-
Title
-
Coevolution of Decision‐Making and Social Environments
-
Author
-
Bruch, Elizabeth
-
Hammond, Ross A.
-
Todd, Peter M.
-
Research Area
-
Cognition and Emotions
-
Topic
-
Decision Making
-
Abstract
-
Social scientists have a longstanding theoretical interest in the relationship between individual behavior and social dynamics. A growing body of work demonstrates that, when human behavior is interdependent—that is, what one person does depends on the past, present, or anticipated future actions of others—there is not a simple or linear relationship between the choices of individuals and their collective consequences. Outside of the academy, policy makers are increasingly aware that well‐intentioned interventions can backfire if they fail to account for how people change their behavior in response to the intervention. This type of problem requires a systematic modeling approach. Our entry provides a brief introduction to a growing body of research that brings together two disparate literatures—studies of decision‐making and studies of the interplay between individuals' decisions and features of the social environment—through dynamic computational modeling. Cognitive scientists characterize human decision‐making under uncertainty using heuristics, rules‐of‐thumb that produce satisfactory choices quickly and with limited information. The heuristics we use and information samples we gather have profound consequences for the choices we make. At the same time, the social context defined by the choices of others feeds back to affect individual decision‐making. In recent years, there has been growing interest in methods such as agent‐based modeling and systems dynamics that can capture the dynamic interplay between individuals' choices and features of the environment. However, historically these approaches have not been grounded in cognitively plausible models of human behavior. We identify areas of high potential for future research, and lay out a preliminary framework to help guide understanding of the decision‐making process and its consequences in different social domains.
-
Identifier
-
etrds0044
-
extracted text
-
Coevolution of Decision-Making
and Social Environments
ELIZABETH BRUCH, ROSS A. HAMMOND, and PETER M. TODD
Abstract
Social scientists have a longstanding theoretical interest in the relationship between
individual behavior and social dynamics. A growing body of work demonstrates
that, when human behavior is interdependent—that is, what one person does
depends on the past, present, or anticipated future actions of others—there is not a
simple or linear relationship between the choices of individuals and their collective
consequences. Outside of the academy, policy makers are increasingly aware that
well-intentioned interventions can backfire if they fail to account for how people
change their behavior in response to the intervention. This type of problem requires
a systematic modeling approach. Our entry provides a brief introduction to a
growing body of research that brings together two disparate literatures—studies of
decision-making and studies of the interplay between individuals’ decisions and
features of the social environment—through dynamic computational modeling.
Cognitive scientists characterize human decision-making under uncertainty using
heuristics, rules-of-thumb that produce satisfactory choices quickly and with
limited information. The heuristics we use and information samples we gather have
profound consequences for the choices we make. At the same time, the social context
defined by the choices of others feeds back to affect individual decision-making.
In recent years, there has been growing interest in methods such as agent-based
modeling and systems dynamics that can capture the dynamic interplay between
individuals’ choices and features of the environment. However, historically these
approaches have not been grounded in cognitively plausible models of human
behavior. We identify areas of high potential for future research, and lay out a
preliminary framework to help guide understanding of the decision-making process
and its consequences in different social domains.
INTRODUCTION
The leading causes of death and disease in the United States are attributed
to behavioral factors such as tobacco use, poor diet and inactivity, alcohol
consumption, risky sexual behavior, and avoidable injuries (Danaei et al.,
2009; Mokdad, Marks, Stroup, & Gerberding, 2004). Behavior is thus central
to the prevention, treatment, and management of diseases and health
Emerging Trends in the Social and Behavioral Sciences. Edited by Robert Scott and Stephen Kosslyn.
© 2015 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.
1
2
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
care, and many interventions are aimed at changing behavior in healthier
directions. However, these interventions are often based on models of
human decision-making that lack empirical support, are difficult to quantify,
or largely ignore the bi-directional feedback between the individual and
the social (for a review of the behavioral theories guiding interventions,
see Glanz & Bishop, 2010). Research that ignores the processes by which
people actually make decisions in real settings or the interdependence of
those decisions may result in mis-specified behavioral models that lead to
incorrect predictions and ineffective policy recommendations.
There is also growing recognition that policy interventions are most likely
to be effective if they adopt ecological or systems perspectives (Huang &
Glass, 2008; Nader et al., 2012; Mabry, Olster, Morgan, & Abrams, 2008;
Sallis, Owen, & Fisher, 2008). The ecological perspective emphasizes that
multiple levels of influence shape behavior (e.g., individual, interpersonal,
organizational, community, and public policy), and the systems perspective
emphasizes the interconnectedness of these levels (including dynamic
feedback between individuals’ actions and their social environment). A
major challenge in solving pernicious social problems is accounting for this
bidirectional feedback between individuals and their environment (Sterman,
2006). In recent years, studies have identified “systems science” methods
such as agent-based modeling (ABM) as a potentially transformative tool
for capturing feedback across multiple levels of analysis (Hammond, 2009;
Luke & Stamatakis, 2012; Mabry, Marcus, Clark, Leischow, & Méndez,
2010). However, the ability of systems science approaches to reach their
full potential in offering new insights about behavior has been limited by
the unmet need for empirically valid models of individual behavior that
can be operationalized computationally and incorporated into the systems
analysis.
Below, we draw on insights from cognitive science and decision theory to
outline an initial framework to help guide the development of more cognitively sophisticated computational models. We identify both key features
of individuals’ choices (and the contexts in which those choices take place)
that may shape the process of decision-making, and also how those decisions
impinge on the current and future decisions of others. We view these as an
initial set of important structural features that determine both what strategies are available for individuals to use in navigating the choice environment,
and also the ways in which the decisions of individuals feed back to shape
the choice environment. We begin with a brief overview of agent-based models aimed at capturing feedback between individuals’ choices and the social
environment. We then review some of the state-of-the-art literature on decision strategies, and how they rely on features of the environment. In the final
section, we outline a framework for exploring the aggregate implications of
Coevolution of Decision-Making and Social Environments
3
“cognitively plausible” decision models as individuals simultaneously react
to and change their social environment.
LINKING INDIVIDUAL BEHAVIOR WITH THE SOCIAL
ENVIRONMENT
Social science has a longstanding interest in the relationship between individuals’ motivations and decisions and large-scale patterns of social organization and change. The “micro–macro problem” concerns how to explicitly
account for the ways in which actions of individuals give rise to social organization and dynamics, rather than assuming that macrolevel phenomena
are simply aggregates of individual characteristics and behavior (Coleman,
1994, p. 197; Granovetter, 1978, p. 1421). The connection between individuals’ actions and their collective consequences would be simple if one could
simply sum over individuals’ intentions or behavior to generate expected
population-level attributes. The problem is that much of human behavior is
interdependent; individuals’ actions often depend on what others are doing.
For example, individuals may directly influence one another through social
norms, peer effects, and other expectations for behavior. Even in situations
where direct social influence plays a minimal role, the alternatives available to people at any given moment may nonetheless depend on the past,
present, or future choices of others. For example, the availability of a job is
often dependent on the decision of its prior occupant to vacate it (Chase,
1991), individuals’ decisions about who to date or marry typically require
reciprocity of affections by the potential mate (Roth & Sotomayor, 1992), and
both require that the options have not already been taken by other searchers
(Todd, 2007).
ABM is a relatively new computational simulation approach specifically
designed to yield insight into how the behavior of decentralized autonomous
actors generate macro-level outcomes of interest, explicitly incorporating
dynamic feedback from macro back to micro. Agent-based models have been
used to study the macro–micro-dynamics governing a host of social processes, ranging from segregation to civil unrest to the outbreak and spread
of disease (Bruch & Mare, 2006; Cioffi-Revilla & Rouleau, 2010; Epstein,
2006; Schelling, 1971). Until recently, however, behavior of individual agents
in an ABM has rarely been grounded in cognitively or neurobiologically
“plausible” rules.1 Typically, agents have been programmed with stylized
rules for behavior, or in some cases with a psychologically unrealistic
statistical model that relates features of the environment to the probability of
1. Exceptions include the work of Todd and colleagues, who explore how heuristics interact with and
create features of the social environment (Todd, Billari, & Simão, 2005; Davis, Todd, & Bullock, 1999); as
well as work by Hammond and Ornstein (2014) and Epstein (2014).
4
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
the agent taking one set of actions or another. The potential has thus largely
been missed for psychologically grounded agent-based models to provide
new insights about how individuals’ choices shape and are shaped by their
social and physical environment. Combining advances in cognitive science
with those in complex systems science can yield a new generation of models
that shed light on both individual behavior and its instantiation in social
environments.
MODELS OF DECISION-MAKING
How do people make decisions? The classical model of decision-making is
the rational actor model endemic to neoclassical economics, which assumes a
fully informed, forward-looking rational actor with unlimited time for information processing (Becker, 1993; Von Neumann & Morgenstern, 2007). However, over the past 40 years, a large body of work has demonstrated that real
people make decisions under conditions of limited time, bounded cognitive
resources, and uncertainty. As a result, people often use heuristics—simple
rules for making a choice or inference—that serve to keep the processing
demands of a decision within the environmental and cognitive limits.
Heuristics are “problem-solving methods that tend to produce efficient
solutions to difficult problems by restricting the search through the space
of possible solutions, on the basis of some evaluation of the structure
of the problem” (Braunstein, 1972, p. 520). While early work from the
“heuristics and biases” school often emphasized the fallibility of human
decision-making (Kahneman, Slovic, & Tversky, 1982), more recent research
on “ecological rationality” shows how individuals’ simple decision strategies
can capitalize on systematic structure in the decision-making environment
(Todd, Gigerenzer, & The ABC Research Group, 2012). In other words, while
people do have limited knowledge and constraints on their ability to process
information, they can nonetheless make good decisions using heuristics that
match (by learning or evolution) to the ways that information is organized in
their environment. Such well-matched heuristics are designed to capitalize
on key features of the decision environment, and so can get away with using
limited information processed in a quick manner.
Heuristics are often composed of building blocks that underlie
decision-making, including: search rules that specify how to seek out
information on available choice alternatives; stopping rules that specify
when a search should be ended; and decision rules that specify how the final
choice is reached (Gigerenzer & Gaissmaier, 2011, p. 456). For example,
some marriage market models consider how people choose a marriage
partner when potential mates can only be explored one at a time, and there
is uncertainty about whether the next person to be encountered will be
Coevolution of Decision-Making and Social Environments
5
better than a currently available partner. One commonly studied heuristic
approach for such challenging sequential choices is to use a “satisficing”
mechanism incorporating building blocks whereby people initially spend
some period of time searching for available options and learning about them,
stop that search after a reasonable amount of time and set an aspiration level
based on what they have experienced, and then decide on the next available
partner encountered who meets that aspiration level (Todd & Miller, 1999;
Todd, Billari, & Simão, 2005).
DECISION TASKS AND DECISION ENVIRONMENTS
Many decision theories emphasize that the heuristics used successfully in
decision-making depend on particular features of the task environment
(Gigerenzer & Gaissmaier, 2011; Payne, Bettman, & Johnson, 1993; Simon,
1990; Todd & Gigerenzer, 2012), so to know what heuristics to build into
psychologically realistic models we must first assess the relevant features
of that environment that shape decision strategies. These features include
attributes of the decision task as well as characteristics of the social and
physical environment in which the decision occurs.2
Key features of decision tasks include:
1. The expected time horizon over which the decision will play out. This
includes the anticipated consequences of the decision as well as the
number of decisions made over a day, year, or lifetime. For example,
decisions about what to eat are made on a daily basis, whereas decisions
about where to live are made on average once every 3–5 years. It follows
that repeated decisions are likely to be governed more by habits and
learning over time than infrequent decisions (Scheibehenne, Mata, &
Todd, 2011). Another dimension of time horizons is the extent to which
decision consequences are immediate or cumulative. For example,
the effects of food choices and physical activity decisions cumulate
gradually over time, with feedback via biological outcomes often
occurring only after a substantial time lag.
2. The extent to which the decision is subject to social influence. This
depends on how much individuals can observe the preferences, strategies, or decision outcomes of others (e.g., exercise choices made in
public vs sleep behaviors done in private). Another aspect of social
influence is the degree to which the successful outcome of a decision is
under the control of the decision maker alone or is affected by others.
2. In this essay, we make an analytical distinction between attributes of decision tasks and attributes
of decision environments. But in practice, because tasks, decisions, and social environments are strongly
intertwined, this distinction may be blurred.
6
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
For example, the successful implementation of a young woman’s
decision to use condoms depends on the cooperation of her partner,
but her choice to take the stairs or the elevator is more individually
determined, both of which will affect her selection of appropriate
decision mechanisms.
3. The decision’s valence in terms of reward seeking versus harm avoidance.
It is well established that human beings are more sensitive to negative
change in their environment than positive change; this is sometimes
referred to as the positive–negative asymmetry effect. In reward-seeking situations, people are typically more tolerant of uncertainty and willing to
take risks (Kahneman & Tversky, 1979). In contrast, when the environment is difficult and dangerous, it is likely that people’s time horizons
will become considerably shorter, changing the heuristics they use.
