*** .495*** Physical perpetration .184 −.017 .562*** .181 .417*** .368** Sexual victimization .584*** .068 .822*** .261 .660*** .351** Sexual perpetration Relationship knowledge –.084 .039 –.211 –.272 –.137 .055 Conflict resolution skills –.177 –.026 –.355** –.260 –.366** –.059 Note: No participants reported sexual abuse perpetration. ** p < .01. ***p < .000. Open in new tab Association between Relationship Skills and DDA Experience Latinx youths in this sample reported high levels of healthy dating relationship knowledge (M = 3.10, SD = 0.60) and conflict resolution skills (M = 3.10, SD = 0.60). T tests examined gender differences, and no significant gender differences were found. See Table 3 for zero-order correlations between healthy dating relationship knowledge, conflict resolution skills, and DDA experience. To correct for multiple tests, we restricted the p value to .01 to minimize Type II error. As expected, there was a significant negative association between both digital direct aggression victimization and digital monitoring and control, and conflict resolution skills. Participants who reported that their partner perpetrated more frequent digital direct aggression or digital monitoring control in their relationship were less likely to use positive conflict resolution skills during an argument with their partner. We split the sample by gender and ran the zero-order correlations again to determine if there were different patterns of associations for Latinx girls and boys. For Latinx girls, the negative association between digital direct aggression and conflict resolution skills remained, r(49) = –.363, p = .010, as did the negative association between digital monitoring and control victimization and conflict resolution skills, r(49) = –.446, p = .001. No significant correlations remained between DDA, healthy relationship knowledge, and conflict resolution skills for boys. Discussion This study explored reports of DDA, associations between online and offline TDA, and healthy relationship knowledge and conflict resolution skills among 70 Latinx high school students from the Central Coast of California. The TDA victimization rates in this convenience sample of Latinx youths were 28.6% for sexual abuse, 40% for physical abuse, and 59.4% for psychological abuse, which were higher than rates reported in a recent meta-analysis of teenage romantic partners (for example, 20% experienced physical dating violence; 10% experienced forced sexual activity) (Wincentak et al., 2017). Similarly, DDA rates of 37.1% for digital sexual abuse, 44.3% for digital direct aggression, and 57.1% for digital monitoring and control in this sample were much higher than 25% of DDA found in past surveys (see, for example, Korchmaros et al., 2013), although other research has found similar rates of these DDA types (see, for example, Reed et al., 2017). There are a few possible reasons for higher-than-expected rates of dating violence and DDA in this sample. First, there might have been selection bias related to how participants were recruited for this study. This sample was drawn from students who elected to, or were directly recruited into, participation in a school-based TDA prevention program. The Latinx youths were drawn from one community, and thus do not represent all Latinx youths in the United States. Prior research found higher rates of offline physical (Wechsler, 2012) and sexual dating violence (Kann et al., 2014) for Latinx students than non-Latinx White students. Thus, it is possible that Latinx youths are more likely than others to experience TDA and DDA. However, before conclusions can be drawn, these findings should be replicated with a representative sample of  Latinx youths. TDA and DDA perpetration rates mirrored or were lower than the rates of victimization in each category. Lower rates of  DDA perpetration compared with victimization is consistent with previous studies (see, for example, Reed et al., 2017; Zweig et al., 2013). Self-reported rates of offline sexual abuse perpetration were lower than rates of victimization; no participants reported perpetuating sexual abuse, despite 28.6% of participants reporting experiencing sexual abuse victimization. This discrepancy suggests that participants drawn to this study are either less likely to engage in sexual abuse perpetration, or more influenced by social desirability for sexual abuse perpetration than for other TDA perpetration behaviors. Gender and DDA There were no statistically significant gender differences in experience of the three subscale types of DDA. However, these results should be interpreted with caution, as inferences were drawn from a relatively small sample size of 70 Latinx youths. Reports of digital sexual abuse victimization and perpetration were trending higher among girls across three of the four individual digital sexual abuse behaviors (with the exception of “Sent a sexual or naked photo/video to others without permission”), but further research with a larger sample size is needed to confirm this trend. There were significant or approaching significant differences in individual DDA behaviors. Girls were significantly more likely than boys to report pressuring their dating partner to sext and to report that their partner sent them a mean or hurtful private message and looked at private information to check up on them without permission. These results were somewhat consistent with the emerging DDA literature among primarily White youths, in the sense that many studies find equal rates of DDA experience for boys and girls. However, DDA research also indicates that sexual DDA may be more likely to be experienced by girls, and DDA may have differential negative emotional and behavioral impacts on girls (Reed et al., 2017). In addition, the very low rates of digital direct aggression reported by boys in the current sample is unique and suggests that factors of sample size or sample selection may have affected results. Further research on Latinx youths is warranted to discern whether DDA gender dynamics among La-tinx youths mirrors or differs from that of other groups. Association between DDA and Offline Dating Violence Results were consistent with prior research that found a link between online DDA and offline TDA experiences (Kernsmith et al., 2018; Reed et al., 2016). Like other cultural groups, this study supported that Latinx youths who experience DDA are also likely experiencing offline abuse, and vice versa (Kernsmith et al., 2018). However, this association was not wholly consistent across all types of online and offline victimization and perpetration. Significant associations were found between several DDA victimization and perpetration subscales and offline dating violence victimization subscales. However, digital sexual abuse perpetration was not associated with any types of offline dating violence victimization and perpetration, and digital direct aggression perpetration was only associated with offline psychological abuse victimization and perpetration. Conversely, there was a strong association between DDA monitoring and control perpetration and all types of offline victimization and perpetration. These results indicate that although DDA is conceptually similar to offline psychological abuse, all forms of DDA were not associated with offline abuse as we expected. It is possible that the characteristics and widespread consequences of digital media do not allow it to neatly map onto offline dating violence. Perhaps DDA provides a unique opportunity for a different subset of youths, particularly Latinx youths, to engage in TDA. Healthy Relationship Knowledge, Conflict Resolution Skills, and DDA Results indicated that boys and girls alike reported strong and healthy dating relationship knowledge and conflict resolution skills. There were significant associations between conflict resolution skills and DDA in the sample overall and for girls separately, but not for boys. These results indicate that less frequent use of positive conflict resolution skills was associated with more experience (both perpetration and victimization) with DDA, particularly for girls. These results might be supported by investigating these associations with a larger sample size, especially for boys. Improving positive conflict resolution skills may prevent unhealthy digital dating behaviors in dating relationships. It is important to note that this finding does not suggest that youths are responsible for preventing their partner from perpetrating offline dating violence or DDA by improving their conflict resolution skills. However, maladaptive relationship communication has been associated with TDA in a study of racially and ethnically diverse urban youths (Rueda, Yndo, Williams, & Shorey, 2018). As Rueda et al. (2018) argued, teaching conflict resolution skills should not be the sole aim of  TDA prevention programs, but it can be a helpful and important component. DDA and Latinx Youths Cultural context may influence dating relationships for Latinx youths (Milbrath, Ohlson, & Eyre, 2009). Future research should explore the influence of cultural socialization around dating, violence, and gender to better understand the unique cultural experience of Latinx youths around DDA. This research should measure youth endorsement of Latinx conceptualizations of gender and gender roles and should also include measures of acculturation. For example, Latinx cultural models of relationships are rooted in cultural mores and romantic care factors such as familismo (that is, the needs of the family take priority over individual needs), and respeto (that is, deference to authority—in this case, the authority of the parent) (Haglund et al., 2019; Malhotra, Gonzalez-Guarda, & Mitchell, 2015; Stein, Gonzalez, Cupito, Kiang, & Supple, 2015). Communication within relationships for Latinx adolescents is also rooted in the gender role expectations and values of machismo (that is, boys as strong protectors and figures of authority in relationships) and marianismo (that is, girls as virginal, unspoiled, and submissive with strong devotion to their family) (Deardorff, Tschann, & Flores, 2008; Rueda & Williams, 2015). Future research should explore how familismo, respeto, machismo, and marianismo might serve as risk or protective factors for DDA experience among Latinx youths. Research indicates that Latinx cultural expectations may be increasingly flexible for more recent generations of Latinx youths in the United States. In a longitudinal study of 246 Mexican-origin participants, Latinx girls demonstrated more declines in their endorsement of rigid gender role attitudes than Latinx boys over time, potentially leading to opportunities for relationship conflict in heterosexual dating relationships if gender expectations do not align (Updegraff, Umaña-Taylor, McHale, Wheeler, & Perez-Brena, 2012). A recent qualitative study of 10 Latinx heterosexual couples found that male teenagers demonstrated high levels of what scholars are calling adaptive machismo, which includes values and activities such as emotional availability, demonstrations of affection, desire to financially care for a female partner, responsibility in child rearing, and responsibility to the community or friends into the individual’s more traditional experience of machismo (Williams & Rueda, 2016). This adaptive machismo may be a protective factor for adolescent relationships and decrease male-partner perpetration of abusive behaviors. Therefore, the emerging literature on Latinx cultural values and the association with dating violence are mixed; future research is needed and should include digital forms of abuse. Limitations Despite the strengths of this study, including an exclusively Latinx sample and in-depth survey of DDA and TDA, there are limitations that should be addressed with future research. First, our participants represent a group of high school students who had dating experience, were seeking participation in a TDA prevention program, and were able to obtain parental consent. This might explain why their ratings of TDA were relatively high; perhaps they were seeking support. In addition, data were cross-sectional with a small sample, limiting the ability to detect variance. Data were collected with one group of Latinx youths from one community and are not representative of all nationalities and cultural groups that represent the U.S. Latinx population. The perpetration scales had lower internal consistency reliability than the victimization scales. It is difficult to determine whether low reliability was due to low numbers of reported perpetration, cultural differences in the experience of DDA, or other factors. It is possible that people who are perpetrators of violence are hesitant to report the full scope of perpetration for social desirability reasons. Finally, this measure has never been used with an exclusively Latinx sample. Future research is needed to better understand the reliability and validity of TDA perpetration scales, particularly for Latinx participants. Implications This study has important implications for social work practitioners and educators working with Latinx youths. DDA is experienced by Latinx youths and likely occurs in a constellation of other dating abuse experiences. The current study suggests that as in other studies of DDA among primarily White samples, there is a strong link between online and offline dating abuse. Therefore, social workers should incorporate this knowledge into their interventions with teenagers who might be experiencing or participating in DDA. To do so, social workers should include digital forms of abuse in any assessments or interventions with Latinx youths involved in TDA. If practitioners and educators only ask students about possible physical, sexual, and psychological abuse without asking about their digital interactions with dating partners, they might be missing key parts of teenagers’ experiences. Safety planning around TDA can also include digital media, such as blocking an abusive partner on social media sites or changing their phone number. Teenagers are unlikely to tell anyone about their dating violence experiences, and when they do tell someone, it is most likely to be a peer (Black & Weisz, 2004; Boldero & Fallon, 1995; Molidor & Tolman, 1998). Social workers should be invested in training teenagers how to respond if a friend tells them about a DDA experience, including how to identify problematic digital dating behavior and how to connect someone with resources. This is particularly important within minority cultural groups, as teenagers might be most likely to access resources that are culturally sensitive. The current study also has important implications for preventing DDA. Schools may want to include healthy relationship knowledge and conflict resolution skills as a universal feature in health or other relevant courses. Knowledge acquisition may not be enough to alter dating behaviors, and conflict resolution skills are likely not sufficient to prevent DDA at its most severe. However, knowledge about healthy relationships may help those experiencing more isolated unhealthy digital dating behaviors, as this abuse may result from lack of relationship skills rather than behaviors motivated by power and control. Practitioners should include DDA and TDA items in schoolwide screenings to proactively identify students who are experiencing dating abuse and connect them to services that may help interrupt the cycle of abuse and help develop healthy relationship skills. We also encourage social workers to include teenagers in DDA prevention efforts whenever possible. The way that youths use and understand the role of digital media in their dating relationships is shaped by their cultural beliefs and norms. Including youths from various cultural backgrounds and identities as peer advocates to prevent TDA in their community would help to ensure that efforts to prevent digital forms of  TDA are generationally and culturally relevant. Further research is needed to understand cultural values, acculturation, and how these may be used in a strengths-based approach to healthy relationship promotion and TDA prevention. References Anderson M. , Jiang J. ( 2018 ). Teens, social media & technology 2018 . Washington, DC : Pew Research Center . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Ball B. , Tharp A. T. , Noonan R. K. , Valle L. A. , Hamburger M. E. , Rosenbluth B. ( 2012 ). Expect Respect support groups: Preliminary evaluation of a dating violence prevention program for at-risk youth . Violence Against Women, 18 , 746 – 762 . doi:10.1177/1077801212455188 Google Scholar Crossref Search ADS WorldCat Black B. , Weisz A. ( 2004 ). Dating violence: A qualitative analysis of Mexican American youths’ views . Journal of Ethnic & Cultural Diversity in Social Work, 13 , 69 – 90 . Google Scholar Crossref Search ADS WorldCat Boldero J. , Fallon B. ( 1995 ). Adolescent help-seeking: What do they get help for and from whom? Journal of Adolescence, 18 , 193 – 209 . Google Scholar Crossref Search ADS WorldCat Bradford K. , Wade Stewart J. , Higginbotham B. J. , Skogrand L. ( 2015 ). The Perceived Relationship Knowledge Scale: An initial validation . Family Relations, 64 , 305 – 318 . doi:10.1111/fare.12116 Google Scholar Crossref Search ADS WorldCat Cutbush S. , Williams J. , Miller S. , Gibbs D. , Clinton-Sherrod M. ( 2018 ). Longitudinal patterns of electronic teen dating violence among middle school students . Journal of Interpersonal Violence . Advance online publication. doi:10.1177/0886260518758326 Google Scholar OpenURL Placeholder Text WorldCat Deardorff J. , Tschann J. M. , Flores E. ( 2008 ). Sexual values among Latino youth: Measurement development using a culturally based approach . Cultural Diversity and Ethnic Minority Psychology, 14 , 138 – 146 . doi:10.1037/1099-9809.14.2.138 Google Scholar Crossref Search ADS WorldCat Dick R. N. , McCauley H. L. , Jones K. A. , Tancredi D. J. , Goldstein S. , Blackburn S. , et al. ( 2014 ). Cyber dating abuse among teens using school-based health centers . Pediatrics, 134 , e1560 – e1567 . doi:10.1542/peds.2014-0537 Google Scholar Crossref Search ADS WorldCat Foshee V. A. , Bauman K. E. , Arriaga X. B. , Helms R. W. , Koch G. G. , Linder G. F. ( 1998 ). An evaluation of Safe Dates, an adolescent dating violence prevention program . American Journal of Public Health, 88 , 45 – 50 . Google Scholar Crossref Search ADS WorldCat Guerra-Reyes L. , Rush B. , Herbenick D. , Dodge B. , Reece M. , Schick V. , et al. ( 2017 ). Sexual health and language dominance among Hispanic/Latino women and men: Analysis of a nationally representative sample . Journal of Immigrant and Minority Health, 19 , 1059 – 1072 . doi:10.1007/s10903-016-0446-7 Google Scholar Crossref Search ADS WorldCat Haglund K. , Belknap R. A. , Edwards L. M. , Tassara M. , Van Hoven J. , Woda A. ( 2019 ). The influence of masculinity on male Latino adolescents’ perceptions regarding dating relationships and dating violence . Violence Against Women, 25 , 1039 – 1052 . doi:10.1177/1077801218808395 Google Scholar Crossref Search ADS WorldCat Henry N. , Powell A. ( 2018 ). Technology-facilitated sexual violence . Trauma, Violence, & Abuse, 19 , 195 – 208 . doi:10.1177/1524838016650189 Google Scholar Crossref Search ADS WorldCat Kann L. , Kinchen S. , Shanklin S. L. , Flint K. H. , Hawkins J. , Harris W. A. , et al. ( 2014 ). Youth Risk Behavior Surveillance—United States, 2013 . Morbidity and Mortality Weekly Report: Surveillance Summaries, 63 ( SS04 ), 1 – 168 . Google Scholar OpenURL Placeholder Text WorldCat Kernsmith P. D. , Victor B. G. , Smith-Darden J. P. ( 2018 ). Online, offline, and over the line: Coercive sexting among adolescent dating partners . Youth & Society, 50 , 891 – 904 . doi:10.1177/0044118X18764040 Google Scholar Crossref Search ADS WorldCat Korchmaros J. D. , Ybarra M. L. , Langhinrichsen-Rohling J. , Boyd D. , Lenhart A. ( 2013 ). Perpetration of teen dating violence in a networked society . Cyberpsychology, Behavior, and Social Networking, 16 , 561 – 567 . doi:10.1089/cyber.2012.0627 Google Scholar Crossref Search ADS WorldCat Malhotra K. , Gonzalez-Guarda R. M. , Mitchell E. M. ( 2015 ). A review of teen dating violence prevention research: What about Hispanic youth? Trauma, Violence, & Abuse, 16 , 444 – 465 . doi:10.1177/1524838014537903 Google Scholar Crossref Search ADS WorldCat McCullough Cosgrove J. , LeCroy C. W. , Fordney M. , Voelkel D. ( 2018 ). Considering the role of acculturation in parent–child communication about sexual health . American Journal of Public Health, 108 ( Suppl. 1 ), S13 – S14 . doi:10.2105/AJPH.2017.304222 Google Scholar Crossref Search ADS WorldCat Milbrath C. , Ohlson B. , Eyre S. L. ( 2009 ). Analyzing cultural models in adolescent accounts of romantic relationships . Journal of Research on Adolescence, 19 , 313 – 351 . doi:10.1111/j.1532-7795.2009.00598.x Google Scholar Crossref Search ADS WorldCat Molidor C. , Tolman R. M. ( 1998 ). Gender and contextual factors in adolescent dating violence . Violence Against Women, 4 , 180 – 194 . Google Scholar Crossref Search ADS WorldCat Olsen E. O. , Vivolo-Kantor A. , Kann L. ( 2017 ). Physical and sexual teen dating violence victimization and sexual identity among U.S. high school students, 2015 . Journal of Interpersonal Violence . Advance online publication. doi:10.1177/0886260517708757 Google Scholar OpenURL Placeholder Text WorldCat Reed L. A. , Tolman R. M. , Ward L. M. ( 2016 ). Snooping and sexting: Digital media as a context for dating aggression and abuse among college students . Violence Against Women, 22 , 1556 – 1576 . doi:10.1177/1077801216630143 Google Scholar Crossref Search ADS WorldCat Reed L. A. , Tolman R. M. , Ward L. M. ( 2017 ). Gender matters: Experiences and consequences of digital dating abuse victimization in adolescent dating relationships . Journal of Adolescence, 59 , 79 – 89 . Google Scholar Crossref Search ADS WorldCat Reed L. A. , Ward L. M. , Tolman R. M. , Lippman J. R. , Seabrook R. ( 2018 ). The association between stereotypical gender and dating beliefs and digital dating abuse perpetration in adolescent dating relationships . Journal of Interpersonal Violence . Advance online publication. doi:10.1177/0886260518801933 Google Scholar OpenURL Placeholder Text WorldCat Rueda H. A. , Nagoshi J. L. , Williams L. R. ( 2014 ). Mexican American and European American adolescents’ dating experiences across the ecosystem: Implications for healthy relationships within an ecodevelopmental framework . Journal of Human Behavior in the Social Environment, 24 , 358 – 376 . doi:10.1080/10911359.2013.831013 Google Scholar Crossref Search ADS WorldCat Rueda H. A. , Williams L. R. ( 2015 ). Mexican American adolescent couples communicating about conflict: An integrated developmental and cultural perspective . Journal of Adolescent Research, 31 , 375 – 403 . doi:10.1177/0743558415584999 Google Scholar Crossref Search ADS WorldCat Rueda H. A. , Yndo M. , Williams L. R. , Shorey R. ( 2018 ). Does Gottman’s marital communication conceptualization inform teen dating violence? Communication skill deficits analyzed across three samples of diverse adolescents . Journal of Interpersonal Violence . Advance online publication. doi:10.1177/0886260518814267 Google Scholar OpenURL Placeholder Text WorldCat Sabina C. , Cuevas C. A. , Cotignola-Pickens H. M. ( 2016 ). Longitudinal dating violence victimization among Latino teens: Rates, risk factors, and cultural influences . Journal of Adolescence, 47 , 5 – 15 . doi:10.1016/J.ADOLESCENCE.2015.11.003 Google Scholar Crossref Search ADS WorldCat Stein G. L. , Gonzalez L. M. , Cupito A. M. , Kiang L. , Supple A. J. ( 2015 ). The protective role of familism in the lives of Latino adolescents . Journal of Family Issues, 36 , 1255 – 1273 . doi:10.1177/0192513X13502480 Google Scholar Crossref Search ADS WorldCat Temple J. R. , Choi H. J. , Brem M. , Wolford-Clevenger C. , Stuart G. L. , Peskin M. F. , Elmquist J. A. ( 2016 ). The temporal association between traditional and cyber dating abuse among adolescents . Journal of Youth and Adolescence, 45 , 340 – 349 . doi:10.1007/s10964-015-0380-3 Google Scholar Crossref Search ADS WorldCat Updegraff K. A. , Umaña-Taylor A. J. , McHale S. M. , Wheeler L. A. , Perez-Brena N. J. ( 2012 ). Mexican-origin youth’s cultural orientations and adjustment: Changes from early to late adolescence . Child Development, 83 , 1655 – 1671 . doi:10.1111/j.1467-8624.2012.01800.x Google Scholar Crossref Search ADS WorldCat Vagi K. J. , O’Malley Olsen E. , Basile K. C. , Vivolo-Kantor A. M. ( 2015 ). Teen dating violence (physical and sexual) among US high school students . JAMA Pediatrics, 169 , 474 – 482 . doi:10.1001/jamapediatrics.2014.3577 Google Scholar Crossref Search ADS WorldCat Wechsler H. ( 2012 ). Youth risk behavior surveillance—United States, 2011 . Morbidity and Mortality Weekly Report, 61 ( SS04 ), 1 – 162 . Retrieved from https://www.cdc.gov/mmwr/preview/mmwrhtml/ss6104a1.htm Google Scholar OpenURL Placeholder Text WorldCat Williams L. R. , Rueda H. A. ( 2016 ). Mexican American adolescent couples’ vulnerability for observed negativity and physical violence: Pregnancy and acculturation mismatch . Journal of Adolescence, 52 , 170 – 181 . doi:10.1016/j.adolescence.2016.08.001 Google Scholar Crossref Search ADS WorldCat Wincentak K. , Connolly J. , Card N. ( 2017 ). Teen dating violence: A meta-analytic review of prevalence rates . Psychology of Violence, 7 , 224 – 241 . doi:10.1037/a0040194 Google Scholar Crossref Search ADS WorldCat Yahner J. , Dank M. , Zweig J. M. , Lachman P. ( 2015 ). The co-occurrence of physical and cyber dating violence and bullying among teens . Journal of Interpersonal Violence, 30 , 1079 – 1089 . doi:10.1177/0886260514540324 Google Scholar Crossref Search ADS WorldCat Zweig J. M. , Dank M. , Yahner J. , Lachman P. ( 2013 ). The rate of cyber dating abuse among teens and how it relates to other forms of teen dating violence . Journal of Youth and Adolescence, 42 , 1063 – 1077 . doi:10.1007/s10964-013-9922-8 Google Scholar Crossref Search ADS WorldCat Zweig J. M. , Lachman P. , Yahner J. , Dank M. ( 2014 ). Correlates of cyber dating abuse among teens . Journal of Youth and Adolescence, 43 , 1306 – 1321 . doi:10.1007/s10964-013-0047-x Google Scholar Crossref Search ADS WorldCat © 2020 National Association of Social Workers This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Voith, Laura, A;Hamler,, Tyrone;Francis, Meredith, W;Lee,, Hyunjune;Korsch-Williams,, Amy

2020 Social Work Research

doi: 10.1093/swr/svaa013

Abstract Social workers, including social work researchers, are called on to challenge social injustices and pursue social change. Research has shown a strong association between trauma and adversity and marginalized populations, exposing the unequal distribution of trauma and its effects throughout society. Given the focus on marginalized populations in social work, the social justice orientation of the field, and the intersection of trauma with marginalized populations, a framework to guide social work researchers in conducting trauma-informed, socially just research with marginalized populations is warranted. Therefore, this article provides a framework integrating trauma theory, trauma-informed principles, and intersectionality as a guide for social work research. The proposed framework is illustrated using a case study of low-income, predominantly African American men recruited from a criminal justice setting, acknowledging facilitators and barriers to implementation. The article concludes with a discussion of the implications for researchers and doctoral student training. Social work practice and research are grounded in social justice. The National Association of Social Workers’ (NASW) Code of Ethics holds social workers accountable to this aim: “Social workers challenge social injustice … pursue social change, particularly with and on behalf of vulnerable and oppressed individuals and groups of people” (NASW, 2017, p. 2). These individuals, groups, and communities experience discrimination and exclusion in social, political, and economic domains because of unequal power relationships, resulting in their marginalization in society (Given, 2008). Thus, negotiating power dynamics on multiple levels is critical for social workers to address social injustice with and on behalf of vulnerable populations. Research on trauma and adversity and the implementation of trauma-informed care (TIC) has been taken up by social work and other helping professions over the past two decades. Research has shown that trauma and adversity are pervasive in the United States (Felitti et al., 1998), with marginalized groups reporting disproportionately higher rates of exposure (Cronholm et al., 2015). Given the strong intersection between trauma and adversity and marginalized populations, the role of social work researchers and the conduct of research with these populations must be considered. Furthermore, social work researchers often conduct studies within institutions that can simultaneously help and contribute to revictimization among individuals and families (for example, health care/hospitals, criminal justice settings, education/schools, child protection), but we do not have strong frameworks guiding socially just, trauma-informed research. Therefore, this article provides a framework integrating trauma theory, trauma-informed principles, and intersectionality as a guide for social work research, using a case study with a sample of marginalized men involved in the criminal justice system. Literature Review Historical Research with Marginalized Populations Research has significantly advanced society; however, serious harms have been perpetrated historically in the name of research with marginalized populations (for example, Nazi human experimentation, the Tuskegee syphilis experiment, radiation testing on human subjects, experimentation on prisoners and children; Bracken-Roche, Bell, Macdonald, & Racine, 2017). Although substantial steps have been taken to prevent researchers from repeating these heinous crimes against humanity, less dubious forms of harm that are relevant for contemporary research remain. For example, conducting research within the criminal justice system presents ethical concerns regarding the civil rights of vulnerable parties ( Jones, 2012), due to the possibility of coercion or exposure to unnecessary risk. In such studies, researchers maintain considerable influence over participants from their knowledge and perceived authority ( Jones, 2012). Moreover, these individuals are far more likely to be members of marginalized groups because of disparities in policing, sentencing, and incarceration among racial and ethnic minorities and individuals occupying a lower socioeconomic status (SES) (Alexander, 2012). Exposure to Adversity and Traumatic Events among Marginalized Populations Individual trauma results from a singular event, series of events, or set of circumstances that are experienced as physically or emotionally harmful or life threatening and that have lasting adverse effects on an individual’s functioning and overall well-being (Substance Abuse and Mental Health Services Administration [SAMHSA], 2014). These events are uncontrollable in nature and characterized by the potential to overwhelm one’s sense of coping skills (Briere & Scott, 2015). Examples include natural disasters, violence, and war (American Psychological Association, 2017); child abuse and household dysfunction (for example, parent with a mental illness or substance abuse issue; Felitti et al., 1998); and community exposures (for example, community violence, bullying, chronic poverty; Cronholm et al., 2015). Studies show that marginalized populations such as racial minorities (Smith & Patton, 2016; Vásquez, Udo, Corsino, & Shaw, 2019), economically disadvantaged groups (Metzler, Merrick, Klevens, Ports, & Ford, 2017; Sherman, 2013), and prisoners (Wolff, Chugo, Shi, Huening, & Frueh, 2015) are more likely to have experienced adverse childhood experiences (ACEs) compared with the general population, with studies demonstrating the connection between these experiences and individual needs and behaviors (for example, Reavis, Looman, Franco, & Rojas, 2013). Moreover, marginalized groups are more likely to experience traumatic events by way of community or sociopolitical violence throughout the life span (Koenen et al., 2017). Researchers assert that the impact of traumatic events is augmented by sociocultural contexts that involve discrimination and structural inequalities, which are a daily reality for some groups (Quiros & Berger, 2015; Seng, Lopez, Sperlich, Hamama, & Reed Meldrum, 2012), pointing to the intersection and complexity of trauma experienced by marginalized populations. Although researchers should not assume that individuals exposed to traumatic events will be traumatized, it is reasonable to anticipate some negative impacts on groups disproportionately exposed to adversity and traumatic events. Reasonably, researchers have questioned whether studies involving trauma-exposed populations risk retraumatizing individuals by probing about traumatic experiences (Carlson, Newman, & Loewenstein, 2003). A meta-analysis on participant reactions to trauma-related research found that participants generally do not experience retraumatization from participation or regret participating in the research, regardless of the type of traumatic events they had experienced ( Jaffe, DiLillo, Hoffman, Haikalis, & Dykstra, 2015). Despite these findings, the authors still emphasized that efforts to mitigate the potential