[Babase] REPSTATS and the unborn

Daphne Onderdonk babase@www.eco.princeton.edu
Mon, 03 May 2004 11:32:21 -0400


Hi all,

This issue is the exact issue I brought up on my email of April 22 (it
wasn't to the whole Babase list, Karl, because it had to do with data we
were pulling together for Joan Silk).  Yes, I discovered that females who
were pregant but had not yet given birth did not have rows in Repstats for
their pregnancies.  I had said then that I didn't really see this as a
problem, just something to be aware of.

Karl, I believe the same is NOT true for lactating females.  What I found
was that lactating females were in there up till the end of time period of
the table, even if their lactating state had not ended.  Correct me if I'm
wrong about that.  This means that days remaining in state isn't meaningful
for lactating females, but days into state is fine.  This seems fine to me,
since lactation is such a relatively prolonged state - if we waited until
it was over to put the lines in Repstas, like for pregnancies, we'd be
missing a lot more data.  Again, it would just be something to be aware of.

About the abortions in Biograph, all of them that are in there do have
birth dates (everyone without an sname has a birth date, and the ones
without snames are the same as the ones with equal birth and statdates).
But I was thinking recently about females who died while they were pregnant
and making sure those pregnancies were getting into Biograph.  I have been
aware of it recently and have been checking for this when I do the
demography update, but I think it would be good to check in the past.
Maybe the best way would be to get all the females who died as adults and
see if they were pregnant when they died, and then see if those pregnancies
are in Biograph.  Or maybe look at the pregnancies in the Pregs table that
don't have resume cycles.  Let me know if this is something I should do
here, or if that's something Steph wants to work on.

Karl, if a female dies while she's pregnant, does she then get lines in
Repstats for her pregnancy (from conception till death)?  I can check this,
but I thought you might know off hand.  If so, then it might make checking
the above easier.

Daphne


"Karl O. Pinc" wrote:

> On 2004.05.01 15:34 Jeanne Altmann wrote:
> > We definitely need to check abortions (including fetal death when mom
> > dies) to be sure we get birth dates).
>
> I imagine you already have birthdates, as this is the only record of
> the date
> of the abortion.
>
> > I don't recall why we made this decision, but we should consider all
> > the implications now.  This decision does mean that repstat will have
> > the wrong stat for many females for approximately the most recent
> > 6mos of updates.  This isn't good, so I'm inclined to change this
> > unless there's a good reason not to.
>
> It's not that you have the wrong stat, you have no row at all and
> so no stat.  This seems reasonable to me.  If you did have a row
> it would have the wrong number for 'days remaining in state'.
> With no row, using SQL, you get no results for 'incomplete'
> pregnancies.  This _sounds_ ok to me.
>
> I'm going to try this with some test data to be sure what I think
> is happening is happening.
>
> Note that there's lots of wierd data in REPSTATS, as the system
> dosen't know that two sequential cycles years apart mean that
> data is missing, not that one cycle is especially long.
> It _does_ know that data is missing when there's a missing tdate
> or ddate in a cycles row, and then rows don't appear.  I've
> been going over this with Steph.  It seems that the first step
> when using this data is to eliminate nonsensical rows.
> Of course, I can make the program do _anything_, if we have
> good things for it to do.  I'm kind of opting for simple and
> understandable, and what we've got now seems to meet that
> criteria, although I'm biased because I've been working with it.
>
> >
> > jeanne
> >
> > At 01:16 PM 5/1/2004 -0500, you wrote:
> >> Hi,
> >>
> >> Looking over the program that populates REPSTATS
> >> I notice that when there's a pregnancy you don't get
> >> P rows in repstats until the offspring has a birthdate.
> >> I belive this was intentional, especially as there
> >> is a 'days remaining' in state column which couldn't
> >> be filled out until the length of the state is known,
> >> but there are implications.  I figured I'd write as
> >> we're starting to work with this data.
> >>
> >> All abortions must have rows in BIOGRAPH with a Birth date.
> >>
> >> The as-yet-unborn don't show in REPSTATS.
> >>
> >> Karl <kop@meme.com>
> >> Free Software:  "You don't pay back, you pay forward."
> >>                  -- Robert A. Heinlein
> >> _______________________________________________
> >> Babase mailing list
> >> Babase@www.eco.princeton.edu
> >> http://www.eco.princeton.edu/mailman/listinfo/babase
> >
> > _______________________________________________
> > Babase mailing list
> > Babase@www.eco.princeton.edu
> > http://www.eco.princeton.edu/mailman/listinfo/babase
>
> Karl <kop@meme.com>
> Free Software:  "You don't pay back, you pay forward."
>                    -- Robert A. Heinlein
>
> _______________________________________________
> Babase mailing list
> Babase@www.eco.princeton.edu
> http://www.eco.princeton.edu/mailman/listinfo/babase