A Structural Learning Model of Cultural Polarization


“Why on earth," asks Steven Pinker in The Blank Slate, "should people's beliefs about sex predict their beliefs about the size of the military? What does religion have to do with taxes? Whence the linkage between strict construction of the Constitution and disdain for shocking art?” A simple toy model of opinion cascades on a dynamic network suggests an answer: People are attracted to those with similar cultural preferences, repelled by those with salient differences, and are positively and negatively influenced by those to whom they are attracted/repelled. The model consists of 100 fully connected agents with initially random traits on 10 cultural dimensions. The axes displayed above are the two principle eigenvectors from the 10-dimensional opinion matrix. The size of the circle indicates the number of agents with that cultural profile. Multiple runs show how the dynamics of social influence, homophily, and xenophobia can lead to cultural polarization, such that agents at each pole have identical traits on all ten cultural dimensions and are maximally dissimilar to the "outgroup" at the opposite pole. Thus, from a random start, all 45 pairwise cultural dimensions come to be highly correlated, but the signs of the correlations are in all cases entirely arbitrary. Incomplete polarization is also possible, with agents aligned on a small number of cross-cutting dimensions. The model's results were recently confirmed in a laboratory experiment with human subjects.





Michael W. Macy
Goldwin Smith Professor of Arts and Sciences
Department of Sociology
Department of Information Science
Director, Social Dynamics Laboratory
Cornell University
Ithaca, NY 14853
Voice: (607) 255-4555
Fax: (607) 202-4913
Email: mwm14@cornell.edu




Click here for Nature feature story by Jim Giles on computational social science at Cornell

Click here for Oct 2011 "Naked Scientist" interview on BBC

Click here for Seti Institute's "Big Picture Science" interview on Public Radio

Click here for March 2014 NY Times Op Ed "The Science of Paying It Forward"

Click here for Guardian article on "Partisan Consumption of Science"

Click here or here for "Science Friday" interview

Click here for our lifestyle-politics.com website



Why does our world not degenerate into the world of Mad Max -- and why does it sometimes seem as if it may? Social order can be imposed "from above" by a global policing mechanism or it can emerge "from below" through local interactions with no centralized coordination. Suppose no member of the population has the ability to identify or impose a global solution. How then is social order possible? My research explores how norms, opinions, emotions, and collective action emerge and spread through local interaction. The motivating problem is one that defines the human condition: the overwhelming need for norms that constrain aggressive and mutually destructive behavior is no guarantee that such rules will emerge or be obeyed by anyone except a few "suckers." On the contrary, norms can even make matters worse, by obligating people to engage in behaviors that are individually and collectively harmful (e.g., pressuring one's neighbors to cheer a "naked emperor"). Our research team uses computational models, laboratory experiments with human subjects, and "big data" from online networks to look for elementary principles of social interaction that may yield clues about possible answers to these and related puzzles about human behavior and social interaction. A wine-tasting study with Robb Willer and Ko Kuwabara showed that participants not only rated a vinegar-tainted wine over one that was untainted but also publicly rated fellow wine tasters higher who agreed. However, when  the ratings were private, participants revealed their true beliefs. The public pressure  for "false enforcement" is a "complex  contagion" that  depends on social reinforcement from multiple neighbors. Follow-up studies with Damon Centola, Vladimir Barash, and Chris Cameron showed how complex contagions depend initially on dense local structure but once the contagion achieves a critical mass, diffusion can leverage the long-range ties characteristic of small world networks. In another small-worlds study (with Nathan Eagle and Rob Claxton) using a nearly complete set of telephone logs from the UK we found that small worlds are associated with economic development. A related study with Patrick Park and Joshua Blumenstock used global online phone and social media networks to reveal "social wormholes" that create high-bandwidth connections across vast network distances. 

