Short-term effects of peer bullying victimization on intimate partner violence

Abstract

Introduction:

While research has established associations between peer bullying victimization and increased risk of intimate partner violence (IPV) from a partner, most research has been cross-sectional and therefore unable to account for prior experiences of bullying and the cyclical nature of the bullying-IPV relationship

Methods:

The goal of our study was to provide a robust evaluation of the effects of physical and psychological peer bullying victimization on IPV victimization using longitudinal data from 2015–2017 that bridges the transition between middle and high school among youth from a predominately Latine agricultural region in California (analysis performed in 2025). We explored if physical and psychological peer bullying increased risk for IPV and assessed contextual factors that may modify the relationship between bullying and IPV to identify intervention targets. To account for this time-varying confounding, we used sequential conditional mean models.

Results:

All forms of bullying from a peer including physical (Adjusted Odds Ratio (AOR) 1.77, 95% Confidence Interval (CI): 1.32, 2.37; p<0.00) and psychological (AOR 1.58 95% CI: 1.10, 2.25; p=0.01) bullying, similarly increased risk of IPV in the short-term after accounting for prior experiences of bullying and IPV. In addition, we found that contextual factors including increased neighborhood disorder and decreased social cohesion modified the relationship between bullying and IPV.

Conclusions:

There is a short-term immediate effect of bully victimization on IPV despite reoccurring and cyclical experiences of both bullying and IPV. Increasing social support from community members is a critical intervention target for preventing IPV among youth.

Introduction:

Peer bullying constitutes a predominant form of violence among adolescents and young adults and takes multiple forms including physical violence, threats to safety, and emotional violence (e.g., social exclusion).¹ Broadly speaking, these forms fall within the categories of physical (e.g., hitting, pushing) or psychological bullying (e.g., emotional, threats). While common, bullying decreases during adolescence, with the highest rates in middle school. Nationwide, 28% of middle schools, 15% of high schools, and 10% of elementary schools report bullying at least once a week.¹ Bullying victimization can have adverse physical and mental health consequences including subsequent engagement in violence and other risk behaviors making it an important target for prevention interventions among adolescents.

More recently, research has examined the role of peer bullying in romantic relationship formation in adolescence, including implications for relationship empowerment (i.e. fostering mutual respect, understanding, and support within a relationship) and intimate partner violence (IPV). While research has established associations between bullying victimization and increased risk of IPV, most research has been cross-sectional and therefore unable to establish temporal associations, adequately control for time-varying confounding, or examine patterns in either factor over time.^2–5 The few existing longitudinal studies have focused on bullying and IPV perpetration, but not victimization..⁶ One study found that adolescents bullied in their teens were twice as likely to experience IPV victimization later in adulthood (20–27 years old).⁵ However, the authors did not examine the short-term effects on IPV during adolescence when bullying is most prevalent, or how various forms of bullying victimization effect different forms of IPV (e.g. physical or psychological). While studies do examine the number of bullying experiences in relation to the outcome of IPV, most do not account for time-varying exposure and confounding in this relationship meaning that someone who is bullied may be more likely to have repeated experiences of bullying and it might be the recurrence that affects risk. Longitudinal studies are needed to examine the relationship between bullying and IPV, considering prior bullying experiences and the cyclical nature of this relationship.

Additionally, there is limited evidence related to peer bullying and partner IPV victimization in areas outside of urban centers. Adolescent and young adults in agricultural regions face unique vulnerabilities and contextual factors such as seasonal employment, migration, and strong extended family ties,^7–9 For example, Salinas, CA is an agricultural hub which produces much of the nation’s lettuce and other crops.^7,10 The area has a predominately Latine population and strong community relationships, but it also faces challenges such as high poverty rates and community violence, which may negatively impact adolescent health. Ecological Systems Theory posits that human development is shaped by a variety of nested environments, ranging from family dynamics to societal and cultural influences.¹¹ Peer bullying might occur because of environments with both negative and positive influences, such as exposure to community violence or community cohesion, and these dynamics can influence behaviors and expectations in relationship formation. Research is needed in agricultural regions to understand how these factors shape bullying and its impact on IPV and health.

The goal of our study was to examine the effects of physical and psychological peer bullying victimization on IPV victimization using longitudinal data that bridges the transition between middle and high school among youth from a predominately Latine agricultural region in California. This two-year period (age 13–15) is a critical developmental period with social and normative transitions/relationship dynamic formation. The study explored if physical and psychological peer bullying increased IPV from a partner and assessed contextual factors that may modify the relationship between bullying and IPV to understand how to target future interventions to reduce bullying and prevent IPV.

