Abstract

Many voters, even when they are motivated to go to the polls or mail in their ballots, tend to vote only in high-salience contests at the top of the ticket and not in state legislative races that appear farther down the ballot. This behavior, termed down-ballot roll-off, has not been studied extensively and is not well understood, particularly at the level of state legislative races. The potential partisan valence of this behavior has also received relatively little attention. SDAN previously conducted an analysis using election results from 10 states during the years 2012-2020 and found that, while down-ballot roll-off commonly affects candidates of both major parties, it occurred much more frequently for Democratic candidates than Republican candidates, and this pattern held even when the analysis was restricted to contested state legislative races only. Here, we build on the previous study, first by replicating the aggregate-level analysis using 2022 election data from the same states, and then conducting more granular analyses to understand how much roll-off is occurring at the district level. We find that the aggregate-level results are consistent with those observed for 2012-2020: Roll-off occurs in the majority of cases, and there is a notable partisan skew such that Democratic candidates at the top of the ticket almost always receive more votes than Democratic state legislative candidates, while the opposite is true for Republican candidates. We observe a similar partisan skew at the district level, where the average district-level roll-off rates vary considerably by party and are higher for Democrats in the majority of cases. These findings are particularly remarkable given how well state legislative Democrats did in 2022—the first midterm election since 1934 where the president’s party did not lose a single legislative chamber. Even in this positive electoral context, Democratic candidates still experienced higher rates of roll-off than Republicans.

Main Takeaways

Aggregate level: We frequently observe cases of down-ballot roll-off at the state chamber level in 2022.

• We find that roll-off occurs even when the analysis is restricted to contested state legislative districts only.
• State legislative Democrats in contested races experienced roll-off 83% of the time.¹ Meanwhile, down-ballot Republicans in contested races experienced roll-off just 13% of the time. 
• We also observed some overages, where state legislative candidates collectively received more votes than their party mates at the top of the ticket. Among candidates in contested races, state legislative Republicans experienced this down-ballot advantage in 87% of the chambers, compared to just 17% of chambers for state legislative Democrats.
• Given how successful state legislative Democrats were in the 2022 elections, it is remarkable that they still experienced higher rates of roll-off than state legislative Republicans.
• Combined, for the years 2012-2022, state legislative Democrats in contested races experienced ballot roll-off 80% of the time (106 of 132 cases), versus just 32% for contested state legislative Republicans (42 of 132 cases).
• Combined, for the years 2012-2022, state legislative Democrats in contested races experienced ballot overages just 20% of the time (26 of 132 cases), versus 52% for contested state legislative Republicans (68 of 132 cases).

District level: In 2022, state legislative roll-off occurred in the majority of districts in our sample.

• Among districts with contested races, we observe Democratic roll-off in 68% of the districts, but we observe Republican roll-off in only 20% of them.
• Average rates of roll-off within districts are also typically higher for Democrats than Republicans.
• The average rates vary depending on the state and the type of race at the top of the ticket (governor vs. U.S. Senate), but overall, Democratic state legislative candidates appear to be more negatively affected by roll-off than Republicans.

Executive Summary

Chamber-level roll-off

As we observed in the 2012-2020 analysis, down-ballot roll-off occurred almost universally across state chambers in the 2022 elections. States included were: AZ, FL, GA, MI, MN, NV, NC, PA, TX, and WI. Of the 30 aggregate data points in this sample, roll-off occurred in nearly every case (the exception is Minnesota, where more votes were cast for state house and state senate candidates in 2022 than for governor). In every other state, regardless of chamber (upper or lower) or type of race at the top of the ticket, more total votes were cast for gubernatorial or U.S. Senate candidates than for state legislative candidates.²

Chamber-level roll-off by party

Different patterns emerge when we analyze roll-off by party. At the chamber level, there are far more cases of roll-off between Democratic candidates at the top and bottom of the ticket than Republican candidates. In fact, we observe roll-off on the Democratic side in every case, while overages are much more common among Republican candidates. That is, Republican state legislative candidates received more total votes than their Republican counterparts at the top of the ticket in most cases (24 of 30, or 80%). We observe Republican roll-off in only 6 cases (20%), and all of those occurred in governor’s races. Of the 13 U.S. Senate data points, Republican state legislative candidates collectively received more votes than the Republican U.S. Senate candidate in every state chamber in our sample. Down-ballot overages, not roll-off, occur consistently among Republican candidates, while the opposite is true for Democratic candidates.

