Search tips
Search criteria 


Logo of jepicomhJournal of Epidemiology and Community HealthVisit this articleSubmit a manuscriptReceive email alertsContact usBMJ
J Epidemiol Community Health. 2007 July; 61(7): 625–630.
PMCID: PMC2465737

Prevalence and concordance of interpersonal violence reports from intimate partners: findings from the Pacific Islands Families Study



Intimate partner violence (IPV) is an increasingly significant public health issue. Few studies investigate both partners' reports of such violence.


To determine the prevalence of IPV perpetration and victimisation and the concordance of both partners' reports of that violence in a representative sample of Pacific couples using a standardised measure.


A cohort of Pacific infants born during 2000 in Auckland, New Zealand, was followed. At 24‐months postpartum, home interviews were conducted separately for mothers and fathers, and experience of IPV within the last 12 months was measured using the Conflict Tactic Scale. Concordance and symmetry between partner's reports was assessed using the κ statistic and McNemar's test, respectively.


The sample included 915 partnered mothers and 698 partnered fathers of which 674 were couples. Over 85% of individuals perpetrated or were victims of verbal aggression. Perpetration and victimisation of physical IPV was reported by 37% and 28% of mothers and 11% and 8% of fathers, respectively. Fathers systematically under‐reported IPV and significant differences emerged in the response distributions from couples with Tongan mothers compared with couples with Samoan or Cook Island Maori mothers.


IPV is common for many Pacific couples and consistent with that reported in other New Zealand groups. Mothers are as likely as fathers to perpetrate and be victims of this violence. However, patterns of reporting were different between mothers and fathers, and between Pacific ethnic groups, which has implications for the understanding and health promotion targeting of domestic violence.

Intimate partner violence (IPV) occurs in all countries, irrespective of social, economic, religious or cultural groups.1 It is generally recognised that the overwhelming burden of partner violence is borne by women at the hands of men.1,2 The World Health Organization's Multi‐country Study on Women's Health and Domestic Violence Against Women3 reported lifetime prevalence of physical or sexual violence, or both, by an intimate partner that ranged from 15% to 71% across 10 diverse countries for women who had ever been partnered.3,4 In a recent New Zealand study, over a third of “ever‐partnered” women in two sites reported physical or sexual violence perpetrated by an intimate partner.5 Although the costs to any one individual can be enormous, the public health burden of IPV is also considerable. It is estimated that domestic violence and rape account for approximately 5% of the total disease burden in women aged 15–44 years in developing countries.6 In 1994, domestic violence was estimated to cost New Zealand between $1.2 and $5.3 billion per year.7

The predominant research focus on IPV casts men in the role of perpetrators and women in the role of victims1,2,8 and, although recognised,1 few studies also examine women's violence towards their partners. An international review of IPV published in 2005 identified five studies which present comparative information of men's and women's involvement as perpetrators of IPV.8 One of these was from the Dunedin Multidisciplinary Health and Development Study, a birth cohort established in the 1970s.9 Among 869 members of the cohort in current or recent intimate relationships, 37% and 27% of women and 22% and 34% of men reported perpetrating and being victims of physical IPV, respectively.9 However, the Dunedin study (and three of the other four studies identified) does not shed light on couples' concordance of reported IPV as respondents‐reported experience was not matched to their partner's.8 The fifth and only study with a design that enabled an investigation into this reporting concordance consisted of a convenience sample of 95 men convicted of IPV and their female partners.10 The authors found that IPV is primarily an asymmetrical problem of men's violence to women, and women's violence does not equal to men's in terms of frequency, severity, consequences and the victim's sense of safety and wellbeing. However, the reliance on convicted men as a sampling unit markedly limits the generalisability of that study's findings.

Since the review,8 a multistage area probability study with a sample representative of married and cohabiting couples from 48 contiguous US states has been published.11 This study investigated unidirectional IPV (where only one partner perpetrates the violence) and bidirectional IPV (where both partners engage in violence) among white, black and Hispanic couples. The authors found that most couples reporting violence engaged in bidirectional IPV and significant ethnic differences emerged. A study that assessed IPV–spousal agreement within 273 couples seeking marital therapy has also appeared in the literature.12 The results revealed low‐to‐moderate levels of agreement for reported IPV. Moreover, husbands and wives reported a lower level of aggression for themselves than their partners attributed to them, though this discrepancy was generally stronger for husbands.