distress of participants in trauma-related research should be considered, given the vulnerability of the participants ( Jaffe et al., 2015). Current guidelines to minimize retraumatization include emphasizing the potential for emotional distress during informed consent, assessment of participant reactions, and close examination of study phases most likely to revictimize participants, such as eligibility screening (Carlson et al., 2003). Collecting resilience information in addition to adverse experiences may enhance the richness of studies, guide analyses on the mediating effects of protective factors, and reduce study attrition rates (Leitch, 2017). Although these suggestions are certainly good practices, the field lacks a robust framework guiding research practice grounded in theoretical and empirical literature. Gaps and Current Study Given the focus on marginalized populations in social work, the social justice orientation of the field, and the intersection of trauma with marginalized populations, a framework to guide researchers in conducting trauma-informed, socially just research (TISJR) with marginalized populations is warranted. Notably, a framework for prioritizing intersectionality in research exists (Murphy, Hunt, Zajicek, Norris, & Hamilton, 2009); however, this framework does not incorporate the robust body of research on trauma prevalence and related best practices. Furthermore, the guiding principles for working with trauma-exposed populations (that is, trauma-informed principles) are most commonly followed in clinical settings rather than research settings. Given the key differences between clinical and research boundaries, further work in this area is necessary. Therefore, we present a TISJR framework and illustrate the application of this framework using a case study of low-income, predominantly African American men recruited from a criminal justice setting, acknowledging facilitators and barriers to implementation. TISJR: Framework Trauma Theory and Trauma-Informed Principles Individuals exposed to a significantly traumatic event may experience alterations in cognition (that is, perceptions of oneself, relationships with others and the world) and memory (for example, intrusive thoughts or fragmented memory), and experience heightened (for example, easily startled, impulse control) or hypo (numbing, low affect or dissociation) physiological arousal (van der Kolk, 2003). Disruptions to one’s physical, cognitive, and psychological ecosystem are normal human responses to traumatic experiences that, if unresolved after several months, can become maladaptive (van der Kolk, 2003). Trauma survivors may respond to events or situations with apparently incongruent reactions, due to dysregulation in multiple systems in the body and the brain (Bloom, 1999). For example, a person who experienced interpersonal violence might have difficulty concentrating at work, become disproportionately angry or fearful when others raise their voices, and avoid social situations that serve as traumatic reminders of the violent experience. Complex trauma results from repeated traumatic or adverse experiences throughout development, particularly from ongoing abuse, neglect, or disturbances in caregiving, but also from persistent poverty, racism, sexism, and other societal-level stressors (Ford & Courtois, 2009). Those exposed to complex trauma are more vulnerable to toxic stress, which can alter the body’s stress response system (for example, fight-or-flight response) and change the architecture of the brain as described in the ecobiodevelopmental framework (Shonkoff et al., 2012). Multiple exposures to interpersonal trauma have consistent and predictable consequences that affect functioning and development across domains, such as intense affects (rage, fear, shame), efforts to avoid the recurrence of these overwhelming emotions through avoidance or controlling behaviors, marked physiological and emotional dysregulation, somatic problems, and difficulty forming and maintaining positive interpersonal relationships (van der Kolk, 2005). Complex trauma is associated with pervasive problems in adulthood such as employment and relationship instability, and behavioral issues such as substance abuse or aggression and violence (Ford & Courtois, 2009). Applying trauma theory in practice shifted the orientation of the clinician from “What’s wrong with you?” to “What happened to you?” This shift in orientation, coined “TIC,” moves away from viewing responses to trauma as a disorder, or manifestation of a behavioral, psychological, or biological dysfunction within an individual, to distress, a normal human response to overwhelming stress (Harris & Fallot, 2001; RYSE Center, 2015). SAMHSA (2014) proposed six guiding tenets for the implementation of TIC in clinical settings: (1) the safety of all people within and served by the agency; (2) trustworthiness and transparency of operations and decisions; (3) peer support and mutual self-help; (4) collaboration and leveling of power differentials across the agency; (5) empowerment, voice, and choice; and (6) addressing cultural, historical, and gender issues that may affect services. Social Justice Recently, researchers have extended the work of TIC to move beyond the knowledge of trauma and associated responses, to instead emphasize interactions that are healing in nature. Healing-centered engagement moves from a clinical to a political perspective and is inherently strengths based, shifting the question from “What happened to you?” to “What is right about you?” (Ginwright, 2015, 2018). For example, the principle “addressing cultural, historical, and gender issues” was recently reframed as intersectionality to undergird the other TIC principles in the context of identity and power dynamics (Bowen & Murshid, 2016), illuminating a politicized understanding of trauma. That is, trauma exposure and related effects are unequally distributed throughout society, with marginalized groups bearing the greatest burden (Bowen & Murshid, 2016). Intersectionality was first introduced by Crenshaw (1989) in reference to the oppression of African American women, to explain more fully how multiple identities and dimensions of experience relative to power dynamics (for example, race and ethnicity, sexuality and gender identity, religion, SES) shape one’s social “location” in society considering the social, cultural, and historical context. This framework has since been extended widely to other populations and applied using a social work lens to practice (Mattsson, 2014), criminal justice work (Glynn, 2016), and research (Murphy et al., 2009). Along those lines, certain groups bear the burden of “historical trauma,” which entails a mass trauma deliberately and systematically inflicted on a target group by a dominant group continuing over an extended period of time, creating a universal experience that derails a population, and results in physical, psychological, social, and economic disparities that persist across generations (for example, slavery) (Sotero, 2006). In healing-centered engagement, intersectionality becomes the lens through which every other principle is implemented. Intersectional approaches to research align with the social work ethical imperative of cultural competence and reflect the ecological, holistic, and multi-lensed approach that is at the core of social work theory and practice (Murphy et al., 2009). Healing-Centered Engagement in Research Taken together, trauma-informed, socially just principles in clinical work can be adapted for research. Applying healing-centered engagement in research requires a shift in orientation resulting in an intentional top-down, systems approach that assesses the structural inequalities that perpetuate trauma throughout all of the involved systems (Bloom, 2018; Bowen & Murshid, 2016). Moreover, this orientation must be applied to all components of the research including the conceptualization, implementation, analysis, and dissemination. The research design should align with the four key assumptions of TIC (SAMHSA, 2014): (1) realizing the widespread impact of trauma and understanding potential paths for healing and recovery; (2) recognizing the signs and symptoms of trauma in all of those involved in the research study (participants and researchers); (3) responding by fully integrating knowledge about trauma into research policies, procedures, and practices; and (4) seeking to actively resist retraumatization. Furthermore, the study participants’ voices should be central to the study not only to resist retraumatization, but also to build platforms for disempowered groups to speak their truth through the application of qualitative or mixed methods and/or principles of community-based participatory research that elevates the expertise of participants’ lived experience. Implementing a trauma-informed approach with a social justice orientation can serve as a framework to guide social work researchers’ engagement with the populations of interest. TISJR: Application of Framework Sociopolitical Context Researchers must first understand the historical, sociopolitical, and cultural context of the population of interest to realize the widespread impact of individual, community, and historical trauma through a social justice lens. To do this, researchers should analyze the interplay of these power dynamics in systems of privilege and oppression relative to the potential impact on various facets of the study, such as the study site location and populations of interest. Conducting this analysis will provide researchers with salient information that will help shape a trauma-informed, socially just study design. To aid in application, refer to the “Pre-Study” guiding questions in the TISJR Framework Application Inventory (see Table 1). Table 1: Trauma-Informed, Socially Just Research (TISJR) Framework Application Inventory Stage of Research . Priority Tenets of TISJR Framework . Guiding Questions for TISJR Framework Application . Pre-study Sociopolitical, cultural, and historical context; peer support; transparency What systems of privilege and oppression at the micro, meso, exo, and macro levels could affect your study (for example, study staff, population, sociocultural/historical context)? How could these dynamics promote or violate the assumptions of healing-centered engagement? Are there any ways to mitigate violations of key assumptions? Study design Safety; transparency; empowerment, voice, and choice; centralization of participants’ identities Does the study design consider the goals of the study while still promoting safety, transparency, and choice among participants? Does the study design consider the goals of the research study while centralizing the lived experiences and identities of participants? Recruitment Safety, transparency, shared power, collaboration Is the recruitment process transparent? Do protocols and procedures reduce power differentials and promote collaboration between participants and those involved directly or indirectly with the study? Is the study team prepared to discuss their social location in the context of the sociopolitical, historical context relative to the social location of participants? Informed consent Empowerment, voice, and choice; transparency How can we promote agency, choice, and control during the informed consent process? Do documents and procedures protect against potential challenges for trauma-exposed populations (for example, cognition overload)? What elements of this process might threaten the safety (psychological, physical) of participants? What changes can be made to make the process more transparent? Data collection Safety, building and maintaining trust What threats to safety can we anticipate for participants and study staff? How can safety and security be addressed while collecting quality data? Is the study team equipped to maintain and promote emotional and behavioral regulation with participants? Post–data collection Safety; empowerment, voice, and choice How will participants who have given their time to share deeply personal information be acknowledged? Which tools to assess participants’ stress levels are most appropriate for the study context? What resources are available to empower participants post-study? Stage of Research . Priority Tenets of TISJR Framework . Guiding Questions for TISJR Framework Application . Pre-study Sociopolitical, cultural, and historical context; peer support; transparency What systems of privilege and oppression at the micro, meso, exo, and macro levels could affect your study (for example, study staff, population, sociocultural/historical context)? How could these dynamics promote or violate the assumptions of healing-centered engagement? Are there any ways to mitigate violations of key assumptions? Study design Safety; transparency; empowerment, voice, and choice; centralization of participants’ identities Does the study design consider the goals of the study while still promoting safety, transparency, and choice among participants? Does the study design consider the goals of the research study while centralizing the lived experiences and identities of participants? Recruitment Safety, transparency, shared power, collaboration Is the recruitment process transparent? Do protocols and procedures reduce power differentials and promote collaboration between participants and those involved directly or indirectly with the study? Is the study team prepared to discuss their social location in the context of the sociopolitical, historical context relative to the social location of participants? Informed consent Empowerment, voice, and choice; transparency How can we promote agency, choice, and control during the informed consent process? Do documents and procedures protect against potential challenges for trauma-exposed populations (for example, cognition overload)? What elements of this process might threaten the safety (psychological, physical) of participants? What changes can be made to make the process more transparent? Data collection Safety, building and maintaining trust What threats to safety can we anticipate for participants and study staff? How can safety and security be addressed while collecting quality data? Is the study team equipped to maintain and promote emotional and behavioral regulation with participants? Post–data collection Safety; empowerment, voice, and choice How will participants who have given their time to share deeply personal information be acknowledged? Which tools to assess participants’ stress levels are most appropriate for the study context? What resources are available to empower participants post-study? Open in new tab Table 1: Trauma-Informed, Socially Just Research (TISJR) Framework Application Inventory Stage of Research . Priority Tenets of TISJR Framework . Guiding Questions for TISJR Framework Application . Pre-study Sociopolitical, cultural, and historical context; peer support; transparency What systems of privilege and oppression at the micro, meso, exo, and macro levels could affect your study (for example, study staff, population, sociocultural/historical context)? How could these dynamics promote or violate the assumptions of healing-centered engagement? Are there any ways to mitigate violations of key assumptions? Study design Safety; transparency; empowerment, voice, and choice; centralization of participants’ identities Does the study design consider the goals of the study while still promoting safety, transparency, and choice among participants? Does the study design consider the goals of the research study while centralizing the lived experiences and identities of participants? Recruitment Safety, transparency, shared power, collaboration Is the recruitment process transparent? Do protocols and procedures reduce power differentials and promote collaboration between participants and those involved directly or indirectly with the study? Is the study team prepared to discuss their social location in the context of the sociopolitical, historical context relative to the social location of participants? Informed consent Empowerment, voice, and choice; transparency How can we promote agency, choice, and control during the informed consent process? Do documents and procedures protect against potential challenges for trauma-exposed populations (for example, cognition overload)? What elements of this process might threaten the safety (psychological, physical) of participants? What changes can be made to make the process more transparent? Data collection Safety, building and maintaining trust What threats to safety can we anticipate for participants and study staff? How can safety and security be addressed while collecting quality data? Is the study team equipped to maintain and promote emotional and behavioral regulation with participants? Post–data collection Safety; empowerment, voice, and choice How will participants who have given their time to share deeply personal information be acknowledged? Which tools to assess participants’ stress levels are most appropriate for the study context? What resources are available to empower participants post-study? Stage of Research . Priority Tenets of TISJR Framework . Guiding Questions for TISJR Framework Application . Pre-study Sociopolitical, cultural, and historical context; peer support; transparency What systems of privilege and oppression at the micro, meso, exo, and macro levels could affect your study (for example, study staff, population, sociocultural/historical context)? How could these dynamics promote or violate the assumptions of healing-centered engagement? Are there any ways to mitigate violations of key assumptions? Study design Safety; transparency; empowerment, voice, and choice; centralization of participants’ identities Does the study design consider the goals of the study while still promoting safety, transparency, and choice among participants? Does the study design consider the goals of the research study while centralizing the lived experiences and identities of participants? Recruitment Safety, transparency, shared power, collaboration Is the recruitment process transparent? Do protocols and procedures reduce power differentials and promote collaboration between participants and those involved directly or indirectly with the study? Is the study team prepared to discuss their social location in the context of the sociopolitical, historical context relative to the social location of participants? Informed consent Empowerment, voice, and choice; transparency How can we promote agency, choice, and control during the informed consent process? Do documents and procedures protect against potential challenges for trauma-exposed populations (for example, cognition overload)? What elements of this process might threaten the safety (psychological, physical) of participants? What changes can be made to make the process more transparent? Data collection Safety, building and maintaining trust What threats to safety can we anticipate for participants and study staff? How can safety and security be addressed while collecting quality data? Is the study team equipped to maintain and promote emotional and behavioral regulation with participants? Post–data collection Safety; empowerment, voice, and choice How will participants who have given their time to share deeply personal information be acknowledged? Which tools to assess participants’ stress levels are most appropriate for the study context? What resources are available to empower participants post-study? Open in new tab The case study took place in Cleveland, a Rust Belt city (that is, a city in the Great Lakes region that experienced deindustrialization and urban decay of a once booming industry; High, 2003). Deindustrialization exacerbated the conditions of one of the most racially segregated cities in the United States (Massey & Tannen, 2015). Given the conditions and history of the city, it is likely that individuals of different races, ethnicities, and socioeconomic standing interact less often, which may be mirrored across researcher and participant, potentially amplifying mistrust of the other. Furthermore, the city of Cleveland experienced a great tragedy when Tamir Rice, a 12-year-old Black boy, was shot and killed by a Cleveland police officer while in a park near his home in 2014. Tamir Rice’s death caused outrage throughout the city and gained national attention in the Black Lives Matter movement (Ransby, 2018). That same year, the Cleveland Police Department (CPD) was investigated by the U.S. Department of Justice Civil Rights Division and the U.S. Attorney’s Office North District of Ohio. Results of this investigation showed that CPD had a systematic pattern of reckless and inappropriate use of deadly and lethal force by officers and that CPD “frequently” violated people’s civil rights (United States of America v. City of Cleveland, 2015). These acts have fueled a deep mistrust of the city’s law enforcement, especially in the Black community (United States of America v. City of Cleveland, 2015). The resulting racial tensions likely negatively affect the sense of physical and psychological safety of the participants, especially the Black men, given the study setting (for example, Sewell, Jefferson, & Lee, 2016). Study Setting In realizing potential trauma exposure with populations of interest, researchers can assess opportunities and challenges to uphold the principle of “do no harm” and possible pathways to mitigate retraumatization vis-à-vis the study setting. Considering the sociopolitical and cultural context, researchers should assess potential study settings, including geographical location, physical environment, and social environment relative to all those involved with the study (that is, participants, researchers, other personnel) that may affect one’s sense of physical or psychological safety, transparency, power dynamics, and sense of empowerment. For example, the case study took place in a criminal justice setting—the county probation office—to enhance accessibility and convenience for participants. However, a power imbalance at this site was evident because of the law, the possession of lethal weapons by police officers, and a sense that one’s freedom is at risk. These inherent power differentials likely influenced the ability to foster collaboration and mutuality among researchers and participants, potentially influenced participants’ sense of safety and empowerment, and perhaps influenced participants’ sense of obligation to comply with the study. Moreover, the criminal justice setting and safety measures of that setting (for example, bailiff, panic buttons in each room, metal detectors at building entrances) likely affected researchers’ sense of safety and level of comfort. Given that trauma has been identified as an underlying mechanism of intimate partner violence perpetration (see Voith, Logan-Greene, Strodthoff, & Bender, 2018), the primary aim of the original study was to investigate trauma exposure and overall health (for example, mental, physical, behavioral) of men participating in batterer intervention programs as it relates to treatment completion and intimate partner violence perpetration. The authors’ university institutional review board approved all study procedures and materials. Sample Study teams must also recognize signs and symptoms of all individuals involved in research. There are multiple avenues that can help researchers gain an understanding of participants’ life experiences that can give context for understanding potential signs and symptoms of trauma among participants. Ranging in accessibility, researchers can access previous trauma-related research with the population of interest, elicit insights from community partners working with the population of interest, collect formative data or use program data, and form an advisory council of individuals representing that population. To illustrate, in the case study participants were predominantly African American or Black, with a high school education and employed, earning less than $20,000 in the past year (see Table 2). Men in the study reported high exposure to ACEs, with an average score of three out of the conventional 10-item checklist (Felitti et al., 1998), over two on an expanded checklist (for example, homelessness, food insecurity) (Mersky, Janczewski, & Topitzes, 2017), and over 2.5 community-level exposures to trauma (for example, witnessed violence, unsafe neighborhood, felt discrimination) (Cronholm et al., 2015). Moreover, nearly 30% of the study sample met the clinical threshold for posttraumatic stress disorder (see Table 2). These experiences, coupled with participants’ demographic characteristics and surveillance by the criminal justice system, suggest that the study’s sample represented men with profoundly marginalized identities, calling for a trauma-informed, socially just approach to research. Table 2: Study Participant Demographics Demographic Characteristic . n (%) . M (SD) . Age (years) (N = 54) 35.72 (11.33) Racial or ethnic group (N = 54)  African American or Black 40 (74.1)  White, Caucasian 4 (7.4)  Native American or Alaska Native 1 (1.9)  Hispanic or Latino 5 (9.3)  Other 4 (7.4) Educational attainment (N = 53)  Less than high school 16 (29.7)  GED/high school 27 (50)  Some college 10 (18.5) Employed in past year (N = 54)  No 6 (11.1)  Yes 48 (88.9) Ever incarcerated (N = 51) 35 (68.6) Ever in solitary confinement (N = 35) 14 (40.0) ACEs  Sum of 10 traditional ACEs items (N = 36) 3.11 (2.61)  Sum of 6 expanded ACEs items (N = 50) 2.34 (1.94)  Sum of 5 community-level ACEs items (N = 50) 2.66 (1.41)  Sum of 21 expanded ACEs items (N = 36) 7.72 (5.00) PTSD total score (N = 46) 25.48 (19.02)  Over clinical cutoff score for PTSD (N = 46) 13 (28.3) Demographic Characteristic . n (%) . M (SD) . Age (years) (N = 54) 35.72 (11.33) Racial or ethnic group (N = 54)  African American or Black 40 (74.1)  White, Caucasian 4 (7.4)  Native American or Alaska Native 1 (1.9)  Hispanic or Latino 5 (9.3)  Other 4 (7.4) Educational attainment (N = 53)  Less than high school 16 (29.7)  GED/high school 27 (50)  Some college 10 (18.5) Employed in past year (N = 54)  No 6 (11.1)  Yes 48 (88.9) Ever incarcerated (N = 51) 35 (68.6) Ever in solitary confinement (N = 35) 14 (40.0) ACEs  Sum of 10 traditional ACEs items (N = 36) 3.11 (2.61)  Sum of 6 expanded ACEs items (N = 50) 2.34 (1.94)  Sum of 5 community-level ACEs items (N = 50) 2.66 (1.41)  Sum of 21 expanded ACEs items (N = 36) 7.72 (5.00) PTSD total score (N = 46) 25.48 (19.02)  Over clinical cutoff score for PTSD (N = 46) 13 (28.3) Notes: ACEs = adverse childhood experiences; PTSD = posttraumatic stress disorder. Open in new tab Table 2: Study Participant Demographics Demographic Characteristic . n (%) . M (SD) . Age (years) (N = 54) 35.72 (11.33) Racial or ethnic group (N = 54)  African American or Black 40 (74.1)  White, Caucasian 4 (7.4)  Native American or Alaska Native 1 (1.9)  Hispanic or Latino 5 (9.3)  Other 4 (7.4) Educational attainment (N = 53)  Less than high school 16 (29.7)  GED/high school 27 (50)  Some college 10 (18.5) Employed in past year (N = 54)  No 6 (11.1)  Yes 48 (88.9) Ever incarcerated (N = 51) 35 (68.6) Ever in solitary confinement (N = 35) 14 (40.0) ACEs  Sum of 10 traditional ACEs items (N = 36) 3.11 (2.61)  Sum of 6 expanded ACEs items (N = 50) 2.34 (1.94)  Sum of 5 community-level ACEs items (N = 50) 2.66 (1.41)  Sum of 21 expanded ACEs items (N = 36) 7.72 (5.00) PTSD total score (N = 46) 25.48 (19.02)  Over clinical cutoff score for PTSD (N = 46) 13 (28.3) Demographic Characteristic . n (%) . M (SD) . Age (years) (N = 54) 35.72 (11.33) Racial or ethnic group (N = 54)  African American or Black 40 (74.1)  White, Caucasian 4 (7.4)  Native American or Alaska Native 1 (1.9)  Hispanic or Latino 5 (9.3)  Other 4 (7.4) Educational attainment (N = 53)  Less than high school 16 (29.7)  GED/high school 27 (50)  Some college 10 (18.5) Employed in past year (N = 54)  No 6 (11.1)  Yes 48 (88.9) Ever incarcerated (N = 51) 35 (68.6) Ever in solitary confinement (N = 35) 14 (40.0) ACEs  Sum of 10 traditional ACEs items (N = 36) 3.11 (2.61)  Sum of 6 expanded ACEs items (N = 50) 2.34 (1.94)  Sum of 5 community-level ACEs items (N = 50) 2.66 (1.41)  Sum of 21 expanded ACEs items (N = 36) 7.72 (5.00) PTSD total score (N = 46) 25.48 (19.02)  Over clinical cutoff score for PTSD (N = 46) 13 (28.3) Notes: ACEs = adverse childhood experiences; PTSD = posttraumatic stress disorder. Open in new tab Research Team Training and Preparation Applying this framework to studies calls for researchers to explicitly draw on trauma-related research and clinical expertise and, using the most rigorous evidence, apply this knowledge to shape policies, procedures, and practices carried out by all those involved in the study (that is, responding). Therefore, in this case study, a consultant with expertise in trauma was contracted to host a four-hour training for the study team. The training defined individual and historical trauma; described how trauma affects individuals, groups, and communities (for example, worldview, adaptive behavior, brain development, triggers); and explored trauma-informed techniques. In addition, each team member explored their own social locations (for example, intersecting identities of race, age, gender, nationality, education; see Table 3) and discussed how those identities might contribute to the dynamics of researcher–participant interactions, which empowered research staff to recognize their own potential exposure to adversity and related responses. Enhancing the awareness of each team members’ social location and biases in these settings was important to increase the research team’s authenticity with participants (Braxton-Newby & Jones, 2014). Finally, it was critical to establish how to interact with participants using a trauma-informed, socially just approach while maintaining our boundaries as researchers (illustrated in the following sections). Table 3: Study Team Composition (N = 5) Gender . Racial/Ethnic Group . Nationality . Age M (SD) . Educational Attainment . Gender queer: 1 Men: 2 Women: 2 African American/Black: 1 Asian: 1 Caucasian/White: 3 North American: 4 South Korean: 1 33.0 (4.95) Master’s: 4 PhD: 1 Gender . Racial/Ethnic Group . Nationality . Age M (SD) . Educational Attainment . Gender queer: 1 Men: 2 Women: 2 African American/Black: 1 Asian: 1 Caucasian/White: 3 North American: 4 South Korean: 1 33.0 (4.95) Master’s: 4 PhD: 1 Open in new tab Table 3: Study Team Composition (N = 5) Gender . Racial/Ethnic Group . Nationality . Age M (SD) . Educational Attainment . Gender queer: 1 Men: 2 Women: 2 African American/Black: 1 Asian: 1 Caucasian/White: 3 North American: 4 South Korean: 1 33.0 (4.95) Master’s: 4 PhD: 1 Gender . Racial/Ethnic Group . Nationality . Age M (SD) . Educational Attainment . Gender queer: 1 Men: 2 Women: 2 African American/Black: 1 Asian: 1 Caucasian/White: 3 North American: 4 South Korean: 1 33.0 (4.95) Master’s: 4 PhD: 1 Open in new tab Study Design Researchers can design studies that range from minimally adhering to the basic principle of “do no harm” to enacting principles that can empower marginalized populations and potentially begin to actively resist social injustice through the process of research. The priority tenets of the TISJR framework that support those aims include prioritizing safety, transparency, and empowerment and centralizing the experiences of participants’ intersectional identities. First, a certificate of confidentiality was obtained to protect the information provided by participants from compelled disclosure, to enhance a sense of safety among participants. Second, each measure included in the survey was assessed for difficulty on a five-point scale, using seven dimensions (that is, number of items, approximate length, administration method, reading difficulty, time frame recall, number of items with emotional weight, and the degree of detail). Based on prior research (Fink, 2012) and clinical advisement, the survey began with higher-intensity questions and then transitioned to lower-intensity questions. Grouping these questions in the beginning of the survey minimized harm that may have occurred from answering potentially triggering questions when participants were fatigued later in the survey. Lower-intensity questions were introduced for the remainder of the survey to provide time for participants to self-regulate, if needed, before completing the survey. Third, participants were empowered to share as much or as little information as they felt comfortable by building in the option of “refuse to answer” (that is, choice) for each item, which was highlighted prior to beginning the survey. Fourth, a private room was provided to complete the survey with a research assistant nearby to provide assistance and clarification. Finally, the survey was self-administered, enhancing privacy of disclosure and reducing potential distress from having to verbally narrate one’s trauma history ( Jaffe et al., 2015). Researchers should see Table 1’s guiding questions under “Study Design” to consider developing simple and more complex policies and practices that adhere to the TISJR framework. Recruitment Transparency, sharing power, and collaboration during the recruitment phase are critical tenets to enhance participants’ sense of safety during the recruitment process (see Table 1). For example, certain procedures were developed and followed that prioritized safety in the case study. First, the role of the probation officers was minimized in the recruitment effort, given the inherent power differential between probation officers