Child Abuse & Neglect 97 (2019) 104167
Contents lists available at ScienceDirect
Child Abuse & Neglect journal homepage: www.elsevier.com/locate/chiabuneg
Trajectories of psychological dating violence perpetration in adolescence
T
Andréanne Lapierrea, Alison Paradisa, Emily Todorova, Martin Blaisb, ⁎ Martine Hébertb,c, a
Psychology Department, UQAM, Montreal, Quebec, Canada Sexology Department, UQAM, Montreal, Quebec, Canada c Canada Research Chair in Interpersonal Traumas and Resilience, Canada b
A R T IC LE I N F O
ABS TRA CT
Keywords: Psychological dating violence Trajectories Risk factors Self-esteem Risky behaviors Peers’ characteristics
Background: More than one adolescent out of three will use psychological dating violence (DV) as a strategy to resolve conflict in romantic relationships, which will have major consequences on their partner’s well-being. However, over time, most adolescents will learn skills to better interact and face conflicts, suggesting that psychological DV rates should decrease over time. Yet, because of individual characteristics and experiences, evolution over time may differ across adolescents. Risk factors for psychological DV have not been examined specifically, even though it is the most common form of dating violence. Objective: This study aimed to explore latent trajectories of psychological DV perpetration and examine risk factors predicting trajectory group membership. Participants and methods: A sub-sample of 449 adolescents who reported being in a dating relationship for the three waves of the [blind for review] completed printed and online self-report questionnaires. Results: Results from the group-based modeling identified 4 trajectories of psychological DV perpetration: absence of violence (30.7%), low violence (61.3%), high descending (4.2%), and moderate elevating (3.8%). Age, DV victimization, exposure to father toward mother violence, low self-esteem, marijuana use, DV victimization of peers, and antisocial behaviors of peers predicted trajectory membership. The model explained 36.7% of the variance. Conclusions: These results support the need for a person-oriented approach to study psychological DV and for developing prevention programs adapted to the specific characteristics of vulnerable youth.
1. Introduction Adolescence is a developmental stage during which numerous physical and psychological changes occur. Early dating experiences create new relational challenges which, combined with unrealistic expectations, may cause conflicts and stress, and lead adolescents to adopt various behaviors to maintain the relationship (Harper & Welsh, 2007), including dating violence (DV). DV is a major public health issue with substantial social costs (Leen et al., 2013; Offenhauer & Buchalter, 2011), and deleterious consequences on the
⁎ Corresponding author at: Département de sexologie, Université du Québec à Montréal, C.P. 8888, Succursale Centre-Ville, Montréal, Québec, H3C 3P8, Canada. E-mail address:
[email protected] (M. Hébert).
https://doi.org/10.1016/j.chiabu.2019.104167 Received 2 October 2018; Received in revised form 22 August 2019; Accepted 27 August 2019 Available online 05 September 2019 0145-2134/ © 2019 Elsevier Ltd. All rights reserved.
Child Abuse & Neglect 97 (2019) 104167
A. Lapierre, et al.
physical, mental, and sexual health of youth (Chiodo et al., 2011; Exner-Cortens, Eckenrode, & Rothman, 2013). In addition, DV is a widespread phenomenon. Approximately 1 in 2 adolescents report having used some form of DV (Hébert, Blais, & Lavoie, 2017; Wincentak, Connolly, & Card, 2017; Ybarra, Espelage, Langhinrichsen-Rohling, Korchmaros, & Boyd, 2016), the most common form being psychological DV perpetration. In fact, according to recent data, the lifetime prevalence of psychological violence perpetrated in dating relationship is 33% for boys and 46% for girls (Ybarra et al., 2016). Similar rates (25% to 49%) have been found with a 12month reference period (Hébert, Lapierre, Lavoie, Fernet, & Blais, 2018; Hébert et al., 2017). Psychological DV consist of any acts intended to threaten, denigrate, deceit, or control, and that destabilizes the dating partner or compromises his/her well-being (Lavoie, Hotton-Paquet, Laprise, & Joyal Lacerte, 2009). Adolescents need to constantly adjust their behaviors facing psychological and interpersonal changes that take place in their life. As adolescents mature, they gradually become better equipped to respond to the interpersonal challenges they face, and they develop skills to negotiate intimate conflicts (Connolly & McIsaac, 2009). We can expect these skills to lead to diminished psychological DV perpetration over the course of adolescence. While trajectories of DV victimization have been documented (Brooks-Russell, Foshee, & Ennett, 2013; Swartout, Cook, & White, 2012), research on the evolution of DV perpetration is particularly scarce. However, adolescence is an important developmental period during which the use of certain abusive dating behaviors can emerge and become established patterns over time. By examining perpetration, we wish to move the focus away from the victims and instead emphasize moral responsibility of the perpetrator. Studies on perpetration may greatly inform practices and prevention programming by identifying characteristics in youth that increase the risk of dating violence perpetration. Thus, the current study aims to explore psychological DV perpetration over time to determine what patterns of perpetration adolescents experience as they get older. Victimization experiences were not considered in the current trajectory analysis. Known to be reported more often than perpetration (Wincentak et al., 2017; Ybarra et al., 2016), victimization had the potential of overshadowing some of the specificities related to the perpetration trajectories. DV victimization was, however, included as a risk factor in the prediction of trajectory membership. 1.1. DV perpetration over the adolescent life course The few available longitudinal studies on perpetration of DV have documented prevalence rates of perpetration across time, but have not shown how these rates evolve from one time point to another (i.e., in examining linear, quadratic or cubic tendencies), or how this evolution possibly differs for subgroups. For example, Temple, Shorey, Fite, Stuart, and Le (2013) explored evolution of girls and boys physical DV perpetration with a single one-year follow-up. They found no differences in trajectories of boys and girls, but did not examine if latent subgroups could show distinct trajectory patterns of perpetration, pathways that may relate to individual or relational characteristics. In recent years, researchers have integrated person-centered approaches to longitudinal analysis to distinguish pathways and better reflect individual heterogeneity. A person-oriented approach assumes that the behaviors vary from one person to another, according to their individual differences and the characteristics of the subgroup to which they belong (Sterba & Bauer, 2010). Some cross-sectional studies (e.g., Chiodo et al., 2011; Hébert, Moreau, Blais, Oussaïd, & Lavoie, 2018; Spencer et al., 2016) have since used this approach to examine profiles of DV victimization and perpetration. These studies have considered different forms of violence (psychological, physical and sexual) to cluster experiences of DV. A few rare studies have explored DV in longitudinal designs relying on a person-oriented approach, but failed to distinguish between victimization and perpetration when examining profiles, which may have clouded our understanding of the course of DV perpetration (Brooks-Russell et al., 2013; Haynie et al., 2013; Miller et al., 2013; Orpinas, Hsieh, Song, Holland, & Nahapetyan, 2013). In studies that did examine specifically perpetration, most chose to concentrate on less prevalent forms of perpetration, such as physical DV (Chang, Foshee, Reyes, Ennett, & Halpern, 2015; Foshee et al., 2013; Orpinas, Nahapetyan, & Truszczynski, 2017). One of these studies conducted by Orpinas et al. (2017) examined physical DV perpetration in teens from grades 6 to 12. Their results showed two distinct trajectories of perpetration over time: a low and stable trajectory and an increasing trajectory. Further, their findings suggested that the trajectories differed in terms of the number of at-risk behaviors and supported the need to adopt a longitudinal person-oriented approach to study DV and risky behaviors in adolescence. Foshee et al. (2013) explored peer characteristics, and analyzed between subjects and within person effects on physical DV perpetration. They found that while adolescents reported perpetrating lower levels of physical DV when they had more friends with pro-social beliefs, they reported higher levels of DV perpetration throughout their adolescence when they were high in social status among their peers. These results suggest that peers’ characteristics may play an important role in DV perpetration over time and that the influence of peers should not be neglected when studying DV perpetration. Chang et al. (2015) published the only other study that examined DV perpetration with a longitudinal person-oriented approach, comparing boys and girls living in low and high disadvantaged neighborhoods. Their results showed that physical DV perpetration is higher for girls who lived in a highly disadvantaged neighborhood. Trajectories from grade 8 to grade 12 were all similarly quadratic, suggesting that, while different levels of violence were reported, similar patterns of evolution were observed. These results are interesting as they showed that DV evolves over the course of adolescence, and that this evolution differs for some individuals. However, these findings relate to physical DV only; whereas psychological DV remains largely unexplored even if it is the most widespread form of DV (Hébert, Lapierre et al., 2018, Hébert, Moreau et al., 2018; Ybarra et al., 2016). Psychological dating violence (DV) is distinct from physical violence theoretically and experientially, as it does not cause physical harm. Instead, psychological DV is conceptualized as a tactic to control or destabilize a dating partner, resulting in an impact on his/ her psychological well-being (Lavoie et al., 2009). In adolescence, psychological DV is the most common form of violence (Ybarra et al., 2016) and its use increases the risk of later perpetrating intimate partner violence (O’Leary, Tintle, & et Bromet, 2014). The use of physical aggression is known to decrease with age for most people (Barker et al., 2007), as the social codes are learned and 2
