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: AW: AW: st: AW: Generating a dropout dummy variable?
From
"Martin Weiss" <[email protected]>
To
<[email protected]>
Subject
RE: AW: AW: st: AW: Generating a dropout dummy variable?
Date
Sun, 27 Jun 2010 23:37:36 +0200
<>
***********
//generate an indicator ==1 if is the last time i see the individual
bys id (year): gen byte lasttime=_n==_N
//Then i need to create the dropout indicator following the rule that if
//last year ==1 and graduated!=1 then dropout==1
gen byte dropout=lasttime & graduated!=1
***********
These two lines could be telescoped into one, but I wanted to show the process more comprehensively.
HTH
Martin
-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Ignacio Martinez
Sent: Sonntag, 27. Juni 2010 23:13
To: [email protected]
Subject: Re: AW: AW: st: AW: Generating a dropout dummy variable?
Now I think I finally know what I need to do.
First I need to generate an indicator ==1 if is the last time i see the
individual (for example for id=1 if the last time i see him is t=2002
then the indicator ==1 in 2002 and ==0 in the other periods)
Then i need to create the dropout indicator following the rule that if
last year ==1 and graduated!=1 then dropout==1
I think this is going to work, but I'm not sure how to code it.
Thanks for the help!!!
-Ignacio
On Sun, 2010-06-27 at 20:22 +0200, Martin Weiss wrote:
> <>
>
>
> "So the code should do something like: when is the first time I see the
> student. Do I see him after the first time. Was he graduated the last
> time i saw him."
>
>
>
> The answers to these questions are quite easily obtained, actually:
>
>
> *************
> //Create data
> clear*
> set obs 1000
> set seed 43550
>
> //id
> gen int id=_n
> //random start year
> gen startyear=1999+irecode(runiform(),0,.2,.5,.7,1)
>
> //expand to # of time periods
> expand 7
>
> bys id: gen int year=_n-1+startyear
> gen byte graduated= /*
> */ runiform()<0.3
>
> drop if runiform()<.3
>
> //throw out twice graduated guys
> bys id: egen mytotal=total(graduated)
> drop if mytotal>2 & !mi(mytotal)
> drop mytotal startyear
> drop if year==2009
>
> xtset id year
>
>
>
> //So the code should do something like:
> //when is the first time I see the student?
> bys id: egen firstyear=min(year)
>
> //Do I see him after the first time?
> by id: gen byte afterfirsttime=year>firstyear & /*
> */ !mi(year)
>
> //Was he graduated the last time I saw him?
> bys id (year): gen byte gradlasttime=(_n==_N)*graduated
>
> l in 1/21, sepby(id) noo ab(20)
> *************
>
>
>
> HTH
> Martin
>
>
> -----Ursprüngliche Nachricht-----
> Von: [email protected] [mailto:[email protected]] Im Auftrag von Ignacio Martinez
> Gesendet: Sonntag, 27. Juni 2010 19:54
> An: [email protected]
> Betreff: Re: AW: st: AW: Generating a dropout dummy variable?
>
> Thanks
> Now i understand better.
>
> I think I explain my self wrong the first time so the code is not doing
> exactly what I need.
>
> My data goes from 2000 to 2008. So is possible that a student enroll in
> any year between 2000 and 2008. If I'm understanding the code right I
> should change the first line to:
> bys id: egen myvar=total(inlist(year,2000,2008))
>
> Is this right?
>
> But now I'm not sure how should I change the second line.
>
> The problem is that some student were enroll for the first time in 2000
> some others in 2001, 2002, etc.
>
> So the code should do something like: when is the first time I see the
> student. Do I see him after the first time. Was he graduated the last
> time i saw him. Or something like that
>
> Thanks a lot for your help.
>
> -Ignacio
>
>
> On Sun, 2010-06-27 at 19:38 +0200, Martin Weiss wrote:
> > <>
> >
> > " About your code... I'm kind of lost on how you create the dropout
> > indicator. Could you explain a little that piece of code?"