and potential participants. Second, researchers openly acknowledged the potential mistrust between participants and study staff based on institutional harms perpetrated against men of color through research (for example, Tuskegee; see George, Duran, & Norris, 2014). By acknowledging historical trauma with participants early in the recruitment process, the study staff were able to hold space to discuss mistrust of these institutions and, by extension, the study team. As representatives of these institutions, researchers are responsible for acknowledging these previous harms and providing opportunities for participants to voice their opinions, and ultimately respect their autonomy to engage or withdraw from the study. During this process, opportunities arose to discuss the social positions (that is, the intersection of race, class, nationality, and gender) that each researcher held in relation to the men in the study. Ethnic and cultural differences must be explored when expressing empathy, to clarify verbal and nonverbal communication (Burkard & Knox, 2004; Henretty & Levitt, 2010). For example, a research team member may begin this conversation by acknowledging their own social identity and bridge that with a potential assumption, such as “As a [White woman] [highly educated Black man] you may wonder what business I have asking questions about your life or how I could understand your experiences.” Opening the door for this discussion by identifying the researcher’s differences and naming a possible assumption prompted participants to explore these differences by commenting and asking the researcher questions (for example, “Did you grow up here?”) that helped identify the researcher’s social location. That is, the men attempted to identify what shared life experiences, values, or pathways they had with the researcher to see if the researcher could understand them and their life experiences. Addressing these issues directly enhanced relationship building between researchers and participants and increased empowerment and voice among participants. These conversations also allowed researchers to express empathy by being genuine with participants. Third, before explaining the study procedures, researchers asked about participants’ previous experience or knowledge of the research process. By acknowledging participants’ experience and incorporating this information into the recruitment process, the researcher and participant began the study on mutual terms. Informed Consent Empowerment, voice, choice, and transparency are priority tenets of the TISJR framework during the informed consent process. Considering that trauma survivors are more likely to experience challenges with memory and cognitive functioning (for example, focusing) (Briere & Scott, 2012), researchers must consider how to respond to these effects through study procedures and practices (see Table 1’s guiding questions). For example, in this case study the consent form was written at an eighth-grade reading level and explained in smaller, digestible chunks of information using mutually understood terms. Before advancing, the researcher and participant would discuss the information and clarify any questions in approachable terms, which enhanced transparency and engagement with participants. Furthermore, a key facet of traumatic experiences is that they are uncontrollable and overwhelming (Briere & Scott, 2015), which can be compounded by the intersectional identities of disempowered groups. Thus, researchers asked participants for permission to provide information or transition to next steps (for example, “Can I tell you about the types of questions we will ask?” “Are you ready to move on to the next part?”); by sharing power, participants maintained control of the pace and process, their autonomy was acknowledged, and mutuality was enhanced (Butler, Critelli, & Rinfrette, 2011). The promotion of safety and transparency is paramount to mitigate the effects of trauma, and researchers can respond with relatively simple changes in procedures and practices. For example, to enhance a sense of safety and transparency in the case study, the researcher provided a paper copy of the survey for participants to look through while explaining the survey content. This allowed participants to have a tangible document that clarified exactly what they would be asked. During this process, the researcher highlighted survey items to prepare participants for more challenging or potentially triggering topics. These items were identified based on content related to trauma (for example, questions regarding violence exposure, ACEs, and trauma symptoms) and feedback from men who completed the survey (for example, sexual behavior and gender norms, employment questions). Using this approach resulted in a longer consenting process; however, transparency was enhanced, and it appeared to result in increased trustworthiness and mutuality. Finally, all participants were provided the option to receive the results of the study to enhance transparency and overall engagement in the research process. Data Collection Applying the TISJR framework during data collection calls for the establishment and maintenance of trust with participants to be a high priority, which ultimately contributes to a sense of safety. Therefore, the research team prioritized consistency and reliability by keeping appointments and being available as requested for a survey. In addition, an effort was made to address the physical needs of participants to enhance their ability to focus and support self-regulation while completing the survey (Perry, 2006). Offering food and a choice in the type of snack further enhanced the men’s sense of emotional safety. Making efforts to ensure the comfort of participants is also conducive to establishing rapport and a beginning sense of trust. Providing sustenance to the men facilitated relaxation while enhancing their ability to focus on the interview (Perry, 2006). Furthermore, certain practices can be adopted by researchers to respond to trauma responses in study participants and enhance a sense of safety. For example, the case study’s research team was mindful of their emotional and behavioral regulation during all phases of the study (for example, modulating tone of voice, nonjudgmental expressions, calm affect). Using a strengths-based approach (that is, empowerment), the research team practiced affirmations with participants before beginning the survey (for example, “I want to acknowledge your willingness to show up for our appointment today”), after the survey (for example, “This survey can be tough for a lot of people, whether it’s the questions asked or the time it takes to answer. I appreciate that you’re sharing your experiences with us”), and during check-ins if participants shared life adversities (for example, “It must have taken great strength for you to share this with me today”). Participants were allotted as much time as needed to complete the survey. Researchers asked participants how they were feeling at several points throughout the survey, to assess participants’ level of comfort. Using this proactive approach, researchers were able to keep a steady assessment of participants’ emotional status and level of stress while completing the survey (that is, safety) and provide opportunities to support self-regulation. For example, observing agitation or discomfort, researchers offered opportunities to take breaks, guide breathing exercises, and conclude the study if necessary, as well as opportunities to discuss emotional or physical responses to the survey. Protocols were also developed for more extreme reactions to triggers, such as suicidal or homicidal thoughts. It is notable that no participants experienced any extreme responses during the course of the study. Researchers should refer to “Data Collection” in Table 1 to assess for potential safety issues relative to their studies and generate ideas to promote this tenet by building and maintaining trust among researchers and participants. Post–Data Collection Researchers should prioritize empowerment, voice, and choice, and safety after participants have completed the data collection (see Table 1). In the case study, for example, additional procedures and practices were followed after the interview to mitigate any potential stress response by participants. That is, researchers conducted a brief check-in with participants, acknowledging the challenging nature of the survey (for example, time, effort, topics), and assessed participants’ level of stress through participants’ verbal cues (for example, comments about the experience and process) and nonverbal cues (for example, physical distress). This provided another opportunity to support the emotion regulation of participants. Each participant was offered the option to participate in a grounding exercise that lasted approximately 60 to 90 seconds. Researchers introduced this exercise as a short activity that can help to “reset” after stressful activities, and then guided the participant through deep breathing. This grounding exercise is based in research that emphasizes the mind–body connection (Varvogli & Darviri, 2011); that is, if the body is relaxed, the mind will also be relaxed. Research has shown that guided deep breathing exercises lasting as little as 60 to 90 seconds can achieve measurable gains in relaxation following acute stressful tasks (Varvogli & Darviri, 2011) and thus aid in the emotional and physical regulation of participants before departing from the interview. To enhance the participants’ systems of support, services were available for free or on a sliding scale, and local area resources tailored to the study sample (that is, low-income, men, racial and ethnic minorities) and area of study (that is, employment, mental health services, physical health services) were provided at the completion of each session. To enhance collaboration and mutuality, researchers offered to conduct “warm handoffs” (that is, researchers calling providers and making a transitioning call to participants). Barriers to Application Applying the TISJR framework will have challenges and, likely, researchers will apply this framework to varying degrees with each study, making accessible changes to study protocols throughout individual studies and improving with each subsequent study. For example, in the case study the individual differences between the research team and the study sample (that is, race, gender, nationality, histories of violence against women) created some obstacles to establishing and maintaining trust, establishing a sense of safety, expressing empathy, and empowerment. For example, participants’ histories of violence against women impeded expressing empathy by way of sharing personal information to establish rapport (that is, social location in the world) because of concerns for safety. Along these lines, the research team found it challenging to express empathy while still being authentic when some men presented a hypermasculine persona or “hardened exterior.” For example, men expressed comments disrespecting women (for example, blaming the victim), challenged female researchers on the basis of gender (for example, flirting), and confronted male researchers on the basis of race and nationality. As a result, we consistently strategized how to establish healthy boundaries with participants, serving as a model for our participants and prioritizing safety, empowerment, and a voice for researchers and participants alike. Applying a trauma and social justice lens, the research team understood that these expressions may have been a manifestation of survival tactics learned by trauma survivors to protect themselves from past and future harm (Mathews, Jewkes, & Abrahams, 2011; Voith et al., 2018). Thus, understanding this behavior, researchers practiced emotional and behavior regulation, while remaining empathetic and authentic, relying on reflective, strengths-based statements such as “I hear that you are really struggling with your partner” or “I appreciate how hard it may feel to share personal information about yourself and am grateful for your courage to do so.” After completing the survey, several men described how some of the questions made them feel stigmatized. Although substantial changes to the survey could not be made at that point, this feedback provides valuable information to this approach: When using healing-centered engagement in research with disempowered populations, the content of the study itself must align with these principles. Researchers should consider whether the research questions, variables of interest, and subsequent measurements reflect both risk and protective factors, and violate or uphold key principles of TIC and social justice, especially resisting retraumatization. Though not part of the original study, the research team recognized that the original design did not align fully with these principles and therefore conceptualized a qualitative component as a follow-up with participants to ensure that their voices were central to the study. Discussion Healing-centered engagement, initially conceptualized in youth work, focuses on using individual and collective strengths and cultural experiences to holistically improve the collective well-being of all individuals within the system (Ginwright, 2018). This approach views trauma not simply as an individual, isolated experience but, rather, highlights how trauma and healing are experienced collectively and emphasizes a systems approach to address it (Ginwright, 2018). Moreover, scholars have indicated that institutions and researchers leading projects should establish strong ties with racial and ethnic minority communities to address the trauma from their social exclusion and exploitation (George et al., 2014; Kanuha, 2000; Yancey, Ortega, & Kumanyika, 2006). This article proposes one approach to addressing this call that is embedded in the research process itself, acknowledging the implicit role of institutions, including research and the academy, in the perpetuation of trauma among disempowered populations. The application of a trauma-informed, socially just framework for social work research is necessary given the orientation of the profession, the contexts and settings in which social work research is conducted, and the focus on disempowered populations. This study demonstrates that it is feasible to implement this framework and provides concrete steps at each stage of study. Notably, barriers to implementation and “lessons learned” are also highlighted, providing recommendations for future researchers to improve on. Limitations Elements of the framework (for example, sense of safety, perception of transparency) were not directly measured, which would have provided more information on the framework effects. The quality of data, however, was assessed including data completeness and the participant’s rationale for incomplete data. Participant’s willingness to provide sensitive information can serve as a proxy for key elements of the framework, such as a sense of safety and trust. The average overall quality data score was 4.38 (SD = 0.92) out of 5. With any surveys that received lower than a 5 or “excellent” (38.5%), researchers probed for rationales of incomplete data. Of those probed for further information, only three people provided answers indicating discomfort, such as “the questions were too sensitive.” Future researchers should include process evaluation assessments, including participant’s perceptions of safety, transparency, trust, collaboration and mutuality, and support systems through the lens of intersectionality. Implications and Future Directions Researchers have questioned the traditional nature of the relationship between the university and the community or the “researcher–researched” relationship (Beebeejaun, Durose, Rees, Richardson, & Richardson, 2015). Although the dominant view of this relationship is “do no harm,” researchers on the vanguard have considered how to reshape the narrative and process of this partnership to move to a more researcher-reflexive practice that favors coproduction of knowledge and solutions with communities (Beebeejaun et al., 2015). Social work researchers are well positioned to help reshape this narrative, given that their work must be informed by the core value of social justice, with the goal of dismantling unequal access to systems and societal institutions (Rountree & Pomeroy, 2010). This case study demonstrates a number of concrete steps that researchers can incorporate into study designs with relative ease to begin reshaping the narrative. For example, researchers can incorporate qualitative or mixed methods to amplify the voices and centralize the experiences of participants’ intersectional identities; research questions and analyses can be person centered, including both risk and protective factors; protocols can be designed to acknowledge and minimize power differentials between researchers and participants to promote collaboration and transparency; and informed consent documents and survey instruments can be presented in ways that minimize cognitive overload and allow participants to maintain control throughout the process (for example, build in “refuse to answer” response option). Researchers can tailor study changes using the key questions in the inventory (see Table 1). Longer-term shifts are also necessary to fully actualize this trauma-informed social justice research framework. First, social work researchers must build strong partnerships with and use the expertise of practitioners and community representatives to understand and interrupt underlying mechanisms affecting the disproportionate rates of trauma exposure and related health disparities among marginalized populations. Researchers also need to understand how societal and organizational structures implicit in the institution of research can perpetuate harm or dismantle the oppression of disempowered groups. To that end, participatory action research methods are recommended when applying healing-centered engagement in research. Second, the measures and methods used to collect data may have to be adapted. The cultural validity of the measurements used in studies with racial and ethnic minorities is questionable, given that assessment norms are often derived from the samples that do not fully represent minority groups (Fisher et al., 2002). Third, social work doctoral programs should consider including student training in the application of the TISJR framework. Reflecting the core value of social justice, social work researchers must consider the institution of research itself and begin to rebuild it from the inside out. Although these shifts may not be easy or immediate, they stand to make the act of research itself a conduit for social justice. References Alexander M. ( 2012 ). The new Jim Crow: Mass incarceration in the age of colorblindness . New York : New Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC American Psychological Association. ( 2017 ). Trauma. Retrieved from http://www.apa.org/topics/trauma/ Beebeejaun Y. , Durose C. , Rees J. , Richardson J. , Richardson L. ( 2015 ). Public harm or public value? Towards coproduction in research with communities . Environment and Planning C: Government and Policy, 33 , 552 – 565 . Google Scholar Crossref Search ADS WorldCat Bloom S. L. ( 1999 , October). Trauma theory abbreviated. In Pennsylvania Attorney General Mike Fisher’s Family Violence Task Force, Final action plan: A coordinated community-based response to family violence ( pp. 1 – 17 ). Harrisburg: Pennsylvania Office of the Attorney General, Family Violence Task Force. OpenURL Placeholder Text WorldCat Bloom S. L. ( 2018 ). The sanctuary model: Trauma-informed, trauma-responsive culture. Retrieved from http://sanctuaryweb.com/PublicHealth/Trauma-ResponsivePrograms/TheSanctuaryModelTrauma-Informed,Trauma-ResponsiveCulture.aspx Bowen E. A. , Murshid N. S. ( 2016 ). Trauma-informed social policy: A conceptual framework for policy analysis and advocacy . American Journal of Public Health, 106 , 223 – 229 . Google Scholar Crossref Search ADS WorldCat Bracken-Roche D. , Bell E. , Macdonald M. E. , Racine E. ( 2017 ). The concept of ‘vulnerability’ in research ethics: An in-depth analysis of policies and guidelines . Health Research Policy and Systems, 15 ( 1 ), Article 8. Google Scholar OpenURL Placeholder Text WorldCat Braxton-Newby M. , Jones T. ( 2014 ). Authentic engagement: Practice with African American males . Journal of African American Studies, 18 , 260 – 265 . Google Scholar Crossref Search ADS WorldCat Briere J. , Scott C. ( 2012 ). Principles of trauma therapy: A guide to symptoms, evaluation, and treatment . Los Angeles: SAGE . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Briere J. , Scott C. ( 2015 ). Complex trauma in adolescents and adults. Psychiatric Clinics, 38 , 515 – 527 . Google Scholar OpenURL Placeholder Text WorldCat Burkard A. W. , Knox S. ( 2004 ). Effect of therapist color-blindness on empathy and attributions in cross-cultural counseling . Journal of Counseling Psychology, 51 , 387 – 397 . Google Scholar Crossref Search ADS WorldCat Butler L. D. , Critelli F. M. , Rinfrette E. S. ( 2011 ). Trauma-informed care and mental health . Directions in Psychiatry, 31 ( 3 ), 197 – 212 . Google Scholar OpenURL Placeholder Text WorldCat Carlson E. B. , Newman E. , Loewenstein R. J. ( 2003 ). Distress in response to and perceived usefulness of trauma research interviews . Journal of Trauma & Dissociation, 4 ( 2 ), 131 – 142 . Google Scholar Crossref Search ADS WorldCat Crenshaw K. ( 1989 ). Demarginalizing the intersection of race and sex: A black feminist critique of antidiscrimination doctrine, feminist theory and antiracist politics. University of Chicago Legal Forum, 1, Article 8. Cronholm P. F. , Forke C. M. , Wade R. , Bair-Merritt M. H. , Davis M. , Harkins-Schwarz M. , et al. ( 2015 ). Adverse childhood experiences: Expanding the concept of adversity . American Journal of Preventive Medicine, 49 , 354 – 361 . Google Scholar Crossref Search ADS WorldCat Felitti V. J. , Anda R. F. , Nordenberg D. , Williamson D. F. , Spitz A. M. , Edwards V. , Marks J. S. ( 1998 ). Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: The Adverse Childhood Experiences (ACE) Study . American Journal of Preventive Medicine, 14 , 245 – 258 . Google Scholar Crossref Search ADS WorldCat Fink A. ( 2012 ). How to conduct surveys: A step-by-step guide . Thousand Oaks, CA : SAGE . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Fisher C. B. , Hoagwood K. , Boyce C. , Duster T. , Frank D. A. , Grisso T. , et al. ( 2002 ). Research ethics for mental health science involving ethnic minority children and youths . American Psychologist, 57 , 1024 – 1039 . Google Scholar Crossref Search ADS WorldCat Ford J. , Courtois C. ( 2009 ). Defining and understanding complex trauma and complex traumatic stress disorders. In Courtois C. Ford J. (Eds.), Treating complex traumatic stress disorders (pp. 13 – 30 ). New York : Guilford Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC George S. , Duran N. , Norris K. ( 2014 ). A systematic review of barriers and facilitators to minority research participation among African Americans, Latinos, Asian Americans, and Pacific Islanders . American Journal of Public Health, 104 ( 2 ), e16 – e31 . Google Scholar Crossref Search ADS WorldCat Ginwright S. ( 2015 ). Hope and healing in urban education: How urban activists and teachers are reclaiming matters of the heart . New York : Routledge . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Ginwright S. ( 2018 ). The future of healing: Shifting from trauma informed care to healing centered engagement. Retrieved from https://medium.com/@ginwright/the-future-of-healing-shifting-from-trauma-informed-care-to-healing-centered-engagement-634f557ce69c Given L. M. (Ed.). ( 2008 ). The SAGE encyclopedia of qualitative research methods. Thousand Oaks, CA : SAGE . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Glynn M. ( 2016 ). Towards an intersectional model of desistance for black offenders . Safer Communities, 15 ( 1 ), 24 – 32 . doi:10.1108/SC-05-2015-0016 Google Scholar Crossref Search ADS WorldCat Harris M. , Fallot R. D. ( 2001 ). Envisioning a trauma-informed service system: A vital paradigm shift . New Directions for Mental Health Services, 2001 ( 89 ), 3 – 22 . Google Scholar Crossref Search ADS WorldCat Henretty J. R. , Levitt H. M. ( 2010 ). The role of therapist self-disclosure in psychotherapy: A qualitative review . Clinical Psychology Review, 30 ( 1 ), 63 – 77 . Google Scholar Crossref Search ADS WorldCat High S. ( 2003 ). Industrial sunset: The making of North America's rust belt, 1969-1984 . Canada : University of Toronto Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Jaffe A. E. , DiLillo D. , Hoffman L. , Haikalis M. , Dykstra R. E. ( 2015 ). Does it hurt to ask? A meta-analysis of participant reactions to trauma research . Clinical Psychology Review, 40 , 40 – 56 . Google Scholar Crossref Search ADS WorldCat Jones J. A. ( 2012 ). Ethical considerations in criminal justice research: Informed consent and confidentiality . Inquiries Journal, 4 ( 8 ), 1 – 2 . Google Scholar OpenURL Placeholder Text WorldCat Kanuha V. K. ( 2000 ). “Being” native versus “going native”: Conducting social work research as an insider [Special Section: Alternative Writing Formats]. Social Work, 45 , 439 – 447 . Google Scholar Crossref Search ADS WorldCat Koenen K. , Ratanatharathorn A. , Ng L. , McLaughlin K. , Bromet E. , Stein D. , et al. ( 2017 ). Posttraumatic stress disorder in the World Mental Health Surveys . Psychological Medicine, 47 , 2260 – 2274 . Google Scholar Crossref Search ADS WorldCat Leitch L. ( 2017 ). Action steps using ACEs and trauma-informed care: A resilience model . Health & Justice, 5 , Article 5. doi:10.1186/s40352-017-0050-5 Google Scholar OpenURL Placeholder Text WorldCat Massey D. S. , Tannen J. ( 2015 ). A research note on trends in black hypersegregation . Demography, 52 , 1025 – 1034 . Google Scholar Crossref Search ADS WorldCat Mathews S. , Jewkes R. , Abrahams N. ( 2011 ). ‘I had a hard life’: Exploring childhood adversity in the shaping of masculinities among men who killed an intimate partner in South Africa . British Journal of Criminology, 51 , 960 – 977 . Google Scholar Crossref Search ADS WorldCat Mattsson T. ( 2014 ). Intersectionality as a useful tool: Anti-oppressive social work and critical reflection . Affilia, 29 ( 1 ), 8 – 17 . Google Scholar Crossref Search ADS WorldCat Mersky J. P. , Janczewski C. E. , Topitzes J. ( 2017 ). Rethinking the measurement of adversity: Moving toward second-generation research on adverse childhood experiences . Child Maltreatment, 22 , 58 – 68 . Google Scholar Crossref Search ADS WorldCat Metzler M. , Merrick M. T. , Klevens J. , Ports K. A. , Ford D. C. ( 2017 ). Adverse childhood experiences and life opportunities: Shifting the narrative . Children and Youth Services Review, 72 , 141 – 149 . Google Scholar Crossref Search ADS WorldCat Murphy Y. , Hunt V. , Zajicek A. M. , Norris A. N. , Hamilton L. ( 2009 ). Incorporating intersectionality in social work practice, research, policy, and education. Washington, DC : NASW Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC National Association of Social Workers. ( 2017 ). Code of ethics of the National Association of Social Workers. Retrieved from https://www.socialworkers.org/About/Ethics/Code-of-Ethics Perry B. D. ( 2006 ). Fear and learning: Trauma-related factors in the adult education process . New Directions for Adult and Continuing Education, 2006 ( 110 ), 21 – 27 . Google Scholar Crossref Search ADS WorldCat Quiros L. , Berger R. ( 2015 ). Responding to the sociopolitical complexity of trauma: An integration of theory and practice .  Journal of Loss and Trauma, 20 ( 2 ), 149 – 159 . Google Scholar Crossref Search ADS WorldCat Ransby B. ( 2018 ). Making all black lives matter: Reimagining freedom in the twenty-first century (Vol. 6). Oakland : University of California Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Reavis J. A. , Looman J. , Franco K. A. , Rojas B. ( 2013 ). Adverse childhood experiences and adult criminality: How long must we live before we possess our own lives? Permanente Journal, 17 ( 2 ), 44 – 48 . Google Scholar Crossref Search ADS WorldCat Rountree M. A. , Pomeroy E. C. ( 2010 ). Bridging the gaps among social justice, research, and practice [Editorial ]. Social Work, 55 , 293 – 295 . Google Scholar Crossref Search ADS WorldCat RYSE Center. ( 2015 ). Disorder versus distress. Retrieved from https://static1.squarespace.com/static/58ece61644024383be911a95/t/593e579b37c58172ed51340c/1497257886219/ACEs_social-location_2015.pdf Seng J. S. , Lopez W. D. , Sperlich M. , Hamama L. , Reed Meldrum C. D. ( 2012 ). Marginalized identities, discrimination burden, and mental health: Empirical exploration of an interpersonal-level approach to modeling intersectionality . Social Science & Medicine, 75 , 2437 – 2445 . Google Scholar Crossref Search ADS WorldCat Sewell A. A. , Jefferson K. A. , Lee H. ( 2016 ). Living under surveillance: Gender, psychological distress, and stop-question-and-frisk policing in New York City . Social Science & Medicine, 159 , 1 – 13 . Google Scholar Crossref Search ADS WorldCat Sherman J. ( 2013 ). Surviving the great recession: Growing need and the stigmatized safety net . Social Problems, 60 , 409 – 432 . Google Scholar Crossref Search ADS WorldCat Shonkoff J. P. , Garner A. S. , Siegel B. S. , Dobbins M. I. , Earls M. F. , McGuinn L. , et al. ( 2012 ). The lifelong effects of early childhood adversity and toxic stress . Pediatrics , 129 ( 1 ), e232 – e246 . Google Scholar Crossref Search ADS PubMed WorldCat Smith J. R. , Patton D. U. ( 2016 ). Posttraumatic stress symptoms in context: Examining trauma responses to violent exposures and homicide death among Black males in urban neighborhoods . American Journal of Orthopsychiatry, 86 , 212 – 223 . Google Scholar Crossref Search ADS WorldCat Sotero M. ( 2006 ). A conceptual model of historical trauma: Implications for public health practice and research . Journal of Health Disparities Research and Practice, 1 ( 1 ), 93 – 108 . Google Scholar OpenURL Placeholder Text WorldCat Substance Abuse and Mental Health Services Administration. ( 2014 ). Trauma-informed care in behavioral health services [Treatment Improvement Protocol (TIP) Series 57, HHS Publication No. (SMA) 13-4801]. Rockville, MD : Author . Google Scholar PubMed OpenURL Placeholder Text Google Preview WorldCat COPAC United States of America v. City of Cleveland, No. 1:15-cv-01046 (N.D. Oh. 2015). Retrieved from https://www.justice.gov/sites/default/files/crt/legacy/2015/05/27/cleveland_agreement_5-26-15.pdf van der Kolk B. A. ( 2003 ). The neurobiology of childhood trauma and abuse . Child and Adolescent Psychiatric Clinics of North America, 12 , 293 – 318 . Google Scholar Crossref Search ADS WorldCat van der Kolk B. A. ( 2005 ). Developmental trauma disorder: Toward a rational diagnosis for children with complex trauma histories . Psychiatric Annals, 35 , 401 – 408 . Google Scholar Crossref Search ADS WorldCat Varvogli L. , Darviri C. ( 2011 ). Stress management techniques: Evidence-based procedures that reduce stress and promote health . Health Science Journal, 5 ( 2 ), 74 – 89 . Google Scholar OpenURL Placeholder Text WorldCat Vásquez E. , Udo T. , Corsino L. , Shaw B. A. ( 2019 ). Racial and ethnic disparities in the association between adverse childhood experience, perceived discrimination and body mass index in a national sample of U.S. older adults . Journal of Nutrition in Gerontology and Geriatrics, 38 , 6 – 17 . Google Scholar Crossref Search ADS WorldCat Voith L. A. , Logan-Greene P. , Strodthoff T. , Bender A. E. ( 2018 ). A paradigm shift in batterer intervention programming: A need to address unresolved trauma . Trauma, Violence, & Abuse. Advance online publication. doi:10.1177/1524838018791268 Google Scholar OpenURL Placeholder Text WorldCat Wolff N. , Chugo M. G. , Shi J. , Huening J. , Frueh B. C. ( 2015 ). Screening for PTSD among incarcerated men: A comparative analysis of computer-administered and orally administered modalities . Criminal Justice and Behavior, 42 , 219 – 236 . Google Scholar Crossref Search ADS WorldCat Yancey A. K. , Ortega A. N. , Kumanyika S. K. ( 2006 ). Effective recruitment and retention of minority research participants . Annual Review of Public Health, 27 ( 1 ), 1 – 28 . Google Scholar Crossref Search ADS WorldCat © 2020 National Association of Social Workers This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Chen,, Zibei;Livermore,, Michelle

2020 Social Work Research

doi: 10.1093/swr/svaa009

