This is the website of Abulsme Noibatno Itramne (also known as Sam Minter). Comments here or emails to me at abulsme@abulsme.com are encouraged... or follow me on Twitter as @abulsme.

Categories

Calendar

September 2016
S M T W T F S
 123
45678910
11121314151617
18192021222324
252627282930  

Electoral College: Has Trump Peaked?

States with new poll data added since the last update: All 50 states, DC, and the Maine congressional districts. Notable changes in Kansas, New Hampshire, New Mexico, and Maine’s 2nd Congressional District.

National Picture

The update today is a strange one. Some indicators move toward Clinton, others move toward Trump. Some of the moves seem to be prompted by outlier data as well, so some skepticism is probably in order. But with the directional mix, one possible conclusion is that we are at or near the top of the recent move toward Trump. As I write this though, we are hours away from the first debate, which may well scramble things again.

The summary changes:

  • The expected case moves from Clinton by 6 EV to Clinton by 8 EV
  • Clinton’s best case improves from Clinton by 178 EV to Clinton by 188 EV
  • Trump’s best case improves from Trump by 98 EV to Trump by 106 EV
  • The tipping point moves from Clinton by 0.8% in NM to Clinton by 0.1% in KS

So two moves in Clinton’s direction (Expected and Clinton Best), and two moves in Trump’s direction (Trump Best and Tipping Point.)

The charts look like this:

chart-298

chart-299

I’ve added a new vertical line marking the moment when Trump shook up his campaign by bringing in Steve Bannon and Kellyanne Conway. Looking back, we can now see this very closely aligns to the moment Trump stopped dropping and started closing the gap with Clinton.

Trump has still not led in my model, but this is about as close as you can get without crossing that line.

The fact that we are now seeing simultaneous moves in both directions might indicate we are near the top of this particular trend toward Trump. But events have the potential to change things quickly… in either direction.

43.4 days until polls start to close on election day…

State Details

For those who are interested in the individual state moves that combine for the national changes above, here they are. I will only add individual commentary when there is something specific worth noting.

Flipping to Clinton

chart-300

I’m putting this one first, because it is one I honestly don’t believe. Every single one of the results showing Clinton ahead since August is from one pollster… Google Consumer Surveys. This pollster has some unorthodox methods, and has produced strings of outliers in a wide variety of states… some favoring Trump, others favoring Clinton. In this case, other than a single Zogby poll in June, every other pollster shows Kansas not only favoring Trump, but favoring Trump by a substantial margin.

Because I am inclusive and try not to exclude polls, three of the five polls in my average in Kansas are from Google at the moment. This leads to a 0.1% Clinton lead in Kansas. If you excluded the Google polls, my average would show Trump leading by 11.6%. I am very very dubious that Kansas is even a close state, let alone that Clinton is ahead. We need more non-Google polls in Kansas to clear up this situation.

Now, astute observers will note that if you flip Kansas back to red, Trump would actually now be leading in the “expected” result. But not so fast. Google causes changes in other states too. To pick out one prominent example, Google has been giving Trump much bigger leads than other pollsters in Florida. My average in Florida currently shows Trump up by 1.1%, but if you removed Google, the average moves to Clinton by 1.3%. So removing Google might flip Kansas, but it would also flip Florida. You’d still have a Clinton lead overall, just in a different way.

The philosophy of this site is to just include everything unless is it obviously fake or fraudulent. The Google polls may be bad… they are just all over the place compared to more traditional pollsters… but they are not an obvious fake or fraud. So I will continue to include them. I don’t pick and choose polls based on if I “believe” them. Google is in the average. So you end up with a very slight Clinton lead… in Kansas.

But Google is clearly adding noise to the model at the moment… which I’m not particularly thrilled about.

Note that Kansas moved all the way from Strong Trump to Weak Clinton in one jump, so it improved both Clinton’s expected case and her best case.

Flipping to Trump

chart-301

This is another state where three of five data points in my average are from Google, and this time the other two are from Reuters/Ipsos, another “50 state” pollster where the numbers can sometimes be way out of line with other polls. In this case Google and Ipsos seem to trend in opposite directions, but it would be really nice to have some other big name pollsters check out New Mexico again before too long, because when all you have are these polls looking at the whole country and then breaking out subsamples by state, you really have a lot of uncertainty on the results. So like Kansas, treat New Mexico with a substantial pile of salt.

Now in reach for Trump

chart-302

No longer in reach for Clinton

chart-303

Interview

The latest Curmudgeon’s Corner podcast presented interviews with three people who do independent electoral college projections based on state polls… Luis Mendez from projectiondesk.com, Darryl Holman from horsesass.org, and me… Sam Minter from electiongraphs.com. We each talked about how we got into this sort of analysis, what makes our sites different, the challenges we face, and more. Take a listen!

