Self-Control

By Chris Fariss

I found the following post at Chris Blattman‘s blog; however, the original post is from yet another blog, The Frontal Cortex.  Is this cross-blogging or blog-crossing?  Anyway, I think the experiments will be of interest to our group.  Enjoy.

For the most part, self-control is seen as an individual trait, a measure of personal discipline. If you lack self-control, then it’s your own fault, a character flaw built into the brain.

However, according to a new study by Michelle vanDellen, a psychologist at the University of Georgia, self-control contains a large social component; the ability to resist temptation is contagious. The paper consists of five clever studies, each of which demonstrates the influence of our peer group on our self-control decisions.

For instance, in one study 71 undergraduates watched a stranger exert self-control by choosing a carrot instead of a cookie, while others watched people eat the cookie instead of the carrot. That’s all that happened: the volunteers had no other interaction with the eaters. Nevertheless, the performance of the subjects was significantly altered on a subsequent test of self-control. People who watched the carrot-eaters had more discipline than those who watched the cookie-eaters.

Be sure to check out the many other interesting posts at Chris Blattman‘s blog.  As for The Frontal Cortex, this was my first visit to the site but it might be worth exploring a bit more.

Slime Mold Leaves Urban Planners Unemployed

By Chris Fariss

The abstract of a recently published report in Science:

Transport networks are ubiquitous in both social and biological systems. Robust network performance involves a complex trade-off involving cost, transport efficiency, and fault tolerance. Biological networks have been honed by many cycles of evolutionary selection pressure and are likely to yield reasonable solutions to such combinatorial optimization problems. Furthermore, they develop without centralized control and may represent a readily scalable solution for growing networks in general. We show that the slime mold Physarum polycephalum forms networks with comparable efficiency, fault tolerance, and cost to those of real-world infrastructure networks—in this case, the Tokyo rail system. The core mechanisms needed for adaptive network formation can be captured in a biologically inspired mathematical model that may be useful to guide network construction in other domains.

Atsushi Tero, Seiji Takagi, Tetsu Saigusa, Kentaro Ito, Dan P. Bebber, Mark D. Fricker, Kenji Yumiki, Ryo Kobayashi, Toshiyuki Nakagaki. 2010. “Rules for Biologically Inspired Adaptive Network Design” Science Vol. 327. no. 5964, pp. 439 – 442 DOI: 10.1126/science.1177894

Human Rights as a Latent Variable

By Chris Fariss

Keith Schnakenberg and I are working on a paper in which we measure the unobservable level of respect for human rights.  We use the same Bayesian model that Shawn Treier and Simon Jackman use to measure the latent level of democracy in their 2008 paper that was published in the American Journal of Political Science.   As with the construction of GRE scores, the ordinal item-response (IRT) model explicitly models the measurement error that results when different component indicators are aggregated together.

Posterior densities for item discrimination parameters for individual physical integrity rights (300 draws). The item discrimination parameter represent the degree to which the item discriminates between states' along the latent human rights variable. Greater values along the x-axis signify greater discrimination by the item.

The data we use to estimate the IRT model is available from the CIRI Human Rights Data Project.  The model allows us to generate point estimates and credible intervals for the latent variable of interest.

Latent variables estimates for all 192 countries in the CIRI dataset in the year 2007. Blue dots are point estimates (posterior means) and red lines are 95% credible intervals.

For those interested, Simon Jackman has posted several slide-shows on his website that demonstrate the IRT model in action.  In our paper we also demonstrate a simple way to include the uncertainty from the estimates in models that include such a measure as an independent variable.  You can download a copy of our paper at SSRN (we will upload the paper soon).  For now, here is the current version of our abstract:

Recent Abstracts

By Darren Schreiber

In this set of abstracts, we get some interesting evidence on why high sensation seekers might be drawn to scary movies (1) and a series of methodological innovations that demonstrate how new advances are going to improve our ability to get useful insights from MRI data (2, 5, 7, 8, 9, 114). We also have some new work illuminating the role of stereotyped racial judgments (which rely on the amgydala) and individuate judgments which use a broad network of regions connected with mentalizing (3). Another article shows that there is variation in the deactivations of some regions of the default mode network in particular types tasks (4). We also learn that the default mode network contains the most heavily functionally connected brain regions, with the posterior cingulate being particular well connected to the rest of the brain (10).