These partisan differences persist even when the analysis is restricted to contested state legislative races only—see Table 1. Races are considered contested if a candidate from each major party was officially nominated and both candidates appeared on the ballot in the general election. On the Democratic side, we still find that gubernatorial candidates received more votes than state legislative candidates in 12 of 17 cases (71%), and every Democratic U.S. Senate candidate received more votes than their state legislative co-partisans (13 of 13 cases). By contrast, Republican gubernatorial candidates received more votes than state legislative candidates in only 4 cases (24%), and, as before, none of the Republican Senate candidates received more votes than Republican state legislative candidates. In short, even in a restricted analysis of contested districts only, we observed Democratic roll-off in 83% of the chambers in our sample, while we observed Republican roll-off in just 13% of the chambers.

Table 1: Cases of state legislative roll-off, by party, contested races only

Percentages of districts with Democratic roll-off vs. Republican roll-off

Measuring the frequency of roll-off another way, by calculating the numbers of districts where state legislative roll-off occurred, we find that roll-off in contested districts with a governor’s race occurred in 327 of 911 districts (35.89%), and in 178 of 608 contested districts with a U.S. Senate race (29.28%). Importantly, we still observe a clear partisan skew. Of the 911 districts in our sample with contested state legislative races and a governor’s race at the top of the ticket:

• Democratic roll-off occurred in 615 of them (or 67.51% of the districts).
• Republican roll-off occurred in far fewer districts—183 of 911, or in just 20.09 percent of the districts.
• In 563 districts, Democratic roll-off but not Republican roll-off occurred; the reverse case is true in only 131 districts.

The pattern is nearly identical for districts with Senate races at the top of the ticket:

• Democratic roll-off occurred in 400 of 608 districts (or 67.59% of the districts).
• Republican roll-off occurs only 20.39 percent of the time, in 124 districts.
• In 355 districts, Democratic roll-off but not Republican roll-off occurred, with the reverse case occurring in only 79 districts.

District-level roll-off rates

We calculate district-level roll-off rates to provide another indicator of how much state legislative roll-off occurred (expressed as the percentage of voters who rolled off), on average, in a district. The roll-off rate is the difference between the number of votes cast in the district for the major-party candidates in the top-of-ticket race (governor; U.S. Senate) and the number of votes cast in the district for major-party candidates in the down-ballot race (state house; state senate). The average, district-level rate of roll-off overall in our sample is 9.86 percent in gubernatorial races and 12.11 percent in U.S. Senate races. In other words, it appears that roughly 10 to 12 percent of voters in a district voted for a candidate at the top of the ticket but did not vote for a state legislative candidate.

As expected, these numbers decrease when we examine contested state legislative races only: For districts with contested races, the average district-level roll-off rates are 2.33% (gubernatorial) and 1.82% (U.S. Senate). 

Those average rates may seem very low at first, but an average is just one metric, and trying to capture how much roll-off is occurring with a single measure for the whole sample obscures important variation in the rates of roll-off—particularly by party or by state. Average, district-level roll-off rates by party and state, for districts that had contested state legislative races only, vary considerably. Comparing these different averages, we still see that district-level roll-off rates tend to be higher for Democrats than Republicans: as high as 14% in PA and 12% in MI. These numbers offer further evidence of the partisan skew in roll-off.

Differences in vote share

Focusing on the sample of districts with a gubernatorial race at the top of the ticket (911 districts with contested state legislative races), Democratic candidates for governor collectively carried 477 districts. However, the Democratic state legislative candidate lost in 59 of those districts (12.4%). By contrast, of the 434 districts that Republican gubernatorial candidates won, the Republican state legislative candidate lost in only 10 of those districts (2.3%). 

