Looking in the mirror, 2020 Democratic nomination polls edition

Monthly since April 2019, I have updated my weighted-adjusted polling averages for the 2020 Democratic presidential nomination. You may read about my aggregation methods here, but a key difference between my algorithm and those used by some other polling aggregators (e.g. RealClearPolitics) is that I use every publicly-available poll (as listed on FiveThirtyEight.com) released since January 1, 2019.

I Voted sticker

This means I do not:

  1. Exclude polls based on “quality,”
  2. Drop polls from the algorithm after a certain period of time, or
  3. Distinguish between polls of adults, registered voters and likely voters.

I do, however, give much more weight to polls from higher-quality pollsters (as measured by FiveThirtyEight) and those released more recently. I weight “quality” by converting grades to numeric equivalents (A+=4.3, A=4.0) then dividing by 4.3. And I weight more recent polls by dividing the number of days between the poll’s field dates midpoint and January 1, 2019 by the number of days between January 1, 2019 and the election being assessed. Finally, I have not seen any appreciable difference in candidate standing based upon what set of respondents is sampled.

Simply put, I would rather collect more information, even of lower “quality” or “outdated,” than less. I would prefer to avoid defending exclusion/inclusion criteria.

My aggregation process also does something no other polling average or selection process does. It yields a single score (national-and-state-weighted weighted-adjusted polling average, or NSW-WAPA) for each 2020 Democratic nomination candidate based upon the fact nominations are won through the accumulation of delegates committed to voting for them at the 2020 Democratic National Convention. These delegates are accrued at the state level, usually based upon the results of that state’s presidential primary or caucuses. NSW-WAPA combines state and national polling averages (WAPA), weighting WAPA from the early state contests of Iowa, New Hampshire, Nevada and South Carolina higher than those from later states (with Super Tuesday states weighted twice as high) and national WAPA lowest of all.

As a brief aside, while I try to keep my personal feelings as far from this site as I can, there are two things even the best political journalists do that annoy me to no end:

  1. They call caucuses simply “caucus.” (Multiple such events, requiring time and effort, each called a “caucus,” will be held on the same day in such states as Iowa and Nevada)
  2. They refer to a complex, multi-stage, months-long nomination process as a “primary.” (“Primary” implies a single, national, one-day event in which every interested party member casts a ballot for the person they want to be their party’s presidential nominee. There is no such event.)

It is, frankly, lazy writing and reporting from people I otherwise respect and who should know better. I understand that “caucuses+ is an awkward word to pronounce, and that “nomination process” is a mouthful, but that is no excuse for inaccuracy and imprecision.

OK, I have put my soapbox back in the closet. Thank you for listening.

**********

Having chastised political journalists, I now look in the mirror myself.

I believe my algorithm approach (modeled to a large extent on the FiveThirtyEight approach) is the appropriate one, because it is both more comprehensive (individual higher quality polls may be outliers while certain lower quality polls may better reflect current preferences) and less prone to fluctuate wildly based on any single fluky poll or set of polls.

Nonetheless, it is important to ask if my algorithmic choices are biasing my NSW-WAPA in some way. By “bias,” I mean the mathematical difference between a calculated value and some platonic ideal “true” value.

There are three ways to think about this question:

  1. What would happen to NSW-WAPA if I excluded “lower quality” polls entirely?
  2. Does my current weighting scheme give outdated polls too much influence on what is essentially a snapshot of current voter preferences?
  3. Am I correct that the type of voters sampled (e., registered vs. likely voters) does not make an appreciable difference in NSW-WAPA?

I altered my algorithm to reflect these three questions then compared the resulting NSW-WAPA to what I typically calculate. The results are summarized in the following sections.

**********

Pollster Quality. As noted, FiveThirtyEight assigns a letter grade to dozens of polling organizations based on how they conduct polls (e.g., do they call cellphones as well as landlines; do they use live callers or recordings, sometimes called “robo-polls;” do they randomly select subjects or are they Internet- or panel-based). They also assign a C+ to pollsters who did appear in their May 2018 update.

I include A+-level pollsters like Monmouth University and Seltzer & Co. (gold standard of Iowa polling, extremely difficult in multi-candidate caucuses), C- (and lower)-level pollsters like Zogby Interactive/JZ Analytics (C), McLaughlin & Associates (C-) and Survey Monkey (D-), as well as all pollsters in between.

On balance, the polling is OK, averaging between B and B-, depending on the location; B is a good midpoint.

But what if I only included pollsters with at least a B rating from FiveThirtyEight (while still weighting as before)?

First, the set of polls drops essentially in half, from:

  • 157 national polls to 62 (34 to 17[1] pollsters)
  • 20 Iowa polls to 10, (10 to six pollsters)
  • 22 New Hampshire polls to 11 (10 pollsters to six pollsters)
  • Five Nevada polls to two (four to two pollsters)
  • 18 South Carolina polls to six (eight to five pollsters)
  • 36 Super Tuesday polls (10 states) to 16 polls, with 0 polls from Alabama, Minnesota, Oklahoma, Tennessee or Virginia
  • 32 polls from 14 other states to 10 polls from six other states (Michigan, Ohio, Florida, Wisconsin, Pennsylvania, Oregon).

