Cognitive procedures for smoking reduction: Symptom attribution versus efficacy attribution