Reading the 2026 Polls: Fake Data Inserted into Campaign Coverage Reveals a Weakness of Our Political System
Editor’s note: Michael W. Traugott is professor emeritus of communication and media and of political science at the University of Michigan and is currently completing his term as interim director of the Center for Political Studies at the Institute for Social Research. A past president of the American Association for Public Opinion Research, he is co-author of The Voter’s Guide to Election Polls, out last year in its sixth edition.
A recent incident of the insertion of fake data into the coverage of the Wisconsin Senate primary, the Nevada gubernatorial race, and the Los Angeles Mayoral primary highlights a weakness in coverage of our campaigns and the outsized role of polls.
The data were supplied by an entity called Meridian Strategies, which has now disappeared from the internet after indicating that its sourcing of fake data was part of a “short-term social experiment examining how purported polling information could enter and spread through the political information ecosystem without independent verification.” Further sleuthing by The Guardian identified a single individual behind the scheme. The news media were clearly wrong to describe the source as a polling organization, a company, or a pollster.
This is not the first time that fake data have appeared in a political campaign. In the 2000 presidential campaign, the Daily Kos website featured poll results from a firm by the name of Research 2000 that were completely made up. In this case, the people at Research 2000 prepared a complete press release for each “poll” that included topline results, a methodology section, and a sample questionnaire. The problem was that no interviews were conducted, and the results were fabricated.
Why do some news organizations fall for fake polls?
Communication researchers who study campaign coverage are well aware of the tendency for news media to engage in horserace coverage where the emphasis is who’s ahead or behind and how the dynamics of the race are changing over time. Polls that measure candidate preferences play right into this “game dynamic” focus; they place a premium on reporting the latest results and how things have changed since the previous measurement, especially as Election Day approaches.
Polls can contribute to an understanding of the forces at work in the electorate that explain support for candidates based on their issue positions or the success of their appeals to specific groups. But too often the media only report the results of the “trial heat” question such as “”If the presidential election were held today, would you vote for Candidate A or Candidate B?”
There is also a difference in the in-house resources a reporter might have in different types of news organizations. Large circulation newspapers with a national focus often have their own polling expert, but local news outlets don’t have such access. Print and web outlets are more likely to have in-house expertise than local television stations.
How to Evaluate New Poll Results
There are some guidelines that journalists and the public can consult and organizations they can ask for guidance. The American Association for Public Opinion Research (AAPOR) has several resources on responsible reporting on polls, as does the Poynter Institute. Journalists can also turn to local universities where there are public opinion and survey methodology experts for consultation and advice on the validity of new poll results.
The key elements are:
- a transparent description of the poll’s methodology
- who sponsored the research
- the population studied
- the sample design and size
- whether the data were weighted and how
- the mode of interviewing
- the dates of the field work, and
- a copy of the questionnaire.
For news readers: Check who conducted a poll, who paid for it, whom they interviewed, when, and how — reputable pollsters disclose all of this. Then remember that every poll carries a margin of error that is wider for subgroups like “young voters” or “independent voters,” than for the full sample of all voters, and that a poll is a snapshot rather than a forecast. Similar trends that appear across many, good polls are more trustworthy than any single, surprising number.
Additional information can be found at the site of the AAPOR Transparency Initiative.
This post was written by Michael Traugott of the Center for Political Studies and prepared for publication by Center for Political Studies communications manager Tevah Platt.
