Abstract
Dating violence is a deleterious relational dynamic that affects around one-third of adolescents. It is well documented that dating violence can persist over time, occurring within the same relationship or across multiple relationships. Self-esteem could be a key factor in understanding the mechanisms of violence continuity, potentially acting as a protective factor during adolescence. Exploring its role in the continuity of dating violence and the potential differences based on whether the violence occurs within the same relationship or across different relationships could inform interventions aimed at preventing repetitive violence and revictimization by multiple partners. This study aimed to (1) examine the association between dating violence measured at two distinct time points, (2) investigate the moderating role of self-esteem in this association, and (3) explore potential variations in the protective role of self-esteem based on whether the violence reoccurred within the same relationship or across different relationships. Sex differences were also examined. A subsample of 1,018 adolescents who reported being involved in a dating relationship at two time points completed self-reported questionnaires. A moderation analysis revealed that dating violence at Time 1 predicted dating violence at Time 2 (6 months later) for both boys and girls. Among girls, self-esteem acted as a protective factor when considering the effect of partner change. Hence, girls who experienced dating violence victimization at Time 1 but had higher self-esteem were more inclined to engage in a subsequent relationship devoid of dating violence at Time 2. However, self-esteem did not emerge as a significant protective factor among boys. Self-esteem appears to be a pivotal factor in promoting engagement in subsequent positive relationships following experiences of dating violence among girls. These findings offer valuable insight that can inform preventive measures and interventions, ultimately contributing to healthier relationship dynamics and improving the well-being of dating violence victims during adolescence.
Adolescence represents a critical developmental period marked by numerous new challenges for young individuals. Notably, the emergence of first dating relationships often occurs during this period (Collins et al., 2009). However, not all adolescents have the opportunity to experience positive dating relationships and some adolescents even experience dating violence. The World Health Organization defines dating violence as “any behavior within an intimate relationship that inflicts physical, psychological, and/or sexual harm or suffering on the individuals involved” (Heise & Garcia-Moreno, 2002, p. 99). A meta-analysis conducted by Wincentak et al. (2017) revealed that 1 in 5 adolescents reported experiencing physical violence and 1 in 10 reported experiencing sexual violence. Psychological violence stands as the predominant manifestation of dating violence, showing prevalence rates that range from 20% to 70% among adolescents (Stonard et al., 2014). Notably, adolescents appear particularly vulnerable to experiencing dating violence, with a higher prevalence observed within this age group compared to young adults (Rubio-Garay et al., 2017).
Dating violence has a propensity to persist over time, characterized by recurring acts of violence by one partner, as well as across multiple dating relationships (Exner-Cortens et al., 2017). The mechanisms explaining why some adolescents escape the continuity of violence over time and across relationships remain understudied. Identifying individual protective factors that can be modified through interventions would be beneficial in guiding intervention for victims and facilitating their engagement in positive subsequent relationships following experiences of victimization. The present study aimed to examine the role of self-esteem as a protective factor against the continuity of violence, comparing its role among adolescents who are repetitively victimized by a single partner to those who experience revictimization by multiple partners.
Continuity of Dating Violence
Continuity of dating violence encompasses both repetitive violence, which refers to multiple episodes of violence within a single relationship, as well as revictimization by multiple partners, a phenomenon in which an individual experiences dating violence sequentially across more than one relationship. The existing literature on violence continuity within dating relationships fails to adequately distinguish between these two phenomena and studies specifically addressing revictimization by multiple partners are scarce (Ørke et al., 2018, 2022).
Research on repetitive violence consistently supports the notion that victims, including adolescents, commonly experience multiple episodes of violence. Furthermore, violence tends to remain stable or escalate over time within the context of the same dating relationship. Indeed, studies underscore that isolated episodes of violence during a dating relationship are relatively uncommon (Basile et al., 2020; Bonomi et al., 2012; Ybarra et al., 2016).
One longitudinal study showed that experiencing dating violence increases the risk of subsequent dating violence during adolescence (Sabina et al., 2016). However, these findings include both adolescents who remained with the same partner and those who changed partners. While no study has specifically explored experiences of dating violence revictimization by different partners during adolescence, several have examined the association between dating violence during adolescence and subsequent experiences of intimate partner violence in adulthood (Cui et al., 2013; Exner-Cortens et al., 2017; Greenman & Matsuda, 2016; Jouriles et al., 2017). A longitudinal study revealed that the likelihood of experiencing intimate partner violence in adulthood is increased by up to two times if an individual was a victim of dating violence during adolescence (Exner-Cortens et al., 2017). This study also found that victimization in dating relationships during adolescence was directly associated with experiencing physical violence in a dating relationship approximately 5 years later and indirectly associated with experiencing physical violence 12 years later (Exner-Cortens et al., 2017).
