Contextual Strength and the Subordinate Bias Effect: Comment on Martin, Vu, Kellas, and Metcalf