Methods

Study Sample

A Crecer:

The Salinas Teen Health study was a prospective cohort study of 599 adolescents enrolled as eighth graders in Salinas, California between November 2015 and March 2017.⁷ Salinas, a small city in California’s central coast, is an established migration destination for agricultural sector employment and home to multiple generations of immigrants, primarily of Mexican origin. The study was designed to examine the social and structural factors that influence Latine adolescents’ well-being during the transition from middle school to high school.

Eligibility criteria included being enrolled in eighth grade, aged 12 to 15 years old, able to speak English or Spanish, intending to live in Salinas for the next year, and willing to provide contact information for a parent who could provide permission for study enrollment. Details of recruitment strategy can be found elsewhere.⁷

Participants completed a computer-based questionnaire every six months over two years of follow-up, for a total of five study visits. Surveys were administered by bilingual interviewers at community-based locations; sensitive questions including those related to sexual health and violence were completed independently via computer-assisted self-interviewing (CASI). Parents provided verbal permission to a member of the research team following a structured protocol. Youth provided written assent at their baseline study visit. The RTI Institutional Review Board approved all study procedures.

Measures

Bullying victimization was assessed using a 6-item version of the Multidimensional Peer Victimization Scale, with self-reported CASI questions at each study visit.¹² Participants reported the frequency of different bullying behaviors experienced in the past six months, with response options of “none,” “once,” or “more than once.” The effect of peer physical violence on IPV was hypothesized to differ from verbal or psychological bullying; therefore, these two forms of bullying were examined separately. Physically bullying was defined by one item in the scale, as self-report of someone “physically hurting you” at least once in the past 6 months.” Psychologically bullying was defined by self-report of one of the following occurring at least once in the past 6 months: “called you names”, “threatened you”, “tried to get you into trouble with your friends”, “made fun of you for some reason”, or “made other people not talk to you” (Cronbach’s alpha= 0.76). The peer bullying questions were asked in relation to “peers” more broadly and did not explicitly state that this should not be a partner. While there may be some overlap with IPV, we expect it would uncommon, and measures of bullying and IPV were at different time points and controlling for prior experiences of both.

IPV victimization was self-reported using CASI and assessed with five questions adapted from the Revised Conflict Tactics Scale.^6,8,13,14 Participants were first asked if they had a boyfriend/girlfriend, someone they considered a main partner or had an intimate relationship with someone who was “more than just a friend” in the past six months. If yes, they were asked: “In the last 6 months, did a boyfriend/girlfriend or someone else you were dating, going out with or hooking up with do any of the following to you: called you names, insulted you or treated you disrespectfully in front of others, swear at you, threatened you with violence or threatened to hurt you, pushed or shoved you, threw something at you that could hurt you?” IPV victimization was defined as an affirmative response to any of these questions. If no romantic relationship was reported, IPV was assumed not to have occurred. Physical IPV victimization was also examined on its own as one variable and was defined by whether a partner pushed, shoved, or threw something at the participant. Frequency or severity of IPV was not assessed in the questionnaire hence the analysis focused on any experience.

Potential covariates were selected based on prior literature, were interviewer-administered and included high school attended, age (years), biological sex (male/female), and mother’s education level (less than high school, high school/GED, more than high school, unknown). Contextual factors that were examined as effect measure modifiers and were selected based on prior research highlighting associations between these factors and adolescent health outcomes.^8–10,15,16 These factors were interviewer administered and included neighborhood disorder (dichotomized at midpoint), community social cohesion, experiences of discrimination, neighborhood resources and engagement outside of school and home. Neighborhood disorder was measured using a validated 11-item scale, dichotomized into high (≥5 events) and low (<5 events). based on prior analyses. ^9,17 Community social cohesion was a continuous mean score with 6 items on a four point Likert scale.¹⁸ Discrimination was measured using a short self-report instrument, the “Experiences of Discrimination” (EOD) measure” which is 9-item EOD scale, scored by frequency of occurrence or by situation counts.¹⁹ Neighborhood resources was an inventory of 9 items and a score was created based on the number of resources (range 0–9).^20,21 Engagement outside of school/home was measured with 3 items on a 4-point Likert scale. Responses were summed to create a total score ranging from 3–12.²² To examine modification, all scales were dichotomized at the midpoint (high vs low) to increase interpretability of the estimates and because this has been done in prior papers from the study.^9,16 Effect estimates for the relationship between bullying and IPV stratified were examined by each modifier.