A substantial proportion of individuals who report experiencing intimate partner violence during adulthood also report victimization by multiple romantic partners, with rates ranging from 23% (Vatnar & Bjørkly, 2008) to 56% (Alexander, 2009) in various studies. To our knowledge, only one study aimed to assess the prevalence of revictimization by multiple partners among adolescents (Bonomi et al., 2012). In this study, which used a random sample design, over one-third of adolescents reported having experienced dating violence by two or more partners. However, a previous study conducted by the same research team, also using a random sample design, found that adult victims primarily reported a single violent relationship since the age of 18 (Thompson et al., 2006). Taken together, these two studies suggest that adolescence may represent a period where revictimization by multiple partners is more prevalent. This could be partly explained by a lack of relationship skills during adolescence which may lead to more conflicts in dating relationships, potentially contributing to engagement in multiple, brief relationships and increased turbulence (Connolly & McIsaac, 2009; Lantagne & Furman, 2017; Meier & Allen, 2009). All of these findings highlight the importance of implementing effective prevention programs during adolescence to promote subsequent engagement in positive relationships following experiences of dating violence.
To this effect, Elder’s life course theory (Elder, 1994; Elder et al., 2003) emphasizes that the end of a violent relationship can serve as a turning point, leading to a different life trajectory, one devoid of dating violence, as a result of deliberate actions and choices. These actions and choices may encompass seeking help and making decisions not to tolerate future violence and are, in part, influenced by the individual’s self-esteem, which is shaped by past experiences of violence.
Self-Esteem as a Protective Factor
Identifying protective factors against violence continuity is of key importance to prevent increased exposure to dating violence and the associated deleterious outcomes (Taquette & Monteiro, 2019). To date, most studies have focused on risk factors, while protective factors have received limited attention (Kuijpers et al., 2012; Ørke et al., 2018). Moreover, most protective factors identified by the scientific community pertain to the social networks surrounding adolescents (e.g., social support and parental supervision; East & Hokoda, 2015; Hébert et al., 2019; Kerig, 2010; Parker et al., 2016) rather than the adolescents themselves. To optimize prevention efforts for this population, it is imperative to identify modifiable individual factors that facilitate subsequent engagement in violence-free relationships for adolescents who have experienced violence in their previous relationships.
Self-esteem refers to an individual’s subjective evaluation of their worth, self-respect, and self-confidence, encompassing their positive or negative opinion of themselves (Sedikides & Gress, 2003). Particularly malleable during adolescence, self-esteem can be influenced by experiences such as puberty, identity development, and new relational experiences (Potard et al., 2016). Indeed, dating relationships have been identified as a factor influencing the self-esteem of adolescents, and experiencing dating violence during adolescence has been associated with lower self-esteem among victims (Boislard & Van de Bongardt, 2017; Dupras & Bouffard, 2012; Taquette & Monteiro, 2019). Moreover, adolescents highly value dating relationships among their peers, which can create pressure to be in a dating relationship. In a recent study, over two-thirds of adolescents expressed a firm intention to remain with their partner despite experiencing dating violence, primarily due to perceived peer pressure to be committed to a partner (Muñoz-Rivas et al., 2022). Because part of adolescent self-esteem depends on this relationship status, young individuals with low self-esteem may be motivated to maintain their relationship at any cost, despite experiencing violence (Balleys, 2016). In addition, these adolescents may encounter more difficulty establishing boundaries in the face of their partner’s behavior, making them more vulnerable targets for perpetrators of violence (Mastromonica, 2012).
In addition to being a consequence of victimization, a meta-analysis on peer victimization concluded that low self-esteem can also predict victimization (Van Geel et al., 2018). In fact, existing literature indicates that self-esteem plays a guiding role in interpersonal behavior (Anthony et al., 2007; Leary, 2004). Individuals with higher self-esteem tend to seek more positive interpersonal relationships with better cost–benefit ratios for themselves (Cameron et al., 2010). This suggests that individuals whose self-esteem has been particularly impacted by past violence may face a higher risk of violence continuity.
