Studies of deductive reasoning have revealed a number of systematic errors made by reasoners. Some of these systematic errors, belief bias and content bias arc the focus of this paper. Two hundred and sixty two university students participated in the study. The results indicate that when the premises of the most simple forms of syllogisms include simple mathematical content (c.g.. numbers), the acceptance of the conclusion as logically following from the premises drops from 9(>-95% when the conclusion is believable, to 25-30% when the conclusion is unbelievable. However, there is no comparable effects when the premises are concerned with non-math content. Thus the paper provides evidence that belief bias differs across contents of reasoning.