Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Social Withdrawal and Romantic Relationships: A Longitudinal Study in Early Adulthood

Social Withdrawal and Romantic Relationships: A Longitudinal Study in Early Adulthood Involvement in romantic relationships is a salient developmental task in late adolescence and early adulthood, and deviations from normative romantic development are linked to adverse outcomes. This study investigated to what extent social withdrawal contributed to deviations from normative romantic development, and vice versa, and the interplay between withdrawal and couples’ relationship perceptions. The sample included 1710 young adults (55–61% female) from the Tracking Adolescents’ Individual Lives Survey cohort and their romantic partners. Data were collected across 4 waves, covering romantic relationships from ages 17 to 29 years. The results showed that higher withdrawal predicted a higher likelihood of romantic non-involvement by adulthood, consistently being single at subsequent waves, and entering one’s first relationship when older. Withdrawal moderately decreased when youth entered their first relationship. Male’s withdrawal in particular affected romantic relationship qualities and dynamics. These results provide new insights into the developmental sequelae of withdrawn young adults’ romantic relationship development. ● ● ● ● ● ● Keywords Dating Early adulthood Late adolescence Longitudinal Relationship quality Romantic relationships Social withdrawal Introduction common, but also shift in their qualities and functions across these ages: they become longer-lasting and a main Developmental Task Theory (Roisman et al., 2004) posits source of support and intimacy (Collibee & Furman, 2015; that involvement in romantic relationships becomes a sali- Lantagne & Furman, 2017). A comprehensive investigation ent developmental task during late adolescence and early of the specific features of romantic relationship develop- adulthood, and delaying or not meeting this task is related to ment that withdrawal affects and is affected by has been adverse outcomes. Indeed, romantic relationships become lacking in the extant literature. This study focused on the increasingly prevalent throughout these ages, with over half longitudinal associations between social withdrawal and of adolescents having been romantically involved by the romantic relationships in late adolescence and early adult- age of 16, and the large majority by early adulthood (Carver hood. Specifically, it investigated to what extent withdrawal et al., 2003). Romantic relationships not only become more contributed to deviations from normative romantic devel- opment and the interplay between withdrawal and couples’ perceptions of the quality of relationships. Three key fea- Supplementary information The online version contains tures were examined as indicators of romantic development: supplementary material available at https://doi.org/10.1007/s10964- involvement (lifetime involvement, current involvement, 021-01469-1. and timing), relationship quantity (number of partners and relationship duration), and relationship quality (commit- * Stefania A. Barzeva s.a.barzeva@umcg.nl ment, satisfaction, support, and conflict; Collins, 2003). Social withdrawal is an umbrella term referring to the Department of Psychiatry, Interdisciplinary Center voluntary self-isolation from familiar and unfamiliar others Psychopathology and Emotion Regulation, University of through the consistent display of solitary behaviors, such as Groningen, University Medical Center Groningen, Groningen, The Netherlands avoiding social interaction and spending excessive time alone (Rubin et al., 2009). The motivation to withdrawal Research Center Adolescent Development, Utrecht University, Utrecht, The Netherlands varies between individuals and differentiates three types of 1234567890();,: 1234567890();,: Journal of Youth and Adolescence (2021) 50:1766–1781 1767 withdrawal: shyness, unsociability, and avoidance (Coplan (Roswell & Coplan, 2013; with an exception in Schmidt & Armer, 2007; Ozdemir et al., 2015). Phenotypic with- et al. 2017). The relations between withdrawal and romantic drawal behaviors overlap across these withdrawal types. relationship quantity have never been tested directly, but The current study uses the term “social withdrawal” to refer one study found that shy young adults date less frequently to the global, multidimensional, behavioral phenotype of (Leck, 2006), suggesting a possible effect of withdrawal on voluntary self-isolation. Previous studies using this global the quantitative features of romantic relationships, such as conceptualization have indicated that 12 to 23% of indivi- the number of partners and the duration of relationships. If duals are persistently withdrawn throughout adolescence, withdrawn youth likewise date less frequently, it would and that withdrawn adolescents report less social affiliation, suggest that they change partners less often, leading to social contact, and social competence, and more anxiety having fewer partners with possibly longer relationship than their non-withdrawn peers (Barzeva et al., 2019a; Tang durations. et al. 2017). The romantic development of withdrawn Entering a romantic relationship for the first time might adolescents and young adults may differ from the normative have an effect on youth’s withdrawal. Developmental Task patterns due to deviations in previous social experiences Theory suggests that when life events are off-time— that are usually conducive to romantic involvement, such as occurring earlier or later than the majority of peers—ado- friendships (Kingery et al., 2010). The sequential stage lescents experience negative social sanctions for deviating theory of heterosexual romantic relationship emergence from the normative pattern of development and receive states that youth develop social-emotional competencies in fewer social resources from peers (Furman & Collibee, the context of same-sex friendships in childhood, and start 2014). When adolescents’ life events become normative in to apply these competencies in mixed-sex peer groups the context of their peers’ experiences, social sanctions may during adolescence (Connolly et al., 2004; Connolly et al., be lifted and more social resources provided, leading to 2000). Mixed-sex groups provide opportunities to learn greater socio-emotional adjustment. Additionally, the first how to approach and interact with opposite-sex peers, who romantic relationship is unique and of particular importance are also potential dating partners. Withdrawn youth, how- to young adults’ socio-emotional adjustment because it ever, do not follow this cascade of development (Nelson represents a shift in social identity, namely as a “girlfriend” et al., 2008). They experience difficulties initiating and or “boyfriend” (Raley et al., 2007). This new identity, and maintaining same-sex friendships, subsequently leading to a the social roles and experiences that come with it, changes smaller mixed-sex peer group. Not participating in a mixed- how young adults perceive themselves and are perceived by sex peer group hinders the development of important social others; when entering a romantic relationship for the first skills for romantic relationship initiation and maintenance, time, young adults feel an increase in autonomy, status, and limits the size of adolescents’ dating pool. belonging, and social support, and may be seen by others as Withdrawn youth’s formative experiences may set the more mature and as a potential dating partner (Raley et al., stage for non-normative or delayed romantic relationship 2007). These provisions may lead to a decrease in social involvement and quantity in late adolescence and beyond withdrawal. Although an intriguing proposal, the effect of (Raley et al., 2007; Shulman & Connolly, 2013). That is, reaching the developmental task of romantic involvement having had fewer opportunities to learn how to approach on youth’s withdrawal has not been tested yet. and interact with opposite-sex peers in adolescence (Barry Despite the possible decrease of withdrawal when et al., 2013; Nelson et al., 2008) is likely to contribute to entering a romantic relationship for the first time, withdrawn anxiety, rejection sensitivity (Gazelle & Druhen, 2009), and adolescents and young adults may differ from their non- avoidance of novel social situations, such as when asking withdrawn peers in the qualities of their romantic relation- someone out on a date, and hence to delay romantic ships. By late adolescence and early adulthood, individuals involvement. While this is a provoking notion, the empirical are better equipped with relationship maintenance skills and work on withdrawn individuals’ romantic involvement has start seeking and entering longer-lasting, more committed, been scarce, has mainly concerned the shyness aspect of supportive, and exclusive relationships than early adoles- withdrawal, and shows somewhat mixed findings. A semi- cents (Shulman & Connolly, 2013). Starting from about 17 nal study of shy children born in the 1920s found that men, years of age, young adults transition from more sporadic but not women, who were shy as children were older than romantic relationship involvements that resemble friend- their non-shy counterparts when they first married (Caspi ships to more intimate and committed relationships with a et al., 1988). More recent studies have found that withdrawn greater dyadic orientation (Seiffge-Krenke, 2003; Shulman and inhibited individuals become romantically involved at & Connolly, 2013). In turn, being involved in a committed an older age than their more sociable peers (Boisvert & romantic relationship is linked to a decline in emotional Poulin, 2016; Meeus et al., 2011), and those who were problems in early adulthood (Meeus et al., 2007). The currently involved were less shy than those who were not transition to more committed relationships may occur later 1768 Journal of Youth and Adolescence (2021) 50:1766–1781 for withdrawn young adults, who have less romantic Sex- and ethnicity-based social norms may contribute to experience due to delays in romantic involvement. Thus, differences between males and females and Western and higher withdrawal might predict lower commitment in non-Western youth in romantic relationship development young adults’ romantic relationships, and lower commit- and the effect of withdrawal. Although males and females ment, in turn, predict higher withdrawal. Furthermore, it is may both experience non-normative romantic development, well documented that the formative qualities of adolescent withdrawn males may be particularly affected because friendships are linked to concurrent and future romantic inhibited behavior violates gender-normative expectations relationship qualities (Collins & Van Dulmen, 2006; Collins of male dominance and assertiveness (Doey et al., 2014). et al., 2009). Withdrawn youth tend to report friendships Additionally, females may enter romantic relationships with characterized by low support and high conflict (Rubin et al., more relationship maintenance skills than males. In child- 2006), which likely translates into less satisfying romantic hood and adolescence, girls tend to prefer involvement in relationships. Empirical studies on the effects of withdrawal more intimate, dyadic relationships (Hall, 2011; Rose & on romantic relationship qualities are rare, however, and Rudolph, 2006), which provide more opportunities than have predominantly focused on the effects of shyness on boys’ friendships to become comfortable self-disclosing, relationship satisfaction. These studies have consistently develop intimacy with others, and learn conflict resolution found that shy adolescents and young adults report lower skills (Giordano et al., 2006). Hence, girls’ friendships satisfaction than more sociable individuals (Luster et al., resemble romantic relationships more closely than boys’ 2013; Nelson et al., 2008; Roswell & Coplan, 2013; Tackett friendships. Because of this, boys – especially withdrawn et al., 2013; except in Schmidt et al., 2017). This effect is ones who lack social experience in general – may be less speculated to occur because of withdrawn individuals’ prepared for maintaining intimate relationships than girls. social inhibition and rejection sensitivity, which leads to Little is known about ethnic differences in social with- decreased responsiveness, self-disclosure, and intimacy drawal and romantic relationship development, but some with their romantic partners (Luster et al., 2013). These differences may be expected due to cultural norms around characteristics might likewise be related to support and inhibited behavior (Chen & Tse, 2008; Coplan et al., 2012) conflict. and dating, especially in involvement and timing (Connolly Another possible reason for the low romantic relation- & McDonald, 2020). Thus, sex and ethnicity were included ship quality ratings of withdrawn youth are negative cog- as covariates in all models to adjust for possible nitive biases. When assessing their relationships, confounding. withdrawn individuals might report low quality because they are more attentive toward the negative aspects of their relationships and recall more negative interactions with The Current Study their romantic partners (Gazelle & Duhen, 2009). In that case, their romantic partners might have more positive There is reason to believe that social withdrawal is linked to perceptions of the relationship. While untested in with- deviations from normative romantic relationship develop- drawn young adults’ romantic relationships, these effects ment, and vice versa, but a direct empirical test of the extent have been found in the romantic relationships of indivi- to which withdrawn youth deviate from—and in which duals with high attachment anxiety, which is characteristic features of—normative romantic development has been of withdrawn individuals’ attachment style (Roswell & lacking. The aim of this study was to investigate long- Coplan, 2013; Rubin et al., 2009). Anxiously attached itudinal associations between social withdrawal and individuals perceive more conflict and less support in their romantic relationship development in late adolescence and romantic relationships than securely attached ones, and early adulthood, a crucial life phase for romantic these biased perceptions are associated with greater emo- relationship-related exploration, decisions, and family for- tional distress and decreased satisfaction with and com- mation. This was done, first, by testing to what extent mitment to their romantic partners (Campbell et al., 2005). withdrawal predicts deviations in romantic involvement Thus, romantic partners’ ratings may provide additionally (lifetime involvement, current involvement, timing) and informative, less biased relationship quality information, relationship quantity (relationship length, the number of and social withdrawal may have a differential effect on partners). Higher withdrawal was hypothesized to predict a self- and partner-reported relationship perspectives, but greater likelihood of never having been involved in a empirical evidence has been lacking so far. In sum, how romantic relationship by early adulthood, being single withdrawal directly affects self- and partner-reported across late adolescence and early adulthood, initiating one’s romantic relationship qualities—commitment, satisfaction, first romantic relationship at an older age, and having support, and conflict—and if these relationship qualities longer-lasting relationships, but with fewer partners affect withdrawal, has remained largely unknown. (Hypothesis 1). Second, this study tested if entering a Journal of Youth and Adolescence (2021) 50:1766–1781 1769 romantic relationship for the first time changes one’s Measures withdrawal levels. Withdrawal was expected to decrease after entering a romantic relationship for the first time, Social withdrawal especially in individuals with high pre-involvement with- drawal (Hypothesis 2). A third set of tests addressed the Social withdrawal was assessed at T4 to T7 with the mean question if withdrawal predicts deviations from romantic of five items from the Adult Self-Report (ASR; Achenbach relationship qualities, and vice-versa, and examined the & Rescorla, 2003) withdrawn scale. In a sample of adults interplay between withdrawal and couple’s perceptions of over the age of 18 years, the ASR withdrawn scale had high their relationship. It was hypothesized that higher with- test-retest reliability and correlated moderately positively drawal will be concurrently and longitudinally associated with measures of anxiety and social introversion (Achen- with lower self-reported and partner-reported commitment, bach & Rescorla, 2003). The five items included were: I satisfaction, and support, and higher conflict, which in turn would rather be alone than with others; I am secretive or will be associated with higher withdrawal (Hypothesis 3). keep things to myself; I refuse to talk; I am too shy or timid; and I keep from getting involved with others. These items were selected based on face validity and previous research Method (Booth-LaForce & Oxford, 2008; Tang et al., 2017), and items have been found to be longitudinally measurement Participants invariant in adolescence and early adulthood (Barzeva et al., 2019a, 2019b). Items were rated on a 3-point scale, with 0 The study included 1710 participants from the prospective, = Not at all, 1 = A little or sometimes, and 2 = Always or population-based cohort Tracking Adolescents’ Individual often true, in the past 6 months. Cronbach’s alpha ranged Lives Survey (TRAILS; www.trails.nl) and their romantic from 0.67 to 0.72 at each wave. For scales with fewer than partners, if applicable (N = 463 to 559 partners per assess- ten items, an internal reliability cutoff of α > 0.60 is con- ment wave). The TRAILS sample was recruited from rural sidered acceptable (Loewenthal, 2004). and urban areas of the North of the Netherlands, and includes individuals born between 1989 and 1991. Data Lifetime and current relationship involvement collection began in 2001 when the participants were approximately 11 years old. Assessments occurred 7 times, Lifetime romantic relationship involvement was assessed at every 2 to 3 years, from ages 11 to 29 years, and 64–96% of T4 to T7 with the item, Have you ever had a boyfriend or the initial sample participated in subsequent assessment girlfriend? (0 = no, 1 = yes). Current romantic relationship waves. More details about the TRAILS recruitment and involvement was assessed at T4 to T7 with the item, Do you assessment procedure have been reported elsewhere (De have a girlfriend or boyfriend at the moment? (0 = no, 1 = Winter et al., 2005; Huisman et al., 2008; Oldehinkel et al., yes). 2015). The current study used data from the last four assessment waves of TRAILS (T4-T7), when participants Timing of first relationship were approximately 19, 22, 26, and 29 years old (55–61% female across waves). About 90% of participants were from The approximate timing of participants’ first romantic an ethnically Dutch background. Table 1 shows the parti- relationship in late adolescence was assessed at T4 by cipant and partner demographics, and descriptive statistics subtracting the duration of their current relationship from of the romantic relationship characteristics at each wave, their age at T4. If a participant did not have a partner at T4, and Table 2 the correlations between variables. an item from an Events History Calendar asking if they had started a romantic relationship with someone in the last two Data Collection Procedure years was used instead. Participants could report the month and year in which they started up to two romantic rela- The Dutch Central Committee on Research Involving tionships, and the date of the earliest reported relationship Human Subjects approved the TRAILS study. Participants was used to calculate their age at that time. For those who provided written consent at T4 to T7. Data was collected by entered a relationship for the first time at T5 to T7, their age means of online questionnaires at these waves. At T5 to T7, at entry was determined by subtracting the length of that participants nominated their current romantic partner. first relationship from their age at that wave. Early romantic Research assistants contacted romantic partners for partici- relationships that ended before the age of 17 years were not pation. When the partners consented to participate, they included, whereas those that continued through the age of were sent online questionnaires via email. 