Abstract As reliance on alternative financial services (AFS) usage continues its exponential expansion among American families, policy debates over banking regulation perdure with limited empirical understanding of how usage affects individuals’ financial lives. Using data from the 2014 Survey of Household Economics and Decisionmaking, this study explored the association between AFS use and financial well-being using a nationally representative sample (N = 5,896). It also examined the role of household income in AFS use and its relation to financial well-being. Results from regression analyses indicated that AFS use was negatively associated with present financial security measured by credit score, making ends meet, subjective financial well-being, and credit card payment, and that future financial security was measured by having an emergency fund and a rainy-day fund. Moreover, the interaction models showed that lower-income groups had the most negative associations between AFS and most financial well-being indicators, suggesting a substantive role of income in exacerbating the negative relationships. This was the first known study linking use of AFS and household financial well-being with a focus on the role of income. The article concludes with a discussion of implications for policy and social work practice. Controversy surrounds alternative financial services (AFS), the financial products and services that operate outside the federal insurance system (Martin & Longa, 2012). Consumer advocates argue that AFS prey on those who are financially vulnerable by charging excessive fees and setting financial traps (Barr, 2002; Caskey, 1994, 2006, 2010). In contrast, industry supporters contend that AFS products simply meet the needs of those underserved by mainstream financial institutions and that the products’ short-term nature and the high-risk profile of clients justify high transaction costs (Elliehausen, 2009, 2011). This controversy is currently being debated at the federal level. In 2017, the Consumer Financial Protection Bureau (CFPB) clarified the federal government’s position in a rule that identified the provision of short- or long-term balloon-payment loans as abusive if they fail to consider the consumer’s ability to pay back the loan. However, at the time of this writing, the CFPB was taking public comments regarding their proposal to rescind this rule (CFPB, 2019). Amid the controversy, empirical knowledge about whether and to what extent AFS use affects household finance remains limited. Legal (see, for example, Austin, 2004) and economic (see, for example, Morse, 2011) analyses of the industry dominate the literature, with only a handful of studies examining the population who use AFS and the relationship between AFS use and financial well-being. Social work’s historic concern about financial deprivation and poverty makes this a particularly troublesome oversight. As the AFS industry continues to grow, it is imperative to understand the role of AFS use in financial lives of individuals and families. Many low- to moderate-income (LMI) families live paycheck to paycheck and need small-dollar loans and quick financial transactions in a market where traditional banks fail to respond to this demand. Although the use of AFS helps some LMI families cope with financial shocks, little is known about how AFS enables these families to weather other financial difficulties and affects families’ long-term financial outcomes. Knowledge regarding the relationship between AFS use and financial well-being is needed to inform banking regulation policies in the future. Using data from the 2014 Survey of Household Economics and Decisionmaking (SHED), this study explored the association between AFS use and financial well-being with a focus on the role of household income. Literature Review Also known as a fringe economy, AFS are defined as a broad range of financial institutions that serve individuals who, for a myriad of reasons, cannot or do not use traditional financial services, such as checking and saving accounts, standard personal loans, and investment portfolios (Caskey, 1994, 1997). AFS products have been classified as either transactional or credit-based (Federal Deposit Insurance Cooperation [FDIC], 2013). Transactional AFS are primarily composed of nonbank check cashing and money orders that facilitate money transferring at fixed rates. For example, check cashers typically charge 1% to 4% of the face value of the check, depending on the check issuers and subject to limitations of state law. Credit-based AFS, in contrast, provide a short-term, small credit that is secured with payday check or personal possession; these include payday loans, payday advances, pawn shop lending, and auto title loans. A payday loan, for example, typically has a two-week term coinciding with a pay cycle and offers a small credit with annual percentage rates (APRs) of 30% to 400%. Pawn lending is also a short-term, secured lending transaction in which the lender typically takes physical possession of the item securing the loan with APRs ranging from 12% to 300%. Auto title lending is similar to pawn lending except that title lenders make short-term loans secured by clear car titles (FDIC, 2013). In contrast to services offered by traditional banking institutions, services offered by AFS providers are not federally insured and are often high-cost. Who Uses AFS AFS use varies along a number of demographic and socioeconomic characteristics (for example, age, race, and household income). Studies show that AFS use was more prevalent among young adults (Burhouse et al., 2014; Lusardi & de Bassa Scheresberg, 2013), racial minorities (Gross, Hogarth, Manohar, & Gallegos, 2012; Rhine & Greene, 2006), individuals with lower levels of education and income (CFPB, 2013; Lim et al., 2014; Lusardi & de Bassa Scheresberg, 2013), and families with dependent children (see, for example, Lusardi & de Bassa Scheresberg, 2013; Pew Charitable Trusts, 2012). For example, Elliehausen (2011) reviewed several data sources (for example, Survey of Consumers) regarding characteristics of AFS users and found that those who were 35 and younger demonstrated a greater likelihood of using AFS products such as payday loans, rent-to-own (RTO) products, and refund anticipation loans. Also, a national study of RTO customers found that African Americans were more likely to use payday loans and RTO than other racial and ethnic groups, controlling for income and educational differences (Lacko, McKernan, & Hastak, 2000). In addition, individuals with low levels of education (for example, less than college) showed a significantly higher frequency of RTO transactions (see, for example, Lacko et al., 2000). Financial Circumstance Income was another important predictor of AFS use, with studies showing that AFS users were disproportionally represented by LMI segments of the population. For example, individuals taking out payday loans were likely to have annual household incomes between $25,000 and $50,000, whereas those who used tax refund anticipation loans were likely to earn between $15,000 to $40,000 (see, for example, Weller & Logan, 2009). In addition to income, researchers found that other indicators of financial conditions (for example, home ownership, credit availability, and incidence of financial shocks) were associated with AFS use. Home ownership was indicative of a person’s connection to traditional banks (through a home loan mortgage) mostly because homeowners had access to mainstream banking options (Shobe, Christy, Givens, & Murphy-Erby, 2013). Relatedly, renters were two to three times more likely to use AFS (Northwood & Rhine, 2018; Shobe et al., 2013) and individuals with low credit limits were three to four times more likely to use AFS (Prager, 2014). In addition, families experiencing financial shocks and economic strains displayed a propensity toward AFS use (Gross et al., 2012; Lim et al., 2014; Lusardi & de Bassa Scheresberg, 2013). Overall, research suggests that AFS provided an easier, timely way to obtain money when household financial circumstances prevented individuals from otherwise meeting their financial needs. Potential Consequence of AFS Use In regard to the relationship between AFS use and financial well-being, the published literature has provided very limited and mixed findings. Approximately one dozen empirical studies examined the association between AFS use and financial well-being, and most focused on payday loans. Studies have linked AFS use to an individual’s amount of debt (Morgan, 2007; Skiba & Tobacman, 2011), likelihood of filing for bankruptcy and home foreclosure (Mayer, 2004; Morgan & Strain, 2007; Morse, 2011; Skiba & Tobacman, 2011), credit records (Bhutta, 2014), job retention, mental health status (Karlan & Zinman, 2010), and material hardship (for example, food consumption, late bill payment) (Melzer, 2011; Morgan, 2007; Zinman, 2010). Several studies investigated the relationship between AFS use and indicators of neighborhood well-being, such as property crime (Morse, 2011). Findings from the corpus of studies on payday loans have indicated a complex relationship. Specifically, some found that access to payday loans was related to elevated rates of bankruptcy (Skiba & Tobacman, 2011), declines in job performance (Carrell & Zinman, 2014), increased difficulty paying bills (Melzer, 2011), and increased likelihood of losing one’s bank account (Campbell, Martínez-Jerez, & Tufano, 2012). In contrast, other studies have indicated that use of payday loans was positively associated with job retention and financial well-being (Zinman, 2010) and a reduced possibility of consumer complaints against lenders (Morgan & Strain, 2007). Two of these latter studies (that is, Morgan & Strain, 2007; Zinman, 2010) indicated that restricting access to payday loans led people to turn to more costly financial behaviors (for example, overdrafting bank accounts and paying bills late). Several studies showed little or no association between payday loans and credit scores or other credit record outcomes (Bhutta, 2014; Bhutta, Skiba, & Tobacman, 2015). Overall, the extant research examining the association between AFS use and financial well-being is limited and inconclusive. Although the body of research provides a preliminary understanding of the impact of AFS use on certain outcomes, it has included only a few financial well-being indicators (for example, home foreclosure, bankruptcy, job retention, bill payment, credit score) and no research has used a theory-based financial well-being framework to systematically investigate well-being correlates of AFS use. Consequently, a considerable knowledge gap remains regarding how AFS use affects financial well-being and whether associations differ by income. To address this gap, this study used a theoretical framework of financial well-being developed by the CFPB to investigate the association between AFS use and financial well-being. According to the CFPB (2015), financial well-being is a state of being wherein a person is able to meet current and ongoing financial obligations, feels secure in their financial future, and is able to make choices that allow enjoyment of life. CFPB further specified that the definition of financial well-being includes the extent to which individuals (a) have control over day-to-day, month-to-month finances; (b) have the capacity to absorb a financial shock; (c) are on track to meet financial goals; and (d) have the financial freedom to make the choices that allow them to enjoy life (CFPB, 2015). The CFPB developed this framework through in-depth, one-on-one interviews with 59 consumers and 30 financial professionals. We chose this financial well-being framework to guide this analysis because of its grounding in expert opinions, literature, and the experiences and voices of consumers. The study had two research questions: (1) Is there a significant association between AFS use and financial well-being indicators? and (2) Do the associations between AFS and financial well-being indicators differ by income? Method This study used data from the 2014 SHED, a nationally representative data set focusing on the economic situation of American households. The Federal Reserve Board’s Division of Consumer and Community Affairs designed the SHED; GfK, an online consumer research company, administered it. Since 2013, the SHED has collected data annually from a representative sample of American households on a range of  household finance topics. These topics included economic fragility, saving, spending, banking, and credit usage. The purpose of SHED was to complement the existing base of knowledge from other data sources such as Survey of Consumer Finances (Larrimore, Arthur-Bentil, Dodini, & Thomas, 2015). Data Collection The SHED data were collected through KnowledgePanel, a probability-based Web panel of randomly sampled households. Specifically, GfK sent invitations to a random selection of residential postal addresses, asking interested individuals to complete a survey online. GfK provided a laptop and Internet access to those who were contacted and interested in participating but did not have the technological means to do so. The average time that respondents needed to complete the survey was approximately 19 minutes. SHED data are publicly available and were downloaded from the Federal Reserve Board Web site for this study. The current study used 2014 SHED data collected from a sample of 5,896 randomly selected American households. SHED used three approaches to collect data: (1) a random selection of 2,190 out of 4,134 respondents who participated in 2013 SHED, (2) a random selection of an additional 4,059 respondents, and (3) a random sample of 2,726 respondents with household income under $40,000. The last approach aimed to improve the precision of estimates for the low-income population (Larrimore, Arthur-Bentil, et al., 2015); 5,896 individuals responded to the e-mail request to participate and completed the survey, with a 65.6 percent completion rate. Researchers adopted several strategies to enhance completion rates, including sending e-mail reminders to nonrespondents and offering raffle, lottery, and monetary incentives to participants for completing the survey. According to the SHED report , the raw distribution of the GfK sample mirrors that of U.S. adults fairly closely (Larrimore, Arthur-Bentil, et al., 2015). The SHED used the 2014 Current Population Survey benchmarks to adjust its sample to match the U.S. population based on geodemographic information including gender, age, race and ethnicity, education, census region, household income, home ownership status, metropolitan area status, and Internet access. Survey Instrument and Measures The 2014 SHED survey instrument contained approximately 100 closed-ended questions on financial well-being, housing, economical fragility, savings and spending, access to credit, education and student loans, and retirement planning. The CFPB designed the survey in consultation with Federal Reserve System staff and outside academics with relevant research backgrounds. Study variables included six financial well-being measures, six AFS use measures, and nine sociodemographic measures. Dependent Variables Dependent variables included two dimensions of financial well-being: present financial security and future financial security. According to the CFPB (2015), present financial security is control over daily finances. The four indicators in this study were self-reported credit score, credit card bill payment, making ends meet, and subjective financial well-being. Specifically, the survey asked respondents to report or guess if their current credit score (such as FICO score) was excellent (coded 5), very good (coded 4), good (coded 3), fair (coded 2), or poor (coded 1). The survey also asked if respondents always paid their monthly credit card bills in full (positive responses were coded 1, negative responses were coded 0). The “making ends meet” measure assessed whether a household’s total spending was more than its income (coded 3), the same as its income (coded 2), or less than its income (coded 1). The subjective financial well-being measure indicated how well the respondent thought they were managing financially these days. Response options included finding it difficult to get by (coded 1), just getting by (coded 2), doing OK (coded 3), and living comfortably (coded 4). Future financial security is the capacity to absorb a financial shock (CFPB, 2015), and was assessed by two measures. Individuals noted whether they had set aside emergency or rainy-day funds that would cover expenses for three months in case of sickness, job loss, economic downturn, or other emergencies. Answers were coded as 1 = yes, 0 = no. The survey also asked individuals about the largest emergency expense they could pay right now using cash or money in their checking or saving account. Options were under $100, $100 to $199, $200 to $299, $300 to $399, and over $400. Responses were coded 1 = over $400, 0 = other. Pairwise correlations between well-being types indicate positive relationships between each of the financial well-being indicators examined in this study (data not shown). Independent Variables The key variable of interest was respondent use of one or more of the six AFS. The six AFS are check cashing service, money order, pawn shop loan, auto title loan, payday or deposit advance, and payday loan. This variable was coded 1 if respondents used one or more AFS, 0 if no AFS were used. The other key independent variable was household income, collected as an ordinal variable with 19 categories. We recoded the income variable into a quartile variable having four equally distributed categories for the interaction models. The quartile income variable included four groups: bottom ($2,500 to $22,500), lower middle ($27,500 to $37,500), upper middle ($45,000 to $80,000), and high-income quartile ($92,500 to $175,000). Control variables included several sociodemographic and household characteristics. Individual-level sociodemographic variables included age, gender, marital status, educational attainment, race, and ethnicity. Respondent age was a continuous variable. The remaining dichotomous variables indicated presence (coded as 1) or absence (coded as 0) of each condition. Four categories comprised the race and ethnicity variable: White, Black, Hispanic, and others (including one or more races). The dichotomous race variable indicated whether the respondent was White. The dichotomous gender variable indicated female or not. Employment status included three groups: employed, unemployed, and retired. The marital status variable indicated whether the respondent was married. The educational attainment variable indicated either (a) high school or GED completion or less, or (b) more education. Household-level characteristics also included dichotomous variables indicating residence in a metropolitan area (or not) and in the South (or not), and a continuous variable, household size, depicting the number of individuals within the household. Analytic Strategy Descriptive and bivariate analyses compared sample characteristics of AFS users to non-AFS users. A series of multivariate regression models predicted factors contributing to each of the six financial well-being indicators. AFS use was the independent variable of interest in all models. Ordered logit regression models were used when the dependent variable was ordinal (for example, self-reported credit score, making ends meet, subjective financial well-being). We used logit regression for each binary dependent variable (credit card payment, rainy-day fund, and emergency fund). Interaction terms assessing AFS use by income quartile were included to examine variations in the associations between AFS use and financial well-being by income. Results Sample Descriptive A small portion (15.04%) of the sample reported using AFS in the past 12 months. Descriptive statistics, presented in Table 1, showed some demographic differences between AFS users and nonusers. For instance, female respondents reported slightly higher rates of AFS use than male respondents did (15.92% and 14.07%, respectively); non-White individuals had more than three times higher percentage of AFS use than their White counterparts did (29.49% and 9.78%, respectively). The portion of AFS users among those with high school education, GED, or less was about 7% higher than among those with more than high school education (19.00% and 12.68%, respectively). Around half as many married respondents (10.86%) used AFS as nonmarried individuals (19.41%). More respondents who were unemployed (24.71%) used AFS than those who were employed (13.22%) and retired (8.54%). Respondents from households in the bottom quartile of the income distribution reported AFS use at a rate of 26.94%, more than twice as high as those in the third and highest quartile (10.41% and 5.04%, respectively). Those living in metropolitan areas (15.36%) used AFS at slightly higher rates than others, as did those living in the South (17.13%). On average, AFS use occurred more for younger individuals (44.81 years old compared with 51.72 years old) with larger families. Table 1: Descriptive Statistics by Alternative Financial Services (AFS) Usage (N = 5,896) . Full Sample . AFS Users . Non-AFS Users . Variable . % . M (SD) . % . M (SD) . % . M (SD) . Gender  Male 48 14.07 85.93  Female 52 15.92 84.08 Race  White 73.30 9.78 90.22  Nonwhite 26.70 29.49 70.51 Education  High school, GED or less 37.30 19.00 81.00  Other 62.70 12.68 87.32 Marital status  Married 51.14 10.86 89.14  Other 41.62 19.41 80.59 Employment  Employed 48.74 13.22 86.78  Unemployed 26.10 24.71 75.29  Retired 25.16 8.54 91.46 Household income  Quartile 1 27.48 26.49 73.51  Quartile 2 23.13 17.02 82.98  Quartile 3 24.76 10.41 89.59  Quartile 4 24.63 5.04 94.96 Dependent child  Yes 25.66 20.61 79.39  No 74.34 13.11 86.89 Residential region  South 35.31 17.13 82.87  Other 64.69 13.89 86.11 Metropolitan area  Yes 82.95 15.36 84.64  No 17.05 13.46 86.54 Bank account ownership  Yes 92.14 13.23 86.77  No 7.86 35.51 64.49 Financial well-being indicator  Satisfactory credit score 67.77 7.29 92.71  Paying credit card bill in full 56.03 8.23 91.77  Making ends meet 77.09 13.02 84.95  Subjective financial well-being 61.88 9.75 90.25  Having rainy day fund to cover three month of expense 45.43 6.92 93.08  Having emergency fund of $400 60.60 7.47 92.53 Age (years) 50.66 (17.39) 44.81 (16.57) 51.72 (17.35) Household size 2.49 (1.37) 2.72 (1.58) 2.46 (1.33) . Full Sample . AFS Users . Non-AFS Users . Variable . % . M (SD) . % . M (SD) . % . M (SD) . Gender  Male 48 14.07 85.93  Female 52 15.92 84.08 Race  White 73.30 9.78 90.22  Nonwhite 26.70 29.49 70.51 Education  High school, GED or less 37.30 19.00 81.00  Other 62.70 12.68 87.32 Marital status  Married 51.14 10.86 89.14  Other 41.62 19.41 80.59 Employment  Employed 48.74 13.22 86.78  Unemployed 26.10 24.71 75.29  Retired 25.16 8.54 91.46 Household income  Quartile 1 27.48 26.49 73.51  Quartile 2 23.13 17.02 82.98  Quartile 3 24.76 10.41 89.59  Quartile 4 24.63 5.04 94.96 Dependent child  Yes 25.66 20.61 79.39  No 74.34 13.11 86.89 Residential region  South 35.31 17.13 82.87  Other 64.69 13.89 86.11 Metropolitan area  Yes 82.95 15.36 84.64  No 17.05 13.46 86.54 Bank account ownership  Yes 92.14 13.23 86.77  No 7.86 35.51 64.49 Financial well-being indicator  Satisfactory credit score 67.77 7.29 92.71  Paying credit card bill in full 56.03 8.23 91.77  Making ends meet 77.09 13.02 84.95  Subjective financial well-being 61.88 9.75 90.25  Having rainy day fund to cover three month of expense 45.43 6.92 93.08  Having emergency fund of $400 60.60 7.47 92.53 Age (years) 50.66 (17.39) 44.81 (16.57) 51.72 (17.35) Household size 2.49 (1.37) 2.72 (1.58) 2.46 (1.33) Open in new tab Table 1: Descriptive Statistics by Alternative Financial Services (AFS) Usage (N = 5,896) . Full Sample . AFS Users . Non-AFS Users . Variable . % . M (SD) . % . M (SD) . % . M (SD) . Gender  Male 48 14.07 85.93  Female 52 15.92 84.08 Race  White 73.30 9.78 90.22  Nonwhite 26.70 29.49 70.51 Education  High school, GED or less 37.30 19.00 81.00  Other 62.70 12.68 87.32 Marital status  Married 51.14 10.86 89.14  Other 41.62 19.41 80.59 Employment  Employed 48.74 13.22 86.78  Unemployed 26.10 24.71 75.29  Retired 25.16 8.54 91.46 Household income  Quartile 1 27.48 26.49 73.51  Quartile 2 23.13 17.02 82.98  Quartile 3 24.76 10.41 89.59  Quartile 4 24.63 5.04 94.96 Dependent child  Yes 25.66 20.61 79.39  No 74.34 13.11 86.89 Residential region  South 35.31 17.13 82.87  Other 64.69 13.89 86.11 Metropolitan area  Yes 82.95 15.36 84.64  No 17.05 13.46 86.54 Bank account ownership  Yes 92.14 13.23 86.77  No 7.86 35.51 64.49 Financial well-being indicator  Satisfactory credit score 67.77 7.29 92.71  Paying credit card bill in full 56.03 8.23 91.77  Making ends meet 77.09 13.02 84.95  Subjective financial well-being 61.88 9.75 90.25  Having rainy day fund to cover three month of expense 45.43 6.92 93.08  Having emergency fund of $400 60.60 7.47 92.53 Age (years) 50.66 (17.39) 44.81 (16.57) 51.72 (17.35) Household size 2.49 (1.37) 2.72 (1.58) 2.46 (1.33) . Full Sample . AFS Users . Non-AFS Users . Variable . % . M (SD) . % . M (SD) . % . M (SD) . Gender  Male 48 14.07 85.93  Female 52 15.92 84.08 Race  White 73.30 9.78 90.22  Nonwhite 26.70 29.49 70.51 Education  High school, GED or less 37.30 19.00 81.00  Other 62.70 12.68 87.32 Marital status  Married 51.14 10.86 89.14  Other 41.62 19.41 80.59 Employment  Employed 48.74 13.22 86.78  Unemployed 26.10 24.71 75.29  Retired 25.16 8.54 91.46 Household income  Quartile 1 27.48 26.49 73.51  Quartile 2 23.13 17.02 82.98  Quartile 3 24.76 10.41 89.59  Quartile 4 24.63 5.04 94.96 Dependent child  Yes 25.66 20.61 79.39  No 74.34 13.11 86.89 Residential region  South 35.31 17.13 82.87  Other 64.69 13.89 86.11 Metropolitan area  Yes 82.95 15.36 84.64  No 17.05 13.46 86.54 Bank account ownership  Yes 92.14 13.23 86.77  No 7.86 35.51 64.49 Financial well-being indicator  Satisfactory credit score 67.77 7.29 92.71  Paying credit card bill in full 56.03 8.23 91.77  Making ends meet 77.09 13.02 84.95  Subjective financial well-being 61.88 9.75 90.25  Having rainy day fund to cover three month of expense 45.43 6.92 93.08  Having emergency fund of $400 60.60 7.47 92.53 Age (years) 50.66 (17.39) 44.81 (16.57) 51.72 (17.35) Household size 2.49 (1.37) 2.72 (1.58) 2.46 (1.33) Open in new tab For all indicators, small portions of AFS users reported high levels of financial well-being. A small minority of those reporting satisfactory credit scores were AFS users (7.29%). Similarly, only 8.23% of those who reported always paying credit card bills in full were AFS users. The highest percentage of AFS users reported an ability to make ends meet (13.02%). The smallest proportions of AFS users had enough emergency savings to cover three months of expense (6.92%) and for a $400 expense (7.47%). AFS and Financial Well-Being Table 2 shows results of multivariate analyses in which we regressed six financial well-being indicators on AFS use while controlling for covariates. Each model consisted of the same set of the control variables. Results from ordered logit regressions showed that AFS users were more likely to report lower credit scores (odds ratio [OR] = 0.35, z = –13.34) and low levels of subjective financial well-being (OR = 0.61, z = –6.73), and less likely to make ends meet (OR = 0.66, z = –5.66). In addition, results from binary logistic regressions suggested that AFS use was negatively associated with full credit card bill payment (OR = 0.73, z = –2.89), having a rainy-day fund that would cover three months of expenses (OR = 0.47, z = –7.70), and having an emergency fund of $400 or more (OR = 0.42, z = –9.46). Table 2: Multivariate Regressions for Effects of Alternative Financial Services (AFS) Use on Financial Well-Being Indicators . Present Security . Future Security . . Credit Score . Making Ends Meet . Subjective Financial Well-Being . Credit Card Bill Payment . Rainy-Day Fund . Emergency Fund . Variable . OR (SE) . OR (SE) . OR (SE) . OR (SE) . OR (SE) . OR (SE) . AFS use 0.35 (.03)(OpenType)"?>*** 0.66 (.05)(OpenType)"?>*** 0.61 (.04)(OpenType)"?>*** 0.73 (.08)(OpenType)"?>** 0.47 (.05)(OpenType)"?>*** 0.42 (.04)(OpenType)"?>*** Gender 1.01 (.05) 1.13 (.06) 0.98 (.05) 1.19 (.08)(OpenType)"?>** 1.09 (.07) 1.25 (.08)(OpenType)"?>*** Age 0.97 (.01)(OpenType)"?>** 0.97 (.01)(OpenType)"?>** 0.94 (.01)(OpenType)"?>*** 0.89 (.01)(OpenType)"?>*** 0.96 (.10)(OpenType)"?>** 0.96 (.01)(OpenType)"?>*** Age2 1.00 (.00)(OpenType)"?>*** 1.00 (.00) 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** Race 1.55 (.10)(OpenType)"?>*** 0.98 (.06) 0.96 (.06) 1.18 (.09)(OpenType)"?>* 0.97 (.07) 1.27 (.09)(OpenType)"?>** Marital status 1.81 (.11)(OpenType)"?>*** 1.06 (.06) 1.54 (.09)(OpenType)"?>*** 1.37 (.10)(OpenType)"?>*** 1.57 (.11)(OpenType)"?>*** 1.49 (.11)(OpenType)"?>*** High school, GED, less 1.42 (.08)(OpenType)"?>*** 0.93 (.05) 1.17 (.06)(OpenType)"?>** 1.17 (.09)(OpenType)"?>* 1.43 (.10)(OpenType)"?>*** 1.45 (.10)(OpenType)"?>*** Unemployed 0.69 (.05)(OpenType)"?>*** 0.79 (.05)(OpenType)"?>*** 0.67 (.04)(OpenType)"?>*** 1.23 (.11)(OpenType)"?>** 0.76 (.06)(OpenType)"?>** 0.62 (.05)(OpenType)"?>*** Retired 3.31 (.47)(OpenType)"?>*** 1.96 (.26)(OpenType)"?>*** 4.92 (.67)(OpenType)"?>*** 1.10 (.19) 3.44 (.56)(OpenType)"?>*** 3.84 (.63)(OpenType)"?>*** Quartile 2 1.49 (.11)(OpenType)"?>*** 1.25 (.09)(OpenType)"?>** 1.27 (.09)(OpenType)"?>** 0.91 (.09) 1.24 (.11)(OpenType)"?>* 1.71 (.15)(OpenType)"?>*** Quartile 3 2.69 (.21)(OpenType)"?>*** 2.12 (.16)(OpenType)"?>*** 2.63 (.20)(OpenType)"?>*** 1.09 (.11) 1.85 (.17)(OpenType)"?>*** 3.00 (.27)(OpenType)"?>*** Quartile 4 4.75 (.42)(OpenType)"?>*** 3.23 (.27)(OpenType)"?>*** 6.24 (.53)(OpenType)"?>*** 1.60 (.17)(OpenType)"?>*** 3.36 (.33)(OpenType)"?>*** 6.33 (.68)(OpenType)"?>*** Household size 0.81 (.02)(OpenType)"?>*** 0.86 (.02)(OpenType)"?>*** 0.85 (.02)(OpenType)"?>*** 0.90 (.03)(OpenType)"?>** 0.80 (.03)(OpenType)"?>*** 0.84 (.03)(OpenType)"?>*** Dependent child 1.08 (.09) 0.95 (.08) 1.03 (.08) 0.85 (.09) 1.04 (.10) 1.12 (.11) Living in metro area 1.11 (.08) 0.84 (.06) 1.00 (.07) 1.07 (.09) 1.24 (.10)(OpenType)"?>** 1.16 (.10) Living in South 1.04 (.06) 0.93 (.06) 0.93 (.05) 1.13 (.09) 0.94 (.07) 0.99 (.08) Bank account user 2.26 (.26)(OpenType)"?>*** 0.98 (.10) 1.37 (.14)(OpenType)"?>** 0.81 (.17) 2.12 (.32)(OpenType)"?>*** 3.33 (.48)(OpenType)"?>*** Model significance LR χ2 (17) = 2,172.71(OpenType)"?>*** LR χ2 (17) = 534.80(OpenType)"?>*** LR χ2 (17) = 1,720.71(OpenType)"?>*** LR χ2 (17) = 348.90(OpenType)"?>*** LR χ2 (17) = 1,214.06(OpenType)"?>*** LR χ2 (17) = 1,703.76(OpenType)"?>*** . Present Security . Future Security . . Credit Score . Making Ends Meet . Subjective Financial Well-Being . Credit Card Bill Payment . Rainy-Day Fund . Emergency Fund . Variable . OR (SE) . OR (SE) . OR (SE) . OR (SE) . OR (SE) . OR (SE) . AFS use 0.35 (.03)(OpenType)"?>*** 0.66 (.05)(OpenType)"?>*** 0.61 (.04)(OpenType)"?>*** 0.73 (.08)(OpenType)"?>** 0.47 (.05)(OpenType)"?>*** 0.42 (.04)(OpenType)"?>*** Gender 1.01 (.05) 1.13 (.06) 0.98 (.05) 1.19 (.08)(OpenType)"?>** 1.09 (.07) 1.25 (.08)(OpenType)"?>*** Age 0.97 (.01)(OpenType)"?>** 0.97 (.01)(OpenType)"?>** 0.94 (.01)(OpenType)"?>*** 0.89 (.01)(OpenType)"?>*** 0.96 (.10)(OpenType)"?>** 0.96 (.01)(OpenType)"?>*** Age2 1.00 (.00)(OpenType)"?>*** 1.00 (.00) 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** Race 1.55 (.10)(OpenType)"?>*** 0.98 (.06) 0.96 (.06) 1.18 (.09)(OpenType)"?>* 0.97 (.07) 1.27 (.09)(OpenType)"?>** Marital status 1.81 (.11)(OpenType)"?>*** 1.06 (.06) 1.54 (.09)(OpenType)"?>*** 1.37 (.10)(OpenType)"?>*** 1.57 (.11)(OpenType)"?>*** 1.49 (.11)(OpenType)"?>*** High school, GED, less 1.42 (.08)(OpenType)"?>*** 0.93 (.05) 1.17 (.06)(OpenType)"?>** 1.17 (.09)(OpenType)"?>* 1.43 (.10)(OpenType)"?>*** 1.45 (.10)(OpenType)"?>*** Unemployed 0.69 (.05)(OpenType)"?>*** 0.79 (.05)(OpenType)"?>*** 0.67 (.04)(OpenType)"?>*** 1.23 (.11)(OpenType)"?>** 0.76 (.06)(OpenType)"?>** 0.62 (.05)(OpenType)"?>*** Retired 3.31 (.47)(OpenType)"?>*** 1.96 (.26)(OpenType)"?>*** 4.92 (.67)(OpenType)"?>*** 1.10 (.19) 3.44 (.56)(OpenType)"?>*** 3.84 (.63)(OpenType)"?>*** Quartile 2 1.49 (.11)(OpenType)"?>*** 1.25 (.09)(OpenType)"?>** 1.27 (.09)(OpenType)"?>** 0.91 (.09) 1.24 (.11)(OpenType)"?>* 1.71 (.15)(OpenType)"?>*** Quartile 3 2.69 (.21)(OpenType)"?>*** 2.12 (.16)(OpenType)"?>*** 2.63 (.20)(OpenType)"?>*** 1.09 (.11) 1.85 (.17)(OpenType)"?>*** 3.00 (.27)(OpenType)"?>*** Quartile 4 4.75 (.42)(OpenType)"?>*** 3.23 (.27)(OpenType)"?>*** 6.24 (.53)(OpenType)"?>*** 1.60 (.17)(OpenType)"?>*** 3.36 (.33)(OpenType)"?>*** 6.33 (.68)(OpenType)"?>*** Household size 0.81 (.02)(OpenType)"?>*** 0.86 (.02)(OpenType)"?>*** 0.85 (.02)(OpenType)"?>*** 0.90 (.03)(OpenType)"?>** 0.80 (.03)(OpenType)"?>*** 0.84 (.03)(OpenType)"?>*** Dependent child 1.08 (.09) 0.95 (.08) 1.03 (.08) 0.85 (.09) 1.04 (.10) 1.12 (.11) Living in metro area 1.11 (.08) 0.84 (.06) 1.00 (.07) 1.07 (.09) 1.24 (.10)(OpenType)"?>** 1.16 (.10) Living in South 1.04 (.06) 0.93 (.06) 0.93 (.05) 1.13 (.09) 0.94 (.07) 0.99 (.08) Bank account user 2.26 (.26)(OpenType)"?>*** 0.98 (.10) 1.37 (.14)(OpenType)"?>** 0.81 (.17) 2.12 (.32)(OpenType)"?>*** 3.33 (.48)(OpenType)"?>*** Model significance LR χ2 (17) = 2,172.71(OpenType)"?>*** LR χ2 (17) = 534.80(OpenType)"?>*** LR χ2 (17) = 1,720.71(OpenType)"?>*** LR χ2 (17) = 348.90(OpenType)"?>*** LR χ2 (17) = 1,214.06(OpenType)"?>*** LR χ2 (17) = 1,703.76(OpenType)"?>*** * p < .05. **p <.01. ***p < .001. Open in new tab Table 2: Multivariate Regressions for Effects of Alternative Financial Services (AFS) Use on Financial Well-Being Indicators . Present Security . Future Security . . Credit Score . Making Ends Meet . Subjective Financial Well-Being . Credit Card Bill Payment . Rainy-Day Fund . Emergency Fund . Variable . OR (SE) . OR (SE) . OR (SE) . OR (SE) . OR (SE) . OR (SE) . AFS use 0.35 (.03)(OpenType)"?>*** 0.66 (.05)(OpenType)"?>*** 0.61 (.04)(OpenType)"?>*** 0.73 (.08)(OpenType)"?>** 0.47 (.05)(OpenType)"?>*** 0.42 (.04)(OpenType)"?>*** Gender 1.01 (.05) 1.13 (.06) 0.98 (.05) 1.19 (.08)(OpenType)"?>** 1.09 (.07) 1.25 (.08)(OpenType)"?>*** Age 0.97 (.01)(OpenType)"?>** 0.97 (.01)(OpenType)"?>** 0.94 (.01)(OpenType)"?>*** 0.89 (.01)(OpenType)"?>*** 0.96 (.10)(OpenType)"?>** 0.96 (.01)(OpenType)"?>*** Age2 1.00 (.00)(OpenType)"?