Experts predict Clinton will win popular vote by 4 points

The PollyVote team has completed its 10th survey of elections experts to forecast the 2016 presidential election. In this survey, conducted between September 28 and 30, and thus after the first presidential debate, 13 academics from a variety of colleges and universities responded.

Similar to the previous round, all respondents expect a Clinton win. However, her lead has narrowed substantially.

Whereas, in late August, the experts expected that Clinton will win the popular vote by 7 points, the new average forecast is 1.3 percentage points lower, at 52.2% of the two-party vote (or about a 4-point margin). The individual forecasts ranged from 51.6% to 53.7%, with a standard deviation of only 0.7 points.

 

Polly thanks the experts who participated in this round, namely

  1. Randall Adkins (University of Nebraska Omaha)
  2. Lonna Rae Atkeson (University of New Mexico)
  3. Keith Gaddie (University of Oklahoma)
  4. John Geer (Vanderbilt University)
  5. Sandy Maisel (Colby College)
  6. Michael Martinez (University of Florida)
  7. Thomas Patterson (Harvard University)
  8. Gerald Pomper (Rutgers University)
  9. David Redlawsk (University of Delaware)
  10. Larry Sabato (University of Virginia)
  11. Michael Tesler (University of California, Irvine)
  12. Charles Walcott (Virginia Tech)

and one expert who preferred to remain anonymous.

Survey of political scientists: Clinton will win 347 electoral votes

The PollyVote team has conducted the second round of its state-level expert survey. According to the experts’ judgment, Hillary Clinton will win 347 electoral votes, compared to 191 for Donald Trump.

This Electoral College forecast is similar to the previous survey’s results, although Clinton’s lead has considerably narrowed in a number of states, particularly Florida, North Carolina, Iowa, Ohio, and Nevada. That said, even if Clinton lost all five states, she would still end up with 273 electoral votes, three more than necessary to win the election.

The following map visualizes the experts’ median estimates regarding Clinton’s chance of winning each state.

In comparison, the latest PollyVote forecast, which also incorporates the results from the expert survey, predicts Clinton to gain 314 electoral votes, compared to 224 for Trump. This difference results from different forecasts for Ohio and North Carolina. While the experts predict Clinton to prevail in these states, the combined PollyVote forecasts currently favors Trump.

Method

We reached out to political scientists across the country and asked them two short questions:

  1. What share of the vote do you expect the nominees to receive in your home state?
  2. What do you think is Hillary Clinton’s chance of winning the election in your home state?

The survey was conducted from September 19 to 22. A total of 653 experts made estimates as requested. The number of experts by state ranged from 2 to 35. The table at the end of this post shows the number of respondents per state as well as the median answer for each question.

Clinton’s predicted
State N Chance of winning Two-party vote
District of Columbia 13 100% 94.6
Vermont 9 100% 66.7
Hawaii 2 100% 69.9
New York 25 99% 64.2
Massachusetts 20 99% 66.7
Maryland 26 99% 63.6
California 23 99% 62.2
Rhode Island 8 97% 62.7
Washington 11 96% 59.1
Illinois 25 95% 60.4
Delaware 11 95% 60.6
Connecticut 21 95% 60.3
Oregon 11 93% 64.0
New Jersey 5 90% 57.4
New Mexico 4 87% 59.6
Minnesota 17 85% 57.4
Colorado 9 80% 57.8
Wisconsin 14 78% 55.6
Virginia 35 75% 56.0
Pennsylvania 15 75% 56.9
Maine 12 75% 54.8
Michigan 14 70% 54.7
New Hampshire 16 64% 55.1
Florida 16 54% 52.7
North Carolina 22 54% 51.9
Nevada 5 52% 51.2
Ohio 9 51% 52.7
Iowa 15 51% 52.8
Alaska 3 45% 43.4
Arizona 19 40% 48.4
Missouri 11 40% 48.9
Georgia 20 35% 48.6
Arkansas 10 30% 44.2
Indiana 14 18% 48.6
Kansas 8 13% 46.1
Texas 24 13% 44.0
South Carolina 13 10% 45.2
North Dakota 5 10% 43.0
Utah 18 10% 43.6
Kentucky 4 10% 44.4
Wyoming 5 8% 44.0
Alabama 7 5% 40.4
West Virginia 8 5% 37.8
Montana 5 5% 47.7
Mississippi 15 5% 44.1
Louisiana 12 5% 43.2
Tennessee 12 3% 42.1
South Dakota 4 3% 42.2
Nebraska 3 2% 45.5
Idaho 11 1% 34.5
Oklahoma 9 1% 31.7

Bio-index model update: Clinton now at 58.3%

Variable 36 in Trump’s bio-index was coded incorrectly. Trump’s brother Freddy died from alcoholism in 1981, at age 41. Due to the variable recoding, the forecast has changed.

