This is kind of an experiment, but given that there won’t be any traditional exit polling today, Pace’s data is all we’re going to have. (And traditional exit polls have, as we’ve learned in recent presidential elections, their own shortcomings.)
“The exit poll will be conducted today via telephone, among registered voters in New York City who voted in today’s general election. Potential respondents will be randomly-selected from a list of registered voters in New York City; respondents will then be screened to ensure that they voted already. Results will be weighted to reflect the actual vote,” Pace Poll director Jonathan Trichter writes.
Trichter acknowledges that this isn’t a perfect system. In particular, it’s not going to predict the outcome (not that we need that) since like other telephone polls, it’s captive to its turnout model.
But Trichter thinks the poll will be useful for seeing how the candidates did within demographic groups.
“The dirty little secret known to all pollsters is that the errors in exit polls horse race predictions don’t affect the accuracy of subgroup analyses, since one key feature of exit polls is that the results are weighted to reflect the actual vote.
“Since we will weight our results to the actual vote, we can say that we know the percentage of Latinos or Catholics who voted for each candidate, even if we may have mis-estimated the percentage of Latinos or Catholics in the overall electorate.”
It’s an experiment, but could be a really valuable tool. I’ll post their findings as I get them tonight and tomorrow. In case you just can’t get enough.