Table 1

Overview of planned methods and analyses for the development of the MUSE Scales

Phase 1: item development
1.1 Domain specificationWe specified domains and sub-domains based on an existing conceptual model of empowerment and a systematic literature review.
1.2 Item generationWe developed candidate items for each sub-domain through a multi-step process, beginning with a systematic literature review and a parallel landscape analysis.
1.3 Face validity and content validity assessmentWe assessed face validity through cognitive interviews (CIs) conducted in two sites (Tiruchirappalli, India and Kampala, Uganda) and content validity through evaluation by four expert reviewers. We also conducted key informant interviews (KIIs) for additional insight.
1.4 Item refinementWe revised items based on field notes from CIs, KIIs and expert feedback, adding, eliminating, combining, and re-phrasing items.
Phase 2: scale development and initial validation
2.1 Content validity assessment for new and modified itemsWe conducted CIs with respondents in Tiruchirappalli and Kampala to assess the refined items.
2.2 Survey participant selection in two sitesWe randomly sampled female participants ages 18+in Tiruchirappalli and Kampala in purposively selected neighbourhoods in coordination with BMGF-funded partners and local government.
2.3 Survey implementation and management in two sitesWe trained enumerators who carried out tablet-based data collection with approximately 1000 women in each of the two sites. Teams also targeted up to 100 women per site for retesting.
2.4 Statistical analysisWe developed an analysis plan a priori, which included the following steps: 2.4 a) initial item reduction; 2.4b) factor extraction and further item reduction using exploratory factor analysis (EFA); 2.4 c) item reduction using item response theory (IRT); 2.4d) dimensionality confirmation using confirmatory factor analysis (CFA); 2.4e) measurement invariance assessment using multiple-group CFA and IRT; 2.4 f) scale scoring; 2.4 g) reliability testing; and 2.4 h) validity testing.
Phase 3: scale evaluation and further validation
3.1 Content validity assessment for new and modified itemsWe will conduct CIs with respondents in two additional locations to assess any items refined or added based on statistical analyses (2.4).
3.2 Survey participant selection in three sitesWe will administer cross-sectional surveys to a validation sample in three cities. The sampling strategy, participant selection strategy, and inclusion criteria are the same as Phase 2.
3.3 Data collection and managementWe will conduct enumerator training and survey implementation using tablet-based data collection.
3.4 Statistical analysisWe will conduct EFAs and IRT on data from each site to evaluate any new items generated in Phase 2, and conduct CFAs to test the factor structures identified in Phase two using the same fit indices and criteria to assess model fit. We will test for measurement invariance at both the group and item level using multiple-group CFA and IRT methods; calculate the reliability coefficient; and test for convergent, discriminant, known-groups, and external criterion validity.