When considering the body of research, self-esteem appears to be associated with dating violence and its continuity (Krahé & Berger, 2017; Van Bruggen et al., 2006; Yolcu & Akbay, 2020). However, while higher self-esteem has been linked to improved social integration, psychological well-being, and resilience (Ayedi et al., 2018), its potential protective role against the continuity of dating violence remains unexplored. It is plausible that adolescents who have experienced dating violence, but who maintained a positive self-esteem, may display reduced vulnerability to revictimization by multiple partners. Similarly, among victims of repetitive violence, those with higher self-esteem may exhibit a greater inclination to end the violent relationship.
Current Study
The first aim of this study is to examine the continuity of dating violence during adolescence using a short-term prospective design. A positive relationship is expected between the reported violence at Time 1 (T1) and Time 2 (T2). The second objective of this study is to examine the potential moderating role of self-esteem in the continuity of dating violence. It is hypothesized that self-esteem will act as a moderator in the association between dating violence victimization at both measurement points. Specifically, it is expected that the experience of dating violence at T1 will be more strongly associated with the experience of dating violence at T2 among adolescents with lower self-esteem. Lastly, the third aim of this study is to explore potential variations in the protective role of self-esteem based on whether the violence is repeatedly experienced by a single partner or inflicted by multiple partners. Given the absence of existing literature to inform hypotheses, this objective of the study is approached in an exploratory manner.
Method
Participants and Procedure
Data for the current study are derived from a large-scale longitudinal study, entitled the Youth Romantic Relationships Survey. A representative sample of 8,194 students in grades 10 to 12 (i.e., from approximately 14 to 17 years old) was recruited at Time 1 (T1), and 5,932 students completed the follow-up survey at Time 2 (T2). The two measurement times were spaced approximately 6 months apart. To be included in the present analysis, adolescents had to partake in T1 and T2 of the study, report being in a dating relationship at both time points, and provide the initials of their partner at each measurement time (n = 1,042). In the current study, a “dating relationship” was defined as “having a boyfriend or a girlfriend” or “dating a guy or a girl.” The provided partner’s initials were meticulously cross-referenced at both time points, ensuring consistency with reported relationship duration at T2 (e.g., participants were excluded if different initials were reported at T1 and T2 but reported being together for more than 6 months at T2). Two indices of relationship duration were considered (i.e., length of the relationship as a continuous variable and participants’ responses to the question “were you with the same partner 6 months ago”). In addition, the time interval between T1 and T2 was considered when less than 6 months passed between data collection points to ensure that participants were not wrongfully eliminated. Only participants with complete coherence were retained in the sample, resulting in the exclusion of 24 participants due to inconsistencies. These indices were used to contrast repeated victimization by the same partner with revictimization by multiple partners. The resulting subsample consisted of 1,018 adolescents aged from 14 to 19 (718 girls, 300 boys; M = 15.62 years, SD = 0.93). Missing data for all study variables ranged from 0.29% to 6.78%. Because the percentage of partial respondents is below 10% on all variables, a pairwise deletion approach was used to manage missing data (Asparouhov & Muthén, 2010; Newman, 2014), leading to 79 participants being excluded from the final analyses. Thus, the analyses were completed on a sample of 939 adolescents. Adolescents had been in a relationship for an average of 9.03 months (SD = 9.18) at T1 and 12.28 months (SD = 9.51) at T2 and mostly described their parents’ ethnic background as Canadian (83.8%). Most participants reported being in heterosexual dating relationships (97.6% at T1 and 96.6% at T2).
Adolescents were recruited from 34 high schools in Québec, Canada. High schools were selected through a stratified cluster sampling procedure. To obtain a representative sample of high school students, schools were classified into eight strata based on their metropolitan geographical region, school status (i.e., public or private), language of instruction (i.e., French or English), and index of social and economic disadvantage. Schools were then randomly selected from a list of eligible schools provided by the Ministry of Education of Québec from 2011 to 2012. Students completed the questionnaires in the fall of 2011 (T1) and approximately 6 months later in the spring of 2012 (T2). Two research assistants were present to explain the study to students and answer any questions. Questionnaire completion took approximately 40 min and could be done in English or French, depending on the main language of instruction in the school. Following completion, students had the option to enter a draw for a chance to win an iTunes gift card (T1) and an iPod Shuffle (T2). This study was approved by the ethics board of the Université du Québec à Montréal.
