Differences in margins postestimation command for tpoisson relative to poisson

[Austin Shermon]

I'm currently using a truncated poisson model with command tpoisson and ll(0) and ul(6) for a dependent variable that is constrained between 0 and 6. I'm trying to estimate the marginal effect of a particular covariate (experience) on the DV at particular levels of another covariate (patent):

[margins, dydx(experience) asobserved at(`patent' = (0 1 2 3 4 5 6 7 8))]

I'm wondering if the margins postestimation with dydx has a different (or even valid) interpretation relative to the margins postestimation with dydx in a traditional Poisson model?

Thanks!

[John Mullahy]

I think the interpretation is the same. The default parameter for each is "n" (which I believe is E[y|x] in both instances), although other parameters can be specified instead.
See margins under

help tpoisson_postestimation

and

help poisson_postestimation