Making Little Progress?
Domestic violence in South Australia: a population survey of males and females Eleonora Dal Grande, Jacqueline Hickling, Anne Taylor Population Research and Outcome Studies, Department of Human Services, South Australia
Abstract Objective: To determine the self-reported prevalence of domestic violence in South Australian adults and to examine the
associated risk factors, demographic
Strategic Planning and Population Health Division, Department of Human Services, South Australia
factors and related health issues using computer-aided telephone interviewing (CATI) methodology.
Sample: A representative random sample
omestic violence is a significant public health problem with a range of important social, economic, political, emotional and physical ramifications. 1-4 The outcomes of domestic violence vary across the mental and physical health spectrum. Mental health effects include depression, anxiety, low self-esteem and post-traumatic stress morbidity.5 Physical outcomes range from bruising or cuts to internal injuries, unconsciousness and death.2,4-6 Domestic violence often has long-term emotional and/or physical effects for those either experiencing it or witnessing it. One important outcome of any violent domestic situation is the flow-on consequence that affects other family members. Children are said to be particularly affected and early exposure to abuse is thought to greatly increase the risk of violent behaviour towards others or themselves in the future.7 It has been documented that when a person experiences child abuse, the risk of victimisation in later life is increased.8-11 This trend has been noted more frequently in women.7,12 There has been considerable debate in the literature as to the definition of domestic violence and various prevalence estimates of domestic violence exist depending on the definitions used.2,6,13-17 Domestic violence is also known as partner abuse and involves abuse among intimate partners, rather than siblings or others. Although most definitions
have concentrated on physical and sexual abuse, the current study also included emotional abuse. Prevalence estimates vary. The Australian Bureau of Statistics (ABS) Women’s Safety Survey published an overall prevalence rate of 7% for both physical or sexual assault in the past 12 months and lifetime (15+ years of age) female partner abuse prevalence of 23%.18,19 However, these figures did not include emotional abuse and was not limited to domestic violence by a partner. Prevalence estimates of domestic violence in the United States range between 5% and 20%, depending on the definition used.20 Similar estimates have been found in the United Kingdom.21 In contrast, figures from Australian police reports of spousal or partner assault are low.15 It is acknowledged that police report figures are underestimates and only include events classified as violent criminal behaviour.15 Previous studies have found a number of risk factors associated with domestic violence.15,17,20 Alcohol and drug abuse, unemployment or intermittent employment, low levels of household income and low levels of education of both the abused and the perpetrator have been shown to be important indicators, as is the history of prior abuse.5,22 Recent termination of the relationship has also been found to be a predictor variable and it has been argued that a relationship other than marriage (for example, former
Correspondence to: Ms Eleonora Dal Grande, Centre for Population Studies in Epidemiology, Department of Human Services, PO Box 287, Rundle Mall, Adelaide, South Australia 5000. Fax: (08) 8226 6244; e-mail: [email protected]
2003 VOL. 27 NO. 5
of South Australian adults aged 18 years and over selected from the Electronic White Pages. Overall, 6,004 interviews were completed (73.1% response rate).
Results: In total, 17.8% of adults in South Australia reported some form of domestic violence by a current or an ex-partner. Demographic factors such as low household income, unemployment or parttime employment and health variables such as poor to fair self-reported health status and alcohol abuse problems were found to have a significant relationship with domestic violence.
Conclusions: Approximately one in five South Australian adults report physical and emotional abuse from current or ex-partners, of whom the majority are women who are separated, divorced or never married and on lower incomes. Telephone interviewing is a cost-effective method of identifying prevalence estimates of domestic violence when compared with data collection by way of police reports or hospital emergency statistics.
