2012 Preconference

 

PRECONFERENCE:  Dynamical Systems and Computational Modeling in Social Psychology

Society for Personality and Social Psychology Meeting

San Diego, CA on Thursday, Jan. 26, 2012.

 

Stephen J. Read (University of Southern California) and Robin Vallacher (Florida Atlantic University)

Email: read@usc.edu

We feel that the time is right to focus the field’s attention on the application of concepts and methods from complexity science, nonlinear dynamical systems, and computational modeling to social psychological processes.  This perspective shows signs of emerging as a major paradigm for the field, with many psychologists developing interesting and innovative lines of theory, empirical research, and computational models.  Because the approach is still relatively new to the field, a wide variety of specific models and methodologies have been developed, each with a somewhat unique set of assumptions and theoretical concerns.  Nonetheless, there are some common denominators, one of which is an emphasis on the emergence of coherence in systems composed of many elements, which can be observed at different levels of human experience (e.g., judgment, affect, self-concept, social relations, group-level processes).  The preconference, then, will showcase both the diversity of approaches—e.g., agent-based models, parallel constraint satisfaction in neural networks, identification of temporal patterns and attractors in time-series data—and the common theme of complexity and coherence for a wide variety of topics. 

We currently have 10 individuals who have agreed to be speakers in this inaugural preconference.  Here is the program. A title and a brief description follows for each talk. 

 

PROGRAM

Time

Title

Speaker

8:30

Introduction

R. Vallacher & S. J. Read

9AM

A Neural Network Model of Motivation and Personality

S. J. Read, V. Droutman, G. Chopra, & L. C. Miller

9:30

Increasing the predictive validity of neural network models of person judgments

B. Monroe, T. Laine, S. Gupta, & I. Farber

10AM

Coffee Break

 

10:30

A Dynamical Systems Approach to Person Construal

J. B. Freeman & N. Ambady

11AM

Mental Dynamism and its Constraints: Finding Patterns in the Stream of Consciousness

R. R. Vallacher & J. L. Michaels

11:30

The intra-individual dynamics of health behavior.

M. Orr, R. Thrush, & D. C. Plaut

Noon

LUNCH

LUNCH

1PM

The Coherence Effect: Blending Cold and Hot Cognitions by Constraint Satisfaction

D. Simon, D. Stenstrom, & S. J. Read

1:30

Emergence and contagion in dynamic network systems

J. D. Westaby

2PM

Evolutionary approaches to group behavior: A producer-scrounger dilemma and its consequences for collective adaptation.

T. Kameda & R. Hastie

2:30

COFFEE BREAK

SPSP registration

3PM

Network Clustering and Group Segregation Effects on Defensive and Non-Defensive Rumor Belief, Polarization, and Self-Organization

N. DiFonzo, M. J. Bourgeois, & J. M. Suls

3:30

Interpersonal synchrony, dynamical view of social relations: Computational model and empirical data

A. Nowak, W. Borkowski, J. Kania, W. Kulesza, K. Lisiecka, & A. Szostek

4PM

Wrap-up: Future plans

Everyone

4:30PM

END

END

 

 

 


 

INVITED SPEAKERS/TALKS

 

Evolutionary approaches to group behavior: A producer-scrounger dilemma and its consequences for collective adaptation.

Tatsuya Kameda (Department of Behavioral Science, Hokkaido University)

Reid Hastie (Booth Graduate School of Business, University of Chicago)

Human work-teams are often composed of a mixture of ‘producers’ who contribute to group endeavors and ‘scroungers’ who try to free-ride on others’ efforts. If free-riders are universally better off than producers in terms of net payoffs (as is commonly thought), how do such mixtures repeatedly and stably emerge in groups, and what consequences does this have for group-level adaptation? In this talk, we will present our recent work to address these questions using evolutionary games, agent-based simulations and behavioral experiments.

 

Mental Dynamism and its Constraints: Finding Patterns in the Stream of Consciousness

Robin R. Vallacher and Jay L. Michaels

Florida Atlantic University

       Social psychology has largely neglected the dynamic properties of mind and action, focusing instead on the immediate impact of independent variables (causes) on a dependent variable (effect).  In recent years, however, the concepts and tools of complexity science and nonlinear dynamical systems have been adapted to capture the time-dependency of psychological processes.  Knowledge of how a process unfolds over time provides insight into the mental or behavioral structures responsible for the observed temporal pattern.  In this presentation, we illustrate how this paradigm has been employed in our lab to identify the link between self-structure and the dynamics of self-evaluative thought. 

 

 

Network Clustering and Group Segregation Effects on Defensive and Non-Defensive Rumor Belief, Polarization, and Self-Organization

Nicholas DiFonzo, Rochester Institute of Technology

Martin J. Bourgeois, Florida Gulf Coast University‚Ä®    Jerry M. Suls, University of Iowa

Belief in rumors that are defensive (ingroup-positive/outgroup-negative) and non-defensive (outgroup-positive/ingroup-negative) were investigated within a theoretical framework integrating attitude polarization and Dynamic Social Impact Theory (DSIT). Three computer-mediated laboratory social network experiments were pooled to test the moderating effects of network clustering (cliquish structure) on segregation-induced defensive polarization, emergence of pockets of rumor belief/disbelief, and group conformity. Twenty-six 16-person networks, each composed of two groups (i.e., 8 Democrats and 8 Republicans, 8 deaf and 8 hearing persons, and 8 women and 8 men), serially discussed nine controversial rumors in lattice (unclustered) or “family” (clustered) structures. Group segregation was varied incrementally from full integration to full segregation (two disconnected groups). Support was found for network clustering as a moderator of the effects of segregation: in clustered versus unclustered configurations, segregation polarized rumor belief in a defensive direction, created distinct pockets of rumor belief/disbelief, and group conformity about truth of the rumors. Network clustering also amplified the effects of disagreement from in- and outgroup neighbors (the Echo Chamber Effect). Results connect identity-based attitude polarization with DSIT and advance knowledge about network structure moderation of intergroup contact effects on rumors.

