2
$\begingroup$

I've got some problems with selecting right model/method in my analysis.

Two groups of animals (differ by "Treatment") were measured from 1-st to 42-nd day, one common value for each group was measured by each day (food consumption per day). I need to compare this groups and I want to use ANCOVA with "Treatment" as fixed effect and "Day number" as covariate in the model.

There are many examples which contain covariates with some sort of randomness in their distribution (like weight, age, IQ, etc., where we can't directly control this variable). But in our experiment we eliminate this variation by measuring groups day by day. Is it correct to use "Day number" as continious covariate?

Maybe the whole model is incorrect and you can suggest right one?

To be more concrete I would like to present the scatter plot of my data:


(source: fastpic.ru)

Best regards!

$\endgroup$
0
0
$\begingroup$

The scatterplot shows clear temporal auto-correlation in the data, which would certainly invalidate some of the statistical assumptions for an ANCOVA. You would be much better off using a general or generalized linear model with an appropriate residual autocorrelation structure to deal with this, otherwise you will not be able to trust the results from your model.

This is a reasonably large topic, and assuming it's not something you know about you would be best spending some time reading and understanding as much of the details as possible before trying anything. It sounds like you are in ecology or a related area, and as an ecologist I would really recommend the book by Zuur et al. 2009 - Mixed Effects Models and Extensions in Ecology with R. This covers exactly these issues, provides great step-by-step ecology-related examples, and is arguably 'easy' to understand, going into some of the maths if you want it, but not being essential.

$\endgroup$
4
  • $\begingroup$ Hi @JupiterM104! thanks for recomendation, what the difference between my data and data from here? Could you specify more precisely what of the statistical assumptions for an ANCOVA are not suitable here? $\endgroup$ Apr 28 '14 at 9:00
  • $\begingroup$ This is the study where I measured food consumption per day for chickens. So there is study for 42 days only, we can't talk about seasons here, I don't thing that we have autocorrelated data... Maybe I'm wrong. How can we check this? $\endgroup$ Apr 28 '14 at 11:05
  • $\begingroup$ The usual assumptions for an ANCOVA or similar general linear models are that the residuals are normally distributed and homoscedastic, and that each observation is independent. Although not the residuals, the clear, curvy shape of your scatterplot indicates almost certain temporal autocorrelation, and therefore non-independence of observations. I don't mean to be rude, but checking model assumptions is quite a basic issue, and will be covered in any decent stats book/many places online, so I would suggest doing some ground work reading and then asking more specific questions if necessary. $\endgroup$ Apr 29 '14 at 10:57
  • $\begingroup$ Ok, thanks, I'll try to go deeper and to read more about autocorrelation and more complex models. $\endgroup$ Apr 29 '14 at 11:07

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.