The Washington PostDemocracy Dies in Darkness

Voting in Florida? Bring a book, it may take a while

Estimated average voter wait time in 2012. Fourteen states were excluded: two because they were vote-by-mail states and a dozen others and D.C. because estimates were imprecise. (Government Accountability Office analysis of 2012 Cooperative Congressional Election Study data)
Placeholder while article actions load

Voters in Florida waited more than a half hour to vote during the 2012 election. The wait in Alaska was a speedy 1.4 minutes.

Those are the findings of new Government Accountability Office report published this week on voter wait times in the last election. Few states had average waits longer than 20 minutes, but in Florida, Maryland and Virginia — the three states with the longest average wait times — about 12 percent of voters waited more than an hour. (Fourteen states were excluded due, mostly, to a lack of precise data.*)

The GAO report looks at more than just voting wait times by state; it’s a nearly 100-page assessment of the who, what, where, why and how of long waits during the 2012 election. Here’s a look at some of its findings.

• How long is too long

While the state data was based on voter-reported wait times*, most of the GAO report is based on its survey of elections officials from hundreds of jurisdictions.

The findings, naturally, all depend on what exactly qualifies as a wait. As far as GAO was concerned, wait times began when a voter entered the line to begin filling out a ballot. Officials in 24 percent of local jurisdictions surveyed felt a 10-minute wait was too long. Thirty percent said 20 minutes was too long and 21 percent said a half hour was too long. (The rest said waits of less than 10 minutes, more than 60 minutes or more than 120 minutes were too long, or did not know or answer.)

• Who faced the longest waits

Of the 338 jurisdictions that answered the GAO’s survey, 18 reported wait times of 60 minutes or more at at least a few polling places. Those jurisdictions with the longest waits had larger populations, lower median ages, and more minorities who speak English as a second language:

  • All but one had populations larger than the median of the jurisdictions surveyed. (Twelve had populations greater than 500,000. Five had populations between 100,000 and 500,000 and one had a population smaller than that.)
  • Fifteen of those jurisdictions had median ages lower than the overall median age of 38.4 years.
  • In 17 jurisdictions, the white population was below the overall median share of 76.4 percent.
  • In 11 of the 18 jurisdictions, non-native English speakers accounted for more than 20 percent of the population. The median for all jurisdictions was a little over 10 percent.

• Waits are not that widespread, election officials say

Most local jurisdictions didn’t collect data because wait times weren’t an issue, GAO found from its survey. Based on that sample, the agency estimates that roughly 4 in 5 jurisdictions nationwide did not collect, get or have wait time data for Election Day 2012. Of those, an estimated 4 in 5 appeared to do so because wait times were not an issue, as far as officials were concerned.

Most officials felt none of their polling places had wait times, as shown in the chart below. Only a small share — an estimated 22 percent — admitted to having a few or more polling places with significant wait times.

• What officials know about voting delays

The GAO also asked officials what wait-time data they collected. From their sample, they estimated that officials in more than 1 in 3 jurisdictions collected data based on observations of voter wait times at polling places. Just under 1 in 3 kept data on how many votes were cast during specific time periods and roughly 1 in 6 collected data on how long polling places stayed open after closing times, when voters checked into polling places using an electronic poll book and how often they complained about waits.

• How waits differed throughout Election Day

Local officials generally reported wait times of 10 minutes or less throughout Election Day, though the share of short waits shrank slightly as Election Day wore on, they said. Meanwhile, the share of voters who had to wait between 11 and 20 minutes grew slightly as the day wore on, according to the after-the-fact reports.

• What made a wait long

GAO asked the officials who reported long waits whether they felt various factors contributed, as depicted in the chart below. Nearly 3 in 4 said long ballots caused long waits. More than 1 in 3 blamed the need to determine voter eligibility (process provisional voters), the training of poll workers and the use of paper poll books. A third said the design and layout of polling places had an effect.

In the end, GAO identified nine factors that affected wait times in 2012, as depicted by the bulleted text in the graphic below. They were:

  • The opportunity to vote before Election Day
  • The type of poll book used
  • Determining voter eligibility
  • What the ballot itself looked like
  • The amount and type of voting equipment
  • The number and layout of polling places
  • The number and training of poll workers
  • Voter education
  • The availability and allocation of resources

* Note regarding state estimates for 2012 voting wait times: GAO left Oregon and Washington out of its state-by-state wait-time analysis because they are vote-by-mail states. Another dozen states and D.C. were omitted due to imprecise data. The missing states are Arkansas, Colorado, Hawaii, Mississippi, Montana, Nevada, New Hampshire, New Mexico, North Dakota, Rhode Island, Utah, and West Virginia.

The estimated wait times were based off an analysis of self-reported data from nearly 55,000 voters collected for the Cooperative Congressional Election Study at Harvard. The GAO borrowed the method for its analysis from Massachusetts Institute of Technology professor Charles Stewart III’s “Waiting to Vote in 2012” journal article.