关键词: 招聘选拔, 工作绩效, 工作记忆, 抑制控制, 基于游戏的测评

Abstract: In personnel selection, the phenomenon in which candidates perform well during the selection process but fail to match expectations in subsequent job performance suggests that selection assessments may not fully capture the role of different cognitive abilities across career stages. Against this backdrop, by introducing a "stage-matching" perspective to address this performance-job mismatch, we conducted two field studies using game-based assessments adapted from classic psychological experimental paradigms, aiming to identify the differential contributions of specific cognitive abilities. Study 1 included 789 graduate applicants in a campus recruitment program at a Fortune Global 500 company. These participants completed a suite of Game-Based Assessments (GBA) specifically designed to measure specific cognitive abilities. The GBA battery comprised four tasks: the Corsi Block Task for working memory, the Mental Rotation Test for spatial ability, the Flanker Task for inhibitory control, and a Maze Task for planning and flexibility. Prior to the main analysis, psychometric evaluations were conducted, confirming the test-retest reliability, structural validity, and criterion-related validity of these game-based measures. Hiring outcomes were drawn from official organizational records. Study 2 tracked 446 newly hired employees from four different companies who had previously completed the GBA during their recruitment and successfully passed the standard interview process. These participants were hired for diverse roles, including management trainees, marketing, and technical positions. Due to the heterogeneity of performance appraisal systems across different organizations and roles, performance metrics were categorized into "high performance" and "low performance" groups based on internal organizational standards to ensure comparability. Their job performance was evaluated by supervisors three to six months after onboarding. This design enabled a robust comparison of the predictive validity of specific cognitive abilities between recruitment and workplace contexts.In Study 1, the group of selected candidates exhibited significantly superior performance on most cognitive indicators compared to non-selected candidates, indicating a broad cognitive advantage among hired individuals. Since the selection stage represents a "maximal performance" context characterized by time pressure, high stakes, and novel information, specific cognitive resources are taxed differently. Logistic regression and XGBoost-SHAP analyses further confirmed working memory accuracy as the strongest predictor of selection outcomes (B = 0.97, OR = 2.64, p < 0.001; SHAP = 0.67), demonstrating that rapid information updating is the most critical factor for success in competitive recruitment settings. In Study 2, high-performance employees similarly outperformed the low-performance group on the majority of cognitive tasks. In contrast to the selection phase, the workplace represents a "typical performance" context characterized by persistence, multi-tasking, and frequent distractions. Regression analyses identified inhibitory control accuracy as the strongest predictor of job performance (B = 0.58, OR = 1.78, p < 0.001; SHAP = 0.32), while other cognitive measures contributed smaller, auxiliary effects, reflecting the essential role of resisting interference and maintaining goal-directed focus during routine professional tasks. The findings of this study not only extend the ability–performance compatibility principle and refine the theory of stage-specific job performance, but also provide empirical support for the validity of cognitive ability assessments in predicting real working performance. These results offer organizations actionable psychological insights for talent selection and management, highlighting how targeted assessment of specific cognitive abilities can inform more effective human resource practices across different career stages.

Key words: recruitment, job performance, working memory, inhibitory control, game?based assessment