CNN Data Guru Warns Trump Support Crashing: 'Absolute Collapse'

[pat_k]

20. I think that pollsters may also be assuming...

... the percentage of R's is bigger than it is.

The "gold standard" for party affiliation has long been gallup's annual number, which is released at the begining of each year based on interviews conducted over the course of the year.

It is tough to find quarterly numbers. They don't publish them -- at least not directly to the public that I can find. I was VERY surprised to see ABC's first quarter estimates from Gallup in that report.

I suspect they've been seeing a surprising downward trend since the felon took office. I think the first quarter numbers were reported because it has now become clear something "strange' is going on -- i.e., a big exodus. And it's been going on since the felon took office.

The thing is, i'm no expert, but it is my understanding that party affiliation is a very stable thing. It doesn't tend to change much within the same cohort. It does shift as young people are added and older people subtracted over longer periods of time.

So, if, on the assumption party affiliation is pretty stable, pollsters use the annual "gold standard" when they extrapolate from samples, rather than adjusting for every quarter, they are over-estimating his support because they are actually missing the "exodus."