Illusory correlation refers to the perceived (but erroneous) relation between stimuli. In social psychology this phenomenon has been related to stereotype formation. Recent research in this area has shown that the perception of illusory correlation between two groups and their behaviours can be a product of understandable attempts to differentiate between these groups. This is due to participants' interpretations of the experimental task and to features that draw attention to group differences. In the first study we showed that the task instructions presented to participants can induce expectations of differences between the stimulus groups, which helped to produce illusory correlation. We also demonstrated that giving precedence to the behaviours, rather than to the groups, resulted in attenuated illusory correlation effects. In the second study we investigated how aspects of the stimulus distributions affected the perception of illusory correlation. In line with the first study, we showed that increasing the focus on the stimulus groups can enhance illusory correlation, whereas focusing on the behaviours can reduce this affect. The present findings support the self-categorization explanation of illusory corrlation that proposes it to be the product of a meaning ful category differentiation process.