Network Society Experiment - Evaluation

well, if nothing else this experiment was a success in that it got me thinking more about the process of doing literature reviews. searching for articles and other secondary sources is often assumed to be a basic skill that all grad students already know how to do equally. of course, searching for, managing, sorting, retrieving and meaningfully representing citations is in itself a skill that can be improved upon and performed in a number of ways. especially at an intsitution with such a terrible library system (from collections to interfaces to support) taking the time to figure out how to best utilize the available tools is a worthwhile endavor. i think i will continue to use the softwares and system i tested in this "experiment."

why do i think this system is worth using?
+ large data sets. it is easier to manage a larger set of citation data using bibliographic tagging tools like citeUlike.
+ platform independence, portability and easy of input. adding citations on any computer with a browswer with a bookmarklet is great. using portable firefox on a flash drive to do it is an added bonus for a grad student who often uses several computers in a day.
+ emergence on the scale of one works for me. i saw connections and patterns i don't think i would have otherwise.
+ the potential of collaboration and collective intelligence. seeing how others organize citation data exposes me to different organization system and new sources. [although i'm not convinced the best rise to the top.] also, the possibility of collaboration, be it in the form of leaving or recieving comment or more substantial interaction with a like minded scholar, always lurks on a "social" bibliographic website.
+ exposing and ignoring disciplinary boundaries. social biobliographic tools should make clear that doing interesting work requires ignoring disciplinary boundaries and expose work that is too bound by a discipline. graphing citation connections seems to have real interesting possibilities for showing how scholars' works are [or are not] cross disciplinary.
+ "pervasive access." having access to my citations and my research notes [via this site] could be a great thing, especially when traveling or not working next to my giant stack of papers and books.
+ writing is good. though no one will read it, even writing this is valuable to me as writing practice and writing is a crucial part of my thinking process.

why do I think this system might not be worth it?
+ the learning curve. i was lucky to have the institutional space to do this for a class. otherwise i would have not put in the time.
+ information overload. "casting a wide net" can make it easier to get bogged down in finding out what is out there and managing data. which can yeild a certain forgetfulness towards actually reading and writing.
+ work load increase. being able to process more data quicker can be seen as related to work speed-ups in academia. i don't want to internalize this pressure, nor am i convinced that processing, managing and understanding more in any way yields better scholarly work.
+ data loss. i don't have a backup to alot of this work; i'm trusting the great server in the sky. i'll have to create regular local back-ups.
+ intellectual property. a friend of mine insightfully pointed out that alot of academics are afraid to share because they feel that thier ideas are all they have. a structure that makes scholars feel this way should be fought tooth and nail, but it is a reasonable concern, especially for graduate students who well know histories of advisors and facultry sniping thier ideas and living off of ther intellectual labor. [however, i think the increased transparancy a system like this might offer could combat that particular issue]
+ a waste of time? it might be ... i'm easily distracted by shiny new ways of doing the same thing that i did before.

- network society experiment - No comments / No trackbacks - § ¶ May 08 06 - 11:50

TouchGraph and visualizing relevance.

TouchGraph is a handy tool for visualizing connectivity. i used TouchGraph Amazon Browser and GraphScholar, a mod of TouchGraph for use on Google Scholar to attempt to map connectivity in way that would allow me to extract something meaningful about relevance.

both are fun and easy to use and yield some interesting connections, but are not as immediately useful as i had hoped.

the TouchGraph Amazon Browser did not yield any particularily useful connections to books i hadn't considered. perhaps this is because i used it at the end of my searching. seems like it would be best used as tool for initial research, especially into an unfamilar field.

GraphScholar is a different beast. i was very excited about using this software (which graphs the "cited by" page on google scholar), but i haven't been able get the results out of it i would like. it delivers similar content to web of science (requires subscription) "cited ref search," but not exactly the same. (a "cited ref search" shows how many times an article or author has been cited and where -- but the interface is frustrating the results are difficult to manage. it is mostly used by "science" and clinical researchers to prove to tenure committees and review boards that their work has had an impact on the field). not only does google scholar rank the "cited by" articles in a different manner than web of science (and the ordering can't be changed, thus the only way to change what you graph is to change the # of citations displayed) it also has different includes/excludes.

what i really want to be able to do is graph connectivity, not popularity or relevance. however, currently i can only get GraphScholar to show me exploding links from certain nodes, not all of the links between a set of nodes. this might be a user error ... i have high hopes for the usefulness of this tool.

