You are not currently logged in.
Access your personal account or get JSTOR access through your library or other institution:
Conditional ANOVA for Mean Differences When Population Variances Are Unknown
The Journal of Experimental Education
Vol. 55, No. 3 (Spring, 1987), pp. 141-148
Published by: Taylor & Francis, Ltd.
Stable URL: http://www.jstor.org/stable/20151687
Page Count: 8
Preview not available
The comparison of population variances as a preliminary test to the ANOVA for unequal sample sizes was examined. To develop the sampling distribution of the conditional F ratio for comparing means, data were generated using computer simulation methodology for several conditions of varying sample sizes, raw score distributions, and population variances. The results indicated that because of low statistical power, preliminary tests for variance equality were considered of marginal value. Even when the significance level for the preliminary test was increased to .10, Type II error rates were unacceptably high. The difference between the conditional F distribution for mean differences and the theoretical F distribution was greatest when sample sizes were small and unequal.
The Journal of Experimental Education © 1987 Taylor & Francis, Ltd.