Introduction

Intimate Partner Violence (IPV) remains a severe problem with serious negative implications for individuals, families, and society. Previous researchers have found that around 42% of women and men experience physical violence over their lifetime, 20% of women and 8% of men experience sexual violence, and almost 50% of women and 45% of men experience psychological violence (Leemis et al., 2022). The rates of IPV are even higher among couples seeking therapy, with studies finding that as many as 50% of couples in clinical settings have a history of IPV (Greene & Bogo, 2002; Jose & O’Leary, 2009), and most Marriage and Family Therapists will be confronted with IPV at some point in their clinical careers (Blasko et al., 2007).

Unfortunately, evidence suggests that therapists sometimes work with couples where IPV is under-identified or present without awareness (Schacht et al., 2009; Stith et al., 1991). Also, community and clinical populations tend to underreport IPV (Szinovacz, 1983; Todahl et al., 2019). In addition, therapists inconsistently assess IPV (Nyame et al. 2013; Schacht et al., 2009; Todahl et al., 2008). However, regardless of whether screening occurs, couples experiencing violence are attending therapy (Blasko et al., 2007). Providing therapy when IPV is unidentified comes with risk (Todahl & Walters, 2011), such as placing violent couples at risk for future harm and neglecting a major factor contributing to relationship satisfaction (Stith et al., 2011). Although the American Association for Marriage and Family Therapy (AAMFT) does not have official standards for assessing and treating IPV, the ethical code requires that therapists protect clients from possible harm (2015). Thus, ethical therapy necessitates prioritizing client safety.

Given the potential consequences of IPV, including injury or death (Breiding et al., 2014), depression and suicidality (Ulloa & Hammett, 2016), alcohol and substance use (Gehring & Vaske, 2017), relationship dissatisfaction and instability (Simmons et al., 2018), and population level costs of $3.6 trillion across victims’ lifetimes in the United States (Peterson et al., 2018), it is important to increase our understanding of factors that influence reporting of IPV in clinical settings. Therefore, the purpose of the present study was to examine whether clients who indicated violence in their relationship via assessment paperwork and reported violence as a primary clinical concern differed in aspects of their clinical presentation from clients who indicated violence via assessment paperwork but did not report IPV as a clinical concern. The findings from the present study have the potential to significantly contribute to the field’s understanding of how different clinical factors might affect the extent which clients consider IPV to be a core relational concern. This would then have important clinical implications regarding clinical factors for therapists to consider when assessing IPV.

Literature Review

Previous researchers have found multiple barriers to reporting IPV to healthcare and mental health professionals. These include whether victims feel the environment is safe to disclose (Heron & Eisma, 2021), fear of retaliation (Heron et al., 2022; Straus & Kaufman-Kantor, 1994), fear of being met with disbelief or minimization (Walker et al., 2020), blaming their partner(s) for the abuse (Chan, 2009), financial dependence (Szinovacz & Egley, 1995), normalizing the abuse (Kimmel, 2002), and victim-blaming (Giles, 2004).

Although there is research exploring factors affecting reporting violence in general, there is a dearth of research on factors that contribute to clients reporting violence as a clinical concern to their therapists. However, the following clinical variables derived from the present study’s sample have been found to be associated with the incidence of IPV in general: relationship satisfaction, mental health, substance use, post-traumatic stress, and childhood violence exposures. Since these factors have been associated with the incidence of IPV, they will be reviewed as they may also contribute to clients reporting violence as a clinical concern in therapy.

Relationship Satisfaction

In their meta-analytic review, Stith and colleagues (2008) found that marital satisfaction is significantly negatively associated with IPV. Also, researchers have found a bidirectional relationship between IPV and relationship satisfaction. Simmons and colleagues (2018) found that IPV has a detrimental effect on relationship satisfaction and found that lower levels of relationships satisfaction can contribute to increased incidence of IPV. Although increases in IPV victimization over time is associated with decreased relationship satisfaction for both men and women, increases in IPV perpetration over time was associated with higher relationship satisfaction in women but was not significantly associated for men. This may reflect the greater control perpetrators feel in the relationship by using violence (Ulloa & Hammett, 2015).

Psychopathology

Additionally, meta-analyses have found that mental health disorders are significant correlates of IPV for both men and women (Spencer et al., 2019). Specifically, previous literature indicates that psychopathology is an influencing factor and outcome for both IPV victimization and perpetration (Chandra et al., 2009; Karakuła-Juchnowicz et al., 2017; Lawrence et al., 2009; Shorey et al., 2012). For example, Chandra and colleagues (2009) found that women reporting IPV had higher scores for depression compared to women without a history of IPV. Shorey and colleagues (2012) found high levels of anxiety in male perpetrators of IPV. Further, those who were involved in psychological IPV experienced an increase in depression and anxiety (Lawrence et al., 2009), while victims of physical IPV experienced an increase in depression only (Karakuła-Juchnowicz et al., 2017).

Substance Use

In their meta-analytic review, Stith and colleagues (2004) found illicit drug use to be a correlate of IPV. Also, research has indicated that IPV victims often report engaging in substance use, and a substantial portion of IPV perpetration occurred in conjunction with perpetrator substance use (Lund, 2014). Additionally, researchers found that 47% of women entering substance abuse treatment were victims of IPV at some point in their lives (Rivera et al., 2015; Schneider et al., 2009). Substance use in this population has often been cited as a coping mechanism (Rivera et al., 2015).

