The Worker Profiling and Reemployment Services (WPRS) program requires states to establish systems for identifying Unemployment Insurance (UI) claimants likely to exhaust their UI benefits and refers them to reemployment services. An evaluation was conducted to assess the reliability of the impact estimates provided in the evaluation of the WPRS program, and to compute revised estimates of the impacts of WPRS programs if a more accurate estimation method could be identified. Data for the evaluation were gathered from the Job Search Assistance Demonstration in Florida, which, beginning in 1995, randomly assigned UI claimants to control groups or treatment groups who received training in job search techniques. The data were to be tested in two phases: In Phase I, using the regression method; and in Phase II, using variants of the matching methods used in other evaluations. The Phase I evaluation found that the linear regression model used in the WPRS evaluation produced accurate impact estimates, while the matched comparison groups tested in this evaluation produced less accurate impact estimates than the linear regression model. Based on the results of Phase I, therefore, it was decided not to proceed to Phase II of the evaluation. (Contains 10 references.) (KC)