Among the Republican power struggles throughout primary campaigns in 2014, one clear winner emerges, albeit quietly — public-opinion polling.
Results of individual polling may have been all over the map in successive GOP primaries, but Real Clear Politics has predicted the winner in every Republican in every race so far, with the exception of Mississippi, which showed a tie. There will be a June 24 runoff in Mississippi, since neither candidate received 50% of the vote
Although it may not have been a ringing endorsement of the polling industry, given widespread results on individual studies, the modest victory of polling in 2014 should come as a welcome relief to an occupation that took a beating in the 2012 White House race. That was the election cycle where polling failed to predict the relative strength of President Barack Obama, writes Patrick O’Connor of the Wall Street Journal.
Seen as imperfect indicators in politics, public opinion polling has become a frustration to candidates, and an advantage for reporters looking to examine every angle in a race.
Even in the best of times — which this cycle is certainly not — polling is most useful when tracking changes in public opinion than for determining specific levels of support for a particular candidate or issue.
Pollsters are struggling to keep up with the rapidly shifting demographic trends, and it has become harder to predict which groups will actually turn out to vote. In addition, it has grown harder to reach people at home; this forces polling firms to rely on data culled from cellphone users, or even complement results with online surveys.
These factors drive up the polling costs, leading several firms and media organizations to rely heavily on automated phone calls that offer lower prices, often at the cost of accuracy. It is the expected consequence of reduced budgets at newspapers and other media outlets.
Furthermore, campaigns are beginning to supplement opinion research with increasingly other, more sophisticated and costly methods for polling voters, which rely on massive capital investments and near constant surveys, which are beyond the reach of most media outlets.
This state of affairs in the polling industry creates an ever-widening knowledge gap between campaigns and the voters they wish to influence
Given the situation, the results at this point in 2014 merit a closer look.
Not only has Real Clear Politics picked winners in each competitive Republican Senate primary, but also they predicted the top three finishers in the Georgia, Nebraska and North Carolina races.
In Mississippi, the RCP polling average on Election Day showed Sen. Thad Cochran tied with state Sen. Chris McDaniel, who finished only 0.7% ahead of the incumbent, which led to a runoff.
Not all of this is good news.
RCP polling averages failed to show the ultimate winners in both Iowa and North Carolina, eclipsing the amount needed for an outright win – 40% in North Carolina and 35% in Iowa. However, both polling indicated both eventual winners were right on the edge, and had the momentum into the day of the primary. Both cases had RCP averages with a large share of undecided voters leading into the primary.
In each of those races, the fence sitters were the deciding factor.
In the Iowa primary, state Sen. Joni Ernst outpaced all public opinion polling by at least 20 percentage points, thanks to an intense tilt in her favor in the last days of the primary campaign. Ben Sasse, the Republican nominee to fill Nebraska’s open Senate seat, also surged towards the end, even as his two chief rivals scarcely budged in polling.
As for Oklahoma, there is not enough data to make an educated guess on which Republican Senate candidate has momentum in the lead up to the June 24 primary — conventional wisdom has former state House Speaker T.W. Shannon edging Rep. James Lankford.
And in Mississippi, the unpredictable nature of runoffs remains a mystery.
From looking at early primary results, the takeaway is that averaging disparate polling still tend to give a good idea of where the races are heading. As O’Connor says, the old political adage applies: ten bad polls equals one good one.