Ahead of the Democratic primary in Virginia, the average poll had Terry McAuliffe getting about 64% among voters who had already decided who to vote for. McAuliffe won 62% tonight. Yet another reminder that the quadrennial freak out over accuracy is dramatically miscalibrated…
Also that a 2-point error on vote share looks a lot different at 64% than 51, but models aren’t going to get you extra juice in competitive elections
To be sure, polls have some big issues! I have written about them. I’m not saying they’re all fine and dandy here — what I’m saying is that the media response to errors can be really disproportionate — misleading, but also misled!

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with G. Elliott Morris

G. Elliott Morris Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @gelliottmorris

10 Jun
Contrary to what you might expect, it’s non-voters, not Trump voters, who are most likely to refuse getting vaccinated for covid-19. This makes for an interesting paradox since voting for Trump is still the strongest predictor of being vax-reluctant. Huh?
The solution is that many of the demographic variables that predict hesitancy *other than* being a Republican also predict non-voting. So the observed non-voting effect is actually a combo of other traits, with huge pools of people, whereas the GOP effect acts alongside them.
The big traits acting here are: not paying attention to the news, not having a degree, being young & living in rural areas. All are correlated with both political activity and hard opposition to the vaccine. (Note I am not counting people with low access as hard vax-reluctant.)
Read 4 tweets
8 Jun
I don’t agree with his decision at all, but the answer might be that Manchin’s looking at better polling data than this
I haven’t seen this End Citizens United poll, but every national poll I’ve found has Republican support for HR 1 — and the provisions in it — way lower than 76%. It’s reasonable to assume support among WV Reps is even lower.

I think HR 1 is a fine bill and that the Senate should pass it.

But if the specific poll being reported on in the underlying MSNBC graphic is anything like this question from the same firm, fielded in March, then I’m inclined to throw it in the trash ecuactionfund.org/wp-content/upl… Image
Read 7 tweets
7 Jun
Democracy Dies in the Senate
To be clear, I do not think HR 1 even comes close to meeting the threats to responsive, representative government that afflict our republic today. But opposing reforms because you want a deal with the minority, anti-democratic faction is not a good way to combat these problems.
Manchin’s opposition to HR 1 could end up being extremely short-sighted, too, depending on how the next 4 years play out. If Rs nuke filibuster in 2022/24, it looks silly in hindsight. Same if R state legislatures and House Reps manufacture a GOP electoral college win in 2024.
Read 6 tweets
3 Jun
This is good but to quibble with @Bencjacobs @dziblatt if GOP leaders were paying attention to the *majority* in polls, they would know the public is against them. The problem comes with abusing polls to reflect the factionalization they have caused. Are the people at fault here?
Yes, that’s the point — if you use polls like you “should” you arrive at a different answer than the GOP is, which means the fault is in how *they’re* analyzing the data
Yeah yeah we all understand each other I’m just on my hobby horse here
Read 4 tweets
2 Jun
Looks like Dems win the NM-01 special by 2-ish pts more than their 2020 margin. The avg swing in specials v 2020/16 is 1-2 pts to Ds.

1 thing ppl might have glossed over is that increasing educational polarization w/ whites could insulate Ds somewhat from the 2010/14 drop-off.
It’s a baseline change, so it’ll change Dem performance on average relative to other factors. How those other factors shake out is really important!
You don’t want to read too much into these specials. But they are, on average, pretty useful for forecasting midterms — just not in isolation. Avg D swing suggests a constant national environment since 2020… So does Congressional ballot polling and Biden’s approval rating polls.
Read 4 tweets
1 Jun
There are (at least) three huge problems with the US & state constitutions re: election subversion. Short thread

1) State legislatures are given power over how elections are conducted and which elections are legitimate—which POTUS electors to send, which contests to certify, &c.
2) Most offices are elected on a plurality-winner district-based electoral system, which incentives zero-sum politics and rewards acting in bad faith and against partisan mutual toleration and forbearance. This is one reason why you have partisan gerrymandering, but also how you
get GA removing full certification power from the Secretary of State, why AZ Senate Rs are voting for a second partisan audit of the 2020 electoral results, which were verified by both bipartisan and non-partisan processes already, and why 75% of R voters Parrott Trump’s big lie.
Read 8 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!