Agency and Resilience Along the Arizona-Sonora Border: How Unauthorized Migrants Become Aware of and Resist Contemporary U.S. Nativist Mobilization

Agency and Resilience Along the Arizona-Sonora Border: How Unauthorized Migrants Become Aware of... Abstract Little is known about the extra-political consequences of contemporary U.S.-based nativist mobilization as well as the resilience unauthorized migrants display in the face of anti-immigrant mobilization along the U.S.-Mexico border. Bringing together social movements and immigration literatures, we examine these interrelated issues using original survey data from the first wave of the Migrant Border Crossing Study. In so doing, we examine: (1) factors influencing repatriated unauthorized migrants’ awareness of nativist mobilization (i.e., Minutemen) along the Arizona-Sonora border, and (2) factors explaining why some migrants would or would not be potentially deterred from attempting future unauthorized crossings if encountering the Minutemen were a possibility. Results from a Heckman probit selection model indicate that higher levels of general, financial, and migration-specific human capital are associated with awareness of the Minutemen, while higher household income and status as an indigenous language speaker predict who would be less likely to be deterred from crossing. We also uncover an interesting paradox: migrants traveling with coyotes were less likely to have heard of Minutemen and more likely to be potentially deterred. Collectively, our results provide insight into the overlooked extra-political consequences of contemporary U.S. nativist mobilization, how resiliency in the face of such a deterrent is structured among repatriated unauthorized migrants, and how seemingly powerless migrant groups can mitigate potential threats initiated by relatively privileged groups of U.S. citizens. immigration, Minutemen, border crossers, social movements, unauthorized migration Debates over immigration reform have waxed and waned but never fallen from center stage in the United States. Passage of restrictionist legislation coupled with deliberation over Senate Bill 744, President Obama’s 2012 and 2014 Executive Orders of Deferred Action for Childhood Arrivals (DACA) and Deferred Action Program for Parents (DAPA), increasing electoral participation of Latinos, and the recent rise to prominence of nativist political candidates signal as much. However, given the focus on policy reform and institutional change, it becomes easy to overlook grassroots mobilization along the U.S.-Mexico border and two key protagonists helping to sustain this drama—unauthorized migrants and mobilized nativists called “Minutemen.” Two shifts in population trends among (un)authorized immigrants set the stage for this drama. First is the growth of the unauthorized immigrant population since 1990, which currently stands around 11 million (Passel and Cohn 2016). Scholars, ironically, attributed this growth to increased border militarization (Cornelius and Lewis 2007; Massey, Durand, and Malone 2002). Second, (un)authorized immigrant populations moved into “new destination” states (Massey and Capoferro 2008), which became ground zero for the passage of anti-immigrant legislation. And in both “new” and “traditional destination” states, residents mobilized against foreign population growth. Recent research has focused on the rise of anti-immigrant mobilization, in particular self-described Minutemen organizations that have moved beyond merely advocating for restrictionist legislation and are known largely for their controversial and visible patrol efforts along the U.S.-Mexico border, as well as their interior enforcement efforts targeting migrants in the workplace (e.g., Doty 2009; Shapira 2013; Stewart, Bendall, and Morgan 2015; Ward 2014, 2016). While these groups are organizationally diverse and their goals multifaceted (Doty 2009; Ward 2016), generally-speaking, they seek to both raise awareness among the public and policy makers about unauthorized migration and the need for increased border enforcement while also leveraging the media spectacles their presence along the border and key interior spaces generates to try and deter migrants from crossing in the first place (Chavez 2008; Doty 2009; Gilchrist and Corsi 2006; Shapira 2013; Ward 2016). Although exceedingly rare, Minutemen have on occasion apprehended unauthorized border crossers and turned them over to U.S. authorities. Thus, Minutemen potentially deter unauthorized crossings by creating a physical presence along the border but also—and primarily—by leveraging the media spectacle of this presence and the uncertainty attached to it regarding potential physical harm that “rogue” nativist border-vigilantes may exercise against unauthorized immigrants. Extant literature on Minutemen contributes to our understanding of right-wing exclusionary mobilization. However, a key perspective remains absent: that of the movement’s targeted opponents—unauthorized migrants. Taking this silence as a starting point, we first ask: What factors help predict whether or not a repatriated unauthorized migrant knows of the Minutemen?1 Such migrants do not have equal access to information about the crossing experience. Focusing attention on informational disparities advances sociological understandings of the social process of border crossing. Additionally, it allows us to assess the nativist movement’s broader influence on the population they sought to deter. Our second research question builds on the first. We ask, what factors explain migrants’ perceptions of their own susceptibility to being deterred from attempting another crossing if encountering Minutemen were a possibility?2 Much attention has been paid to the ways a militarized border (Andreas 2009; Cornelius and Lewis 2007; Dunn 1996; Nevins 2010), labor markets (Cornelius 1998; Cornelius etal. 2010), and crossing conditions (Hagan 2008; Martínez 2016; O’Leary 2009) affect unauthorized crossings, but little is known about how a nation’s mobilized nativist citizenry influences unauthorized repatriated migrants’ perceptions of the viability of clandestine crossings. This provides an additional measure of nativist influence on the movement’s targeted opponents. We uncover findings using our original data set of 415 in-depth surveys with recently repatriated migrants that attempted crossing one of the most heavily traversed regions along the U.S.-Mexico border among unauthorized Mexican migrants—southern Arizona. Using a Heckman probit selection model, which allows us to examine the conditional probability of potential deterrence given prior knowledge of the Minutemen, we found unauthorized repatriated migrants’ awareness of Minutemen was positively associated with being male, age, education, household income, being from northern or central Mexico (relative to west-central Mexico), number of lifetime apprehensions, and having lived in the United States. The use of a coyote (guide), marginalized status, and crossing for the first time were negatively associated with awareness. Second, respondents with higher monthly household income, those who spent more time crossing the border, those with higher cumulative lifetime apprehensions, and those who believed they would cross again were less susceptible to the potential deterrence effect embodied in a mobilized nativist citizenry. Respondents travelling with coyotes were more susceptible to deterrence stemming from a potential Minutemen encounter. Nevertheless, nativist mobilization along the U.S.-Mexico border appears to fail to deter many resilient migrants. Our research bridges literatures on the social processes of migration and social movement consequences. First, we contribute to migration literature by describing the heterogeneity of repatriated unauthorized migrants, a group often erroneously regarded as homogenous (De Genova 2002). We offer a rare survey-based, quantitative analysis focusing on an overlooked dimension of the resource disparities existing among such migrants—knowledge, specifically, knowledge about non-infrastructural, non-environmental deterrents like a xenophobic, mobilized citizenry. In so doing, we reveal how unequal access to information about potential crossing hurdles is structured. Second, we contribute at the intersection of migration and social movements scholarship by offering novel empirical measures of nativist movement influence. Examining two interrelated issues—(1) factors influencing migrants’ awareness of anti-immigrant activism and (2) factors associated with migrants’ perceptions of their own susceptibility to being deterred from future clandestine crossing attempts—provides insight into the extra-political consequences of nativist mobilization. Moreover, by studying nativists’ targets (i.e., unauthorized migrants), our findings reveal how these opponents—often thought of as passive and powerless—exert agency and resist nativist efforts. NATIVIST MOBILIZATION AND ITS CONSEQUENCES Do social movements matter? If so, how? While these remain central issues for social movement scholars, little consensus exists (Amenta etal. 2010). On the one hand, social movements are touted as influential and responsible for important political changes (Baumgartner and Mahoney 2005; Piven 2006). On the other hand, movements often fail to shape social and political landscapes (Giugni 2007; Skocpol 2003). Surveying literature on the expressly political consequences of social movements, Edwin Amenta and associates (2010) concluded that most research on larger movements demonstrates that movements matter. Research on social movement influence, however, focuses overwhelmingly on political consequences. Nevertheless, while movements are frequently geared towards effecting political change, they may also have consequences beyond the political sphere (Earl 2004). Countercultural and self-help movements, for instance, focus largely on expressive action rather than political change (Earl 2004; Snow, Soule, and Kriesi 2004). Moreover, “movement influence” need not be limited to outcomes at the societal level (e.g., policy change). Activism influences participants too, in the short- and long-term (McAdam 1989; Sherkat and Blocker 1997). Yet, despite the significance of understanding the effects of social movement mobilization both within and beyond the political realm, research on nativist mobilization, in particular, largely ignores the extent to which it might influence local, state, and national politics.3 Additionally, the movement’s extra-political consequences remain understudied. Instead, research trends towards providing a fine-grained description of the movement’s origins (Ward 2014, 2016), members’ attitudes (Cabrera and Glavac 2010), movement activities (Elcioglu 2015; Shapira 2013), and cultural frames used to construct anti-immigrant rationalizations (Dove 2010). Sang Kil, Cecilia Menjivar, and Roxanne Doty (2009), for instance, examined how Minutemen targeted employers, protested against officials, and raised awareness. April Dove (2010) explained how Minutemen framed websites to generate collective action resonance. Doty (2009) brought us into the nativist world through interviews and observation. Leo Chavez (2008) argued the inaugural minuteman border campaign served as a media spectacle reaffirming citizenship privileges and fanning the flames of restrictionist legislation initiatives. Emine Elcioglu’s (2015) fieldwork examined how Minutemen challenged and reinforced state authority by enacting nativism through popular sovereignty. Finally, Harel Shapira’s (2013) ethnography shed light on processes of acculturation involved in becoming a minuteman. Despite the significance of these studies, nativist mobilization’s extra-political consequences remain under-examined. In what ways, if any, has nativist mobilization mattered beyond the political sphere? We tackle this issue by linking it to our analyses of repatriated unauthorized migrants’ awareness of nativist mobilization, as well as their perceived susceptibility to such mobilization. In so doing, we extend research on social movement consequences in two respects. First, our questions represent novel conceptualizations of potential nativist movement influence. The contemporary nativist movement seeks to raise awareness about unauthorized migration among both the public and policy makers (Dove 2010; Kil etal. 2009). Additionally, and largely through Minutemen organizations and the media spectacles their surveillance efforts create (Chavez 2008), the movement also seeks to deter migrants from attempting unauthorized crossings in the first place (Doty 2009; Gilchrist and Corsi 2006; Shapira 2013; Ward 2016). Given this latter goal, we ask: how much awareness has been raised among our target migrant population and how is this awareness structured? Second, given Minutemen’s exclusionary rhetoric and border patrol activity, would these migrants potentially be deterred from crossing again if encountering Minutemen were a possibility? Relatedly, how are migrants’ perceptions of their own potential susceptibility shaped by a variety of individual characteristics and experiential factors related to their most recent crossing attempt? Answers provide insight into the under examined extra-political consequences of nativist mobilization. Second, examining potential deterrence reveals how relatively powerless groups like unauthorized migrants can resist mobilization initiated by relatively privileged groups. Examining how migrants’ resiliency is constructed in the face of nativist mobilization emphasizes their agency, which although highlighted by migration scholars, has been largely neglected by the general public and policy makers. Focusing on migrants’ resiliency also provides insight into forces stymying nativist influence. By presenting novel ways to think about how movements matter outside of regular politics, our analysis examines diverse ways that contemporary nativism has and has not exerted influence. NATIVIST MOBILIZATION AT THE U.S.-MEXICO BORDER: AN OVERVIEW Nativist mobilization along the U.S.-Mexico border is not new. In the mid-1800s, vigilante and state-sanctioned groups roamed the borderlands (Spener 2009). In the 1970s and 80s, members of the Ku Klux Klan operated the Klan Border Watch program. Yet, it was not until after the U.S. government’s adoption of its “prevention through deterrence” strategy and the Operation Gatekeeper complex (Andreas 2009; Nevins 2010) that growing numbers of residents began patrolling (Doty 2007). Over the next two decades, the Border Solution Task Force, U.S. Citizen Patrol, Voices of Citizens Together, American Border Patrol, Ranch Rescue, and Civil Homeland Defense emerged. Recently, the Minuteman Project (MMP) and the Minuteman Civil Defense Corps (MCDC) ushered in a new era of media-savvy and better-funded anti-immigrant vigilantism. Of course, these organizations were part of a broader milieu of nativist hostilities unauthorized immigrants faced during this period, particularly in Arizona. The state legislature and governor considered numerous anti-immigrant laws, and the Border Patrol engaged in an aggressive media campaign highlighting enforcement efforts and the dangers of unauthorized immigration. In 2005, Jim Gilchrist and Chris Simcox implored Americans to participate in a month-long “muster” to observe and report unauthorized immigrants entering along the Arizona-Sonora border. These Minutemen sought to raise awareness among the public and policy makers about unauthorized immigration, while also instilling a sense of fear among migrants by signaling to them that unauthorized crossing attempts would be met with force. Soon thereafter, Minutemen focused their attention throughout the United States.4 At the time our survey was conducted (2007-2009), Minutemen were experiencing explosive growth, jumping from 173 chapters in 2008 to 309 just a year later (Beirich 2011). Twenty-one chapters operated in Arizona (second only to California) (SPLC 2010), and most of the high profile musters occurred in southern Arizona (Chavez 2008; Shapira 2013). Nationally, some of the growth during this period was focused around increased interior enforcement. Minutemen began targeting migrants in the workplace and their employers. This also included increased lobbying for reform of local and state immigration policies. By the mid-to-late 2000s, the nativist agenda was normalized as various localities considered or passed restrictive housing ordinances and Arizona’s SB 1070 and Alabama’s HB 56 were debated. However, despite that some chapter growth occurred in communities away from the border, Shapira (2013) found that many individuals involved in border patrol were travelling from such communities into southern Arizona. Thus, despite the growth of interior enforcement during this period, patrol activity still continued at the Arizona-Sonora border well into 2008 and 2009 (MCDC 2009; Shapira 2013), albeit in a reduced capacity relative to 2005-2007. By 2011, nativist mobilization was declining. The Tea Party offered a legitimate space in which anti-immigrant sentiment was stoked behind a veil of traditional conservatism and channeled into political gain for the Right (Skocpol and Williamson 2013). Infighting and internal splintering among major nativist organizations, along with the highly publicized arrests of nativist leaders on murder and drug charges exacerbated the downward spiral. And yet, nativist mobilization has not disappeared. In response to the recent transportation of unauthorized migrant children from Texas to makeshift emergency shelters in Arizona and California, MMP and others launched Operation Normandy in 2014 (Minuteman Project 2014). Finally, “mainstream” nativist decline has also triggered the growth of smaller, radical splinter cell organizations (Neiwert 2013). EXPLAINING AWARENESS AND POTENTIAL DETERRENCE Despite the growth of anti-immigrant mobilization, little is known about how unauthorized migrants’ differential awareness of this potential deterrent is structured. Moreover, by influencing perceptions about the viability of clandestine crossings, it is plausible that a mobilized nativist citizenry represents another roadblock to migrants’ passage. However, only some migrants will be deterred. To what extent do economic and political forces known to significantly structure migration overshadow the influence of nativist mobilization? Below we develop theoretical arguments and present hypotheses to explain these two extra-political consequences of nativist mobilization. Gender Unauthorized migration from Mexico to the United States is gendered (O’Leary 2009). Males are more likely to migrate (Cerrutti and Massey 2001). When they do migrate, women typically migrate for family reunification (Cerrutti and Massey 2001), though not exclusively so. Despite this, unauthorized border crossers continue to be largely male. According to FY 2013 U.S. Border Patrol apprehensions, females constituted 16.5 percent of apprehensions (USBP 2014a). Males also have greater first-hand migration experience. A recent study found that repatriated Mexican migrant women have, on average, 2.9 lifetime unauthorized crossing attempts and 1.7 apprehensions compared to 5.3 crossings and 3.2 apprehensions among men (p < .05) (Slack etal. 2013). Because first-hand migration experience contributes to migrants’ knowledge stores, it shapes expectations for the unauthorized crossing. And these experiences likely shape perceptions of potential hurdles, including anti-immigrant mobilization efforts. Considering men’s greater migration experience, as well as the gendering of unauthorized Mexican migration, we suggest: (H1) Migrant men are more likely than migrant women to have heard of the Minutemen. (H2) Migrant men are less likely to be deterred by a potential encounter relative to migrant women. General Human and Financial Capital General human and financial capital vary among migrants. Greater community-level economic resources play important roles in determining the probability of first-time and repeat migration (Massey and Espinoza 1997). These resources likely impact migrants’ knowledge of and susceptibility to anti-immigrant mobilization. Although some unauthorized migrants received their information about Minutemen through friends and family members in the United States and Mexico (a form of social capital), as well as during the migration process, most got their information from media outlets in the United States, and to a lesser extent, in Mexico5 (Ward and Martínez 2015). Higher levels of household income allow migrants to consume more diverse media. Literacy, operationalized as higher levels of formal educational attainment, may also increase migrants’ access to information about unauthorized crossings, including the possibility of encountering Minutemen. Thus, opportunities to consume a variety of media should increase a migrant’s likelihood of having heard of Minutemen. After all, media visibility was a focus of Minutemen mobilization. Unauthorized migrants also differ with respect to their marginalized status within Mexico. Approximately 9.8 percent of Mexico’s population is indigenous (CDI 2006). This population experiences higher levels of socioeconomic and ethno-racial marginalization and contends with social barriers that disproportionately negatively impact access to border-crossing information (see Martínez, Vandervoet, and Slack 2013). We, thus, hypothesize: (H3) Migrants with greater household income and higher levels of formal education will be more likely to have heard of the Minutemen, whereas indigenous language speakers will be less likely to have heard. Migrants also consider the costs and risks of unauthorized crossings. Given the risks—such as exposure to the elements and apprehension by Border Patrol—many attempts fail. Failure quickly compounds costs. Higher household income should mitigate costs incurred from multiple crossing attempts, thus we expect: (H4) Migrants with greater household income view Minutemen—and the potential for subsequent apprehension and deportation—as less of a risk and, thus, less of a potential deterrent. Region of Origin Mexico’s Consejo Nacional de Población (CONAPO) identified four major migrant sending regions within Mexico: northern, west-central (traditional), central, and southern-southeastern. Constant migration from the west-central region since the U.S. Bracero Program was enacted (1942-1964) has fostered a culture of migration in these communities. Migration is viewed as an important rite of passage, and there are social expectations that young men will migrate (Cornelius and Lewis 2007; Kandel and Massey 2002). A “culture of migration,” and the maturity and density of social networks involved in the migration process from these communities, likely increases awareness of nativist mobilization. Thus: (H5) Migrants from west-central (traditional) areas are more likely to have heard of recent Minutemen efforts when compared to those from other regions of Mexico. The south-southeastern and central regions of Mexico have more recent histories of U.S. migration (Marcelli and Cornelius 2001). Yet, migration is occurring at levels comparable to more established migrant-sending communities (Martell, Pineda, and Tapia 2007). These individuals are more likely to come from indigenous, marginalized communities (CDI 2006). These regions’ relatively recent history of migration, higher levels of poverty and marginalization, and greater distance from the U.S.-Mexico border often mean lower levels of general and migration-specific social and human capital, thus making the journey riskier. Migrants from these regions likely have less knowledge of the hurdles involved in the crossing experience, including anti-immigrant groups. Thus, we hypothesize: (H6) Migrants from southern and central Mexico have less knowledge of anti-immigrant groups relative to those from west-central areas. Migration-Specific Social and Human Capital Migration-specific social capital—operationalized as ties to family members in the United States—has consistently shown to be important in the migration process. These ties increase the odds of making a first trip with or without documents, shape modes of crossing, and assist in securing employment (Martínez 2016; Massey and Espinoza 1997; Singer and Massey 1998). Additionally, migration-specific human capital—operationalized as first-hand migration experience—influences migrants’ modes of crossing (Martínez 2016; Singer and Massey 1998). More first-hand experience also decreases the probability of apprehension and spurs repeat migration (Massey and Espinoza 1997). Lived experience abroad also seems to matter. Place attachment, particularly the subjective understanding of where one’s current home is located, plays an important role in future migration intentions. For instance, recently repatriated Mexican migrants who consider their home to be located in the United States report higher rates of future migration intentions relative to those whose homes are in Mexico (Slack etal. 2015). In a similar vein, a higher degree of affectual ties to the United States—operationalized as having one's home in the United States and having lived in the country for a decade or longer—is a significant and strong predictor of future crossing intentions, even after controlling for conventional measures of economic, human, and migration-specific social capital as well as the specific punitive immigration enforcement programs through which unauthorized immigrants are processed (Martínez, Slack, and Martínez-Schuldt n.d.). Overall, migration-specific social and human capital not only shape migration attempts but also increase access to key information about the unauthorized migration process, either through social ties, lived experience, or place attachment. We, therefore, posit the following hypotheses: (H7) Migrants with greater migration-specific social and human capital are more likely to have heard of the Minutemen. (H8) Migrants who consider their home to be located in the United States have a stronger resolve to migrate despite the possibility of encountering Minutemen. Unauthorized crossings are risky, especially for first timers. The possibility of encountering an anti-immigrant group may be overwhelming for individuals that have never lived or worked in the United States and never attempted a prior crossing. Because more apprehensions bring greater awareness of the potential consequences of being caught by Minutemen and turned over to immigration officials, we hypothesize: (H9) First-time crossers are more likely to be potentially deterred, while a migrant's number of lifetime apprehensions is negatively associated with this outcome. Context of the Last Crossing Experience No two unauthorized crossing experiences are identical. Depending upon where a migrant started the trip, the corridor being traversed, and time of year, trip duration can range from a few hours to two weeks. The majority of unauthorized migrants employ coyotes (human smugglers) to cross the border (Martínez 2016; Singer and Massey 1998; Spener 2009). Migrants rely on coyotes to decrease the physical risks associated with crossing the border and to lower the likelihood of apprehension (Hagan 2008; Spener 2009). Coyotes instill a sense of security because they are subject-matter experts. The foundations of social capital—reciprocity, trust, value introjection, and bounded solidarity—help mitigate against mistreatment and often lead to a mutually beneficial relationship between coyotes and migrants (Portes 1998; Spener 2009). However, this may be changing with the emergence of “border business” coyotes in response to border enforcement initiatives that are funneling unauthorized migration and drug trafficking into the same geographic