There are limited data regarding IPV rates in New Zealand5,13 and a paucity of data relating to resident Pacific peoples.14,15 The Pacific population in New Zealand is now estimated to exceed 284 000,16 Samoans constitute the largest ethnic group (50%) followed by Cook Island Maori (23%) and Tongans (18%), and most live in Auckland.17 This ethnic diversity is manifest in differing cultures, languages and strength of acculturation. However, Pacific peoples have an excess of social, health and economic deprivation.14,18 There is a growing recognition that issues which have a significant impact on Pacific peoples' lives need to be addressed, of which IPV stands out.14,15

The lack of clear theoretical understanding of the causes of IPV1 and the controversial issue of men's and women's symmetrical versus asymmetrical involvement in IPV8 have led calls for studies that examine prevalence, relationships, consequences and risk, and protective factors of violence by intimate partners in different cultural settings by using standardised methods.1,19 Indeed, if the public health approach to violence is based on the rigorous requirement of scientific method that moves from understanding and measuring the problem to finding, implementing and evaluating a solution,1,20 then such studies are essential. Using a standardised measure, this paper aims to determine the prevalence of IPV perpetration and victimisation, and the concordance of both partners' reports of that violence in an ethnically representative sample of Pacific couples.


The Pacific Islands Families study follows a cohort of Pacific infants born at Middlemore Hospital, South Auckland, New Zealand, between 15 March and 17 December 2000. Detailed information about the cohort, and its recruitment and retention procedures is described elsewhere.21 In brief, all potential participants were selected from birth records, where at least one parent was identified as being of Pacific Islands ethnicity and a permanent resident of New Zealand. Information about the study was provided to all potential participants and consent was sought to make a home visit.

Approximately 6 weeks after infants' births, female interviewers of Pacific Islands ethnicity who were fluent in English and a Pacific Islands language visited mothers at their homes. Once eligibility was confirmed and informed consent was obtained, mothers participated in 1‐h interviews concerning family functioning, and the health and development of the child. This interview was conducted in the mother's preferred language. When the children reached their first and second birthdays, all maternal participants were recontacted and revisited by a female Pacific interviewer. Again, consent was obtained before the interview was conducted in the mother's preferred language. At the time of the interview, mothers were asked to give permission for a male Pacific interviewer to contact and interview the father of the child. If permission and paternal contact details were obtained, then a male Pacific interviewer contacted the father to discuss participation in the study. Once informed consent was obtained from the father, the interview was carried out in the father's preferred language.

Intimate partner violence

The experience of both verbal and physical IPV was measured using Form R of the Conflict Tactics Scale (CTS) developed by Strauss.22,23,24 Arguably, the CTS is the current dominant instrument for assessing violence among couples.12 Participants were asked to enter their responses on an answer sheet as the interviewer read out the questions. Firstly, participants reported on their behaviour towards their partner (perpetration), and secondly, they reported about their partner's behaviour towards them (victimisation). The CTS measure of verbal aggression includes six items, the minor physical violence scale includes three items, and the scale of severe physical violence scale includes six items. An individual was considered to be a perpetrator of verbal aggression, minor physical violence or severe physical violence if they reported any of the behaviours during the past 12 months. Likewise, an individual was considered to be a victim of verbal aggression, minor physical violence or severe physical violence if they reported that a partner subjected them to any of the behaviours during the past 12 months. For the purpose of this paper, we report verbal aggression, any physical violence (which is indicated if minor or severe physical violence is admitted) and severe physical violence. Psychometric properties of the CTS scales are robust and have been described by Straus.25 Responses were included in these analyses if participants completed at least 13 of the 15 items of CTS.