Child Abuse & Neglect 97 (2019) 104167
A. Lapierre, et al.
executive functioning competencies increase. For example, the rates of physical aggression observed in kindergarten decrease later in elementary school (Côté, Vaillancourt, LeBlanc, Nagin, & Tremblay, 2006). In a dating violence context, studies have shown that trajectories of physical aggression are absent or low and, for most individuals, stable while increasing for others (Chang et al., 2015; Foshee et al., 2013; Orpinas et al., 2017). Results from longitudinal studies on the prevalence of DV suggest that while most adolescents successfully know how to repress their physical aggressive behaviors they have more difficulties repressing their verbal aggressive behaviors, with rates of psychological DV consistently higher than physical ones (respectively around 33% and 20%, Hébert et al., 2017; Wincentak et al., 2016; Ybarra et al., 2016). Considering that refraining from using psychological violence in relationships may require higher self-regulation skills than physical violence, and that it is not as socially repressible as the later (Courtain & Glowacz, 2018), our hypothesis is that psychological DV may not follow the same patterns of evolution as physical DV. We postulate that since physical aggression decrease over time for most people, psychological violence will also decrease (our hypothesis). However, we think that for some people, psychological violence frequency will remain the same over time, as it is more difficult to refrain from perpetrating psychological violence (e.g., insulting someone, shouting), than physical violence (e.g., slapping, punching). 1.2. Risk factors for psychological DV perpetration Risk factors for DV perpetration (all forms confounded) are well known (Foshee et al., 2013; Garthe, Sullivan, & McDaniel, 2017). Among key factors, insecure attachment (e.g., Ulloa, Martinez-Arango, & Hokoda, 2014), antisocial behaviors (e.g., running away from home, substance use), and affiliation with deviant peers or with peers that are victims of DV (Chiodo et al., 2011; Foshee, Linder, MacDougall, & Bangdiwala, 2001; Foshee et al., 2013; Garthe et al., 2017; Shorey, Rhatigan, Fite, & Stuart, 2011) have been identified. Some distal factors, such as witnessing interparental intimate partner violence in childhood, are still being debated by researchers in the field. For example, while some studies have found a significant effect of witnessing interparental intimate partner violence on DV perpetration (O’Donnell et al., 2006), others have not (Linder & Collins, 2005; Gover, Jennings, Tomsich, Park, & Rennison, 2011). At this time, risk factors already identified for DV have rarely been examined specifically for psychological DV perpetration. Psychological DV may have its own specific risk factors. For example, general theory of violence suggests that when someone feels threatened, he will use aggression in order to protect himself (Berkowitz, 1989, 1993, 2001). People with lower self-esteem usually feel more rapidly threatened in interpersonal context (VanDellen, Campbell, Hoyle, & Bradfield, 2010). Theories about violence and self-esteem suggest that aggression may be triggered by the feelings of shame or inferiority provoked by situations in which others project a depreciated image (for e.g., incompetency) or when the projected image does not correspond to one’s ideal self (Bushman & Baumeister, 1998; Tracy & Robins, 2003). Low self-esteem has been found to be associated with violence perpetration in general (see Ostrowsky, 2010 for a review). Surprisingly, however, it has never been examined in association with DV perpetration specifically. We postulated that psychological DV may be influenced by low self-esteem. 2. Current study Psychological DV is the most common form of DV (Hébert, Lapierre et al., 2018; Hébert, Moreau et al., 2018; Ybarra et al., 2016). To diminish its high prevalence, we need to better understand the risk factors associated with its occurrence. A person-oriented approach offers an optimal strategy to identify characteristics of subgroups at greater risk of using psychological violence in their relationship, which in turn may inform the development of tailored intervention and prevention programs. Therefore, the main objective of this study was to identify latent trajectories of psychological DV perpetration and to examine risk factors associated with the different classes. More specifically, this study aimed at 1) describing prevalence rates across time for psychological DV perpetration; 2) examining latent trajectories of psychological DV to identify subgroups of perpetrators; and 3) identifying predictive risk factors of trajectory membership. Risk factors were included with the rationale that they have been associated with a greater risk for physical or overall DV perpetration, and that they have not yet been examined for psychological DV perpetration only. Our hypothesis was that, as in studies on physical or overall DV, anxious attachment, DV victimization, witnessing intimate partner violence (IPV), alcohol and drug use, antisocial behaviors, DV victimization of peers and antisocial behaviors of peers would be associated with a greater risk of psychological DV perpetration. Therefore, these risk factors should increase the risk of being part of trajectories in which rates of violence are moderate or high. Even though self-esteem has not been examined yet in the DV field, previous research on general violence has shown its association with perpetration (Ostrowsky, 2010; VanDellen et al., 2010). 3. Methods 3.1. Participants The [blind for review] was a longitudinal study, and consisted of a clustered representative sample, stratified according to the socioeconomic, linguistics, and geographic characteristics of [blind for review] adolescent population. The present study used a subsample composed of adolescents who reported being in a dating relationship for all three waves of the study. At Wave 1, 4 095 adolescents reported being in a dating relationship, whereas Wave 2 had 2 329 participants (57% of the original sample), and Wave 3 had 650 (16% of the original sample). From this subsample, only 449 participants had complete valid questionnaires for the variables of interest on all time points, 353 girls (78.6%) and 96 boys, between the ages 14 and 18 (M = 15.4, SD = .85 at Wave 1). The 3
Child Abuse & Neglect 97 (2019) 104167
A. Lapierre, et al.
participants were mainly from French-Canadian origin (85.9%), lived with their parents (56.3%), and had mothers with university (31.5%) or college/professional education (41.5%). In the first Wave, participants were in grades 9 (19.4%), 10 (52.6%), or 11 (28.1%). 3.2. Procedure The study was conducted over 3 time points at 6-month intervals. All participants were volunteers, and provided written consent to participate. The anonymity of their information and data was guaranteed. By partaking in the study, participants were eligible to win different prizes (e.g., iPods Touch, gift-certificates). Self-reported data of the Waves 1 and 2 was gathered in classrooms where participants filled out a paper-form questionnaire. Wave 3 data were gathered through an online version of the same initial questionnaire after participants have been solicited through email. The institutional ethics and research board of the [blind for review] approved this study. 3.3. Measures Psychological violence perpetration was assessed at each time point with self-report questionnaires. All independent variables were measured at Wave 1, with the exception of relationship length which was measured at Wave 3 and dimensions of attachment at Wave 2. The reference period for Wave 1 was “the past 12 months” while it was “the past 6 months” for waves 2 and 3. The H coefficients were used instead of Cronbach’s alpha to better estimate reliability (internal consistency) of the brief measures (McNeish, 2017). 3.3.1. Psychological DV perpetration Three items were adapted from the Conflict in Adolescent Relationship Inventory (CADRI, Wekerle et al., 2009; Wolfe et al., 2001) to measure psychological DV perpetration at Wave 1. Because this study was part of a large national survey and given time constraints associated with in class administration, brief versions were used. Based on past studies (Wekerle et al., 2001, 2009), the authors referred to the validation study (Wolfe et al., 2001) and selected items presenting the highest loadings on the psychological DV subscale. Items were “I said things just to make him/her angry”; “I ridiculed or made fun of him/her in front of others” and “I kept track of who he/she was with and where he/she was”. Participants rated on a 4-point Likert scale (0 = never; 1 = 1 to 2 times; 2 = 3 to 5 times; 3 = 6 times or more) how often they used these behaviors toward their partner in the context of their current romantic relationship. The internal consistency of the scale was satisfactory for the current sample (from H = 0.72 to H = 0.80). For each time point, a total score (0–30) was calculated. 3.3.2. Demographic characteristics Gender, age, and family structure were measured at the first wave with the following items: ‘What is your gender’ (male/female), ‘What is your birth date’ (month, day, year), and ‘Currently, with whom are you living’ (multiple categorical answer). Because most adolescents were living in a two-parent family (56.3%), this last item was dichotomized (living in a two-parent family or not). 3.3.3. Individual risk factors In order to determine if participants witnessed psychological and physical interparental intimate partner violence in childhood, eight items from a French revised version of the Revised Conflicts Tactics Scale (Lussier, 1997; Straus, Hamby, Boney-McCoy, & Sugarman, 1996) were used : four regarding the father’s violence towards the mother, and four concerning the mother’s violence towards the father. Participants had to rate each item on a 3-point Likert scale to report, if they had ever witnessed their father or their mother “insulting, swearing at, shouting at, or screaming at” or “pushed, shoved, slapped, twisted arm of, or throw something that could hurt” their other parent during their lifetime, or if they had ever witnessed their father/mother “threatening to hit or destroy something that belongs to” or “threatening with a knife or a gun, punching, beating or slamming against the wall” their mother/father. The instruction also specified that the terms “father” and “mother” included biological parents, adoptive parents or any other significant adult who played this role (e.g., step-mother, step-father, grandmother, grandfather, foster care parents…). Each scale (father towards mother violence and mother towards father violence) was dichotomized to differentiate adolescents who reported witnessing violence from those who did not, and was considered separately in analyses. In order to measure self-esteem, 5 adapted and translated items of the Self-Description Questionnaire (SDQ, Marsh & O’Neill, 1984) were used. On a 5-points Likert scale, participants rated to what extent they related to each statement (e.g., “Generally, I love myself as I am”; “When I do something, I make it well”). In the present sample, the measure of internal validity was good (H = .92). For the purpose of this study, adolescents were considered to have low self-esteem if their score on the SDQ was higher than the recommended clinical cut-off score (total score > 10) suggested by Statistics Canada (Garriguet, 2005). 3.3.4. Relational characteristics We asked participants about the duration of their relationship at Wave 3. They indicated in months how long they had been with their current boyfriend or girlfriend. To improve skewness this variable was transformed through a square root function. Attachment was then measured using the short French translation (Lafontaine et al., 2016) of the Experiences in Close Relationships questionnaire (ECR; Brennan, Clark, & Shaver, 1998). This validated 12-item questionnaire was answered with a 7-point Likert scale. Internal consistency was H = .84 for anxious scale and H = .81 for avoidant scale. 4
