Abstract:

    The advent of the big data era has made it possible to use data to predict behavior. Game is an important part of human social behavior. The issue of how to use the rich data obtained from games to predict players' knowledge, skills, and traits has attracted increasing attention, resulting in game-based psychological assessment. Game-based psychological assessment (GBPA) refers to evaluating a person's ability, personality, and other psychological characteristics through games or gamified activities. In the early days, it was mainly used to evaluate the effectiveness of education and training, and later it was extended to assessing psychological characteristics. As a new technology, game-based assessment creates a more realistic situation to get more accurate data, which has advantages in form, process, and outcome.

    A paradigm based on evidence-centered design has been developed in game-based assessment to design instruments and conduct empirical studies. It consists of three core components: competence model, task model, and evidence model. Establishing a competency model refers to defining the structure of the target characteristics under a theoretical framework. Establishing an evidence model refers to determining the indicators and scoring rules. Specifically, selecting the indicators associated with the competence model, setting the scoring rules, and further establishing the predictive model. Establishing a task model refers to designing tasks to obtain indicators from them. The game is the assessment task itself in a game-based assessment. Researchers can refine predictive indicators based on existing games for evaluation or design new games according to the purpose of the study. Utilizing evidence-centered design to establish assessment tools provides necessary conditions for further data collection, processing, and reliability and validity testing.

This paradigm has been applied to assessing individual differences in cognitive and non-cognitive abilities. For example, predicting reasoning ability and mathematics performance through Sokoban games, evaluating elderly people with mild cognitive impairment and Alzheimer's disease using “Kitchen and cooking” game, and assessing children's social emotional skills using a novel computer game called Zoo U. Non-cognitive ability reflecting social characteristics and personality traits is more difficult to measure than cognitive ability. Researchers have not reached a consensus about whether the game-based assessment is effective in predicting personality traits.

    However, this technique is still in its infancy. Future research can be further expanded in task design, data mining, and practical application. Non-linear game patterns and multi-player large-scale game designs expand the applicable scope. The attempts in corporate recruitment and clinical assessment and treatment also bring new value to this technique. Game-based assessment has great potential in the field of psychometrics.

Key words: game-based assessment, evidence-centered design, cognitive ability, non-cognitive ability