Another article melds structural and functional connectivity data to provide an exciting methodological innovation (11). I have long hoped that brain imaging statistical analysis would start to do better at accounting for our priors about how the brain is organized, specifically that techniques would be developed to take advantage of our knowledge of structure while we try to understand function. This paper is a great step in this direction and I know that the developers of the FSL package are also making strides towards this goal.

On the topic of connectivity, we have a study that used a structural connectivity technique to demonstrate that the amgydala appears to have at least three distinct subregions (5). The amgydala is also shown to vary in size in non-medicated bi-polar patients, controls, and medicated bipolar patients (6).

Rounding out the remainder, we have a nice paper that provides a possible explanation for why the anterior cingulate appears to be active in reward activity in non-human primates, but seems to be mostly involved in error prediction in humans (12). Another paper demonstrates an important lesson for economists about the distinction between monetary and social rewards (13). While we may anticipate them using the same region (striatum), consumption of those rewards implicates distinct regions. We also have some insight into mental fatigue that confirms the importance of rest and perhaps provides a way of understanding why we feel so depleted after a hard day of academic brain busting (14). The next to last paper tries to synthesize two literatures suggesting alternative functions for the anterior temporal lobe, suggesting that perhaps it is our attempt to integrate conceptual understanding of our social tasks that is activating the region (15). And finally, we’ve got a paper looking at a potential genetic role for face recognition and a demonstration that it is independent of general intelligence (16).

Reporting the News Using Only Twitter and Facebook as Sources

By Jason J. Jones

Five journalists have volunteered to be locked in a house with no access to the outside world except Facebook and Twitter. Their task will be to report the day’s news as best they can using only these sources.

The Washington Post and Ars Technica are optimistic.  Personally, I wouldn’t be surprised if the stories they report end up a little garbled.

Altruism: Reciprocity v. Kinship

by Jaime Settle

A Nature News article highlights the recent finding that reciprocity explains more of the variance in grooming behaviors between primate species than does kinship:

Deriving Social Networks from Mobile Phone Data: Benefits and Challenges

By Jaime Settle

I’m going to present two articles at our meeting next week that both generate social network structure and interaction using data collected from mobile phones.

The Onnela et al. 2007 paper uses the call records from over 4,000,000 people to create a network which includes about 20% of a country’s entire population. The authors explore the relationship between local and global network topology and teh spread of information through the network, concluding that:

“Taken together, weak ties appear to be crucial for maintaining the network’s structural integrity, but strong ties play an important role in maintaining local communities. Both weak and strong ties are ineffective, however, when it comes to information transfer, given that most news in the real simulations reaches an individual for the first time through ties of intermediate strength” (p. 7336)

The Eagle, Pentland and Lazer 2009 paper uses proximity data generated from software embedded in cell phones to track the behavior of 94 students and faculty at a research university. They find that the observational data on proximity can better predict job satisfaction and friendship formation than respondent self-report.

Both articles represent major advances in our ability to collect data on social networks, however, they also raise interesting questions about what we are actually measuring as links between the nodes. I hope our discussion of the contributions and drawbacks of these two approaches will help us think through the operationalization of our own network research.

Microfinance & Social Capital

by Michael Rivera

This is the cool project Alex Hughes mentioned he’s working on.  Enjoy!

Union MicroFinanza (UMF).  Hughes’ brother is running the organization.

