Page 67 - PhD GT

Table 3.1 Sources of bias in self-report and proposed solutions (adapted from Babor et al. 1987b)
Source of bias Proposed solution
Ambiguous role requirements or task definition
Guarantee anonymity / confidentiality Give clear instructions / examples Emphasise scientific role and separation from the clinical process
Specificity of desired responses
Word questions accurately Ask multiple questions and probe questions
Interviewer bias and unreliability
Train interviewers in appropriate interview techniques Standardise and check protocols
Forgetting, telescoping
Increase question length
Give good instructions
Assist with recall, using memory aids and bounded recall
Response distortion
Alert subject to problem
Include recognition questions Use diaries / calendars / time line
Motivation
Get agreement of commitment
Explain purpose and value
Give clear instructions re the procedures, times and other expectations
Give prompts / encouragement / reinforcement and session feedback
Each of these sources of bias was addressed in preparation for the collection of data in the present studies. The procedures put in place for doing so in the validation study are described in Chapters 5 and 6. The extent to which they were addressed in the measures and procedures of the main study is described in Chapter 7.
3.4 Self-report measures of dependence
There are several methods for measuring dependence, some based upon a particular theory of the nature of dependence and some based in factor analysis of an aggregate view of substance related problems. Those chosen for the present discussion are based in the disaggregation of dependence and problems, measure dependence as a continuum and are self-report measures. The first four instruments described measure the alcohol dependence syndrome, one of these includes the adapted version of the scale. Thereafter, a scale designed to measure the drug dependence
55