poisson regression result

[teshager mazengia]

poisson regression result

Today, 02:13
i am doing research on determinants of crop diversity and poisson regression result is as follows:
poisson cereal farmsize gender age hhsize impseed hhplotdis irr awarness livestock i.oxen i.ti e
> du income es numplot productio sccs, vce(robust)

Iteration 0: log pseudolikelihood = -289.59357
Iteration 1: log pseudolikelihood = -289.55671
Iteration 2: log pseudolikelihood = -289.55668
Iteration 3: log pseudolikelihood = -289.55668

Poisson regression Number of obs = 200
Wald chi2(22) = .
Prob > chi2 = .
Log pseudolikelihood = -289.55668 Pseudo R2 = 0.1392

goodness of fit result
poisgof

Deviance goodness-of-fit = 2.75826
Prob > chi2(177) = 1.0000

Pearson goodness-of-fit = 2.713196
Prob > chi2(177) = 1.0000
my question is can i use the estimated model for analysis with missing wald chi² and the above goodness of fit result

[Carlo Lazzaro]

Teshager:
the missing Wald test outcome can depnedent on the non-default standard errors.
However, I agree that the results of -estat gof- are weird.
You should be ebtter off with sharing with the list s what you tyoed (as you did) but also what Stata gave you back (ie, -poisson- outcome table). Please use CODE delimiters for that (see the FAQ).
In addition, you can also share an example/excerpt of your data via -dataex- (see the FAQ again). Thanks.

[teshager mazengia]

thank you when i avoid i. for categorical variables goodness of fit and Wald tests are corrected, but i face inconsistency between tobit and Poisson regression. i used tobit model for diversity index (margalef index), which is left censored. with same explanatory variable i estimate poisson regression. for the two model the coefficient of variable have opposite sign. how can estimate the two model?

[Carlo Lazzaro]

Teshager:
different results returned by different regression models with different regressand are, in fact, not surprising.