sunbelt 2012

Very interesting social network meet up: full of sociologists and full of jokes about physicists – jokes mostly by older folks. Those are my rough notes. I don’t have time to edit them… sorry ;)

- daniele (daniele tweeting)

(warning: i could not cover all presentations i attended).

Strategy and status in online dating (or: How to get a date online). Brilliant (super) brilliant talk by Kevin Lewis. He studied network data from a popular online dating site and looked at questions like “Who sends more or fewer messages to other dating site users? Who receives more or fewer messages? And who is more or less likely to respond to the messages they receive, and to receive responses to the messages they send? “. Very good work with a nice outcome – “gendered status hierarchies that characterize a given social structure”. A must read work.

Happiness as the Duality of Ritual and Belief: Mapping Buddhist Social-Cultural Identities. This was a really good talk! José A Rodríguez and John W Mohr presented their analysis of the process of becoming a Buddhist in a Catholic country (Spain) “by mapping the shared systems of meanings and repertoires of practice in a western Buddhist lay sangha.” They analysed beliefs and of practices by administering a survey with “detailed questions about their life styles, including fine-grained questions about religious beliefs (subsets of beliefs that reflect one’s Buddhist way of feeling and seeing) and the ensemble of practices that a person assembles.”

Concurrent Sex Partner Relations within Sexual Networks of Swingers. Fully packed room! Anne-Marie Niekamp “examined network indicators for the level of concurrency in the sexual ego networks of swingers predicting high potential of STI transmission”. Work with great potential!

Targeting Conflict with Social Network Balance. Ian McCulloh delivered a brilliant presentation in which he proposed “an approach based on balance theory and transitivity to identify non-intuitive centers of gravity to target and eliminate social conflict within an organization. The approach is demonstrated on three examples: a family in divorce, a team of intelligence analysts and information technologists operating in Iraq, and policy implications for the US war in Afghanistan.” He also compared the centrality of the same individuals in two networks centrality(hate net)/centrality(like net) ;)

The “Telefunken Code”: A technique for the collection of anonymized social network data on sensitive topics. This was a brilliantly delivered presentation. Travis Wendel presented a way to “anonymously” collect social network data (which is very important if one is collecting data about as HIV transmission, illicit drug distribution and the functioning of criminal organisations). Using this technique, they interviewed methamphetamine users and distributors in New York City. More specifically: “we asked each participant how many members of the target population’s cell phone numbers he/she knew; if the number

was five or fewer, we elicited information about all of them; if the number was greater, we elicited information about a random sample of five. We asked participants whether each of the last four digits of their (and network members’) cell phone numbers were odd or even, and from 0-4 or from 5-9 (generating the “Telefunken Code”).” In the future, they might ask for more digits, tackle the cognitive difficulty for understanding which number is odd/even, and understand the impact of prepaid (disposable) mobile plans.

To Broker, or Not: The Psychology of Bridge Decay. Eric Gladstone presented a very nice work on how brokers are perceived. They demonstrate that people perceive brokerage roles as relatively more advantageous relative to other network positions. However.. Participants in our studies saw these positions as burdensome, carrying the potential for significant cognitive overload and emotional stress. .. what is necessary to maintain one’s position as a bridge is likely to be detrimental to one’s reputation, which also may explain why the bridge decays over time. ” Here are the findings: 1) non-brokers were rated significantly more trustworthy than brokers; 2) non-brokers were chosen more often than brokers in a game that required trust (investment game); 3) brokers were chosen more often than non-brokers for a trivia game.

Risk, Uncertainty and Tie Strength by Paolo Parigi and Bogdan State. ” Although risk and uncertainty have long been recognized as fundamental factors impacting the formation of social ties, their close study has been confined to the lab. ” To fix that, they analysed data from They found “that ties formed in either high-risk or high-uncertainty conditions are stronger than those formed in no-risk or low-uncertainty circumstances. Uncertainty is not a significant predictor of tie strength when risk is inexistent, however. Our findings show, likewise, that highly-embedded ties are stronger, and that ties are stronger if they are embedded by stronger ties.”

Navigation in semantic memory. Nicole Beckage presented a very neat work on semantic networks (using florida free association norms [nelson et al. 1999]). she considered words being nodes and edges are directed and weighted based on frequency of response. and she looked at how individuals navigate from a starting location in the semantic network to a target word.

More talks (for hard-core readers) ;)

Disruptive Diffusion: Adoption of Microblog Technologies Among Urban Organizations (by Britta Johnson & co.). They “investigate the adoption a of popular microblogging service among five metropolitan areas tasked with coordinating homeland security preparedness for urban populations..