Statistical analysis

Prevalence estimates and CIs were made using the binomial distribution. Concordance between response pairs was measured using the κ statistic. Using Landis and Koch's characterisation,26 κ>0.75 was taken to represent strong agreement, κ<0.40 to represent poor agreement and 0.40[less-than-or-eq, slant]κ[less-than-or-eq, slant]0.75 to represent moderate agreement beyond chance. Symmetry of the discordant observations was compared using McNemar's test.27 Comparison of response distributions between groups was made using Fisher's exact test. SAS V.9.1 was used for all computations and a significance level of α = 0.05 was used to define statistical significance.


Ethical approval was obtained from the National Ethics Committee, the Royal New Zealand Plunket Society and the South Auckland Health Clinical Board.


Overall, 1477 mothers were eligible for the Pacific Islands Families study, 1144 (77%) participated at the 24‐month interview and of these 915 (80%) were currently partnered and eligible for these analyses. 999 men were identified, 757 (76%) agreed to be interviewed at 24 months postpartum, of whom almost all, 754 (99%), were biological fathers of the children in the cohort. For ease of exposition, we shall refer to this group collectively as “fathers” hereafter. Of the 757 fathers participating, 698 (92%) were partnered at the time of the interview and eligible for these analyses. Table 11 presents the distribution of selected sociodemographic variables for eligible mothers and fathers.

Table thumbnail
Table 1 Sociodemographics of maternal (n = 915) and paternal (n = 698) responders in intimate relationships, 24 months postpartum

Prevalence of IPV

Of the eligible 915 maternal and 698 paternal respondents, 893 (98%) and 691 (99%) provided sufficient information to give valid CTS scores, respectively, yielding the results given in table 22.. For all three violence levels considered, both mothers and fathers self‐reported higher prevalences of IPV perpetration than victimisation. However, fathers self‐reported substantially lower prevalences of physical violence perpetration and victimisation than mothers.

Table thumbnail
Table 2 Frequencies, prevalence estimates (%) and associated 95% CI of self‐reported victimisation and perpetration of intimate partner violence any time in the last 12 months for mothers (n = 893) ...

Table 33 gives the relationships between IPV perpetration and victimisation self‐reports for mothers and fathers. There was generally moderate agreement between these self‐reports for mothers and moderate‐to‐strong agreement for fathers. This indicates that if a mother or father reported being a victim of IPV, then she or he was also likely to report perpetrating such violence and vice versa. However, significant asymmetry in discordant responses was seen across all three violence levels for mothers and in two of the three levels for fathers (all p<0.001). Here, both mothers and fathers were significantly more likely to report that they were perpetrators but not victims of IPV than the reverse. No asymmetry was observed for fathers' self‐reports of severe physical violence, a result likely due to the small discordant frequencies associated with this variable.

Table thumbnail
Table 3 Frequencies and percentages of self‐reported responses to both victimisation and perpetration questions of intimate partner violence any time in the previous 12 months for mothers (n = 893) and fathers (n = 691) ...

IPV concordance

Table 44 gives matched self‐reports of IPV that were available for 674 intimate couples. Agreement between mothers' response to victimisation questions and fathers' response to perpetration was poor (κ range –0.02 to 0.17) and significantly asymmetrical. More mothers failed to declare being a victim of verbal aggression when their partner admitted to perpetrating this aggression than the reverse. Conversely, more mothers admitted to being a victim of any or severe physical violence, but had their partner fail to declare perpetrating this violence than the reverse. Of the 185 mothers reporting that they were victims of physical IPV, 147 (79%) partners did not declare perpetrating such violence.

Table thumbnail
Table 4 Frequencies and percentages of the relationship between mothers' response to victimisation questions and fathers' response to perpetration questions of intimate partner violence and mothers' response to perpetration questions and fathers' ...

The pattern of response was similar for mothers' response to perpetration questions and fathers' response to victimisation questions, but agreement was poor (κ range –0.03 to 0.12) and there was significant asymmetry. For each violence level, significantly more mothers admitted to perpetrating IPV but had their partner fail to report this violence than the reverse. Of the 50 fathers reporting that they were victims of physical IPV, 18 (36%) partners did not declare perpetrating such violence.

