New citizen forecast predicts tight race: Clinton 50.5% vs Trump 49.5%

A new ABC News/Washington Post poll provided updated citizen forecasts. In the polls, people were asked: “Regardless of whom you support, if Clinton and Trump are the nominees for president, who would you expect to win: Clinton or Trump?” According to the results, 50% of respondents expected Clinton to win, whereas 40% predicted a Trump victory. Translating these results into two-party vote-shares yield a forecast of 50.5% for Clinton and 49.5% for the Trump.

In comparison, the combined PollyVote currently predicts Clinton to gain 52.0% of the popular two-party vote.

Trump slightly favored in updated DeSart model

Jay DeSart updated his long-range model with April polling data. The most recent forecast slightly favors Trump in a likely match-up against Clinton. In particular, the DeSart model predicts Trump to gain 50.9% of the vote, compared to 49.1% for Clinton. Trump is also predicted to win the electoral college with 295 votes, compared to 243 for Clinton.

The vote-share forecast currently enters the combined PollyVote with a weight of 3%. In comparison, today’s PollyVote forecast predicts the Democrats to gain 53.3% of the vote.

Democrats lead by 6 points in latest expert survey

The Pollyvote team has completed its fifth survey of elections experts to forecast the 2016 presidential election. In the late April survey, conducted between April 27 and 30, 15 academics from a variety of colleges and universities responded.

The mean forecast is that the Democrats will garner 53.2% of the major-party vote (compared to 46.8% for the Republicans), with individual forecasts ranging from a minimum of 49.5% to a maximum of 55.2%. Only one expert expected the Republicans to win a majority of the vote.

The predicted Democratic vote share is thus slightly (0.1 percentage points) higher than in the previous survey, conducted in late March, when the mean expert experts had predicted a vote share of 53.1% for the Democrats and 46.9% for the Republicans.

Polly thanks the experts who participated in this round, namely

and one expert who preferred to remain anonymous.

Experts see Democrats’ lead widen

The Pollyvote team has completed its fourth survey of elections experts to forecast the 2016 presidential election. In the late March survey, conducted between March 28 and 31, 14 academics from a variety of colleges and universities responded.

For the first time in this election cycle, all responding experts expected a Democratic win, with forecasts of the popular vote ranging from a minimum of 50.8% to a maximum of 56.8%. The mean forecast is that the Democrats will garner 53.1% of the major-party vote (compared to 46.9% for the Republicans).

The predicted Democratic vote share is thus more than one percentage point higher than in the previous survey, conducted in late February, when the mean expert experts had predicted a vote share of 52.0% for the Democrats and 48.0% for the Republicans.

Polly thanks the experts who participated in this round, namely

and one expert who preferred to remain anonymous.

Issue models turn in Democrats’ favor

After months of no issue polls, an ABC News/Washington Post survey finally provides information about how people rate the candidates’ ability to handle the issues. In particular, the survey asked participants whom of either Clinton or Trump they would expect to do a better job in handling four issues: terrorism (which is currently seen as the most important issue), the economy, immigration, and a potential international crisis.

Participants strongly favored Clinton on all four issues (see Figure). In comparison, prior surveys that asked people to compare Clinton to a generic Republican candidate tended to favor the Republicans. In addition to the bio-index model, this provides further support for the notion that Trump would likely be a poor choice as a nominee for the Republican Party.

The results from the issue questions are used in two models, namely the issue-index and the big-issue models. While, prior to the release of the survey, both models predicted the Republicans to win the popular vote, their forecasts have turned. The issue-index model predicts the Democrats to gain 52.0% of the vote. The big-issue also favors the Democrats, despite predicting a closer race, at 50.6%.

 

Daily Pennsylvanian covers the PollyVote

The Daily Pennsylvanian talked to Scott Armstrong about the Pollyvote. Here are some of Scott’s key points.

On the motivation to launch the PollyVote in 2004

“What we were trying to do is to demonstrate to the world that if you use what we call ‘evidence based principles forecasting,’ you can improve just about any forecast in the world,” Armstrong said. He chose the presidential election for the purpose of gaining attention from the press and the general public.

On the benefits of combining forecasts:

“Most people think if you look at a forecast, you should try to pick the best one, but that’s false,” Armstrong noted. “In fact, if you use a combined forecast, you can very often, very often, do better than the best component in that combination.”

On the bio-index model‘s poor scoring for Donald Trump:

“We’ve never had anybody so low, it’s astonishing,” Armstrong said. “That should be an indication to the party that they are going to have a real problem if they select Trump.”

PollyVote at APSA 2016

The project team will present the PollyVote at the 2016 Annual Meeting of the American Political Science Association (APSA), which will be held from September 1-4 in Philadelphia, PA. The title of the talk, which is part of a session on Methodological Innovations in Election Forecasting, is Combining forecasts for US presidential elections. PollyVote’s advances since 2004.

In addition, Andreas Graefe and Scott Armstrong will join the Roundtable on Forecasting the 2016 US Presidential & Congressional Elections to give a talk on Forecasting the presidential election based on candidates’ biographies and perceived issue-handling competence. This talk will feature the predictions from the bio-index, the issue-index, and the big-issue model, among others.

Stay updated on the program schedule:

Armstrong challenges Polly

Scott Armstrong bets that he found a method that will help to further increase the accuracy of the combined PollyVote forecast.

In particular, Scott proposes that adding a naive model will improve accuracy when uncertainty is high, such when the election is still far away and when the parties’  eventual nominees are yet to be known. This suggestion is based on the Golden Rule of Forecasting, which recommends considering a naive model if the situation involves uncertainty.

In operational terms, Scott suggests the following:

  • Until the candidates from both parties are nominated on July 28, 2016: add the naive model, which predicts a vote share of 50% for each party, as a seventh component to the PollyVote.
  • From July 29th: move on with the regular PollyVote.

The chart above shows how the forecasts of the original PollyVote and Armstrong’s version that includes the naive model have developed from January until July 28th (after that, the forecasts are of course similar).

We will keep you posted on whether Scott is able to beat Polly for long-term forecasts! Stay updated:

First citizen forecasts available: PollyVote now with forecasts from all six components

Polly added citizen forecasts as the sixth and final component method to its combined forecast. Citizen forecasts, also known as vote expectation surveys, are among the most accurate methods available for forecasting elections.

Citizen forecasts are derived from surveys that include the vote expectation question, which asks respondents who they expect to become president. For example, an  ABC News/Washington Post poll, published this week, asked: “Regardless of whom you support, if Clinton and Trump are the nominees for president, who would you expect to win: Clinton or Trump?” According to the results, 59% of respondents expected Clinton to win, whereas 36% predicted a Trump victory. Translating these results into two-party vote-shares yield a forecast of 51.6% for the Democrats and 48.4% for the Republicans. This is the information Polly uses for her the citizen forecasts component.

Adding the citizen forecasts had little immediate impact on the PollyVote, which currently predicts the Democrats to gain 52.5% of the popular two-party vote.

Compute your own bio-index model forecast

You can now compute your own forecasts based on the bio-index model, which predicts the election outcome based on information about candidates’ biographies. This feature allows you

  1. See how the model forecast would change for different variable coding.
  2. Compare two hypothetical candidates against each other.
  3. See how you would perform as a candidate.

To do so, go to the bio-index model page, scroll all the way down to Table 1, and adjust the variable values in the green highlighted cells. Whenever you change a variable value, the forecast will be updated at the bottom of the table.