Stand up for the facts!
Our only agenda is to publish the truth so you can be an informed participant in democracy.
We need your help.
I would like to contribute
Matthew Dowd: Historic polling pattern points to Trump
One big question in the 2016 race for the White House is when -- or if -- Donald Trump’s campaign will run out of gas. Trump’s poll numbers continue to climb even as he stumbles on questions about policy, recants his past liberal leanings, and makes comments that draw the ire of women and Hispanics.
Given the conventional wisdom that Trump has so far defied some law of political gravity, eyebrows went up when ABC News analyst and GOP political consultant Matthew Dowd declared that Trump has the earmarks of a winner.
"I think Donald Trump, as of today, is the Republican nominee for president," Dowd said on ABC’s This Week on Sept. 6, 2015. "He leads nationally in every single poll for more than two months. He leads every single state, including favorite-son states like Florida, where he leads Jeb Bush. Jeb Bush is third. And any Republican that has led for two months and led every state has won the GOP nomination."
We wanted to check Dowd’s history. Is it accurate that the Republican who has topped the polls nationally and in states that conduct polls this early goes on to win the nomination?
Dowd told us that the numbers back him up, although he didn’t share any hard data. He said he was thinking of every race since Ronald Reagan ran in 1980.
It’s relatively simple to find national polls going back to the summer of 1979. Finding state polls is dicier. Limited Iowa and New Hampshire polls exist, and even those don’t provide the two consecutive months in every election that Dowd said he had in mind. Dowd told us his list of states included ones such as South Carolina, Florida and Louisiana. Early polling in those states is even spottier, especially during the older elections.
But for the polls we could find, we found there were three years when the pattern Dowd identified roughly held up. Our sources include the Roper Center for Public Opinion Research, Polling Report, Selzer and Company (the Iowa Poll firm), the Granite State Poll, the Odum Institute, 538.com, and political scientist Martin Cohen at James Madison University. To keep in step with where we are in the present election, we looked at polls in the spring and summer in years when no incumbent Republican was running.
Source: Roper Center for Public Opinion Research, Polling Report, Odum Institute, et al.
Three candidates in three elections, Ronald Reagan, Bob Dole and George W. Bush, won the pre-September polling trifecta and went on to win the nomination.
For the record, Trump has led the polls in Iowa, New Hampshire and nationally for the past two months. Plus, he’s in the lead in South Carolina, Florida and other states. These days, there’s more polling. But as you go back in time, you find gaps in the month-to-month record for all states. That lack of monthly state data weakens the support for Dowd’s statement somewhat, but the general point holds up in the early primary states and nationally.
A bevy of caveats
The political scientists we reached don’t put much stock in the predictive power of Dowd’s observation.
Martin Cohen co-authored The Party Decides, an assessment of the modern nomination process. Cohen, and others, see early polls as mainly driven by name recognition, a factor that plays itself out as time goes by. Cohen said the pattern Dowd spotted reflects forces that might be less obvious than polls. Cohen told us that endorsements by party leaders is much more telling.
"Most times the poll leader is the endorsement leader," Cohen said. "In the few cases where the poll leader is not the endorsement leader, the poll leader fails. This occurred most notably in 2008 with Giuliani leading and 2012 when a number of candidates equaled or overtook Romney in the polls but didn't come close to his endorsement totals. I think 2016 will turn out this way with Trump losing to someone who has more elite support."
Political scientist Marc Meredith at the University of Pennsylvania discounts the significance of the polling pattern on statistical grounds. The claim’s power rests on only three elections out of six.
"We just have a very little sample to draw upon, and no sample to draw on in this circumstance when we consider the fact that Trump never won a single election," Meredith said. "So drawing on the experiences of Reagan, Dole, and George W. Bush may not be that useful."
Pollster and political scientist Andrew Smith at the University of New Hampshire Survey Center noted that while Trump might have the lead, his numbers are relatively modest.
"I'd be very cautious about claiming that Trump has any sort of ‘lead’ anywhere," Smith said. "I'm not aware that he has even broken the 40 percent mark."
