Non-response

TL;DR

• Occurs when intended participants either refuse to take part or cannot be contacted, reducing the pool of data.
• Can produce a biased sample and threaten the accuracy and validity of research findings.
• Common mitigation strategies include increasing sample size, using multiple recruitment methods, offering incentives, following up with non-respondents, and analyzing non-response bias.

Definition

Non-response refers to the inability or unwillingness of individuals to participate in a research study or survey.

Explanation

Non-response can arise for various reasons and can substantially affect the validity and accuracy of research results. It reduces the number of observations and can cause the sample of respondents to differ systematically from the target population. When those who participate differ from those who do not, the study can produce biased estimates. Researchers use several approaches to limit or assess the impact of non-response, including recruiting more participants, diversifying recruitment channels, providing incentives, conducting follow-ups, and analyzing differences between respondents and non-respondents to detect non-response bias.

Examples

Non-response due to refusal

Some individuals simply refuse to participate because of lack of interest, time constraints, or privacy concerns. For example, a researcher conducting a study on the effects of a new medication may encounter participants who are unwilling to take the medication or provide personal information about their health. This form of non-response can lead to a biased sample if those who agree to participate are more motivated or interested in the topic than those who decline.

Non-response due to inability to reach the respondent

Researchers may be unable to contact intended participants because of incorrect or outdated contact information or because participants are unavailable. For example, a survey on consumer behavior may be sent to a list of email addresses but some emails may bounce back due to invalid addresses. Similarly, during attempts at in-person interviews, researchers may find participants are not at home or unwilling to answer the door. This form of non-response can also produce a biased sample if the reached and willing participants differ from those who are not reached.

Notes or pitfalls

• Non-response can significantly impact the accuracy and validity of research findings.
• If non-respondents differ systematically from respondents, analyses may suffer from non-response bias.
• Researchers must recognize the potential for non-response and apply appropriate strategies to minimize its effect and assess residual bias.

Related terms

• Non-response bias
• Sample
• Representative sample