Item and Person Statistics
Rita Green · Palgrave Macmillan UK eBooks · 2013
The item and person statistics tables which are produced by Winsteps enable test developers to see a number of useful things very quickly about their items and test takers. First, information about the difficulty level (items) and person ability (test takers) measured in logits is supplied. Second, the degree to which the items and persons (test takers) fit the model is provided by means of a statistic called the Infit Mean Square (MNSQ; see below). This tells us how confident we can be in the measures (logits) associated with the persons and the items. Where an item or a person do not fit (where, for example, an item is performing in an unexpected way or where a test taker has produced an unexpected performance) we are able to take this into consideration when making decisions about the test and / or the test taker. These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.