[Babase] Ranker Problem

Lacey Maryott lacey.maryott at duke.edu
Fri Oct 12 09:39:31 EDT 2007


Hi Jun (& al),

        Susan and I used the new Ranker to assist in making our 
decisions while doing male ranks this week.  We found some interesting 
things.  When comparing new ranker to Tabby's ranks she sent us from 
FoxPro ranker, there were discrepencies in actual agonistic 
interactions.  After doing some looking within babase, it seems that the 
new ranker is 'creating' data that don't exist within babase.  I have 
included some specific examples so that we can try to figure out where 
these anomalous data are being generated from.  They were almost all 
interactions between subadults, and some of them were even symmetrical, 
and would have cancelled each other out essentially... but we need to 
figure out what's causing this to happen.  Also, lastly... we found that 
when we selected 'ALM' the same group of individuals was produced as was 
produced when we selected 'ADM'. In other words, infants, juveniles and 
subadults are being selected when we select the ''ADM" ranktype.  Why is 
this?

Anomalous Interactions: None of these interactions can be found in babase

1. Viola's Group June 2005 SEB>VIB and EGO>SEB.  , (but there was one 
listed for VIB>SEB (which makes a lot more sense))

2. Viola's Group Aug 2005 VEB>LOF x1 and YOK>LOF x1

3.Viola's Group Sep 2005 THR>NJU and VEB>WES

4. Linda's Group Nov 2005  LES > MOR and LES > LAW

These are all of the specifics I am coming across for now, but if we 
need to look into it more deeply to find the root of the problem, I can 
regenerate the matrices and compare them with Tabby's to find more examples.

Cheers
Lacey


-- 
Lacey Maryott
Alberts Lab
Department of Biology
Duke University
ph: 919-660-7306
fax: 919-660-7293
Lacey.Maryott at duke.edu 



More information about the Babase mailing list