Background

Intimate Partner Violence (IPV) is a global public health and human rights issue. Its prevalence has been shown to range from 15% to 71% worldwide. Health consequences of IPV vary from physical health to reproductive health, and to mental health in women.

Attitudes towards violence describe the perception about whether a person justifies wife beating or accept wife beating. Positive attitudes towards violence were significantly associated with IPV victimization of wives and perpetration of husbands. Hence, attitudes towards violence may also influence the reporting of IPV victimization by wives and IPV perpetration by husbands.

Data on IPV are often subject to considerable reporting bias, particularly when derived from one-couple responses. Use of couple data can provide more reliable estimate of prevalence than one partner's data. However, couple data may also require additional efforts to resolve divergence in information provided between partners. Cognitive factors, such as attitudes towards violence, may explain the relationship with IPV reporting.

In Nepal, women are traditionally expected to subordinate their husbands. Prevalence of violence is high. Many women believed that a husband was justified in beating his wife under certain circumstances. Such cultural characteristics may influence the reporting of IPV by couples in Nepal differently from other developing countries.

However, there was no study investigating the characteristics and collateral validity of couple data and the factors associated with reporting IPV victimization among women and IPV perpetration among men in the Nepalese cultural context. Hence, this study aimed to examine the divergence between husbands' reports of IPV perpetration and wives' reports of IPV victimization, and to assess whether couples' attitudes towards wife beating are associated with such IPV reporting practices.

Methods

This is a cross-sectional study conducted in the Kathmandu Valley, Nepal during August and September 2011. In total, 717 couples who matched inclusion criteria were analysed. Using two separate questionnaires, each for wives and husbands, questions were asked to collect lifetime experience of physical, sexual, and/or emotional violence from husbands. Attitudes towards violence were measured by their acceptance of physical violence.

The level of agreement and disagreement for physical, sexual, and emotional violence were assessed by percent agreement and Cohen's kappa. I run multiple logistic regressions to examine factors predicting divergence of report. Then, the multilevel modelling was constructed to determine whether individual's attitudes towards violence influence on own reporting of violence towards wives. Lastly, dyadic effects of attitudes towards violence on IPV reporting among couples were assessed by the Actor-Partner Interdependence Model framework.

Results

Levels of agreement in the reporting violence were significantly low within couples based on kappa statistics (p<0.001). In general, husbands reported more violence than wives (physical 11.29% vs.13.52%; sexual 3.77% vs.5.86%; emotional 7.39% vs.7.81%). Marital relationship was strongly associated with divergent reporting. Lower regret for marriage and higher satisfaction with marriage were associated with lower disagreement of reporting.

Attitudes towards violence were positively associated with IPV reporting of couples in physical IPV (β=0.32, p=0.03), emotional IPV (β=0.63, p<0.001), and any IPV (β=0.42, p<0.001), even after controlling for partner's effects. Likewise, wife's mistreatment of husband was also positively associated with IPV reporting of couples in physical IPV (β=1.27, p<0.001), emotional IPV (β=0.67, p=0.023), and any IPV (β=0.77, p=0.001). Partner's attitudes towards violence and partner's report of wife's mistreatment of husband were also associated with IPV reporting in physical IPV (attitudes β=0.40, p=0.003; wife's mistreatment β=1.04, p<0.001), and any IPV (attitudes β=0.29, p=0.015; wife's mistreatment β=0.72, p=0.007).

Interaction terms indicated that the association of respondent's attitudes towards violence with IPV reporting were stronger for husbands' reporting of IPV perpetration than that for wives' reporting of IPV victimization in all type of IPV. On the other hand, the association of partner's report of wife's mistreatment of husbands were stronger among wives than among husbands in any IPV. Violence during childhood and marital satisfactions were all positively associated with reporting of IPV victimization and perpetration by both wives and husbands, respectively.

Conclusion

Intimate partner violence is still a new field of study in Nepal. The present study added valuable information to the growing knowledge of IPV study in Nepal. In the present study, husbands' reports of their violent behaviour towards their wives are largely different from wives' reports of victimization among Nepalese couples residing the two areas. The prevalence of violence is considerably different between one-partner report and couple's report. This result indicates the methodological difficulties of obtaining true experience of IPV and data from one partner would not be a good proxy of IPV experience. Couple data are useful to address these reporting bias as well as factors that predict the divergent reporting.

