Join 3,439 readers in helping fund MetaFilter (Hide)


Know any good spots for social network analysis methods advice?
August 31, 2011 5:09 AM   Subscribe

Where can a fellow (who has access to online academic resources but is 1000mi+ from a quality academic library) get some hands-on advice for gathering data for social network analysis? Bonus substantive questions inside.

I'm off on an extended period of ethnographic fieldwork and would like to use some basic social network analysis to round out my picture of the various institutions I'm researching. SNA is not my primary research method--I want to use it as a heuristic, pointing my interviews and participant observation in unexpected directions, or allowing me to check conclusions from more purely qualitative work.

I figured I could pick this stuff up pretty quickly, which may have been a mistake. What I would like is a good place to 1) learn some practical basics with software (UCINET has been recommended to me, and I will be going through their tutorial) and 2) get a grip on how I should collect my data. I figured there might be an appropriate forum or listserv out there whose archives and hivebrain I could rummage through, but googling hasn't helped me much.

Two substantive questions:
1. It was suggested to me that I could get a good handle on the connections between individuals in the scattering of institutions that I'm interested in by requesting CVs. I could then easily see where people had overlapped at different times in their work histories. Sounds good, but I haven't happened across examples (or better yet, tutorials) of how to go about this. I don't want to reinvent the wheel if I don't have to.
2. I gather it's also a standard method to elicit networks from interview respondents ("Please tell me who you've been in touch with in the past two weeks" etc). Also sounds good to me, and easy to incorporate into my interview schedules--but how do I turn this into a network visualization?

Links to fora, listservs, academic papers, websites, online tutorials and so forth are very welcome. Books, too, but I might have trouble getting my hands on them. While I retain my home institution online access, there are no good physical libraries anywhere near me.
posted by col_pogo to Science & Nature (7 answers total) 11 users marked this as a favorite
 
A few years back, I downloaded Pajek and meant to get a handle on it, but never got around to trying much with it. It has a wiki site with some howtos and background.
posted by knile at 5:38 AM on August 31, 2011


check out: http://www.insna.org/

They have a great listserv with people from around the world in the field. Also they have some pretty good conferences.
posted by fozzie33 at 5:41 AM on August 31, 2011 [2 favorites]


If using CV's, you are either going to need a text miner, or do some basic data processing before inputting the data.
i actually just downloaded this software: http://www.casos.cs.cmu.edu/projects/automap/software.php

to see how it processes Text data. looks pretty cool.

I do SNA for Law Enforcement reason, but am also an academic(adjunct at local school) and enjoy the theory and science behind SNA.
posted by fozzie33 at 6:09 AM on August 31, 2011


and, as for gathering data: my two suggestions:
1. contact the schools of interest, they keep CV's on file... not sure how up-to-date they will be, but they should be in their offices... you can then gather these and process with a program or manually.

2. use a web crawler and look at EDU sites for CV, and download any DOC, DOCX, PDF's, etc. that are on those sites and see what you come up with...
posted by fozzie33 at 6:42 AM on August 31, 2011


INSNA is a great recommendation--you should definitely sign up for SOCNET. Discussions there range from the technical to the ontological, and people ask and answer questions about everything from very beginner-level to advanced social and semantic network analysis. The INSNA.org site linked to above contains sample data sets to play with, as well as links to various software. Also tons of articles as references, most of which you should have online access to. Plus: the annual Sunbelt conference is a blast, so you should absolutely submit your project! Many people, especially SNA beginners, bring works in progress to discuss. You can get feedback from the top network analysis researchers in the world, which is amazing.

I think the CASOS ORA suite is overkill for what you need, and as an entry point into SNA. It's a great package, but has a fairly steep learning curve, especially for a beginner. Pajek is also pretty complex, although it's what I started with; it's pretty useless without the book (not the instruction manual), which you may have trouble getting your hands on in a timely manner. Conventional wisdom is that UCINET is both the most common and the easiest to use of all the research-level applications, and is generally recommended as the first pass.

Unless you have a huge number of CVs in your sample, you don't need to do text mining. Manual entry should be more than sufficient. It's hard to give more specific advice without knowing more about your study. Feel free to email me if you want to toss some ideas around.
posted by Superplin at 8:52 AM on August 31, 2011 [1 favorite]


Thanks for the advice, folks. I will definitely check out SOCNET and am impressed with INSNA (although their site had some great-looking web forums that turned out to be sadly empty).

I am looking at a fairly small network of individuals, maybe 150-200 total, so I don't think I will need anything pitched at analyzing huge networks.

I don't have terribly ambitious goals as far as analysis goes: maybe a little bit of calculating centrality, density, and so forth--but my main goal is visualizing a network rather than doing much formal analysis or statistical testing.

My specific concerns are:
1) in the CV (or résumé) study, that I will end up with an excel file chock-full of entries on individuals' attributes, and no clear sense of how to transform that into a network visualization with UCINET. Does the software have clear tutorials? Are there good examples elsewhere that I might follow?;
2) that I follow good practice in structuring my free-listing/elicitation questions and in subsequently using them to make a visualization. I have found a few papers relating to this (from Brewer, Marsden, and others), but I could do with some resource where I could follow along an example step-by-step.
posted by col_pogo at 10:07 AM on August 31, 2011


you might want to try http://www.cytoscape.org/
posted by fozzie33 at 10:49 AM on August 31, 2011


« Older Need help quick: is there a st...   |  Was this what I think it was?... Newer »
This thread is closed to new comments.