2

I was wondering if it is possible to apply correspondence analysis to a 2x2 contingency table. Since that correspondence analysis is a method for dimensionality reduction, I think it is necessary to have a contingency table with many rows and columns. Is there a minimum value of the number of rows and columns needed to apply correspondence analysis?

Thanks in advance.

Carl
  • 11,532
  • 7
  • 45
  • 102
Ricardo UES
  • 461
  • 1
  • 3
  • 8
  • 3
    (Please read paragraphs on Correspondence analysis in http://stats.stackexchange.com/a/141755/3277.) In a 2x2 contingency table since there is only 2 categories on each side profiles are trivial: to have "similar vs distant" profiles you need to have at least 3 categories. Moreover, the number of dimensions of the biplot will be `min(2-1,2-1)=1. Thus, the results are trivial and of no much interest. – ttnphns Sep 16 '16 at 04:48

0 Answers0