Statistical Analysis

This analysis included participants who completed at least three consecutive study visits to estimate the effect of bully victimization on subsequent IPV risk, accounting for prior experiences. We used chi-square tests to assess potential selection bias by comparing key characteristics of participants not included in the analytic sample

It was hypothesized that there is likely time-dependent confounding in the relationship between bullying and IPV victimization. Prior bullying influences future victimization and IPV risk, and is a confounder of the subsequent bullying and IPV relationship (Appendix Figure 1). Prior IPV experiences also affect the likelihood of future IPV and bullying. To address this time-varying confounding, sequential conditional mean models (SCMMs) were used. SCMMs estimate short-term effects by controlling for prior exposures, outcomes, and covariates. ^23,24 These models, fitted using generalized estimating equations (GEEs) with a logit link and exchangeable correlation matrices, were applied separately for physical and psychological bullying to examine their impact on general and physical IPV risk.

Each SCMM included three lagged variables: two for bully victimization (one from the prior visit and one from the visit a year earlier) and one for IPV from the prior visit. Including the history of exposure in the model (i.e., bullying) adjusts for confounding between future exposures and outcomes. Likewise, including outcome history (i.e., IPV) adjusts for confounding between future exposures and outcomes and is recommended to increase precision of short-term effect.²³ In addition, each model controlled for participant’s sex, age at interview, mother’s highest level of education and recruitment school.

Frequencies of measures were calculated to examine characteristics of the sample and changes over time. Differences in bullying exposures by participant biological sex were assessed using GEE models with a logit link function and unstructured working correlation matrices, adjusting for covariates. The analysis focused on differences by sex given prior literature showing differences by sex.⁸ To examine modification, estimates of the bullying and IPV victimization relationship were stratified by contextual factors by including an interaction term in the model. All analyses were performed using Stata 16 (StataCorp LLC, College Station, TX).

Results

Among the 599 participants enrolled in the cohort, 558 (93%) completed at least three consecutive study visits and were included in this analysis; 526 (88% of cohort) completed all five visits. The participants excluded for having missed visits over time (N=41) were somewhat more likely to have a mother with higher educational attainment (39% vs. 24% had post-secondary education, p=0.10), not to have a parent working in agriculture (63% vs 50%, p=0.10), and have food insecurity at enrollment (15% vs 7%, p=0.08) compared to those included. They did not differ in report of being physically or psychologically bullied at enrollment (p>0.30). In general, youth were second generation (US-born with at least one immigrant parent, 70%) and Latine (95%) of Mexican origin (89%); 53% were female. Half had a parent working in agriculture and 61% were living in crowded housing conditions.

Nearly 19% of eighth graders (N=108) reported being physically harmed by a peer bully at least once in the past six months at enrollment, with 6% experiencing multiple instances.

Approximately 73% reported being psychologically bullied at least once in the six months prior to enrollment and 37% experienced it more than once. Almost all (99%) physically bullied participants also reported psychological bullying. Over time, prevalence of both physical and psychological bullying slowly declined, though new instances were reported by additional participants at each visit. (Figure 1). By the end of the two-year study period, 38% had reported been physically bullied and 91% reported being emotionally bullied at least once. Females were slightly more likely than males to experience emotional bullying (AOR 1.14, 95% CI: 1.05, 1.25, p=0.002;) but less likely to experience physical bullying, though the results were inconclusive (AOR 0.87, 95% CI: 0.67, 1.12; p=0.27).

Figure 1.

The proportion of youth who reported being physically bullied (A.) and psychologically bullied (B.) within the previous 6 months at each study visit (N=558). The dotted line represents the cumulative proportion who ever reported being bullied. Figure A represents youth who reported being physically bullied while Figure B portray experiences of psychological bullying.

At each visit, approximately half of participants (range: 52–57%) reported having a partner in the previous six months. On average, 8.7% of youth per visit reported being a victim of any type of IPV over the previous 6 months (95% CI: 7.3–10.0%), with 3.1% experiencing physical violence from a partner (95% CI: 2.2–3.9%). By the two-year visit, 22% of participants had ever experienced IPV. Those who were physically bullied had nearly twice the odds of being a victim of IPV in the next 6 months (AOR 1.77, 95% CI: 1.32, 2.37; p<0.001); likewise, those who had been psychologically bullied had 1.6 times the odds of IPV in the next 6 months (95% CI: 1.10, 2.25; p=0.01; Table 2). Bullying victimization had a similar effect on risk of physical IPV; odds of physical IPV were 2.7 times the odds of those who were physically bullied (95% CI 1.41, 5.00; p=0.001) and were 1.9 times the odds for those who had been psychologically bullied (95% CI: 0.98, 3.78; p=0.06) compared to those who had not. Bullying in the last l2 months was not associated with IPV outcomes indicating that effects may be through reoccurring bullying experiences in the last 6 months. As a sensitivity analysis, participants who never reported having a partnership were excluded from the model (n=107, 19%; Appendix Table 1); associations between bullying and IPV remained similar.

