Print

Print


Stacy --

You are right about 10 being too small. There is simply not enough room in
a network of 10 people for anything particularly interesting or complex to happen.

As for what will be so large as to cause respondent fatigue, it's not possible to
give a particular number. It depends, among other things, on how dense the
network is. If the network has a total of 120 people and each person in the
network has only half a dozen connections, there shouldn't be much respondent
fatigue as long as the names are organized in a way that makes it easy for each
respondent to locate the ones he or she is connected to.

If the network has a total of 50 people and each person in the network has thirty
or forty connections, you are going to have respondent fatigue, especially if
"multiple measures" means a LOT of complicated measures.

The critical factor for an upper limit that prevents so much respondent
fatigue that the validity of your results are threatened will depend more on how
many connections each member of the network is likely to have and how many
measures they have to fill out for each person they select than on the total number
of people in the network.

A network with 120 will be large enough for a range of interesting phenomena
to be seen. Do you want to just demonstrate an analytic technique or are you
interested in studying real-world phenomena? For the former, use a network
that is big enough to let you demonstrate the abilities of your technique. For
the latter, use a network that is more than "just large enough" to show the kind
of things you are interested in. If you don't, you take the risk of trivializing your
problem or stacking the dice by limiting the range of what is possible.

Personally, I like bigger better than smaller.

Have fun!

Bill

Stacy wrote:

> I'm looking for a few seasoned network researchers who can give me a little
> guidance on choosing an optimal network size that will:
>
> (a) maximize the range in centrality scores I will find within the network,
> from highly central to peripheral, and
> (b) minimize respondent fatigue, as I plan to ask each network member to
> scan the list and spend time filling out multiple measures for each person
> selected.
>
> I am in the fortunate position of choosing size parameters for the networks
> I will analyze, and I'd like to specify a range that has some grounding in
> experience, if possible.  I have potential access to networks that range
> from 10 members up to 120, and I intuitively think that 10 may be too small
> and that 120 may be too big.  A pilot study is in the works to help me come
> up with an optimal range, but I'm wondering if any of you, from experience,
> can suggest a range that will satisfy both conditions.
>
> Best,
>
> Stacy
>
> P.S. I wish to publicly thank all of you who responded to my earlier email
> about using the Web to collect network data. Your ideas and suggestions
> were quite helpful and I'm moving along on this portion of my study. Thanks!

--
Bill Richards
Professor
School of Communication, Simon Fraser University
8888 University Drive, Burnaby, B.C. Canada  V5A 1S6
Phone: 604 291-4119, secretary: 604 291-3687, home: 604 251-3272
fax: 604 291-4024
Web site: http://www.sfu.ca/~richards

Support anti-Spam legislation.
Join the fight http://www.cauce.org/