Cyber Dating Abuse Victimization: Links With Psychosocial Functioning

Participants and Procedure

As part of a larger study on technology and romantic relationships, data were collected from 278 (204 women; 74 men) participants recruited from a mid-sized Canadian university’s psychology undergraduate participant pool, where they viewed a description of the study online and could choose whether to sign up (see Table 1 for demographics). They were compensated for their participation by receiving 1% bonus credit in their psychology course. There were several inclusion criteria: (a) in a romantic relationship for at least 3 months (M = 18.7 months, SD = 15.8) that was not long distance or cohabiting, and (b) in the emerging adulthood age range, between 17 and 25 years (M = 20.5 years, SD = 1.9). A relationship duration of 3 months was required because participants were asked to reflect on the frequency of behaviors in their romantic relationships within the past 3 months as a part of the study. Being in a long-distance relationship or cohabiting with one’s partner would influence the amount that technology is used as a medium of communication, and therefore these participants were excluded. Participants completed anonymous online questionnaires about various forms of IPV victimization and their psychosocial functioning. This research was approved by the university’s research ethics board.

OPEN IN VIEWER

Table 1.

Demographic Information of Participants.

Characteristic	n	%
Relationship status
Dating one person, casually	24	8.6
Dating one person, committed	253	91.0
Dating multiple people, casually	1	0.4
Sexual orientation
Heterosexual/straight	243	87.4
Gay or lesbian	4	1.4
Bisexual	29	10.4
Other	1	0.4
No response	1	0.4
Living situation
In university residence	25	9.0
With parents or guardians	79	28.4
Off campus, with roommates	146	52.5
Off campus, alone	28	10.1
Met current partner through technology
Yes	45	16.2
No	233	83.8
Ethnic background a
White	229	82.4
African	5	1.7
Latino/Hispanic	10	3.4
Indigenous	1	0.4
Middle-Eastern	8	2.9
East Asian	32	11.5
South Asian	11	4.0
Caribbean	1	0.4
Other	5	1.7
Citizen of Canada
Yes	246	88.5
No	32	11.5

a

Individuals could select as many ethnic backgrounds as were applicable to them, and thus the total ethnic background selected sums to greater than the total number of participants, and the percentages do not add up to 100%.

Measures

CDA victimization

The Cyber Aggression in Relationships Scale (CARS; Watkins et al., 2016) was used to measure CDA. There is not yet a well-established measure of CDA victimization. The CARS was chosen after the authors reviewed the literature in this area because it appeared to have good coverage of all the forms of victimization that could occur through technology while using language that was general enough to allow it to be understandable over time as technologies quickly evolve (e.g., it avoided using specific language such as “Facebook”). The CARS victimization scale contains 19 items measuring victimization. Respondents rate the frequency with which they were victimized by CDA behaviors in the past 3 months ¹ on a scale of 0 (“this never happened”) to 6 (“more than 20 times in the past 3 months”). Total scores can range from 0 to 114 (see Table 2 for descriptive statistics). Example items include “My partner sent threatening or harassing messages to me via text or social media” and “My partner asked me online for sexual information about myself when I did not want to tell.” Because the CARS is not a well-established measure, we first conducted a factor analysis to determine if any statistically sound and theoretically coherent factor structure could be established. We began with a confirmatory factor analysis utilizing the same structure as was initially established. When this failed to demonstrate adequate model fit, we conducted several exploratory factor analyses. Through these analyses, no statistically sound and theoretically coherent factor structure could be established, and thus a full scale score was used ² .

OPEN IN VIEWER

Table 2.

Scale Descriptive Statistics and Reliability for Victimization and Psychosocial Variables.

	M	SD	Minimum	Maximum	t Test of Gender Differences	α
CDA victimization a	4.71	6.38	0	26	1.76	.80
Physical victimization b	1.03	2.41	0	10	2.68**	.74
Psychological victimization b	3.57	4.47	0	18	1.73	.71
Sexual victimization c	1.70	2.23	0	9	−0.64	.51
Stalking victimization d	0.47	1.28	0	7	−0.46	.82
Depression e	16.42	10.34	1	46	0.65	.92
Perceived stress f	26.92	6.84	9	46	−1.31	.83
Alcohol use g	6.90	4.91	0	22	2.35*	.87
Fear of partner h	25.96	9.32	19	58	4.61**	.90
Relationship satisfaction i	15.50	3.88	4	21	−2.14*	.85
Quality of life j	102.37	11.47	67	126	−1.17	.90
Perceived social support k	52.60	8.14	29	69	−1.76	.90
Post-traumatic stress l	16.74	13.85	0	59	0.69	.94

