Demystifying polling: a political pollster helps make sense of Connecticut elections
This is part 1 of a 2 part series.
We're not just in the middle of campaign season. We're also in the middle of polling season. With so many political polls released in the months leading up to the November midterm election, it can be hard to square the various results.
One of the polls released in September, which estimated big leads for Democrats in Connecticut's gubernatorial and Senate races, came from Western New England University in Springfield. Tim Vercellotti is the director of WNE's polling Institute, and he spoke with All Things Considered host John Henry Smith about how polling works.
John Henry Smith: Tim, your poll has Connecticut Republican gubernatorial candidate Bob Stefanowski, trailing incumbent Ned Lamont by 15 percentage points. Other polls seem to more or less agree. Now when told of his deficits, Stefanowski told local TV station Fox 61 that he pays attention to his internal polls instead of the external polls. What do you make of Bob Stefanowski, saying that?
Tim Vercellotti: That is a common refrain from candidates, in particular, candidates who are on the losing side in a survey where there's a big gap. No one would want to say, uh, "yeah, you know, our internal polls show me down by double digits as well." And it's not so much the effect it has on voters, but the effect it has on donors. The one way to close the gap is to make sure you've got sufficient money for campaign ads, and field operations, and voter outreach. And it is a concern for candidates if polling shows them behind that that money's going to dry up.
Now, it could also be -- and private campaign pollsters will do this -- that they run a series of scenarios with different messages, different circumstances. And under some circumstances, say a low turnout election, the gap might be smaller. High turnout, the gap might be wider. So it could very easily also be that Mr. Stefanowski's, internal polling was showing a different race than what we are other public polls for showing.
John Henry Smith: I heard one local political pundit talk about how he was suspect of most recent polling results, because in his mind, you and other organizations weren't careful to select only registered voters. With regards to that criticism. What do you make of it?
Tim Vercellotti: Yeah, well, we ask. We say "Are you currently registered to vote at the address where you live or not?" And of that 850 to 766 said they were registered to vote. So then you go from there, the folks who aren't registered? You keep them in to get some demographic data - "How old are you?" "How far in school did you go?" "What are your race and ethnicity?" so that you can, again, get a sense of who everyone is. And, does your whole pool of respondents look like the population under study? And then we asked, "well, what party are you registered with?" And Connecticut's Office of the Secretary of the State, of course, has statistics on the number of voters registered as Democrats, Republicans or unaffiliated. So, you can look at your pool of registered voters and see, okay!
For example, in Connecticut, 20% of registered voters are registered as Republicans. How did our sample do? 37% of registered as Democrats, how do we stack up? And then 42% unaffiliated. So you have these markers, and you can get a sense of how representative your sample is. And if you're off, you can make some mathematical adjustments using weighting there as well to make sure your sample looks more like the population of registered voters.
John Henry Smith: What does your polling say about the [unaffiliated] voter phenomenon here in Connecticut versus other states in New England?
Tim Vercellotti: In a state where Democrats have significantly higher voter registration than Republicans, the Republican candidates for statewide office have to really build up some significant support among the unaffiliated voters. And at least in the polling that we did in September, we weren't seeing that. They were pretty evenly split between the Democratic candidate and the Republican candidate for governor and Senate, which is not good news for the Republican candidates. They've really got to run up some big numbers among those independent voters in Connecticut.