Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: RE: bootstrap gives errors with -rd-


From   Austin Nichols <austinnichols@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: RE: bootstrap gives errors with -rd-
Date   Mon, 14 Nov 2011 10:23:25 -0500

Darjusch D-Ta <altruist81@gmx.de> :
Note that I did not use the n option in my example code, and you
should not. Your first example of code is clearly wrong since it does
not include the assignment variable.  Did you want to reestimate the
optimal bandwidth on each resample, or use the raw data to estimate
optimal bandwidth and apply that to each resample?  My example does
the first.

On Sun, Nov 13, 2011 at 6:08 AM, D-Ta <altruist81@gmx.de> wrote:
> Yes, I use the latest version of -rd-. I also use the right syntax, i.e.
> bootstrap, reps(1999): rd outcome Z, n(50)
>
>
>
> Am 13.11.2011 05:03, schrieb Austin Nichols:
>>
>> D-Ta<altruist81@gmx.de>
>>
>> The help file, refs, and prior posts make clear that the assignment
>> variable Z should always appear last in the variable list:
>>
>> bootstrap, rep(200) : rd Y Z
>>
>> Are you using the latest version?
>>
>> ssc inst rd, replace
>>
>> On Sat, Nov 12, 2011 at 12:43 PM, D-Ta<altruist81@gmx.de>  wrote:
>>>
>>> sorry... -rd- stands for regression discontinuity and is a user written
>>> command by Austin Nichols.
>>>
>>> Darjusch
>>>
>>> Am 11.11.2011 20:09, schrieb Nick Cox:
>>>>
>>>> ... and -rd- is ?
>>>>
>>>> Nick
>>>> n.j.cox@durham.ac.uk
>>>>
>>>> D-Ta
>>>>
>>>> I have the following problem.
>>>>
>>>> I use the -rd- command to estimate a local effect of  (a treatment) X on
>>>> outcomes Y (several ouotcomes, some binary and some continous) at Z=0 in
>>>> the following way:
>>>>
>>>> bootstrap, rep(200) : rd Y X, n(50)
>>>>
>>>> where X=1 if Z>=0 and X=0 if Z<0.
>>>>
>>>> Z takes the values: -12 , -11.5, -11, .... ,-0.5, 0, 0.5, 1, ...,12
>>>> Around the cut-off value I have 15-20 observations in each category of
>>>> Z.
>>>>
>>>> When I run the command I get a lot of bootstrap replications that fail
>>>> (red X). This is more so for binary outcomes! I also tried to use
>>>> -bootstrap, strata(Z) rep(200)-  but the problem remains.
>>>>
>>>> Can someone explain what I do wrong?

*
*   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/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index