Sticking with What (Barely) Worked: A Test of Outcome Bias

Source page