[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

From |
Melonie Sullivan <meloniebeth@yahoo.com> |

To |
statalist@hsphsun2.harvard.edu |

Subject |
Thank you Nick! (re: st: preserving missing values in collapse (sum)) |

Date |
Thu, 25 Oct 2007 08:44:54 -0700 (PDT) |

Thank you so much for your assistance (and your patience). With -fillin-, setting duration=0 when _fillin=1, and then -reshape-, I got exactly what I needed. Whew! Melonie --- n j cox <n.j.cox@durham.ac.uk> wrote: > This response y has not so far appeared in the > drama. > Where does it come from? The same dataset? > > Either way, I think you might make progress by > checking out -reshape-. > > Melonie Sullivan > > Okay, so far so good, thanks. But now how do I get > that information into a matrix of this form - one > line > for each youth: > > youthid y x1 x2 x3 x4 x5 x6 > 11 0 15 41 0 0 . 0 > 12 1 0 13 0 42 0 55 > > where y=dependent variable, x1=duration if group=1, > x2=duration if group=2, etc. If I take your > solution, > then generate x1, x2...., and do a -list- I still > get > a 6x6 matrix of x for each youth that looks like > this: > > youthid x1 x2 x3 x4 x5 x6 > 11 15 . . . . . > 11 . 41 . . . . > 11 . . . . . . > 11 . . . . . . > 11 . . . . . . > 11 . . . . . . > > < intermediate posts> > > > >> I have data on history of placements into > > different > > >> -group-s by -youthid-: there are multiple > > placement > > >> records for each youth. I need to create a > > variable > > >> equal to the sum of -duration- of all > placements > > into > > >> each -group- for each youth. -collapse (sum)- > > seems to be > > >> the appropriate procedure, but it treats > missing > > >> values as zeroes. This causes a problem if a > > youth has > > >> only one placement in a given group with > unknown > > >> duration. Example: > > >> > > >> > > >> youthid group duration > > >> 11 1 15 > > >> 11 1 . > > >> 11 2 31 > > >> 11 2 10 > > >> 11 5 . > > >> 12 2 5 > > >> 12 2 8 > > >> 12 4 42 > > >> 12 6 55 > > >> > > >> I create a duration variable for each group > > (-generate > > >> grp1dur = duration if group==1-, etc.) and > > -collapse > > >> (sum)- by -youthid- and I want to get this: > > >> > > >> youthid 11 12 > > >> grp1dur 15 0 > > >> grp2dur 41 13 > > >> grp3ddur 0 0 > > >> grp4dur 0 42 > > >> grp5dur . 0 > > >> grp6dur 0 55 > > >> > > >> But collapse gives me a zero on grp5dur for > youth > > #11, > > >> though youth #11 had placement in that group, > > albeit > > >> of an unknown duration. The other zeroes are > > correct; > > >> the youth had zero days in that placement > group. > > >> > > >> The problem has been addressed here before, > best > > in > > >> the following post by Nick Cox: > > >> > > >> > > > http://www.stata.com/statalist/archive/2004-07/msg00783.html > > >> > > >> However, this is not solving my particular > > problem, > > >> because my data essentially looks like a big > > stack of > > >> Nick's "toy datasets" -- one for each of 1800 > > youth in > > >> my data. So collapsing by (youthid) gives the > > same > > >> value of Nick's allmissing for each youth, > since > > the > > >> allmissing tags missing durations for groups > > within > > >> youths. > > * > * 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/ > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com * * 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/

**References**:**Re: Re: st: preserving missing values in collapse (sum)***From:*n j cox <n.j.cox@durham.ac.uk>

- Prev by Date:
**st: nl -choice between alternative parametrisations of sigmoid models** - Next by Date:
**Re: st: re: question about Mata programming** - Previous by thread:
**Re: Re: st: preserving missing values in collapse (sum)** - Next by thread:
**st: Quantile regression** - Index(es):

© Copyright 1996–2017 StataCorp LLC | Terms of use | Privacy | Contact us | What's new | Site index |