0
$\begingroup$

I am having a great deal of difficulty understanding how to use the Generalized linear model for my data set. The response variable of interest is hatch success of sea turtles, which is a %. The explanatory variables are environmental variables pertaining to the nest environment (above ground vegetative cover, average daily temperature of nest, organic content of the nest soil, and distance to the mean high tide line at excavation). I was attempting to use a mixed effects linear model, but gave up the pursuit because the distribution of my response variable was non-normal, hence why I am attempting to use a generalized linear model.

My questions are: should I attempt to use a mixed effects Generalized linear model? or Generalized linear model?

which link function should I use? I know this depends on my distribution, which looks exponential (increasing).

what is the difference between variables with fixed effects and variables with random effects?

$\endgroup$
  • $\begingroup$ I am puzzled by "which logit function should I use?" -- as far as I know, there's only one logit function. Do you mean to ask which link function should you use? $\endgroup$ – Glen_b Sep 26 at 3:00
  • $\begingroup$ Yes, that was my mistake. Is logit a link function? $\endgroup$ – Manu Sep 26 at 15:14
  • $\begingroup$ yes, its one of the functions commonly used as a link function, being the natural link for binomial regression. $\endgroup$ – Glen_b Sep 26 at 15:23
  • 1
    $\begingroup$ Mmm. You asked "which link function to use", implying that you know what a link function is. But your new question "what is the logit function for" suggests you don't know anything at all about glms or link functions. No quick answer to that. You ask where to put weights in your code, but we don't know what statistical software you are using, let alone what code you have tried. You may need to allow for overdispersion because the eggs from each turtle are not independent -- they share common causes because they are from the same turtle and are laid in the same spot. $\endgroup$ – Gordon Smyth Sep 28 at 0:55
  • 1
    $\begingroup$ Manu, you need to think about how to ask a clearer question, one more likely to elicit real help for you. Please do that by editing your question rather than adding comments. Comments are not the right place for long discussion or for further questions. Amongst other things, you need to explain what software you are using and, if you don't know how to get started at all with glms, then explain that as well. When you fitted a mixed linear model, what variables were treated as random? Why did you think they should be random? And so on. $\endgroup$ – Gordon Smyth Sep 28 at 0:59

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.