A major constraint that was faced in this research was data inadequacies. Small sample sizes and inadequate work history information in the CPS meant that the dataset could only be used for tabular purposes. While the SIPP provided better work history information, the definition of temporary work was not nearly as rich as the one provided by the CPS, and, again, insufficient sample size meant that only one definition of at-risk workers could be used, rather than the plethora of possible measures. Furthermore, although the work history data in the SIPP improve our ability to obtain good matched comparison groups, the match remains problematic for some of the comparisons. In addition, the differences between temporary help employment estimates derived from household surveys, such as the CPS, and establishment surveys, such as the CES, are troublingly large.