Post-Traumatic Stress Disorder

In their meta-analytic review, Spencer and colleagues (2019) found post-traumatic stress disorder (PTSD) to be a significant correlate of IPV, with PTSD being associated with perpetration in both men and women (Bell & Orcutt, 2009; Hahn et al., 2015; Kendra et al., 2012; Miles-McLean et al., 2021; Price et al., 2014). Many hypothesize this association is due to PTSD arousal symptoms such as “angry outbursts, verbal or physical aggression, reckless behavior, and hypervigilance” (American Psychiatric Association, 2013, p. 272). However, few researchers have examined if PTSD symptom severity is associated with the likelihood of partners’ reporting IPV (Chandra et al., 2009; Marshall et al., 2021). According to Marshall and colleagues (2021), greater severity of PTSD symptoms with male and female partners were associated with females reporting IPV in the relationship. One possibility for the relationship between PTSD severity and IPV reporting is that those with higher susceptibility to PTSD symptoms are more sensitive to perceived threats (Marshall et al., 2021), and women may be more likely to report IPV to seek support (Sayem et al., 2015).

Childhood Violence Exposures

Previous meta-analyses have found childhood abuse to be a significant correlate of IPV (Smith-Marek et al., 2015). Physical child abuse is especially associated with future long-term risks of IPV (Richards et al., 2017; Yan & Karatzias, 2020). Also, Butler and colleagues (2020) found that compared to individuals who experienced no abuse in childhood, those who experienced one form of abuse were more than twice as likely to experience physical abuse in the past year and three times as likely to have experienced IPV and/or sexual violence since age 16. It is hypothesized that child abuse normalizes violence as an acceptable form of conflict management, and that normalization may transfer to adult relationships (Messinger et al., 2021).

Purpose of Study

IPV is common in clinical settings but is often underreported as a primary clinical issue for treatment. If therapists are unaware of IPV, then they could both contribute to future violent incidents and unintentionally undermine the effectiveness of treatment. Thus, it is crucial for therapists to understand potential differences between clients who are experiencing violence and report it as a clinical concern versus those who are experiencing violence but do not report it as a clinical concern in therapy. Previous researchers have found multiple variables that are connected to the incidence of IPV in general, including relationship satisfaction, psychopathology, substance use, post-traumatic stress, and childhood violence exposures. The purpose of this study was to examine these variables to identify potential differences between clients in violent relationships that reported violence as a clinical concern versus clients in violent relationships that did not report IPV as a clinical concern to their therapists. Thus, this study contributes to the literature by expanding research from risk factors for IPV in general to factors that impact clinical populations viewing IPV as a concern worth addressing in therapy.

Methods

Participants

Participants in this study were 1,208 individuals who sought couple or family therapy at a university training clinic located in the southwestern United States. Data collected from 2,172 client constellations that included an adult intimate relationship were reviewed to identify those who indicated IPV in assessment material. All participants were 18 years of age or older and consented to the use of their clinical data for research. All participants were in a relationship where at least one individual indicated violence on the CTS-Short Form.

51% of the sample was female, and 49% of the sample was male. Most participants were white (66.1%), followed by Hispanic (22.6%), biracial (3.7%), and African American (2.7%). The majority of participants identified as heterosexual (92.8%), followed by gay/lesbian (3.6%), and bisexual (2.3%). The mean age of participants was 32 years. The average income level of participants was $20,000 to $29,999, and the highest level of educational attainment was some college. The average length of a romantic relationship was 7.5 years, with most participants being in their first marriage (42.5%).

Procedure

Previously collected de-identified data from years 2008 to 2017 were used in this study. The authors confirm that all data generated or analyzed in this study are included in this article. The data were collected as part of the standard intake process. Before the first therapy session, clients were asked to complete an intake packet that included several written assessments and prompts asking clients to report the concerns they wish to be addressed in therapy. Clients were provided the option to consent to de-identified data being used for research purposes.

The first step of data cleaning retained only clients who presented for therapy with an intimate partner, either as part of a couple or as part of a family constellation. The next step was to review the clinical assessment data and only retain couples where at least one partner indicated IPV on the CTS-Short Form. IPV was defined using the physical assault, psychological aggression, and sexual coercion subscales of the CTS-Short Form. The physical assault and sexual coercion subscales were combined to create an overarching physical violence subscale. Only data from participants in couples where at least one partner indicated IPV were retained for analysis.

Measures

Intimate Partner Violence as a Clinical Concern

Reporting IPV as a clinical concern at intake was measured by participant self-report on the adult intake assessment packet. Participants were asked to select problems that were a concern to them in their intimate relationship, with “physical violence” and “emotional abuse” listed as options. Participants were also asked to respond to an open-ended question on the intake assessment, “What problem(s) would you like help with in therapy?” A coding team consisting of four researchers independently reviewed the intake assessment packets and coded each participant as having reported ‘no violence as a clinical concern,’ ‘psychological violence only as a clinical concern,’ or ‘physical violence as a clinical concern.’ Sexual violence and physical violence reports were combined and conceptualized as physical violence. After independently coding the assessment packet, the coders met to review the codes and resolve any disagreements. Participants who reported both psychological violence and physical violence as a clinical concern on the open-ended question were also coded as ‘physical violence as a clinical concern.’ The results of the coding were used to create corresponding categorical variables.

