Hoisted from the comments, a gentle pushback from the CivicScience folks on our overhyping of their poll of the Pittsburgh Mayoral race. Forgive us CivicScience! We did not mean to get too excited, we are just starved for polls on this race…
Thank you so much for the write-up. We never expected such a reaction like this to our little blog post. We do worry that your headline is a bit misleading however. What it should say is: “A Hypothetical But Highly Unlikely Two-Way Mayor’s Race Would Be Within Single Digits If Every Registered Democrat Turned Out to Vote, Which Everyone Knows Won’t Happen.” Unfortunately, we realize that wouldn’t Tweet so well.
There are a number of reasons we didn’t (and most likely won’t) share any forecast numbers. First, we ran all of our voter analysis one day before a new and viable candidate entered the race. To spend the time crunching our data to forecast an unlikely two-person race made little sense. We’ve already seen in our tracking numbers since the 17th how Michael Lamb’s candidacy is altering the numbers.
Second, we only analyzed REGISTERED Democratic voters, not super voters or even likely voters, which is a monumental distinction. In an off-year local government primary, we all know that registration means very little. The election will come down to a small subset of those registered voters who turn out (on a probably-rainy day in Pittsburgh in May). Do we have a pretty good idea who they will be? Yes we do. But we’re not going to share that now and may not do so publicly at all…which leads to our last point.
We wrestle a lot with releasing political “predictions” to the public, especially in a local race like this where there will be very little if any other public polling data. We’re not a predominantly political shop (the vast majority of our business is focused on consumer marketing and financial market research). The last thing we want to do is be seen as biasing or influencing the race in anyway. What we will probably do, like we did during the Presidential, is release our predictions about 5 minutes after the polls close on election day. In the meantime, we just want to share some fun numbers from time to time, numbers that should have a little bit of insight and hope for all of the participating candidates…not to mention fodder for our friends in the media to write about.