Social desirability bias reflects respondents' propensity to answer what is perceived as socially acceptable, rather than the respondent’s true opinion or behavior.
Social desirability bias can cause research to be vulnerable to over- or under-reporting of certain behaviors. Researchers inquiring about behaviors and consumption will encounter social desirability bias as these behaviors are perceived as socially very desirable or undesirable. When asking about purchases of sustainable or organic products respondents subject to social desirability bias will - in aggregate - over-report this desirable behavior. On the other hand, a survey about consumption of sweets or fast food might result in respondents under-reporting this behavior.
What to watch out for
Leverage indirect questions
To limit social desirability bias avoid asking personally confronting questions about For example, instead of “what do you do in case X happens”, ask “what people usually do when X happens?” This question is less personal and reduces the propensity of the respondent to be subject to social desirability bias.
Assure and emphasize anonymity
For issues potentially subject to social desirability bias, make sure you anonymize your respondents to curb the effect of social desirability bias.
Avoid loaded questions
To minimize social desirability bias you need to be neutral in your questioning. This is not always immediately obvious, though. Watch out for negative connotations e.g. in words like “impulse”, “emotional”, “purposeful” and so forth. Be more descriptive to avoid this.
Measure the effect
When confronted with social desirability bias, be sure to quantify the effect. Only when measuring the social desirability bias in your context can you control for it and make your data useful. Popular methods include the Balanced Inventory of Desirable Responding, the Marlowe-Crowne Social Desirability Scale, and Palhus Deception Scales.