Intimate Partner Violence

IPV was measured using the Conflict Tactics Scale 2 Short Form (CTS2S; Straus & Douglas, 2004). The CTS2S is a condensed version of the Revised Conflict Tactics Scale (CTS2; Straus et al., 1996) and is a 20-item self-report measure assessing conflict between partners. The five subscales of the CTS2S include: negotiation, psychological aggression, physical assault, sexual coercion, and injury. A sample item includes, “My partner insulted or swore or shouted or yelled at me.” Participants responded using an 8-point scale: once in the past year (1), twice in the past year (2), 3–5 times in the past year (3), 6–10 times in the past year (4), 11–20 times in the past year (5), > 20 times in the past year (6), none in the past year but it did happen before (7), or this never happened (0). The CTS2S was used as a screening measure to determine whether the participants were in relationships where IPV had occurred. If at least one partner in a couple reported psychological or physical violence on the CTS2S, then responses from both partners were retained for analysis. For the CTS2S measure, Straus and Douglas (2004) indicated good construct validity. The CTS2S demonstrated good reliability (α = 0.81) in this study.

Psychopathology

Psychopathology was measured using the Global Severity Index (GSI) from the Brief Symptom Inventory (BSI; Derogatis, 1993), with a higher score indicating more distress. The BSI is a shortened version of the SCL-90-R (Derogatis, 1983). The BSI is a 53-item scale that encompasses nine symptom dimensions, including somatization, obsessive-compulsive, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation, and psychoticism. The BSI also encompasses three global indicators of psychopathology, including the GSI, positive symptom distress index, and positive symptom total. The GSI combines the number of symptoms with their distress intensity to create an overarching measure of psychopathology. A sample item includes, “How much have you been bothered or distressed by nervousness or shakiness inside?” Participants responded using a 5-point Likert scale, ranging from not at all (0) to extremely (4). The nine dimensions of psychopathology demonstrate good reliability, with Cronbach’s alphas ranging from 0.71 to 0.85 (Derogatis, 1975). The three global indicators of psychopathology demonstrate good test-retest reliability, with alphas ranging from 0.87 to 0.90 (Derogatis, 1975). The BSI demonstrated good reliability (α = 0.96) in this study.

Relational Adjustment

Relational adjustment was measured using the Revised Dyadic Adjustment Scale (RDAS; Busby et al., 1995). The RDAS is a shortened revision of the DAS. The RDAS is a fourteen-item questionnaire. There are seven first-order subscales in this measure: decision-making, leisure, values, affection, stability, conflict, activities, and discussion. These first-order subscales are combined to form the three second-level subscales that form the highest concept of relational adjustment: consensus, satisfaction, and cohesion (Busby et al., 1995). A sample item includes, “How often do you discuss or consider divorce, separation, or terminating your relationship?” Relationship adjustment was estimated by taking the average of all the items. This scale is scored using a 6-point Likert scale in sections one through three and a 5-point Likert scale in section four. Consensus is measured in section one, ranging from always disagree (0) to always agree (5). Satisfaction is measured in section two, ranging from never (0) to all the time (5). Cohesion is measured in sections three and four, ranging from never (0) to more than once a day (5) in section three and never (1) to every day (5) in section four. A higher number indicates better relational adjustment. Busby and colleagues (1995) found good reliability for the RDAS (α = 0.90). The RDAS demonstrated good reliability (α = 0.86) in this study.

Safety

Individuals’ perceptions of safety were measured using the Self-Assessment of Future Events Scale (SAFE; Smith et al., 2013). The SAFE is a 15-item scale measuring individuals’ perception of the risk for future physical violence, psychological violence, and controlling behaviors in relationships. There are three subscales measured within the SAFE: Verbal/Psychological Safety, Control, and Physical Safety. A sample item includes, “My partner will be physically aggressive towards me.” The SAFE is scored using a 6-point Likert, ranging from extremely unlikely (1) to extremely likely (6). A total score is derived from the sum of all items. Higher total scores are a result of decreased perceptions of safety in the relationship. The SAFE demonstrated good internal reliability (α = 0.88; Smith et al., 2013). The SAFE demonstrated good reliability (α = 0.88) in this study.

Post-Traumatic Stress

Post-traumatic stress was measured using the MPSS-SR (Modified PTSD Symptom Scale; Resnick et al., 1996). The MPSS-SR measures post-traumatic stress via a 17-item self-report test that asks respondents to indicate both severity and frequency of PTSD symptoms. A sample item includes, “Have you had repeated or intrusive upsetting thoughts or recollections of the event(s)? Frequency: not at all (0), once a week or less (1), 2–4 times a week (2), 5 or more times a week (3). Severity: not at all disturbing (A), a little bit disturbing (B), moderately disturbing (C), quite a bit distressing (D), and extremely distressing (E).” Falsetti (1997) tested reliability of the MPSS-SR using Cronbach’s alpha for clinical sample (frequency scale α = 0.93; severity scale α = 0.94). The MPSS-SR demonstrated good reliability (α = 0.96) in this study.

