Predicting NCLEX-RN performance: cross-validating an identified classification procedure

This study was an attempt to cross-validate a procedure from an earlier study for predicting NCLEX-RN performance. In the original study, a discriminant function was developed on 313 baccalaureate nursing graduates (1988 through 1990), which used 15 variables to correctly categorize more than 91 per cent of them on NCLEX-RN performance. This study used 142 subjects who graduated in 1991 and 1992 to evaluate the adequacy of the classification procedure. Subjects in this more recent sample differed significantly from those in the first on several variables, including SAT scores, physiology grades, and nursing grades. Eight-four per cent of the subjects in the cross-validation sample were correctly classified using the original discriminant function. More than 62 per cent of failures and 87 per cent of passes were correctly categorized. These results suggest that it is statistically and ethically sound to use this discriminant function to identify future students in danger of NCLEX-RN failure.