0

I’m running a trial for my supervisor in a fisheries position and it’s been a few years since stats.

I have two groups of marked fish.

Group A: 249 fish. Treated with a dye. 

Group B: 247 fish. Treated with tags.

These were treated, held, then released upstream for a quick mark-recapture effort. We want to see if our fish trap captures the treatment groups at different rates (IE if the fish behave differently).

Say we recapture 3 from group A, 2 from group B.

I am -thinking- this would be a chi-square? Would you compare percentages recaptured for statistical significance, is that viable, and with what test?

Fish Dood
  • 1
  • 1
  • You have two groups and a binary outcome (recaptured or not). This falls squarely into the world of 2x2 table analyses. Chisquare is fine, you could also do a test of proportions as that test would also give you a risk difference estimate (i.e. the difference in probability of recapturing) – Demetri Pananos Mar 17 '21 at 05:13
  • If your recapture probability is very small, you may want to read [this](https://stats.stackexchange.com/questions/14226/given-the-power-of-computers-these-days-is-there-ever-a-reason-to-do-a-chi-squa) question which concerns what to do when the expected cell frequency is below 5 – Demetri Pananos Mar 17 '21 at 05:22

0 Answers0