Computer-based assessment of reading skills

In this paper, we present a recently developed and computer-based instrument for the detailed assessment of reading skills. The theory underlying its construction is van Dijk and Kintsch's (1983) strategy model of text comprehension. The target group of the instrument are adults with a presumably high level of reading ability, for instance university students. Therefore, we do not intend to assess difficulties in reading or achievement in learning to read. Apart from that, the subtests refer to basic cognitive processes of reading but not to metacognitive strategies or standards of comprehension. The instrument is designed for research purposes; we are planning to use it for the measurement of covariates in experiments on text comprehension (see Christmann, Groeben, Flender, Naumann & Richter, 1999). We begin with a short discussion of some advantages and disadvantages of presently available methods for the assessment of reading skills. Next, we will argue that it is reasonable to ground the diagnostic efforts on a hierarchically structured model of component processes of reading like van Dijk and Kintsch's strategy model. Following a brief sketch of the strategy model, we will explain the structure of the instrument we have recently developed and illustrate its subtests by item examples. Finally, we shall report some empirical results concerning the psychometric properties of the instrument, that is its reliability and validity.