>*** 1.00 (.00) 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** Race 1.55 (.10)(OpenType)"?>*** 0.98 (.06) 0.96 (.06) 1.18 (.09)(OpenType)"?>* 0.97 (.07) 1.27 (.09)(OpenType)"?>** Marital status 1.81 (.11)(OpenType)"?>*** 1.06 (.06) 1.54 (.09)(OpenType)"?>*** 1.37 (.10)(OpenType)"?>*** 1.57 (.11)(OpenType)"?>*** 1.49 (.11)(OpenType)"?>*** High school, GED, less 1.42 (.08)(OpenType)"?>*** 0.93 (.05) 1.17 (.06)(OpenType)"?>** 1.17 (.09)(OpenType)"?>* 1.43 (.10)(OpenType)"?>*** 1.45 (.10)(OpenType)"?>*** Unemployed 0.69 (.05)(OpenType)"?>*** 0.79 (.05)(OpenType)"?>*** 0.67 (.04)(OpenType)"?>*** 1.23 (.11)(OpenType)"?>** 0.76 (.06)(OpenType)"?>** 0.62 (.05)(OpenType)"?>*** Retired 3.31 (.47)(OpenType)"?>*** 1.96 (.26)(OpenType)"?>*** 4.92 (.67)(OpenType)"?>*** 1.10 (.19) 3.44 (.56)(OpenType)"?>*** 3.84 (.63)(OpenType)"?>*** Quartile 2 1.49 (.11)(OpenType)"?>*** 1.25 (.09)(OpenType)"?>** 1.27 (.09)(OpenType)"?>** 0.91 (.09) 1.24 (.11)(OpenType)"?>* 1.71 (.15)(OpenType)"?>*** Quartile 3 2.69 (.21)(OpenType)"?>*** 2.12 (.16)(OpenType)"?>*** 2.63 (.20)(OpenType)"?>*** 1.09 (.11) 1.85 (.17)(OpenType)"?>*** 3.00 (.27)(OpenType)"?>*** Quartile 4 4.75 (.42)(OpenType)"?>*** 3.23 (.27)(OpenType)"?>*** 6.24 (.53)(OpenType)"?>*** 1.60 (.17)(OpenType)"?>*** 3.36 (.33)(OpenType)"?>*** 6.33 (.68)(OpenType)"?>*** Household size 0.81 (.02)(OpenType)"?>*** 0.86 (.02)(OpenType)"?>*** 0.85 (.02)(OpenType)"?>*** 0.90 (.03)(OpenType)"?>** 0.80 (.03)(OpenType)"?>*** 0.84 (.03)(OpenType)"?>*** Dependent child 1.08 (.09) 0.95 (.08) 1.03 (.08) 0.85 (.09) 1.04 (.10) 1.12 (.11) Living in metro area 1.11 (.08) 0.84 (.06) 1.00 (.07) 1.07 (.09) 1.24 (.10)(OpenType)"?>** 1.16 (.10) Living in South 1.04 (.06) 0.93 (.06) 0.93 (.05) 1.13 (.09) 0.94 (.07) 0.99 (.08) Bank account user 2.26 (.26)(OpenType)"?>*** 0.98 (.10) 1.37 (.14)(OpenType)"?>** 0.81 (.17) 2.12 (.32)(OpenType)"?>*** 3.33 (.48)(OpenType)"?>*** Model significance LR χ2 (17) = 2,172.71(OpenType)"?>*** LR χ2 (17) = 534.80(OpenType)"?>*** LR χ2 (17) = 1,720.71(OpenType)"?>*** LR χ2 (17) = 348.90(OpenType)"?>*** LR χ2 (17) = 1,214.06(OpenType)"?>*** LR χ2 (17) = 1,703.76(OpenType)"?>*** . Present Security . Future Security . . Credit Score . Making Ends Meet . Subjective Financial Well-Being . Credit Card Bill Payment . Rainy-Day Fund . Emergency Fund . Variable . OR (SE) . OR (SE) . OR (SE) . OR (SE) . OR (SE) . OR (SE) . AFS use 0.35 (.03)(OpenType)"?>*** 0.66 (.05)(OpenType)"?>*** 0.61 (.04)(OpenType)"?>*** 0.73 (.08)(OpenType)"?>** 0.47 (.05)(OpenType)"?>*** 0.42 (.04)(OpenType)"?>*** Gender 1.01 (.05) 1.13 (.06) 0.98 (.05) 1.19 (.08)(OpenType)"?>** 1.09 (.07) 1.25 (.08)(OpenType)"?>*** Age 0.97 (.01)(OpenType)"?>** 0.97 (.01)(OpenType)"?>** 0.94 (.01)(OpenType)"?>*** 0.89 (.01)(OpenType)"?>*** 0.96 (.10)(OpenType)"?>** 0.96 (.01)(OpenType)"?>*** Age2 1.00 (.00)(OpenType)"?>*** 1.00 (.00) 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** Race 1.55 (.10)(OpenType)"?>*** 0.98 (.06) 0.96 (.06) 1.18 (.09)(OpenType)"?>* 0.97 (.07) 1.27 (.09)(OpenType)"?>** Marital status 1.81 (.11)(OpenType)"?>*** 1.06 (.06) 1.54 (.09)(OpenType)"?>*** 1.37 (.10)(OpenType)"?>*** 1.57 (.11)(OpenType)"?>*** 1.49 (.11)(OpenType)"?>*** High school, GED, less 1.42 (.08)(OpenType)"?>*** 0.93 (.05) 1.17 (.06)(OpenType)"?>** 1.17 (.09)(OpenType)"?>* 1.43 (.10)(OpenType)"?>*** 1.45 (.10)(OpenType)"?>*** Unemployed 0.69 (.05)(OpenType)"?>*** 0.79 (.05)(OpenType)"?>*** 0.67 (.04)(OpenType)"?>*** 1.23 (.11)(OpenType)"?>** 0.76 (.06)(OpenType)"?>** 0.62 (.05)(OpenType)"?>*** Retired 3.31 (.47)(OpenType)"?>*** 1.96 (.26)(OpenType)"?>*** 4.92 (.67)(OpenType)"?>*** 1.10 (.19) 3.44 (.56)(OpenType)"?>*** 3.84 (.63)(OpenType)"?>*** Quartile 2 1.49 (.11)(OpenType)"?>*** 1.25 (.09)(OpenType)"?>** 1.27 (.09)(OpenType)"?>** 0.91 (.09) 1.24 (.11)(OpenType)"?>* 1.71 (.15)(OpenType)"?>*** Quartile 3 2.69 (.21)(OpenType)"?>*** 2.12 (.16)(OpenType)"?>*** 2.63 (.20)(OpenType)"?>*** 1.09 (.11) 1.85 (.17)(OpenType)"?>*** 3.00 (.27)(OpenType)"?>*** Quartile 4 4.75 (.42)(OpenType)"?>*** 3.23 (.27)(OpenType)"?>*** 6.24 (.53)(OpenType)"?>*** 1.60 (.17)(OpenType)"?>*** 3.36 (.33)(OpenType)"?>*** 6.33 (.68)(OpenType)"?>*** Household size 0.81 (.02)(OpenType)"?>*** 0.86 (.02)(OpenType)"?>*** 0.85 (.02)(OpenType)"?>*** 0.90 (.03)(OpenType)"?>** 0.80 (.03)(OpenType)"?>*** 0.84 (.03)(OpenType)"?>*** Dependent child 1.08 (.09) 0.95 (.08) 1.03 (.08) 0.85 (.09) 1.04 (.10) 1.12 (.11) Living in metro area 1.11 (.08) 0.84 (.06) 1.00 (.07) 1.07 (.09) 1.24 (.10)(OpenType)"?>** 1.16 (.10) Living in South 1.04 (.06) 0.93 (.06) 0.93 (.05) 1.13 (.09) 0.94 (.07) 0.99 (.08) Bank account user 2.26 (.26)(OpenType)"?>*** 0.98 (.10) 1.37 (.14)(OpenType)"?>** 0.81 (.17) 2.12 (.32)(OpenType)"?>*** 3.33 (.48)(OpenType)"?>*** Model significance LR χ2 (17) = 2,172.71(OpenType)"?>*** LR χ2 (17) = 534.80(OpenType)"?>*** LR χ2 (17) = 1,720.71(OpenType)"?>*** LR χ2 (17) = 348.90(OpenType)"?>*** LR χ2 (17) = 1,214.06(OpenType)"?>*** LR χ2 (17) = 1,703.76(OpenType)"?>*** * p < .05. **p <.01. ***p < .001. Open in new tab Several control variables were associated with financial well-being indicators in the expected direction. In terms of present security, those within higher household income quartiles, who were married, and had full employment or were retired were consistently more likely to report a satisfactory credit score, make ends meet, report higher levels of subjective financial well-being, and pay credit card bills in full each month. Larger household size was negatively associated with all present security indicators. As for future security, individuals who were married, had more than high school education, were employed or retired, or had a high household income were consistently more likely to have a rainy-day fund to cover three months of expenses or to have a $400 emergency fund. Living in the South and having dependent children were insignificant in all models. Of all financial well-being indicators, being female only predicted a higher likelihood of paying credit card bills on time (OR = 1.19, z = 2.74) and having a $400 emergency fund (OR = 1.24, z = 3.48). Interaction of Household Income and AFS Use Descriptive data revealed similar patterns in the proportion of AFS users exhibiting each financial well-being indicator within each income group (table available on request). Over half of all income groups always paid their credit card balances in full monthly, with the top income quartile containing the largest percentage (61.87%). Substantially more of the highest income group had a three-month rainy-day fund (64.80%) than the lowest income quartile group (27.94%). This pattern was the same for having a $400 emergency fund, with 84.44% of the highest income group and 40.00% of the lowest income group having a fund. Previous studies found that AFS use was prevalent among LMI families (Elliehausen, 2005; Elliehausen & Lawrence, 2001; Weller & Logan, 2009); however, little is known about whether or not AFS use evenly affects families with different income levels. In Table 3, interaction models depict how the relationship between AFS and household financial well-being varied by income. These models contained the same set of independent variables as those in Table 2 with the addition of interaction terms for AFS use and income group. Table 3: Results of Multivariate Regressions Predicting Financial Well-Being among Samples by Household Income Quartile . Present Security . Future Security . . Credit Score . Make Ends Meet . Subjective Financial Well-Being . Credit Card Bill Payment . Rainy-Day Fund . Emergency Fund  . Variable . OR (SE) . OR (SE) . OR (SE) . OR (SE) . OR (SE) . OR (SE) . AFS use 0.44 (.05)(OpenType)"?>*** 0.74 (.08)(OpenType)"?>*** 0.85 (.09) 0.92 (.16) 0.60 (.09)(OpenType)"?>** 0.62 (.08)(OpenType)"?>*** Base group: Income 1  Quartile 2 1.54 (.13)(OpenType)"?>*** 1.26 (.10)(OpenType)"?>** 1.38 (.11)(OpenType)"?>*** 0.95 (.10) 1.33 (.13)(OpenType)"?>** 1.56 (.14)(OpenType)"?>***  Quartile 3 2.99 (.26)(OpenType)"?>*** 2.22 (.19)(OpenType)"?>*** 3.06 (.25)(OpenType)"?>*** 1.16 (.13) 1.95 (.19)(OpenType)"?>*** 3.00 (.30)(OpenType)"?>***  Quartile 4 5.13 (.48)(OpenType)"?>*** 3.40 (.30)(OpenType)"?>*** 7.09 (.64)(OpenType)"?>*** 1.69 (.19)(OpenType)"?>*** 3.54 (.37)(OpenType)"?>*** 5.40 (.61)(OpenType)"?>*** Base group: AFS user × income 1  AFS × quartile 2 0.90 (.16) 1.00 (.17) 0.73 (.13) 0.81 (.21) 0.64 (.16) 0.95 (.20)  AFS × quartile 3 0.52 (.10)(OpenType)"?>** 0.76 (.14) 0.43 (.08)(OpenType)"?>*** 0.64 (.18) 0.71 (.18) 0.72 (.17)  AFS × quartile 4 0.62 (.16) 0.57 (.15)(OpenType)"?>* 0.41 (.10)(OpenType)"?>*** 0.62 (.21) 0.70 (.21) 0.42 (.12)(OpenType)"?>**  Gender 1.00 (.05)(OpenType)"?>* 1.12 (.06)(OpenType)"?>* 0.98 (.05) 1.18 (.08)(OpenType)"?>** 1.09 (.07) 1.08 (.07)  Age 0.97 (.00) 0.97 (.01)(OpenType)"?>** 0.93 (.01)(OpenType)"?>*** 0.88 (.01)(OpenType)"?>*** 0.96 (.01)(OpenType)"?>*** 0.96 (.01)(OpenType)"?>***  Age2 1.00 (.00) 1.00 (.00)(OpenType)"?>* 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>***  Race 1.56 (.10)(OpenType)"?>*** 0.98 (.06) 0.97 (.06) 1.19 (.09)(OpenType)"?>* 0.97 (.07) 1.27 (.09)(OpenType)"?>**  Marital status 1.79 (.11)(OpenType)"?>*** 1.05 (.06) 1.53 (.09)(OpenType)"?>*** 1.36 (.10)(OpenType)"?>*** 1.56 (.11)(OpenType)"?>*** 1.47 (.11)(OpenType)"?>***  High school, GED, or less 1.41 (.08)(OpenType)"?>*** 0.93 (.05) 1.16 (.06)(OpenType)"?>** 1.17 (.09)(OpenType)"?>* 1.43 (.10)(OpenType)"?>*** 1.47 (.10)(OpenType)"?>***  Unemployed 0.69 (.05)(OpenType)"?>*** 0.78 (.05)(OpenType)"?>*** 0.67 (.04)(OpenType)"?>*** 1.23 (.11)(OpenType)"?>* 0.76 (.06)(OpenType)"?>*** 0.65 (.05)(OpenType)"?>***  Retired 3.33 (.46)(OpenType)"?>*** 1.97 (.27)(OpenType)"?>*** 4.97 (.67)(OpenType)"?>*** 1.11 (.20) 3.44 (.56)(OpenType)"?>*** 3.56 (.58)(OpenType)"?>***  Household size 0.81 (.02)(OpenType)"?>*** 0.87 (.02)(OpenType)"?>*** 0.85 (.02)(OpenType)"?>*** 0.90 (.03)(OpenType)"?>** 0.80 (.03)(OpenType)"?>*** 0.82 (.03)(OpenType)"?>***  Dependent child 1.07 (.09) 0.95 (.08) 1.02 (.08) 0.84 (.09) 1.04 (.10) 1.15 (.12)  Living in metro area 1.11 (.08) 0.84 (.06)(OpenType)"?>* 1.00 (.07) 1.07 (.09) 1.25 (.10)(OpenType)"?>** 1.10 (.09)  Living in South 1.03 (.06) 0.92 (.06) 0.93 (.06) 1.13 (.09) 0.94 (.07) 1.03 (.08)  Bank account user 2.30 (.27)(OpenType)"?>*** 1.00 (.10) 1.38 (.14)(OpenType)"?>** 0.84 (.18) 2.13 (.33)(OpenType)"?>*** 4.09 (.57)(OpenType)"?>***  Model significance LR χ2 (20) = 2,185.52(OpenType)"?>*** LR χ2 (20) = 541.35(OpenType)"?>*** LR χ2 (20) = 1,746.67(OpenType)"?>*** LR χ2 (20) = 352.33(OpenType)"?>*** LR χ2 (20) = 1,218.21(OpenType)"?>*** LR χ2 (20) = 1,476.27(OpenType)"?>*** . Present Security . Future Security . . Credit Score . Make Ends Meet . Subjective Financial Well-Being . Credit Card Bill Payment . Rainy-Day Fund . Emergency Fund  . Variable . OR (SE) . OR (SE) . OR (SE) . OR (SE) . OR (SE) . OR (SE) . AFS use 0.44 (.05)(OpenType)"?>*** 0.74 (.08)(OpenType)"?>*** 0.85 (.09) 0.92 (.16) 0.60 (.09)(OpenType)"?>** 0.62 (.08)(OpenType)"?>*** Base group: Income 1  Quartile 2 1.54 (.13)(OpenType)"?>*** 1.26 (.10)(OpenType)"?>** 1.38 (.11)(OpenType)"?>*** 0.95 (.10) 1.33 (.13)(OpenType)"?>** 1.56 (.14)(OpenType)"?>***  Quartile 3 2.99 (.26)(OpenType)"?>*** 2.22 (.19)(OpenType)"?>*** 3.06 (.25)(OpenType)"?>*** 1.16 (.13) 1.95 (.19)(OpenType)"?>*** 3.00 (.30)(OpenType)"?>***  Quartile 4 5.13 (.48)(OpenType)"?>*** 3.40 (.30)(OpenType)"?>*** 7.09 (.64)(OpenType)"?>*** 1.69 (.19)(OpenType)"?>*** 3.54 (.37)(OpenType)"?>*** 5.40 (.61)(OpenType)"?>*** Base group: AFS user × income 1  AFS × quartile 2 0.90 (.16) 1.00 (.17) 0.73 (.13) 0.81 (.21) 0.64 (.16) 0.95 (.20)  AFS × quartile 3 0.52 (.10)(OpenType)"?>** 0.76 (.14) 0.43 (.08)(OpenType)"?>*** 0.64 (.18) 0.71 (.18) 0.72 (.17)  AFS × quartile 4 0.62 (.16) 0.57 (.15)(OpenType)"?>* 0.41 (.10)(OpenType)"?>*** 0.62 (.21) 0.70 (.21) 0.42 (.12)(OpenType)"?>**  Gender 1.00 (.05)(OpenType)"?>* 1.12 (.06)(OpenType)"?>* 0.98 (.05) 1.18 (.08)(OpenType)"?>** 1.09 (.07) 1.08 (.07)  Age 0.97 (.00) 0.97 (.01)(OpenType)"?>** 0.93 (.01)(OpenType)"?>*** 0.88 (.01)(OpenType)"?>*** 0.96 (.01)(OpenType)"?>*** 0.96 (.01)(OpenType)"?>***  Age2 1.00 (.00) 1.00 (.00)(OpenType)"?>* 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>***  Race 1.56 (.10)(OpenType)"?>*** 0.98 (.06) 0.97 (.06) 1.19 (.09)(OpenType)"?>* 0.97 (.07) 1.27 (.09)(OpenType)"?>**  Marital status 1.79 (.11)(OpenType)"?>*** 1.05 (.06) 1.53 (.09)(OpenType)"?>*** 1.36 (.10)(OpenType)"?>*** 1.56 (.11)(OpenType)"?>*** 1.47 (.11)(OpenType)"?>***  High school, GED, or less 1.41 (.08)(OpenType)"?>*** 0.93 (.05) 1.16 (.06)(OpenType)"?>** 1.17 (.09)(OpenType)"?>* 1.43 (.10)(OpenType)"?>*** 1.47 (.10)(OpenType)"?>***  Unemployed 0.69 (.05)(OpenType)"?>*** 0.78 (.05)(OpenType)"?>*** 0.67 (.04)(OpenType)"?>*** 1.23 (.11)(OpenType)"?>* 0.76 (.06)(OpenType)"?>*** 0.65 (.05)(OpenType)"?>***  Retired 3.33 (.46)(OpenType)"?>*** 1.97 (.27)(OpenType)"?>*** 4.97 (.67)(OpenType)"?>*** 1.11 (.20) 3.44 (.56)(OpenType)"?>*** 3.56 (.58)(OpenType)"?>***  Household size 0.81 (.02)(OpenType)"?>*** 0.87 (.02)(OpenType)"?>*** 0.85 (.02)(OpenType)"?>*** 0.90 (.03)(OpenType)"?>** 0.80 (.03)(OpenType)"?>*** 0.82 (.03)(OpenType)"?>***  Dependent child 1.07 (.09) 0.95 (.08) 1.02 (.08) 0.84 (.09) 1.04 (.10) 1.15 (.12)  Living in metro area 1.11 (.08) 0.84 (.06)(OpenType)"?>* 1.00 (.07) 1.07 (.09) 1.25 (.10)(OpenType)"?>** 1.10 (.09)  Living in South 1.03 (.06) 0.92 (.06) 0.93 (.06) 1.13 (.09) 0.94 (.07) 1.03 (.08)  Bank account user 2.30 (.27)(OpenType)"?>*** 1.00 (.10) 1.38 (.14)(OpenType)"?>** 0.84 (.18) 2.13 (.33)(OpenType)"?>*** 4.09 (.57)(OpenType)"?>***  Model significance LR χ2 (20) = 2,185.52(OpenType)"?>*** LR χ2 (20) = 541.35(OpenType)"?>*** LR χ2 (20) = 1,746.67(OpenType)"?>*** LR χ2 (20) = 352.33(OpenType)"?>*** LR χ2 (20) = 1,218.21(OpenType)"?>*** LR χ2 (20) = 1,476.27(OpenType)"?>*** Notes: AFS = alternative financial services. LR = likelihood ratio. * p < .05. **p < .01. ***p < .001. Open in new tab Table 3: Results of Multivariate Regressions Predicting Financial Well-Being among Samples by Household Income Quartile . Present Security . Future Security . . Credit Score . Make Ends Meet . Subjective Financial Well-Being . Credit Card Bill Payment . Rainy-Day Fund . Emergency Fund  . Variable . OR (SE) . OR (SE) . OR (SE) . OR (SE) . OR (SE) . OR (SE) . AFS use 0.44 (.05)(OpenType)"?>*** 0.74 (.08)(OpenType)"?>*** 0.85 (.09) 0.92 (.16) 0.60 (.09)(OpenType)"?>** 0.62 (.08)(OpenType)"?>*** Base group: Income 1  Quartile 2 1.54 (.13)(OpenType)"?>*** 1.26 (.10)(OpenType)"?>** 1.38 (.11)(OpenType)"?>*** 0.95 (.10) 1.33 (.13)(OpenType)"?>** 1.56 (.14)(OpenType)"?>***  Quartile 3 2.99 (.26)(OpenType)"?>*** 2.22 (.19)(OpenType)"?>*** 3.06 (.25)(OpenType)"?>*** 1.16 (.13) 1.95 (.19)(OpenType)"?>*** 3.00 (.30)(OpenType)"?>***  Quartile 4 5.13 (.48)(OpenType)"?>*** 3.40 (.30)(OpenType)"?>*** 7.09 (.64)(OpenType)"?>*** 1.69 (.19)(OpenType)"?>*** 3.54 (.37)(OpenType)"?>*** 5.40 (.61)(OpenType)"?>*** Base group: AFS user × income 1  AFS × quartile 2 0.90 (.16) 1.00 (.17) 0.73 (.13) 0.81 (.21) 0.64 (.16) 0.95 (.20)  AFS × quartile 3 0.52 (.10)(OpenType)"?>** 0.76 (.14) 0.43 (.08)(OpenType)"?>*** 0.64 (.18) 0.71 (.18) 0.72 (.17)  AFS × quartile 4 0.62 (.16) 0.57 (.15)(OpenType)"?>* 0.41 (.10)(OpenType)"?>*** 0.62 (.21) 0.70 (.21) 0.42 (.12)(OpenType)"?>**  Gender 1.00 (.05)(OpenType)"?>* 1.12 (.06)(OpenType)"?>* 0.98 (.05) 1.18 (.08)(OpenType)"?>** 1.09 (.07) 1.08 (.07)  Age 0.97 (.00) 0.97 (.01)(OpenType)"?>** 0.93 (.01)(OpenType)"?>*** 0.88 (.01)(OpenType)"?>*** 0.96 (.01)(OpenType)"?>*** 0.96 (.01)(OpenType)"?>***  Age2 1.00 (.00) 1.00 (.00)(OpenType)"?>* 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>***  Race 1.56 (.10)(OpenType)"?>*** 0.98 (.06) 0.97 (.06) 1.19 (.09)(OpenType)"?>* 0.97 (.07) 1.27 (.09)(OpenType)"?>**  Marital status 1.79 (.11)(OpenType)"?>*** 1.05 (.06) 1.53 (.09)(OpenType)"?>*** 1.36 (.10)(OpenType)"?>*** 1.56 (.11)(OpenType)"?>*** 1.47 (.11)(OpenType)"?>***  High school, GED, or less 1.41 (.08)(OpenType)"?>*** 0.93 (.05) 1.16 (.06)(OpenType)"?>** 1.17 (.09)(OpenType)"?>* 1.43 (.10)(OpenType)"?>*** 1.47 (.10)(OpenType)"?>***  Unemployed 0.69 (.05)(OpenType)"?>*** 0.78 (.05)(OpenType)"?>*** 0.67 (.04)(OpenType)"?>*** 1.23 (.11)(OpenType)"?>* 0.76 (.06)(OpenType)"?>*** 0.65 (.05)(OpenType)"?>***  Retired 3.33 (.46)(OpenType)"?>*** 1.97 (.27)(OpenType)"?>*** 4.97 (.67)(OpenType)"?>*** 1.11 (.20) 3.44 (.56)(OpenType)"?>*** 3.56 (.58)(OpenType)"?>***  Household size 0.81 (.02)(OpenType)"?>*** 0.87 (.02)(OpenType)"?>*** 0.85 (.02)(OpenType)"?>*** 0.90 (.03)(OpenType)"?>** 0.80 (.03)(OpenType)"?>*** 0.82 (.03)(OpenType)"?>***  Dependent child 1.07 (.09) 0.95 (.08) 1.02 (.08) 0.84 (.09) 1.04 (.10) 1.15 (.12)  Living in metro area 1.11 (.08) 0.84 (.06)(OpenType)"?>* 1.00 (.07) 1.07 (.09) 1.25 (.10)(OpenType)"?>** 1.10 (.09)  Living in South 1.03 (.06) 0.92 (.06) 0.93 (.06) 1.13 (.09) 0.94 (.07) 1.03 (.08)  Bank account user 2.30 (.27)(OpenType)"?>*** 1.00 (.10) 1.38 (.14)(OpenType)"?>** 0.84 (.18) 2.13 (.33)(OpenType)"?>*** 4.09 (.57)(OpenType)"?>***  Model significance LR χ2 (20) = 2,185.52(OpenType)"?>*** LR χ2 (20) = 541.35(OpenType)"?>*** LR χ2 (20) = 1,746.67(OpenType)"?>*** LR χ2 (20) = 352.33(OpenType)"?>*** LR χ2 (20) = 1,218.21(OpenType)"?>*** LR χ2 (20) = 1,476.27(OpenType)"?>*** . Present Security . Future Security . . Credit Score . Make Ends Meet . Subjective Financial Well-Being . Credit Card Bill Payment . Rainy-Day Fund . Emergency Fund  . Variable . OR (SE) . OR (SE) . OR (SE) . OR (SE) . OR (SE) . OR (SE) . AFS use 0.44 (.05)(OpenType)"?>*** 0.74 (.08)(OpenType)"?>*** 0.85 (.09) 0.92 (.16) 0.60 (.09)(OpenType)"?>** 0.62 (.08)(OpenType)"?>*** Base group: Income 1  Quartile 2 1.54 (.13)(OpenType)"?>*** 1.26 (.10)(OpenType)"?>** 1.38 (.11)(OpenType)"?>*** 0.95 (.10) 1.33 (.13)(OpenType)"?>** 1.56 (.14)(OpenType)"?>***  Quartile 3 2.99 (.26)(OpenType)"?>*** 2.22 (.19)(OpenType)"?>*** 3.06 (.25)(OpenType)"?>*** 1.16 (.13) 1.95 (.19)(OpenType)"?>*** 3.00 (.30)(OpenType)"?>***  Quartile 4 5.13 (.48)(OpenType)"?>*** 3.40 (.30)(OpenType)"?>*** 7.09 (.64)(OpenType)"?>*** 1.69 (.19)(OpenType)"?>*** 3.54 (.37)(OpenType)"?>*** 5.40 (.61)(OpenType)"?>*** Base group: AFS user × income 1  AFS × quartile 2 0.90 (.16) 1.00 (.17) 0.73 (.13) 0.81 (.21) 0.64 (.16) 0.95 (.20)  AFS × quartile 3 0.52 (.10)(OpenType)"?>** 0.76 (.14) 0.43 (.08)(OpenType)"?>*** 0.64 (.18) 0.71 (.18) 0.72 (.17)  AFS × quartile 4 0.62 (.16) 0.57 (.15)(OpenType)"?>* 0.41 (.10)(OpenType)"?>*** 0.62 (.21) 0.70 (.21) 0.42 (.12)(OpenType)"?>**  Gender 1.00 (.05)(OpenType)"?>* 1.12 (.06)(OpenType)"?>* 0.98 (.05) 1.18 (.08)(OpenType)"?>** 1.09 (.07) 1.08 (.07)  Age 0.97 (.00) 0.97 (.01)(OpenType)"?>** 0.93 (.01)(OpenType)"?>*** 0.88 (.01)(OpenType)"?>*** 0.96 (.01)(OpenType)"?>*** 0.96 (.01)(OpenType)"?>***  Age2 1.00 (.00) 1.00 (.00)(OpenType)"?>* 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>*** 1.00 (.00)(OpenType)"?>***  Race 1.56 (.10)(OpenType)"?>*** 0.98 (.06) 0.97 (.06) 1.19 (.09)(OpenType)"?>* 0.97 (.07) 1.27 (.09)(OpenType)"?>**  Marital status 1.79 (.11)(OpenType)"?>*** 1.05 (.06) 1.53 (.09)(OpenType)"?>*** 1.36 (.10)(OpenType)"?>*** 1.56 (.11)(OpenType)"?>*** 1.47 (.11)(OpenType)"?>***  High school, GED, or less 1.41 (.08)(OpenType)"?>*** 0.93 (.05) 1.16 (.06)(OpenType)"?>** 1.17 (.09)(OpenType)"?>* 1.43 (.10)(OpenType)"?>*** 1.47 (.10)(OpenType)"?>***  Unemployed 0.69 (.05)(OpenType)"?>*** 0.78 (.05)(OpenType)"?>*** 0.67 (.04)(OpenType)"?>*** 1.23 (.11)(OpenType)"?>* 0.76 (.06)(OpenType)"?>*** 0.65 (.05)(OpenType)"?>***  Retired 3.33 (.46)(OpenType)"?>*** 1.97 (.27)(OpenType)"?>*** 4.97 (.67)(OpenType)"?>*** 1.11 (.20) 3.44 (.56)(OpenType)"?>*** 3.56 (.58)(OpenType)"?>***  Household size 0.81 (.02)(OpenType)"?>*** 0.87 (.02)(OpenType)"?>*** 0.85 (.02)(OpenType)"?>*** 0.90 (.03)(OpenType)"?>** 0.80 (.03)(OpenType)"?>*** 0.82 (.03)(OpenType)"?>***  Dependent child 1.07 (.09) 0.95 (.08) 1.02 (.08) 0.84 (.09) 1.04 (.10) 1.15 (.12)  Living in metro area 1.11 (.08) 0.84 (.06)(OpenType)"?>* 1.00 (.07) 1.07 (.09) 1.25 (.10)(OpenType)"?>** 1.10 (.09)  Living in South 1.03 (.06) 0.92 (.06) 0.93 (.06) 1.13 (.09) 0.94 (.07) 1.03 (.08)  Bank account user 2.30 (.27)(OpenType)"?>*** 1.00 (.10) 1.38 (.14)(OpenType)"?>** 0.84 (.18) 2.13 (.33)(OpenType)"?>*** 4.09 (.57)(OpenType)"?>***  Model significance LR χ2 (20) = 2,185.52(OpenType)"?>*** LR χ2 (20) = 541.35(OpenType)"?>*** LR χ2 (20) = 1,746.67(OpenType)"?>*** LR χ2 (20) = 352.33(OpenType)"?>*** LR χ2 (20) = 1,218.21(OpenType)"?>*** LR χ2 (20) = 1,476.27(OpenType)"?>*** Notes: AFS = alternative financial services. LR = likelihood ratio. * p < .05. **p < .01. ***p < .001. Open in new tab Compared with individuals in the lowest income quartile, individuals in all other income groups reported higher credit scores. The interaction terms depicted the statistical effect of AFS use on credit scores across income groups. These relationships were complex: the second and fourth quartile revealed statistically similar negative relationships between AFS and credit score as the lowest income group (OR = 0.44). However, the negative effect of AFS on credit score was significantly smaller for the third income group (OR = 0.44 × 0.52 = 0.23). The “making ends meet” interaction model revealed that AFS users were less likely to report having enough resources to cover expenses than non-AFS users, and individuals in higher income groups were still more likely to make ends meet. The interaction terms revealed that the negative effect of AFS use on the ability to make ends meet was the same for those in the second and third income group as it was for the lowest income group (OR = 0.74). However, the negative effect of AFS use on the odds of making ends meet was significantly weaker for those in the highest income group (OR = 0.74 × 0.57 = 0.42). In the subjective financial well-being interaction model, higher incomes continued to predict higher levels of subjective financial well-being, but AFS use among the first and second income quartile was no longer a significant predictor. However, compared with those in the lowest income group, AFS use was associated with lower levels of subjective financial well-being for those in the third (OR = 0.85 × 0.43 = 0.37) and fourth income groups (OR = 0.85 × 0.41 = 0.35). The credit card bill payment interaction model reflected no significant effect of AFS use overall or for any of the income categories. Only the highest income category was significantly more likely to pay their credit card bill on time than the lowest income category (OR = 1.69). The rainy-day fund interaction model indicated that those using AFS were less likely to have enough money set aside to cover three months of expenses than those not using AFS. In addition, individuals in the top three income groups were more likely to have a rainy-day fund than those in the lowest income group. However, interaction terms indicated that the negative effect of AFS was the same for income groups 2, 3, and 4 as it was for group 1. The final $400 rainy-day fund interaction model showed that those using AFS were less likely to have a $400 rainy-day fund than those not using AFS, and individuals in all other income categories were more likely to have a $400 rainy-day fund than those in the lowest income group. AFS use among those in the top income quartile decreased the likelihood that they had a $400 rainy-day fund less so than AFS use among the bottom income group (OR = 0.62 × 0.42 = 0.26). Discussion This study adds to the body of knowledge on the relationship between AFS and financial well-being. Findings demonstrated that AFS use consistently predicted negative financial well-being when controlling for other characteristics. Specifically, AFS predicted reduced levels of well-being across all multivariate models while controlling for demographic characteristics, income, location, and bank account ownership. AFS users had lower credit scores, less ability to make ends meet, lower subjective financial well-being levels, a lower likelihood of paying off credit card bills monthly, and either large or small emergency fund balances. With the mixed findings from previous studies (see, for example, Bhutta, 2014; Karlan & Zinman, 2010; Morgan, 2007; Morse, 2011; Skiba & Tobacman, 2011), this is the first known study linking AFS use and financial well-being with a nationally representative data set; the negative associations indicate that the effect of AFS use could be detrimental to families’ present and future financial security. Interaction models revealed that the relationship between AFS use and financial well-being varied by household income. The negative association between AFS use and three measures of financial well-being disproportionately fell on lower income groups. Compared with the lowest income group, the strength of negative relationship between AFS and credit score was weaker for the third highest income group. Similarly, the negative association between AFS and making ends meet and having a $400 emergency fund was weaker among those in the highest income group, possibly due to the ability of this income group to save over time and that their AFS use may be more a matter of convenience than of  last resort. These results indicate that AFS use in the fourth income group may carry less of a penalty regarding perceived credit score, ability to make ends meet, and having an emergency fund. Whereas prior research has demonstrated a relationship between income and making ends meet, no research has assessed the role of AFS in this relationship. This is the first study to test this relationship empirically. Since the literature has identified LMI individuals as the primary users of AFS (Elliehausen, 2005; Elliehausen & Lawrence, 2001; Weller & Logan, 2009), little attention has been paid to higher income groups. Although common knowledge suggests that AFS are less important in predicting the ability to make ends meet in highest income groups compared with the lowest, the varied findings across income groups and financial well-being indicators were less intuitive. The role of AFS in predicting whether or not individuals were able to make ends meet was the same for those in the second and third quartiles and the lowest income group. This demonstrates that individuals with above-average incomes who rely on AFS also have trouble making ends meet. However, for subjective financial well-being, individuals in the third income quartile display similar patterns as those in the highest income group. Compared with other income groups, higher income groups seem to fare worse on subjective financial well-being when they use AFS. Although no significant relationship between AFS use and subjective financial well-being was found among the lowest two income groups, AFS use among the third and fourth quartiles was associated with a reduced likelihood of having higher levels of subjective financial well-being. Because this measure reflected respondent perceptions, perceived norms may have differed, with AFS use itself perceived as an indicator of financial difficulty among the higher income groups, but not among the lower income groups. Future research should examine these speculations. For the likelihoods of paying credit card bills in full and having a large rainy-day fund, their negative associations between AFS use did not vary by income level. Although AFS use predicted a reduction in the likelihood of having a three-month rainy-day fund across income groups, its relation to credit card payment did not vary by income in the interaction models. This study has several limitations. As with all correlational research, a causal relationship between AFS use and financial well-being was not established. It is possible that those in each income quartile with the worst financial well-being were most likely to use AFS, rather than AFS causing this negative effect. Measurement issues may have also biased findings. The AFS use measure used in this study was binary and did not capture use of each AFS product separately. By treating all AFS products the same, this measure may have obscured heterogeneous effects. In addition, financial well-being was examined using only two of the four central elements (CFPB, 2015), therefore it was subject to weak content validity. Another related issue is missing covariates such as state policy, which is an important factor in determining AFS use. However, we find it neither viable nor meaningful to code and analyze state regulations on six different forms of AFS use in the current study. Future research should consider state policy when focusing on individual AFS use. Another limitation was the generalizability of the findings. Because researchers collected the data online, they may have excluded individuals without Internet access from the sample, although they did recruit respondents by mail and provide computers and Internet access to those who requested technological assistance. Despite using weights to make the sample population match the U.S. population, noncoverage and nonresponse issues might have resulted in unmeasured differences between the sample population and the U.S. population that are not corrected by using weights. However, the comparison of the SHED data to census data and other national data sets conducted by Larrimore, Schmeiser, and Devlin-Foltz (2015) provides reasonable support for the reliability of the estimates of this study. Even with these limitations, this research extended prior work in two important ways. It demonstrated that AFS use was associated with lower levels of financial well-being using multiple measures of present and future financial security. The multidimensionality of the financial well-being measures used by this study allows for a rather comprehensive examination of household financial conditions and its relation to AFS use (CFPB, 2015). This study also contributes to the literature by introducing variations in income as a factor, deepening our understanding of this relationship and interventions required to improve financial well-being. The negative associations between AFS use and financial well-being found here suggest that social workers should advocate for policies that monitor and regulate the availability of AFS and limit any negative effects on financial well-being across all income groups. Particularly, this work provides further support for implementing the new CFPB rules that require lenders to assess the ability of consumers to repay loans (CFPB, 2017). Maintaining this rule is consistent with the reasonable policies and practices for short-term, small-dollar loans promoted by the Office of the Comptroller of the Currency (2018), which identifies the use of data to determine credit-worthiness as a reasonable practice. However, the most recent rule published by the CFPB proposes to rescind the 2017 rule and is currently seeking public comment (CFPB, 2019). Rather than continuing to make policy decisions based on ideology and the power of interest groups, a more prudent response to this problem would be to study the effect of policy changes on financial well-being and to adopt policies that are in the best interest of consumers. The 2017 CFPB rules scheduled to go into effect in August 2019 would provide a good opportunity to study the effect of a policy environment that protects consumer interests. Rescinding these rules before implementation would impede this opportunity. The results also suggest that practitioners working with individuals and families should recognize AFS use as an indicator of financial difficulties for all income groups. Individuals across all income groups in this study have experienced low levels of financial well-being and AFS use correlated with this. Whether AFS use is the cause or result of compromised financial well-being, understanding the role and dangers of these services is important. Practitioners can share information to help them develop financial capabilities. The CFPB provides a plethora of free resources that can help practitioners and consumers navigate the AFS market (CFPB, n.d.). It is important to note the exploratory nature of the current correlational study and its findings on relationship between AFS use and examined financial well-being indicators. To deepen the understanding of AFS use and its relation to financial well-being, future research should track changes in financial well-being measures among AFS users over time. Future research should also examine how variation in type of AFS product and frequency of use affects financial well-being. Researchers should examine the role of different types of AFS products in predicting financial well-being and seek to determine the direction of the relationships. In addition, qualitative research should explore differences in perceptions of and rationales for using AFS across income groups. References Austin R. ( 2004 ). Of predatory lending and the democratization of credit: Preserving the social safety net of  informality in small-loan transactions critical race theory: History, evolution, and new frontiers . American University Law Review, 53 , 1217 – 1258 . Google Scholar OpenURL Placeholder Text WorldCat Barr M. S. ( 2002 ). Access to financial services in the 21st century: Five opportunities for the Bush administration and the 107th congress symposium on poverty and the law . Notre Dame Journal of Law, Ethics & Public Policy, 16 , 447 – 474 . Google Scholar OpenURL Placeholder Text WorldCat Bhutta N. ( 2014 ). Payday loans and consumer financial health . Journal of Banking & Finance, 47 , 230 – 242 . doi:10.1016/j.jbankfin.2014.04.024 Google Scholar Crossref Search ADS WorldCat Bhutta N. , Skiba P. M. , Tobacman J. ( 2015 ). Payday loan choices and consequences . Journal of Money, Credit and Banking, 47 , 223 – 260 . doi:10.1111/jmcb.12175 Google Scholar Crossref Search ADS WorldCat Burhouse S. , Chu K. , Goodstein R. M. , Northwood J. M. , Osaki Y. , Sharma D. ( 2014 ). 