Measurement of Intra-household Resource Control: Exploring the Validity of Experimental Measures

Abstract

We study the validity of experimental methods designed to measure preferences for intra-household resource control among spouses in Ghana and Uganda. We implement two incentivized tasks; (1) a game that measures willingness to pay to control resources, and (2) private and joint dictator games that measure preferences for resource allocation and the extent to which those preferences are reflected in joint decisions. Behavior in the two tasks is correlated, suggesting that they describe similar underlying latent variables. In Uganda the experimental measures are robustly correlated with a range of household survey measures of resource control and women’s empowerment and suggest that simple private dictator games may be as informative as more sophisticated tasks. In Ghana, the experimental measures are not predictive of survey indicators, suggesting that context may be an important element of whether experimental measures are informative.