I have 6 sites with 3 sites as constant food in a feeder and 3 sites as pulsed food in a feeder. For the pulsed food sites, there are days with food and without food.So, I have 2 treatments: constant, pulsed with food, pulsed without food.

I counted birds at 2 locations(near and far from feeder) at each site multiple times. I want to know if treatment predicts the difference in the number seen near and far. So I want to know, is there a significant difference between the differences?

I've thought about doing this problem in a mixed models approach: (number of birds) ~ treatment*location + (1|site)


(number of birds near - number of birds far) ~ treatment + (1|site)

Any advice/guidance on this?

  • $\begingroup$ Is "constant" a treatment? If yes, why do I count 3 treatments rather than the 2 you mentioned? $\endgroup$ May 15, 2019 at 22:13
  • $\begingroup$ Sorry, I didn't explain that well. I have 2 treatments - constant and pulsed. But, for analysis, I have broken the pulsed into pulsed with no food and pulsed with food. I do surveys on days where the pulsed has no food and days the pulsed does have food. $\endgroup$
    – Rachael
    May 16, 2019 at 16:28


Your Answer

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

Browse other questions tagged or ask your own question.