For more information…

This post is an update based on the data on ElectionGraphs.com. Election Graphs tracks both a poll based estimate of the Electoral College and a numbers based look at the Delegate Races. All of the charts and graphs seen in this post are from that site. Additional graphs, charts and raw data can be found there. Follow @ElectionGraphs on Twitter or like Election Graphs on Facebook to see announcements of updates or to join the conversation. For those interested in individual general election poll updates, follow @ElecCollPolls on Twitter for all the polls as they are added. If you find the information in these posts interesting or useful, please consider visiting the tip jar.

@ElecCollPolls tweets from 2016-09-25 (UTC)

  • 02:34:39 Poll Added: Google w/3P in NJ from 2016-09-14 to 2016-09-20 – Clinton 35.60% to Trump 36.19% https://t.co/QqP3qDpgOw
  • 02:34:45 Full 3P results: Google w/3P in NJ from 2016-09-14 to 2016-09-20 – Clinton 35.60% to Trump 36.19% to Johnson 5.48%
  • 02:39:43 Poll Added: Google w/3P in NM from 2016-09-14 to 2016-09-20 – Clinton 30.14% to Trump 29.76% https://t.co/KdHmQkyQ2Q
  • 02:39:48 Full 3P results: Google w/3P in NM from 2016-09-14 to 2016-09-20 – Clinton 30.14% to Trump 29.76% to Johnson 17.90%
  • 02:40:02 Clinton vs Trump state category change: NM has moved from Weak Clinton to Weak Trump https://t.co/KdHmQkyQ2Q
  • 02:54:12 Clinton vs Trump tipping point change: Clinton by 0.8 in NM -> Clinton by 0.1 in KS https://t.co/lyZIv8GQxx
  • 02:54:13 Clinton vs Trump expected case changed: Clinton 278 to Trump 260 -> Clinton 273 to Trump 265 https://t.co/lyZIv8GQxx
  • 03:04:44 Poll Added: Google w/3P in NV from 2016-09-14 to 2016-09-20 – Clinton 32.23% to Trump 33.54% https://t.co/9TiErrtNaj
  • 03:04:49 Full 3P results: Google w/3P in NV from 2016-09-14 to 2016-09-20 – Clinton 32.23% to Trump 33.54% to Johnson 15.06%
  • 03:17:06 Poll Added: Google w/3P in NY from 2016-09-14 to 2016-09-20 – Clinton 41.56% to Trump 28.23% https://t.co/noeC7NWY3h
  • 03:17:11 Full 3P results: Google w/3P in NY from 2016-09-14 to 2016-09-20 – Clinton 41.56% to Trump 28.23% to Johnson 6.76%
  • 05:19:46 Poll Added: Google w/3P in OH from 2016-09-14 to 2016-09-20 – Clinton 33.76% to Trump 32.71% https://t.co/jX5kqhinzT
  • 05:19:51 Full 3P results: Google w/3P in OH from 2016-09-14 to 2016-09-20 – Clinton 33.76% to Trump 32.71% to Johnson 7.79%
  • 05:25:30 Poll Added: Google w/3P in OK from 2016-09-14 to 2016-09-20 – Clinton 27.82% to Trump 44.20% https://t.co/UVY7QUWdfp
  • 05:25:35 Full 3P results: Google w/3P in OK from 2016-09-14 to 2016-09-20 – Clinton 27.82% to Trump 44.20% to Johnson 8.45%
  • 05:55:48 Poll Added: Google w/3P in OR from 2016-09-14 to 2016-09-20 – Clinton 40.19% to Trump 28.11% https://t.co/C4mGGTMTCX
  • 05:55:53 Full 3P results: Google w/3P in OR from 2016-09-14 to 2016-09-20 – Clinton 40.19% to Trump 28.11% to Johnson 7.93%
  • 06:01:38 Poll Added: Google w/3P in PA from 2016-09-14 to 2016-09-20 – Clinton 34.81% to Trump 36.05% https://t.co/lRQ6ICvBZS
  • 06:01:43 Full 3P results: Google w/3P in PA from 2016-09-14 to 2016-09-20 – Clinton 34.81% to Trump 36.05% to Johnson 7.56%