Table 2.

Effect of physical and psychological bullying victimization on IPV victimization over a two-year period

	IPV victimization			Physical IPV victimization
	AOR	95% CI	p-value	AORa	95% CI	p-value
Model 1 (N=558)
Physically bullied, previous 6 months	1.77	(1.32, 2.37)	<0.001	2.66	(1.41, 5.00)	0.002
IPV victimization, previous 6 months	5.91	(4.45, 7.85)	<0.001	2.89	(1.55, 5.38)	0.001
Physically bullied, previous 12 months	1.01	(0.74, 1.37)	0.96	0.95	(0.48, 1.89)	0.89
Model 2 (N=558)
Psychologically bullied, previous 6 months	1.58	(1.10, 2.25)	0.01	1.93	(0.98, 3.78)	0.06
IPV victimization, previous 6 months	5.67	(4.29, 7.50)	<0.001	2.76	(1.50, 5.10)	0.001
Psychologically bullied, previous 12 months	1.43	(0.98, 2.09)	0.07	1.49	(0.74, 2.99)	0.26

^aBoth GEE models adjusted for mother’s education, participant sex and age at each visit, as well as recruitment school

Note: Boldface indicates statistical significance (p<0.05); IPV=Intimate Partner Violence; AOR= Adjusted Odds Ratio; Model 1” and “Model 2” correspond to physical and psychological bullying, respectively

Among those with a higher neighborhood disorder, those who were bullied had much higher odds of IPV (OR 2.08; 95% CI 1.16, 3.72) compared to those who reported less neighborhood disorder (OR 1.03; 95% CI 0.73, 1.45 (Table 3). Similarly, those who reported lower community social cohesion had a higher odds of IPV associated with bullying (OR 2.09; 95% CI 1.19, 3.67) compared to those with higher community social cohesion (OR 1.29; 95% CI 0.79, 2.11). Estimates were not substantially different when stratifying by experience of discrimination, neighborhood resources, and engagement outside of school and home.

Table 3:

Effect of bullying victimization on IPV victimization over a two-year period, stratified by contextual factors.

	Relationship between any bullying and any IPV victimization
Baseline social environment	AORa	95% CI	p-value
Neighborhood disorder
Higher (5–11 events)	2.08	(1.16, 3.72)	0.01
Lower (0–4 events)	1.03	(0.73, 1.45)	0.88
Community social cohesion
Higher (score 3–4)	1.29	(0.79, 2.11)	0.30
Lower (score 1–2.99)	2.09	(1.19, 3.67)	0.007
Experiences of discrimination
Any	1.89	(0.88, 4.07)	0.10
None	1.50	(0.98, 2.30)	0.06
Neighborhood resources
Higher (score 5–9)	1.34	(0.89, 2.01)	0.16
Lower (score 0–4)	1.73	(0.89, 3.36)	0.11
Engagement outside of school and home
Higher (score 7–12)	1.48	(0.89, 2.45)	0.13
Lower (score 3–6)	1.59	(0.91, 2.79)	0.10

^aBoth GEE models adjusted for mother’s education, age at each visit, and recruitment school

Note: Boldface indicates statistical significance (p<0.05); IPV=Intimate Partner Violence; AOR= Adjusted Odds Ratio

Discussion:

In this longitudinal analysis, physical and psychological peer bullying victimization was common and associated with IPV victimization from a partner after accounting for time varying confounding. All forms of bullying from a peer including physical and psychological bullying, similarly increased risk of IPV. There is a short-term immediate effect of bully victimization on IPV despite reoccurring and cyclical experiences of bullying and IPV. Contextual factors including increased neighborhood disorder and decreased social cohesion modified the relationship between bullying and IPV and may be targets for intervention.

