1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

SPSS identification restrictions on EFA models

Discussion in 'Education' started by D. Bontempo, Oct 8, 2018.

  1. D. Bontempo

    D. Bontempo Guest

    I am trying to understand the restrictions used by SPSS to identify EFA models using ML extraction. FOr example the GFI for a 2 factor model of 8 items (shown below) has 13df


    I understand to identify EFA models m-squared restrictions are needed, where m=#of factors. I have also seen the formula [(p−m)2−(p+m)]/2 for the df of the chi-sq test.

    This 2-factor model of 8 items needs 2^2=4 restrictions. We get 3 from fixing the factor var-covar matrix to 1 for factors and 0 for factor covarience.

    How is the additional restriction obtained? 13 df mean one loading appears to be fixed to zero, or some restriction has been placed on a function of loadings. But the unrotated solution has 24 loadings reported, none of which are 0.

    Even a text discussing how to restrict this model for estimation would be appreciated.

    Login To add answer/comment

Share This Page