Psycho-Emotional Violence, Its Association, Co-Occurrence, and Bidirectionality with Cyber, Physical and Sexual Violence

Abstract

Using data from a cross-sectional sample of seventh to twelfth graders who attended northeastern schools between 2011 and 2012, we examined the associations between psycho-emotional dating violence and cyber, physical, and sexual violence Overall, we found that respondents experienced more than one type of dating violence simultaneously, indicating the prevalence of co-occurrence in dating violence. Further, being a victim of psycho-emotional violence was associated with perpetrating similar types of violence, suggesting the bidirectional nature of this type of aggression. Among the covariates, we found that peer bullying, cyber bullying, cell phone usage, the respondents’ age, gender, and grades were risk factors for psycho-emotional violence. The findings may be used to guide dating violence intervention efforts and the development of school based and family-oriented treatment plans.

Navigating dating relationships is a key developmental process for teens as they transition from early to late adolescence (Connolly et al. 2014; Norona et al. 2017). However, dating today may be more complicated than in the past, as teens face challenges associated with electronic devices and online applications (e.g., social media, messengers, and online dating tools) through which they communicate (Lenhart et al. 2015b; Shapiro and Margolina 2014). The Centers for Disease Control and Prevention (CDC; 2018b) define teen dating violence as aggression between dating couples manifested in various forms (i.e., physical, psychological, emotional, and sexual). Although different forms of dating violence have been examined (Foshee et al. 2007; Swan et al. 2008), much remains unknown about how traditional and emerging forms of dating violence (e.g., cyber dating violence) are linked to one another. To date, the pervasiveness of teen dating violence is often examined in a one-way context, in which the respondents are portrayed as victims; less is known about experiences of bidirectionality in which both victimization and perpetration occur concurrently and how experiences with victimization versus perpetration may differ. Further, in the arena of teen dating violence, few studies have investigated the co-occurrence of the various forms of dating and non-dating aggression, and the roles that family and peers may play simultaneously in the midst of this technology revolution and teens’ developmental transition. Extending previous research (Yahner et al. 2015; Zweig et al. 2013b, Zweig et al. 2014), we examined the associations between the most prevalent form of dating violence (i.e., psycho-emotional violence) and emerging dating violence (i.e., cyber dating violence), other traditional forms of dating violence (i.e., physical and sexual), controlling for teens’ social environment (i.e., peer, family, technology usage, and socio-demographic characteristics) while taking into consideration the bidirectionality (i.e., victimization and perpetration) of dating violence. Our study was guided by the following research question: How was psycho-emotional violence related to different forms of dating violence in terms of victimization and perpetration in the presence of other contextual factors? In this study, we used the term “psycho-emotional violence” in dating to broadly capture two types of sub-violence: psychological control and emotional abuse.

Psycho-Emotional Violence in Dating

Psycho-emotional violence involves intentional or unintentional manipulation, threats to physical harm, abandonment, personal safety, derogatory speech, or coercive tactics that denigrate a dating partner, with the aim to induce fear, anxiety, trauma, shame, guilt, powerlessness, and other negative emotions (Ureña et al. 2015). Examples of such violence may be manifested in the use of sarcasm, silent treatment, deceptive tactics, behavioral control, emotional withholding, name-calling, and verbal intimidation (Adamo 2014; Harper et al. 2005; Smith and Segal 2018; Temple et al. 2016). Psychological control and emotional abuse are frequently listed as prevalent forms of dating violence that precede or correlate with physical dating violence (Eshelman and Levendosky 2012; Jouriles et al. 2009). Almost half of teens (47.2%) acknowledged encountering psychological aggression from a dating partner, with victimization rates that range from 14.8% to 47.2% across studies (National Institute of Justice 2016). Nevertheless, there is evidence that the bidirectional form of such violence in which both partners subject each other to adverse emotional treatment is prevalent in teen dating (Barreira et al. 2014). Unfortunately, this form of violence, in the absence of physical scars or injuries, can be subtle and difficult to recognize (Lancer 2017). Medical evidence highlights the connection between psychological control and poor mental health outcomes (e.g., distress, depression, anxiety, loss of identity, low self-esteem) and adverse physical health consequences (e.g., alcohol use, substance use) (Moore et al. 2008; Sackett and Saunders 2001; Shorey et al. 2015; Temple et al. 2016).

Cyber Dating Violence

Cyber dating violence, an emerging form of teen dating violence, concerns the use of electronic devices to monitor, humiliate, threaten, exclude, or harass a dating partner through acts that may comprise sending intimidating messages, posting embarrassing photos or comments, or imposing online surveillance with the intent to injure or harm a dating partner (e.g., Borrajo et al. 2015a; Morelli et al. 2018; Zweig et al. 2013b). Cyber sexual violence, more specifically, typically involves coercion, comments, or gestures that solicit unwanted sexual acts and distribution of sexual images without consent (Dick et al. 2014; Henry and Powell 2015). The prevalence of cyber dating violence, based on research within the past decade, has typically ranged from 10% to 31.5% (for victimization) and from 10.5% to 18.4% (for perpetration) (Borrajo et al. 2015a; Zweig and Dank 2013). Although this form of dating violence is characterized by the absence of physical proximity, geographic boundaries, or face-to-face contact, it has been demonstrated empirically that dating violence that occurs through online communication can be just as devastating as traditional forms of in-person dating violence. In particular, electronic aggression can easily be translated to face-to-face hostility. Stephenson and colleagues (Stephenson et al. 2018) pointed out the many commonalities between cyber abuse and traditional form of psychological violence with online gadgets such as social media offering “a bigger platform to facilitate most psychologically abusive behaviors (p. 133). Further, the mobile nature of the online environment provides easy access to potential victims (e.g., through GPS monitoring), creating more opportunities for diverse forms of monitoring, controlling, isolating, public humiliation (e.g., postings of messages, photos, video recordings), and cyber-stalking (Baker and Carreño 2016; Morelli et al. 2018). Notably, it has been well established that cyber dating violence can induce distress, have serious implications on a teen’s mental health (e.g., self-worth, isolation, withdrawn, depression, suicidality), and increase other behavioral risks (e.g., alcoholic drinking, eating disorder, school challenges) (Bennett et al. 2011; National Society for the Prevention of Cruelty to Children 2018; Smith et al. 2018; Van Ouytsel et al. 2017).

