Social Networks Session Notes

From SocialSoftwareSymposium

Jump to: navigation, search

Notes from session / compiled by Jacob, Carolyn and Songphan.
Feel free to edit/expand/comment/etc.


Experience researching social networks from Anderson, Lampe, and Ellison:

Janna Anderson

Metaverse Roadmap

Imagining the Internet

-Examined early 1990s predictions of what the Internet was going to be and also 2004 and 2006 surveys and scenarios on future development and directions of the Internet.

-2006 survey results: more than half agreed that by 2020, nation-state boundaries will blur and be replaced by city-states, cultural and social groupings; 42% were concerned about predominance of tech over man, leaving humans out of the loop; 56% said VR is bad; 49% said transparency would be a net negative -- i.e., privacy concerns (51% said positive).

-Concern exists over control of internet architecture and establishment of walled gardens by those in power in entrenched institutions and the commensurate stifling of human-networking potential; new technologies disruptive.

-Social Networks are migrating into synthetic worlds - MySpace, Friendster et al. are precursors to a world where it's all represented in immersive virtual worlds like WoW and Second Life where individuals are represented by avatars.

-Metaverse roadmap looks ahead; Metaverse doubling every two years. Ongoing focused effort to work toward the best future possible; a working report is being developed by ASF (Studies Foundation)

-Gaming has pushed the field, beginning with mid-1990s adoption of polygon-based graphics processors in consoles. Starbright World (1995) was the first broadband virtual world; Alpha World (1995) introduced further elements of a full virtual world.

-Ultima Online (1997) was the first big hit for 3D MMORPG - peaked at 250k subscribers in 2003

-Second Life (2003): first persistent virtual world allowing users to retain property rights for virtual objects they create, and first to make it over the hump into sustainable exponential growth. Commerce in VR is in nine figures; by May 2006, 230k downloads.

-Web 2.0 is of a piece with these developments

-2D online worlds are still most popular (i.e., CyWorld); WoW is "the new golf" - a shared social experience where users get together and talk about real-world events and life while engaging in automatic and ritualized behavior.

-Today's browsers are just beginning to process 3D graphics; much work to be done with interfaces, security, etc. May need 3D desktops first.

-AI avatars and agents continue development, much through Hollywood - building automatic actors

-Virtual tourism set to expand; geo-tagging on the rise

-Most VR environments still isolate us, distract us and undereducate us more than they empower us.

-Ubiquitous computing and the participatory panopticon - the lifelog, recording all experience digitally. New social conventions will need to arise to deal with these issues.

-3D isn't always more efficient or intuitive than 2D; 1D and 2D still have their uses.

-Better collaboration tools needed for development of the metaverse.

-Best 3D virtual worlds will recreate the reality of physical space but also make it better by, e.g, violating the laws of physics

-Identity: one model is a single global sign-on, but this is difficult technologically as well as presenting ethical and security challenges.

-Will we choose to spend more and more time in front of boxes and in pretend places?

-3D virtual worlds will enhance real-world cooperative activities


Nicole Ellison and Cliff Lampe

- Focusing specifically on methods – ways to study social network sites – what kinds of data sets are out there, etc. - Data sets – survey of MSU graduates last spring – random sample pulled from MSU registrar’s office (n=286) – first pass, explore

- SNS Methods: SNS Research at MSU

- Intensity of Facebook use – how much time you spend on Facebook, how much it is part of daily practices, perceptions of attitude about being on face book – indicators – bridging social capital

  • Three categories of social capital: bridging, bonding and maintained. Intensity of Facebook use was a predictor for all three, esp. bridging

Web crawling:

  • prove the relationship b/w profile fields, number of friends, and

social capital

  • Web crawling – captured entire MSU Facebook networks and ran some analyses– i.e., does number of friends predict social capital?

First year student survey:

  • social searching vs. social browsing
  • Added questions to first year student survey – n=1440, with a twenty percent responses rate…
  • Do [use Facebook to identify people they already know or to identify new individuals?

Plan:

  • Coming up – cognitive walk-throughs; sit down with users in front of their FaceBook profiles and ask them questions about it
  • interviews(self presentation and formation)

- Vibrant and growing field; evidenced by, for example, 105 abstract submissions for SNS publication ...

Overview of Methods:

Self-report / Survey / Sampling concerns:

  • Who chooses to participate
  • Difficulty of locating non-users (population and sampling frame)
  • Incentivizing non-users
  • Self report bias
  • Combining survey with behavior
  • What population are you trying to generalize to ..

Interviews:

  • Particularly appropriate for early stages when trying to understand a phenomena and don’t know how to word survey items ..
  • Allow probing and contextual information
  • Cognitive walkthroughs; guided interview while looking at some artifact or performing a task
  • Controlled behavior data with the interviews

Experimental Design:

  • Useful for exploring questions of causality
  • Quasi-Experimental
  • Naturally occurring intervention (Pre/post)
  • Privacy setting after and before Facebook implemented NewsFeeds
  • Artificial intervention
  • Privacy setting before and after
  • Have not used it yet ...

Direct Observation / Ethnographic approach:

  • Involves observation over period of time
  • Participant-observation in which researcher engages in activity
  • Capture emergent behavioral data

Content analysis:

  • Hand coding of profiles or other artifacts
  • Automated analysis of text- programs available
  • Richer data about a smaller subset; items that cannot be computer-coded
  • Low entry barrier … however, very labor intensive in terms of analyzing these huge bits of data sets ...

Web-crawling and web log analysis:

  • Goals for starting up web analysis in first place – how was it being used – what fields were being filled out, who was befriending who ….
  • Automated collection of web activity
  • Track profile elements
  • Track friendship links
  • Activity over time -- see relationship network -- are people adding or decreasing amount of information (do people actually ever unfriend people on Facebook)?

Constraints:

  • Reduce effect on server load
  • Work with permission of company/administration
  • Hardware limits -- constrains type of information we can collect
  • Technical constraints -- scripting expertise, database and server setup

Issues with Facebook web analysis:

  • No automatic downloading of content on the site; got permission from lawyers, but not tech guys, so tech guys were like, why are you scrapping our database ….
  • Also, technical constraints with web crawling; technical barrier to writing these things – all scrapping scripts have to be customized for the site …
  • Open fields -- hard to fix over several thousand profiles -- i.e., to just try and find people from Michigan, hard to do – maybe just list city, or high school, or misspell it.
  • Privacy settings
  • Facebook keeps changing the site -- new fields, new services, the administration = vague future
  • Also, got a message a few moths ago that they have to stop crawling --- Facebook is working on a way to give server information to researchers …

- What we have now ---- friendship networks that change over time (people just add friends, nobody de-friends).

Future Directions:

  • Cross-SNS analysis -- how SNS fit into constellation of tools; also, differences between information displayed on Facebook from MySpace …
  • Cross-cultural investigations -- non-US SN sites
  • Longitudinal/life cycle studies
  • Non-users and exiters – what distinguishes them from users? I.e., people who create identities and never go back to them
  • SNS in ubiquitous environments
  • Social networks --defining success; measurements of success
  • Online/offline connection
Personal tools