The Polls versus History

So here we are. Early voting is underway in many states, and we may know the outcome of the election in three weeks. There are a lot of people on both sides who are convinced not only of winning, but of blowing out their opponent. One thing everyone should be able to agree on, is that when the election is over the losing side will have millions of angry voters protesting the outcome.

But it’s strange to see an election this far into the season, where each side has a mix of people in doubt but also people convinced they will win easily. I have been wrong about my expectations before, sometimes right, and sometimes I cannot tell. But experience is a good teacher if you pay attention, and there are lessons that are crying out for consideration. So I am posting this article to post a hypothetical moot court between Joe Biden and Donald Trump for the Presidential Election.

Before going into that, I want to make clear my belief that this election comes down to turnout – if Trump’s supporters stay strong and vote, make the case to their friends and neighbors about who is better for the United States and everyone in it, the President will win re-election. But Biden’s campaign is a technocratic masterpiece, where every one of Biden’s faults and limits is hidden and the media obsesses on painting the President in the foulest colors. If we give up, we lose. If we get complacent, we lose. If we keep our belief in the President to ourselves, we lose. The Democrats mean to steal our future, to kill our energy industries, to reinstall Obamacare, to demand everyone become a meatless prole who mindlessly obeys the state which may or may not let him work, worship, or have an opinion of his own. More is at stake in this election than any I have seen since 1980.

So my moot court starts with the biggest noise in the media – Biden’s big lead in the polls. And as much as I comment about the way polls are weighted and the anti-Trump bias I think is present in so many of them, I recognize that the polls want to get the call right, because this is what they do and failing to get the call right would cost the pollsters a lot. I flinch a bit when I hear someone say they just ignore the polls or believe they are all wrong, because it’s more complex than that. But I do think there are elements of polls which reveal problems for Biden and opportunity for Trump that you do not see in the headlines. I will come back to those in some of Trump’s arguments, but for now I will simply say there are problems with the polls, especially since many of the details have been covered before.

My main counter here is to discuss the difficulty of creating a good, valid poll. Polling firms like to point to the history of polling to demonstrate how reliable they are, but in doing so they accidentally reveal a past to which they cannot return. But first, a brief overview of political opinion polls.

Polling is – for all its pretense – an inexact business, and bias by a polling firm can lead them into bad habits that hurt them later. I have noted before that polls don’t want anyone judging them by any published poll but the last one before an election, and even then they don’t want anything considered but the spread between the top two candidates. If they had been willing to look closely at their polls, the polls which missed in 2016 might have notice that there was a lot of undeclared support sitting in ‘other’ or ‘undecided’ territory. Almost all drastically undercounted Trump’s election support as much as 11 points, which is a serious blunder for any group purporting to seek accurate reflection of the voters’ intentions. My point is that while the polling agencies followed their internal procedures, they did so without really testing their assumptions or validating those procedures.

To put it another way, suppose you run a polling agency and want to put together a national opinion poll on the race for the White House. Let’s say you want 1,200 responses for your poll. In the old days, you would simply get your team to make a buttload of random calls. But those days are gone, and with them, the whole business of opinion polling has changed.

A Washington Post article from 2014 observed that that random calls placed to potential respondents often go unanswered, even after six attempts per number.

The WaPo article blamed most of that on cell phones replacing landlines. The problem, observed the WaPo, is that cell phone users tend to be much younger than landline users, and are different in other demographic characteristics.

WaPo bluntly summed up that landline-centered calling would produce more Republican responses, and cellphone-centered calling would produce more Democrat responses.

Two years later, in 2016, The Huffington Post asked an even more pointed question:

Do Polls Still Work If People Don’t Answer Their Phones?

HuffPo, in February of 2016, noted that most Americans don’t want to take calls from people they do not know, meaning that pollsters – whether calling landlines or cell phones – are likely to be shunned by most voters. The HuffPo article observed most polling firms do not report response rates, but a Pew report in 2012 confirmed that response rates overall had declined less than one fifth the rate of the 1990s.

And in February of last year, Politico reported that phone polling was “in crisis again”.

The big takeaway from the Politico article was the confirmation that many polls would be moving “the lion’s share” of their tracking to online platforms, which essentially ends the purely random nature of respondent pools.

So going back to our scenario of trying to contact 1,200 responses, your polling firm is facing two difficult foe: time and resources. That is, polls have a deadline by which to issue their report, and nobody has an unlimited budget to get those responses. So what polling firms do is to get as many responses as possible that they can use, then they use weighting to balance their results to match demographics.