spaces (see Martínez 2016; Spener 2009). Coyotes can also be thought of as knowledge creators and diffusers. Not only do they know the terrain and most effective crossing strategies, but they are also vehicles through which knowledge about these conditions is transmitted. This includes knowledge about potential hurdles, such as nativist mobilization. By constructing and transmitting more proximate knowledge about crossing hurdles, coyotes shape migrants’ risk perceptions. One of the primary ways in which migrants become aware of Minutemen is during the migration process itself (Ward and Martínez 2015). Thus, we hypothesize: (H10) Migrants who travel with coyotes are more likely to have heard of the Minutemen. (H11) Because migrants use coyotes to reduce social and physical risks associated with unauthorized migration attempts, travelling with a coyote decreases migrants’ susceptibility to being deterred by potentially encountering Minutemen. DATA AND METHODS Using original survey data from the first wave of our Migrant Border Crossing Study (MBCS; N = 415), we examine unauthorized repatriated migrants’ awareness of Minutemen and the extent to which a potential encounter with Minutemen may deter future crossings. The aim of our survey was to better grasp the social processes shaping unauthorized repatriated migrants’ border-crossing experiences through southern Arizona (Martínez etal. 2017). Although our survey approach cannot capture the full spectrum of experiences individuals accumulate during their crossings, it does gauge what respondents knew about Minutemen at that time following their most recent border crossing attempt. Numerous insightful qualitative studies have been conducted on migrants’ crossing experiences; however, there has been less effort to leverage the advantages of a more quantitative, survey approach, especially with respect to the questions addressed in this article. Doing so, however, allows us to generate more precise descriptions of a population often erroneously regarded as homogenous (De Genova 2002) and also allows us to more confidently generalize our results. Our population frame consists of recently repatriated migrants who crossed through and were apprehended within the Tucson Sector, a location that was, at the time of the survey period (2007-2009), a hotbed of nativist mobilization (Chavez 2008; Doty 2009; Shapira 2013; Ward 2014). Results, therefore, generalize to migrants with a crossing, apprehension, and repatriation experience within this sector.6 In an era of increased border and immigration enforcement, the apprehension-repatriation experience is becoming much more commonplace. Unauthorized migrants have never been at higher risk for apprehension and repatriation (Rosenblum 2013). Thus, the typical MBCS respondent is likely more representative than the hypothetical, perfectly evasive migrant never detected by U.S. authorities. During our survey period (2007-2009), 937,608 migrants were apprehended in the Tucson sector (USBP 2014b) and 565,278 Mexican migrants were repatriated to the state of Sonora (CEMLA 2013). Because the Tucson sector accounted for roughly 45 percent of all apprehensions, and Sonora accounted for 33 percent of repatriations to Mexico during this time period, northern Sonora represents an ideal research location (CEMLA 2013; USBP 2014b). We completed surveys between October 2007 and July 2009 in a migrant shelter in Nogales, Sonora. This was an ideal research site because 78 percent of migrants repatriated to Sonora between 2007 and 2009 were repatriated to the city of Nogales (CEMLA 2013). Furthermore, the overwhelming majority of migrants were transported from repatriation sites to this particular shelter—the only one in the city that provided lodging for up to three days—by Mexican federal agents. Once migrants arrived at the shelter for the evening they could not enter and exit the shelter throughout the night due to safety concerns; therefore, most did not typically depart until the following morning. People with local connections or those originally from Sonora might opt to not stay at the shelter because they can presumably go home and/or stay with friends and family. The shelter also provided a safe space for researchers and respondents during a time when Nogales, Sonora, experienced high levels of drug trafficking-related violence (Martínez, etal. 2013; Slack, Martínez, and Vandervoet 2011). Eligible participants were at least 18 years of age and must have attempted an unauthorized crossing along the Arizona-Sonora border, been apprehended by any U.S. authority, and repatriated to Mexico within the past six months at the time of the survey. Ensuring that respondents’ most recent crossing attempt occurred within a comparable geographical space and time allows for comparability between cases. The six-month cutoff reduced retrospective bias. MBCS survey methodology is extensively detailed elsewhere (Martínez etal. 2017). Sampling Potential MBCS participants were randomly selected using a spatial random sampling technique. We divided the shelter into five previously defined spaces and randomly selected every nth person in each area. We opted not to randomly sample from the shelter’s daily sign-in sheet to protect respondents’ anonymity. Instead, we discretely approached potential respondents, introduced ourselves, and invited them to speak with us. We screened those who had been randomly selected for eligibility and invited them to complete the survey. The response rate was 97 percent (Martínez etal. 2017). There was little to occupy migrants’ time while at the shelter (e.g., no televisions or radios). And because shelter guests could not depart until the following morning, many welcomed the opportunity to discuss their experiences, which helps account for the study’s high response rate (Martínez etal. 2017). We applied probability weights calculated from monthly Border Patrol apprehension statistics between 2007 and 2009 to minimize selection bias between shelter goers and people who did not stay in a shelter upon repatriation. Weights were constructed to correct for any discrepancies between the proportion of males and females in the population and our sample, as well as any discrepancies between the proportion of individuals from each of the four Mexican sending regions in the population and our sample. These weights are based on Tucson sector apprehensions for the month each respondent was interviewed during the study time period. Probability weights are used when calculating our descriptive statistics. However, we did not implement them when conducting the inferential analyses (Winship and Radbill 1994). In situations (like ours) where sampling weights are solely a function of independent variables included in the models (and not the dependent variable), Christopher Winship and Larry Radbill recommend using unweighted estimates because they are unbiased, consistent, and produce smaller standard errors.7 The generalizability of the sample and comparisons of respondents’ demographic characteristics to estimates of the study population are discussed elsewhere (Martínez etal. 2017). Demographic characteristics of our sample are highly consistent with findings from similar data sources (for instance, see data from EMIF-Norte, Instituto Nacional de Migración, and U.S. Border Patrol). Minutemen Module Wave I of the MBCS consisted of 12 modules and over 350 questions focused around a range of crossing-related topics, such as: migrants’ most recent border-crossing experiences, experiences while in U.S. custody, and demographic characteristics. We also included a module on migrants’ perceptions of and potential reactions to Minutemen. This module was not included in the second wave of the MBCS, and to our knowledge, no survey of unauthorized repatriated migrants to date has explored these topics. Therefore, the first wave of the MBCS is the best data source for our purposes. We began by asking migrants whether or not they had ever heard of Minutemen. We then debriefed all respondents by reading a standardized script informing them that Minutemen were residents in the United States voluntarily mobilizing to stop unauthorized migrants from crossing. Essentially, all respondents were given the same information before being asked this study’s second outcome of interest (“Would the possibility of encountering the Minutemen deter you from attempting a future unauthorized crossing?”). As such, our second question focuses on a potential—albeit hypothetical—encounter with Minutemen, and not actual encounters, which our data indicate are rare events.8 Instead, we are interested in repatriated migrants perceptions of their own susceptibility to deterrence were such an encounter a possibility. Nevertheless, because people who had previously heard of the group are likely qualitatively distinct from those who had never heard, and because reading a standardized script while being surveyed does not adequately substitute for previously acquired first-hand knowledge of the Minutemen, our analysis employs a Heckman probit model correcting for sample selection associated with having previously heard of the group. This technique helps identify factors predicting knowledge of the group while also correcting for selection bias when examining potential deterrence (Heckman 1976), and ultimately allows us to examine the conditional probability of potential deterrence given “having heard of the Minutemen.” In our case, both the “selection” and “outcome” variables are dichotomous, which warrants the use of a probit model. Missing Data We use multiple imputations (MI) to handle missing observations (Rubin 1987). MI preserves information that listwise deletion would omit. Following John Graham, Allison Olchowski, and Tamika Gilreath (2007), we conducted 20 imputations (m = 20) using the multiple imputation by chained equations (MICE) method (Royston 2009). Each imputation substituted cases with missing information with unbiased plausible values using their predictive distributions in a separate data set. We estimated variable means on each of the imputed data sets, combining the results using “Rubin’s Rules” to yield coefficient estimates and standard errors (Rubin 1987 as cited by Graham etal. 2007). Appendix Table A1 provides descriptive statistics for non-imputed and non-weighted MBCS data as well as the percentage of observations missing for each variable. Most variables contained very little missing data. “Household income” (13 percent) and “home in the U.S.” (17 percent) are exceptions. Both questions were open-ended, which partially explains the relatively higher rate of “missingness.” The monthly household income question may have yielded less missing information had categories been presented. However, we asked for the exact currency amount to be able to conduct non-linear transformations if needed. The home location question was open-ended but also subjective. Some respondents “didn’t know” where they considered home to be, while others refused to answer the question (both instances coded: “missing”). Considering respondents’ immigration status, this is unsurprising. We found similar results using imputed and non-imputed data. However, nearly 100 cases are lost using listwise deletion, which explains minor discrepancies. Because it yields less biased estimates, we use multiple imputations to handle missing observations. RESULTS AND DISCUSSION Descriptive Statistics for Variables Used in Heckman Probit Selection Model Table 1 provides descriptive statistics for independent, selection, and dependent variables. Forty-two percent of the full sample had previously heard of Minutemen, while 43 percent indicated a possible future encounter with the group could potentially deter them from crossing. Table 1. Descriptive Statistics for Dependent and Independent Variables used in the Probit Selection Model (multiply imputed data)    Mean/Percent (full sample)  Mean/Percent (“heard” = 0)  Mean/Percent (“heard” = 1)  Mean/Percent (“deterrence” = 0)  Mean/Percent (“deterrence” = 1)  Dependent (outcome) variable                  Minutemen deter (%)  43  48  35  –  –  Selection variable                  Heard of Minutemen (%)  42  –  –  47  34  General human and financial capital                  Male (%)  87  80  97  93  80   Age (years)                   <25 (%)  24  29  17  23  25    25-34 (%)  41  35  48  40  42    35-44 (%)  24  25  23  28  18    45 + (%)  11  11  12  9  15   Log household income  6.1  5.7  6.6  6.4  5.7   Years of formal education  7.0  6.7  7.4  6.9  7.2   Indigenous language speaker (%)  20  27  10  21  19  Region of origin                  North (%)  14  8  22  15  13   West-Central (“traditional”) (%)  23  25  21  24  23   Central (%)  33  30  36  32  33   South (%)  30  37  20  29  31  Migration-specific social capital                  Family in destination (%)  58  54  63  55  62  Migration-specific human capital                  First crossing (%)  19  29  5  14  25   Apprehensions  4.4  2.8  6.6  5.1  3.4   Have lived in United States (%)  68  54  87  71  63   Current home in United States (%)  17  8  31  21  13  Context of last crossing attempt                  Used a coyote or guide (%)  70  80  57  65  77   Days traveled  2.4  2.5  2.2  2.4  2.5   Perceives crossing as very/ extremely dangerous (%)  81  82  78  80  81  Control                  Possible you will cross again? (%)  51  42  63  59  40  N  404  251  153  236  168     Mean/Percent (full sample)  Mean/Percent (“heard” = 0)  Mean/Percent (“heard” = 1)  Mean/Percent (“deterrence” = 0)  Mean/Percent (“deterrence” = 1)  Dependent (outcome) variable                  Minutemen deter (%)  43  48  35  –  –  Selection variable                  Heard of Minutemen (%)  42  –  –  47  34  General human and financial capital                  Male (%)  87  80  97  93  80   Age (years)                   <25 (%)  24  29  17  23  25    25-34 (%)  41  35  48  40  42    35-44 (%)  24  25  23  28  18    45 + (%)  11  11  12  9  15   Log household income  6.1  5.7  6.6  6.4  5.7   Years of formal education  7.0  6.7  7.4  6.9  7.2   Indigenous language speaker (%)  20  27  10  21  19  Region of origin                  North (%)  14  8  22  15  13   West-Central (“traditional”) (%)  23  25  21  24  23   Central (%)  33  30  36  32  33   South (%)  30  37  20  29  31  Migration-specific social capital                  Family in destination (%)  58  54  63  55  62  Migration-specific human capital                  First crossing (%)  19  29  5  14  25   Apprehensions  4.4  2.8  6.6  5.1  3.4   Have lived in United States (%)  68  54  87  71  63   Current home in United States (%)  17  8  31  21  13  Context of last crossing attempt                  Used a coyote or guide (%)  70  80  57  65  77   Days traveled  2.4  2.5  2.2  2.4  2.5   Perceives crossing as very/ extremely dangerous (%)  81  82  78  80  81  Control                  Possible you will cross again? (%)  51  42  63  59  40  N  404  251  153  236  168  Notes: Cases in which values were missing on the outcome and selection variables were included in the multiple imputation process and then omitted (n = 11) before estimating ”heckprob” procedure in Stata 13; m = 20. Standard errors available upon request. Source: Migrant Border Crossing Survey I, weighted data. Table 1. Descriptive Statistics for Dependent and Independent Variables used in the Probit Selection Model (multiply imputed data)    Mean/Percent (full sample)  Mean/Percent (“heard” = 0)  Mean/Percent (“heard” = 1)  Mean/Percent (“deterrence” = 0)  Mean/Percent (“deterrence” = 1)  Dependent (outcome) variable                  Minutemen deter (%)  43  48  35  –  –  Selection variable                  Heard of Minutemen (%)  42  –  –  47  34  General human and financial capital                  Male (%)  87  80  97  93  80   Age (years)                   <25 (%)  24  29  17  23  25    25-34 (%)  41  35  48  40  42    35-44 (%)  24  25  23  28  18    45 + (%)  11  11  12  9  15   Log household income  6.1  5.7  6.6  6.4  5.7   Years of formal education  7.0  6.7  7.4  6.9  7.2   Indigenous language speaker (%)  20  27  10  21  19  Region of origin                  North (%)  14  8  22  15  13   West-Central (“traditional”) (%)  23  25  21  24  23   Central (%)  33  30  36  32  33   South (%)  30  37  20  29  31  Migration-specific social capital                  Family in destination (%)  58  54  63  55  62  Migration-specific human capital                  First crossing (%)  19  29  5  14  25   Apprehensions  4.4  2.8  6.6  5.1  3.4   Have lived in United States (%)  68  54  87  71  63   Current home in United States (%)  17  8  31  21  13  Context of last crossing attempt                  Used a coyote or guide (%)  70  80  57  65  77   Days traveled  2.4  2.5  2.2  2.4  2.5   Perceives crossing as very/ extremely dangerous (%)  81  82  78  80  81  Control                  Possible you will cross again? (%)  51  42  63  59  40  N  404  251  153  236  168     Mean/Percent (full sample)  Mean/Percent (“heard” = 0)  Mean/Percent (“heard” = 1)  Mean/Percent (“deterrence” = 0)  Mean/Percent (“deterrence” = 1)  Dependent (outcome) variable                  Minutemen deter (%)  43  48  35  –  –  Selection variable                  Heard of Minutemen (%)  42  –  –  47  34  General human and financial capital                  Male (%)  87  80  97  93  80   Age (years)                   <25 (%)  24  29  17  23  25    25-34 (%)  41  35  48  40  42    35-44 (%)  24  25  23  28  18    45 + (%)  11  11  12  9  15   Log household income  6.1  5.7  6.6  6.4  5.7   Years of formal education  7.0  6.7  7.4  6.9  7.2   Indigenous language speaker (%)  20  27  10  21  19  Region of origin                  North (%)  14  8  22  15  13   West-Central (“traditional”) (%)  23  25  21  24  23   Central (%)  33  30  36  32  33   South (%)  30  37  20  29  31  Migration-specific social capital                  Family in destination (%)  58  54  63  55  62  Migration-specific human capital                  First crossing (%)  19  29  5  14  25   Apprehensions  4.4  2.8  6.6  5.1  3.4   Have lived in United States (%)  68  54  87  71  63   Current home in United States (%)  17  8  31  21  13  Context of last crossing attempt                  Used a coyote or guide (%)  70  80  57  65  77   Days traveled  2.4  2.5  2.2  2.4  2.5   Perceives crossing as very/ extremely dangerous (%)  81  82  78  80  81  Control                  Possible you will cross again? (%)  51  42  63  59  40  N  404  251  153  236  168  Notes: Cases in which values were missing on the outcome and selection variables were included in the multiple imputation process and then omitted (n = 11) before estimating ”heckprob” procedure in Stata 13; m = 20. Standard errors available upon request. Source: Migrant Border Crossing Survey I, weighted data. Explanatory variables used in the selection model (“heard of Minutemen”) and outcome model (“potential deterrence”) are arranged in six conceptual groupings. Appendix Table A2 provides survey questions and metrics. The typical case in the inferential analysis is male (87 percent), between the age of 18 and 34 (65 percent combined categories), with seven years of education, reporting about $450 in monthly household income prior to the most recent crossing attempt. Twenty percent of respondents spoke an indigenous language. Fourteen percent were from northern Mexico, 23 percent from west-central Mexico, 33 percent from the central region, and 30 percent from the south. Respondents were relatively experienced border crossers (lifetime mean of 4.4 apprehensions). Roughly 19 percent were first-time crossers. Fifty-eight percent had family in their desired destination and 68 percent had previously lived and/or worked in the United States. Seventeen percent indicated their current home was in the United States, not Mexico. Most (70 percent) used a coyote to cross and, on average, traveled for 2.4 days prior to apprehension. As noted, sample characteristics vary according to prior knowledge of the Minutemen and whether or not respondents said they would be potentially deterred from attempting a future crossing if encountering Minutemen were a possibility (see Table 1, columns 2-5). Inferential Results for Heckman Probit Model with Sample Selection: Explaining “Having Heard” of the Minutemen (Selection Variable) Table 2 provides the coefficients of the Heckman probit model predicting “having heard” and “potential deterrence.”9 The p-value associated with the rho is statistically significant (rejecting rho = 0), suggesting sample selection needed to be corrected. However, for the sake of substantively interpreting the results, in Appendix Table A3 we present the average marginal effects (AMEs) for “having heard” (first column) and the conditional probability of “deterrence” given “having heard” (second column). Figures 1 through 4 provide “marginal effects at representative values” (MERs) for “having heard” by monthly household income, educational attainment, and lifetime apprehensions, and “potential deterrence” for monthly household income (for additional discussion of AMEs and MERs in probit models, see Williams 2012). Table 2. Heckman Probit Selection Model Results for “Heard of the Minutemen” (selection variable) and “Potential Deterrence” (dependent variable)    “Heard of MM”a  “Potential Deterrence”a     (selection)  (outcome)  Variables        General human and financial capital         Male  .64*  –    (.269)      Age (years)          25-34  .38*  –    (.176)       35-44  .22  –    (.201)       45+  .74*  –    (.313)      Log household income  .23***  −.50***    (.068)  (.121)   Years of formal education  .07***  –    (.021)      Indigenous language speaker  −.63***  –    (.195)     Region of origin         North  .64*  –    (.298)      Central  .51*  –    (.214)      South  .26  –    (.193)     Migration-specific social capital         Family in destination  .12  –    (.143)  –  Migration-specific human capital         First crossing  −.67***  .29    (.260)  (.395)   Apprehensions  .03†  −.03†    (.019)  (.017)   Have lived in United States  .48*  –    (.195)      Current home in United States  –  .51       (.306)  Context of last crossing attempt         Used a coyote or guide  −.50**  .58**    (.195)  (.218)   Days traveled  –  −.11†       (.068)   Perceives crossing as very/extremely dangerous  –  −.06       (.228)  Control         Possible that you will cross again?  –  −.51**       (.219)  Uncensored observations / censored observations  251  153  Rho = -.68        LR test of independent equations (rho = 0):         Chi2 (1) = 4.04         P > Chi2 = .0445        N = 404        M = 20 (results presented for 20th imputed data set)           “Heard of MM”a  “Potential Deterrence”a     (selection)  (outcome)  Variables        General human and financial capital         Male  .64*  –    (.269)      Age (years)          25-34  .38*  –    (.176)       35-44  .22  –    (.201)       45+  .74*  –    (.313)      Log household income  .23***  −.50***    (.068)  (.121)   Years of formal education  .07***  –    (.021)      Indigenous language speaker  −.63***  –    (.195)     Region of origin         North  .64*  –    (.298)      Central  .51*  –    (.214)      South  .26  –    (.193)     Migration-specific social capital         Family in destination  .12  –    (.143)  –  Migration-specific human capital         First crossing  −.67***  .29    (.260)  (.395)   Apprehensions  .03†  −.03†    (.019)  (.017)   Have lived in United States  .48*  –    (.195)      Current home in United States  –  .51       (.306)  Context of last crossing attempt         Used a coyote or guide  −.50**  .58**    (.195)  (.218)   Days traveled  –  −.11†       (.068)   Perceives crossing as very/extremely dangerous  –  −.06       (.228)  Control         Possible that you will cross again?  –  −.51**       (.219)  Uncensored observations / censored observations  251  153  Rho = -.68        LR test of independent equations (rho = 0):         Chi2 (1) = 4.04         P > Chi2 = .0445        N = 404        M = 20 (results presented for 20th imputed data set)        Notes: “<25 yrs”; “West-Central” region; and “no family in destination” are the referent groups. Coefficients, significance levels, and standard errors are reported, with the standard errors noted in parentheses. a 1 = yes; 0 = no † p < .10 *p < .05 **p < .01 ***p < .001 (two-tailed tests) Source: Migrant Border Crossing Survey I, unweighted data. Table 2. Heckman Probit Selection Model Results for “Heard of the Minutemen” (selection variable) and “Potential Deterrence” (dependent variable)    “Heard of MM”a  “Potential Deterrence”a     (selection)  (outcome)  Variables        General human and financial capital         Male  .64*  –    (.269)      Age (years)          25-34  .38*  –    (.176)       35-44  .22  –    (.201)       45+  .74*  –    (.313)      Log household income  .23***  −.50***    (.068)  (.121)   Years of formal education  .07***  –    (.021)      Indigenous language speaker  −.63***  –    (.195)     Region of origin         North  .64*  –    (.298)      Central  .51*  –    (.214)      South  .26  –    (.193)     Migration-specific social capital         Family in destination  .12  –    (.143)  –  Migration-specific human capital         First crossing  −.67***  .29    (.260)  (.395)   Apprehensions  .03†  −.03†    (.019)  (.017)   Have lived in United States  .48*  –    (.195)      Current home in United States  –  .51       (.306)  Context of last crossing attempt         Used a coyote or guide  −.50**  .58**    (.195)  (.218)   Days traveled  –  −.11†       (.068)   Perceives crossing as very/extremely dangerous  –  −.06       (.228)  Control         Possible that you will cross again?  –  −.51**       (.219)  Uncensored observations / censored observations  251  153  Rho = -.68        LR test of independent equations (rho = 0):         Chi2 (1) = 4.04         P > Chi2 = .0445        N = 404        M = 20 (results presented for 20th imputed data set)           “Heard of MM”a  “Potential Deterrence”a     (selection)  (outcome)  Variables        General human and financial capital         Male  .64*  –    (.269)      Age (years)          25-34  .38*  –    (.176)       35-44  .22  –    (.201)       45+  .74*  –    (.313)      Log household income  .23***  −.50***    (.068)  (.121)   Years of formal education  .07***  –    (.021)      Indigenous language speaker  −.63***  –    (.195)     Region of origin         North  .64*  –    (.298)      Central  .51*  –    (.214)      South  .26  –    (.193)     Migration-specific social capital         Family in destination  .12  –    (.143)  –  Migration-specific human capital         First crossing  −.67***  .29    (.260)  (.395)   Apprehensions  .03†  −.03†    (.019)  (.017)   Have lived in United States  .48*  –    (.195)      Current home in United States  –  .51       (.306)  Context of last crossing attempt         Used a coyote or guide  −.50**  .58**    (.195)  (.218)   Days traveled  –  −.11†       (.068)   Perceives crossing as very/extremely dangerous  –  −.06       (.228)  Control         Possible that you will cross again?  –  −.51**       (.219)  Uncensored observations / censored observations  251  153  Rho = -.68        LR test of independent equations (rho = 0):         Chi2 (1) = 4.04         P > Chi2 = .0445        N = 404        M = 20 (results presented for 20th imputed data set)        Notes: “<25 yrs”; “West-Central” region; and “no family in destination” are the referent groups. Coefficients, significance levels, and standard errors are reported, with the standard errors noted in parentheses. a 1 = yes; 0 = no † p < .10 *p < .05 **p < .01 ***p < .001 (two-tailed tests) Source: Migrant Border Crossing Survey I, unweighted data. On average, men’s probability of having heard of the Minutemen is 18 percentage points higher than women’s (see Appendix Table A3, column 1). Moreover, the probability of having heard of Minutemen is 11 percentage points higher for respondents 25-34 years old (compared to < 25 years old). Similarly, the probability of awareness is 20 percentage points higher for people age 45+ relative to those under the age of 25. Higher levels of human and financial capital are also associated with knowledge of Minutemen. Indigenous language speakers’ probability of having heard of Minutemen is 17 percentage points lower than non-indigenous language speakers. Respondents with higher household income prior to their most recent crossing and those with more formal educational attainment also had higher probabilities of having heard of the Minutemen (see Figures 1 and 2). Our results support the hypothesized positive associations between gender, income, and education, and prior knowledge of the Minutemen. Figure 1. View largeDownload slide Marginal Effects at Representative Values: “Having Heard of the Minutemen” by Monthly Household Income Figure 1. View largeDownload slide Marginal Effects at Representative Values: “Having Heard of the Minutemen” by Monthly Household Income Figure 2. View largeDownload slide Marginal Effects at Representative Values: “Having Heard of the Minutemen” by Years of Education Figure 2. View largeDownload slide Marginal Effects at Representative Values: “Having Heard of the Minutemen” by Years of Education Region of origin also structured access to information about Minutemen. Findings, however, do not completely support initial hypotheses. On average, northern Mexicans’ probability of awareness of the Minutemen is 18 percentage points higher than respondents from west-central Mexico (see Appendix Table A3, column 1), which is likely due to their proximity to the border. The “border zone” does not begin or end at the international divide but rather spans more than 100 kilometers into northern Mexico and the southwestern United States. Contrary to our hypothesis, however, central Mexicans’ probability of having heard of Minutemen is 14 percentage points higher than those from west-central Mexico. One explanation for this finding is that migration in central Mexico is also now occurring at levels fairly comparable to more established migrant-sending communities (Martell etal. 2007). These contemporary migration patterns may be quickly eroding the general and migration-specific social and human capital advantages long attributed to traditional sending regions. Additionally, perhaps, “sending regions”—as measured in our survey—do not adequately capture with enough precision these changing social processes. Migration-specific social capital—specifically, family ties in one’s destination—was not statistically significantly associated with prior knowledge. However, we did find that greater migration-specific human capital—being a non-first-time crosser, having more lifetime apprehensions, and having previously lived in the United States—was positively associated with prior knowledge of Minutemen. In fact, first-time crossers have a probability of having heard of Minutemen that is 19 percentage points lower than more experienced migrants (see Appendix Table A3, column 1). On the other hand, the probability of having heard of the Minutemen increases as lifetime apprehensions increase (see Figure 3). Figure 3. View largeDownload slide Marginal Effects at Representative Values: “Having Heard of the Minutemen” by Number of Prior Apprehensions Figure 3. View largeDownload slide Marginal Effects at Representative Values: “Having Heard of the Minutemen” by Number of Prior Apprehensions Experience living in the United States is another form of migration-specific human capital explaining differential levels of Minutemen awareness. The probability of awareness is, on average, 13 percentage points higher for migrants who have lived in the United States (versus migrants that have not) (see Appendix Table A3, column 1). While anti-immigrant mobilization efforts and the subsequent 2006 immigration marches across the United States did attract notable media attention throughout Mexico, attention was likely greater in the United States, especially among Latinos. Roughly 86 percent of MBCS respondents who said they had lived in the United States were living there at some point between 2005 and 2007, during both the growth of anti-immigrant mobilization and pro-immigration marches. Repatriated unauthorized migrants’ awareness of nativist groups was likely heightened as a consequence of having lived in the United States during this pivotal period. Among those who had lived in the United States, but not during the 2005-2007 period, it is likely that they were paying close attention to the immigration issue as they prepared for a future crossing prior to being surveyed. Together, findings suggest that migrants’ knowledge about anti-immigrant mobilization increases with first-hand migration experience. Lastly, the context of the last crossing experience helps explain prior knowledge of Minutemen. Surprisingly, we found that coyote users’ probability of having heard of the Minutemen was, on average, 14 percentage points lower compared to respondents not using a guide during their most recent crossing attempt (even when controlling for migration-specific social and human capital) (see Appendix Table A3, column 1). Although coyotes can be seen as producers and transmitters of knowledge about crossings, coyotaje is becoming a more organized informal business along the Sonora-Arizona border (Martínez 2016). While encountering Minutemen is a possibility, it is unlikely—something experienced coyotes understand but migrants may not. Indeed, of our 415 respondents interviewed, only 3 encountered Minutemen while crossing. Savvy coyotes would not want to call attention to an obstacle—especially one that is not likely to arise—that could potentially discourage migrants from using their services. Ultimately, more research is needed to fully explain this phenomenon, especially because coyotes could be plausibly motivated to exaggerate the dangers of the journey in order to emphasize how much their services are needed. Inferential Results for Heckman Probit Model with Sample Selection: The Conditional Probability of “Potential Deterrence” (Outcome Variable) Given “Having Heard” (Selection Variable) We were also interested in identifying factors explaining respondents’ potential deterrence from attempting a future crossing if encountering Minutemen were a possibility when correcting for sample selection for having heard of the group. We address this second question as the “outcome” variable in our Heckman probit selection model (see Table 2, column 2 and Appendix Table A3). First, having spent more time crossing the border and indicating that attempting a future crossing at a later date was a possibility decreased the conditional probability of potential deterrence given prior knowledge of the Minutemen, whereas believing one’s home was in the United States had no effect. Each additional day spent crossing the border was associated with a decrease in the probability of potential deterrence of 4 percentage points, while migrants noting it was possible they may cross again, on average, had probabilities of potential deterrence that were 20 percentage points lower than migrants who did not think or know if they would cross again (see Appendix Table A3, column 2). Migrants travelling for longer periods as well as those indicating they may cross again have likely invested greater resources into clandestine crossings. Second, greater income and being an indigenous language speaker help foster resiliency in the face of anti-immigrant mobilization. Migrants with greater access to financial resources can draw on those resources to facilitate unauthorized crossing attempts, despite a potential encounter with Minutemen that could lead to their removal from the United States (illustrated in Figure 4). In a similar vein, indigenous language speakers had probabilities of potential deterrence that were 12 percentage points lower than non-indigenous language speakers. Indigenous language speakers’ marginalized socioeconomic status in Mexico may operate as a powerful “push” factor compelling them to migrate despite the possibility of encountering a mobilized nativist citizenry. Figure 4. View largeDownload slide Marginal Effects at Representative Values: “Potential Deterrence” by Monthly Household Income Figure 4. View largeDownload slide Marginal Effects at Representative Values: “Potential Deterrence” by Monthly Household Income We uncovered a counterintuitive finding related to coyote use. Traveling with a coyote is associated with higher probabilities of potential deterrence. On average, coyote users’ probability of potential deterrence is 13 percentage points higher than migrants not using a coyote (see Appendix Table A3, column 2). We expected that migrants travelling with coyotes would be less likely to be deterred because guides should mitigate physical and social risks during the journey. Conversely, migrants travelling with coyotes are often relatively less experienced and thus more susceptible to anti-immigrant mobilization. Below, we elaborate on this finding and suggest that future research is needed to disentangle this paradox. CONCLUSION We know little about the extra-political consequences of nativist mobilization. Addressing this gap, we developed novel measures of nativist movement influence incorporating unauthorized repatriated migrants' perceptions of and potential reactions to such mobilization. Our analyses provided insight into the extra-political consequences of contemporary nativist mobilization and, in so doing, addressed long-debated questions: Do social movements (specifically, nativist movements) matter? If so, how (might they intersect with social processes of unauthorized migration)? Contemporary nativist movement influence is partially a story of gaining national media attention and raising awareness (Chavez 2008). Minutemen succeeded in generating awareness of their presence along the U.S.-Mexico border among the very population they sought to deter—nearly half of migrants in our survey had heard of them. However, we do not wish to understate the movement’s real political consequences. For one, nativist mobilization likely helped increase government funding for the U.S. Border Patrol (Doty 2009; Shapira 2013). More broadly, however, nativist organizations succeeded in reframing what was in the realm of possibility for immigration reform. Since 2005, individual and package immigration reforms have frequently included a border militarization component. This can partially be attributed to nativist mobilization normalizing exclusionary politics and strengthening the link between immigration control and the larger conservative political platform.10 However, it always seemed unlikely that nativist mobilization could abate migration generated by powerful economic, social, and political forces. In this regard, our results illuminate how unauthorized repatriated migrants and the border-crossing context intersect to mitigate nativist influence. Financial and migration-specific capital and crossing experiences can shield unauthorized repatriated migrants, making them less susceptible to potential deterrence. By examining how resiliency in the face of potential deterrents is structured among repatriated unauthorized migrants, our work demonstrates, more broadly, how relatively powerless groups—through, for instance, financial and human capital deployment and decisions about how to cross—can mitigate potential threats initiated by relatively privileged groups of U.S. citizens. However, one counterintuitive finding we encountered was that traveling with a coyote was associated with higher probabilities of potential deterrence. Previous research suggests coyotes help mitigate risks during clandestine crossings. It may well be the case that those using guides are relatively less experienced and thus perceive themselves to be more susceptible to nativist mobilization. It is also plausible that those that use coyotes are risk averse. If this is the case, this would explain why they hired coyotes in the first place and why they are likely to be potentially deterred by nativist groups. And because our survey instrument does not probe deeper into the interactions between migrants and coyotes, future research is needed to fully understand this finding. Second, our findings illuminate the inequalities intertwined with the clandestine migration process and reveal how these are structured. Unauthorized migrants do not have equal access to information about non-infrastructural, non-environmental deterrents in the form of a xenophobic and mobilized U.S. citizenry. Forty-two percent of respondents had previously heard of Minutemen. Of course, given how infrequently Minutemen actually encounter border crossers (based on our research and that of Shapira 2013), such mobilization along the border is unlikely to be as consequential as, for instance, the presence of a militarized border patrol or harsh crossing conditions. Nevertheless, discrepancies in prior knowledge of Minutemen are largely explained by higher levels of general human and financial capital, and migration-specific human capital—in the form of greater prior crossing experience, more lifetime apprehensions, and experience living in the United States. Although it should be acknowledged that because our sample comes from repatriated unauthorized migrants—individuals that did not “successfully” cross during their most recent attempt—it may be possible that the sample underrepresents the proportion of individuals with knowledge of the Minutemen, which one might presume would be higher among migrants that were able to cross without being apprehended and deported. Experiences associated with a migrant’s most recent crossing also help explain knowledge disparities. Unauthorized repatriated migrants who crossed without (versus with) coyotes had a higher probability of being aware of Minutemen. This paradoxical finding may be the result of financial pressures encouraging savvy coyotes to protect their business by keeping migrants unaware of potential dangers. With the emergence of “border business” coyotes—triggered by border enforcement initiatives that are funneling unauthorized migration and drug trafficking into the same geographic space—future research will want to investigate the complex and shifting relations between coyotes and their clients. Collectively, our findings illuminate both the structured inequalities—in terms of access to valued information about deterrents—shaping the unauthorized migration process, as well as why some unauthorized repatriated migrants are more or less resilient in the face of potential deterrents to clandestine mobilization. Scholars of migration and, in particular, those interested in the role various forms of capital play in the clandestine migration process, will want to further explore these issues by examining whether the same forms of capital we find important also play a significant role in the acquisition of knowledge about other—potentially more consequential—hurdles in the clandestine migration process (e.g., U.S. Border Patrol, violence/robbery, extreme weather, etc.). Is migrants’ knowledge about and resiliency towards a variety of potential threats shaped by some general set of capital-based factors (i.e., one, two, or three specific forms of capital) and/or are disparities in knowledge and differential levels of resiliency uniquely structured according to some general qualities of the hurdle being faced, like the rate (or probability) of encountering the hurdle, the potential costs associated with running into the hurdle, etc.? Finally, our research has broader implications for the study of social movements, particularly exclusionary right-wing mobilization. Social movement literature largely separates movements into two distinct categories—those targeting the state to effect change and those targeting civil society to transform culture and affirm identities. As such, the relationship between social movements and broader social change is typically assessed in terms of political change, on the one hand, and cultural change on the other (Amenta etal. 2010; Earl 2004). While this analytic grouping makes sense in the context of more progressive mobilization by disadvantaged groups, our research demonstrates that right-wing exclusionary movements too should be evaluated in terms of the extent to which they influence the behavior of targeted social groups. This includes not only the behavior of state actors and (potential) movement participants, but also the movement's targeted opponents, which constitute an important piece of the field in which movements operate. The authors wish to Jeremy Slack, Prescott Vandervoett, Kristin Klingman, Shiras Manning, Paola Molina, Kristen Valencia, Kylie Walzak, Melissa Burham, and Lorenzo Gamboa for their help with the data collection process. They also thank Kraig Beyerlein, Hana Brown, and the blind reviewers of Social Problems for their comments and suggestions. The authors express their gratitude to the owners and staff of the migrant shelter where they conducted their fieldwork as well as to the hundreds of people that took the time to discuss their migration experiences with the research team. This work was supported by the University of Arizona’s Underrepresented Graduate Student Final Project Fund as well as the Programa de Investigación de Migración y Salud (PIMSA), Health Initiative of the Americas at the University of California, Berkley. Direct correspondence to: Daniel E Martínez, School of Sociology, The University of Arizona, Social Sciences Building, Room 400, Tucson, AZ 85721. E-mail: daniel.martinez@arizona.edu. 1 Elsewhere (Ward and Martínez 2015) we discuss how repatriated unauthorized Mexican migrants perceive Minutemen and how they obtain information on nativist mobilization. Migrants get most of their information through U.S.-based media (and to a lesser extent the migration process itself or friends and family). Migrants discussed Minutemen by referencing ethno-racial categories (i.e., white, Chicano, etc.), specific roles or duties (i.e., preventing migrants from crossing, etc.), or by noting negative attitudes towards migrants and Mexicans (i.e., suggesting Minutemen are racist, etc.). 2 We do not address actual deterrence. We are talking about a potential, hypothetical encounter. Although intentions do not perfectly predict behavior, the theory of planned behavior suggests intentions are the most proximate determinant of behavior (Ajzen 1988). 3 Brown’s (2013) examination of the policy consequences of U.S. anti-immigrant mobilization is an exception. Still, her focus remains strictly on political consequences. Minuteman organizations—and the larger movement of which they were a visible part—framed the immigration reform debates of 2006-2007 (and even more recently) as matters of national security. They succeeded in having many of their proposals for increased border enforcement passed (or at least in getting favorable compromises passed). While Minutemen were not the sole force behind these surges, they remained one of the most visible public faces of the contemporary nativist movement. 4 From 2007-2010, MMP and MCDC chapters existed in most states. Of the 319 anti-immigration groups in 2010, 128 were connected with MMP or MCDC (Beirich 2011). 5 Less media attention has been paid to groups such as the Minutemen in Mexican media outlets relative to those in the United States. Rather, Mexican media has tended to focus their efforts on rallying against the expansion of the border wall (Trujeque Díaz 2007). 6 We focus exclusively on Mexican migrants because surveys were conducted in Mexico, and non-Mexicans are not repatriated to Mexico. Moreover, we are not generalizing to migrants in sending communities who have yet to migrate, migrants who were apprehended and repatriated along other regions of the border, or unauthorized migrants who crossed but were never detected by U.S. authorities. Nevertheless, heightened border enforcement efforts over the last few decades have increased the likelihood of apprehension. According to the U.S. Border Patrol, the current probability of apprehension ranges between 67 percent and 86 percent, which is substantially higher than older estimates based on self-report surveys collected through Princeton’s Mexican Migration Project and UC San Diego’s Mexican Migration Field Research Program (Rosenblum 2013:24). 7 Marginal differences exist between weighted and unweighted models (Winship and Radbill 1994), but the key substantive findings remain consistent. 8 To our knowledge no studies provide systematic quantitative data on encounters between Minutemen and unauthorized migrants. Our data suggest encounters are rare. Only three respondents who had heard of the Minutemen indicated they encountered them during their most recent crossing experience, two of which indicated they would not be deterred from a future crossing. 9 We tested for collinearity among variables prior to MI using the variance inflation factor (VIF). Collinearity was not an issue, as none of the VIFs exceeded 2.12 (Menard 1995). 10 Although the militarization of the border was underway prior to the emergence of contemporary nativist mobilization (Dunn 1996), we contend that nativist mobilization efforts further contribute to border militarization. APPENDIX Table A1. Descriptive Statistics for Selected MBCS Variables Used in Analyses (non-imputed and non-weighted data)    Mean/Percent  Std. Dev.  Missing Observations (%)  Dependent variable            Minutemen deter  .54  .610  1.93  Selection variable            Heard of Minutemen  .38  .487  1.20  General human and financial capital            Male  .88  .320  .00   Age (years)             < 25  .30  .458  .00    25-34  .39  .489  .00    35-44  .24  .425  .00    45+  .07  .259  .00   Log household income  6.00  1.153  13.01   Years of formal education  6.93  3.592  2.17   Indigenous language speaker  .21  .410  .48  Region of origin            North  .09  .287  .96   West-Central (“traditional”)  .20  .398  .96   Central  .20  .403  .96   South  .51  .500  .96  Migration-specific social capital            Family in destination  .54  .499  .24  Migration-specific human capital            First crossing  .22  .413  .00   Apprehensions  3.96  5.930  .00   Have lived in United States  .63  .483  .48   Current home in United States  .13  .332  17.34  Context of last crossing attempt            Used a coyote or guide  .72  .448  .96   Days traveled  2.49  1.677  .48   Perceives crossing as very/extremely dangerous  .80  .405  2.65  Control            Possible you will cross again?  .48  .500  4.58  N = 415           Mean/Percent  Std. Dev.  Missing Observations (%)  Dependent variable            Minutemen deter  .54  .610  1.93  Selection variable            Heard of Minutemen  .38  .487  1.20  General human and financial capital            Male  .88  .320  .00   Age (years)             < 25  .30  .458  .00    25-34  .39  .489  .00    35-44  .24  .425  .00    45+  .07  .259  .00   Log household income  6.00  1.153  13.01   Years of formal education  6.93  3.592  2.17   Indigenous language speaker  .21  .410  .48  Region of origin            North  .09  .287  .96   West-Central (“traditional”)  .20  .398  .96   Central  .20  .403  .96   South  .51  .500  .96  Migration-specific social capital            Family in destination  .54  .499  .24  Migration-specific human capital            First crossing  .22  .413  .00   Apprehensions  3.96  5.930  .00   Have lived in United States  .63  .483  .48   Current home in United States  .13  .332  17.34  Context of last crossing attempt            Used a coyote or guide  .72  .448  .96   Days traveled  2.49  1.677  .48   Perceives crossing as very/extremely dangerous  .80  .405  2.65  Control            Possible you will cross again?  .48  .500  4.58  N = 415        Source: Migrant Border Crossing Survey I. View Large Table A1. Descriptive Statistics for Selected MBCS Variables Used in Analyses (non-imputed and non-weighted data)    Mean/Percent  Std. Dev.  Missing Observations (%)  Dependent variable            Minutemen deter  .54  .610  1.93  Selection variable            Heard of Minutemen  .38  .487  1.20  General human and financial capital            Male  .88  .320  .00   Age (years)             < 25  .30  .458  .00    25-34  .39  .489  .00    35-44  .24  .425  .00    45+  .07  .259  .00   Log household income  6.00  1.153  13.01   Years of formal education  6.93  3.592  2.17   Indigenous language speaker  .21  .410  .48  Region of origin            North  .09  .287  .96   West-Central (“traditional”)  .20  .398  .96   Central  .20  .403  .96   South  .51  .500  .96  Migration-specific social capital            Family in destination  .54  .499  .24  Migration-specific human capital            First crossing  .22  .413  .00   Apprehensions  3.96  5.930  .00   Have lived in United States  .63  .483  .48   Current home in United States  .13  .332  17.34  Context of last crossing attempt            Used a coyote or guide  .72  .448  .96   Days traveled  2.49  1.677  .48   Perceives crossing as very/extremely dangerous  .80  .405  2.65  Control            Possible you will cross again?  .48  .500  4.58  N = 415           Mean/Percent  Std. Dev.  Missing Observations (%)  Dependent variable            Minutemen deter  .54  .610  1.93  Selection variable            Heard of Minutemen  .38  .487  1.20  General human and financial capital            Male  .88  .320  .00   Age (years)             < 25  .30  .458  .00    25-34  .39  .489  .00    35-44  .24  .425  .00    45+  .07  .259  .00   Log household income  6.00  1.153  13.01   Years of formal education  6.93  3.592  2.17   Indigenous language speaker  .21  .410  .48  Region of origin            North  .09  .287  .96   West-Central (“traditional”)  .20  .398  .96   Central  .20  .403  .96   South  .51  .500  .96  Migration-specific social capital            Family in destination  .54  .499  .24  Migration-specific human capital            First crossing  .22  .413  .00   Apprehensions  3.96  5.930  .00   Have lived in United States  .63  .483  .48   Current home in United States  .13  .332  17.34  Context of last crossing attempt            Used a coyote or guide  .72  .448  .96   Days traveled  2.49  1.677  .48   Perceives crossing as very/extremely dangerous  .80  .405  2.65  Control            Possible you will cross again?  .48  .500  4.58  N = 415        Source: Migrant Border Crossing Survey I. View Large Table A2. Descriptions for Dependent, Selection, and Independent Variables Used in the Heckman Probit Model    Survey Question  Metric  Dependent/selection variables         Heard of the Minutemen  "Have you ever heard of the Minutemen?"  1 = "yes"; 0 = "no"   Minutemen deter  “Would the possibility of encountering the Minutemen deter you from attempting a future unauthorized crossing?”  1 = "yes"; 0 = "no / don't know"  Independent variables        Male  "Is the respondent male?"     Age (years)         % < 25  "Is the respondent less than 25 years old?"  1 = "yes"; 0 = "no"    % 25-34  "Is the respondent between 25 and 34 years old?"  1 = "yes"; 0 = "no"    % 35-44  "Is the respondent between 35 and 44 years old?"  1 = "yes"; 0 = "no"    % 45+  "Is the respondent more than 44 years old?"  1 = "yes"; 0 = "no"   Household income (in log dollars)  "What was your monthly household income before your last crossing attempt?"  Count (in log dollars)   Education  "How many years of formal education have you completed?"  Count (in years)   Indigenous language speaker  "Do you speak an indigenous language?"  