In government, let hard data judge what works, says Wolfers in NYT Upshot column
He presents findings from the White House Social and Behavioral Sciences Team’s latest report, which piloted and evaluated a number of small tweaks aimed to improve federal policies and programs.
For example, the team sent a payment reminder email to more than 100,000 federal student loan borrowers, which led to a 29.6 percent increase in the fraction of borrowers making a payment within a week.
“Good design doesn't cost much – it’s no more expensive to send an effective email than an ineffective one – and so it is worth experimenting to figure out what works,” says Wolfers.
In another case, researchers sent SMS messages to randomly selected high school graduates to remind them of the next steps to enroll in college. Sixty-eight percent of those students enrolled in college, compared with 65 percent among those who didn’t get any reminders; the effect was largest among low-income and first-generation students. The total cost? About $7 per student.
“It’s not about knowing how to do better, it’s about testing what works," says Wolfers. "Experiment relentlessly, keep what works, and discard what doesn’t. Following this recipe may yield a government that's just like Google: clear, user-friendly and unflinchingly effective.”
Justin Wolfers is a professor of public policy at the Ford School and a professor of economics in the Department of Economics. Wolfers' research interests include labor economics, macroeconomics, political economy, economics of the family, social policy, law and economics, and behavioral economics.