Measurement bias is systematic error introduced when a measurement process produces responses that differ from the true value or construct of interest. In self-report research, small changes to question wording, response scales, or scale descriptors can substantially change response patterns. section 6.2.1
Watch for measurement bias when:
- A survey measures perceptions rather than the phenomenon itself, creating indirect measurement. section 6.2.1
- Researchers compare responses collected with different wording, scales, or descriptors. section 6.2.1
Self-report versus behavioral data
An employee satisfaction survey and user feedback survey capture self-reported perceptions, while interaction logs, version-control history, and helpdesk tickets capture behavioral data. Comparing these sources can clarify what people report versus what they do.