2013 FDIC national survey of unbanked and underbanked households. Washington , DC : Federal Deposit Insurance Corporation, Division of Depositor and Consumer Protection. Retrieved from https://www.fdic.gov/householdsurvey/2013report.pdf Campbell D. , Martínez-Jerez F. A. , Tufano P. ( 2012 ). Bouncing out of the banking system: An empirical analysis of involuntary bank account closures . Journal of Banking & Finance, 36 , 1224 – 1235 . doi:10.1016/j.jbankfin.2011.11.014 Google Scholar Crossref Search ADS WorldCat Carrell S. , Zinman J. ( 2014 ). In harm’s way? Payday loan access and military personnel performance . Review of Financial Studies, 27 , 2805 – 2840 . doi:10.1093/rfs/hhu034 Google Scholar Crossref Search ADS WorldCat Caskey J. ( 1994 ). Fringe banking: Check-cashing outlets, pawnshops, and the poor . New York : Russell Sage Foundation . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Caskey J. ( 1997 ). Lower income Americans, higher cost financial services . Madison, WI : Filene Research Institute Center for Credit Union Research . Retrieved from https://filene.org/assets/pdf-reports/1752-22LowerIncomeAmericans.pdf Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Caskey J. ( 2006 ). Fringe banking and the rise of payday lending. In Bolton P. , Rosenthal H. (Eds.), Credit markets for the poor (pp. 17 – 45) . New York : Russell Sage Foundation. Retrieved from http://works.swarthmore.edu/fac-economics/237 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Caskey J. ( 2010 ). Payday lending: New research and the big question (Working Paper No. 1696019). Rochester , NY : Social Science Research Network. Retrieved from https://papers.ssrn.com/abstract=1696019 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Consumer Financial Protection Bureau. ( 2013 ). Payday loans and deposit advance product: A white paper on initial findings. Retrieved from https://files.consumerfinance.gov/f/201304_cfpb_payday-dap-whitepaper.pdf Consumer Financial Protection Bureau. ( 2015 ). Measuring financial well-being: A guide to using the CFPB financial well-being scale . Washington, DC : Author . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Consumer Financial Protection Bureau. (2017, November 17). Payday, Vehicle Title, and Certain High-Cost Installment Loans (to be codified at 12 C.F.R. pt. 1041). Consumer Financial Protection Bureau. (2019, February 14). Payday, Vehicle Title, and Certain High-Cost Installment Loans (to be codified at 12 C.F.R. pt. 1041). Consumer Financial Protection Bureau. (n.d.). We’re the CFPB. Retrieved from https://www.consumerfinance.gov/ Elliehausen G. ( 2005 ). Consumer use of tax refund anticipation loans . Washington, DC : Georgetown University , Credit Research Center. Retrieved from http://faculty.msb.edu/prog/CRC/pdf/M37.pdf Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Elliehausen G. ( 2009 ). An analysis of consumers’ use of payday loans . Washington, DC : Georgetown University , Credit Research Center. Retrieved from https://www.researchgate.net/profile/Gregory_Elliehausen/publication/237554300_AN_ANALYSIS_OF_CONSUMERS’_USE_OF_PAYDAY_LOANS/links/00b7d5362429f9db10000000.pdf Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Elliehausen G. ( 2011 ). An assessment of consumers’ use of high-rate credit products. Unpublished manuscript. Retrieved from http://web.law.columbia.edu/sites/default/files/microsites/transactional-studies/files/12PDL_Elliehausen.High_rate_credit_products_2011-07.pdf Elliehausen G. , Lawrence E. C. ( 2001 ). Payday advance credit in America: An analysis of customer demand . Washington, DC : Georgetown University, Credit Research Center . Retrieved from http://faculty.msb.edu/prog/CRC/pdf/Mono35.PDF Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Federal Deposit Insurance Cooperation. ( 2013 ). Alternative financial services: A primer. Retrieved from https://www.fdic.gov/bank/analytical/quarterly/2009-vol3-1/fdic140-quarterlyvol3no1-afs-final.pdf Gross M. B. , Hogarth J. M. , Manohar A. M. , Gallegos S. ( 2012 ). Who uses alternative financial services, and why? Consumer Interests Annual, 58 , 2012 – 2057 . Google Scholar OpenURL Placeholder Text WorldCat Karlan D. , Zinman J. ( 2010 ). Expanding credit access: Using randomized supply decisions to estimate the impacts . Review of Financial Studies, 23 , 433 – 464 . doi:10.1093/rfs/hhp092 Google Scholar Crossref Search ADS WorldCat Lacko J. M. , McKernan S.-M. , Hastak M. ( 2000 ). Survey of rent-to-own customers. Washington, DC : Federal Trade Commission, Bureau of Economics . Retrieved from http://www.urban.org/research/publication/survey-rent-own-customers Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Larrimore J. , Arthur-Bentil M. , Dodini S. , Thomas S. ( 2015 ). Report on the economic well-being of U.S. households in 2014. Retrieved from https://www.federalreserve.gov/consumerscommunities/shed_publications.htm Larrimore J. , Schmeiser M. S. , Devlin-Foltz S. ( 2015 ). Should you trust things you hear online? Comparing SHED and Census Bureau survey results. Retrieved from https://www.federalreserve.gov/econresdata/notes/feds-notes/2015/comparing-shed-and-census-bureau-survey-results-20151015.html Lim Y. , Bickham T. , Dinecola C. M. , Broussard J. , Weber B. E. , Gregory A. ( 2014 ). Payday loan use and consumer well-being: What consumers and social workers need to know about payday loans . Journal of Poverty, 18 , 379 – 398 . Google Scholar Crossref Search ADS WorldCat Lusardi A. , de Bassa Scheresberg C. ( 2013 ). Financial literacy and high-cost borrowing in the United States (Working Paper No. 18969). Cambridge, MA : National Bureau of Economic Research . doi:10.3386/w18969 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Martin N. , Longa E. A. ( 2012 ). High-interest loans and class: Do payday and title loans really serve the middle class? (Working Paper No. 2145598). Rochester, NY : Social Science Research Network . Retrieved from https://papers.ssrn.com/abstract=2145598 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Mayer R. ( 2004 ). Payday lending and personal bankruptcy . Consumer Interests Annual, 50 , 76 – 82 . Retrieved from http://ecommons.luc.edu/politicalscience_facpubs/24 Google Scholar OpenURL Placeholder Text WorldCat Melzer B. T. ( 2011 ). The real costs of credit access: Evidence from the payday lending market . Quarterly Journal of Economics, 126 , 517 – 555 . Google Scholar Crossref Search ADS WorldCat Morgan D. P. ( 2007 ). Defining and detecting predatory lending (Working Paper No. 273). New York : Federal Reserve Bank of New York . Retrieved from https://www.econstor.eu/handle/10419/60671 Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Morgan D. P. , Strain M. R. ( 2007 ). Payday holiday: How households fare after payday credit bans (Working Paper No. 309). New York : Federal Reserve Bank of New York . Retrieved from https://www.newyorkfed.org/research/staff_reports/sr309.html Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Morse A. ( 2011 ). Payday lenders: Heroes or villains? Journal of Financial Economics, 102 , 28 – 44 . doi:10.1016/j.jfineco.2011.03.022 Google Scholar Crossref Search ADS WorldCat Northwood J. M. , Rhine S.L.W. ( 2018 ). Use of bank and nonbank financial services: Financial decision making by immigrants and native born . Journal of Consumer Affairs, 52 , 317 – 348 . doi:10.1111/joca.12150 Google Scholar Crossref Search ADS WorldCat Office of the Comptroller of the Currency. ( 2018 ). Installment lending: Core lending principles for short-term, small-dollar installment lending. Retrieved from https://www.occ.gov/static/rescinded-bulletins/bulletin-2018-14.pdf Pew Charitable Trusts. ( 2012 ). Payday lending in America: Who borrows, where they borrow, and why. Retrieved from http://www.pewtrusts.org/∼/media/legacy/uploadedfiles/pcs_assets/2012/pewpaydaylendingreportpdf.pdf Prager R. A. ( 2014 ). Determinants of the locations of alternative financial service providers . Review of Industrial Organization, 45 , 21 – 38 . doi:10.1007/s11151-014-9421-4 Google Scholar Crossref Search ADS WorldCat Rhine S.L.W. , Greene W. H. ( 2006 ). The determinants of  being unbanked for U.S. immigrants . Journal of Consumer Affairs, 40 , 21 – 40 . doi:10.1111/j.1745-6606.2006.00044.x Google Scholar Crossref Search ADS WorldCat Shobe M. , Christy K. , Givens A. , Murphy-Erby Y. ( 2013 ). The fringe economy: Predictors of utilization among low- and moderate-income households . International Journal of Arts & Sciences, 6 ( 3 ), 375 – 389 . Google Scholar OpenURL Placeholder Text WorldCat Skiba P. M. , Tobacman J. ( 2011 ). Do payday loans cause bankruptcy? (Working Paper No. 1266215). Rochester, NY : Social Science Research Network . Retrieved from https://papers.ssrn.com/abstract=1266215 Weller C. , Logan A. ( 2009 ). Measuring middle class economic security . Journal of Economic Issues, 43 , 327 – 336 . Google Scholar Crossref Search ADS WorldCat Zinman J. ( 2010 ). Restricting consumer credit access: Household survey evidence on effects around the Oregon rate cap . Journal of Banking and Finance, 34 , 546 – 556 . Google Scholar Crossref Search ADS WorldCat © 2020 National Association of Social Workers This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Margolis, Rachel H, F;Bellin, Melissa, H;Tsoukleris,, Mona;Unick,, Jay;Kub, Joan, E;Butz, Arlene, M

2020 Social Work Research

doi: 10.1093/swr/svaa010

Abstract Asthma morbidity and mortality are disproportionately higher among African American children. Medication adherence is essential for reducing adverse asthma outcomes in this population. Depressive symptoms, prevalent among mothers of children with asthma, have been linked to medication nonadherence. This longitudinal, multilevel analysis used data from a randomized clinical trial evaluating the efficacy of an environmental control educational intervention to evaluate the relationship between caregiver depressive symptoms (Center for Epidemiologic Studies Depression Scale) and caregiver-reported medication adherence (Medication Adherence Report Scale) in urban African American children with uncontrolled asthma (N = 208) at baseline, in six months, and in 12 months. Nearly a third (31.7%) of caregivers (97% female) had clinically significant depressive symptoms at baseline. A random intercept model showed that the within-caregiver effect of depression predicted lower medication adherence (β = –.079, p = .002) as did time (β = –.413, p < .001); the between-caregiver effect of depression did not (β = –.007, p = .77). Changes in a caregiver’s baseline level of depressive symptoms appear to have a stronger influence on medication adherence than mean level of depressive symptoms. Policy, practice, and further research should address maternal mental health as a key element in the life course of African American children with asthma. Asthma affects 8.3% of all youths under the age of 18, making it one of the most common child health conditions in the United States (Centers for Disease Control and Prevention, 2018). The burden of asthma is particularly profound for urban African American children and their families (Gergen & Togias, 2015). Asthma prevalence is significantly higher among African American children compared with non-Hispanic White children (15.7% versus 7.1%); morbidity rates are disproportionately higher as well (Zahran, Bailey, Damon, Garbe, & Breysse, 2018). Nearly a quarter (22.5%) of African American children with asthma received care for asthma at the emergency department (ED) or an urgent care facility, whereas only 12.2% of White children with asthma did (Zahran et al., 2018). In addition, the asthma-related death rate for African American children is over seven times higher than that of non-Hispanic White children (Arroyo, Chee, Camargo, & Wang, 2018). Poor asthma control contributes to increased ED visits and mortality among African American children with asthma (Arroyo et al., 2018; Rand et al., 2012). Medication adherence, primarily daily controller medication usage, is essential for asthma control and for reducing adverse asthma outcomes in this population (Butz et al., 2017; Rohan et al., 2010; Wood et al., 2018). Asthma self-management is a complex process that involves avoiding triggers, assessing for asthma symptoms, and taking controller medications on a daily basis with the goal of well-controlled asthma (U.S. Department of Health and Human Services [HHS], 2007). For young children with asthma, caregivers encounter a high burden of responsibility for disease management (Munzenberger, Secord, & Thomas, 2010). Depression, which disproportionately affects urban mothers of children with asthma, has been linked to medication nonadherence in other diseases, but this relationship has been understudied in children with asthma and their caregivers (Bartlett et al., 2004; Bauman et al., 2002; Celano et al., 2010; Easter, Sharpe, & Hunt, 2015; Grenard et al., 2011; Kaugars, Klinnert, & Bender, 2004; Wood et al., 2018). Bauman et al. (2002) investigated caregiver factors associated with medication nonadherence in a prospective panel study of 1,199 urban, predominately racial minority children with asthma between the ages of four and nine years. Caregiver psychological distress, measured by the Brief Symptom Inventory (BSI) (DeRogatis & Spencer, 1982), was significantly related to both self-report of nonadherence and risk for nonadherence to asthma management recommendations by a provider (for example, did not fill a prescription, gives less medicine than prescribed, did not use mattress cover, did not obtain prescribed peak flow meter). More than half (58%) of caregivers who admitted more than one kind of nonadherence had BSI scores indicating clinically significant psychological symptoms compared with 42% of caregivers who denied nonadherence. Moreover, 53.7% of caregivers in the “high risk for nonadherence” category had BSI scores that indicated clinically significant psychological distress compared with 46.5% at medium risk and 42.4% at low risk (p < .01). Bartlett et al. (2004) examined how maternal depression influences asthma management in a prospective cohort study with 177 mothers of predominately African American (>85%) school-age children with asthma living in two urban areas. Mothers with high depressive symptoms (that is, a score ≥ 9 on an 11-item version of the Center for Epidemiologic Studies Depression Scale [CES-D] [Radloff, 1977]) were more likely to report that their child had difficulty using their inhaler properly (odds ratio [OR]: 5.0, 95% confidence interval [CI]: 1.3, 18.9), forgot their medications all or most of the time in the past six months (OR: 4.2, 95% CI: 1.4, 12.4), or had missed their asthma medication on two or more days in the past two weeks (OR: 2.6, 95% CI: 1.2, 5.8). Finally, compared with mothers with low depressive symptoms, mothers with high depressive symptoms reported less understanding about their child’s asthma medications (OR: 7.7; 95% CI: 1.7, 35.9). Celano et al. (2010) yielded contrasting results in their longitudinal cohort study of medication adherence assessed by electronic monitoring devices in 143 low-income African American children ages six to 11 years with persistent asthma, finding that baseline caregiver depressive symptoms were not a statistically significant predictor of medication adherence at 12 months. Despite the potential clinical significance of the relationship between level of caregiver depressive symptoms and child asthma treatment adherence, research in this area is inconclusive, dated, and limited in scope. Even less is known about the impact of maternal depressive symptomatology on medication adherence in racial minority children with poorly controlled asthma, a group already at high risk for poor asthma outcomes. The purpose of this longitudinal, multilevel analysis was to evaluate the relationship between caregiver depressive symptoms and child asthma medication adherence in a sample of young, urban African American children with high-risk asthma. Method Data Source This secondary data analysis used data collected from a randomized controlled trial (RCT) evaluating the efficacy of an environmental control behavioral–educational intervention for children with poorly controlled asthma over three time points (baseline, six months, and 12 months) between August 2013 and February 2016 (Butz et al., 2019). Caregivers of children with persistent, uncontrolled asthma were recruited from two large urban hospitals during a pediatric asthma ED visit. All children received allergy testing during the ED visit. The environmental control intervention consisted of nurse home visits for asthma management and targeted environmental control, a clinic visit for asthma education, and review of allergy test results. The control group in the study received standard asthma education during three nurse home visits. The RCT was approved by the institutional review boards of the associated universities. As reported by Butz et al. (2019), a total of 1,554 children and caregivers were screened for study enrollment in the ED. Of the 437 dyads who met eligibility criteria, 215 declined to participate, 222 enrolled in the study, and 204 completed the RCT. The 18 dropouts were split evenly between the intervention and control groups and were unable to be located at their most recent address or phone number. Ninety-four percent of the enrolled participants were African American (N = 208). The study design, data collection methods, and results of the RCT are described in detail by Butz et al. (2019). Measures Medication adherence, the primary outcome of interest, was measured by the five-item self-report Medication Adherence Report Scale (MARS) (Horne, 2004). Items on the Likert-type scale ascertain the mother’s report of lapses in child asthma medication administration including missing, altering, or forgetting doses. Response options are as follows: 1 = very often, 2 = often, 3 = sometimes, 4 = rarely, and 5 = never. Scores on the MARS range from 5 to 25, with higher scores indicating better adherence. A score of 20 points or higher indicates high adherence (Horne & Weinman, 1999). The MARS was previously adapted for use with caregivers of children with asthma (Koster, Raaijmakers, Vijverberg, & Maitland-van der Zee, 2011). Cronbach’s alpha on the MARS in this sample ranged from .75 to .98 across the three time points. Independent Variable Caregiver depressive symp-tomology was measured by the CES-D (Radloff, 1977), a 20-item self-report instrument in which individuals rate how often over the past week they experienced symptoms of depression (that is, sadness, crying spells, poor appetite). The four-point Likert scale includes the following response options for each item: 0 = rarely or none of the time, 1 = some or little of the time, 2 = moderately or much of the time, and 3 = most or almost all the time. Scores on the CES-D range from 0 to 60, with higher scores denoting greater depressive symptoms. The instrument has a cutoff score of 16 or higher, which indicates that an individual may be at risk for clinical depression (Henry, Grant, & Cropsey, 2018). The CES-D has been validated in diverse populations, including African American women (Makambi, Williams, Taylor, Rosenberg, & Adams-Campbell, 2009). Cronbach’s alpha on the CES-D in this sample ranged from .81 to .87 across time points, which is consistent with the psychometrics reported in the literature (Makambi et al., 2009). Covariates Group status (that is, intervention or control group), event time (that is, baseline, six months, or 12 months follow-up), baseline child asthma severity, and several key child and caregiver demographic characteristics were examined for association with medication adherence and included in the model as covariates. Group status was dummy-coded with “control” as the reference group. Event time was treated continuously. Child asthma severity was determined at baseline, according to the National Asthma Education Prevention Program guidelines (HHS, 2007), and included number of symptom days over the past 14 days, number of symptom nights over the past 30 nights, number of days of short-acting beta-agonist use over the past 14 days, and activity limitation over the past two weeks. Asthma severity includes four categories: (1) intermittent, (2) mild persistent, (3) moderate persistent, and (4) severe persistent. Only children with mild, moderate, and severe persistent asthma were included for participation in the larger RCT. Baseline asthma severity was dummy-coded with “moderate” and “severe” entered into the model and “mild” as the reference group. Child demographic data included age (in years) and gender, which was dummy-coded with “female” as the reference group. Caregiver education level—a dichotomous variable with the following response options, high school graduate and did not graduate high school—was dummy-coded with “did not graduate high school” as the reference group. Data Analysis The data included time points (level 1) nested within caregivers of children with asthma (level 2). The three time points at level 1 were baseline (T1), six months (T2), and 12 months (T3). There were 207 caregivers of children with asthma included at level 2. Data analysis was completed in several steps. First, a null model with the dependent variable (medication adherence) was estimated and the intraclass correlation (ICC) was calculated to assess whether the data were appropriate for multilevel modeling. The ICC represents the proportion of variability in medication adherence at level 2 (that is, caregiver) compared with the total variability in medication adherence. Generally, an ICC greater than .10 suggests that the data are clustered and a multilevel model approach is appropriate (Robson & Pevalin, 2016). After determining the ICC, a random intercept model with the independent variable of interest (depressive symptoms) and the covariates predicting medication adherence was estimated. Subsequently, to assess the within-caregiver (that is, comparing mothers with themselves over time) and between-caregiver effects (that is, comparing different mothers with each other over time) of depressive symptoms on medication adherence, two variables were added—the cluster mean of CES-D score (that is, the between-caregiver effect) and the deviation from cluster mean (that is, the within-caregiver effect)—to a third random intercept model. A postestimation command tested the null hypothesis that these two coefficients were the same. Data were analyzed using Stata 15.1 (StataCorp, 2017). Results The study sample (N = 208) comprised African American children ages three to 12 years with persistent and uncontrolled asthma living in a large mid-Atlantic city and their caregivers. The children had a mean age of 6.3 years (SD = 2.7 years) and were mostly male (65%). Asthma morbidity in the sample was high at baseline, with children averaging nearly six symptom days in the past two weeks and seven symptom nights in the past four weeks. Caregivers ranged in age from 18 to 62 years with a mean of 31.5 years (SD = 7.6). Almost all caregivers were female (97%), and most were single (75.6%). Table 1 presents full sociodemographic and baseline morbidity data for all participants. There were no statistically significant differences between the intervention and control groups in any category (Butz et al., 2019). Mean CES-D and MARS scores at each time point are also presented in Table 1. Although mean CES-D scores decreased over time, this change was not statistically significant. Over the three time points, on average, nearly a third of caregivers (29%) met the cutoff score for depression (that is, a score ≥ 16 on the CES-D). Scores on the MARS decreased over time, with a statistically significant difference between T1 and T3 (t = 2.99, p = .001) indicating a decrease in medication adherence across the time points. Table 1: Participant Demographics, Baseline Morbidity, and Mean CES-D and MARS Scores by Time Point Demographic Characteristic . T1 n (%) . T1 M (SD) . T2 M (SD) . T3 M (SD) . Child’s age 6.3 (2.7) Child male 136 (65) Asthma severity  Mild, persistent 55 (26.4)  Moderate, persistent 95 (45.7)  Severe, persistent 58 (27.9) Asthma-related school absences in past 3 months  None 69 (33.3)  1–5 days 104 (50)  6–10 days 24 (11.5)  11–15 days 6 (2.9)  ≥15 days 4 (1.9) Symptom days in past 2 weeks 5.9 (4.8) Symptom nights in past 4 weeks 7.1 (8.8) Caregiver age 31.5 (7.6) Caregiver female 202 (97) Caregiver single 156 (75) Caregiver high school education or more 168 (80.8) Annual household income($)  <10,000 57 (27.4)  10,000–$19,000 38 (18.3)  20,000–$29,000 32 (15.4)  30,000–$39,000 23 (11.1)  ≥40,000 30 (14.4)  Missing/refused 28 (13.4) CES-D 12.95 (11.43) 11.41 (10.69) 10.8 (10.76) MARS 21.88 (3.33) 21.22 (4.36) 20.53 (5.73) Demographic Characteristic . T1 n (%) . T1 M (SD) . T2 M (SD) . T3 M (SD) . Child’s age 6.3 (2.7) Child male 136 (65) Asthma severity  Mild, persistent 55 (26.4)  Moderate, persistent 95 (45.7)  Severe, persistent 58 (27.9) Asthma-related school absences in past 3 months  None 69 (33.3)  1–5 days 104 (50)  6–10 days 24 (11.5)  11–15 days 6 (2.9)  ≥15 days 4 (1.9) Symptom days in past 2 weeks 5.9 (4.8) Symptom nights in past 4 weeks 7.1 (8.8) Caregiver age 31.5 (7.6) Caregiver female 202 (97) Caregiver single 156 (75) Caregiver high school education or more 168 (80.8) Annual household income($)  <10,000 57 (27.4)  10,000–$19,000 38 (18.3)  20,000–$29,000 32 (15.4)  30,000–$39,000 23 (11.1)  ≥40,000 30 (14.4)  Missing/refused 28 (13.4) CES-D 12.95 (11.43) 11.41 (10.69) 10.8 (10.76) MARS 21.88 (3.33) 21.22 (4.36) 20.53 (5.73) Notes: CES-D = Center for Epidemiologic Studies Depression Scale; MARS = Medication Adherence Report Scale. Open in new tab Table 1: Participant Demographics, Baseline Morbidity, and Mean CES-D and MARS Scores by Time Point Demographic Characteristic . T1 n (%) . T1 M (SD) . T2 M (SD) . T3 M (SD) . Child’s age 6.3 (2.7) Child male 136 (65) Asthma severity  Mild, persistent 55 (26.4)  Moderate, persistent 95 (45.7)  Severe, persistent 58 (27.9) Asthma-related school absences in past 3 months  None 69 (33.3)  1–5 days 104 (50)  6–10 days 24 (11.5)  11–15 days 6 (2.9)  ≥15 days 4 (1.9) Symptom days in past 2 weeks 5.9 (4.8) Symptom nights in past 4 weeks 7.1 (8.8) Caregiver age 31.5 (7.6) Caregiver female 202 (97) Caregiver single 156 (75) Caregiver high school education or more 168 (80.8) Annual household income($)  <10,000 57 (27.4)  10,000–$19,000 38 (18.3)  20,000–$29,000 32 (15.4)  30,000–$39,000 23 (11.1)  ≥40,000 30 (14.4)  Missing/refused 28 (13.4) CES-D 12.95 (11.43) 11.41 (10.69) 10.8 (10.76) MARS 21.88 (3.33) 21.22 (4.36) 20.53 (5.73) Demographic Characteristic . T1 n (%) . T1 M (SD) . T2 M (SD) . T3 M (SD) . Child’s age 6.3 (2.7) Child male 136 (65) Asthma severity  Mild, persistent 55 (26.4)  Moderate, persistent 95 (45.7)  Severe, persistent 58 (27.9) Asthma-related school absences in past 3 months  None 69 (33.3)  1–5 days 104 (50)  6–10 days 24 (11.5)  11–15 days 6 (2.9)  ≥15 days 4 (1.9) Symptom days in past 2 weeks 5.9 (4.8) Symptom nights in past 4 weeks 7.1 (8.8) Caregiver age 31.5 (7.6) Caregiver female 202 (97) Caregiver single 156 (75) Caregiver high school education or more 168 (80.8) Annual household income($)  <10,000 57 (27.4)  10,000–$19,000 38 (18.3)  20,000–$29,000 32 (15.4)  30,000–$39,000 23 (11.1)  ≥40,000 30 (14.4)  Missing/refused 28 (13.4) CES-D 12.95 (11.43) 11.41 (10.69) 10.8 (10.76) MARS 21.88 (3.33) 21.22 (4.36) 20.53 (5.73) Notes: CES-D = Center for Epidemiologic Studies Depression Scale; MARS = Medication Adherence Report Scale. Open in new tab Findings from the multilevel models can be found in Table 2. The ICC from the null random intercept model demonstrated that 24.4% of the variance in medication adherence could be explained at the level of the caregiver. When depressive symptoms were added to the model, medication adherence scores were an average of .04 points lower (β = –.040, p = .03), meaning that for every one-point increase on the CES-D (that is, more depressive symptoms), medication adherence decreased. Once the covariates were added to the model, caregiver depressive symptomology remained a statistically significant predictor of medication adherence (β = –.043, p = .02). Time was the only other statistically significant predictor of medication adherence in the model (β = –.403, p < .001) suggesting that for every one-unit increase in time, medication adherence decreases by .403 points. Table 2: Unstandardized Coefficients for Multilevel Regression Models of Medication Adherence (N = 207) Variable . Null Model . Null + CES-D . Null + CES-D + Covariates . Null+ Mean CES-D + Deviation CES-D + Covariates  . Intercept 21.15** 21.65** 22.36** 21.87** CES-D –.040* –.043* — Moderate .451 .428 Severe –.198 –.354 Male .914 .948* High school graduate –.189 –.061 Intervention –.276 –.258 Time –.403** –.413*** Mean CES-D –.007 Deviation CES-D –.079** Level 1 variance 2.29 2.15 2.09 2.11 Level 2 variance 4.03 3.98 3.93 3.89 ICC .244 .225 .231 .226 Variable . Null Model . Null + CES-D . Null + CES-D + Covariates . Null+ Mean CES-D + Deviation CES-D + Covariates  . Intercept 21.15** 21.65** 22.36** 21.87** CES-D –.040* –.043* — Moderate .451 .428 Severe –.198 –.354 Male .914 .948* High school graduate –.189 –.061 Intervention –.276 –.258 Time –.403** –.413*** Mean CES-D –.007 Deviation CES-D –.079** Level 1 variance 2.29 2.15 2.09 2.11 Level 2 variance 4.03 3.98 3.93 3.89 ICC .244 .225 .231 .226 Note: CES-D = Center for Epidemiologic Studies Depression Scale; Mean CES-D = the cluster mean of CES-D score; Deviation CES-D = the deviation from the cluster mean of CES-D score; Time = event time; ICC = intraclass correlation. * p < .05. **p < .01. ***p < .001. Open in new tab Table 2: Unstandardized Coefficients for Multilevel Regression Models of Medication Adherence (N = 207) Variable . Null Model . Null + CES-D . Null + CES-D + Covariates . Null+ Mean CES-D + Deviation CES-D + Covariates  . Intercept 21.15** 21.65** 22.36** 21.87** CES-D –.040* –.043* — Moderate .451 .428 Severe –.198 –.354 Male .914 .948* High school graduate –.189 –.061 Intervention –.276 –.258 Time –.403** –.413*** Mean CES-D –.007 Deviation CES-D –.079** Level 1 variance 2.29 2.15 2.09 2.11 Level 2 variance 4.03 3.98 3.93 3.89 ICC .244 .225 .231 .226 Variable . Null Model . Null + CES-D . Null + CES-D + Covariates . Null+ Mean CES-D + Deviation CES-D + Covariates  . Intercept 21.15** 21.65** 22.36** 21.87** CES-D –.040* –.043* — Moderate .451 .428 Severe –.198 –.354 Male .914 .948* High school graduate –.189 –.061 Intervention –.276 –.258 Time –.403** –.413*** Mean CES-D –.007 Deviation CES-D –.079** Level 1 variance 2.29 2.15 2.09 2.11 Level 2 variance 4.03 3.98 3.93 3.89 ICC .244 .225 .231 .226 Note: CES-D = Center for Epidemiologic Studies Depression Scale; Mean CES-D = the cluster mean of CES-D score; Deviation CES-D = the deviation from the cluster mean of CES-D score; Time = event time; ICC = intraclass correlation. * p < .05. **p < .01. ***p < .001. Open in new tab The third random intercept model included the same covariates as the previous model with the exception of depressive symptoms. In this model the cluster mean of the CES-D score and the deviation from the cluster mean of the CES-D score were entered instead of the grand mean of CES-D score from the prior models, an approach that allows for different within and between effects (Rabe-Hesketh & Skrondal, 2012). Deviation from the cluster mean of CES-D score was statistically significant (β = –.079, p = .002), whereas the cluster mean CES-D score was not (β = –.007, p = .77). In other words, only the within-caregiver effect of depressive symptoms was a statistically significant predictor of medication adherence. The results of the postestimation command (lincom), which formally tests the null hypothesis that these two variables are the same, demonstrated that the within-caregiver effect did indeed differ from the between-caregiver effect (p = .048). In the third model, time also remained a statistically significant predictor of medication adherence (β = –.413, p < .001), with adherence decreasing over time. Finally, child gender was statistically significant (β = .948, p = .047), with caregivers of male children averaging a nearly one-point higher score on the MARS than caregivers of female children. Discussion This multilevel analysis is the first to show that caregiver depressive symptomology contributes to medication nonadherence over time in young, urban African American children with high-risk asthma. These results are consistent with and build on the findings of Bauman et al. (2002) and Bartlett et al. (2004), who measured child asthma treatment adherence by caregiver report; however, the present analysis uniquely examined the relationship between caregiver depressive symptoms and caregiver-reported medication adherence using multilevel modeling over time. One of the benefits of this type of analysis was the ability to separate the between-caregiver and within-caregiver effects of depressive symptoms on treatment adherence, which is important given the high levels of clinically significant symptoms of depression in this population. Indeed, in this sample, at each of the three time points, nearly a third of caregivers scored above the cut point (that is, a score of 16 or higher) on the CES-D. The results of the analysis