Interpretation of average partial effects in tobit models
Posted: Wed Feb 11, 2009 2:40 am
Hello,
the discussion "Marginal effects in Tobit Model" in this forum dealt with the calculation of marginal effects in truncated tobit models. Thanks to the Glenn's instructions, I computed the average partial effects as mentioned in Wooldridge (2009, 591) rather than compiling the marginal effects at the averages.
However, in order to evaluate the important of exognours variables, I am a bit struggeling with the interpretation of these coefficients. For instance, does a discrete variable (ranging from 0 to 3) with an average partial effect of 0.3 have a bigger influence (on average) on the endogenous variable than a discrete variable (ranging from 0 to 7) with an average partial effect of 0.1?
I would appreciate any help, thanks a lot!
Cheers,
Raupel
the discussion "Marginal effects in Tobit Model" in this forum dealt with the calculation of marginal effects in truncated tobit models. Thanks to the Glenn's instructions, I computed the average partial effects as mentioned in Wooldridge (2009, 591) rather than compiling the marginal effects at the averages.
However, in order to evaluate the important of exognours variables, I am a bit struggeling with the interpretation of these coefficients. For instance, does a discrete variable (ranging from 0 to 3) with an average partial effect of 0.3 have a bigger influence (on average) on the endogenous variable than a discrete variable (ranging from 0 to 7) with an average partial effect of 0.1?
I would appreciate any help, thanks a lot!
Cheers,
Raupel