17 were. The timing variable therefore reflects the age at 1770 Journal of Youth and Adolescence (2021) 50:1766–1781 Table 1 Participant, partner, and T4 T5 T6 T7 romantic relationship characteristics Participant characteristics N participants 1679 1634 1523 1223 Age, years M (SD) 19.1 (0.59) 22.3 (0.65) 25.6 (0.60) 28.9 (0.59) Female % 55.3 55.1 56.3 61.2 Dutch % 89.6 89.7 90.6 91.0 Lifetime involvement % 81.8 91.9 96.1 97.0 Current involvement % 53.1 64.4 72.3 71.2 Social withdrawal M (SD) 0.31 (0.34) 0.32 (0.36) 0.38 (0.37) 0.34 (0.36) Partner characteristics N partners – 559 463 479 Age, years M (SD) – 24.0 (3.62) 27.2 (3.46) 30.4 (3.83) Age range – 16.8–47.3 18.8–43.2 19.5–53.3 Female % – 38.3 38.0 42.0 Relationship % Married – 3.4 11.4 22.8 Registered partnership – 1.3 13.8 18.0 Cohabitating – 35.1 47.7 41.5 Not cohabitating – 58.5 27.0 17.7 Relationship characteristics M (SD) Number of partners 2.28 (1.70) 2.33 (1.47) 2.21 (1.28) – Relationship duration, months 19.0 (14.8) 31.9 (24.5) 49.0 (33.6) 70.3 (43.7) Commitment (participant) 5.82 (1.42) 6.27 (1.16) 6.52 (0.89) – Commitment (partner) – 6.36 (0.95) 6.57 (0.82) – Satisfaction (participant) 6.19 (1.08) 6.30 (0.93) 6.32 (0.84) – Satisfaction (partner) – 6.28 (0.89) 6.27 (0.86) – Support (participant) – 3.50 (0.41) 3.50 (0.48) – Support (partner) – 3.36 (0.51) 3.43 (0.46) – Conflict (participant) – 1.35 (0.34) 1.46 (0.39) – Conflict (partner) – 1.51 (0.38) 1.50 (0.37) – Other ethnicities include: Turkish, Moroccan, Surinam, Antillean, Indonesian or Mollucan, and Other not specified The average maximum relationship duration across participants was 54.9 months (SD = 38.8; range = 0.25–192). Scores could range from 1–10 for number of relationships, 1–7 for commitment and satisfaction, 1–4 for support, and 1–3 for conflict. Dashes indicate that the data was not collected at the assessment wave. which participants entered a more serious, attachment-based relationship with your current partner? (in months) at T4 to romantic relationship for the first time. T7. The maximum reported duration across the four waves was used in the analyzes to account for relationships that Number of partners continued throughout multiple assessment waves. The number of romantic partners that participants had had Commitment was assessed at T6 with the item, How often in your life have you had a steady boyfriend or girlfriend? (rated from Romantic relationship commitment was assessed with mean 0 to 10 or more times). of the three-item Commitment subscale of the Investment Model Scale (IMS; Rusbult et al., 1998). Participants Relationship duration completed the IMS at T4 to T6, and their partners at T5 and T6. The three items were: I am focused on the long-term The duration of the current romantic relationship was future of my relationship; I want my relationship to last a assessed with the item, How long have you had a very long time; and I want my relationship to continue Journal of Youth and Adolescence (2021) 50:1766–1781 1771 Table 2 Correlations between variables Variable 1234 567 8910 11 12 13 14 15 16 17 18 19 20 21 22 23 24 1 T4 Withdrawal ** 2 T5 Withdrawal 0.56 ** ** 3 T6 Withdrawal 0.49 0.62 ** ** ** 4 T7 Withdrawal 0.44 0.54 0.60 ** * 5 Timing (age) 0.08 0.05 0.07 0.03 ** ** 6 Number of partners −0.03 0.04 0.03 0.09 −0.07 ** ** ** ** ** ** 7 Max length relation. −0.08 −0.09 −0.13 −0.13 −0.36 −0.15 ** ** 8 T4 S Commitment −0.04 0.02 −0.001 −0.04 −0.16 −0.06 0.19 * * ** ** ** 9 T5 S Commitment −0.06 −0.09 −0.08 −0.4 −0.11 −0.05 0.25 0.24 * * ** ** ** ** ** ** 10 T6 S Commitment −0.07 −0.08 −0.14 −0.12 −0.12 −0.04 0.16 0.22 0.30 ** ** ** ** 11 T5 P Commitment −0.04 −0.08 −0.05 −0.02 −0.05 −0.02 0.22 0.21 0.26 0.13 * ** 12 T6 P Commitment −0.04 −0.05 −0.02 −0.07 −0.02 −0.04 0.08 0.06 0.09 0.12 0.41 ** * * * * ** ** * * 13 T4 S Satisfaction −0.12 −0.04 −0.10 −0.10 −0.09 −0.09 0.14 0.60 0.10 0.12 0.07 0.03 ** ** ** * ** ** ** ** ** ** 14 T5 S Satisfaction −0.10 −0.16 −0.13 −0.10 −0.05 −0.03 0.16 0.14 0.71 0.24 0.19 0.01 0.18 ** ** ** ** * ** ** ** ** ** ** ** 15 T6 S Satisfaction −0.09 −0.15 −0.20 −0.20 −0.07 −0.01 0.07 0.14 0.21 0.68 0.14 0.15 0.15 0.30 ** ** ** * ** * ** ** ** ** 16 T5 P Satisfaction −0.08 −0.12 −0.15 −0.08 0.03 −0.01 0.12 0.13 0.17 0.13 0.64 0.33 0.06 0.25 0.24 * ** ** ** ** ** 17 T6 P Satisfaction 0.003 −0.02 −0.09 −0.12 0.05 0.03 −0.03 −0.02 0.03 0.18 0.36 0.62 0.01 0.01 0.31 0.45 ** ** ** ** ** * ** * ** ** ** ** ** 18 T5 S Support −0.12 −0.14 −0.14 −0.14 −0.01 −0.04 0.17 0.11 0.34 0.10 0.12 0.10 0.18 0.37 0.14 0.12 0.08 * ** ** ** * ** ** ** ** ** * ** ** 19 T6 S Support −0.08 −0.12 −0.16 −0.19 −0.05 −0.02 0.05 0.10 0.18 0.49 0.04 0.06 0.14 0.23 0.58 0.12 0.16 0.30 ** ** ** ** ** ** ** * ** ** ** 20 T5 P Support −0.08 −0.12 −0.07 −0.06 0.04 −0.01 0.06 0.16 0.17 0.05 0.37 0.18 0.18 0.25 0.10 0.46 0.22 0.20 0.08 * * * ** ** ** ** ** ** ** * ** ** 21 T6 P Support −0.12 −0.11 −0.11 −0.14 0.05 −0.03 −0.02 0.12 0.10 0.14 0.24 0.31 0.02 0.04 0.21 0.32 0.38 0.12 0.16 0.42 * * ** ** ** ** ** ** ** ** ** ** ** ** 22 T5 S Conflict 0.09 −0.13 0.15 0.22 0.09 0.06 −0.24 −0.10 −0.25 −0.30 −0.18 −0.21 −0.07 −0.36 −0.25 −0.24 −0.24 −0.31 −0.27 −0.06 −0.02 ** ** ** * ** * * ** ** * ** ** ** ** ** 23 T6 S Conflict 0.02 0.04 0.19 0.18 −0.02 0.10 −0.04 −0.08 −0.10 −0.25 −0.07 −0.10 −0.12 −0.13 −0.36 −0.10 −0.28 −0.11 −0.30 −0.09 −0.15 0.49 * * * ** ** * ** ** ** ** * ** ** 24 T5 P Conflict 0.01 0.11 0.04 0.05 0.04 0.13 −0.14 −0.07 −0.02 −0.15 −0.30 −0.41 −0.11 −0.15 −0.20 −0.44 −0.31 −0.07 −0.20 −0.15 −0.12 0.55 0.21 * ** * ** ** ** ** ** ** ** ** 25 T6 P Conflict −0.004 −0.03 0.06 0.12 −0.10* 0.04 0.03 −0.07 0.03 −0.07 −0.10 −0.24 −0.14 −0.01 −0.25 −0.23 −0.39 −0.03 −0.18 −0.10 −0.22 0.34 0.57 0.42 Relation. = romantic relationship; S = self-reported; P = partner-reported * ** *** p < 0.05, p < 0.01, p < 0.001 1772 Journal of Youth and Adolescence (2021) 50:1766–1781 forever. Items were rated on a 7-point scale, from 1 = standard errors (MLR) estimation, unless otherwise stated. completely disagree to 7 = completely agree (T4-T6 α The analyzes were pre-registered on the Open Science Fra- self- = 0.84, 0.91, 0.90; T5-T6 α = 0.87, 0.87). mework (www.osf.io/j6nrd), and an explanation of all report partner-report deviations can be found in the Supplementary Materials. First, Satisfaction to test if social withdrawal predicted lifetime non-involve- ment, the age at one’s first relationship, maximum relation- Romantic relationship satisfaction was measured by the ship duration, and the number of partners, an unconditional mean of two items of the IMS Satisfaction subscale. Parti- withdrawal latent growth curve model (LGCM), with corre- cipants completed the questionnaire at T4 to T6, and their lated intercept and linear slope from T4 to T6 was specified. partners at T5 and T6. The two items were: I am satisfied Sex and ethnicity were included as covariates, with paths to with my relationship and My relationship gives me what I the withdrawal intercept and slope and to the outcomes. Paths need in terms of intimacy, friendship, etc. Although the from the withdrawal intercept and slope to the outcomes were subscale also included the item My relationship is much specified. The path from the intercept to the outcomes shows better than the relationships of others, alpha was higher how baseline withdrawal levels predicted the outcomes, and when excluding this item due to a lower overall mean rating the path from the slope to the outcomes shows how the of this item compared to the other two, and weak correla- change in withdrawal between 19 and 29 years was associated tions between this and the other two items. Items were rated with the outcomes. Additionally, a cross-lagged panel model on a 7-point scale from 1 = completely disagree to 7 = (CLPM) with the concurrent and longitudinal effects between completely agree (T4-T6 α = 0.73, 0.78, 0.72; T5- social withdrawal and current involvement at T4 to T7 tested self-report T6 α = 0.81, 0.73). if higher withdrawal consistently predicted being single across partner-report late adolescence and early adulthood. The CLPM was spe- Perceived support cified with stability paths of the same variable over time, the within-wave associations between withdrawal and current At T5 and T6, participants and their partners responded to involvement, the cross-lagged paths from one variable at T to items assessing to what extent they felt supported by their the other variable at T , and a path from sex and ethnicity to x+1 romantic partner in specific domains. Items included Deci- T4 withdrawal and current involvement status. Good model sions about work or education; Problems with your health; fitwas defined as a comparative fit index (CFI) > 0.90; stan- Spending your free time and your social contacts; Practical dardized root mean residual (SRMR) < 0.06, and root mean things; and Personal matters that concern you. Items were square error of approximation (RMSEA) < 0.06. rated on a 4-point scale, from 1 = no support, to 4 = a lot of Second, to test if entering a romantic relationship for the support from partner (T5-T6 α = 0.72, 0.83, and first time reduced adolescents’ withdrawal across 19 to 29 self-report α = 0.82, 0.79). The scale represents the mean of years, a dummy-coded a time-varying covariate (TVC), partner-report the items ratings, with higher scores indicating more per- “entering first relationship,” with 0 = has not yet entered a ceived support from one’s romantic partner. relationship and 1 = has entered first romantic relationship (and TVC = 1 for all subsequent waves after entry) was Perceived conflict used. Again, the withdrawal LGCM with sex and ethnicity as covariates was specified, now subsequently allowing the Perceived relationship conflict was assessed at waves T5 and dummy-coded TVC to have an effect on withdrawal per T6. Participants and their partners responded to five items time point. The WLSMV estimator was used due to the about conflict situations occurring in the last 12 months, dichotomous nature of the TVC. A sensitivity analysis including: Fierce discussions between you and your partner; tested if the TVC effect was present within persons within a One person blamed the other strongly; You didn’t talk to each multilevel modeling framework. The difference between the other for a while; Fights got out of hand;and You no longer original, SEM-based model and the multilevel model is that lived together (if applicable). Responses were given on a 3- the multilevel model is a two-level (vs. single-level), uni- point scale, where 1= no, 2= one time, and 3= multiple variate (vs. multivariate) model in which TVCs have ran- times (T5-T6 α = 0.70, 0.72, and α = 0.73, dom effect (vs. fixed effect) coefficients that vary over self-report partner-report 0.70). The scale score represents the mean item ratings, with individuals (vs. over time). This means that in the sensi- higher scores indicating more perceived relationship conflict. tivity model, the effect of entering a romantic relationship for the first time on withdrawal is assumed to be constant Statistical Analyzes across time, and the TVC effect on withdrawal represents the average shift in withdrawal when entering the first Analyzes were conducted in MPlus Version 80.4 (Muthén & romantic relationship. Bayesian estimation with unin- Muthén, 1998–2017) using maximum likelihood with robust formative priors, 100 thousand iterations, and 2 Monte Journal of Youth and Adolescence (2021) 50:1766–1781 1773 Table 3 Goodness-of-fit Model n χ df CFI SRMR RMSEA [95% CI] statistics of all structural equation models Unconditional LGCM 1702 50.6 5 0.96 0.04 0.07 [0.06, 0.09] TVC LGCM 1597 64.5 21 0.98 0.04 0.04 [0.03, 0.05] Involvement CLPM 1710 217.3 24 0.91 0.05 0.07 [0.06, 0.08] Commitment APIM 402 30.2 12 0.94 0.02 0.09 [0.05, 0.13] Satisfaction APIM 402 7.8 12 1 0.01 0.00 [0.00, 0.05] Conflict APIM 402 21.0 12 0.98 0.03 0.06 [0.00, 0.10] Support APIM 402 14.3 12 0.99 0.02 0.03 [0.00, 0.08] n = number of participants included in the analyzes; CFI = comparative fit index; SRMR = standardized root mean residual; RMSEA = root mean square error of approximation. LGCMs = latent growth curve models; TVC = time-varying covariate model (i.e. entering first romantic relationship); CLPM = cross-lagged panel model (current romantic relationship involvement); APIM = actor partner interdependence model. Good model fit for all models was determined by the following criteria: CFI > 0.90; SRMR < 0.06; RMSEA < 0.06 Carlo chains was used; the model stabilized at PSR = to T6 participant satisfaction. T6 participant and partner 10.001. satisfaction were regressed on the two T5 interaction terms. Third, to investigate the longitudinal, bi-directional asso- T6 withdrawal was regressed on T5 withdrawal, participant ciations between withdrawal and romantic relationship qua- and partner satisfaction, and the two interaction terms. All lities, as reported by both the participants and their partners, a predictors were correlated, thereby estimating actor effects series of longitudinal actor-partner interdependence models while controlling for partner effects, and vice versa. The (APIMs; Cook & Kenny, 2005) were performed, with data residual variances of the outcome variables were also cor- from 402 participants who had the same partner across T5 related to control for additional sources of non-independence. and T6. Six same-sex romantic dyads were excluded because Paths from participants’ ethnicity and length of the romantic of the sex grouping of the analyzes, explained below. relationship (at T5) to all T5 predictors were added as con- Separate APIMS were tested for relationship commitment, trols. To correct for the false discovery rate (FDR) in multiple satisfaction, perceived support, and perceived conflict; satis- testing, the Benjamini and Hochberg (1995)method was faction is used to illustrate the modeling procedure. The applied. This method is more powerful and less conservative APIM is designed to account for non-independence of than the Bonferroni procedure. All observed p-values from observations within interpersonal relationships, with the the APIMs per sex (n= 84 for both females and males) were romantic dyad as the unit of analysis instead of the indivi- ranked, and alpha was specified as 0.05. The Benjamini- dual. APIMs produce estimates of actor effects (how much a Hochberg adjusted p-value criterion was p < 0.009 for person’s satisfaction is predicted by their own prior satis- females and p < 0.014 for males. For simplicity, the criterion faction), and partner effects (how much a person’s satisfac- of p < 0.01 was used for all APIM effects. tion is predicted by their partner’s prior satisfaction). The traditional longitudinal APIM was extended in the current study by including participants’ T5 and T6 withdrawal to the Results model and using T5 withdrawal as both an actor and a moderating variable. To test if withdrawal moderated long- Social Withdrawal Predicts Romantic Relationship itudinal actor and partner effects, all T5 predictors were Involvement and Quantity centered, and two interaction terms were created, namely participant’s withdrawal by participant’s satisfaction, and Table 3 depicts the fit of all models. The LGCM (Fig. 1) participant’swithdrawalby partner’s satisfaction. To inves- indicated that having a higher withdrawal intercept sig- tigate differences between females and males, sex was spe- nificantly predicted a higher likelihood of never having cified as a grouping variable, thereby computing separate been involved in a romantic relationship by early adulthood estimates for females and males within the same APIM fra- (β = −0.32, p < 0.001), becoming romantically involved at mework. A grouping approach (instead of distinguishing an older age (β = 0.14, p < 0.001), and having a shorter actors and partners by sex) was necessary because with- maximum romantic relationship duration (β = −0.13, p < drawal data for only one member per dyad was available. 0.001). The withdrawal slope significantly predicted rela- Actor effects were estimated by specifying paths from T5 to tionship duration (β = −0.10, p = 0.029); a steeper increase T6 participant’s satisfaction and T5 to T6 partner’s satisfac- in withdrawal was associated with a shorter maximum tion. Partner effects were estimated by specifying paths from relationship duration. Neither the withdrawal intercept nor T5 participant to T6 partner satisfaction; and from T5 partner the slope predicted the number of romantic partners. The 1774 Journal of Youth and Adolescence (2021) 50:1766–1781 model also showed that participants of non-Dutch origin positive, and those of involvement were moderate and were on average more withdrawn (β = 0.30, p = 0.005), and positive. Within-wave correlations between withdrawal and had a shorter maximum relationship duration (β = −0.24, p involvement were small and negative, indicating a = 0.003) than Dutch-origin ones. Males were older at their decreased likelihood of current romantic involvement for first romantic relationship (β = 0.16, p = 0.010) and had a more withdrawn youth. Cross-lagged paths were significant shorter maximum relationship duration (β = −0.25, p < from withdrawal to current involvement status, but not vice 0.001) than females. versa. This means that across all ages, more withdrawn Figure 2 depicts the significant paths in the withdrawal youth were less likely to have a romantic relationship and current involvement CLPM. The stability correlations approximately three years later, but relationship status, in of withdrawal at consecutive time points were strong and turn, did not predict future withdrawal levels. The Effect of Entering a Relationship for the First Time on Withdrawal The effects of the TVC “entering first romantic relation- ship” on withdrawal were significant and negative at all waves: T4 (β = −0.26, p < 0.001); T5 (β = −0.27, p < 0.001); T6 (β = −0.15, p = 0.043); T7 (β = −0.28, p = 0.008). This means that across all ages, participants who entered a romantic relationship for the first time were less withdrawn than those who had not yet become involved. These results replicated in the within-person effects of the multilevel model. The overall within-person effect of entering a romantic relationship for the first time on with- drawal was moderate and negative (β = −0.33, 95% CI [−0.44, −0.21]), indicating a moderate decrease in with- drawal when an adolescent became involved for the first time. Longitudinal Within-Dyad Associations Between Withdrawal and Self- and Partner-Reported Fig. 1 Significant paths from the Latent Growth Curve Model (LGCM), with the social withdrawal intercept and slope across 19 to Relationship Qualities 29 years predicting lifetime involvement, age at first romantic rela- tionship, and maximum romantic relationship duration, controlling for Table 4 depicts all APIM effects. For females but not males, ethnicity (0 = Dutch-origin, 1 = non-Dutch-origin) and sex (0 = being in a romantic relationship for a longer time was female, 1 = male). No significant predictors of number of romantic partners emerged associated with lower withdrawal and higher self- and Fig. 2 Significant effects in the Cross-Lagged Panel Model (CLPM) of social withdrawal and current romantic relationship involvement (0 = not currently in a relationship, 1 = currently in a relationship) across 19 to 29 years, controlling for ethnicity (0 = Dutch-origin, 1 = non-Dutch-origin) and sex (0 = female, 1 = male). 