69传媒

Researchers who study children and schooling are often constrained by the data sets they use. School district data, for instance, yield valuable information on students鈥� ages, their achievement history, and their educational placements, but they don鈥檛 tell the whole story.

District data systems typically won鈥檛 reveal, for example, whether children are homeless or living in a foster home. They won鈥檛 disclose if children have been exposed to lead paint or whether their mothers dropped out of school. For those kinds of data, researchers usually have to turn to different databases from different agencies.

Now, researchers from the University of Pennsylvania, working with the city of Philadelphia, have found a way to integrate all of that information into a single archive. Known as the Kids Integrated Data System, or KIDS, the archive is thought to be the first of its kind.

The KIDS archive merges data from seven city agencies and the school district of Philadelphia, providing access to all the data the city collects on its youngest residents. The hope is that the project, besides making researchers鈥� jobs easier, will generate information to help agencies improve services for children.

鈥淭his is about the kids and the people working directly with kids,鈥� said John W. Fantuzzo, an education professor at the university鈥檚 graduate school of education and a project leader. 鈥淲e know university researchers aren鈥檛 all Mother Teresa or Albert Schweitzer. We have to prove we can produce practical knowledge, too.鈥�

The project stems in part from work that Mr. Fantuzzo began in the mid-1990s to gather data from different city agencies on a cohort of 19,000 children entering 1st grade in the city schools.

鈥淚t generated good news and bad news,鈥� Mr. Fantuzzo said of the study. 鈥淭he good news was that people who never really worked together actually got together and learned about different systems, and were able to do it in a successful way.鈥�

The bad news, he said, was that the findings revealed disturbingly high rates of mortality, homelessness, abuse, and other risks among Philadelphia鈥檚 children.

Building on that initial effort, Mr. Fantuzzo and two of his university colleagues鈥擠ennis P. Culhane, a professor of social policy, and Trevor R. Hadley, a professor of psychology in psychiatry鈥攄ecided to build a permanent integrated data system. Work began in 1999 with $800,000 from the Philadelphia-based William Penn Foundation, and proceeded from there in fits and starts.

Low Trust at First

Besides developing new technology, the professors had to persuade skeptical city officials and negotiate a complex web of federal and state laws designed to protect children鈥檚 privacy. 鈥淭here鈥檚 a big risk for the city,鈥� said Ronnie L. Bloom, the foundation鈥檚 program director for children, youth, and families funding. 鈥淎 lot of times cities don鈥檛 want to be examined all that closely.鈥�

Agencies also balked because they felt they had been burned by researchers who did what Mr. Fantuzzo calls 鈥渄rive by鈥� studies.

鈥淩esearchers would come and would want to get data from us,鈥� said Annabella Roig, the deputy director for the city鈥檚 division of social services. 鈥淭hey would find out that the city of Philadelphia is not serving its children in some way, and they would publish it and we would hear about it and get upset. There was low trust.鈥�

The University of Pennsylvania researchers allayed the agencies鈥� concerns by creating a forum for scholars and agency officials to meet quarterly to bounce ideas off one another and approve and vet research projects and findings. While city officials cannot censor findings, they can veto research proposals they feel won鈥檛 prove useful to their agencies.

鈥淩esearchers have to agree to be part of a pretty extensive vetting system because that鈥檚 the only way there鈥檚 going to be change,鈥� Mr. Fantuzzo said.

The arrangement doesn鈥檛 mean, as some critics have suggested, that the researchers aren鈥檛 鈥渉olding city agencies feet to the fire,鈥� Ms. Roig says. 鈥淲e鈥檙e not afraid of bad news. We just don鈥檛 want to hear about it post hoc.鈥�

Privacy a Concern

To satisfy federal privacy-protection laws, three administrators at the university鈥檚 Cartographic Modeling Lab 鈥渟crub鈥� the data from city agencies to remove personally identifiable information. Then they use an algorithm created by the researchers to assign identifier numbers to children so that records can be matched across agencies.

Through the lab, statistics are also merged with geographic data to provide a more complete picture of the social environment in which children are raised, Mr. Culhane said.

The system is already yielding useful findings, according to its developers. Mr. Fantuzzo鈥檚 ongoing work with entering students, for instance, has found that the risks children bring with them to school鈥攕uch as poverty, having a poorly educated mother, lead exposure, or inadequate prenatal care鈥攁re all linked to academic and behavioral problems later on. More important, the study also found that attending a formal, center-based child-care program can 鈥減rotect鈥� children from those adverse outcomes.

The findings, which Mr. Fantuzzo has presented to principals and early-childhood educators across the district, helped the city obtain funds to add another 1,000 seats to its district-operated Head Start programs this fall.

Mr. Hadley has found that parents with mental-health problems who enroll in treatment programs were less likely to lose custody of their children than similar groups of nonparticipating adults. Now, he鈥檚 studying autistic children progressing through the system of public services. 鈥淲e鈥檙e finding out what they鈥檙e really getting and from whom,鈥� he said, 鈥渁nd that鈥檚 something that we鈥檝e never been able to do before.鈥