Column: Having skepticism for polls is the best option

  • Posted: Monday, September 1, 2014 12:01 a.m.

The election season traditionally starts on Labor Day. Yet in several key U.S. Senate, races, the horses are already half way around the track.

And for political pundits, the daily polls are the fodder for daily headlines. In the 24-hour news cycle, every poll is of epic significance. What candidates stand for or might actually do if elected are far less newsworthy.

Even outlier polls are breathlessly reported. Accordingly to a recent poll in New Hampshire, Republican Scott Brown is within two points of Democrat Jeanne Shaheen. Stop the presses and start the debate.

Analysts Harry Enten and Sean Trende have already traded conflicting interpretations of this poll online.

And across the continent, another poll has Democrat Mark Begich trailing Republican Dan Sullivan by two points. Shock and awe. Release the journalistic hounds.

This frenzy is driven by 2014’s political narrative: Republicans will retain control of the House of Representatives. Democrats will pick up a governor or two. And what’s truly up for grabs is the U.S. Senate.

Democrats hold a comfortable 55-to-45 margin in the Senate. But among the punditry, there’s a growing belief that Republicans may gain control come November.

These professional prognosticators approach this question either qualitatively or quantitatively.

Those using a qualitative approach examine the political landscape, voting history, candidates’ individual strengths and weaknesses, presidential popularity and campaign finance reports to formulate their guesstimates. Sometimes bat wings and eyes of newt are sprinkled in for good measure.

For example, analyst Charlie Cook thinks 11 Senate races lean toward the Democrats, 16 toward the Republicans, and nine are rated as toss-ups. He sees Republicans gaining four to six seats.

The quantitative approach uses the same inputs, but then grinds them through complicated statistical models to churn out a forecast.

Nate Silver is the reigning analytical wunderkind – almost alone he correctly predicted the 2012 presidential election results – and his latest analysis shows Republicans gaining 5.9 seats.

(Though the 0.9 is a statistical artifact that rounds up for an even six-seat gain, I still want to know what 0.9 of a Senator looks like. What part is the missing 0.1? And where did it go?)

But even quantitative models go wrong. No one, for example, thought House Majority Leader Eric Cantor would blow his June primary election. Or that the GOP would lose two Senate seats in 2012.

Likewise, Silver gave the German World Cup soccer team a 1-in-4,500 chance to score seven or more goals against Brazil. Yet, Germany ultimately crushed Brazil 7-to-1. The statistically improbable can occur.

Complicating the picture are polling quality issues. In this era of cell phones, robocalls, low response rates, partisan questionnaires and growing costs, the soundness of the data is increasingly uncertain. Separating the wheat from the chaff is difficult.

With these limitations in mind, what can we expect in this year’s Senate races? To win a majority, the Republican path to victory winds through three basic electoral templates.

First, win Democratic open seats (where the incumbent is retiring) in either red (Republican leaning) states or purplish swing states. These include West Virginia, Montana, South Dakota, Iowa and Michigan. Republicans will gain at least three of these.

Second, defeat vulnerable Democratic incumbents in similar states. These include Mark Begich (Alaska), Mark Pryor (Arkansas), Mary Landrieu (Louisiana), Kay Hagan (North Carolina), Mark Udall (Colorado) and Jean Shaheen (New Hampshire). In every one of these states, the incumbents’ polling averages are below 50 percent. Republicans must win some of these.

Third, defend Republican held seats where Democrats have a glimmer of hope. These realistically boil down to two: Mitch McConnell (Kentucky) and Georgia’s open seat. Democratic long shots include Pat Roberts (Kansas) and Thad Cochran (Mississippi).

Obviously, the permutations and combinations of possible results are endless. Estimates of five or six seat gains are merely midpoints between a massive Republican wave netting 10 or 11 seats to minimal gains that fall short of control. The number of true toss-ups is large.

And historically, one side or the other typically wins the majority of toss-ups. An extreme result either way is possible.

Up until Election Day, expect well-informed political handicappers to pontificate on the day-to-day ups and downs in these races. They’ll have massive supplies of facts and figures. They’ll claim to know the score.

Wise political handicappers, however, will keep everything in perspective. Polls, models and analyses are useful but fallible. They – and we – should expect the unexpected because nothing is certain until the votes are counted.

Gary Bunker is a former Aiken County Councilman.

Comments { }

Commenting rules: Do not post offensive, racial or violent messages. Responsibility for the statements posted lies with the commenter, not www.aikenstandard.com. Click 'report abuse' for any comments that you feel should be removed from the site. However, www.aikenstandard.com is not obligated to remove any comment posted on the site. Moderators do not have the ability to edit comments. Read the terms of use.