A Case Study with Implicit Class Analysis
Makaleler > Tam Makale · 2011
Yazarlar: Duygu Gungor Culha, Mediha KORKMAZ
JOURNAL OF MEASUREMENT AND EVALUATION IN EDUCATION AND PSYCHOLOGY-EPOD, SCI Expanded İndekslerine Giren Dergi, Uluslararası, 2011, cilt 2, ss 191-199
Özet
Latent structures that can not be directly observed have frequently been studied in social sciences. Most popular latent variable model is factor analysis however variables of research may not always be continuous. This situation increased the interest of latent variable models for discrete variables. After expectation-maximization algorithm was formulized and started to use in software programs, there has been a growing body of research with latent class analysis (LCA). Purpose of this study is to present an example of LCA; there are few examples in literature in our country. Participants of the study were 138 psychology students who were taken statistics course first time. Data were collected with a 15 questions (totally agree, mildly agree, don't agree ordinal scaling level) questionnaire and analyzed with latent class analysis. Analyses were performed with Latent GOLD 4.5. 3-class solution was selected as the most informative model based on information criteria. Parameter estimation-latent class probabilities and conditional probabilities-of this model were presented and interpreted.