95% CI depicted in brackets Journal of Youth and Adolescence (2021) 50:1766–1781 1775 Table 4 Standardized estimates of the Actor-Partner Interdependence Models (APIMs) of participant social withdrawal and participant- and partner-reported commitment, satisfaction, support, and conflict, by participant’s sex Females Males Commitment Satisfaction Support Conflict Commitment Satisfaction Support Conflict Within-wave associations S ↔ P 0.19 0.11 0.24 0.43 0.14 0.20 0.10 0.46 5 5 SW ↔ S −0.06 −0.11 −0.08 0.06 −0.15 −0.24 −0.05 −0.16 5 5 SW ↔ P −0.01 −0.07 −0.10 0.10 −0.05 −0.08 −0.21 0.24 5 5 S ↔ P 0.08 0.32 0.13 0.15 0.30 0.13 0.09 0.51 6 6 SW ↔ S −0.13 −0.10 −0.07 0.17 −0.21 −0.28 −0.25 0.34 6 6 SW ↔ P 0.02 −0.10 −0.01 0.10 −0.17 −0.12 0.12 0.16 6 6 Actor effects S → S 0.34 0.24 0.37 0.52 0.40 0.40 0.19 0.59 5 6 P → P 0.48 0.45 0.51 0.52 0.48 0.57 0.35 0.18 5 6 SW → SW 0.65 0.65 0.65 0.61 0.72 0.72 0.72 1.02 5 6 Partner effects S → P 0.06 0.02 0.02 0.25 −0.06 0.01 0.15 0.32 5 6 SW → P −0.08 −0.02 −0.16 0.05 0.01 0.13 −0.13 0.07 5 6 + + P → S −0.01 0.17 −0.07 0.21 0.08 0.21 −0.05 −0.20 5 6 SW → S 0.05 −0.01 0.02 0.04 −0.26 −0.34 −0.33 0.31 5 6 S → SW 0.01 −0.06 −0.07 0.19 0.04 0.07 −0.10 0.33 5 6 P → SW 0.04 0.01 0.02 −0.12 −0.10 −0.29 −0.03 −0.23 5 6 Interaction effects SW *S → S 0.12 −0.01 −0.11 −0.01 0.26 0.11 0.05 0.19 5 5 6 SW *P → S −0.01 0.15 0.22 0.15 0.03 0.09 −0.27 −0.16 5 5 6 SW *P → P −0.004 0.04 −0.08 −0.02 −0.17 −0.18 −0.17 −0.24 5 5 6 + + SW *S → P 0.03 −0.07 0.11 −0.12 0.11 0.16 −0.13 0.43 5 5 6 SW *S → SW −0.01 −0.06 −0.05 −0.04 −0.02 −0.02 −0.05 0.50 5 5 6 SW *P → SW 0.04 0.09 0.02 0.13 0.03 −0.01 0.08 −0.67 5 5 6 S = self-reported relationship quality at T5; P = partner-reported relationship quality at T5; SW = self-reported social withdrawal at T5; S = 5 5 5 6 self-reported relationship quality at T6; P = partner-reported relationship quality at T6; SW = self-reported social withdrawal at T6; Double- 6 6 headed arrows indicate a correlational path and single-headed arrows a regression path Bold coefficients are significant at p < 0.01; 0.01 < p < 0.05 partner-reported commitment. For males, ethnicity was providing confidence of robust (rather than chance) find- associated with withdrawal, with non-Dutch-origin males ings. Higher initial (T5) withdrawal predicted lower future reporting higher withdrawal than Dutch-origin males. (T6) self-reported commitment, satisfaction, and support for males, indicating that males who were initially more with- Actor and partner effects drawn reported decreased commitment to, satisfaction with, perceived support from their female partners. For both For females and males, the longitudinal actor effects were females and males, self-reported conflict predicted future moderate to strong for withdrawal and all self- and partner- withdrawal, and for females, self-reported conflict predicted reported relationship characteristics, with the exception of partner-reported conflict. Partner-reported satisfaction also males’ partner-reported conflict and self-reported perceived predicted future withdrawal for males. support, which were not significant. This suggests that young adults’ withdrawal and relationship perceptions, and their Interaction effects partners’ relationship perceptions are moderately stable, con- trolling for ethnicity, relationship duration, and partner effects. Simple slope estimates can be found in Table S1, and Seven significant out of 48 possible partner effects simple slope plots of the significant interaction effects in emerged across the APIMs (see Table 4); they showed a Figures S1-S5. Social withdrawal moderated the effect of consistent pattern of the direction and magnitude of effects, T5 to T6 self-reported commitment in males: withdrawn 1776 Journal of Youth and Adolescence (2021) 50:1766–1781 males’ commitment to their female partners was more stable years later across all ages. These results corroborate findings than non-withdrawn males’ commitment. Withdrawal also from previous cross-sectional studies (Roswell & Coplan, moderated several partner effects in males. First, it 2013), and additionally indicate that withdrawn individuals’ enhanced the association between males’ T5 partner- and romantic involvement delays are long-lasting, and that dif- T6 self-reported support: the less perceived support reported ferences between withdrawn and non-withdrawn youth by the female partners of withdrawn males (> +0.50 SD persist into adulthood. The delay of romantic involvement is average withdrawal), the less support these partners subse- likely attributable to withdrawn youth’s deviation from the quently provided to the withdrawn males. For non- formative cascade of development from same-sex friend- withdrawn males, the amount of support received by their ships to mixed-sex peer groups to romantic partnerships partners did not predict future support from their partners. (Nelson et al., 2008). Having had fewer opportunities to Second, withdrawal moderated the association from T5 self- learn how to approach and interact with opposite-sex peers to T6 partner-reported conflict in males. In withdrawn (> in adolescence, withdrawn youth might be particularly +0.50 SD) males, perceiving high levels of relationship anxious in novel romantic situations, or avoid them alto- conflict predicted more future partner-perceived conflict, gether (Barry et al., 2013; Gazelle & Druhen, 2009). Con- whereas non-withdrawn (< 0.50 SD) males’ perceived trary to expectations that withdrawn young adults would conflict did not. Third, withdrawal moderated how partner- have longer-lasting relationships with fewer partners (a reported conflict predicted males’ future withdrawal: having hypothesis that was based on limited research; Leck, 2006), a female partner that perceived high conflict predicted more higher and increasing withdrawal across the early adulthood withdrawal for non-withdrawn males (< −1 SD), and less decade predicted a shorter maximum romantic relationship withdrawal for highly withdrawn males (> +1.4 SD). There duration, and had no effect on the number of romantic was no effect for males that were average on withdrawal. partners. The effect of withdrawal on relationship duration may be due to the fact that withdrawn adolescents were older when they first became romantically involved, leading Discussion to comparatively less possible time for their relationship to have lasted by the point of assessment than non-withdrawn Involvement in romantic relationships is a central devel- young adults’ relationships. It is also probable that the opmental task of late adolescence and early adulthood romantic relationships of withdrawn young adults are more (Roisman et al., 2004). Although previous work has sug- likely to break-up than those of non-withdrawn young gested that withdrawn youth’s formative social experiences adults, leading to shorter-lasting relationships. Because of lead to delays in romantic development, empirical studies of the additional socio-emotional costs of romantic relation- the extent to which withdrawal contributes to deviations ship dissolutions, an empirical test of this latter possibility is from normative romantic development, and in which spe- warranted. cific features, were lacking. To address this gap, this study It is important to note that, while withdrawn and non- tested to what extent withdrawal predicts delays in romantic withdrawn young adults have diverging patterns of romantic involvement and quantity; if entering a romantic relation- involvement, it is not suggested that withdrawn youth should ship for the first time decreases withdrawal; and if with- be involved, especially if they do not want to be. Postponing drawal predicts self- and partner-rated romantic relationship romantic involvement to older ages than their more sociable qualities, and vice versa. Higher withdrawal across the counterparts might even be adaptive for withdrawn indivi- decade of late adolescence and early adulthood predicted duals. Having more time before focusing on “settling down” delays in all aspects of romantic involvement and a shorter with a romantic partner could provide the opportunity to longest-lasting romantic relationship, but did not influence develop in other domains in which withdrawn young adults the number of romantic partners. When an adolescent are also delayed, such as identity development (Barry et al., became romantically involved for the first time, withdrawal 2013), selecting educational and career paths (Hamer & moderately decreased. Despite this decrease, withdrawal Bruch, 1997), reaching higher levels of education and income remained an important factor in couples’ relationship (Nelson et al., 2020; Schmidt et al., 2017), maximizing quality ratings, especially affecting males’ relationship person-environment fit (Shulman & Connolly, 2013), and perceptions and dynamics. These results provide insights developing interpersonal skills in other social relationships into the developmental sequelae of withdrawn adolescents’ (e.g. friends, classmates, colleagues). Catching up to their and young adults’ romantic relationship development. non-withdrawn peers in these domains could then contribute As predicted, higher withdrawal was associated with a to better maintenance of and positive functioning in with- greater likelihood of never having been romantically drawn young adults’ romantic relationships when they do involved, entering a romantic relationship for the first time emerge, and “increase the chances for better provision for the when older, and a greater likelihood of being single three next generation” (Shulman & Connolly, 2013,p.34). Journal of Youth and Adolescence (2021) 50:1766–1781 1777 Withdrawn young adults who do desire a romantic relation- providing and receiving less support, and perceiving more ship but feel unable to initiate contact with potential dating relationship conflict. Females’ withdrawal, in contrast, had partners may nevertheless feel lonely, have low self-esteem, no effect on self- or partner-reported romantic relationship and subsequently withdraw further; hence attaining this goal qualities. These gender disparities may be due to the greater and entering a romantic relationship for the first time can have social acceptance of withdrawal in females, and females marked benefits. being better-prepared for romantic relationships than males. Indeed, young adults’ withdrawal decreased when they The child and adolescent literature suggests that withdrawn entered a romantic relationship for the first time across behaviors are less socially accepted in boys than in girls participants. Based on Developmental Task Theory, with- because inhibited behaviors are viewed as violations of drawn adolescents and young adults experience negative gender-normative expectations of male assertiveness and social sanctions and receive less social support from their dominance (Doey et al., 2014). This seems to apply to peers when they are delayed in becoming romantically romantic partnerships in early adulthood as well: males involved (Furman & Collibee, 2014). When withdrawn appear to be more accepting of their withdrawn female individuals enter their first romantic relationships, they are partners than are females of their withdrawn male partners. no longer off-time in romantic involvement compared to Additionally, females may enter romantic relationships with their peers, and gain the new social identity of “girlfriend” more relationship maintenance skills than males. In child- or “boyfriend”. This shift in identity could lead withdrawn hood and adolescence, girls tend to prefer involvement in young adults to obtain more social status, belonging, and more intimate, dyadic relationships (Hall, 2011; Rose & social support (Raley et al., 2007). Withdrawn young adults Rudolph, 2006), which more closely resemble romantic may be especially sensitive to this shift in identity because relationships than boys’ friendships. Girls’ friendships they are less socially integrated than non-withdrawn young provide more opportunities than boys’ friendships to adults. Especially for them, entering a relationship for the become comfortable self-disclosing, develop intimacy with first time might lead to greater social integration, the others, and learn conflict resolution methods (Giordano development of interpersonal skills, decreased loneliness, et al., 2006), skills which are conducive to maintaining increased self-esteem, and expansion of social networks via intimate relationships. Highly withdrawn males may come the romantic partners, subsequently decreasing withdrawal. to rely on their female partners to take primary responsi- Despite the decrease in withdrawal when entering a bility of the social and emotional aspects of their romantic romantic relationship for the first time, withdrawal remained relationship. Tentative support for this idea is that the a predictor and outcome of several unfavorable romantic results indicated that female partner-perceived conflict relationship qualities. Reiterating the main within-couple predicted less future withdrawal in highly withdrawn males. findings, high withdrawal (1) predicted lower self-reported A withdrawn male may struggle to communicate his feel- commitment, satisfaction, and support in males; (2) was ings and needs to his partner when he feels unhappy in his predicted by higher self-reported conflict in males and relationship (Giordano et al., 2006)—subsequently con- females, higher partner-reported conflict in females, and tributing to other relationship problems and more with- lower partner-reported satisfaction in males; and (3) in drawal—but when his female partner feels unhappy, she males, was associated with interaction patterns in which may be better able to self-disclose, take initiative to resolve partners who perceived less support subsequently provided conflicts, and rebuild intimacy in the relationship (Giordano less support; perceived relationship conflicts predicted et al., 2006; Raley et al., 2007); skills learned in her partner-perceived relationship conflicts; and partner- friendships might contribute to relationship improvements perceived conflicts predicted low males’ future with- and less withdrawal in her male partner. drawal. Taken together, these results indicate that males’ Several limitations should be considered when inter- withdrawal plays a bigger role in the romantic relationship preting these results. First, a broad conceptualization of quality dynamics than females’ withdrawal. These results social withdrawal was used, which did not assess under- are consistent with the theory that the suboptimal romantic lying motivations to withdrawal such as fear of negative relationship qualities of withdrawn individuals are due to evaluation, social disinterest, or peer rejection. The various difficulties with self-disclosing, being responsive, and motivations for withdrawing may be associated with dif- forming intimate bonds with romantic partners (Luster ferent patterns of romantic development. For example, et al., 2013), but suggest that these mechanisms apply pri- unsociable-withdrawn youth have been found to have fewer marily to withdrawn males. Withdrawn males in particular difficulties initiating and maintaining friendships than may struggle to communicate, self-disclose, and form inti- anxious-withdrawn youth (Ladd et al., 2011), and may macy with their partners, subsequently leading to difficul- likewise have fewer difficulties initiating and maintaining ties committing to their partner, feeling less satisfied in the romantic relationships. Yet, withdrawal subtypes overlap relationship and fostering less satisfaction in their partner, and withdrawn youth of any subtype tend to have worse 1778 Journal of Youth and Adolescence (2021) 50:1766–1781 social relationships than non-withdrawn ones (Eggum- same-sex couples, thus limiting the generalizability of the Wilkens et al., 2020; Nelson, 2013). Regardless, future findings to heterosexual young adults. Also limiting the studies could investigate differences in how young adults generalizability was that the sample included young adults from different withdrawal subtypes initiate and maintain from a predominantly Dutch background. Because there is their romantic relationships. cultural variation in romantic relationship development, Second, only romantic relationships that did not end ethnic differences in the associations between withdrawal before the age of 17 years were investigated, and thus the and romantic relationship initiation and maintenance are analyzes did not account for involvement in earlier rela- likely and require more attention in future work. tionships in adolescence. The timing variable therefore Finally, several gaps that could not be addressed remain to reflects the age at which participants entered a more serious, be investigated in future studies. First, a replication of the attachment-based relationship for the first time. This pos- finding that males’ withdrawal particularly affected their sibly neglects formative romantic experiences that set the romantic relationship characteristics is needed. Previous stage for more “adult-like” relationship functioning. In the reports of sex differences in the associations between with- context of withdrawal, this might not have confounded drawal and romantic relationship characteristics have been results heavily because withdrawn young adults were more somewhat mixed. Although there is a theoretical basis to likely to have never been romantically involved and entered expect that males – especially withdrawn ones – face more romantic relationship when older, meaning that these later challenges in their romantic relationships than females, romantic relationships were probably their first ones. directly testing effects of possible mediators (e.g. intimacy, Nevertheless, it would be interesting to explore the effects self-disclosure) would have provided more robust evidence of of withdrawal on romantic relationships from early ado- the proposed mechanisms underlying this association. Sec- lescence, when romantic interests are just emerging and ond, future studies investigating romantic relationships could involvement is more sporadic, to adulthood. include individuals’ desire for having a romantic partner, Third, partner reports were available across only two especially when assessing young adults, who have postponed waves, which may limit the generalizability of the dyadic romantic relationship-related decisions to increasingly older results to young adults who maintained the same partner ages (Arnett, 2000;OECD 2017). It is likely that young across three or more years. Although the duration of these adults, withdrawn and non-withdrawn, who do not particu- relationships was accounted for, young adults tend to larly desire a romantic relationship fare better than those who engage in multiple relationships of various durations across do desire one but are not involved. Relatedly, investigating if early adulthood, which could not be captured. Future stu- there are benefits for withdrawn youth to delay romantic dies could test if results replicate across longer- and shorter- involvement for identity, educational, and career development lasting romantic relationships throughout early adulthood. warrants more attention. Finally, to obtain a more compre- Fourth, it was not possible to investigate to what extent hensive picture of young adults’ romantic relationship both romantic partners’ withdrawal levels affected their development, interpersonal and dyadic behaviors across relationship perceptions and each other’s withdrawal, multiple romantic partners could be investigated. The large because no data about partners’ withdrawal was available. majority of the participants in this study changed partners This is a limitation because both partners’ withdrawal across the late adolescence and early adulthood decade, but influences the quality of the romantic relationship, and there within-dyad dynamics could only be investigated within only may be differential effects between couples with one one relationship. It would be interesting to see how social withdrawn partner and those with both. Partners’ with- withdrawal affects romantic partner selection, if withdrawal drawal levels may interact to predict relationship outcomes, affects relationship functioning in the same way across dif- for better or for worse. On the one hand, having similarly ferent partners, and how withdrawn individuals cope with high withdrawal levels may contribute to greater mutual romantic relationship dissolutions. understanding and acceptance of one another’s inhibited behaviors within couples. On the other hand, withdrawal in both partners may lead to a lower quality relationship Conclusion because both partners may be non-communicative and hinder intimacy development. The lack of partners’ with- The current study provided insights into the links between drawal data also meant that the sample needed to be split by young adults’ withdrawal and romantic relationship devel- sex in order to investigate sex differences in within-dyad opment and a general theoretical framework which can be associations. This would not have been necessary if data on applied in future investigations into the complex social and partners’ withdrawal were available, and not splitting the romantic worlds of withdrawn young