Problems for the cognitive-depletion model of religious interactions