Özet:

One of the most important decisions in the opening factor analysis is to determine the number of factors. Although many methods have been developed for this, parallel analysis is still among the most recommended and used methods. Its frequent use has led to many modifications and the concentration of research related to the analysis. This study was aimed at studying the impact of the methods of factor-extraction used in the analysis of detection factors on the results of the parallel analysis and the Monte Carlo simulation study was carried out. In the simulation work performed with two categories of data sets, the average factor load, the number of substances, the measurement model, the sample size and the conditions of the correlation matrix used were manipulated. The smallest remains applied in the parallel analysis, the basic components, the basic axes, the most probability, the smallest squares unweighted, the smallest squares and the smallest rank factor extraction methods applied in the optimal parallel analysis are compared. The study found that the smallest rank method applied by the optimal parallel analysis method performed with the tetrakoric correlation matrix gave the best results. This is the average factor load. In the 70 conditions, all methods show sufficient performance in the analyses performed with the Pearson correlation matrix while the use of the tetrakoric correlation matrix can be said that other methods except the smallest rank methods applied by the basic components applied by parallel analysis and the optimal parallel analysis extract the excess factor. Research findings suggest optimum parallel analysis with the tetrakoric (polykoric) correlation matrix with the smallest rank method.

Anahtar Kelimeler: