Three procedures used to control Type I error rate in stepwise regression analysis are forward selection, backward elimination, and true stepwise. In the forward selection method, a model of the dependent variable is formed by choosing the single best predictor; then the second predictor which makes the strongest contribution to the prediction of the dependent variable is chosen, controlling for the effects of the first variable. The process continues so that the variable chosen increases the prediction potential, until remaining variables fail to make any contribution. Backward elimination begins with a model containing all predictors; and, at each step, a variable is eliminated if its removal results in the smallest reduction of effectiveness. True stepwise procedure is a variant of forward selection. To test these procedures, a Monte Carlo computer program, written in FORTRAN IV, was prepared. The results support two conclusions: (1) the probability of erroneously forming a regression model increases as a function of the number of predictors; and (2) as the inter-predictor correlation increases, the probability of making errors decreases. Therefore, the number of predictors and the inter-predictor correlation should be considered when attempting to solve an error rate problem. (MH)