Page 214 - PhD GT

dichotomous dependent variable, number of treatment events between intake and three months emerged as the one significant predictor in logistic regression analysis (see Table 9.23).
Table 9.23
Logistic regression analysis for the dependent variable: abstinent or not during the past seven days at three month follow-up: Sample 6a (n=151)
Variable
Age
Sex
Substance
LDQ t1
GHQ t1
SSQ t1
ICSAC t1
CBI t1
treatment t1 to t2
Constant
Beta Standard error
.01 .02 -.82 .47 -.01 .01 .00 .03 -.02 .06 -.04 .04 .08 .04 .00 .01 .20 .09
.15 1.27
df Significance R
1 .52 .00 1 .08 -.07 1 .26 .00 1 .89 .00 1 .80 .00 1 .37 .00 1 .07 .08 1 .92 .00 1 .02 .14
1 .91
When abstinence or not at twelve month follow-up was the dependent variable regressed on the three demographic variables and functioning variables measured at three months, none of the independent variables had correlations with significance <.05 with the dependent variable. Dependence score at three months was the independent variable showing a correlation with abstinence at twelve months which was closest to reaching significance. (p<.07, see Table 9.24).
The second variable referring to the use of the substance was the number of days using in the past week. A difference score was computed for the number of days per week using at three months compared to intake (by subtraction) and at twelve months compared to three months. Three of the independent variables emerged as being significantly correlated with this dependent variable when the other independent variables had been accounted for. These were substance group, dependence at intake and the number of treatment events between intake and three months. The data are presented in Table 9.25.
209