A participant may give a careful, sincere answer that still does not measure what the researcher intended. They remember a recent week as typical, interpret “regularly” differently from the next person or agree with a statement because its wording makes disagreement awkward. Response bias concerns systematic distortion in how answers are produced.

It is different from sampling bias, which concerns who enters the study, and non-response bias, which concerns who answers at all. A well-recruited sample can still respond to poorly designed questions, and a large number of those responses does not remove the distortion.

Consider the pressures built into the question

Social desirability can make some answers easier to give than others. Acquiescence can favour agreement with a statement. Recall problems can affect reports of frequency or timing, while earlier questions can change how later ones are interpreted. These mechanisms should not be collapsed into a claim that participants simply tell researchers what they want to hear.

A fictional study of a household recycling service illustrates the difference. “Do you always recycle correctly?” asks for an idealised judgement without defining correctness. Asking what happened to a particular item on a recent occasion gives the participant a more concrete event to describe, although memory and self-presentation can still matter.

Pew Research Center’s explanation of question wording shows why clear, neutral phrasing requires care. The answer is shaped partly by the instrument through which it is requested.

Make accurate answers easier to provide

Ask one clear question at a time, specify a meaningful period and use language participants understand. Response options should cover plausible experiences, including uncertainty or non-applicability where relevant. Avoid adding praise, blame or the team’s preferred interpretation to the question.

Explain the handling of responses accurately. Confidentiality may reduce some concerns, but it must not be described as anonymity when answers can still be linked to a person. In an interview, a warm but neutral response helps keep an unwelcome answer from becoming a social confrontation.

Pilot the questions by asking people how they understood them and arrived at their answers. This can reveal a problem that an internal wording review misses. It is also useful to examine question order and whether the study encourages a repetitive response pattern.

Investigate disagreement between accounts and behaviour

Comparing self-report with observed or recorded behaviour can reveal an important discrepancy, but the behavioural record also needs scrutiny. The survey may concern a different period, the tracking may omit a device, or the participant may understand the activity differently from the event definition.

Use triangulation to examine those possibilities rather than automatically declaring one source truthful and the other wrong. In the recycling example, a person’s explanation may reveal confusion about a material that a simple count of bin use could never establish.

Report the question, response options and study conditions with consequential findings. Readers should be able to see what people were actually asked before a percentage is translated into a claim about their attitudes or lives.