Application of the Split-Half Reliability (KR-20) Test in Validation of Multiple Choice-Type Testing Instrument to Assess Student Performance in Competence Based Education (CBE) System in Kenya

Reliability is defined as the consistency of results from an administered test. Theoretically each statistical test contains some inherent error which is the portion of the score on the test that is not exactly relevant to the construct that is intended to be measured. Sit-in objective type questions have been embraced at basic and middle levels of education in Kenya for a long time. However universities have only recently embraced objective type questions by use of online platforms due to large number of students and the availability of reliable online platforms. Even when this is a positive development to address the immediate challenges facing universities the reliability of such instruments is paramount. The purpose of evaluation is to assess the essential life skills which include communication, problem solving and adaptability. In this paper the validation of objective type questions testing instrument at basic and tertiary level is validated by use of Split-Half Reliability Testing Instrument. The Split-half reliability determines how much error in a test score is due to poor test construction. The split-half reliability was administered to Grade 7 students at Ekato Junior Secondary School in Western Kenya in 9 subjects under the Competency Based Education (CBE). The computations were done by using Excel Spreadsheet to determine the reliability index through use the Kuder-Richardson formula 20 (KR-20).