Using Generalizability Theory and Many-Facet Rasch Model to Evaluate In-Basket Tests for Managerial Positions

Purpose: This study aimed to analyze interview data collected from a series of in-basket tests during managerial personnel recruitment in a local Chinese company, taking advantage of the use of combination of Generalizability theory (GT) and Many-facet Rasch Model (MFRM), rather than the Classical Test Theory (CTT). Design/Methodology/Approach: Participants included 78 candidates (M(age) = 37.76, SD = 6.36; 35.9% female) interviewed for eight managerial positions in a local Chinese company. Data were collected based on a series of 10 in-basket interview tests, and a 20-item rating scale (the in-basket test rating scale; IBTRS) was developed and piloted, and five expert raters rated the participants on their performance in five aspects (planning, communication and coordination, capital operations and management, analysis and problem solving, and empowerment and controlling). The data were analyzed using a crossed design of p × i × r, where p represents person, i represents item, and r represents rater. Effects of candidate (person), test item, rater, and the interaction of item and rater were examined. Findings: The use of the combination of GT and MFRM was able to provide accurate, comprehensive information over the process of in-basket interview tests. Specifically, GT analysis showed good generalization coefficient and reliability index (0.893 and 0.871, respectively), and the variation of candidates explained most of the total variance (53.22%). The candidates scored high in the dimension of empowerment and controlling. There were differences in the severity of raters. Three raters should be sufficient to ensure good scoring stability. Originality/Value: This study used the combination of GT and MFRM to assess the interview data instead of using a CTT approach.