Physical and Sexual Dating Violence

Physical dating violence encompasses the intentional use of force and restraints to injure or intimidate a dating partner and may involve aggressive contact, such as slapping, hitting, shoving, kicking, or punching (CDC 2017a, c; National Center for Victims of Crime 2012). Sexual violence, in contrast, often entails the use of coercion to lure someone into a range of nonconsensual acts, such as rape, assault, and other unwanted sexual contacts (CDC 2018a; National Center for Victims of Crime 2012). According to the CDC (2018a), approximately 10% of high school students have experienced some form of physical dating violence, and an estimated 11% acknowledged experiencing sexual violence. Differences in rates for perpetration in other research range between 10% and 30% (National Institute of Justice 2016) and between 7.4% and 18.5% for victimization (youth.gov n.d.). Rarely occurring as an isolated event, physical dating violence may accompany other forms of psycho-emotional violence such as degrading, belittling, and chastising to justify the physical acts (Fernández-Fuertes and Fuertes 2010). Increasing evidence indicates that physical victimization in adolescence is linked to a range of adverse psycho-emotional outcomes (e.g., heavy/binge drinking, substance use, depression, smoking, antisocial behaviors, suicidal ideation) that can have effects in adulthood (CDC 2018b; Exner-Cortens et al. 2012; Van Ouytsel et al. 2016). Further, teens who experience sexual victimization in dating relationships tend to exhibit high sexual risk-taking behaviors (e.g., have a greater number of life time sexual partners, unprotected sex, or sex under the influence of alcohol or drugs) (Hipwell et al. 2013).

Peer Bullying

Peer bullying, which may be enacted through various forms of aggression (i.e., physical, verbal, and relational), has garnered much national and empirical attention (CDC 2017b; Strickland 2015). The National Center for Education Statistics (2016) reported that over 1 in 5 students aged 12 to 18 were bullied in school, and estimates for cyber bullying (28%) are similar (Patchin 2016). Research indicates that, aside from going online, three-quarters (75%) of teens reported text messaging and using mobile devices on a daily basis (Lenhart 2015). Although the majority (69%) claim to be treated with social courtesy online (Lenhart et al. 2011), research on the risk and protective factors of such violence indicates that prevalence rates range between 9% to 45% for traditional in-person bullying and 15% to 24% for cyber bullying (PACER’s National Bullying Prevention Center 2017). The deleterious effects of bullying on the victims’ psychosocial development, emotional well-being, and academic performance are documented extensively in the public health and social science literature (Dake et al. 2003; Ortega et al. 2009; Reijntjes et al. 2010; Wolke and Lereya 2015). Bullying may be used to assert dominance and control. Hence, teens who bully others may be more likely to bully their dating partner (Kirchner 2013). There is also empirical support for the notion that dating violence victimization may be more prevalent among teens who experience peer bullying as compared to their non-bullied counterparts (Borrajo et al. 2015a; Yahner et al. 2015). Notably, bullying in any form has been shown in past studies to increase the risk of suicidal ideation, social isolation, anxiety, substance usage, and mental health concerns (StopBullying.gov n.d.; Wolke and Lereya 2015).

Family Domain

Parents play a pivotal role in their teens’ relational development (Ross and Howe 2009; van de Bongardt et al. 2015). From the social control perspective, parents regularly serve as gatekeepers to their children’s dating relationships through management of their daily activities and offering of support (Sabina and Cuevas 2013; Shaffer et al. 2018). Family scholars have explored a substantial number of predictors as means to examine the family linkage to teens’ predisposition to dating violence. In a number of studies, parental low socio-economic status, lower educational attainment or disadvantage has been identified as a risk factor for dating violence (Foshee et al. 2004, 2008; Spriggs et al. 2009). There may be several explanations for this relationship. Households of lower socio-economic status may lack resources to combat teen dating violence. First, when money is tight, financial stress in the household increases belligerent interactions between parents (Paat 2011). There is evidence that children reared in families in which interparental discord is pervasive are likely to hold normative misperceptions of intimate partner violence. From the standpoint of social learning theory, violent parental contacts increase teens’ likelihood to behave more aggressively including in their own dating relationship (Anderson and Kras 2007; Kimball 2016; Office on Women’s Health 2018; UNICEF n.d.). From the social control perspective, as more households have expectations and rules about Internet usage (Lenhart and Madden 2007; Wang et al. 2005), parental monitoring of youths’ online accounts should significantly decrease their teens’ exposure to cyberbullying (Khurana et al. 2015), thereby possibly reducing the psycho-emotional violence linked to online communication with dating partners. With respect to parent-child communication, research shows that high-quality parent-child communication and reasonable amounts of parental monitoring may minimize the risks associated with teens’ dating violence in which psycho-emotional violence is one (Reyes et al. 2016).

The Current Study

To further contribute to our understanding of teen dating violence, we examined the associations between psycho-emotional violence in the forms of psychological control and emotional abuse and other types of sexual and non-sexual dating violence (cyber and physical), controlling for contextual factors critical to teens’ social environment (peer- and cyber-bullying, family factors, teens’ technology usage), and socio-demographic characteristics. In line with our review of the literature, we posed several hypotheses. First, given dating violence is multifaceted and can take on various forms, we hypothesized that other types of dating violence perpetration (i.e., cyber, physical, and sexual) would be positively associated with psycho-emotional violence perpetration (i.e., psychological control and emotional abuse). This effect is independent of the influence of other contextual factors in the model. Likewise, we hypothesized that victimization of other types of dating violence would be positively associated with psycho-emotional violence victimization, controlling for other variables. Given the common bidrectionality in dating violence noted in the review, we anticipated that the influence of psycho-emotional violence would be bidirectional. That is, respondents who experienced psycho-emotional violence victimization in their dating relationship were more likely to perpetrate similar aggressive acts toward their dating partner and vice versa.

