By Joseph R. Gunderson
It’s no secret that nearly everyone seeking election for public office relies on polls. In the 2012 presidential election, Massachusetts Governor Mitt Romney and US Representative Paul Ryan of Wisconsin secured the Republican Party nomination, but failed to defeat incumbent President Barack Obama and Vice President Joe Biden, winning only 206 electoral votes to Obama’s 359. In an interview following the election, Governor Romney admitted that his campaign strategy on election day was flawed, in that he withheld “get-out-to-vote” campaign activity in two key swing states, Florida and Ohio. He mentioned that polling in those states showed that he was comfortably ahead of Obama going into the election, yet he ended up losing both states, losing by over four percent in Ohio. Although winning in Ohio and Florida would not have given Romney the 270 electoral votes needed to win, Romney’s campaign relied on polling data that influenced their election day-strategy. The 2012 election revealed an important attribute about polls: they aren’t always accurate.
The seemingly low polling integrity displayed in 2012, however, has not affected polls’ relevance in the 2016 election set to begin on Tuesday, November 8, 2016. Starting an unprecedented series of crossfire and “politically incorrect” rhetoric since his announcement speech on June 16, 2015, businessman Donald J. Trump took less than a month to oust former Florida Governor Jeb Bush and Wisconsin Governor Scott Walker from the top spots to claim the front-runner position in the race for the Republican Party nomination, according to multiple national polls. By the end of July, to the apparent surprise of nearly all news media outlets, Trump sat atop the GOP field, polling 20.8 percent support according to the Real Clear Politics average of select national polls. His nearest contender, Scott Walker, was a full 7 points behind.
Trump’s rise to the top of the polls became an instant hot topic; given comments he made at his announcement speech regarding illegal immigration, as well as multiple attacks on Senator John McCain of Arizona and his opponents in the race including, but not limited to Jeb Bush, Rick Perry, Hillary Clinton, and Lindsey Graham, the media erupted in utter shock with the release of each successive national poll. Polls including The Economist and Reuters showed Trump garnering twice, if not more than twice the support of his nearest contender in the GOP nomination race, and spurred discussion and analysis from nearly every news network.
Following a questionable performance at the first GOP debate hosted by Fox News, the news media reacted similarly when new national polls not only showed that Trump wasn’t losing support, but actually gaining support, polling above 30 percent in multiple national polls. Trump’s record-setting crowd size at campaign events complimented his massive lead in the key early-voting states of Iowa, South Carolina, and New Hampshire.
Many anticipated that Trump’s poll numbers would fade en masse after coming under fire for his inflammatory comments about fellow GOP candidate Carly Fiorina, coupled with retired neurosurgeon Ben Carson’s recent rise in national polls following the second GOP debate. Sooner or later, news of Trump’s domination got old, and subsequent headlines replaced news of Trump’s rise with polls showing Carson catching up to Trump, even though said polls bore high margins of error and surveyed very few people.
As a general rule, reliable national polls survey a minimum of 800 people via phone interviews, and have a margin of error of five percent or less. Polls conducted by The Economist, Reuters, and FOX News usually meet these general characteristics, surveying as many as 2000 people at times, and even asking those surveyed to name a second-choice candidate in addition to their first-choice. Reliable polls conducted at the state level generally survey a minimum of 450-500 people via telephone interviews and have a margin of error of five percent or less. While less reliable, polls that do not meet the general requirements sometimes make their way into the Real Clear Politics average of national polls, their releases do not go unnoticed by news media networks, and often make for greater publicity since their smaller sample sizes are more easily influenced by isolated inconsistencies that favor a single candidate over another.
To no surprise, such polls were repeatedly used to highlight Trump’s decline as the front-runner, a trend consistent with the medial value placed on Carson’s rise. In mid-October, the news media erupted when Trump lost his lead in Iowa to Carson, which was the first time in over a hundred days that Donald Trump had not taken the front-runner position in any state-wide poll. Every media outlet soon joined in on the bandwagon when a poll released by CBS News on October 27, 2015 showed Carson at 26% and Trump at 22%. While the poll surveyed less than 600 likely GOP voters, and had a margin of error of 6 percent, the media followed it closely. From sun up to sun down the numbers 26 and 22 under their respective portraits dominated the imagery on TV.
