Tag Archives: polling

About Those Polls

Those of us who are obsessed with the upcoming election–now less than a month away–tend to focus on the the daily polling results. I have previously explained why I don’t think today’s polls are particularly predictive–while they can show which way the wind is blowing, I simply don’t trust their “likely voter” assessments. (As I’ve explained, all pollsters have developed methods for determining who is likely to vote–and their polls are almost always based upon the preferences of those “likely” voters–not the entire universe of registered voters.)

I feel reasonably confident that we will see a lot of votes cast by “unlikely” voters this time around.

But there is also something new–and dishonest–that has emerged in the polling during this election cycle, as Simon Rosenberg recently reported in his Hopium Chronicles. (Paywalled) He calls them “Red Wave Pollsters.”

Red Wave Pollsters Stepped Up Their Work This Week – The red wavers stepped up their activity this past week, releasing at least 20 polls across the battlegrounds. It’s a sign that they are worried about the public polling in both the Presidential and the Senate, and have dramatically escalated their efforts to push the polling averages to the right and make the election look redder than it is.

While they released polls in many states this week the states that have received the most red wave polls over the past few weeks are Montana, North Carolina and Pennsylvania. Over the past 10 days, depending on how you characterize the pollsters, they released at least 5 and as many 7 polls in Pennsylvania alone. Their recent flood of polls in NC and PA tipped the Real Clear Politics polling average for each state to Trump, which then in turn got Trump to 281 in their corrupt Electoral College map. Yes, in Real Clear Politics Trump is now winning the election due to their gamesmanship.

I now count 27 Republican or right-aligned entities in the polling averages: American Greatness, Daily Mail, co/efficent, Cygnal, Echelon, Emerson, Fabrizio, Fox News, Insider Advantage, McLaughlin, Mitchell Communications, Napolitan Institute, Noble Predictive, On Message, Orbital Digital, Public Opinion Strategies, Quantus, Rasmussen, Redfield & Wilton, Remington, RMG, SoCal Data, The Telegraph, Trafalgar, TIPP, Victory Insights, Wall Street Journal.

Rosenberg says it’s time for those who publish their analyses of polls to acknowledge the emergence of this type of poll , which he describes as “red wave, right-aligned narrative polling that only exist for a single purpose – to move the polling averages to the right.”

They are exploiting the “toss it in the averages and everything will work out philosophy” of these sites to once again launder these polls and game the averages – and thus our understanding of the election. Party leaders should expect them to keep these polls coming, and keep working the averages until it looks like Trump is winning in all polling averages. It is what they did in 2022, and it worked. They are doing it again this time, and once again it is working as the averages are moving and everyone is treating this movement like an organic rather than a deeply corrupt process.

Simon Rosenberg and Tom Bonier of TargetSmart were the only two pollsters who predicted the non-emergence of the widely hyped “red wave” in the 2022 midterms–a wave that was widely forecast partly on the basis of voting history and partly on the basis of similar “red wave” polling.

Reputable pollsters face a number of daunting challenges–the shift from landlines to mobile phones, the reluctance of many (if not most) people to answer calls when they don’t recognize the number, evidence of the increasing willingness of respondents who do answer to lie… Despite those challenges, nearly all reputable pollsters find Harris ahead nationally by somewhere between 2 and 5 points. While I’m reluctant to rely on their numbers, I do think they demonstrate that the Democratic ticket has the momentum–that the electoral “wind” is blowing in the Harris/Walz direction.

What we do know with certainty is that this will be a turnout election. Early voting is open in most states right now, and the most effective thing we can do is vote early and work to make sure that every Blue voter we know gets to the polls. As the GOTV experts tell us, every solid Democratic vote that is cast early means that the GOTV effort can concentrate on getting those who are farther down the “reliable voter” list to the polls. 

We’re down to the wire, and as the saying goes, the only poll that counts is the one on election day.

Comments

OK–Let’s Talk About Those Polls

Survey research ain’t what it used to be.

