Probing Poverty

Researchers try meta-analysis to find what works, what doesn’t.

In the search for solutions to poverty, there’s more to do than pinpoint programs that work.

Katherine Magnuson, an associate professor at the UW’s Institute for Research on Poverty, is not afraid to tell policymakers to take funding away or to require improvements from programs that clearly are not helping poor children and families as advertised.

“To me, every dollar that’s spent on a program that’s not effective is a dollar not being spent on a program that is effective,” she says. “From a research standpoint, it’s just as important to know what doesn’t work as what does work.”

There’s plenty of proof that early intervention is needed: one study found that three-year-olds from low-income homes have half the vocabulary of their more affluent peers. But Magnuson devotes her efforts to finding evidence of what actually makes a difference for those children and their families, a common theme in the UW institute’s work.

With colleagues from the University of California-Irvine and Harvard, Magnuson is conducting a large meta-analysis — essentially a study of studies — to identify common factors in programs that are effective at helping at-risk children. Their research is tackling a number of assumptions that seem logical, but have not been proven with data, including parental involvement as a necessary ingredient for effective early-childhood programs.

“We’re hoping to untangle whether it does make a difference and in what ways it makes a difference,” she says.

Published in the Fall 2010 issue