External links

Complex analysis of educational measurement data to understand cognitive demands of assessment tasks

Standard - GACR GA25-16951S [Registered results] 2025 - 2027

Principal Investigator: doc. RNDr. Patrícia Martinková, Ph.D.

The project aims to improve understanding of the cognitive demands of test items by proposing methods for complex educational measurement data analysis involving text analysis and equating different test forms.

Educational measurements yield complex data that are not sufficiently harnessed. The project uses available data from secondary school leaving exams, admission tests, and other educational assessments to gain a deeper understanding of cognitive demands of assessment tasks.

It uses and develops methods that combine textual data extracted from item wording, numerical data on item difficulty, qualitative data on the test takers’ perceptions, and content experts’ judgments of task difficulty to investigate to what extent is an automated analysis of assessment data able to identify sources of task difficulty related to classroom teaching and learning.

We expect that semantic-level features of item wording will reflect the cognitive demands of the subject matter closer than syntactic or lexical features, thus enabling a novel use of assessment data to improve student learning. We also expect that the features associated with task difficulty are dependent on the test takers’ ability and propose methods to equate item difficulty estimates under covariate and anchor test equating designs.