The role of self-reporting bias in health, mental health and labor force participation: a descriptive analysis

Abstract

Previous research on male subjects has conjectured that subjective self-reports of health status may lead to an upward bias in the estimated effect of health on labor force participation because subjects who are out of the labor force may be more likely to understate their health status so as to justify their lack of employment. In the descriptive analysis conducted in this article, we compare the effects of mental and physical health status on labor force participation, employing propensity score methods to investigate whether these effects differ for self- and proxy respondents. The authors initially find some evidence that seems to suggest systematic differences between proxy and self-reporters in the effects of health on labor force participation, raising the possibility that self-reporters may be biased in their health assessments. After we control for objective baseline indices of mental and physical health status, however, differences between subjective health assessments and labor force participation become smaller and statistically insignificant. These results suggest that self-reports do not lead to overestimates of the importance of good physical or mental health on labor force participation, after one controls for objective health conditions in the models. Although we conclude that propensity score matching is a useful way to align observations with covariates in estimating the effects of health on labor force participation, we find that the appropriate specification of matching variables—and in particular, the inclusion of objective health measures—is critical for understanding whether self-reporting bias matters in this context.