Lawyers are expensive, and for millions of low-income people across the United States, nonprofits like Greater Boston Legal Services offer the best—perhaps the only—chance at professional help. Staffed by civic-minded attorneys and paid for with public money and private donations, these organizations represent our society’s primary mechanism for making sure that when it comes to civil proceedings, all people, including the very poor, are treated equally before the law.

For decades, it has been an article of faith among those who have devoted their lives to the cause of legal aid that if only the system had more funding, it could do more good and help more people. But lately, a difficult new question is being asked of the legal services community: What evidence do they have that the help they’re offering even makes a difference—and that they’re allocating their scarce resources as effectively as possible?

That is the challenge being laid down by a group of critics led by James Greiner, a professor at Harvard Law School. Greiner believes passionately in providing free legal assistance to the poor, but he is dismayed by what he sees as a lack of data on how it’s delivered and how it affects people’s lives. For all the good they think they’re doing, Greiner argues, the fact is that legal services providers are working off untested assumptions and operating largely in the dark.

Given the noble intentions and tireless dedication of the individuals who staff legal clinics around the country, it feels almost indecent to question the effectiveness of the help they’re providing. But from where Greiner sits, the sanctity of these efforts—and the fact that the system provides for just two attorneys per 10,000 low-income people—is precisely why it matters. In particular, as he sees it, we know very little about how legal service providers ought to determine which clients to take and which to turn down; in many cases, he argues, they may be wasting precious time and money on cases where they’re unlikely to have an impact.

“Most folks have concluded that we’re never going to be able to give a full attorney-client relationship to every person who has a legal problem,” Greiner said last week. “The funding is just never going to be there. So you have to take steps short of that, in an attempt to meet the need.”

For Greiner, the solution is one that has already revolutionized other fields, most notably medicine. He is part of a budding movement that wants to introduce randomized experiments and put hard numbers to what have traditionally been treated as unquantifiable social problems. In medicine, he and his allies point out, such tests have revealed that once-common treatments—hormone replacement therapy, for instance—were not only ineffective but harmful. “The mission,” Greiner said, “is to make law more evidence-based, more rational and scientific.”

Among those who labor in the legal services world, this insistence on randomized testing has been met with skepticism, even defensiveness—particularly after Greiner’s first study on the topic found that one group of people who were offered representation by law students working at the Harvard Legal Aid Bureau were no more likely to prevail in their court cases than people who were not.

Those results came as a jarring surprise to some on the bureau’s staff, and the study sparked consternation among legal aid workers all over the country. Some worried that Greiner’s findings would be used as justification to defund legal aid, while others raised an ethical objection to the very idea of collecting the kind of data he’d like to see: It amounts to experimenting with vulnerable people’s lives.

***

Randomized trials are the gold standard in the world of science, and for good reason: Done right, they can generate highly precise insights about how the world works. To determine the effectiveness of a new drug, researchers give it to some patients and not others; if the people in the first group fare significantly better than the control group, that sends a clear signal about the efficacy of the medicine.