[R] Generalized Ordered Logit in R
B. Longhurst
bl361 at cam.ac.uk
Thu Jan 30 21:42:21 CET 2014
Dear community,
I am replicating a paper (The Effect of Structures and Power on State
Bargaining Strategies, http://tinyurl.com/oyk289c) for a class project.
The author used “gologit” in STATA, and I need the equivalent function
in R. She used a “Generalized Ordered Logit” (see below model
description).
Can you run a “generalized ordered logit” with glm? I could not find the
specification in the help file for the function. Or is there another R
package for it?
The data can be accessed here: http://tinyurl.com/oydwl4d (McKibben_AJPS
Bargaining Strategies data.tab)
The results should look like this:
http://s17.postimg.org/3zldxs133/Brooke.png
Thank-you so much for your help. It is greatly appreciated.
Kind Regards,
Brooke
Model Description from the original paper:
To test the effects of the bargaining structure against more standard
explanations of variation in state bargaining behavior, I estimate two
generalized ordered logit models with clustered standard errors. Model 1
includes lnGDP, focusing on states capabilities as the central measure
of their power, and Model 2 includes the alternative, Voting power
index. While standard ordered logit models might appear to be the
appropriate statistical technique, the data violate the assumption of
proportional odds that underpins these models. In other words, all
independent variables do not exert the same effect across all categories
of the dependent variable. The generalized ordered logit model relaxes
this assumption, allowing the effects of the independent variables to
vary across the different cate- gories of cooperative bargaining
behavior. Standard er- rors were also clustered by negotiation phase to
deal with the fact that each state’s choice of bargaining behavior
within a given bargaining interaction is dependent on the choices of
other states. (These models were estimated using the gologit2 command in
Stata)
More information about the R-help
mailing list