As predicted, Stockton’s LD1 poll was way off the mark. | Rooney

By Matt Rooney
_

Any honest broker with a scintilla of political knowledge saw this one coming a mile away, Save Jerseyans.

Do you remember that October LD1 Stockton University Poll? Showing Republican challenger Mike Testa trailing Democrat incumbent Bob Andrzejczak by 14 points (53% to 39% in favor of the Democrat)? I wrote about the poll’s many sampling errors at that time, pointing out the big issues with Stockton’s methodology including the pollster’s curious decision to assume better-than-2018 turnout for Democrats in a rural, lean-red legislative district carried by Trump, Guadagno and Hugin.

Lo and behold, two weeks later, Mike Testa is now a Senator-elect.

He won by about 7-points, a 21-point swing.

Stockton’s interim polling director John Froonjian told InsiderNJ he’s not sure what happened. He actually sounded pretty damn defensive. Froonjian’s defenses included:

  • His poll in neighboring LD2 was closer to the mark using the same sampling (despite the fact that these are significantly different districts notwithstanding the geographic proximity – e.g., battleground LD2 includes Atlantic City and has a significant Dem registration advantage)
  • The fact that the poll was conducted “more than two weeks before the election. It was a snapshot in time and not a prediction of the vote.”
  • The fact that, “[b]etween the end of the poll and Election Day, there was a major event in the House vote on the impeachment process” which could’ve mobilized GOP voters.
  • The possibility of”a Bradley effect (after former LA Mayor Tom Bradley) regarding Andrzejczak. Would conservative voters feel awkward saying they oppose a wounded war veteran but vote  against him in the booth?”

I could give you a breakdown (and rebuttal) on each of these points, Save Jerseyans. How much time do you have? But besides what I’ve already explained before the votes were tallied (click here), none of the factors above – individually or cumulatively – could have produced a 21-point swing in just two weeks’ time. Not possible!

Team Testa agrees. After seeing Froonjian’s remarks, they issued a rare post-election rebuttal to clear the air.

Testa (R-1)

“While we don’t hold any ill towards the good folks at Stockton, we are disappointed that they just don’t concede what is obvious – that the methodology in their poll related to partisan turnout was dead wrong and what we said at the time was absolutely right,” said Testa Campaign Manager Brittany O’Neill. “LD1 is a +2 Republican district in terms of registration and has averaged +7 in terms of Republican turnout in recent elections. Despite that, the Stockton poll had a sample that was +6 Democratic. We knew instantly the number was wrong and nothing in the final two weeks of the campaign would have precipitated a 20-point swing in our favor. Moreover, we take issue with the idea that they polled the Assembly race properly. Their poll showed Land at 27, Milam at 26, and both Simonsen and McClellan at 22. When you appropriately double that number to base the polling on 50% to win, it showed a 10 and 8 point lead, respectively, for the two Democrats. That wasn’t close to reality either.”

It’s annoying when an adult can’t just man-up and admit fault. It’s even more annoying when the mistake could’ve helped throw an election.

“Releasing that poll caused our campaign and our candidates hours worth of time reassuring grassroots supporters and donors about our prospects for success, and educating them as to why the poll was flawed,” added O’Neill. “That wasn’t fair to our campaign or the voters, and given the competitive congressional elections in South Jersey on the horizon in 2020, we are hopeful that this incident prompts a full vetting and review by both Strockton’s Polling Institute and the media who cover their polls.”

2016’s polling lessons continue to be lost on many in the media/academic world. Are they really that dense? Or are their skewed numbers some species of attempted wish fulfillment? You be the judge. A realignment is underway and they’re not adjusting accordingly. That’s part of it. 

The lesson for Republican/Conservative/Libertarian/Trump voters: VOTE. Ignore the polls. Sometimes they’re right. Sometimes… not so much. The only poll that matters is the one officially recorded by the secretary of state’s office.