Back in 2020, the Harvard Business Review summarized the changes that have diminished polling accuracy. The article described the industry as “living on borrowed time,” and predicted that its increasing errors would not be soon–or easily–corrected.

The basic problem is low response rates. Thanks to caller ID, fewer Americans pick up the phone when a pollster calls, so it takes more calls to reach enough respondents to make a valid sample. It also means that Americans are screening themselves before they pick up the phone.

So even as our ability to analyze data has gotten better and better, thanks to advanced computing and an increase in the amount of data available to analysts, our ability to collect data has gotten worse. And if the inputs are bad, the analysis won’t be any good either.

It now takes 40+ calls to reach just one respondent. And there really is no reliable way to assess how those who do respond differ from those who don’t. (I know my own children do not answer calls if they don’t recognize the phone number–are they representative of an age group? An educational or partisan cohort? I have no idea–and neither do the pollsters.) There are also concerns that those who do respond are disproportionately rural.

These things matter.

A sample is only valid to the extent that the individuals reached are a random sample of the overall population of interest. It’s not at all problematic for some people to refuse to pick up the phone, as long as their refusal is driven by a random process. If it’s random, the people who do pick up the phone will still be a representative sample of the overall population, and the pollster will just have to make more calls.

Similarly, it’s not a serious problem for pollsters if people refuse to answer the phone according to known characteristics. For instance, pollsters know that African-Americans are less likely to answer a survey than white Americans and that men are less likely to pick up the phone than women. Thanks to the U.S. Census, we know what proportion of these groups are supposed to be in our sample, so when the proportion of men, or African-Americans, falls short in the sample, pollsters can make use of weighting techniques to correct for the shortfall.

The real problem comes when potential respondents to a poll are systematically refusing to pick up the phone according to characteristics that pollsters aren’t measuring…. if a group like evangelicals or conservatives systematically exclude themselves from polls at higher rates than other groups, there’s no easy way to fix the problem.

As the article notes, with response rates to modern polls below 15%, it becomes extremely difficult to determine whether systematic nonresponse problems are even happening.

These problems go from nagging to consequential when the characteristics that are leading people to exclude themselves from polls are correlated with the major outcome that the poll is trying to measure. For instance, if Donald Trump voters were more likely to decide not to participate in polls because they’re rigged, and did so in a way that wasn’t correlated with known characteristics like race and gender, pollsters would have no way of knowing.

Then there’s the failure of likely voter models.

People tend to say they’re going to vote even when they won’t. Every major pollster has its own approach to a “likely voter” screen, but they all include a respondent’s previous voting behavior. As long as that behavior stays stable, these models work. But when something generates turnout among voters who have previously been absent, all bets are off. That happened when the Obama campaign energized previously apathetic voters, and since the Dobbs decision overturning Roe v. Wade, we’ve seen evidence of significantly increased registration and turnout among women who hadn’t previously voted.

As the Harvard article noted,

It may be the case that standard sampling and weighting techniques are able to correct for sampling problems in a normal election — one in which voter turnout patterns remain predictable — but fail when the polls are missing portions of the electorate who are likely to turn out in one election but not in previous ones. Imagine that there’s a group of voters who don’t generally vote and are systematically less likely to respond to a survey. So long as they continue to not vote, there isn’t a problem. But if a candidate activates these voters, the polls will systematically underestimate support for the candidate.

Polling is broken, and we need to stop hyperventilating about their results. Remember, Trump has consistently underperformed his polling percentages in every primary thus far this year.
As the saying goes, the only poll that counts is the one on election day.
Comments

Some Soothing Figures

Among the Substack newsletters I regularly receive are those from Heather Cox Richardson–whom I often quote– and Robert Hubbell. One of Hubbell’s recent missives contained some very welcome information and analysis.

Hubbell began with his frequent admonition that Democrats should be confident, but definitely not complacent–that we will need to work hard to turn out every anti-MAGA voter in November. But that said, he made two very important–and comforting–points:

Trump is running his campaign as an incumbent president. He has accomplished a hostile takeover of the Republican Party apparatus. He has threatened to banish any Republican who supports or donates to his opponents. Under those circumstances, anything less than a Soviet-style win of 100% is a failure.