Implications: Domestic violence is a serious public health concern often ‘hidden’ by the lack of appropriate data. This study has shown that domestic violence is frequent and has important social, economic and health consequences. (Aust N Z J Public Health 2003; 27: 543-50)
Submitted: July 2002 Revision requested: December 2002 Accepted: May 2003
AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH
Dal Grande et al.
husband, estranged husband or former boyfriend) between the abuser and the victim is associated with violent domestic situations.23 In Australia, there have been few population-based surveys on domestic violence. Most studies have focused on specific populations including women,18 pregnant women,24 accident and emergency attendees,25 general practice clinics,26 women’s shelters and accommodation services, police and court statistics, and homicide statistics.26 Many of these studies have also been criticised for having insufficient sample sizes and for relying on selfreporting to police or health care services.1 Some estimation of the extent of the problem can be gained using these samples, however they do not give an indication of the health services utilised by abused individuals, or an indication of prevalence rates for the community.1 In addition to these shortcomings, past prevalence studies have not adequately addressed the consequences of domestic violence on mental health and self-esteem or examined the occurrence of revictimisation.12 Nor have they adequately addressed domestic violence against men, which is of recent concern.6,23
Methods As part of a commitment to reduce morbidity and monitor this public health problem, a study was conducted to assess the prevalence, experiences and risk factors associated with interpersonal violence in South Australia.27 For this survey, interpersonal violence included domestic violence, physical and sexual violence experienced as an adult, elder abuse, and child abuse and/or neglect for both men and women. The analysis presented in this paper is limited to domestic violence and its relationships with adult sexual and physical abuse, and child abuse and/or neglect. The data for this research, collected in September and October 1999, used SERCIS (Social, Environmental and Risk Context Information System), a telephone monitoring system designed to collect high-quality health data on large samples of the South Australian population.27,28 The proportion of households in South Australia with a telephone connected is 97%, hence the potential for non-coverage bias was negligible.29 SERCIS utilises the CATI III (computer-assisted telephone interviewing) system to conduct interviews. All households in South Australia with a telephone connected and the telephone number listed in the Electronic White Pages (EWP) were eligible for selection in the sample. A letter introducing the survey was sent to the selected household. The letter did not specify that the health and well-being survey addressed aspects of interpersonal violence. At each selected household, the adult with the last birthday was chosen for interview. Interviews were conducted by trained female health interviewers in English, Italian, Greek and Vietnamese. The overall sample response rate was 73.1%, with 6,004 completed interviews. The domestic violence questions were derived and modified from other interpersonal violence household surveys (Australian Women’s Safety Survey,18 Canadian Violence Against Women 544
Survey30) and to accommodate the inclusion of males in this study. Domestic violence questions were asked of both men and women aged over 18 years and over who were in a relationship or who had a previous significant long-term relationship. To minimise the possibility of a negative reaction from the questions, the interviewers were trained to introduce the topic in such a way that allowed the respondent to withdraw if they had any fear or anxiety about the issues being raised. Respondents were also told that they could postpone the interview to a time more suitable. Referral telephone numbers were provided for people who felt they required advice or counselling as a result of the issues raised. The interviewers were also trained to terminate the interview should they sense any disturbance on the part of the respondent and were also given the opportunity to talk to the study supervisor or coordinator on any issue if they needed or requested to do so. Respondents were asked if a current partner or an ex-partner had ever physically or emotionally abused them. The definition of domestic violence used in this study is described in Table 1. Respondents were also asked whether they had ever been sexually or physically abused by people other than their partners or ex-partners or whether they had experienced abuse or neglect as a child. Questions were asked on elder abuse, however these were not included in the final analysis due to insufficient sample size. Demographic variables included in the analyses were: gender; age; area of residence; country of birth; main language spoken at home; marital status; work status; highest educational attainment; pension from the Department of Social Security; and gross annual household income. To determine general health, all respondents were asked how they rated their general health on a scale from excellent to poor. The four-item CAGE instrument was included, which provides a picture of the person’s drinking that might indicate alcohol abuse and dependence over their lifetime.31 It was originally designed to screen for alcoholism from a clinical study and has been reported to have high face validity and good concurrent validity.32,33 To determine the self-esteem of respondents, the 10-item Rosenberg self-esteem scale was also used34 and respondents answered whether they agreed or disagreed with the general feeling statements. It has been found that the scale is a valid and reliable measure of self-esteem.35 If a respondent indicated an episode of domestic violence additional questions were asked, including the number of relationships, how often these incidents occurred, health effects such as physical injuries and emotional or mental problems as a result of the violence, non-health (i.e. welfare, housing, and/or legal services) and health services contacted, satisfaction with service used, smoking status and mental health status. The questions on smoking were taken from the 1989-90 National Health Survey.36,37 Nonpsychotic psychological disturbance was measured by the General Health Questionnaire (GHQ-12),38,39 which has been found to be an efficient, reliable and valid instrument in Australia.40 The GHQ12 was designed as a population-screening instrument and not intended as a diagnostic tool. It consists of 12 questions and produces a total score ranging from 0 to 12. In this case, a score of 2
AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH
2003 VOL. 27 NO. 5
Making Little Progress?
