Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: RE: Quantile Regression Confidence Intervals
From
Nick Cox <[email protected]>
To
[email protected]
Subject
Re: st: RE: Quantile Regression Confidence Intervals
Date
Thu, 14 Apr 2011 21:55:49 +0100
I'd work from the manual entry, not accessible to me at this moment.
The main point is that -qreg- is based on quite a different idea of
regression and a different way of estimating it.
Nick [not a Professor]
On Thu, Apr 14, 2011 at 7:07 PM, Jeff <[email protected]> wrote:
> In Stata 10:
>
> I found two syntax formulas to produce confidence intervals around an
> ols regression trend line. I would like to know if these two sytax
> formulas work for a quantile regression (median or any of the other
> quantiles.)
>
> First:
> regress price sf
> predict price
> predict se, stdp
> display invttail (30, .05/2)
>
> [Where 30 = degrees of freedom for computing the t value - say t=2]
>
> generate ll = price - 2*se
> generate ul = price + 2*se
>
> I got the above syntax formula from Statistics with Stata by Hamilton
>
> Second:
> regress price sf
> local level = (100-95)200
> generate ll = price - invttail(e(df_r), `level')*se
> generate ul = price + invttail(e(df_r), `level')*se
>
> I got this formula from Prof. Cox.
>
> Question will each of these syntax formulas work with a quantile
> regression on any of the quantiles?
> Are the degrees of freedom different between an ols regression and a
> quantile regression for the different quantiles?
>
*
* 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/