Fisher's exact test - 2x1 versus 2x2 contingency table

awolf

New Member
Hello everyone,

I am having a problem comparing proportions between groups.

My problem: A mouse can enter door A or B. I have two groups. In group 1, 90% of mice enter door A, in group 2, 55% of mice enter door A. I want to test whether there is a difference in the proportions of entering door A between the groups.

I found that a way of testing this is to create a 2x1 contingency table and compare group 1 versus 2. The outcome gives the expected n for each door (n=5 for each), a residual and a p-value.

So far so good. My problem is that in one experiment, 100% of mice enter door 1, and I cannot use the 2x1 table set up in SPSS (it says the value is a constant and the test cannot be done).

An alternative way is to create a 2x2 contingency table and use the chance level (50%) as the second row. However, this test gives me different, higher p-values.

My problem is that I am not sure what exactly the difference is between the tests, and which way is the correct one to answer my question. What does the 2x1 table test exactly compare? If I can use the results from the 2x1 table, how do I deal with the group that chose door A 100%?

Any help would be greatly appreciated!

awolf

New Member
Never mind, I found my mistake and I am realizing my post wasn't very clear