Page 121 - PhD GT

impaired control scale for heroin users, no alternative scale for the measurement of impaired control in this group was found. However, as is discussed in the earlier part of the chapter, impaired control is seen as the central feature of dependence as it is understood in the psychological sense and it was therefore felt appropriate to test construct validity by comparison with sample scores for the measure of dependence, the Leeds Dependence Questionnaire (LDQ) (see Chapter 3). Correlations among scores for the three parts of the ICSdrg and the LDQ are shown in Table 5.2. For this part of the analysis, impaired control items 3 and 7 in the LDQ were removed. In the development of the LDQ, item 3 ‘Do you feel that your need for drugs is too strong to control’ was designed to be an inability to abstain item, whereas ‘Do you feel that you have to carry on once you have started’ was designed to be the impaired control item. However, since both meanings were used and investigated in the Impaired Control Scale, it was thought appropriate to remove both items in the comparison. Each part of the ICSdrg showed a significant correlation with total dependence score (see Table 5.2).
High scores on ICSFC and ICSPC refer to experienced greater difficulty and expected greater difficulty respectively in controlling heroin use. Both these show strong positive correlations with each other and with dependence scores. Attempts to control heroin use (ICSAC) were significantly negatively correlated with failed control (ICSFC) and with expected failed control (ICSPC) as well as with dependence scores.
Table 5.2 Correlations among ICSdrg part scores and the LDQ; Sample 4 (n=118, data missing for 7 respondents).
ICSFC ICSPC LDQTOT (r)
ICSAC (attempted -.44* control)
ICSFC (failed control) _ ICSPC (perceived
control)
* p=.000; ** p=.001
impaired control and inability to abstain items removed.
Construct validity may also be measured by factor analysis; where it is proposed that a single construct is being measured by a scale, factor analysis may be able to demonstrate that scale
-.51* -.32**
.66* .49* _ .49*
LDQTOT (r) refers to the LDQ total score with
109