Research On Multidimensional Irt Equating Problem

Test equating is an important study in education and Psychology; it is very important for fairness and comparability of examination, item banking, teaching quality assessing and computerized adaptive test. However, a great number of educational assessments usually measure more than one trait, in order to adjust scores on these different test forms, MIRT linking procedures should be developed. So far, some researchers have already extended UIRT linking methods to the multidimensional structure. Moreover, the equating solutions which are usually used have shortcomings, for example, lower precise and the extremely long time spent in performing the procedure. In order to solve these problems, this paper bring about a new equating method-LC equating method.In the paper, the MIRT models, IRT equating, MIRT equating are firstly and briefly introduced. Then the basic ideas of the MCMC algorithm and DSY algorithm and the specific procedure to estimate parameters in IRT are also described in detail. Following, by means of monte carlo method to produce simulated data, then investigate the recovery of equating constant by using HB equating method, SL equating method and LC equating method, it shows that the new method performs better. Finally, the advantages of the new method have been pointed out and the approaching research direction is suggested.