Is Respondent Fraud a Crime? An ongoing debate.

April 29, 2026

Jerry Arbittier

I was on an Insights Association committee call recently when someone said something that stopped me mid-thought:

“A respondent who provides false information to qualify for a survey is committing fraud. That is a crime.”

And I just sat with that for a second.

Because honestly? We don’t act like we believe that.

What exactly is synthetic dataWe’ve accepted fraud as a cost of doing business.

When a vendor fabricates respondents wholesale, we call it a crime. Clear line. No debate.

But when an individual lies about their job title to squeeze through a screener, collect an incentive, and corrupt your dataset? We call it a “data quality challenge.” We build filters for it. We budget for it. We move on.

The larger question is, is that right? And that is up for debate.

I made a recent Linkedin Post about it, and there were a lot of folks who thought this would reduce the amount of survey takers in market research, and add to the already diminishing numbers.

While another perspective which was interesting is the legal one. How would we legally prove that this was fraud, and not just a simple overlook, or a desperate measure.

Here’s what’s actually happening in that transaction.

A research buyer pays real money to reach a specific, qualified audience. A respondent lies to access the study. They collect an incentive they didn’t earn. And somewhere downstream, a product gets launched, a policy gets written, or a marketing budget gets allocated, on data that was quietly poisoned at the source.

The harder a respondent is to find, the higher the risk.

Traditional panels put real effort into fraud prevention. Digital fingerprinting. IP checks. Attention traps. Red herring questions. The industry has invested heavily in this, and it’s not nothing.

But the core mechanic hasn’t changed. Someone clicks their job title. They get paid. The screening questions are still multiple choice , and multiple choice is still gameable, no matter how many layers sit around it.

Now think about what happens when you’re recruiting a hard-to-reach audience. Senior B2B decision makers. Niche healthcare professionals. C-suite executives. The incentive to fake it goes up, but the verification stays exactly the same. That gap; between how hard the industry is working and how determined bad actors are is exactly where the problem lives.

So back to the question on that committee call.

Should organizations like the Insights Association push to formally define this behavior as a crime not just an ethical violation?

Imagine putting this in a survey invite: “Providing false information to qualify for this study constitutes fraud and may be prosecuted.”

It would reframe the act, for the casual cheater, for the panel company looking the other way, and for our industry’s own understanding of what’s actually at stake. But obviously it could cause legal complications.

What would researchers want?

I just got back from the IA Data Quality conference, and if I had to distill what researchers are really asking for, it’s this: real identity verification. Not better filters. Not smarter screeners. Actual confirmation that the person taking the survey is who they say they are.

Open-ended questions are a useful signal. When someone has to explain their thinking in their own words rather than click a radio button, the cracks show fast. Short answers. No nuance. No specificity. AI can detect it, and so can a human paying attention.
Having a video on, goes further. On video, you know within seconds whether someone is who they say they are, whether they’re reading off a screen, whether they’ve actually lived the experience they claimed in the screener. Unscripted follow-up questions, the ones participants don’t see coming, are nearly impossible to fake your way through. If a research provider can’t show you respondents on video giving open-ended answers, you’re flying blind.

But both of these are detection tools. They catch fraud after it’s already in the room.

What researchers are pushing for is validated verification before the door opens, a confirmed identity tied to a real person, a real credential, a real LinkedIn profile, a real government ID. Something that makes misrepresentation structurally harder, not just easier to spot.

The technology exists. The will to build it into standard practice hasn’t caught up yet. That will be a real game changer, is what we think about at AOPS.