Next, we hypothesized that the respondents who experienced peer bullying and cyber bullying victimization were more susceptible to psycho-emotional violence victimization. With respect to the family domain, we hypothesized that a higher parent’s educational attainment would be negatively associated with both psycho-emotional violence victimization and perpetration consistent with previous studies (e.g., Foshee et al. 2008). Further, we expected that parental control of teens’ technology usage should lead to a decrease in the frequency of psycho-emotional violence victimization and perpetration since concerned parents can intervene more readily, decreasing the chance of violence. We also hypothesized that positive parent-child relations, measured in the form of parent-child openness, would be negatively associated with psycho-emotional violence in dating. Meanwhile, regular usage of a computer or cell phone was hypothesized to increase the occurrence of psycho-emotional abuse victimization and perpetration, given a teen’s greater contact with electronic devices. Finally, to ensure that any statistical significant associations that we found were distinct from other confounding factors related to the characteristics of the respondents, we controlled for the respondents’ technology usage and socio-demographic characteristics, such as age, gender, racial/ethnic classification, and scholastic performance.

Method

Data for this cross-sectional study came from the Technology, Teen Dating Violence and Abuse, and Bullying in Three States Study that were collected through convenience sampling from over 5000 seventh to twelfth graders who attended ten schools located in five districts in three northeastern states (Pennsylvania, New York, and New Jersey), between 2011 and 2012 (Zweig and Dank 2016). The aim of the study was to examine the relationship between technology (phone, email, and computer usage) and relationship dynamics (dating violence, cyber bullying, and bullying). After securing ethical approval from the university and informed consent from both parents (via email) and students, with the assurance that the data collected would not be linked to the students’ personal identifying information, the survey was conducted anonymously with participating students, using traditional paper-pencil methods. Classroom teachers were trained in procedures to administer the survey, including protocols for assuring confidentiality, distributing the survey questionnaire, and providing referrals. Upon completing the survey questionnaire, students were provided with a business-size card that included providers’ and hotline numbers on domestic violence, sexual assault, and suicide prevention programs. The response rate, calculated based on student absenteeism, refusal/opt out rates, and inconsistent responses, was 84%, yielding a sample size of 5647. The dataset is well-suited for our study because it was able to capture the rich experience of youths in terms of psycho-emotional violence victimization and perpetration, technology usage, and other relevant proximal factors deemed important for teens’ social environment.

Sample Characteristics

In this study, a dating relationship was defined as a teen’s spending time with a boyfriend or girlfriend or someone whom he or she loved and could involve sexual activities (Yahner et al. 2014; Zweig et al. 2013b; Zweig et al. 2014). Slightly over 4000 students reported that they were dating and/or had ever dated. Our analysis is based on heterosexual students who acknowledged being in this type of relationship only, yielding a sample size of 3918. The average age of the respondents was 15.5 (SD = 1.515, range = 11–19). Slightly fewer than half of the sample (49.4%) was male. Slightly over three-fourths of the respondents were white (75.5%), followed by 7.7% Latino, 4.7% black, 2.3% Asian and 9.8% of an other racial group. Regarding educational performance, 61.9% of the respondents acknowledged receiving mostly As or mostly As and Bs in school.

Measures

Dependent Variables: Psychological Control and Emotional Abuse

To measure dating violence experiences, the respondents who reported being in a dating relationship at the time of data collection or in the past year were asked to respond to a series of questions related to psycho-emotional violence behaviors. Using response categories of “never,” “rarely,” “sometimes,” and “very often” to estimate the extent of victimization, the respondents were asked, “In the past year, how often has the person who you currently are dating, or if you are not currently dating, the person who you most recently dated, done any of the following things to you?” To avoid skewing the responses and to increase validity, the respondents were reminded to indicate the frequencies of the violent occurrence only if the dating partner initiated the aggression due to reasons other than self-defense. Similarly, to assess perpetration, the respondents were asked the frequencies with which they projected their aggression, not initiated in response to self-defense (Yahner et al. 2014; Zweig et al. 2013a). Two types of psycho-emotional violence (psychological control and emotional abuse) were modeled as dependent variables from the victims’ and the perpetrators’ perspectives. Psychological control was assessed using a 5-item scale, in regard to how frequently the respondents sustained or perpetuated psychological threats and harassment as a form of control. Emotional abuse was assessed using an 8-item scale that included acts that cause emotional trauma and humiliation (see Table 1 for sample items). Factor analysis of their respective items using the principal components method indicated a one-factor solution with at least 0.65 factor loadings on the factor. Appropriate groupings of items for each scale were selected based on the dimensionality from the results of the factor analyses and consistency/reliability of the sets of items from the reliability analyses. The items in the variables have been demonstrated to have high validity in a number of studies (Zweig et al. 2013a, b, 2014). The Cronbach’s alphas from the reliability analysis for the four scales (psychological control victimization and perpetration, and emotional abuse victimization and perpetration) are at least 0.84, which is considered “excellent” (50% overlapping variance) (see Comrey and Lee 1992): alpha for psychological control victimization = 0.90; alpha for psychological control perpetration = 0.85; alpha for emotional abuse victimization = 0.88; and the alpha for emotional abuse perpetration = 0.84.

Table 1.

