What respondents don’t tell surveys
What happens to people who answer surveys every day and get screened out halfway through?
We talked with 333 French panellists by voice about the surveys they answer.
Including this one.
Get the full report31 pages, in English
333voice conversations
13 minmedian length
10 daysof fieldwork
Screened out mid-survey, respondents learn to get past the filter.
93%
have been screened out mid-survey at least once, 166 out of 178. Among those who answer surveys every day, 94%; among weekly respondents, 42 out of 42.
When you spend half an hour on a survey and earn 10 cents, that’s not right.
Woman, 65+, answers surveys several times a week
27%
adjust their first answers to stay in, 45 out of 165, including 17 who do it “often”. Because screening comes too late, some fit their answers to the filter rather than to their actual situation.
[…] one time I say I’m a partner, another time I say I’m a deputy […] I put “senior executive”, and it worked.
Woman, 45–54, answers surveys every day
55%
skim or tick boxes at random when a survey drags on, 49 out of 89, while 73% rate their own attention 8 out of 10 or higher. And 61% name no topic on which they would be inaccurate.
It doesn’t make you want to be honest all the time. I mean, you want to answer what’s expected of you, and not the truth.
Woman, 55–64, answers surveys every day
13%
would be most candid with a human interviewer, 23 out of 174, compared with 45% for an online questionnaire and 42% for an interview with an AI. With a human, they fear being judged.
I felt I could say absolutely anything, even things […] private or a bit unusual. […] I didn’t feel you could judge me.
Man, 25–34, Île-de-France
333 conversations, one bubble each.
Click a red bubble to hear or read what that person said.Tap a red bubble to hear or read what that person said.
What the panel learned, and what it does to your data
The full report goes further. 31 pages, in English:
- 01Findings on screen-outs, honesty, trust in polls and interview formats
- 02Three panellists’ stories, in their own words, transcribed verbatim
- 03Three opportunities for research agencies, client-side insights teams, panels and platforms
- 04The set-up: PureSpectrum and nabel, the conversation guide and the methodology
Methodology · Mainland France · 1–20 Sept 2026 · Two waves
333 respondents recruited through PureSpectrum in France, all interviewed by voice, in two waves. 218 complete voice conversations with the nabel agent, plus 115 partial ones whose completed answers were kept.
Our latest studies
TrackerConsumer goodsBig brands, too expensive? Not mine→September 2026 · 259 conversations
TrackerRetail and e-commerceAudio excerpt, 0:33Wary of discounts, sold on Amazon→September 2026 · 246 conversations
TrackerHospitality and travelAudio excerpt, 0:29Going away anyway, on a smaller budget→September 2026 · 263 conversations
TrackerFinancial servicesAudio excerpt, 0:32Bank fees: plenty of complaints, few goodbyes→September 2026 · 274 conversations
TrackerTransport and automotiveAudio excerpt, 0:28To the garage, not the dealership→September 2026 · 258 conversations
TrackerTechnology, Media & TelecomAudio excerpt, 0:28Telecom loyalty lasts until the next price rise→September 2026 · 305 conversationsGet the full report31 pages
