Sunday, August 3, 2014

ODE Asks: Is Your Kid a Potential Criminal?

From our August newsletter ()

The Salem Statesman Journal s—“the Department of Corrections, Oregon Youth Authority, Department of Human Services, Oregon Health Authority and Oregon Department of Education have signed an agreement to share data among them, something they are not typically allowed to do.”

The idea, they say, is to spot correlations that can identify children at risk of bad outcomes. Data are already being used to predict which convicts may be more prone to recidivism (repeat offenses). Comparing models for school outcomes to the prison system strikes us as Orwellian. Examples of supposedly relevant data include test scores, low birth weight, and who knows, probably in-school disciplinary information if they want to predict criminal behavior.

The article says “nowhere in this project can the state analyze an individual child.” But they want to be able to say, “This child is exhibiting three crucial behaviors that make him likely to drop out of school.” Now wait a minute. Of course they can analyze an individual child. Correlations are calculated on the basis of individuals’ data on different measures. If they couldn’t identify the child, they couldn’t run the correlations. And if they can identify that a child has certain troubling characteristics, that information had to be retrieved from somewhere, probably from that shared database.

These plans should set off lots of alarm bells. Oregon Save Our Schools tried two years ago to promote legislation that would guarantee parents’ rights to know what information is being collected on their children and to challenge incorrect information. ODE scuttled that effort. We also have a real concern about the self-fulfilling prophecy. If they identify a child with certain behaviors and start treating him like a potential problem, the child could be led to conform to those expectations.

It’s like the state has this great new plaything, a huge database that can correlate all kinds of information on everyone, including our children. The capabilities of technology are outpacing thoughtful policy on data use and the experience to use the technology wisely. And remember that correlation is not destiny.

We support the ethical and appropriate use of data, but the big data moneyball approach is untested and fraught with problems.