No document available.
Abstract :
[en] Assessments of intelligence by means of paper-pencil tests faced several critiques that point to their lack of face validity, insufficient coverage of the definition of intelligence, their sensitivity to the emotional state of the test taker, and the danger of getting outdated. The present paper discusses to what extent these limitations can be overcome by computer-based problem solving scenarios–so-called microworlds. Generally speaking, microworlds are supposed to be highly accepted by test takers, to provide process measures by directly tracing problem solving behavior, and to realize game-like characteristics that may increase test motivation and reduce test anxiety. To capitalize on these potential advantages, we developed the microworld Genetics Lab that was completed by a large, heterogeneous sample of more than 600 Luxembourgish students. Performance scores were derived for students’ problem solving strategies as well as their mental problem representations–important cognitive data which are not accessible with typical paper-pencil tests. Analyses of the psychometric characteristics of the Genetics Lab empirically underscored the construct validity for the derived performance scores. For example, process oriented measures of strategy use were found to possess discriminant validity with respect to grades. Further, acceptance and induced test anxiety of the Genetics lab was explored relative to a paper-pencil measure of intelligence. Our results show that the Genetics Lab is a reliable and valid assessment instrument and emphasize the benefits of using microworlds for assessing intelligence.