Districts that are either brand-new or discontinued due to reapportionment are noted with dashed lines in the appropriate columns. Note that the chart above includes both old and new districts, so there are 443 in total rather than the typical 435.

In the right-most column, we include our best assessment for each new district’s corresponding predecessor district. Caution is warranted, though, because making these assignments is not an exact science—some districts have clear predecessors, while others don’t. Our ultimate goal is to ensure that each new district (except for the eight brand-new seats created by apportionment) has one and only one predecessor district after the 2022 elections are complete, and that each predecessor district is unique—i.e., that no two new districts share a predecessor district. It is possible in some instances that assignments will change as new information becomes available.

Our approach to making these determinations is generally incumbent-centered. Below are the considerations we rely on:

• District numbers can be helpful but they are not dispositive. Just because a new district shares a number with an old district does not necessarily mean the latter can be regarded as the predecessor of the former. Maps are often renumbered, sometimes radically so, as California’s was following the 2010 round of redistricting.
• If an incumbent is seeking reelection, our presumption is to regard the old district they previously represented as the predecessor of the new district they are seeking to represent. Given the high proportion of House members who seek reelection (an averge of 89% between 2006 and 2020), this presumption guides most of our decisions.
• We follow the presumption above even when an incumbent runs for a new district with a number different from that of their old district.
  • For example, following the 2010 round of redistricting, Democratic Rep. Jim Matheson sought reelection in the new UT-04 even though his prior district was the old UT-02. We therefore designated UT-02 as the predecessor of UT-04.
• In the event that an incumbent has not yet announced which district they will run in, we will assign them to the district we believe they are most likely to seek.
  • For example, Democratic Rep. Kurt Schrader could conceivably seek reelection in either OR-05 or OR-06, but his hometown of Canby is located in OR-05, so we have assigned him to that district.
• Similarly, if an incumbent retires but it’s reasonably clear where they would have sought reelection, then we will assign predecessor districts accordingly.
  • For example, following the 2010 round of redistricting, Democratic Rep. Gary Ackerman was planning to run in the new NY-06 before announcing his retirement, so we designated his prior district, the old NY-05, as the predecessor district.
• If two incumbents face off against one another, we designate both districts as predecessors. When one wins, we remove the loser’s district and leave only the winner’s as the predecessor. We follow this procedure for both primaries and general elections.
  • For example, following the 2010 round of redistricting, Rep. Mark Critz, who represented the old PA-12, and Rep. Jason Altmire, who represented the old PA-04, ran against one another in the Democratic primary for the new PA-12. After Critz won the primary, we designated the old PA-12 as the predecessor of the new PA-12.
• If a district was newly created due to reapportionment, then we have not designated any predecessor.
  • For example, Oregon earned a new district following the 2020 census, which is numbered OR-06. Note that “new” districts are not always those with the highest number in their state. This can be due to renumbering (as in the California example above) or because of decisions incumbents make when seeking reelection (as in the Utah example above).
• Conversely, in states that are losing one or more districts due to reapportionment, a commensurate number of old districts will always lack a successor.
  • For example, Pennsylvania lost a district following the 2010 census, resulting in the primary described above between Critz and Altmire. Because Altmire lost, his district, the old PA-04 has no successor.

We generally do not rely on data showing how the population has been redistributed from old districts to new districts (which we calculate separately here) in order to make these assessments. There are two main reasons for this.

First, we believe that an incumbent’s decision about where to seek reelection says more about political continuity between districts. For instance, a representative’s old district might get divided between new districts, one that has 60% of the population of the old district and another that has 40%. The incumbent’s home, however, might wind up in the latter, prompting them to seek reelection there, even though more of their constituents have wound up in the former.

Second, using population figures would often make it appear that two districts share the same predecessor. For example, following court-ordered redistricting in 2018 in Pennsylvania, the new PA-08 and PA-09 both traced their ancestry to the old PA-17 because it provided a plurality of the population for both new districts. In practice, however, the new PA-08 was properly designated the successor of the old PA-17 because Democratic Rep. Matt Cartwright, who represented the latter, chose to seek reelection in the former. (We also determined that new PA-09’s predecessor was the old PA-11.)

We take an incumbent-centered rather than electorate-centered approach to determining predecessor districts because, in the end, elections are won and lost by individual candidates, and that’s the lens through which we most frequently discuss politics. But we still believe it’s important to analyze how electorates flow from old districts to new because, of course, it’s those voters who will decide elections, which is precisely why we calculate the redistribution tables linked above.

In rare instances, our incumbent-centered approach will lead to outcomes where there’s little or even no overlap between a new district and its predecessor. For example, following the 2010 round of redistricting, which saw California’s map radically redrawn, Republican Rep. Gary Miller chose to seek reelection in the new CA-31, even though none of his constituents from the old CA-42 lived there. That fact, though, is the sort that we’d be certain to take note of in any discussion of a race like that.

As you peruse the data in the chart above, note that the figures for the old districts reflect those for the same district number, even if the old district is not the predecessor of the new district with the same number. Districts with a numerical mismatch are flagged in bold in the “Predecessor District” column on the far right. In such cases, it is necessary to manually match up old districts with their successors for a more direct comparison. We will elaborate with an example when one presents itself.

For presidential election results for the 2016 and 2012 presidential elections (under old district lines), click here. For all of our district-level election calculations, click here.