Measures
Dating Violence
A shortened version of the Conflict in Adolescent Dating Relationships Inventory (Wolfe et al., 2001) was used to measure experiences of psychological and physical violence in dating relationships. Adolescents had to indicate the frequency with which their current partner engaged in the described behaviors during a conflict. Participants were asked to refer to the past 12 months at T1 and the past 6 months at T2. At T1, the measure consisted of eight statements with four anchors (0 = never, 3 = six times or more), with three reflecting psychological violence experienced (e.g., “Ridiculed or made fun of you in front of others”), three reflecting physical violence experienced (e.g., “Kicked, hit, or punched you”), and two reflecting threats of physical violence (e.g., “Threatened to hit or throw something at you”). At T2, the same items were included, but an item measuring control through the use of a cell phone was added (i.e., “Used a cell phone to control you [by asking you to tell her/him where you are and what you are doing, making you show her/him your text messages, etc.]”). This questionnaire has been validated with Canadian adolescents (Wolfe et al., 2001).
A shortened version of the Sexual Experiences Survey (Koss et al., 2007) was used at each of the two measurement times to assess experiences of sexual violence. Statements were adapted from the original scale to assess sexual violence experienced in the context of dating relationships. Participants were asked to provide the frequency of behaviors exhibited by their partner, referring to the past 12 months at T1 and the past 6 months at T2. The tool consists of nine statements (e.g., “Tried to have sex with you when you did not want to by using arguments or pressure”) with four anchors ranging from (0 = never, 3 = six times or more).
At each measurement time, a dichotomous score was created based on the absence (0) or presence (1) of violence in dating relationships (i.e., psychological, physical, and sexual). Participants who responded with anything other than “never” to any of the statements regarding experienced violence were classified as having experienced dating violence at that measurement time. This score was retained for the planned analyses.
Partner Change
To ascertain whether adolescents were involved with a different partner between T1 and T2, we utilized the duration of the relationships along with the partners’ initials provided by participants. A dichotomous score was generated based on the consistency between the provided partners’ initials and the relationship duration (0 = different partner, 1 = same partner). For instance, a participant who provided the initials “AB” for their partner at T1 and “AB” again at T2, while asserting that they have been in the same relationship for over 6 months at T2, would receive a score of 1. All participants who provided inconsistent information were excluded from the analyses to avoid potential classification bias (e.g., different initials at T1 and T2, but claiming to be in a relationship for over 6 months).
Self-Esteem
Self-esteem was measured at T1 using the French adaptation of the Self-Description Questionnaire (Marsh, 1988; Statistique Canada, 2007). Participants responded to five statements (e.g., “In general, I like myself the way I am”) from the questionnaire, indicating the extent to which each statement was true for them on a 5-point scale from (0 = false, 4 = true). The overall score, ranging from 0 to 16, was calculated as the average of the scores obtained for each statement, with a higher score indicating higher self-esteem. The retained score for the analyses was a derived dichotomous score (0 = normal, 1 = low), where a score equal to or lower than 10 (lower decile of the weighted distribution) was deemed indicative of low self-esteem and a score higher than 10 was indicative of normal self-esteem (Garriguet, 2005).
Planned Analyses
Chi-square analyses were computed with SPSS 27 (IBM Corp., 2020) to examine the prevalence of each pattern of dating violence experiences among adolescents who reported dating violence at T1 (i.e., repetitive violence, escaped repetitive violence, revictimization by multiple partners, escaped revictimization by multiple partners). The prevalence of adolescents with low and normal self-esteem who were revictimized by multiple partners and those who escaped revictimization were also examined. Relationship duration was considered as a covariate in this study because of its association with the occurrence of dating violence and because adolescent victims of repetitive violence were, at T2, in a relationship for a minimum of 6 months while victims of revictimization by multiple partners were in a relationship for a maximum of 6 months (Herrero et al., 2018; Johnson et al., 2015). Because adolescents were recruited across different schools, intra-class correlation coefficients were calculated to assess the degree of similarity among students within the same school compared to those from different schools. A sex invariance analysis (i.e., an analysis comparing a model constrained to be equal between sexes to an unconstrained model) was conducted to determine if the model presented a better fit when both girls and boys were combined or when they were considered separately. A moderation analysis was then conducted using Mplus 8.0 (Muthén & Muthén, 1998–2017) to address the objectives of the current study. This analysis assessed the role of self-esteem as a protective factor against repetitive violence within the same relationship and revictimization by multiple partners. The regression model included dating violence at T1 (absence/presence), partner change (same partner/different partner), and self-esteem (low/normal) measured at T1 as independent variables in the prediction of dating violence at T2 (absence/presence). Thus, this analysis makes it possible to document the association between dating violence at T1 and dating violence at T2 (objective 1).
