- The Crosstab
- Posts
- Public Polls Are Killing Me... Smalls.
Public Polls Are Killing Me... Smalls.
MassInc out with a real headscratcher
MassInc’s most recent poll probably oversampled the city of Philadelphia by 88%, Allegheny by 20%, under sampled the Republican heartlands by 33%. In short… TRASH. If you can’t get geographic quotas right WHAT ARE YOU DOING!
I know everyone is sick of hearing about the public polling problems, but here we are again with another outfit pumping out a survey that at least on face value seems to have serious, serious issues.
While MassInc/Spotlight did not release any cross tabs (boooo) they did at least release their top line demographics (golf clap)

So two issues here where some clarity would help ALOT. The first is on “Party Identification” and what they mean by that. If they ask some sort of behavioral party question like are you a strong Republican / Democrat / etc then this configuration they have above at 34%D / 38%R / 27%I might, and I repeat, might be defensible. But if we read through their methodology statement it says they set quotas and/or weighted for party registration and they use the term ‘unenrolled’ which is a Massachusetts terms for people not registered with any party. So its unclear to me if they mean 27% are not registered Rs or Ds or if this was a behavior question they asked.
PA is one of the easiest states in America to poll because we register by party and we have a relatively small percentage of non-major party voters.
For years PA elections have roughly been some variation of independents between 9-12% of the vote. They will not under any circumstances be 27% of the vote. I hope to hear from someone at SpotlightPA/MassINC about it.
However, that’s not the major issue. They missed the geographic quotas in PA in a HORRENDOUS fashion assuming their naming conventions in the publicly released docs are what I think they are.
If “Philadelphia” above means the city/county of Philly then they should have gotten 9-10% NOT 17%. I think that’s really the only way to read it since there is a separate “South East” category that I assume is the rest of the media market.
My best guess is this error alone gives democrats about a 2.5 point boost on the statewide ballot.
I am assuming Northeast / Central is the Wilkes-Barre and Harrisburg media markets which they would be under sampling by 4 points.
Then if I assume “Alleghany” (sic) is meant to be just Allegheny county they oversampled it by 20% (should be 10 points not 12). Then there’s the question of what the heck they did with the rest of the West. If we assume “West” is Erie DMA + Johnstown DMA + Pittsburgh DMA outside Allegheny that should be 21% of the vote NOT 13%.
So in essence I think though hard to tell that they more or less gutted respondents in the Republican parts of the state and then wayyyy oversampled Democratic regions.

My face when I looked at the geo breaks in this survey
Mark Harris is founding partner at ColdSpark and is a leading Republican political consultant having worked for clients including Ambassador Nikki Haley, Sen Pat Toomey, Sen Marco Rubio, Rep Alex Mooney, Rep Byron Donalds, Rep Russ Fulcher, Rep Guy Reschenthaler, State Auditor JB McCuskey and others across the country. He’s worked in 48 of the 50 states at this point.