Substance Use

Substance use was measured by asking clients to report, “How often do you drink?” and “How often do you use drugs?” Participants responded using a 6-point Likert scale: never (1), less than once a month (2), about once a week (3), 2 to 3 days per week (4), 4 to 6 days per week (5), or daily (6). A higher number indicated a higher frequency of substance use.

Childhood Maltreatment

Childhood emotional, physical, and sexual abuse was measured by self-report. Individuals answered “yes” or “no” to prompts asking if they have experienced “emotional abuse in childhood,” physical abuse in childhood,” and “sexual abuse in childhood.”

Analysis Plan

The authors ran independent sample ANOVAs to identify whether there were significant differences in the BSI (GSI), total RDAS score, SAFE, total MPSS-SR score, alcohol frequency, and drug frequency for clients who indicated no violence, psychological violence only, and physical violence as clinical concerns. The authors first ran the ANOVAs separately for the male and female partners, compared the results, and where the pattern was the same, we ran a final ANOVA for the total sample. This allowed us to identify gender differences in the results and maintain parsimony in reporting the patterns. See Table 1 for correlations among study variables.

Table 1 Correlation matrix (n = 1,208)
Full size table
Table 2 Distinguishability test (n = 1,208)
Full size table
Table 3 Dyadic multinominal logistic regression analyses- psychological violence only (n = 1,208)
Full size table
Table 4 Dyadic multinominal logistic regression analyses- physical violence (n = 1,208)
Full size table

For the ANOVAs, the authors used the Levene’s test for equality of variance. The BSI (Levene Statistic (2, 1204), = 3.51, p < .05), SAFE (Levene Statistic (2960), = 7.97, p < .001), MPSS-SR (Levene Statistic, (2, 1185), = 5.84, p < .01), and Drug Frequency (Levene Statistic (2, 1167), = 14.20, p < .001) failed the homogeneity of variance tests. In response, the authors ran Welch’s ANOVAs for these variables to provide a robust test of equality of means. Also, the authors ran Games-Howell Post Hoc tests on these variables as well. The Games-Howell Post Hoc test was conducted as it can be used when there are unequal sample sizes and variances between the different groups (Ruxton & Beauchamp, 2008). The RDAS (Levene Statistic (2, 1202) = 0.99, p = .37) and alcohol frequency (Levene Statistic (2, 1185) = 0.96, p = .38) did not violate the homogeneity of variance tests. For the variables that did not violate homogeneity of variance, the authors ran Fisher’s Least Significant Difference (LSD) Post Hoc tests.

Next, the authors used Chi-Square tests to measure the relationships between childhood emotional, physical, and sexual abuse and reporting IPV as a clinical concern. We selected Chi-Square tests to determine whether a relationship existed between the categorical variables of childhood abuse and reporting IPV as a clinical concern. For these three variables, the authors utilized Fisher’s approach for the Post Hoc analysis (Shan & Gerstenberger, 2017).

Before running the dyadic analysis, we examined the data for nonindependence of male and female partners. Finally, a dyadic multinomial logistic regression was run using the MPSS-SR score, BSI score, total RDAS score, SAFE, total childhood maltreatment (emotional abuse, physical abuse, and sexual abuse), alcohol frequency, and drug frequency for both male and female partners to identify the odds of reporting violence (either psychological only or physical) compared to the odds of not reporting violence. We tested whether the actor and partner paths were distinguishable from one another by constraining each to be estimated as the same for both partners and then conducting a Chi-Square difference test. Distinguishability between the partners was identified where the constrained model worsened the fit significantly (Burnham & Anderson, 2004). The final model maintained constraints where the partners were indistinguishable and freely estimated values where the partners were distinguishable. See Tables 3 and 4 for a presentation of the dyadic multinomial logistic regression results.

Results

Demographics

Most individual participants did not report IPV as a clinical concern (77.2%), despite being a member of a couple where at least one of the partners indicated IPV had occurred in the relationship. Psychological violence was reported as a clinical concern by 13.4% of participants, and physical violence was reported as a clinical concern by 9.4% of participants. Female participants reported violence as a clinical concern at a greater rate compared to male participants, with 15.2% of females reporting psychological violence and 10.4% reporting physical violence as clinical concerns. Around 11% of male participants reported psychological violence and 8.4% of male participants reported physical violence as clinical concerns. The mean score for the GSI was 0.88, meaning that participants scored close to “a little bit” when indicating the frequency of psychopathology symptoms. The mean score for the SAFE was 35.25, with possible scores ranging from 15 to 90, indicating a greater level of perceived safety. Almost 70% of participants reported significant relationship distress, reaching a cutoff score of at least 48, and 25.1% of the sample met the cutoff score for PTSD, reaching a cutoff score of 29. Around 14% of participants reported experiencing childhood emotional abuse, 8% reported childhood physical abuse, and 8% reported childhood sexual abuse. Most participants reported infrequent alcohol consumption (around once a month, 35.4%). Most participants reported never engaging in drug use (82.0%).

