I have a small sample in 2x2 contingency table that it is too small even for Fisher. Therefore I would like to use a continuous distribution like in the Barnard's Test. How can I implement it in MS Excel? Thanks!
Is there any particular reason you would want to do it in Excel? R has it in an extremely easy implementation.
https://www.r-statistics.com/2010/02/barnards-exact-test-a-powerful-alternative-for-fishers-exact-test-implemented-in-r/
There is also an online calculator for it:
http://scistatcalc.blogspot.com/2013/11/barnards-test-calculator.html
If you have the exact calculation procedure that is used by Barnard's, I could implement it for you more than likely.
Hi Matthew,
Many thanks for your answer. I know about R and the online calculator (I used it this morning to test one variable).
We need to use MS Excel because we are building a database and I would like to run the test quite often, even once a day, to "see where er are going". With Fisher is very easy, we just click and we have the results in a second. However, we found that when we have few data there are problems. This is known in the literature since Fisher is a discrete distribution. Barnard's Test uses a continuous distribution so this issue is less relevant.
Therefore, I was keen to add a column with the Barnard's Test .
Here the description https://cran.r-project.org/web/packages/Barnard/Barnard.pdf
How would you implement this?
Many thanks!
This appears to be a much more difficult problem than I was envisioning, however, I ran into an article that talks about making calculation of this more efficient.
Mato, A. S., & Andrés, A. M. (1997). Simplifying the calculation of the P-value for Barnard's test and its derivatives. Statistics and Computing, 7(2), 137-143.
The other thing that would be helpful is getting your hands on the execution of method Barnard in that package. If you have that, you can almost certainly convert it over to a macro (which is just a VBA subroutine). VBA is a very simple language to work with, so it shouldn't be much of a problem once you have a procedure to follow.
- Badges
- Science topic