What are we missing? The profile of non-respondents in the Finnish Gambling 2015 survey

Aims: Population-based gambling surveys provide important information about gambling frequency, problems, attitude and opinions of the general population. This information can be used by social and health care professionals, service providers and policy makers. However, low response rates may cause biased findings. The aim was to define the profile of non-respondents in the Finnish Gambling 2015 survey. Methods: The survey sample (N = 7400) was obtained from the national Population Information System and the survey was conducted using computer-assisted telephone interviews (response rate 62%). The study sample including individuals aged 18-74 was linked to administrative registers to obtain socio-demographic information (sex, age, marital status, education, socio-economic status, net income, residential area) about the respondents and the non-respondents. Register-based information was used to build a non-respondent profile for the survey. Results: The non-response was more prevalent among women, 18- to 24-year-olds, non-married, individuals with primary education, unemployed (vs. self-employed and students) and residents in urban areas. When net income was added to the model, the associations of women and unemployed (vs. self-employed) with non-response became non-significant, while the non-response was more prevalent among the lowest quintile of net income. Conclusions: Socio-economic position was associated with lower response rate which may cause bias while studying gambling behaviour of socio-economically vulnerable individuals. Obtaining additional auxiliary information through record linkage to administrative registers and use of more sophisticated methods for controlling bias caused by non-response, such as multiple imputation, would offer information about the impact of non-response to the results related to gambling and gambling problems.