adults. This study sample would have increased power to detect smaller investigated the longitudinal effects of social withdrawal on effects. Grouping by sex also required the exclusion of deviations from normative romantic development in late Journal of Youth and Adolescence (2021) 50:1766–1781 1779 adolescence and early adulthood, and the interplay between Data Sharing and Declaration The datasets analyzed during the cur- rent study are not publicly available but are available from the cor- withdrawal and couples’ relationship perceptions. The results responding author on reasonable request. indicated that withdrawal in late adolescence and early adulthood contributed to delays in romantic relationship Compliance with Ethical Standards involvement, and was associated with certain romantic rela- tionship quantity and quality features. Withdrawn young Conflict of Interest The authors declare no competing interests. adults became romantically involved when they were older, and were more likely to have never been involved by adult- Ethical Approval All procedures performed in studies involving hood, likely due to their heightened anxiety and avoidance of human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 novel social situations. There might be advantages to this Helsinki Declaration and its later amendments or comparable ethical delay, as withdrawn young adults have more time to catch up standards. The Dutch Central Committee on Research Involving to their more sociable peers in other domains. Withdrawn Human Subjects approved the Tracking Adolescents’ Individual Lives youth who nevertheless initiated a romantic relationship for (TRAILS) study. the first time benefited from it: they became less withdrawn. Informed Consent Participants and their romantic partners provided Entering a romantic relationship for the first time may written, informed consent at all waves. improve social integration, interpersonal skills, socio- emotional functioning, and social networks. Regardless of Preregistration The study design, hypotheses, and analyzes were pre- this initial decrease in withdrawal, withdrawal affected the registered on the Open Science Framework (10.17605/OSF.IO/3FJQD). qualities of males’ romantic relationships, possibly due to males’ withdrawal being less socially accepted, and with- Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. drawn males having particular difficulty communicating, self- disclosing, and building intimacy with their partners. Highly Open Access This article is licensed under a Creative Commons withdrawn males may come to rely on their female partners to Attribution 4.0 International License, which permits use, sharing, take primary responsibility of the social and emotional aspects adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the of their romantic relationship, perhaps because females tend source, provide a link to the Creative Commons license, and indicate if to be more socialized for maintaining romantic relationships. changes were made. The images or other third party material in this Many of these proposed mechanisms remain to be directly article are included in the article’s Creative Commons license, unless and empirically tested. Because early adulthood is character- indicated otherwise in a credit line to the material. If material is not ized by decisions about long-term commitments, including included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted those in romantic relationships such as partner selection, use, you will need to obtain permission directly from the copyright cohabitation and marriage, and family formation, continued holder. To view a copy of this license, visit http://creativecommons. investigations into the effects of withdrawal on the develop- org/licenses/by/4.0/. mental tasks of this period of life are warranted. References Acknowledgements This research is part of the TRacking Adoles- cents’ Individual Lives Survey (TRAILS). Participating centers of Achenbach, T. M., & Rescorla, L. A. (2003). Manual for the ASEBA TRAILS include various departments of the University Medical Center Adult Forms & Profiles. Burlington: University of Vermont. and University of Groningen, the University of Utrecht, the Radboud Arnett, J. J. (2000). Emerging adulthood: a theory of development Medical Center Nijmegen, and the Parnassia Bavo group, all in the from the late teens through the twenties. American Psychologist, Netherlands. We are grateful to everyone who participated in this 55, 468–480. https://doi.org/10.1037/0003-066X.55.5.469. research or worked on this project to and make it possible. Barry, C. M., Nelson, L. J., & Christofferson, J. L. (2013). Asocial and afraid: an examination of shyness and anxiety in emerging Authors’ Contributions S.A.B. conceived of the study, participated in adulthood. Journal of Family Studies, 19(1), 2–18. https://doi. its design, performed the statistical analyzes, participated in the org/10.5172/jfs.2012.1979. interpretation of data, and drafted the manuscript; J.S.R. and W.H.J.M. Barzeva, S. A., Meeus, W. H. J., & Oldehinkel, A. J. (2019a). Social participated in the design of the study, interpretation of the data, and withdrawal in adolescence and early adulthood: measurement provided feedback on all drafts of the manuscript; TO participated in issues, normative development, and distinct trajectories. Journal the design and coordination of the study, interpretation of the data, and of Abnormal Child Psychology, 47, 865–879. https://doi.org/10. provided feedback on all drafts of the manuscript. All authors read and 1007/s10802-018-0497-4. approved the final manuscript. Barzeva, S. A., Richards, J. S., Meeus, W. H. J., & Oldehinkel, A. J. (2019b). The social withdrawal and social anxiety feedback loop Funding TRAILS has been financially supported by various grants and the role of peer victimization and acceptance in the pathways. from the Netherlands Organization for Scientific Research (NWO), Development and Psychopathology, 32(4), 1402–1417. https:// ZonMW, GB-MaGW, the Dutch Ministry of Justice, the European doi.org/10.1017/S0954579419001354. Science Foundation, the European Research Council, BBMRI-NL, and Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery the participating universities. rate: A practical and powerful approach to multiple testing. 1780 Journal of Youth and Adolescence (2021) 50:1766–1781 Journal of the Royal Statistical Society: Series B, 57(1), 289-300. De Winter, A. F., Oldehinkel, A. J., Veenstra, R., Brunnekreef, J. A., https://doi.org/10.1111/j.2517-6161.1995.tb02031.x Verhulst, F. C., & Ormel, J. (2005). Evaluation of non-response Boisvert, S., & Poulin, F. (2016). Romantic relationship patterns from bias in mental health determinants and outcomes in a large adolescence to emerging adulthood: Associations with family and sample of pre-adolescents. European Journal of Epidemiology, peer experiences in early adolescence. Journal of Youth and Ado- 20, 173–181. https://doi.org/10.1007/s10654-004-4948-6. lescence, 45,945–958. https://doi.org/10.1007/s10964-016-0435-0. Doey, L., Coplan, R. J., & Kingsbury, M. (2014). Bashful boys and coy Booth-LaForce, C., & Oxford, M. L. (2008). Trajectories of social girls: A review of gender differences in childhood shyness. Sex withdrawal from grades 1 to 6: prediction from early parenting, Roles, 70, 255–266. https://doi.org/10.1007/s11199-013-0317-9. attachment, and temperament. Developmental Psychology, 44(5), Eggum-Wilkens, N. D., An, D., Zhang, L., & Costa, M. (2020). Co- 1298–1313. https://doi.org/10.1037/a0012954. occurrence of and cross-informant agreement on shyness, Campbell, L., Simpson, J. A., Boldry, J., & Kashy, D. A. (2005). Per- unsociability, and social avoidance during early adolescence. Social ceptions of conflict and support in romantic relationships: The role Development, 29,73–88. https://doi.org/10.1111/sode.12398. of attachment anxiety. Journal of Personality and Social Psychol- Furman, W., & Collibee, C. (2014). A matter of timing: developmental ogy, 88(3), 510-531. https://doi.org/10.1037/0022-3514.88.3.510 theories of romantic involvement and psychosocial adjustment. Carver, K., Joyner, K., & Udry, J. R. (2003). National estimates of Developmental Psychopathology, 26, 1149–1160. https://doi.org/ adolescent romantic relationships. In P. Florsheim (Ed.), Ado- 10.1017/S0954579414000182. lescent romantic relations and sexual behavior: theory, research, Gazelle, H., & Druhen, M. J. (2009). Anxious solitude and peer and practical implications (p. 23-56). Lawrence Erlbaum exclusion predict social helplessness, upset affect, and vagal Associates Publishers. regulation in response to behavioral rejection by a friend. Caspi, A., Elder, G. H., & Bem, D. J. (1988). Moving away from the Developmental Psychology, 45(4), 1077–1096. https://doi.org/10. world: life-course patterns of shy children. Developmental Psy- 1037/a0016165. chology, 24(6), 824–831. Giordano, P. C., Manning, W. D., & Longmore, M. A. (2006). Ado- Chen, X., & Tse, H. C. (2008). Social functioning and adjustment in lescent romantic relationships: An emerging portrait of their Canadian-born children with Chinese and European backgrounds. nature and developmental significance. In A. C. Crouter & A. Developmental Psychology, 44, 1184–1189. https://doi.org/10. Booth (Eds.), Romance and sex in adolescence and emerging 1037/0012-1649.44.4.1184. adulthood: Risks and opportunities (pp. 127–150). Mahwah, Collibee, C., & Furman, W. (2015). Quality counts: developmental New Jersey: Lawrence Erlbaum Associates. shifts in associations between romantic relationship qualities and Hall, J. A. (2011). Sex differences in friendship expectations: a meta- psychosocial adjustment. Child Development, 86(5), 1639–1652. analysis. Journal of Social & Personal Relationships, 28, https://doi.org/10.1111/cdev.12403. 723–747. https://doi.org/10.1177/0265407510386192. Collins, W. A. (2003). More than myth: The developmental sig- Hamer, R. J., & Bruch, M. A. (1997). Personality factors and inhibited nificance of romantic relationships during adolescence. Journal of career development: testing the unique contributions of shyness. Research on Adolescence, 13(1), 1–24. https://doi.org/10.1111/ Journal of Vocational Behavior, 50, 382–400. https://doi.org/10. 1532-7795.1301001. 1006/jvbe.1996.1542. Collins, W. A., & Van Dulmen, M. (2006). “The course of true love(s) Huisman, M., Oldehinkel, A. J., de Winter, A., Minderaa, R. B., de …”: Origins and pathways in the development of romantic rela- Bildt, A., Huizink, A. C., & Verhulst, F. C. (2008). Cohort tionships. In A. Booth & A. Crouer (Eds.), Romance and sex in profile: The Dutch ‘TRacking Adolescents’ Individual Lives’ adolescence and emerging adulthood: risks and opportunities Survey’; TRAILS. International Journal of Epidemiology, 37(6), (pp. 63–86). Mahwah, NJ: Erlbaum. 1227–1235. https://doi.org/10.1093/ije/dym273. Collins, W. A., Welsh, D. P., & Furman, W. (2009). Adolescent Kingery, J., Erdley, C., Marshall, K., Whitaker, K., & Reuter, T. romantic relationships. Annual Review of Psychology, 60,631–652. (2010). Peer experiences of anxious and socially withdrawn https://doi.org/10.1146/annurev.psych.60.110707.163459. youth: An integrative review of the developmental and clinical Connolly, J., Craig, W., Goldberg, A., & Pepler, D. (2004). Mixed- literature. Clinical Child & Family Psychology Review, 13(1), gender groups, dating, and romantic relationships in early ado- 91–128. https://doi.org/10.1007/s10567-009-0063-2. lescence. Journal of Research on Adolescence, 14(2), 185–207. Ladd, G. W., Kochenderfer-Ladd, B., Eggum, N. D., Kochel, K. P., & https://doi.org/10.1111/j.1532-7795.2004.01402003.x. McConnell, E. M. (2011). Characterizing and comparing the Connolly, J., Furman, W., & Konarski, R. (2000). The role of peers in friendships of anxious-solitary and unsociable preadolescents. the emergence of heterosexual romantic relationships in adoles- Child Development, 82, 1434–1453. https://doi.org/10.1111/j. cence. Child Development, 71(5), 1395–1408. https://doi.org/10. 1467-8624.2011.01632.x. 1111/1467-8624.00235. Lantagne, A., & Furman, W. (2017). Romantic relationship development: Connolly J., & McDonald, K. (2020) Cross-cultural perspectives on the interplay between age and relationship length. Developmental dating and marriage. Encyclopedia of Child and Adolescent Psychology, 53(9), 1738–1749. https://doi.org/10.1037/dev0000363. Development, Wiley-Blackwell Press. Leck, K. (2006). Correlates of minimal dating. The Journal of Social Cook, W. L., & Kenny, D. A. (2005). The actor-partner inter- Psychology, 146(5), 549–567. https://doi.org/10.3200/SOCP.146. dependence model: A model of bidirectional effects in develop- 5.549-567. mental studies. International Journal of Behavioral Development, Loewenthal, K. M. (2004). An introduction to psychological tests and 29(2), 101–109. https://doi.org/10.1080/01650250444000405. scales. 2nd ed. Hove: Psychology Press. Coplan, R. J., & Armer, M. (2007). A multitude of solitude: A closer Luster, S. S., Nelson, L. J., & Busby, D. M. (2013). Shyness and look at social withdrawal and nonsocial play in early childhood. communication: Impact on self and partner relationship satisfac- Child Development Perspectives, 1,26–32. https://doi.org/10. tion. Journal of Couple & Relationship Therapy, 12, 359–376. 1111/j.1750-8606.2007.00006.x. https://doi.org/10.1080/15332691.2013.836322. Coplan, R. J., Zheng, S., Weeks, M., & Chen, X. (2012). Young Meeus, W. H. J., Branje, S. J. T., van der Valk, I., & de Wied, M. children’s perceptions of social withdrawal in China and Canada. (2007). Relationships with intimate partner, best friend, and Early Child Development and Care, 182, 591–607. https://doi. parents in adolescence and early adulthood: A study of the sal- org/10.1080/03004430.2011.566328. iency of the intimate partnership. International Journal of Journal of Youth and Adolescence (2021) 50:1766–1781 1781 Behavioral Development, 31(6), 569–580. https://doi.org/10. of alternatives, and investment size. Personal Relationships, 5(4), 1177/0165025407080584. 357–391. https://doi.org/10.1111/j.1475-6811.1998.tb00177.x. Meeus, W., Van de Schoot, R., Klimstra, T., & Branje, S. (2011). Shulman, S. & Connolly, J. (2013). The challenge of romantic rela- Personality types in adolescence: Change and stability and links tionships in emerging adulthood: reconceptualization of the field. with adjustment and relationships: A five-wave longitudinal Emerging Adulthood, 1(1), 27–39. https://doi.org/10.1177/ study. Developmental Psychology, 47(4), 1181–1195. https://doi. 2167696812467330. org/10.1037/a0023816. Schmidt, L. A., Tang, A., Day, K. L., Lahat, A., Boyle, M. H., Saigal, Muthén, L. K., & Muthén, B. O. (1998). Mplus User’s Guide. 8th ed. S., & Van Lieshout, R. J. (2017). Personality development within Los Angeles, CA: Muthén & Muthén. 2017. a generational context: life course outcomes of shy children. Nelson, L. J. (2013). Going it alone: Comparing subtypes of with- Child Psychiatry & Human Development, 48, 623–641. https:// drawal on indices of adjustment and maladjustment in emerging doi.org/10.1007/s10578-016-0691-y. adulthood. Social Development, 22, 522–538. https://doi.org/10. Seiffge-Krenke, I. (2003). Testing theories of romantic development 1111/j.1467-9507.2012.00671.x. from adolescence to young adulthood: Evidence of a developmental Nelson, L. J., Jorgensen, N. A., & Clifford, B. N. (2020). Shy and still sequence. International Journal of Behavioral Development, 27(6), struggling: examining the relations between subtypes of social 519–531. https://doi.org/10.1080/01650250344000145. withdrawal and well-being in the 30s. Social Development, 00, Tackett, S. L., Nelson, L. J., & Busby, D. M. (2013). Shyness and 1–17. https://doi.org/10.1111/sode.12486. relationship satisfaction: evaluating the associations between Nelson, L. J., Padilla-Walker, L. M., Badger, S., Barry, C. M., Carroll, shyness, self-esteem, and relationship satisfaction in couples. The J. S., & Madsen, S. D. (2008). Associations between shyness and American Journal of Family Therapy, 41,34–45. https://doi.org/ internalizing behaviors, externalizing behaviors, and relationships 10.1080/01926187.2011.641864. during emerging adulthood. Journal of Youth and Adolescence, Tang, A., Van Lieshout, R. J., Lahat, A., Duku, E., Boyle, M. H., 37, 605–615. https://doi.org/10.1007/s10964-007-9203-5. Saigal, S., & Schmidt, L. A. (2017). Shyness trajectories across OECD. (2017). OECD family database. Retried from Family Database the first four decades predict mental health outcomes. Journal of http://www.oecd.org/els/family/database.htm. Abnormal Child Psychology, 45(8), 1621–1633. https://doi.org/ Oldehinkel, A. J., Rosmalen, J. G. M., Buitelaar, J. K., Hoek, H. W., 10.1007/s10802-017-0265-x. Ormel, J., Raven, D., Reijneveld, S. A., Veenstra, R., Verhulst, F. C., Vollebergh, W. A. M., & Hartman, C. A. (2015). Cohort profile Update: the Tracking Adolescents’ Individual Lives Sur- Stefania A. Barzeva is a PhD researcher at the University of vey (TRAILS). International Journal of Epidemiology, 44, Groningen, University Medical Center Groningen, Interdisciplinary 76–76n. https://doi.org/10.1093/ije/dyu225. Center Psychopathology and Emotion Regulation. Her major research Ozdemir, S. B., Cheah, C. S. L., & Coplan, R. J. (2015). Con- interests include developmental psychopathology, social withdrawal, ceptualization and assessment of multiple forms of social with- adolescence, psychosocial development, and social relationships. drawal in Turkey. Social Development, 24(1), 142–165. https:// doi.org/10.1111/sode.12088. Raley, R. K., Crissey, S., & Muller, C. (2007). Of sex and romance: late adolescent relationships and young adult union formation. Jennifer S. Richards is a Postdoctoral researcher at the University of Journal of Marriage and Family, 69(5), 1210–12226. https://doi. Groningen, University Medical Center Groningen, Interdisciplinary org/10.1111/j.1741-3737.2007.00442.x. Center Psychopathology and Emotion Regulation. Her major research Roisman, G. I., Masten, A. S., Coatsworth, J. D., & Tellegen, A. interests include resilience and mental health development, (2004). Salient and emerging developmental tasks in the transi- intergenerational transmission, family and peer experiences, parent- tion to adulthood. Child Development, 75(1), 123–133. https:// doi.org/10.1111/j.1467-8624.2004.00658.x. child interactions, differential susceptibility, and gene-environment Rose,A.J., &Rudolph,K.D.(2006). Areviewofsex differences in interactions. peer relationship processes: potential trade-offs for the emotional and behavioral development of girls and boys. Psychological Bul- letin, 132,98–131. https://doi.org/10.1037/0033-2909.132.1.98. Wim H. J. Meeus is Emeritus Professor of Adolescent Development at Roswell, H. C., & Coplan, R. J. (2013). Exploring links between Utrecht University and Developmental Psychology at Tilburg shyness, romantic relationship quality, and well-being. Canadian University. His major research interests include adolescence, social Journal of Behavioural Science, 45(4), 287–295. https://doi.org/ development, longitudinal research, personality and identity 10.1037/a0029853. development, social relationships, and problem behavior. Rubin, K. H., Coplan, R. J., & Bowker, J. C. (2009). Social with- drawal in childhood. Annual Review of Psychology, 60, 141–171. https://doi.org/10.1146/annurev.psych.60.110707.163642. Rubin, K. H., Wojslawowicz, J. C., Rose-Krasnor, L., Booth-LaForce, Albertine J. Oldehinkel is a Full Professor at the University of C., & Burgess, K. B. (2006). The best friendships of shy/with- Groningen, University Medical Center Groningen, Interdisciplinary drawn children: prevalence, stability, and relationship quality. Center Psychopathology and Emotion Regulation. Her major research Journal of Abnormal Child Psychology, 34(2), 143–157. https:// interests include developmental psychopathology, affective disorders, doi.org/10.1007/s10802-005-9017-4. adolescence, social development, longitudinal research, parent-child Rusbult, C. E., Martz, J. M., & Agnew, C. R. (1998). The investment relations, and person-environment interactions. model scale: measuring commitment level, satisfaction level, quality http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Youth and Adolescence Springer Journals

Social Withdrawal and Romantic Relationships: A Longitudinal Study in Early Adulthood

Loading next page...
