Measuring the Impact of NIT Experiments on Work Effort

In spite of exhaustive research by many analysts on the data generated by the NIT experiments, uncertainty remains over whether work reductions in the experiments should be considered "large" or "small." The authors of this paper argue that this uncertainty arises in part because different analysts have implicitly measured the responses of different groups of individuals exposed to the NIT treatment. The authors use data from the Seattle-Denver experiment to provide estimates of several measures of average work response, including corrected estimates of some previously proposed measures. They conclude that the corrected estimates show a consistent pattern: as the fraction of NIT recipients rises in any group, the observed reduction in labor supply also rises.