After more than four years of waiting, Iowa caucus-goers finally get their moment in the spotlight Monday night. For the past few weeks, I’ve been looking at past early polls one, two, and more weeks before the Iowa caucus to see if they were any good at predicting who would win in Iowa, New Hampshire and the eventual nomination. The answer: generally not so much. Let’s run through what the polls predicted for those three contests on Iowa caucus day in 2012 and 2008.
2012, Republican national polls: On Jan. 3, 2012, Newt Gingrich led national polls by about two percentage points over Mitt Romney. Despite his national lead going into Iowa, Gingrich finished fourth there and eventually won only two primary contests. Gingrich lost the polling lead nationally the day after Iowa but retook the lead briefly in the last week of January. Mitt Romney also lost the national polling lead to Rick Santorum, but took the lead for good in late February, just before Super Tuesday.
2012, Iowa GOP caucus: The final Iowa caucus polling average predicted a Romney victory, with 23 percent of the vote. Romney did finish within two points of that prediction, but a surging Rick Santorum got the victory. Santorum never led a single Iowa caucus poll, but the polls did show him surging and predicted he would finish third, with about 16 percent. Santorum finished eight points higher than predicted. Besides Santorum, the final polls were all within two percentage points of predicting how much support the other six candidates would get.
2012, New Hampshire GOP Primary: At this point in the 2012 campaign, polls predicted Romney would win New Hampshire in a landslide. Lo and behold, the polls were right. About a week before New Hampshire was due to vote, polling averages were within five or six percentage points of predicting how much support each candidate would get. In a close race, that might make a difference. With Romney predicted to win by 20 points, it didn’t. He led every New Hampshire poll taken that cycle.
2008, GOP national polls: On Jan. 3, 2008 Rudy Giuliani still led national GOP polling, as he had for all of 2007. Giuliani did not win a single primary or caucus in 2008. At that point, John McCain had finally risen into a virtual tie for second place with Mike Huckabee. McCain took the lead a few days after New Hampshire and held it for the rest of the primary. Giuliani dropped out before Super Tuesday less than a month later.
2008, Iowa GOP caucus: The final Iowa caucus polling average in 2008 predicted a narrow Huckabee victory, by three percentage points. Huckabee did win, but with five points more support than predicted, leading to a nine-point victory over Romney. Besides Huckabee, all the other candidates finished within three percentage points of what polls predicted.
2008, New Hampshire GOP Primary: In 2008, the New Hampshire primary was held only five days after Iowa Caucus day. This year, New Hampshire is eight days after Iowa. The day before Iowa, McCain had just taken the lead in New Hampshire from Romney, who had led every New Hampshire poll for five months. McCain was predicted to win by a couple points over Romney, but eventually won by 5.5 percentage points. Every candidate except McCain finished within two or three percentage points of what the polls predicted on Iowa caucus day.
2008, Democratic national polls: On Jan. 3, 2012, Hillary Clinton led national Democratic polling by an astounding 21 percentage points. She continued to lead national polls into February, well after early states like Iowa, New Hampshire, Nevada and South Carolina had voted. Despite her massive early lead in national polls, Clinton lost the Iowa caucus, and eventually the nomination, to Barack Obama.
2008, Iowa Democratic caucus: As Iowans headed for their caucus sites in 2008, the polls predicted an Obama victory by less than two percentage points over Clinton. Obama surged past the prediction to win by about eight percentage points. Clinton’s final support was just as polls predicted, while John Edwards added almost four percentage points. On the other end of the spectrum, Joe Biden and Bill Richardson’s numbers fell significantly below what polls predicted.
2008, New Hampshire Democratic Primary: On the day Iowans voted, New Hampshire polls predicted a seven-point victory for Clinton. These polls turned out to be better predictors than the final polls, which showed Obama jumping out to an eight-point victory after his Iowa caucus upset victory. Clinton eventually won, but only by about three percentage points. Edwards and Richardson both received about how much support as polls predicted.
At this point in 2012 and 2008, polls were accurately predicting the winners of five of nine of the above important races. None of the national polls on Iowa caucus day were able to pick the eventual nominee. Final Iowa polls went two for three.
Keep in mind that the polling industry has struggled since the 2012 election–for example, in the 2014 North Carolina election for United States Senate and the 2015 Kentucky gubernatorial election. Pollsters have struggled abroad, as well, as in the 2015 United Kingdom general election.
After all that, if you must know who’s winning the early polls in 2016, national polls show Donald Trump ahead in the GOP primary by nearly 20 percentage points. In Iowa, Trump leads Cruz by less than five percentage points, but their numbers are declining while Marco Rubio has added six percentage points in the past week. In New Hampshire, Trump has a much larger lead: More than 20 percentage points over Cruz and Kasich.
On the Democratic side, Hillary Clinton leads national polls by nearly 15 percentage points – a large lead, but not as large as at this point in 2008 when she still lost. In Iowa, Clinton and Sanders are separated by only four percentage points. In New Hampshire, Sanders has opened up an 18 percentage point lead over Clinton, similar to the lead Romney had in 2012 before his landslide victory.
Stay tuned and be sure to watch the results roll in from Iowa tonight. It will be interesting to see how long ago the polls were accurately predicting the winner and how close final polling comes to the actual results.
Jason Russell is a commentary writer for the Washington Examiner.
