Complexity and Social Networks Blog

Posted by Alexander Schellong at 12:30 AM | Comments (2)

27 March 2007

DevalPatrick.com

Deval Patrick recently relaunched his campaign website to be a Web 2.0 style website, allowing anyone to post issues and have people vote on them-- see Boston Globe article, which is implicitly critical. It’s an interesting experiment, and notable that it is not an official government website, but still, essentially, a campaign website (e.g., you can donate money to his campaign).

Posted by David Lazer at 10:33 PM | Comments (0)

The International Working Group on Online Consultation and Public Policy Making

I am pleased to announce the launching this week of the NSF-supported (via the Center for Technology and Governance at SUNY) International Working Group on Online Consultation and Public Policy Making. We will be meeting 4 times over the coming year, with the Program on Networked Governance will be hosting the opening meeting of the working group at the Kennedy School this Friday/Saturday. The initial component of the meeting is public (see details below). The overarching questions motivating the working group’s mission will be (a) how to evaluate the policy and other social impacts of online citizen consultation initiatives aimed at influencing actual government decision making, and (b) how the optimal design of such initiatives is affected by cultural, social, legal and institutional context. There will be a variety of outputs from our work, aimed at both academia and the policy world, including a report evaluating the policy and design issues with respect to online deliberation, and a book.

Practicing what we preach, ideas/suggestions on what we should pursue should be posted here, with links to work and resources in this area. I will make sure that anything (reasonable) posted enters the discussion in some fashion (with credit, of course).

The event this Friday:

Peter Shane (Ohio State University) & Stephen Coleman (University of Leeds)
Launching the International Working Group on Online Consultation and Public Policy Making
12:00-1:30, March 30, 2007
Location: Bell Hall (Fifth floor Belfer building, KSG)

Welcoming remarks will be offered by Valerie Gregg, Assistant Director of Development, Digital Government Research Center, Information Sciences Institute, University of Southern California and Peter M. Shane, Jacob E. Davis and Jacob E. Davis Chair in Law, Ohio State University. A keynote address will be presented by Stephen Coleman, Professor of Communication, University of Leeds, who will be discussing "Future Research Directions in Public Online Consultation."

International Working Group on Online Consultation and Public Policy Making

US Co-Chair:
Peter Shane (Center for Interdisciplinary Law and Policy Studies, Ohio State University)

International Co-Chair:
Stephen Coleman (Institute of Communication Studies, University of Leeds, UK)

Working group members:

Steven Balla (George Washington University)
Patrizia Bertini (European Internet Accessibilità Observatory, Italy)
Andrew Chadwick (Royal Holloway College University of London, UK)
Sungsoo Hwang (PhD Candidate University of Pittsburgh)
David Lazer (Program on Networked Governance, Harvard University)
Jeffrey Lubbers (Washington College of Law, American University)
Laurence Monnoyer-Smith (University of Technology Compiègne, France)
Beth Noveck (New York Law School)
Kerrie Oakes (PhD Candidate Griffith University, Australia)
Oren Perez (Bar-Ilan University, Israel)
Polona Pičman (Štefančič University of Ljubljana, Slovenia)
Vincent Price (Annenberg School for Communication, University of Pennsylvania)
Alicia Schatteman (PhD Candidate State University of NJ at Newark)
Peter Strauss (Columbia University)
Scott Wright (De Montfort University, UK)

Posted by David Lazer at 9:17 AM | Comments (3)

25 March 2007

Following the e-mail trail in the US Attorneys controversy

To any social network scholars out there, I will feature any decent network pictures that you come up with from the e-mail data released by the DOJ on the Netgov blog.

One of the notable points (to a social network scholar) of the controversy over the firing of the 8 US attorneys are the DOJ e-mails that have been released .

