Page 233 - PhD GT

score at the follow-up points would be an acceptable measurement. If the amount of change, or in the case of the present study, the nature of change is the focus of the investigation, the use of change scores would seem to be necessary. The question then is how to interpret change scores.
10.4.2 Criteria for interpreting the significance of changes in dependence
As the purpose of the present study was to examine the nature of change in dependence, some criteria for assessing the question of when meaningful change on the key metric could be said to have occurred were required. In Chapter 9, methods for calculating both statistically reliable and clinically significant change were applied in the analysis of the correlates and the predictors of change. In outlining a method for determining the criteria for statistically reliable change, Jacobson et al. (1999) proposed that “the magnitude of the change can be said to be statistically reliable when it exceeds that which could be attributed to chance or measurement error” (Jacobson et al. 1999 p. 300). The reliable change index (RCI) proposed by Jacobson and Truax (1991) for establishing whether statistically reliable change had occurred, described briefly in Chapter 9, involved dividing the magnitude of change during the course of therapy (or the measurement period) by the standard error of the difference score and in this way was designed to distinguish real change from change that would result from fluctuations of an imprecise measuring instrument. Reliable change scores were computed for the follow-up samples and applied in analysis of the predictors of change in dependence. Jacobson et al. (1999) describe this metric as being capable of determining clinical improvement as well as clinical deterioration.
Clinically significant change, according to Jacobson and his colleagues, can be said to have occurred when clients end up within the range that renders them indistinguishable from well-functioning people (Jacobson et al. 1999 p. 300) though they state that depending upon the nature of the condition and how much change it is realistic to expect, this criterion may potentially be excessively stringent. Furthermore, the question of where people end up may be of greater interest as a treatment outcome criterion than a criterion for examining the nature of change as in the case of the present study. However, since the study population was a treatment population, application of this criterion to changes in dependence in the follow-up samples was explored.
Three methods were proposed for demonstrating that clients had moved from the dysfunctional to the functional and these were described in Chapter 9; these methods have differing degrees of stringency and the choice of which one to use may depend upon the availability of normative data. The stringency of these methods is said to be a function of the degree of overlap between the two distributions, the functional or normal and the dysfunctional (Jacobson et al. 1999
228