Re: st: Tabulate summary statistics by percentiles and save output

[Nick Cox <njcoxstata@gmail.com>]

One problem here is where you think Stata is looking for saved results  
such as r(p1). Either non-missing values are accessible because they  
were left behind by some _previous_ command; or Stata will use missing  
values, which in your case will result in lack of restriction on the  
observations used. It's not an error to refer to saved results that  
don't exist.

At least at the start of your first loop no such non-missing values  
are (likely to be) present.

You refer here to several user-written commands but in no case do you  
explain where they come from.

Nick

On 28 Dec 2012, at 11:43, annoporci <annoporci@gmail.com> wrote:

> Dear Satalist,
>
> I wish to tabulate some summary statistics for some percentiles and to
> export the tables to files in tex format. I have never done that  
> before. I
> have read the stata help about esttab, estpost, estout, tabout and  
> tried
> various things that failed.
>
> My best shot so far. Consider:
>
>   sysuse ibm,clear
>   tsset date
>   local variables ibm spx
>
> /// Tabulate moments for different percentiles:
>
>   foreach var of varlist `variables' {
>     summarize `var' if inrange(`var',`=r(p1)',`=r(p10)'), detail
>     summarize `var' if inrange(`var',`=r(p90)',`=r(p100)'), detail
>   }
>
>
> The result is displayed in the Stata window. For reference:
>   p1-p10:   Obs = 493, Kurtosis = 5.05607
>   p90-p100: Obs = 50,  Kurtosis = 3.619203
>
>
> /// Export some summary satistics:
>
> foreach var of varlist `variables' {
>     quietly estpost summarize `var' if inrange(`var',`=r(p1)',`=r 
> (p10)'),
> detail
>     esttab using moments`var'.tex, replace cells("count mean sd  
> skewness
> kurtosis") noobs nodis nonumbers nonotes
>     quietly estpost summarize `var' if inrange(`var',`=r(p90)',`=r 
> (p100)'),
> detail
>     esttab using moments`var'.tex, append cells("count mean sd  
> skewness
> kurtosis") noobs nodis nonumbers nonotes
>   }
>
> There are two issues: one about the display, one about the data.
>
> 1) FIRST issue: my code creates two separate tables stacked on top  
> of each
> other. I would like instead that the second round of data for count,  
> mean,
> sd, skewness, kurtosis be "inserted" into the first table just below  
> the
> first round of data, something like:
>
>                    &       count&        mean&          sd&     
> skewness&
> kurtosis\\
> \hline
> ibm,    p1-p10    &         493&    .0461508&    1.231737&
> .144965&     5.05607\\
> ibm,    p90-p100  &         493&    .0461508&    1.231737&
> .144965&     5.05607\\
>
>
> 2) SECOND issue: my loop is not correctly selecting the percentiles.  
> As you
> can see, the results for the two parts of the loop are identical.
>
> In fact, the very first loop reported above is also incorrect. The  
> total
> number of observations in the dataset is 493, so p1-p10 should have  
> fewer
> than that and not 493 as reported.
>
> I have tried playing around with preserve/restore, I have also tried
> inserting "use ibm,replace" at various stages inside the loops, but
> couldn't work it out.
>
> I may be approaching this from the wrong angle. Any suggestions will  
> be
> much appreciated. Thanks.
>
> Patrick Toche
*
*   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/