Democratic Women Ponder Polling Process

Hillary Clinton, Donald Trump, you, me — and anyone else in the world who reads polls — all were surprised by last week’s election.

That includes James Delorey. The Westport resident is senior vice president of research at the Global Strategy Group. They use tons of sophisticated tools to advise Fortune 500 companies, non-profits — and political campaigns. In other words: He analyzes data and trends for a living.

James Delorey

James Delorey

Like everyone else in the business, he’s moved past shock. Now he’s trying to figure out how so much polling could have been so wrong for so long.

Tomorrow night (Thursday, November 17, 7:30 p.m., 323 Restaurant) he shares his insights in a local forum.

“Why Data Failed Us: A Closer Look at Polling” is sponsored by the Democratic Women of Westport. Clearly, that’s a group groping for answers.

Delorey will answer questions like:

  • Why did so much 2016 polling data miss the mark?
  • Is it possible to accurately poll the electorate in the current era?
  • Did media coverage of the polls impact the election outcome?
  • What was the impact of late-breaking news?
  • What issues drove the election outcome?
  • Could Clinton have done anything differently to ensure victory?

Delorey — who worked at the state level during this election cycle, and helped elect a Democratic governor in West Virginia, while the state went 69% for Trump — was not one of those presidential pollsters who so misread the presidential tea leaves.

But he knows he, his colleagues — and his industry — are in hot water right now.

7 responses to “Democratic Women Ponder Polling Process

  1. Unfortunately can’t make it tomorrow night but would love to read a recap

  2. Not sure the polling error was so large (same case with Brexit). “Shy Trump” voters, “non-response bias”, or simply non-detectable enthusiasm are perhaps a few reasons for a surprise.
    The Economist Magazine writes “there is one family of forecasts that did better: those which ignore both polls and candidates and predict results based exclusively on structural factors like economic performance and incumbency.”

    • a large part of the issue with the polls this year was the “likely voter screen” where pollsters weight the responses they get by how likely that person is to actually go and vote. Like him or not, Trump got people to vote, who have never voted before- people who were probably not included in the “likely voter” category because they may not have been previously registered, or didn’t vote 4 years ago.

      What I found interesting was this map:
      There are far more people who did not vote at all, than who voted for either Trump or Clinton.

      I hope everyone thinks about this for the next 2 and 4 years, and that the Democratic party can come up with ways to speak to these people who did not vote. Did they want to vote but were unable to due to voter ID laws, time/location restrictions, apathy towards the choices, etc? The more people who get engaged in the process, the better our country can be, but we ALL need to listen to these people and see what all political parties can do better.

  3. Robert Bernard

    How was the LATimes/USC poll able to predict that Trump would win by at least 2 points consistently through the whole campaign? That poll used a different question set than most other pollsters.

  4. Peter Gambaccini

    I was actually saying to a friend just a few days before the election, “you know, there’s no law in this country that says you have to tell pollsters the truth.”

  5. The inherent problem with polls, which I’ve never heard anyone discuss, is that they don’t take into account how people have reached a conclusion, e.g. based on emotion or logic. It’s my observation that logical conclusions tend to stick while emotional ones do not. It’s also my observation that people are far more emotional in their actions than logical.

    For example, you could go into a store expecting to buy a certain chair, trip over it and stub your toe, and get so mad you buy a different one. Or you could insist on going to your favorite seafood restaurant for dinner then, once you get there, decide to order steak because, hey, you just feel like eating steak now instead. On the contrary, if you’ve determined you need a particular size screw at the hardware store, the odds are pretty good you aren’t going to change your mind.

    Insofar as the election, I don’t think the polls missed pockets of voters, or that voters were afraid to admit who they really supported. I think the polls didn’t take into account how malleable people generally are, or how committed someone was to actually vote– both emotional variables. I’d love to see a study where pollsters went back to the exact same people they polled and asked how they really voted, how committed they thought they were at the time of the poll, and, if they changed their minds, why.