LAUSR

Poor mental health and mental disorders are amongst the largest causes of disease burden across the globe, and in developed countries mental illness is on the rise. Studies of the predictors and consequences of ill mental health often rely on surveys. However, there is very little evidence of whether or not there are differences in the ways in which individuals with good and poor mental health engage with the survey interview process, and on their subsequent survey interview outcomes. We examine the associations between respondent mental wellbeing and survey interview outcomes using 14 years (2001-2014) of annual, nationally-representative, Australian panel data (n∼200,000) and state-of-the-art multilevel regression models. We find that individuals with poorer mental health and mental disorders are generally more likely than individuals with better mental health to be deemed by interviewers as being suspicious of the study, experiencing issues understanding survey questions, and being uncooperative. We also find that these individuals are comparatively more likely to experience panel attrition, complete interviews featuring higher item-level missing data, and fail to complete/return self-complete questionnaires. While the magnitude of these effects is moderate, our findings suggest that data collectors, researchers and policymakers need to remain cognizant of potential issues emerging from differences in the ways in which individuals with poorer and better mental wellbeing engage in social surveys