The updated vote-share forecasts are 58.3% for Clinton and 41.7% for Trump. This is the highest forecast for Clinton of all components aggregated in the PollyVote.

Clinton with 7-point lead in new expert survey

The PollyVote team has completed its 9th survey of elections experts to forecast the 2016 presidential election. In this survey, conducted between August 29 and 31, 13 academics from a variety of colleges and universities responded.

All respondents expected a Democratic win. The average forecast is that Democratic nominee Hillary Clinton will receive 53.5% of the two-party vote, with individual forecasts ranging from 52.1% to 55.8%. This value is the same as yesterday’s PollyVote prediction and 0.8 percentage points higher than the 52.7% for Clinton in the previous expert survey.

Polly thanks the experts who participated in this round, namely

  1. Randall Adkins (University of Nebraska Omaha)
  2. Lonna Rae Atkeson (University of New Mexico)
  3. John Coleman (University of Minnesota)
  4. George Edwards (Texas A&M University)
  5. John Geer (Vanderbilt University)
  6. Sandy Maisel (Colby College)
  7. Michael Martinez (University of Florida)
  8. Thomas Patterson (Harvard University)
  9. David Redlawsk (University of Delaware)
  10. Larry Sabato (University of Virginia)
  11. Michael Tesler (University of California, Irvine)
  12. Charles Walcott (Virginia Tech)

and one expert who preferred to remain anonymous.

Get better election predictions by combining diverse forecasts

Imagine you are among the legions of pundits and political commentators striving to predict the outcome of November’s presidential election. You’re not just interested in who will win – most citizens can predict that, it turns out. You want to forecast the candidates’ exact vote shares.

You would likely seek insight from the results of political opinion polling. Perhaps you would also look at forecasts from other methods such as betting markets, statistical models or judgments of fellow pundits. But there is a problem: Different methods can offer different forecasts, and it’s hard to determine which one will turn out to be most accurate.

It’s not a good idea to simply rely on methods that have proved reliable in the past. My research has found that forecasting models that were among the most accurate in one election tended to be among the least accurate in the next. But my research has also identified a way to make predictions much better.

Every election is different

One reason that past results aren’t good evaluations of prediction methods is that every election is held in a different context and has its own idiosyncrasies – such as the first woman major-party nominee running against the first reality-TV star nominee. These anomalies are particularly challenging for statistical models, which try to base their predictions on patterns in electoral history.

Another reason is that the conditions under which certain methods are expected to work can change over time. For example, response rates in traditional phone surveys have dropped below 10 percent in recent years. This makes it harder to believe that respondents form a random and representative sample of the population, and raises concerns about whether traditional polls can still meaningfully measure public opinion.

So how should you make your best prediction? Advice from half a century of evidence from forecasting research is clear: Combine forecasts from different methods that rely on different information. The combined forecast is usually more accurate than any single prediction. It’s also often more accurate than even the most on-target individual forecast. And combining distinct predictions avoids the risk of making large errors.

When forecasters can use more information in an objective way, their predictions get better. In individual forecasts, there is always some amount of bias that creeps in, because of what data were used or excluded, and the methods used to analyze them. But when various methods using different data are combined to make a forecast, those biases tend to cancel each other out.

Picking what to include

You, the pundit, now know what to do. But many questions remain. For instance, where can you find different forecasts? And which ones should you trust and include in your combination? And how should you weight the different forecasts?

The good news is that you do not have to make these decisions yourself. In 2004, we developed the PollyVote, an evidence-based formula designed to forecast election outcomes by combining multiple predictions. In particular, the PollyVote system combines results from six different forecasting methods that use various kinds of information: polls, betting markets, expert judgment, citizen forecasts, index models and econometric models.

When combining these different results into the PollyVote forecast, we use a two-step procedure. First, we calculate a combined forecast for each of the six component methods. For example, the PollyVote currently averages results from eight different poll aggregators into one combined polling projection.

In the second step, we average all the combined component forecasts to calculate the final PollyVote prediction. This equalizes the significance of each component method, whether an element includes many forecasts, like polls and statistical models, or only a few, like the lone prediction market dealing with the national popular vote.

