[Babase] Ranks data errors
Susan Alberts
babase@www.eco.princeton.edu
Thu, 3 Feb 2005 17:44:36 -0500
Leah and I have encountered some issues with ranks.dbf in trying to
both use it and update it. I am fairly certain that these problems
did not exist when i last used it to analyze male ranks (for the
queueing paper). I think the problems were introduced when we tried
to increase the number of different rank categories (beyond just
"rank" and "frank") and designated an "all" category that is supposed
to rank everyone together. Here is a brief summary of some of the
problems and some examples of them:
1. Rank order among age-sex classes not consistent (most begin
with the order ADM, JUVM, ADF, JUVF)
o grp 2.20 changes to ADM, ADF, JUVM, JUVF 01/02/96
o grp 2.10 changes to ADM, ADF, JUVM, JUVF 01/04/96
o grp 1.20 changes to ADM, ADF, JUVM, JUVF 01/01/97
2. Duplicate ranks
o NAF, VOR, both rank 18, one F and one M, grp 1.20, 01/04/96
o ALE, KAT, both rank 16, one F and one M, grp 2.10, 01/03/95
o LEW, LIN, both rank 17, one F and one M, grp 2.10, 01/03/95
o LIO, NYO, both rank 18, one F and one M, grp 2.10, 01/03/95
3. Some very low ranks for adult males (ADM) and juvenile males (JUVM)
o LEW, age 5, rank 70, grp 2.0, 01/05/94
o FAV, age 13, rank 71, grp 2.0, 01/03/94
o NEL, age 16, rank 73, grp 2.0, 01/04/94
4. Random males within the low ranking females. Seems to be a
group 2.0 problem. Could the sex be wrong for these individuals?
o WAL, JNB, NYU, grp 2.0, 01/03/92
5. Adult males with in the adult female ranks
o JIT, age 14, rank 7, 01/08/97. Also adult females ranked
5,6,7,8,9,10 etc.
(As an aside, we also found that the sname change for DHOW (adult
male, non-natal) to DOW has not yet been made in ranks (and perhaps
other places in the database) -- perhaps this is already on the list
of "to do" as the change was just made).
I don't know enough about what happened when we made the changes to
know how to interpret the errors or, more importantly, how to correct
them. However, it seems that having Leah enter new data into these
corrupted tables is not the right approach. Can someone shed light or
should we have a conference call?
Susan
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Susan Alberts, Associate Professor
Department of Biology, Duke University, Box 90338, Durham NC 27708
phone 919-660-7272 fax 919-660-7293