June 15: Multiple polls showing Biden up 9 nationally and one showing him down 1 in Iowa nudged the forecast a bit his way. But overall, things haven’t changed that much over the last couple of weeks. State polls have been showing Biden up big for months. projects.economist.com/us-2020-foreca…
June 16: Lots of polls with Biden up nearly 10 points in swing states today. Combined with yet another dip in Trump's approval ratings, the forecast moved ever so slightly toward the Democrat today. A new Civiqs poll pushed AZ to the left of NC (where it belongs, IMO).
June 17: A batch of state-level Change Research polls may have thrown the model for a bone today were it not for several national polls that were clearly good for Biden. The fallout from the unrest continues to push down Trump's chances of re-election.

June 18: Trump is in a preeetty bad position if we’re arguing over which of the polls showing him down 8, 9, or 12 points is right. FWIW state polls have show a slightly closer race that has hovered around Biden +8 for a while now. Keep an eye on the gap. projects.economist.com/us-2020-foreca…
June 19: A Trump +3 poll in NC did little to shake things up.

Biden's current polling margin in...

SC: -8.1
AK: -5.6
TX: -3.3
IA: -2.6
GA: -0.5

OH: 0.4
NC: 1.1
AZ: 2.9
FL: 4.6
PA: 4.9
WI: 5.4
NH: 5.7
NV: 6.9
MI: 7.0
MN: 8.3

National: 8.6

June 22: We're flying blind over the last week w/o new national polls, so new data in Midwestern states will exert a larger-than-typical sway over the forecast. A Biden +1 poll in MI (probably too low) has pushed Biden down by <1% in WI, OH, PA, for ex.

June 23: No huge changes today. We got two new TX & USA PPP polls that showed a similar race to their last polls.

A friend said the other day that it would be fun if I could be more “emo” w these entries, but sometimes there’s just nothing to report!
June 24: The new Biden +14 NYT poll is definitely on the higher end of what we've seen, didn't move the model much. Polling in Wisconsin pushed Biden up a bit there, but the average in the state is still around +6-7 (consistent with 8-9 nationally). projects.economist.com/us-2020-foreca…
June 25: Today was probably one of Trump’s worst days in the polls. We got a load of surveys that showed him down ~7-8 points in the swing states & 8-9 nationally. It’s going to take a lot for him to win in Nov. That’s not impossible but it is improbable.

June 26: It’s starting to look like the Midwest has shifted left relative to the nation as a whole, with changes in polling numbers outpacing national shifts in states such as MI and WI. That’s a partial explanation for Biden’s improved odds this week.

June 29: Some new polls have pushed Trump underwater in GA & OH. I don't think he'll win them, but this is yet another sign of the president's steep uphill climb to re-election. He could still pull off an upset... but his odds are 2-3x worse than in 2016.

(To clarify: I meant I don’t think BIDEN will win GA and OH — our model currently thinks the race will tighten before November and Trump will regain his footing there. But it could go either way! Both are tossups! A Trump win is just the most likely outcome, not the only one...)
June 30: Some movement to Biden from new Pew and Suffolk polls today. Updated Biden margin in our polling averages:

MO: -8
SC: -7
AK: -5
TX: -3
IA: -2
GA: 0
OH: 1
NC: 2
AZ: 3
FL: 5
PA: 6
WI: 6
NH: 6
NV: 7
MI: 8
MN: 10
ME: 12

National: +9

July 1: At this point it's getting hard to imagine Trump winning—IF our data is right. But even though our model simulates a 12pt interval around today's polls (in margin), it's still at 91% Biden today & entering pretty uncomfortably confident territory. projects.economist.com/us-2020-foreca…
I think it's a good model so I'm standing by it... but 91% is pretty high...
July 2: Today’s promising job numbers did little to change our election-day forecast. That’s bc we’re so far beyond “normal” territory for yearly economic change that the model still just sees a dismal outlook for the incumbent. Biden at +8 for e-day vote. projects.economist.com/us-2020-foreca…
July 3: It’s starting to look like Biden may have plateaued in our forecast. National polls are hanging around +9 & the fundamentals forecast is unchanged. That’s not necessarily bad; every day that goes by without a dip increases his chance of winning.

July 6: Rumor has it the Trump campaign is going to target New Mexico in order to offset their apparent losses in the Midwest and Florida. Well, NM slipped into the 99%+ “Safe Democratic” category in our forecast today. So maybe find a new way to win....