Descriptive statistics of variables in analyses

Constructs	Min	Max	Mean	SD	Sample items	Prevalence (%)	Chronicity
Psycho-emotional violence
Psychological control (victim)	0	15	1.294	2.882	1. Would not let me do things with other people	20.064	0.332
					2. Told me I could not talk to a person of the gender I date	18.442	0.332
					3. Made me describe where I was every minute of the day	13.229	0.217
					4. Tried to limit my contact with friends	13.105	0.215
					5. Insisted on knowing who I am with and where I am at all times	13.815	0.242
Emotional abuse (victim)	0	24	1.785	3.641	1. Insulted me in front of others	11.371	0.175
					2. Put down my looks	7.248	0.109
					3. Blamed me for bad things (s)he did	11.294	0.182
					4. Said things to hurt my feelings on purpose	13.980	0.228
					5. Threatened to start dating someone else	10.156	0.166
					6. Did something just to make me jealous	28.791	0.459
					7. Brought up something from the past to hurt me	21.955	0.352
					8. Called me names to put me down or make me feel bad	8.721	0.148
Psychological control (perpetrator)	0	15	0.441	1.512	1. Would not let him/her do things with other people	7.119	0.102
					2. Told him/her he/she could not talk to a person of the gender that he/she dates	9.243	0.142
					3. Made him/her describe where he/she was every minute of the day	5.232	0.072
					4. Tried to limit his/her contact with friends	3.172	0.043
					5. Insisted on knowing who he/she is with and where he/she is at all times	5.889	0.086
Emotional abuse (perpetrator)	0	24	0.732	2.070	1. Insulted him/her in front of others	4.183	0.054
					2. Put down his/her looks	2.211	0.030
					3. Blamed him/her for bad things they did	6.076	0.091
					4. Said things to hurt his/her feelings on purpose	6.134	0.088
					5. Threatened to start dating someone else	4.579	0.066
					6. Did something just to make him/her jealous	17.390	0.245
					7. Brought up something from the past to hurt him/her	8.759	0.124
					8. Called him/her names to put him/her down or make him/her feel bad	2.995	0.045
Cyber dating violence
Cyber violence (victim)	0	21	0.464	1.842	1. Sent threatening text messages to me	5.888	0.089
					2. Sent me instant messages or chats that made me feel scared	3.113	0.051
					3. Wrote nasty things about me on his/her profile page (e.g., on Facebook, MySpace, etc.)	4.739	0.076
					4. Sent me so many messages (like texts, e-mails, chats) that it made me feel unsafe	5.310	0.090
					5. Spread rumors about me using a cell phone, email, IM, web chat, social networking site, etc.	4.744	0.076
					6. Used information from my social networking site to harass me or put me down	3.351	0.054
					7. Threatened to harm me physically through a cell phone, text message, social networking page, etc.	2.404	0.041
Cyber sexual violence (victim)	0	12	0.325	1.293	1. Sent me text messages, email, IM, chats, etc., to have sex or engage in sexual acts with him/her when he/she knew I did not want to	6.795	0.115
					2. Sent me sexual photos or naked photos of himself/herself that he/she knew I did not want	3.534	0.062
					3. Threatened me if I didn’t send a sexual or naked photo of myself	2.373	0.043
					4. Pressured me to send a sexual or naked photo of myself	6.293	0.110
Cyber violence (perpetrator)	0	21	0.143	1.047	1. Sent threatening text messages to him/her	1.679	0.022
					2. Sent him/her instant messages or chats that made him/her feel scared	0.900	0.014
					3. Wrote nasty things about him/her on my profile page (e.g., on Facebook, MySpace)	2.372	0.033
					4. Sent him/her so many messages (like texts, e-mails, chats) that it made him/her feel unsafe	1.381	0.022
					5. Spread rumors about him/her using a cell phone, email, IM, web chat, social networking site, etc.	1.441	0.022
					6. Used information from his/her social networking site to harass him/her or put him/her down	1.111	0.018
					7. Threatened to harm him/her physically using a cell phone, text message, social networking page, etc.	0.811	0.013
Cyber sexual violence (perpetrator)	0	12	0.069	0.631	1. Sent him/her text messages, email, IM, chats, etc., to have sex or engage in sexual acts with me that I knew the person did not want to do	1.230	0.021
					2. Sent him/her sexual photos or naked photos of myself that I knew he/she did not want	0.691	0.013
					3. Threatened him/her if he/she didn’t send a sexual or naked photo of himself/herself	0.570	0.010
					4. Pressured him/her to send a sexual or naked photo of himself/herself	1.621	0.025
Physical dating violence
Physical violence (victim)	0	24	0.926	2.344	1. Scratched me	10.452	0.139
					2. Slapped me	11.576	0.151
					3. Physically twisted my arm	4.115	0.051
					4. Slammed me or held me against a wall	5.043	0.070
					5. Kicked me	7.026	0.088
					6. Bent my fingers	6.569	0.085
					7. Bit me	16.193	0.231
					8. Pushed, grabbed or shoved me	10.869	0.148
Sexual violence (victim)	0	9	0.173	0.793	1. Pressured me to have sex when he or she knew I didn’t want to	5.903	0.081
					2. Forced me to have sex	2.035	0.035
					3. Forced me to do other sexual things that I did not want to do	4.097	0.057
Physical violence (perpetrator)	0	24	0.573	1.832	1. Scratched him/her	6.509	0.080
					2. Slapped him/her	10.057	0.125
					3. Physically twisted his/her arm	1.950	0.025
					4. Slammed him/her or held him/her against a wall	2.513	0.037
					5. Kicked him/her	3.677	0.045
					6. Bent his/her fingers	3.710	0.047
					7. Bit him/her	9.766	0.136
					8. Pushed, grabbed or shoved him/her	7.800	0.101
Sexual violence (perpetrator)	0	9	0.044	0.413	1. Pressured him/her to have sex when I knew (s)he didn’t want to	1.824	0.025
					2. Forced him/her to have sex	0.639	0.011
					3. Forced him/her to do other sexual things that he/she did not want to do	1.047	0.016
Bullying
Peer-bullying (victim)	0	25	1.092	3.111	1. A student wouldn’t be friends with me because other people didn’t like me	NA	NA
					2. A student got other students not to have anything to do with me
					3. A student got their friends to turn against me.
					4. I wasn’t invited to a student’s place because other people didn’t like me.
					5. I was left out of activities with other students.
Cyber bullying (victim)	0	25	0.619	1.993	1. A student sent me a nasty email.	NA	NA
					2. A student sent me an instant message or chat to hurt my feelings.
					3. A student put something on a profile page (like MySpace or YouTube) about me to hurt my feelings.
					4. I was called names I didn’t like through a text message
					5. A student sent me a text message to hurt my feelings.
Family variables
Primary guardian a college graduate	0	1	0.692	0.462	For your primary parent or guardian, what is the highest level of education he or she completed?	NA	NA
Parental control on computer and phone	0	1	0.173	0.379	Do your parents or guardians restrict the amount of time you can spend on the computer?	NA	NA
					How do your parents or guardians restrict what websites you look at?
					Do your parents restrict the amount of time you can spend on the cell phone?
Openness with parents	0	12	6.140	3.170	During the past month, which of the following things have you done with the parent or guardian you spend the most time with or live with most?	NA	NA
					1. Talked about someone you’re dating
					2. Had a talk about a personal problem you were having
					3. Talked about your school work or grades
					4. Talked about other things you’re doing in school
Technology usage
Computer usage	0	12	3.111	2.350	How much time per day do you spend doing the following activities on the computer?	NA	NA
					1. Using social networking sites (e.g., Facebook, Myspace, etc.)
					2. Writing or posting on blogs/microblogs (e.g., tumblr, twitter, livejournal, etc.)
					3. Instant messaging or chatting
Cell phone usage	0	20	8.043	4.156	How much time per day do you spend doing the following things on a cell phone?
					1. Talking
					2. Sending instant messages or participating in chats
					3. Texting
					4. Using social networking sites (e.g., Facebook, Myspace, etc.)
Socio-demographic variables
Age	11	19	15.547	1.515	Respondent’s age	NA	NA
Male	0	1	0.494	0.500	What is your gender?	NA	NA
Race					What is your race/ethnicity?	NA	NA
White	0	1	0.755	0.430	Respondent is a Caucasian/white
Black	0	1	0.047	0.211	Respondent is an African American/black
Latino	0	1	0.077	0.267	Respondent is a Hispanic or Latino(a)
Asian	0	1	0.023	0.149	Respondent is an Asian
Other	0	1	0.098	0.298	Respondent belongs to other categories not listed above
Received As and Bs mostly	0	1	0.619	0.486	In general, what grades do you get in school?	NA	NA