Child Abuse & Neglect 97 (2019) 104167
A. Lapierre, et al.
DV victimization for the last 12 months was assessed at Wave 1 with 8 items from the CADRI for psychological and physical DV (e.g. saying things to make him/her angry; make fun of him/her in front of the others; slapping or pulling his/her hair; pushed, shoved or shook him/her), and 9 items adapted from the Sexual Experiences Survey (SES; Koss et al., 2007) for sexual DV (e.g., Someone kissed, or rubbed up against the private areas of my body without my consent by : 1) using pressure or arguments; 2) using some physical force; 3) serving me alcohol drinks or giving me drugs;). Participants answered the items on a 4-point scale representing the frequency with which each event occurred (never, 1 to 2 times, 3 to 5 times, and 6 times or more). A total score was computed and then transformed through a square root function to improve skewness. The internal consistency of this scale was good (H = 0.90) in this sample.
3.3.5. Risky behaviors Alcohol and drug use in the last 12 months were measured with 3 items adapted from the Grille de dépistage de la consommation problématique d’alcool et de drogues chez les adolescents et les adolescentes (DEP-ADO, Landry, Tremblay, Guyon, Bergeron, & Brunelle, 2004). The items asked participant how often they used: a) cannabis (pot, hashish, etc.), b) other drugs (ecstasy, speed, cocaine, etc.), and c) how often they engaged in heavy drinking of 5 alcoholic beverages or more. Alcohol was measured on a 3-point (“never” to “5 times or more”) and drugs on a 6-point Likert scale (“never use” to “everyday use”). To make findings easily interpretable and because frequencies were highly skewed, each item was dichotomized to differentiate regular users from infrequent users. For heavy drinking, scores higher than 1.12 obtained the score of 1. For marijuana, participants with scores at one SD above the mean (score > 6) obtained the score of 1. For other drugs use, participants who reported having used these substances obtained 1. Antisocial behaviors were assessed with 6 items adapted from the National Longitudinal Survey of Children and Youth (Statistics Canada, 2007) on a 4-point scale (e.g., have you stayed out all night without permission, have you stolen something). Because of highly skewed distribution, the total score was dichotomized. Participants at 1 SD above the sample mean (score > 6) obtained the score of 1 and the others obtained the score of 0. Internal consistency was H = .816 for this scale.
3.3.6. Peers’ characteristics Antisocial behaviors of peers were assessed with five items from the National Longitudinal Survey of Children and Youth (Statistics Canada, 2007) and one item developed by our research team. Participants indicated on a 5-point Liker scale (none to every) how many of their close friends engaged in delinquent activities, for example “smoked cigarettes” (Statistics Canada, 2007) or “dropped out of school” (research team’s item). Continuous score, ranging from 0 to 24 was used in analysis. Internal consistency for this measure was H = .898. Two items adapted from Sears and Byers (2010) were used to assess physical/sexual DV victimization of peers (How many of their close friends have: “been hit or physically hurt by their boyfriend/girlfriend?” or “been constraint with manipulation, blackmail, or physical force by their boyfriend/girlfriend to have a sexual relationship”). Responses were indicated on a 5-point Liker scale ranging from none to every. Internal consistency was H = .816, and a continuous score, ranging from 0 to 8 was used in the analysis. Because of its skewness this variable had to be transformed using a square root computation.
3.4. Analyses First, descriptive statistics were used to summarize data on each variable. Demographic characteristics, adolescent victimization experiences in their relationships, at-risk behaviors, as well as peers’ characteristics are presented in Table 1 Analyses were also performed to examine how the original sample of adolescents who reported being in a dating relationship at Wave 1 differed from the remaining sample at Wave 3, and to explore if variables of interest predicted missingness at Waves 2 and 3. To address the first objective, we estimated the means and prevalence rates of psychological DV perpetration for all measurement points. To examine trajectories of psychological DV perpetration over the three-time points, a semiparametric group-based modeling for longitudinal data implemented in PROC TRAJ, a SAS macro developed by Nagin and colleagues (Nagin, Jones, Passos, & Tremblay, 2016; Jones & Nagin, 2007), was used. This analytic approach assumes that latent subgroups can be identified in the sample. The optimal number of subgroups is determined by comparing the different estimated models (e.g., 2, 3 or 4 trajectories) and their fit indices to determine the one that best fits the data. As suggested by McCoach and Kaniskan (2010), given that only three-time points are considered, trajectories will necessarily be linear (or flat) because only two random effect can be estimated. Therefore, no other trajectory form (such as quadratic or cubic) will be examined. In order to foster the convergence of models, each time point was conceptually recoded into the period to which the participants were referring to when they were answering questions. At Wave 1, participants had to report their psychologically violent behaviors over the past 12 months (coded -1.2), and at waves 2 and 3, over the past 6 months (respectively coded -0.6 and 0). This procedure was suggested by Singer (1998). Finally, the third objective was to predict group membership with different risk factors. After an optimal number of classes was determined, a multinomial logistic regression was used to examine how demographic, individual and relational characteristics, in addition to risky behaviors and characteristics of peers predicted group membership.
5
Child Abuse & Neglect 97 (2019) 104167
A. Lapierre, et al.
Table 1 Characteristics of the sample. Mean (SD) Demographic characteristics Male Age Two-parent family Relationship characteristics Length (months)3 Anxious attachment2 Avoidant attachment2 DV victimization1 Individual risk factors Low self-esteem (0-1)1−4 Father to mother IPV (0-1)1 Mother to father IPV (0-1)1 Risky behaviors Binge drinking(0-1)1−4 Marijuana problem (0-1)1−4 Other drug use (0-1)1 Antisocial behaviors (0-1)1−4 Peer characteristics Phys/sexual DV victimization1 Antisocial behaviors1
%
21.38 15.41 (0.85) 56.13 18.65 (31.10) 20.34 (6.95) 12.67 (6.05) 4.08 (7.17) 34.52 65.18 61.61 17.82 16.26 10.38 13.81 0.38 (0.78) 8.16 (4.54)
Notes. Percentages are presented for dichotomous variables and means for continuous variables. 1 Variables measured at W1, reference period = 12 months. 2 Variables measured at W2, reference period = 6 months. 3 Variable measured at W3, reference period = 6 months. 4 Variables are dichotomized at 1 SD above the mean.