To investigate whether self-esteem acts as a protective factor against the continuity of violence (objective 2), a two-variable interaction term was created (e.g., dating violence T1 × self-esteem T1). A two-variable interaction term allows us to examine how the effect of one independent variable on the dependent variable changes based on the level of another independent variable. The significance of this interaction effect would mean that, among adolescents with low self-esteem, the relationship between dating violence at T1 and T2 is stronger than among adolescents with normal self-esteem. The absence of a significant effect would mean that self-esteem does not act as a protective factor against violence continuity.
To investigate whether self-esteem differentially protects adolescent victims of repetitive violence and of revictimization by multiple partners, a triple interaction term was created to predict dating violence at T2 (e.g., dating violence T1 × self-esteem T1 × partner change). A triple interaction term allows us to explore how the relationship between the dependent variable and one independent variable depends on the joint influence of the other two independent variables. The observation of a three-variable interaction effect would imply that the protective role of self-esteem in the relationship between dating violence at T1 and T2 is stronger among adolescents who have changed partners (i.e., victims of revictimization by multiple partners) than among those who are with the same partner. If self-esteem could not be established as a protective factor against violence continuity, while the triple interaction term was significant, it would mean that self-esteem serves as a protective factor against revictimization by multiple partners, but not against repetitive violence.
Results
Preliminary Analyses
Descriptive analyses
Descriptive analyses revealed that 46.51% of girls and 39.93% of boys experienced violence continuity. Of the total sample, 38.93% of girls and 33.58% of boys were victims of repetitive violence, while 7.58% of girls and 6.34% of boys were victims of revictimization by multiple partners. Of adolescents who matched the criteria for potential repetitive violence (i.e., reported dating violence at T1 and being with the same partner across T1 and T2), 15.76% of girls and 15.09% of boys escaped repetitive violence. Of adolescents who matched the criteria for potential revictimization by multiple partners (i.e., reported dating violence at T1, but changed partners between T1 and T2), 45.75% of girls and 45.16% of boys escaped revictimization. The prevalence of girls and boys with low and normal self-esteem who experienced repetitive violence and revictimization by multiple partners as well as the prevalence of those who escaped repetitive violence and revictimization by multiple partners are presented in Table 1. The intra-class correlation coefficients for all outcome variables were ≤.002. These low intra-class correlation coefficients imply minimal dependency within schools, justifying the use of single-level analysis (Snijders & Bosker, 2012).
Prevalence of Patterns of Dating Violence Experienced by the Self-Esteem of Adolescents Who Experienced Dating Violence at T1.
Sex Invariance
Sex invariance was tested using the chi-square difference testing with weighted least squares (WLSMV) estimation in Mplus 8.0 (Muthén & Muthén, 1998–2017). This test was used to assess whether the constrained model had a significantly better model fit than the unconstrained model. First, separate models for boys and girls were conducted and compared to models with factor loadings constrained to be equal between boys and girls. A marginal p-value was obtained (p = .056), which suggests measurement invariance for factor loadings. The second step aimed to compare separate models for each sex to models with factor loadings and intercepts constrained to be equal between both sexes. The test was significant (p = .023), which indicates that the model fit of the constrained model was significantly worse than the unconstrained model’s fit. Thus, the unconstrained model was retained for the main analyses of the current study (i.e., separate models for girls and boys).
