How was DC’s primary different this year?

Police brutality, protests against police brutality, the pandemic, and curfews all shape this year’s election.

Our post today was written by Pia Deshpande, a writer with the MIT Election Lab.

Joe Biden handily won the District of Columbia’s Democratic primary on June 2. But this time, the most interesting questions to ask weren’t who won? or why? The DC primary and many others were held as Americans, led by Black activists, protested (and continue to protest) police brutality and the unlawful deaths of Black Americans like George Floyd and Breonna Taylor. Cities instituted strict curfews with varying levels of police enforcement (enforcement that often has dangerous consequences for Black Americans). At the same time, the country is still caught in a global pandemic, which has complicated many of the usual processes for election administration and voting. These factors set the 2020 presidential primary immediately apart from primaries held in 2018 or 2016.

This year’s election saw high turnout despite a relatively noncompetitive Democratic presidential primary and a completely noncompetitive Republican presidential primary. Predominantly Black wards voted at high in-person rates on Election Day despite the threat of COVID-19, which is more deadly to Black individuals than other racial groups. Finally, though it is too early to say anything definitive, the data record on the effects of the coronavirus, police brutality, and protests on voting is growing. Political scientists should consider all these things when they conduct historical analyses on these elections.

This article discusses some of those differences above and more. It also focuses on how Black Americans voted, with a certain data limitation: DC does not provide turnout data or absentee ballot data by race. However, the city does provide data by ward, and the racial composition of each ward. Using this as a proxy, we’ll dive into just how voters participated in this primary, and the potential factors political scientists should consider when analyzing this primary in the future as more data become available.

Before we get started, a disclaimer: recent changes to DC’s ballot deadlines mean these results are incomplete. Because many voters did not receive their absentee ballots in time (or at all), DC will continue counting ballots until June 12. That being said, they have published unofficial election results and voting statistics on their website; these are the data we will be using. It’s also important for us to note that in general, political scientists look for the effects of changes by isolating the factor that is changing and holding the rest of the world constant (or at least, trying the best we can to do that). Since this analysis and the data it draws on is new, it’s in many ways incomplete; we can’t know definitively what caused the changes we find yet. But throughout this article, I’ll try and be clear about the uncertainty inherent in our data.

Turnout is higher this year than in 2016 for both parties

Turnout tells us what portion of the population shows up to vote in an election.

There are a few ways to calculate turnout: you can measure it as votes cast divided by the total voting eligible population (which is preferable), or as votes cast divided by the number of registered voters. These two methods will yield different results, with the first always providing a more pessimistic turnout than the second. Using data compiled from DC’s Election Office, this paper will use the second method, as voting-eligible population isn’t calculated separately for political parties.

Let’s first look at turnout broadly in the 2020, 2018, and 2016 primaries (while only the 2016 and 2020 election years have presidential preference primaries, it’s useful to look at all three):

Overall turnout in DC is the highest it’s been since before 2016, with a modest 4% increase in total turnout from the 2016 presidential primary election to the 2020 primary. For Democrats, who also saw turnout grow during these DC primaries, this increase is around 3.8 percentage points.. Republicans saw a large increase in turnout of around 4.6 points (nearly doubling their turnout in 2016), which is mostly due to a large increase in the 2018 primary (when Democratic turnout dipped). This sets the DC Republican primary apart from those of other states, most of which saw a decline in turnout for their Republican primaries or little change at all.

DC primaries are unique in that the city is overwhelmingly Democratic. In 2020 there were 410,260 registered voters in DC — but only 28,357 were Republican. With the Democratic primary providing a chance to members of the party to choose the next nominee (even if the nomination pretty much decided) and no competition on the Republican ballot, it is not hard to understand why turnout for GOP voters was only one-third that of Democrats this year. However, Republican turnout has always been low in DC, even when the 2016 primary did provide them a competitive race to influence.

Helpfully, DC also documents the method people use to vote. The district’s election website reports three different vote types: election day votes (votes in-person on election day), early votes, and absentee (mail-in) votes.

