PRINCIPAL COMPONENT ANALYSIS AND A MULTILINGUAL CONSTRUCT TO DETERMINE THE UNDERGRADUATE MAJOR SELECTION FACTORS
No Thumbnail Available
Date
2020
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
SDU Bulletin: Natural and Technical Sciences
Abstract
In this article, we review mathematics behind well-known Principal Component Analysis from Linear Algebra implemented in various applied fields. As an application, we develop a construct to measure factors that affect college students in their major selection. This is a multilingual construct given in three languages, namely Kazakh, Russian, and English. To this end, we prepare a survey consisting of 27 Likert scale items in three languages and it is conducted among 314 undergraduate students in Kazakhstan. For dimensionality reduction, Principal Component Analysis is carried in python programming language which resulted in 9 major scales with only 22 elements. The overall reliability of the test is calculated to be 0,856. The nine scales are the effect of Uniform National Testing, state grant affect, personal interest affect, skills affect, occupation salary affect, teacher affect, external affect, university cost affect, parent’s affect.
Description
Keywords
Principal Component Analysis, Factor Analysis, Varimax rotation
Citation
Assanbayeva G , Kadyrov Sh. / PRINCIPAL COMPONENT ANALYSIS AND A MULTILINGUAL CONSTRUCT TO DETERMINE THE UNDERGRADUATE MAJOR SELECTION FACTORS / SDU Bulletin: Natural and Technical Sciences / 2020