Psychopathology

There were no differences in the pattern of significance in the analysis comparing the BSI across the report groups between male and female participants. The total sample ANOVA identified that the three groups differed from each other regarding psychopathology, Welch (2, 212.73) = 33.34, p < .001. The effect size was η2 = 0.06. First, the Games-Howell post hoc test identified that individuals who reported psychological violence only as a clinical concern had significantly higher psychopathology compared to individuals who did not report violence as a clinical concern (MD = − 0.36, p < .01). Second, individuals who reported physical violence as a clinical concern had significantly higher psychopathology than individuals who did not report violence as a clinical concern (MD = − 0.45, p < .001). There was no significant difference in psychopathology between individuals who reported physical violence as a clinical concern and those who reported psychological violence only as a clinical concern (MD = − 0.09, p = .61).

Safety

There were no differences in the pattern of significance in the analysis comparing the SAFE across the report groups between male and female participants. The total sample ANOVA identified that the three groups differed from each other regarding perceptions of safety, F(2, 164.50) = 139.19, p = < 0.001. The effect size was η2 = 0.27. First, the Games-Howell post hoc test identified that individuals who reported psychological violence only as a clinical concern had greater perceived lack of safety compared to individuals who did not report violence as a clinical concern (MD = -12.01, p < .001). Second, individuals who reported physical violence as a clinical concern had greater perceived lack of safety than individuals who did not report violence as a clinical concern (MD = -20.65, p < .001) and who reported psychological violence only as a clinical concern (MD = -8.63, p < .001).

Relational Adjustment

There were no differences in the pattern of significance in the analysis comparing the RDAS across the report groups between male and female participants. The total sample ANOVA identified that the three groups differed from each other regarding relational adjustment, F(2, 1203) = 89.55, p < .001. The effect size was η2 = 0.13. First, the Fisher’s LSD post hoc test identified that individuals who reported psychological violence only as a clinical concern had significantly lower relational adjustment compared to individuals who did not report violence as a clinical concern (MD = 7.95, p < .001). Second, individuals who reported physical violence as a clinical concern had significantly lower relational adjustment than individuals who did not report violence as a clinical concern (MD = 10.41, p < .001). When examining differences in the total sample rather than by gender, one additional post hoc test detected a significant difference; individuals who reported physical violence as a clinical concern also had significantly lower relational adjustment than those who reported psychological violence only as a clinical concern (MD = 2.46, p < .05).

Post-Traumatic Stress

There were no differences in the pattern of significance in the analysis comparing the MPSS-SR across the report groups between male and female participants. The total sample ANOVA identified that the three groups differed from each other regarding post-traumatic stress, F (2, 206.95) = 41.56, p < .001. The effect size was η2 = 0.08. First, the Games-Howell post hoc test identified that individuals who reported psychological violence only as a clinical concern had significantly greater post-traumatic stress compared to individuals who did not report violence as a clinical concern (MD = -14.36, p < .001). Second, individuals who reported physical violence as a clinical concern had significantly higher post-traumatic stress than individuals who did not report violence as a clinical concern (MD = -19.32, p < .001). There was no significant difference between individuals who reported physical violence as a clinical concern and those who reported psychological violence only as a clinical concern (MD = -4.96, p = .30).

Substance Use

When examining male participants separately, there was a significant difference in frequency of alcohol use between males who reported psychological violence only as a clinical concern and those who reported physical violence as a clinical concern, F(2, 575) = 2.54, p < .05. The Fisher’s LSD post hoc test identified that male participants who reported physical violence as a clinical concern had significantly greater alcohol frequency than male participants who reported psychological violence only as a clinical concern (MD = 0.59, p < .05). There were no significant differences between the groups when examining female participants separately.

There were no differences in the pattern of significance in the analysis comparing drug frequency across the report groups between male and female participants. The total sample ANOVA identified that the three groups differed from each other regarding drug frequency, F(2, 199.85) = 4.25, p < .05. The effect size was η2 = 0.01. First, the total sample Games-Howell post hoc test identified that individuals who reported physical violence as a clinical concern had significantly more frequent drug use than individuals who did not report violence as a clinical concern (MD = -0.39, p < .05). There was no significant difference in frequency of drug use between individuals who reported psychological violence only as a clinical concern and those who did not report violence as a clinical concern (MD = -0.06, p = .82). There was no significant difference in frequency of drug use between individuals who reported psychological violence only as a clinical concern and those who reported physical violence as a clinical concern (MD = -0.34, p = .09).

Childhood Maltreatment

The first chi-square test indicated a significant relationship between experiencing emotional abuse in childhood and reporting IPV as a clinical concern, X2 (2, N = 1202) = 33.30, p = < 0.001. The second test identified a significant relationship between experiencing physical abuse in childhood and reporting IPV as a clinical concern X2 (2, N = 1202) = 41.02 p = < 0.001. The third test identified a significant relationship between experiencing sexual abuse in childhood and reporting IPV as a clinical concern, X2 (2, N = 1203) = 20.76, p = < 0.001.