The motivating idea behind the organization is that microfinance companies often give group based loans, using social capital as collateral on the loans, but they do it without a clear idea of a.) what social capital is, or b.) how to measure social capital within the target community.  Union MicroFinanza is making an attempt at measuring the social capital in the towns it provides loans to by measuring the social structure.  It is sort of as if UMF is appraising the collateral it takes on the loans; if a town is found to have very weak measures of social capital, UMF can alter its lending practices to ensure full recovery of its capital, and if a town is found to have strong measures of social capital, which Putnam would say is worth more, UMF can make larger investments into the town.

Friendship and marriage ties in 500 person town.  Friendship ties are in black.  Marriage ties in blue.

Problems inherited in modularity-maximization approaches

By Yunkyu Sohn

Since the publication of a landmark paper, which initially introduced an objective measure of good partition, called modularity, modularity maximization approach as a community partition method has been favored by scientists in various disciplines. However, recently, theoretical studies have discovered several serious fundamental problems inhere in the method.

I would like to highlight two major problems among them: i) Resolution limit in community detection: Fortunato and Barthelemy validate that modularity maximization algorithms “fail to identify modules smaller than a scale which depends on the total size of the network and on the degree of interconnectedness of the modules, even in cases where modules are unambiguously defined.” ii) Rugged modularity landscape and picking up a single local optimum: Since a modularity maximization task is known as an NP-hard problem, we use several types of heuristics (spectral, greedy, simulated annealing, random walk or genetic, and etc.) to obtain a single community assignment vector whose modularity value is believed to be located near the global maximum. This approach is based upon an assumption that our objective function (Q(x) where Q is modularity resulted by x community assignment vector) has a single peak so that several community assignment vectors near this peak exhibit very similar community assignments, and thus it is OK to select one of them to know about the best partition (Massen and Doye). However think of a case where the objective function has two peaks (global maxima) at very different locations. In this case, even though we find out a community assignment vector which results a modularity value near the global maximum, the vector doesn’t contain any information about a partition which also results maximal modularity at another location. Unfortunately, several studies have proven that many popular network structures possess very rugged landscapes around the top of their modularity functions (Good et al.; Reichardt and Bornholdt; Sales-Pardo et al.). In other words, the modularity landscapes of those networks have many near-global-optima at various places and we cannot know where our algorithm locates its solution.

A family of solutions (Sales-Pardo et al.; Sawardecker et al.) for this problem use stochastic searching (ex. simulated annealing or random walk) to produce various near-optima community assignment vectors and aggregate the information within these vectors. This approach has an advantage of keeping the information of relevant community assignment vectors having diverse characteristics, yet uses very complicated and lengthy computations to reach the final community assignment vector. Another group of solutions throw away modularity measures and build up noble approach to reveal the best partition (Rosvall and Bergstrom Rosvall and Bergstrom b). The most successful algorithm of this type uses random walks (Rosvall and Bergstrom b) and is shown to work well for several types of artificial networks (Lancichinetti and Fortunato). However since this algorithm dumps the authority of modularity measure, we cannot have objective criteria on a better community assignment vector without prior knowledge for node categories.

The statistical physics community has just started to think about above problems seriously, and no definite answer has been introduced yet. We may have to care much about the stability of our community assignment vectors and refine the available methods to obtain better criteria adjusted for the purpose of the study.

SPSP Political Psychology Pre-Conference.

by Jason J. Jones

A one-day Political Psychology conference will take place in Las Vegas on the opening day of the Society for Personality and Social Psychology Annual Conference.

If anyone wants to go, I was planning on driving up there for SPSP anyway, and you are welcome to join me.  I even have an extra room in the timeshare I have for the weekend.  The pre-conference is on Thursday, Jan. 28th, and SPSP will continue on Friday and Saturday.  (Sunday we break the bank at the blackjack tables.)

Facebook Resources

By Jaime Settle

As is the norm for academic research on novel technologies and phenomena, it seemed like we waited ages for quality research about Facebook using user-generated data instead of self-reported usage data. I was recently steered toward two sites that may be useful for staying informed about the latest trends and data analysis using Facebook:

Facebook Data Team, a group of researchers at Facebook who handle the collection, management, and analysis of Facebook data

Overstated, a blog by Cameron Marlow, an in-house researcher at Facebook whose research focuses on “various aspects of online communities including the diffusion of information across online social networks, access to information and social capital, and the incentives that impact social media production.”