  • 18:09:06 Poll Added: Google w/3P in RI from 2016-09-14 to 2016-09-20 – Clinton 46.40% to Trump 30.90% https://t.co/DVPdKlLx8y
  • 18:09:11 Full 3P results: Google w/3P in RI from 2016-09-14 to 2016-09-20 – Clinton 46.40% to Trump 30.90% to Johnson 8.84%
  • 18:16:19 Poll Added: Google w/3P in SC from 2016-09-14 to 2016-09-20 – Clinton 33.23% to Trump 40.95% https://t.co/8A8ZHLs8Ov
  • 18:16:24 Full 3P results: Google w/3P in SC from 2016-09-14 to 2016-09-20 – Clinton 33.23% to Trump 40.95% to Johnson 7.28%
  • 18:24:19 Poll Added: Google w/3P in SD from 2016-09-14 to 2016-09-20 – Clinton 31.03% to Trump 30.11% https://t.co/jWJggE2w5l
  • 18:24:24 Full 3P results: Google w/3P in SD from 2016-09-14 to 2016-09-20 – Clinton 31.03% to Trump 30.11% to Johnson 13.38%
  • 18:32:33 Poll Added: Google w/3P in TN from 2016-09-14 to 2016-09-20 – Clinton 28.63% to Trump 41.49% https://t.co/MdleB52MeU
  • 18:32:38 Full 3P results: Google w/3P in TN from 2016-09-14 to 2016-09-20 – Clinton 28.63% to Trump 41.49% to Johnson 9.14%
  • 18:40:11 Poll Added: Google w/3P in TX from 2016-09-14 to 2016-09-20 – Clinton 28.94% to Trump 42.35% https://t.co/obkqHjbWX6
  • 18:40:16 Full 3P results: Google w/3P in TX from 2016-09-14 to 2016-09-20 – Clinton 28.94% to Trump 42.35% to Johnson 7.62%
  • 18:46:23 Poll Added: Google w/3P in UT from 2016-09-14 to 2016-09-20 – Clinton 27.93% to Trump 24.07% https://t.co/vcUI7G27bE
  • 18:46:28 Full 3P results: Google w/3P in UT from 2016-09-14 to 2016-09-20 – Clinton 27.93% to Trump 24.07% to Johnson 8.38%
  • 18:46:41 Clinton vs Trump state category change: UT has moved from Solid Trump to Strong Trump https://t.co/vcUI7G27bE
  • 18:55:03 Poll Added: Google w/3P in VA from 2016-09-14 to 2016-09-20 – Clinton 36.26% to Trump 36.84% https://t.co/laE3AtODZg
  • 18:55:08 Full 3P results: Google w/3P in VA from 2016-09-14 to 2016-09-20 – Clinton 36.26% to Trump 36.84% to Johnson 7.64%
  • 19:01:07 Poll Added: Google w/3P in VT from 2016-09-14 to 2016-09-20 – Clinton 43.99% to Trump 18.62% https://t.co/wMilBaw6pK
  • 19:01:12 Full 3P results: Google w/3P in VT from 2016-09-14 to 2016-09-20 – Clinton 43.99% to Trump 18.62% to Johnson 10.38%
  • 19:09:36 Poll Added: Google w/3P in WA from 2016-09-14 to 2016-09-20 – Clinton 36.97% to Trump 28.20% https://t.co/LcbPUSYIY3
  • 19:09:41 Full 3P results: Google w/3P in WA from 2016-09-14 to 2016-09-20 – Clinton 36.97% to Trump 28.20% to Johnson 10.83%
  • 19:15:33 Poll Added: Google w/3P in WI from 2016-09-14 to 2016-09-20 – Clinton 42.22% to Trump 24.49% https://t.co/D2RCcF0rWj
  • 19:15:38 Full 3P results: Google w/3P in WI from 2016-09-14 to 2016-09-20 – Clinton 42.22% to Trump 24.49% to Johnson 8.14%
  • 19:22:13 Poll Added: Google w/3P in WV from 2016-09-14 to 2016-09-20 – Clinton 19.42% to Trump 35.52% https://t.co/4PV8AG3iZn
  • 19:22:18 Full 3P results: Google w/3P in WV from 2016-09-14 to 2016-09-20 – Clinton 19.42% to Trump 35.52% to Johnson 10.50%
  • 19:28:46 Poll Added: Google w/3P in WY from 2016-09-14 to 2016-09-20 – Clinton 30.13% to Trump 45.26% https://t.co/KDFGK1cYpf
