PCA questions

#1
here goes...
1. The Kaiser-Meyer-Olkin Measure of Sampling Adequacy and Bartlett's Test of Sphericity tests are done on the standardised or the original data?
2. Component matrix (as in SPSS) gives the eigenvectors in columns..yes?
3. In a factor analysis, what do factor loadings tell us? Are they equal to eigenvectors?
4. What does the scatter diagram with the final data tell us? I understand that the scatter diagram theoretically means that the original data is now explained in terms of the eigenvectors, but what additional information do we get from it, apart from that obtained by eigenvalues and vectors?
5. The tutorial on PCA by Lindsay I. Smith (2002) lists ‘getting the original data back’ from the principal components as the last step in PCA. Why is this step necessary? Is it o be able to get a new data set that we can now use for multiple regression..or any other statistical analysis we wish to do?
6. Why are the steps following the calculation of eigenvalues and vectors necessary? i.e. getting the new data, scatter diagram, getting the original data back.
7. Will you answer all the question? :D

Till I hear from you.

Regards,
p.