If -Maxy- is ever lower than -Miny-, then this
should be fixed somehow, _not_ worked around,
as it could cause many other problems too.
The appropriate check is either
assert Maxy > Miny
or
assert Maxy >= Miny
depending on what is appropriate.
Nick
[email protected]
Austin Nichols
> Nasreen Khan --
> Ulrich Kohler's suggestion
> gen ryear = Miny+int(uniform()*(Maxy-Miny+1))
> and my suggestion
> gen ryear=Miny+round(uniform()*(Maxy-Miny+1)-.5)
> are equivalent except in some cases where, contrary to your belief
> about the data, Miny>Maxy. In these cases, they both fail to give the
> desired result.
>
> Even better might be the more robust:
> gen ryear = min(Miny,Maxy)+int(uniform()*(abs(Maxy-Miny)+1))
> for this reason:
> clear
> range Miny -10000 10000 20001
> set seed 7777
> gen Maxy=Miny+ int(uniform()*13)-3
> set seed 12345
> gen ryear1=Miny+round(uniform()*(Maxy-Miny+1)-.5)
> set seed 12345
> gen ryear2=Miny+int(uniform()*(Maxy-Miny+1))
> set seed 12345
> gen ryear3=min(Miny,Maxy)+int(uniform()*(abs(Maxy-Miny)+1))
> gen dy=max(Max,Min)-min(Max,Min)+1
> gen r1=(ryear1-min(Maxy,Miny))/abs(Maxy-Miny)
> gen r2=(ryear2-min(Maxy,Miny))/abs(Maxy-Miny)
> gen r3=(ryear3-min(Maxy,Miny))/abs(Maxy-Miny)
> tab r1 if dy==3
> tab r2 if dy==3
> tab r3 if dy==3
*
* 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/