The use of equal weights in combining forecasts is supported by a large body of evidence, including my own research, which shows that the simple average often provides more accurate forecasts than complex approaches to estimating “optimal” combining procedures.

If we knew that a particular method was likely to be most accurate, we could give it more weight when calculating the combined forecast. But again, because the accuracy of each forecast changes over time, it is difficult to know which is best at any given moment. So the safest approach – not to mention the simplest and easiest to understand – is to treat them all as equally likely.

Past performance

Since we launched the PollyVote, it has provided accurate forecasts over the last three presidential elections. On average, across three periods in the election cycle (Election Day eve, one month before and three months before), the average forecast error is less than one percentage point. As far as we know, this record is unsurpassed by any other forecasting formula.

2016 forecasts

In early January, we launched the PollyVote for the 2016 election. Since then, the method always predicted a Clinton victory; since mid-February it has Clinton winning by at least four percentage points. As of this writing, the PollyVote predicts Clinton to gain 53.5 percent of the major party vote, which excludes those voting for third-party candidates. Trump is predicted to get 46.5 percent.

Five of the six component methods included in the PollyVote predict a Clinton victory. The only exception is econometric models, which predict the election outcome based on what are called “fundamentals.” These include the state of the economy, the sitting president’s popularity and the amount of time the incumbent party has been in the White House. But they do not capture candidate characteristics. In predicting a narrow Republican win (50.9 percent for Trump), the econometric models component essentially suggests that the Republicans should have an advantage had they nominated an “average” candidate.

The index models component, on the other hand, includes forecasts that specifically focus on candidate characteristics – such as their prior experience, their leadership skills or their issue-handling competence – and thus incorporate information about the candidates themselves. On average, the index models predict Clinton to win 53.7 percent of the major-party vote. One of these models, which looks at the candidates’ biographies, even has her at 58.8 percent.

This year, for the first time ever, the PollyVote also looks at state predictions to forecast the outcome of the Electoral College. At the time of writing, Clinton is predicted to gain 347 electoral votes, versus 191 for Trump.

Apart from providing accurate forecasts, the PollyVote has another important benefit: education. By collecting and aggregating forecasts from different methods, the platform allows readers to learn about different election forecasting methods and to compare their results. We encourage anyone interested in learning more about the PollyVote to visit www.pollyvote.com, where the election forecast is updated daily and the complete data are publicly available.The Conversation

This article was originally published on The Conversation. Read the original article.

Political scientists predict Clinton will win 347 electoral votes

Asking experts to predict what is going to happen is probably one of the oldest forecasting methods available. When it comes to predicting election outcomes, we expect experts to have experience in reading and interpreting polls, putting them in historical context, and estimating the likely effects of upcoming campaign events. While the judgment of a single expert should be treated with caution, combining the opinions of several experts improves accuracy.

We have conducted expert surveys as part of the PollyVote project since its launch in 2004. In these surveys, we asked experts to predict the national popular vote. As shown in the chart below, their forecasts have always added valuable information and thus contributed to the accuracy of the PollyVote.  On average, adding expert judgment to the combined PollyVote reduced forecast error by 15%.

For this year’s election, we launched our first survey in December 2015. Since then, the experts’ average forecast always had Clinton in the lead, with a predicted share ranging from 52.2% to 55.5% of the major party vote, which excludes votes for third-party candidates. The latest survey conducted in late July predicted Clinton to gain 52.7%, compared to 47.3% for Trump.

Now, for the first time ever, we set out to conduct a state-level expert survey in order to predict the electoral college. For this, we reached out to political scientists across the country and asked for their help. After respondents revealed their home state, they had to answer two short questions:

  1. What share of the vote do you expect the nominees to receive in your home state?
  2. What do you think is Hillary Clinton’s chance of winning the election in your home state?

A total of 678 experts made estimates as requested. The number of experts by state ranged from one to 42. The table at the end of this post shows the number of respondents per state as well as the median answer for each question. We aim to recruit more experts in states with seven or fewer experts in future survey rounds.

Expert surveys don’t have to include many people to produce accurate forecasts. Prior research shows that eight to twelve experts are close to the optimum. For example, Polly’s survey of the popular vote has provided accurate forecasts with a sample of about twelve experts.

The map below shows Clinton’s chance of winning per state. This expert survey predicts that Hillary Clinton will win 347 electoral votes, compared to 191 for Donald Trump. This forecast is in line with those from Daily Kos, NYT Upshot, PredictWise, Sabato, and the combined PollyVote.