  • 19:28:51 Full 3P results: Google w/3P in WY from 2016-09-14 to 2016-09-20 – Clinton 30.13% to Trump 45.26% to Johnson 3.58%
  • 20:54:18 Poll Added: YouGov w/4P in CO from 2016-09-21 to 2016-09-23 – Clinton 40% to Trump 39% https://t.co/ltx8c058ml
  • 20:54:23 Full 4P results: YouGov w/4P in CO from 2016-09-21 to 2016-09-23 – Clinton 40% to Trump 39% to Johnson 7% to Stein 2%
  • 20:57:49 Poll Added: YouGov w/4P in VA from 2016-09-21 to 2016-09-23 – Clinton 45% to Trump 37% https://t.co/3Etyah3iLi
  • 20:57:54 Full 4P results: YouGov w/4P in VA from 2016-09-21 to 2016-09-23 – Clinton 45% to Trump 37% to Johnson 7% to Stein 1%
  • 21:00:15 Poll Added: YouGov w/4P in MO from 2016-09-21 to 2016-09-23 – Clinton 37% to Trump 46% https://t.co/oOXFmNsZWD
  • 21:00:20 Full 4P results: YouGov w/4P in MO from 2016-09-21 to 2016-09-23 – Clinton 37% to Trump 46% to Johnson 5% to Stein 2%
  • 21:00:33 Clinton vs Trump state category change: MO has moved from Weak Trump to Strong Trump https://t.co/oOXFmNboy3
  • 21:08:21 Poll Added: UNH in ME-CD2 from 2016-09-15 to 2016-09-20 – Clinton 34% to Trump 48% https://t.co/YJyY0Nd14q
  • 21:08:37 Clinton vs Trump state category change: ME-CD2 has moved from Weak Trump to Strong Trump https://t.co/YJyY0Nd14q
  • 21:19:51 Clinton best case vs Trump has changed: Clinton 374 to Trump 164 -> Clinton 363 to Trump 175 https://t.co/0AnjoTLmu9
  • 21:29:47 BTW, that should have been 2 separate changes. 374-164 to 364-174 for MO, then to 363-175 for ME-CD2. I went too fast and they combined. :-) in reply to ElecCollPolls
  • 21:32:33 Poll Added: UNH in ME-CD1 from 2016-09-15 to 2016-09-20 – Clinton 50% to Trump 28% https://t.co/VyfKbN2TfC
  • 21:37:56 Poll Added: UNH w/4P [2] in ME-All from 2016-09-15 to 2016-09-20 – Clinton 40% to Trump 36% https://t.co/JMEEqBsdPA
  • 21:38:01 Full 4P results: UNH w/4P [2] in ME-All from 2016-09-15 to 2016-09-20 – Clinton 40% to Trump 36% to Johnson 12% to Stein 3%
  • 21:39:53 Poll Added: UNH [2] in ME-All from 2016-09-15 to 2016-09-20 – Clinton 42% to Trump 37% https://t.co/RiEBtjomaM
  • 21:46:16 Poll Added: Hinckley w/4P in UT from 2016-09-12 to 2016-09-19 – Clinton 25% to Trump 34% https://t.co/ZkqANiNRpE
  • 21:46:21 Full 4P results: Hinckley w/4P in UT from 2016-09-12 to 2016-09-19 – Clinton 25% to Trump 34% to Johnson 13% to Stein 1%
  • 21:46:49 Note the Hinckley Utah poll also included McMullin, who came in at 12%. in reply to ElecCollPolls
  • 21:52:40 Poll Added: Just Win w/3P in WV from 2016-09-08 to 2016-09-10 – Clinton 30% to Trump 57% https://t.co/vBfPJJBbQS
  • 21:52:45 Full 3P results: Just Win w/3P in WV from 2016-09-08 to 2016-09-10 – Clinton 30% to Trump 57% to Johnson 4%
  • 22:09:27 Poll Added: Muhlenberg [3] in PA from 2016-09-19 to 2016-09-23 – Clinton 44% to Trump 41% https://t.co/zBVhdErZDv
  • 22:10:48 Poll Added: Muhlenberg w/Lean [3] in PA from 2016-09-19 to 2016-09-23 – Clinton 44% to Trump 41% https://t.co/rYqBuiIuDN
  • 22:13:08 Poll Added: Muhlenberg w/4P [3] in PA from 2016-09-19 to 2016-09-23 – Clinton 40% to Trump 38% https://t.co/lAoyE8GVKJ
  • 22:13:13 Full 4P results: Muhlenberg w/4P [3] in PA from 2016-09-19 to 2016-09-23 – Clinton 40% to Trump 38% to Johnson 8% to Stein 3%
