The Possible Use of Mixture Model Analysis Technique in the Educational Sciences: A Case Study

Article excerpt

Abstract

In this article, it is explored that the issue about applicability of the Mixture Model analyses technique which can be used in the searches of education sciences. Therefore in this search, for the semester of 2005-2006, the 643 individuals participated in the Private Skill Examination of Physical and Sport Education Department, Education Faculty of Yüzüncü Yil University (YYU). During interview, information was taken with regard to AOOBP, gender, occupation of father and mother, settlement and high school branch from which they graduated. The search data has been gathered by the structured - interview technique. During the interview, the questions about the personal features of the participants were asked and the answers were recorded. The effect level of the variables have indicated the personal features of the registered participants who had the grade of OSS (Student Selection Examination) examination. These effect levels have been solved by Mixture Model which belongs to Latent Class family. This model is a new statistical method for education sciences in Turkey. At the end of this search, 5 sub-classes have been identified and the variables have influenced each identified class from different aspects and levels. Besides, it was observed that the interaction of the variables has been an affective matter in regard of different sides and level.

Key Word

Mixture Model, Latent Class, Education Science, Homogeneity, Parameter Estimation.

It has been known that environment and genetics have influenced the formation of the individual differences. The various skills of the individual during his/her life-time which have been affected by the environment and genetics formed the individual differences. Among the changeable individual differences, it can be counted the level of the improvement of the mental, language, psycho-social, ethical, preliminary knowledge of the individual (Kuzgun & Deryakulu, 2004).

These differences reflect the academic achievement of the individual. In this respect Baltas (2001) emphasized that the variables such as study room, anxiety of parents and the harmony in family life have influenced the success of the individual. In the mental improvement, there are serious points which determine not only genetical factors, but also environmental factors such as social stimulus (Güngörmüs, 2001). The socio-economical and social cultural levels of the parents achieved the stimulus, family life, peers could have influenced the attitude towards the school and success. It is known that the gender constructs an important variable on success. The other factor that influences the success of an individual is related to the education sciences.

In general, it is noticed that there are many environmental factors besides genetical factors that influence the academic success of individual. Especially, the discipline of physical education and sport require some different features originated from biological and environmental factors. However, to have an education opportunity in these departments, it is an obligation to have enough OSS grade. In addition, universities that have differences in application of the rules apply the Private Skill Examination. Thus, what the main variables that influence OSS grade of the individuals has constituted the focus of this study. For instance, it was neccessary to know how do average high-school success grade and/or gender affect the OSS grade? The other ambiguous point is that how the interaction between the level of this sequenced variables is.

In other words, it is not known whether the sample belongs to one or more than one population or not when sampling is performed for the features of one and more than one changeable features. The Mixture Model both determines unobservable heterogeneity of a sample and makes predictions on parameters for each sub-population by finding out what number achieved for sub-population belonged to these populations. …