Key features of decision environments include:
1. The number of alternatives to choose among. When there are only a few
alternatives, decisions are often made through a comparison process
that considers the most important features one at a time until a choice
can be made (e.g., the take-the-best heuristic and others—see Payne
et al., 1993; Rieskamp & Hoffrage, 1999). With more alternatives, a
multi-stage process can be used where each stage reduces the number
of options remaining under consideration. For example, the elimination
by aspects heuristic systematically reduces the number of alternatives
by first eliminating all those that are not good enough on the most
important aspect (e.g., all those restaurants in town that cost more than
$50 per person), then eliminating all those left that are not good enough
on the second aspect (e.g., all those remaining restaurants that are over
ten miles away), and so on until only one option is left (Tversky, 1972).
2. The distribution of alternatives. When satisfactory alternatives are plentiful, the decision maker requires little search or information to decide
among them, as most choices will be good. When good choices are rare,
strategies that search for longer are more appropriate (Fasolo, Hertwig,
Huber, & Ludwig, 2009). The distribution of cue values (which is related
to how informative the cues are) also influences what strategies will
work well (Reimer & Hoffrage, 2012).
3. The extent to which available options depend on the choices of others, and
the degree to which scarce items are replenished. For example, food
options may be sold out at the grocery story, but this could imply high
demand, which usually results in a resupply, calling for choice strategies
that revisit resource locations periodically. In contrast, when two people
marry they are removed from the list of possibilities available to others
Coevolution of Decision-Making and Social Environments
7
for an extended period (Todd, 2007). People may also be sensitive to the
rate of change in their choice environment, as well as the direction of
change (i.e., whether the change is perceived as positive vs negative),
which can favor different heuristics (Dudey & Todd, 2002; Hey, 1982).
4. The redundancy in the environment in terms of the correlation among
feature dimensions of choice alternatives. When there is a high degree
of redundancy, knowing one attribute of a particular alternative tells
the chooser something about its other attributes, so that heuristics
that focus on “one good reason” for making a choice will work well
with a quick information search (Rieskamp & Dieckmann, 2012). In
situations with multiple orthogonal attributes that are important for
choice (e.g., in social settings as when children select friends based
on family background, sex, and mutual interests), heuristics that tally
all those features may be more effective (Fasolo, McClelland, & Todd,
2007).
These components of task environment structure can be used to identify
shared properties of seemingly disparate choice applications, which may
help in the design of more effective policies aimed at changing behavior.
Because features of the decision environment determine what heuristics
will be most efficacious, interventions that succeed in one domain may be
fruitfully applied to another that shares structural features.
THE COEVOLUTION OF DECISION-MAKING
AND THE SOCIAL ENVIRONMENT
In most decision contexts, there is feedback between the choices that individuals make and the environment in which they make them. For example,
there are well-documented peer effects on eating, smoking, medication compliance, exercise, and mate choice in which one’s current choices influence
and are influenced by the witnessed choices of others (Bowers, Place, Todd,
Penke, & Asendorpf, 2012; Crandall, 1988; Lazev, Herzog, & Brandon, 1999;
Todd & Minard, 2014). In addition, the choices of individuals at one time
point can shape what options are available for future individuals. The classic example from social science is neighborhood tipping: Each individual
who leaves a neighborhood because she cannot tolerate its racial composition changes its composition and that of the neighborhood she moves into
(Schelling, 1978). Over time, the neighborhood choices available to others
evolve as a product of previous mobility decisions. This phenomenon can
also be seen in what food products are available for purchase in different
areas, which reflects aggregate demand. Finally, the success of a decision may
depend on the behavior of another person. For example, children are subject
8
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Individual strategies for
navigating choice environments
Features of choice environments
h
Population attributes
Preferences or choices of others
Strategies of others
Social norms/expectations
Gather information
from the social
environment
b
d
Desires
Beliefs
Values
a
e
Choice/
action
e
Structural features of
decision-making environment
Number & variety of choices
Frequency of decision
Accountable to others
Success dependent on others
Figure 1
f
c
Sample from
available
choice options
e
g
Select/apply
decision
strategy
Interaction between individuals and environment.
to the food consumption decisions of their parents and people are more likely
to exercise if they are accountable to a workout partner (Dishman, Sallis, &
Orenstein, 1985; Wardle, Guthrie, Sanderson, Birch, & Plomin, 2001).
Figure 1 illustrates how individual strategies for decision-making interact with features of the environment. The left hand side represents features
of both the social environment (e.g., the demographic make-up of the population, the behaviors of others, and social norms and other expectations)
and the choice domain. The social context (in particular, the choices of others) influences the number and type of options available (a). Individuals will
observe some subset of the social environment (b); these observations of the
social environment may influence their preferences, beliefs, and/or expectations (d). For a particular choice domain, the individual will sample some set
of options from the environment (c) which will—in conjunction with preferences and in some cases observing the decisions of others—determine their
decision-making strategy (e). This strategy results in a particular action outcome (f). Individuals’ choices may affect the choices available to others; for
example in the case of mate choice, a pairing will eliminate those two people as options for others (g). Feedback to the social environment occurs both
because the individual’s choice may change his or her other attributes as a
member of the population (e.g., weight and location) and also because his or
her choice may be observed by others (e.g., eating in a group) (h).
Coevolution of Decision-Making and Social Environments
9
Agent-based models that incorporate realistic decision heuristics to predict
individual-environment interactions have begun to appear in the literature.
For example, psychologically plausible satisficing heuristics have been used
in models of mate choice, where individual agents first engage in an adolescent “dating” period where they meet a succession of potential mates of varying levels of quality, learn how well they can do at attracting those potential
mates, adjust their own aspiration level for the kind of long-term mate they
should seek in the future on the basis of those initial dates (raising their aspirations after successful interactions and lowering them after unsuccessful
ones), and then enter the true mate-choice phase where they make marriage
offers to individuals they encounter who meet their aspirations and get married and removed from the population when the offer is mutual—thereby
changing the choice environment for all those agents still remaining in the
mating market (Todd & Miller, 1999). This model predicts observed demographic patterns of the ages at which people get married (Todd et al., 2005)
and demonstrates the strength of the effect of others’ decisions on one’s own
best choice strategy (Todd, 2007); other models have shown how the social
norms that guide individuals’ choices of an appropriately aged spouse can
evolve (Billari, Prskawetz, & Fürnkranz, 2003). A similar model in different domain shows related effects on the best strategy to use for searching
for a parking space (another type of sequential choice) when the choices
made by other earlier parkers creates the environment—here the spatial layout of available spaces—for later drivers (Hutchinson, Fanselow, & Todd,
2012). In the context of food choice, recent work using agent-based models
grounded in neurobiology describes the influence of food environments on
preferences (Hall, Hammond, & Rahmandad, 2014; Hammond et al., 2012)
and the bi-directional coevolution of body weight and social norms (Hammond & Ornstein, 2014).
FUTURE DIRECTIONS AND CHALLENGES
Agent-based simulation models that put agents with realistic psychological
decision mechanisms into social environments can be very useful in enabling
researchers to learn about aspects of cognition and behavior that are otherwise difficult to study (Todd, 1996). First, such models can provide existence proofs, showing that particular hypothesized cognitive mechanisms
can lead to particular observed patterns of behavior in specific environments.
Second, they can elucidate the dynamics of an interacting population over
time, helping us to understand what mechanisms and conditions can lead
to the appearance or disappearance of particular behaviors or environment
structures. And third, as “runnable thought experiments,” they can help us
explore complex interactions for which our intuitions are usually inadequate,
10
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
and thereby come up with predictions—including for effective interventions
to change behavior—that can be tested with empirical research.
A key challenge for future researchers interested in seeding their dynamic
models with realistic human behavior is identifying what decision rules people are using, and how those rules depend on features of both the decision
task and the social environment. The classic approach for studying how people make decisions is experiments that systematically vary features of both
the decision task and the environment and observe outcomes. Commonly,
computer-based choice situations are created in which people are presented
with two options to choose between, and they can examine different pieces
of evidence for each option (e.g., by clicking on a “show price” or “show
mileage” button for choosing a car) until they have seen enough to decide,
at which point their search, stopping, and decision rules can be assessed
(Bröder, 2012; Payne et al., 1993; Rieskamp & Hoffrage, 1999).
One promising source of information on choice processes is the behavioral
data produced through activities online. Dating websites, housing search
sites, job search sites, Facebook, and other electronic venues provide a
detailed look at how people navigate these decision processes. However,
an open question is the degree to which decision strategies observed online
can be generalized to the same behaviors in other settings. For example,
people visiting online dating sites may be confronted with thousands of
potential mates over a short time period, while in their day-to-day lives they
only encounter potential mates sporadically over an extended period. These
task and environment differences may result in differences in strategy and
selectivity as well (Lenton, Fasolo, & Todd, 2010).
There is a growing demand and opportunity for behaviorally sophisticated
models of individual behavior that can also capture bi-directional feedback
with social dynamics. We believe that advances in computational simulation
and in cognitive science mean the time is ripe for rapid progress in connecting
these two fields (to mutual benefit). This “emerging trend” is already underway, and offers the promise of both new insights into complex human behavior, and the design of more effective and efficient policies and interventions.
REFERENCES
Becker, G. (1993). Nobel lecture: The economic way of looking at behavior. Journal of
Political Economy, 101, 385–409.
Billari, F. C., Prskawetz, A., & Fürnkranz, J. (2003). On the cultural evolution of
age-at-marriage norms. In F. C. Billari & A. Prskawetz (Eds.), Agent-based computational demography: Using simulation to improve our understanding of demographic
behavior (pp. 139–157). Heidelberg, Germany: Physica Verlag.
Bowers, R. I., Place, S. S., Todd, P. M., Penke, L., & Asendorpf, J. B. (2012). Generalization in mate choice copying in humans. Behavioral Ecology, 23, 112–124.
Coevolution of Decision-Making and Social Environments
11
Braunstein, M. L. (1972). Perception of rotation in depth: A process model. Psychological Review, 79(6), 510–524.
Bröder, A. (2012). The quest for take-the-best: Insights and outlooks from experimental research. In P. M. Todd, G. Gigerenzer & The ABC Research Group (Eds.),
Ecological rationality: Intelligence in the world (pp. 216–240). New York, NY: Oxford
University Press.
Bruch, E., & Mare, R. (2006). Neighborhood choice and neighborhood change. American Journal of Sociology, 3, 667–709.
Chase, I. D. (1991). Vacancy chains. Annual Review of Sociology, 1, 133–154.
Cioffi-Revilla, C., & Rouleau, M. (2010). MASON RebeLand: An agent-based model
of politics, environment, and insurgency. International Studies Review, 1, 31–52.
Coleman, J. S. (1994). Foundations of social theory. Cambridge, MA: Harvard University Press.
Crandall, C. S. (1988). Social contagion of binge eating. Journal of Personality and Social
Psychology, 55(4), 588.
Danaei, G., Ding, E. L., Mozaffarian, D., Taylor, B., Rehm, J., Murray, C. J. L., & Ezzati,
M. (2009). The preventable causes of death in the united states: Comparative risk
assessment of dietary, lifestyle, and metabolic risk factors. PLoS Medicine, 6(4),
e1000058.
Davis, J. N., Todd, P. M., & Bullock, S. (1999). Environment quality predicts parental
provisioning decisions. Proceedings of the Royal Society of London B: Biological Sciences, 266, 1791–1797.
Dishman, R. K., Sallis, J. F., & Orenstein, D. R. (1985). The determinants of physical
activity and exercise. Public Health Reports, 100, 158.
Dudey, T., & Todd, P. M. (2002). Making good decisions with minimal information:
Simultaneous and sequential choice. Journal of Bioeconomics, 3, 195–215.
Epstein, J. M. (2014). Agent_Zero: Toward Neurocognitive Foundations for Generative
Social Science. Princeton, NJ: Princeton University Press.
Fasolo, B., Hertwig, R., Huber, M., & Ludwig, M. (2009). Size, entropy, and density: What is the difference that makes the difference between small and large
real-world assortments? Psychology & Marketing, 26, 254–279.
Fasolo, B., McClelland, G. H., & Todd, P. M. (2007). Escaping the tyranny of choice:
When fewer attributes make choice easier. Marketing Theory, 7, 13–26.
Gigerenzer, G., & Gaissmaier, W. (2011). Heuristic decision making. Annual Review
of Psychology, 62, 451–82.
Glanz, K., & Bishop, D. B. (2010). The role of behavioral science theory in development and implementation of public health interventions. Annual Review of Public
Health, 31, 399–418.
Granovetter, M. (1978). Threshold models of collective behavior. American Journal of
Sociology, 83(6), 1420–1443.
Hall, K. D., Hammond, R. A., & Rahmandad, H. (2014). Dynamic interplay among
homeostatic, hedonic, and cognitive feedback circuits regulating body weight.
American Journal of Public Health, e1–e7.
Hammond, R. A. (2009). “Complex systems modeling for obesity research”. Preventing Chronic Disease 6(3), A97.
12
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Hammond, R. A., & Ornstein, J. T. (2014). A model of social influence on body mass
index. Annals of the New York Academy of Science, 1331, 34–42.
Hammond, R. A., Ornstein, J. T., Fellows, L. K., Dube, L., Levitan, R., & Dagher, A.
(2012). A model of food reward learning with dynamic reward exposure. Frontiers
in Computational Neuroscience, 6, 82.