July 7: At this point in the 2016 campaign, Hillary Clinton was at 52% and on a downward trajectory in our model. Contrast that with Joe Biden holding steady around 54-55%, and with a more robust position across the swing states.

July 8: Here are the states our model says will tip the election to either Trump or Biden, and how often that happens:

PA: 20%
FL: 17%
MI: 13%
WI: 10%
NC: 6%
MN: 5%
NH: 4%
AZ: 4%
GA: 4%
TX: 3%

(This is why Trump spending money in Ohio is silly.)

July 9: As of today, the model started incorporating national polls in its "prior" prediction of the popular vote (vs just using "fundamentals"). This will push up Biden's margin a bit in the coming weeks, as polls put him at +10 while fundamental are +7.

July 10: It's tempting to doubt the polls, but let me put Biden's lead this way: even if polls were twice as wrong as they were in 2016 (errors that large are very unlikely, to be sure), he would still probably win the election.

July 13: Our forecast model for the presidential election suggests that Joe Biden needs to win the popular vote by 2.8 percentage points to win the majority of electoral college votes, a very slight dip in the chamber’s pro-Republican bias since 2016. projects.economist.com/us-2020-foreca…
July 14: Today was the first time our model was above 90% for Biden all day. In recent weeks he has really been propelled forward by great (not good, great -- talking implied shifts of 10+ pts nationally) numbers in red states projects.economist.com/us-2020-foreca…
July 15: Need I say more than the linked tweet? One thing our model does (that the site doesn't show you) is calculate how far left/right of the national margin each state leans. Trump has been imploding in MI, PA & FL recently.


July 16: Our model is rapidly approaching uncharted waters. At no point in other 21st-century elections did a candidate lead by so large a national margin 110 days out than Biden does today (+10 points). His state-level position is equally unprecedented.
projects.economist.com/us-2020-foreca… Image
July 17: A new UCLA Nationscape + Democracy Fund poll showed Biden up 9 today. That’s right on the model’s current estimate. Other than that, nothing new to report. Maybe keep an eye on TX? If Biden gets any more favorable polls there it’ll move to tossup. projects.economist.com/us-2020-foreca…
July 20: Every time I say this Biden gets an additional set of promising polls, but nevertheless: Trump's slump has stalled recently. He may be bottoming out. Our forecast appears resistant to hitting Biden +10, probably because of the e-day prior at +7-8. projects.economist.com/us-2020-foreca…
July 21: The average “error” in polls at this point in elections since 1980 is about 8 points. Polls in recent elections have been about twice as accurate. Volatility in 2020 has been closer to the latter, but that’s not a guarantee of a Biden win.

July 22: Our model took in that new Quinnipiac poll in TX but it didn't really change its mind about anything. That's partly bc Q polls have tended to be bluer than the average so far (our model adjusts for this) & bc of some slight tightening nationally.

projects.economist.com/us-2020-foreca… Image
July 23: Several new polls with strong Biden numbers pushed Florida out of our "competitive" categories to a "very likely" Biden victory today. Feels weird to be saying that about Florida, but... here we are. (We might expand the definition.)

(Like arguably "very likely" states are still "competitive" because they are not "safe" any more. Depends on your definition. We might want a more inclusive definition when we're forecasting elections and dealing with things where misses have huge consequences.)
July 24: Nothing really new today. We're spending some time re-doing the way the model handles its historical economic and political fundamentals. It's pretty cool but (as I've been saying) shouldn't change much. We'll share the code soon. Happy weekend! projects.economist.com/us-2020-foreca…
July 27: (Forgot to journal yesterday): There has been some tightening in the race over the last few week, with Biden's national margin slipping about a point off its peak. Our model tries very hard to smooth out noise so I think this is something real.
July 28: I entered 23 new state polls into our spreadsheet today, but the model didn't think they showed a race that was any different. It's always interesting when we get a bunch of data from the bellwethers and nothing changes (nationally, that is). projects.economist.com/us-2020-foreca… Image
July 29: I fed 11 polls into our model today, but nothing really changed. The name of the game right now appears to be stagnation. By this point in past cycles we'd be in convention season, but the campaign is kinda in limbo now and the polls show it.

July 30: We got some much-needed polling from New Hampshire today. It was roughly in line with what we already expected based on the national environment and polls in similar states. Maybe a bit bluer (as has been the case in other northern, white states) projects.economist.com/us-2020-foreca… Image
July 31: It's hard to know what to focus on when the race for POTUS is so steady and there's a severe lack of events. Perhaps this is a good time to recall that campaign effects are usually pretty small anyway, and what we see on TV very rarely matters.