These studies used digital traces of behavior to observe global patterns that would be difficult or even impossible to measure using  conventional sources of observational data such as surveys. For  example, Scott Golder and I analyzed the affective content of global Twitter messages to reveal remarkable similarity in diurnal and seasonal mood variations across diverse cultures. A follow-up study with Minsu Park and colleagues at Spotify used global streaming logs to reveal similar patterns in the emotions people experience through music. Recent research has focused on political polarization. Daniel Della-Posta, Yongren Shi and I wondered why liberals drink lattes, and a study with Bill Shi, Yongren Shi, Fedor Dokshin and James Evans shows how polarization extends even to the consumption of books on popular science.

I have listed below some of my papers that you are welcome to download. 

Michael Macy, Sebastian Deri, Alex Ruch, and Natalie Tong
Science Advances, August, 2019

The Strength of Long-Range Ties in Population Scale Social Networks
Patrick Park, Joshua Blumenstock, and Michael Macy
Science, 362: 1410–13, 2019

Global Music Streaming Data Reveals Diurnal and Seasonal Patterns of Affective Preference
Minsu Park, Jennifer Thom, Sarah Mennicken, Henriette Cramer, and Michael Macy
Nature Human Behaviour, Dec. 2018

The Opacity Problem in Social Contagion
George Berry, Christopher Cameron, Patrick Park, and Michael Macy
Social Networks 56:93-101, 2019

Peer effects on adolescent smoking: Are popular teens more influential?
Juan David Robalino and Michael Macy
Plos One 7/12/2018

Cultural Fault Lines and Political Polarization
Yongren Shi, Kai Mast, Ingmar Weber, Agrippa Kellum, and Michael Macy
Proceedings of the 2017 ACM on Web Science, pp. 213-217

Bots as Virtual Confederates: Design and Ethics
Peter Krafft, Michael Macy, Sandy Pentland
Proceedings of CSCW 2017


The Local Dynamics of Institutional Change
Chris Cameron and Michael Macy
Rationality and Society 2017

Millions of Online Book Co-purchases Reveal the Politicization and Polarization of U.S. Science
F. Bill Shi, Yongen Shi, Fedor Dokshin, James Evans and Michael Macy.
Nature Human Behaviour 2017


Automated Hate Speech Detection and the Problem of Offensive Language
Tom Davidson, Dana Warmsly, Ingmar Weber, and Michael Macy.
Proceedings of the ICWSM 2017.


Complex Contagions and the Diffusion of Popular Twitter Hashtags in Nigeria
Clay Fink, Aurora Schmidt, Vlad Barash, Chris Cameron, and Michael Macy.
Social Network Analysis and Mining, December 2016, 6:1


Investigating the Observability of Complex Contagion in Empirical Social Networks
Clay Fink, Aurora Schmidt, Vlad Barash, John Kelly, Chris Cameron, and Michael Macy
Proceedings of ICWSM 2016


Measuring Structural Similarity in Large Online Networks
Yongren Shi and Michael Macy
Social Science Research 2016

Embracing Cultural Diversity: Online Social Ties in Distributed Workgroups
Wei Dong, Kate Ehrlich, Michael Macy, and Michael Muller.
Proceedings of CSCW 2016

The Center Cannot Hold
Daniel DellaPosta and Michael Macy
In Order on the Edge of Chaos, Cambridge University Press, 2016.

Why Do Liberals Drink Lattes?
Daniel Della Posta, Yongren Shi, and Michael Macy
American Journal of Sociology, March, 2015

The Mesh of Civilizations in the Global Network of Digital Communication
Bogdan State, Patrick Park, Ingmar Weber, and Michael Macy
PLOS ONE, May 29, 2015

The Social Contagion of Anti-Social Behavior
Milena Tsvetkova and Michael Macy
Sociological Science, Feb. 10, 20


Big Theory: A Trojan Horse for Economics?
Michael Macy
Review of Behavioral Economics 2014


The Signal Importance of Noise
Michael Macy and Milena Tsvetkova
Sociological Methods and Research 2015


Digital Footprints: Opportunities and Challenges for Social Research
Scott Golder and Michael Macy
Annual Review of Sociology, Vol. 40, 2014


