A U.S. House forecaster’s guide

Most independent analysts now believe Democrats are extremely likely to retake the U.S. House of Representatives this fall. As we assess their chances and the number of seats they might pick up, it’s helpful to recall that not all districts are created equal. Marginal House seats divide roughly into four groups, only three of which are fertile ground for Democratic wins. That will likely limit their gains in the House and make any Democratic majority small and tenuous, unless the political ground shifts further in their direction.

The first and most important group involves 25 seats in districts Hillary Clinton carried in 2016 and currently represented by Republicans. To win House control, Democrats must carry the large majority of these seats. They fall into two parts—one dominated by high-income, highly educated suburban whites and one dominated by Hispanic voters. Democrats need a net gain of 23 House seats to obtain a majority: They’ll need to win at least half, and probably two-thirds, of these seats to be sure of that.

The second group involves 21 seats in districts Barack Obama first carried in 2012 and then Donald Trump carried in 2016. Democrats hold nine of these, and all of the Republican targets for pickups to offset their losses elsewhere are among these seats. Republicans hold 12 and Democrats are currently targeting eight. These districts are dominated by white, rural, blue-collar voters that national media tend to overlook. Democrats probably need a net gain of three to five of these seats if they are to retake the House.

The third group involves 10 seats in districts Romney carried in 2012 and Trump carried in 2016 with less than 50 percent of the vote. This is swing territory, with voters who tend to be more habitually Republican and less well-off and less educated than those in the first group. These places also tend to have very few racial minorities. To win these districts in November, Democrats will need either an unusually high turnout from their own voter base or strong support from voters who did not support Clinton. Democrats need to win at least four of these seats to ensure they gain a House majority.

The final group consists of 27 seats in districts won by both Romney and Trump, with Trump receiving no more than about 52 percent of the vote, or districts Trump carried with an 11 percent margin or less. These districts are safely Republican in most years, but they might be in play this year if the Democrats can repeat their outstanding recent performances in special elections: Margins in federal special elections in the Trump era have shifted about 11 percent in Democrats’ direction, compared with the 2016 presidential election margins in those same special election districts. If Democrats can replicate that shift in this November’s midterms, they could win almost all of the 27 seats in this group. If they are to have a “blue tsunami,” an overwhelming wipeout win similar to what Republicans scored in the 2010 midterms, Democrats will need to pick up some of these seats along with the vast majority of seats in the other three groups.

Assigning House seats to these groups makes it relatively easy on election night to forecast the number of Democratic gains: If early poll results show them winning almost all of the seats in the first two groups, and half or more of the seats in the third group, then Democrats will surely retake the House, perhaps by a wide margin. On the other hand, if Democrats are winning only about half of the Clinton-carried seats and only a couple of the Obama-Trump seats, we should expect Republicans narrowly to retain control.

Forecasting elections is an art, not a science. Using this breakdown as a guide, though, makes a difficult task easier.