[Babase] MEMBERS comparison - Karl question
Karl O. Pinc
babase@www.eco.princeton.edu
Tue, 27 Sep 2005 16:19:33 +0000
On 09/27/2005 10:11:47 AM, Catherine Markham wrote:
> Karl,
>
> I have a question for you about MEMBERS interpolation regarding death
> dates. The documentation states the following:
>
> "... 14 day interpolation limit applies when the individual is dead.
> When there are no absences after the last census and there are more
> than 14 days between the last census and the Statdate the individual
> is placed in the unknown group from the 15th day through the day of
> death."
>
> That makes sense to me, but I don't think I had also consciously
> considered the other way an individual can be placed in Group 9 on
> the death date - it didn't occur to me until looking into the records
> for ATO.
>
> ATO is a female censused in Weaver's group as present on 25 May 2001
> and absent on the next group census day on 29 May 2001 (this absence
> was entered in CENSUS). Her death date is 27 May 2001. The old
> MEMBERS placed her in Weaver's group on this day, the new MEMBERS
> places her in the unknown group on this day. The new MEMBERS is
> looking at the absence on 29 May 2001 no differently than it would
> for, say, any male who disappears from a study group - it essentially
> considers ATO's group to be 9 on the 29th. So it turns out that the
> 27th is a perfect mid-point between presence in Weaver's and
> placement in Group 9. This makes the day's group placement more or
> less a coin toss - it could have been Weaver's or it could have been
> (and in this case it was) Group 9.
>
> Did I follow the logic of that correctly?
Yes. Exactly. In fact, this caught me too when I went to review
the members comparison. IIRC I'd explicitly put in an example
like this in the documentation, but even so I'd forgotten again
until it came up in real data.
Karl <kop@meme.com>
Free Software: "You don't pay back, you pay forward."
-- Robert A. Heinlein