Predictor weights estimated by using multiple linear regression are biased when there is restriction in the range (RR) of the dependent variable. Standardized multiple regression yields partial correlations as weights for the predictors, and these can be corrected for range difference between calibration and application samples. However, additional bias occurs in weights due to unreliability in the dependent variable. Five cohorts (four graduation cohorts and one 3-year cohort) of veterinary medicine applicants were studied to assess the effectiveness of adjusting regression weights prior to estimating future graduate grade point average (GGPA). The cohorts were from the College of Veterinary Medicine at Michigan State University (East Lansing), with application years ranging from 1981 to 1985 and graduation years ranging from 1985 to 1989, respectively. Accepted and rejected applicants were included. Three graduation cohorts (1985, 1986, and 1987) were used to select and weight predictors. These weighted predictors were used to estimate future GGPA for the two remaining cohorts (1988 and 1989), which were used for validity testing of the estimated beta weights. The correction appeared to impact the rank order validity of only one cohort, but even this finding was not statistically significant. Creation of a simulated data set also showed larger validity coefficients when RR was corrected; however, the margin of difference was not statistically significant. Two situations were identified where correction for range may be desirable: (1) in cases that are highly homogeneous, and (2) for predictor weights that are highly homogeneous. Six data tables are included. (RLC)