I’m not sure how this happened, but apparently I’m the last person Frank Graves tweeted to before he left the building:

I had made an observation that EKOS was finding Green voter intentions to be much higher than the actual election result, while Nanos (among others) was closer. I believe this result comes, in part, because Nanos does not prompt poll respondents potential party choices in their question, which more accurately reflects what goes on in the voting booth (yes, the party names are on the ballot, but EKOS wasn’t prompting for the Rhinoceros Party or the Marxist-Leninists as well). These methodological differences can create substantial gaps in the results, especially for a small party like the Greens, where voter responses are driven just as much as how they want the Green Party to look in the polls as their actual voting intentions. The lack of actual observations to compare polls to elections makes it hard to make the case decisively, but I think EKOS does need to consider this factor.

I’m sure it’s merely a timing thing, but it’s weird nonetheless to feel like I’m holding the last correspondence someone sent before falling into the Witness Protection Program. This is, in all honesty, completely messed up. Public opinion polling is overwhelming an apolitical affair. Pollsters follow elections and run voter intention polls as a form of proving their art, as to demonstrate their methodology to commercial clients, the true cash cows of polling. They have little to no incentive to bugger up their political polls. While I do find the increased demand for pollsters to act as commentators to be unfortunate, the rationale for them filling that demand is the same: increased exposure for commercial clients. That anyone would harass a pollster for other people’s responses is, not only obviously criminal and heinous, but also idiotic and delusional.