Hey, J. D. (1982). Search for rules for search. Journal of Economic Behavior and Organization, 3, 65–81.
Huang, T. T. K., & Glass, T. A. (2008). Transforming research strategies for understanding and preventing obesity. JAMA, 300(15), 1811–1813.
Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under
risk. Econometrica, 47(2), 263–292.
Lazev, A. B., Herzog, T. A., & Brandon, T. H. (1999). Classical conditioning of environmental cues to cigarette smoking. Experimental and Clinical Psychopharmacology,
7, 56.
Lenton, A. P., Fasolo, B., & Todd, P. M. (2010). Who is in your shopping cart? Expected
and experienced effects of choice abundance in the online dating context. In N.
Kock (Ed.), Evolutionary psychology and information systems research: A new approach
to studying the effects of modern technologies on human behavior (pp. 149–167). New
York, NY: Springer.
Mabry, P. L., Marcus, S. E., Clark, P. I., Leischow, S. J., & Méndez, D. (2010). Systems
science: a revolution in public health policy research. American Journal of Public
Health, 100, 1161–1163.
Mokdad, A. H., Marks, J. S., Stroup, D. F., & Gerberding, J. L. (2004). Actual causes
of death in the United States, 2000. Journal of the American Medical Association, 10,
1238–1245.
Nader, P. R., Huang, T. T. K., Gahagan, S., Kumanyika, S., Hammond, R. A., &
Christoffel, K. K. (2012). Next steps in obesity prevention: Altering early life systems to support healthy parents, infants, and toddlers. Childhood Obesity, 8(3),
195–204.
Reimer, T., & Hoffrage, U. (2012). Ecological rationality for teams and committees:
Heuristics in group decision making. In P. M. Todd, G. Gigerenzer & the ABC
Research Group (Eds.), Ecological rationality: Intelligence in the world (pp. 335–359).
New York, NY: Oxford University Press.
Rieskamp, J., & Dieckmann, A. (2012). Redundancy: Environment structure that simple heuristics can exploit. In P. M. Todd, G. Gigerenzer & The ABC Research Group
(Eds.), Ecological rationality: Intelligence in the world (pp. 187–215). New York, NY:
Oxford University Press.
Rieskamp, J., & Hoffrage, U. (1999). When do people use simple heuristics, and how
can we tell?. In G. Gigerenzer, P. M. Todd & The ABC Research Group (Eds.), Simple
heuristics that make us smart (pp. 141–167). New York, NY: Oxford University Press.
Roth, A. E., & Sotomayor, M. A. O. (1992). Two-sided matching: A study in game-theoretic
modeling and analysis. Cambridge, England: Cambridge University Press.
Sallis, J., Owen, N., & Fisher, E. (2008). Ecological models of health behavior. In
Glanz, K., B. K. Rimer, and K. Viswanath (Eds.), Health behavior and health education: Theory, research, and practice, (pp. 465–486). San Francisco, CA: John Wiley
& Sons, Inc.
Coevolution of Decision-Making and Social Environments
13
Scheibehenne, B., Mata, J., & Todd, P. M. (2011). Older but not wiser—Predicting
spouse’s preferences gets worse with age. Journal of Consumer Psychology, 21,
184–191.
Schelling, T. C. (1971). Dynamic models of segregation. Journal of Mathematical Sociology, 2, 143–186.
Schelling, T. C. (1978). Micromotives and macrobehavior. New York, NY: WW Norton
& Company.
Simon, H. A. (1990). Invariants of human behavior. Annual Review of Psychology, 41(1),
1–20.
Sterman, J. D. (2006). Learning from evidence in a complex world. American Journal
of Public Health, 3, 505–514.
Todd, P. M. (1996). The causes and effects of evolutionary simulation in the behavioral sciences. In R. Belew & M. Mitchell (Eds.), Adaptive individuals in evolving
populations: Models and algorithms (pp. 211–224). Reading, MA: Addison-Wesley.
Todd, P. M. (2007). Coevolved cognitive mechanisms in mate search: Making decisions in a decision-shaped world. In J. P. Forgas, M. G. Haselton & W. von Hippel
(Eds.), Evolution and the social mind: Evolutionary psychology and social cognition (pp.
145–159 (Sydney Symposium of Social Psychology series)). New York, NY: Psychology Press.
Todd, P. M., Billari, F. C., & Simão, J. (2005). Aggregate age-at-marriage patterns from
individual mate-search heuristics. Demography, 42, 559–574.
Todd, P. M., & Gigerenzer, G. (2012). Ecological Rationality: Intelligence in the World.
Oxford, England: Oxford University Press.
Todd, P. M., & Miller, G. F. (1999). From pride and prejudice to persuasion: Satisficing
in mate search. In G. Gigerenzer, P. M. Todd & The ABC Research Group (Eds.),
Simple heuristics that make us smart (pp. 287–308). New York, NY: Oxford University
Press.
Todd, P. M., & Minard, S. L. (2014). Simple heuristics for deciding what to eat. In
S. D. Preston, B. Knutson & M. Kringelbach (Eds.), The interdisciplinary science of
consumption. Cambridge, MA: MIT Press.
Tversky, A. (1972). Elimination by aspects: A theory of choice. Psychological Review,
79(4), 281–299.
Von Neumann, J., & Morgenstern, O. (2007). Theory of games and economic behavior
(Commemorative ed.). Princeton, NJ: Princeton University Press.
Wardle, J., Guthrie, C., Sanderson, S., Birch, L., & Plomin, R. (2001). Food and activity
preferences in children of lean and obese parents. International Journal of Obesity &
Related Metabolic Disorders, 25(7), 971–977.
FURTHER READING
Agent-Based Modeling
Epstein, J. (2006). Generative social science: Studies in agent-based computational modeling.
Princeton, NJ: Princeton University Press.
14
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Macy, M., & Willer, R. (2002). From factors to actors: Computational sociology and
agent-based modeling. Annual Review of Sociology, 28, 143–166.
Miller, J., & Page, S. (2009). Complex adaptive systems: An introduction to computational
models of social life: An introduction to computational models of social life. Princeton,
NJ: Princeton University Press.
Systems Science and Public Policy
Luke, D. A., & Stamatakis, K. A. (2012). Systems science methods in public health:
Dynamics, networks, and agents. Annual Review of Public Health, 33, 357–376.
Mabry, P. L., Marcus, S. E., Clark, P. I., Leischow, S. J., & Mendez, D. (2010). Systems
science: a revolution in public health policy research. American Journal of Public
Health, 100, 1161–1163.
Mabry, P. L., Olster, D. H., Morgan, G. D., & Abrams, D. B. (2008). Interdiciplinarity
and systems science to improve population health: a view from the NIH Office of
Behavioral and Social Sciences Research. American Journal of Preventive Medicine,
35(S2), 11–24.
Heuristic Decision-Making
Gigerenzer, G., Hertwig, R., & Pachur, T. (2011). Heuristics: The foundations of Adaptive
Behavior. New York, NY: Oxford University Press.
Gigerenzer, G., Todd, P. M., & The ABC Research Group (1999). Simple heuristics that
make us smart. New York, NY: Oxford University Press.
Kahneman, D., Slovic, P., & Tversky, A. (Eds.) (1982). Judgment under uncertainty:
heuristics and biases. Cambridge, England: Cambridge University Press.
Payne, J., Bettman, J., & Johnson, E. (1993). The adaptive decision maker. Cambridge,
England: Cambridge University Press.
Co-Evolution of Heuristic Strategies and Social Environments
Hammond, R. & Ornstein, J. (2014). A model of social influence on body weight.
Annals of the New York Academy of Science. Forthcoming.
Hutchinson, J. M. C., Fanselow, C., & Todd, P. M. (2012). Car parking as a game
between simple heuristics. In P. M. Todd, G. Gigerenzer & The ABC Research
Group (Eds.), Ecological rationality: Intelligence in the world (pp. 454–484). New York,
NY: Oxford University Press.
Todd, P. M., Billari, F. C., & Simão, J. (2005). Aggregate age-at-marriage patterns from
individual mate-search heuristics. Demography, 42(3), 559–574.
ELIZABETH BRUCH SHORT BIOGRAPHY
Elizabeth Bruch, PhD is an Assistant Professor in sociology and complex
systems, and Affiliate of the Population Studies Center at the Institute for
Coevolution of Decision-Making and Social Environments
15
Social Research. Much of her work blends statistical and agent-based methods to examine the relationship between individuals’ decisions about where
to live and patterns of residential segregation. She is also working on a project
exploring how individuals’ mate search strategies and willingness to settle
intersect with demographic constraints to generate dating or marriage market dynamics in US metro areas. She has served as a consulting editor for the
American Journal of Sociology, and on the editorial board for Sociological
Methodology. Her work has received the Gould Prize from the American
Journal of Sociology, as well as best article prizes from the Mathematical
and Urban Sociology sections of American Sociological Association. She is
a member of the National Institute of Health’s Network on Inequality, Complexity, and Health.
ROSS A. HAMMOND SHORT BIOGRAPHY
Ross A. Hammond, PhD is a Senior Fellow in Economic Studies at the
Brookings Institution, where he is also the Director of the Center on Social
Dynamics & Policy. His primary area of expertise is modeling complex
dynamics in economic, social, and public health systems using methods
from complexity science. His current research topics include obesity etiology
and prevention, food systems, tobacco control, behavioral epidemiology,
crime, corruption, segregation, and decision-making. Hammond received
his BA from Williams College and his PhD from the University of Michigan. He has authored numerous scientific articles, and his work has been
featured in New Scientist, Salon, The Atlantic Monthly, Scientific American,
and major news media. Hammond was recently appointed to the Institute
of Medicine/National Research Council committee Framework for Assessing
the Health, Environmental, and Social Effects of the Food System, and is both a
Public Health Advisor at the National Cancer Institute and an Advisory
Special Government Employee at the FDA Center for Tobacco Products. He
serves on the editorial boards of the journals Behavioral Science & Policy and
Childhood Obesity, and has been a member of the NIH-funded research networks MIDAS (Models of Infectious Disease Agent Study), ENVISION (part
of the National Collaborative on Childhood Obesity Research), and NICH
(Network on Inequality, Complexity, and Health). Hammond currently
holds appointments at the Harvard School of Public Health, Washington
University, and University of Michigan. He has been a consultant to the
World Bank, the Asian Development Bank, the Food and Drug Administration, the Institute of Medicine, and the National Institutes of Health. He
has taught computational modeling at Harvard, the University of Michigan,
Washington University, the National Cancer Institute, and the NIH/CDC
Institute on Systems Science and Health.
16
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
PETER M. TODD SHORT BIOGRAPHY
Peter M. Todd grew up in Silicon Valley, studied mathematics and electronic
music at Oberlin College, received an MPhil in computer speech and
language processing from Cambridge University, and developed neural
network models of the evolution of learning for his PhD in psychology at
Stanford University. In 1995, he moved to Germany to help found the Center
for Adaptive Behavior and Cognition (ABC), based at the Max Planck Institute for Human Development in Berlin. The Center’s work was captured in
the book Simple Heuristics That Make Us Smart (Gigerenzer, Todd, and The
ABC Research Group; Oxford, 1999); the sequel, Ecological Rationality: Intelligence in the World, covering information-environment structures and their
impact on decision-making, came out in 2012, along with a book on search
behavior, Cognitive Search: Evolution, Algorithms, and the Brain (Todd, Hills,
and Robbins, eds.; MIT Press). Todd moved to Indiana University in Bloomington as Professor of Cognitive Science, Psychology, and Informatics in 2005
and set up the ABC-West lab there (http://www.indiana.edu/∼abcwest/).
His ongoing research interests cover the interactions between and coevolution of decision-making and decision environments, focusing on the
ways that people and other animals search for resources—including mates,
information, and food—in space and time.
RELATED ESSAYS
Models of Revealed Preference (Economics), Abi Adams and Ian Crawford
Choice Architecture (Psychology), Adrian R. Camilleri and Rick P. Larrick
Behavioral Economics (Sociology), Guy Hochman and Dan Ariely
Emotion and Decision Making (Psychology), Jeff R. Huntsinger and Cara Ray
Against Game Theory (Political Science), Gale M. Lucas et al.
From Individual Rationality to Socially Embedded Self-Regulation (Sociology), Siegwart Lindenberg
Event Processing as an Executive Enterprise (Psychology), Robbie A. Ross and
Dare A. Baldwin
-
Coevolution of Decision-Making
and Social Environments
ELIZABETH BRUCH, ROSS A. HAMMOND, and PETER M. TODD
Abstract
Social scientists have a longstanding theoretical interest in the relationship between
individual behavior and social dynamics. A growing body of work demonstrates
that, when human behavior is interdependent—that is, what one person does
depends on the past, present, or anticipated future actions of others—there is not a
simple or linear relationship between the choices of individuals and their collective
consequences. Outside of the academy, policy makers are increasingly aware that
well-intentioned interventions can backfire if they fail to account for how people
change their behavior in response to the intervention. This type of problem requires
a systematic modeling approach. Our entry provides a brief introduction to a
growing body of research that brings together two disparate literatures—studies of
decision-making and studies of the interplay between individuals’ decisions and
features of the social environment—through dynamic computational modeling.
