The Effect of Health Status, Nutrition, and Some Other Factors on Low School Performance Using Induction Technique

Mevlüt Türe ¹, Zekeriya Aktürk ², imran Kurt ³, Nezih Dağdeviren ⁴

¹Department of Biostatistics, Medical Faculty of Trakya Üniversity, Edirne
²Department of Family Medicine, Faculty of Trakya University, Edirne
³Department of Biostatistic, Medicine Faculty of Trakya University, Edirne
⁴Departments of Family Medicine, Medical Faculty of Trakya University, Edirne

Pages : 28-38

Abstract

Objectives: We investigated the effect of some hypothetical factors (nutrition, health indicators, risk behaviors, personal characteristics, and family indicators) on academic achievement using Logistic Regression (LR) and the Chi-squared Automatic Interaction Detection (CHAID) method.

Study Design: Participants were 873 secondary school or high school students selected randomly from a total of 12,150 students after stratification according to school populations in Edirne, in 2003.

Results: The sensitivity, positive predictivity, and specificity rates were 61.19%, 67.70%, and 74.25% for CHAID, and 50.00%, 64.29%, and 75.69% for LR, respectively. Father's educational level was the most important factor in the CHAID method. Smoking status, time reserved for homework, and nutrition were the other important factors predicting low school performance according to the CHAID method.

Conclusion: The classification tree algorithm can be used in risk analysis and target segmentation for academic achievement management. Our results may contribute to developing guidelines for those involved in secondary school and high school education.