This is how science progresses: One researcher comes up with a hypothesis, which others question and test. But shaken-baby cases are haunted by the enormous repercussions of getting it wrong — the conviction of innocent adults, on the one hand, and on the other, the danger to children of missing serious abuse. In one study, researchers looked into the deaths of five children who had head injuries that initially were misjudged to be accidents and found that four of them could have been prevented if an earlier pattern of abuse had been detected.

By Emily Bazelon, New York Times Magazine