While most evidence focuses on violence perpetration, these findings align with studies showing that peer bullying victimization increases the likelihood of IPV victimization.^2,5,25 There are several mechanisms that may link bullying to an increased risk of IPV victimization. Based on Cognitive-Behavioral Theory, bullying can shape how individuals perceive social interactions and may lead individuals to internalize negative self-perceptions or views on control in relationships, making youth more vulnerable to tolerating or perpetrating abusive behaviors.^26–29 Repeated exposure to aggression can also lead to desensitization, making individuals more likely to engage in or tolerate violence.^26,28,30 Bullying can cause emotional distress, including anxiety, depression, and anger, and can disrupt the formation of healthy interpersonal attachments.^27,28 Ecological Systems Theory suggests that adolescents who are victimized by peers may be more likely to be in an environment where violence has become habituated or normalized making them less likely to leave violent relationships with a partner.²

This study adds to the evidence base by examining shorter-term longitudinal effects of bullying on IPV and different types of bullying and IPV victimization. The relationship between bullying and IPV is complex and different forms of violence may be interrelated and reoccur over time. The results did determine that prior experiences of bullying increased risk of IPV after accounting for the prior bullying and IPV victimization. The strength of this study is that it uses multiple measures to account for the complex and likely intertwined dynamic between these two forms of violence, adjusting for time-dependent confounding. The findings indicate that intervening, even in the short term, will have important consequences in terms of preventing IPV victimization and breaking the cycle of violence.

The results showed that that the relationship between peer bullying and IPV was stronger among those with high neighborhood disorder and nearly nonexistent in those with low disorder.

Therefore, high-disorder environments may be key targets for programs aimed at preventing bullying and IPV. In contrast, high community social cohesion weakened the bullying-IPV association. Social support is critical for preventing IPV, particularly in youth who may have less power to advocate for themselves in relationships. Discrimination and community resources were not moderators indicating that active positive community support may be more influential than negative experiences or simply increasing access to resources. Examples of interventions to increase community cohesion and decrease social disorder are youth sports programs, and the Safe Routes to School (SRTS) program in California, which improved infrastructure to encourage walking or biking to school.^31,32 Similar programs could be implemented in communities like Salinas to decrease bullying and prevent IPV.^31–33

Limitations

The study has several limitations. First, the phrasing of the bullying question did not distinguish intimate relationships and could capture IPV experiences. However, one would then expect to see collinearity among these two measures causing model convergence issues and making it difficult to see independent associations with future IPV. Hence, the results can be attributed to effects to peer bullying. Physical peer bullying was also assessed using only one item but was part of a validated scale. Second, IPV was not measured at baseline and did not assess severity/frequency of IPV. The authors were unable to control for IPV at baseline. Third, the study included only participants with three consecutive time points, and slight differences in education and food insecurity may limit generalizability. However, those excluded did not differ in bullying at enrollment or IPV during the study. Lastly, neighborhood variables were based on individual perception of the neighborhood. Future studies should use community-level measures of these constructs. Community factors like cohesion and neighborhood disorder may also differ in urban settings with contextual differences such as having less tight knit communities; results may not be generalizable to those settings.

Conclusions:

Physical and psychological peer bullying were linked to IPV victimization, even after accounting for prior experiences. Intervening early is crucial to prevent IPV and break the cycle of violence. The relationship between bullying and IPV victimization was stronger in youth who reported high neighborhood disorder, making these environments key targets for intervention. Additionally, high community social cohesion weakened the association between peer bullying and IPV, highlighting the importance of social support in preventing IPV, especially for youth who may have less autonomy and power in relationships.

Table 1.

Characteristics of analytic sample at enrollment. A Crecer: Salinas Teen Health Study, 2015–2019.

	N	%
Total	558	(100)
Age, years - mean, median (range)	13.3, 13	(12–15)
Female	294	(53)
Immigrant Generation
1st: born outside the US	68	(12)
2nd: US-born and at least one parent born outside the US	393	(70)
3rd+: both parents and youth US-born	88	(16)
US-born, generation unknown	9	(2)
Latinx	528	(95)
Mexican origin	495	(89)
Mother’s education
Less than high school	238	(43)
High school/GED	169	(30)
More than high school	133	(24)
(Unknown)	18	(3)
Living with
Both parents, at least part time	417	(75)
Mother only	124	(22)
Other	17	(3)
At least one parent works in agriculture	279	(50)
At least one parent moves for work during the year	83	(15)
Crowded housing conditions	337	(61)
Food insecurity (hunger), past 6 months	40	(7)
Receipt of government assistance past 6 monthsa	293	(53)
Experienced any discrimination	161	(27)
Social environment
Neighborhood disorder events in past year (range 0–11) – mean, median (IQR)	4.9, 5	2–7
Community social cohesion (range 1–4) – mean, median (IQR)	2.7, 2.8	2.4–3.0
Level of engagement in activities outside of school/home (range 0–9) – mean, median (IQR)	4.4, 4	2–6

GED= general equivalency diploma; US = United States; IQR= interquartile range

^aIncludes Medi-Cal, unemployment benefits, food stamps

Data availability:

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.