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.

Trial registration: https://ClinicalTrials.gov/NCT03124589, registered 19 January 2017.

社会期望的反应是一种反应偏差，会影响自我报告的准确性。当问题涉及敏感话题时尤其有可能，例如赌博态度、问题和行为。从亚马逊的 mTurk 众包平台招募了 321 名参与者的样本，用于调查在线赌博干预的随机对照试验。来自这项研究的数据被用来检查随着时间的推移，社会期望的反应对赌博自我报告措施的影响。在基线时，自欺欺人 (SD) 是社会理想反应的一个因素，在男性中明显高于女性，并且在家庭年收入超过 20,000 美元的人群中明显高于女性。控制人口变异，进行了混合效应模型检查社会期望响应因素 [SD 和印象管理 (IM)] 与时间之间的关系。在男性中，随着时间的推移，IM 与 NORC DSM-IV 赌博问题筛查 (NODS) 和赌博症状评估量表 (G-SAS) 分数之间存在显着的相互作用。换句话说，与 IM 分数较低的男性相比，IM 分数较高的男性在从基线到 6 个月的随访期间 NODS 和 G-SAS 分数的变化较小。在任何模型中，女性之间或整个样本之间都没有显着的交互作用。控制社会理想反应对成瘾研究中自我报告措施的充分记录的影响应被视为一种有助于减少错误和提高有效性的方法。

试验注册：https://ClinicalTrials.gov/NCT03124589，2017 年 1 月 19 日注册。