Independent Variables

Physical and Cyber Dating Violence

Each of the dating violence variables was constructed to measure the intensity of the victimization and perpetration. Items were selected based on results from factor analyses and reliability analyses using previously validated items (Zweig et al. 2013b). Using response categories from “never” to “very often,” cyber violence, conceptualized as violent behaviors imposed by electronic means from a dating partner, was assessed using a 7-item scale. Using the same response categories, cyber sexual violence used a 4-item scale that captures aggressive online behaviors perpetrated by a dating partner in a sexual nature. In addition, using response categories from “never happened” to “happened 10 or more times,” physical violence, implicating aggression, assault, and force, was assessed through an 8-item scale while sexual violence was evaluated using a 3-item scale that consisted of items about forced or coercive sexual acts. Prior to scale construction, items for all the scales were coded so that higher values indicated higher frequencies of violent acts. In addition, factor analyses revealed a one-factor solution with at least 0.68 factor loadings on their respective factor for each respective scale and alphas from the reliability analyses ranged from 0.77 to 0.89 (see Table 1 for sample items): alpha for physical violence victimization = 0.85; alpha for physical violence perpetration = 0.85; alpha for sexual violence victimization = 0.77; and alpha for sexual violence perpetration = 0.77. The high internal reliability of the scales shows that there is a high correlation between the items and their respective groupings. The items from the variables have also been tested in other studies such as Zweig et al. (2013a, b, 2014).

Control Variables

Bullying

Bullying variables captured two forms of victimization: peer bullying and cyber bullying, with each one using a 6-point response scale that ranged from “never” to “every day.” Peer bullying victimization was assessed using a 5-item scale that concerned the respondents’ peers’ attempts to isolate and shun the respondent from his or her circle of friends. The cyber bullying scale was constructed from 5 items that refer to the frequency with which the respondents endured bullying through electronic means. The reliability alphas for peer bullying and cyber bullying were 0.92 and 0.85, respectively. The items from the variables have been tested in Zweig et al. (2013a, b) and Yahner et al. (2015).

Family Variables

Three family variables were incorporated into each model. The primary parent (or guardian) as a college graduate was used as a proxy for the respondents’ family socio-economic status. The dichotomous variable was coded as 1 if the respondents’ primary parent or guardian was a college graduate, and 0 otherwise. Parental control comprised three types of parental monitoring of the respondents’ cell phone and computer usage: (a) restricting the time that respondents were on their computer and cell phone; (b) using filtering software for the computer; and (c) expecting the computer to be situated in an open/public family area. When any form of parental monitoring was in place, the variable was coded 1, and 0 otherwise. A 4-point response scale (“Never” to “Often”) for openness with parents, which reflected how often the respondents talked with their parent(s) about intimate issues, was also developed, for which a higher score indicated greater readiness of respondents to confide in their parents (alpha = 0.73).

Technology Usage

Using a 5-point response scale (“No time” to “More than 6 hours”), items related to computer usage captured the extent to which the respondents spent time accessing social networking sites, writing for blogs or microblogs, or instant messaging. Cell phone usage concerned the frequency using a 6-point response scale (“No time” to “More than 7 hours”) with which respondents used their cell phones for talking, texting, messaging, chatting, and accessing social network sites. Reliability alphas for these scales were 0.67 and 0.69, respectively.

Socio-Demographic Variables

We also controlled for a number of socio-demographic characteristics. Age was measured in years. Gender was coded as 1 for male and 0 for female. Race/ethnicity comprised four dummy variables (black, Latino, Asian, and Other), with white as the reference category. Because educational performance is a significant part of the respondents’ life, and their educational performance can influence their socialization and other areas, respondents who received mostly As or “mostly As and Bs” were coded as 1, while others who had lesser academic performance were coded as 0.

Data Analysis

Our analyses were structured in several steps. First, we examined the distribution of psycho-emotional violence victimization and perpetration variables in terms of their prevalence (i.e., the percentage of respondents who had the experience) and chronicity (i.e., the average number of times that respondents had the experience), presented as descriptive statistics (see Kaukinen et al. 2012). Next, multiple regression with full information maximum likelihood estimation was utilized to examine the associations between psycho-emotional violence variables and other dating violence perpetration and victimization variables as well as other contextual variables. Multiple regression analysis is a statistical procedure that assesses the relationship between a dependent variable and a set of independent variables, while other variables are held constant (Keith 2015; Konasani and Kadre 2015). Unlike listwise deletion, which can introduce a biased sample that leads to inaccurate results, the full information maximum likelihood estimation procedure increases the efficiency of parameter calculations, using all available information in the presence of missing data (Enders 2010). Modeling was conducted in which each set of victimization and perpetration variables (based on two types of psycho-emotional violence: psychological control and emotional abuse) was modeled as either the dependent or independent variables in separate models. This resulted in four separate multiple regression models for two types of psycho-emotional violence.

Results

Psychological Control

The results presented in Table 1 indicate that being isolated from doing things with others was the most common form of psychological control victimization, as 20% of the respondents reported being victimized. Almost 10% of the respondents acknowledged telling their dating partner not to talk to a person of opposite gender. Over one quarter (i.e., 28.8%) of the respondents revealed that their dating partner did something just to make them jealous and slightly less than one fifth (i.e., 17.4%) of them admitted behaving similarly to their dating partner.