 
/lp/springer-journals/social-withdrawal-and-romantic-relationships-a-longitudinal-study-in-908raieLPM
Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2021
ISSN
0047-2891
eISSN
1573-6601
DOI
10.1007/s10964-021-01469-1
Publisher site
See Article on Publisher Site

Abstract

Involvement in romantic relationships is a salient developmental task in late adolescence and early adulthood, and deviations from normative romantic development are linked to adverse outcomes. This study investigated to what extent social withdrawal contributed to deviations from normative romantic development, and vice versa, and the interplay between withdrawal and couples’ relationship perceptions. The sample included 1710 young adults (55–61% female) from the Tracking Adolescents’ Individual Lives Survey cohort and their romantic partners. Data were collected across 4 waves, covering romantic relationships from ages 17 to 29 years. The results showed that higher withdrawal predicted a higher likelihood of romantic non-involvement by adulthood, consistently being single at subsequent waves, and entering one’s first relationship when older. Withdrawal moderately decreased when youth entered their first relationship. Male’s withdrawal in particular affected romantic relationship qualities and dynamics. These results provide new insights into the developmental sequelae of withdrawn young adults’ romantic relationship development. ● ● ● ● ● ● Keywords Dating Early adulthood Late adolescence Longitudinal Relationship quality Romantic relationships Social withdrawal Introduction common, but also shift in their qualities and functions across these ages: they become longer-lasting and a main Developmental Task Theory (Roisman et al., 2004) posits source of support and intimacy (Collibee & Furman, 2015; that involvement in romantic relationships becomes a sali- Lantagne & Furman, 2017). A comprehensive investigation ent developmental task during late adolescence and early of the specific features of romantic relationship develop- adulthood, and delaying or not meeting this task is related to ment that withdrawal affects and is affected by has been adverse outcomes. Indeed, romantic relationships become lacking in the extant literature. This study focused on the increasingly prevalent throughout these ages, with over half longitudinal associations between social withdrawal and of adolescents having been romantically involved by the romantic relationships in late adolescence and early adult- age of 16, and the large majority by early adulthood (Carver hood. Specifically, it investigated to what extent withdrawal et al., 2003). Romantic relationships not only become more contributed to deviations from normative romantic devel- opment and the interplay between withdrawal and couples’ perceptions of the quality of relationships. Three key fea- Supplementary information The online version contains tures were examined as indicators of romantic development: supplementary material available at https://doi.org/10.1007/s10964- involvement (lifetime involvement, current involvement, 021-01469-1. and timing), relationship quantity (number of partners and relationship duration), and relationship quality (commit- * Stefania A. Barzeva s.a.barzeva@umcg.nl ment, satisfaction, support, and conflict; Collins, 2003). Social withdrawal is an umbrella term referring to the Department of Psychiatry, Interdisciplinary Center voluntary self-isolation from familiar and unfamiliar others Psychopathology and Emotion Regulation, University of through the consistent display of solitary behaviors, such as Groningen, University Medical Center Groningen, Groningen, The Netherlands avoiding social interaction and spending excessive time alone (Rubin et al., 2009). The motivation to withdrawal Research Center Adolescent Development, Utrecht University, Utrecht, The Netherlands varies between individuals and differentiates three types of 1234567890();,: 1234567890();,: Journal of Youth and Adolescence (2021) 50:1766–1781 1767 withdrawal: shyness, unsociability, and avoidance (Coplan (Roswell & Coplan, 2013; with an exception in Schmidt & Armer, 2007; Ozdemir et al., 2015). Phenotypic with- et al. 2017). The relations between withdrawal and romantic drawal behaviors overlap across these withdrawal types. relationship quantity have never been tested directly, but The current study uses the term “social withdrawal” to refer one study found that shy young adults date less frequently to the global, multidimensional, behavioral phenotype of (Leck, 2006), suggesting a possible effect of withdrawal on voluntary self-isolation. Previous studies using this global the quantitative features of romantic relationships, such as conceptualization have indicated that 12 to 23% of indivi- the number of partners and the duration of relationships. If duals are persistently withdrawn throughout adolescence, withdrawn youth likewise date less frequently, it would and that withdrawn adolescents report less social affiliation, suggest that they change partners less often, leading to social contact, and social competence, and more anxiety having fewer partners with possibly longer relationship than their non-withdrawn peers (Barzeva et al., 2019a; Tang durations. et al. 2017). The romantic development of withdrawn Entering a romantic relationship for the first time might adolescents and young adults may differ from the normative have an effect on youth’s withdrawal. Developmental Task patterns due to deviations in previous social experiences Theory suggests that when life events are off-time— that are usually conducive to romantic involvement, such as occurring earlier or later than the majority of peers—ado- friendships (Kingery et al., 2010). The sequential stage lescents experience negative social sanctions for deviating theory of heterosexual romantic relationship emergence from the normative pattern of development and receive states that youth develop social-emotional competencies in fewer social resources from peers (Furman & Collibee, the context of same-sex friendships in childhood, and start 2014). When adolescents’ life events become normative in to apply these competencies in mixed-sex peer groups the context of their peers’ experiences, social sanctions may during adolescence (Connolly et al., 2004; Connolly et al., be lifted and more social resources provided, leading to 2000). Mixed-sex groups provide opportunities to learn greater socio-emotional adjustment. Additionally, the first how to approach and interact with opposite-sex peers, who romantic relationship is unique and of particular importance are also potential dating partners. Withdrawn youth, how- to young adults’ socio-emotional adjustment because it ever, do not follow this cascade of development (Nelson represents a shift in social identity, namely as a “girlfriend” et al., 2008). They experience difficulties initiating and or “boyfriend” (Raley et al., 2007). This new identity, and maintaining same-sex friendships, subsequently leading to a the social roles and experiences that come with it, changes smaller mixed-sex peer group. Not participating in a mixed- how young adults perceive themselves and are perceived by sex peer group hinders the development of important social others; when entering a romantic relationship for the first skills for romantic relationship initiation and maintenance, time, young adults feel an increase in autonomy, status, and limits the size of adolescents’ dating pool. belonging, and social support, and may be seen by others as Withdrawn youth’s formative experiences may set the more mature and as a potential dating partner (Raley et al., stage for non-normative or delayed romantic relationship 2007). These provisions may lead to a decrease in social involvement and quantity in late adolescence and beyond withdrawal. Although an intriguing proposal, the effect of (Raley et al., 2007; Shulman & Connolly, 2013). That is, reaching the developmental task of romantic involvement having had fewer opportunities to learn how to approach on youth’s withdrawal has not been tested yet. and interact with opposite-sex peers in adolescence (Barry Despite the possible decrease of withdrawal when et al., 2013; Nelson et al., 2008) is likely to contribute to entering a romantic relationship for the first time, withdrawn anxiety, rejection sensitivity (Gazelle & Druhen, 2009), and adolescents and young adults may differ from their non- avoidance of novel social situations, such as when asking withdrawn peers in the qualities of their romantic relation- someone out on a date, and hence to delay romantic ships. By late adolescence and early adulthood, individuals involvement. While this is a provoking notion, the empirical are better equipped with relationship maintenance skills and work on withdrawn individuals’ romantic involvement has start seeking and entering longer-lasting, more committed, been scarce, has mainly concerned the shyness aspect of supportive, and exclusive relationships than early adoles- withdrawal, and shows somewhat mixed findings. A semi- cents (Shulman & Connolly, 2013). Starting from about 17 nal study of shy children born in the 1920s found that men, years of age, young adults transition from more sporadic but not women, who were shy as children were older than romantic relationship involvements that resemble friend- their non-shy counterparts when they first married (Caspi ships to more intimate and committed relationships with a et al., 1988). More recent studies have found that withdrawn greater dyadic orientation (Seiffge-Krenke, 2003; Shulman and inhibited individuals become romantically involved at & Connolly, 2013). In turn, being involved in a committed an older age than their more sociable peers (Boisvert & romantic relationship is linked to a decline in emotional Poulin, 2016; Meeus et al., 2011), and those who were problems in early adulthood (Meeus et al., 2007). The currently involved were less shy than those who were not transition to more committed relationships may occur later 1768 Journal of Youth and Adolescence (2021) 50:1766–1781 for withdrawn young adults, who have less romantic Sex- and ethnicity-based social norms may contribute to experience due to delays in romantic involvement. Thus, differences between males and females and Western and higher withdrawal might predict lower commitment in non-Western youth in romantic relationship development young adults’ romantic relationships, and lower commit- and the effect of withdrawal. Although males and females ment, in turn, predict higher withdrawal. Furthermore, it is may both experience non-normative romantic development, well documented that the formative qualities of adolescent withdrawn males may be particularly affected because friendships are linked to concurrent and future romantic inhibited behavior violates gender-normative expectations relationship qualities (Collins & Van Dulmen, 2006; Collins of male dominance and assertiveness (Doey et al., 2014). et al., 2009). Withdrawn youth tend to report friendships Additionally, females may enter romantic relationships with characterized by low support and high conflict (Rubin et al., more relationship maintenance skills than males. In child- 2006), which likely translates into less satisfying romantic hood and adolescence, girls tend to prefer involvement in relationships. Empirical studies on the effects of withdrawal more intimate, dyadic relationships (Hall, 2011; Rose & on romantic relationship qualities are rare, however, and Rudolph, 2006), which provide more opportunities than have predominantly focused on the effects of shyness on boys’ friendships to become comfortable self-disclosing, relationship satisfaction. These studies have consistently develop intimacy with others, and learn conflict resolution found that shy adolescents and young adults report lower skills (Giordano et al., 2006). Hence, girls’ friendships satisfaction than more sociable individuals (Luster et al., resemble romantic relationships more closely than boys’ 2013; Nelson et al., 2008; Roswell & Coplan, 2013; Tackett friendships. Because of this, boys – especially withdrawn et al., 2013; except in Schmidt et al., 2017). This effect is ones who lack social experience in general – may be less speculated to occur because of withdrawn individuals’ prepared for maintaining intimate relationships than girls. social inhibition and rejection sensitivity, which leads to Little is known about ethnic differences in social with- decreased responsiveness, self-disclosure, and intimacy drawal and romantic relationship development, but some with their romantic partners (Luster et al., 2013). These differences may be expected due to cultural norms around characteristics might likewise be related to support and inhibited behavior (Chen & Tse, 2008; Coplan et al., 2012) conflict. and dating, especially in involvement and timing (Connolly Another possible reason for the low romantic relation- & McDonald, 2020). Thus, sex and ethnicity were included ship quality ratings of withdrawn youth are negative cog- as covariates in all models to adjust for possible nitive biases. When assessing their relationships, confounding. withdrawn individuals might report low quality because they are more attentive toward the negative aspects of their relationships and recall more negative interactions with The Current Study their romantic partners (Gazelle & Duhen, 2009). In that case, their romantic partners might have more positive There is reason to believe that social withdrawal is linked to perceptions of the relationship. While untested in with- deviations from normative romantic relationship develop- drawn young adults’ romantic relationships, these effects ment, and vice versa, but a direct empirical test of the extent have been found in the romantic relationships of indivi- to which withdrawn youth deviate from—and in which duals with high attachment anxiety, which is characteristic features of—normative romantic development has been of withdrawn individuals’ attachment style (Roswell & lacking. The aim of this study was to investigate long- Coplan, 2013; Rubin et al., 2009). Anxiously attached itudinal associations between social withdrawal and individuals perceive more conflict and less support in their romantic relationship development in late adolescence and romantic relationships than securely attached ones, and early adulthood, a crucial life phase for romantic these biased perceptions are associated with greater emo- relationship-related exploration, decisions, and family for- tional distress and decreased satisfaction with and com- mation. This was done, first, by testing to what extent mitment to their romantic partners (Campbell et al., 2005). withdrawal predicts deviations in romantic involvement Thus, romantic partners’ ratings may provide additionally (lifetime involvement, current involvement, timing) and informative, less biased relationship quality information, relationship quantity (relationship length, the number of and social withdrawal may have a differential effect on partners). Higher withdrawal was hypothesized to predict a self- and partner-reported relationship perspectives, but greater likelihood of never having been involved in a empirical evidence has been lacking so far. In sum, how romantic relationship by early adulthood, being single withdrawal directly affects self- and partner-reported across late adolescence and early adulthood, initiating one’s romantic relationship qualities—commitment, satisfaction, first romantic relationship at an older age, and having support, and conflict—and if these relationship qualities longer-lasting relationships, but with fewer partners affect withdrawal, has remained largely unknown. (Hypothesis 1). Second, this study tested if entering a Journal of Youth and Adolescence (2021) 50:1766–1781 1769 romantic relationship for the first time changes one’s Measures withdrawal levels. Withdrawal was expected to decrease after entering a romantic relationship for the first time, Social withdrawal especially in individuals with high pre-involvement with- drawal (Hypothesis 2). A third set of tests addressed the Social withdrawal was assessed at T4 to T7 with the mean question if withdrawal predicts deviations from romantic of five items from the Adult Self-Report (ASR; Achenbach relationship qualities, and vice-versa, and examined the & Rescorla, 2003) withdrawn scale. In a sample of adults interplay between withdrawal and couple’s perceptions of over the age of 18 years, the ASR withdrawn scale had high their relationship. It was hypothesized that higher with- test-retest reliability and correlated moderately positively drawal will be concurrently and longitudinally associated with measures of anxiety and social introversion (Achen- with lower self-reported and partner-reported commitment, bach & Rescorla, 2003). The five items included were: I satisfaction, and support, and higher conflict, which in turn would rather be alone than with others; I am secretive or will be associated with higher withdrawal (Hypothesis 3). keep things to myself; I refuse to talk; I am too shy or timid; and I keep from getting involved with others. These items were selected based on face validity and previous research Method (Booth-LaForce & Oxford, 2008; Tang et al., 2017), and items have been found to be longitudinally measurement Participants invariant in adolescence and early adulthood (Barzeva et al., 2019a, 2019b). Items were rated on a 3-point scale, with 0 The study included 1710 participants from the prospective, = Not at all, 1 = A little or sometimes, and 2 = Always or population-based cohort Tracking Adolescents’ Individual often true, in the past 6 months. Cronbach’s alpha ranged Lives Survey (TRAILS; www.trails.nl) and their romantic from 0.67 to 0.72 at each wave. For scales with fewer than partners, if applicable (N = 463 to 559 partners per assess- ten items, an internal reliability cutoff of α > 0.60 is con- ment wave). The TRAILS sample was recruited from rural sidered acceptable (Loewenthal, 2004). and urban areas of the North of the Netherlands, and includes individuals born between 1989 and 1991. Data Lifetime and current relationship involvement collection began in 2001 when the participants were approximately 11 years old. Assessments occurred 7 times, Lifetime romantic relationship involvement was assessed at every 2 to 3 years, from ages 11 to 29 years, and 64–96% of T4 to T7 with the item, Have you ever had a boyfriend or the initial sample participated in subsequent assessment girlfriend? (0 = no, 1 = yes). Current romantic relationship waves. More details about the TRAILS recruitment and involvement was assessed at T4 to T7 with the item, Do you assessment procedure have been reported elsewhere (De have a girlfriend or boyfriend at the moment? (0 = no, 1 = Winter et al., 2005; Huisman et al., 2008; Oldehinkel et al., yes). 2015). The current study used data from the last four assessment waves of TRAILS (T4-T7), when participants Timing of first relationship were approximately 19, 22, 26, and 29 years old (55–61% female across waves). About 90% of participants were from The approximate timing of participants’ first romantic an ethnically Dutch background. Table 1 shows the parti- relationship in late adolescence was assessed at T4 by cipant and partner demographics, and descriptive statistics subtracting the duration of their current relationship from of the romantic relationship characteristics at each wave, their age at T4. If a participant did not have a partner at T4, and Table 2 the correlations between variables. an item from an Events History Calendar asking if they had started a romantic relationship with someone in the last two Data Collection Procedure years was used instead. Participants could report the month and year in which they started up to two romantic rela- The Dutch Central Committee on Research Involving tionships, and the date of the earliest reported relationship Human Subjects approved the TRAILS study. Participants was used to calculate their age at that time. For those who provided written consent at T4 to T7. Data was collected by entered a relationship for the first time at T5 to T7, their age means of online questionnaires at these waves. At T5 to T7, at entry was determined by subtracting the length of that participants nominated their current romantic partner. first relationship from their age at that wave. Early romantic Research assistants contacted romantic partners for partici- relationships that ended before the age of 17 years were not pation. When the partners consented to participate, they included, whereas those that continued through the age of were sent online questionnaires via email. 