The number of polls overall drops from 293 (national, 28 states) to 117 (national, 15 states). However, the overall quality rises from B/B- to A-/B+ (precisely the student I was at Yale).

Table 1 compares NSW-WAPA with and without “lower quality” pollsters for the 21 announced Democratic candidates:

Table 1: NSW-WAPA for declared 2020 Democratic presidential nomination candidates with and without pollsters rated B- or lower by FiveThirtyEight

Candidate All Polls Pollster Rating≥B Difference
Biden 27.5 29.8 2.27
Sanders 16.0 15.8 -0.20
Warren 13.2 12.5 -0.73
Harris 9.2 8.8 -0.40
Buttigieg 7.5 6.5 -1.05
O’Rourke 2.6 2.7 0.13
Booker 2.2 2.2 -0.01
Klobuchar 1.5 1.5 -0.02
Yang 1.1 1.0 -0.10
Gabbard 0.9 0.7 -0.17
Steyer 0.6 0.2 -0.41
Castro 0.6 0.6 -0.02
Gillibrand 0.5 0.5 -0.05
Delaney 0.4 0.4 0.02
Bennet 0.3 0.3 -0.03
Williamson 0.3 0.3 0.04
Ryan 0.2 0.2 -0.04
Bullock 0.2 0.1 -0.07
de Blasio 0.1 0.0 -0.11
Messam 0.0 0.1 0.02
Sestak 0.0 0.0 0.01
DK/Other 14.1 15.0 1.16

On average, there is no appreciable difference (-0.04) based on the two criteria. Regardless of direction, the average candidate shift is just 0.28.

Most of that comes from former Vice President Joe Biden, who has a 2.27 higher NSW-WAPA (29.8) in the higher-quality polls than among all polls. While I did not examine the data this way, this would imply a NSW-WPA of “just” 25.0 using only the lower-quality polls, though he would still clearly be in first place, about nine points ahead of Vermont Senator Bernie Sanders (~16.2). No other candidate does appreciably better in the higher-quality polls, with the possible exception of former Texas United States House of Representatives (“House”) member Beto O’Rourke (2.75 vs. 2.62).

In fact, a number of candidates fared worse in the higher-quality polls relative to all polls, most notably South Bend, Indiana Mayor Pete Buttigieg (-1.05), Massachusetts Senator Elizabeth Warren (-0.73) and billionaire activist Tom Steyer (-0.41). However, the only one of the three whose relative positioning would change is Steyer; using only the highest-quality polls, he drops from 11th place to 17th place.

Curiously, the number of respondents selecting “don’t know/not sure” or an unlisted candidate rises 1.16 to 15.1 when only higher-quality polls are analyzed.

Overall, however, while my aggregation method may slightly underrate Biden’s position (and percentage not choosing a listed candidate) and slightly overrate the positions of Buttigieg, Warren and Steyer, these differences are fairly minor.

**********

Poll recency. One simple way to down-weight older polls faster is to square the weight. For example, in my current algorithm, a poll whose field midpoint is August 22 (weight=0.417) is weighted four times as much as a poll whose field midpoint is February 28 (weight=0.104). Squaring each value, however, gives the more recent poll 16 times more weight (0.174 to 0.011).

Table 2 compares NSW-WAPA using more gradual down-weighting to more rapid down-weighting for the 21 announced Democratic candidates:

Table 2: NSW-WAPA for declared 2020 Democratic presidential nomination candidates, simple time weighting vs. squared time weighting

Candidate Simple time weight Squared time weight Difference
Biden 27.5 26.3 -1.22
Sanders 16.0 15.2 -0.75
Warren 13.2 13.7 0.48
Harris 9.2 9.1 -0.07
Buttigieg 7.5 7.5 -0.05
O’Rourke 2.6 2.2 -0.40
Booker 2.2 2.0 -0.23
Klobuchar 1.5 1.4 -0.15
Yang 1.1 1.1 0.00
Gabbard 0.9 0.9 0.01
Steyer 0.6 0.7 0.10
Castro 0.6 0.6 -0.01
Gillibrand 0.5 0.5 0.00
Delaney 0.4 0.4 -0.01
Bennet 0.3 0.3 0.03
Williamson 0.3 0.3 0.02
Ryan 0.2 0.2 -0.01
Bullock 0.2 0.2 0.02
de Blasio 0.1 0.1 0.01
Messam 0.0 0.0 0.00
Sestak 0.0 0.0 0.00
DK/Other 14.1 16.2 2.25

If anything, the differences are even smaller here: although the mean “actual” difference is -0.11, the mean shift, regardless of direction, was only 0.17.

Table 2 does suggest Warren is rising faster (13.7 vs. 13.2) than my more-conservative algorithm shows, while Biden (-1.22), Sanders (-0.75) and O’Rourke (-0.40) are dropping faster; New Jersey Senator Cory Booker and Minnesota Senator Amy Klobuchar also appear to be losing ground recently[2]. Meanwhile, the proportion not choosing any listed candidate seems to be increasing, suggesting greater volatility in the race than my algorithm suggests.

Again, however, these differences are minor.