1 = "yes"; 0 = "no"   North  "Is the respondent from the north region of Mexico?"  1 = "yes"; 0 = "no"   West-Central (“traditional”)  "Is the respondent from the west-central region of Mexico?"  1 = "yes"; 0 = "no"   Central  "Is the respondent from the central region of Mexico?"  1 = "yes"; 0 = "no"   South  "Is the respondent from the south region of Mexico?"  1 = "yes"; 0 = "no"   Family in destination  "Do you have family in your desired destination?"     First crossing  "Was your last crossing attempt your first?"  1 = "yes"; 0 = "no"   Number of lifetime apprehensions  "How many times have you been apprehended by any U.S. authority, including your most recent apprehension?"  Count (number of times)   Have lived in United States  "Have you ever lived or worked in the United States?"     Current home in United States  "Is your current home in the United States?"  1 = "yes"; 0 = "no"   Used a coyote or guide  "Did you use a coyote or guide during your last crossing attempt?"  1 = "yes"; 0 = "no"  Days traveled  "How many days did you spend traveling in the desert before being apprehended or picked up to proceed to the next leg of the journey?"  Count (number of days)  Perceives crossing as very/extremely dangerous  "Respondent perceives the crossing as ‘very’ or ‘extremely dangerous’ after their most recent attempt."  1 = "yes"; 0 = "no"  Possible you will cross again?  "Is it possible that you will cross the border again without papers?"  1 = "yes"; 0 = "no"     Survey Question  Metric  Dependent/selection variables         Heard of the Minutemen  "Have you ever heard of the Minutemen?"  1 = "yes"; 0 = "no"   Minutemen deter  “Would the possibility of encountering the Minutemen deter you from attempting a future unauthorized crossing?”  1 = "yes"; 0 = "no / don't know"  Independent variables        Male  "Is the respondent male?"     Age (years)         % < 25  "Is the respondent less than 25 years old?"  1 = "yes"; 0 = "no"    % 25-34  "Is the respondent between 25 and 34 years old?"  1 = "yes"; 0 = "no"    % 35-44  "Is the respondent between 35 and 44 years old?"  1 = "yes"; 0 = "no"    % 45+  "Is the respondent more than 44 years old?"  1 = "yes"; 0 = "no"   Household income (in log dollars)  "What was your monthly household income before your last crossing attempt?"  Count (in log dollars)   Education  "How many years of formal education have you completed?"  Count (in years)   Indigenous language speaker  "Do you speak an indigenous language?"  1 = "yes"; 0 = "no"   North  "Is the respondent from the north region of Mexico?"  1 = "yes"; 0 = "no"   West-Central (“traditional”)  "Is the respondent from the west-central region of Mexico?"  1 = "yes"; 0 = "no"   Central  "Is the respondent from the central region of Mexico?"  1 = "yes"; 0 = "no"   South  "Is the respondent from the south region of Mexico?"  1 = "yes"; 0 = "no"   Family in destination  "Do you have family in your desired destination?"     First crossing  "Was your last crossing attempt your first?"  1 = "yes"; 0 = "no"   Number of lifetime apprehensions  "How many times have you been apprehended by any U.S. authority, including your most recent apprehension?"  Count (number of times)   Have lived in United States  "Have you ever lived or worked in the United States?"     Current home in United States  "Is your current home in the United States?"  1 = "yes"; 0 = "no"   Used a coyote or guide  "Did you use a coyote or guide during your last crossing attempt?"  1 = "yes"; 0 = "no"  Days traveled  "How many days did you spend traveling in the desert before being apprehended or picked up to proceed to the next leg of the journey?"  Count (number of days)  Perceives crossing as very/extremely dangerous  "Respondent perceives the crossing as ‘very’ or ‘extremely dangerous’ after their most recent attempt."  1 = "yes"; 0 = "no"  Possible you will cross again?  "Is it possible that you will cross the border again without papers?"  1 = "yes"; 0 = "no"  Source: Migrant Border Crossing Survey I. View Large Table A2. Descriptions for Dependent, Selection, and Independent Variables Used in the Heckman Probit Model    Survey Question  Metric  Dependent/selection variables         Heard of the Minutemen  "Have you ever heard of the Minutemen?"  1 = "yes"; 0 = "no"   Minutemen deter  “Would the possibility of encountering the Minutemen deter you from attempting a future unauthorized crossing?”  1 = "yes"; 0 = "no / don't know"  Independent variables        Male  "Is the respondent male?"     Age (years)         % < 25  "Is the respondent less than 25 years old?"  1 = "yes"; 0 = "no"    % 25-34  "Is the respondent between 25 and 34 years old?"  1 = "yes"; 0 = "no"    % 35-44  "Is the respondent between 35 and 44 years old?"  1 = "yes"; 0 = "no"    % 45+  "Is the respondent more than 44 years old?"  1 = "yes"; 0 = "no"   Household income (in log dollars)  "What was your monthly household income before your last crossing attempt?"  Count (in log dollars)   Education  "How many years of formal education have you completed?"  Count (in years)   Indigenous language speaker  "Do you speak an indigenous language?"  1 = "yes"; 0 = "no"   North  "Is the respondent from the north region of Mexico?"  1 = "yes"; 0 = "no"   West-Central (“traditional”)  "Is the respondent from the west-central region of Mexico?"  1 = "yes"; 0 = "no"   Central  "Is the respondent from the central region of Mexico?"  1 = "yes"; 0 = "no"   South  "Is the respondent from the south region of Mexico?"  1 = "yes"; 0 = "no"   Family in destination  "Do you have family in your desired destination?"     First crossing  "Was your last crossing attempt your first?"  1 = "yes"; 0 = "no"   Number of lifetime apprehensions  "How many times have you been apprehended by any U.S. authority, including your most recent apprehension?"  Count (number of times)   Have lived in United States  "Have you ever lived or worked in the United States?"     Current home in United States  "Is your current home in the United States?"  1 = "yes"; 0 = "no"   Used a coyote or guide  "Did you use a coyote or guide during your last crossing attempt?"  1 = "yes"; 0 = "no"  Days traveled  "How many days did you spend traveling in the desert before being apprehended or picked up to proceed to the next leg of the journey?"  Count (number of days)  Perceives crossing as very/extremely dangerous  "Respondent perceives the crossing as ‘very’ or ‘extremely dangerous’ after their most recent attempt."  1 = "yes"; 0 = "no"  Possible you will cross again?  "Is it possible that you will cross the border again without papers?"  1 = "yes"; 0 = "no"     Survey Question  Metric  Dependent/selection variables         Heard of the Minutemen  "Have you ever heard of the Minutemen?"  1 = "yes"; 0 = "no"   Minutemen deter  “Would the possibility of encountering the Minutemen deter you from attempting a future unauthorized crossing?”  1 = "yes"; 0 = "no / don't know"  Independent variables        Male  "Is the respondent male?"     Age (years)         % < 25  "Is the respondent less than 25 years old?"  1 = "yes"; 0 = "no"    % 25-34  "Is the respondent between 25 and 34 years old?"  1 = "yes"; 0 = "no"    % 35-44  "Is the respondent between 35 and 44 years old?"  1 = "yes"; 0 = "no"    % 45+  "Is the respondent more than 44 years old?"  1 = "yes"; 0 = "no"   Household income (in log dollars)  "What was your monthly household income before your last crossing attempt?"  Count (in log dollars)   Education  "How many years of formal education have you completed?"  Count (in years)   Indigenous language speaker  "Do you speak an indigenous language?"  1 = "yes"; 0 = "no"   North  "Is the respondent from the north region of Mexico?"  1 = "yes"; 0 = "no"   West-Central (“traditional”)  "Is the respondent from the west-central region of Mexico?"  1 = "yes"; 0 = "no"   Central  "Is the respondent from the central region of Mexico?"  1 = "yes"; 0 = "no"   South  "Is the respondent from the south region of Mexico?"  1 = "yes"; 0 = "no"   Family in destination  "Do you have family in your desired destination?"     First crossing  "Was your last crossing attempt your first?"  1 = "yes"; 0 = "no"   Number of lifetime apprehensions  "How many times have you been apprehended by any U.S. authority, including your most recent apprehension?"  Count (number of times)   Have lived in United States  "Have you ever lived or worked in the United States?"     Current home in United States  "Is your current home in the United States?"  1 = "yes"; 0 = "no"   Used a coyote or guide  "Did you use a coyote or guide during your last crossing attempt?"  1 = "yes"; 0 = "no"  Days traveled  "How many days did you spend traveling in the desert before being apprehended or picked up to proceed to the next leg of the journey?"  Count (number of days)  Perceives crossing as very/extremely dangerous  "Respondent perceives the crossing as ‘very’ or ‘extremely dangerous’ after their most recent attempt."  1 = "yes"; 0 = "no"  Possible you will cross again?  "Is it possible that you will cross the border again without papers?"  1 = "yes"; 0 = "no"  Source: Migrant Border Crossing Survey I. View Large Table A3. Average Marginal Effects for Heckman Probit Selection Model Explaining “Potential Deterrence” (outcome) and “Having Heard of the Minutemen” (selection) Variables  Pr  Pr  (heard = 1)  (deterrence = 1 / heard = 1)  General human and financial capital         Male  .18*  .12    (.074)  (.088)   Age (years)          25-34  .11*  .07    (.048)  (.053)    35-44  .06  .04    (.056)  (.044)    45+  .20*  .14    (.085)  (.089)   Log household income  .06***  −.15***    (.018)  (.039)   Years of formal education  .02***  .01    (.005)  (.008)   Indigenous language speaker  −.17***  −.12†    (.052)  (.071)  Region of origin         North  .18*  .12    (.082)  (.086)   Central  .14*  .10    (.058)  (.066)   South  .07  .05    (.053)  (.046)  Migration-specific social capital         Family in destination  .03  .02    (.040)  (.030)  Migration-specific human capital         First crossing  −.19**  −.01    (.071)  (.143)   Apprehensions  .01†  −.01    (.005)  (.007)   Have lived in United States  .13*  .09    (.053)  (.057)   Current home in United States  –  .20       (.109)  Context of last crossing attempt         Used a coyote or guide  −.14**  .13†    (.044)  (.079)   Days traveled  –  −.04†       (.024)   Perceives crossing as very/extremely dangerous  –  −.02       (.090)  Control         Possible that you will cross again?  –  −.20**       (.075)  N = 404      M = 20 (results presented for 20th imputed data set)        Variables  Pr  Pr  (heard = 1)  (deterrence = 1 / heard = 1)  General human and financial capital         Male  .18*  .12    (.074)  (.088)   Age (years)          25-34  .11*  .07    (.048)  (.053)    35-44  .06  .04    (.056)  (.044)    45+  .20*  .14    (.085)  (.089)   Log household income  .06***  −.15***    (.018)  (.039)   Years of formal education  .02***  .01    (.005)  (.008)   Indigenous language speaker  −.17***  −.12†    (.052)  (.071)  Region of origin         North  .18*  .12    (.082)  (.086)   Central  .14*  .10    (.058)  (.066)   South  .07  .05    (.053)  (.046)  Migration-specific social capital         Family in destination  .03  .02    (.040)  (.030)  Migration-specific human capital         First crossing  −.19**  −.01    (.071)  (.143)   Apprehensions  .01†  −.01    (.005)  (.007)   Have lived in United States  .13*  .09    (.053)  (.057)   Current home in United States  –  .20       (.109)  Context of last crossing attempt         Used a coyote or guide  −.14**  .13†    (.044)  (.079)   Days traveled  –  −.04†       (.024)   Perceives crossing as very/extremely dangerous  –  −.02       (.090)  Control         Possible that you will cross again?  –  −.20**       (.075)  N = 404      M = 20 (results presented for 20th imputed data set)        Notes: “<25 yrs”; “West-Central” region; and “no family in destination” are the referent groups. Average marginal effects (AME), significance levels, and standard errors are reported, with the standard errors noted in parentheses. † p < .10 *p < .05 **p < .01 ***p < .001 (two-tailed tests) Source: Migrant Border Crossing Survey I, unweighted data View Large Table A3. Average Marginal Effects for Heckman Probit Selection Model Explaining “Potential Deterrence” (outcome) and “Having Heard of the Minutemen” (selection) Variables  Pr  Pr  (heard = 1)  (deterrence = 1 / heard = 1)  General human and financial capital         Male  .18*  .12    (.074)  (.088)   Age (years)          25-34  .11*  .07    (.048)  (.053)    35-44  .06  .04    (.056)  (.044)    45+  .20*  .14    (.085)  (.089)   Log household income  .06***  −.15***    (.018)  (.039)   Years of formal education  .02***  .01    (.005)  (.008)   Indigenous language speaker  −.17***  −.12†    (.052)  (.071)  Region of origin         North  .18*  .12    (.082)  (.086)   Central  .14*  .10    (.058)  (.066)   South  .07  .05    (.053)  (.046)  Migration-specific social capital         Family in destination  .03  .02    (.040)  (.030)  Migration-specific human capital         First crossing  −.19**  −.01    (.071)  (.143)   Apprehensions  .01†  −.01    (.005)  (.007)   Have lived in United States  .13*  .09    (.053)  (.057)   Current home in United States  –  .20       (.109)  Context of last crossing attempt         Used a coyote or guide  −.14**  .13†    (.044)  (.079)   Days traveled  –  −.04†       (.024)   Perceives crossing as very/extremely dangerous  –  −.02       (.090)  Control         Possible that you will cross again?  –  −.20**       (.075)  N = 404      M = 20 (results presented for 20th imputed data set)        Variables  Pr  Pr  (heard = 1)  (deterrence = 1 / heard = 1)  General human and financial capital         Male  .18*  .12    (.074)  (.088)   Age (years)          25-34  .11*  .07    (.048)  (.053)    35-44  .06  .04    (.056)  (.044)    45+  .20*  .14    (.085)  (.089)   Log household income  .06***  −.15***    (.018)  (.039)   Years of formal education  .02***  .01    (.005)  (.008)   Indigenous language speaker  −.17***  −.12†    (.052)  (.071)  Region of origin         North  .18*  .12    (.082)  (.086)   Central  .14*  .10    (.058)  (.066)   South  .07  .05    (.053)  (.046)  Migration-specific social capital         Family in destination  .03  .02    (.040)  (.030)  Migration-specific human capital         First crossing  −.19**  −.01    (.071)  (.143)   Apprehensions  .01†  −.01    (.005)  (.007)   Have lived in United States  .13*  .09    (.053)  (.057)   Current home in United States  –  .20       (.109)  Context of last crossing attempt         Used a coyote or guide  −.14**  .13†    (.044)  (.079)   Days traveled  –  −.04†       (.024)   Perceives crossing as very/extremely dangerous  –  −.02       (.090)  Control         Possible that you will cross again?  –  −.20**       (.075)  N = 404      M = 20 (results presented for 20th imputed data set)        Notes: “<25 yrs”; “West-Central” region; and “no family in destination” are the referent groups. Average marginal effects (AME), significance levels, and standard errors are reported, with the standard errors noted in parentheses. † p < .10 *p < .05 **p < .01 ***p < .001 (two-tailed tests) Source: Migrant Border Crossing Survey I, unweighted data View Large REFERENCES Ajzen Icek. 1988. Attitudes, Personality, and Behavior . Milton Keynes, UK: Open University Press. Amenta Edwin, Caren Neal, Chiarello Elizabeth, Su Yang. 2010. “The Political Consequences of Social Movements.” Annual Review of Sociology  36: 287- 307. Google Scholar CrossRef Search ADS   Andreas Peter. 2009. Border Games.  Ithaca, NY: Cornell University Press. Baumgartner Frank, Mahoney Christine. 2005. “Social Movements, the Rise of New Issues, and the Public Agenda.” Pp. 65- 86 in Routing the Opposition , edited by Meyer D., Jenness V., Ingram H. Minneapolis: University of Minnesota Press. Beirich Heidi. 2011. “The Year in Nativism, 2010.” Intelligence Report , Spring, p. 141. The Southern Poverty Law Center, Montgomery, AL. Brown Hana. 2013. “Race, Legality, and the Social Policy Consequences of Anti-Immigration Mobilization.” American Sociological Review  78( 2): 290- 314. Google Scholar CrossRef Search ADS   Cabrera Luis, Glavac Sonya. 2010. “Minutemen and Desert Samaritans.” Journal of Ethnic and Migration Studies  36( 4): 673- 95. Google Scholar CrossRef Search ADS   Centro de Estudios Migratorios Latinoamericanso (CEMLA). 2013. “Eventos de Repatriación de Mexicanos Desde Estados Unidos, Según Entidad Federativa y Punto de Recepción, 1995-2012.” Unidad de Política Migratoria, Secretaría de Gobernación, México D.F. Cerrutti Marcela, Massey Douglas S. 2001. “On the Auspices of Female Migration from Mexico to the United States.” Demography  38( 2): 187- 200. Google Scholar CrossRef Search ADS PubMed  Chavez Leo R. 2008. The Latino Threat.  Stanford, CA: Stanford University Press. Comisión Nacional Para el Desarrollo de los Pueblos Indígenas (CDI). 2006. Indicadores Sociodemográficos de la Población Indígena 2000-2005 . Retrieved July 20, 2017 (www.cdi.gob.mx/cedulas/sintesis_resultados_2005.pdf). Cornelius Wayne A. 1998. “The Structural Embeddedness of Demand for Mexican Immigrant Labor.” Pp. 115- 44 in Crossings , edited by Suarez-Orozco M. Cambridge, MA: Harvard University Press. Cornelius Wayne A., Fitzgerald David, Fischer Pedro Lewin, Muse-Orlinoff Leah. 2010. Migration and the U.S. Economic Crisis.  San Diego, CA: Center for Comparative Immigration Studies. Cornelius Wayne A., Lewis Jessa M. 2007. Impacts of Border Enforcement on Mexican Migration . San Diego, CA: Center for Comparative Immigration Studies. De Genova Nicholas P. 2002. “Migrant ‘Illegality’ and Deportability in Everyday Life.” Annual Review of Anthropology  31: 419- 47. Google Scholar CrossRef Search ADS   Doty Roxanne L. 2007. “States of Exception on the Mexico-U.S. Border.” International Political Sociology  1: 113- 37. Google Scholar CrossRef Search ADS   Doty Roxanne L. 2009. The Law Into Their Own Hands , Tucson: The University of Arizona Press. Dove April L. 2010. “Framing Illegal Immigration at the U.S.-Mexican Border.” Pp. 199- 237 in Research in Social Movements, Conflicts, and Change , edited by Coy P. Bingley, UK: Emerald Group Publishing Limited. Google Scholar CrossRef Search ADS   Dunn Timothy. 1996. The Militarization of the U.S.-Mexico Border . Austin: University of Texas Press. Earl Jennifer. 2004. “The Cultural Consequences of Social Movements.” Pp. 508- 30 in The Blackwell Companion to Social Movements , edited by Snow D., Soule S., Kriesi H. Malden, MA: Blackwell Publishing. Elcioglu Emine. 2015. “Popular Sovereignty on the Border.” Ethnography  16( 4): 438- 62. Google Scholar CrossRef Search ADS   Giugni Mario. 2007. “Useless Protest?” Mobilization  12: 53- 77. Gilchrist Jim, Corsi Jerome. 2006. Minutemen: The Battle to Secure America's Borders.  Los Angeles, CA: World Ahead Publishing. Graham John W., Olchowski Allison E., Gilreath Tamika D. 2007. “How Many Imputations are Really Needed?” Prevention Science  8: 206- 13. Google Scholar CrossRef Search ADS PubMed  Hagan Jacqueline M. 2008. Migration Miracle . Cambridge, MA: Harvard University Press. Heckman James J. 1976. “The Common Structure of Statistical Models of Truncation, Sample Selection and Limited Dependent Variables and a Simple Estimator for Such Models.” Annals of Economic and Social Measurement  5: 475- 92. Kandel William, Massey Douglas S. 2002. “The Culture of Mexican Migration.” Social Forces  80( 3): 981- 1004. Google Scholar CrossRef Search ADS   Kil Sang Hea, Menjivar Cecilia, Doty Roxanne Lynne. 2009. “The Border Spectacle” Pp. 297- 312 in Immigration, Crime, and Justice , edited by McDonald W. Bingley, UK: Emerald Group Publishing Ltd. Google Scholar CrossRef Search ADS   Marcelli Enrico A., Cornelius Wayne A. 2001. “The Changing Profile of Mexican Migrants to the United States” Latin American Research Review  36( 3): 105- 31. Martell Alpha, Pineda Maribel, Tapia Luis 2007. “The Contemporary Migration Process.” Pp. 49- 72 in Mayan Journeys , edited by Cornelius W., Fitzgerald D., Fischer P. San Diego, CA: Center for Comparative Immigration Studies. Martínez Daniel E. 2016. “Coyote Use in an Era of Heightened Border Enforcement: New Evidence from the Arizona-Sonora Border Region.” Journal of Ethnic and Migration Studies  42( 1): 103- 19. Google Scholar CrossRef Search ADS   Martínez Daniel E., Slack Jeremy, Beyerlein Kraig, Vandervoet Prescott, Klingman Kristin, Molina Paola, Manning Shiras, Burham Melissa, Walzak Kylie, Valencia Kristen, Gamboa Lorenzo. 2017. “The Migrant Border Crossing Study: A Methodological Overview.” Population Studies  71( 2): 249- 64. Google Scholar CrossRef Search ADS PubMed  Martínez Daniel E., Slack Jeremy, Martínez-Schuldt Ricardo. n.d. “Return Migration in the Age of the Unauthorized Permanent Resident: A Quantitative Assessment of Migration Intentions Post-Deportation.” Unpublished Manuscript. Martínez Daniel E., Vandervoet Prescott, Slack Jeremy. 2013. “Methodological Challenges and Ethical Concerns of Researching Marginalized and Vulnerable Populations: Evidence from Firsthand Experiences of Working with Undocumented Migrants.” Pp. 101- 20 in Uncharted Terrains: New Directions in Border Research Methodology, Ethics, and Practice , edited O’Leary A., Deeds C., Whiteford S. Tucson: University of Arizona Press. Massey Douglas, Durand Jorge, Malone Nolan. 2002. Beyond Smoke and Mirrors . New York: Russell Sage Foundation. Massey Douglas S., Capoferro Chiara. 2008. “The Geographic Diversification of American Immigration.” Pp. 25- 50 in New Faces in New Places , edited by Massey Douglas S. New York: The Russell Sage Foundation. Massey Douglas S., Espinoza Kristin E. 1997. “What's Driving Mexico-U.S. Migration?" The American Journal of Sociology  102( 4): 939- 99. Google Scholar CrossRef Search ADS   McAdam Doug. 1989. “The Biographical Consequences of Activism.” American Sociological Review  54( 5): 744- 60. Google Scholar CrossRef Search ADS   Minuteman Civil Defense Corps Project (MCDC). 2009. “Please Mark Your Calendars with Our 2009 Muster Schedule.” Retrieved March 5, 2014 (www.minutemanhq.com/state/read.php?chapter=AZ&sid=1023). Minuteman Project, The. 2014. “Operation Normandy.” Retrieved June 7, 2016 (http://baesic.net/minutemanproject/). Menard Scott. 1995. Applied Logistic Regression Analysis . Thousand Oaks, CA: Sage. Neiwert David. 2013. “Smaller, More Radical Border Militias Patrolling Arizona Desert.” The Southern Poverty Law Center, Montgomery, AL. Nevins Joseph. 2010. Operation Gatekeeper and Beyond . New York: Routledge. O’Leary Anna Ochoa. 2009. “Mujeres en el Cruce.” Journal of the Southwest  51( 4): 523- 42. Google Scholar CrossRef Search ADS   Passel Jeffrey S., Cohn D’Vera. 2016. “Overall Number of Unauthorized Immigrants Holds Steady Since 2009: Decline in share from Mexico mostly offset by growth from Asia, Central America and sub-Saharan African.” Pew Research Center, Washington, DC. Piven Francis F. 2006 Challenging Authority.  Lanham, MD: Rowman & Littlefield. Portes Alejandro. 1998. “Social Capital” Annual Review of Sociology  24: 1- 24. Google Scholar CrossRef Search ADS   Rosenblum Marc R. 2013. Border Security: Immigration Enforcement Between Ports of Entry . Washington, DC: Congressional Research Service. Royston Patrick. 2009. “Multiple Imputation of Missing Values.” The Stata Journal  9: 466- 77. Rubin Donald B. 1987. Multiple Imputation for Nonresponse in Surveys . New York: Wiley. Google Scholar CrossRef Search ADS   Shapira Harel. 2013. Waiting for Jose . Princeton, NJ: Princeton University Press. Google Scholar CrossRef Search ADS   Sherkat Darren, Blocker T. Jean. 1997. “Explaining the Political and Personal Consequences of Protest.” Social Forces  75( 3): 1049- 70. Google Scholar CrossRef Search ADS   Singer Audrey, Massey Douglas S. 1998. “The Social Process of Undocumented Border Crossing Among Mexican Migrants.” International Migration Review  32( 3): 561- 92. Google Scholar CrossRef Search ADS PubMed  Skocpol Theda. 2003. Diminished Democracy . Norman: University of Oklahoma Press. Skocpol Theda, Williamson Vanessa. 2013. The Tea Party and the Remaking of Republican Conservatism.  Oxford, UK: Oxford University Press. Slack Jeremy, Martínez Daniel, Vandervoet Prescott. 2011. “Methods of Violence: Researcher Safety and Adaptability in Times of Conflict.” Practicing Anthropology  33( 1): 33- 37. Google Scholar CrossRef Search ADS   Slack Jeremy, Martínez Daniel E., Whiteford Scott, Emily Peiffer. 2013. “In the Shadow of the Wall: Family Separation, Immigration Enforcement and Security.” The Center for Latin American Studies, University of Arizona, Tucson, AZ. Retrieved July 20, 2017 (http://las.arizona.edu/sites/las.arizona.edu/files/UA_Immigration_Report2013web.pdf). Slack Jeremy, Martínez Daniel E., Whiteford Scott, Peiffer Emily. 2015. “In Harm’s Way: Family Separation, Immigration Policy, and Security on the U.S.-Mexico Border.” Journal on Migration and Human Security  3( 2): 109- 28. Google Scholar CrossRef Search ADS   Snow David, Soule Sarah, Kriesi Hanspeter. 2004. “Mapping the Terrain.” Pp. 3- 16 in Blackwell Companion to Social Movements , edited by Snow D., Soule S., Kriesi H. Malden, MA: Blackwell Publishing. Google Scholar CrossRef Search ADS   Stewart Julie, Bendall Michele Enciso, Morgan Charlie V. 2015. “Jobs, Flags, and Laws: How Interests, Culture, and Values Explain Recruitment into the Utah Minuteman Project.” Sociological Perspectives  58( 4): 627- 48. Google Scholar CrossRef Search ADS   Southern Poverty Law Center (SPLC). 2010. “Intelligence Report.” Spring. Retrieved July 20, 2017 (www.splcenter.org/intelligence-report?f[0]=field_intel_report_issue%3A11689). Spener David. 2009. Clandestine Crossings . Ithaca, NY: Cornell University Press. Trujeque Díaz José Antonio. 2007. “Minuteman Project: Segregación y Activismo Antimigratorio.” Andamios  2( 6): 137- 72. U.S. Border Patrol (USBP). 2014a. “United States Border Patrol Apprehensions by Gender - Fiscal Year 2013 (Oct. 1st through Sept. 30th).” Retrieved August 18, 2017 (www.cbp.gov/sites/default/files/documents/U.S.%20Border%20Patrol%20Fiscal%20Year%202013%20Profile.pdf). U.S. Border Patrol (USBP). 2014b. “United States Border Patrol Southwest Border Sector Total Illegal Alien Apprehensions by Fiscal Year (Oct. 1st through Sept. 30th).” Retrieved August 18, 2017 (www.cbp.gov/sites/default/files/assets/documents/2016-Oct/BP%20Southwest%20Border%20Sector%20Apps%20FY1960%20-%20FY2016.pdf). Ward Matthew. 2014. “They Say Bad Things Come in Threes: How Economic, Political and Cultural Shifts Facilitated Contemporary Anti-Immigration Activism in the United States.” Journal of Historical Sociology  27( 2): 263- 92. Google Scholar CrossRef Search ADS   Ward Matthew. 2016. “Opportunity, Resources, and Threat: Explaining Local Nativist Organizing in the United States.” Sociological Perspectives : 1- 20. Retrieved July 20, 2017 (DOI: 10.1177/0731121416655994). Ward Matthew, Martínez Daniel E. 2015. “Know Your Enemy: How Unauthorized Migrants Learn About and Perceive Anti-Immigrant Mobilization in the United States.” Migration Letters  12( 2): 50- 66. Williams Richard. 2012. “Using the Margins Command to Estimate and Interpret Adjusted Predictions and Marginal Effects.” The Stata Journal  12( 2): 308- 31. Winship Christopher, Radbill Larry. 1994. “Sampling Weights and Regression Analysis.” Sociological Methods & Research  23( 2): 230- 57. Google Scholar CrossRef Search ADS   © The Author 2017. Published by Oxford University Press on behalf of the Society for the Study of Social Problems. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Social Problems Oxford University Press

Agency and Resilience Along the Arizona-Sonora Border: How Unauthorized Migrants Become Aware of and Resist Contemporary U.S. Nativist Mobilization

Loading next page...