Domestic violence in South Australia
or more was considered to detect a person having a mental health condition. Suicidal ideation was measured using four questions from the GHQ-2838,39 where a positive response to any of the four questions was considered indicative of suicidal ideation.41 Since smoking status, GHQ-12 and GHQ suicide ideation were only asked of people who reported domestic violence, these data were compared with independent South Australian population samples (smoking,42 GHQ-12,40 and GHQ suicidal ideation41). It should be noted that these population comparisons are of the whole South Australian population and therefore may include both those who would report domestic violence together with those who would not report interpersonal violence. The data were weighted by area, age, gender and probability of selection in the household using the ABS Estimated Resident Population for 1997.44 Data were then analysed using SPSS Version 10.0. Univariate analyses were conducted for males, females and persons in total who had ever experienced domestic violence compared with those who had not. All independent variables that were statistically significant at the 0.25 level in each of the univariate analysis were entered into a logistic regression analysis.45 Once a satisfactory model was obtained, tests for interaction were performed on likely combinations of variables. Interaction terms were entered into the final model to determine if a statistically significant improvement in the model was obtained. The
presence of confounders was also assessed during the modelling process.
Results The overall prevalence of domestic violence was 17.8% (95% CI 16.8-18.8, n=976). A total of 12.0% (95% CI 11.2-12.9, n=660) of respondents reported a physically violent or abusive relationship and 14.1% (95% CI 13.2-15.1, n=775) reported an emotionally abusive relationship (see Table 1). A small number of respondents (n=7) refused to answer one or more of the questions assessing domestic violence. Females were significantly more likely (22.9%) than males (12.1%) to have experienced any form of domestic violence. A majority of respondents who reported some form of domestic violence reported that the situation occurred on more than one occasion (75.9%). Persons who reported domestic violence were significantly more likely to be aged between 35 and 54 years, be separated, divorced or never married, be either employed part-time or unemployed, have a gross annual household income of less than $40,000, to report their health as being fair or poor, have medium self-esteem and alcohol abuse problems (see Table 2). In addition, they were significantly more likely to be current smokers, to have a mental health problem and to have suicidal thoughts (see Table 3). Overall, 38.6% (n=216) of respondents who reported forms of
Table 1: Prevalence of physical violence or abuse by a current partner or an ex-partner, by category of physical violence or emotional abuse. Variable
Males n %
Females n %
Persons n %
Physical abuse Ever hurt physically; for example kicked, choked, pushed or hit with their fist or anything else that could hurt them
Ever threatened to use or has used a gun or knife to harm them
Ever forced into any sexual activity when they did not want to
Any of the three above Emotional abuse
Social abuse such as restricting social freedom and isolating from family and friends
Economic abuse such as withholding money or deliberately giving inadequate funds for household needs
Other abuse such as threats and intimidation, constantly being called names or humiliated
Any of the four above
Spiritual abuse such as denying the practice of religion
Domestic violence or abuse Some form of domestic violence or abuse Physical violence and abuse only Emotional abuse only
Both physical violence and abuse and emotional abuse
Other domestic violence or abuse
No domestic violence or abuse reported
0.5 87.9 ∧ 100.0
77.1 ∨ 100.0
Note: ∧ ∨ Statistically significantly higher or lower (χ2 test, p