Yes, it's hazardous. -panels- expects users know what they're doing. I
see that it is better not to rely on that. The problem, however, is
that -panels- is "blind" to what the called command does: It does not
know which variables to check.
I will think about how to make -panels- aware of the variables that
will be used by the called command, so that they can be checked
(although there will probably be no general solution that works with
all commands). Let me know if you have any ideas.
ben
On Thu, Oct 22, 2009 at 11:23 AM, Martin Weiss <[email protected]> wrote:
>
> <>
>
> Very good idea!
>
> Just playing around with it: Could the sentence in the help file "Variables
> referred to in command are assumed to be constant within panels." be backed
> up by a check on the part of the command whether that is really the case?
>
>
> *************
> sysuse auto, clear
> //first attempt
> panels rep: su gear
> //second attempt
> panels rep: su gear
> *************
>
> gives a different result for every execution. This is in keeping with the
> description in the help file, but could be a bit of a hazard for some
> users...
>
> HTH
> Martin
>
>
> -----Ursprüngliche Nachricht-----
> Von: [email protected]
> [mailto:[email protected]] Im Auftrag von Ben Jann
> Gesendet: Mittwoch, 21. Oktober 2009 22:20
> An: statalist
> Betreff: st: New -panels- package available from SSC
>
> Thanks to Kit Baum, a new package called -panels- is now available
> from the SSC Archive. Type
>
> . ssc install panels
>
> to install.
>
> -panels- has two purposes:
>
> (1) Quickly count panels (groups) in a dataset and display some basic
> information about the sizes of the panels. You can also use -panels-
> simply to count the number of distinct (unique) values of a variable.
> See Cox (2007) for some thoughts on counting panels.
>
> Cox, N. J. (2007). Speaking Stata: Counting groups, especially panels.
> The Stata Journal 7(4):571-581.
> http://www.stata-journal.com/article.html?article=dm0033
>
> (2) Compute statistics based on panel units (instead of individual
> observations). -panels- can be used as a prefix command to other Stata
> commands to apply them to panel units instead of individual
> observations. This is useful, for example, if you want to compute
> frequency
> distributions or summary statistics for panel characteristics.
>
> Examples:
>
> . webuse nlswork, clear
> (National Longitudinal Survey. Young Women 14-26 years of age in 1968)
>
> . panels idcode
>
> Number of obs = 28534
> Number of groups = 4711
>
> Obs per group: min = 1
> avg = 6.056888
> max = 15
>
> . panels idcode if race==2
>
> Number of obs = 8051
> Number of groups = 1325
>
> Obs per group: min = 1
> avg = 6.076226
> max = 15
>
> . panels idcode: tabulate race
>
> 1=white, |
> 2=black, |
> 3=other | Freq. Percent Cum.
> ------------+-----------------------------------
> 1 | 3,329 70.66 70.66
> 2 | 1,325 28.13 98.79
> 3 | 57 1.21 100.00
> ------------+-----------------------------------
> Total | 4,711 100.00
>
> ben
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