Three Fisher post hoc tests were conducted to identify the direction of the significant relationships. A Bonferroni adjustment was implemented to account for the increased potential for type one error. The significance level was divided by six, leading to an adjusted significance level of p < .0083. The first test identified that male participants were more likely to report psychological violence (adjusted residual = 3.6, p < .001) as a clinical concern if they had experienced childhood emotional abuse and less likely to report IPV as a clinical concern if they had not experienced childhood emotional abuse (adjusted residual = -4.3, p < .001). Also, the first test identified that female participants were more likely to report physical violence (adjusted residual = 2.8, p < .01) as a clinical concern if they had experienced childhood emotional abuse and less likely to report IPV as a clinical concern if they had not experienced childhood emotional abuse (adjusted residual = -3.7, p < .001). The second test identified that male participants were more likely to report psychological (adjusted residual = 4.7, p < .001) and physical violence (adjusted residual = 3.8, p < .001) as a clinical concern if they had experienced childhood physical abuse and less likely to report IPV as a clinical concern if they had not experienced childhood physical abuse (adjusted residual = -6.4, p < .001). Also, the second test identified that female participants were more likely to report physical violence (adjusted residual = 3.2, p < .01) as a clinical concern if they had experienced childhood physical abuse and less likely to report IPV as a clinical concern if they had not experienced childhood physical abuse (adjusted residual = -2.7, p < .01). The third test identified that male participants were more likely to report psychological violence (adjusted residual = 2.9, p < .01) as a clinical concern if they had experienced childhood sexual abuse and less likely to report IPV as a clinical concern if they had not experienced childhood sexual abuse (adjusted residual = -3.2, p < .01). Also, the third test identified that female participants were less likely to report IPV as a clinical concern if they had not experienced childhood sexual abuse (adjusted residual = -2.8, p < .01).

Dyadic Multinomial Logistic Regression Analyses

Using a test for independence of the data, we found that all predictors and the report of violence as a clinical concern outcome variable were all significantly associated, and therefore non-independent. All independent variable (BSI [r = 2.92. p < .001], RDAS [r = .60, p < .001], MPSS-SR [r = .36, p < .001], SAFE [r = .50, p < .001], childhood psychological abuse [r = .15, p < .001], childhood physical abuse [r = .18, p < .001], childhood sexual abuse [r = .14, p < .001], frequency of alcohol use [r = .43, p < .001], and frequency of drug use [r = .38, p < .001]) had significant correlations between the partners that ranged from small to large correlations (Cohen, 1992). The report of violence as a clinical concern for each partner at intake was also significantly associated χ2 (4) = 147.48, p < .001.

We ran the dyadic multinomial logistic regression to identify the comparison model, and the individually constrained actor and then partner paths to determine if there was distinguishability between the partners on the strength of those associations. We identified three distinguishable actor paths and one distinguishable partner path in the analysis (see Table 2). The final model had a Loglikelihood value = -405.36, AIC = 898.72, BIC = 1075.54.

Three significant predictors emerged when comparing males who did not report violence as a clinical concern to males who did report psychological violence as a clinical concern (Table 3). A one-unit increase in males’ relationship satisfaction was associated with 4% lower odds of males reporting psychological violence as a clinical concern (OR = 0.96). In other words, the more satisfied men were in their relationships the less likely they were to report psychological violence as a clinical concern. A one-unit increase in males’ perceived lack of safety was associated with 8% greater odds of males reporting psychological violence as a clinical concern (OR = 1.08). Thus, men who felt unsafe were more likely to report psychological violence as a clinical concern. A one unit-increase in females’ perceived lack of safety was associated with 7% greater odds of males reporting psychological violence as a clinical concern (OR = 1.07). Therefore, female partners’ perceived lack of safety contributed to their male partners being more likely to report psychological violence as a clinical concern. In the final model, where indistinguishable paths were constrained, females’ relationship satisfaction no longer emerged as a significant predictor of males reporting psychological violence as a clinical concern (OR = 1.03).

Five significant predictors emerged when comparing males who did not report violence as a clinical concern to males who did report physical violence as a clinical concern (Table 4). In the final model where indistinguishable paths were constrained, males’ relationship satisfaction emerged as an additional significant predictor of males reporting physical violence as a clinical concern. A one-unit increase in males’ relationship satisfaction was associated with 4% lower odds of males reporting physical violence as a clinical concern (OR = 0.96). In other words, the more satisfied men were in their relationships the less likely they were to report physical violence as a clinical concern. A one-unit increase in males’ perceived lack of safety was associated with 7% greater odds of males reporting physical violence as a clinical concern (OR = 1.07). Thus, men who felt unsafe were more likely to report physical violence as a clinical concern. A one-unit increase in males alcohol frequency was associated with 33% lower odds of males reporting physical violence as a clinical concern (OR = 0.67). Therefore, men who consumed alcohol more frequently were less likely to report physical violence as a clinical concern. Males experiencing childhood physical abuse was associated with 269% greater odds of males reporting physical violence as a clinical concern (OR = 3.69). This indicates that men who had experienced childhood physical abuse were more likely to report physical violence as a clinical concern. A one-unit increase in females’ alcohol frequency was associated with 62% greater odds of males reporting physical violence as a clinical concern (OR = 1.62). Thus, female partners’ alcohol frequency contributed to their male partners being more likely to report physical violence as a clinical concern.