So, against that backdrop, Trump’s loss of 40% of the vote in the South Carolina primary is devastating. It is particularly bad because he lost 40% in a state that is more favorable to him than almost any state in the union—because of its strong presence of white, older, evangelical voters (60% of voters are white evangelical or born-again Christians). Losing 40% of the vote under those circumstances should send shockwaves through the Republican establishment.

As Hubbell quoted Axios:

If America were dominated by old, white, election-denying Christians who didn’t go to college, former President Trump would win the general election in as big of a landslide as his sweep of the first four GOP contests.

It’s not. That’s why some top Republicans are worried about the general election in November, despite Trump’s back-to-back-to-back-to-back wins in Iowa, New Hampshire, Nevada and South Carolina.

Or, as Josh Marshall of Talking Points Memo put it, Face It: This is a Weak Showing for Trump in South Carolina.

It is not merely that Trump lost 40% of the vote. It is also that 50% of those voters said they would not vote for Trump if he became the nominee—which translates into 25% of Republicans who will not vote for Trump!

One quibble: it translates into 25% of the Republicans who went to the polls in the primary who will not vote for Trump in the General. Some of those voters will stay home in November, but that percentage probably is also predictive of the percentage who didn’t vote in the primary but who will vote in the General.

Hubbell’s most reassuring–and eye-opening– analysis, however, was his discussion of contemporary polling, and its demonstrable bias.

Polls do not “predict” outcomes of races; rather, they predict ranges of outcomes at different levels of confidence. But on average and over time, polls should cluster around the actual outcomes. That is not happening with polling regarding Trump.

Instead, the polling averages have consistently overstated Trump’s support—something the media and pollsters have ignored or excused. At some point, they should simply admit that their polling models are broken and overstate support for Trump.

Adam Carlson posted the following on Twitter, comparing Trump’s average margin of victory predicted by 538.com versus the actual margin of victory by which Trump won the first three GOP primaries:

In Iowa:

  • Final 538.com Average: Trump +37
  • Final Result: Trump +30

In New Hampshire:

  • Final 538.com Average: Trump +18
  • Final Result: Trump +11

In South Carolina:

  • Final 538.com Average: Trump +28
  • Final Result: Trump +20

Notice a pattern? The average of 538.com’s polls overstated Trump’s support by at least 7 percentage points in three primaries.

I will add that 538.com is probably the most credible of all the polling sites.

Since Hubbell’s post, Michigan held its primary, and the trend held. Trump won by roughly 42 points; the final 538 polling average had him winning by 57 points, an underperformance of some 15 points.

Hubbell is undoubtedly correct when he says that when polls show a consistent bias, there is likely to be a flaw in the methodology that warrants skepticism. Here, that flaw consistently overstates Trump’s support. As he concludes,

My point is that we should ignore the polls. We should not delude ourselves, but neither should we trust polls that consistently overstate Trump’s support. Just keep working hard and ignore the uncritical, breathless reporting about polls that have shown consistent bias in favor of Trump.

I share these little rays of sunshine to remind you–and remind myself–that the future will be what we make it. Nothing is certain–certainly not the polls.

As unsettling as it is, we live in a time where there simply are no comforting political verities, no outcomes we can confidently predict. Polls are inaccurate, artificial intelligence is creating misleading messages, media fragmentation and online propaganda encourage confirmation bias…The list goes on.

We need to “power through” this very confusing environment, separating wheat from chaff to the extent possible. We also need to reassure ourselves that, since most Americans are sane, we need to GET OUT THE VOTE.

Comments

Beyond Cherry-Picking

A recent essay from The New Republic addressed a question that constantly bedevils me: why do people firmly believe things that are demonstrably false?

