Neighborhood political yard-sign survey suggests Biden is in trouble (Election 2020 – Street Scenes #3)

(This was written October 30-31; the survey was mid-October; observations on which I decided to do to full survey were made in August and September.)

After census work ended in mid-October, I conducted my own census, a political yard sign census, around a certain neighborhood in Arlington, Virginia. It is heavily white, heavily educated and credentialed and with new in-mover necessarily having to be pretty high income as since the early 2000s it has had ballooning land prices. It is quite strongly Democratic for a white neighborhood.

I walked past 478 occupied homes (houses, mainly; some townhouses) and noted what kind or yard signs they had, if any, whether there were Halloween decorations, an whether there was a US flag. This was a personal research project ahead of the election, to see if I might spot anything that might signal the election result, either before or after:

Political Signs:

• Biden signs alone,
• Biden signs with other political signs,
• Other political signs without Biden signs, or
• No political signs of any kind.

Halloween decorations:

• Pumpkins alone,
• Pumpkins and other Halloween paraphernalia, or
• No Halloween decorations.

Flags:

• US flag flying
• Other flag/banner flying
• No flag of any kind

13.1% of occupied homes had a Biden sign, of which 7.5% had a Biden sign alone (no other accompanying political sign) and 5.6% having a Biden sign along with at least one other political sign, including “Black Lives Matter” and a particularly common long list of doctrinal statements. (Earlier versions had existed before but in June this year after the protests, riots, and looting, some of which I witnessed personally, the signs began to be replaced with a version that led with “We believe… Black Lives Matter,” followed by the other doctrinal statements of belief below it.)

Another 6.1% of occupied homes had some kin of political or quasi-political sign but no Biden sign, 0.8% pro-Trump or proxy pro-Trump, and 5.3% anti-Trump or other.

Of these, only 0.8% appeared to be pro-Trump either directly (0.4%) or by proxy (0.4%). In the latter I include a house with a “Blue Line” flag, the flag being flown to support police.

The other 5.3% had left-wing signs such as the doctrinal confession aforementioned and various other common ones such as the “Hate Has No Home Here” sign. Another common one was “You are Welcome Here,” which explains to mean the resident is pro-immigration and implies they also support those illegally resident in the USA.

Often, the homes in this 5.3% set that had left-wing slogans but passed on the Biden signs had a more radical and angry and even obscene tenor: “Tuck Frump.” One had “Profits are a Crime Against Humanity.” About three homes had “A Woman’s Place is in the Resistance” featuring a picture of a Star Wars character (I think a reference to a 2010s-era Star Wars movie which put women as the protagonists). Lots said Dump Trump and had various kinds of anti-Trump slogans. There were other miscellaneous signs that those who followed the various political news cycles would recognize to also be connected to the Orange Man, such as “Save the Post Office.”

Rearranging the categories a little:

• Homes advertising support for left-wing causes d’jour (with or without a Biden sign alongside) were at 10.9%.
• Those advertising support for Biden without any co-endorsement of any left-wing cause d’jour were at 7.5%.
• Those advertising possible Trump support were at only 0.8%.
• (People without any political signs were at 80.8%.)

Using these groups as proxies for Left, Center, and Right, respectively, we have a ratio in the large sample of 14:9:1, or 23:1 on the expected Biden-Trump vote in this area. The problem is, this is far higher a ratio than the actual result is going to be; 23 to 1 is a blowout on par with one-party stats that have rubber-stamp elections. Sure enough it can hold in isolated geographic zones, but not in this one: The precinct’s result in November 2016 was strongly tilted Left, but nowhere near 23:1. The Left-to-Right voting ratio in 2026 was 26:10 (Hillary+Stein vs. Trump+Johnson+McMullin). The Hillary:Trump margin alone was 37:10.

But what about the four-in-five majority who put up no political yard signs in 2020? The 2016 voting rate was 26:10:18 Left:Right:Nonvoter. So not quite a majority voted Hillary. This is hardly the kind of total dominance you’d expect from a casual walk through the area with an eye on yard signs, or the data as presented in this post, with its 230:10 Left:Right political yard sign ratio.

That is the first lesson I take from this. In certain areas, a tipping point is reached an a status quo asserts itself, and people in the minority begin to self-censor. This is the Arlington I have known. It is run by smug, self-satisfied, and self-righteous people who cannot fathom any opposition, and never see any. Truer bubble people you seldom find.

I’ve often wondered when the tipping point was in Arlington. In decade terms, I think it was the 1980s. There was still a conservative element in the 1970s, although by fading fast as new arrivals kept coming in, from near and far. Starting in the 1980s, the apparent public consensus was established and conservatives began to disappear from public discourse; I expect you could find similar versions of this finding of a laughably lopsided sign ratio back to the 1990s at least. Having been a relatively regular observer of Arlington starting at that time, that aligns with my memory. There was still some degree of conservative strength but it was only seen in the privacy of polling booths and with very occasional civic questions of the moment, but starting the 1980s they had an unbeatable voting coalition given the type of person coming into Arlington, and they had a coming one-party state. The one-party state is all I’ve ever known of Arlington, 1990s to present.

