Working Paper: NBER ID: w21317
Authors: John Cawley; Anna Choi
Abstract: One of the most robust findings in health economics is that higher-educated individuals tend to be in better health. This paper tests whether health disparities across education are to some extent due to differences in reporting error across education. We test this hypothesis using data from the pooled National Health and Nutrition Examination Survey (NHANES) Continuous for 1999-2012, which include both self-reports and objective verification for an extensive set of health behaviors and conditions, including smoking, obesity, high blood pressure, high cholesterol and diabetes. \nWe find that college graduates are more likely to give false negative reports of obesity and high total cholesterol; one possible explanation for this is social desirability bias. However, college graduates are also significantly less likely to give false positive reports of smoking, obesity, high cholesterol, and diabetes. Because there are far more truly negative people (who are less likely to give a false positive report) than truly positive people (who are more likely to give a false negative report), we find that college graduates report their health significantly more accurately overall.
Keywords: health disparities; education; reporting error; self-reports; NHANES
JEL Codes: I1; I12; I14; I20; I24; I3
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| education (I29) | health reporting accuracy (I10) |
| college graduates (I23) | accuracy in reporting smoking status (C83) |
| college graduates (I23) | accuracy in reporting obesity (C83) |
| college graduates (I23) | accuracy in reporting high blood pressure (I10) |
| college graduates (I23) | false negative reports for obesity (I12) |
| college graduates (I23) | false negative reports for high total cholesterol (Y10) |