  • 22:17:56 Poll Added: JMC w/4P in GA from 2016-09-20 to 2016-09-22 – Clinton 38% to Trump 44% https://t.co/ecTnW76XLl
  • 22:18:01 Full 4P results: JMC w/4P in GA from 2016-09-20 to 2016-09-22 – Clinton 38% to Trump 44% to Johnson 5% to Stein 1%
  • 22:20:59 Poll Added: Landmark w/3P in GA from 2016-09-21 to 2016-09-22 – Clinton 42.9% to Trump 46.5% https://t.co/kgSAyZEXP6
  • 22:21:04 Full 3P results: Landmark w/3P in GA from 2016-09-21 to 2016-09-22 – Clinton 42.9% to Trump 46.5% to Johnson 5.7%
  • 23:11:30 Poll Added: GQR w/4P [3] in NC from 2016-09-10 to 2016-09-19 – Clinton 41% to Trump 39% https://t.co/8rF2MkEnU4
  • 23:11:35 Full 4P results: GQR w/4P [3] in NC from 2016-09-10 to 2016-09-19 – Clinton 41% to Trump 39% to Johnson 9% to Stein 3%
  • 23:12:10 Poll Added: GQR w/Lean w/4P [3] in NC from 2016-09-10 to 2016-09-19 – Clinton 44% to Trump 40% https://t.co/i2bdrhzSPC
  • 23:12:15 Full 4P results: GQR w/Lean w/4P [3] in NC from 2016-09-10 to 2016-09-19 – Clinton 44% to Trump 40% to Johnson 9% to Stein 3%
  • 23:14:25 Poll Added: GQR [3] in NC from 2016-09-10 to 2016-09-19 – Clinton 48% to Trump 45% https://t.co/vfYDl8wtQp
  • 23:18:17 Poll Added: GQR w/4P [3] in NV from 2016-09-10 to 2016-09-19 – Clinton 42% to Trump 41% https://t.co/FN532GTnl5
  • 23:18:22 Full 4P results: GQR w/4P [3] in NV from 2016-09-10 to 2016-09-19 – Clinton 42% to Trump 41% to Johnson 6% to Stein 2%
  • 23:20:18 Poll Added: GQR w/Lean w/4P [3] in NV from 2016-09-10 to 2016-09-19 – Clinton 42% to Trump 42% https://t.co/5tLOL4I2fZ
  • 23:20:23 Full 4P results: GQR w/Lean w/4P [3] in NV from 2016-09-10 to 2016-09-19 – Clinton 42% to Trump 42% to Johnson 7% to Stein 3%
  • 23:21:58 Poll Added: GQR [3] in NV from 2016-09-10 to 2016-09-19 – Clinton 49% to Trump 46% https://t.co/0OO3UU2JN1
  • 23:25:06 Poll Added: GQR w/4P [3] in OH from 2016-09-10 to 2016-09-19 – Clinton 38% to Trump 39% https://t.co/IZ1MIshHjh
  • 23:25:11 Full 4P results: GQR w/4P [3] in OH from 2016-09-10 to 2016-09-19 – Clinton 38% to Trump 39% to Johnson 10% to Stein 4%
  • 23:26:46 Poll Added: GQR w/Lean w/4P [3] in OH from 2016-09-10 to 2016-09-19 – Clinton 39% to Trump 41% https://t.co/wnRn1FcOJx
  • 23:26:51 Full 4P results: GQR w/Lean w/4P [3] in OH from 2016-09-10 to 2016-09-19 – Clinton 39% to Trump 41% to Johnson 11% to Stein 4%
  • 23:29:03 Poll Added: GQR [3] in OH from 2016-09-10 to 2016-09-19 – Clinton 46% to Trump 46% https://t.co/DtAUiUp3w1
  • 23:32:12 Poll Added: GQR w/4P [3] in PA from 2016-09-10 to 2016-09-19 – Clinton 45% to Trump 37% https://t.co/ecABISZgNt
  • 23:32:17 Full 4P results: GQR w/4P [3] in PA from 2016-09-10 to 2016-09-19 – Clinton 45% to Trump 37% to Johnson 8% to Stein 3%
  • 23:33:53 Poll Added: GQR w/Lean w/4P [3] in PA from 2016-09-10 to 2016-09-19 – Clinton 46% to Trump 38% https://t.co/oZzUyHwfi6
  • 23:33:58 Full 4P results: GQR w/Lean w/4P [3] in PA from 2016-09-10 to 2016-09-19 – Clinton 46% to Trump 38% to Johnson 8% to Stein 3%
  • 23:35:48 Poll Added: GQR [3] in PA from 2016-09-10 to 2016-09-19 – Clinton 51% to Trump 42% https://t.co/noOKL4yZ7o

@abulsme tweets from 2016-09-25 (UTC)