Next, matched self‐reports of IPV were partitioned across the three major maternal ethnic groups (Samoan, Tongan and Cook Island Maori) and compared (table 55).). Across all violence levels, there were significant differences in couples' distributions of concordance and discordance IPV responses between ethnic groups (all had p<0.001). In particular, distributions of concordance and discordance for couples with a Tongan mother were substantially different from those couples with a Samoan or Cook Island Maori mother.

Table thumbnail
Table 5 Frequencies and percentages of the relationship between mothers' response to victimisation questions and fathers' response to perpetration questions of intimate partner violence (IPV) and mothers' response to perpetration questions and ...

Informative attrition or refusal

When investigating maternal replies for those with partners who participated (n = 674) against those with partners who did not participate in the 24‐month postpartum interview (n = 212), there was no difference in whether mothers reported being a victim or perpetrator of IPV for any of the levels considered (all p>0.05). However, when comparing maternal responses for those with partners who declined to complete the CTS questions (n = 7) against all others (n = 886), there was no difference in whether mothers reported being a victim of verbal abuse, 7 (100%) versus 764 (86%), p = 0.60; but significant differences emerged in self‐reports of any physical violence victimisation, 5 (71%) versus 245 (28%), p = 0.02 and severe physical violence victimisation, 4 (57%) versus 132 (15%), p = 0.01.

When comparing paternal responses for those whose partners participated (n = 674) against those with partners who declined to complete the CTS questions (n = 17), there was no difference in whether fathers reported being a victim or perpetrator of IPV for any of the levels considered (all p>0.05).


Over 85% of Pacific mothers or fathers resident in New Zealand in intimate relationships admitted to being victims or perpetrators of verbal aggression at least once within the previous year. For women, the reported perpetration and victimisation rates of physical partner violence were 37% and 28%, respectively, and were very similar to those reported in the Dunedin study of 37% and 27%, respectively.9 However, this similarity was not apparent in fathers' self‐reports, with perpetration and victimisation rates of physical partner violence measured at 11% and 8%, respectively, much lower than 22% and 34% reported previously.9 The level of under‐reporting by fathers is forcefully demonstrated in the analyses that compare the concordance of responses between partners. Of mothers reporting that they were victims of physical partner violence, 79% of their male partners did not declare perpetrating such violence, whereas of the fathers reporting victimisation, only 36% of their female partners did not declare perpetrating this physical violence. The lower perpetration and victimisation rates for fathers is probably owing to the failure by many men to disclose physical violence in the presence of a male interviewer for reasons that stem from Pacific culture, customs or attitudes together with issues of gender.28

When looking at individual reports of IPV, there was moderate‐to‐strong agreement between victimisation and perpetration responses for both mothers and fathers. Like the US study,11 this implies that most couples reporting violence engage in bidirectional rather than unidirectional IPV. The dynamics of partner violence are varied and extremely complex1,19 but, put simply, they are usually instigated by one individual, which can lead to retaliation or self‐defence by their partner.

Important differences were seen in the distributions of concordance and discordance between maternal and paternal responses of IPV between the three major ethnic groups. This finding has two important implications. Firstly, an understanding of ethnic differences to IPV and perhaps culturally specific prevention strategies are needed to decrease IPV rates, and secondly, IPV rates need to be monitored within the Pacific Island ethnic groups separately rather than collapsed under a “Pacific” banner. Here, a broad Pacific label would serve to disguise the heterogeneity of IPV reports within the Pacific population resident in New Zealand.29

The Pacific Islands Families study follows a large, representative cohort over time, and has achieved a relatively small and non‐differential attrition rate to date.21 The study also employs a reliable standardised scale of IPV, the CTS.22,23,24 The distribution of IPV responses was not different between those who had partners participating in the study and those having partners not participating in this study, suggesting that IPV was not a factor for partner's attrition. However, the distribution of IPV responses was significantly worse for those mothers who had partners participating in the study but declined to answer the CTS questions. Omission of these fathers is likely to bias the prevalence estimates of IPV downwards, thereby underestimating the true IPV rates for men. However, the number of participants declining or providing insufficient information to the CTS questions was small, implying that any inherent participation bias would be negligible.

