Fake polls expose vulnerabilities in US elections ahead of midterms

Fabricated surveys are adding to tensions in the country’s increasingly polarised political environment.

Experts warn that fake pollsters could weaken trust in election data [File: Stephen Maturen/Getty Images via AFP]

Fabricated political polling has forced its way into the spotlight during the midterm election period in the United States.

Revelations that a small number of fraudulent firms have released fictitious results – some of which were picked up by news outlets and polling trackers – have raised concerns about how easily false information can acquire credibility.

list 1 of 3 What polls say as Trump reverses course on cancelled campaign rallies

list 2 of 3 Trump reverses course, reinstates cancelled campaign rallies

list 3 of 3 Trump administration renews bid to block Alaska election results

Those datasets can be used to cast doubt on the electoral process. And when bad actors are unmasked, their existence can be leveraged to undermine voter confidence in election information overall.

The latest example occurred earlier this month, when Wedgewood Polls released a statement on social media admitting its statistics about a bevy of battleground states – like Texas, Ohio, Michigan, Kansas and Maine – were fake.

Wedgewood described its work as “an experimental project” designed to highlight the “vulnerabilities, lack of oversight and overall low barrier to entry” in the US polling industry.

“Having previously worked in the polling and research sector, we watched as the public’s trust in polling eroded alongside a rise in unverified, low-quality methodology,” the firm said in a statement posted to its social media account, which has now been deleted.

It justified its false data as a test to see how easy it would be for fabricated polling to gain traction among voters and the media.

“The short answer: Far too easy,” Wedgewood said.

But Wedgewood’s admission came after its surveys had already circulated through the political information ecosystem, reaching prominent media outlets like The New York Times.

The incident was the second such scandal in the 2026 election cycle, following the exposure of another operation, Median Strategies, which acknowledged in August it had also faked results as a “short-term social experiment”.

In an increasingly polarised US political environment, the incidents have roiled a polling industry – and a public – struggling to adapt to an ever-changing information landscape.

How fake polls gained credibility

Wedgewood, which launched in June, published surveys covering competitive races across several states. Its results attracted attention from political campaigns, commentators and organisations that compile polling data.

The New York Times said its polling tracker had listed 15 Wedgewood surveys between June and October. The newspaper subsequently removed them.

In the wake of the scandal, it explained that the firm had answered questions about its personnel, methodology and funding before its results were included.

While the Times said Wedgewood’s surveys were not incorporated into the newspaper’s own presidential approval or national congressional polling averages, the case revealed a vulnerability in the process by which polling information becomes established as credible.

Polling aggregators help journalists and voters make sense of a crowded field of surveys by collecting results from multiple research organisations. Their value depends partly on their ability to distinguish reliable research from poorly conducted or fraudulent work.

Wedgewood’s success in passing some of those checks suggests that transparency on paper is not necessarily proof that research has been conducted as claimed.

According to FiftyPlusOne, a polling and forecasting website that investigated the firm, inconsistencies in the raw data and the technical identifiers associated with survey responses helped raise suspicions about its authenticity.

The investigation prompted Wedgewood to acknowledge that most of its results were fabricated.

Though Wedgewood has denied trying to mislead the public, some pollsters are not convinced, suggesting the group could was up to something more nefarious.

“We were investigating Wedgewood for faking their data,” said data journalist G Elliott Morris, who was involved in FiftyPlusOne’s probe. “They are trying to cover up for their fraud with the ‘we were doing a social experiment’ excuse.”

How fake polling has impacted elections

The consequences of fake polling extend far beyond the polling industry.

Mallory McMorrow, who ran unsuccessfully in Michigan’s Democratic Senate primary, has argued that a Wedgewood survey showing her support in the single digits damaged her campaign.

McMorrow said the poll left her volunteers demoralised, and journalists repeatedly contacted her to ask whether she should abandon the race.

Donations, she added, also evaporated, with Democratic supporters looking to other campaigns to invest.

Michigan State Senator Mallory McMorrow, a former candidate for the US Senate, stands with supporters in Detroit on April 19 [Jose Juarez/AP Photo]

In a social media post , McMorrow described the poll from Wedgewood – and another from the firm Mitchell Research and Communications – as a “one-two punch” that “killed her campaign”. She suspended her Senate bid in July.