Three forms of dating violence victimization (emotional abuse, cyber violence, and physical violence) were associated with a greater occurrence of psychological control victimization. Although respondents who reported imposing psychological control on a dating partner were more likely to be victimized (b = 0.531, p < 0.001), emotional abuse and cyber violence perpetration were associated with a decrease in such occurrence (b = −0.073, p < 0.01 for emotional abuse; b = −0.235, p < 0.001 for cyber violence), controlling for other variables in the model. In terms of technology usage, greater cell phone usage was associated with respondents’ being more likely to face psychological control victimization (b = 0.031, p < 0.01). No statistically significant associations were noted between bullying, family, or socio-demographic variables and psychological control victimization (Model 1).

Having experienced psychological control from a dating partner was associated with perpetration in a similar fashion (b = 0.185, p < 0.001). The victimization experience of cyber violence and physical violence, however, was associated with the opposite effect (b = −0.100, p < 0.001 for cyber violence and b = −0.046, p < 0.001). Three forms of dating violence perpetration (emotional abuse, cyber violence, and physical violence) were associated with a positive association with psychological control perpetration. In terms of peer aggression, experiencing cyber bullying victimization was associated with perpetration of psychological control (b = 0.030, p < 0.05). As associated with increases in age, the older respondents, they were more likely to exert control over their dating partner (b = 0.033, p < 0.05). No statistically significant associations were noted between family variables, technology usage, gender, race/ethnicity, or scholastic achievement and psychological control perpetration (Model 2) (See Table 2).

Table 2.

Multiple regression analyses with psychological control as the dependent variable

Variables	Psychological Control
	Model 1					Model 2
	(Victim)					(Perpetrator)
	b	(SE)	95% CI			b	(SE)	95% CI
Intercept	−0.107	(0.374)	−0.840	–	0.627	−0.589	(0.223)**	−1.026	–	−0.152
Psycho-emotional violence
Psychological control (victim)						0.185	(0.010) ***	0.165	–	0.204
Emotional abuse (victim)	0.410	(0.014) ***	0.382	–	0.437	−0.011	(0.010)	−0.030	–	0.007
Psychological control (perpetrator)	0.531	(0.029) ***	0.474	–	0.587
Emotional abuse (perpetrator)	−0.073	(0.023) **	−0.119	–	−0.027	0.248	(0.013) ***	0.223	–	0.273
Cyber dating violence
Cyber violence (victim)	0.202	(0.028) ***	0.146		0.257	−0.100	(0.017) ***	−0.132	–	−0.067
Cyber sexual violence (victim)	0.060	(0.037)	−0.012	–	0.132	−0.020	(0.023)	−0.063	–	0.022
Cyber violence (perpetrator)	−0.235	(0.059) ***	−0.351	–	−0.119	0.213	(0.035) ***	0.144		0.282
Cyber sexual violence (perpetrator)	0.000	(0.099)	−0.195	–	0.195	−0.086	(0.060)	−0.203	–	0.032
Physical dating violence
Physical violence (victim)	0.069	(0.021) **	0.028	–	0.110	−0.046	(0.012) ***	−0.070	–	−0.022
Sexual violence (victim)	0.014	(0.057)	−0.098	–	0.126	−0.027	(0.034)	−0.094	–	0.039
Physical violence (perpetrator)	0.018	(0.028)	−0.037	–	0.072	0.146	(0.016) ***	0.115	–	0.177
Sexual violence (perpetrator)	−0.157	(0.101)	−0.355	–	0.041	0.078	(0.063)	−0.044	–	0.201
Bullying
Peer-bullying (victim)	0.018	(0.013)	−0.008	–	0.044	−0.004	(0.008)	−0.019	–	0.012
Cyber bullying (victim)	−0.007	(0.022)	−0.049	–	0.036	0.030	(0.013) *	0.005	–	0.055
Family variables
Primary guardian a college graduate	−0.095	(0.081)	−0.253	–	0.063	−0.053	(0.048)	−0.146	–	0.040
Parental control on computer and phone	0.046	(0.090)	−0.132	–	0.223	0.031	(0.054)	−0.075	–	0.137
Openness with parents	0.000	(0.012)	−0.024	–	0.023	0.005	(0.007)	−0.009	–	0.018
Technology usage
Computer usage	−0.020	(0.018)	−0.056	–	0.016	0.000	(0.011)	−0.023	–	0.022
Cell phone usage	0.031	(0.011) **	0.010	–	0.052	0.010	(0.007)	−0.003	–	0.022
Socio-demographic variables
Age	0.014	(0.023)	−0.031	–	0.059	0.033	(0.014) *	0.006	–	0.059
Male	−0.009	(0.074)	−0.154	–	0.136	0.054	(0.044)	−0.032	–	0.141
Race
Black	−0.116	(0.161)	−0.431	–	0.199	0.003	(0.097)	−0.187	–	0.192
Latino	0.241	(0.129)	−0.012	–	0.493	−0.069	(0.078)	−0.221	–	0.084
Asian	−0.112	(0.234)	−0.569	–	0.346	−0.019	(0.138)	−0.289	–	0.251
Other	−0.049	(0.114)	−0.273	–	0.175	−0.009	(0.069)	−0.144	–	0.127
Received As and Bs mostly	−0.010	(0.074)	−0.155	–	0.136	−0.010	(0.044)	−0.097	–	0.076
R-Squared	0.564					0.446

This table presents multiple regression models with unstandardized coefficients, standard errors, and 95% confidence interval

* refers to p < 0.05, ** refers to p < 0.01, *** refers to p < 0.001

Source: Technology, Teen Dating Violence and Abuse, and Bullying in Three States, 2011–2012

b = unstandardized coefficients; SE = standard errors; CI = confidence interval of b’s

Standard errors are in parentheses

N = 3918

Emotional Abuse

In another form of psycho-emotional violence (i.e., emotional abuse), we found that all five forms of dating violence victimization (psychological control, cyber violence, cyber sexual violence, physical violence, and sexual violence) were positively associated with experiencing emotional abuse victimization from a dating partner respectively, controlling for other variables in the model. Further, respondents who reported being emotionally abusive as a dating partner were more likely to encounter similar emotional dating abuse themselves (b = 0.432, p < 0.001). Having perpetrated cyber sexual violence and sexual violence, however, had the opposite effect (b = −0.412, p < 0.001 for cyber sexual violence and b = −0.277, p < 0.05 for sexual violence). In addition, experiencing cyber bullying victimization increased the likelihood of emotional abuse victimization (b = 0.093, p < 0.001). Compared to their female counterparts, male respondents were less likely to experience emotional abuse victimization (b = −0.369, p < 0.001). In terms of their scholastic performance, respondents who received mostly As and Bs were less likely to experience emotional abuse victimization than were their counterparts who did not, scoring 0.18 unit lower, on average (b = −0.180, p < 0.05). No statistically significant associations were noted between family variables, technology usage, age, or race/ethnicity and emotional abuse victimization (Model 3).

