Pres-by-CD: We hit Maryland for our project to calculate the 2016 presidential results for all 435 congressional districts nationwide. You can find our complete data set here, which we're updating continuously as the precinct-level election returns we need for our calculations become available.
Maryland’s was one of the few congressional maps drawn exclusively by Democrats. Hillary Clinton handily carried the same seven congressional districts that Barack Obama took in 2012, while Donald Trump easily won the conservative 1st District along the Eastern Shore. The suburban D.C. 6th District was the closest seat, but it still backed Clinton by a 56-40 margin, an improvement on Obama’s 55-43 win here. Democratic Rep. John Delaney had a close call during the 2014 GOP wave and only won 50-48; however, Delaney won 56-40 last year, matching the top of the ticket. Clinton took at least 60 percent of the vote in the other six Democratic-held seats.
While Maryland’s results may not be the most interesting we’ve calculated, there’s one big issue with the state’s data we want to address, and it’s one that applies to other states as well. Maryland, like many states, allows residents to cast ballots before Election Day, both at early voting locations and by traditional absentee voting (we’ll collectively call these both “early votes" for ease). However, in the results they provide to the public, Maryland counties do not assign these early votes to any particular precinct, meaning we don’t know which congressional districts they belong to.
It’s a totally shoddy approach, because the state does know which precincts these votes were cast in. How else could they tally the results of House races? For instance, Anne Arundel County (home of Annapolis) is split between four different congressional districts. Election officials naturally have to make sure early votes cast in Anne Arundel are tabulated in the proper congressional district, but they simply don’t make this data public.
Unfortunately, Maryland is not the only state that refuses to break out early votes by precinct. But what makes the Old Line State particularly problematic is that its proportion of unassigned early votes is far higher than in any other state where we’ve calculated these results (the same was true in 2012). We've contacted Maryland’s state and county election boards to try to find more precise data multiple times, but we’ve repeatedly been informed that this data does not exist—even though we know it has to.
So what do we do? We have a formula we use in every state that has unassigned early votes; it allocates those votes between districts based on the how much of each county is in each district, and how the portion of each county in a given district voted on Election Day (where we do have breakdowns by precinct). This is an imperfect and imprecise method, and we wish we didn’t have to use it, but election officials leave us no choice. Luckily, in this case, none of Maryland’s eight congressional districts were at all close in the presidential race, so we’re confident that our results have correctly identified the winner in each district.
P.S. Confusingly, Maryland does have data files that break down just the Election Day vote by congressional and state legislative district, while ignoring the many early and absentee votes. However, since those files leave out over 40 percent of the vote, they aren’t very useful.