This study also suggested the importance of collecting characteristics from both couples. Both couple's attitudes towards violence, violence during childhood, wife's mistreatment of husband as well as marital satisfaction were all positively associated with reporting of IPV victimization by wives and IPV perpetration by husbands. These cognitive factors, which cannot be obtained from one-partner data, may lead to the divergence by increasing or decreasing reporting of wives and husbands.

In conclusion, researchers should collect data from both wives and husbands whenever possible if the aim is to find the prevalence of violence that is more accurate or factors associated with divergent reporting among Nepalese couples. True risk factors of IPV and complete pictures of IPV would not be obtained until all the factors that influenced the divergent IPV reporting are identified. Future research should continue to investigate cognitive, behavioural, cultural, and methodological components of measurements to resolve the reliability issues and improve IPV reporting.

Keywords: Intimate Partner Violence, dyadic, attitudes towards violence, acceptance of wife beating

本研究はネパール、カトマンズに在住する1年以上結婚している女性とその配偶者717組を対象に夫から妻に対する暴力に関して実施した横断研究である。収集データを用いて夫婦間のデータの相違、及び暴力に対する態度と暴力の申告との関連を分析し、下記の結果を得ている。

1. 本研究の質問票は、WHOのMulti-country study for domestic violence against womenを基盤にして作成した。夫から妻への暴力は、身体的、性的および心理的暴力に分けて質問した。夫と妻それぞれの回答を比較したところ、すべての暴力タイプにおいて夫自身の妻への暴力行為の報告の方が、妻自身の夫からの暴力の被害報告よりも多いという結果が得られた(身体的暴力11.29% vs.13.52%;性的暴力 3.77% vs.5.86%;心理的暴力 7.39% vs.7.81%)。

2. 夫から妻への暴力に対する報告に関して、夫と妻それぞれのデータをカッパ係数を用いて解析したところ、すべての暴力タイプにおいてカッパ係数が有意に低いことが示された。これによって夫婦間のデータの一致度は低いという結果が示されたことになる。

3. 夫と妻それぞれの年齢、教育、結婚への満足度、夫婦間の年齢・教育・収入差および世帯収入と夫婦間のデータの相違との関連を多重ロジスティック回帰分析によって解析したところ、夫と妻それぞれの結婚への満足度が、一貫してすべての暴力タイプの夫婦間データの相違に有意に関連していることが示された。結婚への満足度がより強く、結婚への後悔の頻度が少ない夫婦の方が、夫婦間のデータがより一致している傾向が示された。

4. 暴力に対する態度と暴力の申告についての関連はマルチレベルモデリングを用いて解析した。交互作用として夫と妻を項目として追加した。その結果、暴力に対する態度は身体的(β=0.32, p=0.03)、心理的(β=0.63, p<0.001)及びいずれかの暴力(β=0.42, p<0.001)の申告と有意に関連があることが示された。また、この関連は夫においてより強く表れていた。暴力に対し肯定的な態度を持つ夫は、妻への暴力行為をより多く申告する傾向があることが示された。

5. 配偶者の暴力に対する態度と個人の暴力の申告との関連を行為者-パートナー 相互依存性モデルを用いて解析した。その結果、暴力に対する配偶者の肯定的な態度は個人の身体的暴力(β=0.40, p=0.003)といずれかの暴力(β=0.29, p=0.015)の申告と有意な関連があった。この関連は夫と妻に等しくみられることが交互作用の解析より示唆された。

以上、本研究はカトマンズに在住する一年以上結婚している夫婦において、夫から妻への暴力に関する夫婦間のデータの一致度は低く、夫婦それぞれの暴力に対する態度は夫婦の暴力の申告と関連があることを示した。本研究は、まだ歴史の浅い夫婦間暴力研究においてデータ収集法とデータ分析・解釈のあり方について重要な貢献をなすものであり、学位の授与に値するものと考えられる。