Inter-group groomings *more* common than agonisms? was Re: [Babase] Total numbers by interact type

Niki Learn nlearn at princeton.edu
Mon Apr 19 16:58:24 EDT 2010


Jeanne left already.  I'm sure many are errors.  Probably most or all of the
female-female ones are errors and there is probably at least a similar
number in the male-male that are as well.  I think some of them may be real
though.  I will send the small number of examples I've looked up so far
under separate email.

1.  Makes sense to me.  When we get the demog note thing sorted out we
should be able to identify any real ones that have notes on them.
2.  I don't think Laurence cares about inter-group interactions.  They were
just in our way.
3.  Yes, Jeanne and I were discussing the need for a warning system before
the babase call.  Don't think it came up during the call though.
4.  Excellent.

-----Original Message-----
From: babase-bounces at eeblistserv.Princeton.EDU
[mailto:babase-bounces at eeblistserv.Princeton.EDU] On Behalf Of Susan Alberts
Sent: Monday, April 19, 2010 4:36 PM
To: The Baboon Database Project
Subject: Re: Inter-group groomings *more* common than agonisms? was Re:
[Babase] Total numbers by interact type

So, my guess is that the vast majority of these inter-group ad libs  
are errors and I think we will not be able to solve the mystery of why  
the numbers (absolute and relative and grooming versus agonism) vary  
so much over time.  It's probably more important to try to fix the  
problem for the future rather than work out the past patterns.

Here is what I propose. Jeanne, please weigh in.

1. Given that we know that inter-group groomings happen extremely  
rarely, we should *not* pursue these errors now.

2. Given that Laurence really wants to know about some of the apparent  
inter-group agonisms, we should pursue at least some of the ones she  
is focused on, to get a sense of what fraction are "real". I'm  
guessing it's a tiny fraction, so you have the problem of deciding  
whether to invest a lot of time in the 350 in order to find the  
handful of real ones. That's a call for JA and LG and you, I think,  
although I am happy to provide input.

3. For the future, I think we need to have a warning system so that  
when the data are uploaded, a warning occurs when there are inter- 
group ad libs.

4. I will reiterate with the team that it is essential that they  
always record an inter-group encounter when they inter-group agonisms  
so that we can cross-reference these, and also that they make a demog  
note when a male is hanging around, so that these interactions with  
lone males are clear in provenance. However, I do think that they are  
generally quite good at this, so I think our main thing will be to  
pick them out when they come up and prevent them from going in. Looks  
like it happens at a rate of several every month.

Other thoughts?

Susan




On Apr 19, 2010, at 4:21 PM, Niki Learn wrote:

> Hmm...yes, for 2000-2008a the grooming are higher.  I was referring  
> to the
> overall numbers across all years.  So there is a difference between  
> all data
> and recent data on that score.  Yes, across all years .62% of  
> groomings are
> between groups and 1.29% of agonisms are between groups.  I've no  
> idea why
> the more recent data does not match this overall pattern.  Those  
> years have
> 30% of the inter-group grooming, which is a startlingly high number.
> Perhaps this is a result of more grooming data being collected over  
> time and
> individual baboons showing up in more grooming pairs per average on  
> a given
> date in recent years?  Dunno.  After I get you the agonisms  
> examples, I can
> look at grooming data too if we think that would be helpful.
>
> -----Original Message-----
> From: babase-bounces at eeblistserv.Princeton.EDU
> [mailto:babase-bounces at eeblistserv.Princeton.EDU] On Behalf Of Susan  
> Alberts
> Sent: Monday, April 19, 2010 4:07 PM
> To: The Baboon Database Project
> Subject: Inter-group groomings *more* common than agonisms? was Re:  
> [Babase]
> Total numbers by interact type
>
> Thanks Niki.
>
> I am not quite sure I follow these numbers.  It looks to me like 358
> agonisms were inter-group and 520 groomings were inter-group in
> 2000-2008? Is that correct? This looks like inter-group groomings are
> actually higher (both in numbers and in proportions) than agonisms.
>
> I think you are also showing that if you look across all years then
> inter-group groomings are *less* common than inter-group agonisms (is
> this the .62% versus 1.29%?).
>
> I think it's hard to compare with just these data or get a good sense
> of what is going on, except that to me these data strongly suggest
> that most of these interactions are probably errors.
>
> It is hard to know though, and hard to know quite how to proceed. I do
> think it makes sense to look into a handful of them.
>
> Susan
>
>
>
>
> On Apr 19, 2010, at 3:34 PM, Niki Learn wrote:
>
>> Okay, I will take a look at some of the cases and see what I find.
>> Meanwhile:
>>
>> You also asked about how the numbers for grooming compared, with the
>> idea that groomings are less likely to actually occur between groups
>> than are agonisms.  That in fact appears to be the case, both by raw
>> numbers and by proportion of the total.  And MCEs are lower still.
>
> I follow these numbers.
>
>
>>
>> Using Karl's warning query he sent earlier, with the following line
>> or lines added (right before "order by"), I got these results (I am
>> not familiar with these interact tables and so have not attempted to
>> add gender to these queries - I have been using actor_actees) out of
>> the 5689 total mismatches Karl reported:
>>
>> and interact.act in ('A', 'AS', 'OS', 'DS')
>>                3883 agonisms (all years) - this is 1.29% of all
>> agonisms
>>
>> and interact.act in ('A', 'AS', 'OS', 'DS')
>> and interact.date >= '2000-01-01'
>> and interact.date <= '2008-06-30'
>>                358 agonisms (2000 - 2008a only)
>>
>> Total agonisms in babase = 300,066; total in 2000-2008a = 183,862;  
>> adult
> male to adult male = 10,996
>
> This last number is a typo I assume.
>>
>>
>> and interact.act = 'G'
>> 1705 groomings (all years) - this is 0.62% of all groomings
>>
>> and interact.act = 'G'
>> and interact.date >= '2000-01-01' and interact.date <= '2008-06-30'
>>                520 groomings (2000 - 2008a only)
>>
>> Total grooming in babase = 274,601; total in 2000-2008a = 104,477
>>
>>
>> and interact.act in ('P', 'R')
>>                3 approaches/requests to groom (all three were in
>> 2000 - 2008a) - this is 0.38% of total
>>
>> Total approaches and requests to groom in babase = 787
>>
>>
>> and interact.act in ('M', 'C', 'E')
>>                98 MCEs (all years) - this is 0.30% of all MCEs
>>
>> and interact.act in ('M', 'C', 'E')
>> and interact.date >= '2000-01-01'
>> and interact.date <= '2008-06-30'
>>                22 MCEs (2000 -2008a only)
>>
>> Total MCEs in babase = 32,708; total in 2000-2008a = 10,953
>> _______________________________________________
>> Babase mailing list
>> Babase at www.eco.princeton.edu
>> http://www.eco.princeton.edu/mailman/listinfo/babase
>
> --------------------------------------------------------
> Susan Alberts, Dept of Biology, Duke University, Box 90338, Durham NC
> 27708, 919-660-7272 (Ph), 919-660-7293 (Fax)
>
>
>
>
>
> _______________________________________________
> Babase mailing list
> Babase at www.eco.princeton.edu
> http://www.eco.princeton.edu/mailman/listinfo/babase
>
> _______________________________________________
> Babase mailing list
> Babase at www.eco.princeton.edu
> http://www.eco.princeton.edu/mailman/listinfo/babase

--------------------------------------------------------
Susan Alberts, Dept of Biology, Duke University, Box 90338, Durham NC  
27708, 919-660-7272 (Ph), 919-660-7293 (Fax)




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



More information about the Babase mailing list