Premium
Prediction of Human Relations Training Outcome by Traditional and Nontraditional Selection Indices
Author(s) -
ANTHONY WILLIAM A.,
GORMALLY JAMES,
MILLER HENRY
Publication year - 1974
Publication title -
counselor education and supervision
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.608
H-Index - 40
eISSN - 1556-6978
pISSN - 0011-0035
DOI - 10.1002/j.1556-6978.1974.tb00851.x
Subject(s) - psychology , selection (genetic algorithm) , variance (accounting) , contrast (vision) , predictive validity , test (biology) , linear regression , regression analysis , graduate students , variety (cybernetics) , outcome (game theory) , statistics , applied psychology , clinical psychology , artificial intelligence , pedagogy , computer science , mathematics , paleontology , accounting , mathematical economics , business , biology
A variety of traditional and nontraditional selection indices were used to predict the outcome of a graduate‐level course in counseling which was specifically designed to improve the students' human relations skills. Multiple correlations of .75 to .80 were found between a measure of training outcome and various combinations of predictors. A series of multiple linear regression analyses indicated that the traditional selection indices (GPA, GRE‐Verbal, GRE‐Quantitative, Miller Analogies Test) accounted for a very minor amount of the variance in ratings of the final taped interview. In contrast, a newly developed selection measure (trainability index) accounted for almost all of the predictable variance. The need for a new and more efficient predictive model for graduate‐level counselor training is discussed.