Respondents who had to endure emotional abuse themselves were more likely to be emotionally abusive to their dating partner (b = 0.210, p < 0.001); however, the experience of cyber violence and physical violence victimization was associated with a decrease of such occurrence (b = −0.075, p < 0.01 for cyber violence; b = −0.067, p < 0.001 for physical violence). Having experiencing psychological control victimization and perpetrating cyber sexual violence, however, were inversely associated with emotional abuse perpetration (b = −0.044, p < 0.01 for psychological control victimization and b = −0.270, p < 0.01 for cyber sexual violence perpetration). We also found that four types of dating violence perpetration (psychological control, cyber violence, physical violence, and sexual violence) were positively related to perpetrating emotional abuse, as was the experience of peer bullying victimization (b = 0.026, p < 0.05), but that lower cell phone usage might serve as a risk reduction factor (b = −0.017, p < 0.05). Finally, although violent tendencies increased with age (b = 0.052, p < 0.01), males were less likely to be emotionally abusive than were their female counterparts (b = −0.163, p < 0.01). No statistically significant associations were found between family variables, race/ethnicity, or scholastic achievement and emotional abuse perpetration (Model 4) (See Table 3).

Table 3.

Multiple regression analyses with emotional abuse as the dependent variable

Variables	Emotional Abuse
	Model 3					Model 4
	(Victim)					(Perpetrator)
	b	(SE)	95% CI			b	(SE)	95% CI
Intercept	0.143	(0.416)	−0.672	–	0.957	−0.533	(0.294)	−1.109	–	0.043
Psycho-emotional violence
Psychological control (victim)	0.507	(0.018) ***	0.473	–	0.542	−0.044	(0.014) **	−0.071	–	−0.016
Emotional abuse (victim)						0.210	(0.012) ***	0.186	–	0.233
Psychological control (perpetrator)	−0.040	(0.034)	−0.106	–	0.026	0.429	(0.022) ***	0.385	–	0.472
Emotional abuse (perpetrator)	0.432	(0.024) ***	0.384	–	0.480				–
Cyber dating violence
Cyber violence (victim)	0.493	(0.030) ***	0.434	–	0.552	−0.075	(0.022) **	−0.118	–	−0.033
Cyber sexual violence (victim)	0.219	(0.041) ***	0.139		0.299	−0.003	(0.029)	−0.058		0.053
Cyber violence (perpetrator)	−0.086	(0.066)	−0.215	–	0.043	0.443	(0.045) ***	0.355	–	0.532
Cyber sexual violence (perpetrator)	−0.412	(0.112) ***	−0.630	–	−0.193	−0.270	(0.078) **	−0.423	–	−0.118
Physical dating violence
Physical violence (victim)	0.182	(0.023) ***	0.137	–	0.228	−0.067	(0.016) ***	−0.098	–	−0.036
Sexual violence (victim)	0.181	(0.063) **	0.057	–	0.304	−0.039	(0.044)	−0.126	–	0.047
Physical violence (perpetrator)	−0.055	(0.031)	−0.115	–	0.005	0.242	(0.020) ***	0.202	–	0.282
Sexual violence (perpetrator)	−0.277	(0.118) *	−0.508	–	−0.045	0.230	(0.081) **	0.071	–	0.390
Bullying
Peer-bullying (victim)	−0.006	(0.015)	−0.036	–	0.023	0.026	(0.010) *	0.006	–	0.046
Cyber bullying (victim)	0.093	(0.024) ***	0.046	–	0.140	0.008	(0.017)	−0.025	–	0.041
Family variables
Primary guardian a college graduate	−0.009	(0.090)	−0.186	–	0.167	0.059	(0.062)	−0.062	–	0.179
Parental control on computer and phone	−0.069	(0.101)	−0.266	–	0.128	−0.082	(0.071)	−0.222	–	0.057
Openness with parents	−0.002	(0.013)	−0.028	–	0.024	0.005	(0.009)	−0.013	–	0.022
Technology usage
Computer usage	−0.011	(0.021)	−0.051	–	0.030	−0.005	(0.015)	−0.034	–	0.024
Cell phone usage	0.006	(0.012)	−0.018	–	0.030	−0.017	(0.009) *	−0.034	–	0.000
Socio-demographic variables
Age	0.035	(0.025)	−0.015	–	0.085	0.052	(0.018) **	0.017	–	0.087
Male	−0.369	(0.083) ***	−0.531	–	−0.207	−0.163	(0.058) **	−0.277	–	−0.049
Race
Black	−0.065	(0.180)	−0.418	–	0.288	0.126	(0.127)	−0.123	–	0.376
Latino	−0.225	(0.144)	−0.507	–	0.056	0.163	(0.102)	−0.038	–	0.363
Asian	0.128	(0.260)	−0.383	–	0.638	0.068	(0.181)	−0.288	–	0.423
Other	−0.032	(0.127)	−0.281	–	0.218	0.168	(0.091)	−0.011	–	0.346
Received As and Bs mostly	−0.180	(0.083) *	−0.342	–	−0.018	−0.033	(0.058)	−0.148	–	0.081
R-Squared	0.657					0.499

This table presents regression models with unstandardized coefficients, standard errors, and 95% confidence interval

* refers to p < 0.05, ** refers to p < 0.01, *** refers to p < 0.001

Source: Technology, Teen Dating Violence and Abuse, and Bullying in Three States, 2011–2012

b, unstandardized coefficients; SE, standard errors; CI, confidence interval of b’s

Standard errors are in parentheses

N = 3918

Collectively, approximately 45 to 66% of the variance in respondents’ psycho-emotional violence can be attributed to the variables in the models.

