July 15 Poll Report

            Good morning everybody.   Two big notes: 

(1)  There was a shocking result from yesterday’s poll, and

(2)  The campaign has entered a fascinating and unexpected phase.

The shocking result from yesterday’s poll was that Howard Schultz got 14% of the vote in a contest against three sitting United States Senators—Elizabeth Warren, Amy Klobuchar and Cory Booker.  We would have predicted, based on previous polling, that Schultz would get 4%.   With 1400 people responding to the poll, more or less (1,391), that means he got about 140 votes more than we would have expected him to get.

The most obvious explanation for this anomalous result would be that (a) the poll got re-tweeted into a nest of Schultz supporters, or (b) there was some other organized effort to get Schultz supporters to respond to the poll.  There are other possible explanations; Schultz may have received a nice write-up in a magazine or on a web site somewhere that I didn’t see, or he may have done an interview somewhere in which he did well; I don’t know.  Whatever it was, it was one of the most unexpected outcomes of a vote that I have had in almost 100 of these polls, and—combined with another outcome—it shakes up the standings to a meaningful extent. 

These are the results from yesterday’s poll:

            To explain a little more, there are three positioning points for each candidate from each poll.  Schultz getting 14% vs. Warren getting 62% indicates that, of the total support for Schultz and Warren, Schultz should get about

14/76ths, and Warren about 62/76ths.   The current estimate of Schultz’ support (after this poll is included) is 166 points; the current estimate for Warren is 1809 points, a total of 1975 points, or 1,975 voters out of each 10,000.  Out of those 1,975 points, Schultz gets 14/76ths, or 364, and Warren gets 62/76ths, or 1611.  These are "position estimates".  These position estimates are then combined with all of the other position estimates that we have for those candidates.  We have 20 other position estimates for Schultz, 23 others for Warren, so Schultz’s 364 positioning estimate doesn’t pull him up to 364, but it does pull him up some meaningful distance. 

            Before yesterday, we had 18 position estimates for Schultz, which ranged from 70 to 149.   The three position estimates resulting from yesterday’s poll are 294 (vs. Booker), 301 (vs. Klobuchar) and 364 (vs. Warren).   This also causes us, to an extent, to re-evaluate the earlier polls, and the result of that is that Schultz’ Support Scores leaps from 109 (yesterday) to 166 (today).   It is one of the largest one-day leaps in the Support Scores since I have been doing this, and all the more shocking because Schultz was the only candidate in the system who was marked in Gray yesterday, meaning that his Support Score had declined 25% in the previous month. 

            Also yesterday, we removed from the data the poll of May 25, which was Kamala Harris (55%), Bernie Sanders (23%), Andrew Yang (16%) and Steve Bullock (7%).  The removal of that poll is also a fairly high-impact event, since Yang is much stronger now then he was then.  That poll showed Yang’s support at 29% of Harris’s support; it is now over 50% of Harris’ support.  That poll showed Yang behind Sanders; in recent polling he is ahead of Sanders.   So removing that poll pushes Yang up 51 points, almost as many as Schultz is up, although it isn’t anything like as large a gain as a percentage of the candidate’s standing. 

            The combined impact of Schultz moving up 57 points and Yang moving up 51 points rattles the standings to a somewhat unusual extent.  Of the 26 candidates being polled, 15 are either up or down far enough today that we would ordinarily take note of it—as opposed to last Saturday, when only two candidates were up or down enough to make a note of it.  It is secondary effects, which I explain in reference to sports’ teams.   Suppose that, in the first week of the 2019 NFL season, the Washington Redskins were beat the Philadelphia Eagles, in Philadelphia and with a healthy Carson Wentz.   That would be a very surprising result, as the Eagles are heavy favorites, and we would mark the Eagles way, way down because they got beat at home by an inferior team.  But then suppose that the Redskins beat the Cowboys in the second week of the season, and beat the Bears in the third week.   Then we’d have to say "Oh; I guess the Redskins are just better than we thought they were."   But that means, then, that the Eagles’ loss to the Redskins in the first week was not, in fact, as shocking as it appeared to be at the time.   We have to evaluate the previous game—the Eagles’ loss to the Redskins—in light of the subsequent information.

            The same thing in the polls; mathematically, we re-evaluate all of the polls each day in light of the new information that has been added to the system.   That makes other candidates look a little stronger or a little weaker than we thought they were.   Since yesterday:

            Howard Schultz is up 57 points as a result of yesterday’s poll.

            Andrew Yang is up 51 points as a result of the removal from the relevant data of the poll of May 25.