**********

Likely vs. registered voters. If I were only analyzing national polls, this might make a meaningful difference; of the 157 national polls, just 88 limited their sample to likely voters. And while most eliminated polls are from lower-quality pollsters, it also eliminates polls from Monmouth (A+), CNN/SSRS (A), IBD/TIPP (A-), Quinnipiac University (A-) and Reuters/Ipsos (B+).

However, just two polls (both by Gravis Marketing) in total from Iowa, New Hampshire, Nevada and South Carolina are of registered voters. Moreover, just 18 polls from every other state combined (mostly from Pennsylvania and Texas) were not of registered voters, meaning the number of polls analyzed only drops from 293 to 204, primarily from the lowest-weighted polls (national).

Not surprisingly, there is barely any difference between NSW-WAPA using all polls and only those of likely voters. The mild exception is Steyer dropping from 11th to 15th place (0.62 to 0.39), while the percentage not choosing a listed candidate drops from 14.1 to 13.4.

All combined. Just for fun, I limited the polls being analyzed only to those which were from pollsters with at least a B rating AND sampled likely voters, AND I used the squared time weight. This left 32 national polls, 10 Iowa polls, 11 New Hampshire polls, two Nevada polls, six South Carolina polls, 12 Super Tuesday polls and six polls from all other states (Ohio was dispatched), for a total of just 79 polls.

Table 3: NSW-WAPA for declared 2020 Democratic presidential nomination candidates, comparing original algorithm to most restrictive

Candidate All Polls Pollster Rating≥B Difference
Biden 27.5 29.8 2.29
Sanders 16.0 15.4 -0.61
Warren 13.2 13.4 0.12
Harris 9.2 8.7 -0.50
Buttigieg 7.5 6.8 -0.80
O’Rourke 2.6 2.2 -0.36
Booker 2.2 2.1 -0.14
Klobuchar 1.5 1.4 -0.16
Yang 1.1 1.0 -0.04
Gabbard 0.9 0.8 -0.12
Steyer 0.6 0.2 -0.38
Castro 0.6 0.5 -0.05
Gillibrand 0.5 0.5 -0.05
Delaney 0.4 0.5 0.06
Bennet 0.3 0.3 -0.01
Williamson 0.3 0.4 0.09
Ryan 0.2 0.1 -0.05
Bullock 0.2 0.1 -0.06
de Blasio 0.1 0.0 -0.12
Messam 0.0 0.1 0.02
Sestak 0.0 0.0 0.01
DK/Other 14.1 15.2 1.26

As with limiting polls to those from pollsters with a B rating or better, the average “actual” difference is -0.04, while the average shift, regardless of direction, is 0.29. Other than Biden (+2.29) and “unlisted/unsure” (+1.26), no candidate did measurably better when limiting the analysis to these 79 national and state polls. Buttigieg (-0.80), Sanders (-0.61), Harris (-0.50) and Steyer (-0.38) all did somewhat worse, but—again—only Steyer’s rank changed (11th to 17th).

Conclusion. While my algorithm may somewhat understate Biden’s strength—and overestimate Steyer’s—while not fully capturing the rate at which Warren is gaining support, the overall differences are so minor I see no reason to alter it.

Until next time…

Postscript. For those who are curious, here is a comparison, as of August 28, 2019, between NSW-WAPA and the RealClearPolitics (RCP) averages, combining national and state polls, using my weighting scheme. I exclude New York Senator Kirsten Gillibrand—who dropped out on August 28—and Miramar, FL Mayor Wayne Messam, who is not included in the RCP averages; if a candidate’s average was not listed by RCP (i.e., it was less than 0.5%), I assigned her/him a value of 0.25%.

Table 4: Comparing NSW-WAPA to RealClearPolitics averages for declared 2020 Democratic presidential nomination candidates

Candidate All Polls RealClearPolitics Difference
Biden 27.5 27.8 0.3
Sanders 16.0 15.9 -0.1
Warren 13.2 15.3 2.1
Harris 9.2 10.7 1.5
Buttigieg 7.5 6.4 -1.1
O’Rourke 2.6 1.8 -0.8
Booker 2.2 2.0 -0.2
Klobuchar 1.5 1.8 0.3
Yang 1.1 1.5 0.4
Gabbard 0.9 1.4 0.5
Steyer 0.6 2.0 1.4
Castro 0.6 1.0 0.4
Delaney 0.4 0.5 0.1
Bennet 0.3 0.3 0.0
Williamson 0.3 0.3 0.0
Ryan 0.2 0.4 0.2
Bullock 0.2 0.3 0.1
de Blasio 0.1 0.4 0.3
Sestak 0.0 0.2 0.2
DK/Other 14.1 10.1 -4.0

Other than underestimating the polling strength of Warren, California Senator Kamala Harris and Steyer—and overestimating the strength of O’Rourke—these differences are minimal; the much lower percentage not choosing a listed candidate in the RCP average could easily result from assigning 0.25% (likely too high) to unlisted candidates. The average actual difference is just 0.1, and the average difference, regardless of direction, is 0.7.

The next comprehensive update will come just before the September 12, 2019 debate.

[1] Emerson College, Monmouth, GBAO, CNN/SSRS, ABC News/Washington Post, Quinnipiac, YouGov, Suffolk University, WPA Intelligence, YouGov Blue, NBC News/Wall Street Journal, Perry Undem-YouGov, SurveyUSA, Public Policy Polling (PPP), IBD/TIPP, Reuters/Ipsos, Fox News

[2] This may be illusory for Booker. Taking a simple average of national polls, Booker was at 1.7% between the first and second Democratic debates, but he has risen to 2.7% since then.

