Washington's welfare recipients are spending less time on assistance, working more and earning more than they were in the first years of welfare reform, according to a study being released today that tracks 3,300 families on welfare.

Yet the same families still struggle to meet their basic needs: More than one-third reported eating less than they should because they didn't have enough money, and even more said they sometimes couldn't pay for housing or utilities.

"It's a mixed bag — are people better off than they were before welfare reform? Yes, probably. Are they well off? Maybe not," said Marieka Klawitter, a University of Washington associate professor of public affairs who is overseeing the state's most comprehensive look at welfare-reform efforts.

Advocates for the poor have a different take on the findings: Getting a job doesn't necessarily mean families are better off than when they were only getting a monthly welfare check. They also fault the way the study was conducted, saying it skews the overall picture.

Researchers are tracking thousands of Washington families from the welfare rolls to evaluate WorkFirst, Washington's answer to the national mandate to reform welfare. The 4-year-old program requires welfare recipients to work and puts a five-year lifetime limit on aid.

The five-year study is a collaborative effort of the UW, Washington State University and four state agencies that oversee WorkFirst.

The data being released today is based on state records and phone interviews with welfare families. The data analyzes their health, employment, food and housing needs and time on welfare.

Researchers are following 3,000 families selected randomly in March 1999; they conducted interviews with nearly 2,000 of them in early 2001, but the other 1,000 couldn't be found. In 2000, 1,334 people were added to the study to help evaluate how characteristics of families on welfare change over time.

The interviews were conducted more than a year ago and don't capture what's happened to low-income families since Washington's recession began and its unemployment rate rose to one of the highest in the nation. Researchers did study more recent state records as part of the study.

But those involved with the study say it still offers valuable insight into what has been called the government's greatest social experiment in decades.

About 56,000 Washington residents are on welfare today, 42 percent fewer than five years ago.

About 60 percent of the 1999 group and half of the 2000 group were employed in early 2001, figures that had risen steadily over time, according to the study.

Median wages increased over time to more than $8 an hour for the 1999 group and $7.25 an hour for the 2000 group. (By comparison, the median hourly wage was $11.63 for all working women between the ages of 20 and 40, according to the 2000 Washington Population Survey. Most welfare clients are women.)

As of February — three years after the study began — more than three-quarters of the 1999 group had left welfare. The 2000 group appears to be leaving welfare at faster rates.

"Those trends of more employment and fewer families on welfare slowed but didn't stop this fall," Klawitter said. "That was a big surprise — I worried it would reverse (because of the recession), but we didn't see that."

The 2000 group was working an average of 30 hours a week, but the 1999 group's work week grew to 38 hours.

"WorkFirst works in the sense that it seems to connect more clients than ever before with the labor market and brings more income into those families. But there are lots of problems that remain," said Greg Weeks, the study director, who works for the state Employment Security Department.

Among other key findings:

• About one-third of welfare recipients surveyed used a food bank or soup kitchen in the past year. About 20 percent said their children had eaten less than they should have for lack of money.

• About 42 percent of the 2000 group couldn't pay rent, mortgage or utility bills at some point in the past year, and a quarter had lost phone service. Seven percent had lost their housing.

The figures were somewhat lower for the 1999 group. Researchers speculate the 2000 group may be worse off because the financial problems that put them on welfare are more recent.

• About one-third of both groups rated their health as fair or poor, but they reported 90 percent of their preschool- and school-aged children were in good health.

• More than 90 percent of the children received grades of C or better in the past year.

• About 90 percent of children and 80 percent of adults have health insurance, most often Medicaid. About 40 percent of the 1999 group who were working had health insurance, sick leave and vacation pay on their most recent job. Fewer in the 2000 group had those benefits.

• Almost half of those surveyed reported working evenings or weekends, and about a third said their hours changed weekly.

Advocates for the poor say the study's contradictory findings don't surprise them.

"This promise of a better job and a career is not being fulfilled," said Jean Colman of the Welfare Rights Organizing Coalition.

At $8 an hour and 35 hours a week, a worker makes $1,100 a month — just above the federal poverty line for a family of two.

"It's not a self-sufficiency wage, it's not a living wage," Colman said.

Even if families are bringing in more money by working than from welfare checks, in some ways they may be worse off, said Tony Lee of the Fremont Public Association, a social-service organization.

Their new income may disqualify them from food stamps and other subsidies but may not be enough to cover basic living expenses, he said.

Lee also argues the study itself is flawed. Conducting the survey primarily by phone means the sample is skewed, he said. "Automatically, you're getting a group that's better off, that speaks English."

And he believes the results are further distorted because the state doesn't know what happened to one-third of the 3,000 participants from 1999. The missing people are likely to be worse off, perhaps homeless or unable to afford a phone, he said.

Weeks said researchers went to great lengths to find all the original participants, but "with increased employment, people have the necessity and the means to move around more."

Jolayne Houtz: 206-464-3122 or jhoutz@seattletimes.com.