I have 25 years of nest initiation dates, I used quantile regression to look for changes in the distribution over time, as well as to look in detail to early and late breeders. My model would be something like: nest date ~ year. Since the environment where the species nest is highly variable, there is a lot of fluctuation in nest initiation dates. Do I need to center my observations (i.e. my response variable laying date) for each year before I run the quantile regression to get rid of the noise of annual variation? (i.e. that some years all birds nest late and other all birds nest early).

UPDATE: With and without centering I get opposite results. Which one is right?enter image description here

enter image description here

  • $\begingroup$ Centering won't get rid of noise. It will just mess up your estimates of trend. $\endgroup$ – Nick Cox Jan 26 '18 at 15:13
  • $\begingroup$ Centering is generally used to 1) reduce multicollinearity for interaction terms and 2) make 0 a meaningful value for your predictor (in the event that you're using a scale that does not have a possible/meaningful 0). Nick Cox is right that it won't help you with reducing that variability. $\endgroup$ – TPM Jan 26 '18 at 15:31

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.