Having conducted the survey, however, the Fox Study concludes that the results are unreliable, and discards many of the responses in favor of ad hoc corrections based on a subset of the data which more closely match the authors’ a priori expectations.