Experts’ Electoral College forecast

Overview of results

Clinton’s predicted
State N Chance of winning Two-party vote
District of Columbia 18 100% 94.6
Vermont 11 99% 66.7
New York 24 99% 64.2
Massachusetts 11 99% 66.7
Maryland 26 99% 63.6
California 24 99% 62.2
Illinois 27 98% 60.4
Washington 11 97% 59.1
Rhode Island 8 97% 62.7
Minnesota 16 95% 57.4
Delaware 7 95% 60.6
Connecticut 22 95% 60.3
New Jersey 4 93% 57.4
Wisconsin 14 90% 55.6
Oregon 11 90% 64.0
New Mexico 6 90% 59.6
Virginia 42 87% 56.0
Colorado 10 86% 57.8
Michigan 8 85% 54.7
Pennsylvania 18 84% 56.9
Maine 16 83% 54.8
New Hampshire 11 80% 55.1
Ohio 9 75% 52.7
Hawaii 1 75% 69.9
Florida 21 75% 52.7
North Carolina 22 63% 51.9
Nevada 1 57% 51.2
Iowa 23 57% 52.8
Georgia 34 45% 48.6
Arizona 21 43% 48.4
Missouri 9 40% 48.9
Indiana 15 40% 48.6
Wyoming 2 30% 44.0
Kansas 11 30% 46.1
South Carolina 15 25% 45.2
North Dakota 3 25% 43.0
Utah 12 23% 43.6
Arkansas 15 20% 44.2
Alabama 6 18% 40.4
West Virginia 8 15% 37.8
Nebraska 10 15% 45.5
Montana 7 15% 47.7
Kentucky 7 15% 44.4
Texas 21 10% 44.0
Tennessee 11 10% 42.1
Mississippi 8 10% 44.1
Louisiana 12 8% 43.2
South Dakota 3 5% 42.2
Oklahoma 12 4% 31.7
Idaho 11 3% 34.5
Alaska 3 0% 43.4

Clinton leads by more than 5 points in new expert survey

The Pollyvote team has completed its 8th survey of elections experts to forecast the 2016 presidential election. In the late July survey, conducted between July 29 and 31, 12 academics from a variety of colleges and universities responded.

All but one responding experts expected a Democratic win. The average forecast is that Democratic nominee Hillary Clinton will receive 52.7% of the two-party vote, with individual forecasts ranging from 48.5% to 55.4%. The expert forecast is thus 1.5 percentage points lower than in the previous survey, when the mean forecast was that Clinton will receive 54.2%.

Polly thanks the experts who participated in this round, namely

  1. Randall Adkins (University of Nebraska Omaha)
  2. Lonna Rae Atkeson (University of New Mexico)
  3. John Coleman (University of Minnesota)
  4. George Edwards (Texas A&M University)
  5. Sandy Maisel (Colby College)
  6. Michael Martinez (University of Florida)
  7. Thomas Patterson (Harvard University)
  8. Gerald Pomper (Rutgers University)
  9. David Redlawsk (Rutgers University)
  10. Larry Sabato (University of Virginia)
  11. Michael Tesler (University of California, Irvine)
  12. Charles Walcott (Virginia Tech)

New high for Democrats in PollyVote expert survey

The Pollyvote team has completed its seventh survey of elections experts to forecast the 2016 presidential election. In the late June survey, conducted between June 28 and 30, 12 academics from a variety of colleges and universities responded.

All responding experts expected a Democratic win, with forecasts of the popular vote ranging from a minimum of 52.0% to a maximum of 57.0%. The mean forecast is that the Democrats will garner 54.2% of the major-party vote (compared to 45.8% for the Republicans).

The predicted Democratic vote share is thus 1.8 percentage points higher than in the previous survey, conducted in late May, when the mean expert experts had predicted a vote share of 52.4%. The current results also show the highest vote share for the Democrats in all expert surveys conducted since January.

Polly thanks the experts who participated in this round, namely

and one expert who preferred to remain anonymous.

Experts see Democrats’ lead narrowing to 5 points

The Pollyvote team has completed its sixth survey of elections experts to forecast the 2016 presidential election. In the late May survey, conducted between May 29 and 31, 13 academics from a variety of colleges and universities responded.

All but one responding experts expected a Democratic win. The average expert forecast for the Democratic share of the two-party vote was 52.4%, compared to 47.6% for the Republicans.

The predicted Democratic vote share is thus 0.8 percentage points lower than in the previous survey, conducted in late April, when the mean expert experts had predicted a vote share of 53.2%.

Polly thanks the experts who participated in this round, namely

and one expert who preferred to remain anonymous.