17 were. The timing variable therefore reflects the age at 1770 Journal of Youth and Adolescence (2021) 50:1766–1781 Table 1 Participant, partner, and T4 T5 T6 T7 romantic relationship characteristics Participant characteristics N participants 1679 1634 1523 1223 Age, years M (SD) 19.1 (0.59) 22.3 (0.65) 25.6 (0.60) 28.9 (0.59) Female % 55.3 55.1 56.3 61.2 Dutch % 89.6 89.7 90.6 91.0 Lifetime involvement % 81.8 91.9 96.1 97.0 Current involvement % 53.1 64.4 72.3 71.2 Social withdrawal M (SD) 0.31 (0.34) 0.32 (0.36) 0.38 (0.37) 0.34 (0.36) Partner characteristics N partners – 559 463 479 Age, years M (SD) – 24.0 (3.62) 27.2 (3.46) 30.4 (3.83) Age range – 16.8–47.3 18.8–43.2 19.5–53.3 Female % – 38.3 38.0 42.0 Relationship % Married – 3.4 11.4 22.8 Registered partnership – 1.3 13.8 18.0 Cohabitating – 35.1 47.7 41.5 Not cohabitating – 58.5 27.0 17.7 Relationship characteristics M (SD) Number of partners 2.28 (1.70) 2.33 (1.47) 2.21 (1.28) – Relationship duration, months 19.0 (14.8) 31.9 (24.5) 49.0 (33.6) 70.3 (43.7) Commitment (participant) 5.82 (1.42) 6.27 (1.16) 6.52 (0.89) – Commitment (partner) – 6.36 (0.95) 6.57 (0.82) – Satisfaction (participant) 6.19 (1.08) 6.30 (0.93) 6.32 (0.84) – Satisfaction (partner) – 6.28 (0.89) 6.27 (0.86) – Support (participant) – 3.50 (0.41) 3.50 (0.48) – Support (partner) – 3.36 (0.51) 3.43 (0.46) – Conflict (participant) – 1.35 (0.34) 1.46 (0.39) – Conflict (partner) – 1.51 (0.38) 1.50 (0.37) – Other ethnicities include: Turkish, Moroccan, Surinam, Antillean, Indonesian or Mollucan, and Other not specified The average maximum relationship duration across participants was 54.9 months (SD = 38.8; range = 0.25–192). Scores could range from 1–10 for number of relationships, 1–7 for commitment and satisfaction, 1–4 for support, and 1–3 for conflict. Dashes indicate that the data was not collected at the assessment wave. which participants entered a more serious, attachment-based relationship with your current partner? (in months) at T4 to romantic relationship for the first time. T7. The maximum reported duration across the four waves was used in the analyzes to account for relationships that Number of partners continued throughout multiple assessment waves. The number of romantic partners that participants had had Commitment was assessed at T6 with the item, How often in your life have you had a steady boyfriend or girlfriend? (rated from Romantic relationship commitment was assessed with mean 0 to 10 or more times). of the three-item Commitment subscale of the Investment Model Scale (IMS; Rusbult et al., 1998). Participants Relationship duration completed the IMS at T4 to T6, and their partners at T5 and T6. The three items were: I am focused on the long-term The duration of the current romantic relationship was future of my relationship; I want my relationship to last a assessed with the item, How long have you had a very long time; and I want my relationship to continue Journal of Youth and Adolescence (2021) 50:1766–1781 1771 Table 2 Correlations between variables Variable 1234 567 8910 11 12 13 14 15 16 17 18 19 20 21 22 23 24 1 T4 Withdrawal ** 2 T5 Withdrawal 0.56 ** ** 3 T6 Withdrawal 0.49 0.62 ** ** ** 4 T7 Withdrawal 0.44 0.54 0.60 ** * 5 Timing (age) 0.08 0.05 0.07 0.03 ** ** 6 Number of partners −0.03 0.04 0.03 0.09 −0.07 ** ** ** ** ** ** 7 Max length relation. −0.08 −0.09 −0.13 −0.13 −0.36 −0.15 ** ** 8 T4 S Commitment −0.04 0.02 −0.001 −0.04 −0.16 −0.06 0.19 * * ** ** ** 9 T5 S Commitment −0.06 −0.09 −0.08 −0.4 −0.11 −0.05 0.25 0.24 * * ** ** ** ** ** ** 10 T6 S Commitment −0.07 −0.08 −0.14 −0.12 −0.12 −0.04 0.16 0.22 0.30 ** ** ** ** 11 T5 P Commitment −0.04 −0.08 −0.05 −0.02 −0.05 −0.02 0.22 0.21 0.26 0.13 * ** 12 T6 P Commitment −0.04 −0.05 −0.02 −0.07 −0.02 −0.04 0.08 0.06 0.09 0.12 0.41 ** * * * * ** ** * * 13 T4 S Satisfaction −0.12 −0.04 −0.10 −0.10 −0.09 −0.09 0.14 0.60 0.10 0.12 0.07 0.03 ** ** ** * ** ** ** ** ** ** 14 T5 S Satisfaction −0.10 −0.16 −0.13 −0.10 −0.05 −0.03 0.16 0.14 0.71 0.24 0.19 0.01 0.18 ** ** ** ** * ** ** ** ** ** ** ** 15 T6 S Satisfaction −0.09 −0.15 −0.20 −0.20 −0.07 −0.01 0.07 0.14 0.21 0.68 0.14 0.15 0.15 0.30 ** ** ** * ** * ** ** ** ** 16 T5 P Satisfaction −0.08 −0.12 −0.15 −0.08 0.03 −0.01 0.12 0.13 0.17 0.13 0.64 0.33 0.06 0.25 0.24 * ** ** ** ** ** 17 T6 P Satisfaction 0.003 −0.02 −0.09 −0.12 0.05 0.03 −0.03 −0.02 0.03 0.18 0.36 0.62 0.01 0.01 0.31 0.45 ** ** ** ** ** * ** * ** ** ** ** ** 18 T5 S Support −0.12 −0.14 −0.14 −0.14 −0.01 −0.04 0.17 0.11 0.34 0.10 0.12 0.10 0.18 0.37 0.14 0.12 0.08 * ** ** ** * ** ** ** ** ** * ** ** 19 T6 S Support −0.08 −0.12 −0.16 −0.19 −0.05 −0.02 0.05 0.10 0.18 0.49 0.04 0.06 0.14 0.23 0.58 0.12 0.16 0.30 ** ** ** ** ** ** ** * ** ** ** 20 T5 P Support −0.08 −0.12 −0.07 −0.06 0.04 −0.01 0.06 0.16 0.17 0.05 0.37 0.18 0.18 0.25 0.10 0.46 0.22 0.20 0.08 * * * ** ** ** ** ** ** ** * ** ** 21 T6 P Support −0.12 −0.11 −0.11 −0.14 0.05 −0.03 −0.02 0.12 0.10 0.14 0.24 0.31 0.02 0.04 0.21 0.32 0.38 0.12 0.16 0.42 * * ** ** ** ** ** ** ** ** ** ** ** ** 22 T5 S Conflict 0.09 −0.13 0.15 0.22 0.09 0.06 −0.24 −0.10 −0.25 −0.30 −0.18 −0.21 −0.07 −0.36 −0.25 −0.24 −0.24 −0.31 −0.27 −0.06 −0.02 ** ** ** * ** * * ** ** * ** ** ** ** ** 23 T6 S Conflict 0.02 0.04 0.19 0.18 −0.02 0.10 −0.04 −0.08 −0.10 −0.25 −0.07 −0.10 −0.12 −0.13 −0.36 −0.10 −0.28 −0.11 −0.30 −0.09 −0.15 0.49 * * * ** ** * ** ** ** ** * ** ** 24 T5 P Conflict 0.01 0.11 0.04 0.05 0.04 0.13 −0.14 −0.07 −0.02 −0.15 −0.30 −0.41 −0.11 −0.15 −0.20 −0.44 −0.31 −0.07 −0.20 −0.15 −0.12 0.55 0.21 * ** * ** ** ** ** ** ** ** ** 25 T6 P Conflict −0.004 −0.03 0.06 0.12 −0.10* 0.04 0.03 −0.07 0.03 −0.07 −0.10 −0.24 −0.14 −0.01 −0.25 −0.23 −0.39 −0.03 −0.18 −0.10 −0.22 0.34 0.57 0.42 Relation. = romantic relationship; S = self-reported; P = partner-reported * ** *** p < 0.05, p < 0.01, p < 0.001 1772 Journal of Youth and Adolescence (2021) 50:1766–1781 forever. Items were rated on a 7-point scale, from 1 = standard errors (MLR) estimation, unless otherwise stated. completely disagree to 7 = completely agree (T4-T6 α The analyzes were pre-registered on the Open Science Fra- self- = 0.84, 0.91, 0.90; T5-T6 α = 0.87, 0.87). mework (www.osf.io/j6nrd), and an explanation of all report partner-report deviations can be found in the Supplementary Materials. First, Satisfaction to test if social withdrawal predicted lifetime non-involve- ment, the age at one’s first relationship, maximum relation- Romantic relationship satisfaction was measured by the ship duration, and the number of partners, an unconditional mean of two items of the IMS Satisfaction subscale. Parti- withdrawal latent growth curve model (LGCM), with corre- cipants completed the questionnaire at T4 to T6, and their lated intercept and linear slope from T4 to T6 was specified. partners at T5 and T6. The two items were: I am satisfied Sex and ethnicity were included as covariates, with paths to with my relationship and My relationship gives me what I the withdrawal intercept and slope and to the outcomes. Paths need in terms of intimacy, friendship, etc. Although the from the withdrawal intercept and slope to the outcomes were subscale also included the item My relationship is much specified. The path from the intercept to the outcomes shows better than the relationships of others, alpha was higher how baseline withdrawal levels predicted the outcomes, and when excluding this item due to a lower overall mean rating the path from the slope to the outcomes shows how the of this item compared to the other two, and weak correla- change in withdrawal between 19 and 29 years was associated tions between this and the other two items. Items were rated with the outcomes. Additionally, a cross-lagged panel model on a 7-point scale from 1 = completely disagree to 7 = (CLPM) with the concurrent and longitudinal effects between completely agree (T4-T6 α = 0.73, 0.78, 0.72; T5- social withdrawal and current involvement at T4 to T7 tested self-report T6 α = 0.81, 0.73). if higher withdrawal consistently predicted being single across partner-report late adolescence and early adulthood. The CLPM was spe- Perceived support cified with stability paths of the same variable over time, the within-wave associations between withdrawal and current At T5 and T6, participants and their partners responded to involvement, the cross-lagged paths from one variable at T to items assessing to what extent they felt supported by their the other variable at T , and a path from sex and ethnicity to x+1 romantic partner in specific domains. Items included Deci- T4 withdrawal and current involvement status. Good model sions about work or education; Problems with your health; fitwas defined as a comparative fit index (CFI) > 0.90; stan- Spending your free time and your social contacts; Practical dardized root mean residual (SRMR) < 0.06, and root mean things; and Personal matters that concern you. Items were square error of approximation (RMSEA) < 0.06. rated on a 4-point scale, from 1 = no support, to 4 = a lot of Second, to test if entering a romantic relationship for the support from partner (T5-T6 α = 0.72, 0.83, and first time reduced adolescents’ withdrawal across 19 to 29 self-report α = 0.82, 0.79). The scale represents the mean of years, a dummy-coded a time-varying covariate (TVC), partner-report the items ratings, with higher scores indicating more per- “entering first relationship,” with 0 = has not yet entered a ceived support from one’s romantic partner. relationship and 1 = has entered first romantic relationship (and TVC = 1 for all subsequent waves after entry) was Perceived conflict used. Again, the withdrawal LGCM with sex and ethnicity as covariates was specified, now subsequently allowing the Perceived relationship conflict was assessed at waves T5 and dummy-coded TVC to have an effect on withdrawal per T6. Participants and their partners responded to five items time point. The WLSMV estimator was used due to the about conflict situations occurring in the last 12 months, dichotomous nature of the TVC. A sensitivity analysis including: Fierce discussions between you and your partner; tested if the TVC effect was present within persons within a One person blamed the other strongly; You didn’t talk to each multilevel modeling framework. The difference between the other for a while; Fights got out of hand;and You no longer original, SEM-based model and the multilevel model is that lived together (if applicable). Responses were given on a 3- the multilevel model is a two-level (vs. single-level), uni- point scale, where 1= no, 2= one time, and 3= multiple variate (vs. multivariate) model in which TVCs have ran- times (T5-T6 α = 0.70, 0.72, and α = 0.73, dom effect (vs. fixed effect) coefficients that vary over self-report partner-report 0.70). The scale score represents the mean item ratings, with individuals (vs. over time). This means that in the sensi- higher scores indicating more perceived relationship conflict. tivity model, the effect of entering a romantic relationship for the first time on withdrawal is assumed to be constant Statistical Analyzes across time, and the TVC effect on withdrawal represents the average shift in withdrawal when entering the first Analyzes were conducted in MPlus Version 80.4 (Muthén & romantic relationship. Bayesian estimation with unin- Muthén, 1998–2017) using maximum likelihood with robust formative priors, 100 thousand iterations, and 2 Monte Journal of Youth and Adolescence (2021) 50:1766–1781 1773 Table 3 Goodness-of-fit Model n χ df CFI SRMR RMSEA [95% CI] statistics of all structural equation models Unconditional LGCM 1702 50.6 5 0.96 0.04 0.07 [0.06, 0.09] TVC LGCM 1597 64.5 21 0.98 0.04 0.04 [0.03, 0.05] Involvement CLPM 1710 217.3 24 0.91 0.05 0.07 [0.06, 0.08] Commitment APIM 402 30.2 12 0.94 0.02 0.09 [0.05, 0.13] Satisfaction APIM 402 7.8 12 1 0.01 0.00 [0.00, 0.05] Conflict APIM 402 21.0 12 0.98 0.03 0.06 [0.00, 0.10] Support APIM 402 14.3 12 0.99 0.02 0.03 [0.00, 0.08] n = number of participants included in the analyzes; CFI = comparative fit index; SRMR = standardized root mean residual; RMSEA = root mean square error of approximation. LGCMs = latent growth curve models; TVC = time-varying covariate model (i.e. entering first romantic relationship); CLPM = cross-lagged panel model (current romantic relationship involvement); APIM = actor partner interdependence model. Good model fit for all models was determined by the following criteria: CFI > 0.90; SRMR < 0.06; RMSEA < 0.06 Carlo chains was used; the model stabilized at PSR = to T6 participant satisfaction. T6 participant and partner 10.001. satisfaction were regressed on the two T5 interaction terms. Third, to investigate the longitudinal, bi-directional asso- T6 withdrawal was regressed on T5 withdrawal, participant ciations between withdrawal and romantic relationship qua- and partner satisfaction, and the two interaction terms. All lities, as reported by both the participants and their partners, a predictors were correlated, thereby estimating actor effects series of longitudinal actor-partner interdependence models while controlling for partner effects, and vice versa. The (APIMs; Cook & Kenny, 2005) were performed, with data residual variances of the outcome variables were also cor- from 402 participants who had the same partner across T5 related to control for additional sources of non-independence. and T6. Six same-sex romantic dyads were excluded because Paths from participants’ ethnicity and length of the romantic of the sex grouping of the analyzes, explained below. relationship (at T5) to all T5 predictors were added as con- Separate APIMS were tested for relationship commitment, trols. To correct for the false discovery rate (FDR) in multiple satisfaction, perceived support, and perceived conflict; satis- testing, the Benjamini and Hochberg (1995)method was faction is used to illustrate the modeling procedure. The applied. This method is more powerful and less conservative APIM is designed to account for non-independence of than the Bonferroni procedure. All observed p-values from observations within interpersonal relationships, with the the APIMs per sex (n= 84 for both females and males) were romantic dyad as the unit of analysis instead of the indivi- ranked, and alpha was specified as 0.05. The Benjamini- dual. APIMs produce estimates of actor effects (how much a Hochberg adjusted p-value criterion was p < 0.009 for person’s satisfaction is predicted by their own prior satis- females and p < 0.014 for males. For simplicity, the criterion faction), and partner effects (how much a person’s satisfac- of p < 0.01 was used for all APIM effects. tion is predicted by their partner’s prior satisfaction). The traditional longitudinal APIM was extended in the current study by including participants’ T5 and T6 withdrawal to the Results model and using T5 withdrawal as both an actor and a moderating variable. To test if withdrawal moderated long- Social Withdrawal Predicts Romantic Relationship itudinal actor and partner effects, all T5 predictors were Involvement and Quantity centered, and two interaction terms were created, namely participant’s withdrawal by participant’s satisfaction, and Table 3 depicts the fit of all models. The LGCM (Fig. 1) participant’swithdrawalby partner’s satisfaction. To inves- indicated that having a higher withdrawal intercept sig- tigate differences between females and males, sex was spe- nificantly predicted a higher likelihood of never having cified as a grouping variable, thereby computing separate been involved in a romantic relationship by early adulthood estimates for females and males within the same APIM fra- (β = −0.32, p < 0.001), becoming romantically involved at mework. A grouping approach (instead of distinguishing an older age (β = 0.14, p < 0.001), and having a shorter actors and partners by sex) was necessary because with- maximum romantic relationship duration (β = −0.13, p < drawal data for only one member per dyad was available. 0.001). The withdrawal slope significantly predicted rela- Actor effects were estimated by specifying paths from T5 to tionship duration (β = −0.10, p = 0.029); a steeper increase T6 participant’s satisfaction and T5 to T6 partner’s satisfac- in withdrawal was associated with a shorter maximum tion. Partner effects were estimated by specifying paths from relationship duration. Neither the withdrawal intercept nor T5 participant to T6 partner satisfaction; and from T5 partner the slope predicted the number of romantic partners. The 1774 Journal of Youth and Adolescence (2021) 50:1766–1781 model also showed that participants of non-Dutch origin positive, and those of involvement were moderate and were on average more withdrawn (β = 0.30, p = 0.005), and positive. Within-wave correlations between withdrawal and had a shorter maximum relationship duration (β = −0.24, p involvement were small and negative, indicating a = 0.003) than Dutch-origin ones. Males were older at their decreased likelihood of current romantic involvement for first romantic relationship (β = 0.16, p = 0.010) and had a more withdrawn youth. Cross-lagged paths were significant shorter maximum relationship duration (β = −0.25, p < from withdrawal to current involvement status, but not vice 0.001) than females. versa. This means that across all ages, more withdrawn Figure 2 depicts the significant paths in the withdrawal youth were less likely to have a romantic relationship and current involvement CLPM. The stability correlations approximately three years later, but relationship status, in of withdrawal at consecutive time points were strong and turn, did not predict future withdrawal levels. The Effect of Entering a Relationship for the First Time on Withdrawal The effects of the TVC “entering first romantic relation- ship” on withdrawal were significant and negative at all waves: T4 (β = −0.26, p < 0.001); T5 (β = −0.27, p < 0.001); T6 (β = −0.15, p = 0.043); T7 (β = −0.28, p = 0.008). This means that across all ages, participants who entered a romantic relationship for the first time were less withdrawn than those who had not yet become involved. These results replicated in the within-person effects of the multilevel model. The overall within-person effect of entering a romantic relationship for the first time on with- drawal was moderate and negative (β = −0.33, 95% CI [−0.44, −0.21]), indicating a moderate decrease in with- drawal when an adolescent became involved for the first time. Longitudinal Within-Dyad Associations Between Withdrawal and Self- and Partner-Reported Fig. 1 Significant paths from the Latent Growth Curve Model (LGCM), with the social withdrawal intercept and slope across 19 to Relationship Qualities 29 years predicting lifetime involvement, age at first romantic rela- tionship, and maximum romantic relationship duration, controlling for Table 4 depicts all APIM effects. For females but not males, ethnicity (0 = Dutch-origin, 1 = non-Dutch-origin) and sex (0 = being in a romantic relationship for a longer time was female, 1 = male). No significant predictors of number of romantic partners emerged associated with lower withdrawal and higher self- and Fig. 2 Significant effects in the Cross-Lagged Panel Model (CLPM) of social withdrawal and current romantic relationship involvement (0 = not currently in a relationship, 1 = currently in a relationship) across 19 to 29 years, controlling for ethnicity (0 = Dutch-origin, 1 = non-Dutch-origin) and sex (0 = female, 1 = male). 