#misdis: A Social Network Diffusion Model of Misinformation and Disinformation (by Natascha Karlova and Karen Fisher). “The rise of social media, such as Twitter, Facebook, and Tumblr, has made the dissemination and diffusion of misinformation and disinformation easier and faster”. The authors “describe distinctions between misinformation and disinformation, and provide a conceptual and social model of how both diffuse through personal social networks afforded by social media”

Spatial Networks and the Perception of “Neighborhood”. Unfortunately, I missed John Hipps’ presentation. The abstract looks interesting: “What is the spatial footprint of a neighborhood, and how does that relate to the spatial distribution of residents’ social networks? We use data from the Twin Communities Network Study, which explores the spatial distribution of social ties in two communities covering 13 census tracts in Southern California, to address these questions. The survey asks residents to define which of their network ties live in the “same neighborhood”, which provides us implicit information on the spatial distribution of their neighborhood. We then compare this perceived neighborhood spatial footprint among the residents living in the same geographic area. We assess whether the number of neighbors known is related to the size of the perceived neighborhood, and assess whether each respondent’s perceived cohesion with the neighborhood is related to the spatial footprint of the neighborhood. We also assess whether the degree of cohesion with the larger city is related to the spatial footprint of the neighborhood.” Going through John’s publications, I notice this article “Measuring ‘neighborhood’: Constructing network neighborhoods” Super cool topic. I found interesting this claim in the paper “A key focus of this study is attempting to optimally cluster the observed social networks in space. To our knowledge, this has not been done in the literature. ” I need to read the paper ;)

Stratified identities and variable distances.

Matteo Magnani presented a work in which they are studying the same individuals in multiple networks, say, in Twitter and Youtube.

Emergence of Scale Free Architectures in the Ktunaxa Online Language Community. : Christopher Horsethief presented preliminary results from a Native American online speech community. Very nice work

Network-Relevant Personality and the Agency Question. In his talk, Ron Burt assumes that “consistency across the role-specific networks reveals the person’s network-relevant personality, a recurring network style; she brings to the roles she plays”. In other words, if one takes the same individuals across different projects and look at their network structure, then one should find that the more one individual’s network structure is consistent across projects, the more personality matters in shaping social networks. This is a very special way of thinking about this stuff… not sure

An Examination of the Effectiveness of the APAN 2010 Haiti Earthquake Network for Disaster Response Coordination (by Gail Fann Thomas & co.). “This study examines coordination efforts for the 2010 Haiti earthquake international relief efforts using the U.S. Department of Defense’s (DoD’s)All Partners Access Network (APAN) 2010 Haiti Earthquake Responders’ forum database. We analyze real-time interactions of 271 participating organizations during the response period from 13 January to 6 June 2010 using semantic network analysis. We answer the question: Did APAN provide an effective online environment for disaster response coordination during the 2010 Haiti earthquake international relief effort?”. As for methodology, they used “Lexical Link Analysis (LLA) to extract semantic networks and expose themes and topics discussed between the actors over time.”

Mapping the global performance art network 1975 – 2005. : Rebecca Clunn presented a longitudinal analysis of “an unbound global performance art network which is thought to have commenced c1975 and consist of more than 3000 artists”. She plans to have an online resource soon.

Comparing networks : an approach grounded in a theory of social processes. Emmanuel Lazega “explored a theory-grounded comparison between substantively similar networks measured across different types of social settings. Substantive similarity is defined as measurement of the same social process.”

How do brokers broker? Eric Quintane offered “a new image of the broker as an individual who swiftly identifies and moves across structural holes during the short period in which these are open, rather than one who keeps holes open in the long run in order to benefit from the resulting positional advantage” Very nice work.

An Examination of the International Internet with Multiple Measures
Authors: George Barnett. He presented an interesting  graph of residuals (Countries with negative residuals separated by countries with positive residuals)

Do we befriend those whom our friends like? András Vörös presented a paper on triadic closure -” The more friends with positive attitudes towards a certain peer will increase the probability of the individual befriending him or her. Likewise, negative feelings of friends will decrease the chances for forming a friendship tie between the individual and the disliked peer.”

Do you tweet like a Republican? Concept networks as an indicator of voter behavior: Maksim Tsvetovat. They hypothesise “the word choice of individual in non-political speech will carry predictive power in relation to voting behavior. For example, any individual consistently using the word “Obamacare” can be immediately classified as a conservative. From this noise, I construct linguistic networks for specific candidates, locations, news outlets”. The basic idea is that Republican voters will use language a way similar to Republican candidates (mirroring hypothesis). They identified words used by anyone and words (trademark words) who are used by those with a specific political affiliation.

Beautiful people, beneficial networks. Kathleen O’Connor. “That attractive people enjoy relatively better work outcomes compared to other people is well and widely documented in the organizational psychology and management literatures. How and why this occurs is less clear. .. , the more attractive the respondent, the larger his/her network, and the greater proportion of weaker ties he/she reported. Both of these characteristics of networks could help attractive people gain access to important information and new ideas, for instance”. This work reminded me of Catherine Hakim’s book “Erotic Capital: The Power of Attraction in the Boardroom and the Bedroom”

Tutorials: martina & co. presented a tutorial on how to use ergm models in R. the tool looks really good. warning: it might become quite slow for network with large number of nodes. tom valente delivered a tutorial about how to determine which network alterations would maximise network cohesion. he mentioned a paper by cross, borgatti, parker 2003 that i need to look up.


Comments are closed.