Cross Country Analysis of the Mathematical Achievement Scores with TIMSS Data


5th International Conference on Data Science and Applications (ICONDATA’22), Fethiye, Turkey, 07 September 2022

  • Publication Type: Conference Paper / Summary Text
  • City: Fethiye
  • Country: Turkey
  • Middle East Technical University Affiliated: Yes


Providing qualified education and raising well-educated students are the core responsibilities to achieve success in various areas such as art, science, and culture for countries. The significant paths to do this are tracking education systems periodically, specifying educational quality level, and enhancing it by considering various techniques such as other countries’ educational policies. One of the preferable and straightforward ways to see educational improvements is to use students’ exam scores. At this point, for similar purposes, the International Mathematics and Science Trends Survey (TIMSS) is conducted every four years for 70 countries. The survey regularly makes exams for fourth and eighth-grade students and collects their mathematics and science achievement scores. The scores gathered in the TIMSS database are utilized to determine the similarities and differences, educational levels of the countries, and the most valuable factors that affect the students' achievement scores. In this study, we aim to draw attention to critical and essential points of educational achievement factors with the last updated TIMSS data for three countries from different success rankings. By doing so, Linear Mixed Model is taken into account to consider the variation between- and also within- observations due to the dependent structure of the nested dataset. Distinct analysis ideas and paths are run for the sample of mathematics scores of the fourth-grade students living in a developed country England and developing countries, Turkey and South Africa. The affecting factors of students' mathematics achievement are determined as gender, born status, an emotional factor, mathematical tendency factor, SES score, country, and the interactions between country and covariates (except country and SES score).