ON POLLS AND PREDICTIONS
(Reflections on this Political Season)
by Gennady Shkliarevsky
“There are lies, damned lies, and statistics” is a phrase of unknown origin that was popularized by Mark Twain in Chapters from My Autobiography. The phrase refers to the growing power that numbers have on public consciousness—the widespread belief that statistical numbers tell us something that is objectively real and true.
The phrase, however, conveys something contrary to this belief. It basically suggests that numbers can be the worst possible liars; and they are. We all remember how 2016 all political analysts and pundits predicted a landslide victory for Hillary Clinton. The reality proved to be different and different—and by a lot, not just by some fractional smithereens.
Many supporters of then-president-elect Donald Trump ridiculed this gross mistake in predicting the outcome. No one could explain why this mistake could have occurred.
The pollsters remained generally silent about this strange discrepancy. Clinton supporters could find no better explanation for it than that Trump and the Republicans stole the election. They simply could not believe that their numbers lied. Yet the facts stared them in the face.
One possibility that has not been considered by either side is that the entire statistical method may be profoundly flawed. For those on either side of the aisle, this possibility is simply too iconoclastic, too subversive because it undermines the firm belief in the objective power of numbers. Since only but a few minds, viewed as nothing but eccentric, dared to question this belief, it has remained largely intact.
In this election season, just as in the previous ones, polls and statistics continue to hold a firm grip on the public mind; and battles over statistics continue to rage uncontained.
The reason for the factual vulnerability of predictions based on statistics is very simple. Many of us know very well that when we look at objects from different angles, they look different. Numbers are also objects—mental objects; and, like all other objects, they also depend on one’s point of view—an angle or perspective from which we view them. We do not know what to do with this dependence because it involves our subjectivity. Subjectivity appears to be ineluctable; its impact seems to be impossible to eliminate.
The reason for this rather long preface is the fact that a similar situation emerges again in the current election season. Again, as in 2016, we have multiple polls, statistics, and analyses that convincingly demonstrate Biden’s significant lead over Donald Trump.
Number wizards are more cautious now, however. Andrea Benjamin, a professor from the University of Oklahoma, is optimistic about numbers, including sales of books on race-related problems in America; she sees a trend to the left that may benefit the progressives in November. Yet her conclusions are not particularly sanguine. In her interview for the NYTimes, Benjamin intimates: “Let me be clear, we need real change and real reform to try to start chipping away at these outcomes and I am less optimistic that we are really ready to make those changes.”
Yet the number game continues. Analysts persist in using their polls to make predictions. Thomas B. Edsall—one of the big guns in the NYTimes who uses “hard data” for his predictions—has published a piece this week based on interviews with a number of prominent scholars who analyze a variety of statistical data that they have amassed and interpret them for the readers. They all argue that based on their data, American voters are leaning leftwards—the fact that, in their view, it is very likely to affect the outcome in November and benefit the progressives.
They all base their predictions on polls they conducted. As I have shown, outcomes of polls depend very much on assumptions that pollsters use in formulating their questionnaires. In this case, pollsters have made an assumption that the touchstone for determining the drift of the electorate is issues related to “white privilege” and policing.
Their numbers show that more Americans, particularly among the young but not only, today feel that white Americans are privileged by comparison with black Americans, which is what liberals argue. Based on these numbers, pollsters conclude that the electorate is leaning to the left, which should benefit Biden.
As to policing, the pollsters find that while people across the entire spectrum, including black Americans, support police, many of them also feel that some reforming and weeding out “racist cops” is in order. In their view, this growing opposition to “systemic racism” will translate into more support for progressive liberals and their candidates.
The numbers are good and they do warrant the conclusions. The assumptions on which the polls are based are, however, questionable. They do not include one of the main issues that drives today’s politics. The issue that has shaped the current political crisis is the attitude toward elite rule. This is not to say that the issues related to race and policing are unimportant. This is to say, however, that in the current political climate, the attitude toward elites underlies these issues.
Indeed, if you look at the electoral campaign of 2016, the rebellion against the elite rule was the main cause for the rise of Donald Trump. It also created the phenomenon of Bernie Sanders. Although the progressive establishment was able to quell Sanders’s rebellion among their rank and file, the cost of this suppression was substantial.
