CNN misleads viewers by prematurely declaring Joe Biden the Democratic presidential frontrunner
by Alison Hartson, TYT Army
A controversy was stirred up over the weekend when news sources, including The Washington Times and The Young Turks, criticized the reporting of a recent CNN poll used to determine public support for presidential candidates. While the overall methodology of the poll appears to be credible, there are questionable factors that create doubt about the relevancy of some of the results. Of more concern is the bias with which the poll findings were reported.
Jennifer Agiesta, CNN Polling Director, states:
“A new CNN poll conducted by SSRS after Biden’s announcement on Thursday shows 39% of voters who are Democrats or Democratic-leaning independents saying he is their top choice for the nomination, up from 28% who said the same in March.”
Agiesta reports this poll as though it is conclusive and indicative of overall support for Biden. She did not disclose the important fact that the results of the poll actually represent Democratic or Democratic-leaning independent voters who are age 50 or older.
While SSRS did poll ages 18-49, they were unable to get a representative sample size that can be reliably reported. To their credit, SSRS does indicate this in their 57 page report on page 16:
“Crosstabs on the following pages only include results for subgroups with a minimum n=125 unweighted cases. While interviews were conducted among a representative sample of the adult population of the United States, results for subgroups with fewer than n=125 unweighted cases are not displayed and instead are denoted with ‘N/A’ because they are too small to be projectable to their true values in the population.”
This key fact should have been translated into clear and simple language and then reported by CNN for their audience to easily understand. Instead, Agiesta buried this caveat in the final paragraph of her article:
“The CNN Poll was conducted by SSRS April 25 through April 28 among a random national sample of 1,007 adults reached on landlines or cellphones by a live interviewer. Results for the full sample have a margin of sampling error of plus or minus 3.8 percentage points, for the subsample of 411 Democrats and Democratic-leaning independents who are registered to vote, it is 5.9 points.”
This explanation leaves out the key fact that the subgroup that was measured is statistically skewed since reliable data was unavailable for ages 18-49. These details were not discussed at all in CNN’s television report titled “Joe Biden solidifies frontrunner status.” Instead, they continued to repeat that “Most Democrats,” “Potential Democratic Voters,” and “Democrats” prefer Joe Biden.
Disappointingly, other media replicated this reporting, such as The Hill and Politico.
What is of concern regarding the poll itself are two specific questions that could bias the results. A question was asked about whether respondents had a favorable or unfavorable opinion about presidential candidates: “We’d like to get your overall opinion of some people in the news. As I read each name, please say if you have a favorable or unfavorable opinion of these people – or if you have never heard of them.”
Seven candidates were presented for this question:
- Joe Biden
- Pete Buttigieg
- Kirsten Gillibrand
- Tim Ryan
- Eric Swalwell
- Seth Moulton
- Wayne Messam
Lead candidates such as Bernie Sanders, Elizabeth Warren, Beto O’Rourke, Kamala Harris, Cory Booker, and Tulsi Gabbard were left out.
CNN and other media outlets reported a different question, one in which 24 candidates were available to choose from: “I’m going to read a list of people who may be running in the Democratic primaries for president in 2020. After I read all the names, please tell me which of those candidates you would be most likely to support for the Democratic nomination for president in 2020, or if you would support someone else.“
Why this discrepancy? Why did SSRS ask a question only about the 7 candidates listed above? Were the questions of this poll randomized? Were the questions about those 7 candidates asked before the question about which of the 24 candidates the respondents preferred? The answers to these questions matter because they can influence the respondents’ answers.
There was also a question about Biden that was not asked of any other candidate: “How much do you feel you know about the positions Joe Biden took when he served in the Senate – a great deal, a fair amount, just a little, or nothing at all?”
Why did SSRS choose to ask this question only about Biden? Even if the reason is that he had recently announced his candidacy for president, asking a question solely focused on one candidate creates the possibility of influencing the results of the whole survey even if statisticians don’t consider this to be what is known as a push poll.
It is also worth noting that the Democratic/Leaning Democratic voters age 50 and older support progressive policies. This received little to no attention in the media. Support for these policies could indicate the possibility for Democratic/Leaning Democratic voters to support more progressive candidates who already champion these policies.
Question: “How important is it to you that the Democratic candidate for president supports [insert policy]? Very important, somewhat important, not too important, or not at all important?
Net Results for “Very Important” and “Somewhat Important”:
- 96% Taking aggressive action to slow the effects of climate change
- 91% Providing health insurance for all Americans through the government, a plan sometimes called “Medicare- for-all”
- 85% Taking executive action if Congress fails to pass stricter gun laws
- 78% Making public colleges tuition-free
- 71% Impeaching Donald Trump
- 63% Paying reparations to the descendants of enslaved people
- 60% Restoring voting rights for all convicted felons, regardless of the severity of their crime or whether they have finished serving their sentence
Results for “Very Important”:
- 82% Taking aggressive action to slow the effects of climate change
- 75% Providing health insurance for all Americans through the government, a plan sometimes called “Medicare- for-all”
- 65% Taking executive action if Congress fails to pass stricter gun laws
- 52% Making public colleges tuition-free
- 43% Impeaching Donald Trump
- 31% Paying reparations to the descendants of enslaved people
- 28% Restoring voting rights for all convicted felons, regardless of the severity of their crime or whether they have finished serving their sentence
While 66 percent of respondents said that progressive policies are extremely/very important to them in the candidate they vote for, 92 percent said that it is extremely/very important that the candidate has a good chance of beating Donald Trump. Could that move people away from progressive candidates if they believe that centrist candidates like Biden have a better chance of beating Trump regardless of their policy positions? Were the answer options for this question randomized to increase reliability of these results? These are questions that poll experts in the media should be asking.
Statisticians work hard to ensure that the methodology of polls and the reporting of them is done so in a transparent, scientific manner. The average voter, however, cannot be expected to be an expert on polling methodology, let alone comb through the polls that they rely on the media to translate. It is incumbent upon our media outlets to be aware of their biases, be honest about them, and work towards removing them from their analysis as much as is humanly possible. Anything less is going to appear to be manufactured consent, especially when generalizations are made without clear explanations for a general audience.
Contributor: Leigh Schmitt
Editor: Arun Ravendhran