The problem there, is that invariably you are going to have some demographics where your responses are way below what you want, and while you can weight the responses heavier to make up the impact, the additional, unintended effect you get from that is that any quirky character of that smaller response is magnified. What I mean is not that the polls can be recalculated to say something else, but the polls demonstrated a, well, wobble to their output. A careful look should be applied to current polls, even when they look good for Biden. Take a close look at Trump’s support among Black voters in come polls, for example, or his rising support among seniors in others. Ask why so many polls insist on weighting Republican participation below 30% of all respondents – does anyone really think Republican voters will sit out this election?

Also, we’re well into October, plenty far for polls to be basing their results on Likely Voters. Yet a surprising number of polls are all still using Registered Voters only. That suggests they are having difficulty reaching a sufficient number of responses from Likely Voters. And those who are calling their respondents ‘Likely Voters’ often do so on no test beyond asking if the respondent plans to vote. Common tests like asking if they voted in the primary election or in the last election were simply not even asked. Not asking those basic questions says something in itself.

Now consider the support levels for Trump reported in those polls, anywhere from 87 % down to 76% from ‘Republican’ voters. Democrats can believe in those numbers, because they like to believe in groups like the ‘Lincoln Republicans’, even though close inspection reveals the Lincoln Republicans are just Never Trumpers left over from 2016, who got repackaged and financed by Democrats to sell a story, but they have failed to collect interest, let along numbers. A stronger retort to the claim of lower same-party support for Trump comes from what some call ‘the only real poll’. That is, the primary elections from this spring.

Opinion polls are nice but don’t really have the credibility of actual election results. Democrats want to point to the 2018 elections, but that claim fails on two vital points – first, that the midterm elections were not about the White House race as we see now, so a lot of the attention was on local, state and smaller-office races; and two, that the primary elections – for both Democrats and Republicans – happened this year during the pandemic conditions. I will not mention for now how Biden fared in the primaries, although that is relevant here, but I will note that even though Trump was never in doubt to claim the GOP nomination for his re-election, crowds of Republican voters showed up to vote for him, and Trump collected 94% support in his primary campaign. And Primary voters are definitely going to vote in the fall.

That disparity in poll reports is difficult to explain conclusively, but there is a scenario which seems likely to me. Polls are conducted using either volunteers, like college students, or using relatively low-paid employees who follow a script. In each case, the people taking the poll are usually relatively inexperienced and just want to get the job done as soon as possible. We know from studies done by polling groups that landline phones are far less likely to answer than cell phones, and the most prevalent respondent tends to be urban and young. This means that a lot of polling groups will fill up quickly with responses from young voters in cities like New York and Los Angeles, and especially along the coasts. Least likely to answer are older professionals like plumbers and electricians. It’s easy to say ‘Republicans’ don’t support Trump if your poll focuses on big cities and a younger respondent pool than your presented image. That’s just one obvious example of how polls can mislead, even if they don’t think they are doing that.

You can believe the headlines, or you can look closer at the internal data.

Now on to Trump’s first point – voter enthusiasm. I don’t mean the boat rallies or parades or lawn signs, by the way, because Trump had Hillary beat on rally attendance and lawn signs/bumper stickers back in 2016, but he did not crush her in the overall vote tally. What matters here in terms of voter enthusiasm is the roughly 13 point lead for Trump in enthusiasm in the same polls giving Biden a double-digit lead in headlines. If you simply apply logic to that, this strongly implies that the support for Biden in those polls is very shaky.

Frankly, the closest thing to a response by Biden supporters I have seen, is a collection of forum comments claiming voters can’t wait to get rid of Trump. It’s possible I guess, but it would be the quietest demonstration of anger I can remember on this scale.

Biden’s second attack is the COVID response. Democrats have tried to blame the President for all the deaths associated with the virus, plus their own economic shutdowns and despotic rulings, like Governor Whitmer not only shutting down many stores, but when allowing a store to be open, disallowing purchase of anything she deemed ‘non-essential’, while allowing ‘essential’ purchases like lottery tickets or alcohol.

At first, those attacks seemed to stick a bit to Trump, whose casual behavior regarding the virus has indeed been less than ideal at times. But over time, voters have been able to see for themselves that Trump has done as much as anyone to fight the virus’ spread, while focusing on keeping the economy working. Trump has rebuilt the economy – again – even as Democrats refused to help protect jobs and industries. Biden’s COVID tactics have failed.