Interrogating memory: The Beatles, wax museums and a diner mystery solved

To the extent my writing over the last three years has a theme (or perhaps even a brand), it is what I call interrogating memory.

At one level, this is just a fancy term for “fact-checking,” as in looking through my elementary school report cards (I am missing the one for third grade[1]) to confirm my fourth-grade teacher was named Ms. Goldman, only to discover she was my fifth-grade teacher and her name was “R. Goldberg.”

Quick story.

On the first day of fifth grade at Lynnewood Elementary School, my new teacher called me up to her desk. Ms. Goldberg, an attractive woman with an unwavering platinum blonde permanent, was curious about my father, whose name she had seen was David Louis Berger. We quickly established (most likely through his age and being raised in West Philadelphia) they had been in the same confirmation class at Congregation Beth El in 1951. It was also clear from the way she spoke about him (my aunt once wrote me, “He really was lovable you know”) she had a serious crush on him. I do not recall how I reacted, or what my father said when I told him.

Still, knowing it was fifth, not fourth, grade and that her surname was Goldberg, not Goldman, does not materially alter the story: my teacher had known and liked my father when they were teenagers.

The thing is, however, I pulled out those report cards in the process of reassessing an entirely different memory, one that better exemplifies the complexity of interrogating memory.

As a child and young teen, I hated The Beatles (or, at least, refused to succumb to the pressure to love them). And until a few weeks ago, I believed this disdain stemmed from my active resistance to being told what to like and what not to like. My attitude from a very young age was that I will decide for myself what I like and do not like, thank you very much.

My proof, other than my own memory?

I was certain that mixed in with otherwise glowing comments from my elementary school teachers on my report cards was a common phrase along the lines of “does not like to follow directions.”

But when I pulled out my five surviving report cards from Lynnewood, this sentiment was far less ubiquitous than I had remembered. Mrs. Virginia Hoeveler did begin her extensive (and humbly flattering) comments, dated June 13, 1973, by noting I initially had “difficulty conforming to a classroom situation,” though I quickly adjusted. She also added a postscript: “Matt is quite the ‘individual – he likes to do his ‘own thing.’ “

Five months later (November 7, 1973), Ms. C. Edwards—who broke the heart of every boy in my second-grade class when she became Mrs. C. Stevenson at the end of the school year (many of us attended the wedding, sitting in a mezzanine area of the church, overlooking the ceremony, stage left)—wrote, “Matt sometimes gets carried away with his intelligence. He seems to feel that he doesn’t need to follow directions.”

Ouch.

Still, as of June 1, 1974, I had “become much more social with [my] peers.” Good to know I was ceasing to be a curmudgeon at seven years old.

But…that is it. I have no third grade report card, neither Miss Nichols nor R. Goldberg wrote more than a token sentence or two, and Mr. Bianco (a good-looking man who wore platform shoes and was smitten with my mother) merely noted I would have had an “O” (Outstanding) instead of an “S” (Satisfactory) in Social Studies but for too many missed assignments.

Oh.

The point is, my memory was not, strictly speaking, incorrect; there were comments along the lines of “does not like to follow directions.” It was just that they were confined to first and second grades, when I was apparently still adjusting socially and academically to a formal classroom environment.

Here is the kicker, though. Even before I pulled out those report cards, I had already concluded my aversion to structured guidance was not why I had hated The Beatles (which I no longer do; quite the contrary, in fact[2]). Or, at least, it was not the only reason.

Just bear with me while I wax rhapsodic about Atlantic City, New Jersey.

I spent the summers of 1974 and 1975 living with my mother and our dog—a Keeshond named Luvey—in Penthouse A (really, just one of two slightly larger rooms with two queen beds and a walk-in closet sharing a small semi-circular concrete balcony overlooking the pool) of the Strand Motel in Atlantic City. On weekends, my father would drive the roughly 80 miles from our home in Havertown, Pennsylvania (just west of Philadelphia) to join us.

Luvey in Atlantic City August 1974 2.jpg

The Strand Motel, which sat between the Boardwalk and Pacific Avenues, and between Providence and Boston Avenues, was knocked down around 1979 as part of the construction of the Golden Nugget Casino (which, after many name changes, closed in 2014). I am reasonably certain this photograph was taken in the lounge directly below the penthouses one of those two summers; my father is the silver-haired man in the blue jacket sitting at the bar, while the left side of my mother’s face is just visible on the right (her natural red hair was back).

Scan0011.jpg

Those two summers, I spent my days wandering up and down Pacific Avenue (either on foot, or riding a jitney for 35 cents) and the Boardwalk. By myself, at the ages of seven and eight, that is; I cannot imagine that happening today. I especially loved going into the lobby of every motel and hotel along the roughly three miles of roads/Boardwalk in my purview to collect one of each pamphlet available in the large wooden racks there. During the winter, I would dump them onto my parents’ bed and rummage through them, wishing I was back in Atlantic City.

