2020 Arizona Drop Off & Roll Off Data Analysis Takeaways

Climate Cabinet Action
5 min readAug 9, 2023

By: Andrew Hong


Recently, Climate Cabinet Action published an analysis of how ticket-splitting (or “drop-off”) and undervoting (“roll-off”) rates changed the down-ballot political calculus for climate organizers in North Carolina. Now, we’ve chosen to provide the same analysis for Arizona, another crucial state that presents a huge strategic opportunity for the climate movement. Take a look below at what we found.

Drop-off / Ticket-Splitting Details

Topline: Arizona ticket-splits by voting Democratic top-ticket and GOP down-ballot at a moderate rate (1–3% drop in Dem. support down-ballot), but at larger rates than North Carolina. White, older, Republicans ticket-split while Hispanic, young, Democrats don’t. Data revealed that Hispanic voters disliked Biden.

There are moderate rates of ticket-splitting (drop-off) in Arizona in the direction of Democrats top-ticket and Republicans down-ballot. There was a 2.5% drop in Democratic support in State Senate races from Democratic U.S. Senator Mark Kelly’s margins (in contested State Senate districts). This is different from North Carolina where there was no significant statewide ticket-splitting (statewide state legislature totals were typically <0.5% different from U.S. Senate results). However this margin is only 0.6% when replacing Senator Kelly’s numbers with President Joe Biden’s performance. In other words, Arizonans disliked Biden more than Kelly, thus GOP down-ballot voters split their tickets more in the U.S. Senate race to vote for Democrat Mark Kelly (but not Biden).

The gap between Kelly and Biden’s performance comes largely from Hispanic voters that soured on Biden. In precincts with over 75% Hispanic registered voters, Kelly outperformed Biden by an average of 3.6%. These same Hispanic precincts voted for Trump by 7.7% more in 2020 than in 2016, too. This follows trends nationwide of Hispanic voters souring on Biden (and/or warming to Trump) in 2020, especially prevalent in Texas and Florida.

Further, the only areas that consistently split their tickets with Trump top-ticket and Democrats down-ballot were these Hispanic-majority precincts. This is similar, and more pronounced, to Black voter trends in North Carolina.

Like in North Carolina, we see older voters who split their tickets at a higher rate, supporting top-ticket Democrat candidates while voting Republican down-ballot. However there is little evidence that young voters ticket-split the other direction (GOP-top, Dem-down) in Arizona like we saw strikingly in North Carolina. Instead, younger precincts did not ticket-split but voted for Democrats at roughly equal rates up and down the ballot.

Democrat neighborhoods (by voters’ party registration) voted for Democrats in top-ticket (U.S. Senate) and local races (ex: State Senate) at pretty similar rates, but in very Republican neighborhoods a substantive number of voters voted for Democrat Mark Kelly for U.S. Senate but remained Republican in state races (in 2020).

Roll-Off / Undervoting Details

Topline: Arizona voters skipped down-ballot races more than North Carolina. About 4.3% of Presidential voters skipped State Senate races (when contested). Age was the biggest predictor.

Compared to North Carolina, Arizona has higher roll-off (undervoting) rates from top-ticket to down-ballot races. About 4.3% of voters who voted for President did not vote for State Senate races, and about 3.5% of voters (v.s. 1.9% in North Carolina) who voted for U.S. Senate did not vote for State Senate races (when throwing out uncontested races).

There is a stronger association between age and roll-off rates in Arizona than in North Carolina, with younger voters skipping down-ballot races more. There was a 40% increase in roll-offs (skipping down-ballot races) in younger-than-average precincts than older precincts. In precincts with a below average median age, 4.9% of voters who voted for President did not vote in their State Senate race. For U.S. Senate voters, that number was about 3.9%. In precincts with an above average median age, about 3.5% of voters who voted for President did not vote in their State Senate race. For U.S. Senate voters, that number was about 2.9%.

Like in North Carolina, Arizona had no association between race and roll-off rates: white, nonwhite, and Hispanic voters skipped down-ballot races at similar rates. And, Democratic and Republican-leaning precincts also skipped down-ballot races at similar rates.


There were much starker contrasts of ticket-splitting and skipping down-ballot races in Arizona than compared to North Carolina. This highlights two key takeaways for climate activists interested in mobilizing Arizona for climate action:

  1. It is even more important to look at ticket-splitting and undervoting data in Arizona than in North Carolina to get a fuller picture of the state and local political dynamics across the state.
  2. There is more work to be done at a local level in Arizona to both (1) decrease undervoting rates and (2) vote for progressive climate action not just for federal races, but also down-ballot state elections.

Arizona is a politically dynamic and exciting state for climate action. The electorate is diversifying, becoming more progressive, and is a potential epicenter of solar energy and innovative water preservation policies that can help lead the United States to a better future. We get there by looking carefully at the data we have to out-strategize anti-climate politicians and their special interests to flip the Arizona State Legislature to a pro-climate state government.



Climate Cabinet Action

Moneyball for climate politics. We help downballot candidates run, win, and legislate on the climate crisis. climatecabinet.org