Four main messages emerge from the study of subjective survey data. First, a large experimental literature by and large supports economists' skepticism of subjective questions. Second, put in an econometric framework, these findings cast serious doubts on attempts to use subjective data as dependent variables, because the measurement error appears to correlate with a large set of characteristics in behaviors. Third, these data may be useful as explanatory variables. Finally, the empirical work suggests that subjective variables are useful in practice for explaining differences in behavior across individuals. Changes in answers to these questions, however, do not appear useful in explaining changes in behavior.