Fake News Outlets Are Now Pushing Fake Polls for The Kamala Narrative

In the September edition of Strategic Intelligence, I provided a detailed statistical study of why recent polls of the presidential race have been rigged in favor of Kamala Harris in a way designed to maintain a narrative that she is a young, energetic new kid on the block instead of just another tired old politician with little to offer.

A recent example was The Hill, which ran a headline: Harris leading Trump by 7 points: Poll. Wow.

RealClearPolitics had Harris ahead by 1.7 points on the same day. How does that happen? You have to consider the polling methodology to see what’s going on.

The Hill headline was based on a poll conducted by Fairleigh Dickinson University. The poll was conducted among “registered voters” using only 801 respondents and the sampling error was plus or minus 3.5 percentage points. That poll is garbage. Here’s why.

To conduct a statistically valid poll you need at least 1,300 respondents. That will give you a decent result. Anything less has a large margin of error and is of little value. The sample of “registered voters” is also garbage because almost half of registered voters don’t actually vote.

The only polls worth considering are conducted among “likely voters” based on some measure of likelihood such as whether they actually voted in any of several recent elections. Beyond that, there are statistical games that are hidden from view such as the skew in poll respondents.

Assume voter registration in a certain jurisdiction is 27% Democrat, 27% Republican and 46% Independent. What some pollsters do is to create a sample of, say, 35% Democrats, 25% Republicans and 40% Independents instead of using the actual distribution.

When you oversample Democrats, the result naturally overstates support for the Democratic candidate. That’s huge in a tight race. Even if you narrow down your sample to likely voters, you still have to adjust for what are called “low propensity” and “high propensity” voters.

The high propensity voters are those most engaged in politics who watch practically every headline and twist and turn in the race. Low propensity voters are those who pay little or no attention and may or may not turn out on Election Day even if they fall in the “likely” category. The difference matters in a high turnout election where even those not particularly focused on politics feel there is a lot at stake, and they should make the effort to vote.

The 2024 election is definitely one of those crucial elections where low propensity voters turn out and those voters tend to be more supportive of Trump (blue collar, no college, most hurt by inflation, etc.). The polls are not well-calibrated to identify and reach out to low propensity voters and therefore systematically underestimate Trump’s actual support.

There’s a lot more in the article and we hope you get to read the entire piece. No sooner had we distributed our article to subscribers then we got confirmation of our analysis from Zero Hedge, which published this article. It makes the same points about oversampling that we made and warns Kamala Harris supporters that they’re making a huge mistake if they believe these media polls.

It’s also good news for Trump supporters. You have to look beyond the fake news to see what’s going on. Right now, it looks like Trump is maintaining a lead in what is admittedly a tight race.

Corporate leaders and institutional fiduciaries looking to incorporate state of the art predictive analytics to their risk mitigation and strategic analysis should click the link to learn more about Raven Predictive Analytics®.