Skip to main content
Menu
How Much Can Really Change Before Iowa Votes

The tremendous buildup to the Iowa caucuses is about to culminate in real voting. We’re less than a week away. But anything can still happen … there’s still plenty of time … right? Sort of. It’s true that even the final Iowa polls are sometimes way off. But it’s also true that every caucus winner since 1980 was either within about 10 percentage points of the leader or showing at least some momentum in the polls by this point.

Right now, only two candidates on both the Democratic and Republican side are even close to the lead. In the Republican race, Donald Trump seems to have some momentum over Ted Cruz, though Cruz remains within striking distance. On the Democratic side, Hillary Clinton continues to hold a small lead over Bernie Sanders, though the race remains as close as it did a week ago.

More Politics

How does that compare to past races? Let’s time travel to a week before the caucuses in each open election since 1980 (we don’t have any polls for this time period for the Democratic caucuses in 1984 and 1992): What did a 10-day polling average show, and how did that average compare to the results?

1980 Republicans: Although the Des Moines Register poll had Ronald Reagan leading with 26 percent, Reagan’s support was down from the 50 percent he held a little over a month before the caucuses. Meanwhile, eventual winner George H.W. Bush had climbed from 1 percent in August 1979 to 14 percent in December to 17 percent one week before the caucuses. He rode this momentum to beat Reagan on caucus night, 32 percent to 30 percent.

1988 Democrats: Every single poll taken had Dick Gephardt leading; he averaged 30 percent and finished with 31 percent. The polls did have Michael Dukakis ahead of Paul Simon by 4 percentage points for second place, though it was Simon who ended up ahead of Dukakis for second by 5 points. Either way, it was not a big jump.

1988 Republicans: The polls pegged the eventual winner, Bob Dole, with a clear lead over the competition. Dole averaged 44 percent and won 37 percent. In one of the largest predictive errors in our data set, second-place finisher Pat Robertson averaged only 9 percent in the polls but ended up with 25 percent thanks to a strong performance among religious conservatives.

1996 Republicans: A University of Iowa survey conducted during this stretch had Bob Dole ahead with 22 percent. He won with 26 percent. The surprise — just like in 1988 — came from the candidate appealing to religious conservatives, this time Pat Buchanan. Buchanan was well back in the polls with only 4 percent, but he finished second with 23 percent. Also finishing strong was Lamar Alexander, who climbed from 6 percent in the polls to 18 percent in the final results.

2000 Democrats: Al Gore led the average of polls over Bill Bradley by 20 percentage points a week before the caucuses. He won by 26 percentage points.

2000 Republicans: George W. Bush was leading Steve Forbes in the polls 45 percent to 22 percent. The actual result was closer: 41 percent to 31 percent; Bush’s margin was cut in half, but it was never in doubt.

2004 Democrats: This was the race that moved the most. John Kerry was in third place in the polls with 16 percent, while Howard Dean was leading the field at 27 percent. John Edwards averaged just 11 percent a week before the caucuses, but that was up from 5 percent in December. The final result: Kerry 37 percent, Edwards 33 percent, Dean 17 percent.

2008 Democrats: Hillary Clinton led Barack Obama in the average of polls, 32 percent to 30 percent, though a number of polls had Obama ahead. He eventually won 35 percent to 30 percent over Clinton. But the biggest surprise was John Edwards, who was averaging just 17 percent and came in second with 31 percent.

2008 Republicans: Mike Huckabee held a clear lead over Mitt Romney, 32 percent to 25 percent. On election night, it was Huckabee with 34 percent to Romney’s 25 percent.

2012 Republicans: You can see Mitt Romney and Rick Santorum gaining momentum during this stretch. While the entire 10-day average had Romney at 20 percent and Santorum at 7 percent, a CNN/Time survey ending seven days before the caucuses gave Romney 25 percent and Santorum 16 percent. The final result: Santorum 24.54 percent to Romney’s 24.51 percent. Third-place finisher Ron Paul almost perfectly matched his support in the polls (22 percent) by ending up with 21 percent.

So what does all this mean? Historical polls have been off by a lot — by enough that it wouldn’t be shocking (or even surprising) if Cruz, who has consistently run second to Trump in recent Iowa polls, prevailed in the Hawkeye State. Here’s what would be shocking: If someone not named Cruz, Trump, Clinton or Sanders carried Iowa. Then again, there have been plenty of shocks in the 2016 campaign already.

Harry Enten is a senior political writer and analyst for FiveThirtyEight.

Filed under , , , , ,

Comments Add Comment