The whole idea behind the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 -- the formal name of the biggest reform of the federal welfare system in decades -- was to push more people off the federal dole and into the work force. By that measure, it undeniably worked: Welfare rolls have declined by about half since 1996, with much of the decline attributable to the policy changes, and employment rates have grown for many of those groups historically well-represented on welfare.

But from the very beginning, some advocates for low-income Americans and for higher education feared that that job gains might come at a cost, particularly in terms of access to a postsecondary education and the financial and other benefits that often accrue from it. Although rules governing the law vary from state to state, all of them -- to varying degrees -- significantly limit the extent to which time spent in a classroom or training count toward "work" requirements.

A new study released by the National Bureau of Economic Research seeks to measure the extent to which those policies have reduced the educational attainment of a key constituent of welfare programs -- low-educated single mothers. The study, conducted by researchers at Bentley University, Rider University, and the University of Medicine and Dentistry of New Jersey, finds "robust and convincing evidence" that federal welfare reform reduced the college attendance of high-school-educated unmarried mothers, aged 24 to 49, by between 20 and 25 percent.

"The more heavily the women are working, and the more women are living in states that require a lot of work and don't count education as work, the less likely they are to be in school," said Hope Corman, a professor of economics at Rider University and one of three authors of the study. "They can't seem to do both, so 'work first' comes at the expense of education."

The researchers focus their study not merely on actual welfare recipients but on what they call "women at risk of being on public assistance," which they define, in the case of college going behavior, as unmarried mothers ages 24-49 with less than a college education. (The reason the scholars examine this group, as opposed to women who are actual recipients of welfare benefits, is because the latter group would ebb and flow based on the policies over time. "There are a lot of people [these policies] might be affecting who might never show up on welfare rolls," said Corman.) The researchers compare the target group to a control group of women who are similarly situated -- in terms of their age, education levels and marital status -- but do not have children and are thus generally ineligible for the welfare rolls.

Examining the representation of the two groups from 1992 to 2005, as states experimented with their approaches to the Aid to Families with Dependent Children (which was eventually replaced by block grants under the Temporary Assistance for Needy Families program), the researchers found that "being in the high risk group and living in a state that had [sought a waiver to adopt work-related requirements under AFDC] made you 2 percentage points less likely to go to college, and about 1.1 percentage points less likely to go to college full time," said Corman.

Because about 9 percent of the total population of high-school-educated single mothers were enrolled in college, the two-percentage-point reduction meant a drop of about 22 percent in the number of such women in college, and a comparable decline in the proportion of such women enrolled in college full time. The drop was steepest, not surprisingly, in states that had less lenient policies on which educational activities counted as "work," and for how long, with dips as large as four percentage points in some states.

The full study is available for purchase from the National Bureau of Economic Research for $5.