Panel conditioning
Panel conditioning is the change in panel members' answers or behavior that comes from taking part in surveys repeatedly, rather than from any real change in their views. It can bias results from long-running panels and tracking studies.
How it works
Repeated participation can change people in several ways. They learn which answers lead to shorter surveys and give those. They become more informed about a topic because they have been asked about it. They start to think about a brand or issue more than they otherwise would, or they rush through familiar questions. Any of these makes panelists less like the population they represent.
Pew Research Center tested for it in its own panel and found that conditioning did not contribute significant error to its estimates; the risk grows where members answer the same questions often. Researchers limit conditioning by capping how often members are surveyed, rotating members out after a period, varying question order and comparing results from long-standing members with newer recruits.
A worked example
A car brand's tracking panel asks members about its advertising every month. After a year, long-standing members report 30 percent higher ad awareness than new recruits with the same profile. The brand caps each member at 4 waves a year and reports trends from fresh members.
Related
Automate this in Pulse
Conditioning comes from answering the same questionnaire repeatedly, so recruiting first-time participants to the brief's profile for each Pulse wave keeps practiced answers out of the data.