> >
> >
> >
> >
> >
> > *************
> > //get "dropout" indicator
> >
> > //Do I see "id" in 2000 and 2001?
> > bys id: egen myvar=total(inlist(year,2000,2001))
> >
> > //How many times did I see "id" between 2002 and 2006?
> > bys id: egen myvar2=total(inrange(year,2002,2006))
> >
> > //dropout is 1 if I saw "id" in 2000 and 2001
> > //and did not see him again afterwards
> > //and, in 2001, his "graduated" was not 1
> > bys id: gen dropout=((myvar==2)*(myvar2==0))& /*
> > */ ((graduated!=1)*(year==2001))
> > *************
> >
> >
> > See this thread for the tricks re "conditions": http://www.stata.com/statalist/archive/2010-06/msg00985.html
> >
> >
> >
> > HTH
> > Martin
> >
> >
> > -----Ursprüngliche Nachricht-----
> > Von: [email protected] [mailto:[email protected]] Im Auftrag von Ignacio Martinez
> > Gesendet: Sonntag, 27. Juni 2010 19:33
> > An: [email protected]
> > Betreff: Re: st: AW: Generating a dropout dummy variable?
> >
> > About the second part:
> > What I was trying to say is that if i see a student in the year 2000 and
> > i don't see him in 2001 because he is taking a break but in the year
> > 2005 he is back in my sample I don't want to mark him as drop out but I
> > want to generate a variable that says that he took a 4 year break.
> >
> > About your code... I'm kind of lost on how you create the dropout
> > indicator. Could you explain a little that piece of code?
> >
> > Thanks a lot
> >
> > -Ignacio
> >
> >
> > On Sun, 2010-06-27 at 19:01 +0200, Martin Weiss wrote:
> > > <>
> > >
> > > The first part could be:
> > >
> > >
> > > *************
> > >
> > > //Create data
> > > clear*
> > > set obs 1000
> > > set seed 43550
> > >
> > > //id
> > > gen int id=_n
> > >
> > > //expand to # of time periods
> > > expand 7
> > >
> > > bys id: gen int year=_n+1999
> > > gen byte graduated= /*
> > > */ cond(runiform()<0.3 & year==2001,1,0)
> > >
> > > drop if runiform()<.7
> > >
> > > xtset id year
> > >
> > > //get "dropout" indicator
> > > bys id: egen myvar=total(inlist(year,2000,2001))
> > > bys id: egen myvar2=total(inrange(year,2002,2006))
> > > bys id: gen dropout=((myvar==2)*(myvar2==0))& /*
> > > */ ((graduated!=1)*(year==2001))
> > > drop myvar*
> > >
> > > //take a look
> > > bys id: egen anydropout=max(dropout)
> > > l if anydropout, sepby(id)
> > > *************
> > >
> > >
> > > I do not understand the second part. Why would dropout be 1 in 2000 if you saw the guy in 2000?
> > >
> > >
> > > HTH
> > > Martin
> > >
> > > -----Ursprüngliche Nachricht-----
> > > Von: [email protected] [mailto:[email protected]] Im Auftrag von Ignacio Martinez
> > > Gesendet: Sonntag, 27. Juni 2010 15:54
> > > An: [email protected]
> > > Betreff: st: Generating a dropout dummy variable?
> > >
> > > Hi,
> > >
> > >
> > > My panel has the following variables: Year (from 2000 to 2008), ID,
> > > Graduated (an indicator =1 if the student graduated that year) .
> > >
> > > I want to generate a dropout variable. If I see ID==1111 in year==2000
> > > and 2001 but not after that and graduated !=1 in 2001 I want dropout==1
> > > in 2001 . The only other detail is that if a see someone in 2000 and
> > > then I see him again in 2006 with graduated=1 I don't want dropout =1 in
> > > 2000 I want other variable that is break=5 (that student took 5 year of
> > > break)
> > >
> > >
> > > Thanks for the help
> > >
> > >
> > > -Ignacio
> > >
> > >
> > > *
> > > * 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/
>
>
> *
> * 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/