• The Civitas Partisan Index (CPI) is a measure of the partisan tendency of state legislative districts
  • Although the CPI is not a predictive model, someone using the CPI alone would have correctly predicted the outcome of 98.8% of races assigned a partisan lean
  • The two districts (of 162 assigned a partisan lean) the CPI got wrong were both won by incumbents in areas that are trending toward their party

NOTE: Lawsuits may affect the outcome of some races. If there are any changes, we will update this post.

With the 2024 (almost) behind us, it is time to assess how the 2024 Civitas Partisan Index (CPI) performed as a tool to evaluate the partisan lean of North Carolina’s General Assembly districts.

What Is the Civitas Partisan Index?

The CPI measures the partisan leanings of North Carolina state legislative districts, similar to the Cook Political Report’s Partisan Voting Index (PVI), which measures the partisan leanings of congressional districts.

The 2024 CPI used data from the North Carolina State Board of Elections to calculate the precinct-level results of the ten 2020 Council of State races and six 2022 statewide judicial races. The Republican votes of sixteen races were added together in each state legislative district and divided by the total two-party votes for those sixteen races to get a Republican percentage in each district. That percentage was then subtracted from the Republican two-party statewide average for those sixteen races (51.36%) to find the CPI rating for each district.

Someone using the CPI alone would have correctly predicted the outcome in 98.8% of races assigned a partisan lean.

The CPI score lists the party the district tends to favor, followed by a number indicating its relative lean toward that party. For example, a district that tends to vote about 3 percent more Democratic than the state average would have a CPI score of D+3.

We also assigned each district a rating based on its CPI score. We rate districts with a CPI of 0–1 in either direction as a toss-up, 2–5 as lean, 6–9 as likely, and ten or more as safe seats for the respective parties. Districts with a 0 rating have a partisan lean of less than half a percentage point. The D+3 district in the example above would be rated lean Democratic.

Click here to find a fuller description of our methodology.

How Well Did the CPI Do?

The CPI is not a predictive model since it does not measure other factors that affect election outcomes, such as candidate quality and political environment.

However, as seen in Table 1, a person using nothing except the CPI would have correctly predicted the result in 115 of the 117 North Carolina House districts where the CPI assigned a partisan tendency, a 98.3% accuracy rate. The table does not include 37 safe Democratic and 39 safe Republican House districts.

Table 1: Winning North Carolina House races by Civitas Partisan Index rating, excluding 37 safe Democratic and 39 safe Republican districts

Dr. Chris Cooper at Western Carolina University produced a scatterplot comparing House CPI scores with election results. I am reposting it here with his permission. Note that the X and Y axes are different scales, reflecting the fact that winning candidates tend to outperform the district’s baseline tendency. That is natural since the CPI incorporates data from statewide races that both overperformed and underperformed in the district.

Figure 1: Scatterplot comparing North Carolina House CPI scores and 2024 election results. Reposted with permission from Chris Cooper.

The CPI was perfect in the Senate. Using only CPI ratings, a person could have successfully predicted the outcome of all 45 races in districts assigned a partisan lean. Part of the reason for that perfect accuracy is that there are relatively few “lean” Senate districts (10 percent compared to 19 percent of House districts).

Table 2: Winning NC Senate races by Civitas Partisan Index rating, excluding 16 safe Democratic and 19 safe Republican districts.

Again, Chris Cooper’s scatterplot comparing CPI scores and election results is illustrative.

Figure 2: Scatterplot comparing North Carolina Senate CPI scores and 2024 election results. Reposted with permission from Chris Cooper.

Overall, the CPI was correct in 160 of 162 districts assigned a partisan lean (98.8 percent). Even if you also consider the toss-up districts, someone using only the CPI would have correctly predicted the outcome of 164 of 170 races (96.5 percent).

What About the Two Races the CPI Got Wrong?

Someone using the CPI to predict election outcomes in districts assigned a partisan lean would have gotten two races wrong: House districts 25 and 115.

House District 25 includes most of Nash County in eastern North Carolina. Republican Allen Chesser first won the seat in the state’s 2022 red-wave election. Like much of eastern North Carolina, Nash County has been trending Republican. The CPI rated the district D+2 (lean Democratic), but Chesser held on to the seat on November 5, winning 48.8 to 47.9 percent.

House District 115 is in Buncombe County, which has been trending Democratic. After court-ordered redistricting, the formally safe Democratic district became R+2 (lean Republican). Incumbent Democrat Lindsey Prather, who, like Chesser, is serving her first term in the House, won 51.5 to 48.5 percent.

The races had several things in common that help explain why the election result did not match the CPI rating. First, both races had CPI ratings that the minimum required to be assigned a partisan leaning (+2). Other factors in common are:

  • Incumbency
  • Counties that are trending their way
  • Candidates reputed to be good campaigners

Those are the kinds of things a predictive model would include. However, they are not part of the Civitas Partisan Index, which only incorporates currently available election data.

Still, 160 out of 162 isn’t bad.