4. Results 4.1. Descriptive 4.1.1. Characteristics of the sample Characteristics of the sample are presented in Table 1 The participants, predominantly girls, lived in a two-parent family, and reported having witnessed interparental intimate partner violence in their childhood. On average, they reported to be in a dating relationship with their partner for 19 months. Regarding risky behaviors at 1 SD above the sample mean, 18% of adolescents reported binge drinking, 16% marijuana problems, and 14% antisocial behaviors. Finally, 10% of adolescent reported to have ever used drugs other than marijuana. Chi-square and t-test analysis were performed with an alpha level of 0.05 to examine differences from adolescents who reported being a dating relationship at Wave 1, and participants remaining at Wave 3. While there were no significant differences for psychological DV perpetration, teenagers that participated at Wave 3 were more likely to be women (X2 (1) = 118.14, p < .001), to be in a dating relationship for a shorter amount of time at Wave 1 (t (848) = -2.57, p = .010), to have reported higher attachment avoidance (t (2735) = 4.94, p < .001). Also, they had higher rates of witnessing father to mother IPV (X2 (1) = 7.04, p = .008) and mother to father IPV (X2 (1) = 7.54, p = .006), binge drinking (X2 (1) = 17.50, p < .001), marijuana problem (X2 (1) = 27.45, p < .001), other drugs use (X2 (1) = 14.43, p < 0.001), and antisocial behaviors (X2 (1) = 13.47, p < .001). Finally, adolescents who participated in Wave 3 were older (t (1034) = 2.70, p = .007). However, there were no significant differences between Wave 1 and Wave 3 participants regarding their family situation, their DV victimization, their self-esteem, their attachment anxiety, their peers’ dating victimization, and their peers’ antisocial behaviors. We also performed binomial logistic regression to predict the data missingness at Waves 2 and 3, in two separated models. Among all the variables, only a few predicted the missingness. Being older (B = 0.23, OR = 1.25, p < .001), antisocial behaviors (B = -0.20, OR = .82, p = .036), and peers’ antisocial behaviors (B = .05, OR = 1.05, p < .001) predicted attrition from first to second wave, while attrition at Wave 3 was only predicted by being a male (B = .67, OR = 1.95, p < .001), and binge drinking (B = 0.59, OR = 1.80, p = .034).
4.1.2. Prevalence of psychological DV perpetration over time Even if the rates of psychological DV perpetration are similar over time (W1 = 51.9%; W2 = 48.6%; W3 = 50.6%), the SDs for psychological DV perpetration are above the means (W1, M = 2.04 [SD = 3.12]; W2, M = 1.56 [SD = 2.86]; W3, M = 1.71 [SD = 3.03]), suggesting a high variability in psychological DV reported. This supports the need of using a person-oriented approach to better understand the trajectories of psychological DV perpetration.
6
Child Abuse & Neglect 97 (2019) 104167
A. Lapierre, et al.
Table 2 Model Fit Results for Psychological DV. # classes
BIC
SABIC
AIC
L
1 2 3 4 5 6
−2514.61 −2456.53 −2435.35 −2428.68 −2424.81 −2433.46
−2512.96 −2453.23 −2430.41 −2422.09 −2416.57 −2423.57
−2506.80 −2440.91 −2411.93 −2397.45 −2385.77 −2386.61
−2503.80 −2434.91 −2402.93 −2385.45 −2370.77 −2368.61
Notes. n = 1347 for BIC and n = 449 for adjusted BIC. Models with 5 and 6 classes failed to converge.Perpetration Trajectories.
4.2. Identification of psychological DV perpetration trajectories Several models were tested with semiparametric group-based modeling, from 1 to 6 classes, to identify the one that best represented the data (presented in Table 2). Fit of the models were compared using the Bayesian Information Criterion (BIC, Schwarz, 1978), the sample-size-adjusted Bayesian information criterion (SABIC, Sclove, 1987), the Akaike Information Criterion (AIC, Akaike, 1987), and the log-likelihood value. A comparison of model fit indices suggested that models of 4 and 5 classes, with lower BIC, SABIC and AIC, and lower log likelihood were preferred. However, models composed of 5 classes failed to converge, suggesting instability. Therefore, the 4-class model was chosen to represent trajectories of psychological DV perpetration, providing a parsimonious model with the best fit to the data. The first trajectory was named Absence of perpetration (ß = -0.28). This trajectory comprised 30.8% of the sample and reported no psychological DV perpetration. The second trajectory was named Low perpetration (ß = -0.30) and contained 61.3% of the sample. The majority of adolescents in this trajectory reported isolated events of psychological DV perpetration across waves, with no variation across time (non-linear association, p > .05). Adolescents in this trajectory reported only isolated acts of psychological DV with a mean score of 2 on the CADRI. For example, they reported committing one type of psychological DV behaviors 3–5 times during the reference period (Wave 1 = 12 months; Waves 2 and 3 = 6 months). The third trajectory was named High descending (ß = -8.04, p < 0.001) and included 4.2% of the sample. At Wave 1, this subgroup reported a mean score of 11, but showed an important decrease over time. Finally, the fourth trajectory was named Moderate elevating and regrouped 3.8% (ß = 4.92, p < 0.001) of the sample. This subgroup of adolescents started with a mean score of approximately 6.25 and increased up to 8.75 at Wave 2 and then up to 12.0 at Wave 3. All group-based trajectories are illustrated in Fig. 1. 4.3. Risk factors predicting group membership in trajectories To better differentiate experiences of psychological DV perpetration in adolescents over time, we next examined risk factors predicting group membership when controlling for sociodemographic characteristics. The Absence of perpetration trajectory was used as a reference category in order to determine the distinguishing factors for adolescents who exerted psychological DV relative to those who did not. Overall, the model, presented in Table 3, explained 36.8% of the variance. 4.3.1. Demographic characteristics Group membership did not differ by gender or family structure. Regarding age, adolescents who were one year older were 3.87 times more likely to be classified in the High descending trajectory and 2.22 times more likely to be in the Moderate elevating trajectory, suggesting that the older you are, the more likely you are to engage in psychological DV. It is interesting to note that these two linear
Fig. 1. Trajectories of Psychological Dating Violence Perpetration. 7
Child Abuse & Neglect 97 (2019) 104167
A. Lapierre, et al.
Table 3 Risk Factors Predicting Group Membership in Trajectories of Psychological DV Perpetration.
Demographic characteristics Male Age Two-parent familyL Relationship characteristics Length3 Anxious attachment2 Avoidant attachment2 DV victimization1 Individual risk factors Low self-esteem (0-1)1−4 Father to mother IPV (0-1)1 Mother to father IPV (0-1)1 Risky behaviors Binge drinking(0-1)1−4 Marijuana problem (0-1)1−4 Other drug use (0-1)1 Antisocial behaviors (0-1)1−4 Peer characteristics DV victimization1 Antisocial behaviors1
Low
High descending
Moderate elevating
n = 292 (61.3%)
n = 14 (4.2%)
n = 15 (3.8%)
ß
OR
ß
OR
ß
OR
−0.15 0.28 −0.36
0.86 1.32t 0.70
−0.39 1.35 −0.47
0.68 3.87** 0.63
0.66 0.80 −0.59
1.94 2.22* 0.55
< -.01 0.04 −0.02 0.94
1.00 1.04† 0.99 2.56***
−0.17 0.03 0.05 2.18
0.85 1.03 1.05 8.85***
0.01 0.11 −0.02 1.19
1.01 1.12* 0.98 3.30***
0.46 −0.05 0.35
1.59 0.95 1.42
0.52 2.67 −0.85
1.69 14.40* 0.43
1.35 1.62 −1.11
3.88* 5.05† 0.33
−0.34 1.08 −0.03 −0.39
0.71 2.93* 0.98 0.68
−0.20 0.27 −1.17 1.16
0.82 1.32 0.31 3.19
−1.05 1.21 −0.33 1.85
0.35 3.36 0.72 6.36†
0.09 0.10
1.09 1.11**
−1.91 0.19
0.15* 1.21†
0.10 −0.04
1.10 0.97
Notes. Nagelkerke R2 = .37. 16 cases deleted due to missing values. 1 Variables measured at W1, reference period = 12 months. L Variable measure at W1, lifetime reference period. 2 Variables measured at W2, reference period = 6 months. 3 Variable measured at W3, reference period = 6 months. Reference category = Group 3 (absence of perpetration, n = 112, 30.8%). 4 Variables are dichotomized at 1 SD above the mean. * p < .05, ** p < .01 *** p < .001 †p < 0.10.