Voting absentee (or voting by mail) has become a controversial subject this election cycle, as national attention on it grows because of the coronavirus pandemic. Many Republican legislators, including the President, have publicly denounced voting by mail, claiming that it will help Democrats far more than Republicans in upcoming elections.

So far, there has been little evidence for this claim. Looking at DC, for example, the assertion that Democrats are helped more by voting-by-mail than Republicans doesn’t hold up. The graph below illustrates the percent of ballots cast in each party’s primary that were categorized as absentee. In this case, absentee ballots consistently make up a larger percentage of Republican votes than Democratic votes each primary since 2016. In fact, in 2016, around 19% of Republican votes were absentee. This year, during the pandemic, when voting by mail was the norm for voters of both parties, the number of Republicans who took advantage of absentee ballots still outstripped Democratic voters.

Predominantly Black wards see lower turnout, are more likely to vote in-person

An analysis of turnout would not be complete without examining whether or not turnout differed depending on race. This is especially true as we consider the effects of an increased police presence, police violence, and the coronavirus pandemic on Black Americans.

Unfortunately, DC does not provide data on turnout by race. However, they do provide election data and demographic information for each ward, which we can use as a rough proxy. First, let’s take a broad look at the ethnic composition of DC’s wards (as defined by the data on the DC government website):

The district as a whole, on average, is around 42% White, 45% Black, and 4% Asian (“Hispanic” is not included by the DC government here because it is considered an ethnicity and not a race). The individual wards, though, can deviate significantly from this city-wide average. Wards 7 and 8 are 92% Black, for example, while Ward 3 is only 5% Black.

The visual above depicts some of the large demographic swings from ward to ward. For a more granular look at the data, take a look at the table below:

With these demographics in mind, let’s examine levels of turnout in 2020 for each DC ward:

The ward with the highest turnout, Ward 3, is overwhelmingly White. The wards with the lowest turnout, Ward 7 and Ward 8, are overwhelmingly Black. Ward 4 is the only jurisdiction that is not majority white that saw a turnout greater than 30% — this ward, along with Ward 2, had a competitive primary that probably had a positive effect on turnout this year.

So, those are the numbers from 2020. How do they differ from previous primary elections? When we look at the data, we can see that all wards saw an increase in turnout compared to the 2016 presidential primary, but some wards saw a larger increase than others.

Ward 2 and Ward 4 have the highest increases in turnout at a considerable 8 percentage point rise (remember from above that Wards 2 and 4 had very close races this year). Ward 2 is the second Whitest ward in DC. Ward 4 is 31% White, 47% Black, and 15% some other race. Wards 7 and 8 have the smallest increases.

Looking beyond overall turnout, we already know that voting was markedly different this year because of the pandemic. So let’s now take a look at how each ward voted. In 2016 and 2018, there is not much variation in what percentage of ballots are cast in-person by ward. In 2020, stark patterns emerge.

Ward 3 (the Whitest ward in DC, with approximately 82% of its residents identifying as White) saw the lowest rate of in-person voting. In stark contrast, Wards 7 and 8 (which are 92% Black) saw the highest rates of in-person voting. In Ward 8, the ward with the highest percentage of in-person ballots, around 66% of votes were cast in-person on Election Day.

It’s too early to know what explains these discrepancies in voting methods. We don’t have poll data yet that asks voters why they chose to go to the vote in-person. However, we do know that to get an absentee ballot in DC, voters need to apply for them in advance and receive them in the mail with enough time to fill out the ballot (correctly, otherwise the ballot may be tossed) and send it back to be counted.

It will be a while before we get a comprehensive picture of how DC managed this primary, but there are already accounts of some voters in Ward 4 waiting past midnight to cast their ballots. Many voters have also complained that they never received their ballot in the mail. The confusion even prompted DC councilmember Elissa Silverman to write a letter to the DC Board of Elections, expressing her concerns about the number of DC voters that had contacted her office to report that they had requested an absentee ballot but never received one:

“In the last 24 hours, I have received over 380 emails from DC voters who requested an absentee ballot, did not receive one, and then learned they can receive a ballot via email. Right now, at 5 pm, I am receiving a request almost every minute for an emailed ballot.”

