How do women’s empowerment metrics measure up? A comparative analysis

Research has identified women’s empowerment as a critical factor for nutritional outcomes and a priority area for understanding women’s mental health status. At the same time, there is no consensus on how empowerment should be measured. The surrounding debate has produced several empowerment metrics that are widely used, yet we know little about whether they can be substituted for one another or their respective strengths and weaknesses. Using data collected from a single sample of women from rural, northern Kenya, we compare five empowerment metrics: The Project-level Women’s Empowerment in Agriculture Index (pro-WEAI) and associated Health and Nutrition Module (HN), Women’s Empowerment in Nutrition Index (WENI), Women’s Empowerment in Livestock Index (WELI), and the Survey Based Women’s Empowerment Index (SWPER). The metrics have shared theoretical origins and are commonly used in the food, nutrition and health spaces to study rural women’s lives across low- and middle-income countries. We examine the metrics’ characteristics, distributions, pairwise correlations and capacity of each metric to predict outcomes often associated with the concept of empowerment: body mass index (BMI) and the Center for Epidemiologic Studies Depression Scale (CES-D). We find striking differences between these common empowerment metrics. The metrics’ correlations with one another are highly variable as are the predictive capacities for both outcomes. Further, our analysis finds that the choice of metric can dramatically influence which individuals are identified as empowered. In sum, our results suggest that while these metrics are used in remarkably similar ways to understand rural women’s empowerment and its consequences, unless they are computed with many identical survey questions, the metrics do not capture the same underlying concept and are not interchangeable. We recommend that our work be replicated elsewhere and caution should be taken when implementing and interpreting research using these metrics, as findings may be highly sensitive to the choice of metric.