Effects of the two methods for oblique transformation of morphological factors implemented in the SPSS program package

Gustav, Bala

Glasnik Antropološkog društva Jugoslavije

35

89

94

0351-1480

https://scindeks.ceon.rs/article.aspx?query=ISSID%26and%263347&page=0&so...

1999-2021/11/03/12:22:29

The aim of this paper is to compare two methods implemented in the SPSS 8.0 program package which are often used to define the ftrst and second-order morphological factors, with the purpose to check the findings about the behaviour of correlation matrix of latent factors (Momirović, 1998a, 1998b). Effects of the methods were demonstrated by the analysis of first and second-order morphological factors in a sample of 217 students of the Faculty o/ Physical Education in Novi Sad. In this sample of subjects, the sample of 12 anthropometrical variables, which are most frequently used at sportsmen, were applied. The analysis wa.s made by direct oblimin (Jenrich and Sampson, 7966) and promax (Hendrickson and White, 1964) methods implemented in the SPSS 8.0 program package. According to the analysis, it could be concluded that the Promax method gave significantly better solution, than the direct oblimin method. These ftndings did not give sufficient proof that the promax was implemented correctly in the SPSS 8.0 program package, so it is necessary to extend the investigation bv other professional statistical programs. Namely, the problem is not in confrdence or unconfidence to the analysed methods, but in their ineorrect implementation in the most popular stutistical program package SPSS, as Momirović proved for direct oblimin method.