Alexandra de Montrichard
>
> I am somewhat new to stata and I need some help with a
> problem I can't
> seem to resolve.
>
> I am working with a database of students for whom I have quarterly
> grades for 5 years (1995-2000). I also have the start &
> completion date
> for a class that all students took a one point during those
> five year. I
> need to figure out how many quarters of grades I have for
> each student
> before & after they took this class & how many are
> consecutive quarters
> of grades. I've created a 0-1 variable for each quarter (0 if the
> student did not get any grades that quarter, 1 if they received a
> grade), converted the class start & exit date into quarter
> format & used
> egen to count of the number of quarters in which each
> student received
> grades
> . egen byte gradenbr = neqany (grade951-grade004), v (1)
> However, I now need to figure out a way of counting the number of
> quarters with grades both before & after the class as well
> as a way to
> know how many consecutive quarters with grade I have after
> completion of
> the class. I though I could use egen combined with if to create a
> gnbrpost variable of the number of quarters with grades
> after completion
> of the class:
> . egen byte gnbrpost = neqany (grade951-grade004) if (exitdq==
> q(1995q1)), v (1)
> . egen byte gnbrpost = neqany (grade952-grade004) if (exitdq==
> q(1995q2)), v (1)
> . egen byte gnbrpost = neqany (grade953-grade004) if (exitdq==
> q(1995q3)), v (1) etc.....
>
> but I can't figure out how to replace the gnbrpost variable
> each time I
> run a new egen calculation plus I'm sure there must be a
> better way of
> doing this, which would also allow me to know how many consecutive
> quarters of grades I have.
In one sentence: -reshape- to -long- and convert your structure
into a panel data structure. Then use e.g. -tsspell-. from SSC.
Your columns
grade951-grade004
should be mapped into one variable -grade-, another
-quarter- and another -student-.
. reshape long grade, i(student) j(quarter) string
-quarter- as a string variable is not handy for time series work.
. gen year = real( cond(substr(quarter,1,1)=="9",
"19" + substr(quarter,1,2),
"20" + substr(quarter,1,2)) )
. gen qtr = real(substr(quarter,-1,1))
. gen q = yq(year, qtr)
. format q %tqq_Cy
. tsset student q
Now you need to divide each student's record into
before, during and after taking a course. One
way to do it is using -tsspell- from SSC. There
are several examples in its help.
Nick
[email protected]
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
