I can think of three approaches that could be considered, and the correct
one will probably depend on the details of your study.
1. Kappa statistics are commonly used in medicine to assess a diagnostic
procedure with a gold standard (radiological images are the gold standard
for diagnosis of artery blockage, we could contrast how often medical
history can make the correct diagnosis). Kappa would be covered in most
2. Canonical correlation is used to correlate two subsets of variables in
a dataset. This assumes (as I recall) multivariate normaility but I am
guessing it is probably robust to violations of this assumption (aren't
all normal distribution methods robust?). Canonical correlation would be
covered in a multivariate statistics text book.
3. Mantel statistics might be appropriate since they wourk on different
matrices. Good references would be Legendre and Legendre book on Numerical
Ecology and Smouse, Long, and Sokal, Systematic Zoology, 1986,
William D. Shannon, Ph.D.
Assistant Professor of Biostatistics in Medicine
Division of General Medical Sciences and Biostatistics
Washington University School of Medicine
Campus Box 8005, 660 S. Euclid
St. Louis, MO 63110
e-mail: [log in to unmask]
web page: http://ilya.wustl.edu/~shannon
On Mon, 11 Nov 2002, Ufuk Yildirim wrote:
> Hi All,
> I am a doctoral candidate studying in the UK. As part of my research, I use card sorting tasks to investigate underlying cognitive structures of prospective teachers within a subject domain. The result of this card sorting task is a symmetrical matrix with 56 rows and 56 columns. What I want to do is to correlate two matrices which represents an expert's categorisation of the cars and a subject's categorisation. Here I have the problem. I do not know how to do this. Can you help me in solving my problem? I am reasonably familiar with SPSS. Is it possible to do this with SPSS? Or do I need another program?
> Thank you very much for your interest and time.
> Ufuk YILDIRIM