Anxiety sensitivity (AS; the fear of anxiety-related sensations) has been proposed as a risk factor for the development of panic disorder. The present study involved a conceptual replication of Ehlers' (1993, Behaviour Research and Therapy, 31, 269-278) study on childhood learning experiences and panic attacks, but also extended her work by investigating the relationship between early learning experiences and the development of AS, in a non-clinical sample. A sample of 551 university students participated in a retrospective assessment of their childhood and adolescent instrumental and vicarious learning experiences with respect to somatic symptoms (i.e. anxiety and cold symptoms, respectively) using an expanded version of Ehler's (1993) Learning History Questionnaire. AS levels were assessed using the Anxiety Sensitivity Index, and panic history was obtained using the Panic Attack Questionnaire, Revised. Contrary to hypotheses, the learning experiences of high AS individuals were not found to be specific to anxiety symptoms, but involved parental reinforcement of sick-role behavior related to somatic symptoms in general. High AS subjects reported both more anxiety and cold symptoms prior to age 18 than individuals with lower levels of AS. In addition, both cold and anxiety symptoms elicited more special attention and/or instructions from parents for high AS individuals to take special care of themselves. These findings are contrasted with the results for self-reported panickers who reported more learning experiences (modeling and parental reinforcement) specific to anxiety-related symptoms, than the non-panickers. The results suggest that higher-than-normal levels of AS may arise from learning to catastrophize about the occurrence of bodily symptoms in general, rather than anxiety-related symptoms in particular.