 
/lp/ou_press/agency-and-resilience-along-the-arizona-sonora-border-how-unauthorized-WS610Ntu7P
Publisher
University of California Press
Copyright
© The Author 2017. Published by Oxford University Press on behalf of the Society for the Study of Social Problems. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
0037-7791
eISSN
1533-8533
D.O.I.
10.1093/socpro/spx028
Publisher site
See Article on Publisher Site

Abstract

Abstract Little is known about the extra-political consequences of contemporary U.S.-based nativist mobilization as well as the resilience unauthorized migrants display in the face of anti-immigrant mobilization along the U.S.-Mexico border. Bringing together social movements and immigration literatures, we examine these interrelated issues using original survey data from the first wave of the Migrant Border Crossing Study. In so doing, we examine: (1) factors influencing repatriated unauthorized migrants’ awareness of nativist mobilization (i.e., Minutemen) along the Arizona-Sonora border, and (2) factors explaining why some migrants would or would not be potentially deterred from attempting future unauthorized crossings if encountering the Minutemen were a possibility. Results from a Heckman probit selection model indicate that higher levels of general, financial, and migration-specific human capital are associated with awareness of the Minutemen, while higher household income and status as an indigenous language speaker predict who would be less likely to be deterred from crossing. We also uncover an interesting paradox: migrants traveling with coyotes were less likely to have heard of Minutemen and more likely to be potentially deterred. Collectively, our results provide insight into the overlooked extra-political consequences of contemporary U.S. nativist mobilization, how resiliency in the face of such a deterrent is structured among repatriated unauthorized migrants, and how seemingly powerless migrant groups can mitigate potential threats initiated by relatively privileged groups of U.S. citizens. immigration, Minutemen, border crossers, social movements, unauthorized migration Debates over immigration reform have waxed and waned but never fallen from center stage in the United States. Passage of restrictionist legislation coupled with deliberation over Senate Bill 744, President Obama’s 2012 and 2014 Executive Orders of Deferred Action for Childhood Arrivals (DACA) and Deferred Action Program for Parents (DAPA), increasing electoral participation of Latinos, and the recent rise to prominence of nativist political candidates signal as much. However, given the focus on policy reform and institutional change, it becomes easy to overlook grassroots mobilization along the U.S.-Mexico border and two key protagonists helping to sustain this drama—unauthorized migrants and mobilized nativists called “Minutemen.” Two shifts in population trends among (un)authorized immigrants set the stage for this drama. First is the growth of the unauthorized immigrant population since 1990, which currently stands around 11 million (Passel and Cohn 2016). Scholars, ironically, attributed this growth to increased border militarization (Cornelius and Lewis 2007; Massey, Durand, and Malone 2002). Second, (un)authorized immigrant populations moved into “new destination” states (Massey and Capoferro 2008), which became ground zero for the passage of anti-immigrant legislation. And in both “new” and “traditional destination” states, residents mobilized against foreign population growth. Recent research has focused on the rise of anti-immigrant mobilization, in particular self-described Minutemen organizations that have moved beyond merely advocating for restrictionist legislation and are known largely for their controversial and visible patrol efforts along the U.S.-Mexico border, as well as their interior enforcement efforts targeting migrants in the workplace (e.g., Doty 2009; Shapira 2013; Stewart, Bendall, and Morgan 2015; Ward 2014, 2016). While these groups are organizationally diverse and their goals multifaceted (Doty 2009; Ward 2016), generally-speaking, they seek to both raise awareness among the public and policy makers about unauthorized migration and the need for increased border enforcement while also leveraging the media spectacles their presence along the border and key interior spaces generates to try and deter migrants from crossing in the first place (Chavez 2008; Doty 2009; Gilchrist and Corsi 2006; Shapira 2013; Ward 2016). Although exceedingly rare, Minutemen have on occasion apprehended unauthorized border crossers and turned them over to U.S. authorities. Thus, Minutemen potentially deter unauthorized crossings by creating a physical presence along the border but also—and primarily—by leveraging the media spectacle of this presence and the uncertainty attached to it regarding potential physical harm that “rogue” nativist border-vigilantes may exercise against unauthorized immigrants. Extant literature on Minutemen contributes to our understanding of right-wing exclusionary mobilization. However, a key perspective remains absent: that of the movement’s targeted opponents—unauthorized migrants. Taking this silence as a starting point, we first ask: What factors help predict whether or not a repatriated unauthorized migrant knows of the Minutemen?1 Such migrants do not have equal access to information about the crossing experience. Focusing attention on informational disparities advances sociological understandings of the social process of border crossing. Additionally, it allows us to assess the nativist movement’s broader influence on the population they sought to deter. Our second research question builds on the first. We ask, what factors explain migrants’ perceptions of their own susceptibility to being deterred from attempting another crossing if encountering Minutemen were a possibility?2 Much attention has been paid to the ways a militarized border (Andreas 2009; Cornelius and Lewis 2007; Dunn 1996; Nevins 2010), labor markets (Cornelius 1998; Cornelius etal. 2010), and crossing conditions (Hagan 2008; Martínez 2016; O’Leary 2009) affect unauthorized crossings, but little is known about how a nation’s mobilized nativist citizenry influences unauthorized repatriated migrants’ perceptions of the viability of clandestine crossings. This provides an additional measure of nativist influence on the movement’s targeted opponents. We uncover findings using our original data set of 415 in-depth surveys with recently repatriated migrants that attempted crossing one of the most heavily traversed regions along the U.S.-Mexico border among unauthorized Mexican migrants—southern Arizona. Using a Heckman probit selection model, which allows us to examine the conditional probability of potential deterrence given prior knowledge of the Minutemen, we found unauthorized repatriated migrants’ awareness of Minutemen was positively associated with being male, age, education, household income, being from northern or central Mexico (relative to west-central Mexico), number of lifetime apprehensions, and having lived in the United States. The use of a coyote (guide), marginalized status, and crossing for the first time were negatively associated with awareness. Second, respondents with higher monthly household income, those who spent more time crossing the border, those with higher cumulative lifetime apprehensions, and those who believed they would cross again were less susceptible to the potential deterrence effect embodied in a mobilized nativist citizenry. Respondents travelling with coyotes were more susceptible to deterrence stemming from a potential Minutemen encounter. Nevertheless, nativist mobilization along the U.S.-Mexico border appears to fail to deter many resilient migrants. Our research bridges literatures on the social processes of migration and social movement consequences. First, we contribute to migration literature by describing the heterogeneity of repatriated unauthorized migrants, a group often erroneously regarded as homogenous (De Genova 2002). We offer a rare survey-based, quantitative analysis focusing on an overlooked dimension of the resource disparities existing among such migrants—knowledge, specifically, knowledge about non-infrastructural, non-environmental deterrents like a xenophobic, mobilized citizenry. In so doing, we reveal how unequal access to information about potential crossing hurdles is structured. Second, we contribute at the intersection of migration and social movements scholarship by offering novel empirical measures of nativist movement influence. Examining two interrelated issues—(1) factors influencing migrants’ awareness of anti-immigrant activism and (2) factors associated with migrants’ perceptions of their own susceptibility to being deterred from future clandestine crossing attempts—provides insight into the extra-political consequences of nativist mobilization. Moreover, by studying nativists’ targets (i.e., unauthorized migrants), our findings reveal how these opponents—often thought of as passive and powerless—exert agency and resist nativist efforts. NATIVIST MOBILIZATION AND ITS CONSEQUENCES Do social movements matter? If so, how? While these remain central issues for social movement scholars, little consensus exists (Amenta etal. 2010). On the one hand, social movements are touted as influential and responsible for important political changes (Baumgartner and Mahoney 2005; Piven 2006). On the other hand, movements often fail to shape social and political landscapes (Giugni 2007; Skocpol 2003). Surveying literature on the expressly political consequences of social movements, Edwin Amenta and associates (2010) concluded that most research on larger movements demonstrates that movements matter. Research on social movement influence, however, focuses overwhelmingly on political consequences. Nevertheless, while movements are frequently geared towards effecting political change, they may also have consequences beyond the political sphere (Earl 2004). Countercultural and self-help movements, for instance, focus largely on expressive action rather than political change (Earl 2004; Snow, Soule, and Kriesi 2004). Moreover, “movement influence” need not be limited to outcomes at the societal level (e.g., policy change). Activism influences participants too, in the short- and long-term (McAdam 1989; Sherkat and Blocker 1997). Yet, despite the significance of understanding the effects of social movement mobilization both within and beyond the political realm, research on nativist mobilization, in particular, largely ignores the extent to which it might influence local, state, and national politics.3 Additionally, the movement’s extra-political consequences remain understudied. Instead, research trends towards providing a fine-grained description of the movement’s origins (Ward 2014, 2016), members’ attitudes (Cabrera and Glavac 2010), movement activities (Elcioglu 2015; Shapira 2013), and cultural frames used to construct anti-immigrant rationalizations (Dove 2010). Sang Kil, Cecilia Menjivar, and Roxanne Doty (2009), for instance, examined how Minutemen targeted employers, protested against officials, and raised awareness. April Dove (2010) explained how Minutemen framed websites to generate collective action resonance. Doty (2009) brought us into the nativist world through interviews and observation. Leo Chavez (2008) argued the inaugural minuteman border campaign served as a media spectacle reaffirming citizenship privileges and fanning the flames of restrictionist legislation initiatives. Emine Elcioglu’s (2015) fieldwork examined how Minutemen challenged and reinforced state authority by enacting nativism through popular sovereignty. Finally, Harel Shapira’s (2013) ethnography shed light on processes of acculturation involved in becoming a minuteman. Despite the significance of these studies, nativist mobilization’s extra-political consequences remain under-examined. In what ways, if any, has nativist mobilization mattered beyond the political sphere? We tackle this issue by linking it to our analyses of repatriated unauthorized migrants’ awareness of nativist mobilization, as well as their perceived susceptibility to such mobilization. In so doing, we extend research on social movement consequences in two respects. First, our questions represent novel conceptualizations of potential nativist movement influence. The contemporary nativist movement seeks to raise awareness about unauthorized migration among both the public and policy makers (Dove 2010; Kil etal. 2009). Additionally, and largely through Minutemen organizations and the media spectacles their surveillance efforts create (Chavez 2008), the movement also seeks to deter migrants from attempting unauthorized crossings in the first place (Doty 2009; Gilchrist and Corsi 2006; Shapira 2013; Ward 2016). Given this latter goal, we ask: how much awareness has been raised among our target migrant population and how is this awareness structured? Second, given Minutemen’s exclusionary rhetoric and border patrol activity, would these migrants potentially be deterred from crossing again if encountering Minutemen were a possibility? Relatedly, how are migrants’ perceptions of their own potential susceptibility shaped by a variety of individual characteristics and experiential factors related to their most recent crossing attempt? Answers provide insight into the under examined extra-political consequences of nativist mobilization. Second, examining potential deterrence reveals how relatively powerless groups like unauthorized migrants can resist mobilization initiated by relatively privileged groups. Examining how migrants’ resiliency is constructed in the face of nativist mobilization emphasizes their agency, which although highlighted by migration scholars, has been largely neglected by the general public and policy makers. Focusing on migrants’ resiliency also provides insight into forces stymying nativist influence. By presenting novel ways to think about how movements matter outside of regular politics, our analysis examines diverse ways that contemporary nativism has and has not exerted influence. NATIVIST MOBILIZATION AT THE U.S.-MEXICO BORDER: AN OVERVIEW Nativist mobilization along the U.S.-Mexico border is not new. In the mid-1800s, vigilante and state-sanctioned groups roamed the borderlands (Spener 2009). In the 1970s and 80s, members of the Ku Klux Klan operated the Klan Border Watch program. Yet, it was not until after the U.S. government’s adoption of its “prevention through deterrence” strategy and the Operation Gatekeeper complex (Andreas 2009; Nevins 2010) that growing numbers of residents began patrolling (Doty 2007). Over the next two decades, the Border Solution Task Force, U.S. Citizen Patrol, Voices of Citizens Together, American Border Patrol, Ranch Rescue, and Civil Homeland Defense emerged. Recently, the Minuteman Project (MMP) and the Minuteman Civil Defense Corps (MCDC) ushered in a new era of media-savvy and better-funded anti-immigrant vigilantism. Of course, these organizations were part of a broader milieu of nativist hostilities unauthorized immigrants faced during this period, particularly in Arizona. The state legislature and governor considered numerous anti-immigrant laws, and the Border Patrol engaged in an aggressive media campaign highlighting enforcement efforts and the dangers of unauthorized immigration. In 2005, Jim Gilchrist and Chris Simcox implored Americans to participate in a month-long “muster” to observe and report unauthorized immigrants entering along the Arizona-Sonora border. These Minutemen sought to raise awareness among the public and policy makers about unauthorized immigration, while also instilling a sense of fear among migrants by signaling to them that unauthorized crossing attempts would be met with force. Soon thereafter, Minutemen focused their attention throughout the United States.4 At the time our survey was conducted (2007-2009), Minutemen were experiencing explosive growth, jumping from 173 chapters in 2008 to 309 just a year later (Beirich 2011). Twenty-one chapters operated in Arizona (second only to California) (SPLC 2010), and most of the high profile musters occurred in southern Arizona (Chavez 2008; Shapira 2013). Nationally, some of the growth during this period was focused around increased interior enforcement. Minutemen began targeting migrants in the workplace and their employers. This also included increased lobbying for reform of local and state immigration policies. By the mid-to-late 2000s, the nativist agenda was normalized as various localities considered or passed restrictive housing ordinances and Arizona’s SB 1070 and Alabama’s HB 56 were debated. However, despite that some chapter growth occurred in communities away from the border, Shapira (2013) found that many individuals involved in border patrol were travelling from such communities into southern Arizona. Thus, despite the growth of interior enforcement during this period, patrol activity still continued at the Arizona-Sonora border well into 2008 and 2009 (MCDC 2009; Shapira 2013), albeit in a reduced capacity relative to 2005-2007. By 2011, nativist mobilization was declining. The Tea Party offered a legitimate space in which anti-immigrant sentiment was stoked behind a veil of traditional conservatism and channeled into political gain for the Right (Skocpol and Williamson 2013). Infighting and internal splintering among major nativist organizations, along with the highly publicized arrests of nativist leaders on murder and drug charges exacerbated the downward spiral. And yet, nativist mobilization has not disappeared. In response to the recent transportation of unauthorized migrant children from Texas to makeshift emergency shelters in Arizona and California, MMP and others launched Operation Normandy in 2014 (Minuteman Project 2014). Finally, “mainstream” nativist decline has also triggered the growth of smaller, radical splinter cell organizations (Neiwert 2013). EXPLAINING AWARENESS AND POTENTIAL DETERRENCE Despite the growth of anti-immigrant mobilization, little is known about how unauthorized migrants’ differential awareness of this potential deterrent is structured. Moreover, by influencing perceptions about the viability of clandestine crossings, it is plausible that a mobilized nativist citizenry represents another roadblock to migrants’ passage. However, only some migrants will be deterred. To what extent do economic and political forces known to significantly structure migration overshadow the influence of nativist mobilization? Below we develop theoretical arguments and present hypotheses to explain these two extra-political consequences of nativist mobilization. Gender Unauthorized migration from Mexico to the United States is gendered (O’Leary 2009). Males are more likely to migrate (Cerrutti and Massey 2001). When they do migrate, women typically migrate for family reunification (Cerrutti and Massey 2001), though not exclusively so. Despite this, unauthorized border crossers continue to be largely male. According to FY 2013 U.S. Border Patrol apprehensions, females constituted 16.5 percent of apprehensions (USBP 2014a). Males also have greater first-hand migration experience. A recent study found that repatriated Mexican migrant women have, on average, 2.9 lifetime unauthorized crossing attempts and 1.7 apprehensions compared to 5.3 crossings and 3.2 apprehensions among men (p < .05) (Slack etal. 2013). Because first-hand migration experience contributes to migrants’ knowledge stores, it shapes expectations for the unauthorized crossing. And these experiences likely shape perceptions of potential hurdles, including anti-immigrant mobilization efforts. Considering men’s greater migration experience, as well as the gendering of unauthorized Mexican migration, we suggest: (H1) Migrant men are more likely than migrant women to have heard of the Minutemen. (H2) Migrant men are less likely to be deterred by a potential encounter relative to migrant women. General Human and Financial Capital General human and financial capital vary among migrants. Greater community-level economic resources play important roles in determining the probability of first-time and repeat migration (Massey and Espinoza 1997). These resources likely impact migrants’ knowledge of and susceptibility to anti-immigrant mobilization. Although some unauthorized migrants received their information about Minutemen through friends and family members in the United States and Mexico (a form of social capital), as well as during the migration process, most got their information from media outlets in the United States, and to a lesser extent, in Mexico5 (Ward and Martínez 2015). Higher levels of household income allow migrants to consume more diverse media. Literacy, operationalized as higher levels of formal educational attainment, may also increase migrants’ access to information about unauthorized crossings, including the possibility of encountering Minutemen. Thus, opportunities to consume a variety of media should increase a migrant’s likelihood of having heard of Minutemen. After all, media visibility was a focus of Minutemen mobilization. Unauthorized migrants also differ with respect to their marginalized status within Mexico. Approximately 9.8 percent of Mexico’s population is indigenous (CDI 2006). This population experiences higher levels of socioeconomic and ethno-racial marginalization and contends with social barriers that disproportionately negatively impact access to border-crossing information (see Martínez, Vandervoet, and Slack 2013). We, thus, hypothesize: (H3) Migrants with greater household income and higher levels of formal education will be more likely to have heard of the Minutemen, whereas indigenous language speakers will be less likely to have heard. Migrants also consider the costs and risks of unauthorized crossings. Given the risks—such as exposure to the elements and apprehension by Border Patrol—many attempts fail. Failure quickly compounds costs. Higher household income should mitigate costs incurred from multiple crossing attempts, thus we expect: (H4) Migrants with greater household income view Minutemen—and the potential for subsequent apprehension and deportation—as less of a risk and, thus, less of a potential deterrent. Region of Origin Mexico’s Consejo Nacional de Población (CONAPO) identified four major migrant sending regions within Mexico: northern, west-central (traditional), central, and southern-southeastern. Constant migration from the west-central region since the U.S. Bracero Program was enacted (1942-1964) has fostered a culture of migration in these communities. Migration is viewed as an important rite of passage, and there are social expectations that young men will migrate (Cornelius and Lewis 2007; Kandel and Massey 2002). A “culture of migration,” and the maturity and density of social networks involved in the migration process from these communities, likely increases awareness of nativist mobilization. Thus: (H5) Migrants from west-central (traditional) areas are more likely to have heard of recent Minutemen efforts when compared to those from other regions of Mexico. The south-southeastern and central regions of Mexico have more recent histories of U.S. migration (Marcelli and Cornelius 2001). Yet, migration is occurring at levels comparable to more established migrant-sending communities (Martell, Pineda, and Tapia 2007). These individuals are more likely to come from indigenous, marginalized communities (CDI 2006). These regions’ relatively recent history of migration, higher levels of poverty and marginalization, and greater distance from the U.S.