August 3: We got some new polls out of the Midwest and a Fox national survey today. Overall, it looks like the race may have tightened half a point or so over the last month, but we need some more data to be sure. Polls are really all over the place. projects.economist.com/us-2020-foreca…
Apologies — this Shaw & Co and Beacon Research poll wasn’t conducted for Fox (their usual sponsor) but apparently for a different group called the Dream Corps. thedreamcorps.org/wp-content/upl…
August 4: Texas is looking pinker and pinker every day...

projects.economist.com/us-2020-foreca… Image
August 5: Our adjusted polling averages on this day in 2020 and 2016:

National: Biden +10 (Clinton +5)

IA: -1 (5)
GA: 0 (-3)
OH: 0 (4)
TX: 1 (-12)
NC: 2 (-1)
AZ: 5 (-3)
FL: 6 (2)
PA: 7 (5)
WI: 8 (9)
MI: 9 (9)
NV: 9 (4)
MN: 10 (9)
NH: 11 (3)

It is very interesting that Biden is doing 5 points better nationally than Clinton was, but is up an average of just 3 points in swing states.

Another big diff is that our prior at this point in 2016 was for a popular vote tie. Today, our prior is that Biden would win by 7.
Our model’s projected popular vote forecast based on state numbers, and the weighted average of state polls over the last few weeks implies a national margin of about 9.8 points (that’s before doing stuff like pollster, mode, and population adjustments).
August 7: People are talking a lot about tightening in Biden’s polling margin, but our model ain’t showing any. State-level polls have suggested a steady, 9-10-pt lead over the past few weeks, & the higher-quality national polls are also pretty stable.

August 10: A few new polls pushed Trump's odds up just a touch in the Midwest today (and thus in the country as a whole too). As I shared earlier, Biden is doing about 8 points better than Clinton's final 2016 margins both nationally and in the states.

projects.economist.com/us-2020-foreca… Image
There's a very real lack of polling right now, though. You can tell just by looking at the popular vote chart. Compare the mass of polls around the line now to the beginning of June. Image
August 11: Polls from Monmouth and Marquette suggested a race that was tightening just a tiny bit today. We're talking half a point or so. Overall, Biden's polling margins versus the 2016 results still suggest he's by 9pts or so.


August 12: Roughly what our forecast would have said if we ran it on this day in...

2008: Obama +6 nationally +/-9, 199-426 EVs, 84% to win
2012: Obama +4, 189-390 EVs, 80%
2016: Clinton +3, 185-387 EVs, 77%

2020: Biden +9, 218-431 EVs, 88%

The Kamala bump in CA today: ;) ImageImageImage
August 13: We got two new high-quality polls today — from Fox and Pew — that suggest about 0.5-1pt of tightening since a peak in Biden’s margin in late June. But less uncertainty as we get closer to the election offsets a bit of the dip in his odds projects.economist.com/us-2020-foreca… Image
August 14: Here are the states our model thinks are most likely to tip the electoral college to either candidate—ie our "tipping-point index":

PA: 22%
FL: 17
MI: 14
WI: 8
TX: 5
MN: 5
NC: 5
AZ: 4
VA: 4
GA: 4
NV: 3
NM: 3
NH: 2
CO: 2

All else: ≤1%

August 17: Biden heads into the DNC with slightly luffing sails. He has lost about a point of support nationally since June, but is still running ahead of every challenger's position since at least 1948. Trump is running out of time to turn things around. projects.economist.com/us-2020-foreca… Image
Slight correction: I was going off the claim That Biden is doing better than Carter 1976 from this article, but wrong I think. See this series from Gallup — Carter was clearly up by a larger margin than Biden. (So, better than 12 of last 13 challengers...) cnn.com/2020/08/16/pol… Image
August 18: I didn't input any new polls into the model today (at least not yet). The S&P record will be factored in by tomorrow's run, but it won't substantially change anything (the stock market has historically been pretty unrelated to election results). projects.economist.com/us-2020-foreca…
August 19: We’re kind of flying blind in the Midwest right now. What limited polling we have (n=1 in Wisconsin) suggests a largely stable or slightly tighter race, but our model would welcome some new data there after the convention!

projects.economist.com/us-2020-foreca… ImageImageImageImage
August 20: Last wk we got several national polls that suggested tightening & state polls that suggested stability. Today we got the opposite! Biden up 13 natly from Echelon and tied, up 4 in MN, PA from others. (This is why models need measurement error.)