 

Emergence and contagion in dynamic network systems

James D. Westaby

Columbia University, Teachers College

This session describes how dynamic network theory and its new methodological paradigm are related to dynamical processes and computational modeling.  This new metatheory is the first to comprehensively explain how social network roles influence goal pursuit and emergent processes at individual, group, organizational, and international levels of analysis.  The theory substantively extends past social network theories by infusing goal nodes directly into social networks, which allows the theory to explain a host of new network regulation, emergent, and contagion processes.  It also illustrates how traditional contagion propositions in social network theory, such as hubs with high centrality should generate diffusion and effective change, can be inaccurate in various settings, because they do not fully account for the finite motivational underpinnings in social networks.  Dynamic network theory provides this motivational understanding.  The theory and its methodologies are highly amenable to computational modeling and a variety of applications are illustrated.

 

Interpersonal synchrony, dynamical view of social relations: Computational model and empirical data.

Andrzej Nowak, Wojciech Borkowski, Jaroslaw Kania, Wojciech Kulesza, Karolina Lisiecka, Agnieszka Szostek

Institute for Social Studies

University of Warsaw

 

Interpersonal synchrony is a dynamical manifestation of social relationship. Computer simulations are used to explore how patterns of synchronization in dyads and social groups are related to parameters describing the individuals and their relationships. Rich sets of dynamical phenomena emerge as individuals synchronize their behavior and internal states. Computational results are related to empirical data showing the relevance of synchrony for interpersonal relations.

 

The Coherence Effect: Blending Cold and Hot Cognitions by Constraint Satisfaction

Dan Simon (USC Gould Law School), Douglas Stenstrom (CSLA), and Stephen J. Read (University of Southern California)

In previous work, we found experimental evidence of the "Coherence Effect," by which reasoning and decision-making tasks are driven by a coherence-maximizing constraint satisfaction process.  In our current studies, we extend this research to encompass also hot cognitions, namely, motivation, anticipated valence and emotion.  Experimental data and a computational model will be presented. 

 

A Dynamical Systems Approach to Person Construal

Jonathan B. Freeman & Nalini Ambady

Tufts University

A theoretical approach to person construal will be discussed. It sees the perception of other people as being accomplished by a dynamical system involving continuous interaction between social categories, stereotypes, high-level cognitive states, and the low-level processing of facial, vocal, and bodily cues. This system permits lower-level sensory perception and higher-order social cognition to dynamically coordinate across multiple interactive levels of processing to give rise to stable person construals. Using computational modeling and real-time behavioral data (computer-mouse movements), I will explain how such a dynamical system can account for major research findings. The implications of a dynamic and interactive person construal system will be discussed, as well as the importance of taking dynamical systems approaches to social psychological phenomena more broadly.

 

The intra-individual dynamics of health behavior.

Mark G. Orr*, Roxanne Thrush*, David C. Plaut#

*Columbia Univ.    #Carnegie Mellon Univ.

Some of the most successful models of health behavior are variants of expectancy-value theory (e.g., Theory of Planned Behavior).  Expectancy-value theory (a social psychological theory) postulates that behavior stems in large part from the aggregation of beliefs to attitudes to intention to behavior.  We expand on current theory by developing a computational model (using artificial neural networks) of the psychological processes involved in 1) the learning of beliefs and attitudes from others, and 2) the dynamics of the interface between the individual's behavior and the behaviors of others.  Discussion will center around our assumptions, our findings and where further data is needed to validate the model.

 

A Neural Network Model of Motivation and Personality Captures Dynamics in Social Interaction

Stephen J. Read, Vita Droutman, Gurveen Chopra, and Lynn C. Miller

University of Southern California

Human social interaction is driven by the dynamics of changing motivational systems as interactions and situations change and evolve.  We will show how a recently presented neural network model of human personality and motivation can be used to model the behavioral and motivational dynamics of a dyadic interaction and can simulate the behavior of an individual in a social interaction. 

 

 

 

Increasing the predictive validity of neural network models of person judgments
 
Brian Monroe*#, Tei Laine*, Swati Gupta*, & Ilya Farber*
 
* Institute of High Performance Computing; Agency for Science, Technology, and Research (Singapore)
# National University of Singapore
 

Previous network models of impression judgments have been useful tools, able to explain phenomena such as how stereotypes combine with individuating information to produce trait judgments, how contradictory attributes can be resolved into a coherent inference by using goal- and belief-based knowledge structures, simultaneous activation of category memberships, and accounting for proposed multi-stage models of person perception in terms of more parsimonious parallel processing via coherence mechanisms.  Starting with the foundation of a network architecture, we focus more on the predictive than the explanatory ability of networks: on refining a model to increase the fidelity of performance vs. actual human judgments.  In doing so we identify architectural and processing features that appear to be necessary to account for and also identify limitations of existing models.  We also compare the new model’s predictions against a large set of human judgments to validate its performance.

 


 

Preliminary Schedule

Time

Speaker

Title

8:30

Introduction

 

9AM

 

 

9:30

 

 

10AM

Coffee Break

 

10:30

 

 

11AM

 

 

11:30

 

 

Noon

LUNCH

LUNCH

1PM

 

 

1:30

 

 

2PM

 

 

2:30

COFFEE BREAK

Pick up registration materials

3PM

 

 

3:30

 

 

4PM

Wrap-up

 

4:30PM

END

END