2

I have performed a PCA with some data sets. Prior to performing PCA, I tested the applicability of PCA with KMO and Bartlett's Sphericity tests. I got some critics stating that I have to add "my decision level of factor loadings" instead of KMO or BS results!

Can anyone tell me what "decision level of factor loadings" means?

amoeba
  • 93,463
  • 28
  • 275
  • 317
igo
  • 21
  • 2
  • 1
    `I tested the applicability of PCA with KMO` To note about KMO: [PCA itself needs not](http://stats.stackexchange.com/q/48497/3277) "good" KMO. It is factor analysis which needs it. So, I assume that you are doing FA. But you have chosen to do in by means of PCA (which isn't a true FA method). – ttnphns Jul 10 '16 at 12:12
  • 1
    You ought to yell out onto those critics what they mean specifically by the decision level. Usually factor analysts, when factor validating a being designed questionnaire, set for themselves a loading threshold bolow which an item cannot be eligible for the construct. Typically, value about 0.4 is the threshold. However, it all depends on the data. In some research fields, data are so high correlated that a threshold could be chosen 0.6, for example. – ttnphns Jul 10 '16 at 12:24
  • Perhaps ask the critics to explain? – mdewey Jul 11 '16 at 13:03

0 Answers0