The Impact of Respondent Burden on Current Drinker Rates

Increasing response burden in alcohol surveys combined with filter questions to exclude abstainers, results in systematically missing data in questions on alcohol consumption as abstainers are not required to answer them.

The aim of the current study is to assess the impact of responder burden on current drinker rates in a large scale Australian survey.

23,855 Australian adults completed the National Drug Strategy Household Survey in 2013 and answered increasingly complex questions on alcohol consumption.

Although 80% of respondents stated that they had consumed alcohol in the past 12 months, the current drinker rate appears to be 78% excluding, or 74% including missing data if taken from the quantity frequency measure. When respondents are then asked to give more detailed responses in a graduated frequency measure, current drinker rates appear to be at 75% or 73%, excluding or including missing data.

The rate of abstention in alcohol survey research is artificially inflated when more complex survey methods are used. Excluding missing data only partially corrects for this. Given that more sensitive analyses are usually performed on more detailed survey questions, rates of abstention and consumption should be adjusted to account for systematically missing data.