I'm analysing some roe deer biological traits with a gam in mgcv. One of my variable is a factor with nine levels (i.e represents a combination of litter size and sex: M, F, FF, MF and so on) and my response variable is total litter size weight. I'd like to apply a post hoc test in order to compare F vs M. Many methods or r package are avalaible but only for models of class "lm", "lme" or "GLM".

can someone tell me if there is a method?

  • $\begingroup$ What do you mean when you say "compare"? Joint significance of several variables? $\endgroup$ Feb 21, 2017 at 15:18
  • $\begingroup$ Basically i want to compare the mean of two levels of variable sequence (M, F, MF, FF, MM) and see if the difference is statistically significant. For example, M vs F $\endgroup$ Feb 21, 2017 at 19:19
  • $\begingroup$ How do you define such a "mean" if you have continuous covariates? $\endgroup$
    – Roland
    Feb 22, 2017 at 8:09
  • $\begingroup$ My response variable is total_litter_mass and Sequence is one of my covariates. Nevertheless, Sequence have 9 levels and the model take as reference level F but i am not interested if there is difference between F vs MF. Basically i would like to know if is possible to apply the same concept applied for orthogonal contrast. $\endgroup$ Feb 22, 2017 at 8:35


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.