This episode illustrates the siren call of e-mail. Here is a set of people who have an enormous incentive to keep their interactions untraceable, and yet so much of their communication is via e-mail. It’s just too convenient. I am sure that truly sensitive issues are much more likely to be conducted via face to face or phone, but it is hard to anticipate what will become sensitive. Thus, for example, one e-mail highlights the AG’s presence at a meeting about the firings, contradicting some of his later statements. It would have been hard to anticipate that this would be an issue when the e-mail was sent.

Our lives, in short, are becoming increasingly recorded by the (nonhuman) network—via e-mail, via mass transit cards, via phone records. What are the implications for social science? As I have discussed before, and will focus on in a series of entries in a month or so, the implications are potentially revolutionary for our understandings of collective human behavior. Whether the academy is poised to seize the day is another story (something else I will be examining).

Posted by David Lazer at 9:16 PM | Comments (1)

23 March 2007

Sharing knowledge about vacations

One of the exercises in my Building Organizational Social Capital Class to illustrate the benefits and challenges of accessing the “knowledge in the network” was to have students post where they have vacation relevant knowledge on a Google maps mash up (at www.socialight.com). The way socialight works is that you can put “stickies” up on a global map in particular destinations—e.g., if you know about London, you put a sticky on London with some comments. Everyone who belongs to your channel can see the stickies in the map and can thus see what is known, and since each sticky belongs to an individual, you can see who knows what. Finally, since this is embedded within a social network type of site, you can see who knows whom. The idea was to see whether individuals in the class who were going on vacation (this is spring break at Harvard) had classmates with expertise on their destination. The nice thing, from a pedagogical point of view, is that geographic knowledge (unlike many other domains) can easily (and literally) be visually mapped in a way that everyone understands.

A few interesting lessons:

(1) There’s a lot of knowledge in the network—in my class of about 30 for every single destination that people had, there were quite a few people in the classroom who had substantial knowledge about that destination, even for fairly esoteric destinations, like Jordan.

(2) This knowledge is fairly invisible—most students did not know about the locations that other students had expertise in.

(3) Face to face still beats the Internet—just asking people where they were going and who knew about those locations proved vastly more effective at linking individuals than having people look at the map. Some of this has to do with the still developing socialight interface—e.g., the fact that only 10 stickies can be seen at a time proved to be a problematic constraint when you have well over 100 stickies on the map. Further, there is no way to search geographically for stickies. Of course, face to face is often not possible, which presumably is the power of Web 2.0.

(4) The value of the stickies was not so much in conveying knowledge (even though people put some interesting information on the stickies) as conveying who had knowledge.

(5) Boundaries can be important to facilitate sharing of knowledge. Unfortunately, a number of students received rather offensive spam from the site. This reflected the fact that we could not block entry to the channel we created, which meant that we had some interlopers. If we had long term aspirations for the mashup, this might well have scuttled any such plans.

(6) (not so much a lesson, but an idea) It would be neat to be able to highlight particular geographic areas and find out who in your network knows about those areas (this relates directly to research that Nosh Contractor has done on network and knowledge).

I will let you know in some future comment on this posting whether any students came away with knowledge that proved useful in their vacations…

Posted by David Lazer at 8:50 AM | Comments (1)

18 March 2007

Digital Life and Design Conference 2007 - Follow up: Video of Online Social Networking Panel Discussion

In case you have nothing to do this Sunday, here is a short follow up on the DLD conference 2007 which I noted in an earlier entry. There was a panel with Erik Wachtmeister (asmallworld), Lars Hinrichs (xing) and Matt Cohler (facbook) which covered various aspects of social networking platforms (i.e. business models, future). Here is a link to the full video of the DLD social network panel discussion "The Link Society" moderated by former Alando and Jamba founder Oliver Samwer. In order to watch the video please click "Monday - January 22" on the navigation bar on the right, scroll down to "10:30 am The Link Society" and then just click on "Play video".

Posted by Alexander Schellong at 12:00 AM | Comments (1)

16 March 2007

Quick survey: How do you search for information?