-Mexico border often mean lower levels of general and migration-specific social and human capital, thus making the journey riskier. Migrants from these regions likely have less knowledge of the hurdles involved in the crossing experience, including anti-immigrant groups. Thus, we hypothesize: (H6) Migrants from southern and central Mexico have less knowledge of anti-immigrant groups relative to those from west-central areas. Migration-Specific Social and Human Capital Migration-specific social capital—operationalized as ties to family members in the United States—has consistently shown to be important in the migration process. These ties increase the odds of making a first trip with or without documents, shape modes of crossing, and assist in securing employment (Martínez 2016; Massey and Espinoza 1997; Singer and Massey 1998). Additionally, migration-specific human capital—operationalized as first-hand migration experience—influences migrants’ modes of crossing (Martínez 2016; Singer and Massey 1998). More first-hand experience also decreases the probability of apprehension and spurs repeat migration (Massey and Espinoza 1997). Lived experience abroad also seems to matter. Place attachment, particularly the subjective understanding of where one’s current home is located, plays an important role in future migration intentions. For instance, recently repatriated Mexican migrants who consider their home to be located in the United States report higher rates of future migration intentions relative to those whose homes are in Mexico (Slack etal. 2015). In a similar vein, a higher degree of affectual ties to the United States—operationalized as having one's home in the United States and having lived in the country for a decade or longer—is a significant and strong predictor of future crossing intentions, even after controlling for conventional measures of economic, human, and migration-specific social capital as well as the specific punitive immigration enforcement programs through which unauthorized immigrants are processed (Martínez, Slack, and Martínez-Schuldt n.d.). Overall, migration-specific social and human capital not only shape migration attempts but also increase access to key information about the unauthorized migration process, either through social ties, lived experience, or place attachment. We, therefore, posit the following hypotheses: (H7) Migrants with greater migration-specific social and human capital are more likely to have heard of the Minutemen. (H8) Migrants who consider their home to be located in the United States have a stronger resolve to migrate despite the possibility of encountering Minutemen. Unauthorized crossings are risky, especially for first timers. The possibility of encountering an anti-immigrant group may be overwhelming for individuals that have never lived or worked in the United States and never attempted a prior crossing. Because more apprehensions bring greater awareness of the potential consequences of being caught by Minutemen and turned over to immigration officials, we hypothesize: (H9) First-time crossers are more likely to be potentially deterred, while a migrant's number of lifetime apprehensions is negatively associated with this outcome. Context of the Last Crossing Experience No two unauthorized crossing experiences are identical. Depending upon where a migrant started the trip, the corridor being traversed, and time of year, trip duration can range from a few hours to two weeks. The majority of unauthorized migrants employ coyotes (human smugglers) to cross the border (Martínez 2016; Singer and Massey 1998; Spener 2009). Migrants rely on coyotes to decrease the physical risks associated with crossing the border and to lower the likelihood of apprehension (Hagan 2008; Spener 2009). Coyotes instill a sense of security because they are subject-matter experts. The foundations of social capital—reciprocity, trust, value introjection, and bounded solidarity—help mitigate against mistreatment and often lead to a mutually beneficial relationship between coyotes and migrants (Portes 1998; Spener 2009). However, this may be changing with the emergence of “border business” coyotes in response to border enforcement initiatives that are funneling unauthorized migration and drug trafficking into the same geographic spaces (see Martínez 2016; Spener 2009). Coyotes can also be thought of as knowledge creators and diffusers. Not only do they know the terrain and most effective crossing strategies, but they are also vehicles through which knowledge about these conditions is transmitted. This includes knowledge about potential hurdles, such as nativist mobilization. By constructing and transmitting more proximate knowledge about crossing hurdles, coyotes shape migrants’ risk perceptions. One of the primary ways in which migrants become aware of Minutemen is during the migration process itself (Ward and Martínez 2015). Thus, we hypothesize: (H10) Migrants who travel with coyotes are more likely to have heard of the Minutemen. (H11) Because migrants use coyotes to reduce social and physical risks associated with unauthorized migration attempts, travelling with a coyote decreases migrants’ susceptibility to being deterred by potentially encountering Minutemen. DATA AND METHODS Using original survey data from the first wave of our Migrant Border Crossing Study (MBCS; N = 415), we examine unauthorized repatriated migrants’ awareness of Minutemen and the extent to which a potential encounter with Minutemen may deter future crossings. The aim of our survey was to better grasp the social processes shaping unauthorized repatriated migrants’ border-crossing experiences through southern Arizona (Martínez etal. 2017). Although our survey approach cannot capture the full spectrum of experiences individuals accumulate during their crossings, it does gauge what respondents knew about Minutemen at that time following their most recent border crossing attempt. Numerous insightful qualitative studies have been conducted on migrants’ crossing experiences; however, there has been less effort to leverage the advantages of a more quantitative, survey approach, especially with respect to the questions addressed in this article. Doing so, however, allows us to generate more precise descriptions of a population often erroneously regarded as homogenous (De Genova 2002) and also allows us to more confidently generalize our results. Our population frame consists of recently repatriated migrants who crossed through and were apprehended within the Tucson Sector, a location that was, at the time of the survey period (2007-2009), a hotbed of nativist mobilization (Chavez 2008; Doty 2009; Shapira 2013; Ward 2014). Results, therefore, generalize to migrants with a crossing, apprehension, and repatriation experience within this sector.6 In an era of increased border and immigration enforcement, the apprehension-repatriation experience is becoming much more commonplace. Unauthorized migrants have never been at higher risk for apprehension and repatriation (Rosenblum 2013). Thus, the typical MBCS respondent is likely more representative than the hypothetical, perfectly evasive migrant never detected by U.S. authorities. During our survey period (2007-2009), 937,608 migrants were apprehended in the Tucson sector (USBP 2014b) and 565,278 Mexican migrants were repatriated to the state of Sonora (CEMLA 2013). Because the Tucson sector accounted for roughly 45 percent of all apprehensions, and Sonora accounted for 33 percent of repatriations to Mexico during this time period, northern Sonora represents an ideal research location (CEMLA 2013; USBP 2014b). We completed surveys between October 2007 and July 2009 in a migrant shelter in Nogales, Sonora. This was an ideal research site because 78 percent of migrants repatriated to Sonora between 2007 and 2009 were repatriated to the city of Nogales (CEMLA 2013). Furthermore, the overwhelming majority of migrants were transported from repatriation sites to this particular shelter—the only one in the city that provided lodging for up to three days—by Mexican federal agents. Once migrants arrived at the shelter for the evening they could not enter and exit the shelter throughout the night due to safety concerns; therefore, most did not typically depart until the following morning. People with local connections or those originally from Sonora might opt to not stay at the shelter because they can presumably go home and/or stay with friends and family. The shelter also provided a safe space for researchers and respondents during a time when Nogales, Sonora, experienced high levels of drug trafficking-related violence (Martínez, etal. 2013; Slack, Martínez, and Vandervoet 2011). Eligible participants were at least 18 years of age and must have attempted an unauthorized crossing along the Arizona-Sonora border, been apprehended by any U.S. authority, and repatriated to Mexico within the past six months at the time of the survey. Ensuring that respondents’ most recent crossing attempt occurred within a comparable geographical space and time allows for comparability between cases. The six-month cutoff reduced retrospective bias. MBCS survey methodology is extensively detailed elsewhere (Martínez etal. 2017). Sampling Potential MBCS participants were randomly selected using a spatial random sampling technique. We divided the shelter into five previously defined spaces and randomly selected every nth person in each area. We opted not to randomly sample from the shelter’s daily sign-in sheet to protect respondents’ anonymity. Instead, we discretely approached potential respondents, introduced ourselves, and invited them to speak with us. We screened those who had been randomly selected for eligibility and invited them to complete the survey. The response rate was 97 percent (Martínez etal. 2017). There was little to occupy migrants’ time while at the shelter (e.g., no televisions or radios). And because shelter guests could not depart until the following morning, many welcomed the opportunity to discuss their experiences, which helps account for the study’s high response rate (Martínez etal. 2017). We applied probability weights calculated from monthly Border Patrol apprehension statistics between 2007 and 2009 to minimize selection bias between shelter goers and people who did not stay in a shelter upon repatriation. Weights were constructed to correct for any discrepancies between the proportion of males and females in the population and our sample, as well as any discrepancies between the proportion of individuals from each of the four Mexican sending regions in the population and our sample. These weights are based on Tucson sector apprehensions for the month each respondent was interviewed during the study time period. Probability weights are used when calculating our descriptive statistics. However, we did not implement them when conducting the inferential analyses (Winship and Radbill 1994). In situations (like ours) where sampling weights are solely a function of independent variables included in the models (and not the dependent variable), Christopher Winship and Larry Radbill recommend using unweighted estimates because they are unbiased, consistent, and produce smaller standard errors.7 The generalizability of the sample and comparisons of respondents’ demographic characteristics to estimates of the study population are discussed elsewhere (Martínez etal. 2017). Demographic characteristics of our sample are highly consistent with findings from similar data sources (for instance, see data from EMIF-Norte, Instituto Nacional de Migración, and U.S. Border Patrol). Minutemen Module Wave I of the MBCS consisted of 12 modules and over 350 questions focused around a range of crossing-related topics, such as: migrants’ most recent border-crossing experiences, experiences while in U.S. custody, and demographic characteristics. We also included a module on migrants’ perceptions of and potential reactions to Minutemen. This module was not included in the second wave of the MBCS, and to our knowledge, no survey of unauthorized repatriated migrants to date has explored these topics. Therefore, the first wave of the MBCS is the best data source for our purposes. We began by asking migrants whether or not they had ever heard of Minutemen. We then debriefed all respondents by reading a standardized script informing them that Minutemen were residents in the United States voluntarily mobilizing to stop unauthorized migrants from crossing. Essentially, all respondents were given the same information before being asked this study’s second outcome of interest (“Would the possibility of encountering the Minutemen deter you from attempting a future unauthorized crossing?”). As such, our second question focuses on a potential—albeit hypothetical—encounter with Minutemen, and not actual encounters, which our data indicate are rare events.8 Instead, we are interested in repatriated migrants perceptions of their own susceptibility to deterrence were such an encounter a possibility. Nevertheless, because people who had previously heard of the group are likely qualitatively distinct from those who had never heard, and because reading a standardized script while being surveyed does not adequately substitute for previously acquired first-hand knowledge of the Minutemen, our analysis employs a Heckman probit model correcting for sample selection associated with having previously heard of the group. This technique helps identify factors predicting knowledge of the group while also correcting for selection bias when examining potential deterrence (Heckman 1976), and ultimately allows us to examine the conditional probability of potential deterrence given “having heard of the Minutemen.” In our case, both the “selection” and “outcome” variables are dichotomous, which warrants the use of a probit model. Missing Data We use multiple imputations (MI) to handle missing observations (Rubin 1987). MI preserves information that listwise deletion would omit. Following John Graham, Allison Olchowski, and Tamika Gilreath (2007), we conducted 20 imputations (m = 20) using the multiple imputation by chained equations (MICE) method (Royston 2009). Each imputation substituted cases with missing information with unbiased plausible values using their predictive distributions in a separate data set. We estimated variable means on each of the imputed data sets, combining the results using “Rubin’s Rules” to yield coefficient estimates and standard errors (Rubin 1987 as cited by Graham etal. 2007). Appendix Table A1 provides descriptive statistics for non-imputed and non-weighted MBCS data as well as the percentage of observations missing for each variable. Most variables contained very little missing data. “Household income” (13 percent) and “home in the U.S.” (17 percent) are exceptions. Both questions were open-ended, which partially explains the relatively higher rate of “missingness.” The monthly household income question may have yielded less missing information had categories been presented. However, we asked for the exact currency amount to be able to conduct non-linear transformations if needed. The home location question was open-ended but also subjective. Some respondents “didn’t know” where they considered home to be, while others refused to answer the question (both instances coded: “missing”). Considering respondents’ immigration status, this is unsurprising. We found similar results using imputed and non-imputed data. However, nearly 100 cases are lost using listwise deletion, which explains minor discrepancies. Because it yields less biased estimates, we use multiple imputations to handle missing observations. RESULTS AND DISCUSSION Descriptive Statistics for Variables Used in Heckman Probit Selection Model Table 1 provides descriptive statistics for independent, selection, and dependent variables. Forty-two percent of the full sample had previously heard of Minutemen, while 43 percent indicated a possible future encounter with the group could potentially deter them from crossing. Table 1. Descriptive Statistics for Dependent and Independent Variables used in the Probit Selection Model (multiply imputed data)    Mean/Percent (full sample)  Mean/Percent (“heard” = 0)  Mean/Percent (“heard” = 1)  Mean/Percent (“deterrence” = 0)  Mean/Percent (“deterrence” = 1)  Dependent (outcome) variable                  Minutemen deter (%)  43  48  35  –  –  Selection variable                  Heard of Minutemen (%)  42  –  –  47  34  General human and financial capital                  Male (%)  87  80  97  93  80   Age (years)                   <25 (%)  24  29  17  23  25    25-34 (%)  41  35  48  40  42    35-44 (%)  24  25  23  28  18    45 + (%)  11  11  12  9  15   Log household income  6.1  5.7  6.6  6.4  5.7   Years of formal education  7.0  6.7  7.4  6.9  7.2   Indigenous language speaker (%)  20  27  10  21  19  Region of origin                  North (%)  14  8  22  15  13   West-Central (“traditional”) (%)  23  25  21  24  23   Central (%)  33  30  36  32  33   South (%)  30  37  20  29  31  Migration-specific social capital                  Family in destination (%)  58  54  63  55  62  Migration-specific human capital                  First crossing (%)  19  29  5  14  25   Apprehensions  4.4  2.8  6.6  5.1  3.4   Have lived in United States (%)  68  54  87  71  63   Current home in United States (%)  17  8  31  21  13  Context of last crossing attempt                  Used a coyote or guide (%)  70  80  57  65  77   Days traveled  2.4  2.5  2.2  2.4  2.5   Perceives crossing as very/ extremely dangerous (%)  81  82  78  80  81  Control                  Possible you will cross again? (%)  51  42  63  59  40  N  404  251  153  236  168     Mean/Percent (full sample)  Mean/Percent (“heard” = 0)  Mean/Percent (“heard” = 1)  Mean/Percent (“deterrence” = 0)  Mean/Percent (“deterrence” = 1)  Dependent (outcome) variable                  Minutemen deter (%)  43  48  35  –  –  Selection variable                  Heard of Minutemen (%)  42  –  –  47  34  General human and financial capital                  Male (%)  87  80  97  93  80   Age (years)                   <25 (%)  24  29  17  23  25    25-34 (%)  41  35  48  40  42    35-44 (%)  24  25  23  28  18    45 + (%)  11  11  12  9  15   Log household income  6.1  5.7  6.6  6.4  5.7   Years of formal education  7.0  6.7  7.4  6.9  7.2   Indigenous language speaker (%)  20  27  10  21  19  Region of origin                  North (%)  14  8  22  15  13   West-Central (“traditional”) (%)  23  25  21  24  23   Central (%)  33  30  36  32  33   South (%)  30  37  20  29  31  Migration-specific social capital                  Family in destination (%)  58  54  63  55  62  Migration-specific human capital                  First crossing (%)  19  29  5  14  25   Apprehensions  4.4  2.8  6.6  5.1  3.4   Have lived in United States (%)  68  54  87  71  63   Current home in United States (%)  17  8  31  21  13  Context of last crossing attempt                  Used a coyote or guide (%)  70  80  57  65  77   Days traveled  2.4  2.5  2.2  2.4  2.5   Perceives crossing as very/ extremely dangerous (%)  81  82  78  80  81  Control                  Possible you will cross again? (%)  51  42  63  59  40  N  404  251  153  236  168  Notes: Cases in which values were missing on the outcome and selection variables were included in the multiple imputation process and then omitted (n = 11) before estimating ”heckprob” procedure in Stata 13; m = 20. Standard errors available upon request. Source: Migrant Border Crossing Survey I, weighted data. Table 1. Descriptive Statistics for Dependent and Independent Variables used in the Probit Selection Model (multiply imputed data)    Mean/Percent (full sample)  Mean/Percent (“heard” = 0)  Mean/Percent (“heard” = 1)  Mean/Percent (“deterrence” = 0)  Mean/Percent (“deterrence” = 1)  Dependent (outcome) variable                  Minutemen deter (%)  43  48  35  –  –  Selection variable                  Heard of Minutemen (%)  42  –  –  47  34  General human and financial capital                  Male (%)  87  80  97  93  80   Age (years)                   <25 (%)  24  29  17  23  25    25-34 (%)  41  35  48  40  42    35-44 (%)  24  25  23  28  18    45 + (%)  11  11  12  9  15   Log household income  6.1  5.7  6.6  6.4  5.7   Years of formal education  7.0  6.7  7.4  6.9  7.2   Indigenous language speaker (%)  20  27  10  21  19  Region of origin                  North (%)  14  8  22  15  13   West-Central (“traditional”) (%)  23  25  21  24  23   Central (%)  33  30  36  32  33   South (%)  30  37  20  29  31  Migration-specific social capital                  Family in destination (%)  58  54  63  55  62  Migration-specific human capital                  First crossing (%)  19  29  5  14  25   Apprehensions  4.4  2.8  6.6  5.1  3.4   Have lived in United States (%)  68  54  87  71  63   Current home in United States (%)  17  8  31  21  13  Context of last crossing attempt                  Used a coyote or guide (%)  70  80  57  65  77   Days traveled  2.4  2.5  2.2  2.4  2.5   Perceives crossing as very/ extremely dangerous (%)  81  82  78  80  81  Control                  Possible you will cross again? (%)  51  42  63  59  40  N  404  251  153  236  168     Mean/Percent (full sample)  Mean/Percent (“heard” = 0)  Mean/Percent (“heard” = 1)  Mean/Percent (“deterrence” = 0)  Mean/Percent (“deterrence” = 1)  Dependent (outcome) variable                  Minutemen deter (%)  43  48  35  –  –  Selection variable                  Heard of Minutemen (%)  42  –  –  47  34  General human and financial capital                  Male (%)  87  80  97  93  80   Age (years)                   <25 (%)  24  29  17  23  25    25-34 (%)  41  35  48  40  42    35-44 (%)  24  25  23  28  18    45 + (%)  11  11  12  9  15   Log household income  6.1  5.7  6.6  6.4  5.7   Years of formal education  7.0  6.7  7.4  6.9  7.2   Indigenous language speaker (%)  20  27  10  21  19  Region of origin                  North (%)  14  8  22  15  13   West-Central (“traditional”) (%)  23  25  21  24  23   Central (%)  33  30  36  32  33   South (%)  30  37  20  29  31  Migration-specific social capital                  Family in destination (%)  58  54  63  55  62  Migration-specific human capital                  First crossing (%)  19  29  5  14  25   Apprehensions  4.4  2.8  6.6  5.1  3.4   Have lived in United States (%)  68  54  87  71  63   Current home in United States (%)  17  8  31  21  13  Context of last crossing attempt                  Used a coyote or guide (%)  70  80  57  65  77   Days traveled  2.4  2.5  2.2  2.4  2.5   Perceives crossing as very/ extremely dangerous (%)  81  82  78  80  81  Control                  Possible you will cross again? (%)  51  42  63  59  40  N  404  251  153  236  168  Notes: Cases in which values were missing on the outcome and selection variables were included in the multiple imputation process and then omitted (n = 11) before estimating ”heckprob” procedure in Stata 13; m = 20. Standard errors available upon request. Source: Migrant Border Crossing Survey I, weighted data. Explanatory variables used in the selection model (“heard of Minutemen”) and outcome model (“potential deterrence”) are arranged in six conceptual groupings. Appendix Table A2 provides survey questions and metrics. The typical case in the inferential analysis is male (87 percent), between the age of 18 and 34 (65 percent combined categories), with seven years of education, reporting about $450 in monthly household income prior to the most recent crossing attempt. Twenty percent of respondents spoke an indigenous language. Fourteen percent were from northern Mexico, 23 percent from west-central Mexico, 33 percent from the central region, and 30 percent from the south. Respondents were relatively experienced border crossers (lifetime mean of 4.4 apprehensions). Roughly 19 percent were first-time crossers. Fifty-eight percent had family in their desired destination and 68 percent had previously lived and/or worked in the United States. Seventeen percent indicated their current home was in the United States, not Mexico. Most (70 percent) used a coyote to cross and, on average, traveled for 2.4 days prior to apprehension. As noted, sample characteristics vary according to prior knowledge of the Minutemen and whether or not respondents said they would be potentially deterred from attempting a future crossing if encountering Minutemen were a possibility (see Table 1, columns 2-5). Inferential Results for Heckman Probit Model with Sample Selection: Explaining “Having Heard” of the Minutemen (Selection Variable) Table 2 provides the coefficients of the Heckman probit model predicting “having heard” and “potential deterrence.”9 The p-value associated with the rho is statistically significant (rejecting rho = 0), suggesting sample selection needed to be corrected. However, for the sake of substantively interpreting the results, in Appendix Table A3 we present the average marginal effects (AMEs) for “having heard” (first column) and the conditional probability of “deterrence” given “having heard” (second column). Figures 1 through 4 provide “marginal effects at representative values” (MERs) for “having heard” by monthly household income, educational attainment, and lifetime apprehensions, and “potential deterrence” for monthly household income (for additional discussion of AMEs and MERs in probit models, see Williams 2012). Table 2. Heckman Probit Selection Model Results for “Heard of the Minutemen” (selection variable) and “Potential Deterrence” (dependent variable)    “Heard of MM”a  “Potential Deterrence”a     (selection)  (outcome)  Variables        General human and financial capital         Male  .64*  –    (.269)      Age (years)          25-34  .38*  –    (.176)       35-44  .22  –    (.201)       45+  .74*  –    (.313)      Log household income  .23***  −.50***    (.068)  (.121)   Years of formal education  .07***  –    (.021)      Indigenous language speaker  −.63***  –    (.195)     Region of origin         North  .64*  –    (.298)      Central  .51*  –    (.214)      South  .26  –    (.193)     Migration-specific social capital         Family in destination  .12  –    (.143)  –  Migration-specific human capital         First crossing  −.67***  .29    (.260)  (.395)   Apprehensions  .03†  −.03†    (.019)  (.017)   Have lived in United States  .48*  –    (.195)      Current home in United States  –  .51       (.306)  Context of last crossing attempt         Used a coyote or guide  −.50**  .58**    (.195)  (.218)   Days traveled  –  −.11†       (.068)   Perceives crossing as very/extremely dangerous  –  −.06       (.228)  Control         Possible that you will cross again?  –  −.51**       (.219)  Uncensored observations / censored observations  251  153  Rho = -.68        LR test of independent equations (rho = 0):         Chi2 (1) = 4.04         P > Chi2 = .0445        N = 404        M = 20 (results presented for 20th imputed data set)           “Heard of MM”a  “Potential Deterrence”a     (selection)  (outcome)  Variables        General human and financial capital         Male  .64*  –    (.269)      Age (years)          25-34  .38*  –    (.176)       35-44  .22  –    (.201)       45+  .74*  –    (.313)      Log household income  .23***  −.50***    (.068)  (.121)   Years of formal education  .07***  –    (.021)      Indigenous language speaker  −.63***  –    (.195)     Region of origin         North  .64*  –    (.298)      Central  .51*  –    (.214)      South  .26  –    (.193)     Migration-specific social capital         Family in destination  .12  –    (.143)  –  Migration-specific human capital         First crossing  −.67***  .29    (.260)  (.395)   Apprehensions  .03†  −.03†    (.019)  (.017)   Have lived in United States  .48*  –    (.195)      Current home in United States  –  .51       (.306)  Context of last crossing attempt         Used a coyote or guide  −.50**  .58**    (.195)  (.218)   Days traveled  –  −.11†       (.068)   Perceives crossing as very/extremely dangerous  –  −.06       (.228)  Control         Possible that you will cross again?  –  −.51**       (.219)  Uncensored observations / censored observations  251  153  Rho = -.68        LR test of independent equations (rho = 0):         Chi2 (1) = 4.04         P > Chi2 = .0445        N = 404        M = 20 (results presented for 20th imputed data set)        Notes: “<25 yrs”; “West-Central” region; and “no family in destination” are the referent groups. Coefficients, significance levels, and standard errors are reported, with the standard errors noted in parentheses. a 1 = yes; 0 = no † p < .10 *p < .05 **p < .01 ***p < .001 (two-tailed tests) Source: Migrant Border Crossing Survey I, unweighted data. Table 2. Heckman Probit Selection Model Results for “Heard of the Minutemen” (selection variable) and “Potential Deterrence” (dependent variable)    “Heard of MM”a  “Potential Deterrence”a     (selection)  (outcome)  Variables        General human and financial capital         Male  .64*  –    (.269)      Age (years)          25-34  .38*  –    (.176)       35-44  .22  –    (.201)       45+  .74*  –    (.313)      Log household income  .23***  −.50***    (.068)  (.121)   Years of formal education  .07***  –    (.021)      Indigenous language speaker  −.63***  –    (.195)     Region of origin         North  .64*  –    (.298)      Central  .51*  –    (.214)      South  .26  –    (.193)     Migration-specific social capital         Family in destination  .12  –    (.143)  –  Migration-specific human capital         First crossing  −.67***  .29    (.260)  (.395)   Apprehensions  .03†  −.03†    (.019)  (.017)   Have lived in United States  .48*  –    (.195)      Current home in United States  –  .51       (.306)  Context of last crossing attempt         Used a coyote or guide  −.50**  .58**    (.195)  (.218)   Days traveled  –  −.11†       (.068)   Perceives crossing as very/extremely dangerous  –  −.06       (.228)  Control         Possible that you will cross again?  –  −.51**       (.219)  Uncensored observations / censored observations  251  153  Rho = -.68        LR test of independent equations (rho = 0):         Chi2 (1) = 4.04         P > Chi2 = .0445        N = 404        M = 20 (results presented for 20th imputed data set)           “Heard of MM”a  “Potential Deterrence”a     (selection)  (outcome)  Variables        General human and financial capital         Male  .64*  –    (.269)      Age (years)          25-34  .38*  –    (.176)       35-44  .22  –    (.201)       45+  .74*  –    (.313)      Log household income  .23***  −.50***    (.068)  (.121)   Years of formal education  .07***  –    (.021)      Indigenous language speaker  −.63***  –    (.195)     Region of origin         North  .64*  –    (.298)      Central  .51*  –    (.214)      South  .26  –    (.193)     Migration-specific social capital         Family in destination  .12  –    (.143)  –  Migration-specific human capital         First crossing  −.67***  .29    (.260)  (.395)   Apprehensions  .03†  −.03†    (.019)  (.017)   Have lived in United States  .48*  –    (.195)      Current home in United States  –  .51       (.306)  Context of last crossing attempt         Used a coyote or guide  −.50**  .58**    (.195)  (.218)   Days traveled  –  −.11†       (.068)   Perceives crossing as very/extremely dangerous  –  −.06       (.228)  Control         Possible that you will cross again?  –  −.51**       (.219)  Uncensored observations / censored observations  251  153  Rho = -.68        LR test of independent equations (rho = 0):         Chi2 (1) = 4.04         P > Chi2 = .0445        N = 404        M = 20 (results presented for 20th imputed data set)        Notes: “<25 yrs”; “West-Central” region; and “no family in destination” are the referent groups. Coefficients, significance levels, and standard errors are reported, with the standard errors noted in parentheses. a 1 = yes; 0 = no † p < .10 *p < .05 **p < .01 ***p < .001 (two-tailed tests) Source: Migrant Border Crossing Survey I, unweighted data. On average, men’s probability of having heard of the Minutemen is 18 percentage points higher than women’s (see Appendix Table A3, column 1). Moreover, the probability of having heard of Minutemen is 11 percentage points higher for respondents 25-34 years old (compared to < 25 years old). Similarly, the probability of awareness is 20 percentage points higher for people age 45+ relative to those under the age of 25. Higher levels of human and financial capital are also associated with knowledge of Minutemen. Indigenous language speakers’ probability of having heard of Minutemen is 17 percentage points lower than non-indigenous language speakers. Respondents with higher household income prior to their most recent crossing and those with more formal educational attainment also had higher probabilities of having heard of the Minutemen (see Figures 1 and 2). Our results support the hypothesized positive associations between gender, income, and education, and prior knowledge of the Minutemen. Figure 1. View largeDownload slide Marginal Effects at Representative Values: “Having Heard of the Minutemen” by Monthly Household Income Figure 1. View largeDownload slide Marginal Effects at Representative Values: “Having Heard of the Minutemen” by Monthly Household Income Figure 2. View largeDownload slide Marginal Effects at Representative Values: “Having Heard of the Minutemen” by Years of Education Figure 2. View largeDownload slide Marginal Effects at Representative Values: “Having Heard of the Minutemen” by Years of Education Region of origin also structured access to information about Minutemen. Findings, however, do not completely support initial hypotheses. On average, northern Mexicans’ probability of awareness of the Minutemen is 18 percentage points higher than respondents from west-central Mexico (see Appendix Table A3, column 1), which is likely due to their proximity to the border. The “border zone” does not begin or end at the international divide but rather spans more than 100 kilometers into northern Mexico and the southwestern United States. Contrary to our hypothesis, however, central Mexicans’ probability of having heard of Minutemen is 14 percentage points higher than those from west-central Mexico. One explanation for this finding is that migration in central Mexico is also now occurring at levels fairly comparable to more established migrant-sending communities (Martell etal. 2007). These contemporary migration patterns may be quickly eroding the general and migration-specific social and human capital advantages long attributed to traditional sending regions. Additionally, perhaps, “sending regions”—as measured in our survey—do not adequately capture with enough precision these changing social processes. Migration-specific social capital—specifically, family ties in one’s destination—was not statistically significantly associated with prior knowledge. However, we did find that greater migration-specific human capital—being a non-first-time crosser, having more lifetime apprehensions, and having previously lived in the United States—was positively associated with prior knowledge of Minutemen. In fact, first-time crossers have a probability of having heard of Minutemen that is 19 percentage points lower than more experienced migrants (see Appendix Table A3, column 1). On the other hand, the probability of having heard of the Minutemen increases as lifetime apprehensions increase (see Figure 3). Figure 3. View largeDownload slide Marginal Effects at Representative Values: “Having Heard of the Minutemen” by Number of Prior Apprehensions Figure 3. View largeDownload slide Marginal Effects at Representative Values: “Having Heard of the Minutemen” by Number of Prior Apprehensions Experience living in the United States is another form of migration-specific human capital explaining differential levels of Minutemen awareness. The probability of awareness is, on average, 13 percentage points higher for migrants who have lived in the United States (versus migrants that have not) (see Appendix Table A3, column 1). While anti-immigrant mobilization efforts and the subsequent 2006 immigration marches across the United States did attract notable media attention throughout Mexico, attention was likely greater in the United States, especially among Latinos. Roughly 86 percent of MBCS respondents who said they had lived in the United States were living there at some point between 2005 and 2007, during both the growth of anti-immigrant mobilization and pro-immigration marches. Repatriated unauthorized migrants’ awareness of nativist groups was likely heightened as a consequence of having lived in the United States during this pivotal period. Among those who had lived in the United States, but not during the 2005-2007 period, it is likely that they were paying close attention to the immigration issue as they prepared for a future crossing prior to being surveyed. Together, findings suggest that migrants’ knowledge about anti-immigrant mobilization increases with first-hand migration experience. Lastly, the context of the last crossing experience helps explain prior knowledge of Minutemen. Surprisingly, we found that coyote users’ probability of having heard of the Minutemen was, on average, 14 percentage points lower compared to respondents not using a guide during their most recent crossing attempt (even when controlling for migration-specific social and human capital) (see Appendix Table A3, column 1). Although coyotes can be seen as producers and transmitters of knowledge about crossings, coyotaje is becoming a more organized informal business along the Sonora-Arizona border (Martínez 2016). While encountering Minutemen is a possibility, it is unlikely—something experienced coyotes understand but migrants may not. Indeed, of our 415 respondents interviewed, only 3 encountered Minutemen while crossing. Savvy coyotes would not want to call attention to an obstacle—especially one that is not likely to arise—that could potentially discourage migrants from using their services. Ultimately, more research is needed to fully explain this phenomenon, especially because coyotes could be plausibly motivated to exaggerate the dangers of the journey in order to emphasize how much their services are needed. Inferential Results for Heckman Probit Model with Sample Selection: The Conditional Probability of “Potential Deterrence” (Outcome Variable) Given “Having Heard” (Selection Variable) We were also interested in identifying factors explaining respondents’ potential deterrence from attempting a future crossing if encountering Minutemen were a possibility when correcting for sample selection for having heard of the group. We address this second question as the “outcome” variable in our Heckman probit selection model (see Table 2, column 2 and Appendix Table A3). First, having spent more time crossing the border and indicating that attempting a future crossing at a later date was a possibility decreased the conditional probability of potential deterrence given prior knowledge of the Minutemen, whereas believing one’s home was in the United States had no effect. Each additional day spent crossing the border was associated with a decrease in the probability of potential deterrence of 4 percentage points, while migrants noting it was possible they may cross again, on average, had probabilities of potential deterrence that were 20 percentage points lower than migrants who did not think or know if they would cross again (see Appendix Table A3, column 2). Migrants travelling for longer periods as well as those indicating they may cross again have likely invested greater resources into clandestine crossings. Second, greater income and being an indigenous language speaker help foster resiliency in the face of anti-immigrant mobilization. Migrants with greater access to financial resources can draw on those resources to facilitate unauthorized crossing attempts, despite a potential encounter with Minutemen that could lead to their removal from the United States (illustrated in Figure 4). In a similar vein, indigenous language speakers had probabilities of potential deterrence that were 12 percentage points lower than non-indigenous language speakers. Indigenous language speakers’ marginalized socioeconomic status in Mexico may operate as a powerful “push” factor compelling them to migrate despite the possibility of encountering a mobilized nativist citizenry. Figure 4. View largeDownload slide Marginal Effects at Representative Values: “Potential Deterrence” by Monthly Household Income Figure 4. View largeDownload slide Marginal Effects at Representative Values: “Potential Deterrence” by Monthly Household Income We uncovered a counterintuitive finding related to coyote use. Traveling with a coyote is associated with higher probabilities of potential deterrence. On average, coyote users’ probability of potential deterrence is 13 percentage points higher than migrants not using a coyote (see Appendix Table A3, column 2). We expected that migrants travelling with coyotes would be less likely to be deterred because guides should mitigate physical and social risks during the journey. Conversely, migrants travelling with coyotes are often relatively less experienced and thus more susceptible to anti-immigrant mobilization. Below, we elaborate on this finding and suggest that future research is needed to disentangle this paradox. CONCLUSION We know little about the extra-political consequences of nativist mobilization. Addressing this gap, we developed novel measures of nativist movement influence incorporating unauthorized repatriated migrants' perceptions of and potential reactions to such mobilization. Our analyses provided insight into the extra-political consequences of contemporary nativist mobilization and, in so doing, addressed long-debated questions: Do social movements (specifically, nativist movements) matter? If so, how (might they intersect with social processes of unauthorized migration)? Contemporary nativist movement influence is partially a story of gaining national media attention and raising awareness (Chavez 2008). Minutemen succeeded in generating awareness of their presence along the U.S.-Mexico border among the very population they sought to deter—nearly half of migrants in our survey had heard of them. However, we do not wish to understate the movement’s real political consequences. For one, nativist mobilization likely helped increase government funding for the U.S. Border Patrol (Doty 2009; Shapira 2013). More broadly, however, nativist organizations succeeded in reframing what was in the realm of possibility for immigration reform. Since 2005, individual and package immigration reforms have frequently included a border militarization component. This can partially be attributed to nativist mobilization normalizing exclusionary politics and strengthening the link between immigration control and the larger conservative political platform.10 However, it always seemed unlikely that nativist mobilization could abate migration generated by powerful economic, social, and political forces. In this regard, our results illuminate how unauthorized repatriated migrants and the border-crossing context intersect to mitigate nativist influence. Financial and migration-specific capital and crossing experiences can shield unauthorized repatriated migrants, making them less susceptible to potential deterrence. By examining how resiliency in the face of potential deterrents is structured among repatriated unauthorized migrants, our work demonstrates, more broadly, how relatively powerless groups—through, for instance, financial and human capital deployment and decisions about how to cross—can mitigate potential threats initiated by relatively privileged groups of U.S. citizens. However, one counterintuitive finding we encountered was that traveling with a coyote was associated with higher probabilities of potential deterrence. Previous research suggests coyotes help mitigate risks during clandestine crossings. It may well be the case that those using guides are relatively less experienced and thus perceive themselves to be more susceptible to nativist mobilization. It is also plausible that those that use coyotes are risk averse. If this is the case, this would explain why they hired coyotes in the first place and why they are likely to be potentially deterred by nativist groups. And because our survey instrument does not probe deeper into the interactions between migrants and coyotes, future research is needed to fully understand this finding. Second, our findings illuminate the inequalities intertwined with the clandestine migration process and reveal how these are structured. Unauthorized migrants do not have equal access to information about non-infrastructural, non-environmental deterrents in the form of a xenophobic and mobilized U.S. citizenry. Forty-two percent of respondents had previously heard of Minutemen. Of course, given how infrequently Minutemen actually encounter border crossers (based on our research and that of Shapira 2013), such mobilization along the border is unlikely to be as consequential as, for instance, the presence of a militarized border patrol or harsh crossing conditions. Nevertheless, discrepancies in prior knowledge of Minutemen are largely explained by higher levels of general human and financial capital, and migration-specific human capital—in the form of greater prior crossing experience, more lifetime apprehensions, and experience living in the United States. Although it should be acknowledged that because our sample comes from repatriated unauthorized migrants—individuals that did not “successfully” cross during their most recent attempt—it may be possible that the sample underrepresents the proportion of individuals with knowledge of the Minutemen, which one might presume would be higher among migrants that were able to cross without being apprehended and deported. Experiences associated with a migrant’s most recent crossing also help explain knowledge disparities. Unauthorized repatriated migrants who crossed without (versus with) coyotes had a higher probability of being aware of Minutemen. This paradoxical finding may be the result of financial pressures encouraging savvy coyotes to protect their business by keeping migrants unaware of potential dangers. With the emergence of “border business” coyotes—triggered by border enforcement initiatives that are funneling unauthorized migration and drug trafficking into the same geographic space—future research will want to investigate the complex and shifting relations between coyotes and their clients. Collectively, our findings illuminate both the structured inequalities—in terms of access to valued information about deterrents—shaping the unauthorized migration process, as well as why some unauthorized repatriated migrants are more or less resilient in the face of potential deterrents to clandestine mobilization. Scholars of migration and, in particular, those interested in the role various forms of capital play in the clandestine migration process, will want to further explore these issues by examining whether the same forms of capital we find important also play a significant role in the acquisition of knowledge about other—potentially more consequential—hurdles in the clandestine migration process (e.g., U.S. Border Patrol, violence/robbery, extreme weather, etc.). Is migrants’ knowledge about and resiliency towards a variety of potential threats shaped by some general set of capital-based factors (i.e., one, two, or three specific forms of capital) and/or are disparities in knowledge and differential levels of resiliency uniquely structured according to some general qualities of the hurdle being faced, like the rate (or probability) of encountering the hurdle, the potential costs associated with running into the hurdle, etc.? Finally, our research has broader implications for the study of social movements, particularly exclusionary right-wing mobilization. Social movement literature largely separates movements into two distinct categories—those targeting the state to effect change and those targeting civil society to transform culture and affirm identities. As such, the relationship between social movements and broader social change is typically assessed in terms of political change, on the one hand, and cultural change on the other (Amenta etal. 2010; Earl 2004). While this analytic grouping makes sense in the context of more progressive mobilization by disadvantaged groups, our research demonstrates that right-wing exclusionary movements too should be evaluated in terms of the extent to which they influence the behavior of targeted social groups. This includes not only the behavior of state actors and (potential) movement participants, but also the movement's targeted opponents, which constitute an important piece of the field in which movements operate. The authors wish to Jeremy Slack, Prescott Vandervoett, Kristin Klingman, Shiras Manning, Paola Molina, Kristen Valencia, Kylie Walzak, Melissa Burham, and Lorenzo Gamboa for their help with the data collection process. They also thank Kraig Beyerlein, Hana Brown, and the blind reviewers of Social Problems for their comments and suggestions. The authors express their gratitude to the owners and staff of the migrant shelter where they conducted their fieldwork as well as to the hundreds of people that took the time to discuss their migration experiences with the research team. This work was supported by the University of Arizona’s Underrepresented Graduate Student Final Project Fund as well as the Programa de Investigación de Migración y Salud (PIMSA), Health Initiative of the Americas at the University of California, Berkley. Direct correspondence to: Daniel E Martínez, School of Sociology, The University of Arizona, Social Sciences Building, Room 400, Tucson, AZ 85721. E-mail: daniel.martinez@arizona.edu. 1 Elsewhere (Ward and Martínez 2015) we discuss how repatriated unauthorized Mexican migrants perceive Minutemen and how they obtain information on nativist mobilization. Migrants get most of their information through U.S.-based media (and to a lesser extent the migration process itself or friends and family). Migrants discussed Minutemen by referencing ethno-racial categories (i.e., white, Chicano, etc.), specific roles or duties (i.e., preventing migrants from crossing, etc.), or by noting negative attitudes towards migrants and Mexicans (i.e., suggesting Minutemen are racist, etc.). 2 We do not address actual deterrence. We are talking about a potential, hypothetical encounter. Although intentions do not perfectly predict behavior, the theory of planned behavior suggests intentions are the most proximate determinant of behavior (Ajzen 1988). 3 Brown’s (2013) examination of the policy consequences of U.S. anti-immigrant mobilization is an exception. Still, her focus remains strictly on political consequences. Minuteman organizations—and the larger movement of which they were a visible part—framed the immigration reform debates of 2006-2007 (and even more recently) as matters of national security. They succeeded in having many of their proposals for increased border enforcement passed (or at least in getting favorable compromises passed). While Minutemen were not the sole force behind these surges, they remained one of the most visible public faces of the contemporary nativist movement. 4 From 2007-2010, MMP and MCDC chapters existed in most states. Of the 319 anti-immigration groups in 2010, 128 were connected with MMP or MCDC (Beirich 2011). 5 Less media attention has been paid to groups such as the Minutemen in Mexican media outlets relative to those in the United States. Rather, Mexican media has tended to focus their efforts on rallying against the expansion of the border wall (Trujeque Díaz 2007). 6 We focus exclusively on Mexican migrants because surveys were conducted in Mexico, and non-Mexicans are not repatriated to Mexico. Moreover, we are not generalizing to migrants in sending communities who have yet to migrate, migrants who were apprehended and repatriated along other regions of the border, or unauthorized migrants who crossed but were never detected by U.S. authorities. Nevertheless, heightened border enforcement efforts over the last few decades have increased the likelihood of apprehension. According to the U.S. Border Patrol, the current probability of apprehension ranges between 67 percent and 86 percent, which is substantially higher than older estimates based on self-report surveys collected through Princeton’s Mexican Migration Project and UC San Diego’s Mexican Migration Field Research Program (Rosenblum 2013:24). 7 Marginal differences exist between weighted and unweighted models (Winship and Radbill 1994), but the key substantive findings remain consistent. 8 To our knowledge no studies provide systematic quantitative data on encounters between Minutemen and unauthorized migrants. Our data suggest encounters are rare. Only three respondents who had heard of the Minutemen indicated they encountered them during their most recent crossing experience, two of which indicated they would not be deterred from a future crossing. 9 We tested for collinearity among variables prior to MI using the variance inflation factor (VIF). Collinearity was not an issue, as none of the VIFs exceeded 2.12 (Menard 1995). 10 Although the militarization of the border was underway prior to the emergence of contemporary nativist mobilization (Dunn 1996), we contend that nativist mobilization efforts further contribute to border militarization. APPENDIX Table A1. Descriptive Statistics for Selected MBCS Variables Used in Analyses (non-imputed and non-weighted data)    Mean/Percent  Std. Dev.  Missing Observations (%)  Dependent variable            Minutemen deter  .54  .610  1.93  Selection variable            Heard of Minutemen  .38  .487  1.20  General human and financial capital            Male  .88  .320  .00   Age (years)             < 25  .30  .458  .00    25-34  .39  .489  .00    35-44  .24  .425  .00    45+  .07  .259  .00   Log household income  6.00  1.153  13.01   Years of formal education  6.93  3.592  2.17   Indigenous language speaker  .21  .410  .48  Region of origin            North  .09  .287  .96   West-Central (“traditional”)  .20  .398  .96   Central  .20  .403  .96   South  .51  .500  .96  Migration-specific social capital            Family in destination  .54  .499  .24  Migration-specific human capital            First crossing  .22  .413  .00   Apprehensions  3.96  5.930  .00   Have lived in United States  .63  .483  .48   Current home in United States  .13  .332  17.34  Context of last crossing attempt            Used a coyote or guide  .72  .448  .96   Days traveled  2.49  1.677  .48   Perceives crossing as very/extremely dangerous  .80  .405  2.65  Control            Possible you will cross again?  .48  .500  4.58  N = 415           Mean/Percent  Std. Dev.  Missing Observations (%)  Dependent variable            Minutemen deter  .54  .610  1.93  Selection variable            Heard of Minutemen  .38  .487  1.20  General human and financial capital            Male  .88  .320  .00   Age (years)             < 25  .30  .458  .00    25-34  .39  .489  .00    35-44  .24  .425  .00    45+  .07  .259  .00   Log household income  6.00  1.153  13.01   Years of formal education  6.93  3.592  2.17   Indigenous language speaker  .21  .410  .48  Region of origin            North  .09  .287  .96   West-Central (“traditional”)  .20  .398  .96   Central  .20  .403  .96   South  .51  .500  .96  Migration-specific social capital            Family in destination  .54  .499  .24  Migration-specific human capital            First crossing  .22  .413  .00   Apprehensions  3.96  5.930  .00   Have lived in United States  .63  .483  .48   Current home in United States  .13  .332  17.34  Context of last crossing attempt            Used a coyote or guide  .72  .448  .96   Days traveled  2.49  1.677  .48   Perceives crossing as very/extremely dangerous  .80  .405  2.65  Control            Possible you will cross again?  .48  .500  4.58  N = 415        Source: Migrant Border Crossing Survey I. View Large Table A1. Descriptive Statistics for Selected MBCS Variables Used in Analyses (non-imputed and non-weighted data)    Mean/Percent  Std. Dev.  Missing Observations (%)  Dependent variable            Minutemen deter  .54  .610  1.93  Selection variable            Heard of Minutemen  .38  .487  1.20  General human and financial capital            Male  .88  .320  .00   Age (years)             < 25  .30  .458  .00    25-34  .39  .489  .00    35-44  .24  .425  .00    45+  .07  .259  .00   Log household income  6.00  1.153  13.01   Years of formal education  6.93  3.592  2.17   Indigenous language speaker  .21  .410  .48  Region of origin            North  .09  .287  .96   West-Central (“traditional”)  .20  .398  .96   Central  .20  .403  .96   South  .51  .500  .96  Migration-specific social capital            Family in destination  .54  .499  .24  Migration-specific human capital            First crossing  .22  .413  .00   Apprehensions  3.96  5.930  .00   Have lived in United States  .63  .483  .48   Current home in United States  .13  .332  17.34  Context of last crossing attempt            Used a coyote or guide  .72  .448  .96   Days traveled  2.49  1.677  .48   Perceives crossing as very/extremely dangerous  .80  .405  2.65  Control            Possible you will cross again?  .48  .500  4.58  N = 415           Mean/Percent  Std. Dev.  Missing Observations (%)  Dependent variable            Minutemen deter  .54  .610  1.93  Selection variable            Heard of Minutemen  .38  .487  1.20  General human and financial capital            Male  .88  .320  .00   Age (years)             < 25  .30  .458  .00    25-34  .39  .489  .00    35-44  .24  .425  .00    45+  .07  .259  .00   Log household income  6.00  1.153  13.01   Years of formal education  6.93  3.592  2.17   Indigenous language speaker  .21  .410  .48  Region of origin            North  .09  .287  .96   West-Central (“traditional”)  .20  .398  .96   Central  .20  .403  .96   South  .51  .500  .96  Migration-specific social capital            Family in destination  .54  .499  .24  Migration-specific human capital            First crossing  .22  .413  .00   Apprehensions  3.96  5.930  .00   Have lived in United States  .63  .483  .48   Current home in United States  .13  .332  17.34  Context of last crossing attempt            Used a coyote or guide  .72  .448  .96   Days traveled  2.49  1.677  .48   Perceives crossing as very/extremely dangerous  .80  .405  2.65  Control            Possible you will cross again?  .48  .500  4.58  N = 415        Source: Migrant Border Crossing Survey I. View Large Table A2. Descriptions for Dependent, Selection, and Independent Variables Used in the Heckman Probit Model    Survey Question  Metric  Dependent/selection variables         Heard of the Minutemen  "Have you ever heard of the Minutemen?"  1 = "yes"; 0 = "no"   Minutemen deter  “Would the possibility of encountering the Minutemen deter you from attempting a future unauthorized crossing?”  1 = "yes"; 0 = "no / don't know"  Independent variables        Male  "Is the respondent male?"     Age (years)         % < 25  "Is the respondent less than 25 years old?"  1 = "yes"; 0 = "no"    % 25-34  "Is the respondent between 25 and 34 years old?"  1 = "yes"; 0 = "no"    % 35-44  "Is the respondent between 35 and 44 years old?"  1 = "yes"; 0 = "no"    % 45+  "Is the respondent more than 44 years old?"  1 = "yes"; 0 = "no"   Household income (in log dollars)  "What was your monthly household income before your last crossing attempt?"  Count (in log dollars)   Education  "How many years of formal education have you completed?"  