One of those pamphlets was actually a red-covered brochure for Louis Tussaud’s Wax Museum, then located at 1238 Boardwalk (yes, the Boardwalk is considered a road for mailing purposes), roughly halfway between North Carolina and South Carolina Avenues.

I do not know why I suddenly recalled this wax museum a few weeks ago (which was opened by Madame Tussaud’s somewhat less-talented great-grandson). Perhaps it was researching my book, and thinking about how we stopped summering down the shore (as those of us raised near Philadelphia say) in 1976, just before the casinos started being built, effectively ending “my” Atlantic City. Along those lines, I have reflected a great deal this summer on how much my wife Nell and our daughters love spending much of the summer on Martha’s Vineyard, and how much, frankly, I do not. And I have concluded no longer spending summers in Atlantic City, even as it was inexorably changing (for the worst, in my opinion)[3], was a deeply painful occurrence I have yet fully to process. But, the result is a silly jealousy of Nell’s childhood (and current) summer home.

Or, Louis Tussaud’s Wax Museum came to mind for no other reason than the 1953 Vincent Price vehicle House of Wax was recently on TCM OnDemand (I did not get a chance to watch it).

Regardless, what I specifically recalled about that slightly tacky museum was that one of the first tableaus you saw when you entered from the Boardwalk was of The Beatles circa 1964. Walking by the four wax figures, I would hear “I Want to Hold Your Hand” playing; perhaps songs like “She Loves You” played as well. In fact, now that I interrogate that memory, the point of the tableau may have been to reproduce their historic February 9, 1964 appearance on The Ed Sullivan Show.

I could not tell you what other tableaus I saw in Louis Tussaud’s because, frankly, the only other thing I clearly remember is the Chamber of Horrors.

Again, I was seven or eight years old when I viewed those displays, some of which were particularly gory and graphic. This nostalgic video includes two of them: a low-quality rendition of the Lon Chaney version of the Phantom of the Opera and a gruesome Algerian Hook (speaks for itself, despite being misspelled in the video).

As an aside, the photograph in the video of the Boardwalk in front of Steel Pier in the summer of 1974 was like stepping out of a TARDIS: that is the Atlantic City I remember. To be fair, I preferred Million Dollar Pier, whose Tilt-a-Whirl I would foolishly ride every weekday, around 12:30 in the afternoon, after eating a slice of pizza from a little stand just where Arkansas Avenue meets the Boardwalk. Seeing that photograph was both exhilarating and painful; I may have known Atlantic City at the very end of its family-resort glory, but I loved being there.

Returning to the Chamber of Horrors, I was both terrified and fascinated by the scenes it depicted. If memory serves, they also included Lee Harvey Oswald being shot by Jack Ruby on November 24, 1963. As deeply unsettling as they were, I could not stop poring over the photographs of those displays in my souvenir booklet back home in Havertown.

But rather than admit they scared the bleepity-frick out of me, I displaced that emotion onto the completely banal and non-threatening (if mildly creepy, in the way all wax figures are mildly creepy) wax renditions of John, Paul, George and Ringo. Simply because they were what I saw before I entered the Chamber of Horror, which truly did scare me. This may not be quite what Sigmund Freud meant by a “screen memory,” but the concept is broadly the same.

In some ways, “interrogating memory” is like the love child of psychoanalytic technique (patiently probing memories to get at any underlying meaning) and the epistemological underpinnings of epidemiology (questioning and verifying everything, putting all data points into context—usually chronological), raised on a steady diet of persistence and a genuine love of history.

Or, to put it even more simply, it is using every technique in your critical toolbox to answer the question, “Hold on a minute, did that really happen that way, then, in that place?”

*********

Speaking of persistence, I may have solved a mystery I first identified here:

Memory 2: One Saturday night in 2002, 2003 or 2004, I took a meandering night drive. Somewhere in Montgomery County, north of Philadelphia, I found myself driving on a “road with a route number.” I then turned left onto a different “road with a route number” to explore further; I may have intended to find this latter road from the start. Sometime later, I find a 24-hour diner (on weekends, at least); I park and enter. I am almost certain I walked up a few concrete steps to do so. It was clean and kind of “retro-modern;” despite my sense of a great deal of black and white in the décor, I also feel like there was a fair amount of neon and chrome. I sat at a small-ish counter (curved?) in a separate room to the right as you entered (there were some booths behind me); in front of me may have been glass shelving stacked high to the ceiling. Behind me and to the left was a large glass window through which I can look down onto an asphalt-covered parking area with at most a few spaces. The diner itself is sort of tucked into a dark urban commercial corner, almost as though it jutted out from an adjoining building. I do not recall what I ordered or what I was reading, or whether I even liked the diner or not. I never returned there, and I can no longer recall the name of the diner or its precise location.

In the post, I concluded I had almost certainly turned north on Route 152 from Business Route 202 that night, eventually wending through the Montgomery County towns of Chalfont, Briarwyck, Silverdale, Perkasie, Sellersville and Telford (where Route 152 ends at Route 309). It was just that none of these towns had the sort of urban-feeling center in which my memory placed the diner.

Frustrated in my efforts to find a diner that fit the necessary criteria, I concluded thus:

I have a sinking suspicion this particular eatery has since closed; this was 15 or so years ago, after all. Or else I have simply mixed up an intersection from one drive with a diner I happened upon in another—though I highly doubt it. What remains mystifying is how this late-night restaurant could have made such an impression on me—yet I have no idea where it is/was or what its name is/was.

