st: Re: zscores in xt data

[Kit Baum <baum@bc.edu>]

If you want to create standardized variables over subsets of xt data, 
it is not clear to me that calculating them for each time period would 
be the obvious way to do it. The major source of variation in the data 
might be along the cross-sectional dimension. Consider xt data on US 
cities: the variation in, say, mean population among the 20 largest 
cities is much greater than the variation over the last 10 years in 
mean city size. So you might want to consider doing the z-scoring by i, 
rather than by t.

Kit Baum, Boston College Economics
http://ideas.repec.org/e/pba1.html

On Jan 3, 2005, at 2:33 AM, statalist-digest wrote:


> I would like to ask for advice in using standardized variables 
> (z-scores) in
> cross-sectional time series analysis (xtreg).
>
> Should z-scored be calculated over the entire sample, or should 
> different
> z-scores be calculated for each time period?
>
> The argument for calculating z-scores by time period would focus on 
> keeping
> variables "within context."
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/