Five methods for computing significant individual client change and improvement rates: Support for an individual growth curve approach.

Abstract

Interest has been renewed in methods for determining individual client change. Currently, there are at least 4 pretreatment-posttreatment (pre-post) difference score methods. A 5th method, based on a random effects model and multiwave data, represents a growth curve approach and was hypothesized to be more sensitive to detecting significant (p < .05) change than the pre-post methods. The change rates produced by the 5 methods were compared in a sample of 73 older outpatients with 3 to 5 assessments per client on a measure of well-being (H. J. Dupuy, 1977). Results indicated that the growth curve approach improvement rate was the highest (68.5%). The growth curve and the Edwards-Nunnally (63.0%) methods produced significantly (p < .05) higher improvement rates than the other 3 methods, with 1 exception.