Throughout its history, the information systems (IS) discipline has engaged in extensive self-examination, particularly with regard to its apparent diversity. Our overall objective in this study is to better understand the diversity in IS research, and the extent to which diversity is universal across journals that publish IS research. We developed a classification system that comprises five key characteristics of diversity (reference discipline, level of analysis, topic, research approach, and research method) based on a review of prior literature. We then examined articles over a five-year period, from 1995 to 1999, in five journals acknowledged as the top journals of the field, at least in North America. Analyses reveal considerable diversity in each of the key characteristics. Perhaps not surprisingly, the research approach used is more focused with most studies being conducted using hypothetico-deductive approaches, whereas reference discipline is perhaps the most diverse of the characteristics examined. An interesting finding is that IS itself emerged as a key reference discipline in the late 1990s. The 'Journal of Management Information Systems' and 'Information Systems Research' publish articles displaying the greatest diversity, and 'MIS Quarterly' and 'Decision Sciences' publish articles that focus on subsets of the field. Our research provides a foundation for addressing the direction that diversity in the IS discipline takes over time. In the shorter term, researchers can use our classification system as a guide to writing abstracts and selecting key words, and the findings of our journal analyses to determine the best outlet for their type of research.