I’ll note that the online pollsters and pollsters who adjust their data for partisanship/past vote have had very stable numbers for the last 3 months. YouGov especially, which since May has been at Biden +8-10 on average after every new survey. That tells you something...
One thing I’m fairly confident about is that using polls in a model requires a fairly up-to-date reading of literature on data quality, and if you’re not letting the model know that eg some pollsters weight by party, some don’t weight by edu then you’re prone to larger errors.
(That goes for modelers who use polls from MTurk without adjustment, too)
August 21: Some material tightening in the Midwest over the past couple of days projects.economist.com/us-2020-foreca…
August 24: The RNC starts as Trump is down 9-10 points in the polls, about twice as large as his deficit at this point in 2016. I’m watching the programming this week for clues at how he’s trying to turn things around (and not seeing much yet).

August 25: Today's model updates gave Joe Biden some of his best numbers of the campaign—a 10 point lead nationally and +7 in PA, the likeliest "tipping-point" state. RNC speakers have so far shown little sign that Trump knows how to turn things around. projects.economist.com/us-2020-foreca… ImageImageImageImage
August 26 (late): Biden’s chances in model ticked down a hair today on the back of a few only so-so polls for the former VP. projects.economist.com/us-2020-foreca…
August 27: Our model shows Biden is running ahead of Clinton in the electoral college, but how voters react to the DNC and RNC over the next two weeks will be a big tell about how robust that lead is. projects.economist.com/us-2020-foreca…
August 28: Today brought both a few new online polls and new “fundamentals” data for our forecast. The former shows stability nationally and in the swing states, and the latter an ever-so-slightly improved election-day prediction for Trump (<0.1 points). projects.economist.com/us-2020-foreca…
(There was also the Trafalgar [IVR] poll in Michigan, which people shouldn't discard but also shouldn't put too much weight on, and an Emerson College poll in Massachusetts that we're not counting because of demonstrable data quality concerns)
August 31: After ingesting seven new polls today, our model thinks the race has moved ~1pt to Trump over the last week, landing where it was before Biden's modest DNC bounce. We should get more (& better) polls tomorrow.

Biden is still the clear favorite. projects.economist.com/us-2020-foreca…
For the sake of keeping my model journal updated in this thread, here are some updates for Sept 1 & 2, which didn't make it in on time...
Sept 1: Trump's position in the electoral college *HAS* improved over the last 2 months. Relative to his June-Jul low, his numbers are up maybe 2 points on margin. But with little time left to make up ground, our model still sees him as the clear underdog projects.economist.com/us-2020-foreca… ImageImageImageImage
Sept 2: Our model has taken in 3 new Fox polls, and the result is stability. Biden is +9 nationally (in the two-party vote), about where he has been over the past month. The Midwest has tightened a bit v June/July, but that started before the RNC/Kenosha.

Sept 3: The big news today came in FL, where a new poll from Quinnipiac showed Biden with just a 3-point lead. That, plus yesterday's higher POTUS approval ratings pushed Biden's odds of winning down by 2 points even tho national averages remain unchanged. projects.economist.com/us-2020-foreca… Image
Sept 4: The polls haven't tightened all that much, if at all, nationally, but (a) close swing-state polls, (b) good economic data & (c) higher POTUS approval polls have all pushed our forecast of Biden's chances into the low 80s over the last week.

Sept 8: Despite all of the hullabaloo about that Marist poll in Florida today, there was no significant movement in our forecast. Trump enters the fall campaign 8 points down nationally and behind by 5 in the tipping-point states—down (big), but not out. projects.economist.com/us-2020-foreca…
Let this be a reminder that all the breathless punditry and overreactions over one poll that you see on Twitter are just that. Evolutions in the campaign tend to happen steadily over time, usually evidence by multiple polls and across multiple states.
Both Florida and the national race are closer than they were in June/July. But I don't see much evidence that FL has gone from a 2-3 point race to a 1-2-point one over the last couple of weeks. (Though I will certain proclaim such tightening if our model reveals it.)
Sept 9: A batch of 20 new polls today prompted a slight reversion toward Biden in our model, but broadly speaking the polls are about the same as they were 1-2 weeks ago — though tightening in FL is one clear exception. projects.economist.com/us-2020-foreca…
Sept 10: After some modest tightening in August (esp the past 3 weeks), the presidential race seems to have settled in around Biden +8 for now. Here is his estimated support in the 4 likeliest "tipping-point" states. (FL & PA are make-or-break for Trump.)
projects.economist.com/us-2020-foreca… ImageImageImageImage
Sept 14: Not a lot going on in the forecast over the last week. This is perhaps a good time to remind you that other models should converge toward ours, which has (correctly, so far) forecast much less #TIGHTENING relative to others, as Nov approaches. projects.economist.com/us-2020-foreca…
Just to be clear, we *do* forecast the race to tighten by about 1 point on average between now and election day, though things can obviously deviate from that path. We explore a slightly smaller range of outcomes v other forecasters.