Main Analyses
Main Effects and Moderation Effects
The results of the moderation analyses among girls and boys are presented in Table 2. The main effects revealed that, after accounting for the duration of the relationship, dating violence at T1 predicted dating violence at T2 for both girls and boys. Partner change was a significant predictor of dating violence at T2 among boys while being marginally associated with dating violence at T2 among girls. Self-esteem was not found to be predictive of dating violence at T2 for both girls and boys. Exploring the moderation effects, the two-way interaction term (dating violence T1 × self-esteem T1) was not significantly associated with dating violence at T2, among girls and boys. This means that self-esteem does not moderate the association between dating violence at T1 and T2. However, the three-way interaction term (dating violence T1 × self-esteem T1 × partner change) was significantly and positively associated with dating violence at T2, but only among girls. Hence, self-esteem moderated the association between dating violence at T1 and T2, only among adolescent girls who reported having changed partners between both measurement points. The model explained 30.7% of the variance of dating violence T2 among girls and 28.3% among boys.
Moderation Effects of Self-Esteem and Partner Change on the Relationship Between Dating Violence at T1 and T2 Among Girls and Boys.
Note. SE = standard error.
Simple Effects
To better understand the significant moderation effect observed on the triple interaction term, analyses of simple effects were conducted (i.e., specific associations between both time points of dating violence within each level of self-esteem and partner change). Among adolescents who reported normal self-esteem and being with the same partner throughout T1 and T2, dating violence at T1 predicted dating violence at T2 for both girls (β = .58, p < .001) and boys (β = .49, p < .001). Among adolescents who reported low self-esteem and being with the same partner throughout T1 and T2, dating violence at T1 significantly predicted dating violence at T2, but only for girls (β = .49, p < .001). This association was marginal for boys (β = .35, p = .068). For adolescents with normal self-esteem but who reported having changed partners between T1 and T2, dating violence at T1 was not associated with dating violence at T2 for girls (β = .06, p = .614) and for boys (β = .19, p = .312 for boys,). Lastly, for adolescents who reported low self-esteem and partner change between T1 and T2, dating violence at T1 was predictive of dating violence at T2 among girls (β = .32, p = .016), but marginally associated with dating violence at T2 among boys (β = .50, p = .082).
Discussion
While the field of dating violence research has expanded significantly in recent decades, a notable scarcity persists in terms of understanding protective factors against violence continuity. Moreover, by failing to differentiate between individuals who experience repetitive violence within the same relationship and those who are revictimized by multiple partners, previous research might have inadvertently obscured significant variations in the mechanisms that contribute to the continuity of violence (Kuijpers et al., 2012; Ørke et al., 2018). To address these gaps, this study aimed to investigate the association between dating violence victimization across two measurement points. This study also aimed to examine the potential protective role of self-esteem in this association. Lastly, the study aimed to determine whether the role of self-esteem varied based on whether the dating violence was perpetrated repeatedly by a single partner or by multiple partners.
The findings partially supported our hypotheses. Indeed, the results revealed a positive association between dating violence at T1 and dating violence at T2. This finding is in line with previous studies establishing that dating violence tends to persist over time and across relationships (Exner-Cortens et al., 2017; Sabina et al., 2016). In fact, 46.51% of girls and 39.93% of boys experienced violence continuity (38.93% and 33.93% experiencing repetitive violence within the same relationship and 7.58% and 6.34% experiencing revictimization by multiple partners, for girls and boys, respectively). Our findings underscore that, even among adolescents, the likelihood of experiencing violence continuity within dating relationships is substantial, underscoring the imperative to explore protective factors in this association.
Contrary to our initial hypotheses, self-esteem did not moderate the association between dating violence at T1 and dating violence at T2. This suggests that self-esteem cannot be unequivocally established as a protective factor against violence continuity, encompassing both repetitive violence within the same relationship and revictimization by multiple partners. However, when considering partner change, self-esteem did moderate the association between dating violence at both times among girls. Specifically, self-esteem played a protective role in preventing revictimization by multiple partners among girls. In other words, girls who experienced dating violence at T1, but exhibited elevated levels of self-esteem, were more likely to engage in a subsequent dating relationship devoid of dating violence at T2. These results highlight the existence of distinct mechanisms underlying repetitive violence within the same relationship and revictimization within different relationships in adolescence (Kuijpers et al., 2012; Ørke et al., 2018).
