Examination of Teacher Candidates’ Intercultural Sensitivity Levels by CART Analysis

The aim of this study is to examine the intercultural sensitivity levels of teacher candidates using CART analysis and to develop a predictive model using machine learning algorithms. Additionally, this study provides a framework for understanding the relationship between internet usage and intercultural sensitivity. The participants comprised 416 teacher candidates enrolled in the education faculty of a state university in Southeast Anatolia, Turkey, during the fall semester of the 2022–2023 academic year. The study was conducted on a voluntary basis. A relational screening model was employed to assess the intercultural sensitivity levels. Subsequently, a two-step cluster analysis integrated with CART classification was performed, identifying 9 nodes that explain the intercultural sensitivity levels of participants. Among the teacher candidates, 43.56% displayed a “high” level of intercultural sensitivity, 40.26% showed a “medium” level, and 16.17% exhibited a “low” level. The variable “class” emerged as the primary determinant for the cluster of students reporting daily internet usage times of “0–1 hr” or “1–3 hr.” “Gender” was identified as the most influential variable in explaining the cluster of students whose class was “2,” and daily internet usage time fell within the range of “0–1 hr” or “1–3 hr.” To predict intercultural sensitivity, five machine learning algorithms were utilized, with the Naive Bayes algorithm achieving the highest accuracy at 69.0%. Based on these findings, the study recommends implementing effective teacher training programs aligned with the observed data patterns.