Assessing and controlling social desirability bias in self-reported survey research: common practices, recommendations, and an application in cyberbullying research

Social desirability (SD) bias poses a significant challenge in self-reported, survey-based research, leading to the overreporting of socially desirable behaviours and the underreporting of undesirable ones. While previous IS studies have attempted to address SD bias through various strategies—such as prevention, assessment, and control—many have treated it as a procedural formality rather than developing a precise and rigorous approach to managing it effectively. To bridge this gap, this study examines common practices and provides up-to-date recommendations across seven key areas: (1) prevention and assessment, (2) the use of SD scales for detection and control, (3) controlling SD bias in research variables, (4) interpreting SD bias through correlations, (5) its impact on model validity, (6) its effects on causal relationships, and (7) the application of self-deception enhancement and impression management scales for addressing specific biases. Through a comprehensive literature review of prior IS research from these seven perspectives and an empirical examination of the contextual suitability of SD scales in online harassment bystander behaviour, this study offers a more structured and effective approach to detecting and controlling SD bias. By improving the reliability of self-reported data in socially (un)desirable contexts, this research strengthens the validity and methodological rigour of IS findings.