Abstract
This study investigates the association between family relationships, anger, alcohol use, and self-reported intimate partner violence (IPV). Participants were 55 male prisoners who completed a survey about their family relationships, anger, alcohol use, and aggression. Exposure to parental IPV predicted rates of self-reported perpetration of IPV, suggesting the importance of understanding more about the developmental pathways to IPV if effective prevention, intervention, and assessment strategies are to be developed for use with this high-risk population.
The term Intimate Partner Violence (IPV) is used in this study to refer to acts of violence that occur between people who have, or have had, an intimate relationship (Morgan & Chadwick, 2009). It has been suggested that rates of IPV are particularly high in prison populations, both in terms of violence perpetrated before imprisonment and violence that occurs following release, although surprisingly little is known about the intervention needs of this group or the features of effective prison-based behavior change programs (see Day, Richardson, Bowen, & Bernardi, 2014). In light of evidence that programs do not lead to significant reductions in IPV re-offending (see Babcock, Green, & Robie, 2004; Feder & Wilson, 2005), even when delivered in the community to those who are potentially at much lower risk, there is a need to understand more about IPV in prison populations. This relates to prevalence as well as those factors that are associated with onset and development.
Although a large number of risk factors are known to be associated with IPV (see Abramsky et al., 2011), key risk factors consistent with the World Health Organization’s bio-psycho-social conceptualization of the violent offending pathway include family-of-origin IPV, abusive fathering, trait anger, and alcohol use (WHO, 2002). Witnessing and experiencing violence as a child is, for example, thought to lead to a greater tolerance of violence in adulthood (see Delsol & Margolin, 2004) and may be associated with both angry emotion and alcohol use (Klostermann & Fals-Stewart, 2006; Norlander & Eckhardt, 2005). In fact, anger is one of the most consistently identified risk factors, regardless of the methodological paradigm used (see Fals-Stewart, 2003; Norlander & Eckhardt, 2005) and although the exact aetiological role of alcohol in IPV has yet to be established, there is extensive empirical evidence that alcohol use often precedes or accompanies violence (see Cunradi, Caetano, & Schafer, 2002; Field, Caetano, & Nelson, 2004; Morgan & McAtamney, 2009). One study reported that 35% of all IPV incidents involve alcohol (Marcus & Braaf, 2007), with victims of severe IPV (or severe injury as a result of IPV) more likely to report having partners who frequently consume high levels of alcohol or exhibit symptoms of alcohol dependency.
A recent systematic review of longitudinal studies that have prospectively investigated childhood and/or adolescent predictors of IPV identified child and adolescent abuse, family-of-origin risks, child and adolescent behavioral problems, adolescent peer risks, and socio-demographic risks as significant predictors of IPV (Costa et al., 2015), although none of the 25 studies reviewed involved prison populations. Theoretically, IPV predictors may differ for this population given the lower prevalence of what Holtzworth-Munroe and Stuart (1994) refer to as “Family-Only Violence” (male-perpetrated domestic violence that occurs primarily in response to environmental triggers and where perpetrators do not typically endorse the use of violence and exhibit empathy and positive attitudes toward women; see also Holtzworth-Munroe, Stuart, & Hutchinson, 1997).
The aim of this pilot study is to examine the relationship between key risk factors for IPV (exposure to parental violence, quality of paternal relationship, trait anger, and alcohol use) and self-reported perpetration of IPV in adulthood in a sample of male prisoners. Specifically, it is hypothesised that childhood exposure to parental IPV, poor paternal relationships, trait anger, and level of alcohol consumption will predict self-reported IPV.
Method
Participants
A total of 55 prisoners in a medium security Australian prison (aged between 19 and 54 years, M = 33.98, SD = 8.45 years) participated in the study. Participants were selected to be representative of the wider prison population as IPV is not recorded as a specific offense under local legislation. Nearly three quarters (73%) reported convictions for violent offenses, with more than two thirds (68%) of these reporting that alcohol was involved at the time the offense was committed. One quarter (27%) reported that they did not have a father as a primary carer, defined in relation to the response to the following question: “During your first 16 years, who were your main carers?” Of these, the majority (80%) did not witness parental violence. One quarter (n = 13; 24%) of prisoners reported witnessing parental violence, with only 3 reporting that they witnessed parental violence without having a father as a primary carer (none of these reported IPV in their relationships). Of the 13 prisoners who reported witnessing parental violence, 8 reported perpetrating IPV (62%). Nearly 1 in 3 (31%; n = 17) disclosed a history of IPV, with nearly a quarter (24%) reporting father to mother IPV.
