E-mail communication is often abused for conducting social engineering attacks including spamming, phishing, identity theft and for distributing malware. This is largely attributed to the problem of anonymity inherent in the standard electronic mail protocol. In the literature, authorship attribution is studied as a text categorization problem where the writing styles of individuals are modeled based on their previously written sample documents. The developed model is employed to identify the most plausible writer of the text. Unfortunately, most existing studies focus solely on improving predictive accuracy and not on the inherent value of the evidence collected. In this study, we propose a customized associative classification technique, a popular data mining method, to address the authorship attribution problem. Our approach models the unique writing style features of a person, measures the associativity of these features and produces an intuitive classifier. The results obtained by conducting experiments on a real dataset reveal that the presented method is very effective.