Discussion

Dating violence is a critical issue during teens’ developmental phase as many teens report they are currently dating, have dated, or desire to be in a romantic relationship (Child Trends 2015; Lenhart et al. 2015a). Empirical evidence suggests that such relationships, which help teens to develop a sense of identity, refine resolution skills, and build emotional support, will have a long-lasting impact on their subsequent personal development in adulthood (Norona et al. 2017; van de Bongardt et al. 2015). Despite the fact that much research has been conducted on teen dating violence, our understanding of how psycho-emotional violence may be related to other forms of violence in teens’ social environment remains unclear. Less frequently examined is the empirical link between other forms of peer aggression, both online and offline, as the influence of dating violence is most often examined independent of peer aggression. This study addressed the aforementioned concern by incorporating different forms of dating violence from the perspective of victims and perpetrators. In line with previous research (Hamby et al. 2012; Zweig et al. 2013a), our study showed that respondents often confronted more than one type of dating violence simultaneously. Several scenarios are possible. For example, an emotionally abusive verbal altercation may be an indication of impending physical violence (Eshelman and Levendosky 2012), while sexual coercion may be accompanied by a series of verbal threats. In addition, exacerbation of offline violence may be related to online violence (e.g., Baker and Carreño 2016), following jealousy-provoking online networking. We believe that face-to-face altercations can similarly spark digital harassment and vice versa.

With few exceptions, our study shows that victimization or perpetration of one form of dating violence was linked to psycho-emotional violence victimization or perpetration respectively, attesting support for the co-occurrence of different forms of dating violence (e.g., Hamby et al. 2012; Sears et al. 2007; Yahner et al. 2015). The fact that victimization was associated with perpetration of a similar fashion suggests the potential bidirectional nature of teen dating violence, in line with other studies (Barreira et al. 2014; Taylor and Sullivan 2017). In terms of non-dating aggression, our study aligns with other studies that found that experience with peer bullying and/or cyber bullying were risk factors for both dating violence victimization and perpetration (Borrajo et al. 2015a; Yahner et al. 2015). Research has shown that bullying behaviors, whether victimization or perpetration, share acts with teen dating violence, which is essentially a form of bullying that occurs in a dating relationship (Hertzog et al. 2016).

When the family domain was taken into account, we did not find any family factors in the model to be associated with the participants’ experience with psycho-emotional violence. This finding was unexpected as other studies have shown that teens who keep their parents informed of relational difficulties and potential violence may be able to seek help from their parents more readily compared to their counterparts who do not (e.g., Sabina and Cuevas 2013; Shaffer et al. 2018). Additionally, there is evidence from previous studies that parents of higher educational attainment may be more involved in their children’s daily activities and, thus, more aware of the social challenges that teens confront, or likely to incorporate expert-guided advice into their instrumental parenting (Lareau 2011; Roksa and Potter 2011) to shield their children from violence in dating.

With respect to technology usage, today’s technology both strengthens relationships and facilitates abuse among dating partners. Empirical evidence posited that teens’ daily communication may take place through a mixture of online gadgets and networking sites (e.g., social media, instant messengers, chatrooms, and blogs). In particular, not only are digital devices being utilized in initiating, maintaining, and dissolving a relationship, they are being used to solicit sexual contacts (e.g., sexting) and inflict different forms of aggression/control (e.g., verbal, emotional, and psychological) (Baker and Carreño 2016). In our study, we found that regular cell phone usage was positively associated with psychological control victimization. Nevertheless, not all technological usage was associated with adverse outcomes; we found that cell phone usage helped to mitigate the risks of emotional abuse perhaps because regular phone calls may help to facilitate better communication.

Lastly, our study found that male and female respondents had differing experiences of psycho-emotional violence in dating. In particular, we found that females as compared to males were more predisposed to emotional abuse, both as victims and perpetrators. We suspect that female participants may be more sensitive to cues related to emotional pain and have been socialized to express their feelings freely, unlike their male counterparts, who are applauded for not expressing their emotions as freely (Chaplin 2015; Schmitt 2015).

Implications for Practice

Because teen dating violence often involves multiple forms of abuse, adults who work with teens (e.g., parents, school teachers, counselors) should be provided with intervention materials that help to identify early signs of abuse, are tailored to teens’ needs, and prevent further escalation. Signs of psycho-emotional violence may be subtle compared to other forms of dating violence; however, school personnel can utilize digital technology to educate teens on how to set healthy boundaries and reduce social isolation (Zweig et al. 2013b). Further, teens should be informed of the potential emotional impact of their online and offline behaviors. Relationship building, media literacy, and Internet safety workshops can be extended to teach teens how to initiate positive communication, cope with jealousy, identify warning signs, promote adequate boundaries, and preserve healthy relationships. Given that psycho-emotional violence frequently occurs bidirectionally, prevention and intervention programs must focus on the context of mutual violence (Borrajo et al. 2015b).

Limitations of the Study

Our study has presented the dynamics of and associations among psycho-emotional violence and various types of dating violence victimization and perpetration. Nevertheless, several study limitations should be noted. First, the respondents were recruited using convenience sampling in which the invitation was extended only to teens who were in attendance at school during the day of data collection; thus, the results may not be generalizable to other samples. Next, given that our sample comprises predominantly white students from small cities in Pennsylvania, rural New York, and suburban New Jersey, our findings may not capture the experiences of students from other geographical locations or those of racial/ethnic minorities. Additionally, our findings are limited to the experiences of heterosexual teens. Like other studies, our data were self-reported via survey administration and thus may be subject to social desirability and recall bias. Despite the study was conducted anonymously, it is possible that some of the teen respondents might have felt embarrassed about disclosing, been in denial of current abuse, harbored guilt of past victimization, or feared repercussions of their disclosure, thus affecting their responses. Finally, because the study was cross-sectional, we cannot establish causation or distinct pathways between victimization and perpetration. Overall, however, our study findings have merit and contribute to the small pool of current literature on bidirectionality and co-occurrence in understanding of teen dating violence.

Funding

This study effort was supported by the UTEP BUILDing SCHOLARS NIH Award # RL5GM118969.

Compliance with Ethical Standards

Conflict of Interest

The authors declare that they have no conflict of interest.