I have personally worked a few campaigns and have been close to other operatives who worked other campaigns. There was a time when I saw fit to question the accuracy of a poll, and because of my position in the campaign, was able to talk to the director of this major newspaper's polling department. He and I had several in-depth conversations about polling in general and his in particular. That does not make me an expert bit I do have some insight into how campaigns use polls.
Yet campaigns make their decisions on internal polling, ...
Not entirely. The campaigns that have the money to spend on internal polling use them for a variety of reasons. In some cases, the reason has to do with the campaign manager's job security. If the candidate does not like what he or she sees in the public polling, or if the staff is demoralized by what the external polls say, some managers will hire a company to do a private poll (internal poll), which will be structured to produce the result the candidate and staff want to see. A happy candidate does not fire the campaign manager. A motivated staff keeps working.
If external polls were accurate, they would accurately predict where candidates would spend their money and their time.
That would be true if poll numbers were the only thing to consider, but that is never the case. A host of variables determine where candidates spend their money and time. External and internal polling is certainly considered, but so are the subjective judgements of of the candidate and the campaign leadership team. And those can be flawed.
For example, right now, and for the past couple weeks, the polls show Trump behind Biden in numerous issues and several key states. How is Trump spending his time? At rallies where he speaks not to voters who may change their minds to vote for Trump, but to voters who long ago made up their minds. He is doing nothing to demonstrate leadership in Washington (like working to get another stimulus package through Congress). He's not holding highly visible meetings with suburban women (who the polls say he is losing) or other key groups. Instead, he's doing his rally stand-up routine to entertain his base and nurture his narcissism. He's not letting the polls guide his time management decisions. He's basking in the praise of his supporters. Flawed judgement.
For example, external polling averages show Biden with a commanding 7 point lead in Pennsylvania. If that were accurate, Biden and Company would be spending little if any time there, and even less money. But look what's happening. Biden had had events there 3 times in the last two weeks, the campaign is throwing cash at television advertising in the state, and if that's not enough, Obama campaigns in PA this week for Biden.
That too may be a case of flawed judgement. I have no special insights into the Biden campaign but an equally plausible explanation is Biden went to Hillary school. To this day, Democrats are warning against the complacency that contributed to Hillary's 2016 defeat. Biden may be working PA hard because PA is a big electoral vote state and he does not want to make Hillary's Michigan mistake. Today, the winning percentage of people say they are voting for Biden. But will they actually turn out to vote? The Democrats learned a hard lesson in 2016. Take nothing for granted.
You don't do any of that stuff if you have a 7 point lead. You do it, however, if you within a point either way.
That would be true if polling was the only factor in play. But as I said above, there are many factors in play.
Trump's had a lot of events there, too. Not just him, but his suurogates, as well. It would seem the internal polling for Biden and Trump are telling both campaigns the same thing.
Again, I have no inside knowledge of the Biden or Trump campaign (and neither do you). It could well be that both Trump and Biden disbelieve the polls and instead simply have their eyes on the big electoral college prize PA is. If polls alone drove campaign decisions, Trump would not go to CA. But he went. Why did he do that? He did it because of another factor there in play, the ability to raise funds from rich Trump supporters there.
Another indicator that internal polls are more accurate than external polls is the rolling changes in Party Affiliation (Gallup tracks that weekly). It's the percentage of people who consider themselves Democrat or Republican, and the percentage of Independents who lean one way or the other. Currently, Democrats are down 2 over previous tracking, Republicans are about the same, and Independents are leaning Republican by 4. That means people are leaving the Democrat Party and not necessarily for the Republican Party, but rather for Independents and are leaning Right. That flies in the face of public polling, but matches up closely with the hints about what internal polling shows.
You're drawing a distinction without a difference. There are lots of Republicans who STRONGLY identify themselves as Republicans, but are voting for Biden. In 2016, there were lots of Democrats who strongly identified themselves as Democrats, but voted for Trump. While it can be generally presumed that a person who identifies ones self with one party or another will vote for that party's presidential candidate, that is not always the case. It is an unfounded leap to suggest that this general presumption somehow validates the accuracy of internal polls.
All of the public polls will tighten up by election day, so they can save face and hang on to a modicum of credibility.
If the polls are going to tighten, they better hurry. In the last three weeks, they have widened in Biden's favor, and only two weeks remain before Election Day. A summary review of the polls in 2016 show no pollster face saving is necessary. National polls measure public opinion and voter intent, not the electoral college result. In 2016 the polls (taken as a whole) were spot-on predicting Hillary's popular vote win (
see this post). In two weeks, plus a few days for final vote counting, we'll know how the pollsters did this time around.
www.cnn.com
