Eliminating Aggregation Bias when Estimating Treatment Effects on Combined Outcomes with Applications to Quality of Life Assessment

Researchers are often interested in combined measures such as overall ratings, indices of physical or mental health, or health-related quality-of-life (HRQoL) outcomes. Such measures are typically composed of two or more underlying discrete variables. I show that estimating the effect of a treatment on the combined measure is biased with non-random treatment selection. I provide a solution to this problem by adopting an alternative estimator that first estimates treatment effects on the underlying variables and then combines these effects into an overall effect on the combined outcome of interest.