Poster #95 - A Latent Class Analysis of Cyber Dating Violence among Adolescents

• In Event: Poster Session 02

Abstract

Although smartphones have become the primary form of communication within adolescent dating relationships, they have also created greater opportunities for control and monitoring within these relationships (Baker & Carreno, 2016). These experiences have led to a new form of dating violence known as cyber dating violence (CDV) (Zweig et al., 2014). Research with adolescents has focused on identifying the prevalence and associations with CDV victimization and perpetration (Peskin et al., 2017). However, a reliance on variable-centered analyses in past research fails to identify groups or classes of adolescents who may be particularly vulnerable to distress (Thulin et al., 2021). The current study addresses this gap in the literature by using a person-centered approach to identify different classes of CDV victimization and perpetration, and to examine how these classes are associated with demographic characteristics and a range of feelings and emotions (e.g., subjective invulnerability, shame proneness, psychosomatic symptoms and emotional problems).

Participants included 591 adolescents (Mage = 15.84, 51.8% girls, 87.3% heterosexual) recruited by a Canadian research firm between September and October 2020. A subsample of 219 adolescents who reported being in a current relationship were used for all analyses. Participants were asked to self-report their demographics, experiences with CDV victimization and perpetration (9-items each; Borrajo et al., 2015), subjective invulnerability using a modified version of the adolescent invulnerability scale (6-items; Lapsley & Hill, 2010), psychosomatic symptoms (9-items; Haugland et al., 2001), shame proneness (15-items; Tangney et al., 1991), and emotional problems (8-items; Ma, 2007).

First, the optimal number of latent classes were determined based on model fit indices, class counts, and interpretability (see Table 1). As shown in Figure 1, a three-class solution fit the data best and resulted in a no/low CDV class (n = 153; 69.9%), a moderate CDV class (n = 45; 20.5%) and a high CDV class (n = 21; 9.6%). Interestingly, the moderate CDV class included higher rates of CDV victimization than perpetration. Second, associations between latent class membership and covariates of age and gender were examined. However, no significant differences in age or gender were found across the latent classes. Third, latent class membership was used to predict distal outcomes of subjective invulnerability, shame proneness, psychosomatic symptoms and emotional problems controlling for the covariates. The high CDV class (M = 4.92) reported greater feelings of subjective invulnerability (M = 3.47 for moderate CDV and M = 3.22 for no/low CDV), greater feelings of shame proneness (M = 4.50 for high CDV compared to M = 3.90 for moderate CDV and M = 3.22 for no/low CDV), greater psychosomatic symptoms (M = 4.41 for high CDV compared to M = 2.79 for moderate CDV and M = 2.28 for no/low CDV), and greater emotional problems (M = 3.52 for high CDV compared to M = 2.94 for moderate CDV and M = 2.28 for no/low CDV). The findings suggest that adolescents involved in unhealthy relationships characterized by engagement in and experiences of CDV are more likely to experience a range of negative feelings and emotions.

Authors

• Brett Holfeld, Grenfell Campus, Memorial University of Newfoundland
  Presenting Author

• Rachel Baitz, University of British Columbia (UBC)

• Alexa Martin-Storey, Universite de Sherbrooke

• Wendy Craig, Queen’s University