- network society experiment, research tools - No comments / No trackbacks - § ¶ May 08 06 - 11:32

folksonomy on a scale of one

i was surprised how well "folksonomy on the scale of one" worked for me. by tagging data and having the capacity to represent it in an intuitive to understand way i am came across categories in the literature i think i would have otherwise missed.

even more intersting was finding other people writing on this. which as soon as i find the damn articles i'll post and write more. can't believe i didn't add them to del.icio.us. yeesh.

- network society experiment, research tools - No comments / No trackbacks - § ¶ May 08 06 - 10:46

del.icio.us

although very handy, del.icio.us does not seem to be a useful research tool for literature reviews or bibliographic management.

links to journal articles rarely have any other tags, nor to books from amazon. the interface is not conducive to bibliographic management and citeUlike does function like “a del.icio.us for academics.” however, del.icio.us does appear to be very useful for tracking news stories [and finding related stories i would otherwise miss], so when it comes to developing a methodology for tracking press on FOSS, it will be very useful.

i also tested the following “social bookmarking”/folksonomy sites:

+ http://www.connotea.org/ : connotea was too focused on clinical research. there was almost no crossover with any of the test books i used. looks okay if you’re doing clinical research, but i don’t like the interface as much as citeUlike. also appears to only support RIS, not BibTeX

+ http://www.blinklist.com/ blinklist had the same problems of del.icio.us, not so good for scholarly work. also, not certain about mindvalley labs … looks like the sort of site that might have develop a more annoying interface when it goes out of beta.

+ http://www.librarything.com/ i like the principle of librarything and the interface doesn’t seem bad, but it lacks a good interface for articles and only allows 200 entries for free. “rise of the network society” was in 25 people’s libraries, but was surprisingly not useful for discovering new sources or categories. perhaps this is b/c it is not geared specifically at academics like citeUlike.

+ here’s an actual review of social bookmarkig toos.

jjpeters, jake - network society experiment, research tools - No comments / No trackbacks - § ¶ May 08 06 - 01:08

book reviews - useful?

i found a surprising number of book reviews of Manuel Castells trilogy … 34 to be exact.

i’m not certain exactly what i can do with this data, but i think there is something interesting here. almost half the review are surprisingly long, 5+ pages. somewhat to my surprise, sociology is the discipline with the most reviews, surpasing all other disciplinary journals by far, with what might loosly be called communications journals coming in a distant second.

had i not been casting a wide net and using citeUlike to manage these citations i would have either missed or discarded the book reviews. question is, are they of any use to a literature review?

i think that for my purposes they might be very useful, as i am trying to pull out the categories that are used to understand the socio-ecnomic changes that are understood to constitute the “network society.” book reviews seem to be forum where scholars might be willing to be staight forward with what they consider to be at stake. this may give me an interesting overview of some of the literature, insight into disciplinary boundaries and an idea about how the “network society” is considered relevant.

i’ve got the DOI or .pdf for almost all of the book reviews, i think i may review them before the articles to get a framework for what’s going on in this literature.

jjpeters, jake - network society experiment - No comments / No trackbacks - § ¶ May 08 06 - 01:06

categories for the "network society"

from the 99 articles i tagged in citeUlike some surprising and some not surprising categories emerged.

categories i already had in mind:
+ governance
+ cities
+ digital divide
+ cooperation
+ community
+ nation-state
+ public space

categories i might have guessed:
+ information capitalism
+ civil society
+ ecology
+ gender
+ power

categories i don’t think i would have thought of otherwise:
+ education
+ war
+ planning

jjpeters, jake - network society experiment, research tools - No comments / No trackbacks - § ¶ May 08 06 - 01:01

an experiment in creating a lit review

this post details my attempts to find a "new and improved" research method for creating literature reviews and reading lists.

