Re: st: Vargranger

[Nick Cox <njcoxstata@gmail.com>]

The syntax of -display- is defined in the help as

  display [display_directive [display_directive [...]]]

which is to be interpreted as the command -display- followed by zero
or more display_directives. The term display_directive is defined
immediately. (Asking to -display- zero display_directives makes sense,
as -display- by itself yields a blank line.)

The examples of -display- given in the help include

. display as text "mean of mpg = " as result r(mean)

showing in this case that literal text and a saved result can be
shown. Use of formats is indicated as a possibility in the help. It's
to be expected that you'd need to look at other sections of the help
and/or the manuals to find out more.

In my recent example R was explained to be a matrix. Earlier in the
thread I gave an example of putting saved results into a matrix.

Nick
njcoxstata@gmail.com

On 8 July 2013 13:18, Ioannis Bournakis <I.Bournakis@mdx.ac.uk> wrote:

> Sorry Nick, I know...this question might look basic but it is quite
> confusing indeed. I could not find any explicit example in the help display.
>
> In you example below, what does R[1,1] declare? Is it the name of a matrix
> previously defined or just stands from return.

Nick Cox

> Your reply was sent in HTML. See
>
> http://hsphsun3.harvard.edu/cgi-bin/lwgate/STATALIST/archives/statalist.1307/date/article-225.html
>
> for how your email will appear to the majority of the members of this
> list  The FAQ explains that you should not send formatted posts.
>
> As before, why not read the help to find out what you need? -help
> display- makes it quite explicit that you can use -display- to display
> two or more things.
>
> For example, if a statistic were held in the first column and a
> P-value in the third column of a matrix R then something like
>
> . di  %9.3f  R[1,1]  "     " %05.3f R[1,3]
>     0.065     0.978
>
> shows some possibilities. In general, you will need to specify (a)
> literal text (b) numeric results (c) display formats for a decent
> display. (The numbers here have no bearing on your example,
> naturally.)

On 8 July 2013 11:55, Ioannis Bournakis <I.Bournakis@mdx.ac.uk> wrote:

>> That was helpful, thanks.  Is there any possibility to report two
>> statistics
>> in the same row suing the di [a,b] command.
>>
>> For instance, I am given this output after the vargranger  command:
>>
>> Granger causality Wald tests
>> +------------------------------------------------------------------+
>> Equation           Excluded    chi2     df Prob > chi2
>> --------------------------------------+---------------------------
>> D_trade        D.ind_share       13.282     4     0.010
>> D_trade        D.ser_share       18.164     4     0.001
>> D_trade                ALL       27.893     8     0.000
>> --------------------------------------+---------------------------
>> D_ind_share           D.trade   3.5946     4    0.464
>> D_ind_share        D.ser_share   150.71     4    0.000
>> D_ind_share                ALL   175.65     8    0.000
>> --------------------------------------+---------------------------
>> D_ser_share            D.trade   8.2083     4    0.084
>> D_ser_share        D.ind_share   41.003     4    0.000
>> D_ser_share                ALL   54.064     8    0.000
>> +------------------------------------------------------------------+
>>
>> I want chi2 and p-value of the first equation to appear together (or to be
>> reported with a single command) and likewise for the rest equations.
>> Typing
>> di [a,b] provides only only statistic value as a scalar.
>
> Nick Cox
>
>> It is a good idea to read the help when using any command.
>>
>> http://www.stata.com/help.cgi?vargranger documents that -vargranger-
>> leaves r-class results in its wake. After running -vargranger-
>>
>> mat results = r(gstats)
>>
>> will copy the set of statistics to a matrix, after which you can
>> display or use one or more elements as desired, e.g.
>>
>> di results[1,1]
>
> On 8 July 2013 10:04, Ioannis Bournakis <I.Bournakis@mdx.ac.uk> wrote:
>
>>> I estimate a VAR model for a panel of countries and after the var
>>> command,
>>> I perform a granger causality test using the vargranger command.
>>> My question is how to save the chi-square value of the vargranger
>>> command.
>>> When I type ereturn list, it gives me all the information associated with
>>> the var  estimation but not the chi2 value for vargranger.
>>>
>>> I want to save the results from vargranger and then report them in a
>>> table.
*
*   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/