The Cochran Armitage Trend Test is used in genetic association studies where you expect a trend between the groups viz. wild, hetero and homozygous. But if any of the counts in the 2x3 table are less than 5, will the CATT test be valid? If I'm technically wrong in my question, please correct me.
Cochran Armitag Trend (CAT) test is an asymptotic test. Permutation-based or Monte-Carlo-based CAT test is suggested for the case of a small sample size.
CAT has beken applied for clinical trailer to test for dose reponse curve and other medical data with oftenly abnormally distributed. The approach for trend test bebas es on nonparametrical metod i.e. Khi square multilevel comparison or by logistric procedure. The samples in the classificattion or treatment groups might turn to be small, but does not necessarily overriding the rule of 5 as minimum, since the minimum si actually mathematically the grand total fo 13 or 15, not the n of each cell. COchran Armitage Trend test has been comparison using Freq Proc, Multitest Proc and Logistic Proc, including permutasi on and the results are apprecibly similar. Any way, the CAT si handy and helpful to Finder appropriate application..
Pleasure, refer to the attached paper on CAT.
Article The size of the CochranArmitage trend test in 2×C contingency tables
Pleasure, Finder attached file for further consideration on CAT
Article Cochran-Armitage Test versus Logistic Regression in the Anal...
The folllowing was mistransmitted, as related discussion
Article On estimation of the variance in Cochran-Armitage trend test...
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics