Print

Print


*****  To join INSNA, visit http://www.insna.org  *****

I would certainly recommend average degree over density as it places far fewer demands on large networks. Assuming people have larger networks by knowing people in more contexts rather than more people in a single context, average degree will not increase much with network size. 

I assume the networks will have many isolates and components, in which case share in the largest component would be worth comparing. You might also consider some deductive/descriptive metrics such as the number of people with degree > 4, based on some specific logic related to the process being studied. 

Dr Bernie Hogan
Oxford Internet Institute
University of Oxford

Sent from my iPhone

On Aug 8, 2013, at 3:27 PM, Juergen Pfeffer <[log in to unmask]> wrote:

> ***** To join INSNA, visit http://www.insna.org *****
> wrt diversity index: I prefer the Brillouin Index:
> http://www.pisces-conservation.com/sdrhelp/index.html?brillouind.htm
> Just use (or create) nominal groups (species) of your data. In an ego-network N is the number of alters, if you want to use 5 attributes for one index, just use 5*N. 
> 
> A great book for diversity indices: 
> Anne E. Magurran, Measuring Biological Diversity
> 
> 
> 
> 
> 
> On 8/8/2013 6:08 AM, Daryl Sweet wrote:
>> ***** To join INSNA, visit http://www.insna.org *****
>>  
>> Hi all,
>>  
>> I have two fairly basic personal network analysis questions Ifd appreciate some advice on:
>>  
>> 1)Can anyone recommend the best way to compare density across a large number of personal networks when the size of each network is different? Is there a way to control for size when considering density? Or is it best to use degree? We know from other studies that density can be important in our study population, however we collected networks open endedly so network size varies from c.6-65. I was considering whether splitting network size into quartiles would control for size impacting density.
>>  
>> 2) Ifm looking for a good diversity index score. I have info  for age, gender and relationship (amongst other attributes) for every alter in each ego network. Ideally I would like one diversity score, perhaps which combined Standard deviation (for age), EI index (for gender) and number of different relationship types into an overall diversity value. Ifm not sure that this makes sense however.
>>  
>> With thanks in advance,
>> 
>> Daryl
>>  
>>  
>> Daryl Sweet
>> Researcher
>> The McPin Foundation
>> 32-36 Loman Street
>> London SE1 0EH
>>  
>> T: 02079227874
>> www.mcpin.org
>> www.facebook.com/McPinFoundation
>> @McPinFoundation
>>  
>> We are a charity registered in England 1117336.
>>  
>>  
>> _____________________________________________________________________ SOCNET is a service of INSNA, the professional association for social network researchers (http://www.insna.org). To unsubscribe, send an email message to [log in to unmask] containing the line UNSUBSCRIBE SOCNET in the body of the message.
> 
> 
> -- 
> 
> Juergen Pfeffer, Assistant Research Professor
> School of Computer Science
> Carnegie Mellon University
> Pittsburgh, USA
> www.pfeffer.at
> _____________________________________________________________________ SOCNET is a service of INSNA, the professional association for social network researchers (http://www.insna.org). To unsubscribe, send an email message to [log in to unmask] containing the line UNSUBSCRIBE SOCNET in the body of the message.

_____________________________________________________________________
SOCNET is a service of INSNA, the professional association for social
network researchers (http://www.insna.org). To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.