As I said, though, a key element of interrogating memory is persistence, so the other night I resolved to trace my possible route that night, starting at the intersection of Routes 152 and Business 202, using StreetView on Google Maps.

Patiently clicking the forward arrow, waiting less patiently for the photographs to resolve on my computer screen, I made my virtual way through Chalfont and Briarwyck and Silverdale and Perkasie into Sellersville. I took a few wrong turns along the way (Route 152, like many state routes, has a habit of randomly turning left or right onto a different street), but always righted myself.

After getting lost multiple times at a particularly tricky five-way intersection, I continued along South Main Street, heading away from the center of Sellersville. In that confusing way of state routes, by following “North” Route 152, I actually travelled south. After passing a few scattered two-story brick houses and local businesses, a large (for the area) parking lot appeared on my left.

In the middle of the lot was a light gray single-story building with a double-sloped roof. The front of the building was a two-story structure from which short flights of concrete steps, under red awnings, protruded. Above each awning was a lighted sign, white with red letters, reading “A & N DINER.” A yellow road sign embedded in the asphalt just beyond the sidewalk read “A & N DINER/ FAMILY RESTAURANT / OPEN 24 HOURS”; with “HAPPY LABOR DAY” spelled out in removable black plastic letters just below that.

Say what now? How did I miss this 24-hour diner in my extensive search?

Something about it seemed vaguely familiar, especially adjusting for the fact these September 2018 photographs were taken during the day, while my drive occurred at night, when the A & N Diner would have been brightly lit in the darkness. I clicked on the map’s icon to learn it is no longer open 24 hours. If that change occurred between Labor Day 2018 and early March 2019, that would explain why I could not find it searching for “24 hour restaurants.”

Scrolling through the accompanying photographs, I observed a small counter area to the left as you entered. One photograph showed five dark pink (almost gray) leather-covered stools bolted to the floor. To the left of the counter was a window, which another photograph confirmed overlooked the parking lot. And the wall one faced sitting at the counter might be the one I recalled—the glass shelving could easily have been replaced since I was (possibly) there in 2003 or 2004 (or existed only in my memory).

The only problem was that this was hardly the urban downtown my memory insisted housed the diner. However, I may have an explanation for that.

One of the classes I took in the first semester of my biostatistics Master’s program at Boston University School of Public Health was on probability theory. While I earned an A on the first of three exams (which comprised ~90% of the final grade), I bombed the second exam. Forget getting an A in the class; I was simply hoping to salvage a B with the final exam. Sometime after that disastrous second exam, say in November 2005, I had a powerful dream. In that dream, in which I learned I did in fact earn an A, it was night. The dark second floor room in which I stood extended far behind me as I stared out a large bay window; perhaps I was in bed first, it is all a bit fuzzy 14 years later. Below me was an urban corner with low buildings, lit by a single street lamp; a kind of brick culvert was off to my right.

This dream made such an impression on me, I still remember it relatively clearly nearly 14 years later. It is possible I mixed up looking out the window into the dark parking lot at the A & N Diner with looking out the window at the urban street corner in the dark in my dream. Why, I could not begin to tell you…unless the former somehow got worked into the latter? I would have to drive to the A & N Diner at night to be certain.

Another slight variation is that I recall the diner being on my right, but I would have approached it from the left that night. That could easily be explained, however, if I parked on the opposite side of the building (putting the diner on my right as I entered) and/or if I drove past it at first, decided to stop in for a snack, and turned around, thus placing the building to my right as I drove to it again.

There is one additional small point of confirmation. In my memory, the diner is shiny and new. Well, a little digging on the invaluable Newspapers.com uncovered a February 2000 article in the NEWS-HERALD of Perakasie, PA[4]. The gist of the article is that Nicholas and Vasso Scebes had assumed control of Angelo’s Family Restaurant on January 31, 2000, renaming it A & N Diner and Family Restaurant.

The key passage is this:

“Later this month, the manager said, they hope to be settled in enough to change the environment of the restaurant, starting with the interior wall colors, which are currently a bright two-tone lime green. Vasso said that’s the first thing regulars asked to have changed.”

Later in the article, Vasso avowed her intention to “clean up this place and make it respectable.”

If those renovations were completed sometime in 2000, they could well have seemed “shiny and new” three or four years later, when a young man out for a meandering night drive almost certainly stopped in with his book for a meal and lots of decaffeinated coffee, black.

For the record, dreams sometimes do come true. I studied intensely for the final exam, and earned something like a 92. Great, I thought, that will get me a solid B in the course. When I learned I had actually received an A, I e-mailed the professor to make sure he had not made a mistake. No, he said, he thought well enough of my participation in the class to essentially “throw out” the middle exam as an unfortunate outlier. Oh, I replied, thank you very much.

Until next time…

[1] Itself a curious slip of memory, as I originally wrote (from memory) “fourth grade.” I only pulled out these report cards to review a week or two ago.

[2] I am even listening to Abbey Road as I edit this post.

[3] This shift is beautifully rendered in Louis Malle’s 1980 film Atlantic City.