Three significant predictors emerged when comparing females who did not report violence as a clinical concern to females who did report psychological violence as a clinical concern (Table 3). In the final model where indistinguishable paths were constrained, females’ relationship satisfaction emerged as an additional significant predictor. A one-unit increase in females’ relationship satisfaction was associated with 4% lower odds of females reporting psychological violence as a clinical concern (OR = 0.96). This indicates that the more satisfied women were in their relationship the less likely they were to report psychological violence as a clinical concern. A one-unit increase in females’ perceived lack of safety was associated with 12% greater odds of females reporting psychological violence as a clinical concern (OR = 1.12). Thus, the more unsafe women felt the more likely they were to report psychological violence as a clinical concern. A one-unit increase in males’ perceived lack of safety was associated with 7% greater odds of females reporting psychological violence as a clinical concern (OR = 1.07). Therefore, males’ perceived lack of safety contributed to their female partners being more likely to report psychological violence as a clinical concern. In the final model where indistinguishable paths were constrained, men’s experience of childhood emotional abuse no longer emerged as a significant predictor of women reporting psychological violence as a clinical concern (OR = 1.68); women’s post-traumatic stress no longer emerged as a significant predictor of their own report of psychological violence as a clinical concern (OR = 1.02).

Two significant predictors emerged when comparing females who did not report violence as a clinical concern to females who did report physical violence as a clinical concern (Table 4). A one-unit increase in females’ relationship satisfaction was associated with 4% lower odds of females reporting physical violence as a clinical concern (OR = 0.96). This indicates that the more satisfied women were in their relationships the less likely they were to report physical violence as a clinical concern. A one-unit increase in females’ perceived lack of safety was associated with 7% greater odds of females reporting physical violence as a clinical concern (OR = 1.07). Thus, the more unsafe women felt the more likely they were to report physical violence as a clinical concern. Finally, in the final model where indistinguishable paths were constrained, females’ post-traumatic stress no longer emerged as a significant predictor of their own report of physical violence as a clinical concern (OR = 1.01).

The results also indicate that three actor effects and one partner effect significantly differed in the strength of associations between male and female participants (Table 2). Starting with the actor effects, perceived lack of safety for men demonstrated greater odds of reporting psychological violence as a clinical concern compared to women. Second, experiencing childhood physical abuse was only a significant predictor for reporting physical violence as a clinical concern for men. Third, alcohol frequency was only a significant predictor for reporting physical violence as a clinical concern for men. For the partner effect, women’s alcohol frequency predicted men’s report of physical violence as a clinical concern, but not the opposite.

Discussion

Overall, the results from the ANOVA and chi-square analyses demonstrated that participants who reported psychological violence only and physical violence as clinical concerns were more distressed than those who did not report violence as a clinical concern. This remained true for male and female participants across all variables, except for alcohol and drug frequency affecting reporting violence as a clinical concern for male participants only. Thus, it appears that men and women demonstrated similar patterns of significance regarding the variables that affect reporting violence as a clinical concern. This greater level of distress is not surprising as previous literature demonstrates connections between IPV and psychopathology (Chandra et al., 2009; Karakuła-Juchnowicz et al., 2017; Lawrence et al., 2009; Shorey et al., 2012), relationship satisfaction (Hammett, 2017; Simmons et al., 2018), post-traumatic stress (Hahn et al., 2015; Miles-McLean et al., 2021; Price et al., 2014), substance use (Lund, 2014; Rivera et al., 2015; Schneider et al., 2009), and childhood maltreatment (Richards et al., 2017; Yan & Karatzias, 2020). Although further research is needed to confirm the following, it is logical to assume that people who are experiencing greater distress as a result of IPV may be more likely to report IPV as a clinical concern. It may also imply that clients experiencing lower levels of distress may not view IPV as a primary concern in their relationships, lending credence to the value of comprehensive screening for IPV in all clinical couples, regardless of perceived distress.

The dyadic multinomial logistic regression analyses indicated that there were both actor and partner effects regarding males and females reporting psychological violence and physical violence as clinical concerns. Starting with reporting psychological violence as a clinical concern, relationship satisfaction and perceived safety were the only significant predictors. Males’ relationship satisfaction and perceived safety emerged as significant predictors for their own report of psychological violence as a clinical concern (i.e., actor effect). Further, females’ perceived safety emerged as a significant predictor of males’ report of psychological violence as a clinical concern (i.e., partner effect). Also, females’ relationship satisfaction and perceived safety emerged as significant predictors for their own report of psychological violence as a clinical concern (i.e., actor effect); males’ perceived safety emerged as a significant predictor for females’ report of psychological violence as a clinical concern (i.e., partner effect).

For reporting physical violence as a clinical concern, relationship satisfaction, perceived safety, childhood physical abuse, and alcohol frequency emerged as significant predictors. Relationship satisfaction and perceived safety emerged as significant predictors for both males’ and females’ own reports of physical violence as a clinical concern (i.e., actor effects). Further, males experiencing childhood physical abuse emerged as a significant predictor for their own report of physical violence as a clinical concern (i.e., actor effect). Additionally, males’ alcohol frequency and females’ alcohol frequency predicted males’ report of physical violence as a clinical concern (i.e., actor and partner effects).

