Application of Many-facet Rasch Model in In-basket Tests in China

The purpose of this paper is to utilize many-facet Rasch model (MFRM) to examine raters' severity/leniency, internal consistency, dimension difficulty, and examinees' ability level in order to further discuss the biases in the assessment center rating. Research questions are tested on a sample of 138 examinees who participated in the in-basket test and 6 raters who rated the test in 2010. The raters were divided into trained and untrained group to finish the work independently. An MFRM analysis was conducted to test the research questions. MFRM can estimate examinees' ability value independently of all types of variances, thus providing insights into the assessment field. MFRM also provides analysis of rater's effects, assessment dimension difficulty, and bias in assessment center test.