Just days later, an IBD poll released showed Trump at 23 and Carson at 28, which made headlines once more, with NBC News taking the opportunity to resurface claims Trump had made about Ford a few weeks earlier to imply a causation for Trump’s seeming poll slip. Interestingly, reporters hardly mentioned the poll’s small sample size of 402 likely Republican voters and high margin of error. More polls making headlines that week included a CBS News poll, showing Carson leading Trump 29 to 23, and a McClatchy poll, with Carson leading Trump 24 to 23. All of these polls, despite their small sample sizes and high margins of error, were included in the RCP average, and spurred more excitement when the average of the week’s polls showed Carson leading Trump 24.8 to 24.6, making him the “official” front-runner in what became a monumental end to Trump’s reign at the top.
Interestingly, Reuters released a poll the same day that the CBS poll made headlines, and sampled over 900 registered Republican voters, and showed Trump leading the GOP field at 35.7% support, over 15 points above Carson. Later that same week, The Economist released a poll that sampled nearly 2000 GOP voters which showed Trump’s support safely above 30%. None of these polls made headlines nor were they included in the RCP average. This is not to say that Trump is the deserving front-runner of the election for the GOP nomination, but merely to suggest that polls are used more often to entertain a story or shared belief, as opposed for strategic campaign purposes.
What is most striking about the brief slip in Trump’s polling is that the CBS poll was not the first poll to show that Carson was the front-runner. In fact as early as October 1, an IBD poll was released, showing Carson leading Trump 24 to 17 percent. While this poll was by no means accurate, surveying just over 300 likely voters, the news media did not purport it, likely because it still made for profitable media to marvel at Trump’s success amid his brash comments.
While Carson’s lead in the RCP averages lasted only a few days, almost no polls were released following the first week of November, keeping Trump and Carson locked in a fragile tie, separated by 0.4 percent.
However, the tragic terrorist attacks on Paris on November 13 changed everything. Ann Coulter declared the next day that “Last night, Donald Trump became the President of the United States,” sounding off a brief era of Trump’s rise amid an increased desire for the wall he has promised to build at the southern border should he secure the nomination and win the general election. The accuracy of this sentiment is unknown, but after a two-week respite in poll releases, a series of polls, Bloomberg, PPP, and ABC/Wash Post showed Trump as the clear favorite among Republican voters. Most notable was the Fox News poll released on November 19, showing Trump leading the GOP field at 28 percent support, ten points ahead of Carson.
While the polls that show Trump on top are all more reliable than those that were used to actualize Carson’s dominance in terms of sample size and margin of error, the sequence in which they were portrayed brings about another defining characteristic regarding polls. Accuracy does not assure a wider release by the media. The public accessibility of polls is largely due to what will make the most headlines or draw the most viewer interest.
While current polls show that Trump has regained a healthy lead over Carson in nearly every state, these polls are signaling the rise of another candidate who has largely gone under the radar: Texas Senator Ted Cruz. Many Republicans have come under fire from President Obama for their calls for a stronger strategy to combat ISIS and more cautionary measures on the intake of Syrian refugees. It would come as no surprise if a Cruz surge became the new mainstream talk of polls, given that he has claimed second place in a CBS poll conducted among likely Republican voters in Iowa. Although the poll has a hefty 6.2% margin of error, its mainstream attention could signal the arrival of a Cruz-favoring polling era.
With the election close to a year away, the polls so far have seemingly served little purpose other than making for profitable headlines. However, with Super Tuesday just around the corner, and a series of debates approaching, the polls will likely come under increased scrutiny as voters begin to make their decisions as to who they want to be the next President of the United States. Because polls also determine debate placement, added attention as to which polls are used becomes a responsibility of the news network hosting the debate. At the recent GOP debate hosted by Fox Business Network, Governor Mike Huckabee of Arkansas and Governor Chris Christie of New Jersey were disqualified from the main debate stage because of low polling numbers. Trump still took center-stage despite Carson’s recent takeover as the official frontrunner.
Although many polls are put out there for publicity’s sake, their relevance is taken into consideration in the context of more serious matters. As Iowa is set to caucus on Monday, February 1, 2016, many of the polls to come will place some candidates over another, and will ultimately influence campaign decisions as they did back in 2012. A 500-person sample does not accurately reflect the views of millions of people in every state. Small polls do not ensure that the sampled people will actually make the effort to vote. Yet, polls do matter, and they will very likely determine campaign strategy despite their inherent inaccuracies.