Silverman also expressed a concern that many voters may not have known about the emailed ballot option, and therefore may not have voted because of their concerns about going to an in-person polling place.

Black DC residents make up a plurality of COVID cases and a majority of COVID deaths

Continuing our ward-by-ward analysis, remember that the COVID-19 pandemic has been a primary concern of voters and election officials this year, especially the risk of in-person voting. Wards 7 and 8 (overwhelmingly Black jurisdictions) voted in-person this year at much higher levels than other wards. This puts those voters at a higher risk of infection by the coronavirus.

Political scientists often discuss voting as a cost-benefit analysis conducted by the voter. Someone may feel that because they live in a swing-state, that means their vote is more likely to count (which makes it seem more valuable), but it may be very difficult for them to get to a polling station (which makes the cost of voting seem higher). The added threat of the coronavirus could be enough to convince voters to stay home, if they were planning on voting in person.

Vote-by-mail (VBM) measures adopted in many states seek to remedy this additional cost, and protect the health of voters and poll workers alike. It’s too early to know for sure how effective these measures will be. They provide constituents an opportunity to vote safely from home, but the unfamiliarity of a VBM ballot may result in some voters not applying on time or making disqualifying mistakes. Therefore, coronavirus remains an impediment to voting that should be taken seriously.

It’s also important to note that COVID is has been more deadly to Black Americans. Looking at the week-by-week data provided by DC through the date of the primary (June 2), we can see that Black DC residents often make up a plurality of COVID cases. Black DC residents also have seen a 1,274% increase in COVID cases from the start of data collection on April 5 to June 2, the third largest percent increase of any racial group (American Indians/Alaskan Natives and DC residents who identify as multiracial see even larger percent increases). White residents have seen a comparably smaller increase of 910%.

Black residents also make up the vast majority of COVID deaths in the city. (Eagle-eyed readers will note that the category “Hispanic” is included in the data for coronavirus deaths but not for cases. This is a function of the data provided by DC. In their cases data, they examine ethnicity and race separately. The only categories listed under ethnicity were “Hispanic or Latino,” “Not Hispanic or Latino,” “Unknown,” or Refused to Answer.)

These data provide an incomplete picture of the virus’ permeation into DC wards, but the picture painted here is still valuable. There are likely far more cases and COVID-related deaths than the ones recorded by the city (this is why it is important to consider testing data when we can, but currently DC doesn’t provide testing data disaggregated by race). All this means that the health risk posed by COVID increases the cost of going to the polls for Black Americans by a significant amount; more so than for other Americans.

COVID has affected more than the country’s public health. With the start of the pandemic, the U.S. fell into an economic recession sometime in late February according to the National Bureau of Economic Research. This may motivate people to vote for change at the ballot box, or it may make voting (and taking time off work to do it) even harder. While it’s too soon to draw any conclusions on this front, it will be an important issue for future analysis.

Local politicians feared that curfew dampened Black turnout

A day before the June 2 election, Mayor Bowser instituted a 7pm curfew in response to looting, joining several other mayors who try to stem the destruction of property amidst a largely peaceful anti-racist protest.

All curfews that have been implemented have made exceptions for essential workers and those Americans trying to exercise an essential liberty. For example, DC’s Mayor Bowser’s curfew order allowed residents who were trying to vote to do so.

However, even if a curfew still allows for voting, that does not mean that Americans will continue to turn out to the polls at the same rate they would have otherwise. Some voters, particularly Black Americans, may fear violence enacted upon them by the police whether or not they are actually violating curfew. In DC, many feared that the curfew laws were overly vague and could lead to selective enforcement, allowing the police department to perpetuate well-documented policing biases against Black individuals.

