Validity and reliability of teacher-made tests: Case study of year 11 physics in Nyahururu District of Kenya

Abstract

This study was carried out to establish the factors influencing the validity and reliability of teacher made tests in Kenya. It was conducted in Nyahururu District of Laikipia County in Kenya. The study involved 42 teachers and 15 key informants selected from teachers holding various positions of academic responsibilities in their schools in Nyahururu District. A mixed descriptive survey research design was applied. Data was collected through questionnaires and interviews with key informants. Quantitative data analysis was applied to survey data collected via questionnaires. The frequency distribution was described while data from interviews were qualitatively analyzed. The findings of the study revealed that the experience of teachers, training on test construction and analysis, level of education, use of Bloom’s taxonomy, moderation of tests and length of tests have an effect on validity and reliability of the tests. Also these factors have a varying influence on the validity and reliability of teacher-made tests. Experienced teachers who had prior training in testing and therefore applied a number of these factors in their test construction tended to design tests with higher validity and reliability than their counterparts without such training. It was concluded that teacher-made tests are generally valid and reliable. The study recommended that teacher training on test construction and analysis needs to be enhanced in order to raise tests validity and reliability.

Keywords: Teacher-made tests, test design, validity, reliability.