The study of the impact of various question formats and their combination variability on test results

Abstract

The aim of the research was to identify and examine the characteristics of the impact of four question types and their combination in polymorphism tests on the testing output indicators for secondary school students. Students in the 8th and 9th grades (N=164) participated in the study that was done in the field of studying the native (Ukrainian) language. It was possible to demonstrate the existence of the effect of factor "question form" on the test scores (impact power is up to 35%) and to identify the combination of question forms that is characterized by the highest diagnostic accuracy thanks to the tools (criterion-oriented author's tests, questionnaire), and methods of results' analysis (factor analysis, dispersion analysis, etc.) used in the research process. According to research findings, creating a polymorphic linguistic test is advised. If there is an even distribution of these questions throughout the test, the use of multiple-choice questions with one correct answer combined with text and constructed-response questions can guarantee the clarity of test output indicators (with the least impact of question formats on them). The study's drawbacks are identified, and the likelihood that they will be addressed by further research on the most effective test framework for empirical validity is supported.

Keywords

• Educational testing
• test format
• question forms
• question combination variability
• item analysis
• scores