Procedure
After ethical approval for this study was granted by a university Human Research Ethics Committee, prisoners were invited to complete a hard copy of the survey in small groups in a communal area. All responses were anonymous and confidential, and both verbal and written assurances were given that they would have no bearing on their current treatment or management. The researchers were present to explain the purpose of the research, answer any questions, and assist those with poor literacy skills who requested help. Data collection in the prison took place over 1 day. Completion of the survey was voluntary, and no compensation was offered in lieu of participation.
Measures
The survey comprised a number of measures as follows:
IPV
Participants were asked to rate their aggressive and violent behavior toward an intimate partner. These ratings were made in relation to a relationship in the 12 months prior to incarceration. Items were modified from the Revised Conflict Tactics Scale (CTS2; Straus, Hamby, Boney-McCoy, & Sugarman, 1996) to measure the frequency of psychological aggression and physical assault. For example, participants were asked to report the number of times in the past year they had “insulted or swore at” or “pushed or shoved” their partner. Scores from the relevant items were totalled to produce a prevalence score (if one or more of the acts on the scale was reported) and a chronicity score (the number of times the act(s) occurred). In the current study, Cronbach’s alphas for the Psychological Aggression and Physical Assault scales of the CTS2 were .80 and .95.
Paternal IPV
A subscale of the Adult-Recall Version of the Revised Conflict Tactics Scale (CTS2-CA; Straus, 1999) was used to assess the participant’s perception of their parent’s behavior toward each other during childhood. Only those items considering the behavior of a father toward a mother were used in the analysis. Participants used a 3-point Likert-type response scale (never, occasionally, frequently) to indicate the extent to which statements such as “throw, hit or smash something” or “choke or strangle” depicted their father’s behavior toward their mother. Behaviors were scored as either acts of “verbal aggression” or “severe violence.” Cronbach’s alpha for the subscales of the CTS2-CA was .95.
Fathering
Participants were asked to use a 4-point Likert-type response scale (not at all true, slightly true, moderately true, extremely true) to indicate the extent to which 15 statements described their father’s behavior toward them (e.g., “ignored me,” “over controlling of me,” or “physically violent or abusive of me”). Participants were also asked whether they were raised by biological parents and who the most significant role model was during the first 16 years of their life. The Measure of Parental Style (Parker, Tupling, & Brown, 1979) was used to measure three aspects of perceived parenting style: “indifference,” “abuse,” and “over-control.” Although this scale assesses perceptions of both parental relationships, only the items relating to fathering were used in the analysis. Scores were totalled to produce an overall score for each of the three aspects of parenting. The Measure of Parental Style is considered a valid and reliable measure, with Cronbach’s alphas for the Indifference, Abuse, and Over-Control being .93, .92, and .76, respectively (Parker et al., 1997). Cronbach’s alphas for Indifference, Abuse, and Over-Control subscales in the current sample were .88, .94, and .76, respectively.
Anger
The State-Trait Anger Expression Inventory–2 (STAXI-2; Spielberger, 1999) purports to quantify the experience, expression, and control of anger. Only the 10-item Trait Anger subscale of the STAXI-2 was used, which asks respondents to indicate how they generally feel or react in relation to statements such as “I am quick tempered” using a 4-point Likert-type response scale (almost never, sometimes, often, almost always). The STAXI-2 and the Trait Anger scale have been shown to be valid, reliable measures for use with prison populations (Etzler, Rohrmann, & Brandt, 2014), and the Cronbach’s alpha for the Trait Anger scale in this study was .87.
Alcohol consumption
Participants were asked to report how many days a week, on average, they consumed alcoholic beverages and the number of standard alcoholic drinks consumed during this time. Participants were asked to provide responses based on alcohol consumption prior to incarceration. The usual number of standard drinks when drinking (alcohol use) was the variable used throughout the analyses, following the methodology described by Miller, Hargreaves, Curtis, and Zinkiewicz (2013).
Demographics
Finally, participants answered some questions about their age, cultural background, and marital status. They were asked about the nature of their current offense and the number of past convicted offenses that had involved violence or alcohol.