Note. CDA = cyber dating abuse; PTSD = post-traumatic stress disorder.

a

Cyber Aggression in Relationships Scale.

b

Conflict Tactics Scale–Revised.

c

Sexual Experiences Survey.

d

Stalking Victimization Questionnaire.

e

Center for Epidemiological Studies Depression Scale.

f

Perceived Stress Scale.

g

Alcohol Use Disorders Identification Test.

h

Fear of Partner Scale.

i

Couples Satisfaction Index.

j

World Health Organization Quality of Life Scale, Brief.

k

Social Support Appraisals.

l

PTSD Checklist for DSM-5.

*

p < .05. **p < .01.

Data Analysis

Significant Associations With CDA Victimization

The finding that CDA victimization is uniquely associated with alcohol use for both men and women corroborates several studies which have demonstrated the same association (Bennett et al., 2011; Van Ouytsel et al., 2016). Despite the consistency of this link, little theorizing has been done as to why this might be the case. One possible theory is that alcohol may be used as a coping mechanism for victimization (Schenk & Fremouw, 2012), and another is that being involved in abusive dating relationships is associated with a range of high-risk behaviors (Van Ouytsel et al., 2017). Although both explanations seem plausible, neither addresses why alcohol use may be uniquely related to CDA victimization. One potential explanation is that both technology use and alcohol use are frequent among this age group (White et al., 2006; Subrahmanyan, Reich, Waechter, & Espinoza, 2008), and thus CDA may be common, and the dominant means of coping with such victimization and other forms of distress may be alcohol use. It is theorized that, due to a lack of social cues, there is greater potential for ambiguity and misunderstandings in technological communication (Runions, Shapka, Dooley, & Modecki, 2013), and thus victims may be more able to explicitly avoid acknowledgment of this form of victimization. As alcohol use is an avoidant means of coping, it is possible that it is used more to manage experiences of CDA victimization rather than in-person victimization, as an attempt to minimize or avoid thinking about the possibly ambiguous experience. Further investigations should work to identify potential mechanisms of this association and incorporate the perspectives of victims on why this link exists.

CDA victimization uniquely predicted fear of partner among women only, along with a range of in-person victimization variables (sexual, physical, and psychological IPV victimization). It appears that IPV victimization overall is much more strongly related to fear of partner for women than men, with only physical IPV victimization being a significant predictor of fear of partner for men. Thus, it could be that CDA victimization is related to fear of partner for women only because experiencing any type of partner abuse in an intimate relationship elicits more fear among women than it does among men. It could also be that these technological behaviors are associated with later in-person victimization experiences (i.e., pressure to send sexual photos through technology may be followed by pressure to perform sexual acts in-person), and that these CDA victimization experiences may signify risk of future danger to women. For men, the lack of association between CDA victimization and fear of partner could be because the danger associated with in-person IPV is more imminent and widespread (i.e., could lead to immediate physical harm). Thus, greater fear may be associated with these victimization experiences as opposed to CDA victimization, which provides more physical distance between the victim and perpetrator. As mentioned, it is easier to misinterpret technological communication than face to face (Runions et al., 2013), and thus it may be possible that men view these experiences more lightly than in-person victimization experiences. Furthermore, there is evidence that adolescent boys view CDA victimization experiences more lightly than girls (Reed et al., 2017). Future research should further investigate these gender differences, with special attention paid to mechanisms that may explain this differential association.

Nonsignificant Associations With CDA Victimization

CDA was not a unique predictor of depression, stress, relationship satisfaction, quality of life, social support, or post-traumatic stress for men or women. Although one previous study found a link between CDA and depression (Zweig et al., 2014), this study failed to control for in-person victimization. In the cyberbullying literature, there is evidence that higher odds of depression occur only among those who both perpetrate and are victimized by cyberbullying, with no increased odds among those who are solely victims (Selkie, Kota, Chan, & Moreno, 2015). Although a different form of victimization, it suggests that technological victimization alone may not be sufficient to be associated with depressive symptoms. Thus, it could be that the previously identified association between CDA victimization and depression is better explained by concurrent in-person victimization experiences.

