[babase] Morris (MOR) and interpolation rules in members table + incomplete censuses!

Karl O. Pinc kop at meme.com
Fri Aug 21 11:18:21 EDT 2009


On 08/21/2009 08:58:20 AM, Niki Learn wrote:
> On 08/20/2009 04:04:42 PM, Susan Alberts wrote:
> > 
> > > On a related question, Jeanne and Susan, we have in this update a 
> 
> > > lot of incomplete census data.  ....
> > > In the case of current data generated by Courtney (and probably  
> > > older data generated by others), this means that sometimes males 
> > who  
> > > move around a lot are verified as absent on an incomplete census 
> > day
> > > though they had been present on the previous full census day (and 
> 
> > > may be either present or absent on the next full census day). 
> With 
> > > the way uploads work I cannot enter such incomplete census data 
> > with
> > > the full census data because every individual listed in the sheet 
> 
> > > must be marked either present or absent and we don't have that
> data 
> > > for every individual on those dates.  I could, if desired, either 
> 
> > > use a demog note to mark them as present with an unknown group 
> > (9.0)
> 
> > This seems like a good option in these cases. But as you noted  
> > earlier, we currently don't use "not here" info in demog notes.
> Would 
> > doing so require any structural or any other changes to data entry
> or 
> > babase?
> 
> The documentation does not say you can't,
> and it implies that you can so I assume you can.
> If you try it and it does not work we can change
> something.  Be sure to set the Cen column to false
> so the CENUSU row is deleted if the DEMOG row is deleted.
> 
> Niki writes:
> Well, I could do that, yes, but it could cause trouble.  For example 
> I
> was
> just looking at Liberty toward the end of yesterday.  He was in Omo's
> group
> on the last census in April (on the 29th) and then did not show up in
> the
> group in May so I was thinking I would have to enter a zero for him
> somehow
> for at least 1 May, which was the first census day of May, to make
> sure that
> he did not get interpolated in the group in May.  But then I realized
> that I
> didn't have to because he showed up in Nyayo's on 2 May.  According 
> to
> the
> interpolation rules he would then be interpolated as present with
> Nyayo's
> group on 1 May anyway.  And from what I read it doesn't matter if he
> is
> marked absent in Omo's on 1 May, he would still be interpolated as
> present
> in Nyayo's - but I'm not sure that would be the case if we marked him
> as
> present with group 9 - wouldn't he then be listed as in group 9 on 1
> May
> rather than in Nyayo's (which may well have been where he was)?

I was thinking you'd mark him absent from Omo's group, not present
in the unknown group.

Whether it "doesn't matter" is a matter of knowing the interpolation
rules, which we might even change someday.   It's better not to have
to think about it and have the data in place that will do the right
thing no matter what.

> not
> censused idea seems safer to me, plus it eliminates the need to use
> demography notes to mark baboons present during incomplete censuses,
> which
> would also be handy.

The "not censused idea" still means entering an absence for the one
census day in the month that you would omit the "not censused code"
from the census sheet.  In the case of Liberty and Omo's group
you'd mark him absent ("O") on the first of May and the rest of the 
census days in the month you'd mark him "N".  He'd then have an
absence on the first of the month and no census records for
the rest of the month for that group.  Come to think of it this is
not much an improvement in data entry.   You'd expend the same
amount of work marking him absent for all the census days of the month,
and this could result in a more accurate interpolation if, for
example he moved back into Omo's group on the 1st of July.
(We have not only a "end of the month" problem but also a
"beginning of the month" problem.  Right?)

Do we still want an "N" or should we just enter absent, O, instead?
> 
> Karl, I have no idea what you mean by setting the census column to
> false...
> Is this something I should know?

Probably.  See CENSUS.Cen:
https://papio.biology.duke.edu/babase_system_html/ch03s13.html#cen

Note that when there's possibility of
confusion I try to follow the conventions of the documentation
in writing these emails.
Table names in upper case.  Column names beginning with
a capital letter.  So when you see Cen, you know it's a
column name.

> 
> 
> _______________________________________________
> Babase mailing list
> Babase at www.eco.princeton.edu
> http://www.eco.princeton.edu/mailman/listinfo/babase
> 
> 




Karl <kop at meme.com>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein




More information about the Babase mailing list