Many of us spend a considerable amount of time every day on searching for various kinds of information required to do our job. I find it quite fascinating to observe how individuals go about this task in different ways. In some professions (or organizations?) there seems to be a set of unwritten norms guiding search behavior. Then, of course, how we search depends on our personal traits and preferences. If you generally are not the person who picks up the phone and calls somebody as soon as a question arises, you will probably use alternative ways to retrieve the answer you are looking for - such as sending a request to a listserv, or googling a term. Furthermore, we often use a combination of sources and media for a single query, sometimes without even realizing this "sequencing behavior". Finally, our search patterns most likely vary according to the type of knowledge we are looking for, as well as other parameters. These are some of the trends that emerge from an NSF-sponsored research project I'm conducting with David Lazer and Ines Mergel, and more about this will be posted here soon. But let me now turn to you, my fellow bloggers (and whoever else feels like answering, of course!), and ask you to respond to this quick survey about your own search behavior:

1. What are the three types of questions that come up most frequently on your job?
2. How do you generally start off your search for information on these questions (in terms of source and medium)?
3. Does your search behavior vary according to the type of questions you have? How and why?

You might find that you engage in much more complex search behaviors than you realize. Or not. Be humorous about your answer, if you wish. Thanks for playing along!

Posted by Maria Binz-Scharf at 10:04 AM | Comments (6) | TrackBack

15 March 2007

The Venn diagram of "terrorist" and "network"

Alex's recent post on Dark Networks started me thinking about how and why network structures evolve. This turned into a conversation with my friend and colleague Jon Lindsey, a PhD candidate in Political Science at MIT and an intelligence officer in the Navy soon to deploy in Iraq. He suggested that the network form is actual quite fragile with respect to organizational forces: given the opportunity, many of these organizations will grow quite hierarchical with the standard bureaucracies we would expect to see else where.

Sure enough, from the Combating Terrorism Center at West Point comes a project called "Harmony and Disharmony: Exploiting Al-Qa'ida's Organizational Vulnerabilities". The full report (116 pp, pdf) makes for an interesting read, but I am fascinated by the translated original sources that are included. I was particularly drawn to the Al-Qaeda employment contract, with details as mundane as:

The bachelor Mujahed qualifies for a round trip ticket to his country after one year from
joining the organization. He can take a one month vacation. He doesn't get reimbursed if
the ticket is not used, but he has the right to change it to a ticket to perform
the pilgrimage. This period starts from the date of joining AL-QAEDA.

The married Mujahed and his family qualify for round trip tickets to their country of origin
after two years, and one month vacation. Tickets can not be reimbursed if unused.

This seems as bureaucratic as any employment agreement that I've signed, with the exception of the organizational goals. This document was discovered in Afghanistan, where the the administration of the organization appears to have become more mundane that the "4th Generation warfare" theory might suggest. Absent the constraints of a hostile, ubiquitous surveillance and law enforcement state, Al Qaeda starts to look a little like The Office.

At the same time, the demand for a network analysis approach for understanding and combating terrorism might be coming even more important. The Washington Post reports that

With new plots surfacing every month, police across Europe are arresting significant numbers of women, teenagers, white-skinned suspects and people baptized as Christians -- groups that in the past were considered among the least likely to embrace Islamic radicalism.

The demographics of those being arrested are so diverse that many European counterterrorism officials and analysts say they have given up trying to predict what sorts of people are most likely to become terrorists. Age, sex, ethnicity, education and economic status have become more and more irrelevant.

Absent the ability to use profiling to detect targets of interest, the type of data that the NSA was accused of collecting (albeit illegally) might be very useful for prevention and threat containment.

