In Quest for IS Research Relevance: Using Interrupted Time Series Methods in IS Post-Intervention Research

Abstract

The lack of demonstrated intervention effectiveness has often prevented Information Systems research from achieving practical relevance. The IS field tends to mainly provide IS practitioners with a range of constructs, reference and contextualized theories, or else IT artifact-enriched models; often forgetting that its core mission is to point towards the best approaches and practices that practitioners shall implement. This paper attempts to reignite the intervention analysis stream of IS research in the form of interrupted time series (ITS) research designs, arguing that this would strengthen the connection between IS research and practice. ITS research designs in the specific IS context are introduced, followed by a presentation of four important ITS analytical techniques namely difference-in-differences (DID), interventional autoregressive integrated moving average (I-ARIMA), segmented regression, and Bayesian Structural Time-Series (BSTS) analysis. An illustration then follows, aiming at demonstrating how ITS designs can be implemented. Finally, we provide guidance to IS researchers to help them identify and frame potential IS intervention research projects. Elements of discussion about the advantages of ITS designs when conducting IS intervention research are then discussed, and we elaborate on the contributions of this research for IS scholarship.