Last updated: 7 October 2026
Quick Answer: Response bias is a systematic tendency for survey respondents to answer inaccurately, so recorded answers drift from what they actually think or do. It comes from the question, the setting or the respondent; Pew Research Center reports some people agree with statements by default. Neutral wording, forced choices and self-administered online modes reduce it.
Response bias is the gap between what people tell a survey and what is true for them. It is not lying, mostly. People agree to be polite, round their memories, avoid awkward admissions and speed up when bored, and each habit pushes answers in a predictable direction.
That predictability is what makes it dangerous. Random slips cancel out across a large sample; response bias does not, so a bigger sample measures the bias more precisely instead of removing it.
What Are the Types of Response Bias?
The main types of response bias are acquiescence, social desirability, extreme and midpoint responding, recall bias, order effects and satisficing. Each has a different cause, so each needs a different fix. The table runs from biases rooted in the question to those rooted in the respondent.
| Type | What happens | Typical fix |
|---|---|---|
| Question order | Earlier questions color later answers | Randomize or order general before specific |
| Acquiescence | People agree with statements by default | Offer a choice between two statements |
| Social desirability | People give the answer that sounds better | Self-administered modes and neutral framing |
| Recall bias | People misremember past events | Shorter recall periods and specific prompts |
| Extreme responding | People favor the end points of scales | Fully labeled scales and within-person analysis |
| Midpoint responding | People park on "neutral" to avoid thinking | A "don't know" option kept apart from the scale |
| Satisficing and speeding | People rush and pick the first acceptable answer | Shorter surveys and gentle prompts |
Acquiescence has its own guide with worked examples, covered in biased survey questions and how to fix them.
What Is an Example of Response Bias?
A clear response bias example comes from Pew Research Center's experiments on question format. Its guide to writing survey questions reports that less educated and less informed respondents are more likely to agree with agree-disagree statements, and that this acquiescence is stronger when an interviewer is present.
Pew's fix changed the results: offering a forced choice between two alternative statements produced a very different overall result from the agree-disagree version of the same values question. The underlying opinion did not change. The question format did.
How Is Response Bias Different From Sampling and Nonresponse Bias?
Response bias concerns how people answer, while sampling bias concerns who is selected and nonresponse bias concerns who declines to take part. All three are forms of survey bias, the broader idea of what is bias in research: any systematic error that pushes results away from the truth.
The difference matters for fixing them. Better recruitment reduces sampling bias, follow-ups with people who stop answering reduce nonresponse bias, and only better questions and better interviewing conditions reduce response bias.
What Is Recall Bias?
Recall bias is a systematic error caused by people remembering past events inaccurately or incompletely. The University of Oxford's Catalogue of Bias notes that it grows with longer recall periods and that undesirable habits, such as smoking or unhealthy eating, tend to be underreported.
In market research it shows up when a survey asks how many times someone bought coffee out last month. Asking about a shorter, specific window, such as the last seven days, and anchoring it to events people remember reduces the drift.
What Is Extreme Response Bias?
Extreme response bias is a tendency to choose the end points of a rating scale, such as "strongly agree" or "strongly disagree", regardless of how strongly someone actually feels. Its opposite, midpoint responding, parks answers on "neutral". Both vary by person, culture and topic, which complicates comparisons across groups and countries.
Fully labeled scales make the meaning of each point clearer and reduce guessing. When comparing groups known to use scales differently, compare each person's answers with their own average, or compare rankings of items rather than raw scores.
How Do You Reduce Response Bias in Surveys?
You reduce response bias in surveys by fixing the question first, then the setting, then the respondent's effort. Most of the gains come before fieldwork, in wording and design, rather than in cleaning data afterward.
A practical checklist:
- Use neutral wording: No loaded adjectives or assumed answers.
- Offer forced choices: Two balanced statements instead of agree-disagree.
- Keep it short: Fatigue drives speeding and straight-lining.
- Randomize order: Rotate items and options so order effects average out.
- Shorten recall windows: Ask about specific, recent periods.
- Prompt rushing respondents: In six web experiments, Frederick Conrad and colleagues found that a message to slow down reduced speeding without a noticeable rise in breakoffs.
- Pretest: Run a few cognitive interviews to see how people actually read each question.
Does the Interview Mode Change Response Bias?
Yes, the interview mode changes response bias, because a human interviewer adds social pressure that self-administered modes remove. In a Pew Research Center mode experiment, phone respondents reported higher satisfaction with family and social life than web respondents, by 18 and 14 points.
No mode removes response bias; each trades one pressure for another. Self-administered and automated modes reduce the urge to please a person, but can invite rushing. AI-moderated interviews add a different safeguard: they follow each rating with a probe for the reason, so a suspicious answer comes with an explanation that can be checked. How much latitude a system has to probe varies by tool.
Mode also decides who answers. WhatsApp-native interviews run as text inside a chat, with voice notes supported and no link or app, and AI phone interviews reach people who prefer to talk. Alchemic publishes 60+ languages including Spanish and Hindi, so questions are asked in the language people think in. Once it has the client's brief, Alchemic designs and tailors the discussion guide to handle these risks before fielding, rather than leaving that work to the buyer.