95% CI depicted in brackets Journal of Youth and Adolescence (2021) 50:1766–1781 1775 Table 4 Standardized estimates of the Actor-Partner Interdependence Models (APIMs) of participant social withdrawal and participant- and partner-reported commitment, satisfaction, support, and conflict, by participant’s sex Females Males Commitment Satisfaction Support Conflict Commitment Satisfaction Support Conflict Within-wave associations S ↔ P 0.19 0.11 0.24 0.43 0.14 0.20 0.10 0.46 5 5 SW ↔ S −0.06 −0.11 −0.08 0.06 −0.15 −0.24 −0.05 −0.16 5 5 SW ↔ P −0.01 −0.07 −0.10 0.10 −0.05 −0.08 −0.21 0.24 5 5 S ↔ P 0.08 0.32 0.13 0.15 0.30 0.13 0.09 0.51 6 6 SW ↔ S −0.13 −0.10 −0.07 0.17 −0.21 −0.28 −0.25 0.34 6 6 SW ↔ P 0.02 −0.10 −0.01 0.10 −0.17 −0.12 0.12 0.16 6 6 Actor effects S → S 0.34 0.24 0.37 0.52 0.40 0.40 0.19 0.59 5 6 P → P 0.48 0.45 0.51 0.52 0.48 0.57 0.35 0.18 5 6 SW → SW 0.65 0.65 0.65 0.61 0.72 0.72 0.72 1.02 5 6 Partner effects S → P 0.06 0.02 0.02 0.25 −0.06 0.01 0.15 0.32 5 6 SW → P −0.08 −0.02 −0.16 0.05 0.01 0.13 −0.13 0.07 5 6 + + P → S −0.01 0.17 −0.07 0.21 0.08 0.21 −0.05 −0.20 5 6 SW → S 0.05 −0.01 0.02 0.04 −0.26 −0.34 −0.33 0.31 5 6 S → SW 0.01 −0.06 −0.07 0.19 0.04 0.07 −0.10 0.33 5 6 P → SW 0.04 0.01 0.02 −0.12 −0.10 −0.29 −0.03 −0.23 5 6 Interaction effects SW *S → S 0.12 −0.01 −0.11 −0.01 0.26 0.11 0.05 0.19 5 5 6 SW *P → S −0.01 0.15 0.22 0.15 0.03 0.09 −0.27 −0.16 5 5 6 SW *P → P −0.004 0.04 −0.08 −0.02 −0.17 −0.18 −0.17 −0.24 5 5 6 + + SW *S → P 0.03 −0.07 0.11 −0.12 0.11 0.16 −0.13 0.43 5 5 6 SW *S → SW −0.01 −0.06 −0.05 −0.04 −0.02 −0.02 −0.05 0.50 5 5 6 SW *P → SW 0.04 0.09 0.02 0.13 0.03 −0.01 0.08 −0.67 5 5 6 S = self-reported relationship quality at T5; P = partner-reported relationship quality at T5; SW = self-reported social withdrawal at T5; S = 5 5 5 6 self-reported relationship quality at T6; P = partner-reported relationship quality at T6; SW = self-reported social withdrawal at T6; Double- 6 6 headed arrows indicate a correlational path and single-headed arrows a regression path Bold coefficients are significant at p < 0.01; 0.01 < p < 0.05 partner-reported commitment. For males, ethnicity was providing confidence of robust (rather than chance) find- associated with withdrawal, with non-Dutch-origin males ings. Higher initial (T5) withdrawal predicted lower future reporting higher withdrawal than Dutch-origin males. (T6) self-reported commitment, satisfaction, and support for males, indicating that males who were initially more with- Actor and partner effects drawn reported decreased commitment to, satisfaction with, perceived support from their female partners. For both For females and males, the longitudinal actor effects were females and males, self-reported conflict predicted future moderate to strong for withdrawal and all self- and partner- withdrawal, and for females, self-reported conflict predicted reported relationship characteristics, with the exception of partner-reported conflict. Partner-reported satisfaction also males’ partner-reported conflict and self-reported perceived predicted future withdrawal for males. support, which were not significant. This suggests that young adults’ withdrawal and relationship perceptions, and their Interaction effects partners’ relationship perceptions are moderately stable, con- trolling for ethnicity, relationship duration, and partner effects. Simple slope estimates can be found in Table S1, and Seven significant out of 48 possible partner effects simple slope plots of the significant interaction effects in emerged across the APIMs (see Table 4); they showed a Figures S1-S5. Social withdrawal moderated the effect of consistent pattern of the direction and magnitude of effects, T5 to T6 self-reported commitment in males: withdrawn 1776 Journal of Youth and Adolescence (2021) 50:1766–1781 males’ commitment to their female partners was more stable years later across all ages. These results corroborate findings than non-withdrawn males’ commitment. Withdrawal also from previous cross-sectional studies (Roswell & Coplan, moderated several partner effects in males. First, it 2013), and additionally indicate that withdrawn individuals’ enhanced the association between males’ T5 partner- and romantic involvement delays are long-lasting, and that dif- T6 self-reported support: the less perceived support reported ferences between withdrawn and non-withdrawn youth by the female partners of withdrawn males (> +0.50 SD persist into adulthood. The delay of romantic involvement is average withdrawal), the less support these partners subse- likely attributable to withdrawn youth’s deviation from the quently provided to the withdrawn males. For non- formative cascade of development from same-sex friend- withdrawn males, the amount of support received by their ships to mixed-sex peer groups to romantic partnerships partners did not predict future support from their partners. (Nelson et al., 2008). Having had fewer opportunities to Second, withdrawal moderated the association from T5 self- learn how to approach and interact with opposite-sex peers to T6 partner-reported conflict in males. In withdrawn (> in adolescence, withdrawn youth might be particularly +0.50 SD) males, perceiving high levels of relationship anxious in novel romantic situations, or avoid them alto- conflict predicted more future partner-perceived conflict, gether (Barry et al., 2013; Gazelle & Druhen, 2009). Con- whereas non-withdrawn (< 0.50 SD) males’ perceived trary to expectations that withdrawn young adults would conflict did not. Third, withdrawal moderated how partner- have longer-lasting relationships with fewer partners (a reported conflict predicted males’ future withdrawal: having hypothesis that was based on limited research; Leck, 2006), a female partner that perceived high conflict predicted more higher and increasing withdrawal across the early adulthood withdrawal for non-withdrawn males (< −1 SD), and less decade predicted a shorter maximum romantic relationship withdrawal for highly withdrawn males (> +1.4 SD). There duration, and had no effect on the number of romantic was no effect for males that were average on withdrawal. partners. The effect of withdrawal on relationship duration may be due to the fact that withdrawn adolescents were older when they first became romantically involved, leading Discussion to comparatively less possible time for their relationship to have lasted by the point of assessment than non-withdrawn Involvement in romantic relationships is a central devel- young adults’ relationships. It is also probable that the opmental task of late adolescence and early adulthood romantic relationships of withdrawn young adults are more (Roisman et al., 2004). Although previous work has sug- likely to break-up than those of non-withdrawn young gested that withdrawn youth’s formative social experiences adults, leading to shorter-lasting relationships. Because of lead to delays in romantic development, empirical studies of the additional socio-emotional costs of romantic relation- the extent to which withdrawal contributes to deviations ship dissolutions, an empirical test of this latter possibility is from normative romantic development, and in which spe- warranted. cific features, were lacking. To address this gap, this study It is important to note that, while withdrawn and non- tested to what extent withdrawal predicts delays in romantic withdrawn young adults have diverging patterns of romantic involvement and quantity; if entering a romantic relation- involvement, it is not suggested that withdrawn youth should ship for the first time decreases withdrawal; and if with- be involved, especially if they do not want to be. Postponing drawal predicts self- and partner-rated romantic relationship romantic involvement to older ages than their more sociable qualities, and vice versa. Higher withdrawal across the counterparts might even be adaptive for withdrawn indivi- decade of late adolescence and early adulthood predicted duals. Having more time before focusing on “settling down” delays in all aspects of romantic involvement and a shorter with a romantic partner could provide the opportunity to longest-lasting romantic relationship, but did not influence develop in other domains in which withdrawn young adults the number of romantic partners. When an adolescent are also delayed, such as identity development (Barry et al., became romantically involved for the first time, withdrawal 2013), selecting educational and career paths (Hamer & moderately decreased. Despite this decrease, withdrawal Bruch, 1997), reaching higher levels of education and income remained an important factor in couples’ relationship (Nelson et al., 2020; Schmidt et al., 2017), maximizing quality ratings, especially affecting males’ relationship person-environment fit (Shulman & Connolly, 2013), and perceptions and dynamics. These results provide insights developing interpersonal skills in other social relationships into the developmental sequelae of withdrawn adolescents’ (e.g. friends, classmates, colleagues). Catching up to their and young adults’ romantic relationship development. non-withdrawn peers in these domains could then contribute As predicted, higher withdrawal was associated with a to better maintenance of and positive functioning in with- greater likelihood of never having been romantically drawn young adults’ romantic relationships when they do involved, entering a romantic relationship for the first time emerge, and “increase the chances for better provision for the when older, and a greater likelihood of being single three next generation” (Shulman & Connolly, 2013,p.34). Journal of Youth and Adolescence (2021) 50:1766–1781 1777 Withdrawn young adults who do desire a romantic relation- providing and receiving less support, and perceiving more ship but feel unable to initiate contact with potential dating relationship conflict. Females’ withdrawal, in contrast, had partners may nevertheless feel lonely, have low self-esteem, no effect on self- or partner-reported romantic relationship and subsequently withdraw further; hence attaining this goal qualities. These gender disparities may be due to the greater and entering a romantic relationship for the first time can have social acceptance of withdrawal in females, and females marked benefits. being better-prepared for romantic relationships than males. Indeed, young adults’ withdrawal decreased when they The child and adolescent literature suggests that withdrawn entered a romantic relationship for the first time across behaviors are less socially accepted in boys than in girls participants. Based on Developmental Task Theory, with- because inhibited behaviors are viewed as violations of drawn adolescents and young adults experience negative gender-normative expectations of male assertiveness and social sanctions and receive less social support from their dominance (Doey et al., 2014). This seems to apply to peers when they are delayed in becoming romantically romantic partnerships in early adulthood as well: males involved (Furman & Collibee, 2014). When withdrawn appear to be more accepting of their withdrawn female individuals enter their first romantic relationships, they are partners than are females of their withdrawn male partners. no longer off-time in romantic involvement compared to Additionally, females may enter romantic relationships with their peers, and gain the new social identity of “girlfriend” more relationship maintenance skills than males. In child- or “boyfriend”. This shift in identity could lead withdrawn hood and adolescence, girls tend to prefer involvement in young adults to obtain more social status, belonging, and more intimate, dyadic relationships (Hall, 2011; Rose & social support (Raley et al., 2007). Withdrawn young adults Rudolph, 2006), which more closely resemble romantic may be especially sensitive to this shift in identity because relationships than boys’ friendships. Girls’ friendships they are less socially integrated than non-withdrawn young provide more opportunities than boys’ friendships to adults. Especially for them, entering a relationship for the become comfortable self-disclosing, develop intimacy with first time might lead to greater social integration, the others, and learn conflict resolution methods (Giordano development of interpersonal skills, decreased loneliness, et al., 2006), skills which are conducive to maintaining increased self-esteem, and expansion of social networks via intimate relationships. Highly withdrawn males may come the romantic partners, subsequently decreasing withdrawal. to rely on their female partners to take primary responsi- Despite the decrease in withdrawal when entering a bility of the social and emotional aspects of their romantic romantic relationship for the first time, withdrawal remained relationship. Tentative support for this idea is that the a predictor and outcome of several unfavorable romantic results indicated that female partner-perceived conflict relationship qualities. Reiterating the main within-couple predicted less future withdrawal in highly withdrawn males. findings, high withdrawal (1) predicted lower self-reported A withdrawn male may struggle to communicate his feel- commitment, satisfaction, and support in males; (2) was ings and needs to his partner when he feels unhappy in his predicted by higher self-reported conflict in males and relationship (Giordano et al., 2006)—subsequently con- females, higher partner-reported conflict in females, and tributing to other relationship problems and more with- lower partner-reported satisfaction in males; and (3) in drawal—but when his female partner feels unhappy, she males, was associated with interaction patterns in which may be better able to self-disclose, take initiative to resolve partners who perceived less support subsequently provided conflicts, and rebuild intimacy in the relationship (Giordano less support; perceived relationship conflicts predicted et al., 2006; Raley et al., 2007); skills learned in her partner-perceived relationship conflicts; and partner- friendships might contribute to relationship improvements perceived conflicts predicted low males’ future with- and less withdrawal in her male partner. drawal. Taken together, these results indicate that males’ Several limitations should be considered when inter- withdrawal plays a bigger role in the romantic relationship preting these results. First, a broad conceptualization of quality dynamics than females’ withdrawal. These results social withdrawal was used, which did not assess under- are consistent with the theory that the suboptimal romantic lying motivations to withdrawal such as fear of negative relationship qualities of withdrawn individuals are due to evaluation, social disinterest, or peer rejection. The various difficulties with self-disclosing, being responsive, and motivations for withdrawing may be associated with dif- forming intimate bonds with romantic partners (Luster ferent patterns of romantic development. For example, et al., 2013), but suggest that these mechanisms apply pri- unsociable-withdrawn youth have been found to have fewer marily to withdrawn males. Withdrawn males in particular difficulties initiating and maintaining friendships than may struggle to communicate, self-disclose, and form inti- anxious-withdrawn youth (Ladd et al., 2011), and may macy with their partners, subsequently leading to difficul- likewise have fewer difficulties initiating and maintaining ties committing to their partner, feeling less satisfied in the romantic relationships. Yet, withdrawal subtypes overlap relationship and fostering less satisfaction in their partner, and withdrawn youth of any subtype tend to have worse 1778 Journal of Youth and Adolescence (2021) 50:1766–1781 social relationships than non-withdrawn ones (Eggum- same-sex couples, thus limiting the generalizability of the Wilkens et al., 2020; Nelson, 2013). Regardless, future findings to heterosexual young adults. Also limiting the studies could investigate differences in how young adults generalizability was that the sample included young adults from different withdrawal subtypes initiate and maintain from a predominantly Dutch background. Because there is their romantic relationships. cultural variation in romantic relationship development, Second, only romantic relationships that did not end ethnic differences in the associations between withdrawal before the age of 17 years were investigated, and thus the and romantic relationship initiation and maintenance are analyzes did not account for involvement in earlier rela- likely and require more attention in future work. tionships in adolescence. The timing variable therefore Finally, several gaps that could not be addressed remain to reflects the age at which participants entered a more serious, be investigated in future studies. First, a replication of the attachment-based relationship for the first time. This pos- finding that males’ withdrawal particularly affected their sibly neglects formative romantic experiences that set the romantic relationship characteristics is needed. Previous stage for more “adult-like” relationship functioning. In the reports of sex differences in the associations between with- context of withdrawal, this might not have confounded drawal and romantic relationship characteristics have been results heavily because withdrawn young adults were more somewhat mixed. Although there is a theoretical basis to likely to have never been romantically involved and entered expect that males – especially withdrawn ones – face more romantic relationship when older, meaning that these later challenges in their romantic relationships than females, romantic relationships were probably their first ones. directly testing effects of possible mediators (e.g. intimacy, Nevertheless, it would be interesting to explore the effects self-disclosure) would have provided more robust evidence of of withdrawal on romantic relationships from early ado- the proposed mechanisms underlying this association. Sec- lescence, when romantic interests are just emerging and ond, future studies investigating romantic relationships could involvement is more sporadic, to adulthood. include individuals’ desire for having a romantic partner, Third, partner reports were available across only two especially when assessing young adults, who have postponed waves, which may limit the generalizability of the dyadic romantic relationship-related decisions to increasingly older results to young adults who maintained the same partner ages (Arnett, 2000;OECD 2017). It is likely that young across three or more years. Although the duration of these adults, withdrawn and non-withdrawn, who do not particu- relationships was accounted for, young adults tend to larly desire a romantic relationship fare better than those who engage in multiple relationships of various durations across do desire one but are not involved. Relatedly, investigating if early adulthood, which could not be captured. Future stu- there are benefits for withdrawn youth to delay romantic dies could test if results replicate across longer- and shorter- involvement for identity, educational, and career development lasting romantic relationships throughout early adulthood. warrants more attention. Finally, to obtain a more compre- Fourth, it was not possible to investigate to what extent hensive picture of young adults’ romantic relationship both romantic partners’ withdrawal levels affected their development, interpersonal and dyadic behaviors across relationship perceptions and each other’s withdrawal, multiple romantic partners could be investigated. The large because no data about partners’ withdrawal was available. majority of the participants in this study changed partners This is a limitation because both partners’ withdrawal across the late adolescence and early adulthood decade, but influences the quality of the romantic relationship, and there within-dyad dynamics could only be investigated within only may be differential effects between couples with one one relationship. It would be interesting to see how social withdrawn partner and those with both. Partners’ with- withdrawal affects romantic partner selection, if withdrawal drawal levels may interact to predict relationship outcomes, affects relationship functioning in the same way across dif- for better or for worse. On the one hand, having similarly ferent partners, and how withdrawn individuals cope with high withdrawal levels may contribute to greater mutual romantic relationship dissolutions. understanding and acceptance of one another’s inhibited behaviors within couples. On the other hand, withdrawal in both partners may lead to a lower quality relationship Conclusion because both partners may be non-communicative and hinder intimacy development. The lack of partners’ with- The current study provided insights into the links between drawal data also meant that the sample needed to be split by young adults’ withdrawal and romantic relationship devel- sex in order to investigate sex differences in within-dyad opment and a general theoretical framework which can be associations. This would not have been necessary if data on applied in future investigations into the complex social and partners’ withdrawal were available, and not splitting the romantic worlds of withdrawn young