The psychometric equivalence of two alternate forms of the Controlled Oral Word Association Test.

The equivalence of forms CFL and PRW of the Controlled Oral Word Association Test (COWAT; Benton, de Hamsher, & Sivan; 1994) was examined in samples of healthy college students. Both novel words and qualitative scores (e.g., clusters, switches) were examined using null hypothesis statistical testing (NHST) and inferential confidence intervals (Tryon, 2001). Study 1 used a within-subjects design (n = 66) to examine form equivalence in persons who completed versions CFL and PRW. No mean differences were found for any index, and only one (mean cluster size) differed when examining the standard deviations across forms. Inferential confidence intervals were constructed and Rg (maximum probable difference) values were small and, when compared to prior research, less than the standard error of measure associated with using a single form. In Study 2, we employed a between-subjects design (n = 208) and found that the use of each version generated comparable normative data and correlations among scores. These studies suggest that versions CFL and PRW are best viewed as alternate (rather than parallel) forms at this time; however, they are likely to produce comparable results for most purposes. We discuss the use of inferential confidence intervals to examine form equivalence and directions for future work.