Page 211 - PhD GT

The change in dependence at twelve months was computed by subtracting the total dependence score at twelve months from the total dependence score derived at three months and regressing this dependent variable on the three intake demographic variables of age, sex and substance, the social and psychological functioning variables at three months and the number of treatments between three months and twelve months. The results of this analysis are presented in Table 9.21.
Table 9.21
Linear regression analysis for the dependent variable: change in dependence between three months and twelve months: Sample 6c (n=101)
Unstandardized Coefficients
B
-2.03
-3.86E-02 .81 -6.75E-03 .78
-.12 -1.14E-02 -.15 -4.55E-02 8.03E-02
Standardized t Sig.
Coefficients Std. Error Beta
(Constant)
age
sex
substance
LDQ t2
GHQ t2
SSQ t2
ICSAC t2
CBI t2 treatment t2 - t3
6.53 .12 2.30 .03 .16 .36 .19 .22 .05 .19
-.31 .76 -.04 -.32 .75 .03 .35 .72 -.03 -.23 .81
.62 4.94 .001 -.05 -.34 .73 -.01 -.06 .95 -.06 -.69 .49 -.08 -.87 .39 .04 .43 .67
Once again, dependence at three months was the only variable shown significantly to predict the change in dependence score between three months and twelve months.
9.4.1 Predicting reliable and significant change in dependence
The sample used for preliminary examination of the potential predictors of reliable and significant change in dependence at three months was Sample 6a (n = 151). This sample consisted of all the participants in Sample 6 who were also seen at three month follow-up (see Chapter 7). On the basis of the change score calculated for each of the participants in this sample by subtracting total dependence score at three months from total dependence score at intake, a statistically reliable change score was calculated using the formula devised by Jacobson and Truax (1991) and described in Section 9.1 above. A dichotomous measure of reliable change was assigned to each participant, with 1 to denote reliable change and 0 to denote the absence of such change. This was used in the analysis as the dependent variable. Independent variables were 3 pretreatment characteristics: age, sex and primary substance, 5 independent variables referring to psychological and social
206