Using a Table of Specifications to improve teacher-constructed traditional tests: an experimental design

Abstract

We investigated if instruction on a Table of Specifications (TOS) would influence the quality of classroom test construction. Results should prove informative for educational researchers, teacher educators, and practising teachers interested in evidenced-based strategies that may improve assessment-related practices. Fifty-three college undergraduates were randomly assigned to an experimental (exposed to the TOS strategy) and a comparison condition (no specific strategy support) and given materials for an instructional unit to use to construct a classroom test. Results of a multivariate analysis of covariance suggested that students exposed to the TOS strategy constructed a test with higher test content evidence but not response process evidence scores. Furthermore, we found that treatment participants were able to accurately complete the TOS tool and choose items that reflected the subject matter specified in the TOS tool. However, they experienced difficulty selecting items at the cognitive level specified in the TOS tool.

Keywords:

• validity
• summative assessments
• teacher-made tests