Abstract: Using data from the 2008 Survey of Internal Migration and Health in China, this study examines the reliability and validity of selfrated general health for the Chinese population. Results show that selfrated general health is a highly reliable measure of individual health. Two repeated measures of selfrated general health in the survey are quite consistent and the difference between the two answers reflects random variations rather than any systematic biases. Nonetheless, there is also some evidence that selfrated general health is likely to be affected by question orders in a survey.
In addition, this study examines the validity and potential reporting bias of selfrated general health by fitting Hopit models. Results show that selfrated general health is a valid summary measure of individual’s selfperceived and known health conditions, although it does not reflect bodily functional changes that can hardly be perceived. The response of selfrated general health is strongly correlated with respondent’s chronic medical conditions, the occurrence of acute illness, selfperceived pains/discomfort, insomnia symptoms and depression; however, it is only weakly correlated with objective biometrics such as blood pressure and lung capacity.
Finally, it is worth noting that there exists complicated reporting heterogeneity in selfrated general health among different social groups. Due to differences in rating standard, expectation and cognitive capability regarding health, different social groups respond to the question of selfrated general health differently. After controlling for all the specific measures of individual health in the data, older respondents tend to underestimate their true health status, while those better educated and respondents with higher family incomes tend to overestimate their true health conditions. The existence of reporting heterogeneity is likely to hamper the crosspopulation comparability of selfrated general health.

Key words: reliability , validity , reporting heterogeneity, self-rated general health