[Limdep Nlogit List] possibility of two-staged ordered probit with random effects

Grossman, Philip J. pjgrossman at stcloudstate.edu
Sat Apr 26 02:26:42 EST 2008

I hope someone can help me.

The data I need to analyze is from a two-stage experiment.  In the first stage, a subject makes a prediction about which of 6 gambles (the gambles are linear in expected payoff and risk) each of the other n-1 participants selected for themselves.  So each subject makes n-1 predictions.  The predictor bases his predictions on a given set of individualized information about the others.  In the second stage, additional information is provided about the subjects and each subject is given the opportunity to revise their initial prediction.

The model I want to estimate (if it is even possible) is

(1)    P1 = X1b1 + e1

(2)    P2 = P1*b2 + X2b3 + e2


P1 = the first stage prediction
X1 = information about other participants provided in stage one
P2 = the second stage prediction
P1* = the predicted P1 from equation 1
X2 = new information provided in stage two

Is this even doable with Limdep and if so how?

Thanks in advance for any help provided.

Philip Grossman

More information about the Limdep mailing list