ABSTRACT
It wou ld be a useful starting point to consider that there exists a family of Rasch models for measurement (Andrich, 1988; Masters & Wright, 1984; Fischer & Molenaar, 1995; Rost, 1996). In this chapter the authors introduce the use of the simplest model, the model for analyzing dichotomous data. This was, in fact, the model with which Georg Rasch d id his initial work. Since then, the procedures for performing dichotomous Rasch analysis have been developed further by a number of researchers (foremost among them Ben Wright from Chicago; Wright & Stone, 1979, and G. H . Fischer, 1974; Fischer & Molenaar, 1995), whereas others have extended the basic Rasch model to include analysis of Likert-type rating scales (David Andrich, Perth, Australia, 1978a, 1978b, 1978c; Earling Andersen, a student of Rasch, 1977), responses that could be given partial credit (Geoff Masters, Melbourne, Australia, 1982), and testing situations in which many facets other than just person and item needed to be measured (Mike Linacre, Chicago, 1989). In each case, many researchers justly could claim that they contributed to the development of the Rasch-family models. This is not the place to attempt a full picture of some of the rich history of the Rasch approach. Fischer and Molenaar (1999) provide a useful summary of the contributions of several key European contributors. However, the researchers cited are certainly amongst the most energetic proponents of Rasch measurement models. Each of these models from the Rasch family is more Complex than the preceding one, but has the basic dichotomous model at its core. Therefore, i f the researcher chooses not to use an added feature of any more Complex model, it then collapses to the preceding, simpler model.
