Generating Research Questions from Digital Trace Data: A Machine-Learning Method for Discovering Patterns in a Dynamic Environment

Digital trace data derived from organizations' information systems represent a wealth of possibilities for analyzing decision-making processes and organizational performance. While data-mining methods have advanced considerably over recent years, organizational process research has rarely analyzed this type of trace data with the objective of better understanding organizations' decision-making processes. However, accurately tracking decision-making actions via digital trace data can produce numerous applications that represent new and unexplored opportunities for IS research. The paper presents a novel method developed to combine quantitative process mining approaches with a variance perspective. Its viability is demonstrated by looking at teams' decision patterns from a dynamic business-simulation game. This exploratory data-driven method represents a promising starting point for translating complex raw process data into interesting research questions connected with dynamic decision-making environments.