Classification Based on Hierarchical Linear Models: The Need for Incorporation of Social Contexts in Classification Analysis

• Published in: i-manager's journal on educational psychology Vol. 3; no. 1; pp. 34 - 42
• Main Authors: K. Vaughn, Brandon, Wang, Dr. Qiu
• Format: Journal Article
• Language: English
• Published: i-manager Publications 15.07.2009
• Subjects: Classification; Comparative Analysis; Educational Research; Foreign Countries; Hierarchical Linear Modeling; Prediction; Psychological Studies; Regression (Statistics); School Districts; Thailand; Thailand
• ISSN: 0973-8827; 2230-7141
• DOI: 10.26634/jpsy.3.1.183

Many areas in educational and psychological research involve the use of classification statistical analysis. For example, school districts might be interested in attaining variables that provide optimal prediction of school dropouts. In psychology, a researcher might be interested in the classification of a subject into a particular psychological construct. The purpose of this study was to investigate alternative procedures to classification other than the use of discriminant and logistic regression analysis. A classification rule utilizing Hierarchical Linear Modeling (HLM) was derived and examined, with a following example which will show the benefit for using such an approach by comparing the hit rates to those of a logistic regression analysis. Specifically, a real data set on retention of Thailand students (7,516 students in 356 schools) was investigated using a reduced logistic regression, full logistic regression, and multilevel model. The results show that a multilevel approach increases the level of correct classification. Suggestions for practical use are considered.