Cogito Ergo Non Serviam
Why the Polls are Probably Wrong
The American general election is the Tuesday of next week. The polls are showing a close race between Vice President Harris and former President and current convict Donald Trump. This has a great many Democrats and Republicans worried. Regardless of position on the political spectrum, both sides are baffled that it is so close when the other side is so clearly awful. The fact is that polls only claim to be correct within the margin or error, and even with that proviso, they are right 95% of the time not 100%. This time around, few are paying attention to this fact. Yet, it is quite possible that the polls are not as good as that. In fact, the probability is they are off by more than their margins of error.
The first thing one must understand is that the polls are not used by the campaigns to see who is winning and who is losing as their primary function. Instead, they are used to find out what is working and what is not and with whom. They help refine messaging and where to deploy resources.
The reason behind that stems from the knowledge they have about how pollsters do their jobs. A poll needs to decide whom to poll. The easiest is registered voters. Everyone who is registered is fair game, and that is a group who can be defined objectively. That is different from likely voters, which is determined by guess and by golly. It is subjective. And that is where error comes in.
To decide who their likely voters are, a pollster will start with the actual electoral results of the previous election for a given office. People who voted in the last election are likely to vote in this one, too. But in the four years since the last election, a bunch of kids grew up to be eligible to vote while others have passed away. So they cannot use the exact same criteria.
So, they start making assumptions. When Mr. Obama ran in 2008, it was safe to presume black voters would turn up in greater numbers than they did in 2004 for the Bush-Kerry contest.
The real problem for all pollsters is first time voters. They have never voted before by definition, so how does one figure out who they will be and for whom they will vote? In years passed, the 18-24 demographic just did not turn up at the ballot box often. Getting their voting intentions wrong carried very little penalty because they were too few to matter greatly.
The kids coming of age now are much more politically aware and active. Whether it is the general lack of opportunity or the active shooter drills they have practiced since kindergarten does not matter for now. What matters is that they are turning up and voting in greater numbers than before. Getting them wrong now results in bigger misses. They tend to vote left rather than right. If a pollster has their turnout rate at 30% and they vote at 45%, Democrats are going to be under-represented.
The campaigns do not care. The campaigns just compare the result to the previous poll and check cross-tabs for greater insight. They talk about being up 3% with blue collar voters or down 1% with business owners.
The press and public misuse the polls when they talk about who is winning or losing. If they get first time voters wrong, they could well forecast the wrong outcome.
What factors have changed the electorate from 2020? A black and Asian woman is on the ballot. A felon is on the ballot. Mr. Trump has his own record to defend and Ms. Harris is saddled with Mr. Biden's. So who is the incumbent?
This journal has some experience with polls and votes. This one feels like it is not tight at all. The Harris-Walz team is probably a good 5% ahead nationwide. It will defeat the Trump-Vance ticket by 10 million votes in the popular count. It will have more than 300 electoral votes. The polls do not show that. Either this journal is wrong (not the first time) or the polls are (not the first time either). One will know next week.
© Copyright 2024 by The Kensington Review, Jeff Myhre, PhD, Editor. No part of this publication may be reproduced without written consent. Produced using Ubuntu Linux.
Kensington Review Home
|