Destiny Wells’ campaign released the results of a poll a few days ago.
The poll said that Wells, the Democratic candidate for Indiana attorney general, and Jennifer McCormick, the Democrats’ gubernatorial candidate, were within striking distance of their Republican opponents, Indiana Attorney General Todd Rokita and U.S. Sen. Mike Braun, R-Indiana, respectively.
The survey results found that Wells trailed Rokita, 41% to 44% — and she actually climbed into the lead, 50-42, when those polled were informed about Rokita’s record in office. (He’s been reprimanded by the Indiana Supreme Court Disciplinary Commission and the Indiana Supreme Court once already and is the subject of two other ethics investigations by the commission.)
The margin in the McCormick-Braun race, the poll said, was even tighter — with Braun up, 41% to 39% while Libertarian Donald Rainwater trails far behind.
Given that the poll had a margin of error of 4%, that means both races are too close to call.
If the poll can be believed, that is.
That’s a big “if.”
There are at least two reasons I’m skeptical of these results.
First, this survey greatly resembles a push poll. Push polls are presented as surveys to those being asked for their preferences, but then become a means of advancing a campaign’s or cause’s message.
Push polls can be effective — remarkably effective, in fact — as political tools, but they are less reliable as measures of public opinion.
Second, I’m dubious about polls in general these days.
This is not a reflection on Lake Research Partners, a well-respected pollster, albeit one with decided progressive leanings. I’m sure they worked hard to secure as accurate a sample as possible for Wells, their client, because giving her wildly inaccurate numbers wouldn’t help her or her campaign in any way.
But the fact is that polling now is much more complicated and therefore much more difficult than it was a generation ago.
That is why, again and again, election results have confounded both pollsters and the public. For almost 20 years now, the actual vote has varied — often by large margins — from the polls taken close to election day.
This year, the gap could be even wider than usual.
There are several reasons for this.
One is that the science of sampling has struggled to keep pace with the explosive growth of technology — particularly communications technology. In the days when our telephone conversations took place exclusively through landlines, it was relatively easy to determine samples of demographic groups to arrive at a representative survey.
Now that both people and phones are more mobile, though, this has become harder and harder to do. My grown children both have cellphones with Indiana area codes, but neither has lived in the Hoosier state for years. Pollsters can and do try to correct for this, but doing so takes time, money and increasing levels of skill.
Another factor that pollsters don’t like to talk about is that a growing number of people lie when called as part of a poll. No one can say definitively why this is, but it accounts for the fact that, for example, former President Donald Trump significantly under-polled in the key battleground states of Michigan and Wisconsin in 2020. For whatever reason, many people surveyed intended to vote for him but didn’t say so.
And then there’s this: Every poll depends upon past turnouts to determine the models for measuring present sentiment. Pollsters look to previous elections to decide how the sample should be determined.
It’s a method that works well in placid times, but not so well when the nation is in a rapidly evolving state of turmoil.
That’s when many new voters without histories of partisan allegiance enter the process — and their entry disrupts the models.
This has happened at least three times in the past 16 years — in 2008 when Barack Obama became president, in 2016 when Trump scored an upset victory, and this year, when Vice President Kamala Harris has emerged with surprising, even startling force.
Brand new voters — particularly brand-new enthusiastic voters — make everything hard to predict.
It’s possible the poll Destiny Wells’ campaign conducted is right on the money. It’s also possible that it isn’t within a country mile of being accurate.
The one thing that is definite is that it doesn’t matter all that much.
The only poll that counts for much is the one taken on election day.
John Krull is director of Franklin College’s Pulliam School of Journalism and publisher of TheStatehouseFile.com, a news website powered by Franklin College journalism students, where this commentary originally appeared. The opinions expressed by the author do not reflect the views of Franklin College.
