Comparing Modes and Samples in Experiments : When Lab Subjects Meet Real People

In this essay we assess the internal and external validity of experiments with online subjects. The assessments are based on tax compliance experiments in which subjects earn real money, are subject to a tax, and can lie about their earnings. We identify the mode effects that affect internal validity by comparing treatment effects in lab and online versions of the experiments. Secondly, we compare the treatment effects across different online (MTurk and U.K. CESS Online) and lab subject pools to determine whether these might affect the external validity of the experimental results