Controlling the judge variable in grading essay-type items: an application of Rasch analyses to the recruitment exam for Korean public school teachers.

The purpose of this paper is to show how the Rasch measurement model can be used to control the effects of judge variable on the grading of essay-type items in the recruitment test for Korean teachers. Special attention is given to two aspects of judges' involvement in the grading. One is to identify a way to minimize the variation of grading due to judge severity. The other concern is to figure out a way to reduce the number of judges without threatening objectivity of ability estimates. Results from the FACETS analyses tell us not only how much grading standards vary among judges and how to adjust them but also it produces comparably reliable ability estimates with fewer judges.