Although a useful and commonly used tool for assessing IPV in epidemiological studies, the brief CTS instrument is not without fault.30 The CTS uses 12‐month retrospective self‐report of IPV, which may suffer from recall bias and subjective or selective interpretations of violence. Furthermore, the instrument does not identify who initiated the violence, the intention, context and motivation of the violence nor does it provide information on the cause, complexity or consequence of IPV.31 Research suggests that the consequences of partner violence differs between men and women, as does the motivations for perpetrating it. Women are more likely to be injured, experience more severe forms of violence, and where violence by women occurs, it is more likely to be in the form of self‐defence.1

IPV is an increasingly significant and important public health issue in New Zealand19 and worldwide.1 Heeding calls from researchers and policy makers,1,8,19 we present prevalence and concordance of IPV reports in an ethnically representative sample of Pacific couples resident in New Zealand. These measures will be used to begin to inform the development of the multisectorial and collaborative approaches needed to reduce the prevalence of this violence. However, to adequately understand and prevent IPV, future research within the Pacific Islands population is required. Risk factors for IPV such as time lived in New Zealand, level of acculturation, and level of family or community networks need to be identified and understood. Research into the microsequencing of violence and the roles of initiation, resistance and self‐defence by sex is needed. Similarly, research that provides an understanding of the impact that IPV exposure has on the functioning of the family and the child's development is needed, and is vital to prevent current and future individual and society costs. Moreover, a deeper understanding of Pacific family structure and function within the context of New Zealand and how it relates to IPV needs investigation and articulation. Understanding the similarities and disparities between indigenous and immigrant Pacific peoples' beliefs, functions and behaviours coupled with knowledge of risk factors and the consequence of IPV is required before culturally sensitive and appropriate approaches can be fully developed to reduce the occurrence of this violence.

What is already known

Intimate partner violence is global, prevalent and enormously costly to individuals and society. To date, only three critical published studies have been expressly designed to explore individuals' reports of partner violence that could be related to their partner's; one in a convenience sample of convicted men, one in a convenience sample of couples seeking marital therapy and one in a representative US sample. Research findings are contradictory, with the first study showing that it is overwhelmingly men who perpetrate violence against women partners whereas the other two studies reveal that women are as likely as men to perpetrate violence against their partner.

What this paper adds

Using a standardised measure, this paper reports prevalence of intimate partner violence (IPV) in an ethnically representative sample of Pacific couples and describes the symmetry of both partners' reports of that violence. Men systematically under‐reported both their perpetration and victimisation of IPV. However, women are as likely as men to perpetrate violence against their partner.

Policy implications

Worldwide, prevention strategies have been developed on the premise that it is men who perpetrate violence against women. Future research and understanding is required to determine the motivation and consequence of women's violence towards their male partners.


The Pacific Islands Families (PIF) study is supported by grants awarded from the Foundation for Science, Research and Technology, the Health Research Council of New Zealand and the Maurice and Phyllis Paykel Trust. The authors gratefully acknowledge the families who participated in the study, the Pacific Peoples Advisory Board and the other members of the PIF research team.


CTS - Conflict Tactics Scale

IPV - intimate partner violence


Competing interests: None.


