A recent HEATED article features a poll released by Data for Progress that ignores just about every established best practice of polling work and instead built a survey designed to push their preferred climate litigation narrative and convince readers that the public supports this campaign.

It’s an unsurprising tactic from HEATED, which is heavily-linked with climate activists and recently acquired Climate Liability News, an activist “reporting” website that receives financial backing from the Rockefellers and Climate Communications and Law, who are pushing climate lawsuits.

While Data for Progress claims that respondents support prosecuting fossil fuel companies over climate change, a review of the poll’s question and methods clearly shows the results are incredibly flawed. The poll was released shortly after the attorneys general for Minnesota and D.C. filed lawsuits against fossil fuel producers for allegedly misleading consumers about the role the companies’ products play in causing climate change.

The obfuscation around the poll’s methodology may explain why activist-run outlets are the only ones reporting on it, because mainstream media would never give credibility to a poll conducted like this.

Faulty from the Start

From the very beginning, it’s clear the poll wasn’t intended to gather honest opinions from respondents, but rather to push a narrative. The poll’s main question reads: “Do you think fossil fuel corporations should be held responsible and pay damages for their deception and harms of climate change?”

Wording is everything in polling and that “question” is actually posed as a fact intended to establish in the minds of respondents that energy companies have caused “deception and harm.”

This automatically biases the results by skewing the impression of respondents before they even have a chance to answer the questions. The poll completely ignores respondents’ level of trust in the companies by failing to ask if they believe these companies have acted improperly in the first place.

Flawed Collection Process

This poll easily could have been manipulated because it was conducted using only online web panels. The vast majority of major news outlets will not report on data solely collected from an online web panel and usually require a mixture of phone responses in the collection process in order to have solid, representative data.

The poll is also conveniently missing a confidence level. While the margin of error is listed as +/- 2.7%, that’s misleading because the level of miscalculations is unknown without a confidence level.

No Background to Support Findings

The poll makes plenty of bold claims but offers up no background to support its findings.

There are no topline results available for the entire survey linked to the poll. The poll doesn’t contain any external links, except for one that simply opens the same version of the poll in a separate webpage tab. Without the full topline results, it is impossible to view what questions led into the poll, which could lead to biased responses. Data for Progress has hid the questions that are essential in better understanding the survey and the root of the responses.

Although the poll features typical methodology language, there is no information listing basic demographic information or voting history of the sample. Without this data, there is no way of known if the sample was balanced. From the information provided, the poll could have been sampled from a homogeneous group of 1,303 individuals, thus skewing the data.

The Data For Progress website also does not feature any information regarding this poll and there are no polls on the website that show the same, or even similar, results to this poll.

Pulled Out of Thin Air

This poll lacks virtually all the necessary elements to produce accurate data. Between a biased question, lack of topline results, missing background information, and flawed collection process, it is impossible to gather accurate data. To top it off, HEATED limits viewers ability to analyze or even read the poll data by not providing any external links or additional information beyond what is listed in the article. The only insights that can be gathered from this poll is what data looks like when it is pulled out of thin air.