An easy way to boost midterm-year turnout in Utah

We could expect a permanent 6.7% boost in Utah’s midterm-year turnout levels if we permanently moved gubernatorial elections to midterm years.

Utah’s lieutenant governor declared that 2010 had the best turnout for a midterm election in Utah since 1994. Although I have criticized that claim, I agree that 2010’s turnout was better than we might have expected if Utah’s turnout had continued the downward trend discussed in this post. Something happened in 2010 that helped midterm-year turnout. What was it?

It was the presence of a gubernatorial race on the ballot.

Usually, Utah elects its governors in presidential years, one of only 9 states to do so. Most states–34 of them–elect their governors in midterm years.¹ In 2010, Utah held a special gubernatorial election because Jon Huntsman resigned early in his second term. And it appears that this special election pushed up turnout in Utah.

I gathered turnout data (via Michael McDonald) for every state from 1980 through 2008. (Official 2010 data aren’t available from every state yet.) Then, I ran a bit of analysis to see how these timing decisions influence turnout.² Here’s what I found.

In a midterm election year without a governor’s race or senate election at the top of the ticket, expect turnout of 36.2%. (That’s averaging across 50 states.)

Add a senate race: Turnout rises by 4.9 percent.
Add a governor’s race: Turnout rises by 6.7 percent.
Add both races: Turnout rises by 8.2 percent.

In a presidential election year without a governor’s race or senate election, expect turnout to be 21.9 percentage points higher (i.e. 58.1%) than in a midterm year.

Add a senate race: No change.
Add a governor’s race: No change.
Add both races: No change.³

In other words, we could expect a permanent 6.7% boost in Utah’s midterm-year turnout levels if we permanently moved gubernatorial elections to midterm years. This change would not hurt presidential-year turnout at all.

Footnotes

Note 1: Two states hold their elections every two years (New Hampshire and Vermont), so I suppose that 11 states hold presidential-year elections and 36 states hold midterm-year elections if we count them. The remaining states hold odd-year gubernatorial elections.
Note 2: Statistical details for nerds: I used a random-effects generalized least squares model. These findings are mostly driven by a “within” effect (over time within each state) rather than a “between” effect, supporting the interpretation I give here.
Note 3: More accurately, adding these different races does not make a statistically significant (i.e. outside the margin of error) change. Technically, adding a senate race is predicted to boost turnout by 0.3%, adding a governor’s race drops it by 0.8%, and adding both drops it by 0.4%, but these are all statistically indistinguishable from zero.

An easy way to boost midterm-year turnout in Utah

Possibly related posts:

Possibly related posts (automatically generated)

About Adam Brown

Who are we?

Hear about new posts

Archives

Meta