Most previous studies relating autism and humor have compared responses to jokes or cartoons in autistic children and controls. The present study used the Humor Appreciation Measure (HAM, Rawlings 2008), which measures responses to hypothetical real life situations, as well as jokes; and employed the Autism-spectrum Quotient (AQ, Baron-Cohen et al. 1995), a self-report instrument which assumes that autism exists on a continuum. In an undergraduate sample (N = 126), substantial correlations were reported between the total AQ score and the "unpleasant/aversive" ratings of potentially humorous, non-violent situations involving other people and the self. Correlational and regression analysis indicated that the strongest relationships with humor variables involved the Attention Switching sub-scale. The overall importance of Attention Switching was further examined by correlating individual sub-scale items with relevant humor variables. It is argued that the correlation between humor and Attention Switching is dependent on participants' relative tendency to avoid situations associated with the novel or unpredictable.