OpenEpi, Version of December 2003

Test Results for R by C Table Module

 

 

Dataset #

Data

Compared with

Result

RC1

983,383,2892

679,416,2625

134,84,570

Snedecor and Cochran, Statistical Methods, Seventh ed., Iowa State, 1980, p. 210

Agree with chi square of 40.54, p < .0000001, df=4

 

RC2

43,11

3,18

Epi Info  chi sq=27.2225

Oswego/Vanilla by Ill

P=1.3505e-6

 

http://www.fourmilab.ch/rpkp/experiments/analysis/chiCalc.html

for this chi sq value, p=1.8126e-7

 

RxC module gives 27.22

P=1.813e-7

Chi Square agrees (27.22) with Epi Info, but p value for 1 d.f. differs.  RbyC does agree with the p-value from the url given, however.

P=0.0000001813

 

 

Generated a 5 x 5 table that gave chi square of 45.5, with 16 df.  P value was .0001162 and site above gave 0.0001

Generated another table with 4 d.f.  chi sq 0 .4688.  chiCalc gave p=0.9764, Rxc gave 0.9765.  Hence the p from chi square appears to work correctly.

 

 

By Andy Dean.  Agreement is considered satisfactory if the first two significant figures agree, when a number, such as a p-value, is small.

 

Conclusions: R x C gives correct results for the Snedecor/Cochran 3 x 3 table. The calculation of p with 1 df differs from Epi Info, Version 3, by an order of magnitude, but agrees with p from a website in Switzerland.  It may be advisable to look into Epi Info’s p-value calculator when p is very small.

 

R x C appears ready for use.