Cognitive scientists characterize human decision-making under uncertainty using
heuristics, rules-of-thumb that produce satisfactory choices quickly and with
limited information. The heuristics we use and information samples we gather have
profound consequences for the choices we make. At the same time, the social context
defined by the choices of others feeds back to affect individual decision-making.
In recent years, there has been growing interest in methods such as agent-based
modeling and systems dynamics that can capture the dynamic interplay between
individuals’ choices and features of the environment. However, historically these
approaches have not been grounded in cognitively plausible models of human
behavior. We identify areas of high potential for future research, and lay out a
preliminary framework to help guide understanding of the decision-making process
and its consequences in different social domains.
INTRODUCTION
The leading causes of death and disease in the United States are attributed
to behavioral factors such as tobacco use, poor diet and inactivity, alcohol
consumption, risky sexual behavior, and avoidable injuries (Danaei et al.,
2009; Mokdad, Marks, Stroup, & Gerberding, 2004). Behavior is thus central
to the prevention, treatment, and management of diseases and health
Emerging Trends in the Social and Behavioral Sciences. Edited by Robert Scott and Stephen Kosslyn.
© 2015 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.
1
2
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
care, and many interventions are aimed at changing behavior in healthier
directions. However, these interventions are often based on models of
human decision-making that lack empirical support, are difficult to quantify,
or largely ignore the bi-directional feedback between the individual and
the social (for a review of the behavioral theories guiding interventions,
see Glanz & Bishop, 2010). Research that ignores the processes by which
people actually make decisions in real settings or the interdependence of
those decisions may result in mis-specified behavioral models that lead to
incorrect predictions and ineffective policy recommendations.
There is also growing recognition that policy interventions are most likely
to be effective if they adopt ecological or systems perspectives (Huang &
Glass, 2008; Nader et al., 2012; Mabry, Olster, Morgan, & Abrams, 2008;
Sallis, Owen, & Fisher, 2008). The ecological perspective emphasizes that
multiple levels of influence shape behavior (e.g., individual, interpersonal,
organizational, community, and public policy), and the systems perspective
emphasizes the interconnectedness of these levels (including dynamic
feedback between individuals’ actions and their social environment). A
major challenge in solving pernicious social problems is accounting for this
bidirectional feedback between individuals and their environment (Sterman,
2006). In recent years, studies have identified “systems science” methods
such as agent-based modeling (ABM) as a potentially transformative tool
for capturing feedback across multiple levels of analysis (Hammond, 2009;
Luke & Stamatakis, 2012; Mabry, Marcus, Clark, Leischow, & Méndez,
2010). However, the ability of systems science approaches to reach their
full potential in offering new insights about behavior has been limited by
the unmet need for empirically valid models of individual behavior that
can be operationalized computationally and incorporated into the systems
analysis.
Below, we draw on insights from cognitive science and decision theory to
outline an initial framework to help guide the development of more cognitively sophisticated computational models. We identify both key features
of individuals’ choices (and the contexts in which those choices take place)
that may shape the process of decision-making, and also how those decisions
impinge on the current and future decisions of others. We view these as an
initial set of important structural features that determine both what strategies are available for individuals to use in navigating the choice environment,
and also the ways in which the decisions of individuals feed back to shape
the choice environment. We begin with a brief overview of agent-based models aimed at capturing feedback between individuals’ choices and the social
environment. We then review some of the state-of-the-art literature on decision strategies, and how they rely on features of the environment. In the final
section, we outline a framework for exploring the aggregate implications of
Coevolution of Decision-Making and Social Environments
3
“cognitively plausible” decision models as individuals simultaneously react
to and change their social environment.
LINKING INDIVIDUAL BEHAVIOR WITH THE SOCIAL
ENVIRONMENT
Social science has a longstanding interest in the relationship between individuals’ motivations and decisions and large-scale patterns of social organization and change. The “micro–macro problem” concerns how to explicitly
account for the ways in which actions of individuals give rise to social organization and dynamics, rather than assuming that macrolevel phenomena
are simply aggregates of individual characteristics and behavior (Coleman,
1994, p. 197; Granovetter, 1978, p. 1421). The connection between individuals’ actions and their collective consequences would be simple if one could
simply sum over individuals’ intentions or behavior to generate expected
population-level attributes. The problem is that much of human behavior is
interdependent; individuals’ actions often depend on what others are doing.
For example, individuals may directly influence one another through social
norms, peer effects, and other expectations for behavior. Even in situations
where direct social influence plays a minimal role, the alternatives available to people at any given moment may nonetheless depend on the past,
present, or future choices of others. For example, the availability of a job is
often dependent on the decision of its prior occupant to vacate it (Chase,
1991), individuals’ decisions about who to date or marry typically require
reciprocity of affections by the potential mate (Roth & Sotomayor, 1992), and
both require that the options have not already been taken by other searchers
(Todd, 2007).
ABM is a relatively new computational simulation approach specifically
designed to yield insight into how the behavior of decentralized autonomous
actors generate macro-level outcomes of interest, explicitly incorporating
dynamic feedback from macro back to micro. Agent-based models have been
used to study the macro–micro-dynamics governing a host of social processes, ranging from segregation to civil unrest to the outbreak and spread
of disease (Bruch & Mare, 2006; Cioffi-Revilla & Rouleau, 2010; Epstein,
2006; Schelling, 1971). Until recently, however, behavior of individual agents
in an ABM has rarely been grounded in cognitively or neurobiologically
“plausible” rules.1 Typically, agents have been programmed with stylized
rules for behavior, or in some cases with a psychologically unrealistic
statistical model that relates features of the environment to the probability of
1. Exceptions include the work of Todd and colleagues, who explore how heuristics interact with and
create features of the social environment (Todd, Billari, & Simão, 2005; Davis, Todd, & Bullock, 1999); as
well as work by Hammond and Ornstein (2014) and Epstein (2014).
4
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
the agent taking one set of actions or another. The potential has thus largely
been missed for psychologically grounded agent-based models to provide
new insights about how individuals’ choices shape and are shaped by their
social and physical environment. Combining advances in cognitive science
with those in complex systems science can yield a new generation of models
that shed light on both individual behavior and its instantiation in social
environments.
MODELS OF DECISION-MAKING
How do people make decisions? The classical model of decision-making is
the rational actor model endemic to neoclassical economics, which assumes a
fully informed, forward-looking rational actor with unlimited time for information processing (Becker, 1993; Von Neumann & Morgenstern, 2007). However, over the past 40 years, a large body of work has demonstrated that real
people make decisions under conditions of limited time, bounded cognitive
resources, and uncertainty. As a result, people often use heuristics—simple
rules for making a choice or inference—that serve to keep the processing
demands of a decision within the environmental and cognitive limits.
Heuristics are “problem-solving methods that tend to produce efficient
solutions to difficult problems by restricting the search through the space
of possible solutions, on the basis of some evaluation of the structure
of the problem” (Braunstein, 1972, p. 520). While early work from the
“heuristics and biases” school often emphasized the fallibility of human
decision-making (Kahneman, Slovic, & Tversky, 1982), more recent research
on “ecological rationality” shows how individuals’ simple decision strategies
can capitalize on systematic structure in the decision-making environment
(Todd, Gigerenzer, & The ABC Research Group, 2012). In other words, while
people do have limited knowledge and constraints on their ability to process
information, they can nonetheless make good decisions using heuristics that
match (by learning or evolution) to the ways that information is organized in
their environment. Such well-matched heuristics are designed to capitalize
on key features of the decision environment, and so can get away with using
limited information processed in a quick manner.
Heuristics are often composed of building blocks that underlie
decision-making, including: search rules that specify how to seek out
information on available choice alternatives; stopping rules that specify
when a search should be ended; and decision rules that specify how the final
choice is reached (Gigerenzer & Gaissmaier, 2011, p. 456). For example,
some marriage market models consider how people choose a marriage
partner when potential mates can only be explored one at a time, and there
is uncertainty about whether the next person to be encountered will be
Coevolution of Decision-Making and Social Environments
5
better than a currently available partner. One commonly studied heuristic
approach for such challenging sequential choices is to use a “satisficing”
mechanism incorporating building blocks whereby people initially spend
some period of time searching for available options and learning about them,
stop that search after a reasonable amount of time and set an aspiration level
based on what they have experienced, and then decide on the next available
partner encountered who meets that aspiration level (Todd & Miller, 1999;
Todd, Billari, & Simão, 2005).
DECISION TASKS AND DECISION ENVIRONMENTS
Many decision theories emphasize that the heuristics used successfully in
decision-making depend on particular features of the task environment
(Gigerenzer & Gaissmaier, 2011; Payne, Bettman, & Johnson, 1993; Simon,
1990; Todd & Gigerenzer, 2012), so to know what heuristics to build into
psychologically realistic models we must first assess the relevant features
of that environment that shape decision strategies. These features include
attributes of the decision task as well as characteristics of the social and
physical environment in which the decision occurs.2
Key features of decision tasks include:
1. The expected time horizon over which the decision will play out. This
includes the anticipated consequences of the decision as well as the
number of decisions made over a day, year, or lifetime. For example,
decisions about what to eat are made on a daily basis, whereas decisions
about where to live are made on average once every 3–5 years. It follows
that repeated decisions are likely to be governed more by habits and
learning over time than infrequent decisions (Scheibehenne, Mata, &
Todd, 2011). Another dimension of time horizons is the extent to which
decision consequences are immediate or cumulative. For example,
the effects of food choices and physical activity decisions cumulate
gradually over time, with feedback via biological outcomes often
occurring only after a substantial time lag.
2. The extent to which the decision is subject to social influence. This
depends on how much individuals can observe the preferences, strategies, or decision outcomes of others (e.g., exercise choices made in
public vs sleep behaviors done in private). Another aspect of social
influence is the degree to which the successful outcome of a decision is
under the control of the decision maker alone or is affected by others.
2. In this essay, we make an analytical distinction between attributes of decision tasks and attributes
of decision environments. But in practice, because tasks, decisions, and social environments are strongly
intertwined, this distinction may be blurred.
6
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
For example, the successful implementation of a young woman’s
decision to use condoms depends on the cooperation of her partner,
but her choice to take the stairs or the elevator is more individually
determined, both of which will affect her selection of appropriate
decision mechanisms.
3. The decision’s valence in terms of reward seeking versus harm avoidance.
It is well established that human beings are more sensitive to negative
change in their environment than positive change; this is sometimes
referred to as the positive–negative asymmetry effect. In reward-seeking situations, people are typically more tolerant of uncertainty and willing to
take risks (Kahneman & Tversky, 1979). In contrast, when the environment is difficult and dangerous, it is likely that people’s time horizons
will become considerably shorter, changing the heuristics they use.
Key features of decision environments include:
1. The number of alternatives to choose among. When there are only a few
alternatives, decisions are often made through a comparison process
that considers the most important features one at a time until a choice
can be made (e.g., the take-the-best heuristic and others—see Payne
et al., 1993; Rieskamp & Hoffrage, 1999). With more alternatives, a
multi-stage process can be used where each stage reduces the number
of options remaining under consideration. For example, the elimination
by aspects heuristic systematically reduces the number of alternatives
by first eliminating all those that are not good enough on the most
important aspect (e.g., all those restaurants in town that cost more than
$50 per person), then eliminating all those left that are not good enough
on the second aspect (e.g., all those remaining restaurants that are over
ten miles away), and so on until only one option is left (Tversky, 1972).
2. The distribution of alternatives. When satisfactory alternatives are plentiful, the decision maker requires little search or information to decide
among them, as most choices will be good. When good choices are rare,
strategies that search for longer are more appropriate (Fasolo, Hertwig,
Huber, & Ludwig, 2009). The distribution of cue values (which is related
to how informative the cues are) also influences what strategies will
work well (Reimer & Hoffrage, 2012).
3. The extent to which available options depend on the choices of others, and
the degree to which scarce items are replenished. For example, food
options may be sold out at the grocery story, but this could imply high
demand, which usually results in a resupply, calling for choice strategies
that revisit resource locations periodically. In contrast, when two people
marry they are removed from the list of possibilities available to others
Coevolution of Decision-Making and Social Environments
7
for an extended period (Todd, 2007). People may also be sensitive to the
rate of change in their choice environment, as well as the direction of
change (i.e., whether the change is perceived as positive vs negative),
which can favor different heuristics (Dudey & Todd, 2002; Hey, 1982).
4. The redundancy in the environment in terms of the correlation among
feature dimensions of choice alternatives. When there is a high degree
of redundancy, knowing one attribute of a particular alternative tells
the chooser something about its other attributes, so that heuristics
that focus on “one good reason” for making a choice will work well
with a quick information search (Rieskamp & Dieckmann, 2012). In
situations with multiple orthogonal attributes that are important for
choice (e.g., in social settings as when children select friends based
on family background, sex, and mutual interests), heuristics that tally
all those features may be more effective (Fasolo, McClelland, & Todd,
2007).
These components of task environment structure can be used to identify
shared properties of seemingly disparate choice applications, which may
help in the design of more effective policies aimed at changing behavior.
Because features of the decision environment determine what heuristics
will be most efficacious, interventions that succeed in one domain may be
fruitfully applied to another that shares structural features.