I’m not talking about questions that are simply unprovable, like “is there a God?” I’m talking about aspects of our common experience about which there is ample data from credible sources. The linked article, for example, looks into the widespread belief that America’s economy is struggling, when all of the data confirms the opposite.

In the article, Timothy Noah dubs the journalism tracking such unsupportable beliefs as the “Folklore Beat,” and provides examples:

Covid vaccines are unnecessary. Foreign aid constitutes two-thirds of the federal budget. Donald Trump won the election. Schoolteachers are trying to turn your children gay or trans. Little green men visited Area 51, and the military doesn’t want you to know.

Noah is impatient with the media’s tendency to report respectfully on the people espousing those beliefs.

I’ll grant that when misconceptions acquire a large following (though seldom a majority one), that’s news. But is it really necessary to hand a megaphone to every street-corner blowhard in America? News organizations don’t do this because they believe what the blowhards say. They do it because they’re sensitive—too sensitive, if you ask me—to any accusation that they’re out of touch with John Q. Public. And while it’s certainly necessary to document ways in which macroeconomic data fails to capture the complexities of everyday life, particularly with respect to economic inequality, how many times do I have to hear some uninformed fool expound on how President Joe Biden is mishandling the economy? He can’t prove it; he’s not trying to prove it; he just feels that way, and we mustn’t disrespect feelings.

When it comes to the economy, polling continues to show much of the public unhappy with Biden’s performance–although, as Noah notes, “the Wall Street titans on whom Biden wishes to raise taxes maintain a higher opinion of Biden’s economic stewardship than the public at large.”

Perhaps that’s because the rich watch economic matters more closely.

Speaking of the rich, Morgan Stanley recently quadrupled its prior estimate of GDP growth for the first six months of this year, and doubled its GDP growth prediction for October–December 2023, signaling that its economists no longer anticipate a recession. But only a paltry 20 percent of respondents to a CNBC  survey released the same week agreed that the economy was excellent or good. Other polls have returned similar results.

The New Republic essay enumerated the truly excellent economic facts of life–employment and paychecks up, inflation down, manufacturing returning to the U.S., etc.–and then considered reasons for the public’s evident dismissal of excellent economic news.

As with so many aspects of American life today, the answer turns out to be partisanship.

In 2016, Gallup polled voters on the economy one week before the election and one week after. During the week preceding the election, with President Obama in the White House and Hillary Clinton widely expected to win, only 16 percent of Republicans thought the economy was improving, compared to 61 percent of Democrats. One week after the election, fully 49 percent of Republicans suddenly thought the economy was improving, compared to only 46 percent of Democrats. Note how much greater this post-election swing was for Republicans: 33 percentage points, compared to 15 for Democrats….

How does voter opinion differ according to party identification on Biden’s handling of jobs and unemployment? So much so as to render the 47–48 percent figure meaningless. Among Democrats, 84 percent approve, in rough approximation to objective reality. Among Republicans, only 15 percent do.

Inflation? Only 5 percent of Republicans approve of how Biden handled that, as against 71 percent of Democrats. If the judgments of both remain less favorable than on jobs and unemployment, that’s because inflation, though greatly diminished, remains above the Fed’s target level of 2 percent (though if you ask me, 3 percent inflation is pretty low).

The inescapable conclusion is that when you ask somebody whether the economy is doing well, you won’t get an answer about the economy. You may not even get an answer about that individual’s personal experience (which may or may not reflect broader economic trends as compared to one, two, or 10 years ago). Most of the time, you’d be better off just asking, “Are you a Democrat or a Republican?” Because these days, that determines how people—especially Republicans—feel about pretty much everything. If the man on the street sounds like a blowhard, hyper-partisanship explains why. The rest is just noise.

Partisan polarization has overwhelmed reason. Tribalism now dictates interpretations of reality. And of course, thanks to the Internet and social media, it’s easy to find “evidence” to support your preferred version of even the most unlikely “facts.”

Welcome to Never-Never Land.

