Panel fatigue and the case for blended sample

A panel is a population of people who have agreed to answer surveys, which is not the same population as the one a study is usually about. The longer someone has been on a panel, the less like a first-time respondent they answer — faster, more consistently, and with a better sense of what qualifies them for the next study.

What fatigue looks like in data

• Shorter open ends with more generic content
• Higher qualification rates on screeners with obvious qualifying answers
• Lower variance on attitudinal batteries
• Faster completion times that still clear speeder thresholds

Blending helps, if it is honest

Drawing from several sources dilutes any single source's idiosyncrasies. It also introduces source effects of its own, which is why the blend has to be recorded and reported rather than treated as an implementation detail.

We report source composition with every dataset. On trackers we hold it stable between waves, because a changing blend is indistinguishable from a changing market once it is in the trend line.

The limits

Blending cannot fix a population that no source reaches. Where the audience is genuinely unreachable through panel supply, the answer is a different method — telephone, face to face, or direct outreach — not more panels.