A study of the predictability and reliability of graduate grades awarded in the College of Business Administration at the University of Tennessee
The purpose of this study was to analyze the validity of undergraduate grade point average, Admission Test for Graduate Study in Business scores, and other preadmission data as predictors of graduate grade point average. The study also included an estimation of the reliability of graduate grades awarded in the College of Business Administration.
Data were compiled for 157 students who had taken the Admission Test for Graduate Study in Business and who had received Master of Business Administration degrees from the University of Tennessee between Fall Quarter, 1971, and Fall Quarter, 1973. Data compiled for each student included social security number, sex, age when he took his first graduate course, number of quarters elapsed between receiving undergraduate degree and taking first graduate course, undergraduate grade point average, Admission Test for Graduate Study in Business scores, and grades for all graduate courses taken. These data were punched on IBM cards and simple correlation, stepwise regression, and intraclass correlation analyses were performed using appropriate computer programs.
Statistically significant validity coefficients were found for undergraduate grade point average and Admission Test for Graduate Study in Business scores as predictors of graduate grade point average. Multiple correlation coefficients for the combination of these two predictors were statistically significant and large enough to indicate some degree of usefulness in the selection of graduate students. Regression equations determined offer only limited use due to the relatively small portion of GGPA total variation accounted for by the predictor variables.
Graduate grade point averages for students with business-related undergraduate degrees were much better predicted from undergraduate grade point average than were students with nonbusiness undergraduate degrees.
The reliability coefficient computed by the intraclass correlation formula for average grades was statistically significant and agreed well with previous coefficients determined by the intraclass formula.
Restriction of range in the criterion variable due to the awarding of a high percentage of A's and B's to graduate students is believed to have caused an underestimation of reliability and validity coefficients in this study.
Thesis74.W794.pdf
4.09 MB
Adobe PDF
189f87116353fec9efe2fbb9021170aa