State databases designed to measure contract spending, parolee histories, Medi-Cal reimbursements and nearly 30 other pieces of information are unreliable or incomplete, according to a study by State Auditor Elaine Howle.
The audit [PDF] was released last month but has recently become more relevant to the discussion of government transparency, as open government advocates have launched into a vocal debate on whether transparency is truly valuable if the public data being released is unreliable.
In its study of 84 state data sets, the state auditor's office found errors in 31 of them. Thirty-four data sets were found reliable, and the rest could not be assessed due to incomplete information.
The report specifically points out that errors in a state database that tracks inspections of radioactive storage facilities could result in too many or too few inspections being performed. It also points out errors in the state's sex offender tracking system, including outdated and incorrect information on registered sex offenders.
The issue of data accuracy is hardly new. The Sunlight Foundation, which advocates for transparency primarily at the federal level, recently revealed gaping holes in the primary public database of federal contracts, finding that $1.3 trillion in grant spending in 2009 alone was somehow missing or incomplete.
O'Reilly Radar did a pretty good job of summing up the current state of the discussion a couple weeks ago:
"(Sunlight Foundation Executive Director Ellen) Miller's strong speech prodded open source advocate Gunnar Hellekson to respond that government doesn't look good naked, which is to say that transparency through open government is a messy, iterative process that inevitably reveals some ugly truth in the process. If people look back at how far open government has come, as Derek Willis wrote, the perspective shifts. In responding to Tom Lee's post on open government carrots and sticks, Hellekson borrowed from the open source world to describe Sunlight's role as a welcome patch.
"Data accuracy is a big concern for reporters, who, not coincidentally, are also typically advocates of open government policies. These days, most of our biggest investigations are based at least in part on computer databases. That's one of the reasons why we at California Watch have our very own data analyst on staff.
What's funny is, the way we're taught to analyze government data, we often go into the process assuming it's wrong. Only after we do our own audit checks and take great pains to assure ourselves that the data is accurate do we start to flesh out publication-worthy conclusions.
