The syllogistic evaluation task paradigm requires participants to assess whether a conclusion is logically determined by its premises. The inability to ignore beliefs when attempting to comply with this requirement is the phenomenon known as 'belief bias'. This research programme primarily examines this phenomenon and is motivated by conflicting accounts of how it arises. This research also examines the relative believability of the logical quantifiers which are used to construct the syllogistic task. Current theories of belief bias differ in respect of whether the premises or the conclusion are thought to guide the reasoning process. Explicit attempts to focus participants' attention on the conclusion or the premises indicated that either method of reasoning was plausible. Subsequent research demonstrated that these methods of reasoning are fairly paradigm specific. Participants appear to utilise the premises to guide the reasoning process only when there is no conclusion to evaluate (the production task paradigm). The presence of the conclusion tends to evoke conclusion based reasoning and in turn tends to promote a greater reliance on beliefs. In general, across the experiments participants appeared to be fairly competent at evaluating conclusions. However, their responses appeared to be based on whether a conclusion was consistent with the premises as opposed to whether it was logically determined by them. Belief bias actually appeared to have a beneficial effect on logical reasoning as the presence of unbelievable conclusions appeared to motivate participants to search for the logically correct response. The Mental Model theory of reasoning provided a useful framework in which to describe the results. However, additional assumptions were needed to accommodate the notion, that when presented with a conclusion, participants utilise it to guide the construction of a mental representations of the premises.