Computer-Aided Text Analysis of Corporate Disclosures: Demonstration and Evaluation of Two Approaches

Abstract

The volume of corporate disclosure is constantly growing and increasing attention is paid to the systematic exploration of its highly informative textual content. Manual analyses, however, are quickly reaching their capacity limits when exploring large collections of texts. Computer-aided text analyses are therefore becoming increasingly important in order to overcome the information overload. In accounting research, however, the corresponding possibilities and limitations of such computer-based analyses are hardly discussed. This paper addresses this knowledge gap and pursues the goal of demonstrating the use of computer-aided text analysis approaches and providing concrete recommendations of “dos” and “don'ts” for their application. Within the framework of a case study, two text analysis strategies – dictionary and statistical approach – are practically applied, documented and subsequently discussed. In conclusion, computer-based processes have proven to be an efficient means for coping with large text collections. Furthermore, the combined use of both text analysis approaches has proven advantageous since they complement each other and compensate for each other's weaknesses. The combination of quantitative results related to thematic categories (dictionary approach) as well as the exploration of new content patterns (statistical approach) provides a more comprehensive picture with regard to the presentation of corporate disclosure.