Posted by Allan Friedman at 11:44 AM | Comments (2)

14 March 2007

Social Networks and Communication Neworks

The University of Toronto’s NetLab has been doing some exciting research on how to measure social networks and communication behavior. Their recent conference paper, “Collecting Social Network Data to Study Social Activity-Travel Behavior: An Egocentric Approach,” discusses new methods of collecting data about social network, travel behavior, and the use of communication technologies. This is exciting research because it shows how to effectively measure two important elements of social life – the cognitive dimension of perceiving the existence of social ties, and the behavioral dimension of interaction that actually occurs with social ties. Moreover, this research incorporates multiple types of communication, including communication that occurs in-person, telephone, and email. The advantage of this approach is that rather collecting data about only certain kinds of ties or ways of interacting – such as the General Social Survey’s question about “those with whom you discuss important matters” – measuring both the cognitive and behavioral elements of social ties gives a more comprehensive understanding of the extent to which social life exists in America and how it actually occurs.

Posted by Jeff Boase at 12:30 AM | Comments (0)

13 March 2007

Small worlds-- the degrees of separation between Cambridge, MA, and Fargo, ND

I recently conducted a small world experiment in my networks class, selecting an individual in Fargo, North Dakota as the target recipient. I selected Fargo for much the same reason that Milgram selected a small town in Nebraska: psychologically and sociologically it is about as far away as from Cambridge, Mass, as is conceivable and still be in the US. My students were instructed to send an e-mail (with explanation) to someone they knew, who was supposed to send the e-mail to someone they knew, and so on, until it reached the target.

The results: for 25 students (and about 40 attempts) 4 chains were completed, plus a fifth chain completed from a split from one of the completed chains (i.e., one of the students hit the target twice with a single e-mail because one of the recipients down chain sent the e-mail to multiple others). The number of hops varied from 3 (a student in the class who happens to be from North Dakota), to 8), with an average of 5.25.

A number of interesting observations (some consistent with existing lit on small worlds—e.g., see Stanley Milgram, and Duncan Watts on small worlds, and John Kleinberg on navigation in small worlds):

1) These results confirm the basic intuition that we live in a small world (i.e., where a small number of degrees of separation is typical). Even for the majority of students who did not have completed chains, presumably there are a maximum of only 4 hops away from this individual in Fargo, since they have one classmate who is just three jumps away.

2) The navigation of the completed e-mails through the network was strikingly efficient, reflecting the crude but effective cognitive representations of participants of the macro- societal network. People used a variety of heuristics for choosing who to send the e-mail to: do I know someone from North Dakota? Do I know someone who is well connected and likely to know someone from the Midwest? Etc. The resulting paths were likely not the optimum paths, but couldn't have been far off. Even assuming that the world is “small” it is a remarkable (if understandable) thing that these e-mails could find a reasonably short path through the network. If only Boston roads were this navigable….

3) Information about the target improves the navigation through the network. To illustrate this, I varied the information that students were given—some were just told name and city of the target, and others were told name, city, and profession. Three of the four completed chains were for the second condition. Further, in the fourth case, someone along the way looked the guy up and incorporated information about where he worked into the e-mail. Nothing statistically significant, but notable.

4) The credibility of the message was essential in pushing it through the network. I suspect that if I had done this experiment 5 years ago, chain completions would have been higher. The vast majority of e-mails that people get now are junk. Further, everyone has received hoax e-mails forwarded on by acquaintances. One guesses that there was a concern not just that it was a hoax, but that it would be embarrassing to forward a hoax e-mail to someone else.

5) People relied on strong ties in sending the message. Interestingly, given the literature on the role of weak ties in disseminating information, when I polled students on whether they sent e-mails to close friends as compared to acquaintances, 90% of the students indicated that they sent the e-mail to close friends. This follows directly from point 4: their concern was to send it to someone who, in turn would forward it on.

6) Friends are helpful, but friends of friends far less so. Participants in chains were instructed to cc the originator, thus we have data on incomplete chains. Interestingly, almost everyone reported that the first person they sent the e-mail to forwarded it on (following from point 4); but there was a big (~50%) drop off at the next jump. My intuition is that this reflects more broadly on the epidemiology of information.

7) The act of “using” the network affects the structure of the network. In reading the chains of e-mails, it was striking to me how people used this e-mail to reconnect with someone—e.g., “I wouldn’t normally send this on, but it seemed like a good excuse to see how you were doing.”

