THE EFFECT OF CREAMING ON PLACEMENT RATES UNDER THE JOB-TRAINING PARTNERSHIP ACT

The authors investigate the degree to which ''creaming''-nonrandom selection of participants-in Job Training Partnership Act (JTPA) Title II-A programs is responsible for the high placement rates in those programs. An analysis of data from Tennessee JTPA agencies, in conjunction with Current Population Survey data, shows that creaming does take place, especially through non-selection of those handicapped by poor education or poor health. The extent of creaming, however, is not as large as some critics have suggested: the 71 % placement rate in Tennessee would fall only to about 62% if participants were randomly selected from among the economically disadvantaged population eligible for training. In contrast, targeting only high school dropouts for training-which would have the virtue of serving a group with particularly large barriers to employment-would reduce success rates by nearly one-quarter.