Is the Ames Straw Poll Useless?
This August, Republican presidential hopefuls will converge on Ames, Iowa, for one of the most idiosyncratic and entertaining rituals of the Republican nominating contest – the Ames Straw Poll. The nonbinding contest, which has been described as a hybrid of a county fair and a political convention, typically draws most of the Republican presidential field and in recent years has drawn well over 10,000 voters. The event features a straw poll of the attendees, and candidates who perform well in this poll are rewarded with media attention and are often regarded as early frontrunners in the Iowa caucus. While Iowa does not always pick the eventual nominee, a solid finish in the nation’s first presidential caucus can give candidates momentum and exposure at a critical early point in the primary.
But not every Iowa mover and shaker is a fan of the poll. In 2012, after Michele Bachmann had won the 2011 poll only to sink like a stone in the caucus, Iowa Gov. Terry Branstad summed up much of the conventional wisdom on the poll, saying, “I think the straw poll has outlived its usefulness. It has been a great fundraiser, but I think its days are over.”
But has the straw poll outlived its usefulness? To answer this question, we pulled together data from every Ames Straw Poll back to the first one in May 1979 and examined how well those results predicted the final results of the Iowa caucus.
Our analysis calls into question the usefulness of the upcoming August poll. At the moment, both the invisible primary (the contest for support of high-dollar donors and party elite) and the visible primary appear be chaotic and unsettled. We found that in such contests, the Ames Straw Poll had little predictive power. In years when the primary was settled fairly early, Ames had a decent track record of predicting Iowa caucus results.
When Ames Was Predictive – 1980, 1988, 1996, 2000
In 1979 there were actually two Ames Straw polls – one in May and another in October. The May poll did not include all of the candidates, but it predicted the final results well – George H.W. Bush won, Ronald Reagan came in second, Howard Baker placed third, John Connally came in at fourth, and Phil Crane came in fifth at both Ames and the Iowa caucus.
The October poll was not as effective as a predictor because Bob Dole over-performed and Ronald Reagan underperformed.
Dole’s surge was due to a concerted effort by his campaign rather than a growth in his popularity in the state. After his lackluster showing in the May poll, Dole reportedly bought $20,000 worth of tickets. That’s enough for 400 supporters at $50 a plate – a sizeable number considering Bush won the October poll with 519 votes. While Dole only received 215 votes in the final tally, that total likely represents his campaign’s efforts and spending rather than his actual popularity in the state.
The October poll also missed Reagan’s strong second-place finish in the caucus, so its overall accuracy in the 1980 cycle was decent – the May poll was quite accurate while the October poll had a couple important misses.
In 1987, the poll did a decent job of predicting the final results.
Televangelist Pat Robertson won the Ames poll, but Bob Dole surged ahead of him to win the Iowa caucus. Both did well in Iowa, so the 1987 Ames Straw poll results mapped well (albeit not perfectly) onto the 1988 Iowa caucus results.
In 1995, the poll only had one big miss – then-Texas Sen. Phil Gramm.
At Ames, Gramm tied then-Kansas Senator and eventual nominee Bob Dole, but Gramm’s momentum came to a grinding halt when conservative firebrand Pat Buchanan beat him in the Louisiana caucus. Five days after that embarrassing defeat, Gramm placed fifth in the Iowa caucus. (Ames also technically missed Steve Forbes, but we can’t count that as an inaccuracy – Forbes officially announced his run in September 1995, and the poll happened in August).
The 1999 Ames straw poll was also a pretty good predictor of that year’s final Iowa caucus results.
The poll put then-North Carolina Sen. Elizabeth Dole in third, but she dropped out of the race before the Iowa caucus due to poor fundraising. Aside from that, Ames results mapped pretty well onto the final Iowa caucus results. While there was some shuffling around near the bottom of the field, it predicted a strong Bush win with a decent showing by Forbes and the rest of the field (except Dole) trailing substantially.
In each of the four elections -- 1980, 1988, 1996 and 2000 -- there was a solid frontrunner among the party elite and the voters, and in each case Ames was a decent predictor of overall Iowa results. Reagan, George H.W. Bush, Dole and George W. Bush were all arguably the early favorites. Each of them received at most one serious challenger – George H.W. Bush opposed Reagan from the center, Dole ran a serious campaign against then-Vice President George H.W. Bush and McCain represented the party’s moderate wing against George W. Bush – but each time the frontrunner managed to put away these challenges relatively quickly. In other words, when the primary was relatively settled and predictable, Ames predicted the overall Iowa results reasonably well.
When Ames Got it Really Wrong –2008 and 2012
In 2008, the Ames sent some false signals about the final results in Iowa.
Former Arkansas Gov. Mike Huckabee gained significant momentum from his second-place finish at Ames. Former Massachusetts Gov. Mitt Romney, on the other hand, dropped from a first-place finish at Ames to a second-place finish in the Iowa caucus. McCain skipped Ames and still made a decent showing in the Iowa caucus. Ames also missed former Tennessee Sen. Fred Thompson’s third-place finish, but that’s understandable – like Forbes in 1995, Thompson did not declare his candidacy until September and thus was not an official candidate during the poll.
The straw poll also had a major winnowing effect in the 2008 cycle. Then-Kansas Sen. Sam Brownback and former Wisconsin Gov. Tommy Thompson both dropped out following the poll. Then-Colorado Congressman Tom Tancredo dropped out of the race in December 2007. This was a bad sign for the poll’s predictive ability – Huckabee received 18.1 percent of the vote but went on to win the contest, but Brownback and Tancredo, who were on Huckabee’s heels (at 15.3 percent and 13.7 percent, respectively), dropped out of the race before the caucus even took place.
In 2012, Ames failed to accurately predict almost any of the salient features of the race.
Former Texas Congressman Ron Paul was the only candidate who performed well in both the Iowa caucus and the Ames Straw Poll. There was no discernible relationship between any of the other candidates’ performance in Ames and their performance in the Iowa caucus (aside from Jon Huntsman and Thad McCotter, who performed poorly in Ames and the Iowa caucus).
In both cases, there was a reasonably deep field without a clear heir apparent. The slow trickle of early endorsements from Republican senators, governors and congressmen in both 2007 and 2011 suggest that neither McCain nor Romney won the invisible primary as quickly as some of their predecessors.
And in the visible primary, most of the race had yet to fully shake out. At the time of the Ames Straw Poll in 2007, Iowa voters had not fully discovered Huckabee, and Rudy Giuliani was still leading in national polls. The 2012 cycle was even more of a rollercoaster – in August of 2011 then-Texas Gov. Rick Perry had just begun to gain popularity and both Rick Santorum and Newt Gingrich had yet to catch fire (which they both did twice before Romney sewed up the nomination).
In short, the race was still pretty chaotic and unsettled in August of 2007 and 2011, so the predictive power of the straw poll was low.
Will the Poll Be Useful in 2016?
Based on this data and the current state of the field, the 2015 Ames Straw Poll will probably have little predictive power.
The Republican field looks deeper than in many previous cycles, and a number of potential candidates are already competing in the invisible primary. There is also no heir apparent. There is no sitting Republican vice president (such as George H.W. Bush in 1988) or clearly acceptable second-place finisher from 2012’s nominating contest (as Reagan was in 1980). It’s possible that one candidate will catch fire early or that many candidates will choose not to run, but for now the Republican presidential primary looks chaotic. Unless that changes, don’t pay attention to Ames.