How to generate McFadden's R sqare for a choice model

[Woohyun Cho]

Dear all,

I wonder if Stata can generate McFadden's R sqare for a choice model.

Thank you very much in advance.

Respectfully,
Woohyun Cho

[William Lisowski]

For choice models, R²-like statistics are commonly called pseudo R-squareds (McFadden originally called his rho squared), and it is my understanding that the statistic reported by Stata's logistic and logit commands as "Pseudo R2" is McFadden's.

There's a nice discussion of pseudo R-squareds at http://www.ats.ucla.edu/stat/mult_pk..._RSquareds.htm

[Bob Fitzgerald]

Long and Freese's spost13 has a "fitstat" program that, among measures of model fit, reports McFadden's R square.

Bob

[Woohyun Cho]

Thank you, William Lisowski and Bob Fitzgerald. Your help is greatly appreciated. Fitstata did not work after I ran 'asclogit' command. But let me try to figure it out.
regards,
Woohyun