The use of multi-national web surveys for comparative analysis: Lessons from the European Web Survey on Drugs (2019)

Background: Most comparative drug policy analyses utilise measures of drug use, often from general population surveys (GPS). However, the limitations of GPS are well-recognised, including the small numbers of people who use illicit drugs sampled. Web surveys offer a potential solution to such issues. Therefore EMCDDA conducted a study to assess the potential for using such surveys to supplement information obtained from GPS.
Methods: The European Web Survey on Drugs (EWSD) asked about use of cannabis, amphetamines, cocaine and MDMA in 14 countries from 2016 to 2018. Each participant country translated the questionnaire as necessary and devised its own sampling strategy. Individuals aged 18+, resident in the participant country, who had used one or more of the drugs covered by the survey in the past 12 months were included in the analysis. Participation was anonymous and voluntary.
Results: More than 40,000 people completed the survey, with recruitment mostly through social media. Larger samples of users of all drug types than found in GPS were generally obtained. However, the respondent profiles differed markedly between countries, e.g. the proportion aged 18-24 ranged from 30% to 80%. The results relating to use showed both inter-country similarities and differences, e.g. mean daily amounts of cocaine used varied between countries but increases in amounts used with increased frequency of use were similar. Price data showed good external validity.
Conclusion: Web surveys offer the possibility of collecting information from large numbers people who use illicit drugs quickly and cheaply and can fill important gaps in our knowledge of patterns of use, particularly by recreational users. However, they also have limitations. Standardising questionnaires and approaches to data cleaning and analysis facilitates comparisons between countries but obtaining comparable samples may be challenging. Multinational surveys need to balance standardisation of methods with responsiveness to differing country contexts; our collaborative model does this.