How would you interpret negative correlations between factors in factor analysis, and should they ever be negative?
I did FA with oblique rotation (Oblimin) on 30 variables which were intended to measure the same dimension, and I extracted 4 (correlated) factors. All the variables were previously recoded in order to express the same side of the measured dimension. Yet, some of factors have negative correlations that I don't know how to interpret. Also, it may be important that some of the loadings from the pattern matrix were negative (which I also am not sure how to interpret). I would expect them all to be positive, since the items were recoded...
(Don't know if it matters, but I am using SPSS statistics, and the extraction method was maximum likelihood.)
I hope somebody more experienced in factor analysis can help me out.