A study was undertaken to determine whether the canonical correlation could exhibit predictive efficiency regarding success in a graduate program. Two criteria and 8 predictors were established as the two variable sets to be correlated. A single significant coefficient of .624 between the two variates indicated that 38.9% of the variance of one set was common to the other variate. However, when subjected to cross-validation, the canonical correlation shrank to a non-significant value. Thus a lack of predictive utility in a population other than the standardization group was indicated.
Get full access to this article
View all access options for this article.
References
1.
Nie, N., Hull, C., Jenkins, J., Steinbrenner, K., and Bent, D.Statistical package for the social services. New York: McGraw-Hill, 1975.
2.
Tatsuoka, M.Selected topics in advanced statistics: Validation studies. Champaign, Illinois: Institute for Personality and Ability Testing, 1976.
3.
Weiss, D.Canonical correlation in counseling psychology research. Journal of Counseling Psychology, 1972, 19, 241-252.
4.
Wiggins, N. and Kohen, E.Man versus model of man revisited: The forcasting of graduate school success. Journal of Psychology, 1971 , 19, 100-106.