trajectories are in opposite directions, suggesting that as adolescents get older, some of them use more psychological DV (Moderate elevating trajectory) while others use less (High descending trajectory). 4.3.2. Relational characteristics Relationship length and an avoidant attachment style were not found to be associated with trajectory membership, whereas an anxious attachment style was associated with a higher risk of being in the Low and Moderate elevating trajectories. Being victimized in one’s dating relationship at Wave 1 was associated with increased risk of membership in all groups that use psychological DV, with a greater effect on the High descending trajectory (OR = 8.85). 4.3.3. Risky behaviors Youth that were regularly binge drinking showed no greater risk of being a member of the psychological DV groups. However, those with a regular use of marijuana were more likely to be in the Low violence trajectory (OR = 2.93), suggesting that people with high consumption of marijuana will use violence occasionally over time. Other drug use (e.g., cocaine, speed) did not predict group membership, but was overall reported by only a small proportion of the sample. Reporting antisocial behaviors was only marginally associated with engagement in the Moderate elevating trajectory. More specifically, adolescents who reported antisocial behavior at 1 SD (or more) above the sample mean were 6.36 times more likely to be included in that trajectory. 4.3.4. Peers’ characteristics High descending trajectory membership was predicted by having peers who experience DV and marginally predicted by peers’ antisocial behaviors. Peers’ antisocial behaviors also predicted membership to the Low psychological DV trajectory: a one-unit increase in peers’ antisocial behaviors raises the risk of being classified in Low violence trajectory of 11%. 4.4. Distinguishing between the descending and the elevating groups To better understand what differentiates adolescents in the Moderate elevating trajectory from adolescents in the High descending trajectory of psychological DV perpetration, the same risk factors were entered in a multinomial logistic regression model but this time the High descending trajectory was taken as a reference category. Results showed that none of the demographic characteristics distinguished the probability of membership in these two classes. With regards to relational characteristics, only DV victimization at the Wave 1 predicted a lower risk of classification in the Moderate elevating trajectory (OR = 0.37, p = .005). Self-esteem, witnessing IPV in childhood, alcohol, marijuana, and other drugs use, and antisocial behaviors failed to predict membership to these groups. Peers’ victimization in their dating relationship (OR = 7.42, p = .034) predicted higher probability of classification in the Moderate 8
Child Abuse & Neglect 97 (2019) 104167
A. Lapierre, et al.
elevating class. Finally, peers’ antisocial behaviors (OR = 0.80, p =0.075) did not predict a lower probability of classification in this group. 5. Discussion The current study aimed to identify psychological DV perpetration trajectories, as well as risk factors that could predict membership to these trajectories. First, we examined the prevalence of psychological DV perpetration over time. Rates presented are consistent with other studies that have assessed the lifetime and past 12-month prevalence of psychological DV perpetration (Hébert, Lapierre et al., 2018; Hébert, Moreau et al., 2018; Ybarra et al., 2016). 5.1. Findings on psychological DV perpetration trajectories and their implications Four psychological DV perpetration trajectories were identified in our sample: Absence of violence, Low violence, High descending, and Moderate elevating. Membership rates of violent trajectories (all trajectories other than Absence of violence) converged with results reported in previous studies suggesting similar rates of psychological DV perpetration (approximately 30%, Hébert, Lapierre et al., 2018; Hébert, Moreau et al., 2018; Ybarra et al., 2016). However, it appears that, fortunately, for most adolescents (61.3%), psychological DV perpetration occurs quite infrequently (low violence trajectory) with one or two events reported in a 6-month period and can thus be defined as situational. When the Absence of violence trajectory and the Low violence trajectory are clustered together, only 8% of the sample reported perpetrating psychological DV at a more frequent rate and are defined as being in Moderate elevating or High descending trajectories. The use of violence for adolescents in these two trajectories could be seen as a recurrent strategy aimed at resolving conflict and generating interactions with a dating partner. Results suggest that psychological DV as a recurrent conflict solving strategy is less prevalent than the situational psychological DV seen in the Low violence trajectory yet potentially more damageable. Clustering the Absence of violence and the Low violence trajectories together is not indicated, however, as these two trajectories differ significantly in risk factors. For example, teens who rarely engage in psychological DV have specific characteristics, such as marijuana use and antisocial behaviors that distinguish them from those who do not perpetrate psychological DV. These findings should be taken into account when studying psychological DV because perpetration may occur in different contexts, and involve individuals with various characteristics, which in turn may generate different consequences for relationship quality and both partners’ well-being. To our knowledge, no other study had yet examined psychological DV perpetration with a group modeling approach. However, in studies on other forms of DV, similar trajectories have been found (Chang et al., 2015; Foshee et al., 2013; Orpinas et al., 2017). For example, low and stable trajectories, increasing trajectories, and decreasing trajectories have already been observed for physical violence (Chang et al., 2015). Chang et al. (2015) found that as adolescents get older, they perpetrate less physical dating violence. However, we did not observe this pattern for psychological DV perpetration. Actually, our results suggested that being older was associated similarly with the High descending trajectory and the Moderate elevating trajectory. A possible explanation for this finding is the developmental increase of inhibition capacity in adolescence (Tamm, Menon, & Reiss, 2002). In fact, although a higher inhibition capacity may result in less physical dating violence, this capacity may not be sufficient to result in a decrease in psychological DV. Psychological DV is the most accepted form of violence in youths (Courtain & Glowacz, 2018), and therefore its reduction over time may require a greater ability than the withholding of impulsive physical behaviors. In contrast, a higher inhibition capacity may result in less physical dating violence given that physical violence is highly socially repressed. This hypothesis should be explored in further studies. 5.2. Implications of identifying risk factors predicting trajectory membership 5.2.1. Individual and relational characteristics According to our results, the majority of risk factors that had been identified in the DV perpetration literature to date (age, anxious attachment, witnessing father to mother IPV, DV victimization, marijuana use, antisocial behaviors and peers’ characteristics) seem to also play a role when examining psychological DV perpetration specifically. Exposure to interparental intimate partner violence in childhood increased the risk of adolescent psychological DV perpetration, suggesting in accordance with Bandura’s (1971) social learning theory that interpersonal communication skills in dating relationships are learned early in childhood and may be repeated over the course of adolescence. Our results showed that low self-esteem was the best predictor for being involved in the increasing trajectory of perpetration, which confirms our hypothesis based on the general violence theory (Berkowitz, 1989, 1993, 2001). Adolescents with low self-esteem may be prone to feel personally threatened and may engage in a psychological DV strategy as a response to this threat. Observed elevation in the Moderate elevating trajectory suggests that low self-esteem may prevent adolescents to develop efficient strategies to resolve conflict, while other adolescents, with high self-esteem may come to develop efficient communication skills and conflict resolution strategies as they interact with their partner over time. Self-esteem thus appears to be a key risk factor in the variation of psychological DV over time and should further be explored. Examining the two trajectories in which psychological DV perpetration was reported (Moderate elevating and High descending trajectories); attachment did not reveal to be a major risk factor. Anxious attachment predicted a slightly higher risk of being part of the Moderate elevating group while avoidant attachment did not predict group membership. Our results suggest that insecure attachment may not play the same role when examining psychological DV specifically in contrast to findings of past studies considering 9
