AGA Study Finds Prediction-Market Users Overestimate Their Maths Skills

Research commissioned by the American Gaming Association says gamblers score higher on financial literacy than non-gamblers, and that prediction-market users are nearly as confident as sports bettors but test like non-gamblers. The trade body is using it against the exchanges.

• Gamblers scored an average of 3.93 out of five on a financial literacy index against 3.72 for non-gamblers, and 41% were highly financially literate against 27%, according to research commissioned by the American Gaming Association (AGA)
• Prediction-market users were nearly as confident in their maths as sports bettors, but performed more like non-gamblers when tested, the AGA says
• Casino players scored highest on the Positive Play Scale, a measure of responsible gambling behaviour, and financial literacy was linked to higher scores
• The survey covered 3,149 US adults recruited online through Prolific from 15 to 21 July, and was conducted by two University of North Carolina researchers
• The AGA says the results show the danger of prediction markets presenting sports wagers as investing against peers

The Casino Lobby's Case Against the Exchanges, in Numbers

The American Gaming Association released research on Friday 25 September that it says challenges "conventional wisdom about gamblers". According to the AGA's release, gamblers scored an average of 3.93 out of five on a widely used financial literacy index, against 3.72 for non-gamblers, a statistically significant difference, and 41% of gamblers qualified as highly financially literate against 27% of non-gamblers.

The second finding is the one aimed at the exchanges. Sports bettors showed the greatest mathematical confidence and the highest objective ability, the AGA says, while prediction-market users expressed nearly as much confidence but performed more in line with non-gamblers when their skills were tested. "The research highlights the risks of 'prediction markets' marketing sports betting as investing against 'peers' and the dangers of misleading consumers into thinking they just have to be smart to win," said David Forman, the AGA's vice president of research.

The study also reports that participants with moderate or high financial literacy scored significantly higher on the Positive Play Scale, a measure of responsible gambling behaviour, and that casino players scored higher than every other segment, which the AGA attributes to "longstanding exposure to responsible gaming resources within mature, regulated environments".

The researchers surveyed 3,201 US adults online through Prolific from 15 to 21 July 2026. After screening, the sample was 3,149: 537 sports bettors, 574 prediction-market users, 557 casino players, 495 iGaming users and 986 non-gamblers. The AGA commissioned the work; it was conducted by Colin López of IN Research & Analytics and Jackson Sears of Matrix Consulting Group, both co-founders of the Betting, Experience, and Trading in Sports Research Center at the University of North Carolina at Chapel Hill. The full report and questionnaire were not published with the release.

The Method Supports a Contrast, Not a Population Estimate

The AGA quotes a margin of error of plus or minus two points, but Prolific is an opt-in online panel, and a margin of error assumes a random sample. The groups add up exactly to the total, which suggests participants were sorted into single categories, although many Americans bet on sport and play casino games. The comparisons between segments are the useful part: within the same panel and the same test, prediction-market users' confidence outran their scores by more than sports bettors'. That is a plausible and interesting finding. It is not evidence about how literate the typical American gambler is, and the release's headline claim that gamblers are more financially literate than non-gamblers should be read with that limitation.

The Finding Fits What the Exchanges' Own Data Suggests

The overconfidence result sits comfortably with evidence from outside the AGA. Kalshi's own trading data showed long-shot parlays making up 38% of one Saturday's volume, and takers lost money on its record NFL weekend. Those are the patterns an overconfident retail user would produce. The AGA's interest is plain: it represents the licensed operators that the exchanges compete with, and it wants the exchanges regulated as gambling. That does not make the finding wrong, but it does mean the claim that matters most, that exchange marketing misleads users about their edge, needs independent research before it carries weight with regulators.

An industry body has produced numbers for an argument it was already making. The comparison between sports bettors and prediction-market users is the part worth testing further; the rest is advocacy.