A lack of significant relationship between CDA victimization and other well-being domains, such as stress and relationship satisfaction, is consistent with previous research. There is prior evidence that no association exists between stress and CDA victimization (Carlson et al., 2015). Although no specific studies have investigated associations between CDA victimization and relationship satisfaction, even evidence of an association between in-person IPV victimization and relationship satisfaction is mixed (e.g., Amanor-Boadu et al., 2011; Kaura & Lohman, 2007). Thus, it is perhaps unsurprising that CDA victimization was not associated with outcomes such as stress and relationship satisfaction in the current study.

This was the first quantitative study examining the link between CDA victimization and post-traumatic stress symptoms, quality of life, and social support. Although there is a well-established relationship between post-traumatic stress symptoms and in-person IPV victimization (e.g., Amanor-Boadu et al., 2011), it is understandable that this association might not extend to CDA victimization. Although CDA may be distressing, the distance between the communicators is greater through technology, meaning the imminent threat to bodily harm is necessarily less when being victimized through technology as opposed to in person. As post-traumatic stress symptoms are predicated on an event that causes imminent threat to bodily integrity, it is likely that those experiences that are more able to engender such a sense of threat (i.e., in-person experiences) are more strongly related to post-traumatic stress symptoms.

There is less reason to expect differences in association between CDA and in-person victimization for outcomes such as quality of life and social support, although neither showed an association with CDA victimization when controlling for in-person victimization experiences in the current study. However, it is notable that the current study found less of an association between multiple forms of IPV victimization and both social support and quality of life than has been reported previously (Katerndahl et al., 2013; Leung et al., 2005). These differences could be due to the nature of the intimate relationships in the current study (i.e., dating as opposed to married or cohabiting) and the nature of the relationship violence experienced (i.e., mainly low levels of violence in the current study). Future research should seek to identify the impacts of CDA in different types of relationships, and in the presence of more severe forms of in-person victimization, to determine whether these factors influence the association with CDA victimization.

With the exception of alcohol use and women’s fear of their partners, the current study overwhelmingly suggests that CDA does not predict negative psychosocial functioning above what is predicted by in-person victimization. Overall, very few studies have examined the psychosocial associations of CDA, and even fewer have done so while controlling for in-person IPV victimization. Although individuals tend to self-report impacts of experiencing different forms of technological victimization when asked explicitly (Mishna et al., 2018), these associations may be better explained by, or in conjunction with, other forms of victimization. Furthermore, there are a number of personal characteristics, such as emotional intelligence, coping skills, and optimism, which are related to differential impacts of victimization (Jenaro, Flores, & Frías, 2018). Thus, it could be that some individuals are particularly adversely affected by CDA victimization, but that the associations depend on victim characteristics as well. Future research should seek to identify potential moderators of CDA associations to determine whether there is a population that is particularly affected by this form of abuse.

There could be several reasons that CDA victimization largely did not predict psychosocial functioning above in-person IPV victimization. First, in both the current study and others, the overlap between in-person and CDA victimization is substantial (Marganski & Melander, 2015; Temple et al., 2016). It may be that, amid a range of victimization experiences, in-person experiences are more impactful than technological. This is important, as there appears to be a veritable media panic about the impacts of technology on relationships (Jones, 2017; Rinaldi, 2015). The discrepancy between media perceptions of the negative impacts of technology and empirical evidence has been highlighted in the cyberbullying literature (Olweus, 2012). The results of the current study similarly suggest not only that CDA does not create many new victims but also that it may not lead to more negative psychosocial functioning outside of other victimization experiences. This does not mean that CDA should be disregarded or trivialized. The overlap between in-person and CDA victimization suggests that CDA is a commonly used abuse tactic within high-risk relationships. Instead, as opposed to CDA being treated as a unique and new form of victimization which may require novel forms of help, the current results suggest that CDA forms part of a constellation of abuse, and that it may be best to focus prevention and treatment efforts on all forms of abuse, rather than focusing solely on CDA and failing to address the many other forms of abuse that may co-occur in relationships. Similarly, research should conceptualize these forms of abuse concurrently (Taylor & Xia, 2018).