THE COEVOLUTION OF DECISION-MAKING
AND THE SOCIAL ENVIRONMENT
In most decision contexts, there is feedback between the choices that individuals make and the environment in which they make them. For example,
there are well-documented peer effects on eating, smoking, medication compliance, exercise, and mate choice in which one’s current choices influence
and are influenced by the witnessed choices of others (Bowers, Place, Todd,
Penke, & Asendorpf, 2012; Crandall, 1988; Lazev, Herzog, & Brandon, 1999;
Todd & Minard, 2014). In addition, the choices of individuals at one time
point can shape what options are available for future individuals. The classic example from social science is neighborhood tipping: Each individual
who leaves a neighborhood because she cannot tolerate its racial composition changes its composition and that of the neighborhood she moves into
(Schelling, 1978). Over time, the neighborhood choices available to others
evolve as a product of previous mobility decisions. This phenomenon can
also be seen in what food products are available for purchase in different
areas, which reflects aggregate demand. Finally, the success of a decision may
depend on the behavior of another person. For example, children are subject
8
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Individual strategies for
navigating choice environments
Features of choice environments
h
Population attributes
Preferences or choices of others
Strategies of others
Social norms/expectations
Gather information
from the social
environment
b
d
Desires
Beliefs
Values
a
e
Choice/
action
e
Structural features of
decision-making environment
Number & variety of choices
Frequency of decision
Accountable to others
Success dependent on others
Figure 1
f
c
Sample from
available
choice options
e
g
Select/apply
decision
strategy
Interaction between individuals and environment.
to the food consumption decisions of their parents and people are more likely
to exercise if they are accountable to a workout partner (Dishman, Sallis, &
Orenstein, 1985; Wardle, Guthrie, Sanderson, Birch, & Plomin, 2001).
Figure 1 illustrates how individual strategies for decision-making interact with features of the environment. The left hand side represents features
of both the social environment (e.g., the demographic make-up of the population, the behaviors of others, and social norms and other expectations)
and the choice domain. The social context (in particular, the choices of others) influences the number and type of options available (a). Individuals will
observe some subset of the social environment (b); these observations of the
social environment may influence their preferences, beliefs, and/or expectations (d). For a particular choice domain, the individual will sample some set
of options from the environment (c) which will—in conjunction with preferences and in some cases observing the decisions of others—determine their
decision-making strategy (e). This strategy results in a particular action outcome (f). Individuals’ choices may affect the choices available to others; for
example in the case of mate choice, a pairing will eliminate those two people as options for others (g). Feedback to the social environment occurs both
because the individual’s choice may change his or her other attributes as a
member of the population (e.g., weight and location) and also because his or
her choice may be observed by others (e.g., eating in a group) (h).
Coevolution of Decision-Making and Social Environments
9
Agent-based models that incorporate realistic decision heuristics to predict
individual-environment interactions have begun to appear in the literature.
For example, psychologically plausible satisficing heuristics have been used
in models of mate choice, where individual agents first engage in an adolescent “dating” period where they meet a succession of potential mates of varying levels of quality, learn how well they can do at attracting those potential
mates, adjust their own aspiration level for the kind of long-term mate they
should seek in the future on the basis of those initial dates (raising their aspirations after successful interactions and lowering them after unsuccessful
ones), and then enter the true mate-choice phase where they make marriage
offers to individuals they encounter who meet their aspirations and get married and removed from the population when the offer is mutual—thereby
changing the choice environment for all those agents still remaining in the
mating market (Todd & Miller, 1999). This model predicts observed demographic patterns of the ages at which people get married (Todd et al., 2005)
and demonstrates the strength of the effect of others’ decisions on one’s own
best choice strategy (Todd, 2007); other models have shown how the social
norms that guide individuals’ choices of an appropriately aged spouse can
evolve (Billari, Prskawetz, & Fürnkranz, 2003). A similar model in different domain shows related effects on the best strategy to use for searching
for a parking space (another type of sequential choice) when the choices
made by other earlier parkers creates the environment—here the spatial layout of available spaces—for later drivers (Hutchinson, Fanselow, & Todd,
2012). In the context of food choice, recent work using agent-based models
grounded in neurobiology describes the influence of food environments on
preferences (Hall, Hammond, & Rahmandad, 2014; Hammond et al., 2012)
and the bi-directional coevolution of body weight and social norms (Hammond & Ornstein, 2014).
FUTURE DIRECTIONS AND CHALLENGES
Agent-based simulation models that put agents with realistic psychological
decision mechanisms into social environments can be very useful in enabling
researchers to learn about aspects of cognition and behavior that are otherwise difficult to study (Todd, 1996). First, such models can provide existence proofs, showing that particular hypothesized cognitive mechanisms
can lead to particular observed patterns of behavior in specific environments.
Second, they can elucidate the dynamics of an interacting population over
time, helping us to understand what mechanisms and conditions can lead
to the appearance or disappearance of particular behaviors or environment
structures. And third, as “runnable thought experiments,” they can help us
explore complex interactions for which our intuitions are usually inadequate,
10
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
and thereby come up with predictions—including for effective interventions
to change behavior—that can be tested with empirical research.
A key challenge for future researchers interested in seeding their dynamic
models with realistic human behavior is identifying what decision rules people are using, and how those rules depend on features of both the decision
task and the social environment. The classic approach for studying how people make decisions is experiments that systematically vary features of both
the decision task and the environment and observe outcomes. Commonly,
computer-based choice situations are created in which people are presented
with two options to choose between, and they can examine different pieces
of evidence for each option (e.g., by clicking on a “show price” or “show
mileage” button for choosing a car) until they have seen enough to decide,
at which point their search, stopping, and decision rules can be assessed
(Bröder, 2012; Payne et al., 1993; Rieskamp & Hoffrage, 1999).
One promising source of information on choice processes is the behavioral
data produced through activities online. Dating websites, housing search
sites, job search sites, Facebook, and other electronic venues provide a
detailed look at how people navigate these decision processes. However,
an open question is the degree to which decision strategies observed online
can be generalized to the same behaviors in other settings. For example,
people visiting online dating sites may be confronted with thousands of
potential mates over a short time period, while in their day-to-day lives they
only encounter potential mates sporadically over an extended period. These
task and environment differences may result in differences in strategy and
selectivity as well (Lenton, Fasolo, & Todd, 2010).
There is a growing demand and opportunity for behaviorally sophisticated
models of individual behavior that can also capture bi-directional feedback
with social dynamics. We believe that advances in computational simulation
and in cognitive science mean the time is ripe for rapid progress in connecting
these two fields (to mutual benefit). This “emerging trend” is already underway, and offers the promise of both new insights into complex human behavior, and the design of more effective and efficient policies and interventions.
REFERENCES
Becker, G. (1993). Nobel lecture: The economic way of looking at behavior. Journal of
Political Economy, 101, 385–409.
Billari, F. C., Prskawetz, A., & Fürnkranz, J. (2003). On the cultural evolution of
age-at-marriage norms. In F. C. Billari & A. Prskawetz (Eds.), Agent-based computational demography: Using simulation to improve our understanding of demographic
behavior (pp. 139–157). Heidelberg, Germany: Physica Verlag.
Bowers, R. I., Place, S. S., Todd, P. M., Penke, L., & Asendorpf, J. B. (2012). Generalization in mate choice copying in humans. Behavioral Ecology, 23, 112–124.
Coevolution of Decision-Making and Social Environments
11
Braunstein, M. L. (1972). Perception of rotation in depth: A process model. Psychological Review, 79(6), 510–524.
Bröder, A. (2012). The quest for take-the-best: Insights and outlooks from experimental research. In P. M. Todd, G. Gigerenzer & The ABC Research Group (Eds.),
Ecological rationality: Intelligence in the world (pp. 216–240). New York, NY: Oxford
University Press.
Bruch, E., & Mare, R. (2006). Neighborhood choice and neighborhood change. American Journal of Sociology, 3, 667–709.
Chase, I. D. (1991). Vacancy chains. Annual Review of Sociology, 1, 133–154.
Cioffi-Revilla, C., & Rouleau, M. (2010). MASON RebeLand: An agent-based model
of politics, environment, and insurgency. International Studies Review, 1, 31–52.
Coleman, J. S. (1994). Foundations of social theory. Cambridge, MA: Harvard University Press.
Crandall, C. S. (1988). Social contagion of binge eating. Journal of Personality and Social
Psychology, 55(4), 588.
Danaei, G., Ding, E. L., Mozaffarian, D., Taylor, B., Rehm, J., Murray, C. J. L., & Ezzati,
M. (2009). The preventable causes of death in the united states: Comparative risk
assessment of dietary, lifestyle, and metabolic risk factors. PLoS Medicine, 6(4),
e1000058.
Davis, J. N., Todd, P. M., & Bullock, S. (1999). Environment quality predicts parental
provisioning decisions. Proceedings of the Royal Society of London B: Biological Sciences, 266, 1791–1797.
Dishman, R. K., Sallis, J. F., & Orenstein, D. R. (1985). The determinants of physical
activity and exercise. Public Health Reports, 100, 158.
Dudey, T., & Todd, P. M. (2002). Making good decisions with minimal information:
Simultaneous and sequential choice. Journal of Bioeconomics, 3, 195–215.
Epstein, J. M. (2014). Agent_Zero: Toward Neurocognitive Foundations for Generative
Social Science. Princeton, NJ: Princeton University Press.
Fasolo, B., Hertwig, R., Huber, M., & Ludwig, M. (2009). Size, entropy, and density: What is the difference that makes the difference between small and large
real-world assortments? Psychology & Marketing, 26, 254–279.
Fasolo, B., McClelland, G. H., & Todd, P. M. (2007). Escaping the tyranny of choice:
When fewer attributes make choice easier. Marketing Theory, 7, 13–26.
Gigerenzer, G., & Gaissmaier, W. (2011). Heuristic decision making. Annual Review
of Psychology, 62, 451–82.
Glanz, K., & Bishop, D. B. (2010). The role of behavioral science theory in development and implementation of public health interventions. Annual Review of Public
Health, 31, 399–418.
Granovetter, M. (1978). Threshold models of collective behavior. American Journal of
Sociology, 83(6), 1420–1443.
Hall, K. D., Hammond, R. A., & Rahmandad, H. (2014). Dynamic interplay among
homeostatic, hedonic, and cognitive feedback circuits regulating body weight.
American Journal of Public Health, e1–e7.
Hammond, R. A. (2009). “Complex systems modeling for obesity research”. Preventing Chronic Disease 6(3), A97.
12
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Hammond, R. A., & Ornstein, J. T. (2014). A model of social influence on body mass
index. Annals of the New York Academy of Science, 1331, 34–42.
Hammond, R. A., Ornstein, J. T., Fellows, L. K., Dube, L., Levitan, R., & Dagher, A.
(2012). A model of food reward learning with dynamic reward exposure. Frontiers
in Computational Neuroscience, 6, 82.
Hey, J. D. (1982). Search for rules for search. Journal of Economic Behavior and Organization, 3, 65–81.
Huang, T. T. K., & Glass, T. A. (2008). Transforming research strategies for understanding and preventing obesity. JAMA, 300(15), 1811–1813.
Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under
risk. Econometrica, 47(2), 263–292.
Lazev, A. B., Herzog, T. A., & Brandon, T. H. (1999). Classical conditioning of environmental cues to cigarette smoking. Experimental and Clinical Psychopharmacology,
7, 56.
Lenton, A. P., Fasolo, B., & Todd, P. M. (2010). Who is in your shopping cart? Expected
and experienced effects of choice abundance in the online dating context. In N.
Kock (Ed.), Evolutionary psychology and information systems research: A new approach
to studying the effects of modern technologies on human behavior (pp. 149–167). New
York, NY: Springer.
Mabry, P. L., Marcus, S. E., Clark, P. I., Leischow, S. J., & Méndez, D. (2010). Systems
science: a revolution in public health policy research. American Journal of Public
Health, 100, 1161–1163.
Mokdad, A. H., Marks, J. S., Stroup, D. F., & Gerberding, J. L. (2004). Actual causes
of death in the United States, 2000. Journal of the American Medical Association, 10,
1238–1245.
Nader, P. R., Huang, T. T. K., Gahagan, S., Kumanyika, S., Hammond, R. A., &
Christoffel, K. K. (2012). Next steps in obesity prevention: Altering early life systems to support healthy parents, infants, and toddlers. Childhood Obesity, 8(3),
195–204.
Reimer, T., & Hoffrage, U. (2012). Ecological rationality for teams and committees:
Heuristics in group decision making. In P. M. Todd, G. Gigerenzer & the ABC
Research Group (Eds.), Ecological rationality: Intelligence in the world (pp. 335–359).
New York, NY: Oxford University Press.
Rieskamp, J., & Dieckmann, A. (2012). Redundancy: Environment structure that simple heuristics can exploit. In P. M. Todd, G. Gigerenzer & The ABC Research Group
(Eds.), Ecological rationality: Intelligence in the world (pp. 187–215). New York, NY:
Oxford University Press.
Rieskamp, J., & Hoffrage, U. (1999). When do people use simple heuristics, and how
can we tell?. In G. Gigerenzer, P. M. Todd & The ABC Research Group (Eds.), Simple
heuristics that make us smart (pp. 141–167). New York, NY: Oxford University Press.