Over 40 elected DC officials called on Mayor Bowser to move the city’s curfew later on the night of the election, because they feared the measure would keep “communities of color” at home out of fear of police violence. Curfews have had a long history of being used to victimize Black individuals or quell anti-racist movements, and it is possible that this curfew kept Black voters home. There are multiple reports of voters waiting in line being told to go home by police officers alleging that they were violating curfew, even though voting was allowed past Bowser’s cutoff time of 7pm. It is also possible that Black voters may have stayed home in fear of police violence and the possibility of increased police interaction because of the curfew. However, until we have race-specific turnout data for that analysis, it will be difficult to know more.

Data on police violence during protests is sparse, but growing

As protests against police brutality have swept across the country, DC has struggled as unmarked federal agents and out-of-district national guard units have stationed in the district, though Mayor Bowser has requested their removal. The inclusion of these out-of-district police officers and agents only serve to hugely increase the presence of police in DC, and with it the fear of police brutality.

Unfortunately, data on how this heightened presence and possibility of violence might affect an ongoing election is especially sparse. The protests are incredibly new, and data collection is imperfect at best.

There are a few resources to pull from. Greg Doucette has compiled video evidence of instances of police brutality in this Twitter thread. Jason Miller has taken these instances and sorted them into a spreadsheet, and Manuchehr Aminian has visualized the data. These resources list 18 instances of police brutality during protests in DC.

There are limitations to this data. There are almost certainly instances of police violence that were not captured on film, and therefore are not part of this dataset. The instances are also not easily mapped onto a particular location within DC (so, no ward-level analysis). However, when we do have more complete data, it will be incredibly important to consider heightened police presence and police violence in DC as both a motivator and a deterrent to go to the polls.

Takeaways

• DC Wards 7 and 8 (which are approximately 92% Black) voted in-person at extremely high rates this year. Ward 3 (which is approximately 82% White) voted absentee at the highest rate out of all the eight wards.
• Black Americans make up a plurality of DC COVID cases and a majority of DC COVID deaths. Voting in-person presents an even greater risk to them than Americans of other racial groups.
• The wards with the lowest turnout this primary were predominantly Black (Wards 7 and 8; Ward 5 is a close third). The ward with the highest turnout is Ward 3, which is predominantly White.
• Every ward in DC saw an increase in turnout this primary compared to 2016. However, we see the smallest increase in predominantly Black Ward 8. The second smallest increase in turnout is seen in Ward 1, which is majority White (though not as White as Ward 2 or 3).
• The wards with the largest increase in turnout are Wards 2 and 4. Ward 2 is also predominantly White (69%) and has the largest percent of Asian individuals of any ward (9%). Ward 4 is 47% Black and 31% White.
• Data on the protests, police brutality during them, and the effects of curfews implemented because of both, are limited right now. However, all three of these factors affect Black Americans more than any other racial group in the United States and may significantly impact turnout. Once data becomes available in a more reliable way, it should be included in analysis. For now, it should be included as a potential factor that we just don’t have the best data to measure.

Though some of these takeaways are specific to DC, political scientists will need to keep them in mind as they begin to study other primaries that have happened and will happen with the backdrop of a pandemic, protests, and police brutality (all of which affect Black Americans disproportionately). COVID will be a factor in every primary (to varying extents), and has proven more deadly to Black Americans than other races. Fear of police violence and ongoing protests will also likely be a factor in other elections this month. It is unclear how these things will affect election results, Black turnout, and access to the polls, but researchers should pay attention to them.

Political scientists should also pay attention to who votes absentee and who does not. In this analysis, we found that Black-predominant wards voted in-person much more frequently than other wards. This put them at a higher risk of COVID infection and death. We don’t know if this pattern will repeat itself across the country, but the implications of it in DC alone are serious and worth further study.

Pia Deshpande is a writer with the MIT Election Lab. She is a graduate of Columbia University.

The MIT Election Data + Science Lab is dedicated to applying scientific principles to how elections are studied and administered, with the aim of improving the democratic experience for all U.S. voters. For more updates, follow us on Twitter and sign up for our quarterly newsletter.