Two experiments were conducted to evaluate progress from use of recurrent selection in 'Iowa Stiff Stalk Synthetic' (BSSSCO) corn (Zea mays L.) population. Both experiments evaluated interpopulation reciprocal recurrent selection (RRS) and intrapopulation half-sib (HT) and S2 per se (S) selection methods. BSSSCO was the common parental population of the improved populations. The objectives of the experiments were to estimate the genetic variances and genetic correlations among a random set of 50 S2 progenies for each of five populations and to estimate the change in the means of populations improved by different recurrent selection methods. Response to selection was estimated by the use of a genetic model that permitted the effects of genetic drift to be separated from the effects of selection;After nine cycles of RRS, the genetic variance decreased for four out of five traits evaluated, relative to the unimproved population. Seven cycles of HT and three cycles of S selection resulted in a decrease in the genetic variance for two traits and an increase in three traits, relative to the unimproved population. RRS, HT, and S selection methods, however, increased the mean grain yield of a random set of S2 progenies relative to the original BSSSCO population. Estimates of genetic correlations were not consistent in magnitude or sign among populations. This study suggests that genetic variance decreased using RRS but not in the population developed from the combined effects of HT and S selection. Lack of precision in estimating the genetic variances may have resulted in this discrepancy between selection methods;The genetic model used to separate the observed response to selection into effects due to genetic drift and selection suggested that there were no significant differences for grain yield among the RRS, HT or S selection methods due to the effects of genetic drift. The differences in effects due to selection when evaluating the populations per se, populations selfed, and population crosses for response to selection were nonsignificant between methods of selection for grain yield. The relative ease in forming HT and S families, relative to forming RRS families, suggests that HT or S selection would be preferred to using RRS.