Do People Report Happiness Accurately? A Method for Validating Survey Responses Without Ground-Truth Benchmarks

We introduce a method for validating survey responses in contexts where ground-truth benchmarks are unavailable. The method involves the measurement of “self-reported misreporting” (SRM). We clarify the inferences one can draw about actual misreporting from SRM under various combinations of assumptions. Then we document the accuracy of SRM through an experiment involving an established paradigm for measuring dishonesty. Applying the method to happiness data, we show that SRM is substantial and varies systematically across experiences. We also confirm that those who expect to exaggerate happiness report greater happiness contemporaneously. We conclude that misreporting of subjective well-being is significant and systematic.