Analyzing unstructured text data: Using latent categorization to identify intellectual communities in information systems

In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier's archiving and manuscript policies are encouraged to visit: a r t i c l e i n f o a b s t r a c t The Information Systems field is structured by the research topics emphasized by communities of journals. The Latent Categorization Method categorized and automatically named IS research topics in 14,510 abstracts from 65 Information Systems journals. These topics were clustered into seven intellectual communities based on publication patterns. The technique develops categories from the data itself, it is replicable, is relatively insensitive to the size of the text units, and it avoids many of the problems that frequently accompany human categorization. As such LCM provides a new approach to analyzing a wide array of textual data. Since the inception of the field, Information Systems (IS) researchers have questioned the identity of the field and have attempted to map the field's intellectual organization. More recently, researchers have proposed normative frameworks of what constitutes the " core " of IS research [7,50]. However, some researchers are concerned that such a normative perspective might be overly constraining and that the " the vibrancy of the information systems discipline lies in its porous boundaries, or the relatively large grey area that characterizes its domain, and the interests of the information systems research community " [52, p. 1]. This current research agrees with the view presented by DeSanctis [18] that the identity of IS lies in the action of systematic inquiry and the interactions within communities of practice. In this research, we are guided by Benbasat and Zmud's [7] statement; " …the primary way in which a scholarly discipline signals its boundaries — and in doing so, its intellectual core — is through the topics that populate discipline-specific research activities " (p. 184). However in contrast to Benbasat and Zmud's [7] normative approach, we argue that the identification of the field's intellectual core is best achieved through empirical examination of what topics researchers actually investigate and publish in IS journals worldwide. We establish a historical view of the discipline [39] by determining the topics of intellectual interest to IS researchers over a 13 year period. We view IS as a " subject complex " which Apter [3] defines …