Re: st: dbetafit vs margins

[Maarten Buis <maartenlbuis@gmail.com>]

Note that the Statalist FAQ asks you to specify where you got user
written software from. There is a good reason for that: There are
often multiple versions of user written software floating around on
Statalist, and it is in your interest that we are talking about the
same thing... I will assume that you are refering to the -dbetafit-
which is part of the -betafit- package which you downloaded from the
SSC archive.

-dbetafit- was written for people who have Stata version < 11, and
thus don't have access to -margins-. So, in your case I would just use
-margins-. I have been able to replicate the differences you report,
and I will look into them more when I have time.

-- Maarten (co-author of the -betafit- package)

On Wed, Jan 29, 2014 at 5:46 PM, Cristina Sechel <cs998@york.ac.uk> wrote:
> I'm using betafit to estimate a model with a proportional dependent
> variable and I want to obtain various marginal effects. At first I used
> dbetafit to obtain the marginal effects of the continuous independent
> variables, but the output shows up in scientific notation and that is
> problematic for me because most of my estimates are very small and they get
> cut off.
>
> I then decided to use margins because I like the format of the output a lot
> better and I can see the estimate to many more decimal points. I also like
> that I can use marginsplot after margins and it's just overall a more
> versatile command than dbetafit, so I'd rather use margins.
>
> But the results of dbetafit and margins do not exactly match and I don't
> understand why. For continuous variables, the estimates are the same but
> the standard errors are different. And for dummy variables, the estimates
> do not match at all. Does anyone know why this is happening?
>
> The dbetafit code I used is:
>
> dbetafit, at(D1 0 D2 0)
>
> The margins code is:
>
> margins, dydx(*) at((mean) X1 X2 X3 (zero) D1 D2)
>
> where X1, X2, X3 are the continuous independent variables, and D1, D2 are
> the dummies.
>
> Thanks,
> Cristina
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/



-- 
---------------------------------
Maarten L. Buis
WZB
Reichpietschufer 50
10785 Berlin
Germany

http://www.maartenbuis.nl
---------------------------------
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/