Abstract:

The objective of this study is to compare the methods of the factor extract used in the analysis of the detection factor in terms of the variance described and the number of factor extracted according to the conditions of the various sample size and common variance level. This work is a simulation (symulation) work. In this context, the data to be used in the study, using the SAS program, were produced in the form of correlation matrix with 1000 replications according to the sample size (100, 500), the common variance levels (high (0.6-0.8), wide (0.2-0.8), low (0.2-0.4)), the number of fixed factors (3), the number of fixed variables (20). As comparisons result, the common variance increases the sample size for the data produced as high or low, the described variance amount also increases. However, the common variation has seen that the total variation described has decreased as the sample size for the widely produced data group increases. In all the cases analyzed the total variance was observed as the highest-explaining method's basic component analysis and the lowest-explaining method's image factor method. According to comparisons made in terms of factor loads, the basic component analysis made the highest factor loads for each situation. On the other hand, the image factor method has the lowest factor load. There is no difference in the number of factors between the methods for data in which a common variation is produced in a broad or high interval. For groups where the common variance is low and the sample size is small, the image factor method has not been able to extract the desired number of factors. Therefore, it is not recommended to use the image factor method for groups where the number of samples and the common variance are low.

Keywords: