3
$\begingroup$

I'm using lme() from the lme4 package to set up my model. The experiment analyzes the effect of six different fertilizer treatments on yield, each treatment has four replicates in a randomized block design and was carried out 11 years.

The model I tryed fitting was:

fit.model <- lme(yield ~ treatment * year + (1|Block))

After reading various posts on CrossValidated I am now unsure if I can consider year as a fixed effect, most of the research papers I have read in my field take it as random.

I am mainly interested in the effect of treatment on yield over all years but also every single year, because of that I thought year should be a fixed effect.

If you need more information, just ask me for. Thanks for your help!

Improve this question
$\endgroup$

1 Answer 1

Reset to default
2
$\begingroup$

Going by what I understood from your wording, year is a continuous variable that goes from 1 to 11.

If that's right, then it's a fixed effect. It could be a random effect if you sampled years randomly (from the population of years), which does not seem to be the case here.

Improve this answer
$\endgroup$
2
  • $\begingroup$ I imagine that the differences in harvest from year to year is mostly random (amount of sun, rain, temperatures, etc) and has little to do with some sort of fixed effect. What is possible that is correlation between different adjacent years due to climate effects that span a wider range than a single year. $\endgroup$
    – Sextus Empiricus
    Commented Jan 8, 2021 at 14:41
  • $\begingroup$ I can follow both of your answers, thanks but what if I want to look at the differences in each single year with emmeans than I would have to take it as fixed to use the by argument in emmeans? $\endgroup$
    – Censkey
    Commented Feb 2, 2021 at 9:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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