The Rasch Model for Item Analysis
Generate an AI Snapshot to get a quick, structured summary of this paper.
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
Abstract
The so-called Rasch model, now widely employed for item analysis, is only one of a complete family of models described by Rasch in his 1960 text. All may be properly called "Rasch Models" since they share a common feature which Rasch labeled "specific objectivity". This is a property of most measurement systems which requires that the comparison of any two objects that have been measured shall not depend upon which measuring instrument or instruments were used. It is a familiar feature of many everyday physical measurements (length, time, weight, etc.). In the context of mental testing, it means that the comparison of two individuals who have been tested should be independent of which items were included in the tests. Traditional test analysis based on "true scores" does not have this property since "scores" on one test cannot be directly compared to "scores" on another. The particular virtues of specific objectivity and the conditions needed to achieve it are discussed later in this chapter.