That beings me to Trump’s second point – voter expectation polls. Polls have commonly asked voters not only who they plan to vote for, but who they expect to win the election. This matters because when the two types of question are in conflict, the expectation question is more often correct. To be fair, I have started to see a few polls where voters say they expect Biden to win, but overall Trump continues to win this question by about 6-7 points.

Biden’s third point is the Recession. When he first started attacking Trump on the economic conditions, Biden was well aware that the economy is historically the most damaging issue for an incumbent President if a recession happens. But the fact that Biden abandoned the attack by mid-summer was no coincidence. The plain fact is that, often ignored by the media, Trump has rebuilt the economy and re-energized the job market to accomplish results consistently better than expectations.

Trump’s third point is his set of accomplishments. Democrats don’t want voters to consider that Trump’s multiple Nobel Prize nominations are real evidence of solid and successful foreign policy, and most people will admit they are glad to have a President who has not started any new wars. The jobs I have already discussed, but it’s still impressive to observe that Trump did that twice. Trump’s judicial appointments have been difficult for Democrats to attack on any rational basis, and the recent hearings for Amy Barrett have revealed a serious fear by Democrats that their attack-dog tactics are not only failing to prevent confirmation, but make the Democrats look irrational and desperate, which are certainly a political bonus for the President.

Biden’s fourth point is the contrast of personality. This is where the media’s despicable bias has helped Biden a lot. If you pay attention, you will see that Biden has a thin-skin and often resorts to insults when he gets frustrated. Yet the media hides countless outbursts, including racist and sexist comments by Biden. A lot of Biden’s favorability advantage depends on voters believing false claims against Trump, while never learning about Biden’s crass character.

Trump’s fourth point is his minority outreach. It’s laughable, in context, to consider how often the media sells the false image of Trump as a ‘racist’, given the President’s many efforts to help minorities. I will let the New York Post make the argument here:

“African Americans were experiencing the best economy we have ever seen: Unemployment for our racial group was the lowest in recorded history, black wages were rapidly increasing for the first time in decades, and people who’d been out of work long-term were being hired and suddenly able to take their families on vacations for the first time in years.
The Trump policies made it possible.”

“[Trump’s] recent police-reform executive order, the First Step Act, released thousands of people from jail (90 percent of whom were black). He has promoted “opportunity zones” that incentivized private investment into marginalized communities, and also increased federal funding to historically black colleges and universities by 17 percent — a total exceeding $100 million, more than any president in history.”

Trump has plainly done more to help minority Americans than any President in decades.

Biden’s fifth and final point is the New Young Voter. There will be millions of new voters this year, and many of them will be young people voting for the first time. It’s difficult to say how much excitement young voters will find in Joe Biden’s campaign, but it is true that Democrats have often benefited the most from new young voters.

Trump’s fifth point is the New Older Voter. An interesting thing about the New Older Voter, is that the media often ignores them, as if they do not exist. The New Older Voter is someone who may have voted in the past but lost interest, or they may be someone who never found voting worth the trouble. If reports from the Trump Campaign can be believed, almost 15% of people attending Trump rallies say they have not voted before, but will do so this year. If that is true, these voters are likely not showing up in polls, and this group could prove to be an ace in Trump’s hand.

Trump’s final point is the ‘Shy Trump Voter’, although I think that name is misleading. Certainly the polls in 2016 undercounted Trump support, but it’s uncertain as to why the polls missed between 6 and 11 percent, pretty consistently all year. The polls say they have corrected for that but that claim is dubious, but it’s also difficult to consider the idea that a lot of Trump voters are not telling anyone they support Trump, when so many Trump supporters are waving flags and making lots of noise.

One possibility I think is happening, is the Late Decider showing up instead of the Shy Trump Voter. Exit polls from CNN and the New York Times in 2016 revealed that as many as sixteen percent of voters made up their minds in the last few days of the election, and most of the late deciders broke for Trump. Polls this year insist that there are almost no undecided voters left, but if you go back to 2016 you will see comparable numbers to the same question then. The reason is that polls only get responses from voters who are willing to discuss their opinion, and people who are not ready to decide are not ready to talk to polls.

I think this is a hidden advantage for Trump. Some people insult late deciders for taking their time, but I see those voters as rational, careful people who want to avoid making emotional choices, and Biden depends on emotional choices. If a lot of undecided voters are still out there, the longer they look the more I think they will lean towards supporting the President and not the former Vice-President.

But in the end, you need to vote, and you need to get your friends, family and neighbors to vote. We are in a fight with stupendous risks and consequences. It’s not just about politics.

This Is A BFD, As Hidin Biden Would Say
New York Post Publishes Smoking Gun Evidence Of Biden Corruption