本文主要目的在介紹IRT等化法在題庫建立之應用的基本邏輯和程序。IRT等化法優於傳統等化法，主要原因是只要所搜集的資料符合某種IRT模式，那兩組參數之間即具有直線關係。 一般常用的等化設計有單組設計、等組設計及共同試題設計。IRT參數聯結常用的方法有平和標準差法、強韌平均數和標準差法、反覆強韌平均數和標準差法、特徵曲線法等，其中以特徵曲線法最適用在2和3參數模式。

　　　　　The purpose of this paper is to introduce the basic logic and the procedures of applying IRT equating methods for item banking. In comparison with classical equating methods, IRT methods are preferred, because the linear relationship between tow parameter sets is provided by the theory. When item and ability parameters are unknown, linear adjustments are necessary to related the ability and item parameters across subgroups. The commonly used equating designs. (1) the single group design, (2) the equivalent group design, and (3) the anchor test design are discussed. The common scale for the parameters can be established by the following methods: the mean and sigma method, the robust mean and sigma method, the iterative robust mean and sigma method and the characteristics curve method. The characteristic curve method appears to be the most appropriate method for the two and three parameter models.