The Social Contagion of Generosity
Milena Tsvetkova and Michael Macy
PLOS One Feb. 13, 2014: 9(2): e87275


Statistical Mechanics and Social Sciences
Santo Fortunato, Michael Macy, Sidney Redner
Journal of Statistical Physics 2013, 151:1-8


Critical Phenomena in Complex Contagions
Vladimir Barash, Chris Cameron, and Michael Macy
Social Networks 2012, 34:451-61


Diurnal and Seasonal Mood Vary with Work, Sleep and Daylength Across Diverse Cultures
Scott Golder and Michael Macy
Science 2011, 333:1878-81


Small Worlds and Cultural Polarization
Andreas Flache and Michael Macy
Journal of Mathematical Sociology 2011, 34: 146-76


Local Convergence and Global Diversity: From Interpersonal to Social Influence
Andreas Flache and Michael Macy
Journal of Conflict Resolution 2011, 55: 970-95


Network Diversity and Economic Development
Nathan Eagle, Rob Claxton, and Michael Macy
Science 2010, 328: 1029-1031


Computational Social Science
David Lazer et al.
Science 2009, 323: 712-723


The False Enforcement of Unpopular Norms
Rob Willer, Ko Kuwabara, and Michael Macy
American Journal of Sociology 2009, 115:451-490


Neighborhood Chance and Neighborhood Change
Arnout Van de Rijt, David Siegel, and Michael Macy
American Journal of Sociology 2009, 114:1166-80


Complex Contagions and the Weakness of Long Ties
Damon Centola and Michael Macy
American Journal of Sociology 2007, 113:702-34


Culture, Identity, and Structure in Social Exchange: A Web-based Trust Experiment in the U.S. and Japan
Kuwabara, K., R. Willer, M. Macy, R. Mashima, S. Terai, and T. Yamagishi
Social Psychology Quarterly, 2007, 70:461-79


Collective Action and the Empirical Content of Stochastic Learning Models
M. Macy and A. Flache
American Journal of Sociology, 2007, 112: 1546-54


Cascade Dynamics of Complex Propagation
Damon Centola, Victor M. Eguiluz, and Michael Macy
Physica A 2007, 374: 449-456


Power and Dependence in Intimate Exchange
Arnout van de Rijt, and Michael Macy
Social Forces 2006, 84:1455-70


Ethnic Preferences and Residential Segregation: Theoretical Explorations Beyond Detroit
Michael Macy and Arnout van de Rijt
Journal of Mathematical Sociology 2006, 30: 275-88


The Emperor's Dilemma: A Computational Model of Self-Enforcing Norms
Damon Centola, Robb Willer, and Michael Macy
American Journal of Sociology 2005, 110:1009-40


Social Life in Silico: The Science of Artificial Societies
Damon Centola and Michael Macy
Handbook of Group Research and Practice 2005, pp. 273-2812


Power, Identity, and Collective Action in Social Exchange
Brent Simpson and Michael Macy
Social Forces, June, 2004.


Polarization in Dynamic Networks: A Hopfield Model of Emergent Structure
Michael Macy, James Kitts, Andreas Flache, and Steve Benard
Dynamic Social Network Modeling and Analysis, National Academy Press, 2003


Learning Dynamics in Social Dilemmas
Michael Macy and Andreas Flache
Proceedings of the National Academy of Sciences, May 14, 2002.


Stochastic Collusion and the Power Law of Learning
Andreas Flache and Michael Macy
Journal of Conflict Resolution, October, 2002


From Factors to Actors: Computational Sociology and Agent-Based Modeling
Michael Macy and Robert Willer
Annual Review of Sociology, Vol. 28, 2002


Trust and Market Formation in the U.S. and Japan
Michael Macy and Yoshimichi Sato
Proceedings of the National Academy of Sciences, April, 2002


Social Simulation
Michael W. Macy
In N. Smelser and P. Baltes, eds., International Encyclopedia of the Social and Behavioral Sciences, Elsevier, 2002


Dependence and Cooperation in Fuzzy Dilemmas: The Effects of Environmental and Endowment Uncertainty
R. Thomas Boone and Michael W. Macy
In R. Suleiman, D. Budescu, & D. Messick, eds., ContemporaryPsychological Research on Social Dilemmas
Cambridge University Press, 2002.


