Page 74 - PhD GT

studies.
In order to examine the question of whether the scale was relatively free of socio-cultural
influence, it was tested in English and Irish samples giving initial support to the claim for its neutrality. It has since been used in Brazil, thus lending further support to its cross cultural applicability. Follow-up studies demonstrating its sensitivity to change over time are not reported; for such a purpose to be pursued it could be argued that it would be necessary to alter the instruction “Think about your most recent drinking habits...” (Raistrick et al. 1983 p. 94) to one that referred to a specific period of time.
3.4.4 The Ethanol Dependence Syndrome Scale (EDS)
Another scale which purports to measure some, but not all of the elements of the alcohol dependence syndrome, referred to as the “provisional” elements of the syndrome (Babor 1996 p. 98), the EDS is a 16 item scale tapping five major alcohol dependence syndrome elements, namely salience, impaired control, tolerance, withdrawal and withdrawal relief. Narrowing of drinking repertoire and reinstatement elements were not included because they were deemed to be too difficult to operationalise. Items are rated on a scale measuring frequency of occurrence during the past three months. Babor (1996) cites studies evaluating the internal consistency, factor structure and cross cultural applicability of the scale, its use as a measure to define alcoholic sub-types and in studies of predictors of relapse. Babor (1996) also reported a study of its test-retest reliability using both a structural method proposed by Nunnally (1967) and some features of a dynamic approach proposed by Del Boca et al. (1994), in a sample of participants in Project MATCH (Project MATCH Research Group 1993). Nunnally’s method consists of internal consistency evaluation, correlation with other measures and repeated administrations with the same respondents, whereas the approach of Del Boca and her colleagues is to focus upon the conditions of responding and the reporting requirements. Respondent variables including demographic data, measures of sociopathy, assessments of social desirability response set and cognitive status were proposed as were interviewer ratings of motivation and other psychological states known to affect the reliability of responding. Del Boca further proposed the collection of interviewer demographic data but in the event, Babor (1996) reported on the way in which discrepancy scores between test and re-test results correlated with respondent characteristics only. Interviewer ratings and characteristics were not included.
62