More Tuesday Takeaways

In the process of staying up until 3am Wednesday morning watching election returns I had a plethora of thoughts. In yesterday’s article I never got to polling and turnout. Today I’d like to say a bit about both.

At this point a turnout discussion is based on mainly incomplete and/or anecdotal evidence. The reality is that good, solid, national turnout data won’t be available for several weeks. It is the crosstabs that will really tell the story. I’m especially interested in the sub-demographic of white women. Did they show up in larger numbers and did they break for the Democrats? If so, I have to assume Dobbs was the motivating factor.

To this point, based almost exclusively on anecdotal evidence, it appears the youth vote was up. I have to think that was choice motivated but, again the cross tabs will be helpful in verifying the anecdote and the motivation. If turnout was up, especially among young voters that speaks well for the future of democracy in America.

Now we get into the much-misunderstood area of polling. We all remember the famous 1948 election headline about Dewey defeating Truman. The reality is that Harry Truman won the 1948 election. The polls, along with a deadline, influence the headline and were wrong mainly because of faulty sampling.

If you really knew how to read the polls, they were actually very accurate in 2022. The good ones told you they didn’t know who was going to win. That doesn’t mean they are not challenged. Too many polls are not worth paying attention to. Any with a small sample size need to be immediately discarded. Many pollsters are making a good living basically feeding campaigns the answers they want to hear. Their samples are heavily skewed to align with the candidate’s philosophy. If I only survey my friends I’m always going to be viewed as a good guy.

That is not to say that polling is easy. These days the most cost and time efficient way to conduct a poll is via telephone. With Caller ID the response rates have dropped. Many are in the 1 and 2% areas. That means you might have to dial 100 numbers to get one complete survey. I, personally received several calls from polling outfits and never answered any of them.

Another factor is just how honest and forthcoming the person will be with a pollster. Paraphrasing Nicole Wallace, if a woman is deleting her period tracker from her phone do you honestly expect her to be forthcoming with a telephone pollster?

The Bradley Effect (in which a person won’t admit to racial prejudice in a conversation with a pollster) still lives in more than one form. It is assumed that many Trump supporters will not admit to it in a polling call. They want to mess with (PG rating to be considered) the pollster, who they view as part of a Deep State/mainstream media conspiracy, or they simply don’t want to be viewed as stupid which deep down, at least in this regard, they know they are.

There are also voters of all persuasions who simply do not want to talk to pollster because they really hold their vote as personal and secret. These do not affect the quality of polls. They just make polling more difficult.

Perhaps the most critical criterion for a reliable poll is its sample. Was it likely voters? Registrations don’t vote. If the person is unlikely to vote I really don’t care about them. A good sample resembles and approximates the turnout. Turnout demographics change with each election. Does your sample include a representative percentage of women, high school dropouts, people of color, etc.? More challenging is how does the pollster know that before the election? The short answer is they don’t; nobody does.

Polls are not always definitive answer machines. Sometimes the most honest answer is, “I don’t know.” Lately that has been the answer a lot of the good ones have been giving.

This article is the property of and its content may not be used without citing the source. It may not be reproduced without the permission of Larry Marciniak.