Page 128 - PhD GT

were negatively worded and not on the basis of the content of the items (indeed content was very similar to those items not presented in this way) suggests that respondents either did not understand the questions or that there was a consistency in the choice of response to tick, a response set, that was independent of the question being asked. There was, however, nothing to distinguish the negatively worded items in the item analysis, examining inter-item correlations and item total correlations (see Appendix 6 and 7).
No such finding was reported in the original validation study (Heather et al. 1993). However, exactly the same finding was reported in the 1998 validation study (Heather et al. 1998). Heather suggested that a possible remedy would be to change the order of the questions so that the negatively worded items occurred earlier in the scale (as it is they occur at the 4th, 6th, 7th, and 10th items of both Parts 2 and 3) and may thereby prevent the response from being set in this way.
At the outset, the authors of the original Impaired Control Scale were concerned to avoid the problem of whether or not there was intention to control as discussed by Chick (1980a, 1980b). In Chick's standardised interview schedule designed to measure the alcohol dependence syndrome, there was one question in which reference was made to “sticking to a limit”, and one in which the respondent had to decide whether he or she had been “what you would call drunk” (Chick, 1980a, pp. 184-185). A number of difficulties were referred to by Chick: the large number of ‘not applicable’ responses to these items and the question of whether, if the individual did get drunk, had they had an initial intention to do so, had they changed their mind once drinking had commenced or had some internal mechanism, beyond the control of the drinker, caused the continued drinking? In designing the Impaired Control Scale to take these problems into account, Heather et al. (1993) did two things: i) they included a short section with questions on whether the respondent had attempted to control their use or not; ii) they worded the items in all parts of the scale in such a way that they referred to being able to stop once started or to being able to resist starting without making any assumption about intention. In the validation of the adapted instrument and in the test re-test reliability study, some of the respondents who had been abstinent for longer than the three months enquired about found the whole of Part 2 of the scale to be not applicable and this finding was repeated in the main study. These respondents had no problem completing Part 3 of the scale.
Problems of response bias, coupled with the irritation expressed by a small number of respondents at the length and seemingly repetitious nature of the items render the utility of this scale for routine use in some doubt. Furthermore, the marked difference in the number of individuals completing the different parts of the scale raises questions of relevance of the items at different stages of the dependence career. For the purpose of the present study, the psychometric
122