[Limdep Nlogit List] ordered probit with selection

[Blane Lewis]

Hello,

I'm running an ordered probit with selection. The first round probit and ordered probit (without selection) work well. The ordered probit with selection converges (exit status = 0) and the estimates look reasonable. The second round probit is problematic however--all coefficients have t values exactly equal to zero. I've scaled the variables well so I don't think that is the issue. Any suggestions on where to look for the problem?

Thanks in advance,

Blane