“By that point, we took a risk and we spend our money. We put our ads up. They were working. But the damage was done. And I was drowned out by outside spending and the narrative that we had completely collapsed,” she wrote.

Experts say McMorrow’s account illustrates a broader problem: Even when a poll does not accurately measure public opinion, its findings can still influence the decisions of people who believe it does.

The Wedgewood controversy demonstrates how fraudulent data can become ammunition in an already polarised political environment.

Researchers at the University of Washington’s Center for an Informed Public tracked how discussion of Wedgewood’s admission spread across social media.

Initially, much of the conversation was concentrated among polling experts and people who followed election data. But the story subsequently gained traction among influential political commentators and partisan accounts.

Some posts focused on the fact that prominent outlets and polling aggregators had listed Wedgewood’s results. Others used the revelation to question polling more broadly, particularly surveys suggesting that Democrats were competitive in closely contested races.

The researchers found the controversy became intertwined with claims that polling was systematically misleading voters or understating support for particular political parties.

“Within 24 hours, the right-wing media ecosystem had succeeded in drawing attention to the story and setting some of the most salient frames: that poll numbers around the midterms, especially poll numbers favoring Democrats, cannot be trusted,” the Center for an Informed Public wrote in its analysis.

President Donald Trump, who has regularly made false claims about US election integrity, seized on the revelations to criticise the Times for including Wedgewood’s surveys in its polling tracker.

“The New York Times got caught,” Trump wrote in a Truth Social post . “Everything about The Failing New York Times is ‘ARTIFICIAL’.”

He also suggested the Wedgewood case would help his $15bn defamation suit against the Times.

“Our powerful Lawsuit against them is getting stronger and stronger everyday,” he added.

No laws exist in the US that criminalise fake polls outright.

Some advocacy groups and law scholars have argued that the publication of fake polls constitute commodities and wire fraud, warranting federal prosecution.

“Deterring this sort of fraud before it can happen – and raising awareness of it – is vital,” Tyler Yeargain wrote for the Missouri Law Review in 2020.

He argued that enforcement of election-related laws against fake pollsters has “not been fully utilized”.

“Prosecution would serve a strong federal interest,” he wrote. “The interest at stake is weightier than concerns over how fake polls can shade the accuracy of an academic endeavor in prediction markets. It is instead to deter other wrongdoers, the sort who would weaponize fake polls to manipulate elections.”

Brendan Carr, the head of the Federal Communication Commission (FCC), which regulates broadcasts in the US, said there has been “a significant surge” in fake polls during the 2026 midterms, urging networks not to amplify them. Otherwise, Carr said, they could lose their licence to operate.

President Donald Trump listens as Federal Communications Commission Chairman Brendan Carr speaks at a campaign rally in Wasilla, Alaska [File: Lindsey Wasson/AP Photo]

“Broadcasters are prohibited by the terms of their federal licenses from engaging in News Distortion or promoting Broadcast Hoaxes,” Carr wrote on social media. “It is important that broadcasters comply with their public interest obligations.”

But the FCC’s sole Democratic commissioner, Anna Gomez, disagreed with Carr’s assessment.

“Pollsters who intentionally mislead the public should answer for it, but broadcasters who report publicly available polling data should never be punished when a poll is wrong,” she wrote on X. “FCC rules on news distortion or hoaxes do not apply to a station that reports this data in good faith.”

Without a clear legal remedy against fake polls, experts argue the onus is placed on media outlets and voters themselves to evaluate the quality of the election data they see.

Mark Baldassare, the survey director of the Public Policy Institute of California (PPIC), said there are ways to quickly spot which surveys are real – and which are not.

“Look for transparency,” Baldassare told PPIC’s Stephanie Barton in an interview. “A credible survey will list its mission, the purpose of its polling, and its sponsors or funders.”

“A poll guided by partisan groups or campaigns that is publicly released may have a message to relay to the public.”