Results
The means, standard deviations, and ranges for all variables are presented in Table 1 (below). Spearman’s ρ correlations, calculated to examine the strength of the association between risk factors and IPV, revealed a moderate significant positive correlation between parental IPV and self-reported IPV. Trait anger, alcohol use, and the three fathering styles were all positively correlated with self-reported IPV.
Spearman’s ρ Correlations Between IPV, Parental IPV, Indifference, Abuse, Over-Control, Trait Anger and Alcohol Use.
Note. IPV = intimate partner violence.
p < .05. **p < .01.
To test the hypothesis that parental IPV, abuse, over-control, indifference, trait anger, and alcohol use would predict self-reported IPV, a binary logistic regression was undertaken (see Table 2). The Hosmer–Lemeshow Goodness-of-Fit Test was non-significant, χ2(8, N = 39) = 1.94, p > .05, indicating that the level of error was small and the model was a good fit for the data. The variance in perpetration of IPV accounted for was relatively good with the model explaining 33% of variance in the data. Classification was biased in favour of those who reported not perpetrating IPV, with 92.6% of these predicted correctly, but only 58.3% of those who reported perpetration of IPV correctly identified. Only exposure to parental IPV significantly predicted self-reported IPV, with those participants 34 times more likely to report perpetration of IPV (B = 3.54). Participants who reported higher levels of trait anger were also more likely to report IPV (B = 0.18).
Logistic Regression Analysis Predicting Perpetration of IPV as a Function of Parental IPV, Abuse, Over-Control, Trait Anger, and Alcohol Use in Prisoners.
Note. IPV = intimate partner violence.
p < .05.
Discussion
This study investigated the association between specific developmental and shorter-term risk factors and self-reported IPV in male prisoners. Given that prisoners are often considered to represent a particularly high-risk group for IPV, the aim of this study was to understand more about the prevalence of IPV as well as to identify those factors associated with IPV that could potentially inform the development of effective prison behavior change programs. Around one in three medium security prisoners reported perpetrating IPV, and experiences of parental IPV and high alcohol consumption were common. Both childhood exposure to parental IPV and trait anger predicted self-reported IPV, although the former was the most powerful predictor.
Although this is clearly an exploratory study involving a small self-selecting sample with low statistical power, these findings do suggest that a significant proportion of men in prison do perpetrate IPV. Although replication with larger samples across different security classifications is needed, this suggests that prison-based IPV programs are required in a country in which this type of program is relatively uncommon (Heseltine, Sarre, & Day, 2011). Given the potential for under-reporting biases associated with impression management when self-report measures are used, the true prevalence rates may be even higher. Nonetheless, we emphasise that substantial caution is required before the findings of this study can be considered to be reliable, given the very small sample on which conclusions are drawn.
Understanding the antecedents of IPV in this population will continue to be important from an intervention perspective, and the contribution of this study is to report preliminary data to suggest that family-of-origin experiences are formative in this population and, therefore, relevant to efforts to change behavior. Although recollections of experiences such as family-of-origin IPV and pre-prison alcohol use may lack reliability, self-report measures of the type used in this study offer the only means to collect information about key variables. Clearly, prospective and longitudinal research designs involving large samples are preferable, although these studies are both costly and time-consuming, and introduce different threats to validity (including those associated with differential attrition and inconsistent recording practices; see Smith, Ireland, Thornberry, & Elwyn, 2008). Nonetheless, the scope of this study is limited, and future work should extend the focus of enquiry to include mother–father violence or violence exposure in other relationships, such as with step-parents or other caregivers.
Given that trait anger or alcohol use did not mediate IPV risk in this study, a sharper focus on those developmental factors that influence the acquisition of violent behavior when developing behavior change programs may be warranted. This is consistent with Stith et al.’s (2000) suggestion that both alcohol and anger regulation issues become significant issues for IPV perpetration only in the context of poor mental health. It may also be, however, that different developmental predictors and risk factors are relevant for different types of perpetrators and that treatment outcomes will improve when interventions are matched to specific subgroups. Clearly, further research involving larger samples is required to establish possible within-group differences.
In conclusion though, this study draws attention to the need to routinely screen for IPV on reception to prison to establish the need for specialist rehabilitation programming, and describes some measures that have potential utility in understanding more about the antecedents of IPV in this particular population.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.