ERIC Number: ED255571
Record Type: RIE
Publication Date: 1985-Apr-2
Reference Count: 0
Power Analysis of Selected Parametric and Nonparametric Tests for Heterogeneous Variances in Non-Normal Distributions.
Olejnik, Stephen F.; Algina, James
The present investigation developed power curves for two parametric and two nonparametric procedures for testing the equality of population variances. Both normal and non-normal distributions were considered for the two group design with equal and unequal sample frequencies. The results indicated that when population distributions differed only in scale, the Klotz procedure consistently provided the most sensitive test of variance equality. With symmetric distributions which differ in both mean and scale the Klotz procedure with mean aligned data was the most powerful procedure. When sampled distributions were skewed and differed in mean as well as scale the O'Brien or Brown-Forsythe procedures were preferred. (Author)
Publication Type: Speeches/Meeting Papers; Reports - Research
Education Level: N/A
Authoring Institution: N/A
Identifiers: Type I Errors; Variance (Statistical)
Note: Paper presented at the Annual Meeting of the American Educational Research Association (69th, Chicago, IL, March 31-April 4, 1985).