A STUDY WAS MADE OF RETRAINING PROGRAMS IN FOUR LABOR MARKETS IN CONNECTICUT, UNDER BOTH THE AREA REDEVELOPMENT ACT AND STATE SPONSORSHIP. THE PURPOSE WAS TO DETERMINE THE BENEFITS AND COSTS TO WORKERS, THE GOVERNMENT, AND THE ECONOMY. THE STATE HAD PIONEERED SUCH PROGRAMS SO A SAMPLE COULD BE STUDIED OF THOSE WHO HAD LONG WORK-HISTORIES AFTER RETRAINING. THE TRAINEES WERE DIVIDED INTO SIX GROUPS--THOSE WHO COMPLETED THE COURSE WITH AND WITHOUT EMPLOYMENT, THOSE WHO WITHDREW WITH AND WITHOUT EMPLOYMENT, AND THOSE WHO REFUSED RETRAINING WITH AND WITHOUT EMPLOYMENT. THREE CONTROL GROUPS WERE TRAINEES WHO, WITHOUT EMPLOYMENT, COMPLETED THE COURSE, WITHDREW, OR REFUSED IT. THE BENEFIT-COST RATIO FOR THE INDIVIDUALS (SUCH AS REDUCED ANNUAL UNEMPLOYMENT AND INCREASED ANIMAL WAGE VERSUS INCOME LOST DURING TRAINING AND HIGHER INCOME TAX AFTERWARDS) WAS NOT AS HIGH AS THE BENEFIT TO THE GOVERNMENT (SUCH AS REDUCED COSTS OF UNEMPLOYMENT AND PUBLIC ASSISTANCE) AND TO THE ECONOMY (THE INCREASE IN GROSS NATIONAL PRODUCT AND REDUCED AGGREGATE UNEMPLOYMENT LEVEL). IT WAS FELT THAT THE BENEFICIAL EFFECTS OF THE RETRAINING PROGRAMS MIGHT BE LOWERED BY THE ADDITION OF THE MANPOWER DEVELOPMENT AND TRAINING ACT PROVISION FOR TRAINING THE HARD CORE UNEMPLOYED. (EB)