If you look at the current campaign, we also see strong anti-elite sentiments among Americans. This sentiment sustains and even strengthens the base of support for President Trump and the Republicans. But this attitude is not characteristic just for conservative voters. If you look at Black Lives Matter and Antifa—the two major players in the political arena today—they both reveal strong anti-elitists attitudes. Black Lives Matter directs its attacks against the establishment in general making little distinction between progressives and conservatives. BLM, for example, has filed a lawsuit against YouTube and Goods alleging racist practices. In recent protests, extremists attacked the CNN headquarters in Atlanta.
For extremists, progressive liberals and Joe Biden represent “white privilege” just as much as Donald Trump and the Republicans do. They often criticize Biden and his campaign for their white their “white sins” and demand that the Democratic candidate should make amends. They do not settle for symbolic gestures, such as removing statues and portraits. Extremists demand that the progressive electoral agenda should include their demands for free healthcare, free education, and no police. Moreover, Black Lives Matter is also very critical of the progressive black leaders for failing to deliver any tangible benefits for the black community. You do not see the likes of Jesse Jackson and Al Sharpton appearing alongside the leading figures of BLM. BLM does not invite them for speeches and joint appearances.
Pollsters’ conclusion that there are more white Americans, particularly among the young, who voice their protest against “white privilege” is not lying. The numbers bear it out. However, the inference that pollsters also make that this shift means greater support for the progressive liberal cause has no justification. The whites who speak against “white privilege” do not joint Biden’s campaign. They give their support to extremist organizations such as BLM and Antifa, which does not necessarily translate into support for the Democratic Party. Leaders of Antifa are more like Joseph Alcott, aka Joseph Martin who has openly acknowledged that he is an active communist who hates America and seeks its destruction than those who work for the Biden campaign. If the Democrats think that they can hijack the extremist movements to serve their purposes, they better think again. Both BLM and Antifa know their strength and intend to use it to lead the progressives, not to follow them. They have recently warned Biden that he should cave in on key issues important to the radicals or pay the price.
The prevalent attitudes toward policing also do not suggest that they mean more support for the progressives. Progressives follow Antifa and BLM that demand dismantling the police, which makes them in the eyes of many moderates and independents opponents to law and order. Poll data indicate that Americans in general across the board are in favor of law and order and do not favor dismantling or even defunding the police. Even black Americans do not favor eliminating the police. They are only 7.3% less likely than whites to support the police. Latinos and Asian Americans are even more in favor of the police than whites are.
The conclusion that the above warrants is this: The shifts that polls register are real but the numbers are deceptive. These shifts do not translate into greater support for progressive liberals and Biden and are likely to work against them.
So what are the likely predictions that this conclusion supports?
The way things are going now we can safely assume that protests and riots will continue, which puts Biden and the Democrats in a tight spot. If they support the protestors, moderates and independents will see them as opponents of law and order. If the Democrats lose this cohort to Trump, they may just pack their suitcases and go home. They will be extremely unlikely to win. If they do not support ongoing protests and riots, the extremists will see them as their enemies and will be unlikely to boost their chances of winning.
At the same time, President Trump and the Republicans will be regarded as proponents of law and order and are likely to preserve and perhaps even expand their base. In fact, as protests continue to rage, the support for President Trump has reached new heights as witnessed, for example, by the parade of Floridians on Trump’s birthday and the enormous response to his rally in Tulsa, Oklahoma, with a number of request for ticket topping a million people. The optimistic predictions for economic recovery will undoubtedly provide a powerful boost to the Trump campaign. Finally, President Trump will have at his disposal vast resources available to him as head of the federal government.
The extremists and their radical tactics are not likely to bring any kind of normalcy to areas under their control. One important reason is the fact that their networks operate on the level of local interactions. In order to sustain themselves, these networks have to create a global level. In other words, they have to create a hierarchy. The fact, however, is that they are averse to hierarchies and fear them. Without creating such a global level of organization, radical networks will be unable to sustain themselves and solve the numerous problems they face.
Solutions for these problems will require regulation, optimization, and rationalization that can only be accomplished at the global level. In the absence of the global level, difficulties, tensions, and privations in “free zones” controlled by the radicals are likely to increase.
The result will be the erosion of the support that extremists currently enjoy, as many supporters will start abandoning them. Extremist groups will fade away on their own, as it happened during the Maidan movement in Ukraine. The federal government will only have to come in and clear the mess.
These are the predictions that logically follow from what progressive pollsters claim are true numbers. These numbers are true; they indicate real shifts in America. I wholeheartedly agree with these pollsters. In contrast to them, however, I see that these shifts are likely to work for President Trump and the Republicans.
Gennady Shkliarevsky is a Professor Emeritus of history at Bard College.