Component Loadings of PCA in Factor Module Are Incorrect
Component Loadings of PCA in Factor Module Are Incorrect
I found that the "Component Loadings" of the Principal Component Analysis (PCA) in the Factor Module are incorrect compared to the output of LISREL. In fact, the listed "Component Loadings" of PCA in jamovi are numerically identical to the "Correlations between Variables and Principal Components" in the output of LISREL.
Re: Component Loadings of PCA in Factor Module Are Incorrect
hi,
the pca in jamovi is based on the psych R package, so if you wanted to understand the difference, it might be worth starting there -- but i wouldn't assume the results are incorrect just because they are different.
kind regards
jonathon
the pca in jamovi is based on the psych R package, so if you wanted to understand the difference, it might be worth starting there -- but i wouldn't assume the results are incorrect just because they are different.
kind regards
jonathon
Re: Component Loadings of PCA in Factor Module Are Incorrect
Thank you for the information of the 'psych' package! ── The pca() of the 'jmv' package (i.e., jamovi) is based on the 'psych' package of R. The "Component Loadings" of the Principal Component Analysis (PCA) in the Factor Module are incorrect if one is looking for the standard principal component loadings. The "Component Loadings" listed in jamovi or the output of the principal() function of the 'psych' package are the eigenvectors multiplied by the square root of the eigenvalues, as the "pseudo-factor loadings". I suggest listing both the standard principal component loadings and the pseudo-factor loadings like LISREL and providing more detailed HELP information as in the 'psych' package of R.
Re: Component Loadings of PCA in Factor Module Are Incorrect
let me pass this over to ravi.
cheers
cheers