adults. This study sample would have increased power to detect smaller investigated the longitudinal effects of social withdrawal on effects. Grouping by sex also required the exclusion of deviations from normative romantic development in late Journal of Youth and Adolescence (2021) 50:1766–1781 1779 adolescence and early adulthood, and the interplay between Data Sharing and Declaration The datasets analyzed during the cur- rent study are not publicly available but are available from the cor- withdrawal and couples’ relationship perceptions. The results responding author on reasonable request. indicated that withdrawal in late adolescence and early adulthood contributed to delays in romantic relationship Compliance with Ethical Standards involvement, and was associated with certain romantic rela- tionship quantity and quality features. Withdrawn young Conflict of Interest The authors declare no competing interests. adults became romantically involved when they were older, and were more likely to have never been involved by adult- Ethical Approval All procedures performed in studies involving hood, likely due to their heightened anxiety and avoidance of human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 novel social situations. There might be advantages to this Helsinki Declaration and its later amendments or comparable ethical delay, as withdrawn young adults have more time to catch up standards. The Dutch Central Committee on Research Involving to their more sociable peers in other domains. Withdrawn Human Subjects approved the Tracking Adolescents’ Individual Lives youth who nevertheless initiated a romantic relationship for (TRAILS) study. the first time benefited from it: they became less withdrawn. Informed Consent Participants and their romantic partners provided Entering a romantic relationship for the first time may written, informed consent at all waves. improve social integration, interpersonal skills, socio- emotional functioning, and social networks. Regardless of Preregistration The study design, hypotheses, and analyzes were pre- this initial decrease in withdrawal, withdrawal affected the registered on the Open Science Framework (10.17605/OSF.IO/3FJQD). qualities of males’ romantic relationships, possibly due to males’ withdrawal being less socially accepted, and with- Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. drawn males having particular difficulty communicating, self- disclosing, and building intimacy with their partners. Highly Open Access This article is licensed under a Creative Commons withdrawn males may come to rely on their female partners to Attribution 4.0 International License, which permits use, sharing, take primary responsibility of the social and emotional aspects adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the of their romantic relationship, perhaps because females tend source, provide a link to the Creative Commons license, and indicate if to be more socialized for maintaining romantic relationships. changes were made. The images or other third party material in this Many of these proposed mechanisms remain to be directly article are included in the article’s Creative Commons license, unless and empirically tested. Because early adulthood is character- indicated otherwise in a credit line to the material. If material is not ized by decisions about long-term commitments, including included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted those in romantic relationships such as partner selection, use, you will need to obtain permission directly from the copyright cohabitation and marriage, and family formation, continued holder. To view a copy of this license, visit http://creativecommons. investigations into the effects of withdrawal on the develop- org/licenses/by/4.0/. mental tasks of this period of life are warranted. References Acknowledgements This research is part of the TRacking Adoles- cents’ Individual Lives Survey (TRAILS). Participating centers of Achenbach, T. M., & Rescorla, L. A. (2003). Manual for the ASEBA TRAILS include various departments of the University Medical Center Adult Forms & Profiles. Burlington: University of Vermont. and University of Groningen, the University of Utrecht, the Radboud Arnett, J. J. (2000). Emerging adulthood: a theory of development Medical Center Nijmegen, and the Parnassia Bavo group, all in the from the late teens through the twenties. American Psychologist, Netherlands. We are grateful to everyone who participated in this 55, 468–480. https://doi.org/10.1037/0003-066X.55.5.469. research or worked on this project to and make it possible. Barry, C. M., Nelson, L. J., & Christofferson, J. L. (2013). Asocial and afraid: an examination of shyness and anxiety in emerging Authors’ Contributions S.A.B. conceived of the study, participated in adulthood. Journal of Family Studies, 19(1), 2–18. https://doi. its design, performed the statistical analyzes, participated in the org/10.5172/jfs.2012.1979. interpretation of data, and drafted the manuscript; J.S.R. and W.H.J.M. Barzeva, S. A., Meeus, W. H. J., & Oldehinkel, A. J. (2019a). Social participated in the design of the study, interpretation of the data, and withdrawal in adolescence and early adulthood: measurement provided feedback on all drafts of the manuscript; TO participated in issues, normative development, and distinct trajectories. Journal the design and coordination of the study, interpretation of the data, and of Abnormal Child Psychology, 47, 865–879. https://doi.org/10. provided feedback on all drafts of the manuscript. All authors read and 1007/s10802-018-0497-4. approved the final manuscript. Barzeva, S. A., Richards, J. S., Meeus, W. H. J., & Oldehinkel, A. J. (2019b). The social withdrawal and social anxiety feedback loop Funding TRAILS has been financially supported by various grants and the role of peer victimization and acceptance in the pathways. from the Netherlands Organization for Scientific Research (NWO), Development and Psychopathology, 32(4), 1402–1417. https:// ZonMW, GB-MaGW, the Dutch Ministry of Justice, the European doi.org/10.1017/S0954579419001354. Science Foundation, the European Research Council, BBMRI-NL, and Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery the participating universities. rate: A practical and powerful approach to multiple testing. 1780 Journal of Youth and Adolescence (2021) 50:1766–1781 Journal of the Royal Statistical Society: Series B, 57(1), 289-300. De Winter, A. F., Oldehinkel, A. J., Veenstra, R., Brunnekreef, J. A., https://doi.org/10.1111/j.2517-6161.1995.tb02031.x Verhulst, F. C., & Ormel, J. (2005). Evaluation of non-response Boisvert, S., & Poulin, F. (2016). Romantic relationship patterns from bias in mental health determinants and outcomes in a large adolescence to emerging adulthood: Associations with family and sample of pre-adolescents. European Journal of Epidemiology, peer experiences in early adolescence. Journal of Youth and Ado- 20, 173–181. https://doi.org/10.1007/s10654-004-4948-6. lescence, 45,945–958. https://doi.org/10.1007/s10964-016-0435-0. Doey, L., Coplan, R. J., & Kingsbury, M. (2014). Bashful boys and coy Booth-LaForce, C., & Oxford, M. L. (2008). Trajectories of social girls: A review of gender differences in childhood shyness. Sex withdrawal from grades 1 to 6: prediction from early parenting, Roles, 70, 255–266. https://doi.org/10.1007/s11199-013-0317-9. attachment, and temperament. Developmental Psychology, 44(5), Eggum-Wilkens, N. D., An, D., Zhang, L., & Costa, M. (2020). Co- 1298–1313. https://doi.org/10.1037/a0012954. occurrence of and cross-informant agreement on shyness, Campbell, L., Simpson, J. A., Boldry, J., & Kashy, D. A. (2005). Per- unsociability, and social avoidance during early adolescence. Social ceptions of conflict and support in romantic relationships: The role Development, 29,73–88. https://doi.org/10.1111/sode.12398. of attachment anxiety. Journal of Personality and Social Psychol- Furman, W., & Collibee, C. (2014). A matter of timing: developmental ogy, 88(3), 510-531. https://doi.org/10.1037/0022-3514.88.3.510 theories of romantic involvement and psychosocial adjustment. Carver, K., Joyner, K., & Udry, J. R. (2003). National estimates of Developmental Psychopathology, 26, 1149–1160. https://doi.org/ adolescent romantic relationships. In P. Florsheim (Ed.), Ado- 10.1017/S0954579414000182. lescent romantic relations and sexual behavior: theory, research, Gazelle, H., & Druhen, M. J. (2009). Anxious solitude and peer and practical implications (p. 23-56). Lawrence Erlbaum exclusion predict social helplessness, upset affect, and vagal Associates Publishers. regulation in response to behavioral rejection by a friend. Caspi, A., Elder, G. H., & Bem, D. J. (1988). Moving away from the Developmental Psychology, 45(4), 1077–1096. https://doi.org/10. world: life-course patterns of shy children. Developmental Psy- 1037/a0016165. chology, 24(6), 824–831. Giordano, P. C., Manning, W. D., & Longmore, M. A. (2006). Ado- Chen, X., & Tse, H. C. (2008). Social functioning and adjustment in lescent romantic relationships: An emerging portrait of their Canadian-born children with Chinese and European backgrounds. nature and developmental significance. In A. C. Crouter & A. Developmental Psychology, 44, 1184–1189. https://doi.org/10. Booth (Eds.), Romance and sex in adolescence and emerging 1037/0012-1649.44.4.1184. adulthood: Risks and opportunities (pp. 127–150). Mahwah, Collibee, C., & Furman, W. (2015). Quality counts: developmental New Jersey: Lawrence Erlbaum Associates. shifts in associations between romantic relationship qualities and Hall, J. A. (2011). Sex differences in friendship expectations: a meta- psychosocial adjustment. Child Development, 86(5), 1639–1652. analysis. Journal of Social & Personal Relationships, 28, https://doi.org/10.1111/cdev.12403. 723–747. https://doi.org/10.1177/0265407510386192. Collins, W. A. (2003). More than myth: The developmental sig- Hamer, R. J., & Bruch, M. A. (1997). Personality factors and inhibited nificance of romantic relationships during adolescence. Journal of career development: testing the unique contributions of shyness. Research on Adolescence, 13(1), 1–24. https://doi.org/10.1111/ Journal of Vocational Behavior, 50, 382–400. https://doi.org/10. 1532-7795.1301001. 1006/jvbe.1996.1542. Collins, W. A., & Van Dulmen, M. (2006). “The course of true love(s) Huisman, M., Oldehinkel, A. J., de Winter, A., Minderaa, R. B., de …”: Origins and pathways in the development of romantic rela- Bildt, A., Huizink, A. C., & Verhulst, F. C. (2008). Cohort tionships. In A. Booth & A. Crouer (Eds.), Romance and sex in profile: The Dutch ‘TRacking Adolescents’ Individual Lives’ adolescence and emerging adulthood: risks and opportunities Survey’; TRAILS. International Journal of Epidemiology, 37(6), (pp. 63–86). Mahwah, NJ: Erlbaum. 1227–1235. https://doi.org/10.1093/ije/dym273. Collins, W. A., Welsh, D. P., & Furman, W. (2009). Adolescent Kingery, J., Erdley, C., Marshall, K., Whitaker, K., & Reuter, T. romantic relationships. Annual Review of Psychology, 60,631–652. (2010). Peer experiences of anxious and socially withdrawn https://doi.org/10.1146/annurev.psych.60.110707.163459. youth: An integrative review of the developmental and clinical Connolly, J., Craig, W., Goldberg, A., & Pepler, D. (2004). Mixed- literature. Clinical Child & Family Psychology Review, 13(1), gender groups, dating, and romantic relationships in early ado- 91–128. https://doi.org/10.1007/s10567-009-0063-2. lescence. Journal of Research on Adolescence, 14(2), 185–207. Ladd, G. W., Kochenderfer-Ladd, B., Eggum, N. D., Kochel, K. P., & https://doi.org/10.1111/j.1532-7795.2004.01402003.x. McConnell, E. M. (2011). Characterizing and comparing the Connolly, J., Furman, W., & Konarski, R. (2000). The role of peers in friendships of anxious-solitary and unsociable preadolescents. the emergence of heterosexual romantic relationships in adoles- Child Development, 82, 1434–1453. https://doi.org/10.1111/j. cence. Child Development, 71(5), 1395–1408. https://doi.org/10. 1467-8624.2011.01632.x. 1111/1467-8624.00235. Lantagne, A., & Furman, W. (2017). Romantic relationship development: Connolly J., & McDonald, K. (2020) Cross-cultural perspectives on the interplay between age and relationship length. Developmental dating and marriage. Encyclopedia of Child and Adolescent Psychology, 53(9), 1738–1749. https://doi.org/10.1037/dev0000363. Development, Wiley-Blackwell Press. Leck, K. (2006). Correlates of minimal dating. The Journal of Social Cook, W. L., & Kenny, D. A. (2005). The actor-partner inter- Psychology, 146(5), 549–567. https://doi.org/10.3200/SOCP.146. dependence model: A model of bidirectional effects in develop- 5.549-567. mental studies. International Journal of Behavioral Development, Loewenthal, K. M. (2004). An introduction to psychological tests and 29(2), 101–109. https://doi.org/10.1080/01650250444000405. scales. 2nd ed. Hove: Psychology Press. Coplan, R. J., & Armer, M. (2007). A multitude of solitude: A closer Luster, S. S., Nelson, L. J., & Busby, D. M. (2013). Shyness and look at social withdrawal and nonsocial play in early childhood. communication: Impact on self and partner relationship satisfac- Child Development Perspectives, 1,26–32. https://doi.org/10. tion. Journal of Couple & Relationship Therapy, 12, 359–376. 1111/j.1750-8606.2007.00006.x. https://doi.org/10.1080/15332691.2013.836322. Coplan, R. J., Zheng, S., Weeks, M., & Chen, X. (2012). Young Meeus, W. H. J., Branje, S. J. T., van der Valk, I., & de Wied, M. children’s perceptions of social withdrawal in China and Canada. (2007). Relationships with intimate partner, best friend, and Early Child Development and Care, 182, 591–607. https://doi. parents in adolescence and early adulthood: A study of the sal- org/10.1080/03004430.2011.566328. iency of the intimate partnership. International Journal of Journal of Youth and Adolescence (2021) 50:1766–1781 1781 Behavioral Development, 31(6), 569–580. https://doi.org/10. of alternatives, and investment size. Personal Relationships, 5(4), 1177/0165025407080584. 357–391. https://doi.org/10.1111/j.1475-6811.1998.tb00177.x. Meeus, W., Van de Schoot, R., Klimstra, T., & Branje, S. (2011). Shulman, S. & Connolly, J. (2013). The challenge of romantic rela- Personality types in adolescence: Change and stability and links tionships in emerging adulthood: reconceptualization of the field. with adjustment and relationships: A five-wave longitudinal Emerging Adulthood, 1(1), 27–39. https://doi.org/10.1177/ study. Developmental Psychology, 47(4), 1181–1195. https://doi. 2167696812467330. org/10.1037/a0023816. Schmidt, L. A., Tang, A., Day, K. L., Lahat, A., Boyle, M. H., Saigal, Muthén, L. K., & Muthén, B. O. (1998). Mplus User’s Guide. 8th ed. S., & Van Lieshout, R. J. (2017). Personality development within Los Angeles, CA: Muthén & Muthén. 2017. a generational context: life course outcomes of shy children. Nelson, L. J. (2013). Going it alone: Comparing subtypes of with- Child Psychiatry & Human Development, 48, 623–641. https:// drawal on indices of adjustment and maladjustment in emerging doi.org/10.1007/s10578-016-0691-y. adulthood. Social Development, 22, 522–538. https://doi.org/10. Seiffge-Krenke, I. (2003). Testing theories of romantic development 1111/j.1467-9507.2012.00671.x. from adolescence to young adulthood: Evidence of a developmental Nelson, L. J., Jorgensen, N. A., & Clifford, B. N. (2020). Shy and still sequence. International Journal of Behavioral Development, 27(6), struggling: examining the relations between subtypes of social 519–531. https://doi.org/10.1080/01650250344000145. withdrawal and well-being in the 30s. Social Development, 00, Tackett, S. L., Nelson, L. J., & Busby, D. M. (2013). Shyness and 1–17. https://doi.org/10.1111/sode.12486. relationship satisfaction: evaluating the associations between Nelson, L. J., Padilla-Walker, L. M., Badger, S., Barry, C. M., Carroll, shyness, self-esteem, and relationship satisfaction in couples. The J. S., & Madsen, S. D. (2008). Associations between shyness and American Journal of Family Therapy, 41,34–45. https://doi.org/ internalizing behaviors, externalizing behaviors, and relationships 10.1080/01926187.2011.641864. during emerging adulthood. Journal of Youth and Adolescence, Tang, A., Van Lieshout, R. J., Lahat, A., Duku, E., Boyle, M. H., 37, 605–615. https://doi.org/10.1007/s10964-007-9203-5. Saigal, S., & Schmidt, L. A. (2017). Shyness trajectories across OECD. (2017). OECD family database. Retried from Family Database the first four decades predict mental health outcomes. Journal of http://www.oecd.org/els/family/database.htm. Abnormal Child Psychology, 45(8), 1621–1633. https://doi.org/ Oldehinkel, A. J., Rosmalen, J. G. M., Buitelaar, J. K., Hoek, H. W., 10.1007/s10802-017-0265-x. Ormel, J., Raven, D., Reijneveld, S. A., Veenstra, R., Verhulst, F. C., Vollebergh, W. A. M., & Hartman, C. A. (2015). Cohort profile Update: the Tracking Adolescents’ Individual Lives Sur- Stefania A. Barzeva is a PhD researcher at the University of vey (TRAILS). International Journal of Epidemiology, 44, Groningen, University Medical Center Groningen, Interdisciplinary 76–76n. https://doi.org/10.1093/ije/dyu225. Center Psychopathology and Emotion Regulation. Her major research Ozdemir, S. B., Cheah, C. S. L., & Coplan, R. J. (2015). Con- interests include developmental psychopathology, social withdrawal, ceptualization and assessment of multiple forms of social with- adolescence, psychosocial development, and social relationships. drawal in Turkey. Social Development, 24(1), 142–165. https:// doi.org/10.1111/sode.12088. Raley, R. K., Crissey, S., & Muller, C. (2007). Of sex and romance: late adolescent relationships and young adult union formation. Jennifer S. Richards is a Postdoctoral researcher at the University of Journal of Marriage and Family, 69(5), 1210–12226. https://doi. Groningen, University Medical Center Groningen, Interdisciplinary org/10.1111/j.1741-3737.2007.00442.x. Center Psychopathology and Emotion Regulation. Her major research Roisman, G. I., Masten, A. S., Coatsworth, J. D., & Tellegen, A. interests include resilience and mental health development, (2004). Salient and emerging developmental tasks in the transi- intergenerational transmission, family and peer experiences, parent- tion to adulthood. Child Development, 75(1), 123–133. https:// doi.org/10.1111/j.1467-8624.2004.00658.x. child interactions, differential susceptibility, and gene-environment Rose,A.J., &Rudolph,K.D.(2006). Areviewofsex differences in interactions. peer relationship processes: potential trade-offs for the emotional and behavioral development of girls and boys. Psychological Bul- letin, 132,98–131. https://doi.org/10.1037/0033-2909.132.1.98. Wim H. J. Meeus is Emeritus Professor of Adolescent Development at Roswell, H. C., & Coplan, R. J. (2013). Exploring links between Utrecht University and Developmental Psychology at Tilburg shyness, romantic relationship quality, and well-being. Canadian University. His major research interests include adolescence, social Journal of Behavioural Science, 45(4), 287–295. https://doi.org/ development, longitudinal research, personality and identity 10.1037/a0029853. development, social relationships, and problem behavior. Rubin, K. H., Coplan, R. J., & Bowker, J. C. (2009). Social with- drawal in childhood. Annual Review of Psychology, 60, 141–171. https://doi.org/10.1146/annurev.psych.60.110707.163642. Rubin, K. H., Wojslawowicz, J. C., Rose-Krasnor, L., Booth-LaForce, Albertine J. Oldehinkel is a Full Professor at the University of C., & Burgess, K. B. (2006). The best friendships of shy/with- Groningen, University Medical Center Groningen, Interdisciplinary drawn children: prevalence, stability, and relationship quality. Center Psychopathology and Emotion Regulation. Her major research Journal of Abnormal Child Psychology, 34(2), 143–157. https:// interests include developmental psychopathology, affective disorders, doi.org/10.1007/s10802-005-9017-4. adolescence, social development, longitudinal research, parent-child Rusbult, C. E., Martz, J. M., & Agnew, C. R. (1998). The investment relations, and person-environment interactions. model scale: measuring commitment level, satisfaction level, quality

Journal

Journal of Youth and AdolescenceSpringer Journals

Published: Jul 12, 2021

Keywords: Dating; Early adulthood; Late adolescence; Longitudinal; Relationship quality; Romantic relationships; Social withdrawal

References