TEST VALIDITY USING MULTIPLE FACET RASCH MODELS ON LEADERLESS GROUP DISCUSSION FOR EMPLOYEE PERFORMANCE EVALUATION

Abstract

The Multi-Side Rasch Model (MFRM) of Item Response Theory (IRT) is applied to performance appraisal. The purpose of this study is to utilize the Many Facet Rasch (MFRM) model in evaluating the validity of the LGD simulation. In this study, ratings are influenced by various factors among which the jury plays the most important role. There are 3 validators judment who assess the LGD simulation draft with 6 criteria such as Situation, mandatory, Actor, Role, Task, systematics and writing format, and the instruction. Data analysis uses the MFRM model, which can estimate the test taker's ability score independently of all variances, thus providing insight into the assessment area. This study mainly examines issues that include content simulation, judges, rating scales and deviations from assessments in the Leaderless Group Discussion (LGD) in personnel assessment to improve the effectiveness and stability of the assessment. This study adopted the Minifacets software, the MFRM computer statistics program to define many aspects of content simulation, assessment dimensions to analyze content simulation, rater severity, interrater reliability, dimensional difficulty and rating scale. The results showed that this subjectively scored exhibited robust unidimensionality. Furthermore, discrepancies in ratings between expert raters were also identified and modeled by the many-facet Rasch model. The implications for applying the many-facet Rasch model in validation LGD simualation at the tertiary level were discussed.

Keywords: leaderless group discussion, many-facet Rasch model,  personnel assessment.