Child Abuse & Neglect 97 (2019) 104167
A. Lapierre, et al.
all forms of DV clustered together (e.g., Ulloa et al., 2014). In regard to alcohol use, our results differ from those of Shorey, Stuart, Moore, and McNulty (2014) which is, to our knowledge, the only study which looked at psychological DV perpetration distinctively. The authors found effects of alcohol use, number of drinks, and heavy drinking on psychological perpetration while in the current study, binge drinking showed no effect. Our diverging results may be explained by the fact that in Shorey et al. (2014)’ study, alcohol use and psychological DV were measured with daily diary entries, whereas in our study, binge drinking was only measured at Wave 1 as an indicator of risky behavior. Therefore, in Shorey et al. (2014), alcohol disinhibition effect on DV may have been captured while it was not in the current study. Thus, previous binge drinking (measured as a risky behavior) was found to have no effect, while binge drinking on the day where partners interact is associated with psychological DV perpetration. Further daily diary studies should examine psychological DV perpetration to better understand what are the key factors involved when violence is used. Marijuana use predicted being in the Low violence group. To our knowledge, no other study examined psychological DV perpetration and drug use. Finally, our result also showed that DV victimization is the only factor that distinguishes between the Absence of violence trajectory and all the other trajectories, highlighting the possible involvement of mutuality in psychological DV. This result is consistent with previous research on general DV which showed that mutuality is present in most cases of DV (Chiodo et al., 2011; Fernández-González, Calvete, & Orue, 2017). Psychological DV may therefore be perpetrated as a defense against victimization, but then, who is the first to throw a dig and how does this dynamic begin? The question of psychological DV mutuality over time in adolescence should be further explored to better understand if victimization and perpetration patterns for a same individual co-vary in time. Daily diary studies and multilevel modeling examining longitudinal data in this fields may help to answer these questions. It must be noted that, in our 12-month period study, some adolescents might have experienced more than one relationship. Therefore, given that it is difficult to know if the victimization was sustained in the same relationship in which the psychological DV was used. These results should be interpreted with caution. 5.2.2. Peers’ characteristics Concerning characteristics of peers, antisocial behaviors increased the risk of being in the Low violence trajectory in accordance with previous literature on DV (see Garthe et al., 2017 for a meta-analysis). Physical/sexual DV victimization of peers lowered the risk of being part of the High descending trajectory, which suggests that having a victimized friend may sensitize youth to deleterious effect of perpetration and prevent them from using it as a strategy to resolve conflicts. This result diverged from studies on overall DV (e.g., Foshee et al., 2001) in which frequenting friend victims was shown to increase the risk of perpetration. It may be explained by the fact that youth who used psychological DV may be more sensitive to acquaintances’ victimization experiences than those who use other forms of DV. When perpetrating physical or psychological DV, youth may normalize violence and therefore, having a friend victim is not a protective factor anymore. Finally, it must also be noted that self-esteem, marijuana and antisocial behaviors failed at predicting membership to moderate elevating trajectory but nevertheless showed odd ratios of approximately 2. Thus, it is possible that an increase in power would have produced significant relationships. 5.3. Limits of the current study This study presents some limitations. First, conclusions and hypotheses surrounding mutuality of DV in relationships should be taken with caution. Indeed, from one time point to another, it was impossible to know if adolescents experience a change in partner. Thus, DV victimization from a partner at Wave 1 may not have been experienced in the same relationship context as psychological DV perpetration at waves 2 or 3. Second, it is possible that prevalence and effect sizes estimated for antisocial behaviors, peers’ antisocial behaviors and the different forms of victimization and perpetration were influenced by the fact that they were measured using brief instruments. In addition, exposure to IPV was measured by asking participants if they had ever witnessed interparental violence, which implies that only eye witnessing has been taken into account and that other forms of IPV victimization such as hearing, being conscious of, or having been told, that mother/father exert violence and control toward the other parent were omitted. Third, our study only examined certain forms of psychological DV as we only included the items from the measure with the highest factor loading (Wolfe et al., 2001). The limited number of item used may have failed to capture other psychological DV perpetration behavior and this may have resulted in an overinflated rate for group membership in the absence of psychological violence and in the low psychological violence trajectories. Fourth, the results are based on self-report measures, which may have induced a desirability bias. However, our research team implemented procedures to reduce this effect, such as the assurance of anonymity. Finally, attrition bias is a potential problem in this study. While our subsample was derived from a provincial representative sample of adolescents, only those who reported being in a dating relationship for all three waves of the study were selected. Therefore, it is important to note that some adolescents are not included in later waves and thus, not retained in the study, because the status of their dating relationship had changed over the course of the study (e.g., they had broken up). Nevertheless, some attrition could also reflect participants’ disengagement and may have been higher between waves 2 and 3 because of changes in the data collection method (i.e., by email instead of an in-class paper form). Furthermore, adolescents who participated in all three waves presented some differences on risk factors from those who did not. Because of these patterns of attrition some of the findings in this study should be interpreted carefully and additional research will be needed to draw a clearer portrait of the factors that increase the probability of membership in the different psychological DV trajectories. Despite the significant attrition rate over time, a main strength of this study remains its longitudinal design. The fact that psychological DV, did not differ significantly between those who remained in the study at Wave 3 and those who did not, suggest that it is unlikely that the attrition observed biased our trajectories substantially. Therefore, the data in the current study offers some important insights and contributions to the field by identifying 10
Child Abuse & Neglect 97 (2019) 104167
A. Lapierre, et al.
trajectories of psychological DV perpetration and frequencies of these patterns in adolescents. 5.4. Implications for practice and future research Psychological DV perpetration is a common form of violence, used by one person out of three (Ybarra et al., 2016), but associated with major consequences, particularly when it is experienced in addition to physical violence (Eshelman & Levendosky, 2012). Our results shed light on the differential experiences of psychological DV perpetration over time, and demonstrated that risk factors predict membership to specific perpetration trajectories. Our results revealed the existence of four distinct profiles of adolescent perpetrators, and that individual, relational, and behavioral characteristics contributed to the prediction of the trajectory group membership. Prevention programs should consider the various profiles of perpetrators according to their characteristics and the way they use psychological DV over time. Depending on the profiles, different factors and contexts can lead adolescents to perpetrate dating violence, and all situations should not be similarly addressed by prevention efforts. For example, adolescents who witnessed IPV were more likely to be part of the High descending trajectory. Because of their frequent but diminishing use of psychological DV as a strategy to resolve conflicts, it would be particularly helpful for them to be part of an early intervention program to foster efficient communication skills in an interpersonal setting. In return, adolescents with anxious attachment were more likely to be part of the Moderate elevating trajectory and could benefit from an intensive program (to interrupt the trajectory), in which strategies to better communicate needs and relational fears, and abilities to expose vulnerabilities and share intimacy can be learned and developed. Finally, adolescents with marijuana consumption problems were more likely to use psychological DV in a low and stable trajectory. Because their use of violence can be considered situational (given the low frequency), these adolescents could benefit from an educational program oriented on attitudes about violence, on personal strength, as well as on available alternatives to resolve conflict. Moreover, in future research and prevention program development, attention should be paid to peers’ characteristics, and dating victimization because these were the only two factors in our model that distinguished between being at risk for belonging to the High descending and the Moderate elevating trajectories. Furthermore, dating victimization was the only factor differentiating the absence of psychological DV from the presence of psychological DV perpetration, suggesting that victims of DV should receive support to identify nonviolent strategies. Therefore, to interrupt the trajectory of psychological DV perpetration, components should be integrated in teen programs to address communication skills and conflict resolution strategies, as well as to address peer pressure. 6. Conclusion Psychological DV perpetration has rarely been studied despite it being a prevalent issue in adolescence. Our study provides a longitudinal perspective on psychological DV perpetration and supports the need for a person-oriented approach to study violence in youth. Risk factors such as age, anxious attachment, low self-esteem, witnessing father to mother IPV, DV victimization, marijuana problems, physical/sexual DV victimization of peers, and peers’ antisocial behaviors successfully predicted membership to trajectories, with greater effects for DV victimization, witnessing IPV, and self-esteem. These results emphasize the importance of further examining self-esteem and violence mutuality in future research, and of developing programs for vulnerable youth according to their profiles. Violence use is not a stable behavior across adolescence and, before it becomes chronic, it should be studied considering variations over time, as well as according to developmental stages, with longitudinal design over several years. Adolescents should be supported throughout their development of interpersonal skills to better learn how to express their needs and feelings in a relationship, as well as to be more receptive to their partners’ ones. Hopefully, policy makers and social health agencies will orient their efforts in preventing dating violence earlier, to use programs adapted for different perpetrator profiles, and finally to disrupt trajectories of violence, particularly the elevating one, before it becomes a crystallized way of interacting with an intimate partner. Funding This work was supported by the Canadian Institutes of Health Research (grant number 103944). Declaration of Competing Interest The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Acknowledgments The authors wish to thank the teenagers and the school personnel that participated in the Youths’ Romantic Relationship Project. Our thanks are also extended to Catherine Moreau for project coordination. References Akaike, H. (1987). Factor analysis and AIC. In Author (Ed.). Selected papers of Hirotugu Akaike (pp. 371–386). New-York, NY: Springer. Bandura, A. (1971). Social learning theory. Morristown, N.J: General Learning Press. Barker, E. D., Séguin, J. R., White, H. R., Bates, M. E., Lacourse, E., Carbonneau, R., & Tremblay, R. E. (2007). Developmental trajectories of male physical violence and
