Dealing with social desirability bias Flashcards
main paper take aways:
Nederhof 1995=
* Social desirability affects the validity of experimental and survey research findings
* Can be regarded as the result of 2 separate factors:
1. Self-deception
2. Other deception
* Modes of coping with social desirability bias
1. Comprises 2 methods- The use of social desirability scales and the rating of item desirability (focus on detection and measurement of social desirability bias)
2. Comprises 7 methods; use of forced choice items, the randomised response technique, the bogus pipeline, self administration of the questionnaire, the selection of interviewers, and the use of proxy subjects. (focus on methods to reduce or prevent effects of social desirability bias)
What is social desirability bias: Paulhus, 1980, social desirability is regarded as a distortion of responses in a socially desirably direction which is a resultant of two factors: self deception and other deception.
Self deception- the respondent actually believes a statement to be true of him or herself, even though it is inaccurate (Millham and Kellogg, 1980)
Other deception- purposely misrepresent the truth as a form of impression management motivated by a desire to avoid evaluation (Milham and Kellogg 1980)
using social desirability scales:
MC SCAKE- The scale is reliable, showing high test-retest and internal consistency correlations, but its validity has been questioned. Critics argue that it measures a broader need for social approval rather than social desirability alone.
Researchers identified two components within the MC scale: self-deception (unconsciously believing in one’s socially desirable responses) and other-deception (consciously aiming to impress others
coping with social desirability
Nederhof 1995=
Rating Item Desirability:
Participants rate the desirability of items (e.g., happiness) on a scale, but this method is time-consuming for large item sets and prone to inconsistency when judged by others. Clancy (1972) found item desirability unrelated to MC scale scores, and this approach is considered basic.
Forced-Choice Items:
Forced-choice items reduce social desirability bias by requiring respondents to choose between equally desirable options, making it harder to give overly favorable responses. This method, introduced by Humm and Wadsworth (1939), resembles neutral questioning techniques.
Randomized Response Technique (RRT):
RRT uses randomization (e.g., coin flips) to allow respondents to answer sensitive questions anonymously, reducing social desirability bias. Bradburn and Sudman (1979) demonstrated that this technique encourages honesty by ensuring individual responses cannot be traced.
Self-Administered Questionnaires:
Self-administered questionnaires, as outlined by Bradburn and Sudman (1979), reduce social desirability bias by offering privacy, anonymity, and reduced social pressure, often through digital formats that further enhance respondent honesty.
The Bogus Pipeline:
The bogus pipeline, developed by Jones and Sigall (1971), uses a fake detection device to convince participants their true responses can be measured, prompting honesty on sensitive topics. While effective, it’s controversial due to its deceptive nature.
Selecting Interviewers:
Carefully selecting interviewers (Scott 1968) can reduce social desirability bias by matching participant demographics, training interviewers in neutrality, and building rapport, making respondents feel less judged and more comfortable.
Proxy Subjects:
Using proxy subjects (Sudman and Bradburn 1974) involves asking participants about the behaviors or attitudes of others rather than themselves, reducing personal exposure and pressure to conform, especially for sensitive topics.
other ways to combat social desirability bias
Ensure Anonymity and Confidentiality
When participants feel their responses cannot be traced back to them, they are more likely to provide honest answers. Researchers should reassure participants that their data will remain confidential (Fisher, 1993)
More biases-
Design Biases
Order Effects=This occurs when the sequence in which questions, tasks, or conditions are presented influences participants’ responses. For example:Participants might perform differently depending on whether they encounter an easy or difficult task first.Earlier questions can influence answers to later ones due to priming or fatigue.
Ceiling/Floor Effect of Task=These effects occur when a task is too easy or too difficult, limiting the ability to differentiate between participants’ abilities:
Ceiling Effect: Everyone scores near the top of the scale because the task is too easy.
Floor Effect: Everyone scores near the bottom because the task is too hard.
Participants: Sampling, Selection, and Individual Differences
Sampling Bias: Occurs when the sample isn’t representative of the population being studied (e.g., over-reliance on one demographic group).
Selection Bias: Arises when participants are non-randomly chosen, which can skew results.
Individual Differences: Variability in participants’ characteristics (e.g., age, education, or prior knowledge) may influence outcomes and introduce bias.
Implementation Biases
Researcher Bias
This occurs when a researcher unintentionally (or intentionally) influences the study’s results due to expectations, personal beliefs, or behavior. Examples include: Subtly leading participants to a desired answer.
Interpreting ambiguous results in line with their hypothesis.
Response Bias-Happens when participants alter their responses to align with what they think is expected or socially acceptable rather than being truthful. Types include:
Social Desirability Bias: Participants give responses they think are more favorable.
Acquiescence Bias: A tendency to agree with statements regardless of content.
2 examples of social desirability bias
Over-reporting of self-reported charitable giving (Lee & Sargent, 2011)
Students’ under-reporting of alcohol consumption and risky drinking (Davis et al., 2010)
