On the application of analysis of variance to GSR data: II. Some effects of the use of inappropriate measures.

Abstract

50 boys and 50 girls were tested at 13.5 and retested at 17.5 years of age for GSR to words rated pleasant, indifferent, or unpleasant. GSR measures included resistance change, conductance change, and the logs of these changes. "The findings of this study justify the following conclusions: . . . The four GSR measures examined in this study cannot be used inter-changeably, since when applied independently to the same original scores, some measures yielded highly significant differences for a given comparison, whereas others did not . . … The variability among the obtained F-values arises from the scale characteristics of the four measures as they are applied to the original or derived scores. It may be assumed that those measures which violate the basic assumptions underlying the analysis of variance technique yield biased F-values." (PsycINFO Database Record (c) 2006 APA, all rights reserved)