The article applies a "fixed-effect" methodology to evaluate the training impacts of the Vocational Rehabilitation program. Two central issues are identification of an acceptable comparison group and adjusting for sources of selection-bias. The use of program dropouts as a comparison group is examined with a modified Hausman test. The results suggest spending on VR may be ore cost-effective in view of the high public cost of serving the disabled in a full dependency mode. These estimates of VR training impacts are compared with the performance of CETA, a public program serving the non-disabled which has been examined extensively using the fixed-effect estimator.
Dean, David H. and Robert C. Dolan. 1988. "Selection-Bias Correction in an Evaluation of the Vocational Rehabilitation Program." E.C.R.S.B. 88-8. Robins School of Business White Paper Series. University of Richmond, Richmond, Virginia.