Investigation of Variables Explaining Mathematics Literacy in PISA 2018 Turkey and China Samples Through C4.5 Decision Tree Algorithm

Investigation of Variables Explaining Mathematics Literacy in PISA 2018 Turkey and China Samples Through C4.5 Decision Tree Algorithm

Zeynep Begümhan ÖZCAN & Sevda ÇETİN

ABSTRACT
The aim of this study is to identify the factors that most affect students' mathematical literacy in the PISA 2018 Turkey and China samples using the C4.5 algorithm. The analyses were conducted on the datasets of 5959 students from the Turkey sample and 12058 students from the China sample. The independent variables of the study are 22 student questionnaire items that are considered depending on the constructs of self-perception, adaptability, school belonging, family support and school environment. After organising the data set, the consistency ratio for the dependent variable, mathematical literacy, was calculated for both samples to evaluate the performance of the C4.5 classification. The decision tree method was then applied and reported separately for each sample. As a result of the analyses, the items that had the most impact on mathematics performance in the Turkey PISA 2018 application were determined as "I am proud that I have achieved something", "My parents support my educational efforts and achievements", and "I feel that I can manage many things at the same time" under the variables of self-perception and family support. In the Chinese sample, the items that most affected mathematics performance were “My parents support my educational efforts and achievements” and “Students seem to compete with each other” under the variables of family support and school environment.

KEYWORDS: Data Mining, C4.5 Algorithm, PISA 1018, Mathematical Literacy

DOI : [PDF]