projects.economist.com/us-2020-foreca… Image
That smaller uncertainty is mostly caused by the fact that we account for the role that polarization has played in decreasing volatility in modern election cycles, a finding that others have ignored but has (so far!) turned out to be right.
We include some extra uncertainty as a safeguard against the relationship between polarization and volatility breaking down, but based on what we've seen over the course of 2020 so far, I'm not necessarily sweating the (statistical) assumptions we've made with our model.
Sept 15: Biden had one of his best days in the polls in a while today, though nothing was different enough from state averages to move any needles significantly. Every day that passes without big movement toward Trump will decrease his odds of winning.

Sept 16: Above-average polling for Biden today has nearly restored his position in the electoral college to its pre-convention high point. Florida remains a weak spot in an otherwise robust grouping of states that would very likely put him over the edge.

• • •

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

Nov 17
trafalgar group has the highest empirical bias relative to other pollster of over 500 firms we have data from. literally the most biased pollster in america. and we warned people that that bias could carry forward and mess with averages in a good year for polls. we were right!!
if someone has a bad record, is opaque about their methods and then tells you they “want to be the elon musk of” that thing, that’s a 🚩 🚩
FWIW i think rob cahaly’s explanation is probably wrong. it’s at least as plausible IMO that his heavily R-leaning stratified sample is at fault. but the point is really that there’s no way to know unless he provides methods details about how this stuff works — and he doesn’t!
Read 6 tweets
Nov 11
Looking at these newest Senate results and tbh if I’m a news network decision desk, I’m probably getting pretty close to calling AZ for the Dems
and honestly NV is looking good too, but it’s a lot closer and we don’t have super reliable data on the number of outstanding ballots in the counties around Las Vegas and Reno
If I'm doing my math right, if this keeps up (and depending on Clark tallies) CCM will win by a little under 2 points
Read 4 tweets
Nov 10
Polls just had one of their best elections ever. Based on our (preliminary) estimates, our averages in competitive states have an absolute error of just 2 points on margin — a little under half the expected error — with average *bias* undershooting Dems.

This chart shows bias in each state in 2022 vs 2020. Take out FL and there is a uniform decrease in bias across states. Good!
Though if you're looking for reasons to be anxious: much like 2018 there is still a positive slope here, indicating *some* persistent partisan nonresponse
In competitive Senate races with more than 2 polls, this is the second-best performance for pollsters since 1998
Read 4 tweets
Nov 10
this is a completely useless exercise if you’re not imputing results in uncontested seats. you do not capture structural advantages when aggregate totals are missing like 50 seat observations!
One issue is that the popular vote is much likelier to end up < R+1 than at R+2 or more. See here
Another error is that you can’t use House election results to write about structural advantages because there is so much else going on in the vote totals (even if you impute uncontested seats) that is not about *structure*.
Read 4 tweets
Oct 30
If you adjust for the percent of votes outstanding in each state in Brazil right now*, you get a result of 50.1 Lula - 49.9 Bolsonaro --
which is to say, it looks like Lula could still pull this off, but it's going to be another close round!
The * here is that doing this assumes that state voting patterns will be constant as new votes come in, which is not true. So it's an estimate with a wide margin of error!
Now getting closer to a 1-pt victory for Lula. (I don't have precinct-level results but if you did the same thing at that level, I'm sure you'd get a different result)
Read 5 tweets
Oct 28
gun to my head i would bet that polls are overestimating democrats again, rather than republicans. but such predictions don’t have a great historical track record…
still the clues here are mostly constant across polls and have helped anticipate recent errors (low n tho)
remember, registered republicans were more likely than democrats to answer polls for much of 2020, including during the summer boost in response rates from Dem activists — so even well designed polls that stratify &/or weight by party don’t solve nonresponse by vote intention!
i said “gun to my head” for a reason!
Read 4 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

Don't want to be a Premium member but still want to support us?

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

Donate via Paypal

Or Donate anonymously using crypto!


0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy


3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us on Twitter!