These sites might generate interesting new ideas and reveal relevant trends that could improve our exploration of the effects of Facebook on political behavior and the transmission of political information. For example, it seems that there was a trend among Facebook users who supported the idea of health care reform to post a very specific status message this fall, leading to a sharp increase in the use of the word “health care” in status messages (much beyond the increase observed during the 2008 election.

The online network of the Underwear Bomber

By Darren Schreiber

The Computational Legal Studies blog has a visualization of the network of online activity of the underwear bomber:

Visualizing the Gawaher Interactions of Umar Farouk Abdulmutallab, the Christmas Day Bomber from Computational Legal Studies on Vimeo.

James on the Colbert Report

By Darren Schreiber

Our very own James Fowler was on the Colbert Report on Thursday night.  Congratulations James!

Recent Abstracts (Part 2 of 2)

By Darren Schreiber

This selection contains some critical new work on extinguishing fear (18, 19, 20) that will be crucial for treating the veterans returning from our many conflicts overseas and for reducing the huge social costs of PTSD. Elizabeth Phelps has been doing some truly impressive work over the years, but this might turn out to be one of her most interesting discoveries. I’m also fascinated by the possibility the treating depression with SSRI’s implicates changes in personality, not merely mental illness (6). Given the role of personality in both political behavior and in general well-being, this work has some interesting social implications. It is also heartening to find additional evidence that measures of subjective well-being correspond well to objective indicators (9). And, since I always relish a nice hike in one of San Diego’s many fantastic parks, I’m glad to know that it might be having a positive effect on my generosity and seeking out of intrinsic rewards (13). On a darker note, the idea that insurgency has some predictable elements to it is an intriguing possibility (3, 4).

I was also interested to read about a new book on the coming era of “big data” (8) and since it was free online I’ve already downloaded and read some of the chapters that looked interesting. On the topic of the future of science, Nature’s asked some scholars to speculate about where their fields will be in a decade (17) and some exciting examples for future work in observing behavior (16) and modeling (15) are presented. I was also fascinated to read more about the possibility that the role of a gene might depend on who you inherited it from (12). When a colleague at the Santa Fe Institute was telling me about this over the summer, I was incredulous.

Recent Abstracts (Part 1 of 2)

By Darren Schreiber

There have been so many interesting things published since my last update that I decided to split the list into two emails. In this collections, we’ve got the gamut from genes to neurotransmitters to brain to behavior. Crisan (1), Long (7), and Zak (13) show that neurochemistry (with serotonin and testosterone) alter economic decision making and Gan (20) provides evidence of the role for dopamine in encoding rewards. Decision-making is further illuminated with work on the nucleus accumbens (2), the temporo-parietal junction (12), and the orbitofrontal cortex (15). The insula and its roles in both internal (18) and social (3) perceptual process are also explored. Other highlights include discussions of the neural architecture of social reasoning (9), pain (10), attention (17), and Bayesian inference (19). I also think that the implications of new work on psychopaths (8) is fascinating as we try to synthesize our intuitions about legal and moral responsibility with work on the neuropathology that underpins this disease.

Stuff you should know

by Robert Bond

When I was back home over the holidays a friend pointed me toward the Stuff You Should Know podcast. The podcast is, apparently, consistently one of the top 10 downloaded from iTunes. I checked it out and they are pretty interesting and entertaining. What really caught my eye, though, was a podcast from a while back about whether or not quitting smoking is contagious. While they discuss research from this paper, they don’t always get all the details right (they call the city where the data were collected “Farmington” instead of Framingham, for instance), but it is entertaining and informative nonetheless. They also have discussions of altruism, the ultimatum game, and pirates, among many other things.