A Potential Outcomes Approach To Assess Causality In Information Systems Research

Difficulties in conducting experiments in commercial settings make it necessary to use observational data to study the effect of technological interventions. Since assignment to treatment group cannot be assumed to be random in observational data, the challenge lies in assessing when one can infer causality in such quasi-experimental studies. This chapter reviews recent theoretical and methodological advances in econometrics and statistics to suggest applicability of a potential outcomes approach to assess the causal effect of a particular treatment on outcomes of interest in the information systems literature.