That pattern held even when polling disagreed. With at least $220 million on 2026 midterm contracts, news networks like CNN now feature these odds. Here’s what the findings mean for bettors.

What the Prediction Markets Voters See Really Measure
On a prediction market, people trade shares linked to a question, like who wins a Senate race. When the event settles, a correct share pays $1 and a wrong one pays nothing. Because of this, the price reads like a probability, so a 67-cent share signals roughly a 67% chance.
Polls vs Prediction Markets at a Glance
Pollsters ask a slice of the electorate about their voting plans, using methods refined over decades. In contrast, market prices only show what traders will pay.
| Feature | Traditional Polls | Prediction Markets |
| Source of the figure | Answers from a group of voters | Prices set by people trading shares |
| What it tracks | Voting intentions | What bettors expect to happen |
| Money behind it | No money wagered | Real money, with backers not disclosed |
| Voters who call it accurate | 37% | 19% |
| Voters who say it reflects the public | 55% | 29% |
Inside the Voter Experiment
The study used a simple experiment to isolate the effect of market odds. Every participant, surveyed in September 2026, judged five matchups between invented congressional hopefuls. Every profile listed party, age, race and career background, plus win chances from NBC News polls and Kalshi.
Researchers randomized the poll figures and the Kalshi odds in each matchup. As a result, they could see which source voters followed whenever those signals clashed.
How Much the Odds Moved Voters
Higher Kalshi odds pulled voters toward a candidate, while lower odds pushed them away. The table compares each result with a candidate holding 50% odds.
| Kalshi Odds | Change in Voters Picking That Candidate |
| 84% | +20 points |
| 67% | +10 points |
| 33% | -11 points |
| 16% | -19 points |
Notably, the odds moved opinions as strongly as polls and outweighed every candidate trait, from race to work history.
Voters Distrust Prediction Markets but Still Follow Them
The biggest finding is the gap between what voters believe and how they act. Just 19% rated election markets as reliable, yet the prediction markets voters doubt still shaped their picks. When researchers cautioned some voters about these markets, their reliance on those odds held steady.
Smaller races face the greatest risk. Few state and down-ballot races get regular polls, so a market price may be the only number voters see.
Should Prediction Markets Face Stricter Oversight?
A clear majority of voters supports more regulation.
- 48% back stronger regulation than exists now
- 23% would ban certain kinds of markets
- 16% would shut every prediction market down
- 14% oppose any regulation
Meanwhile, legal challenges keep multiplying. Minnesota tried to outlaw these platforms, but a US district judge paused the ban days before its start date. Federal oversight faces pressure too, as recent criticism of the CFTC’s role shows.
The federal ProphetX approval shows new platforms keep arriving. For now, states and the CFTC still dispute whether US gambling laws apply to these markets.
How Bettors Should Read These Odds
The prediction markets voters see on screen reflect bets other people placed, not forecasts. Before you trust those odds, run these checks:
- Look at trading volume, since thin markets move on a few big trades
- Compare the odds with recent polls from known pollsters
- Ask who gains if the price shifts before ballots are cast
- Check whether the platform answers to a financial regulator or a state gaming board
News-driven markets bring different risks, as Kalshi’s LeBron James market showed when odds jumped minutes before the public report. If you prefer traditional wagers, state-licensed sports betting sites run under a separate set of rules. Either way, set deposit limits and time-outs before you bet.
The Bottom Line
Major networks now air prediction market prices, and they sway expectations even among skeptics. Until courts and regulators settle the rules, read every on-screen price as a bet, not a poll.