Combined, these results illustrate that self-esteem does not serve as a protective factor against repetitive violence. In cases of repetitive violence within the same relationship, the capacity of self-esteem to alter trajectories could be limited, notably because of the entanglement into a power dynamic. The power dynamic at play and the emotional commitment felt by victims toward their partners might contribute to a distorted perspective on their experiences, even in the presence of higher self-esteem (Heron et al., 2022). Adolescents who are victimized repeatedly within the same relationship might not fully realize that their relationship is violent, or they could minimize their victimization experiences (Fernet et al., 2019). It is also well documented that ending a violent relationship is a challenging process for victims, notably due to the hope of their partner changing, the social pressure to be in a committed relationship, fear of retaliation from the perpetrator after the cessation of the relationship as well as the stigmatization and social isolation associated with dating violence (Heron et al., 2022; Muñoz-Rivas et al., 2022). Therefore, the mere presence of a higher self-esteem might not be sufficient to prompt victims to terminate the relationship. Thus, victimization continues throughout the relationship. It is possible that a cumulation of protective factors is necessary to interrupt repetitive violence (e.g., self-efficacy to deal with violence, social support, and parental supervision; Greenman & Matsuda, 2016).
On the other hand, the termination of a relationship provides victims with the opportunity to heal and retrospectively assess their previous dating experience. Within this context, self-esteem functions as a protective factor, preventing girls from reengaging in a subsequent violent relationship. This phenomenon could be attributed, at least in part, to the newly obtained absence of a power dynamic inherent to dating violence, creating space for self-esteem to exert its influential role in guiding interpersonal relationship selection (Anthony et al., 2007; Cameron et al., 2010; Leary, 2004). Notably, individuals boasting higher self-esteem tend to seek out more positive relationships, a tendency possibly heightened in girl victims of dating violence as a result of their desire to avoid revictimization by multiple partners. Those with elevated self-esteem may perceive themselves as deserving of healthier dating relationships in the future, consequently adopting a more cautious approach when selecting partners (Cameron et al., 2010).
Elder’s life course theory (Elder, 1994; Elder et al., 2003) lends further insight into this result. Indeed, the theory posits that the termination of a dating relationship can serve as a pivotal juncture, steering individuals toward alternative life paths that exclude dating violence through deliberate actions and decisions, such as refusing tolerance of violence in the future. These actions and choices are determined by underlying mechanisms, such as self-esteem. Thus, the termination of a violent relationship should be comprehended as a turning point wherein interventions aiming at increasing self-esteem can play a crucial role in fostering reengagement in healthier relationships following the experience of dating violence.
The moderating effect of self-esteem did not yield statistical significance in relation to revictimization by multiple partners among boys. However, simple effects provided a noteworthy insight into the interplay between self-esteem, partner change, and the association between dating violence at T1 and T2. Interestingly, the observed association between dating violence at T1 and T2 did not persist among both boys and girls who reported normal self-esteem and indicated a change in partners between measurement points. This intriguing finding raises the possibility that the moderating influence of self-esteem on the association between dating violence at T1 and T2 might not have been fully captured among boys who changed partners. It is plausible that the reduced statistical power resulting from a smaller sample size of boys could have contributed to the inability to detect a significant moderation effect in this subgroup. Moreover, for two of the simple effects observed, results were significant among girls, but marginal among boys. The presence of marginal results further suggests that, in a larger sample of boys, such effects could have been detected.
Another explanation for the absence of a moderation effect among boys pertains to the fact that girls’ self-esteem is typically more affected by dating violence than boys’ (Jankowiak et al., 2021; Marcos et al., 2020). The banalization of violence experienced by boys may potentially contribute to reduce the effect of dating violence on their self-esteem (Jankowiak et al., 2021). Moreover, girls may assign greater value to dating relationships compared to boys (Cava et al., 2023). Consequently, this divergence could potentially lead to girls’ self-esteem being more susceptible to the impact of dating violence (Cava et al., 2023). Therefore, it is possible that boys’ self-esteem is more dependent on other influential factors (e.g., sports and athletic performance; Dupras & Bouffard, 2012) and thus remain at higher levels during adolescence (Jankowiak et al., 2021), reducing the ability to observe a significant effect. It is also plausible that self-esteem, as an isolated protective factor, has limited efficacy in mitigating revictimization by multiple partners among boys because of their inclination to swiftly enter new dating relationships as a coping mechanism following break-ups (Shimek & Bello, 2014). Such rapid relationship transitions may result in less discerning partner choices, consequently elevating the likelihood of involvement in subsequent violent relationships (Carter-Snell, 2015). This phenomenon, possibly stemming from the lack of constructive coping mechanisms and robust support networks, might impede the moderating effect of self-esteem on revictimization by multiple partners.
