The Ambiguity of Frequent Survey Participation: Is “Hyperactivity” a Signal of Professional Fraud?

The idea that panelists sign up, receive well-matched survey invitations, and thoughtfully complete only a handful of surveys each month is, unfortunately, a myth. In reality, a disproportionate share of interviews is generated by a small group of highly active respondents. A CASE4Quality study found that just 3 percent of devices generated 19 percent of all survey completions, with 40 percent of them completing over 100 surveys per day.

There are several reasons for this: Sample has become commoditized, prices are low, and so are participation incentives. At the same time, recruiting and retaining panelists is increasingly difficult. Respondents are frequently screened out after investing time in qualifying questions, receive minimal compensation, and are asked to complete long, repetitive, and often frustrating surveys.

To compensate for this imbalance, industry practices such as routing have become standard. Panelists who fail to qualify or have already completed a survey are redirected to others. Sample is also pooled across multiple panels to meet cost, speed, and geographic requirements.

The result is a system in which occasional participation barely pays off. For respondents motivated by financial incentives, meaningful returns require frequent and strategic participation (e.g., higher-paying B2B studies).

But does frequent participation equal fraud? Not necessarily. Frequency is a risk indicator, not a verdict. Some highly active respondents are genuinely engaged and intrinsically motivated. However, when participation is primarily driven by monetary incentives, the system creates pressure to maximize efficiency. In this sense, not all frequent respondents are fraudsters, yet professional fraud relies on sustained, high-frequency activity to be viable. This is reflected in an ESOMAR Congress case where a Venezuelan fraudster reported completing over 3,000 surveys per month.

So how can researchers draw the line between frequent and fraudulent participation?

1. Define and Enforce What is Acceptable

To start, prevent repeated access to the same survey. Combining multiple software and hardware signals enables the creation of a robust device identifier (“digital fingerprint”), which helps detect and block duplicate participation. This is particularly important because frequent respondents typically operate across multiple panels, meaning that monitoring activity within a single source is not sufficient. Ideally, participation patterns should be assessed across the broader panel ecosystem. Moreover, digital fingerprints can be used to maintain continuously updated blacklists of known fraudsters.

In addition, define realistic participation thresholds by assessing how many surveys are attempted and completed within a given time period. There is no universal cutoff for what constitutes “too frequent,” so benchmarks should be tailored to the target population. Factors such as time availability and financial situation matter, individuals with flexible schedules may legitimately complete more surveys than full-time professionals. However, beyond a certain level of activity, human limitations suggest an increasing likelihood of automation (e.g., click and phone farms or bots).

Finally, the attempt-to-completion ratio is a useful signal. Professional fraudsters are skilled at adapting their answers to qualify, so frequent survey takers who consistently pass screeners across diverse studies and topics should be treated with caution.

2. Apply a Funnel Approach, Not Isolated Checks

A useful analogy for survey quality assurance is the Swiss cheese model. Each check acts as a barrier to fraudsters, but every layer has holes. Only by combining multiple layers can these gaps be covered and fraudsters effectively filtered out.

However, this model has an important limitation. It tends to treat each check in isolation. In practice, exclusion decisions are rarely black and white. Many cases fall into grey areas where a single indicator is not sufficient. Reliable decisions emerge only when multiple signals are evaluated together.

For example, a respondent may exhibit a number of prior completions that appears suspicious (e.g., 30 in the last 24 hours). On its own, however, this might not be sufficient grounds for exclusion and requires validation through additional checks. If technologies commonly associated with fraud are also detected, the case can become clearer and the respondent may be blocked from entering the survey (e.g., 30 completions combined with open developer tools). But even then, ambiguity can remain. VPN usage, for instance, was once a strong fraud indicator, but as VPNs are now built into mainstream browsers like Firefox and widely available for free, it is no longer a reliable standalone signal.

In such cases, the appropriate response is not immediate exclusion, but progression. Allow the respondent to enter the survey and gather additional evidence through modern in-survey checks, such as AI-generated content detection, coherence checks, and input pattern analysis. By combining pre-survey signals with observed in-survey behavior, researchers can build a more complete picture and make more confident exclusion decisions.

3. Use Covert, Frictionless Checks

One of the key challenges in ensuring data quality is making sure that only qualified individuals enter a survey. Fraudulent respondents, by definition, misrepresent themselves to match required target profiles.

An approach is identity verification using facial recognition, IDs, phone, or bank validation. While useful, these methods have limitations. They are typically point-in-time checks that provide only limited information (e.g., age or gender), often insufficient for eligibility screening. Moreover, determined fraudsters can circumvent them using stolen or synthetic identities, deepfake images, virtual phone numbers, or mule bank accounts. The financial incentive to do so is significant. The Venezuelan fraudster mentioned earlier reported earning up to $2,500 per month from survey fraud, which is roughly ten times the average urban salary in Venezuela.

At the same time, increasing participation burden and requesting sensitive data could discourage legitimate respondents, especially when incentives are low and privacy concerns are high, thereby risking the loss of the right participants. A more effective approach, therefore, may be not to verify identities directly, but to evaluate whether responses align with a respondent’s panel profile by comparing anonymized sociodemographic, psychographic, and behavioral attributes with survey answers.

This approach is privacy-preserving and scalable. Most importantly, it removes friction for genuine respondents while making it significantly harder for fraudsters to succeed, as consistency across multiple dimensions is difficult to fake.

Industry Implications

Fraud detection technology has matured. The three approaches described above can be combined into an automated, scalable system that helps researchers separate high-frequency participation from fraudulent participation much more reliably than any single check. However, the industry still lacks clear accountability for applying these controls.

Panels are often assumed to ensure data quality, but frequent survey-takers are the economic backbone of their business. Expecting panels to remove them creates an inherent conflict of interest. As a result, decisions about respondent quality may be influenced, consciously or not, by commercial considerations.

Effective fraud detection requires a fundamental shift in responsibility. These methods should be implemented by independent entities whose business models do not depend on retaining highly active respondents, such as non-panel owning agencies and sample aggregators, or in-house research teams, not by panel providers themselves. Where intermediaries are involved, full transparency in how fraudulent behavior is identified and handled is essential.

Only when quality control is separated from commercial interests can inclusion and exclusion decisions be made objectively, and only then can sample buyers truly trust the integrity of their data.