捏造的调查结果加剧了该国日益两极分化的政治环境的紧张局势。

专家警告：虚假民调机构可能会削弱人们对选举数据的信任。[图片：Stephen Maturen/Getty Images via AFP]

在美国中期选举期间，捏造的政治民调成为了关注焦点。

少数欺诈公司发布虚假结果的曝光——其中一些结果被新闻媒体和民调追踪机构报道——引发了人们对虚假信息如何轻易获得可信度的担忧。

列表 1/3：民调显示，特朗普改变取消竞选集会的决定

列表 2/3：特朗普改变策略，恢复取消的竞选集会

特朗普政府再次试图阻止阿拉斯加州选举结果（共3项，此为第3项）

这些数据集可能被用来质疑选举过程。一旦不法分子被揭露，他们的存在就可能被用来削弱选民对选举信息的整体信任。

最近的一个例子发生在本月初，当时韦奇伍德民调公司在社交媒体上发表声明，承认其关于德克萨斯州、俄亥俄州、密歇根州、堪萨斯州和缅因州等一系列摇摆州的统计数据是虚假的。

韦奇伍德将其工作描述为“一个实验性项目”，旨在突出美国民意调查行业的“脆弱性、缺乏监管和总体准入门槛低”的问题。

“我们之前在民意调查和研究领域工作，目睹了公众对民意调查的信任度随着未经证实、低质量方法的增加而下降，”该公司在其社交媒体账户上发布的一份声明中表示，该账户现已被删除。

它辩称，其虚假数据是为了测试捏造的民意调查有多容易在选民和媒体中获得支持。

韦奇伍德说：“简而言之：太容易了。”

但韦奇伍德的承认是在其调查结果已经在政治信息生态系统中传播开来，并被《纽约时报》等知名媒体报道之后做出的。

这是 2026 年选举周期中发生的第二起此类丑闻。此前，另一家名为 Median Strategies 的公司也曾被曝光，该公司在 8 月份承认，他们也曾伪造选举结果，并将其作为一项“短期社会实验”。

在美国日益两极分化的政治环境下，这些事件搅乱了民意调查行业和公众，他们都在努力适应不断变化的信息环境。

虚假民调如何获得可信度

韦奇伍德于6月成立，发布了涵盖多个州竞争激烈选区的民调报告。其结果引起了政治竞选活动、评论员和民调数据汇编机构的关注。

《纽约时报》称，其民调追踪系统在6月至10月期间列出了15项韦奇伍德民调。该报随后删除了这些民调。

丑闻曝光后，该公司解释说，在公布业绩之前，该公司已经回答了有关其人员、方法和资金的问题。

虽然《泰晤士报》表示韦奇伍德的调查并未纳入该报自己的总统支持率或全国国会民调平均值，但此案揭示了民调信息被确立为可信度的过程存在漏洞。

民调聚合器通过收集多家研究机构的结果，帮助记者和选民理解大量调查数据。它们的价值部分取决于它们区分可靠研究与粗制滥造或存在欺诈行为的研究的能力。

韦奇伍德公司成功通过了其中一些审核，这表明纸面上的透明度并不一定能证明研究已经按照声称的方式进行。

据民意调查和预测网站 FiftyPlusOne 对该公司进行调查后发现，原始数据和与调查回复相关的技术标识符存在不一致之处，这引发了人们对调查真实性的怀疑。

调查促使韦奇伍德承认其大部分结果都是捏造的。

尽管韦奇伍德否认试图误导公众，但一些民意调查专家并不相信，认为该集团可能在策划更邪恶的事情。

“我们当时正在调查韦奇伍德公司伪造数据的事件，”参与FiftyPlusOne调查的数据记者G·艾略特·莫里斯说，“他们试图用‘我们当时在做社会实验’的借口来掩盖他们的欺诈行为。”

虚假民调如何影响选举

虚假民调的后果远远超出民调行业本身。

在密歇根州民主党参议员初选中落败的马洛里·麦克莫罗认为，韦奇伍德民调显示她的支持率只有个位数，这损害了她的竞选活动。

麦克莫罗表示，民调结果令她的志愿者士气低落，记者们也多次联系她，询问她是否应该退出竞选。

她补充说，捐款也随之消失，民主党的支持者们开始将资金投入到其他竞选活动中。

密歇根州参议员、前美国参议员候选人马洛里·麦克莫罗于4月19日在底特律与支持者站在一起。[Jose Juarez/AP Photo]