Count (in years)   Indigenous language speaker  "Do you speak an indigenous language?"  1 = "yes"; 0 = "no"   North  "Is the respondent from the north region of Mexico?"  1 = "yes"; 0 = "no"   West-Central (“traditional”)  "Is the respondent from the west-central region of Mexico?"  1 = "yes"; 0 = "no"   Central  "Is the respondent from the central region of Mexico?"  1 = "yes"; 0 = "no"   South  "Is the respondent from the south region of Mexico?"  1 = "yes"; 0 = "no"   Family in destination  "Do you have family in your desired destination?"     First crossing  "Was your last crossing attempt your first?"  1 = "yes"; 0 = "no"   Number of lifetime apprehensions  "How many times have you been apprehended by any U.S. authority, including your most recent apprehension?"  Count (number of times)   Have lived in United States  "Have you ever lived or worked in the United States?"     Current home in United States  "Is your current home in the United States?"  1 = "yes"; 0 = "no"   Used a coyote or guide  "Did you use a coyote or guide during your last crossing attempt?"  1 = "yes"; 0 = "no"  Days traveled  "How many days did you spend traveling in the desert before being apprehended or picked up to proceed to the next leg of the journey?"  Count (number of days)  Perceives crossing as very/extremely dangerous  "Respondent perceives the crossing as ‘very’ or ‘extremely dangerous’ after their most recent attempt."  1 = "yes"; 0 = "no"  Possible you will cross again?  "Is it possible that you will cross the border again without papers?"  1 = "yes"; 0 = "no"     Survey Question  Metric  Dependent/selection variables         Heard of the Minutemen  "Have you ever heard of the Minutemen?"  1 = "yes"; 0 = "no"   Minutemen deter  “Would the possibility of encountering the Minutemen deter you from attempting a future unauthorized crossing?”  1 = "yes"; 0 = "no / don't know"  Independent variables        Male  "Is the respondent male?"     Age (years)         % < 25  "Is the respondent less than 25 years old?"  1 = "yes"; 0 = "no"    % 25-34  "Is the respondent between 25 and 34 years old?"  1 = "yes"; 0 = "no"    % 35-44  "Is the respondent between 35 and 44 years old?"  1 = "yes"; 0 = "no"    % 45+  "Is the respondent more than 44 years old?"  1 = "yes"; 0 = "no"   Household income (in log dollars)  "What was your monthly household income before your last crossing attempt?"  Count (in log dollars)   Education  "How many years of formal education have you completed?"  Count (in years)   Indigenous language speaker  "Do you speak an indigenous language?"  1 = "yes"; 0 = "no"   North  "Is the respondent from the north region of Mexico?"  1 = "yes"; 0 = "no"   West-Central (“traditional”)  "Is the respondent from the west-central region of Mexico?"  1 = "yes"; 0 = "no"   Central  "Is the respondent from the central region of Mexico?"  1 = "yes"; 0 = "no"   South  "Is the respondent from the south region of Mexico?"  1 = "yes"; 0 = "no"   Family in destination  "Do you have family in your desired destination?"     First crossing  "Was your last crossing attempt your first?"  1 = "yes"; 0 = "no"   Number of lifetime apprehensions  "How many times have you been apprehended by any U.S. authority, including your most recent apprehension?"  Count (number of times)   Have lived in United States  "Have you ever lived or worked in the United States?"     Current home in United States  "Is your current home in the United States?"  1 = "yes"; 0 = "no"   Used a coyote or guide  "Did you use a coyote or guide during your last crossing attempt?"  1 = "yes"; 0 = "no"  Days traveled  "How many days did you spend traveling in the desert before being apprehended or picked up to proceed to the next leg of the journey?"  Count (number of days)  Perceives crossing as very/extremely dangerous  "Respondent perceives the crossing as ‘very’ or ‘extremely dangerous’ after their most recent attempt."  1 = "yes"; 0 = "no"  Possible you will cross again?  "Is it possible that you will cross the border again without papers?"  1 = "yes"; 0 = "no"  Source: Migrant Border Crossing Survey I. View Large Table A2. Descriptions for Dependent, Selection, and Independent Variables Used in the Heckman Probit Model    Survey Question  Metric  Dependent/selection variables         Heard of the Minutemen  "Have you ever heard of the Minutemen?"  1 = "yes"; 0 = "no"   Minutemen deter  “Would the possibility of encountering the Minutemen deter you from attempting a future unauthorized crossing?”  1 = "yes"; 0 = "no / don't know"  Independent variables        Male  "Is the respondent male?"     Age (years)         % < 25  "Is the respondent less than 25 years old?"  1 = "yes"; 0 = "no"    % 25-34  "Is the respondent between 25 and 34 years old?"  1 = "yes"; 0 = "no"    % 35-44  "Is the respondent between 35 and 44 years old?"  1 = "yes"; 0 = "no"    % 45+  "Is the respondent more than 44 years old?"  1 = "yes"; 0 = "no"   Household income (in log dollars)  "What was your monthly household income before your last crossing attempt?"  Count (in log dollars)   Education  "How many years of formal education have you completed?"  Count (in years)   Indigenous language speaker  "Do you speak an indigenous language?"  1 = "yes"; 0 = "no"   North  "Is the respondent from the north region of Mexico?"  1 = "yes"; 0 = "no"   West-Central (“traditional”)  "Is the respondent from the west-central region of Mexico?"  1 = "yes"; 0 = "no"   Central  "Is the respondent from the central region of Mexico?"  1 = "yes"; 0 = "no"   South  "Is the respondent from the south region of Mexico?"  1 = "yes"; 0 = "no"   Family in destination  "Do you have family in your desired destination?"     First crossing  "Was your last crossing attempt your first?"  1 = "yes"; 0 = "no"   Number of lifetime apprehensions  "How many times have you been apprehended by any U.S. authority, including your most recent apprehension?"  Count (number of times)   Have lived in United States  "Have you ever lived or worked in the United States?"     Current home in United States  "Is your current home in the United States?"  1 = "yes"; 0 = "no"   Used a coyote or guide  "Did you use a coyote or guide during your last crossing attempt?"  1 = "yes"; 0 = "no"  Days traveled  "How many days did you spend traveling in the desert before being apprehended or picked up to proceed to the next leg of the journey?"  Count (number of days)  Perceives crossing as very/extremely dangerous  "Respondent perceives the crossing as ‘very’ or ‘extremely dangerous’ after their most recent attempt."  1 = "yes"; 0 = "no"  Possible you will cross again?  "Is it possible that you will cross the border again without papers?"  1 = "yes"; 0 = "no"     Survey Question  Metric  Dependent/selection variables         Heard of the Minutemen  "Have you ever heard of the Minutemen?"  1 = "yes"; 0 = "no"   Minutemen deter  “Would the possibility of encountering the Minutemen deter you from attempting a future unauthorized crossing?”  1 = "yes"; 0 = "no / don't know"  Independent variables        Male  "Is the respondent male?"     Age (years)         % < 25  "Is the respondent less than 25 years old?"  1 = "yes"; 0 = "no"    % 25-34  "Is the respondent between 25 and 34 years old?"  1 = "yes"; 0 = "no"    % 35-44  "Is the respondent between 35 and 44 years old?"  1 = "yes"; 0 = "no"    % 45+  "Is the respondent more than 44 years old?"  1 = "yes"; 0 = "no"   Household income (in log dollars)  "What was your monthly household income before your last crossing attempt?"  Count (in log dollars)   Education  "How many years of formal education have you completed?"  Count (in years)   Indigenous language speaker  "Do you speak an indigenous language?"  1 = "yes"; 0 = "no"   North  "Is the respondent from the north region of Mexico?"  1 = "yes"; 0 = "no"   West-Central (“traditional”)  "Is the respondent from the west-central region of Mexico?"  1 = "yes"; 0 = "no"   Central  "Is the respondent from the central region of Mexico?"  1 = "yes"; 0 = "no"   South  "Is the respondent from the south region of Mexico?"  1 = "yes"; 0 = "no"   Family in destination  "Do you have family in your desired destination?"     First crossing  "Was your last crossing attempt your first?"  1 = "yes"; 0 = "no"   Number of lifetime apprehensions  "How many times have you been apprehended by any U.S. authority, including your most recent apprehension?"  Count (number of times)   Have lived in United States  "Have you ever lived or worked in the United States?"     Current home in United States  "Is your current home in the United States?"  1 = "yes"; 0 = "no"   Used a coyote or guide  "Did you use a coyote or guide during your last crossing attempt?"  1 = "yes"; 0 = "no"  Days traveled  "How many days did you spend traveling in the desert before being apprehended or picked up to proceed to the next leg of the journey?"  Count (number of days)  Perceives crossing as very/extremely dangerous  "Respondent perceives the crossing as ‘very’ or ‘extremely dangerous’ after their most recent attempt."  1 = "yes"; 0 = "no"  Possible you will cross again?  "Is it possible that you will cross the border again without papers?"  1 = "yes"; 0 = "no"  Source: Migrant Border Crossing Survey I. View Large Table A3. Average Marginal Effects for Heckman Probit Selection Model Explaining “Potential Deterrence” (outcome) and “Having Heard of the Minutemen” (selection) Variables  Pr  Pr  (heard = 1)  (deterrence = 1 / heard = 1)  General human and financial capital         Male  .18*  .12    (.074)  (.088)   Age (years)          25-34  .11*  .07    (.048)  (.053)    35-44  .06  .04    (.056)  (.044)    45+  .20*  .14    (.085)  (.089)   Log household income  .06***  −.15***    (.018)  (.039)   Years of formal education  .02***  .01    (.005)  (.008)   Indigenous language speaker  −.17***  −.12†    (.052)  (.071)  Region of origin         North  .18*  .12    (.082)  (.086)   Central  .14*  .10    (.058)  (.066)   South  .07  .05    (.053)  (.046)  Migration-specific social capital         Family in destination  .03  .02    (.040)  (.030)  Migration-specific human capital         First crossing  −.19**  −.01    (.071)  (.143)   Apprehensions  .01†  −.01    (.005)  (.007)   Have lived in United States  .13*  .09    (.053)  (.057)   Current home in United States  –  .20       (.109)  Context of last crossing attempt         Used a coyote or guide  −.14**  .13†    (.044)  (.079)   Days traveled  –  −.04†       (.024)   Perceives crossing as very/extremely dangerous  –  −.02       (.090)  Control         Possible that you will cross again?  –  −.20**       (.075)  N = 404      M = 20 (results presented for 20th imputed data set)        Variables  Pr  Pr  (heard = 1)  (deterrence = 1 / heard = 1)  General human and financial capital         Male  .18*  .12    (.074)  (.088)   Age (years)          25-34  .11*  .07    (.048)  (.053)    35-44  .06  .04    (.056)  (.044)    45+  .20*  .14    (.085)  (.089)   Log household income  .06***  −.15***    (.018)  (.039)   Years of formal education  .02***  .01    (.005)  (.008)   Indigenous language speaker  −.17***  −.12†    (.052)  (.071)  Region of origin         North  .18*  .12    (.082)  (.086)   Central  .14*  .10    (.058)  (.066)   South  .07  .05    (.053)  (.046)  Migration-specific social capital         Family in destination  .03  .02    (.040)  (.030)  Migration-specific human capital         First crossing  −.19**  −.01    (.071)  (.143)   Apprehensions  .01†  −.01    (.005)  (.007)   Have lived in United States  .13*  .09    (.053)  (.057)   Current home in United States  –  .20       (.109)  Context of last crossing attempt         Used a coyote or guide  −.14**  .13†    (.044)  (.079)   Days traveled  –  −.04†       (.024)   Perceives crossing as very/extremely dangerous  –  −.02       (.090)  Control         Possible that you will cross again?  –  −.20**       (.075)  N = 404      M = 20 (results presented for 20th imputed data set)        Notes: “<25 yrs”; “West-Central” region; and “no family in destination” are the referent groups. Average marginal effects (AME), significance levels, and standard errors are reported, with the standard errors noted in parentheses. † p < .10 *p < .05 **p < .01 ***p < .001 (two-tailed tests) Source: Migrant Border Crossing Survey I, unweighted data View Large Table A3. Average Marginal Effects for Heckman Probit Selection Model Explaining “Potential Deterrence” (outcome) and “Having Heard of the Minutemen” (selection) Variables  Pr  Pr  (heard = 1)  (deterrence = 1 / heard = 1)  General human and financial capital         Male  .18*  .12    (.074)  (.088)   Age (years)          25-34  .11*  .07    (.048)  (.053)    35-44  .06  .04    (.056)  (.044)    45+  .20*  .14    (.085)  (.089)   Log household income  .06***  −.15***    (.018)  (.039)   Years of formal education  .02***  .01    (.005)  (.008)   Indigenous language speaker  −.17***  −.12†    (.052)  (.071)  Region of origin         North  .18*  .12    (.082)  (.086)   Central  .14*  .10    (.058)  (.066)   South  .07  .05    (.053)  (.046)  Migration-specific social capital         Family in destination  .03  .02    (.040)  (.030)  Migration-specific human capital         First crossing  −.19**  −.01    (.071)  (.143)   Apprehensions  .01†  −.01    (.005)  (.007)   Have lived in United States  .13*  .09    (.053)  (.057)   Current home in United States  –  .20       (.109)  Context of last crossing attempt         Used a coyote or guide  −.14**  .13†    (.044)  (.079)   Days traveled  –  −.04†       (.024)   Perceives crossing as very/extremely dangerous  –  −.02       (.090)  Control         Possible that you will cross again?  –  −.20**       (.075)  N = 404      M = 20 (results presented for 20th imputed data set)        Variables  Pr  Pr  (heard = 1)  (deterrence = 1 / heard = 1)  General human and financial capital         Male  .18*  .12    (.074)  (.088)   Age (years)          25-34  .11*  .07    (.048)  (.053)    35-44  .06  .04    (.056)  (.044)    45+  .20*  .14    (.085)  (.089)   Log household income  .06***  −.15***    (.018)  (.039)   Years of formal education  .02***  .01    (.005)  (.008)   Indigenous language speaker  −.17***  −.12†    (.052)  (.071)  Region of origin         North  .18*  .12    (.082)  (.086)   Central  .14*  .10    (.058)  (.066)   South  .07  .05    (.053)  (.046)  Migration-specific social capital         Family in destination  .03  .02    (.040)  (.030)  Migration-specific human capital         First crossing  −.19**  −.01    (.071)  (.143)   Apprehensions  .01†  −.01    (.005)  (.007)   Have lived in United States  .13*  .09    (.053)  (.057)   Current home in United States  –  .20       (.109)  Context of last crossing attempt         Used a coyote or guide  −.14**  .13†    (.044)  (.079)   Days traveled  –  −.04†       (.024)   Perceives crossing as very/extremely dangerous  –  −.02       (.090)  Control         Possible that you will cross again?  –  −.20**       (.075)  N = 404      M = 20 (results presented for 20th imputed data set)        Notes: “<25 yrs”; “West-Central” region; and “no family in destination” are the referent groups. Average marginal effects (AME), significance levels, and standard errors are reported, with the standard errors noted in parentheses. † p < .10 *p < .05 **p < .01 ***p < .001 (two-tailed tests) Source: Migrant Border Crossing Survey I, unweighted data View Large REFERENCES Ajzen Icek. 1988. Attitudes, Personality, and Behavior . Milton Keynes, UK: Open University Press. Amenta Edwin, Caren Neal, Chiarello Elizabeth, Su Yang. 2010. “The Political Consequences of Social Movements.” Annual Review of Sociology  36: 287- 307. Google Scholar CrossRef Search ADS   Andreas Peter. 2009. Border Games.  Ithaca, NY: Cornell University Press. Baumgartner Frank, Mahoney Christine. 2005. “Social Movements, the Rise of New Issues, and the Public Agenda.” Pp. 65- 86 in Routing the Opposition , edited by Meyer D., Jenness V., Ingram H. Minneapolis: University of Minnesota Press. Beirich Heidi. 2011. “The Year in Nativism, 2010.” Intelligence Report , Spring, p. 141. The Southern Poverty Law Center, Montgomery, AL. Brown Hana. 2013. “Race, Legality, and the Social Policy Consequences of Anti-Immigration Mobilization.” American Sociological Review  78( 2): 290- 314. Google Scholar CrossRef Search ADS   Cabrera Luis, Glavac Sonya. 2010. “Minutemen and Desert Samaritans.” Journal of Ethnic and Migration Studies  36( 4): 673- 95. Google Scholar CrossRef Search ADS   Centro de Estudios Migratorios Latinoamericanso (CEMLA). 2013. “Eventos de Repatriación de Mexicanos Desde Estados Unidos, Según Entidad Federativa y Punto de Recepción, 1995-2012.” Unidad de Política Migratoria, Secretaría de Gobernación, México D.F. Cerrutti Marcela, Massey Douglas S. 2001. “On the Auspices of Female Migration from Mexico to the United States.” Demography  38( 2): 187- 200. Google Scholar CrossRef Search ADS PubMed  Chavez Leo R. 2008. The Latino Threat.  Stanford, CA: Stanford University Press. Comisión Nacional Para el Desarrollo de los Pueblos Indígenas (CDI). 2006. Indicadores Sociodemográficos de la Población Indígena 2000-2005 . Retrieved July 20, 2017 (www.cdi.gob.mx/cedulas/sintesis_resultados_2005.pdf). Cornelius Wayne A. 1998. “The Structural Embeddedness of Demand for Mexican Immigrant Labor.” Pp. 115- 44 in Crossings , edited by Suarez-Orozco M. Cambridge, MA: Harvard University Press. Cornelius Wayne A., Fitzgerald David, Fischer Pedro Lewin, Muse-Orlinoff Leah. 2010. Migration and the U.S. Economic Crisis.  San Diego, CA: Center for Comparative Immigration Studies. Cornelius Wayne A., Lewis Jessa M. 2007. Impacts of Border Enforcement on Mexican Migration . San Diego, CA: Center for Comparative Immigration Studies. De Genova Nicholas P. 2002. “Migrant ‘Illegality’ and Deportability in Everyday Life.” Annual Review of Anthropology  31: 419- 47. Google Scholar CrossRef Search ADS   Doty Roxanne L. 2007. “States of Exception on the Mexico-U.S. Border.” International Political Sociology  1: 113- 37. Google Scholar CrossRef Search ADS   Doty Roxanne L. 2009. The Law Into Their Own Hands , Tucson: The University of Arizona Press. Dove April L. 2010. “Framing Illegal Immigration at the U.S.-Mexican Border.” Pp. 199- 237 in Research in Social Movements, Conflicts, and Change , edited by Coy P. Bingley, UK: Emerald Group Publishing Limited. Google Scholar CrossRef Search ADS   Dunn Timothy. 1996. The Militarization of the U.S.-Mexico Border . Austin: University of Texas Press. Earl Jennifer. 2004. “The Cultural Consequences of Social Movements.” Pp. 508- 30 in The Blackwell Companion to Social Movements , edited by Snow D., Soule S., Kriesi H. Malden, MA: Blackwell Publishing. Elcioglu Emine. 2015. “Popular Sovereignty on the Border.” Ethnography  16( 4): 438- 62. Google Scholar CrossRef Search ADS   Giugni Mario. 2007. “Useless Protest?” Mobilization  12: 53- 77. Gilchrist Jim, Corsi Jerome. 2006. Minutemen: The Battle to Secure America's Borders.  Los Angeles, CA: World Ahead Publishing. Graham John W., Olchowski Allison E., Gilreath Tamika D. 2007. “How Many Imputations are Really Needed?” Prevention Science  8: 206- 13. Google Scholar CrossRef Search ADS PubMed  Hagan Jacqueline M. 2008. Migration Miracle . Cambridge, MA: Harvard University Press. Heckman James J. 1976. “The Common Structure of Statistical Models of Truncation, Sample Selection and Limited Dependent Variables and a Simple Estimator for Such Models.” Annals of Economic and Social Measurement  5: 475- 92. Kandel William, Massey Douglas S. 2002. “The Culture of Mexican Migration.” Social Forces  80( 3): 981- 1004. Google Scholar CrossRef Search ADS   Kil Sang Hea, Menjivar Cecilia, Doty Roxanne Lynne. 2009. “The Border Spectacle” Pp. 297- 312 in Immigration, Crime, and Justice , edited by McDonald W. Bingley, UK: Emerald Group Publishing Ltd. Google Scholar CrossRef Search ADS   Marcelli Enrico A., Cornelius Wayne A. 2001. “The Changing Profile of Mexican Migrants to the United States” Latin American Research Review  36( 3): 105- 31. Martell Alpha, Pineda Maribel, Tapia Luis 2007. “The Contemporary Migration Process.” Pp. 49- 72 in Mayan Journeys , edited by Cornelius W., Fitzgerald D., Fischer P. San Diego, CA: Center for Comparative Immigration Studies. Martínez Daniel E. 2016. “Coyote Use in an Era of Heightened Border Enforcement: New Evidence from the Arizona-Sonora Border Region.” Journal of Ethnic and Migration Studies  42( 1): 103- 19. Google Scholar CrossRef Search ADS   Martínez Daniel E., Slack Jeremy, Beyerlein Kraig, Vandervoet Prescott, Klingman Kristin, Molina Paola, Manning Shiras, Burham Melissa, Walzak Kylie, Valencia Kristen, Gamboa Lorenzo. 2017. “The Migrant Border Crossing Study: A Methodological Overview.” Population Studies  71( 2): 249- 64. Google Scholar CrossRef Search ADS PubMed  Martínez Daniel E., Slack Jeremy, Martínez-Schuldt Ricardo. n.d. “Return Migration in the Age of the Unauthorized Permanent Resident: A Quantitative Assessment of Migration Intentions Post-Deportation.” Unpublished Manuscript. Martínez Daniel E., Vandervoet Prescott, Slack Jeremy. 2013. “Methodological Challenges and Ethical Concerns of Researching Marginalized and Vulnerable Populations: Evidence from Firsthand Experiences of Working with Undocumented Migrants.” Pp. 101- 20 in Uncharted Terrains: New Directions in Border Research Methodology, Ethics, and Practice , edited O’Leary A., Deeds C., Whiteford S. Tucson: University of Arizona Press. Massey Douglas, Durand Jorge, Malone Nolan. 2002. Beyond Smoke and Mirrors . New York: Russell Sage Foundation. Massey Douglas S., Capoferro Chiara. 2008. “The Geographic Diversification of American Immigration.” Pp. 25- 50 in New Faces in New Places , edited by Massey Douglas S. New York: The Russell Sage Foundation. Massey Douglas S., Espinoza Kristin E. 1997. “What's Driving Mexico-U.S. Migration?" The American Journal of Sociology  102( 4): 939- 99. Google Scholar CrossRef Search ADS   McAdam Doug. 1989. “The Biographical Consequences of Activism.” American Sociological Review  54( 5): 744- 60. Google Scholar CrossRef Search ADS   Minuteman Civil Defense Corps Project (MCDC). 2009. “Please Mark Your Calendars with Our 2009 Muster Schedule.” Retrieved March 5, 2014 (www.minutemanhq.com/state/read.php?chapter=AZ&sid=1023). Minuteman Project, The. 2014. “Operation Normandy.” Retrieved June 7, 2016 (http://baesic.net/minutemanproject/). Menard Scott. 1995. Applied Logistic Regression Analysis . Thousand Oaks, CA: Sage. Neiwert David. 2013. “Smaller, More Radical Border Militias Patrolling Arizona Desert.” The Southern Poverty Law Center, Montgomery, AL. Nevins Joseph. 2010. Operation Gatekeeper and Beyond . New York: Routledge. O’Leary Anna Ochoa. 2009. “Mujeres en el Cruce.” Journal of the Southwest  51( 4): 523- 42. Google Scholar CrossRef Search ADS   Passel Jeffrey S., Cohn D’Vera. 2016. “Overall Number of Unauthorized Immigrants Holds Steady Since 2009: Decline in share from Mexico mostly offset by growth from Asia, Central America and sub-Saharan African.” Pew Research Center, Washington, DC. Piven Francis F. 2006 Challenging Authority.  Lanham, MD: Rowman & Littlefield. Portes Alejandro. 1998. “Social Capital” Annual Review of Sociology  24: 1- 24. Google Scholar CrossRef Search ADS   Rosenblum Marc R. 2013. Border Security: Immigration Enforcement Between Ports of Entry . Washington, DC: Congressional Research Service. Royston Patrick. 2009. “Multiple Imputation of Missing Values.” The Stata Journal  9: 466- 77. Rubin Donald B. 1987. Multiple Imputation for Nonresponse in Surveys . New York: Wiley. Google Scholar CrossRef Search ADS   Shapira Harel. 2013. Waiting for Jose . Princeton, NJ: Princeton University Press. Google Scholar CrossRef Search ADS   Sherkat Darren, Blocker T. Jean. 1997. “Explaining the Political and Personal Consequences of Protest.” Social Forces  75( 3): 1049- 70. Google Scholar CrossRef Search ADS   Singer Audrey, Massey Douglas S. 1998. “The Social Process of Undocumented Border Crossing Among Mexican Migrants.” International Migration Review  32( 3): 561- 92. Google Scholar CrossRef Search ADS PubMed  Skocpol Theda. 2003. Diminished Democracy . Norman: University of Oklahoma Press. Skocpol Theda, Williamson Vanessa. 2013. The Tea Party and the Remaking of Republican Conservatism.  Oxford, UK: Oxford University Press. Slack Jeremy, Martínez Daniel, Vandervoet Prescott. 2011. “Methods of Violence: Researcher Safety and Adaptability in Times of Conflict.” Practicing Anthropology  33( 1): 33- 37. Google Scholar CrossRef Search ADS   Slack Jeremy, Martínez Daniel E., Whiteford Scott, Emily Peiffer. 2013. “In the Shadow of the Wall: Family Separation, Immigration Enforcement and Security.” The Center for Latin American Studies, University of Arizona, Tucson, AZ. Retrieved July 20, 2017 (http://las.arizona.edu/sites/las.arizona.edu/files/UA_Immigration_Report2013web.pdf). Slack Jeremy, Martínez Daniel E., Whiteford Scott, Peiffer Emily. 2015. “In Harm’s Way: Family Separation, Immigration Policy, and Security on the U.S.-Mexico Border.” Journal on Migration and Human Security  3( 2): 109- 28. Google Scholar CrossRef Search ADS   Snow David, Soule Sarah, Kriesi Hanspeter. 2004. “Mapping the Terrain.” Pp. 3- 16 in Blackwell Companion to Social Movements , edited by Snow D., Soule S., Kriesi H. Malden, MA: Blackwell Publishing. Google Scholar CrossRef Search ADS   Stewart Julie, Bendall Michele Enciso, Morgan Charlie V. 2015. “Jobs, Flags, and Laws: How Interests, Culture, and Values Explain Recruitment into the Utah Minuteman Project.” Sociological Perspectives  58( 4): 627- 48. Google Scholar CrossRef Search ADS   Southern Poverty Law Center (SPLC). 2010. “Intelligence Report.” Spring. Retrieved July 20, 2017 (www.splcenter.org/intelligence-report?f[0]=field_intel_report_issue%3A11689). Spener David. 2009. Clandestine Crossings . Ithaca, NY: Cornell University Press. Trujeque Díaz José Antonio. 2007. “Minuteman Project: Segregación y Activismo Antimigratorio.” Andamios  2( 6): 137- 72. U.S. Border Patrol (USBP). 2014a. “United States Border Patrol Apprehensions by Gender - Fiscal Year 2013 (Oct. 1st through Sept. 30th).” Retrieved August 18, 2017 (www.cbp.gov/sites/default/files/documents/U.S.%20Border%20Patrol%20Fiscal%20Year%202013%20Profile.pdf). U.S. Border Patrol (USBP). 2014b. “United States Border Patrol Southwest Border Sector Total Illegal Alien Apprehensions by Fiscal Year (Oct. 1st through Sept. 30th).” Retrieved August 18, 2017 (www.cbp.gov/sites/default/files/assets/documents/2016-Oct/BP%20Southwest%20Border%20Sector%20Apps%20FY1960%20-%20FY2016.pdf). Ward Matthew. 2014. “They Say Bad Things Come in Threes: How Economic, Political and Cultural Shifts Facilitated Contemporary Anti-Immigration Activism in the United States.” Journal of Historical Sociology  27( 2): 263- 92. Google Scholar CrossRef Search ADS   Ward Matthew. 2016. “Opportunity, Resources, and Threat: Explaining Local Nativist Organizing in the United States.” Sociological Perspectives : 1- 20. Retrieved July 20, 2017 (DOI: 10.1177/0731121416655994). Ward Matthew, Martínez Daniel E. 2015. “Know Your Enemy: How Unauthorized Migrants Learn About and Perceive Anti-Immigrant Mobilization in the United States.” Migration Letters  12( 2): 50- 66. Williams Richard. 2012. “Using the Margins Command to Estimate and Interpret Adjusted Predictions and Marginal Effects.” The Stata Journal  12( 2): 308- 31. Winship Christopher, Radbill Larry. 1994. “Sampling Weights and Regression Analysis.” Sociological Methods & Research  23( 2): 230- 57. Google Scholar CrossRef Search ADS   © The Author 2017. Published by Oxford University Press on behalf of the Society for the Study of Social Problems. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Journal

Social ProblemsOxford University Press

Published: Sep 11, 2017

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off