The Influence of Social Desirability on Sexual Behavior Surveys: A Review

Research indicates pervasive and often extreme misreporting in self-reported behaviors, particularly influenced by social desirability.
Social desirability leads respondents to over-report culturally desired behaviors and under-report undesirable ones.
This paper reviews the role of socially desirable responding in sexual behavior research, noting the lack of a reliable gold standard for validating self-reports in this area.
Indirect evidence suggests significant under-reporting of sensitive sexual behaviors despite the anonymity of surveys.
There is a recommendation that sex researchers include a measure of social desirability when analyzing results from sexual behavior surveys.

Self-reports have been the primary method for assessing sexual behaviors from Kinsey's study in 1948 onwards, influencing public policy regarding reproductive and sexual health.
Significant weaknesses in self-report surveys include:
- Use of convenience samples
- Volunteer bias (Bogaert, 1996; Strassberg & Lowe, 1995)
- Question structure, terminology used, and low participation rates (Catania et al., 1990a, 1990b; Fenton et al., 2001)
- Flaws due to memory bias and erroneous recollection (Archer et al., 2015; Catania et al., 1990a, 1990b)
Response bias is particularly concerning, where answers are not based merely on content, but influenced by social desirability to gain approval (Paulhus, 1991).

Social desirability has been well-established in various research areas, particularly health.
Commonly documented under-reporting includes:
- Energy intake by more than 30% (Archer et al., 2013; Lissner et al., 2007; Subar et al., 2003)
- About 10-14% of subjects are extreme under-reporters (Ferrari et al., 2002)
- Increased under-reporting correlates with increasing BMI (Braam et al., 1998).
Under-reporting is not limited to false recall, as individuals with obesity under-report high-calorie food inventories (King et al., 2016).
Statistical analysis finds significant correlations between actual under-reporting and social desirability measures (Hebert et al., 2001, 2002; Scagliusi et al., 2003, 2009; Taren et al., 1999; Tooze et al., 2004).
Historical analysis finds that about half of self-reported data could be inaccurate across disciplines (Bernard et al., 1984).

Self-reported behaviors in sexual research lack factual checks, yet there is indirect evidence showing biases influenced by social desirability.
Early studies note high correlations of desirability bias with behavioral items (Cruse, 1965; Edwards, 1970).
A study found a correlation of 0.70 indicating strong influence in sexual behaviors (Galbraith et al., 1974).
Misreporting examples:
- Same-sex sexual experiences rose to 22% under the threat of polygraph testing (Clark & Tiffit, 1966).
- Trust in researchers led to increased admissions of anal intercourse among women (Bolling, 1976; Bolling & Voeller, 1987).
Self-reports often misrepresent behaviors like condom usage (Davoli et al., 1992; Zenilman et al., 1995).
Evidence exists showing high social desirability correlating with self-reports of behaviors like always using condoms and admitting extramarital affairs (Rao et al., 2017; Zapien, 2017).

Factor analysis identifies two components:
1. Self-deception: The participant honestly believes their responses portray a positive self-impression.
2. Impression management: Deliberate misrepresentation of responses to paint a favorable image to others.
The former is less amenable to change; the latter adjusts with social pressures.
Misreporting extends to self-reported penis size, where men report an average of 6.0–6.4 inches while actual measures yield 5.36 inches (King, 2021).
When examining the correlation with desire for social approval, a study indicated men’s responses exaggerated estimations concerning penis size (King et al., 2019).

Recent studies show self-reported behaviors might be as flawed as other areas indicated by self-reports (Schmitt, 2017).
Issues such as height and BMI inaccuracies noted in the CDC’s Youth Risk Behavior Survey (YRBS) raise doubt on the reliability of self-reported sexual data (Brener et al., 2003).
For instance, records revealed that 51% of adolescents denied having had STIs, while medical data uncovered instances of multiple infections.

Effective procedures to minimize response bias include:
- Anonymity in surveys: Self-administered questionnaires yield better accuracy (Durant & Carey, 2000; Robertson et al., 2018).
- Technological adaptations allow improved reporting via computers, offering increased privacy (McCallum & Peterson, 2012).
- Potential utility of indirect questioning to mitigate bias (Fisher, 1993), though less applicable for direct sexual behavior inquiries.
Utilizing established measures of social desirability, such as the Marlowe-Crowne Scale, could standardize research practices (Crowne & Marlowe, 1960).
Data adjustments through logistic regression estimations could enhance accuracy on reported behaviors (Gibson et al., 1999).
When presenting sexual behavior statistics, median measures are recommended to reduce effects from extreme reporting (Alexander & Fisher, 2003; Conley et al., 2011; Pedersen et al., 2002).

Social desirability causes significant misreporting in sexual behavior surveys, paralleling other fields.
Ethical considerations emphasize the need for improved protocols in gathering accurate self-reports to inform public policy effectively.
Understanding demographic differences associated with social desirability bias remains critical for enhanced survey validity and accuracy.