Christopher W. Ryan
> > Thanks. I must still not understand how -pctile- and
> > -xtile- work. This code
> > gave me sort of a U-shaped distribution of grades:
> > highest-scoring students
> > got fail, and so did the lowest-scoring. In the middle
> > were honors, high pass,
> > and pass.
> >
> > I tried this (adding in a "conditional" grade which I
> > forgot to mention) :
> >
> > use "C:\quarantine\studentprogress.dta"
> > collapse (sum) pointsearned, by(studentid)
> > gsort -pointsearned
> > list
> > gen str10 finalgrade="pass"
> > replace finalgrade="honors" if _n<=(.15*_N)
> > replace finalgrade="high pass" if _n>(.15*_N) & _n<=(.25*_N)
> > replace finalgrade="conditional" if pointsearned<300 &
> > pointsearned>199
> > replace finalgrade="fail" if pointsearned<200
> > list
> >
> > Clumsier, but it seemed to work.
>
> Dangerous code.
>
> This could give students different final grades
> even though they got the same points earned,
> because you could be splitting ties quite
> arbitrarily!
To show the point, I graded those cars:
sysuse auto, clear
gsort -mpg
gen str10 finalgrade="pass"
replace finalgrade="honors" if _n<=(.15*_N)
replace finalgrade="high pass" if _n>(.15*_N) & _n<=(.25*_N)
replace finalgrade="conditional" if mpg<15 & mpg>13
replace finalgrade="fail" if mpg<13
The last two lines are immaterial, so long as the thresholds lie below
the upper quartile.
This was the first and only example I tried, and there is
one "student" (car) with 25 mpg who should be filing an appeal.
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/