1. Krug E G, Dahlberg L L, Mercy J A. et alWorld report on violence and health. Geneva: World Health Organization 2002, (accessed 1 Feb 2007)
2. World Health Organization Violence against women: a priority health issue. Geneva: Department of Women's Health and Development, World Health Organization, 1997, (accessed 1 Feb 2007)
3. World Health Organization WHO Multi‐country study on women's health and domestic violence against women. Geneva: World Health Organization, 2005
4. Fathalla M F. When home is no longer safe: intimate‐partner violence. Lancet 2005. 3661910–1911.1911 [PubMed]
5. Fanslow J, Robinson E. Violence against women in New Zealand: prevalence and heath consequences. N Z Med J 2004. 1171173
6. World Bank World development report: investing in health. Washington DC: Oxford University Press, 1993
7. Snively S. The New Zealand economic cost of family violence. Wellington: Coopers and Lybrand for Family Violence Unit, Department of Social Welfare, 1994
8. Krahe B, Bieneck S, Moller I. Understanding gender and intimate partner violence from an international perspective. Sex Roles 2005. 52807–827.827
9. Magdol L, Moffitt T E, Caspi A. et al Gender differences in partner violence in a birth cohort of 21‐year‐olds: bridging the gap between clinical and epidemiological approaches. J Consult Clin Psychol 1997. 6568–78.78 [PubMed]
10. Dobash R P, Dobash R E. Women's violence to men in intimate relationships: working on a puzzle. Br J Criminol 2004. 44324–349.349
11. Caetano R, Ramisetty‐Mikler S, Field C A. Unidirectional and bidirectional intimate partner violence among white, black, and Hispanic couples in the United States. Violence Vict 2005. 20393–406.406 [PubMed]
12. Simpson L E, Christensen A. Spousal agreement regarding relationship aggression on the Conflict Tactics Scale‐2. Psychol Assess 2005. 17423–432.432 [PubMed]
13. Koziol‐McLain J, Gardiner J, Batty P. et al Prevalence of intimate partner violence among women presenting to an urban adult and paediatric emergency care department. N Z Med J 2004. 1171174
14. Koloto A H, Sharma S. The needs of Pacific women when they are victims of family violence. Soc Policy J N Z 2005. i2684–96.96
15. Paterson J, Feehan M, Carter S. et al Intimate partner violence among a cohort of pacific women living in New Zealand. J Interpers Violence. In press
16. Statistics New Zealand National Pacific population projections [2001(base)‐2021]: hot off the press Wellington: Statistics New Zealand, 2003
17. Statistics New Zealand Pacific progress: a report on the economic status of Pacific peoples in New Zealand. Wellington: Statistics New Zealand, 2002
18. Ministry of Health Te Orau Ora: Pacific mental health profile. Wellington: Ministry of Health, 2005
19. Ministry of Social Development Te Rito: New Zealand family violence prevention strategy. Wellington: Ministry of Social Development, 2002.‐policy/te‐rito.pdf (accessed 1 Feb 2007)
20. Rothman K J, Greenland S. Modern epidemiology. 2nd edn. Philadelphia: Lippincott‐Raven, 1998
21. Paterson J, Tukuitonga C R, Abbott M. et al Pacific Islands families first two years of life study: design and methodology. N Z Med J 2006. 119U1814
22. Straus M A. Leveling, civility and violence in the family. J Marriage Fam 1974. 3613–29.29
23. Straus M A. Measuring intrafamily conflict and violence: the Conflict Tactics (CT) scales. J Marriage Fam 1979. 4175–88.88
24. Straus M A. Measuring intrafamily conflict and violence: the Conflict Tactics (CT) Scales. In: Straus MA, Gelles RJ, eds. Physical violence in American families: risk factors and adaptions to violence in 8,145 families New Brunswick. NJ: Transaction, 1990. 29–47.47
25. Straus M A. The Conflict Tactics Scale and its critics: an evaluation and new data on validity and reliability. In: Straus MA, Gelles RJ, eds. Physical violence in American families. New Brunswick, NJ: Transaction, 1990. 49–73.73
26. Landis J, Koch G. The measurement of observer agreement for categorical data. Biometrics 1977. 33159–174.174 [PubMed]
27. Agresti A. Categorical data analysis. New York: Wiley, 1990
28. Amnesty International Women confronting violence (AI Index: ASA 43/001/2004). Solomon Islands: Amnesty International, 2004,$File/ASA4300104.pdf (accessed 1 Feb 2007)
29. Tukuitonga C, Finau S A. The health of Pacific people in New Zealand up to the early 1990's. Pac Health Dialog 1997. 459–67.67
30. Kimmel M S. “Gender Symmetry” in domestic violence: a substantive methodological research review. Violence Against Women 2002. 81332–1363.1363

Articles from Journal of Epidemiology and Community Health are provided here courtesy of BMJ Group