Building School Data Systems: The California Way Not to Do It
If it were only that easy.
Nearly two years ago, in A Byte At the Apple Rethinking Education Data for the Post-NCLB Era, I noted the two decades of struggle California had with developing its school data system. In particular, the California Longitudinal Pupil Achievement Data System — which was supposed to combine nearly all of the state’s disparate databases — had a particularly troubled history. It took five years for CALPADS to make it from legislative intent to begin full development in 2008. It took another year for the system to become somewhat operational. Even then, it wouldn’t be the fully comprehensive data system that policymakers, parents and schools needed in order to improve the quality of education for their students.
These days, CALPADS is serving no use for anyone at all. Because it has been shut down. As reported yesterday by John Fensterwald, state Superintendent Jack O’Connell put the system on hiatus after months of glitches — largely caused by state bureaucrats and computer giant IBM (which built out the system on the state’s behalf) — that have made it difficult for school districts to provide and access data. As the state education department’s consultant, Sabot Technologies bluntly points out in its assessment: “the overall [technology] architecture is sound… Instead, Sabot finds that the system implementation includes anomalies, errors and defects throughout.”
Certainly this shutdown will further hinder the delivery of timely data about student progress. But, in all honesty, CALPADS should probably be scrapped altogether. Not because of technical issues, but because the data system is too-narrowly focused on helping the state and school district meet No Child compliance, not on providing useful data. Even if CALPADS was fully operational, schools and researchers still couldn’t track the long-term performance of individual English Language Learner students (or even determine if they are being fully-mainstreamed into regular classes). The lack of a universal identity number for each student means that student progress can’t be tracked once they enter college; it also means that universities can’t easily access high school student data. Even with the state’s decision to finally integrate CALPADS data with that from the state’s teacher data as part of the effort to tie teacher evaluations to student performance, CALPADS problems means this may not happen for at least another year.
The structural problems underlying CALPADS sheds light on an even bigger problem: An byzantine educational governance system — including a state board of education appointed by the governor, a state education department headed up by an elected superintendent and state universities and community colleges led by different boards at nearly every level — that complicates the development of a fully-unified school data system. Thanks to the sparring matches between each of the politicians and bureaucrats (along with the lack of leadership overall by McConnell and Gov. Arnold Schwarzenegger), data system integration is even less likely to happen now. Which means that parents can’t get the data they need to make smart decisions in shaping the educational destinies of their children — and teachers can’t use data smartly in shaping their classroom instruction.
This, by the way, isn’t just a California problem. Although Florida has succeeded in developing a truly longitudinal school data system, other states are plagued by similar versions of California’s unwieldy school governance and paucity of leadership. It will take more than annual surveys by the Data Quality Campaign to shame states into fully addressing those problems. It is another reason why school reformers, grassroots activists (and business groups such as chambers of commerce) must work together to make data quality (and other elements of the reform agenda) a reality.
