Modeling the Development of Student Assessment


  • John O. Anderson


The study reported in this article was an attempt to investigate how teachers evaluate the achievement of their students. The study was based on a dataset created from portfolios containing achievement products such as written assignments and tests and background information for a simulated student in a grade 8 language arts class. The contents of the portfolio were controlled in terms of achievement level of products and the background of the student. As part of an undergraduate teacher education course in classroom assessment, 147 teacher candidates graded the components of an assigned portfolio over a 10-week period and reported a final grade. These scores and grades were the basis for an investigation of the structure underlying the evaluation of achievement by these teacher candidates. The model developed fitted the data well and had resonance with commonsense views on the kinds of factors that could be expected to affect the decisions teachers make about marking student assignments and tests.






Special Section