Vote-share comparisons using U.S. Senate races yield similar results (n = 608 districts). Democratic candidates for the Senate carried 305 districts in this sample, and Democratic state legislative candidates lost in 27 of those districts (8.9%). Republican state legislative candidates lost in only 6 of the 303 districts that Republican Senate candidates won (2.0%). In the vast majority of districts, candidates of the same party prevailed at the top and bottom of the ballot, but in cases of divergence, more Democratic state legislative candidates suffered losses than Republicans.

 Votes to majority: The electoral consequences of roll-off

To illustrate the impact of the votes “lost” to roll-off, we selected 3 competitive, Republican-controlled state legislative chambers and determined the number of additional seats Democrats would have needed to win in 2022 in order to take the majority. We then calculate the total number of votes it would have taken to win those seats, using the districts where Democratic state legislative candidates lost by the narrowest margins. Finally, we calculate the amount of roll-off, or how many more votes the Democratic gubernatorial candidate received than the Democratic state legislative candidates, in those same districts. The comparison is displayed in Table 2.

Table 2: Democratic Roll-off vs. Democratic Vote Margins in 2022

These examples show concretely why roll-off is important: To flip the PA state senate in 2022, Democrats would have needed to win 4 more seats. In the 4 races that Democratic state senate candidates came closest to winning, the candidates lost by a combined total of 35,790 votes. In these same districts, 40,808 more votes were cast for the Democratic candidate for governor (Josh Shapiro) than for the Democratic state senate candidates. In the AZ Senate and WI House, the roll-off totals do not exceed the vote margins, but it remains clear that meaningful numbers of votes are being lost to roll-off.

Conclusion

Sister District’s previous 2012-2020 analysis found evidence of widespread down-ballot roll-off, across multiple states and top-of-ticket races, and showed that Democratic candidates were disproportionately affected by this behavior. Bringing this analysis up to date with results from the 2022 elections, we find again that roll-off occurs in the majority of cases. Furthermore, the partisan skew in roll-off at the chamber level persists—in almost every case, Democratic candidates at the top of the ticket received more votes than Democratic state legislative candidates, while the exact opposite is true for Republican candidates. At the district level, we also see that Democratic state legislative candidates are at a disadvantage; the average district-level roll-off rates vary considerably by party and are typically higher for Democrats than Republicans. Lastly, there were more districts where the Democratic state legislative candidate lost but the Democratic candidate at the top of the ticket won; on the Republican side, such cases are extremely rare.

In future projects, we will build on these findings and analyze the connections between rates of state legislative roll-off and structural or district-level attributes. For example, the variation in roll-off may be associated with certain candidate characteristics (e.g., incumbency, gender, race/ethnicity) or the demographic composition of the district, such as the percentage of young voters or voters of color. Institutional features potentially play an important role as well; states that give all registered voters the option of voting by mail, for instance, might have lower rates of roll-off because people have more time to fill out their ballots. We have demonstrated the patterns of roll-off at the district and chamber levels, but additional analyses are needed to explain what drives them.

For more details about the data, analysis, and results, a longer report can be found here. If you have questions about this study, please email Jillian.

¹ The unit of analysis here is the state legislative chamber, not districts or candidates: In 83% of the chambers, or cases, we observed roll-off at the state chamber level for Democratic candidates.

²This analysis uses district data that have been aggregated up to the state chamber level: 1) We take the total number of votes cast for state legislative candidates (major party) in a given state chamber; 2) we add up the number of votes cast for major-party candidates at the top of the ticket (gubernatorial or U.S. Senate) from each state legislative district; and 3) we compare the chamber-level totals from 1 and 2: If more votes were cast for candidates at the top of the ticket than for state legislative candidates, this counts as an instance of down-ballot roll-off; if more votes were cast for state legislative candidates, this counts as a down-ballot overage.