Maarten and Alan, thank you immensely for your help and for reading
through my question!!!
---begin---
sysuse auto
qreg price weight length foreign
adjust, by(foreign)
predict [test], xb
sum [test], d
// note that the mean as reported by the sum command is the same as the
value reported for the adjust command.
My question, however, is whether I should report the 50% (median) from the sum
command, or the mean from the sum command, IF I want to report the “adjusted
median value”.
by foreign, sort: ci [test]
---end---
What I want to do:
Run quantile regression model for price, controlled for
weight, length and foreign
Determine what the median price is, by foreign.
Where I am confused:
Running adjust after the qreg command will provide adjusted
values based on the qreg command (in other words, this will provide predicted
median prices).
But, the adjust command (i.e. ‘adjust, by(foreign)’), will
provide an output that states the “mean of the predicted medians from the qreg
model”.
What I don’t quite get is that I should be reporting the “mean
of the predicted medians”? Or should I
report the “median of the predicted median”?
I should note that the mean and median of the “predicted medians” are
very close in value, as should be expected as the predicted distribution is not
skewed.
// note that the mean as reported by the sum command is the same as the
value reported for the adjust command.
My question, however, is whether I should report the 50% (median) from the sum
command, or the mean from the sum command, IF I want to report the “adjusted
median value”.
This will be my final attempt as I really appreciate all
your help and don’t want to bother you anymore!
Thanks!
----------------------------------------
> Date: Wed, 14 Jan 2009 18:59:47 +0000
> From: [email protected]
> Subject: RE: st: How to interpret and summarize adjusted medians
> To: [email protected]
>
> --- I M wrote:
>> But I also wanted to verify that even though the commmand is qreg
>> (for quantile regression), I should use the mean to report the
>> statistics, and not the 50th percentile?
>
> both the mean and the median are reasonable measures of central
> tendency, which one you choose is up to you. You can get the adjusted
> median prices while keeping weight and length at their median rather
> than mean values in the following way:
>
> *------------ begin example ----------
> sysuse auto
> qreg price weight length foreign
> sum weight if e(sample), detail
> local med_w = r(p50)
> sum weight if e(sample), detail
> local med_l = r(p50)
>
> adjust weight=`med_w' length=`med_l', by(foreign)
> *----------- end example ----------------
> (For more on how to use examples I sent to the Statalist, see
> http://home.fsw.vu.nl/m.buis/stata/exampleFAQ.html )
>
> Hope this helps,
> Maarten
>
>
>
> -----------------------------------------
> Maarten L. Buis
> Department of Social Research Methodology
> Vrije Universiteit Amsterdam
> Boelelaan 1081
> 1081 HV Amsterdam
> The Netherlands
>
> visiting address:
> Buitenveldertselaan 3 (Metropolitan), room N515
>
> +31 20 5986715
>
> http://home.fsw.vu.nl/m.buis/
> -----------------------------------------
>
>
>
> *
> * 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/
