DEVELOPMENT AND VALIDATION OF INSTRUMENT FOR ASSESSING STATISTICAL LITERACY IN OF POST GRADUATE STUDENTS

Abstract

This study aimed at the development and validation of statistical literacy scale for postgraduate (PG) students in faculties of education of public universities in South East, Nigeria. One hundred and twenty (120) items were generated, guided by a table of specifications. These constituted the first draft of the instrument. This draft was given to two experts in Educational Measurement and Evaluation and an expert in statistics, all from Nnamdi Azikiwe University Akwa. They were requested to critically examine the items of the draft instrument in terms of coverage and quality of items. After this initial validation, 110 items survived. These items made up the Second draft of the instrument. Six research questions were raised to guide the study. Instrumentation design was adopted. The study was carried out in South East Nigeria, with the population made up of all 2969 postgraduate students offered admission in 2020/2021 session in the public universities in South East, Nigeria. A sample of 1,000 students, was used for the study, made up of 400 that were used for trial- testing and 600 that were used for ascertainment of reliability and validity. Each of the two samples was obtained, using multi-stage sampling technique. After the initial validation, the 110 items that survived were administered on 400 students. Scores obtained from the students were subjected to item analysis. Consequently, the difficulty, discrimination and distracter indices for each item were computed. Forty items, which were found to be acceptable in terms of each of the three indices constituted the final instrument. This final instrument was then administered on the second sample of 600 postgraduate students. scores obtained from this second testing were used to ascertain evidences of reliability and validity for the instrument. The result showed that a reliability coefficient of 0.87 was obtained when Kuder –Richardson 20 method was used while a coefficient of 0.81 was obtained when split-half method was used. Scores obtained with the instrument converged with those obtained with Statistical Reasoning Test (SRT) and diverged with those obtained using Test Anxiety Questionnaire {TAQ}. Based on the above, it was concluded that the new instrument (Postgraduate Students’ Statistical Literacy Scale) was reliable and valid.