Posted by David Lazer at 8:04 PM | Comments (8)

Enterprise Social Networking Software

IBM has announced to launch its Lotus Connections software in the first half of 2007 and Cisco buys the technology assets of tribe.net. It seems as if social networking software has become an important business line within large software vendors.

From a researcher's perspective it makes sense for firms to connect their employees through social networking software. Finding information, locating experts and spotting project relevant knowledge effectively are promises social software seems to able to hold. If not, why would people be interested in paying annual membership fees on platforms such as xing.com or linkedin.com.

At the same time, software vendors haven't got much to offer than whitepapers, prototypes, or other studies. A persistent question software vendors might be struggling with thus is: What is the USP of online social networking software why is it worth a client's effort to go through a massive data migration effort from several expert or knowledge management databases to a consistent social networking platform?

Here are some arguments/talking points that might help:
- Validation through Existing Models: The success of Xing.com and LinkedIn.com as two prominent examples of popular professional social networking platforms shows that managers and practitioners are willing to spend time and money in locating contacts, knowledge, and information within social networks
- Tie Characteristics and Performance: Studies in the management literature have shown that the characteristics of ties among managers and employees can have strong effects on the firm's or a managers performance (Hansen 1999, Moran 2005, Obstfeld 2005, references see below)
- Privacy Concerns: People are willing to publish their profiles online (as it can be observed on prominent Web 2.0//online social networking sites and as described by Ines Mergel in her prior blog). Hence, people are used to publishing their profiles online, have experience with it and might ranke the expected benefits higher than potential data privacy concerns.

Hence, why should making ties among people within firms visible NOT help these people to become more effective or productive over time?

Related Literature:
Hansen, M. T. 1999. The Search-Transfer Problem: The Role of Weak Ties in Sharing Knowledge across Organization Subunits. Administrative Science Quarterly, 44(1): pp. 82-111.
Moran, P. 2005. Structural vs. Relational Embeddedness: Social Capital and Managerial Performance. Strategic Management Journal, 26(12): 1129-1151.
Obstfeld, D. 2005. Social Networks, the Tertius Iungens Orientation, and Involvement in Innovation. Administrative Science Quarterly, 50: 100–130.

Posted by Thomas Langenberg at 8:09 AM | Comments (0) | TrackBack

12 March 2007

Microbias and Macroperformance

This is an abstract of todays PNG/CCCSN seminar with Daniel Diermeier (Kellogg School of Management, Northwestern University). We encourage you to discuss his presentation via comments on the blog.

"We use agent-based modeling to study collective problem solving in complex social networks where information aggregation and consensus building is modeled as the density classification task. We show that simple individual aggregation rules in conjunction with complex interaction patterns are highly efficient in solving the density classification task. We then investigate the effect of conservatism and partisanship on classification efficiency in large populations. We find that conservative agents enhance the populations’ ability to efficiently solve the density classification task despite large levels of noise in the system. In contrast, we find that the presence of even a small fraction of partisans holding the minority position will result in deadlock or a consensus on an incorrect answer. Our results provide a possible explanation for the emergence of conservatism and suggest that even low levels of partisanship can lead to significant social costs."

Here are the related publications:
Global Coordination in Modular Networks
Efficient system-wide coordination in noisy environments

Posted by Bernie Cahill at 11:59 AM | Comments (0)

10 March 2007

Dark Networks – The international network of the Red Army Fraction (RAF)

German researchers recently published an in depth analysis of the RAF, Germany’s terror group of the 1970s. While this topic has been studied well in the past, their work allows for deeper insights into the RAF’s international network. Apparently, the RAF was well connected to other European and Palestinian terrorist groups. They shared stolen weapons, received training in the Middle East or gained access to safe zones. For example, a hand grenade stolen by the RAF in a raid on a German based US Army depot was later used by Carlos, the Venezuelan terrorist. Cooperating Palestinian groups were even secretly supplied by the KGB officially endorsed by Breschnew, the former chairman of the communist party. Despite that, from ideological point of view RAF’s cooperation with the Palestinians was somewhat paradox as the RAF justified their actions with antifascism or members called for sympathy/support of the "left" with Israel in the Six-Day war.