This study is the first to examine the protective role of self-esteem against violence continuity. Furthermore, it provides an additional contribution by distinguishing between victims of repetitive violence within the same relationship and those experiencing revictimization by multiple partners. The results of this study contribute to a deeper understanding of explanatory mechanisms (i.e., self-esteem) that can facilitate re-engagement in a violence-free relationship following a prior violent relationship.
However, this study has some limitations. Notably, while the use of dichotomous total scores of dating violence was necessary to offer valuable insight into this study’s objectives, it is important to acknowledge that this approach might obscure crucial information related to the experience of dating violence, including its frequency and forms (i.e., psychological, physical and sexual violence). Future research could benefit from utilizing continuous variables (i.e., frequency of victimization experiences) and delving into the different forms of violence experienced by adolescents. Using this more nuanced approach, researchers could explore whether self-esteem’s influence is similar across various typologies of dating violence or if it is more pronounced in specific forms of violence experienced. It could also be of interest to gather more context on the termination of relationships (e.g., which partner decided to end the relationship) and explore protective mechanisms against the perpetration of dating violence in multiple relationships, as this study did not consider the reciprocity that may exist in dating violence dynamics (Fernández-González et al., 2020). Considering dating violence mutuality may provide a more comprehensive understanding of the mechanisms at play, as it is possible that the protective effect of self-esteem against revictimization by multiple partners varies depending on whether the violence is mutual or one-sided. Moreover, when interpreting the current results, it is important to keep in mind that some adolescents who participated in the study might have been dating each other, potentially resulting in both parties reporting victimization for the occurrence of the same conflict. While recognizing the potential statistical bias introduced by such interdependence, this study was anonymous and did not use a dyadic design, it was therefore impossible to match data of both partners.
Most of the participants in the current study were girls of Canadian ethnicity who reported being in heterosexual dating relationships. Further inquiry into the role of self-esteem as a protective factor against violence continuity among boys, and adolescents representing a range of socioeconomic classes, ethnicities as well as sexual and gender identities is also warranted. The acquisition of a larger, more inclusive, and diverse sample could offer a more comprehensive understanding of the mechanisms at play across a broader spectrum of identities. Addressing these limits in forthcoming research is crucial to ultimately inform prevention initiatives and increase the generalizability of the current findings.
Future research should also explore additional individual and modifiable factors to achieve a more comprehensive understanding of protective factors against violence continuity that could be targeted in interventions with adolescents (e.g., adolescents’ self-efficacy to deal with dating violence and their attitudes toward dating violence).
Conclusion
The findings of this study revealed that self-esteem serves as a protective factor against revictimization by multiple partners among adolescent girls. This highlights a new potential avenue for interventions aimed at preventing revictimization. Interventions should focus on enhancing adolescent girls’ self-esteem, particularly among those who have experienced dating violence. The termination of a dating relationship marks a pivotal moment for the implementation of these interventions. By bolstering self-esteem, adolescent girls may be better equipped to navigate dating relationships and seek more positive dating relationships in the future. Recognizing self-esteem’s role in reducing the likelihood of revictimization by multiple partners emphasizes the importance of fostering empowerment and resilience for adolescent girls who have been victims of dating violence. It is crucial to emphasize that the findings of this study should never be interpreted in a way to attribute blame on the victims for the violence inflicted upon them. The sole responsibility for dating violence always rests with the perpetrator (Hamby & Grych, 2016). Moreover, we advocate for future research endeavors to distinguish between victims of repetitive violence and victims of revictimization by multiple partners. This differentiation is crucial as these two phenomena seem to diverge in terms of risk factors, consequences, and as evidenced by this study, protective factors as well (Kuijpers et al., 2012; Ørke et al., 2018). Overall, the identification of self-esteem as a protective factor against revictimization by multiple partners offers valuable insight that can guide preventive measures and interventions, ultimately contributing to healthier relationship dynamics and improved well-being for victims.
Footnotes
Acknowledgments and Credits
The authors wish to thank the teenagers and the school personnel who participated in the Youths’ Romantic Relationships Survey. Our thanks also extend to Catherine Moreau for project coordination and Manon Robichaud for data management.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interests with respect to the authorship and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research and/or authorship of this article: This work was supported by the Canadian Institutes of Health Research (Grant # 103944).