[4] Baum, Charles W., “New family takes over operation of former Angelo’s in Sellersville,” NEWS-HERALD (Perkasie, PA), February 16, 2000, pg. 3.

August 2019 update: 2020 Democratic presidential nomination and general election polling

It has been just over two weeks since the second Democratic presidential nomination debates, so it is time for an updated assessment of the relative position of the 23 declared candidates remaining. Former Alaska Senator Mike Gravel ended his campaign on August 6, 2019, and it appears former Colorado John Hickenlooper will end his bid on August 15, 2019.

To learn how I calculate NSW-WAPA (national-and-state-weighted weighted-adjusted polling average), please see here[1]. Note that I recently altered my methodology slightly: within my post-early-state weighted average of each candidate’s WAPA, I now weight the nine states[2] scheduled to hold their nomination contests on March 3, 2019 (“Super Tuesday”) twice as much as all subsequent contests[3]

And, as usual, here is the August 2019 lighthouse photograph in my Down East 2019 Maine Lighthouses wall calendar.

Aug 2019 lighthouse.JPG

Table 1 below aggregates data from all national and state-level polls publicly released since January 1, 2019, including:

  • 149 national polls (including 32 weekly Morning Consult tracking polls)
  • 19 Iowa Caucuses polls
  • 22 New Hampshire Primary polls
  • 4 Nevada Caucuses polls
  • 18 South Carolina polls
  • 35 Super Tuesday polls[4]
  • 33 polls from 13 other states.[5]

This makes a total of 280 polls, up from 247 in the last update.

Table 1: National-and-state-weighted WAPA for declared 2020 Democratic presidential nomination candidates

Candidate National IA NH NV SC Post-SC NSW-WAPA
Biden 29.5 23.6 23.3 30.2 36.4 27.3 27.9 (-0.7)
Sanders 16.7 15.1 19.0 19.1 12.6 15.3 16.4 (-0.1)
Warren 10.4 13.1 13.5 18.0 9.1 12.6 13.2 (+0.6)
Harris 8.6 9.7 8.7 8.6 9.9 9.7 9.2 (–)
Buttigieg 5.6 9.1 9.1 8.0 4.9 6.7 7.7 (-0.3)
O’Rourke 4.1 2.6 2.2 3.1 1.8 5.3 2.8 (-0.3)
Booker 2.5 2.5 1.6 1.3 3.5 1.5 2.2 (-0.2)
Klobuchar 1.3 2.7 1.5 1.1 0.6 1.0 1.5 (–)
Yang 1.0 0.7 1.3 1.5 0.7 0.7 1.1 (–)
Gabbard 0.7 0.6 1.4 1.1 0.3 0.5 0.84 (+0.11)
Castro 0.9 0.7 0.2 1.0 0.1 1.0 0.59 (–)
Gillibrand 0.5 0.6 0.7 0.4 0.3 0.3 0.50 (+0.09)
Delaney 0.3 0.9 0.5 0.00 0.3 0.2 0.43 (–)
Steyer 0.03 0.1 0.4 1.0 0.4 0.1 0.40 (+0.30)
Inslee 0.4 0.4 0.2 0.3 0.1 0.3 0.26 (+0.06)
Williamson 0.2 0.05 0.3 0.4 0.3 0.1 0.24 (+0.07)
Bennet 0.2 0.3 0.2 0.00 0.2 0.4 0.20 (+0.04)
Ryan 0.3 0.1 0.3 0.00 0.2 0.2 0.16 (-0.01)
Bullock 0.2 0.3 0.00 0.00 0.1 0.1 0.10 (+0.03)
de Blasio 0.3 0.05 0.00 0.00 0.1 0.1 0.04
Moulton 0.1 0.04 0.1 0.00 0.00 0.03 0.03
Messam 0.00 0.00 0.00 0.00 0.1 0.04 0.03
Sestak 0.00 0.1 0.00 n/a 0.00 0.06 0.02
DK/Other 14.6 15.8 15.0 4.4 16.9 16.0 13.7 (+0.4)

There has been little substantive change in the relative standing of the 23 remaining candidates over the last two-three weeks, despite some short-term effects from the second round of debates (see below). Former Vice President Joe Biden remains the nominal frontrunner (27.9), primarily because of his dominant position in South Carolina primary polls; his weighted average of 36.4% is well ahead of Vermont Senator Bernie Sanders, California Senator Kamala Harris and Massachusetts Senator Elizabeth Warren. By contrast, the race is much closer in polling for the Iowa Caucuses and New Hampshire Primary; in these first two contests, Biden is only averaging 23-24%, with Sanders close behind at 15-19% and Warren at 13-14%. Harris and South Bend, Indiana Mayor Pete Buttigieg are not much further behind, hovering around 9%.

These five candidates continue to dominate the race overall, albeit with Biden continuing to decline while Warren continues her steady ascent (up from 8.5% in early June to 13.2% now), capturing just under three-quarters of the support of those polled. Just behind these five are four other candidates with an NSW-WAPA of 1.0 or higher: former Texas member of the United States House of Representatives (“Representative”) Beto O’Rourke, New Jersey Senator Cory Booker, Minnesota Senator Amy Klobuchar and entrepreneur Andrew Yang. Perhaps not surprisingly, these are also the only nine candidates to have qualified for the next round of Democratic presidential nomination debates (September 12-13, 2019). Overall these nine candidates account for 81.9% of currently-declared Democratic nomination preferences. Factor in 13.7%s[6] undecided or choosing an unlisted candidate, that means the remaining 14 candidates are divvying up just 4.4% between them.