11
Child Abuse & Neglect 97 (2019) 104167
A. Lapierre, et al.
theft: Relations to neurocognitive performance. Archives of General Psychiatry, 64(5), 592–599. https://doi.org/10.1001/archpsyc.64.5.592. Berkowitz, L. (1989). Frustration-aggression hypothesis: Examination and reformulation. Psychological bulletin. US: American Psychological Associationhttps://doi.org/10. 1037/0033-2909.106.1.59. Berkowitz, L. (1993). Aggression: Its causes, consequences, and control. Aggression: Its causes, consequences, and control. New York, NY, England: Mcgraw-Hill Book Company. Berkowitz, L. (2001). On the formation and regulation of anger and aggression: A cognitive-neoassociationistic analysis. Emotions in social psychology: Essential readings. New York, NY, US: Psychology Press. Brennan, K. A., Clark, C. L., & Shaver, P. R. (1998). Self-report measurement of adult attachment: An integrative overview. Attachment theory and close relationships. New York, NY, US: Guilford Press46–76. Brooks-Russell, A., Foshee, V. A., & Ennett, S. T. (2013). Predictors of latent trajectory classes of physical dating violence victimization. Journal of Youth and Adolescence, 42(4), 566–580. https://doi.org/10.1007/s10964-012-9876-2. Bushman, B. J., & Baumeister, R. F. (1998). Threatened egotism, narcissism, self-esteem, and direct and displaced aggression: Does self-love or self-hate lead to violence? Journal of Personality and Social Psychology, 75(1), 219–229. Retrieved from https://psycnet.apa.org/buy/1998-04530-016. Chang, L.-Y., Foshee, V. A., Reyes, H. L. M., Ennett, S. T., & Halpern, C. T. (2015). Direct and indirect effects of neighborhood characteristics on the perpetration of dating violence across adolescence. Journal of Youth and Adolescence, 44(3), 727–744. https://doi.org/10.1007/s10964-014-0190-z. Chiodo, D., Crooks, C. V., Wolfe, D. A., McIsaac, C., Hughes, R., & Jaffe, P. G. (2011). Longitudinal prediction and concurrent functioning of adolescent girls demonstrating various profiles of dating violence and victimization. Prevention Science: The Official Journal of the Society for Prevention Research, 13(4), 350–359. https://doi.org/10.1007/s11121-011-0236-3. Connolly, J. A., & McIsaac, C. (2009). Romantic relationships in adolescence. October 30Handbook of adolescent psychologyhttps://doi.org/10.1002/9780470479193. adlpsy002005. Côté, S., Vaillancourt, T., LeBlanc, J. C., Nagin, D. S., & Tremblay, R. E. (2006). The development of physical aggression from toddlerhood to pre-adolescence: A nation wide longitudinal study of Canadian children. Journal of Abnormal Child Psychology, 34(1), 68–82. https://doi.org/10.1007/s10802-005-9001-z. Courtain, A., & Glowacz, F. (2018). Exploration of dating violence and related attitudes among adolescents and emerging adults. Journal of Interpersonal Violence. https://doi.org/10.1177/0886260518770185 Advance online publication. Eshelman, L., & Levendosky, A. A. (2012). Dating violence: Mental health consequences based on type of abuse. Violence and Victims, 27(2), 215–228. Exner-Cortens, D., Eckenrode, J., & Rothman, E. (2013). Longitudinal associations between teen dating violence victimization and adverse health outcomes. Pediatrics, 131(1), 71–78. https://doi.org/10.1542/peds.2012-1029. Fernández-González, L., Calvete, E., & Orue, I. (2017). Adolescent dating violence stability and mutuality: A 4-year longitudinal study. Journal of Interpersonal Violence. https://doi.org/10.1177/0886260517699953 Advance online publication. Foshee, V. A., Benefield, T. S., Reyes, H. L. M., Ennett, S. T., Faris, R., Chang, L.-Y., ... Suchindran, C. M. (2013). The peer context and the development of the perpetration of adolescent dating violence. Journal of Youth and Adolescence, 42(4), 471–486. https://doi.org/10.1007/s10964-013-9915-7. Foshee, V. A., Linder, F., MacDougall, J. E., & Bangdiwala, S. (2001). Gender differences in the longitudinal predictors of adolescent dating violence. Preventive Medicine, 32(2), 128–141. https://doi.org/10.1006/pmed.2000.0793. Garriguet, D. (2005). Relations sexuelles précoces. (Research Report No. 82-003-XIF)Retrieved from Statistic Canada websitehttp://publications.gc.ca/Collection-R/ Statcan/82-003-XIF/0030482-003-XIF.pdf. Garthe, R. C., Sullivan, T. N., & McDaniel, M. A. (2017). A meta-analytic review of peer risk factors and adolescent dating violence. Psychology of Violence, 7(1), 45–57. https://doi.org/10.1037/vio0000040. Gover, A. R., Jennings, W. G., Tomsich, E. A., Park, M., & Rennison, C. M. (2011). The influence of childhood maltreatment and self-control on dating violence: A comparison of college students in the United States and South Korea. Violence and Victims, 26(3), 296–318. https://doi.org/10.1891/0886-6708.26.3.296. Harper, M. S., & Welsh, D. P. (2007). Keeping quiet: Self-silencing and its association with relational and individual functioning among adolescent romantic couples. Journal of Social and Personal Relationships, 24(1), 99–116. https://doi.org/10.1177/0265407507072601. Haynie, D. L., Farhat, T., Brooks-Russell, A., Wang, J., Barbieri, B., & Iannotti, R. J. (2013). Dating violence perpetration and victimization among U.S. adolescents: Prevalence, patterns, and associations with health complaints and substance use. The Journal of Adolescent Health: Official Publication of the Society for Adolescent Medicine, 53(2), 194–201. https://doi.org/10.1016/j.jadohealth.2013.02.008. Hébert, M., Blais, M., & Lavoie, F. (2017). Prevalence of teen dating victimization among a representative sample of high school students in Quebec. International Journal of Clinical and Health Psychology, 17(3), 225–233. https://doi.org/10.1016/j.ijchp.2017.06.001. Hébert, M., Lapierre, A., Lavoie, F., Fernet, M., & Blais, M. (2018). La violence dans les relations amoureuses des jeunes [violence in youths’ relationships]. In J. Laforest, P. Maurice, & L. M. Bouchard (Eds.). Rapport québécois sur la violence et la santé (pp. 118–129). Montreal: Institut national de santé publique du Québec. Hébert, M., Moreau, C., Blais, M., Oussaïd, E., & Lavoie, F. (2018). A three-steph gendered latent class analysis on dating victimization profiles. Psychology of Violence, 9(5), 504–516. https://doi.org/10.1037/vio0000225 Advance online publication. Jones, B. L., & Nagin, D. S. (2007). Advances in group-based trajectory modeling and an SAS procedure for estimating them. Sociological Methods & Research, 35(4), 542–571. https://doi.org/10.1177/0049124106292364. Koss, M. P., Abbey, A., Campbell, R., Cook, S., Norris, J., Testa, M., ... White, J. (2007). Revising the SES: A collaborative process to improve assessment of sexual aggression and victimization. Psychology of Women Quarterly, 31(4), 357–370. https://doi.org/10.1111/j.1471-6402.2007.00385.x. Lafontaine, M. F., Brassard, A., Lussier, Y., Valois, P., Shaver, P. R., & Johnson, S. M. (2016). Selecting the best items for a short-form of the experiences in close relationships questionnaire. European Journal of Psychological Assessment, 32(2), 140–154. https://doi.org/10.1027/1015-5759/a000243. Landry, M., Tremblay, J., Guyon, L., Bergeron, J., & Brunelle, N. (2004). La Grille de dépistage de la consommation problématique d’alcool et de drogues chez les adolescents et les adolescentes (DEP-ADO): développement et qualités psychométriques. Drogues, Santé et Société, 3(1), 19–37. Lavoie, F., Hotton-Paquet, V., Laprise, S., & Joyal Lacerte, F. (2009). ViRAJ: Programme de prévention de la violence dans les relations amoureuses chez les jeunes et de promotion des relations égalitaires. Deuxième édition révisée. Québec: Université Laval. Leen, E., Sorbring, E., Mawer, M., Holdsworth, E., Helsing, B., & Bowen, E. (2013). Prevalence, dynamic risk factors and the efficacy of primary interventions for adolescent dating violence: An international review. Aggression and Violent Behavior, 18(1), 159–174. https://doi.org/10.1016/J.AVB.2012.11.015. Linder, J. R., & Collins, W. A. (2005). Parent and peer predictors of physical aggression and conflict management in romantic relationships in early adulthood. Journal of Family Psychology. https://doi.org/10.1037/0893-3200.19.2.252 Linder, Jennifer Ruh: Department of Psychology, Linfield College, Unit #A575, 900 Southeast Baker Street, McMinnville, OR, US, 97128, jlinder@linfield.edu: American Psychological Association. Lussier, Y. (1997). Questionnaire sur la résolution des conflits conjugaux. Trois-Rivières: Université Du Québec à Trois-Rivières. Marsh, H. W., & O’Neill, R. (1984). Self description questionnaire III: The construct validity of multidimensional self-concept ratings by late adolescents. Journal of Educational Measurement, 21(2), 153–174. Retrieved from http://www.jstor.org/stable/1434540. McCoach, D. B., & Kaniskan, B. (2010). Using time-varying covariates in multilevel growth models. Frontiers in Psychology, 1(17), 1–12. https://doi.org/10.3389/fpsyg. 2010.00017. McNeish, D. (2017). Thanks coefficient alpha, we’ll take it from here. Psychological Methods, 23(3), 412–433. https://doi.org/10.1037/met0000144. Miller, S., Williams, J., Cutbush, S., Gibbs, D., Clinton-Sherrod, M., & Jones, S. (2013). Dating violence, bullying, and sexual harassment: Longitudinal profiles and transitions over time. Journal of Youth and Adolescence, 42(4), 607–618. https://doi.org/10.1007/s10964-013-9914-8. Nagin, D. S., Jones, B. L., Passos, V. L., & Tremblay, R. E. (2016). Group-based multi-trajectory modeling. Statistical Methods in Medical Research, 27(7), 2015–2023. https://doi.org/10.1177/0962280216673085. O’Donnell, L., Stueve, A., Myint-U, A., Duran, R., Agronick, G., & Wilson-Simmons, R. (2006). Middle school aggression and subsequent intimate partner physical violence. Journal of Youth and Adolescence, 35(5), 693–703. https://doi.org/10.1007/s10964-006-9086-x. O’Leary, K. D., Tintle, N., & et Bromet, E. (2014). Risk factors for physical violence against partners in the U.S. Psychology of Violence, 4(1), 65–77. https://doi.org/10. 1037/a0034537.
