Reply to Trump wants Ann Selzer punished for her Iowa poll Predicition.

Forum Index » Political Discussion

Reply to "Trump wants Ann Selzer punished for her Iowa poll Predicition."

Subject:

Emoticons





More smilies

Text Color: Font: Close Marks

[quote=Anonymous][quote=Anonymous]I am a Harris-supporting Dem and I do think Selzer might have put a finger on the scales in that poll, in retrospect. Though not in the way people are alleging. I think it has to do with how she identifies voters as "likely" and that her method discounts Trump voters who may respond to interviewers questions about whether or not they will vote with sarcasm or caustic negativity. Unlike other pollster's, Selzer doesn't weight her likely voter sampling by looking at things like past voting history. She takes them at their word -- if a voter says they are going to vote, they are a likely voter. If they don't say this or sounds as though they are still not sure if they are going to vote, they are not a likely voter.

This has been an effective way of measuring likely voters in Iowa in the past but Trump voters often have a different posture about voting than supporters of other candidates (of both parties) have had in the past. As a result I think Selzer under-sampled Trump supporters by excluding them on the question of whether or not they were likely to vote, and this resulted in a poll result that greatly overestimated support for Harris in the Iowa electorate.

I also think this sampling error became more pronounced for the poll shortly before the election because Selzer's interviews were likely looking for very strong indications of plans to vote in the week before Election Day, whereas when she polled the state in June and in September (both polls showed Trump well ahead) she likely accepted any indication that they planned to vote (even if lukewarm) as qualifying for them to be included as a likely voter.

Pollsters hate Trump even when they like him because it's very hard to gauge actual support for him because of the nature of his support and his ability to turn out unlikely voters. The polls that "accurately" predicted election results did so by baking in prior experiences with Trump voters in order to weight the polls in his favor. For instance many pollsters have changed how they characterize hang-ups in phone polls -- previously a respondent who said "I'm ******* voting for Trump" and then hung up or terminated a text poll would have been left out of poll results as "nonresponsive." Now most pollsters will list that respondent as voting for Trump even though they technically didn't answer any polling questions and didn't provide answers to questions like how likely they are to vote or how they voted in the last election.

It will be interesting to see what happens when Trump is no longer running, to see if polling firms can regain their footing. I will note here that the polls were very accurate at predicting outcomes of Senate and House races as well as ballot initiatives.

Anyway it is a shame this was Selzer's last poll. She was good at her job but she didn't adjust to the Trump electorate and she paid for that error with this black mark on an otherwise stellar record over nearly 40 years. A shame.

Trump can go suck his thumb in the corner about this one. He will lose this.[/quote]

isn't it amazing what a whiny little b**** he is even after winning an election? i've honestly never met such a whiny little b**** in all my life. amazing anyone thinks he represents masculinity. maybe for an undersocialized four year old.[/quote]

Options

	Disable HTML in this message
	Disable BB Code in this message
	Disable smilies in this message

Review message

Search Recent Topics Hottest Topics