Common types of response bias
- Social desirability bias: people give the answer that makes them look good, such as reporting more exercise than they do.
- Acquiescence bias: people tend to agree with statements, whatever the statement says.
- Extreme or central tendency: some people always choose the ends of a scale, and others always choose the middle.
- Order effects: the position of a question or an option changes the answer. The first options in a list are often chosen more.
- Leading questions: wording that suggests the expected answer, such as "How much did you enjoy our new design?".
Example
A survey asks "Don't you agree that our support is fast?". The wording pushes people toward Yes, so the results show more satisfaction than really exists.
How to reduce response bias
- Use neutral wording, and ask about one thing per question.
- Use balanced scales with the same number of positive and negative options.
- Mix positively and negatively worded statements in a Likert scale.
- Change the option order between respondents where the order does not matter.
- Let people answer anonymously when the topic is sensitive.
- Keep the survey short, because tired respondents answer less carefully.