A review of Understanding Society questionnaires identified our usage of the Audit-C alcohol consumption questions differed somewhat from the questions the NHS uses for its measures. They differ in how the surveys ask about alcohol in either drinks or units, with the latter being a specific amount of alcohol. Understanding Society surveys have asked those who have had a drink in the past 12 months about how often they have a drink with alcohol, how many drinks they have on day they have alcohol, and how often they have 6+ (for women) or 8+ units (for men) of alcohol on days they are drinking, a drinks-drinks-units pattern.
The NHS asks similar questions but asks for them in terms of drinks-units-units.
Given people may not think in terms of units, drinks may be more intuitive and conform to how respondents formulate responses. One purpose of this experiment is to determine whether asking about drinks or units alters responses significantly. To address this question as clearly as possible, the NHS version was compared to a version asking these same questions in a drinks-drinks-drinks pattern.
For respondents completing the survey with a face-to-face interviewer, this module was asked in self-completion module, to minimise social desirability bias.
The controlling variable, determining which of the two sets of questions respondents received, equally and randomly allocated at the household-level, is on record o_hhsamp_ip:
ff_alcoholw15 (1/2each, allocation stratified by sampleorig ff_gridmodew15 ff_incentw15 ff_infcarew15)
1 = NHS version
2 = Drinks-only version
The variables affected by this experiment are in the file o_indresp_ip:
auditc1, auditc2, auditc3, auditc4a, auditc5a, auditc4b, auditc5b