            John Delaney is up 22 points as a secondary effect, a re-evaluation of the previous polls in light of the new evidence, and also Bill de Blasio is up 19, Marianne Williamson up 9, Julian Castro up 8 and Michael Bennet up 6, all of those being secondary effects. 

            But if somebody is up, somebody else must be down, so also, since yesterday:

            Donald Trump is down 6 points as a secondary effect.

            Bernie Sanders is down 15 as a result of the removal of the May 25^th poll from the data.

            Pete Buttigieg is down 17 as a secondary effect, and also Joe Biden is down 18.

            Amy Klobuchar is down 23 points as a result of yesterday’s poll.

            Cory Booker is down 26 points as a result of his horrendous performance in yesterday’s poll, finishing well below Howard Schultz.  Booker went into the poll at 343, Schultz at 109.  Schultz beat Booker.  

            Kamala Harris is down 34 points as a result of the removal of the May 25^th poll.  This confirms, as many other polls have shown, that Senator Harris did NOT get a major bump out of attacking Joe Biden in the June 28 debate.

            And Elizabeth Warren is down 56 points as a result of yesterday’s poll.

            These are the updated standings:

            Andrew Yang has passed Bernie Sanders in this poll, and Booker has fallen behind Beto O’Rourke. 

            Elizabeth Warren’s Support Score (1809) is the lowest it has been since June 20, and is down about 10% from her peak.  We can clearly conclude that her surge in this poll is over, but not that her supporters are leaving her in any significant number.   It is possible that the peak measurements were not exactly accurate, or perhaps that she had an organized support network feeding supporters into my polls before other candidates did, but other candidates are catching up, or perhaps just that the math is a little bit hinky, and isn’t working for her at the moment.   It’s not clear.   She is likely to receive a little boost tomorrow from the removal of the May 26^th poll from the data, and she will also benefit a little bit (probably) from the removal of the May 28^th poll, two days later.  She is still, in my opinion, in the lead in this race.

            In the chart above, nine candidates are highlighted in green, meaning that their Support Score is up 33% in the last 30 days, while NO candidate is highlighted in Gray, indicating that his/her Support Score is down 25% in the last 30 days.   You might think that that is impossible, that nine candidates are growing rapidly while no one is collapsing, but actually it isn’t.  There are two reasons that it isn’t.   First, in recent weeks I have eliminate from the data four candidates or potential candidates—the guy from California who ended his campaign, whose name I actually can’t remember at the moment, plus Wayne Messam, Jeff Flake and Stacey Abrams.    It’s not actually nine up and none down; it is nine up and four out.  The support of the candidates no longer in the poll goes somewhere else. 

            But larger than that, my second point in the intro is that the campaign has entered a fascinating and un-expected phase.  The stronger candidates are losing support, while the weaker candidates are gaining support.   When Elizabeth Warren loses 10 points, that’s nothing to her; a 25% loss of support for her would be 450 points.   But if 10 points go from Elizabeth Warren to Seth Moulton, that would be an 11% gain for Seth Moulton.  

            Support in my polls is drifting away from ALL of the leading candidates—Warren, Biden, Buttigieg, Sanders—and toward the lesser candidates.  Until June 27, it was exactly the opposite.   When support consolidates around the leading candidates, the standard deviation of the Support Scores goes up.   From late May until June 27—the date of the first debate—the standard deviation went up almost every day, going from a low of 316 up to 443.

            But since June 27, the standard deviation of the support scores has gotten consistently lower, dropping now to 388, its lowest point since June 13. 

            People are learning about the lesser candidates, learning about Delaney and de Blasio and Williamson and Ryan—and some people like them.   That is where we SHOULD be now, and that is where we are.   Elizabeth Warren had a bandwagon rolling, but it is too early to have your bandwagon rolling.   We’re six months from Iowa; we should, at this point, still be in a sorting-out phase, examining the minor candidates. 

            I think that the fate of three candidates has been sealed:  Beto O’Rourke, Cory Booker and Bernie Sanders.   I think they have been examined and rejected by the public, and I doubt that they can recover.   O’Rourke and Booker, in my opinion, will have to drop out of the race, probably before October 1, I am guessing.  Sanders, because he has a passionate base, will be able to hang in the race until the votes in Iowa and Vermont, and obviously he will do well in Vermont, but then he will quickly sink into irrelevance. 

            But the candidates who are in Green in the chart—they’re alive; I mean, obviously they are long shots, but they’re not dead shots. 

            Until the first debate, I had one or two candidates in Green, and a long list in Gray.   The debates turned that around, directing attention from the media favorites to the lesser candidates. 

            Thanks for reading.