FAKTÖR ANALİZİNDE BELİRLİ KOŞULLAR ALTINDA FAKTÖR SAYISI BELİRLEME YÖNTEMLERİNİN KARŞILAŞTIRILMASI
xmlui.mirage2.itemSummaryView.MetaDataShow full item record
Factor analysis is one of the multivariate statistical techniques used in the field of health, sport, education, and social sciences about especially scale development studies. In this technique, determining of retained the number of factors is more important than the estimation of factor loadings and rotation methods. There are some methods for determining number of factors such as Kaiser criteria (K1), cumulative explained variance and scree graph are known utmost, but extracted factor numbers in these methods aren’t accurate in every situation. In except of these methods, there are Horn’s parallel analysis and Velicer’s minimum average partial correlation (MAP) used in recent years. In this study, we compare the performance of K1, PA and MAP in determining of the number of factors for continuous variables and Likert scale in conditions that the number of variables (6, 12, 18, 24, 30), sample size (30, 50, 100, 200, 300, 500), mean loading (0,10-0,90) and the population number of factors (1, 2, 3 ve 4). In conclusion, MAP and PA methods are more accurate, K1 used in most statistical software has poor performance in terms of the number of factors except for the number of the variable is six. Also, K1 is most affected by the change of sample size according to the others. Objective and modern methods as MAP ve PA are recommended instead of subjective methods, such as scree graph, Jolliffe criteria, cumulative explained variance, and K1.