Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: heteroskedasticity between groups, interpretation of -sdtest-


From   Nick Cox <njcoxstata@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: heteroskedasticity between groups, interpretation of -sdtest-
Date   Mon, 17 Dec 2012 11:12:17 +0000

If your main interest is in the regression, the results of -sdtest-
are at best marginal (pun intended) and at worst irrelevant.

The main issue is whether heteroskedasticity is present given the
predictor variables you are using. It is often best to explore this
graphically, by looking at scatter plots or residual plots.

Nick

On Mon, Dec 17, 2012 at 11:01 AM, Laura R. <laura.roh@googlemail.com> wrote:

> I have estimated a regression in which the cases stem from two surveys
> that have been conducted in different years. I put a dummy variable
> for the survey year into the regression
>
> Now my questions:
> 1. Do I still have to worry about group heteroskedasticity, i.e. that
> the estimations are somehow biased?
>
> 2. Does it matter if variances between these two groups are
> homo/heterogenous for the descritive statistics? (I mean, when I
> describe the mean of the dependent variable in these two years, and
> make comparisons)
>
> 3. I did a test for heterogeneity between the variances between the
> groups (-sdtest-), but how can I interpret the results?
>
> E.g., I did: sdtest depvar, by(survey_1)
>
> For example, I received this result:
>     Ha: ratio < 1               Ha: ratio != 1                 Ha: ratio > 1
>   Pr(F < f) = 0.8246         2*Pr(F > f) = 0.3508           Pr(F > f) = 0.1754
>
>
> For some different data, this result:
>     Ha: ratio < 1               Ha: ratio != 1                 Ha: ratio > 1
>   Pr(F < f) = 1.0000         2*Pr(F > f) = 0.0000           Pr(F > f) = 0.0000
*
*   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/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index