Men’s and women’s exposure and perpetration of partner violence: an epidemiological study from Sweden

The data collection for this population-based, cross-sectional study took place in January 2009. A questionnaire was sent to a randomly selected sample of 505 men and 502 women, aged 18–65 by Statistics Sweden. The questionnaire contained items on violent acts, controlling behaviours and additional socio-demographic and psychosocial variables.

The response rate was 56.5% (n = 282) for women and 43.5% for men (n = 217), and for the total sample it was 49.6% (n = 499). Of the non-respondents, the majority were unmarried compared to the respondents (54.7% vs. 40.1%); further 12.8% were divorced and 20.3% were born in a foreign country while corresponding figures for the respondents were 10.4% and 13.4%, respectively.

A further 75 individuals had to be excluded as they did not respond to any of the violence behaviour questions. The final sample consisted of 424 individuals, 173 (40.8%) men and 251 (59.2%) women. The age range was 19 to 65, with an average age of 44.3 (SD = 13.53) for men and 42.8 (SD = 13.59) for women.

Prevalence, frequency and associations were calculated for violence perpetration and exposure.

For the bi-and multivariate analysis, due to the small sample size, acts of PA were merged with the SC acts (labelled as PA/SC). Further, the time-related variables ‘past year’ and ‘earlier in life’ were merged into ‘lifetime prevalence’. Bivariate and multivariate analyses were used to investigate the associations between independent (socio- demographic- and psychosocial variables) and dependent variables (PA/SC) by use of odds ratios (OR) with statistical significance determined at the 95% confidence interval level.

In the logistic regression analyses, variables statistically significant in the bivariate analyses were entered one-by-one in a stepwise fashion for causal chain relationship and confounding analysis. As no of these variables showed high inter-correlation, they were all used in the multivariate analysis.

The Statistical Package for the Social Sciences (SPSS) version 17.0 was used for all statistical analyses.

The Regional Ethics Review Board in Gothenburg gave approval for this research project, reference number 527–08.

The majority of men (59.5%) and women (50.6%) had only basic education with less than 13 years of schooling (Table 2). Most of the respondents were married, cohabiting or living in registered partnerships. More women than men (37.8% vs. 35.3%) reported having poor social support.

We found that more men than women (11% vs. 8.0%) reported exposure to one or more acts of PA during the past year (Table 1). For the time period earlier in life, however, more women (15.9%) than men (11.0%) reported exposure to such acts. SC affected women to a higher extent than men since only one man reported exposure to such an act in the past year compared to eight women. For earlier in life, 9.6% of the women and 3.5% of the men reported exposure to such acts.

Men reported perpetrating PA to the same magnitude for past year (8.1%) as for earlier in life (8.1%), whereas 5.2% of the women reported the use of PA towards their partner in the past year and 11.6% for earlier in life (Table 3). More men than women (5.2% vs. 0.8%) reported the use of SC against their partner in the past year.

Over the lifetime, women were more exposed to controlling behaviours than men, however a considerable proportion of the men also reported such exposure (41.4% and 37.0% respectively) (Table 4). Interestingly, for three of the items, men and women’s exposure were of similar size but for two of the items (item 1 and 3) women reported exposure to a considerably higher extent than men, both mirroring activities taking place outside of the partners’ presence. Women to a larger extent also reported scores on all five items (7.6% vs. 2.9%), which in our interpretation is a sign of being under severe control from the partner.

As a considerable proportion of the men were exposed to PA/SC, we also wanted to investigate whether these men also used any form of violence against their partner. We found that of those exposed, 63.9% (n=23) also reported perpetration of PA/SC in their lifetime. For women this relationship was different as 39.4% (n=26) of those exposed also used PA/SC towards their partner.

For lifetime exposure to PA/SC among men, statistically significant risk factors were attributable to the younger age group (OR 3.19; 1.36 – 7.49), being single or unmarried (OR 3.57; 1.36 – 7.49), and having a relationship of less than three years duration (OR 4.85; 1.94 – 12.12) (Table 5). For women, the most prominent risk factor was to be single or unmarried (OR 4.30; 2.29 – 8.09) and being in a relationship of short duration (OR 2.60; 1.26 - 5.38). Unemployment, low household income and having poor social network and support also proved statistically significant for women (Table 5).

Results from the regression analysis for violence exposure over the lifetime are displayed in Table 5. For men, a relationship of less than three years duration remained the only statistically significant risk factor for PA/SC (OR 4.20; 1.15 – 15.40) while poor social support (OR 2.79; 1.31 – 5.92) and being single or unmarried (OR 3.10; 1.06 – 9.12) remained statistically significant risk factors for women.

