When media coverage doesn't include the numerical magnitude of a scientific study's effect, the risk of readers developing biases increases significantly, according to a new University of Michigan study.