Unfortunately, the two volumes called “Die RAF und der linke Terrorismus (The RAF and the left wing terrorism)” by Kraushaar (Ed.) are only available in German at the moment.

Posted by Alexander Schellong at 12:04 AM | Comments (0) | TrackBack

7 March 2007

Bayesian Models of Social Networks and Text with Application to Political, Legal and Bibliometric Data

This is an abstract of this weeks PNG/CCCSN seminar with Andrew McCallum (University of Massachusetts, Amherst). We encourage you to discuss his presentation via comments on the blog.

"The field of social network analysis studies mathematical models of patterns in the interactions between people or other entities. In this talk I will present several recent advances in generative, probabilistic modeling of networks and their per-edge attributes. The Author-Recipient-Topic model discovers role-similarity between entities by examining not only network connectivity, but also the words communicated on on those edges; I'll demonstrate this method on a large corpus of email data subpoenaed as part of the Enron investigation. The Group-Topic model discovers groups of entities and the "topical" conditions under which different groupings arise; I'll demonstrate this on coalition discovery from many years worth of voting records in the U.S. Senate and the U.N. I'll conclude with further examples of Bayesian networks successfully applied to relational data, as well as discussion of their applicability to trend analysis, expert-finding and bibliometrics."

Here is a link to Andrew's talk: "Bayesian Models of Social Networks and Text with Application to Political, Legal and Bibliometric Data"

Posted by Bernie Cahill at 10:15 AM | Comments (0)

1 March 2007

Control Your Online Public Profile Using Social Networking Platforms

I just taught a segment in David Lazer's Social Networking class at Harvard on how people can analyze and visualize their social networks. David invited the whole class to join him on LinkedIn and we noticed that a couple of students were hesitant to join due to security concerns. We have a very mixed audience of MPP, MPA, Midcareer and PhD students from all kinds of different industries - some of them from the military and security area. One of the students asked me: "Can you give me one good reason why I should join any of the social networking sites?" - given the background and affiliations of some of the students, I couldn't come up with an argument why people should join - on the contrary I understand that some people need to keep a low public profile, so that not too much of their private information or details about their CV will become publicly available.

So I started to think about what are reasons why I have all my information uploaded to all kinds of websites? I have a Flickr page, an openBC/Xing profile, a LinkedIn profile, a personal website, a corporate website and post on my own blog and on our blog at the Kennedy School. Am I too open to give away this much information? On the other hand, I am not working in the military or security area, right?

It turns out that there are ways to control what people can find out about you. I talked with Bill Liao, the co-founder of Xing (formerly openBC) about this issue and he pointed me to the people finder search engine ZoomInfo. It is a search engine that gives summaries of people (Find tab) or let's you create a more detailed profile online, so that recruiters, etc. can find you easier (BeFound tab). Controlling what you actually want other people to find about you comes with a price: pro version for $49/month. But it is definitely one way to control what information can be found about you and also a way to manipulate your online information.

I tried it and was surprised about the result (Remember, I have a at least seven different pages directly connected with my name where I actively produce content). Here is the result:

View image

So there are four entries - one with the direct link to my Kennedy School subpage, but the others are from older sources tracking some of my (past) academic activities. That's about it. Google on the other hand finds 13.200 different entries.

Another way of controlling what is found by Archive.com or Google seems to be to ask thems to take down some of your indexed information and not display it when people search for your name.

What are your thoughts on the dangers of having your information publicly available on social networking platforms? Are there any measures you take to avoid having too much information available for the rest of the world?

(hm... guess I just created another piece of publicly available information)

Posted by Ines Mergel at 2:27 PM | Comments (1)