Regarding exposure to PA during the past year, there is no Swedish population-based study available that estimates men’s exposure, while our findings for women concur with earlier reports [21] However, a study from Straus and colleagues [17] based on a sample of students found that more men (49%) than women (31%) reported exposure to PA during past year which is similar to our findings although our estimates were considerably lower in size. We found however that earlier in life estimates of PA where higher for women than for men. We also noted that women reported exposure to acts considered to be “severe”; for example, being “slammed against the wall,” “grabbed,” “burned or scaled” and “choked” more often than men, which is consistent with previous findings indicating that men use serious assaults to a larger extent than women [14]. Since women mostly are physically disadvantaged [19] we hypothesize that women may feel more threatened than men and subsequently recall bias would be less in women than in men. Such a hypothesis may also explain why men’s reported exposure to physical assault for the period earlier in life was lower than expected (it was of the same size as for the past year) . Support for this reasoning is also found in a study by Dobash and Dobash [8] where men described women’s violence as “insignificant,” “comical” and “ludicrous” (pp 340), hereby demonstrating that the violence was less frightening to them and less import to recall.

Somewhat to our surprise, we found that more women than men reported perpetration of PA for the time period ‘earlier in life’. Other studies describe self-defence as an important motive for woman’s violence [9, 22]. Due to gender norms, women’s use of violence is further seen as less acceptable and may evoke shame in women, therefore women may remember their own use of violence to a higher extent than men [23]. However, since this particular study did not explore the motives behind the acts it is difficult to draw any further conclusions.

Throughout our analyses, more women than men reported exposure to SC during the past year and earlier in life and this is consistent with earlier research [4, 9, 10, 13]. It is widely known that women exposed to SC (with or without physical abuse) are more likely to suffer serious health problems compared to women physically abused but not exposed to any kind of SC [24, 25]. This indicates the need for further efforts regarding both prevention and interventions.

More women than men were exposed to isolating control. According to earlier findings, women are mainly exposed to a combination of violent acts and coercive control tactics such as threats and intimidation while men are the prime perpetrators of such violence in heterosexual relationships [5, 14, 15]. Traditional gender norms, power structures and gender inequalities between men and women are core factors that give rise to coercive control tactics [14, 26]. We did, however, also find that men reported exposure to isolating control which points to the existence of female perpetrators [5, 15]. It may be that exposure to coercive control tactics differ in characteristics and long-term consequences for men and women and therefore future investigations, both qualitative and quantitative, should explore the dynamics and emotions surrounding coercive control tactics.

In summary, we found in this study that both men and women are exposed but also use physical violence and controlling behaviours within their relationship, thus suggesting gender symmetry. Such findings have been published previously, mainly in the USA and other Anglo-Saxon countries [15, 17]. However, as the motives were not further explored, we are not able to state with accuracy the reason for women’s use of violence, i.e. in self-defence or also in aggression.

Similar to other studies of women, poor social support [27] and belonging to the younger age group were significant risk factors for exposure to violence [28, 29].

Good social support is commonly found to be a protective factor against IPV and against the recurrence of IPV for women exposed earlier in life [27] and this is also what we found in this study. Social support contributes to making women feel valued, enhances their self-esteem and functions as a practical resource to assist when exposed to violence [27, 30, 31]. On the other hand, being young and exposed to poor social support signals being in a vulnerable position that may hamper women’s help-seeking.

For men, due to the small sample size, we found that short duration of present relationship was the only factor that remained statistically significant after controlling for socio-demographic and psychosocial factors. However, pointing in the same direction are the findings in the bivariate analyses, which suggest that also young age and not being married or in a stable relationship may contribute to increase the risk of being exposed to partner physical assault and/or sexual coercion. These assumptions do however need to be further investigated in larger samples.

A major strength of this study is that the data comes from a randomised population-based sample of men and women in Sweden. To our knowledge, this was the first Swedish population-based study to explore both exposure and perpetration of IPV and its socio-demographic risk factors among men and women. Another strength in the current study is the use of the CTS2, a well-known violence instrument that is validated and used globally [17].

Among the limitations, the most notable is the rather low response rate and the internal drop- out rate related to the violence items, especially for men. Since the majority of external and internal drop-outs were unmarried, it might be that they had no experience of intimate relationships and therefore were not further motivated to answer the violence-related items. Further, those individuals most exposed and/or perpetrating violence might have been reluctant to fill in the questionnaire because of shame or fear of being identified, which has been found in an earlier study [5]. We believe also that this questionnaire may not be ideal for data collections through mailed questionnaires as the violence items are profuse and rather detailed in content. Another limitation is that the low number of respondents reduced the power of the analysis.

Since the data is based on self-reports of exposure and perpetration, care must be taken in comparing the results from this study with police reports or clinical records as only the most serious cases in terms of injuries or ill-health will be registered in such reports and not the kind of ‘every day’ violence that goes unnoticed by authorities.

Past year prevalence figures are considered to be more accurate than earlier in life figures due to a reduced recall bias. This is obvious in the current study, particularly for men, as earlier in life estimates are lower or equal to estimates related to past year exposure apart from SC exposure. This might be explained by higher recall bias in men being less threatened by the violence and abuse exercised by women than vice-versa.

Since this is a cross-sectional study, it is not possible to make definite statements about the direction of the causal relationships. Lack of social support, for example, can be a risk factor for IPV exposure, but might also be the result of prolonged IPV exposure.