suggest that changes in a caregiver’s baseline level of depressive symptoms influence medication adherence more than a caregiver’s mean level of depression. In other words, the caregiver’s own personal trajectory of depression symptoms (the within-caregiver effect) has a greater impact on medication adherence than how that caregiver’s average level of depression symptoms compares with the average level of depression symptoms of another caregiver (the between-caregiver effect). The clinical significance of this finding underscores the importance of screening for maternal depressive symptoms at each encounter. Indeed, beyond an initial assessment of caregivers’ depressive symptomology, health care providers working with this population need to establish a relationship with the caregivers, develop an understanding of the child and family’s social context, regularly assess caregivers’ depressive symptomatology, and respond with tailored interventions including linkages to community resources for ongoing support. The finding that adherence decreases over time only reinforces the need to monitor both caregiver depressive symptoms and medication adherence at each medical encounter rather than only once at baseline. In addition, it is important to note that medication adherence decreased over time in both the intervention and the control groups, even though all families received some type of in-home services. This finding suggests that caregivers of urban African American children with high-risk asthma may need more intensive services, prolonged engagement with services, or both, to enhance child asthma medication adherence. Social support—which includes the provision of empathy and understanding, guidance, affection, and material resources by family, friends, or professionals (Guidry, Aday, Zhang, & Winn, 1997; Holt-Lunstad & Uchino, 2015)—is associated with fewer symptoms of depression among low-income urban mothers of young, racial minority children (Bellin et al., 2018; Manuel, Martinson, Bledsoe-Mansori, & Bellamy, 2012; Siefert, Finlayson, Williams, Delva, & Ismail, 2007). Future studies should investigate the possible protective effect of social support to caregivers on child asthma medication adherence as well as the optimal type, frequency, and duration of that support. Finally, another interesting finding was that caregivers of male children, on average, have a nearly one point better adherence score compared with caregivers of female children. In a qualitative study of low-income African American mothers caring for children with sickle cell disease, Hill and Zimmerman (1995) found caregiving differences based on the child’s gender. Specifically, the authors found that mothers of sons saw them as sicker and consequently put more effort into their caregiving than did mothers of daughters. It is possible that the mothers in our sample may be enacting a similar pattern of gender-based caregiving, which would explain why they reported better adherence for male children; however, this relationship is tentative and merits further examination. This study has limitations to be considered. First, the outcome variable, medication adherence, was measured via caregiver self-report. Self-report measures, such as the MARS, are more likely to overestimate adherence compared with objective measures of adherence due to social desirability and memory biases (Stirratt et al., 2015). This overestimation tendency likely explains why scores on the MARS reflected high levels of adherence even though asthma morbidity was also high among the children in the sample. These results also suggest a ceiling effect, which should be further explored in future studies. Despite the limitations associated with measuring medication adherence via self-report, the advantages may outweigh the disadvantages of electronic medication adherence measures with high-risk populations such as urban African American children with persistent and uncontrolled asthma. Indeed, measuring medication adherence via self-report is likely the most pragmatic method of assessing adherence in the clinical context given that it is low cost, noninvasive, minimally burdensome for the patient, and flexible and easy to administer (Stirratt et al., 2015). Objective measures of medication adherence (that is, electronic monitors, pharmacy records, pill counters) are costly and require time and expertise in acquiring data. Given the advantages of self-reported medication adherence, it is vital that future research rigorously test self-report medication adherence measures for caregivers of young, urban African American children with high-risk asthma. In addition to the limitation of measuring adherence by caregiver self-report, it is also important to keep in mind that given the focused sample (caregivers of urban African American children with high-risk asthma) the results of this study may not generalize to a broader range of caregivers of children with asthma. Furthermore, as most of the caregivers in this sample were single women, additional research is needed to explore the relationship between caregiver depressive symptoms and medication adherence in other caregiver populations including fathers (or other male caregivers) and partnered caregivers. In summary, within-caregiver depressive symptoms (that is, changes in caregiver’s baseline depressive symptomology), as opposed to between-caregiver depressive symptoms (that is, caregiver’s mean level of depressive symptomology), contribute to medication nonadherence and have implications for health care providers working with African American children with persistent and uncontrolled asthma. A growing body of literature already acknowledges that health care providers need to attend to the psychosocial context of urban African American children with high-risk asthma, including caregiver mental health status, to optimize management of this chronic health condition (Bellin et al., 2017, 2018; Kub et al., 2018; Pak & Allen, 2012). The results of the analysis underscore the need for ongoing assessment of caregiver depressive symptoms as a potentially important target for improving medication adherence and reducing asthma morbidity in this vulnerable population. References Arroyo A.J.C. , Chee C. P. , Camargo C. A. Jr. , Wang N. E. ( 2018 ). Where do children die from asthma? National data from 2003 to 2015 . Journal of Allergy and Clinical Immunology: In Practice, 6 , 1034 – 1036 . doi:10.1016/j.jaip.2017.08.032 Google Scholar Crossref Search ADS WorldCat Bartlett S. J. , Krishnan J. A. , Riekert K. A. , Butz A. M. , Malveaux F. J. , Rand C. S. ( 2004 ). Maternal depressive symptoms and adherence to therapy in inner-city children with asthma . Pediatrics, 113 , 229 – 237 . Google Scholar Crossref Search ADS WorldCat Bauman L. J. , Wright E. , Leickly F. E. , Crain E. , Kruszon-Moran D. , Wade S. L. , Visness C. M. ( 2002 ). Relationship of adherence to pediatric asthma morbidity among inner-city children . Pediatrics, 110 , e6 . Google Scholar Crossref Search ADS WorldCat Bellin M. H. , Collins K. S. , Osteen P. , Kub J. , Bollinger M. E. , Newsome A. , et al. ( 2017 ). Characterization of stress in low-income, inner-city mothers of children with poorly controlled asthma . Journal of Urban Health, 94 , 814 – 823 . doi:10.1007/s11524-017-0162-1 Google Scholar Crossref Search ADS WorldCat Bellin M. H. , Newsome A. , Lewis-Land C. , Kub J. , Mudd S. S. , Margolis R. , Butz A. M. ( 2018 ). Improving care of inner-city children with poorly controlled asthma: What mothers want you to know . Journal of Pediatric Health Care, 32 , 387 – 398 . doi:10.1016/j.pedhc.2017.12.009 Google Scholar Crossref Search ADS WorldCat Butz A. M. , Bollinger M. E. , Ogborn J. , Morphew T. , Mudd S. S. , Kub J. E. , et al. ( 2019 ). Children with poorly controlled asthma: Randomized controlled trial of a home-based environmental control intervention. Pediatric Pulmonology, 54 , 245 – 256 . doi:10.1002/ppul.24239 Google Scholar Crossref Search ADS WorldCat Butz A. , Morphew T. , Lewis-Land C. , Kub J. , Bellin M. , Ogborn J. , et al. ( 2017 ). Factors associated with poor controller medication use in children with high asthma emergency department use . Annals of Allergy, Asthma & Immunology: Official Publication of the American College of Allergy, Asthma, & Immunology, 118 , 419 – 426 . doi:10.1016/j.anai.2017.01.007 Google Scholar Crossref Search ADS WorldCat Celano M. P. , Linzer J.F.S. , Demi A. , Bakeman R. , Smith C. O. , Croft S. , Kobrynski L. J. ( 2010 ). Treatment adherence among low-income, African American children with persistent asthma . Journal of Asthma, 47 , 317 – 322 . doi:10.3109/02770900903580850 Google Scholar Crossref Search ADS WorldCat Centers for Disease Control and Prevention. ( 2018 ). Most recent asthma data. Retrieved from https://www.cdc.gov/asthma/most_recent_data.htm DeRogatis L. R. , Spencer P. M. ( 1982 ). Administration and procedures: Brief Symptom Inventory manual I . Baltimore : Clinical Psychometric Research . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Easter G. , Sharpe L. , Hunt C. J. ( 2015 ). Systematic review and meta-analysis of anxious and depressive symptoms in caregivers of children with asthma . Journal of Pediatric Psychology, 40 , 623 – 632 . Google Scholar Crossref Search ADS WorldCat Gergen P. J. , Togias A. ( 2015 ). Inner city asthma . Immunology and Allergy Clinics of North America, 35 ( 1 ), 101 – 114 . doi:10.1016/j.iac.2014.09.006 Google Scholar Crossref Search ADS WorldCat Grenard J. L. , Munjas B. A. , Adams J. L. , Suttorp M. , Maglione M. , McGlynn E. A. , Gellad W. F. ( 2011 ). Depression and medication adherence in the treatment of chronic diseases in the United States: A meta-analysis . Journal of General Internal Medicine, 26 , 1175 – 1182 . doi:10.1007/s11606-011-1704-y Google Scholar Crossref Search ADS WorldCat Guidry J. J. , Aday L. A. , Zhang D. , Winn R. J. ( 1997 ). The role of informal and formal social support networks for patients with cancer . Cancer Practice, 5 , 241 – 246 . Google Scholar OpenURL Placeholder Text WorldCat Henry S. K. , Grant M. M. , Cropsey K. L. ( 2018 ). Determining the optimal clinical cutoff on the CES-D for depression in a community corrections sample . Journal of Affective Disorders, 234 , 270 – 275 . doi:10.1016/j.jad.2018.02.071 Google Scholar Crossref Search ADS WorldCat Hill S. A. , Zimmerman M. K. ( 1995 ). Valiant girls and vulnerable boys: The impact of gender and race on mothers’ caregiving for chronically ill children . Journal of Marriage and Family, 57 , 43 – 53 . doi:10.2307/353815 Google Scholar Crossref Search ADS WorldCat Holt-Lunstad J. , Uchino B. ( 2015 ). Social support and health . In Glanz K. , Rimer B. , Viswanath K. (Eds.), Health behavior: Theory, research, and practice (pp. 183 – 204) . Hoboken, NJ : John Wiley & Sons . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Horne R. ( 2004 ). The Medication Adherence Report Scale . Brighton, UK : University of Brighton. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Horne R. , Weinman J. ( 1999 ). Patients’ beliefs about prescribed medicines and their role in adherence to treatment in chronic physical illness . Journal of Psychosomatic Research, 47 , 555 – 567 . Google Scholar Crossref Search ADS WorldCat Kaugars A. S. , Klinnert M. D. , Bender B. G. ( 2004 ). Family influences on pediatric asthma . Journal of Pediatric Psychology, 29 , 475 – 491 . doi:10.1093/jpepsy/jsh051 Google Scholar Crossref Search ADS WorldCat Koster E. S. , Raaijmakers J.A.M. , Vijverberg S.J.H. , Maitland-van der Zee A.-H. ( 2011 ). Inhaled corticosteroid adherence in paediatric patients: The PACMAN cohort study . Pharmacoepidemiology and Drug Safety, 20 , 1064 – 1072 . doi:10.1002/pds.2228 Google Scholar Crossref Search ADS WorldCat Kub J. , Bellin M. H. , Butz A. , Bollinger M. E. , Lewis-Land C. , Osteen P. ( 2018 ). The chronicity of depressive symptoms in mothers of children with asthma . Western Journal of Nursing Research, 40 , 1581 – 1597 . doi:10.1177/0193945917705858 Google Scholar Crossref Search ADS WorldCat Makambi K. H. , Williams C. D. , Taylor T. R. , Rosenberg L. , Adams-Campbell L. L. ( 2009 ). An assessment of the CES-D scale factor structure in black women: The Black Women’s Health Study . Psychiatry Research, 168 , 163 – 170 . doi:10.1016/j.psychres.2008.04.022 Google Scholar Crossref Search ADS WorldCat Manuel J. I. , Martinson M. L. , Bledsoe-Mansori S. E. , Bellamy J. L. ( 2012 ). The influence of stress and social support on depressive symptoms in mothers with young children . Social Science & Medicine, 75 , 2013 – 2020 . Google Scholar Crossref Search ADS WorldCat Munzenberger P. , Secord E. , Thomas R. ( 2010 ). Relationship between patient, caregiver, and asthma characteristics, responsibility for management, and indicators of asthma control within an urban clinic . Journal of Asthma, 47 , 41 – 45 . doi:10.3109/02770900903395226 Google Scholar Crossref Search ADS WorldCat Pak L. , Allen P. J. ( 2012 ). The impact of maternal depression on children with asthma . Pediatric Nursing, 38 , 11 – 19 , 30. Google Scholar OpenURL Placeholder Text WorldCat Rabe-Hesketh S. , Skrondal A. ( 2012 ). Multilevel and longitudinal modeling using Stata, Volume 1: Continuous responses (3rd ed.). College Station, TX : Stata Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Radloff L. S. ( 1977 ). The CES-D scale: A self-report depression scale for research in the general population . Applied Psychological Measurements, 1 , 385 – 401 . Google Scholar Crossref Search ADS WorldCat Rand C. S. , Wright R. J. , Cabana M. D. , Foggs M. B. , Halterman J. S. , Olson L. , et al. ( 2012 ). Mediators of asthma outcomes . Journal of Allergy and Clinical Immunology, 129 (Suppl. 3), S136 – S141 . doi:10.1016/j.jaci.2011.12.987 Google Scholar Crossref Search ADS WorldCat Robson R. , Pevalin D. ( 2016 ). Multilevel modeling in plain language . London : SAGE . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Rohan J. , Drotar D. , McNally K. , Schluchter M. , Riekert K. , Vavrek P. , et al. ( 2010 ). Adherence to pediatric asthma treatment in economically disadvantaged African-American children and adolescents: An application of growth curve analysis . Journal of Pediatric Psychology, 35 , 394 – 404 . doi:10.1093/jpepsy/jsp074 Google Scholar Crossref Search ADS WorldCat Siefert K. , Finlayson T. L. , Williams D. R. , Delva J. , Ismail A. I. ( 2007 ). Modifiable risk and protective factors for depressive symptoms in low-income African American mothers . American Journal of Orthopsychiatry, 77 , 113 – 123 . Google Scholar Crossref Search ADS WorldCat StataCorp. ( 2017 ). Stata statistical software: Release 15 [Computer software]. College Station, TX : Author. Stirratt M. J. , Dunbar-Jacob J. , Crane H. M. , Simoni J. M. , Czajkowski S. , Hilliard M. E. , et al. ( 2015 ). Self-report measures of medication adherence behavior: Recommendations on optimal use . Translational Behavioral Medicine, 5 , 470 – 482 . doi:10.1007/s13142-015-0315-2 Google Scholar Crossref Search ADS WorldCat U.S. Department of Health and Human Services. ( 2007 ). Expert Panel Report 3 (EPR3): Guidelines for the diagnosis and management of asthma (NIH Publication No. 70-4051). Bethesda, MD : National Institutes of Health, National Asthma Education and Prevention Program . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Wood B. L. , Brown E. S. , Lehman H. K. , Khan D. A. , Lee M. J. , Miller B. D. ( 2018 ). The effects of caregiver depression on childhood asthma: Pathways and mechanisms . Annals of Allergy, Asthma & Immunology, 121 , 421 – 427 . doi:10.1016/j.anai.2018.06.031 Google Scholar Crossref Search ADS WorldCat Zahran H. S. , Bailey C. M. , Damon S. A. , Garbe P. L. , Breysse P. N. ( 2018 ). Vital signs: Asthma in children—United States, 2001 – 2016 . Morbidity & Mortality Weekly Report, 67 ( 5 ), 149 – 155 . Google Scholar OpenURL Placeholder Text WorldCat © 2020 National Association of Social Workers This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Dell, Nathaniel, A;Vidovic, Kristina, R;Huang,, Jin;Pelham,, Michelle

2020 Social Work Research

doi: 10.1093/swr/svaa007

Abstract The objective of this study was to identify the relationship of emotional reactivity with depression and anxiety symptoms among adults diagnosed with a serious mental illness (SMI) and to explore gender differences in these relationships. Cross-sectional data were collected from intensive case management services recipients (N = 150). Hierarchical multiple regression was used to identify the associations of self-reported positive emotions, fear, sadness, and anger to depression and anxiety, while testing the interaction of gender with emotional response. Compared with men, women reported significantly higher depressive and anxiety symptoms and greater reactivity to sadness and fear. Emotional response variables explained 35.5% of the variance in depression and 38.7% in anxiety. Gender did not moderate the relationship between emotional response and depression; however, gender did moderate the relationship between reactivity to positive emotions and anxiety. Self-reported emotional response may provide clinicians with insight into the severity and presentation of co-occurring anxiety and depressive symptoms among adults with SMI. Increasing the experience of positive emotions among women with SMI may contribute to reduced anxiety symptoms. Therapists and rehabilitation counselors may consider the interplay between mood and anxiety symptoms and emotional response styles to reduce the burden of psychiatric distress among people with SMI. Anxiety and depressive symptoms have been found to be prevalent among people diagnosed with serious mental illness (SMI) (Sala et al., 2012; Temmingh & Stein, 2015). Karpov and colleagues (2016) found among people with schizophrenia spectrum, bipolar, and depressive disorders that 40.2% to 55.6% of respondents reported frequent or constant anxiety (p. 3). Depression among people with schizophrenia varies by the stage and severity of the illness, with an estimated 40% in the early stage and 20% among those with chronic schizophrenia (Upthegrove, Marwaha, & Birchwood, 2017). Support for greater prevalence of anxiety and depressive disorders among people with schizophrenia spectrum disorders—compared with those without this diagnosis—also comes from the National Epidemiological Survey on Alcohol and Related Conditions (McMillan, Enns, Cox, & Sareen, 2009). Within clinical samples, anxiety and depressive symptoms appeared strongly correlated, independent of primary diagnostic categories (Karpov et al., 2016). Gender differences are associated with variation in the prevalence of depressive and anxiety symptoms among those with SMI. In the general population, higher rates of generalized anxiety disorder (GAD) and mood disorders have been found in women (McLean, Asnaani, Litz, & Hofmann, 2011; Vesga-López et al., 2008). Higher anxiety and depressive symptoms have also been found among women with an SMI. Women diagnosed with schizophrenia spectrum disorders may show higher levels of depressive symptoms (Abel, Drake, & Goldstein, 2010). In the case of bipolar disorder, women have been found to show greater predominance of depressive polarity rather than (hypo)manic polarity (Nivoli et al., 2011). Differences in emotional response have been hypothesized for both depressive and anxiety disorders. Emotional response may relate to major depressive disorder (MDD) in at least three distinct ways: increased emotional reactivity to negative emotional cues (negative potentiation); reduced reactivity to positive emotional cues (positive attenuation); or emotion context insensitivity (ECI), which is when individuals experience reduced reactivity to both positive and negative emotional cues (Bylsma, Morris, & Rottenberg, 2008). A meta-analytic review found the ECI hypothesis to offer the most succinct account of emotional response in MDD, characterized by reduced emotional reactivity to both positively and negatively valenced cues, with the reduction larger for positive cues than for negative cues (Bylsma et al., 2008). The ECI view of emotional reactivity in MDD is supported by laboratory data but has mixed findings in natural settings (Bylsma, Taylor-Clift, & Rottenberg, 2011). GAD may be characterized by four significant deficits in emotional experience and regulation compared with those not diagnosed with GAD: (1) greater emotional arousal, (2) difficulties describing and clarifying emotional experiences, (3) holding irrational beliefs about the consequences of both negative and positive emotions, and (4) difficulties with self-soothing (Mennin, Heimberg, Turk, & Fresco, 2005). People experiencing GAD may have difficulty adjusting to emotional experiences and limited success with controlling or suppressing emotions (Barr, Kahn, & Schneider, 2008; Campbell-Sills, Barlow, Brown, & Hofmann, 2006; Mennin et al., 2005). Although Mennin and colleagues (2005) specified those deficits in relation to GAD, these may also be found in social anxiety disorder, with individuals experiencing lower positive expression, less attentiveness to emotions, and difficulty describing emotions (Barr et al., 2008; Salovey, Stroud, Woolery, & Epel, 2002; Turk, Heimberg, Luterek, Mennin, & Fresco, 2005). Orsillo, Theodore-Oklota, Luterek, and Plumb (2007) developed the Emotional Reactivity and Numbing Scale (ERNS) to measure restricted or exaggerated emotional response among people diagnosed with posttraumatic stress disorder (PTSD). Whereas emotional numbing refers to a restricted experience of emotional arousal, emotional reactivity refers to intense experiences of emotional arousal. Both emotional numbing and emotional reactivity are interpreted as automatic responses to potentially traumatic or threatening stimuli. People experiencing emotional numbing may lack positive emotions (in the case of anhedonia) and may have limited emotional response to other emotional experiences (for example, fear, anger, sadness). Although the ERNS tests emotional response patterns in the context of PTSD, it has not been widely explored in relation to mood and anxiety-related disorders. Whereas emotional response varies in relation to depressive and anxiety disorders, gender differences in emotional response have also been reported (Deng, Chang, Yang, Huo, & Zhou, 2016). Women may express stronger subjective emotions, whereas men may express stronger physiological reactions to different emotion-invoking stimuli (Deng et al., 2016). At least in the context of the United States, gender-differentiated norms, which arise from economic and political conditions, specify that women are more emotionally responsive than men (Grossman & Wood, 1993). On the other hand, men are more likely to express anger whereas women express dysphoria in response to stress (McRae, Ochsner, Mauss, Gabrieli, & Gross, 2008). Given previously observed gender differences in anxiety and depressive symptoms, and the unclear relationship of emotional response to both gender and mental illness, the present study assesses for gender differences in relationship between emotional response and psychiatric distress among people with SMI. For social workers, the relationship of anxiety and depressive symptoms, gender, and emotional response among people with SMI are important to consider for several reasons. Self-reported emotional response may be useful for social workers assessing clients’ emotional well-being beyond the discrete symptoms of specific psychiatric disorders. Subjective report of emotional response may relate to concepts such as emotional regulation, emotional expression, and emotional processing. Gross (1998) summarized emotional regulation as “the processes by which individuals influence which emotions they have, when they have them, and how they experience and express these emotions” (p. 275). Maladaptive emotional responses may indicate that one has misinterpreted important information about oneself or others (Pos, Greenberg, Goldman, & Korman, 2003). Furthermore, maladaptive responses may contribute to greater psychiatric distress and poorer outcomes in therapy (Angelakis & Nixon, 2015). Treatment may be less efficacious if client engagement with emotions is restricted or if the emotions are overly active (Rauch & Foa, 2006). The importance of emotional processing in therapy is recognized across different theoretical perspectives (for example, psychoanalytic, cognitive–behavioral) (Pos et al., 2003). In addition, social workers may, by identifying gender differences in emotional response and mental health, address gender disparities in the treatment of depressive and anxiety symptoms. Depression and anxiety comorbid with other mental health conditions can contribute to poorer treatment outcomes. The present study has the following three aims, which consider the relationship between emotional response and mental health, gender differences among these variables, and the limited use of ERNS outside of a PTSD context: (1) extend the use of ERNS to the measurement of emotional response within a psychiatric outpatient sample, (2) explore gender differences in emotional response styles and psychiatric distress, (3) identify the association of emotional response with depression and anxiety symptoms while testing whether gender moderates the relationship of emotional response and depression and anxiety symptoms. Regarding aim 1, we hypothesized that ERNS subscales would display adequate internal consistency and would correlate weakly to moderately with depressive and anxiety symptoms. With aim 2, we hypothesized that women would endorse more severe depressive and anxiety symptoms. For aim 3, we hypothesized that lower reactivity to positive emotions would be most strongly related to depressive symptoms, whereas reactivity to fear and anger would relate more strongly to anxiety symptoms overall, and that gender would moderate the relationship between emotional response and psychiatric distress. Method The present study was conducted at an urban community mental health center in the midwestern United States. The agency’s institutional review board approved the study’s research procedures. Signed informed consent was collected from each participant. Participants Participants (N = 150) receiving outpatient psychiatric services were convenience sampled from teams providing intensive case management and integrated treatment for co-occurring disorders. Exclusion criteria for participating consisted of having organic brain conditions, currently experiencing a mental health–related crisis, unable to provide informed consent, and being unable to communicate in English. Listwise deletion was used, as fewer than 5% of cases were excluded from the analysis and Little’s “missing completely at random” test was nonsignificant (p > .05). Data Collection Procedures Between May and September 2018, participants were recruited through case manager referrals and the agency’s psychosocial rehabilitation center. The agency’s research staff conducted in-person interviews, generally lasting between 30 and 45 minutes. In-person interviews were conducted so that research staff could assist respondents in interpreting questions, if any confusion or lack of clarity was reported. Research staff, consisting of MSWs, were not involved in direct service provision. Participants were provided a $5 gift card for their time. Current primary diagnosis, which is established by an agency-employed psychiatric prescriber, was obtained after receiving participant consent. Instrumentation In addition to age, gender, race, marital status, and education level, the following self-report measures were collected from participants. Emotional Reactivity and Numbing The ERNS is a 62-item scale that measures numbing and hyperactivity to positive emotions (26 items), sadness (11 items), anger (11 items), fear (nine items), and general emotions (eight items). The ERNS measures typical performance (that is, how respondents generally react to different emotionally provocative situations) and includes both negatively and positively worded items. Participants rate each item on a Likert scale ranging from 1 = not typical of me to 5 = entirely typical of me. In addition to veteran samples (Gackenbach, Ellerman, & Hall, 2011; Orsillo et al., 2007), ERNS has been used to measure emotional response in trauma-exposed youth offender samples (Bennett & Kerig, 2014; Bennett, Modrowski, Kerig, & Chaplo, 2015; Kerig, Bennett, Thompson, & Becker, 2012; Kerig, Chaplo, Bennett, & Modrowski, 2016) and in a trauma-exposed university student sample (Gackenbach, Darlington, Ferguson, & Boyes, 2013). Previous studies typically noted adequate to good Cronbach’s alpha for each subscale. Orsillo and colleagues (2007) reported good test–retest reliability and construct validity. All subscales except for the general subscale were used in this analysis, as we wished to test the relationship of specific emotional responses to psychiatric distress rather than a more diffuse general emotion construct. Depressive Symptoms The Patient Health Questionnaire-9 (PHQ-9) is a nine-item scale measuring depressive symptoms; higher scores indicate more severe depression, with a score greater than 9 suggested as a cutoff for a probable major depressive episode (Kroenke, Spitzer, & Williams, 2001). However, in psychiatric samples a higher cutoff score of 13 or 14 may be warranted (Beard, Hsu, Rifkin, Busch, & Björgvinsson, 2016). The PHQ-9 had good internal consistency in the present study (α = .87). Anxiety Symptoms The General Anxiety Disorder-7 (GAD-7) (Spitzer, Kroenke, Williams, & Löwe, 2006) is a seven-item scale used to measure anxiety symptoms. A score greater than 10 has been suggested as a cutoff for probable GAD, with higher scores indicating more severe anxiety. The GAD-7 had good internal consistency in this sample (α = .92). Data Analysis To address aim 1, internal consistencies for ERNS subscales are presented and correlations among ERNS subscales are examined—as is the correlation of each subscale with anxiety and depressive symptoms. For aim 2, gender differences in emotional response and psychiatric distress are explored using independent samples t tests and chi-square tests. For aim 3, hierarchical multiple regression was conducted to (a) identify the association of emotional response styles to anxiety and depressive symptoms and (b) test whether gender and emotional response interacted to explain unique variance in depressive and anxiety symptoms. Two multiple regression models were conducted, one for depression and one for anxiety symptoms. Emotional response variables were entered in the first step, followed by gender and primary diagnostic category, and then the interaction of gender with each emotional response variable. Regression models were examined for multicollinearity, normality of error terms, and heteroscedasticity. Influential points were identified by examining standardized residuals and Cook’s d. Results Sample Characteristics Over half (55.33%, n = 83) of the sample identified as male; one participant identified as a transgender woman. Participants primarily identified as African American (61.3%, n = 92), with 37.33% (n = 56) identifying as White and 1.33% (n = 2) identifying as another racial group. The mean age was 58.42 (SD = 5.86) years. Primary diagnoses fell into five general categories: nearly 40% (n = 61) were diagnosed with a schizophrenia spectrum disorder, 34.67% (n = 52) with MDD, 17.33% (n = 26) with bipolar disorders, and 7.33% (n = 11) with other diagnoses (such as personality disorders or PTSD). Just less than one-third of participants (n = 48) reported less than a high school education, 33.33% (n = 50) reported having a high school education, and 34.67% (n = 52) had more than a high school education—whether vocational/trade school, some college, or a college degree. The sample was primarily unmarried or never married (68.00%, n = 102), followed by 24.00% (n = 36) divorced/separated, 4.00% (n = 6) currently married, and 4.00% (n = 6) widowed. Aim 1: Emotional Response in an SMI Sample In the present study, the positive emotion (α = .87), sadness (α = .77), anger (α = .82), and fear (α =.77) subscales of the ERNS had adequate to good internal consistency. The mean, standard deviation, minimum and maximum scores for each ERNS subscale are as follows: positive emotion (M = 101.83, SD = 15.05, min = 65, max = 127), sadness (M = 42.40, SD = 7.36, min = 15, max = 55), fear (M = 20.76, SD = 5.83, min = 6, max = 30), and anger (M = 37.26, SD = 7.32, min = 15, max = 53). As shown in Table 1, reactivity to positive stimuli was moderately correlated with reactivity to sadness, weakly associated with fear, and uncorrelated with anger. Sadness was moderately related to fear and anger. Fear and anger were weakly correlated. Table 1: Correlations among Depression, Anxiety, and Emotional Reactivity Variables . 1 . 2 . 3 . 4 . 5 . 6 . 1. PHQ-9 1 2. GAD-7 .80*** 1 3. ERNS positive  emotion –.32*** –.30*** 1 4. ERNS sadness .11 .18* .58*** 1 5. ERNS fear .30*** .32*** .23** .59*** 1 6. ERNS anger .21* .31*** .14 .43*** .37*** 1 . 1 . 2 . 3 . 4 . 5 . 6 . 1. PHQ-9 1 2. GAD-7 .80*** 1 3. ERNS positive  emotion –.32*** –.30*** 1 4. ERNS sadness .11 .18* .58*** 1 5. ERNS fear .30*** .32*** .23** .59*** 1 6. ERNS anger .21* .31*** .14 .43*** .37*** 1 Notes: PHQ-9 = Patient Health Questionnaire-9; GAD-7 = General Anxiety Disorder-7; ERNS = Emotional Reactivity and Numbing Scale. * p < .05. **p < .01. ***p < .001. Open in new tab Table 1: Correlations among Depression, Anxiety, and Emotional Reactivity Variables . 