An interesting finding from the dyadic multinomial logistic regression analyses was that higher relationship satisfaction was associated with decreased odds of reporting psychological and physical violence as clinical concerns for both men and women. Although further research is needed to confirm the following, the authors hypothesize that those who are more satisfied with their relationships might not view violence as a primary concern because the more content they are overall with the relationship the less likely they are to worry about behaviors that might not be viewed as relevant to the quality of the relationship. Another interesting finding is that men who experienced childhood physical abuse were 269% more likely to report physical violence as a clinical concern. This finding is in contrast to some researchers who have found that experiencing childhood abuse decreases the likelihood of people finding violence problematic in intimate relationships (Messinger et al., 2021). It is possible that male participants in this study demonstrated a heightened awareness of the problematic nature of violence in their relationships, potentially due to their own experience of physical abuse in childhood.

The presence of actor and partner effects highlights that both individual characteristics influence participants’ own reports of violence as a clinical concern and individuals’ partners’ characteristics also influence each other’s reports of violence as a clinical concern. Thus, it appears that relational dynamics affect the perception of violence as a clinical concern. This suggests that reports of violence as a clinical concern should be examined in context of the relationship, not just the individual. Finally, finding that relationship satisfaction, perceived safety, alcohol frequency, and childhood physical abuse emerged as significant predictors of reporting violence as a clinical concern suggests that these factors should be examined in future studies.

Overall, the dyadic multinomial logistic regression results provide initial evidence that higher levels of distress regarding lack of safety, childhood physical abuse, and more frequent alcohol consumption in women affect the reporting of violence as a clinical concern. It is also possible that individuals with more severe levels of IPV are more likely to exhibit distress. Also, the dyadic multinomial logistic regression results tentatively indicate that being more satisfied in one’s intimate relationship might reduce partners’ perspective that IPV is a primary concern in the relationship. Finally, the significant predictors for reporting psychological violence versus physical violence as clinical concerns varied. Thus, these findings suggest that therapists should consider how factors contributing to distress and satisfaction in the relationship might affect reporting violence as a clinical concern and how they might do so in different ways.

Clinical Implications

Previous researchers have found that MFTs inconsistently utilize systematic IPV screening procedures (Schacht et al., 2009). This is concerning as many couples in therapy have a history of IPV in their relationship (Greene & Bogo, 2002). Also, most couples do not report IPV to their therapists (Greene & Bogo, 2002). Even if clients do indicate violence in their relationships, this does not mean they will report it as a concern that they want to address in therapy. For example, only 22.8% of our sample reported violence as a clinical concern, even though our entire sample had indicated violence in their relationship on the CTS2S. Thus, if most couples do not report violence, and MFTs inconsistently screen for IPV, therapists will fail to identify some cases where IPV is present. This places these couples at risk for future violence and undermines the effectiveness of therapy. Also, even if clients do indicate experiencing violence in their relationship, they might unintentionally minimize its importance by not viewing it as a concern they want to address in therapy. Thus, this might inhibit therapists’ work to eliminate violence in the relationship because these clients do not prioritize this goal. Overall, it is imperative for therapists to understand factors that may contribute to reporting IPV as a clinical concern to effectively address the cessation of violence in couples experiencing IPV.

Systematic screening should include both written assessments and separate interviews with all partners (Stith et al., 2011). For further information on how to systematically screen for IPV, please refer to Stith and colleagues (2011) and Todahl and colleagues (2020). Also, developing safety plans at the onset of therapy and assessing if couple’s therapy is a safe option will help improve the safety of couples experiencing violence (Bradford, 2010; Lechtenberg et al., 2015). Along with the necessity of systematic screening, these results also highlight the importance of therapists not contributing to the potential minimization of violence from couples who report minimal levels of distress. It is crucial that therapists always prioritize safety, even if couples experiencing IPV do not initially perceive the violence as a major concern.

Limitations and Future Directions

Our data relies on self-report intake paperwork to determine whether couples reported IPV as a clinical concern. However, it is possible that clients could have reported IPV as a clinical concern in later sessions. Therefore, it is possible that some couples experiencing IPV may have been inadvertently excluded from the study and could differ in some meaningful way from the participants included in the analyses. In terms of external validity, our sample was taken from a university training clinic. Therefore, the participants in our sample may not be representative of the general population, thus limiting generalizability. Additionally, there may have been other variables that our study did not account for that may have contributed to a partner’s report of IPV as a clinical concern. Therefore, future research looking at other variables is needed in order to gain a more holistic picture of contributing factors to reporting IPV as a clinical concern.

Conclusion

In conclusion, our study contributes to existing literature by identifying factors that contribute to clients who are experiencing IPV to report it as a clinical concern in therapy. These factors included psychopathology, relational adjustment, post-traumatic stress, lack of safety, alcohol use, drug use, and childhood maltreatment. The dyadic multinominal logistic regression analyses demonstrated that clients experiencing greater distress might be more likely to report IPV as a clinical concern while clients with greater relationship satisfaction might be less likely to report violence as a clinical concern. Also, the dyadic multinominal logistic regression analyses demonstrated both actor and partner effects, suggesting that relational dynamics affect the perception of IPV as a clinical concern. This provides tentative support for examining factors that contribute to reporting violence as a clinical concern in the context of the dyad, rather than just on an individual level. In sum, it is possible that couples experiencing IPV but are presenting as less distressed might be less likely to report the violence as a clinical concern. This emphasizes the need for systematic screening and therapists maintaining the cessation of violence as a priority for therapy, even if clients do not initially view the IPV as a concern they wish to address in therapy.