Roth, A. E., & Sotomayor, M. A. O. (1992). Two-sided matching: A study in game-theoretic
modeling and analysis. Cambridge, England: Cambridge University Press.
Sallis, J., Owen, N., & Fisher, E. (2008). Ecological models of health behavior. In
Glanz, K., B. K. Rimer, and K. Viswanath (Eds.), Health behavior and health education: Theory, research, and practice, (pp. 465–486). San Francisco, CA: John Wiley
& Sons, Inc.
Coevolution of Decision-Making and Social Environments
13
Scheibehenne, B., Mata, J., & Todd, P. M. (2011). Older but not wiser—Predicting
spouse’s preferences gets worse with age. Journal of Consumer Psychology, 21,
184–191.
Schelling, T. C. (1971). Dynamic models of segregation. Journal of Mathematical Sociology, 2, 143–186.
Schelling, T. C. (1978). Micromotives and macrobehavior. New York, NY: WW Norton
& Company.
Simon, H. A. (1990). Invariants of human behavior. Annual Review of Psychology, 41(1),
1–20.
Sterman, J. D. (2006). Learning from evidence in a complex world. American Journal
of Public Health, 3, 505–514.
Todd, P. M. (1996). The causes and effects of evolutionary simulation in the behavioral sciences. In R. Belew & M. Mitchell (Eds.), Adaptive individuals in evolving
populations: Models and algorithms (pp. 211–224). Reading, MA: Addison-Wesley.
Todd, P. M. (2007). Coevolved cognitive mechanisms in mate search: Making decisions in a decision-shaped world. In J. P. Forgas, M. G. Haselton & W. von Hippel
(Eds.), Evolution and the social mind: Evolutionary psychology and social cognition (pp.
145–159 (Sydney Symposium of Social Psychology series)). New York, NY: Psychology Press.
Todd, P. M., Billari, F. C., & Simão, J. (2005). Aggregate age-at-marriage patterns from
individual mate-search heuristics. Demography, 42, 559–574.
Todd, P. M., & Gigerenzer, G. (2012). Ecological Rationality: Intelligence in the World.
Oxford, England: Oxford University Press.
Todd, P. M., & Miller, G. F. (1999). From pride and prejudice to persuasion: Satisficing
in mate search. In G. Gigerenzer, P. M. Todd & The ABC Research Group (Eds.),
Simple heuristics that make us smart (pp. 287–308). New York, NY: Oxford University
Press.
Todd, P. M., & Minard, S. L. (2014). Simple heuristics for deciding what to eat. In
S. D. Preston, B. Knutson & M. Kringelbach (Eds.), The interdisciplinary science of
consumption. Cambridge, MA: MIT Press.
Tversky, A. (1972). Elimination by aspects: A theory of choice. Psychological Review,
79(4), 281–299.
Von Neumann, J., & Morgenstern, O. (2007). Theory of games and economic behavior
(Commemorative ed.). Princeton, NJ: Princeton University Press.
Wardle, J., Guthrie, C., Sanderson, S., Birch, L., & Plomin, R. (2001). Food and activity
preferences in children of lean and obese parents. International Journal of Obesity &
Related Metabolic Disorders, 25(7), 971–977.
FURTHER READING
Agent-Based Modeling
Epstein, J. (2006). Generative social science: Studies in agent-based computational modeling.
Princeton, NJ: Princeton University Press.
14
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Macy, M., & Willer, R. (2002). From factors to actors: Computational sociology and
agent-based modeling. Annual Review of Sociology, 28, 143–166.
Miller, J., & Page, S. (2009). Complex adaptive systems: An introduction to computational
models of social life: An introduction to computational models of social life. Princeton,
NJ: Princeton University Press.
Systems Science and Public Policy
Luke, D. A., & Stamatakis, K. A. (2012). Systems science methods in public health:
Dynamics, networks, and agents. Annual Review of Public Health, 33, 357–376.
Mabry, P. L., Marcus, S. E., Clark, P. I., Leischow, S. J., & Mendez, D. (2010). Systems
science: a revolution in public health policy research. American Journal of Public
Health, 100, 1161–1163.
Mabry, P. L., Olster, D. H., Morgan, G. D., & Abrams, D. B. (2008). Interdiciplinarity
and systems science to improve population health: a view from the NIH Office of
Behavioral and Social Sciences Research. American Journal of Preventive Medicine,
35(S2), 11–24.
Heuristic Decision-Making
Gigerenzer, G., Hertwig, R., & Pachur, T. (2011). Heuristics: The foundations of Adaptive
Behavior. New York, NY: Oxford University Press.
Gigerenzer, G., Todd, P. M., & The ABC Research Group (1999). Simple heuristics that
make us smart. New York, NY: Oxford University Press.
Kahneman, D., Slovic, P., & Tversky, A. (Eds.) (1982). Judgment under uncertainty:
heuristics and biases. Cambridge, England: Cambridge University Press.
Payne, J., Bettman, J., & Johnson, E. (1993). The adaptive decision maker. Cambridge,
England: Cambridge University Press.
Co-Evolution of Heuristic Strategies and Social Environments
Hammond, R. & Ornstein, J. (2014). A model of social influence on body weight.
Annals of the New York Academy of Science. Forthcoming.
Hutchinson, J. M. C., Fanselow, C., & Todd, P. M. (2012). Car parking as a game
between simple heuristics. In P. M. Todd, G. Gigerenzer & The ABC Research
Group (Eds.), Ecological rationality: Intelligence in the world (pp. 454–484). New York,
NY: Oxford University Press.
Todd, P. M., Billari, F. C., & Simão, J. (2005). Aggregate age-at-marriage patterns from
individual mate-search heuristics. Demography, 42(3), 559–574.
ELIZABETH BRUCH SHORT BIOGRAPHY
Elizabeth Bruch, PhD is an Assistant Professor in sociology and complex
systems, and Affiliate of the Population Studies Center at the Institute for
Coevolution of Decision-Making and Social Environments
15
Social Research. Much of her work blends statistical and agent-based methods to examine the relationship between individuals’ decisions about where
to live and patterns of residential segregation. She is also working on a project
exploring how individuals’ mate search strategies and willingness to settle
intersect with demographic constraints to generate dating or marriage market dynamics in US metro areas. She has served as a consulting editor for the
American Journal of Sociology, and on the editorial board for Sociological
Methodology. Her work has received the Gould Prize from the American
Journal of Sociology, as well as best article prizes from the Mathematical
and Urban Sociology sections of American Sociological Association. She is
a member of the National Institute of Health’s Network on Inequality, Complexity, and Health.
ROSS A. HAMMOND SHORT BIOGRAPHY
Ross A. Hammond, PhD is a Senior Fellow in Economic Studies at the
Brookings Institution, where he is also the Director of the Center on Social
Dynamics & Policy. His primary area of expertise is modeling complex
dynamics in economic, social, and public health systems using methods
from complexity science. His current research topics include obesity etiology
and prevention, food systems, tobacco control, behavioral epidemiology,
crime, corruption, segregation, and decision-making. Hammond received
his BA from Williams College and his PhD from the University of Michigan. He has authored numerous scientific articles, and his work has been
featured in New Scientist, Salon, The Atlantic Monthly, Scientific American,
and major news media. Hammond was recently appointed to the Institute
of Medicine/National Research Council committee Framework for Assessing
the Health, Environmental, and Social Effects of the Food System, and is both a
Public Health Advisor at the National Cancer Institute and an Advisory
Special Government Employee at the FDA Center for Tobacco Products. He
serves on the editorial boards of the journals Behavioral Science & Policy and
Childhood Obesity, and has been a member of the NIH-funded research networks MIDAS (Models of Infectious Disease Agent Study), ENVISION (part
of the National Collaborative on Childhood Obesity Research), and NICH
(Network on Inequality, Complexity, and Health). Hammond currently
holds appointments at the Harvard School of Public Health, Washington
University, and University of Michigan. He has been a consultant to the
World Bank, the Asian Development Bank, the Food and Drug Administration, the Institute of Medicine, and the National Institutes of Health. He
has taught computational modeling at Harvard, the University of Michigan,
Washington University, the National Cancer Institute, and the NIH/CDC
Institute on Systems Science and Health.
16
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
PETER M. TODD SHORT BIOGRAPHY
Peter M. Todd grew up in Silicon Valley, studied mathematics and electronic
music at Oberlin College, received an MPhil in computer speech and
language processing from Cambridge University, and developed neural
network models of the evolution of learning for his PhD in psychology at
Stanford University. In 1995, he moved to Germany to help found the Center
for Adaptive Behavior and Cognition (ABC), based at the Max Planck Institute for Human Development in Berlin. The Center’s work was captured in
the book Simple Heuristics That Make Us Smart (Gigerenzer, Todd, and The
ABC Research Group; Oxford, 1999); the sequel, Ecological Rationality: Intelligence in the World, covering information-environment structures and their
impact on decision-making, came out in 2012, along with a book on search
behavior, Cognitive Search: Evolution, Algorithms, and the Brain (Todd, Hills,
and Robbins, eds.; MIT Press). Todd moved to Indiana University in Bloomington as Professor of Cognitive Science, Psychology, and Informatics in 2005
and set up the ABC-West lab there (http://www.indiana.edu/∼abcwest/).
His ongoing research interests cover the interactions between and coevolution of decision-making and decision environments, focusing on the
ways that people and other animals search for resources—including mates,
information, and food—in space and time.
RELATED ESSAYS
Models of Revealed Preference (Economics), Abi Adams and Ian Crawford
Choice Architecture (Psychology), Adrian R. Camilleri and Rick P. Larrick
Behavioral Economics (Sociology), Guy Hochman and Dan Ariely
Emotion and Decision Making (Psychology), Jeff R. Huntsinger and Cara Ray
Against Game Theory (Political Science), Gale M. Lucas et al.
From Individual Rationality to Socially Embedded Self-Regulation (Sociology), Siegwart Lindenberg
Event Processing as an Executive Enterprise (Psychology), Robbie A. Ross and
Dare A. Baldwin
Coevolution of Decision-Making
and Social Environments
ELIZABETH BRUCH, ROSS A. HAMMOND, and PETER M. TODD
Abstract
Social scientists have a longstanding theoretical interest in the relationship between
individual behavior and social dynamics. A growing body of work demonstrates
that, when human behavior is interdependent—that is, what one person does
depends on the past, present, or anticipated future actions of others—there is not a
simple or linear relationship between the choices of individuals and their collective
consequences. Outside of the academy, policy makers are increasingly aware that
well-intentioned interventions can backfire if they fail to account for how people
change their behavior in response to the intervention. This type of problem requires
a systematic modeling approach. Our entry provides a brief introduction to a
growing body of research that brings together two disparate literatures—studies of
decision-making and studies of the interplay between individuals’ decisions and
features of the social environment—through dynamic computational modeling.
Cognitive scientists characterize human decision-making under uncertainty using
heuristics, rules-of-thumb that produce satisfactory choices quickly and with
limited information. The heuristics we use and information samples we gather have
profound consequences for the choices we make. At the same time, the social context
defined by the choices of others feeds back to affect individual decision-making.
In recent years, there has been growing interest in methods such as agent-based
modeling and systems dynamics that can capture the dynamic interplay between
individuals’ choices and features of the environment. However, historically these
approaches have not been grounded in cognitively plausible models of human
behavior. We identify areas of high potential for future research, and lay out a
preliminary framework to help guide understanding of the decision-making process
and its consequences in different social domains.
INTRODUCTION
The leading causes of death and disease in the United States are attributed
to behavioral factors such as tobacco use, poor diet and inactivity, alcohol
consumption, risky sexual behavior, and avoidable injuries (Danaei et al.,
2009; Mokdad, Marks, Stroup, & Gerberding, 2004). Behavior is thus central
to the prevention, treatment, and management of diseases and health
Emerging Trends in the Social and Behavioral Sciences. Edited by Robert Scott and Stephen Kosslyn.
© 2015 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.
1
2
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
care, and many interventions are aimed at changing behavior in healthier
directions. However, these interventions are often based on models of
human decision-making that lack empirical support, are difficult to quantify,
or largely ignore the bi-directional feedback between the individual and
the social (for a review of the behavioral theories guiding interventions,
see Glanz & Bishop, 2010). Research that ignores the processes by which
people actually make decisions in real settings or the interdependence of
those decisions may result in mis-specified behavioral models that lead to
incorrect predictions and ineffective policy recommendations.
There is also growing recognition that policy interventions are most likely
to be effective if they adopt ecological or systems perspectives (Huang &
Glass, 2008; Nader et al., 2012; Mabry, Olster, Morgan, & Abrams, 2008;
Sallis, Owen, & Fisher, 2008). The ecological perspective emphasizes that
multiple levels of influence shape behavior (e.g., individual, interpersonal,
organizational, community, and public policy), and the systems perspective
emphasizes the interconnectedness of these levels (including dynamic
feedback between individuals’ actions and their social environment). A
major challenge in solving pernicious social problems is accounting for this
bidirectional feedback between individuals and their environment (Sterman,
2006). In recent years, studies have identified “systems science” methods
such as agent-based modeling (ABM) as a potentially transformative tool
for capturing feedback across multiple levels of analysis (Hammond, 2009;
Luke & Stamatakis, 2012; Mabry, Marcus, Clark, Leischow, & Méndez,
2010). However, the ability of systems science approaches to reach their
full potential in offering new insights about behavior has been limited by
the unmet need for empirically valid models of individual behavior that
can be operationalized computationally and incorporated into the systems
analysis.
