Prediction Levels of Teacher-Made Tests: A Case of Ordinary Level Geography in Highfield, Zimbabwe

The study sought to establish prediction levels of teacher-made tests in final o rdinary l evel examinations as a way of assessing the predictive validity of teacher - made tests. Specifically, the investigation focused on the extent to which teacher-made tests in Geography predict pupil performance in final examinations. Data was collected from six Geography teachers and mark records of sixty pupils selected from two secondary schools using stratified random sampling. This survey research design was used in this study for having a true representative sample. Positive and significant relationship between teacher made test results and final examination results were found in all cases. Significant differences in performance were also observed among the four streamed classes. Among the recommendations was the need to ensure that teachers are well-grounded in measurement theory in order to help them not only set valid tests, but also know how to reduced “ error of measurement ” when scoring tests. K EY WORD : Final e xamination, prediction levels , secondary schools, teacher–made tests, geography, scores, and tests . IKHTISAR : Artikel ini berjudul “ Tingkat an Prediksi Tes Buatan Guru : Sebuah Kasus Tingkat an Biasa pada Bidang Geografi di Wilayah Highfield , Zimbabwe ”. Penelitian ini berusaha untuk menetapkan tingkat an prediksi tes buatan guru dalam ujian akhir t ingkat an b iasa sebagai cara untuk menilai validitas prediktif terhadap tes buatan guru . Secara khusus , penyelidikan difokuskan pada sejauh mana tes buatan guru dalam bidang Geografi dapat memprediksi kinerja murid dalam ujian akhir. Data dikumpulkan dari enam guru Geografi dan catatan buatan enam puluh murid terpilih dari dua sekolah menengah dengan menggunakan sampel acak bertingkat . Desain penelitian survei yang digunakan dalam penelitian ini dimaksudkan untuk memiliki sampel yang benar -benar representatif . Hubungan positif dan signifikan antara hasil tes buat an guru dan hasil ujian akhir ditemukan dalam semua kasus. Perbedaan yang signifikan dalam kinerja juga diamati antara empat kelas secara berurutan . Diantara rekomendasi nya adalah perlu untuk memastikan bahwa guru cukup beralasan untuk melaksanakan teori pengukuran dalam rangka membantu mereka tidak hanya men ata tes yang valid, tetapi juga menge tahu i bagaimana mengurangi “ kesalahan pengukuran ” ketika men skor tes . KATA KUNCI : Ujian akhir, tingkat an prediksi, sekolah menengah, tes buatan guru , pelajaran geografi , skor, dan tes. About the Authors: Mbokochena Edmore , Ph.D. is a Lecturer at the Department of Sociology WUA (Women’s University in Africa), Harare, Zimbabwe; Mushoriwa D. Taruvinga , Ph.D. is a Lecturer at the Department of Educational Foundations and Management UNISWA (University of Swaziland), Swaziland; and Mudzengerere F . Hamilton , Ph.D. is Research Fellow at the Faculty of the Built Environment NUST (National University of Science and Technology), Bulawayo, Zimbabwe. Corresponding author is: eddiembok@gmail.com How to cite this article? Edmore, Mbokochena, Mushoriwa D. Taruvinga & Mudzengerere F. Hamilton. (2013). “Prediction Levels of Teacher-Made Tests: A Case of Ordinary Level Geography in Highfield, Zimbabwe” in ATIKAN: Jurnal Kajian Pendidikan , Vol.3(1) Juni, pp.1-8. Bandung, Indonesia: Minda Masagi Press owned by ASPENSI in Bandung and FKIP UNSUR in Cianjur, West Java, ISSN 2088-1290. Chronicle of the article: Accepted (April 21, 2013); Revised (May 20, 2013); and Published (June 15, 2013).