[Babase] Psion Miscoding Issue - JEANNE PLEASE READ
Susan Alberts
babase@www.eco.princeton.edu
Tue, 12 Apr 2005 15:30:04 -0500
Thanks very much Catherine.
A few observations:
A. Regarding age of juvs that get sampled, I dont
think we want a lower age limit. In the list of
"errors" catherine sent (samples that took place
before 6 months of age), it looks as though most
or all cases involve animals being started a few
weeks or a month or two early. My hope is that
the field team are using their judgement in
deciding when to start an infant -- the point of
these samples has always been that infants and
juvs need to get focal sampled once they start
being off their mothers a lot so that we don't
lose their IDs.It looks to me that this is what
is happening -- some animals are consistently
appearing in the sample list early (WUA for
instance). I think we should just ask the field
team about this and if they sometimes start a kid
a bit early, we should let it happen. I can put
this on my list of things to talk to them about
this summer.
B. Regarding upper limits, I am also not sure
that we want to be very strict about this. For
instance, looking at Catherine's list of "adults
in jpsamps", they fall into 3 categories:
1. Snames belong to adult females who are 3 or
more years into adulthood. These are clearly
errors and need to be tossed, but are a minority
in this list (roughly 1/4).
2. Snames belong to young males who were sampled
as juveniles for a few months after their
official TE dates (ie they were sampled as
subadults). Because enlargement is difficult to
see and we only assess it once a month, this is a
conservative error. These constitute most of the
"errors" in this category and we don't want to
toss these.
3. Snames that belong to young females (MBE, WRA,
VEX, VIG, KIW) who were sampled as juveniles for
several months after menarche. Again, I am not
sure this is a big problem.
Here is ia possible rule of thumb: if the animal
is sampled as a juv within one year of attaining
Testicular enlargement or menarche, we just
accept the sample. Otherwise toss it.
C. In response to Catherine's question about
whether we still want to discard samples in the
three categories she lists below, I would say YES
for the first two types but NO for the third type
for 75% of them. Toss only thoses where the
sample occurred more than a year after maturity.
D. Clearly the error problem is not as great as
it seemed once we accept some post-maturity
animals being sampled as juvs for a while.
E. Regarding dates that the errors occurred. Of
the ones that we clearly need to consider errors
there is not a clear pattern that these happened
in the early days. It looks ongoing. I think we
need to compile a summary and I talk about it
with the team. Especially non-adult females
sampled as adult females and males sampled as
adult females. We need to get some insight into
how to prevent this.
Susan
>Hi,
>
>I wanted to follow-up with the Psion miscoding
>issues Daphne mentioned at the BaBase meeting
>last Friday. Three types of errors were
>discussed:
>
>(1) male samples in the fpsamps table (8 out of 15,269 samples)
>(2) non-adult female samples in the fpsamps table (56 out of 15,269 samples)
>(3) adults in the jpsamps table (96 out of 8,100 samples)
>
>It occurred to me this afternoon that a fourth
>possible error would be young baboons (less than
>6 months) in the the jpsamps table. A total of
>81 out of 8,100 samples had this problem. Note
>that this count does not include cases where
>observation date < (birth + 180 days) if the
>individual turned 180 days at some point during
>the observation month.
>
>Attached are two documents. The Word document
>provides the queries that I did following
>Daphne's notes/instructions to determine the
>error counts. The Excel worksheet lists the
>errors themselves (it contains four worksheets,
>one for each error type). The observation date
>for each record is listed if Susan and Jeanne
>are still interested in seeing when these
>mistakes were made (if I remember right, you
>wanted to double check these dates before we
>moved forward with the clean-up).
>
>Regarding the first three error types, we had
>decided to remove these records from BaBase.
>Does that still sound like a good plan to
>everyone? As for the fourth error type
>(youngsters less than 6 months in jpsamps),
>Jeanne pointed out that this is a different sort
>of error and not necessarily another case where
>the records should be deleted. Any suggestions
>of how to handle or do you simply want to leave
>these records in the database?
>
>Thanks, everyone.
>
>Catherine
>
>P.S. I know the attachments won't be archived
>on the listserv, but couldn't think of a way to
>get around it this time with the Excel
>spreadsheets. At least we'll have the email
>exchange well documented and I'll eventually
>include details of any changed/deleted records
>in the BaBase correction log.
>
>
>Attachment converted: cisticola:Psion Miscoding
>Issues.xls (XLS4/«IC») (0007ADFF)
>Attachment converted: cisticola:Psion Miscoding
>Issues.doc (WDBN/«IC») (0007AE00)
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Susan Alberts, Associate Professor
Department of Biology, Duke University, Box 90338, Durham NC 27708
phone 919-660-7272 fax 919-660-7293