In high school, Melissa Edwards woke up at 6:30 a.m. to catch a bus. It was dark, she was tired and the school's 7:15 a.m. morning bell forced the St. Louis teen to eat lunch at 10:45.
Her brother was incredulous. "I thought that this couldn’t possibly be good," said Finley Edwards, a Colby College economist, of his sister's predawn start. But when Edwards sought data on the topic, he couldn’t find any. So he ran the numbers himself.
On Thursday, the Harvard journal Education Next will release Edwards' findings that show that later start times, which usually allow teens more sleep, boost test scores significantly. The Economics of Education Review will publish a longer version of the study.
"Start times really do matter," Edwards said. "We can see clear increases of academic performance from just starting school later."
Edwards found that students who started middle schools an hour later in Wake County, North Carolina, saw their standardized test scores increase by 2.2 percentile points in math, and 1.5 percentile points in reading on average. The impact was greater for older students. Starting school an hour later had further benefits, he found: 12 fewer minutes of television-watching per day; nine more minutes devoted to homework per week; and an average of 1.3 fewer absences than other students.
A long trail of biological research attests to the effects sleep can potentially have on school work. A Brown University study found that due to different circadian rhythms, as children age it becomes harder to fall asleep earlier. Still, most schools in the U.S. start at 8 a.m. Twenty percent of students start at 7:45 or earlier. Teens get fewer than seven hours and 15 minutes of sleep per school night on average.
"Even though it seemed obvious, there just wasn’t any good empirical evidence that made a direct link between start times or sleep," Edwards said.
As standardized test scores become more important to teacher evaluations and school funding in districts nationwide, a few researchers -- including Edwards -- say later start times for adolescents is a relatively simple solution.
"In these tough times, you may have a useful investment like getting rid of tiered bussing [to start school later] and it gets passed over because there's a limited budget," said Jonah Rockoff, a Columbia University economist who also has researched the topic.
According to Edwards' research, the lowest performers saw the greatest boost when school started later.
"It's a targeted method," Edwards said. "The score changes are comparable to other differences we might see: The effects of starting school one hour later are similar to the increasing grades we see when kids' parents have more education." Drawbacks to later starting times, though, include adults having to recalibrate their schedules, and limits to after-school athletic programs.
Rockoff called Edwards' paper "an important piece of the puzzle of the evidence on early start times," though he said its results would have been more compelling if he randomly placed students into trials, rather than basing it on data from the school district that already existed. Edwards then used mathematical formulas to obtain results.
In recent years, there have been several attempts to delay school starting times. In 2005, Rep. Zoe Lofgren (D-Calif.) introduced a congressional resolution that recommended high schools across the country start at 9 a.m. -- but it died in committee.
At an April school board meeting, Ann Arbor's assistant superintendent of instruction raised the possibility of starting high school later. According to Annarbor.com, her proposal was met with doubts about the lack of the tangible benefits of starting school later, concerns about tweaking bus times, and cost.
In Fairfax County, Va., where most schools start at 7:20 a.m., school board member Sandy Evans has tried since 2004 to delay the morning bell. Now, Fairfax is revisiting the topic.