![Sample our Behavioral Sciences journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/110801/TOC-Subject-Sample-2021-Behavioral-Sciences.jpg"})
Abstract
The equivalence of forms CFL and PRW of the Controlled Oral Word Association Test (COWAT; Benton, de Hamsher, & Sivan; Citation1994) 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, Citation2001). 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.
ACKNOWLEDGEMENT
The authors acknowledge and express thanks to Laura Evans-Blake, Lindsay Clampitt, Eva Hanouskova, Kristen Hess, Sara Parham, and Meg Williams for their assistance with data collection and to Emily Bennert and Misty Deason for their assistance with scoring COWAT protocols. Additionally, the authors thank Warren Tryon for his advice regarding data analysis and reporting.
Notes
*p < .001.
Rg = The maximum probable difference associated with using different versions.
*p < .05 Using F-test.
a p < .05 Using Levene's Test of Homogeneity of Variances.
Rg = The maximum probable difference associated with using different versions.
Note. Values in the upper diagonal represent correlations among COWAT indices for version CFL, while values for version PRW are shown in the lower diagonal.
Persevs. = Total number of perseverations across all three trials.
*p > .01.
**p > .001.