[Babase] RANKS - prepare for conference call

Susan Alberts babase@www.eco.princeton.edu
Thu, 24 Feb 2005 17:49:01 -0500


HI all,

Jeanne, Catherine, Susan and Leah came up with a plan for how to deal 
with ranks. We would like to discuss it with Karl and Daphne in a 
conference call next Weds Mar 2 at 11 am EST. Will this work for 
every one?


Here is a document compiled by me and Leah reflecting the plan that 
the four of us came up with. The proposed goals for overal structure 
of babase ranks file and implementation shoud form the basis of our 
conversation on Weds.

Thanks,
Susan

Proposed Changes to Ranks File
Conversation between Duke and Princeton, February 23, 2005
Leah Gerber and Susan Alberts

Problems that led to this discussion
o	The "ALL" ranking is currently not consistent or complete.
o	It is useful information which has not been used to it's full 
potential as of yet
o	There are a number of errors in ranks, especially for adult males.

Goals for overall structure of babase ranks file
o	We agreed that the "default" or basic structure of the ranks 
file in babase should be one in which the males are all ranked in one 
block from highest ranking adult to youngest male, and the females 
are all ranked in a separate block from the highest ranking adult to 
the youngest male. I.e., there should be a "female ranks" and a "male 
ranks".
o	Ranks within age-sex classes (e.g. only juvenile females or 
only subadult males) could then be easily extracted, by extracting 
the animals within a class and keeping them in the same relative 
order they were in in the mixed-class ranking.
o	This basic default structure does not reflect our ideal 
system. Our ideal system is one in which all the animals in a group 
are ranked relative to each other in one continuous ranking based on 
their actual dyadic interactions; for ranks within age-sex classes we 
extract the animals within any given class and keep them in the same 
relative order.
o	However, this ideal may be difficult to attain because of the 
labor involved, and even if feasible as a strategy from now on, 
updating the early years to reflect this is a huge task. Ranking 
within sexes is feasible for all years.

Implementation
o	At Duke, Susan and Leah will rank subadult males for all 
years, and this will allow us to produce an "all males" rank in which 
juvenile males are simply ranked by age (by "ranker"), for years 
through 1995.
o	Jeanne and Catherine will do the same at Princeton for 
females (this is essentially already done).
o	This will give us within-sex blocks up through 1995.
o	We will also explore what is involved in actually approaching 
our ideal of doing "all ranks". In order to determine the time 
commitment for keeping this category updated and also assessing how 
much it is used for research purposes we propose to calculate "ALL" 
ranks from 1995-2005.
o	This time period is especially useful because it corresponds 
to 5 years of hormone analysis.
o	"ALL" will include juvenile females and juvenile males, and 
will incorporate interspersed sexes. This will allow us to see how 
juvenile males work up through the adult females.
o	Ten years of  "ALL" ranks will be determined initially and it 
will later be decided whether it is worthwhile to continue with this 
type of ranking or revert back to the ranking by sex class that 
exists before 1995.

Technical Details and Problems

1.	Susan and Leah don't understand what role "ranker" will play 
in this process and need clarification on this. Leah's upcoming trip 
to Princeton will help with this. In particular, can ranker can 
create "ALL" with interspersed sexes as well as the previous "by sex 
class" method ?
2.	Princeton will calculate the ranks for adult and juvenile females
3.	Duke will calculate the ranks for adult and juvenile males
4.	After both are done independently, Princeton sends the female 
ranks to Duke and Susan

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Susan Alberts, Associate Professor 
Department of Biology, Duke University, Box 90338, Durham NC 27708 
phone 919-660-7272  fax 919-660-7293