Validation of a new codependency assessment instrument

Cynthia Celeste Valentin, Marquette University

Abstract

This study concerned the development of an instrument to empirically validate the existence of the construct of co-dependency. Its objective was to establish whether or not a common set of reactions and/or symptoms is experienced by those who co-exist with a chemically dependent partner. This was accomplished by systematically analyzing the most frequently cited variables, hypothesized to be part of the construct of co-dependency via a new co-dependency validation instrument. To date, no such instrument exists. The procedure consisted of an initial content validation process whereby a panel of seven field experts evaluated 18 variables and 243 test items. The evaluation processes included ranking, weighing and rating measures. The results were analyzed and used to develop a 90 question questionnaire which was deemed as representative of the co-dependency domain. This instrument was then subjected to the reliability, construct and criterion-related validation study. In the second phase, 178 experimental subjects and a cluster sample of 30 controls completed the questionnaire. The statistical analyses, performed for the quantitative determination regarding content validity, included mean and standard deviation scores. The instrument's reliability was evaluated using Cronbach's Alpha Reliability Coefficient, construct validity was established via Alpha Factor Analysis, and criterion-related validity was determined by a Stepwise Discriminant Analysis approach. The results from the Alpha Reliability Coefficient measure for the 90 item test for 140 subjects was.97. The reliability results for 14 out of 18 test scores ranged from.74 to.89. The construct validity approach of Alpha Factor Analysis, utilizing a varimax rotation method, resulted in a 2 factor solution. Factor I was explained in terms of an "internal dimension" of co-dependency, whereas Factor II was explained as representing an "external dimension." The discriminant analysis approach to establish criterion-related validity resulted in the selection of eleven variables which met the solution criterion. The classification summary reported an overall 81.25% correct classification. This study accomplished its intent, i.e. to establish the reliability, content, construct, and criterion-related validity of a new co-dependency instrument.

Recommended Citation

Cynthia Celeste Valentin, "Validation of a new codependency assessment instrument" (January 1, 1990). Dissertations (1962 - 2010) Access via Proquest Digital Dissertations. Paper AAI9035638.
http://epublications.marquette.edu/dissertations/AAI9035638

Share

COinS