'In Search of Excellence': Fads, Success Stories, and Adaptive Emulation*
David Strang and Michael Macy
Best Paper Proceedings of the 1999 Academy of Management Conference, Chicago,IL
American Journal of Sociology, July, 2001.
*HTML Preprint not identical to published version


Collective Action and Power Inequality: Coalitions in Exchange Networks*
Brent Simpson and Michael Macy
Social Psychology Quarterly, March, 2001.


Structural Learning: Attraction and Conformity in Task-Oriented Groups
James Kitts, Michael Macy, and Andreas Flache
Computational and Mathematical Organization Theory, 1999, vol. 5(2):129-45.


Unlocking the Doors to Prisoners Dilemma: Dependence, Selectivity, and Cooperation
R. Thomas Boone and Michael Macy
Social Psychology Quarterly 1999, 62: 32-52.


The Evolution of Trust and Cooperation between Strangers: A Computational Model*
Michael Macy and John Skvoretz
American Sociological Review, October, 1998
Presented at the Sante Fe Institute, August 6, 1996


Social Order and Emergent Rationality
Michael Macy
In A. Sica, ed. Whatis Social Theory: The Philosophical Debates, 1998, Blackwell.


Dependence and Cooperation in the Game of Trump
R. Thomas Boone and Michael Macy
Advances in Group Processes, vol. 15, 1998


Social Order in an Artificial World
Michael W. Macy
Journal of Artificial Societies and Social Simulation, January, 1998.


The Weakness of Strong Ties II: Collective Action Failure in a Self-Organizing Social Network
Michael Macy , James Kitts, and Andreas Flache
Presented at American Sociological Association, Toronto, August 11, 1997.


Identity, Interest, and Emergent Rationality: An Evolutionary Synthesis
Michael Macy
Rationality and Society, vol. 9, 1997.


The Weakness of Strong Ties: Collective Action Failure in a Highly Cohesive Group
Andreas Flache and Michael W. Macy
Journal of Mathematical Sociology, June, 1996


Natural Selection and Social Learning in Prisoner's Dilemma: Co-adaptationwith Genetic Algorithms and Artificial Neural Networks
Michael W. Macy
Sociological Methods and Research, Vol 25, August, 1996, pp. 103-137


Beyond Rationality in Models of Choice
Michael W. Macy and Andreas Flache
Annual Review of Sociology, Vol. 21, 1995


PAVLOV and the Evolution of Cooperation: An Experimental Test
Michael W. Macy
Social Psychology Quarterly, June, 1995


Artificial Social Intelligence
William Bainbridge, Edward Brent, David Heise, Michael Macy, Barry Markovsky, & John Skvoretz
Annual Review of Sociology, Vol. 21, 1995


Once Upon a Time There Was a Suboptimal Equilibrium
Michael W. Macy
The Agora, June, 1996


Cowardly Lions: Genetic Programming or Social Learning?
Michael W. Macy
The Agora, December, 1995


Social Class
Michael W. Macy
The Encyclopedia of Language and Linguistics


Backward-Looking Social Control
Michael W. Macy
American Sociological Review, 1993, Vol. 58:819-36.


Chains of Cooperation: Threshold Effects in Collective Action
Michael W. Macy
American Sociological Review, 1991, Vol. 56:730-47.


Learning to Cooperate: Stochastic and Tacit Collusion in Social Exchange
Michael W. Macy
American Journal of Sociology, 1991, Vol. 97:808-43.


Learning Theory and the Logic of Critical Mass
Michael W. Macy
American Sociological Review, 1990, Vol. 55:809-26.


Value Theory and the Golden Eggs: Appropriating the Magic of Accumulation
Michael W. Macy
Sociological Theory, 1988, Vol. 6: 131-52.


Subpages (1): .