The linear logistic model with relaxed assumptions (LLRA) was developed for measuring changes in qualitative data. It assumes item-specific person parameters and thus does not require homogeneous items to be presented to the persons at two points in time. The hybrid variant of this model maintains the multidimensionality of the person parameters, but it allows for different sets of items each of which is presented only once. In the model, a Rasch homogeneous item at t 2 with possibly differing difficulty corresponds to each item at t 1 . A short description of both models is followed by a first application of the hybrid LLRA to empirical data from a study on text comprehension. This example not only serves to demonstrate possible results when applying the LLRA, but is also used to outline the principle of hypothesis testing and model controls.Index terms: dichotomous data, linear logistic model, measuring change, Rasch model, text comprehension.The quantification of change has become a topic of great interest in many fields. Traditionally, change is defined in terms of differences of values measured at different points in time. Applying this concept to the social sciences (see Harris, 1963) would require a natural metric of those persons' scores which are used to infer the changes. However, most of the observations collected in the social sciences are qualitative, and thus lack a natural metric. On the other hand, quantitative statements about change are desired which, by the traditional method, cannot be obtained other than by an artificial quantification of the observations.To overcome these problems, special models can be formulated for measuring change in qualitative data. One of these models, the &dquo;linear logistic model with relaxed assumptions&dquo; (LLRA;Fischer, 1972Fischer, , 1976; for similar models see Breslow, 1976;Duncan, 1985aDuncan, , 1985bKoch, Landis, Freeman, Freeman, & Lehnen, 1977;Marascuilo & Serlin, 1979;Plewis, 1981), is based on Rasch's (1960/1980) concept of specific objectivity which is realized in his well-known one-parameter logistic model for item analysis. In contrast to the latter model, the LLRA does not require homogeneous categorical items, that is, items which measure one and the same latent trait.
The LLRAIn the simplest case, there are two groups of persons-the experimental group and the control group. The reactions of all persons to the same k dichotomous items (e.g., presence vs. absence of a certain APPLIED PSYCHOLOGICAL MEASUREMENT