Comments

Worse Than Trump

Ron Desantis (Taliban: Florida) has announced his campaign for President in a widely-mocked, glitch-filled Twitter appearance, so it may be time to re-read a February New Republic essay “A DeSantis Presidency Could Be Even Worse Than Trump.”

The sub-head is brutal: “Donald Trump was and is a lazy, ignorant narcissist. The Florida governor is a smart, motivated, very right-wing Catholic who wants to remake America as he imagines God wants it to be.”

The article preceded DeSantis’ precipitous plunge in the polls, and takes seriously the possibility of DeSantis capturing the GOP nomination. Back in February, before the general public became more familiar with DeSantis’ agenda (autocratic) and personality (wooden/off-putting/missing), the Florida Governor was viewed as relatively sane and intelligent, at least compared to Trump, and many in the public  assumed that meant DeSantis would be a better president than Trump.

The following paragraphs from the essay reminded me of a gloomy conversation I had with a friend the weekend after the 2016 election. Her most hopeful prediction–which turned out to be very accurate–was that Trump’s manifest incompetence would prevent him from actually enacting much of the damage he was promising.

DeSantis differs from Trump in several ways. While Trump couldn’t care less about “critical race theory” or transgender people, and simply throws stuff into speeches at his rallies that get the biggest reaction, DeSantis is a deeply conservative Catholic and a true believer in the culture wars he engages in. The other key difference is that DeSantis is a Harvard- and Yale-educated lawyer, while Trump skated through a bachelor’s in business where one of his professors called him “the dumbest goddamn student I ever had.”

The damage Trump was able to do was limited by his lack of discipline, ignorance of how the system worked, laziness, and lack of motivation. He is simply a narcissist who likes feeling rich, powerful, and important. DeSantis, however, is none of these things. He is not lazy. He has discipline, motivation, and an intimate knowledge of how to use the system to get what he wants. DeSantis fully intends to remake America the way he believes God would want it to be, and his knowledge of law and governmental structure allows him to do it on a scale, and with a precision, that Trump could only dream about.

The essay noted that DeSantis was pursuing what it called “one of the most aggressively authoritarian agendas in the country” by using two primary strategies: “capturing the referees and strategic ambiguity.”

DeSantis quietly packed both the Florida Board of Medicine and the New College Board of Trustees with ideological fellow travelers to bend institutions to their will. The Board of Medicine now includes campaign donors, Catholics who substitute the Vatican’s positions for that of professional medical organizations, and proponents of conversion therapy, while the surgeon general of Florida is an anti-vaxxer. 

The strategy is to move the decision-making process out of the public spotlight by giving important decision-making authority to people who can’t be held accountable at the ballot box. (Diane Ravich recently noted that DeSantis had gutted Florida’s open records law–another ploy to keep that “capture of the referees” hidden from public view.)

And that “strategic ambiguity”?

DeSantis has made persistent war on “woke” education–from public schools to state universities.

The DeSantis administration swore up and down that the “Florida Parental Rights in Education Act” (the “Don’t Say Gay” law) was simply there to protect vulnerable young children from being exposed to dangerous or obscene ideas, images, or writing. In reality, it was deliberately vague and overly broad. When schoolteachers and librarians reached out for guidance on what is allowed, they were met with silence by the DeSantis administration. This left them with the choice: Do we remove everything from school libraries, or do we risk the potential legal consequences of annoying his administration?

Back when we had a legitimate Supreme Court, employment of this tactic was repeatedly struck down for creating a “chilling effect,” that violated the First Amendment.

The article lists a truly terrifying number of ways a DeSantis administration could use these strategies in support of a radically Rightwing  culture war agenda. I would encourage you to click through and read it all if I thought this little martinet had a realistic chance to be President, or even the Republican nominee. But since February, when this essay was published, Americans have learned a lot about Mr. DeSantis.

Current polling reflects the public’s response to DeSantis’ repetitive attacks on  “wokeness” and the strategic stupidity of picking a stubborn, petty fight with Disney (Florida’s largest employer and largest tourism draw), among other unforced errors.

And really, no one wants to have a beer with this guy.

Comments