Examines how the autoregressive earnings model developed by Ashenfelter to evaluate job-training programs can produce badly biased estimates of the magnitude and temporal pattern of program effects. Finds that the decay in Ashenfelter's estimated training effect for men was produced by a time-varying bias in his model, and presents a new, more…

Meta-analytic techniques are employed to find if the impacts of government funded training programs on earnings grew or deteriorated with time. Further, studies are recommended to understand the cost-effectiveness of training programs.

Evaluation of the Work Incentive Program (WIN) indicates that this employment and supportive services program has been very effective in serving welfare recipients with poor work histories on a cost-effective basis. Subsidized public employment is particularly effective in increasing employment and earnings. (MF)

This article develops a production frontier model for the Employment Service and assesses the relative efficiency of the 51 State Employment Security Agencies in attaining program outcomes close to that frontier. This approach stands in contrast to such established practices as comparing programs to their own previous performance. (Author/CT)

The study finds (1) that vocational-technical training taken under the GI Bill raises earning capacity more than does similar training taken under Manpower Development and Training Act (MDTA) and Comprehensive Employment and Training Act (CETA) and (2) that the earnings effect of training are significantly greater for blacks. (EM)