Bookmark and Share

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: Dropping observations and creating a balanced panel


From   Matthew Krauchunas <[email protected]>
To   [email protected]
Subject   Re: st: Dropping observations and creating a balanced panel
Date   Wed, 11 Aug 2010 11:11:54 -0400

Eric,

Please do not get me wrong.  I appreciate your help and just mentioned
the minor typo in case someone comes across this problem later and
wants to follow the solution you and Martin provided.

Thanks again.

V/R,
Matt

On Wed, Jul 28, 2010 at 10:10 AM, Eric Booth <[email protected]> wrote:
> <>
> Yeah, it was a typo, I meant to type "firstyear" instead of "year" in that last line.
>
> Also, you could also use -fillin- (from SSC) to create balanced panels after dropping
> panels that are problematic (start after 2003 & with closure | chow).
>
> cap which fillin
> if _rc ssc install fillin, replace
> fillin oshpd_fac_no year
>
> ~ Eric
>
> ________________________________________
> From: [email protected] [[email protected]] on behalf of Martin Weiss [[email protected]]
> Sent: Wednesday, July 28, 2010 8:48 AM
> To: [email protected]
> Subject: AW: st: Dropping observations and creating a balanced panel
>
> <>
>
> "I am off to try and figure out the duplicates or some similar command."
>
>
>
> You can find the problematic observations thus:
>
> *************
> sysuse nlswork,clear
> bys idcode: gen int number=_N
> su number, mean
> cou if number!=r(max)
> *************
>
>
>
> HTH
> Martin
>
> -----Ursprüngliche Nachricht-----
> Von: [email protected]
> [mailto:[email protected]] Im Auftrag von Matthew
> Krauchunas
> Gesendet: Mittwoch, 28. Juli 2010 15:27
> An: [email protected]
> Betreff: Re: st: Dropping observations and creating a balanced panel
>
> Hi Eric,
>
> I had to tweak the code a little bit, but you got me 95% of the way there!
>
> drop if firstyear>2002 & !mi(year)
>
> instead of drop if year>2003 & !mi(year)
>
> My only problem now is that I would expect a balanced panel (i.e., one
> divisible by 6), but that is not the case as I have 3,971 observations
> remaining.  Also, it will not let me xtset the data.  Instead it gives
> me a "repeated time values within panel error."  I am off to try and
> figure out the duplicates or some similar command.
>
> Thanks,
> Matt
>
> On Tue, Jul 27, 2010 at 9:53 PM, Eric Booth <[email protected]> wrote:
>> <>
>>
>> Note:  I added facility 206010741 to your example data:
>> *********!
>> clear
>> inp double(oshpd_fac_no)    year    chow    closure
>> 206010734    2002    0    0
>> 206010734    2003    0    0
>> 206010734    2004    0    0
>> 206010734    2006    0    0
>> 206010734    2006    1    0
>> 206010734    2008    0    0
>> 206010744    2002    0    0
>> 206010744    2003    0    0
>> 206010744    2004    0    0
>> 206010744    2007    0    0
>> 206010744    2008    0    0
>> 206010747    2002    0    0
>> 206010747    2003    0    0
>> 206010747    2004    0    0
>> 206010747    2006    0    0
>> 206010747    2007    0    0
>> 206010747    2008    0    0
>> 206010741    2007    0    0
>> 206010741    2004    0    0
>> 206010741    2005    0    0
>> 206010741    2006    0    0
>> end
>>
>> ** 1. drop all the observations in each of the panels if there was a chow
> or closure**
>> bys oshpd_fac_no: egen any_chow = max(chow)
>> bys oshpd_fac_no: egen any_closure = max(closure)
>> drop if 1==any_closure | 1==any_chow
>>
>> **2.  drop if I have all nursing homes that opened after 2003 **
>> bys oshpd_fac_no: egen firstyear = min(year)
>> drop if year>2003 & !mi(year)
>> *********!
>>
>> ~ Eric
>>
>> __
>> Eric A. Booth
>> Public Policy Research Institute
>> Texas A&M University
>> [email protected]
>> Office: +979.845.6754
>>
>>
>> On Jul 27, 2010, at 4:52 PM, Matthew Krauchunas wrote:
>>
>>> Hello,
>>>
>>> I am studying nursing homes (2002-2004 & 2006-2008) and I have a two
>>> part problem that I need to solve in order to create a balanced panel.
>>> First, I have nursing homes that have closed or changed ownership
>>> (chow) during my study period.  I need to  drop all the observations
>>> in each of the panels (panel identifier is the oshpd_fac_no) if there
>>> was a chow or closure?  Second, I have nursing homes that opened after
>>> 2003 and therefore I would not have all the observations for these
>>> panels.  How can I identify and remove these panels as well?  Does
>>> anyone have any ideas?
>>>
>>>
>>> oshpd_fac_no    year    chow    closure
>>> 206010734    2002    0    0
>>> 206010734    2003    0    0
>>> 206010734    2004    0    0
>>> 206010734    2006    0    0
>>> 206010734    2006    1    0
>>> 206010734    2008    0    0
>>> 206010744    2002    0    0
>>> 206010744    2003    0    0
>>> 206010744    2004    0    0
>>> 206010744    2007    0    0
>>> 206010744    2008    0    0
>>> 206010747    2002    0    0
>>> 206010747    2003    0    0
>>> 206010747    2004    0    0
>>> 206010747    2006    0    0
>>> 206010747    2007    0    0
>>> 206010747    2008    0    0
>>>
>>> Thank you,
>>> Matt
>>> *
>>
>>
>>
>> *
>> *   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/
>
>
> *
> *   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/
>

*
*   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 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index