Hi everybody,
I want to examine the relationship of wood density with tree growth rates expressed as (i) diameter growth rates and (ii) biomass growth rates. In your opinion, what would be the best way to compare these two relationships, for example in terms of strength and/or the slope. Any advice on R packages or functions would be very valuable.
To run Linear Mixed models I usually use the lme4 package, lme function. Two options for the best model selection are AIC or likelihood ratio test (ANOVA). The lowest AIC is going to be the best model, but make sure that for comparing models they should be fixed with the same method, e.g., "REML". We can not compare a model fixed with REML with another fixed with ML for example. When we are choosing the model we are interested in the best structure of fixed effects, but before you should select the correct random-effect because it surely will affect your standard errors and the slopes. So first select well your random effect and then you can compare different models to select the best-fixed effect.
so you can use aic or anova, any way look the aic of the final models is important because it gives you an idea of what model is more informative when you include factors. you always can double-check with the dredge function to obtain an AIC results table.
Johanna Andrea Martínez-Villa Thank you for your insightful comments, Johanna!
Do you have experience comparing the slopes of mixed models? I have compared the slopes of major axis regressions, using the "smatr" package, but I don't know how do it in the case of mixed models.
This paper should help answer your question.
Schielzeth H. 2010 Simple means to improve the interpretability of regression coefficients. Methods Ecol. Evol. 1, 103–113. doi:10.1111/j.2041-210X.2010.00012.x
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Writing