**********

In the previous update, I assessed the short-term impact of the first round of Democratic presidential nomination debates by comparing support for each candidate in polls conducted by the same pollster within one month prior to, and just after, those debates. Meeting these criteria for the second round of debates are six national polls[7] and one Texas poll[8]. For ease of presentation, Table 2 presents data only for the 12 candidates with an NSW-WAPA of 0.5 or higher (including Hawaii Representative Tulsi Gabbard, former Secretary of Housing and Urban Development Julián Castro and New York Senator Kirsten Gillibrand). Values listed are simple arithmetic averages (with the Texas poll change weighted twice the changes in national polls); weighting by pollster quality or time between polls made little difference.

Table 2: Average change in polls from the same pollster before and after July 2019 Democratic presidential debates:

Candidate National TX Weighted Average
Biden -1.0 +4.5 +0.4
Sanders +1.8 +4.0 +2.4
Warren +1.8 +1.0 +1.6
Harris -3.5 -3.0 -3.4
Buttigieg +0.3 +3.0 +1.0
O’Rourke -0.3 -14.5 -3.9
Booker +0.8 +1.0 +0.9
Klobuchar -0.5 0 -0.4
Yang -0.3 +2.0 +0.3
Gabbard +0.3 0 +0.3
Castro -0.2 +2.0 +0.4
Gillibrand -0.2 0 -0.1
DK/Other +1.0 -1.0 +0.5

Examined this way, support for Harris—who had risen 7.7 percentage points (“points”) following the June debates—dropped fully 3.4 points following the July 2019 Democratic debates. O’Rourke also declined significantly (-3.9 points), but that was almost exclusively due to an astonishing 14.5-point drop (from 38% to 23.5%) in the Texas poll. The largest post-July-debate increases were for Booker (+0.9), Buttigieg (+1.0), Warren (+1.6) and Sanders (+2.4); no other candidate saw her/his support shift by more than 0.4 points in either direction. Finally, the percentage not choosing a listed candidate increased slightly.

**********

To the extent that the polling for the 2020 presidential election between a named Democrat and Republican Donald J. Trump changed, it is due to the modestly-increased likelihood (55.7%) that someone other than Biden (who would hypothetically beat Trump nationally by 8.4 points) and Sanders (by 5.2 points) will be the 2020 Democratic presidential nominee. Thus, once you weight for the likelihood of being the nominee, the Democrat would beat Trump by 3.6 points. This is actually slightly higher than the median Democratic presidential margin (+3.0 points) in the previous six presidential elections, which include three elections with an incumbent seeking reelection and three elections with no incumbent. However, once you exclude Biden and Sanders, the margin over Trump decreases to 0.7 points; Warren would hypothetically win by 1.5 points and Harris by 1.0 points, while Buttigieg would lose by 1.5 points.

Still, given that state-level results actually determine the winner of a presidential election (via the Electoral College), it is more informative to look to those polls, where they are publicly-available. Using my 3W-RDM, a measure of how much more or less Democratic a state’s voting is relative to the nation as a whole, this polling[9] implies Democrats would win the national popular vote by between 2.6 (excluding former Vice President Joe Biden and Vermont Senator Bernie Sanders) and 5.6 (including Biden and Sanders) points on average. Most encouraging to Democrats should be the polls from North Carolina (R+6.0) and Texas (R+15.3), which show a very close race, implying a 6-7-point win and a 12-14-point win nationally for Democrats, respectively; these polls confirm strong opportunities for Democrats in the southeast and southwest. By contrast, however, a few polls from Democratic-leaning Maine (D+5.9) and Nevada (D+2.0) imply Democrats would lose nationwide by 2-5 points. Those remain the exceptions, however, to what continues to be encouraging news for Democrats in 2020.

Until next time…

[1] Essentially, polls are weighted within areal units (nation, state) by days to the nominating contest and pollster quality to form a unit-specific average, then a weighted average is taken across Iowa (weight=5), New Hampshire (5), Nevada (4), South Carolina (4), the time-weighted average of all subsequent contests (2) and nationwide (1).

[2] Alabama, California, Colorado, Massachusetts, Minnesota, North Carolina, Oklahoma, Tennessee, Texas, Virginia

[3] As of this writing, I have at least one poll from (in chronological order) Maine, Michigan, Mississippi, Missouri, Ohio, Washington, Arizona, Florida, Georgia, Wisconsin, Pennsylvania, Indiana and Oregon

[4] Primarily from California (14) and Texas (9)

[5] Primarily Florida (9) and Pennsylvania (5)

[6] This does include polls that limit the number of candidates queried.

[7] Morning Consult Tracking, HarrisX, Change Research, Quinnipiac University, YouGov, Reuters/Ipsos

[8] University of Texas at Tyler

[9] From Pennsylvania, New Hampshire, Wisconsin, Michigan, North Carolina, Texas, Iowa, Arizona, South Carolina, Minnesota, Nevada, Massachusetts, Florida, New York, Kentucky, Maine, Ohio.