They are crowing with victory and have, of late, been on a tear of name-changing, statue-toppling, and memorial defacing. The image you get is one of a wine-and-cheese Taliban. (Their latest target is Lee Highway; they put up a website for people to vote on new names for Lee Highway, something no one asked for. Almost all the many options were ridiculous. They are again trying to push “Loving.” They refused to offer a None of the Above option in their poll or a Keep the Name option or a face-saving “Symbolically rename it for another Lee” and not the great general, the most famous man to ever come out of Arlington.)

In any case, back to the neighborhood political sign survey I conducted. I decided to survey flags flown in front of the same houses once I perceived a possible lopsided sign ratio way beyond the voting ratio. My idea was that US flags might be a proxy for Trump support or Right voting (in 2016, they also could’ve voted for Trump, Johnson, or McMullin to be counted as a Right voter).

8.8% of occupied homes had US flags flying or otherwise displayed at time of visit, of which 1% had a Biden sign and 7.8% had no Biden sign.

• Homes advertising support for left-wing causes d’jour (with or without a Biden sign alongside) were at 10.9%.
• Those advertising support for Biden without any co-endorsement of any left-wing cause d’jour were at 7.5%.
• Those advertising possible Trump support were at only 0.8%.
• Those flying US flags without a Biden sign or a left-wing activist sign were ca. 7.2%.

Ratio of the first two categories (18.4%) against the second two (8%) puts us at a 23:10 Left:Right ratio, which is near the actual voting ratio in 2016, which was 26:10 Left:Right.

So finding two is that US flags are a proxy for conservative voting, also not the shock of the century but interesting how well the numbers line up. Maybe a coincidence.

Homes with a Biden sign would occasionally also have a US flag. I would interpret these people to be older people, mid-twentieth-century liberals, of which Arlington has had plenty for a long time.

Homes with left-oriented or activist signs but without a Biden sign almost never had a US flag. I would interpret these people to represent the new Democratic Party, the one that has rejected its once important populist streak and reoriented to being a party with close ties to academia and the well-credentialed broad elite, which again is certainly well represented in Arlington and in the neighborhood I surveyed.

I didn’t write down perfect notes in all cases on this, but I don’t think more than one of the houses in this grouping (activist signs without Biden signs) had a US flag, and I am sure many would, if asked, say they don’t want to fly the flag because it offends them because reasons. (I am thinking of the Profits are Evil sign-poster.)

Finding Three is that support for Biden is unenthusiastic. This was the most interesting of all and the reason I wanted to do this survey, especially coming off three+ months of hard census work doing things a lot like this. This finding is also why I am typing away at this on November 1, 48 hours before the election results start to come in.

You often hear anecdotal reports about Trump signs versus Biden signs, but how to test these rigorously? I don’t have access to comparisons over time and space, but I do have access to a possibly ‘control’ dataset: Halloween decorations.

Halloween is definitely now part of the USA’s civic religion, with many people taking it as seriously as religious people took religious holidays in days of yore. And a Martian anthropologist would definitely think Halloween something to do with the state religion. For this reason I’d say those who put up Halloween decorations are civically minded and civically engaged, with the caveat that of course it’s a ‘holiday’ primarily for children. Celebrating Halloween is effectively an expression of defacto civic patriotism. Some of these decorations people do are very involved, almost embarrassing amounts of time, money, imagination, effort, and energy go into some of these.

I noted how many houses had pumpkins alone (17.8%), how many had Halloween paraphernalia or decorations besides pumpkins (16.1%) and how many had neither (66.1%) as of the time I collected this data on Oct. 13-15.

19.2% of homes had any kind of political sign up; 33.9% had Halloween decorations up. There were noticeably more homes with Halloween decorations of some kind, even 2.5 weeks before Halloween, than had any kind of political sign.

This, along with the finding of so many anti-Trump signs or other activist signs in comparison to actual pro-Biden signs, leads me to believe people are unenthusiastic about Biden. This doesn’t matter in Virginia, which will certainly be in the D column Tuesday night; Virginia is now a “solid blue” state, entirely because of Northern Virginia and migration into is from elsewhere, domestic and international. The process I described of how Arlington flipped from being once competitive to becoming a dreary one-party state with party bosses controlling who wins each time was ongoing elsewhere as well, driven by essentially the same things it was in Arlington but following by twenty or thirty years. Virginia voted Obama in 2008 and 2012, then, even as so much of Middle America broke for Trump, doubled down on Hillary.

The election is looking very close and could still go either way, and no one knows what to believe. I think this survey of political yard signs gives the insight that Biden support is weaker than it may seem, and the odds must be more like 50-50 than the 90-10 they are talking about.

Yes, there is a large core that just hates Trump with a passion, and this is also reflected in the distribution of political signs. There were more anti-Trump signs than pro-Biden signs in the neighborhood. But beyond this angry and agitated core, will marginals in the middle go for Biden? Given the softness of his support even in such a strongly Democratic-voting area as the precinct of Arlington I surveyed, I can only expect the key states will see lackluster support for him, too. But maybe he doesn’t seen it, and maybe anti-Trump feeling alone, and the anti-Trump drumbeat we hear on the news, could be enough to win. It could.

It turned out Trump’s real odds of victory in 2016 were around 50-50, and he won because he came up heads in more Midwest states than not (all within about a point). The same could happen in 2020.

I remember November 8th, 2016. I was in Washington and saw a lot, many things to remember. Glimpses of what was to come in the mass outbreak of politicized violence and politicized(?)-looting in mid-2020.