Averaging cross-corrleations between repetitions of two variables was evaluated as a method for over-coming limited (small) subject sample sizes. Analysis revealed greatest utility when reliabilities were low and numbers of replications were large. Simulation revealed less biased and variable estimates based upon averaging correlations than upon corresponding correlaions of averages (sums). Averaging correlations is recommended when subject samples are limited.
Get full access to this article
View all access options for this article.
References
1.
BittnerA. C.Jr.Averaged correlations between parallel measures: Reliability estimation. Proceedings of the Eighth Psychology in the DoD Symposium, USAF Academy, Colorado Springs, April, 1982.
2.
DunlapW. P.JonesM. B.BittnerA. C.Jr.Average correlations versus correlated averages. Manuscript submitted for publication, 1982.
3.
JonesM. B.KennedyR. S.BittnerA. C.Jr.A video game for performance testing. American Journal of Psychology, 1982, 94, 143–152.
4.
KendallM.StuartA.The advanced theory of statistics, Vol. 1 (Fourth Ed.). New York: MacMillian, 1977.
5.
SchmittF. L.HunterJ. E.Statistical power criterion-related validity studies. Journal of Applied Psychology, 1976, 61, 473–485.
6.
SchimttF. L.HunterJ. E.Development of a general solution to the problem of validity generalization. Journal of Applied Psychology, 1977, 62, 529–540.
7.
SteigerJ. H.Tests for comparing elements of a correlation matrix. Psychological Bulletin.1980, 87, 245–251.
8.
TrattnerM. H.O'LearyB. S.Sample size for specified statistical power in testing for differential validity. Journal of Applied Psychology, 1980, 65, 127–134.
