The research community would lose its access to sensitive information from California's state-run programs under proposed legislation announced this week, a reaction to the penetration earlier this year of a university system housing personal data on over 1m participants in a state program.
But researchers warn the proposal could cripple a broad range of important research in economics and social sciences.
The bill, proposed by state senator Debra Bowen, would prohibit state agencies from giving researchers any personally identifiable information on Californians, including names, Social Security numbers, dates of birth, and addresses, said a spokesperson for the lawmaker, who plans to introduce the legislation Monday.
"Social Security numbers are the key to identity theft, and state agencies ought to be guarding them like the Hope diamond, not handing them out like holiday stocking stuffers," Bowen said in a statement announcing the proposal. "Sometimes I really wonder why common sense seems to be in such short supply."
The proposal was prompted by an incident on August 1st in which a computer intruder took advantage of a known vulnerability to crack a system at the University of California, Berkeley being used by Candace Howes, a visiting economics professor from Connecticut College conducting research into the effectiveness of California's In-Home Supportive Services program. The program pays a modest hourly wage to workers who provide in-home care for hundred of thousands of low-income elderly, blind and disabled people.
The compromised system housed a database with the names, addresses, phone numbers, Social Security numbers and dates of birth of everyone who provided or received care under the program since 2001 - some 1.4m people, according to state officials and Howes. It's not known whether the intruder accessed the data, but in October, California's Health and Human Services Agency issued a public warning on the breach, and urged potential victims to take precautions against identity theft.
The state had shared the data with Howes under a California law that allows agencies to provide sensitive information to academic researchers under a confidentiality agreement. If the Bowen bill passes, such data would have to be shorn of all personally identifiable information first.
Professors: Important Research at Risk/
Howes doesn't downplay the intrusion. "It's really a tragedy that a very large number of mostly poor people, many of them disabled, have to go through the anxiety of worrying about this," she says. But she argues that researchers still need access to information like Social Security numbers because it serves as a unique identifier, without which the data could not be meaningfully analyzed.
Currently, researchers are expected to swap sensitive information for "pseudo identifiers" before working with the data -- the August hack attack came while Howes was performing that process, she says. But state agencies don't have the resources to do that sanitizing themselves, so the Bowen bill would effectively cut the research community off entirely, she argues.
"I think it would shut down a lot of really important research that's being done trying to make people's lives better, including research on poverty, income distribution, employment, health services, and anything in social services, including welfare and Medicare," says Howes.
Howes says that in August she lacked the expertise to know that her system might be vulnerable to intruders, and she now supports more rigorous security standards for researchers and universities working with sensitive data, including air gaps for computers housing information like Social Security numbers. "But I don't think precluding access to this kind of data is really the answer to the problem," says Howes.
Some corners of academia have already developed stringent methodologies for protecting personal identifiers, says political science professor Henry Brady, a director at the U.C. Data Archive and Technical Assistance center, which works with government census data. Brady advocates the creation of centralized "data enclaves" at every university, where sensitive information is pooled and guarded by those with information security expertise. "We know how to make these data safe," he says. "We know how to do things so there's essentially zero chance of anybody getting that kind of data."
"Often you need to know personal identifiers to do research of fundamental importance," says Brady. "If this bill passes we won't be able to do evaluations of all kinds of government reforms and their consequences."
Senator Bowen was not available for comment. A spokesperson said that her office did not extensively investigate the effect that the proposal would have on researchers, but that the issue might be taken up during the legislative process.
Carlos Ramos, assistant secretary of California's Health and Human Services Agency, said he couldn't comment on the proposed law until it's submitted to the state legislature. Meanwhile, the FBI's investigation into the computer intrusion that sparked the controversy is ongoing. "They've not been able to share with us any results at this point," says Ramos.