Help with Fisher's Exact Test

We're working a paper and the editor is asking me to use FET to calculate if the difference in capture between High Traps and Low Traps is statistically significant. We have two years of data. In our first attempt we have the data as Species 1: 1 specimen in HT and 10 specimens in LT. Jamovi is giving me a p=0.91. On Species 2: 10 specimens in HT and 2 in LT and Jamovi is giving me a p=0.15. This is bugging me because how can such similar data have such different results?

Our second attempt we had Species 1 in 2011 with 1 specimen in HT, and 7 specimens in LT; species 1 in 2012 with 0 specimens in HT and 3 specimens in LT for which the result was p=1.000. Species 2 in 2011 with 0 specimens in both HT and Lt and 2012 10 specimens in HT and 2 specimens in LT with also with p=1.000.

I understand that the N value is low and that's why we're required to use FET. We're just not sure if we're setting the data correctly and if this results make sense or if we're missing something.


Well-Known Member
The normal approach would be to use chi square. Probably with these low numbers you would use the continuity correction and move the observed values 0.5 towards the centre. In Excel -