麦克莫罗在社交媒体上发帖称，韦奇伍德民调公司和米切尔研究与传播公司发布的另一项民调结果，如同“组合拳”，彻底“扼杀了她的竞选”。她于7月暂停了参议员竞选。

“到了那个时候，我们冒了险，花了钱，投放了广告。广告确实有效。但损失已经造成了。外部资金的投入和我们彻底崩溃的说法淹没了我的声音，”她写道。

专家表示，麦克莫罗的描述说明了更广泛的问题：即使民意调查不能准确衡量公众舆论，其结果仍然会影响那些相信民意调查结果的人的决定。

韦奇伍德事件表明，在政治两极分化的环境下，虚假数据会如何成为攻击的弹药。

华盛顿大学知情公众中心的研究人员追踪了关于韦奇伍德认罪的讨论如何在社交媒体上传播。

起初，讨论主要集中在民调专家和关注选举数据的人士之间。但随后，这一事件也引起了有影响力的政治评论员和党派人士的关注。

一些帖子关注的是，一些知名媒体和民调汇总网站都刊登了韦奇伍德的调查结果。另一些帖子则借此机会对民调本身提出了更广泛的质疑，尤其是那些显示民主党在竞争激烈的选区中具有竞争力的民调。

研究人员发现，这场争议与以下说法交织在一起：民意调查系统性地误导选民，或低估对特定政党的支持。

“在 24 小时内，右翼媒体生态系统成功地将人们的注意力吸引到这个故事上，并设定了一些最突出的框架：中期选举前后的民调数字，尤其是有利于民主党的民调数字，是不可信的，”知情公众中心在其分析中写道。

经常对美国选举公正性发表虚假言论的唐纳德·特朗普总统抓住这些爆料，批评《泰晤士报》将韦奇伍德的调查纳入其民调追踪器。

特朗普在Truth Social上发帖写道：“《纽约时报》被抓住了把柄。《纽约时报》的一切都是‘虚假的’。”

他还表示，韦奇伍德案将有助于他起诉《纽约时报》的150亿美元诽谤案。

“我们针对他们的强有力诉讼每天都在变得更加有力，”他补充道。

美国目前没有任何法律直接将虚假民调定为犯罪。

一些倡导团体和法律学者认为，发布虚假民意调查构成商品欺诈和电信欺诈，应受到联邦起诉。

“在发生此类欺诈行为之前就加以阻止，并提高人们对此类欺诈行为的认识，至关重要，”泰勒·耶尔根在 2020 年为《密苏里法律评论》撰文写道。

他认为，针对虚假民调机构的选举相关法律的执行力度“尚未得到充分利用”。

他写道：“提起诉讼符合联邦政府的重大利益。这关乎的利益远比虚假民调如何影响预测市场的学术准确性更为重要。其目的是为了震慑其他不法分子，那些企图利用虚假民调操纵选举的人。”

美国联邦通信委员会（FCC）主席布伦丹·卡尔表示，2026年中期选举期间虚假民调数量“显著激增”，他敦促各电视台不要传播这些虚假民调。卡尔警告说，否则，这些电视台可能会被吊销运营执照。

在阿拉斯加州瓦西拉举行的一场竞选集会上，美国总统唐纳德·特朗普聆听联邦通信委员会主席布伦丹·卡尔的讲话。[图片：林赛·沃森/美联社]

卡尔在社交媒体上写道：“联邦许可证条款禁止广播公司歪曲新闻或散布广播骗局。广播公司必须履行其维护公共利益的义务。”

但联邦通信委员会唯一的民主党委员安娜·戈麦斯不同意卡尔的评估。

“故意误导公众的民调机构应该为此负责，但报道公开民调数据的广播公司在民调出错时绝不应该受到惩罚，”她在X网站上写道。“联邦通信委员会关于新闻歪曲或骗局的规定不适用于出于善意报道这些数据的电视台。”

由于没有明确的法律手段来应对虚假民调，专家认为，评估他们所看到的选举数据的质量的责任就落在了媒体和选民自己身上。

加州公共政策研究所 (PPIC) 的调查主任马克·巴尔达萨雷 (Mark Ba​​ldassare) 表示，有一些方法可以快速识别哪些调查是真实的，哪些不是。

“要看透明度，”巴尔达萨雷在接受PPIC的斯蒂芬妮·巴顿采访时说。“一份可信的调查会列出其使命、民意调查的目的以及赞助商或资助者。”

“由党派团体或竞选活动主导并公开发布的民意调查可能包含向公众传达的信息。”