Trump, Romney, and Data Denial Syndrome
http://spectator.org/trump-romney-and-data-denial-syndrome/Few supporters of Donald Trump will take kindly to the suggestion that he has anything in common with Mitt Romney. It is nonetheless true that the former suffers from a significant handicap that also hobbled the latter—overzealous partisans who flatly refuse to acknowledge very real challenges facing their candidate. The most obvious example involves an affliction that can be styled, “Data Denial Syndrome.” The primary symptoms of this condition include paranoia about the predispositions of public opinion pollsters and an irrational belief that there is a highly motivated, yet oddly silent, group of voters who support the GOP nominee.
During 2012, Data Denial Syndrome reached pandemic proportions. An astonishing number of normally sensible analysts predicted that Mitt Romney would easily beat the president despite consistent polling data portending a narrow victory for Obama. Most of these prognostications were based on the erroneous belief that pollsters were understating Romney’s support using models that included implausible Democrat turnout estimates. More than a few told us Romney would win in a landslide. Newt Gingrich, for example, predicted that the former Massachusetts governor would bury Obama beneath a “53 percent-plus” share of the popular vote.
Romney received only 47.7 percent, yet Gingrich failed to learn anything from his five-point whiff. This year he’s at it again. Recently, the former House Speaker denounced an ABC News-Washington Post poll as irretrievably biased against Donald Trump. Ironically, that very survey was among the most accurate of the 2012 election cycle, projecting that Obama would win by 3 points (he won by 3.9%). Gingrich was, of course, not the only member of the “ignore the polls, Mitt will win” club. Other members included Ann Coulter, Larry Kudlow, Jeffrey Lord, Dick Morris, John Nolte, as well as radio personalities Rush Limbaugh and Sean Hannity.
Not coincidentally, the folks listed above are now Trump supporters and most are still engaged in magical thinking about the polls and attacking anyone who declines to drink the Kool-Aid. Now, before going any further, I’d like to preemptively dispose of a complaint I sometimes get from Trump supporters who dislike what I write about him. This objection usually goes as follows: Catron does fine on health care, but he falls flat when he strays from his area of expertise. For the record, statistical analysis is a big part of my “day job.” And, at the risk of seeming immodest, I predicted in early 2012 that Romney hadn’t a prayer of beating Obama.
At that time, most of the professional pundit class rejected this view as morbidly pessimistic. James Taranto, for example, diagnosed it in the Wall Street Journal as a kind of “malaise” and added the following unfortunate observation: “We have a higher opinion of Romney than Catron does … we think his current adversity is making him a stronger candidate.” After Romney lost, Taranto explained why he and other pundits were wrong: “This time it was we who turned out to be in a media bubble.” In other words, they were so blinded by the Beltway mindset that they couldn’t see objective data that were obvious to mere working stiffs like yours truly.
And many Romney cum Trump supporters are still trapped in the bubble. Despite the ominous fact that Trump has been ahead of Hillary in only 1 of the last 30 national polls listed on the RealClearPolitics average, his supporters insist that the other 29 surveys should be ignored because some understated his support during the primaries. This is just more magical thinking. In reality, Trump underperformed his polling numbers in more than half of the primary states he won. Moreover, the people who cast votes for Trump in the Republican primaries constitute a much smaller and far less diverse group of voters than make up the general electorate.
Comparing national voter surveys to state primary polls tells one little of value except that some pollsters can get it wrong in an idiosyncratic campaign with more than a dozen contenders for a nomination. Yet Trump’s supporters believe that erroneous primary calls by some analysts means their general election prognostications are worthless. Rush Limbaugh recently used this reasoning to advise his listeners that FiveThirtyEight’s Nate Silver can be disregarded when he says Trump has a 20 percent chance of beating Hillary. Never mind that, in 2012, Silver’s state-by-state predictions were 100 percent accurate. Forget that he got 48 of 50 right in 2008.
Another fantasy the Donald’s supporters rely on to explain away his terrible numbers involves the elusive “shy Trump voter.” Newt Gingrich claims that, if Trump is within 5 points of Hillary Clinton in the polls, he’ll beat her because there are at least that many closet Trump supporters. Here’s how the former House Speaker explains it: “You get much better results for Trump for example in a computerized online poll than a telephone poll because people don’t want to tell the pollster something they think is not socially acceptable.” Gingrich has it exactly backwards. Computerized online surveys have tended to overstate the Donald’s support.
All of which means that, despite the Romney defeat, Data Denial Syndrome still infects many conservatives. Some, however, were cured by that debacle. A notable example is Matthew Continetti, who recently published a good column on why he was wrong in 2012 and that it is unwise to ignore the challenges Trump and the GOP face this year: “Around this time in 2012, I believed Mitt Romney had a good chance of beating President Obama…. I scoured every new poll and piece of economic data for nuggets that confirmed my biases.” In the end, Continetti realized why he was wrong and is determined not to repeat the mistake with Trump.
Unlike Gingrich, Limbaugh, and countless others, Continetti has recovered from Data Denial Syndrome and is returning to the basics: presidential job approval, the unemployment rate, and the generic congressional ballot. These fundamentals do not bode well for Trump. The Donald and his supporters are running out of time to face this reality. The election is 4 months away. Trump’s polling numbers are awful, his campaign’s financial situation is worse, and the political fundamentals are against him. Data Denial Syndrome is deadly, and it is about to kill another GOP presidential campaign if Trump and his supporters don’t take the cure.
http://spectator.org/trump-romney-and-data-denial-syndrome/
