Men, Marriage, and Health: Using Longitudinal Data to Examine Selection Bias

Angela Campbell, Pennsylvania State University

The literature has established that married men have better health than men who never marry, divorce or who are widowed. However, the causal pathway between health and marriage has not been established empirically. This paper uses a sample of 3391 men living in the United Kingdom to examine the question, “are healthy men both more likely to marry and more likely to be healthy later in life?” I hypothesize that healthier men are selected into marriage and that marriage is not significantly associated with self-rated general health over time. Fixed and random effects models are estimated to examine the data longitudinally. The fixed effects model shows that marriage is not associated with general health over time, suggesting that healthier men are selected into marriage. The random effects model shows an association of marriage with general health over time (p<.01). However, this is most likely due to the omitted variable bias.