Statalist The Stata Listserver

[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: lnskew0 and bcskew0

From   Maarten buis <[email protected]>
To   [email protected]
Subject   Re: st: lnskew0 and bcskew0
Date   Thu, 19 Apr 2007 21:54:32 +0100 (BST)

--- [email protected] wrote:
> How certain can I be that lnskew0 and bcskew0 have normalized the 
> distribution for the variable have transformed with them (a percent)?
> After using these transformations, the histograms are still skewed.  
> What could be the problem? Is it possible that they don't work in all
> instances?

you can find the skewness by typing -sum variable name, detail-. 

There are distributions for which it won't work too well, for instance
if your distribution has a spike (one number that is very very common).
If your variable is a percentage than obvious candidates for spikes are
zero and a 100%. 

What these commands try to do is to find a transformation that brings a
statistic called skewness, one number, as close to zero as possible.
This statistic is defined in reference manual entry for -summarize-.
Having that statistic close to zero is a necessary but not sufficient
condition for a variable to be normal(ized). 

Furthermore, if variable is the dependent variable in a model than you
do not want that variable to be normal, but the residuals. So looking
at these type of statistics is not a good idea, you should be looking
at the residuals of your model. Moreover, you say that it is a
percentage. In that case you don't want to normalize (use -regress-) at
all, since there are better ways: -betafit- and -glm, family(binomial)
robust-. These are described in: 

Hope this helps,

Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434

+31 20 5986715

Yahoo! Mail is the world's favourite email. Don't settle for less, sign up for
your free account today* 
*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index