2

First of all there is another question about it here, but it has no answers unfortunately...And also here, but it is more general about mixed models and GEE, while my question is more specific...

So, my question is simle but also very general: Are GEE and GLS the same model ?

For instance, if i run a model with gls() + AR(1) structure and a GEE +AR(1) structure, will these 2 correspond to the same model ?

And if not, why ? In which part do they differ ?

Thanks

GiannisZ
  • 195
  • 1
  • 12
  • Are you sure ? Because they talk about mixed models in general vs GEE...But my question is more specific about GLS ( no random parts ). And also, i do not have the reputation yet to comment there.... :( – GiannisZ Jul 24 '18 at 14:16
  • Possible duplicate of [When to use generalized estimating equations vs. mixed effects models?](https://stats.stackexchange.com/questions/16390/when-to-use-generalized-estimating-equations-vs-mixed-effects-models) – kjetil b halvorsen Jul 24 '18 at 18:58
  • i explained myself in the previous comment....someone else also made the same comment which he probably deleted.... – GiannisZ Jul 24 '18 at 19:06

1 Answers1

0

For the case of linear models (i.e., marginal models with an identity link function), the generalized least squares (GLS) estimator of $\beta$ discussed in Chapter 4 can be considered a special case of the GEE approach.

Fitzmaurice, Garrett M., Nan M. Laird, and James H. Ware. Applied longitudinal analysis. Vol. 998. John Wiley & Sons, 2012.

user343233
  • 1