Below, we draw on insights from cognitive science and decision theory to
outline an initial framework to help guide the development of more cognitively sophisticated computational models. We identify both key features
of individuals’ choices (and the contexts in which those choices take place)
that may shape the process of decision-making, and also how those decisions
impinge on the current and future decisions of others. We view these as an
initial set of important structural features that determine both what strategies are available for individuals to use in navigating the choice environment,
and also the ways in which the decisions of individuals feed back to shape
the choice environment. We begin with a brief overview of agent-based models aimed at capturing feedback between individuals’ choices and the social
environment. We then review some of the state-of-the-art literature on decision strategies, and how they rely on features of the environment. In the final
section, we outline a framework for exploring the aggregate implications of
Coevolution of Decision-Making and Social Environments
3
“cognitively plausible” decision models as individuals simultaneously react
to and change their social environment.
LINKING INDIVIDUAL BEHAVIOR WITH THE SOCIAL
ENVIRONMENT
Social science has a longstanding interest in the relationship between individuals’ motivations and decisions and large-scale patterns of social organization and change. The “micro–macro problem” concerns how to explicitly
account for the ways in which actions of individuals give rise to social organization and dynamics, rather than assuming that macrolevel phenomena
are simply aggregates of individual characteristics and behavior (Coleman,
1994, p. 197; Granovetter, 1978, p. 1421). The connection between individuals’ actions and their collective consequences would be simple if one could
simply sum over individuals’ intentions or behavior to generate expected
population-level attributes. The problem is that much of human behavior is
interdependent; individuals’ actions often depend on what others are doing.
For example, individuals may directly influence one another through social
norms, peer effects, and other expectations for behavior. Even in situations
where direct social influence plays a minimal role, the alternatives available to people at any given moment may nonetheless depend on the past,
present, or future choices of others. For example, the availability of a job is
often dependent on the decision of its prior occupant to vacate it (Chase,
1991), individuals’ decisions about who to date or marry typically require
reciprocity of affections by the potential mate (Roth & Sotomayor, 1992), and
both require that the options have not already been taken by other searchers
(Todd, 2007).
ABM is a relatively new computational simulation approach specifically
designed to yield insight into how the behavior of decentralized autonomous
actors generate macro-level outcomes of interest, explicitly incorporating
dynamic feedback from macro back to micro. Agent-based models have been
used to study the macro–micro-dynamics governing a host of social processes, ranging from segregation to civil unrest to the outbreak and spread
of disease (Bruch & Mare, 2006; Cioffi-Revilla & Rouleau, 2010; Epstein,
2006; Schelling, 1971). Until recently, however, behavior of individual agents
in an ABM has rarely been grounded in cognitively or neurobiologically
“plausible” rules.1 Typically, agents have been programmed with stylized
rules for behavior, or in some cases with a psychologically unrealistic
statistical model that relates features of the environment to the probability of
1. Exceptions include the work of Todd and colleagues, who explore how heuristics interact with and
create features of the social environment (Todd, Billari, & Simão, 2005; Davis, Todd, & Bullock, 1999); as
well as work by Hammond and Ornstein (2014) and Epstein (2014).
4
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
the agent taking one set of actions or another. The potential has thus largely
been missed for psychologically grounded agent-based models to provide
new insights about how individuals’ choices shape and are shaped by their
social and physical environment. Combining advances in cognitive science
with those in complex systems science can yield a new generation of models
that shed light on both individual behavior and its instantiation in social
environments.
MODELS OF DECISION-MAKING
How do people make decisions? The classical model of decision-making is
the rational actor model endemic to neoclassical economics, which assumes a
fully informed, forward-looking rational actor with unlimited time for information processing (Becker, 1993; Von Neumann & Morgenstern, 2007). However, over the past 40 years, a large body of work has demonstrated that real
people make decisions under conditions of limited time, bounded cognitive
resources, and uncertainty. As a result, people often use heuristics—simple
rules for making a choice or inference—that serve to keep the processing
demands of a decision within the environmental and cognitive limits.
Heuristics are “problem-solving methods that tend to produce efficient
solutions to difficult problems by restricting the search through the space
of possible solutions, on the basis of some evaluation of the structure
of the problem” (Braunstein, 1972, p. 520). While early work from the
“heuristics and biases” school often emphasized the fallibility of human
decision-making (Kahneman, Slovic, & Tversky, 1982), more recent research
on “ecological rationality” shows how individuals’ simple decision strategies
can capitalize on systematic structure in the decision-making environment
(Todd, Gigerenzer, & The ABC Research Group, 2012). In other words, while
people do have limited knowledge and constraints on their ability to process
information, they can nonetheless make good decisions using heuristics that
match (by learning or evolution) to the ways that information is organized in
their environment. Such well-matched heuristics are designed to capitalize
on key features of the decision environment, and so can get away with using
limited information processed in a quick manner.
Heuristics are often composed of building blocks that underlie
decision-making, including: search rules that specify how to seek out
information on available choice alternatives; stopping rules that specify
when a search should be ended; and decision rules that specify how the final
choice is reached (Gigerenzer & Gaissmaier, 2011, p. 456). For example,
some marriage market models consider how people choose a marriage
partner when potential mates can only be explored one at a time, and there
is uncertainty about whether the next person to be encountered will be
Coevolution of Decision-Making and Social Environments
5
better than a currently available partner. One commonly studied heuristic
approach for such challenging sequential choices is to use a “satisficing”
mechanism incorporating building blocks whereby people initially spend
some period of time searching for available options and learning about them,
stop that search after a reasonable amount of time and set an aspiration level
based on what they have experienced, and then decide on the next available
partner encountered who meets that aspiration level (Todd & Miller, 1999;
Todd, Billari, & Simão, 2005).
DECISION TASKS AND DECISION ENVIRONMENTS
Many decision theories emphasize that the heuristics used successfully in
decision-making depend on particular features of the task environment
(Gigerenzer & Gaissmaier, 2011; Payne, Bettman, & Johnson, 1993; Simon,
1990; Todd & Gigerenzer, 2012), so to know what heuristics to build into
psychologically realistic models we must first assess the relevant features
of that environment that shape decision strategies. These features include
attributes of the decision task as well as characteristics of the social and
physical environment in which the decision occurs.2
Key features of decision tasks include:
1. The expected time horizon over which the decision will play out. This
includes the anticipated consequences of the decision as well as the
number of decisions made over a day, year, or lifetime. For example,
decisions about what to eat are made on a daily basis, whereas decisions
about where to live are made on average once every 3–5 years. It follows
that repeated decisions are likely to be governed more by habits and
learning over time than infrequent decisions (Scheibehenne, Mata, &
Todd, 2011). Another dimension of time horizons is the extent to which
decision consequences are immediate or cumulative. For example,
the effects of food choices and physical activity decisions cumulate
gradually over time, with feedback via biological outcomes often
occurring only after a substantial time lag.
2. The extent to which the decision is subject to social influence. This
depends on how much individuals can observe the preferences, strategies, or decision outcomes of others (e.g., exercise choices made in
public vs sleep behaviors done in private). Another aspect of social
influence is the degree to which the successful outcome of a decision is
under the control of the decision maker alone or is affected by others.
2. In this essay, we make an analytical distinction between attributes of decision tasks and attributes
of decision environments. But in practice, because tasks, decisions, and social environments are strongly
intertwined, this distinction may be blurred.
6
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
For example, the successful implementation of a young woman’s
decision to use condoms depends on the cooperation of her partner,
but her choice to take the stairs or the elevator is more individually
determined, both of which will affect her selection of appropriate
decision mechanisms.
3. The decision’s valence in terms of reward seeking versus harm avoidance.
It is well established that human beings are more sensitive to negative
change in their environment than positive change; this is sometimes
referred to as the positive–negative asymmetry effect. In reward-seeking situations, people are typically more tolerant of uncertainty and willing to
take risks (Kahneman & Tversky, 1979). In contrast, when the environment is difficult and dangerous, it is likely that people’s time horizons
will become considerably shorter, changing the heuristics they use.
Key features of decision environments include:
1. The number of alternatives to choose among. When there are only a few
alternatives, decisions are often made through a comparison process
that considers the most important features one at a time until a choice
can be made (e.g., the take-the-best heuristic and others—see Payne
et al., 1993; Rieskamp & Hoffrage, 1999). With more alternatives, a
multi-stage process can be used where each stage reduces the number
of options remaining under consideration. For example, the elimination
by aspects heuristic systematically reduces the number of alternatives
by first eliminating all those that are not good enough on the most
important aspect (e.g., all those restaurants in town that cost more than
$50 per person), then eliminating all those left that are not good enough
on the second aspect (e.g., all those remaining restaurants that are over
ten miles away), and so on until only one option is left (Tversky, 1972).
2. The distribution of alternatives. When satisfactory alternatives are plentiful, the decision maker requires little search or information to decide
among them, as most choices will be good. When good choices are rare,
strategies that search for longer are more appropriate (Fasolo, Hertwig,
Huber, & Ludwig, 2009). The distribution of cue values (which is related
to how informative the cues are) also influences what strategies will
work well (Reimer & Hoffrage, 2012).
3. The extent to which available options depend on the choices of others, and
the degree to which scarce items are replenished. For example, food
options may be sold out at the grocery story, but this could imply high
demand, which usually results in a resupply, calling for choice strategies
that revisit resource locations periodically. In contrast, when two people
marry they are removed from the list of possibilities available to others
Coevolution of Decision-Making and Social Environments
7
for an extended period (Todd, 2007). People may also be sensitive to the
rate of change in their choice environment, as well as the direction of
change (i.e., whether the change is perceived as positive vs negative),
which can favor different heuristics (Dudey & Todd, 2002; Hey, 1982).
4. The redundancy in the environment in terms of the correlation among
feature dimensions of choice alternatives. When there is a high degree
of redundancy, knowing one attribute of a particular alternative tells
the chooser something about its other attributes, so that heuristics
that focus on “one good reason” for making a choice will work well
with a quick information search (Rieskamp & Dieckmann, 2012). In
situations with multiple orthogonal attributes that are important for
choice (e.g., in social settings as when children select friends based
on family background, sex, and mutual interests), heuristics that tally
all those features may be more effective (Fasolo, McClelland, & Todd,
2007).
These components of task environment structure can be used to identify
shared properties of seemingly disparate choice applications, which may
help in the design of more effective policies aimed at changing behavior.
Because features of the decision environment determine what heuristics
will be most efficacious, interventions that succeed in one domain may be
fruitfully applied to another that shares structural features.
THE COEVOLUTION OF DECISION-MAKING
AND THE SOCIAL ENVIRONMENT
In most decision contexts, there is feedback between the choices that individuals make and the environment in which they make them. For example,
there are well-documented peer effects on eating, smoking, medication compliance, exercise, and mate choice in which one’s current choices influence
and are influenced by the witnessed choices of others (Bowers, Place, Todd,
Penke, & Asendorpf, 2012; Crandall, 1988; Lazev, Herzog, & Brandon, 1999;
Todd & Minard, 2014). In addition, the choices of individuals at one time
point can shape what options are available for future individuals. The classic example from social science is neighborhood tipping: Each individual
who leaves a neighborhood because she cannot tolerate its racial composition changes its composition and that of the neighborhood she moves into
(Schelling, 1978). Over time, the neighborhood choices available to others
evolve as a product of previous mobility decisions. This phenomenon can
also be seen in what food products are available for purchase in different
areas, which reflects aggregate demand. Finally, the success of a decision may
depend on the behavior of another person. For example, children are subject
8
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Individual strategies for
navigating choice environments
Features of choice environments
h
Population attributes
Preferences or choices of others
Strategies of others
Social norms/expectations
Gather information
from the social
environment
b
d
Desires
Beliefs
Values
a
e
Choice/
action
e
Structural features of
decision-making environment
Number & variety of choices
Frequency of decision
Accountable to others
Success dependent on others
Figure 1
f
c
Sample from
available
choice options
e
g
Select/apply
decision
strategy
Interaction between individuals and environment.
to the food consumption decisions of their parents and people are more likely
to exercise if they are accountable to a workout partner (Dishman, Sallis, &
Orenstein, 1985; Wardle, Guthrie, Sanderson, Birch, & Plomin, 2001).
Figure 1 illustrates how individual strategies for decision-making interact with features of the environment. The left hand side represents features
of both the social environment (e.g., the demographic make-up of the population, the behaviors of others, and social norms and other expectations)
and the choice domain. The social context (in particular, the choices of others) influences the number and type of options available (a). Individuals will
observe some subset of the social environment (b); these observations of the
social environment may influence their preferences, beliefs, and/or expectations (d). For a particular choice domain, the individual will sample some set
of options from the environment (c) which will—in conjunction with preferences and in some cases observing the decisions of others—determine their
decision-making strategy (e). This strategy results in a particular action outcome (f). Individuals’ choices may affect the choices available to others; for
example in the case of mate choice, a pairing will eliminate those two people as options for others (g). Feedback to the social environment occurs both
because the individual’s choice may change his or her other attributes as a
member of the population (e.g., weight and location) and also because his or
her choice may be observed by others (e.g., eating in a group) (h).
