[Limdep Nlogit List] Binary probit

[Steven Yen]

I would like to estimate the binary probit model without setting the 
scale parameter (s) to unity. The likelihood contributions would be 
CDF(-x'b/s) and CDF(x'b/s). Any suggestions as to other 
identification restrictions to make the parameter estimates unique? I 
was told the Bayesian do it, but I have not identified a proper 
reference. Any lead would be appreciated.