Validity Threats in Quantitative Data Collection With Games: A Narrative Survey

Abstract

Background. Games are increasingly used to collect scientific data. Some suggest that game features like high cognitive load may limit the inferences we can draw from such data, yet no systematic overview exists of potential validity threats of game-based methods. Aim. We present a narrative survey of documented and potential threats to validity in using games for quantitative data collection. Method. We combined an unsystematic bottom-up literature review with a systematic top-down application of standard validity threat typologies to games to arrive at a systematisation of game-characteristic validity threats. Results. We identify three game characteristics that potentially impact validity: Games are complex systems, impeding the predictable control and isolation of treatments. They are rich in unwanted variance and diversity. And their social framing can differ from and interact with the framing of research studies or non-game situations they are supposed to represent. The diversity of gamers and their differences to general populations bring further complications. Discussion and Conclusions. The wealth of potential validity threats in game-based research is met by a dearth of systematic methodological studies, leading us to outline several future research directions.