What respondent history reveals about data quality

What 10x more respondent history is showing us 

Data Quality Co-op's database grew 10x in the first half of 2026, significantly expanding the industry's shared view of respondent quality. Clients are now more likely than not to encounter respondents for whom DQC already has historical quality signals.

That history is giving us a much fuller picture of who is participating in research. We can identify respondents with a record of consistently poor survey behavior, see suspicious patterns that may not show up in a single real-time fraud check and recognize people who consistently contribute good data.

Recent DQC data puts some numbers around each of those groups.

The 6% you can't see

More than 6% of survey starts across the Data Quality Co-op (DQC) network come from respondents with a documented history of consistently poor survey behavior or, as we call them, "Keyboard Mashers."

Many of these respondents would pass traditional pre-survey screening. They are not necessarily flagged by device-based fraud tools because the issue is their behavior, not their technology. Their patterns only become visible when participation history is connected across suppliers, studies and organizations.

That is where shared quality intelligence matters. As the DQC network grows, we gain more visibility into respondent behavior over time, making it possible to identify known poor-quality participants before they enter a study and influence results, across the entire research ecosystem.

When suspicious respondents look good enough to pass

Over the past 30 days, 14% of respondents seen by DQC have been classified as Incognito Operators, our term for respondents who consistently show suspicious behavior outside of surveys. The name refers to their pattern of behavior, not simply to respondents using a browser's incognito mode.

The device history of these respondents helps illustrate a limitation of real-time fraud detection. Among Incognito Operators, 29% show device-score swings of 40 points or more across sessions. Their scores can move above and below the threshold used to determine whether they are allowed into a study. A respondent who gets blocked in one session may look good enough to pass in another, even though their history shows a much more suspicious pattern.

Moving the threshold could catch more of these respondents, but it could also increase false positives among legitimate participants. DQC provides another signal: what has been observed about that respondent over time. By connecting device scores across sessions, DQC can identify patterns that a single real-time check may miss and help researchers make better-informed decisions about who enters their studies.

Seeing the industry's most valuable participants

Insights professionals invest heavily in identifying bad respondents. While it makes sense to mitigate fraud risk, it is only half of the equation. Data Quality Co-op also identifies respondents who consistently contribute good data over time.

Gold Standard respondents, who perform consistently well across DQC client surveys, are the most valuable resource in insights. These respondents consistently contribute high-quality data, driving high-value insights and improving industry unit economics (it's cheaper to retain than recruit!). They help sample providers build stronger panels while giving brands and agencies another source of evidence about the quality of research.

Among respondents with an established history, they represent 22.0% of known respondents and account for 29.5% of survey starts.

Until now, the industry has had no reliable way to identify these respondents across suppliers and studies. By bringing together participation history from across the industry with DQC, researchers can recognize respondents with an established record of reliable participation.

A broader view of who is entering research

These numbers also show how much variation can sit behind a respondent who, at the point of entry, may look like anyone else. Some have a documented record of poor survey behavior. Some move above and below fraud thresholds depending on the session. Others have repeatedly demonstrated that they contribute good data.

As DQC sees more respondents across more studies, that history becomes available before the next survey begins. Researchers can use it alongside the checks they already have to know more about the people entering their studies and make decisions with more than a single session to go on.

Learn how DQC can add respondent history to your data quality strategy.