Actions Speak Louder Than Words: An exploration of game play behavior and results from traditional assessments of individual differences

In this paper, we describe the results of an exploratory pilot study examining the use of a video game as an assessment tool. This is part of a larger project aimed at identifying the components and mechanics of accurate and engaging assessment games. A critical step towards this goal is to understand what player behaviors might naturally be associated with traditionally measured human variables (e.g. cognitive and non-cognitive individual differences). We use the custom game, Food for Thought, and its in-game logging capabilities to examine relationships between different play behaviors, like planning and level retrying, with results from traditionally measured psychosocial, multitasking, and personality variables. In a sample of 30 undergraduate students, we found 10 statistically significant correlations between in-game player behavior and measures from traditional assessments (ACT Engage, SynWin, and the BFI-44). These preliminary results suggest promise in understanding human individual differences through the exploration of player behaviors.