Last month, the Oregon Supreme Court concluded that a ballot measure was so confusing, the attorney general would have to rewrite it.
Opponents of a proposed health-care tax plan argued that the measure's obtuse language and run-on sentences would mislead voters, and the court largely agreed.
Such legal disputes over ballot language are common. This year alone, lawsuits have been filed over more than a dozen measures, says Josh Altic, who oversees the Ballot Measures Project at Ballotpedia, a nonpartisan website that provides explanations and background information about ballot measures.
“One of the reasons we exist is because people have a hard time understanding what these ballot measures do,” Altic says.
This year, Ballotpedia tried to assess the complexity of ballot measures by using readability tests that estimate the level of education needed to understand a document. Of the 27 measures on ballots in nine states this year, the average question would require a graduate-level education to read and understand, the website found.
Election officials and voting rights experts say complex ballot measures are a problem because the average American adult reads at about a seventh- or eighth-grade level. What's more, political scientists from Georgia State University (GSU) analyzed more than 1,200 ballot measures from 1997 to 2007 and found that voters were more likely to skip complex questions.
Like Ballotpedia, the GSU researchers used the Flesch-Kincaid readability test, which measures difficulty by taking into account the number of words in a sentence and the number of syllables in a word. They similarly concluded that in order to understand the average ballot measure, a voter would need roughly a year more than a college education.
Some election officials are trying to address the complexity of ballot language on their own.
When Rhode Island Secretary of State Nellie Gorbea assumed office in 2015, one of her first priorities was to revamp a voter information handbook. She says her state's handbook used to describe each referendum in "legalese" and didn't explain what approving and rejecting each measure would mean.
Now, the handbook explains the ramifications of each option and uses simpler language, active voice, personal pronouns and shorter sentences to describe each measure. It also still includes the legally required ballot language that voters will see on election day.
The changes might best be illustrated by an example. In 2016, Question 1 asked whether voters wanted to approve or reject a proposed state-operated casino in the town of Tiverton. The official ballot language asked:
Shall an act be approved which would authorize a facility owned by Twin River-Tiverton, LLC, located in the Town of Tiverton at the intersection of William S. Canning Boulevard and Stafford Road, to be licensed as a pari-mutuel facility and offer state-operated video-lottery games and state-operated casino gaming, such as table games?
The handbook translated that to this:
Your vote to “Approve” this question means you want to allow a new state-operated casino, including video-lottery games and table games, to be built in Tiverton, at the intersection of William S. Canning Boulevard and Stafford Road.
Some states, such as Alabama and Pennsylvania, have laws requiring plain-language explanations of ballot measures.
In 2014, Alabama created a Fair Ballot Commission, staffed with state government officials, private citizens and lawyers. Two months before an election, the commission has to publish ballot statements online and in printed materials available at county offices. The descriptions must be in "plain, nontechnical language and in a clear and coherent manner using words with common and every day meaning that are understandable to the average reader."
Alabama state Rep. Steve McMillan says he sponsored the legislation to create the commission after hearing from frustrated constituents who couldn't decipher the questions.
"There’s a general feeling amongst the public that the way the constitutional amendments are worded," he says, "it’s very difficult to understand what they’re trying to accomplish."
Altic, with Ballotpedia, notes that some measures need more explanation than others, and it's not always a matter of using smaller words and shorter sentences. Ballot questions about familiar topics, such as whether to legalize marijuana, are less likely to trip up voters than an obscure and complicated bond proposal. The least intuitive ones, he says, tend to be veto referendums, where the ballot asks voters whether they want to approve or reject the repeal of state legislation.
"It’s never clear," he says. "Whenever they are on the ballot, it’s a huge struggle for us."
Efforts like the ones in Alabama and Rhode Island can help -- but only so much. It's unclear how many people actually read their voter handbook before heading to the polls. When their grade depends on it, many students don't study for a test until the night before. When people aren't being graded on their vote, will they study at all?