Logistic regression predictions of the probability of a dropout correlated strongest with SCHTYPE (r= 0.323), while OLS predicted values correlated 0.298. The difference between the OLS and Logistic probability correlations with SCHTYPE was not statistically significant (t = 0.89, df = 597, p >0.05). Logistic and OLS estimates correlated 0.957, and the rank correlation between the OLS and logistic estimates was 0.992.