The market research industry has invested in solving problems. Sophisticated tools exist to detect fraud, researchers have developed increasingly complex sampling methodologies, and the push for better, faster, more representative data has never been stronger.
But while the industry has focused on optimizing platforms, processes, and technology, it has largely overlooked one of the most fundamental drivers of data quality: the respondent’s experience.
Research participants (who power each survey, panel, and AI-driven interview) are being asked to do more than ever before. And in most cases, without fair compensation.
That gap between what’s being asked and what’s being paid is respondent burden. And it’s distorting the data the industry depends on.
As a rewards and incentives provider to the market research industry, Virtual Incentives understands the connection between fair compensation and quality data.
The Quiet Rise of Respondent Burden
A decade ago, participating in a research study meant answering a set of multiple-choice questions. The format was simple, low-stakes, and relatively quick.
Today, showing up to a study looks very different. Depending on the project, respondents might need to do one or several of the following:
- Verify their identity
- Turn on a camera
- Upload receipts
- Commit to multiple sessions
- Navigate in-depth AI-powered conversations
Every one of these requirements adds friction to the respondent experience. And when participation gets harder without better compensation, people drop out or don’t start at all. Incentives can help, but only when they reflect the level of effort required.
That’s precisely where the industry has fallen short.
Not All Burden Is Created Equal
Respondent burden is no longer one-dimensional. It has three distinct dimensions. Each dimension affects willingness to participate and the level of effort required. And all three should factor into how firms determine incentives for research participants.
- Time or the length of the interview (LOI) has been the default metric for decades. For straightforward studies, it’s a reasonable starting point, but it’s no longer sufficient as the only factor to consider when pricing respondent burden.
- Cognitive load is the most underestimated dimension. Traditionally, qualitative research focused on cognitively demanding studies (in-depth interviews, open-ended exploration, probing dialogue), and it was compensated accordingly. However, AI-driven tools now impose the same cognitive demands on quantitative surveys, and the industry isn’t compensating for it.
- Intrusion is growing faster than most people in the industry realize. Driven by the need to combat fraud and the desire for richer data, researchers are asking for personal information (identity verification, video, facial coding, and, increasingly, emotion analytics). The next frontier of AI in research is emotionally intelligent questioning: systems that analyze facial expressions in real time, tying emotional data to verbal responses in the way a skilled qualitative researcher would. This development is powerful for research, but very personal for respondents.
The common thread across all three dimensions: the industry has raised the bar for what it asks of respondents without increasing compensation. The shift becomes clear when you map research types against the two factors the industry has been ignoring.
The Industry Disconnect: Who Bears the Cost?
Every development that has increased respondent burden (video, identity verification, AI probing, emotion analytics) has been driven by client benefits: richer insights, lower fraud, and deeper data.
For respondents, the same developments require more effort, more risk, and less privacy. While the burden has increased, the incentive has not changed.
This imbalance persists largely because of a broken feedback loop. In the earlier days of face-to-face, postal, and telephone research, project managers evaluated surveys before they were fielded. If a study was too burdensome or poorly designed, a quality control process flagged it before respondents ever saw it.
That process has largely disappeared from commercial research. Today, research participants are routed to studies, and performance is measured by conversion rates. Research buyers routinely field high-burden surveys, and respondents bear the cost.
How Respondent Burden Creates Hidden Bias in Your Data
This is where the problem moves from a fairness issue to a data quality issue.
Respondents willing to participate in high-burden, low-incentive surveys are not representative of the general population. They tend to be more tolerant of friction, more comfortable sharing personal data, more willing to appear on camera, and more capable of sustaining high cognitive effort. Many are experienced “professional respondents” who have learned to navigate the system as a points-maximization exercise.
That subset then dominates the data, and no amount of post-hoc weighting fully corrects it.
Research also suggests that fairly compensated respondents provide more thoughtful, considered answers, meaning incentives not only affect who participates but also the quality of their responses.
As Frank Kelly, Market Research Practice Lead at Virtual Incentives, puts it:
“Right now, we’ve disenfranchised the general public from the research process. To re-enfranchise them, we need a framework that’s fair and compensates respondents according to their contribution.“
The pool of willing participants has narrowed because participation has become increasingly unrewarding. Re-enfranchising that broader population starts with making participation feel fair. The research supports that case and points toward the framework that needs to address it.
What the Research Tells Us About Incentives (And What It Doesn’t)
The academic literature on survey incentives is substantial, largely concentrated in the public opinion and government research space, where the stakes of an unrepresentative sample are existential.
The evidence shows that increasing incentives boosts response rates, offsets higher burden, and lowers project costs; while increasing cognitive effort accelerates participant disengagement and dropout.
At approximately $6–7 per interviewer contact, the cost of chasing non-respondents adds up fast. A more generous upfront incentive often saves money in the long run.
What the research doesn’t yet tell us is how to price incentives according to burden type. There is no validated model for the premium of identity verification, camera use, or sustained AI-driven cognitive effort. These decisions are made every day largely by instinct. That’s why Virtual Incentives is developing a REWARD Framework.
A Better Way Forward: Introducing the REWARD Framework
Raising incentives across the board would be an improvement, but the better solution requires a more systematic approach to pricing respondent burden: one that accounts for all three dimensions.
Consider how survey programming already works. Survey programmers have long used complexity factors to determine the cost of designing a study. A survey with complex logic and custom scripting costs more to build than a simple linear questionnaire. That scoring system exists because complexity is real, quantifiable, and drives cost.
Respondent burden works the same way.
The REWARD Framework (Respondent Effort Willingness Assessment Research Design) is a proposed standardized approach to assessing burden across all three dimensions (time, cognitive load, and intrusion) and translating that assessment into appropriate incentive levels. It operates as a simple scoring system. Panel operators evaluate research projects against a checklist of burden factors, and the system assigns points accordingly. The cumulative score determines the appropriate incentive level for that study. It’s designed to be fast, automatable, and easy to integrate into existing project scoping workflows.
Academic research has explored points-based systems that assign weighted scores to different question types based on their burden. What’s missing is a practical, standardized application of that thinking in commercial research, and that’s exactly what the REWARD Framework is designed to provide.
The framework is currently in development. Virtual Incentives is designing a conjoint study to model how respondents value different types of research participation and what compensation they expect in return. This research will provide the empirical foundation needed to move the REWARD Framework from a concept to a validated, practical tool.
Better Data Starts With Solving Respondent Burden
The industry has invested heavily in data quality: fraud detection, sampling methodology, and research technology. But data quality starts with the people providing the data. And right now, the system isn’t treating them fairly.
The first step is recognizing that respondent burden is a complex problem and needs a structured solution. Research firms are asking for real work and value. And the compensation should reflect that.
The REWARD Framework is one proposed path toward that standard, but getting there requires an industry conversation. If you’re interested in contributing to the development of this framework or exploring how it might apply to your research practice, contact Virtual Incentives to start the conversation.