Counterintuitive polling data suggests unpredictable race

If Louisianans wondering what is up with this year’s state races across the country and at home have learned, it’s that it’s hard to predict what’s going to transpire in the state’s gubernatorial contest.

Earlier this week, voters delivered verdicts in elections in Kentucky and Virginia that defied expectations and polling, bringing a Republican ticket home in the Bluegrass State and keeping the statehouse in GOP hands in the Old Dominion. The seeming surprise of it all matches that observed at present in the Bayou State.

In the last couple of days, governor’s contest polls with varying partisan backers and records of accuracy (one historically overestimating Democrat strength, another this year having continually showed different results from others that were more consistent with the actual outcome) all gave Democrat state Rep. John Bel Edwards a lead over Republican Sen. David Vitter, even enough margin for an outright win regardless of which way the undecided portions would swing. Such results are entirely counterintuitive from the general election results, where Republican candidates lead the Democrats running by 15 percent.

It would seem inexplicable that the state’s voters, which only last year sent the alleged moderate although really unabashedly liberal former Sen. Mary Landrieu to defeat in her reelection bid by 12 points, at this juncture would prefer the alleged moderate although really unabashedly liberal Edwards to Vitter, even as Edwards has run an even more intense liberal-wolf-in-conservative-sheep’s-clothing campaign that did she. More incredibly, Edwards has been outspent even more dramatically than was Landrieu, yet appears to be doing better.

Still, that polling in the other two states’ contests underestimated, and considerably, the support that Republican candidates would receive may indicate a failing common to polling in this cycle when it comes to final general elections (although Louisiana’s technically is a general election runoff, it has the same characteristics as a general election featuring two major party nominees). Drawing upon the Kentucky experience, one analyst tried to extrapolate what could be happening in Louisiana.

His conclusions: some poll respondents not willing to admit they would vote for the Republican and a return to partisan backgrounds on election day. He also notes a difference in that Vitter’s long-time presence in state politics may more stringently define him and make it more difficult to come from 17 points back. However, three other important observations were not made.

Vitter’s supposed well-defined candidate image also includes something shared with the victorious Kentucky Republican: he is viewed by many as a political outsider regardless of his long political career. He certainly is relative to state government, having not held one of its offices for over 15 years. As Republicans in particular become less hesitant to embrace the “outsider” in their vote choices, this will accrue to his advantage, and already his campaign tries to disseminate this theme.

In addition, from what information they have released, these polls lean on a turnout model replicating the general election. While in statewide, nonfederal elections white Democrats typically turn out at a rate higher than Republicans, this gap regardless of the level and of office contested (presidential, senatorial, gubernatorial), GOP turnout picks up in runoffs relative to that of white Democrats (and about washes with black turnout). In other words, these polls may be oversampling white Democrats.

Finally, specific to statewide gubernatorial elections, unlike with those for federal office, turnout increases from general election to runoff. These additional numbers also help Republicans; not only do they make up a higher proportion of voters in the runoff, they also magnify that by having more show up in an electorate that generally expands.

Also, to amplify a point made by the national writer but which has no numbers to back it up, because Vitter has been so vilified throughout the race, this may produce under-polling of him. The state precedent goes back to the 1990-91 period, where former state Rep. David Duke turned in election day totals dramatically higher than even recent polling.

Keep in mind as well that polls represent a snapshot of opinion at a certain time, and while correlated significantly with eventual results when close to an election, they can be off considerably. Given that these polls all show fairly substantial numbers of undecided voters in an environment historically where turnout increases in the runoff, support for Edwards in particular probably is soft, as he has employed a blank slate strategy that does not allow embedding of loyalty such as Vitter has built up over the years. If the Vitter campaign can demonstrate the many instances where the image Edwards has tried to build deviates from his record to the electorate, a significant portion of the less-informed electorate will abandon Edwards by election day.

Simply put, Edwards is not leading this race by double-digits; he may not be leading it at all in two weeks. Yet he does appear to have made it a lot closer than anybody imagined, probably including him. That tells us something interesting about the relative maturation of the Louisiana electorate and evolution of its political culture – subjects for future investigation.