12
Child Abuse & Neglect 97 (2019) 104167
A. Lapierre, et al.
Offenhauer, P., & Buchalter, A. (2011). Teen dating violence: A literature review and annotated bibliography (Research Report No 235863)Retrieved from the National Criminal Justice Reference Service websitehttps://www.ncjrs.gov/pdffiles1/nij/grants/235368.pdf. Orpinas, P., Hsieh, H.-L., Song, X., Holland, K., & Nahapetyan, L. (2013). Trajectories of physical dating violence from middle to high school: Association with relationship quality and acceptability of aggression. Journal of Youth and Adolescence, 42(4), 551–565. https://doi.org/10.1007/s10964-012-9881-5. Orpinas, P., Nahapetyan, L., & Truszczynski, N. (2017). Low and increasing trajectories of perpetration of physical dating violence: 7-Year associations with suicidal ideation, weapons, and substance use. Journal of Youth and Adolescence, 46(5), 970–981. https://doi.org/10.1007/s10964-017-0630-7. Ostrowsky, M. K. (2010). Are violent people more likely to have low self-esteem or high self-esteem? Aggression and Violent Behavior, 15(1), 69–75. https://doi.org/10. 1016/j.avb.2009.08.004. Sclove, S. L. (1987). Application of model-selection criteria to some problems in multivariate analysis. Psychometrika, 52(3), 333–343. https://doi.org/10.1007/ BF02294360. Sears, H. A., & Byers, E. S. (2010). Adolescent girls’ and boys’ experiences of psychologically, physically, and sexually aggressive behaviors in their dating relationships: Co-occurrence and emotional reaction. Journal of Aggression, Maltreatment & Trauma, 19(5), 517–539. https://doi.org/10.1080/10926771.2010. 495035. Schwarz, G. (1978). Estimating the dimension of a model. The Annals of Statistics, 6(2), 461–464. https://projecteuclid.org/download/pdf_1/euclid.aos/1176344136. Shorey, R. C., Rhatigan, D. L., Fite, P. J., & Stuart, G. L. (2011). Dating violence victimization and alcohol problems: An examination of the stress-buffering hypothesis for perceived support. Partner Abuse, 2(1), 31–45. https://doi.org/10.1891/1946-6560.2.1.31. Shorey, R. C., Stuart, G. L., Moore, T. M., & McNulty, J. K. (2014). The temporal relationship between alcohol, marijuana, angry affect, and dating violence perpetration: A daily diary study with female college students. Psychology of Addictive Behaviors: Journal of the Society of Psychologists in Addictive Behaviors, 28(2), 516–523. https://doi.org/10.1037/a0034648. Singer, J. D. (1998). Using SAS PROC MIXED to fit multilevel models, hierarchical models, and individual growth models. Journal of Educational and Behavioral Statistics, 23(4), 323–355. https://doi.org/10.3102/10769986023004323. Spencer, R. A., Renner, L. M., & Clark, C. J. (2016). Patterns of dating violence perpetration and victimization in US young adult males and females. Journal of Interpersonal Violence, 31(5), 2576–2597. Statistics Canada (2007). National longitudinal survey of children and youth: Cycle 8 survey instruments, 2008/2009. Book 1 – Contact, household and exit, parent, child and youth components. Ottawa, ON: Canada. Sterba, S. K., & Bauer, D. J. (2010). Matching method with theory in person-oriented developmental psychopathology research. Development and Psychopathology, 22(2), 239–254. https://doi.org/10.1017/S0954579410000015. Straus, M. A., Hamby, S. L., Boney-McCoy, S., & Sugarman, D. B. (1996). The revised conflict tactics scales (CTS2): Development and preliminary psychometric data. Journal of Family Issues, 17(3), 283–316. https://doi.org/10.1177/019251396017003001. Swartout, K. M., Cook, S. L., & White, J. W. (2012). Trajectories of intimate partner violence victimization. The Western Journal of Emergency Medicine, 13(3), 272–277. https://doi.org/10.5811/westjem.2012.3.11788. Tamm, L., Menon, V., & Reiss, A. L. (2002). Maturation of brain function associated with response inhibition. Journal of the American Academy of Child and Adolescent Psychiatry, 41(10), 1231–1238. https://doi.org/10.1097/00004583-200210000-00013. Temple, J. R., Shorey, R. C., Fite, P., Stuart, G. L., & Le, V. D. (2013). Substance use as a longitudinal predictor of the perpetration of teen dating violence. Journal of Youth and Adolescence, 42(4), 596–606. https://doi.org/10.1007/s10964-012-9877-1. Tracy, J. L., & Robins, R. W. (2003). "Death of a (Narcissistic) Salesman:" An Integrative Model of Fragile Self-Esteem. Psychological Inquiry, 14(1), 57–62. Retrieved from https://www.jstor.org/stable/1449043. Ulloa, E. C., Martinez-Arango, N., & Hokoda, A. (2014). Attachment anxiety, depressive symptoms, and adolescent dating violence perpetration: A longitudinal mediation analysis. Journal of Aggression, Maltreatment & Trauma, 23(6), 652–669. https://doi.org/10.1080/10926771.2014.920452. VanDellen, M. R., Campbell, W. K., Hoyle, R. H., & Bradfield, E. K. (2010). Compensating, resisting, and breaking: A meta-analytic examination of reactions to selfesteem threat. Personality and Social Psychology Review, 15(1), 51–74. https://doi.org/10.1177/1088868310372950. Wekerle, C., Leung, E., Wall, A. M., MacMillan, H., Boyle, M., Trocme, N., & Waechter, R. (2009). The contribution of childhood emotional abuse to teen dating violence among child protective services-involved youth. Child Abuse & Neglect, 33(1), 45–58. https://doi.org/10.1016/j.chiabu.2008.12.006. Wekerle, C., Wolfe, D. A., Hawkins, D. L., Pittman, A.-L., Glickman, A., & Lovald, B. E. (2001). The value and contribution of youth self-reported maltreatment history to adolescent dating violence: Testing a trauma mediational model. Development and Psychopathology, 13, 847–871. https://doi.org/10.1016/j.chiabu.2008.12. 006. Wincentak, K., Connolly, J., & Card, N. (2017). Teen dating violence: A meta-analytic review of prevalence rates. Psychology of Violence, 7(2), 224–241. https://doi. org/10.1037/a0040194. Wolfe, D. A., Scott, K., Reitzel-Jaffe, D., Wekerle, C., Grasley, C., & Straatman, A. L. (2001). Development and validation of the conflict in adolescent dating relationships inventory. Psychological Assessment, 13(2), 277–293. Ybarra, M. L., Espelage, D. L., Langhinrichsen-Rohling, J., Korchmaros, J. D., & Boyd, D. (2016). Lifetime prevalence rates and overlap of physical, psychological, and sexual dating abuse perpetration and victimization in a national sample of youth. Archives of Sexual Behavior, 45(5), 1083–1099. https://doi.org/10.1007/s10508016-0748-9.
13