1 . 2 . 3 . 4 . 5 . 6 . 1. PHQ-9 1 2. GAD-7 .80*** 1 3. ERNS positive  emotion –.32*** –.30*** 1 4. ERNS sadness .11 .18* .58*** 1 5. ERNS fear .30*** .32*** .23** .59*** 1 6. ERNS anger .21* .31*** .14 .43*** .37*** 1 . 1 . 2 . 3 . 4 . 5 . 6 . 1. PHQ-9 1 2. GAD-7 .80*** 1 3. ERNS positive  emotion –.32*** –.30*** 1 4. ERNS sadness .11 .18* .58*** 1 5. ERNS fear .30*** .32*** .23** .59*** 1 6. ERNS anger .21* .31*** .14 .43*** .37*** 1 Notes: PHQ-9 = Patient Health Questionnaire-9; GAD-7 = General Anxiety Disorder-7; ERNS = Emotional Reactivity and Numbing Scale. * p < .05. **p < .01. ***p < .001. Open in new tab Greater reactivity to positive stimuli shared a weak inverse relationship with depressive symptoms, whereas reactivity to fear and anger shared a weaker, but significant relationship with depressive symptoms. It is interesting that reactivity to sadness was uncorrelated with depressive symptoms in this sample. Similarly, reactivity to positive stimuli also shared a weak inverse relationship with anxiety symptoms, whereas reactivity to fear and anger shared a weak positive relationship. Reactivity to sadness was very weakly correlated with anxiety symptoms. Aim 2: Gender Differences in Psychiatric Distress and Emotional Response Mean depression scores were moderately high (M = 8.81, SD = 7.00), with those identifying as women reporting higher scores (M = 10.25, SD = 7.02) compared with those identifying as men [M = 7.69, SD = 6.81; t(146) = –2.238, p = .027, δ = –0.37]. Probable major depressive episode was found in 41.22% (n = 61) of the sample, with significant differences by gender consistent with the differences in mean scores [χ2(1) = 5.885, p = .015, V = .199]. Similarly, anxiety scores were high overall (M = 7.62, SD = 6.71), with scores higher among women (M = 9.54, SD = 7.27) compared with men [M = 6.07, SD = 5.81; t(148) = –3.245, p = .002, δ = –0.53). Probable GAD was found in 32.00% of the sample (n = 48), with significant differences by gender also consistent with mean differences [χ2(1) = 11.329, p = .001, V = .275]. Regarding emotional response variables, no significant gender differences were observed for positive emotions [t(145) = –1.452, p = .149] or anger [t(146) = –1.316, p = .190]. However, women reported greater reactivity to sadness (M = 44.17, SD = 5.96) compared with men [M = 41.00, SD = 8.07; t(147) = –2.662, p = .009, δ = –0.44]. Similarly, women reported greater reactivity to fear (M = 22.28, SD = 5.66) compared with men [M = 19.58, SD = 5.72; t(146) = –2.862, p = .005, δ = –0.47]. Aim 3: Association of Emotional Response with Depression and Anxiety The ERNS sadness subscale was excluded from the regression analyses as sadness was uncorrelated with depression and moderately strongly correlated with reactivity to positive emotions, which may call aspects of its construct validity into question. In addition, when the regression analyses were initially conducted, sadness did not explain significant variance in anxiety or depressive symptoms when holding the fear, anger, and positive emotional reactivity subscales constant. Multicollinearity was unproblematic when including only positive emotion, anger, and fear subscales, with variance inflation factor (VIF) scores ranging from 1.12 to 2.58. When sadness was included in the model, the VIF for sadness was 4.57. Error terms were normally distributed, indicated by inspection of Q–Q plots and nonsignificant Shapiro–Wilkes tests (p > .05). Significant Breusch–Pagan tests indicated that heteroscedasticity was problematic in the model predicting anxiety but not depressive symptoms; therefore, robust standard errors were estimated for the model predicting anxiety. In each model, similar amounts of variance were predicted in scores for depression (R2 = .355) and anxiety (R2 = .387) (see Table 2). The similarities in variance explained across models may result from the strong correlation between PHQ-9 and GAD-7 scores (r = .80, p < .001). Table 2: Relationship of Emotional Numbing and Reactivity Subscales to Anxiety and Depressive Symptoms Depressive Symptom (N = 143) . Block 1 . Block 2 . Block 3 . β (SE) . B . β (SE) . B . β (SE) . B . Positive emotion –0.194 (0.036)*** –.411 –0.214 (0.034)*** –.431 –0.199 (0.048)*** –.422 Fear 0.400 (0.097)*** .334 0.352 (0.095)*** .294 0.413 (0.133)** .344 Anger 0.131 (0.077)† .136 0.135 (0.073)† .141 0.065 (0.099) .068 Diagnosis Primary SSA (reference) Primary MDD 3.779 (1.143)** .255 3.852 (1.150)** .260 Primary BPD 3.229 (1.436)* .175 3.431 (1.452)* .186 Other primary 4.425 (1.951)* .169 4.350 (1.979)* .166 Female 1.715 (1.041) .122 1.747 (1.050)† .124 Female × positive –0.042 (0.070) –.060 Female × fear –0.134 (0.193) –.072 Female × anger 0.164 (0.151) .109 Intercept 15.571 (4.150)*** 16.291 (4.105)*** 15.372 (5.398)** F(df) 15.85 (3, 139)*** 10.26 (7, 135)*** 7.27 (10, 132)*** R2 (ΔR2) 0.255 0.347 (0.092)*** 0.355 (0.008) Depressive Symptom (N = 143) . Block 1 . Block 2 . Block 3 . β (SE) . B . β (SE) . B . β (SE) . B . Positive emotion –0.194 (0.036)*** –.411 –0.214 (0.034)*** –.431 –0.199 (0.048)*** –.422 Fear 0.400 (0.097)*** .334 0.352 (0.095)*** .294 0.413 (0.133)** .344 Anger 0.131 (0.077)† .136 0.135 (0.073)† .141 0.065 (0.099) .068 Diagnosis Primary SSA (reference) Primary MDD 3.779 (1.143)** .255 3.852 (1.150)** .260 Primary BPD 3.229 (1.436)* .175 3.431 (1.452)* .186 Other primary 4.425 (1.951)* .169 4.350 (1.979)* .166 Female 1.715 (1.041) .122 1.747 (1.050)† .124 Female × positive –0.042 (0.070) –.060 Female × fear –0.134 (0.193) –.072 Female × anger 0.164 (0.151) .109 Intercept 15.571 (4.150)*** 16.291 (4.105)*** 15.372 (5.398)** F(df) 15.85 (3, 139)*** 10.26 (7, 135)*** 7.27 (10, 132)*** R2 (ΔR2) 0.255 0.347 (0.092)*** 0.355 (0.008) Anxiety Symptoms (N = 145)a . Block 1 . Block 2 . Block 3 . β (SE) . B . β (SE) . B . β (SE) . B . Positive emotion –0.181 (0.030)*** –.398 –0.201 (0.031)*** –.442 –0.142 (0.039)*** –.313 Fear 0.361 (0.084)*** .314 0.314 (0.082)*** .273 0.341 (0.107)** .296 Anger 0.232 (0.064)*** .253 0.237 (0.063)*** .257 0.200 (0.080)* .217 Diagnosis Primary SSA (reference) 1.480 (1.088) .104 1.585 (1.086) .111 Primary MDD 2.455 (1.29)† .138 2.708 (1.297)* .152 Primary BPD 3.389 (1.847)† .133 3.725 (1.843)* .146 Other primary 2.530 (1.006)* .187 2.575 (1.013)* .190 Female Female × positive –0.129 (0.062)* –.194 Female × fear –0.086 (0.157) –.049 Female × anger 0.089 (0.121) .063 Intercept 9.927 (3.890)* 10.451 (4.109)* 5.343 (5.057) F(df) 26.22 (3, 141)*** 14.77 (7, 137)*** 10.38 (10, 134)*** R2 (ΔR2) 0.293 0.365 (0.071)** 0.387 (0.022) Anxiety Symptoms (N = 145)a . Block 1 . Block 2 . Block 3 . β (SE) . B . β (SE) . B . β (SE) . B . Positive emotion –0.181 (0.030)*** –.398 –0.201 (0.031)*** –.442 –0.142 (0.039)*** –.313 Fear 0.361 (0.084)*** .314 0.314 (0.082)*** .273 0.341 (0.107)** .296 Anger 0.232 (0.064)*** .253 0.237 (0.063)*** .257 0.200 (0.080)* .217 Diagnosis Primary SSA (reference) 1.480 (1.088) .104 1.585 (1.086) .111 Primary MDD 2.455 (1.29)† .138 2.708 (1.297)* .152 Primary BPD 3.389 (1.847)† .133 3.725 (1.843)* .146 Other primary 2.530 (1.006)* .187 2.575 (1.013)* .190 Female Female × positive –0.129 (0.062)* –.194 Female × fear –0.086 (0.157) –.049 Female × anger 0.089 (0.121) .063 Intercept 9.927 (3.890)* 10.451 (4.109)* 5.343 (5.057) F(df) 26.22 (3, 141)*** 14.77 (7, 137)*** 10.38 (10, 134)*** R2 (ΔR2) 0.293 0.365 (0.071)** 0.387 (0.022) Notes: SSA = schizophrenia/schizoaffective disorder; MDD = major depressive disorder; BPD = bipolar disorder.aSample sizes vary due to missing data. † p < .01. *p < .05. **p < .01. ***p < .001. Open in new tab Table 2: Relationship of Emotional Numbing and Reactivity Subscales to Anxiety and Depressive Symptoms Depressive Symptom (N = 143) . Block 1 . Block 2 . Block 3 . β (SE) . B . β (SE) . B . β (SE) . B . Positive emotion –0.194 (0.036)*** –.411 –0.214 (0.034)*** –.431 –0.199 (0.048)*** –.422 Fear 0.400 (0.097)*** .334 0.352 (0.095)*** .294 0.413 (0.133)** .344 Anger 0.131 (0.077)† .136 0.135 (0.073)† .141 0.065 (0.099) .068 Diagnosis Primary SSA (reference) Primary MDD 3.779 (1.143)** .255 3.852 (1.150)** .260 Primary BPD 3.229 (1.436)* .175 3.431 (1.452)* .186 Other primary 4.425 (1.951)* .169 4.350 (1.979)* .166 Female 1.715 (1.041) .122 1.747 (1.050)† .124 Female × positive –0.042 (0.070) –.060 Female × fear –0.134 (0.193) –.072 Female × anger 0.164 (0.151) .109 Intercept 15.571 (4.150)*** 16.291 (4.105)*** 15.372 (5.398)** F(df) 15.85 (3, 139)*** 10.26 (7, 135)*** 7.27 (10, 132)*** R2 (ΔR2) 0.255 0.347 (0.092)*** 0.355 (0.008) Depressive Symptom (N = 143) . Block 1 . Block 2 . Block 3 . β (SE) . B . β (SE) . B . β (SE) . B . Positive emotion –0.194 (0.036)*** –.411 –0.214 (0.034)*** –.431 –0.199 (0.048)*** –.422 Fear 0.400 (0.097)*** .334 0.352 (0.095)*** .294 0.413 (0.133)** .344 Anger 0.131 (0.077)† .136 0.135 (0.073)† .141 0.065 (0.099) .068 Diagnosis Primary SSA (reference) Primary MDD 3.779 (1.143)** .255 3.852 (1.150)** .260 Primary BPD 3.229 (1.436)* .175 3.431 (1.452)* .186 Other primary 4.425 (1.951)* .169 4.350 (1.979)* .166 Female 1.715 (1.041) .122 1.747 (1.050)† .124 Female × positive –0.042 (0.070) –.060 Female × fear –0.134 (0.193) –.072 Female × anger 0.164 (0.151) .109 Intercept 15.571 (4.150)*** 16.291 (4.105)*** 15.372 (5.398)** F(df) 15.85 (3, 139)*** 10.26 (7, 135)*** 7.27 (10, 132)*** R2 (ΔR2) 0.255 0.347 (0.092)*** 0.355 (0.008) Anxiety Symptoms (N = 145)a . Block 1 . Block 2 . Block 3 . β (SE) . B . β (SE) . B . β (SE) . B . Positive emotion –0.181 (0.030)*** –.398 –0.201 (0.031)*** –.442 –0.142 (0.039)*** –.313 Fear 0.361 (0.084)*** .314 0.314 (0.082)*** .273 0.341 (0.107)** .296 Anger 0.232 (0.064)*** .253 0.237 (0.063)*** .257 0.200 (0.080)* .217 Diagnosis Primary SSA (reference) 1.480 (1.088) .104 1.585 (1.086) .111 Primary MDD 2.455 (1.29)† .138 2.708 (1.297)* .152 Primary BPD 3.389 (1.847)† .133 3.725 (1.843)* .146 Other primary 2.530 (1.006)* .187 2.575 (1.013)* .190 Female Female × positive –0.129 (0.062)* –.194 Female × fear –0.086 (0.157) –.049 Female × anger 0.089 (0.121) .063 Intercept 9.927 (3.890)* 10.451 (4.109)* 5.343 (5.057) F(df) 26.22 (3, 141)*** 14.77 (7, 137)*** 10.38 (10, 134)*** R2 (ΔR2) 0.293 0.365 (0.071)** 0.387 (0.022) Anxiety Symptoms (N = 145)a . Block 1 . Block 2 . Block 3 . β (SE) . B . β (SE) . B . β (SE) . B . Positive emotion –0.181 (0.030)*** –.398 –0.201 (0.031)*** –.442 –0.142 (0.039)*** –.313 Fear 0.361 (0.084)*** .314 0.314 (0.082)*** .273 0.341 (0.107)** .296 Anger 0.232 (0.064)*** .253 0.237 (0.063)*** .257 0.200 (0.080)* .217 Diagnosis Primary SSA (reference) 1.480 (1.088) .104 1.585 (1.086) .111 Primary MDD 2.455 (1.29)† .138 2.708 (1.297)* .152 Primary BPD 3.389 (1.847)† .133 3.725 (1.843)* .146 Other primary 2.530 (1.006)* .187 2.575 (1.013)* .190 Female Female × positive –0.129 (0.062)* –.194 Female × fear –0.086 (0.157) –.049 Female × anger 0.089 (0.121) .063 Intercept 9.927 (3.890)* 10.451 (4.109)* 5.343 (5.057) F(df) 26.22 (3, 141)*** 14.77 (7, 137)*** 10.38 (10, 134)*** R2 (ΔR2) 0.293 0.365 (0.071)** 0.387 (0.022) Notes: SSA = schizophrenia/schizoaffective disorder; MDD = major depressive disorder; BPD = bipolar disorder.aSample sizes vary due to missing data. † p < .01. *p < .05. **p < .01. ***p < .001. Open in new tab In the first model, emotional response variables predicted 25.5% of the variance in depressive symptoms; only anger did not significantly contribute to the model, although a trend (p < .10) was observed that was attenuated in subsequent steps. Gender and primary diagnosis explained an additional 9.2% of the variance (p < .001); however, the interaction of gender with emotional response variables was insignificant. In the second model, emotional response variables explained 29.3% of the variance in anxiety symptoms. Gender and primary diagnosis explained an additional 7.1% of the variance (p < .01). Only the interaction of gender with positive emotional reactivity was significantly associated with decreased anxiety symptoms (B = –.194, p < .05). Discussion The present study explored gender differences in the relationship between self-reported emotional reactivity and psychiatric distress in an outpatient psychiatric sample. The first aim was to extend the use of ERNS to an outpatient psychiatric sample and to test the relationship of ERNS subscales to other measures of psychiatric distress. This sample differs from the veteran, juvenile adolescent, and college samples from which ERNS scores have been collected in several ways. Participants were middle-age or older adults involved with outpatient psychiatric services who were diagnosed primarily with schizophrenia spectrum disorders, bipolar disorder, or MDD. Within this sample, internal consistency of the ERNS subscales was adequate to good, which is comparable to Orsillo et al.’s (2007) initial validation study using a primarily middle-age (M = 49.54, SD = 10.61) veteran sample. To our knowledge, this study is the first to explore how ERNS subscales correlated with one another—outside of the juvenile adolescent studies (Kerig, Bennett, Chaplo, Modrowski, & McGee, 2016; Kerig et al., 2012). It was hypothesized that greater reactivity to positive emotions would be only weakly correlated with more negatively valenced emotions such as fear, anger, and sadness. Greater reactivity to positive emotions was not hypothesized to be inversely correlated with sadness, fear, or anger because people with a heightened experience of positive emotions would not necessarily experience numbness toward other emotional experiences. The hypothesis was mainly supported, with the exception of sadness having correlated moderately strongly with the positive emotion subscale; this finding was unanticipated, given the weaker correlations of positive reactivity to fear and anger. The positive emotion subscale differs from the other three in at least one important way. The subscales for fear, sadness, and anger purport to measure either restricted or heightened responses to different stimuli. Higher scores on the fear, sadness, and anger subscales are theorized to indicate greater distress. The positive subscale may measure emotional numbing, but the measurement of excessive or exaggerated positive emotional response is called into question. Whereas reactivity in the context of the other subscales could be considered maladaptive in certain contexts, it is not obvious how this is the case in the positive emotion subscale. Future research should consider the extent to which higher scores on each subscale indicate trait hyperreactivity to different situations. Although ERNS aims to measure the subjective experience of different emotions, respondents must be able both to identify different emotional responses and also be willing to disclose whether such emotions are experienced. Because the expression of emotion differs from the affective experience itself, varying ways in expressing emotion are important to consider as potential confounding variables when interpreting ERNS scores. Future research using ERNS may need to explore to what degree respondents’ tendency to suppress emotional experiences may relate to the endorsement of emotional reactivity or numbness. Additional research is required to establish other aspects of validity in this population. Cognitive interviewing may be used to identify problematic items and to understand how well participants comprehend the questions, how they decide to respond, and how to identify the recall processes participants used to respond. Studies drawing on larger samples could then test the factor structure of ERNS. Despite the additional work required to validate ERNS in the SMI population, reactivity to both fear and anger, and numbness to positive emotions correlated significantly with anxiety and depressive symptoms. The sadness subscale was weakly correlated with anxiety, but uncorrelated with depressive symptoms. The relationship of emotional reactivity to depressed mood is complex: People diagnosed with depression may show less reactivity to negative events compared with those in the nondepressed control group; however, Kahn and colleagues (2019) found a curvilinear relationship between facial expressions (of both happiness and sadness) and depressive symptoms, with increased reactivity at lower levels of depression, but diminished reactivity at high levels of depression. Although behavioral reactions to sad cues may significantly vary at different levels of depressive symptoms, in the present study self-reported reactivity to sadness cues was uncorrelated with depressive symptoms, and a curvilinear relationship was not indicated from the inspection of augmented component plus residual plots. Although the present study assesses the interaction of gender and emotional response as they relate to anxiety and depressive symptoms, roughly 40% of participants were diagnosed with schizophrenia spectrum disorders. It is important to note that, although depressive symptoms were significantly lower for people diagnosed with schizophrenia spectrum disorders, mean anxiety scores did not vary significantly by primary diagnosis. We did not restrict the sample to those with GAD or MDD only, as these disorders frequently co-occur with other mental illnesses, and the present sample may be comparable in the diversity of primary diagnoses to those in other community mental health centers. Consistent with previous literature, women reported stronger anxiety and depressive symptoms compared with men. The main effects for gender remained significant only for the regression model predicting anxiety symptoms, but not depressive symptoms. In addition, women reported greater reactivity to fear and sadness compared with men. This finding is consistent with Orsillo and colleagues’ (2007) study in which women also reported greater reactivity to fear [t(77) = –2.95, p = .002] and to sadness [t(77) = –2.34, p = .011]. Many possibilities exist for the differences in emotional response found here: Echoing Kimerling, Allen, and Duncan (2018), we understand the observed gender differences in emotional response, depressive, and anxiety symptoms as an interaction of biological and social factors. Despite the significant bivariate association between gender and depressive symptoms, the interaction of gender with emotional response variables did not significantly predict depressive symptoms when controlling for primary diagnosis and the main effects of emotional response. It is not surprising that people with MDD, bipolar disorder, and other diagnoses (primarily PTSD) had significantly higher depressive symptoms. Lower reactivity to positive emotions was the strongest predictor of depressive symptoms, followed by greater reactivity to fear. The emotional context insensitivity model—which posits that people experiencing depression will display reduced reactivity to both positive and negative cues—is supported by laboratory experiments examining behavioral responses to emotional cues (Bylsma et al., 2008); however, our results are inconsistent with this model. Rather than claiming that our results counter the ECI model, it is helpful to describe how the results of the present study differ from the results of Bylsma and colleagues’ (2008) meta-analysis. First, the present study takes place in a natural, rather than a laboratory, setting. Respondents were those with SMI, rather than people with MDD compared with controls without MDD. In addition, typical emotional response was only self-reported by participants rather than derived from behavioral observation. Exploring the possible disjunction between perceived emotional reactivity and behavioral measures may advance the theoretical understanding of emotional response in the context of depressive symptoms. In the model predicting anxiety, the main effects for fear, anger, and positive emotional reactivity significantly contributed to the model, along with female gender. Only the relationship between positive emotional reactivity and anxiety symptoms was significantly moderated by gender, with increased positive emotional reactivity among women more strongly covaried with reduced anxiety symptoms compared with men. Within an evolutionary framework, the responses of limited emotional reactivity to positive emotions and increased reactivity to fear and anger could be linked to defensive strategies to cope with aversive or threatening situations (Dixon, 1998; Moskowitz, 2004). Whereas emotional numbness in depression or anxiety could be understood as arrested flight—a defensive strategy to cope with aversive or threatening situations—fear and anger reactivity could relate to flight and fight mechanisms, respectively. Limitations The present study is limited by its cross-sectional nature and relatively small sample size. Given the design, we cannot strongly claim that low reactivity to positive emotions and higher reactivity to anger and fear contributes to mood and anxiety symptoms rather than being a manifestation of these conditions. Although the measures of anxiety and depressive symptoms conformed to Diagnostic and Statistical Manual of Mental Disorders (fifth edition) criteria for major depressive episode and GAD (American Psychiatric Association, 2013), no structured clinical interviews were conducted. Furthermore, other clinician-rated or close informant–rated measures could provide relevant information regarding the validity of the self-reported emotional response in ERNS. In addition, the relationship between emotional response and psychiatric distress may be different for adults without SMI or those diagnosed with SMI who are not receiving intensive case management services. Conclusion Findings from the current study have implications for social work practice in community mental health systems. Understanding clients’ emotional response styles adds important information about the experience of the person served, which informs and may transcend the discrete symptoms of mental illness. Extant literature suggests that the success of any consultation (that is, engagement, assessment, planning, and intervention) depends on how the client and health professional communicate (Gask & Usherwood, 2002; Morrison, 2007). Mental health professionals should maintain awareness of their clients’ emotional response during such interactions (Sheppard, 1993). Furthermore, clinicians working with the SMI population may consider how gender differences in emotional response could require different intervention strategies for promoting emotion regulation to cope with distressing mood and anxiety symptoms. Competence in gender-sensitive social work strategies that address emotional, psychological, social, and behavioral factors is important for social workers to maintain. Acknowledging how people with SMI respond to everyday occurrences has important implications for prompt identification of needs and intervention—such as fostering positive coping strategies or promoting engagement in psychotherapy or pharmacological treatment. Furthermore, researchers may consider the biosocial conditions contributing to differences in emotion regulation and mental health to support improved clinical intervention for mood and anxiety symptoms for people with SMI. References Abel K. M. , Drake R. , Goldstein J. M. ( 2010 ). Sex differences in schizophrenia . International Review of Psychiatry, 22 , 417 – 428 . Google Scholar Crossref Search ADS WorldCat American Psychiatric Association. ( 2013 ). Diagnostic and statistical manual of mental disorders (5th ed.). Arlington, VA : Author . Angelakis S. , Nixon S. ( 2015 ). The comorbidity of PTSD and MDD: Implications for clinical practice and future research . Behaviour Change, 32 ( 1 ), 1 – 25 . doi:10.1017/bec.2014.26 Google Scholar Crossref Search ADS WorldCat Barr L. K. , Kahn J. H. , Schneider W. J. ( 2008 ). Individual differences in emotion expression: Hierarchical structure and relations with psychological distress . Journal of Social and Clinical Psychology, 27 , 1045 – 1077 . doi:10.1521/jscp.2008.27.10.1045 Google Scholar Crossref Search ADS WorldCat Beard C. , Hsu K. J. , Rifkin L. S. , Busch A. B. , Björgvinsson T. ( 2016 ). Validation of the PHQ-9 in a psychiatric sample . Journal of Affective Disorders, 193 , 267 – 273 . Google Scholar Crossref Search ADS WorldCat Bennett D. C. , Kerig P. K. ( 2014 ). Investigating the construct of trauma-related acquired callousness among delinquent youth: Differences in emotion processing . Journal of Traumatic Stress, 27 , 415 – 422 . doi:10.1002/jts.21931 Google Scholar Crossref Search ADS WorldCat Bennett D. C. , Modrowski C. A. , Kerig P. K. , Chaplo S. D. ( 2015 ). Investigating the dissociative subtype of posttraumatic stress disorder in a sample of traumatized detained youth . Psychological Trauma: Theory, Research, Practice, and Policy, 7 , 465 – 472 . doi:10.1037/tra0000057 Google Scholar Crossref Search ADS WorldCat Bylsma L. M. , Morris B. H. , Rottenberg J. ( 2008 ). A meta-analysis of emotional reactivity in major depressive disorder . Clinical Psychology Review, 28 , 676 – 691 . Google Scholar Crossref Search ADS WorldCat Bylsma L. M. , Taylor-Clift A. , Rottenberg J. ( 2011 ). Emotional reactivity to daily events in major and minor depression . Journal of Abnormal Psychology, 120 ( 1 ), 155 – 167 . Google Scholar Crossref Search ADS WorldCat Campbell-Sills L. , Barlow D. H. , Brown T. A. , Hofmann S. G. ( 2006 ). Acceptability and suppression of negative emotion in anxiety and mood disorders . Emotion, 6 , 587 – 595 . doi:10.1037/1528-3542.6.4.587 Google Scholar Crossref Search ADS WorldCat Deng Y. , Chang L. , Yang M. , Huo M. , Zhou R. ( 2016 ). Gender differences in emotional response: Inconsistency between experience and expressivity . PlOS One, 11 ( 6 ), e0158666 . Google Scholar Crossref Search ADS WorldCat Dixon A. K. ( 1998 ). Ethological strategies for defense in animals and humans: Their role in some psychiatric disorders . British Journal of Medical Psychology, 71 , 417 – 445 . Google Scholar Crossref Search ADS WorldCat Gackenbach J. , Darlington M. , Ferguson M. , Boyes A. ( 2013 ). Video game play as nightmare protection: A replication and extension . Dreaming, 23 , 97 – 111 . doi:10.1037/a0032455 Google Scholar Crossref Search ADS WorldCat Gackenbach J. , Ellerman E. , Hall C. ( 2011 ). Video game play as nightmare protection: A preliminary inquiry with military gamers . Dreaming, 21 , 221 – 245 . doi:10.1037/a0024972 Google Scholar Crossref Search ADS WorldCat Gask L. , Usherwood T. ( 2002 ). The consultation . British Medical Journal, 324 , 1567 – 1569 . Google Scholar Crossref Search ADS WorldCat Gross J. J. ( 1998 ). The emerging field of emotion regulation: An integrative review . Review of General Psychology, 2 , 271 – 299 . Google Scholar Crossref Search ADS WorldCat Grossman M. , Wood W. ( 1993 ). Sex differences in intensity of emotional experience: A social role interpretation . Journal of Personality and Social Psychology, 65 , 1010 – 1022 . doi:10.1037//0022-3514.65.5.1010 Google Scholar Crossref Search ADS WorldCat Kahn J. H. , Cox D. W. , Simons K. J. , Hamlet A. N. , Hodge B. J. , Lawell K. J. ( 2019 ). Nonlinear effect of depression symptoms on the time course of emotional reactivity . Motivation and Emotion, 4 , 625 – 635 . Google Scholar Crossref Search ADS WorldCat Karpov B. , Joffe G. , Aaltonen K. , Suvisaari J. , Baryshnikov I. , Näätänen P. , et al. ( 2016 ). Anxiety symptoms in a major mood and schizophrenia spectrum disorders . European Psychiatry, 37 , 1 – 7 . Google Scholar Crossref Search ADS WorldCat Kerig P. K. , Bennett D. C. , Chaplo S. D. , Modrowski C. A. , McGee A. B. ( 2016 ). Numbing of positive, negative, and general emotions: Associations with trauma exposure, posttraumatic stress, and depressive symptoms among justice-involved youth . Journal of Traumatic Stress, 29 , 111 – 119 . doi:10.1002/jts.22087 Google Scholar Crossref Search ADS WorldCat Kerig P. K. , Bennett D. C. , Thompson M. , Becker S. P. ( 2012 ). “Nothing really matters”: Emotional numbing as a link between trauma exposure and callousness in delinquent youth . Journal of Traumatic Stress, 25 , 272 – 279 . doi:10.1002/jts.21700 Google Scholar Crossref Search ADS WorldCat Kerig P. K. , Chaplo S. D. , Bennett D. C. , Modrowski C. A. ( 2016 ). “Harm as harm”: Gang membership, perpetration trauma, and posttraumatic stress symptoms among youth in the juvenile justice system . Criminal Justice and Behavior, 43 , 635 – 652 . doi:10.1177/0093854815607307 Google Scholar Crossref Search ADS WorldCat Kimerling R. , Allen M. C. , Duncan L. E. ( 2018 ). Chromosomes to social contexts: Sex and gender differences in PTSD . Current Psychiatry Reports, 20 ( 12 ), 114 . Google Scholar Crossref Search ADS WorldCat Kroenke K. , Spitzer R. L. , Williams J. B. ( 2001 ). The PHQ‐9: Validity of a brief depression severity measure . Journal of General Internal Medicine, 16 , 606 – 613 . Google Scholar Crossref Search ADS WorldCat McLean C. P. , Asnaani A. , Litz B. T. , Hofmann S. G. ( 2011 ). Gender differences in anxiety disorders: Prevalence, course of illness, comorbidity and burden of illness . Journal of Psychiatric Research, 45 , 1027 – 1035 . Google Scholar Crossref Search ADS WorldCat McMillan K. A. , Enns M. W. , Cox B. J. , Sareen J. ( 2009 ). Comorbidity of Axis I and II mental disorders with schizophrenia and psychotic disorders: Findings from the National Epidemiologic Survey on Alcohol and Related Conditions . Canadian Journal of Psychiatry, 54 , 477 – 486 . Google Scholar Crossref Search ADS WorldCat McRae K. , Ochsner K. N. , Mauss I. B. , Gabrieli J. J. , Gross J. J. ( 2008 ). Gender differences in emotion regulation: An fMRI study of cognitive reappraisal . Group Processes & Intergroup Relations, 11 , 143 – 162 . Google Scholar Crossref Search ADS WorldCat Mennin D. S. , Heimberg R. G. , Turk C. L. , Fresco D. M. ( 2005 ). Preliminary evidence for an emotion dysregulation model of generalized anxiety disorder . Behaviour Research and Therapy, 43 , 1281 – 1310 . doi:10.1016/j.brat.2004.08.008 Google Scholar Crossref Search ADS WorldCat Morrison T. ( 2007 ). Emotional intelligence, emotion and social work: Context, characteristics, complications and contribution . British Journal of Social Work, 37 , 245 – 263 . Google Scholar Crossref Search ADS WorldCat Moskowitz A. K. ( 2004 ). “Scared stiff”: Catatonia as an evolutionary-based fear response . Psychological Review, 111 , 984 – 1002 . Google Scholar Crossref Search ADS WorldCat Nivoli A. M. , Pacchiarotti I. , Rosa A. R. , Popovic D. , Murru A. , Valenti M. , et al. ( 2011 ). Gender differences in a cohort study of 604 bipolar patients: The role of predominant polarity . Journal of Affective Disorders, 133 , 443 – 449 . Google Scholar Crossref Search ADS WorldCat Orsillo S. M. , Theodore-Oklota C. , Luterek J. A. , Plumb J. ( 2007 ). The development and psychometric evaluation of the Emotional Reactivity and Numbing Scale . Journal of Nervous and Mental Disease, 195 , 830 – 836 . doi:10.1097/NMD.0b013e318156816f Google Scholar Crossref Search ADS WorldCat Pos A. E. , Greenberg L. S. , Goldman R. N. , Korman L. M. ( 2003 ). Emotional processing during experiential treatment of depression . Journal of Consulting and Clinical Psychology, 71 , 1007 – 1016 . Google Scholar Crossref Search ADS WorldCat Rauch S.A.M. , Foa E. B. ( 2006 ). Emotional processing theory (EPT) and exposure therapy for PTSD . Journal of Contemporary Psychotherapy, 36 , 61 – 65 . doi:10.1007/s10879-006-9008-y Google Scholar Crossref Search ADS WorldCat Sala R. , Goldstein B. I. , Morcillo C. , Liu S. M. , Castellanos M. , Blanco C. ( 2012 ). Course of comorbid anxiety disorders among adults with bipolar disorder in the US population . Journal of Psychiatric Research, 46 , 865 – 872 . Google Scholar Crossref Search ADS WorldCat Salovey P. , Stroud L. R. , Woolery A. , Epel E. S. ( 2002 ). Perceived emotional intelligence, stress reactivity, and symptom reports: Further explorations using the Trait Meta-Mood Scale . Psychology and Health, 17 , 611 – 627 . Google Scholar Crossref Search ADS WorldCat Sheppard M. ( 1993 ). Client satisfaction, extended intervention and interpersonal skills in community mental health . Journal of Advanced Nursing, 18 , 246 – 259 . doi:10.1046/j.1365-2648.1993.18020246.x Google Scholar Crossref Search ADS WorldCat Spitzer R. L. , Kroenke K. , Williams J. B. , Löwe B. ( 2006 ). A brief measure for assessing generalized anxiety disorder: The GAD-7 . Archives of Internal Medicine, 166 , 1092 – 1097 . Google Scholar Crossref Search ADS WorldCat Temmingh H. , Stein D. J. ( 2015 ). Anxiety in patients with schizophrenia: Epidemiology and management . CNS Drugs, 29 , 819 – 832 . Google Scholar Crossref Search ADS WorldCat Turk C. L. , Heimberg R. G. , Luterek J. A. , Mennin D. S. , Fresco D. M. ( 2005 ). Emotion dysregulation in generalized anxiety disorder: A comparison with social anxiety disorder . Cognitive Therapy and Research, 29 ( 1 ), 89 – 106 . doi:10.1007/s10608-005-1651-1 Google Scholar Crossref Search ADS WorldCat Upthegrove R. , Marwaha S. , Birchwood M. ( 2017 ). Depression and schizophrenia: Cause, consequence, or trans-diagnostic issue? Schizophrenia Bulletin, 43 , 240 – 244 . Google Scholar OpenURL Placeholder Text WorldCat Vesga-López O. , Schneier F. , Wang S. , Heimberg R. , Liu S. M. , Hasin D. S. , Blanco C. ( 2008 ). Gender differences in generalized anxiety disorder: Results from the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC) . Journal of Clinical Psychiatry, 69 , 1606 – 1616 . Google Scholar Crossref Search ADS WorldCat © 2020 National Association of Social Workers This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)