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 at the end of May, and its replacement, statalist.org is already up and running.


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

Re: st: event study ado-file?


From   Nick Cox <njcoxstata@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: event study ado-file?
Date   Mon, 18 Feb 2013 22:18:29 +0000

No. Try it yourself with the auto data and -rep78-.

Nick

On Mon, Feb 18, 2013 at 10:12 PM, László Sándor <sandorl@gmail.com> wrote:
> Thanks again, Nick and Maarten. I'll go with -statsby-. One question
> still: Will it crash if there are subsets of the -by()- variables
> without observations (missing years)? It would be nice if I did not
> need a complicated set of nested loops to run through all meaningful
> combinations.
>
> Thanks!
>
> On Mon, Feb 18, 2013 at 10:52 AM, Maarten Buis <maartenlbuis@gmail.com> wrote:
>> There is no way we can know, as we don't have your dataset.
>>
>> Anyhow programming a new command that works quicker costs a
>> non-trivial amount of time too, so it becomes a tricky optimization
>> problem to find the quickest solution. Actually it is not that tricky;
>> I would say that in most situations you will loose time when
>> programming a new command. I program a lot, but if you add all the
>> time spent programming and subtracted the time gained from quicker
>> running programs I am sure I lost a considerable amount of time with
>> that strategy (I had fun while programming, though...).
>>
>> I would say, just try -statsby- out. If it is too slow you'll notice
>> that quick enough...
>>
>> -- Maarten
>>
>> On Mon, Feb 18, 2013 at 4:11 PM, László Sándor <sandorl@gmail.com> wrote:
>>> Thanks, Nick, as always.
>>>
>>> So you think -statsby- is fast enough for this. I hoped not to
>>> preserve my (big) data too much (= ever), and was lazy to check what
>>> statsby actually does.
>>>
>>> Thanks again!
>>>
>>> Laszlo
>>>
>>> On Mon, Feb 18, 2013 at 9:54 AM, Nick Cox <njcoxstata@gmail.com> wrote:
>>>> As you say, this is a standard kind of graph. I would write a do-file
>>>> separating out the data management from the graphs. I've found
>>>> -statsby- very useful for confidence intervals and there was an essay
>>>> to that effect in
>>>>
>>>> SJ-10-1 gr0045  . . . . . . . . . . . . . Speaking Stata: The statsby strategy
>>>>         . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  N. J. Cox
>>>>         Q1/10   SJ 10(1):143--151                                (no commands)
>>>>         demonstrates the use of statsby to prepare a reduced
>>>>         dataset for subsequent graphing
>>>>
>>>> I am not clear that you need think about this as requiring an ado-file.
>>>>
>>>> Nick
>>>>
>>>> On Mon, Feb 18, 2013 at 2:31 PM, László Sándor <sandorl@gmail.com> wrote:
>>>>
>>>>> Before I code up my own ado-file, I was wondering if there was
>>>>> something on this out there (for Stata 12 MP for Windows, if it
>>>>> matters).
>>>>>
>>>>> I need "standard" graphs of estimates with confidence interval bands
>>>>> for various periods. I am happy to define periods myself, so this
>>>>> would be an event study if I shift periods to be around 0 (time of
>>>>> treatment) and other periods show pre- and post-treatment treatment
>>>>> effects. Or I could also use this for a chronology of treatments: an
>>>>> outcome defined for all years (e.g. income in the given year) and
>>>>> plotting all these treatment effects for treatment occurring at
>>>>> different times.
>>>>>
>>>>> E.g. the first graph here, but with CI bands:
>>>>> http://obs.rc.fas.harvard.edu/chetty/value_added.html

*
*   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/


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