Lastly, Dennis Goldford, a political scientist at Drake University, said the number of Republican candidates in this cycle makes it harder to draw lessons from the past.
"The larger the field -- if the increased number includes realistic candidacies and not just actual or potential flashes in the pan -- the more difficult it is to predict a likely outcome," Goldford said.
Dowd said that any Republican who has led in the polls for two months running at both the state and national levels had gone on to win the nomination. Dowd did not provide the polling data he used, and the historical record is thin across all the states he had in mind. Going back to the 1980 election, there does not seem to be two consecutive months of polling in the spring and summer even in the states with the most polls, Iowa and New Hampshire. But the record does show the sort of early polling dominance leading to nomination success that Dowd described.
Our principle is that it’s up to the speaker to prove his claim. There’s some substance to the statement, but not enough to fully back it up. We rate this claim Half True.
ABC News, This Week with George Stephanopoulos, Sept. 6, 2015
Pew Research Center, How Reliable Are the Early Presidential Polls?, Feb. 14, 2007
Public Perspective, When Frontrunners Emerge, April 2000
Polling Report, White House 2016: Republican Nomination
Polling Report, Preference for Republican Nominee - 1999
Polling Report, Preference for Republican Nominee - 2007
Polling Report, Preference for Republican Nominee - 2011
Associated Press, Public Opinion Magazine poll, April 10, 1979, via Nexis
United Press International, Poll: Bush leads in New Hampshire and Iowa, April 30, 1987, via Nexis
Associated Press, Dole, Bush Virtually Even In Iowa Poll; Hart Holds Wide Lead, Feb. 9, 1987, via Nexis
United Press International, Dole leads in latest Iowa Poll, Aug. 25, 1995, via Nexis
Associated Press, Bush zooms to big lead in Iowa race, June 28, 1999, via Nexis
The Frontrunner, Clinton, Romney open wide leads in Iowa, Aug, 22, 2007, via Nexis
Associated Press, Iowa poll shows tight Democratic race, Romney ahead on GOP side, Aug. 9, 2007, via Nexis
Associated Press, Poll: Iowa caucus voters favor Edwards, Romney, July 27, 2007, via Nexis
538, Candidates In Donald Trump’s Position Have A Terrible Track Record, Aug. 14, 2015
538, Keep Calm And Ignore The 2016 ‘Game Changers’, Sept. 2, 2015
University of New Hampshire Survey Center, New Hampshire primary
Email interview, Matthew Dowd, ABC news analyst, Sept. 8, 2015
Email interview, Michelle Yeoman, research assistant, Selzer and Company, Iowa Poll archives, Sept. 8, 2015
Email interview, Thu-Mai Lewis Christian, archival research associate, Odum Institute for Research in Social Science, University of North Carolina at Chapel Hill, Sept. 8, 2015
Email interview, Kathleen Weldon, research manager, Roper Center for Public Opinion Research, University of Connecticut, Sept. 8, 2015
Email interview, Martin Cohen, associate professor of political science, James Madison University, Sept. 9, 2015
Email interview, Marc Meredith, associate professor of political science, University of Pennsylvania, Sept. 8, 2015
Email interview, Andrew Smith, associate professor of practice and director, University of New Hampshire Survey Center, Sept. 8, 2015
Email interview, Dennis Goldford, professor of political science, Drake University, Sept. 8, 2015
Email interview, John Sides, associate professor of political science, George Washington University, Sept. 8, 2015
Email interview, Andrew Gelman, professor of statistics and political science, Columbia University, Sept. 8, 2015
Email interview, Dennis Goldford, professor of political science, Drake University
Email interview, David Moore, founding director, University of New Hampshire Survey Center
Read About Our Process
Browse the Truth-O-Meter
More by Jon Greenberg
Matthew Dowd: Historic polling pattern points to Trump
Support independent fact-checking.
Become a member!
In a world of wild talk and fake news, help us stand up for the facts.