ERIC Number: ED233070
Record Type: RIE
Publication Date: 1983-Apr-11
Reference Count: 0
Unorthodox Uses of Correlation Coefficients.
Kraemer, Helena Chmura
While the potential use of correlation methods has spread to wide contexts and these uses have permeated behavioral research, their scope remains the orthodox applications: to test independence, and to measure association. It is well worth noticing that problems having no obstensible connection to dependence, independence, or association have been resolved by use of correlation theory. Three examples of unorthodox uses are: (1) compare means, (2) compare variances, and (3) properly prepare for and interpret profile analysis of variance results. In place of comparing the mean difference in two independent normal populations having equal variances, one might report the effect size (the mean difference expressed in units of within-group standard deviation) approximated by a correlation coefficient. In samples of two normal distributions, if the data were totally paired, one could also test the equality of variances with a correlation coefficient. To interpret profiles from analysis of variance, a discarded F-test for items is equivalent to correlation between paired subjects' responses as a measure of the similarity of "shape" of two profiles. (CM)
Publication Type: Speeches/Meeting Papers; Reports - Research
Education Level: N/A
Authoring Institution: N/A
Identifiers: Data Interpretation; Effect Size; F Test
Note: Paper presented at the Annual Meeting of the American Educational Research Association (67th, Montreal, Quebec, April 11-15, 1983).