STEVE INSKEEP, HOST:
Here's a reality of public policy. You know, government programs that affect the economy or public safety or individual's lives. The reality is when we debate public policy; we often have no idea what we're talking about. Even members of Congress may sometimes have no idea because they can't. We do not analyze public policy with anywhere close to the scientific rigor with which we analyze the efficacy of drugs or the safety of cars. We don't study it in that way. NPR is taking a regular look at issues related to poverty. And today, NPR's Shankar Vedantam is here to tell us about new research into a poverty program that has long been the subject of lots of debate. Hi, Shankar.
SHANKAR VEDANTAM, BYLINE: Good morning, Steve.
INSKEEP: What's the program?
VEDANTAM: Well, the program is welfare. And it's been at the center of this intense political crossfire. Liberals have often argued that welfare produces positive effects; conservatives often say it's useless or even counterproductive. Now, there's lots of evidence you can marshal to support either position, but a lot of the evidence lacks the kind of rigor that even marketers now bring to consumer products. And the most important lack in the evidence is it lacks what scientists would call a control group. When we say that welfare works or welfare doesn't work, we have not randomly assigned families who receive welfare and compared the outcomes to families who've been randomly assigned not to receive welfare. I was speaking with economist Anna Aizer. She's at Brown University, and she explained to me that having a control group is really important.
ANNA AIZER: That's really key to understanding the impact of cash welfare - is knowing what would've happened had these children not received it. And if you don't know that, you really can't say anything about the impact of cash welfare on child outcomes.
INSKEEP: I guess that's why we don't have any idea what we're talking about often in public policy, you can't do a control group. You can't say I'm going to give some people a benefit and withhold it from others - that doesn't even seem ethical.
VEDANTAM: Yeah, there are ethical concerns, Steve. But also, it's important to remember that not studying policy rigorously also has ethical implications because if you're not spending your money wisely, that's going to have consequences and ethical consequences. Anna Aizer recently realized there was a way to study the long-term effects of welfare because it turns out there was a time people were selected and rejected for the program. It wasn't scientific randomization, but it came very close.
INSKEEP: And it was a long time ago, so we have a chance to find out what happened over time?
VEDANTAM: That's exactly right. When welfare started originally a century ago, there was a program known as the Mothers Pension Program. It ran between 1911 and 1935. It was administered by the states, not the federal government. And when the Great Depression struck, a lot of money dried up, and Aizer found many states couldn't cover all the moms who needed help. So they picked some and rejected others. The moms were mostly white. The family's who did get the money and didn't get the money were nearly identical. After tracking who got the money and who didn't, Aizer and her colleagues tracked down all the kids of all these moms, in 14 states, and followed what happened to them decades and decades down the road. And what she found was that kids whose families had received the welfare lived one year longer, on average, than identical kids whose families did not receive welfare. And there were also clues in the data as to why this happened.
AIZER: Compared to the children who are rejected, those whose mothers were accepted by the Mothers Pension Program stay in school almost a half a year longer and earn about 15 percent more in young adulthood, suggesting that perhaps the mortality effects are operating through increased schooling and income in adulthood.
INSKEEP: OK, so the finding was welfare helped families, which might sound obvious to some of our listeners but just is not to others. This is a subject of debate. But are these big differences that she's describing?
VEDANTAM: I asked Aizer that question and she asked me to remember that the metric she was studying was how long people actually live - it's the ultimate scientific metric. And by that regard, the findings were actually really unusual. Here she is again.
AIZER: You have to remember that, on average, moms received this pension for about three years and the pension was relatively modest. So we're talking a program that provided a pretty modest amount of money for a pretty short amount of time. And to have large, long-term effects like that is pretty remarkable.
VEDANTAM: I think it's worth mentioning, Steve, that this study doesn't necessarily settle all of the modern debates that we have about welfare. But, to me, the value here is in applying scientific tools to understanding public policy in order to make it smarter. It really makes no sense that marketers selling toys have better data on what works and what doesn't than policy makers who are spending billions and billions of dollars.
INSKEEP: Shankar, thanks.
VEDANTAM: Thank you, Steve.
INSKEEP: That's NPR's Shankar Vedantam, who regularly joins us to talk about social science research. You can find him on Twitter @HiddenBrain. Find this program @MorningEdition and @NPRinskeep. Transcript provided by NPR, Copyright NPR.