Examining Change in Self-Reported Gambling Measures Over Time as Related to Socially Desirable Responding Bias

Abstract

Socially desirable responding is a response bias that can affect the accuracy of self-reports. It is especially likely when questions address sensitive topics, such as gambling attitudes, problems and behaviours. A sample of 321 participants were recruited from Amazon's mTurk crowdsourcing platform for a randomized controlled trial investigating an online gambling intervention. Data from this study was used to examine the influence of socially desirable responding on gambling self-report measures over time. At baseline, self-deception (SD), a factor of socially desirable responding, was significantly higher among males than females and among those with household incomes greater than $20,000 per year. Controlling for demographic variability, mixed effects models examining the relationship between socially desirable responding factors [SD and impression management (IM)] and time were conducted. Among males, there were significant interactions between IM with the NORC DSM-IV screen for gambling problems (NODS) and the gambling symptom assessment scale (G-SAS) scores respectively, over time. In other words, males with higher IM scores, demonstrated less change in NODS and G-SAS scores from baseline to 6-month follow-up compared to males with lower IM scores. There were no significant interactions in any models among females or among the full sample. Controlling the well-documented effect of socially desirable responding on self-reported measures in addictions research should be considered as a method to help reduce error and improve validity. Future research should continue to examine the effect of this bias on gambling measures over time and in each gender.