Damned if you do . . . .

In response to my post about how a social worker whose job is to discern the wishes of people who die alone probably killed the creatures the deceased cared about most (her dogs), I received an email pointing me to another horror story, this one perpetrated by the Department of Children's and Family Services (or whatever DCFS stands for) in Chicago against the parents of a baby with unexplained broken ribs. It makes riveting reading on the Lag Liv blog, authored by a University of Chicago law student and mother. (Go to the October Archives and start reading at Oct. 5. The saga continues for months.) To make a long story short, this is pretty clearly a case of bureaucrats run amok. A doctor and various social workers conclude that the baby's injuries are the result of abuse at his parents' hands---based only on the injuries---and then stick to that conclusion even as the evidence mounts that the injuries arose during delivery. But read it for yourself for the full horror show.

The story, in my view, undercuts a familiar canard. Children's protective services, the canard goes, are stuck between a rock and a hard place: If they intervene when they shouldn't, they cause heartbreak to parents and children alike (and take a child from a safe environment to one that may actually be dangerous, foster care). Yet if the agency waits too long, it makes itself complicit in actual cases of abuse. Damned if they do; damned if they don't.

The reason this picture is a canard is that most of the truly horrific cases---including cases of overzealousness and gross indifference---are simply obvious. The Chicago law student/mother is not objecting to a protocol that, while generally effective at protecting children, is irreducibly overinclusive. She objects, quite rightly, to a defective protocol. Yes, in theory there are circumstances where decreasing type 1 errors leads to more type 2 errors and vice-versa. But in the real world, organizations that tend to make type 1 errors also tend to make type 2 errors. To say in response to people like the Chicago law student/mom, "the type 1 error in your case is regrettable but the justified cost of avoiding horrific type 2 errors" is often just an excuse for failing to take steps to try to reduce both kinds of error.

Posted by Mike Dorf