validity of data arrangement in a 2x2 table for fishers exact test.

#1
Hi,

I have two conditions (A and B) to which the subjects has been exposed to. There are two responses to those conditions yes or no. I can arrange the data in a 2x2 table in two different ways:

View attachment 5211


The P value for fishers exact test is 0.0028.

View attachment 5213

This table for fishers exact test gives me P value of 1.

My question is, is the arrangement of the data in table 1 valid enough to be used?

Any response would be greatly appreciated.

Thank you.
 
Last edited:

Dason

Ambassador to the humans
#2
You second table has "B no" twice and your totals don't match. Are these supposed to be representing the same data?
 

Dason

Ambassador to the humans
#4
Your totals don't match still though. I'm really not sure how you're equating these two tables.
 
#5
It is the same data. But the second table is looking at interactions/overlap between the two variables The first table directly informs the response to each variable on its own. This is why the totals do not match. I would like to present my data in the form of table 1 but I am not sure if the arrangement of data in table 1 is valid. Thank you for looking at the question earlier.
 
Last edited:
#6
When I look at the interaction between the two variables and report that in the table, then table 2 ends up having fewer numbers than table 1. So in total I have 76 responses. When I look at the response to each variable separately then it is 38 to each variable giving the total of 76 responses (A+B).

In the second table I am looking at responses that are common to A and B and that is why the numbers in table 2 are smaller because any response that is unique to only one variable is excluded and not counted for.

I can send my entire table if necessary?

Thanks again for your time.