Im trying to determine the correct analysis for paired data with a grouping variable.

I have 3 fish in each of 4 treatments, so 12 fish in total. Treatments are that the fish are fed at one of four different time points: 24h, 48h, 72h, 120h. However, it suits my study better to consider Time as a factor, so lets call Time = A, B, C and D.

From each individual fish I have repeated sampling: Nitrogen in blood is sampled at the beginning of the experiment (T=0) and then again 3 hours after their feeding treatment time (T= A,B,C,D).

I would like to know 1) if any of the time treatments causes a signifcant change in blood Nitrogen, and 2) which treatments have blood Nitrogen significantly different from one another (A vs B vs C vs D).

If I take the difference in Nitrogen between the two time points within each fish, and then run a one-way anova between treatments, presumably this only tells me if means of treatments are different, but I dont think this tells me if any of the differences in Nitrogen between time points within a fish (within each treatment) are signifcantly different from zero?

What would be the appropriate way to analyse this data to test my two questions?


You should include a repeated-measures (or within-subject) factor (pre- vs. post-delay).

This results in a mixed ANOVA with two factors, and you can test both factors using the same model.

Note: 3 fish per condition seems to be very small number. Are you serious about the research question?

  • $\begingroup$ Thanks Jank. 3 fish should be enough as preliminary tests suggest between fish variation is small and treatment effects large. $\endgroup$ Feb 21 '15 at 16:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.