the goal was to figure out the best categories for understanding literature on the "Network Society." i have no previous knowledge of this literature and had never heard of the concept or Manuel Castells [whose writings popularized the idea and term] before starting graduate school. currently my project is concerned with technology and the state, conceptions of justice and figuring out what is "at stake" in free/open-source software movements, thus, i am looking for a "ways in" to this literature that are conducive to the larger project.

this project is currently on hold, why i work with my advisor to figure out if network society literature is likely to be as large of a part of my project as i had earlier imagined.

my frustration with bibliographic management is rooted in 7 years of working in academic medical-research libraries -- i am frequently frustrated by difficulties in managing large sets of biliographic data. i have have used endnote, reference manager and refworks to various levels of frustration and success. here i attempt to use endnote/refmanager, the BibTeX format, folksonomy websites, connectivity visualization software and blogging software to make the research process function as i would like.

my process:

1) search for relevant citations. with the hope of being able to manage a larger set of citation data than i usually can, i casted a wide net, pulling in aprox. 150 citations from the following databases:

+ Expanded Academic ASAP
+ Gender Watch
+ ISI Web of Knowledge Social Science Abstracts
+ Quick Search Form
+ ISI Web of Knowledge Social Sciences Citation Index
+ ISI Web of Knowledge Arts & Humanities Citation Index
+ ArticleFirst
+ Dissertation Abstracts
+ Humanities Abstracts
+ ABI/INFORM Business
+ EconLit
+ Current Research At University of Southern California
+ NewsBank Access UN
+ RAND Documents
+ Official Documents of the United Nations
+ PolicyFile

these citations were exported from the databases in a variety of formats that would work with endnote, then sorted for duplicates, leaving me with 134 likely looking citations.

2) get the citations from endnote to citeUlike.org.

i highly reccomend posting directly citeUlike if the database is supported by citeUlike or exporting in BibTeX format and then importing into citeUlike instead of using endnote/refworks as an intermediary.

i had to create an endnote export filter [downloadable here] to get the articles out of endnote. after the export they still required a good deal of "cleaning." definetly keeping everything in BibTex format next time.

3) read abstracts, tag articles and attach DOIs or .pdfs. i pulled up abstracts for nearly every article and frequently read portions of the actual paper before tagging with categories. this process whittled 134 citations down to 99 usable citations, 34 of which are book reviews and 9 of which are books. see what i've tagged "network society" in citeUlike.

thanks to UCLA's library i was able to get full text for approximately 85% of the articles (USC's e-journal collection and library in general is almost unusable -- i've heard "UCLA is USC's library" many times).

4) see what others are doing, grep for categories and importance. i found that folksonomy works well on a scale of one [see this post], but it was also useful to see how others tagged the same things i tagged.

unfortunaltly, none of my articles were tagged by anyone else [which may have to do with how i imported my citations], but i added books on the network society specifically for this purpose. 3 of 9 were tagged by other folk, whose research interested intersected with mine in interesting ways. i picked up a few books i hadn't heard of and am now using citeUlike's "watchlist" to keep track of categories used by 4 other users.

from this list of citations i then used TouchGraph Amazon Browser to see the connectivity my books [based on what amazon's customers bought with, looked at after or listed with particular books] and GraphScholar to look at citation webs and connections of key articles. this was not as useful as i thought it would be, but perhaps only because i don't yet know what to do with this data. [more on this]

5) pick key categories, read or re-read articles, and write. in progress. for categories, see this post. as i read the articles, i'll use this site to write-up my thoughts on individual articles and the categories. it's easy to link directly to individual citations in my citeUlike library, so i will be able to eliminate endnote from this process. see results [partial] of an annotated bibliography generated in endnote here -- its not linked to the citation and much less dynamic then i'd like.

6) final product? an annotated bibliography posted on this site -- always accessible, easy to change and reorganize, the basis for my quals questions and literature review for my dissertation proposal and eventually dissertation [or anything else].

my evalution of this process.

jjpeters, jake - network society experiment, research tools - No comments / No trackbacks - § ¶ May 07 06 - 15:30