Second, definitional issues of what behaviors constitute CDA, and the related issue of consistent measurement of CDA, make it difficult to draw conclusions with too much certainty. It has been noted that a lack of consensus about the definition and measurement of CDA has plagued this area of research, at times with each individual study seemingly developing its own way to operationalize CDA (Brown & Hegarty, 2018). Although measure development is often done in consultation with those who have experienced CDA, there seems to be a discrepancy between what researchers perceive as abuse and what those who use technology in their relationships constitute as abuse. For example, there is evidence that at least some of the behaviors which researchers consider cyberstalking are considered normal forms of relationship maintenance, and at times could even be considered prosocial and appropriate among adolescents (Howard, Debnam, & Strausser, 2017; Lucero, Weisz, Smith-Darden, & Lucero, 2014). It is possible that these behaviors do, in fact, constitute abuse or relationship dysfunction, and that those who are experiencing them do not view them that way because abuse has been normalized within their relationship in the same way that in-person forms of abuse can be normalized (Temple et al., 2016). However, it is also possible that as technology develops, new norms are also developing, and that some behaviors that at one time seemed shockingly intrusive and damaging are now a normal, mundane, nonimpactful aspect of living in our technological world. Thus, it is possible that few associations were found in the current study because at least a subset of the items designed to measure CDA are simply not as impactful to relationships as researchers originally thought. If research accumulates demonstrating a lack of impact of victimization by specific technological behaviors, researchers must consider the possibility that if participants do not appear bothered by particular behaviors, and results suggest that they are not being adversely affected, then the behaviors themselves may not be abusive. Future researchers would do well to focus on clearly defining what is considered a normative, relationship maintaining behavior and what constitutes some form of abuse toward a partner through technology.

Finally, it could be that the current study combines individuals in multiple different forms of abusive relationships, thus making it difficult to differentiate psychosocial associations. Individuals in relationships characterized by intermittent violence in the context of an argument experience less severe impacts than those in relationships where violence is a means of dominance and control (Whitaker, 2013). As the current study did not distinguish between these motives, it is not possible to determine which pattern of violence participants are experiencing. It is possible that a subset of participants experienced more severe impacts of both in-person and CDA victimization due to the controlling nature of the violence. Future research should investigate coercive control to determine how this interacts both with forms of violence and with psychosocial outcomes.

Limitations

Several limitations to the current study must be noted. First, the cross-sectional nature of the study means that we are not able to determine whether CDA victimization causes the effects we discovered or whether certain aspects of psychosocial functioning (e.g., higher alcohol use) lead one to be more vulnerable to victimization. The use of cross-sectional data is a noted limitation of CDA research overall (Taylor & Xia, 2018), and thus future research should employ longitudinal methods to determine the nature of these relationships. Furthermore, the sample was recruited from a university undergraduate population, and was predominantly female. A lack of diversity in samples is another noted limitation of CDA research (Taylor & Xia, 2018). Similarly, in the current study, it is unclear how these results would generalize to emerging adults who do not attend university, as well as to students who choose not to study psychology at university. It is also unclear how such results would generalize to individuals who are in different age groups (e.g., adolescent, adulthood) or in different forms of romantic relationships (e.g., long distance, cohabiting). University samples are also limited in diversity in a number of ways, including but not limited to culture, ethnicity, and socioeconomic status. Thus, it is unclear to what extent the results of the current study would generalize to other populations. Although gender effects were tested in the current sample, there was not sufficient power to detect small or medium effects among men, and thus the effects reported for men should be treated with caution. The sample also comprised only those who are currently in a romantic relationship, and CDA victimization was reported only for the current relationship. This limitation may be especially important, as much of the danger of CDA could come after relationship dissolution, as technology allows for continued and potentially problematic contact after relationship dissolution has occurred. There are also limitations endemic to the methods of this study. The context and motives for various abusive behaviors were not collected in the current study. Failure to account for this difference in motives (i.e., the result of an outburst of anger versus part of a larger pattern of control) could have masked impacts that are not uniform across couples (Whitaker, 2013). Finally, the novelty of this research area entails that there are very few, if any, replicated well-validated and reliable measures of CDA victimization and perpetration. Although we sought to utilize measures that had a strong theoretical basis and preliminary evidence of reliability and validity (i.e., the CARS), these measures do not have the empirical evidence to unequivocally support their valid measurement. Furthermore, there is no consensus on what, if any, “subtypes” of CDA exist. For example, the measure of CDA used in the current study originally grouped items into sexual, psychological, and cyberstalking subtypes (Watkins et al., 2016). Another instead grouped items into a “direct aggression” and “control/monitoring” subtype (Borrajo et al., 2015). We did not find statistical support for either of such distinctions through our factor analysis of the CARS. It will be important for future research to determine and define any subtypes of CDA before these subtypes are further applied conceptually in research, as researchers run the risk of creating artificial distinctions that are not supported by the lived experiences of participants. Similarly, our measure of stalking, while used previously in a national study of IPV, has also not been widely validated. To our knowledge, there is not a well-validated measure of intimate partner stalking, and thus we selected a measure that was developed by a reputable source and that appeared to cover behaviors of interest.