Understanding Political Polls: Is the GOP Underweighted?
Trump Demands CNN Apologize for a Poll Showing Biden in the Lead
On June 10, I noted Trump Demands CNN Apologize for a Poll Showing Biden in the Lead
Trump Hires Own Pollster
For months I have been reading the same thing: There is bias in the polls, many say on purpose.
Are the Polls Missing Republican Voters?
Nate Cohn asks and answers Are the Polls Missing Republican Voters?
There are many reasons the polls might ultimately be wrong in November, as many state polls were four years ago, but there’s no serious evidence that the polls are systematically missing Republican voters. There’s more evidence to the contrary — that the polls represent Republicans just fine, and President Trump still trails.
We know that registered Democrats outnumber registered Republicans in Florida by a few points, even if we don’t know whether those same voters would identify as Democrats or Republicans today.
Pollsters can take advantage of this if they use voter registration files, a large data set of every registered voter. Most private pollsters and a handful of high-quality public ones use voter registration files as the foundation of their polling, and these pollsters can use a respondent’s party registration to help ensure that their sample is representative. They can make sure that 37 percent of their respondents in Florida, for instance, are registered Democrats. In states where voters do not have the opportunity to register with a party, a pollster can use a voter’s primary participation history in a similar way.
Surprisingly, registered Republicans were actually more likely than registered Democrats to respond to the Times/Siena survey. This seemingly noteworthy difference can be explained by well-known demographic biases in polling: Older, rural and white voters are likelier than young, urban and nonwhite voters to respond to surveys. After these factors were controlled for, Republicans were no likelier than Democrats to respond to the survey.
Mr. Trump’s problem wasn’t the number of people who said they voted for him last time: It was that only 86 percent of those who said they voted for him last time said they would do so again.
Why Don’t Polls Include the Same Number of Republicans and Democrats?
Nate Cohn provided one look above, here is another from Pew Research.
Newcomers to polling sometimes assume that if you are asking Americans questions about politics, it’s only fair to include an equal number of Republicans and Democrats. While this notion makes some sense on the surface, it’s based on a misunderstanding of what polling is intended to do. The goal of a national political survey isn’t to artificially even the playing field. It’s to represent groups in their actual proportions within the country. And a wide range of evidence shows that there are more Democrats than Republicans in the United States today.
In the 31 states that register voters by party, the number of Americans registered as Democrats outnumbered those registered as Republicans by nearly 12 million in mid-2018. Registration with a party is an imperfect measure of whether a person currently identifies with that party, especially in the South. But nationwide comparisons find that registration is closely associated with self-identification, and the two appear to be growing more correlated over time.
In addition, commercial voter files – which attempt to predict the partisanship of voters in all 50 states and the District of Columbia based on information such as address, race, age and primary vote history – indicate that there are more Democratic registered voters than Republican registered voters in the country today.
So what does all this mean for consumers of political polling? In general, poll watchers today should expect to see more Democrats than Republicans in a national survey, particularly one designed to reflect the views of U.S. adults (as opposed to registered or “likely” voters). These partisan breakdowns don’t necessarily favor one side or the other; rather, they reflect the population in question as accurately as possible.
Getting it Totally, Yet Typically Wrong
There are hundreds more similar Tweets.
The subject of weighting came up today in a Twitter Thread by Nate Cohn.
- You can't just reweight a poll by changing the proportions in the crosstabs, since that can throw off other variables. You've got to weight on everything at once.
- As an example, many noticed our AZ poll had like Trump>10 on '16 recalled vote. But you can't just unskew it 5 points to the left. Our poll, as it was, had the right number of reg Ds/Rs; unskewing on the '16 vote tab would have meant giving Dems a reg advantage in AZ
- If you just give more weight to the GOP in the sample, you just made your sample whiter than it should be, or maybe now you have too few young voters, or too few people outside the South, and so on.
Finally, there are initial polls and weighted polls. It is not impossible to get the correct number of respondents by age, sex, education, and political affiliations.
That is why pollsters release weighted and unweighted responses.
Pollsters Not Purposely Biased
The independent pollsters try to get it right, including those for Fox News, rated A- by Nate Silver at FiveThirtyEight. By contrast McLaughlin is rated C/D.
Beware of party sponsored polls which have one purpose only. They are not unbiased.
Were the Polls Wrong in 2016?
They were not great but almost all of the were within margins of error.
The Midwest state polls did make one mistake in regards to education. Poll samples likely overweighted educated voters.
The turnout had a large number of white voters without a college degree. Trump appealed to those voters and won enough of them to flip key states.
Trump had a huge surge at the last moment due to a blockbuster announcement by FBI director James Comey.
FiveThirtyEight comments The Comey Letter Probably Cost Clinton The Election , So why won’t the media admit as much?
Trump won Michigan, Pennsylvania, and Wisconsin by less than 1 percentage point.
The Comey effect was not reflected in the polls at the state level but it swung the election.
Had Trump not won those states, we would not have all this bickering today about the polls being wrong.
Trump is on the Short End of the Enthusiasm Gap
Another misconception is that Trump supporters are more enthusiastic than Biden supporters.
On the surface, that is true, but only if one looks at positive enthusiasm.
If you take negative enthusiasm into account Trump is on the Short End of the Enthusiasm Gap.
Not a 2016 Replay
For many months, I have commented this is not a replay of 2016. Trump was never well liked, but many people in both parties despised Hillary.
Trump won in 2016 for five reasons.
- People hated Hillary more than Trump
- Trump was an unknown risk that swing voters were willing to take a chance on
- Hillary ran a very poor campaign
- A fantastic campaign slogan: Make America Great Again
- At the last minute, Comey rescued trump with an attack on Hillary and her email server.
Points 1-4 got Trump into the ballpark. It still took Point 5 to cross the finish line.
Despite all of those things Trump barely won as measured by tiny margins of victory in several critical swing states.
The Man People Hate to Hate
Biden is a man people hate to hate. Trump and Hillary are candidates people love to hate.
There is no enthusiasm gap for Trump, except in the negative sense.
Many people despised Hillary, and rightfully so.
The election is now less than three months away. It is not "early" but nor is it "too late".
Trump could still win but it is an extremely uphill battle at this points. Polls will likely narrow, just not by enough.