학술논문

Monitoring sustainable development goal 5.2: Cross-country cross-time invariance of measures for intimate partner violence
Document Type
article
Source
PLoS ONE, Vol 17, Iss 6 (2022)
Subject
Medicine
Science
Language
English
ISSN
1932-6203
Abstract
Background The persistence and impacts of violence against women motivated Sustainable Development Goal (SDG) 5.2 to end such violence. Global psychometric assessment of cross-country, cross-time invariance of items measuring intimate partner violence (IPV) is needed to confirm their utility for comparing and monitoring national trends. Methods Analyses of seven physical-IPV items included 377,500 ever-partnered women across 20 countries (44 Demographic and Health Surveys (DHS)). Analyses of five controlling-behaviors items included 371,846 women across 19 countries (42 DHS). We performed multiple-group confirmatory factor analysis (MGCFA) to assess within-country, cross-time invariance of each item set. Pooled analyses tested cross-country, cross-time invariance using DHSs that showed configural invariance in country-level multiple-group confirmatory factor analysis (MGCFAs). Alignment optimization tested approximate invariance of each item set in the pooled sample of all datasets, and in the subset of countries showing metric invariance over at least two repeated cross-sectional surveys in country-level MGCFAs. Results In country-level MGCFAs, physical-IPV items and controlling-behaviors items functioned equivalently in repeated survey administrations in 12 and 11 countries, respectively. In MGCFA testing cross-country, cross-time invariance in pooled samples, neither item set was strictly equivalent; however, the physical-IPV items were approximately invariant. Controlling-behaviors items did not show approximate cross-country and cross-time invariance in the full sample or the sub-sample showing country-level metric invariance. Conclusion Physical-IPV items approached approximate invariance across 20 countries and were approximately invariant in 11 countries with repeated cross-sectional surveys. Controlling-behaviors items were cross-time invariant within 11 countries but did not show cross-country, cross-time approximate invariance. Currently, the physical-IPV item set is more robust for monitoring progress toward SDG5.2.1, to end IPV against women.