Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
| From | "Venkiteshwaran, Vinod" <Vinod.Venkiteshwaran@tamucc.edu> |
| To | "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu> |
| Subject | st: Ordered Probit Interpretation |
| Date | Fri, 26 Aug 2011 13:55:08 +0000 |
I am running an ordered probit model with pooled cross-sectional data. The ordered dependent variable, Y_it has 7 categories that sometimes change within each i. THe variable of interest is X_it. In the results X_it enters the model with a highly significant negative coefficient implying that the marginal effect of a change in X_it is that it reduces(increases) the probability of the highest(lowest) category of Y_it. I then transformed the Y_it variable into changes and ranked the changes, Y*_it. That is if the change is upward, downward or no change, where the order is 3,2 and 1 respectively. I re-run my ordered probit model with Y*_it as my dependent variable and the same X_it. In this model X_it no longer has a significant coefficient. Econometrically how should this insignificance be interpreted? A note about my dataset and the X_it variable; I have a panel of firms across twenty years where each firm also belongs to a specific industry, j. The X_it variable is actually constant for each j,t. That is all firms within an industry-year have the same X_it. Dr.Vinod Venkiteshwaran Assistant Professor of Finance OCNR 352, College of Business Texas A&M University Corpus Christi TX 78412 Ph: 361-825-3619 Email: vinod.venkit@tamucc.edu * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/