Coevolution of Decision-Making and Social Environments
9
Agent-based models that incorporate realistic decision heuristics to predict
individual-environment interactions have begun to appear in the literature.
For example, psychologically plausible satisficing heuristics have been used
in models of mate choice, where individual agents first engage in an adolescent “dating” period where they meet a succession of potential mates of varying levels of quality, learn how well they can do at attracting those potential
mates, adjust their own aspiration level for the kind of long-term mate they
should seek in the future on the basis of those initial dates (raising their aspirations after successful interactions and lowering them after unsuccessful
ones), and then enter the true mate-choice phase where they make marriage
offers to individuals they encounter who meet their aspirations and get married and removed from the population when the offer is mutual—thereby
changing the choice environment for all those agents still remaining in the
mating market (Todd & Miller, 1999). This model predicts observed demographic patterns of the ages at which people get married (Todd et al., 2005)
and demonstrates the strength of the effect of others’ decisions on one’s own
best choice strategy (Todd, 2007); other models have shown how the social
norms that guide individuals’ choices of an appropriately aged spouse can
evolve (Billari, Prskawetz, & Fürnkranz, 2003). A similar model in different domain shows related effects on the best strategy to use for searching
for a parking space (another type of sequential choice) when the choices
made by other earlier parkers creates the environment—here the spatial layout of available spaces—for later drivers (Hutchinson, Fanselow, & Todd,
2012). In the context of food choice, recent work using agent-based models
grounded in neurobiology describes the influence of food environments on
preferences (Hall, Hammond, & Rahmandad, 2014; Hammond et al., 2012)
and the bi-directional coevolution of body weight and social norms (Hammond & Ornstein, 2014).
FUTURE DIRECTIONS AND CHALLENGES
Agent-based simulation models that put agents with realistic psychological
decision mechanisms into social environments can be very useful in enabling
researchers to learn about aspects of cognition and behavior that are otherwise difficult to study (Todd, 1996). First, such models can provide existence proofs, showing that particular hypothesized cognitive mechanisms
can lead to particular observed patterns of behavior in specific environments.
Second, they can elucidate the dynamics of an interacting population over
time, helping us to understand what mechanisms and conditions can lead
to the appearance or disappearance of particular behaviors or environment
structures. And third, as “runnable thought experiments,” they can help us
explore complex interactions for which our intuitions are usually inadequate,
10
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
and thereby come up with predictions—including for effective interventions
to change behavior—that can be tested with empirical research.
A key challenge for future researchers interested in seeding their dynamic
models with realistic human behavior is identifying what decision rules people are using, and how those rules depend on features of both the decision
task and the social environment. The classic approach for studying how people make decisions is experiments that systematically vary features of both
the decision task and the environment and observe outcomes. Commonly,
computer-based choice situations are created in which people are presented
with two options to choose between, and they can examine different pieces
of evidence for each option (e.g., by clicking on a “show price” or “show
mileage” button for choosing a car) until they have seen enough to decide,
at which point their search, stopping, and decision rules can be assessed
(Bröder, 2012; Payne et al., 1993; Rieskamp & Hoffrage, 1999).
One promising source of information on choice processes is the behavioral
data produced through activities online. Dating websites, housing search
sites, job search sites, Facebook, and other electronic venues provide a
detailed look at how people navigate these decision processes. However,
an open question is the degree to which decision strategies observed online
can be generalized to the same behaviors in other settings. For example,
people visiting online dating sites may be confronted with thousands of
potential mates over a short time period, while in their day-to-day lives they
only encounter potential mates sporadically over an extended period. These
task and environment differences may result in differences in strategy and
selectivity as well (Lenton, Fasolo, & Todd, 2010).
There is a growing demand and opportunity for behaviorally sophisticated
models of individual behavior that can also capture bi-directional feedback
with social dynamics. We believe that advances in computational simulation
and in cognitive science mean the time is ripe for rapid progress in connecting
these two fields (to mutual benefit). This “emerging trend” is already underway, and offers the promise of both new insights into complex human behavior, and the design of more effective and efficient policies and interventions.
REFERENCES
Becker, G. (1993). Nobel lecture: The economic way of looking at behavior. Journal of
Political Economy, 101, 385–409.
Billari, F. C., Prskawetz, A., & Fürnkranz, J. (2003). On the cultural evolution of
age-at-marriage norms. In F. C. Billari & A. Prskawetz (Eds.), Agent-based computational demography: Using simulation to improve our understanding of demographic
behavior (pp. 139–157). Heidelberg, Germany: Physica Verlag.
Bowers, R. I., Place, S. S., Todd, P. M., Penke, L., & Asendorpf, J. B. (2012). Generalization in mate choice copying in humans. Behavioral Ecology, 23, 112–124.
Coevolution of Decision-Making and Social Environments
11
Braunstein, M. L. (1972). Perception of rotation in depth: A process model. Psychological Review, 79(6), 510–524.
Bröder, A. (2012). The quest for take-the-best: Insights and outlooks from experimental research. In P. M. Todd, G. Gigerenzer & The ABC Research Group (Eds.),
Ecological rationality: Intelligence in the world (pp. 216–240). New York, NY: Oxford
University Press.
Bruch, E., & Mare, R. (2006). Neighborhood choice and neighborhood change. American Journal of Sociology, 3, 667–709.
Chase, I. D. (1991). Vacancy chains. Annual Review of Sociology, 1, 133–154.
Cioffi-Revilla, C., & Rouleau, M. (2010). MASON RebeLand: An agent-based model
of politics, environment, and insurgency. International Studies Review, 1, 31–52.
Coleman, J. S. (1994). Foundations of social theory. Cambridge, MA: Harvard University Press.
Crandall, C. S. (1988). Social contagion of binge eating. Journal of Personality and Social
Psychology, 55(4), 588.
Danaei, G., Ding, E. L., Mozaffarian, D., Taylor, B., Rehm, J., Murray, C. J. L., & Ezzati,
M. (2009). The preventable causes of death in the united states: Comparative risk
assessment of dietary, lifestyle, and metabolic risk factors. PLoS Medicine, 6(4),
e1000058.
Davis, J. N., Todd, P. M., & Bullock, S. (1999). Environment quality predicts parental
provisioning decisions. Proceedings of the Royal Society of London B: Biological Sciences, 266, 1791–1797.
Dishman, R. K., Sallis, J. F., & Orenstein, D. R. (1985). The determinants of physical
activity and exercise. Public Health Reports, 100, 158.
Dudey, T., & Todd, P. M. (2002). Making good decisions with minimal information:
Simultaneous and sequential choice. Journal of Bioeconomics, 3, 195–215.
Epstein, J. M. (2014). Agent_Zero: Toward Neurocognitive Foundations for Generative
Social Science. Princeton, NJ: Princeton University Press.
Fasolo, B., Hertwig, R., Huber, M., & Ludwig, M. (2009). Size, entropy, and density: What is the difference that makes the difference between small and large
real-world assortments? Psychology & Marketing, 26, 254–279.
Fasolo, B., McClelland, G. H., & Todd, P. M. (2007). Escaping the tyranny of choice:
When fewer attributes make choice easier. Marketing Theory, 7, 13–26.
Gigerenzer, G., & Gaissmaier, W. (2011). Heuristic decision making. Annual Review
of Psychology, 62, 451–82.
Glanz, K., & Bishop, D. B. (2010). The role of behavioral science theory in development and implementation of public health interventions. Annual Review of Public
Health, 31, 399–418.
Granovetter, M. (1978). Threshold models of collective behavior. American Journal of
Sociology, 83(6), 1420–1443.
Hall, K. D., Hammond, R. A., & Rahmandad, H. (2014). Dynamic interplay among
homeostatic, hedonic, and cognitive feedback circuits regulating body weight.
American Journal of Public Health, e1–e7.
Hammond, R. A. (2009). “Complex systems modeling for obesity research”. Preventing Chronic Disease 6(3), A97.
12
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Hammond, R. A., & Ornstein, J. T. (2014). A model of social influence on body mass
index. Annals of the New York Academy of Science, 1331, 34–42.
Hammond, R. A., Ornstein, J. T., Fellows, L. K., Dube, L., Levitan, R., & Dagher, A.
(2012). A model of food reward learning with dynamic reward exposure. Frontiers
in Computational Neuroscience, 6, 82.
Hey, J. D. (1982). Search for rules for search. Journal of Economic Behavior and Organization, 3, 65–81.
Huang, T. T. K., & Glass, T. A. (2008). Transforming research strategies for understanding and preventing obesity. JAMA, 300(15), 1811–1813.
Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under
risk. Econometrica, 47(2), 263–292.
Lazev, A. B., Herzog, T. A., & Brandon, T. H. (1999). Classical conditioning of environmental cues to cigarette smoking. Experimental and Clinical Psychopharmacology,
7, 56.
Lenton, A. P., Fasolo, B., & Todd, P. M. (2010). Who is in your shopping cart? Expected
and experienced effects of choice abundance in the online dating context. In N.
Kock (Ed.), Evolutionary psychology and information systems research: A new approach
to studying the effects of modern technologies on human behavior (pp. 149–167). New
York, NY: Springer.
Mabry, P. L., Marcus, S. E., Clark, P. I., Leischow, S. J., & Méndez, D. (2010). Systems
science: a revolution in public health policy research. American Journal of Public
Health, 100, 1161–1163.
Mokdad, A. H., Marks, J. S., Stroup, D. F., & Gerberding, J. L. (2004). Actual causes
of death in the United States, 2000. Journal of the American Medical Association, 10,
1238–1245.
Nader, P. R., Huang, T. T. K., Gahagan, S., Kumanyika, S., Hammond, R. A., &
Christoffel, K. K. (2012). Next steps in obesity prevention: Altering early life systems to support healthy parents, infants, and toddlers. Childhood Obesity, 8(3),
195–204.
Reimer, T., & Hoffrage, U. (2012). Ecological rationality for teams and committees:
Heuristics in group decision making. In P. M. Todd, G. Gigerenzer & the ABC
Research Group (Eds.), Ecological rationality: Intelligence in the world (pp. 335–359).
New York, NY: Oxford University Press.
Rieskamp, J., & Dieckmann, A. (2012). Redundancy: Environment structure that simple heuristics can exploit. In P. M. Todd, G. Gigerenzer & The ABC Research Group
(Eds.), Ecological rationality: Intelligence in the world (pp. 187–215). New York, NY:
Oxford University Press.
Rieskamp, J., & Hoffrage, U. (1999). When do people use simple heuristics, and how
can we tell?. In G. Gigerenzer, P. M. Todd & The ABC Research Group (Eds.), Simple
heuristics that make us smart (pp. 141–167). New York, NY: Oxford University Press.
Roth, A. E., & Sotomayor, M. A. O. (1992). Two-sided matching: A study in game-theoretic
modeling and analysis. Cambridge, England: Cambridge University Press.
Sallis, J., Owen, N., & Fisher, E. (2008). Ecological models of health behavior. In
Glanz, K., B. K. Rimer, and K. Viswanath (Eds.), Health behavior and health education: Theory, research, and practice, (pp. 465–486). San Francisco, CA: John Wiley
& Sons, Inc.
Coevolution of Decision-Making and Social Environments
13
Scheibehenne, B., Mata, J., & Todd, P. M. (2011). Older but not wiser—Predicting
spouse’s preferences gets worse with age. Journal of Consumer Psychology, 21,
184–191.
Schelling, T. C. (1971). Dynamic models of segregation. Journal of Mathematical Sociology, 2, 143–186.
Schelling, T. C. (1978). Micromotives and macrobehavior. New York, NY: WW Norton
& Company.
Simon, H. A. (1990). Invariants of human behavior. Annual Review of Psychology, 41(1),
1–20.
Sterman, J. D. (2006). Learning from evidence in a complex world. American Journal
of Public Health, 3, 505–514.
Todd, P. M. (1996). The causes and effects of evolutionary simulation in the behavioral sciences. In R. Belew & M. Mitchell (Eds.), Adaptive individuals in evolving
populations: Models and algorithms (pp. 211–224). Reading, MA: Addison-Wesley.
Todd, P. M. (2007). Coevolved cognitive mechanisms in mate search: Making decisions in a decision-shaped world. In J. P. Forgas, M. G. Haselton & W. von Hippel
(Eds.), Evolution and the social mind: Evolutionary psychology and social cognition (pp.
145–159 (Sydney Symposium of Social Psychology series)). New York, NY: Psychology Press.
Todd, P. M., Billari, F. C., & Simão, J. (2005). Aggregate age-at-marriage patterns from
individual mate-search heuristics. Demography, 42, 559–574.
Todd, P. M., & Gigerenzer, G. (2012). Ecological Rationality: Intelligence in the World.
Oxford, England: Oxford University Press.
Todd, P. M., & Miller, G. F. (1999). From pride and prejudice to persuasion: Satisficing
in mate search. In G. Gigerenzer, P. M. Todd & The ABC Research Group (Eds.),
Simple heuristics that make us smart (pp. 287–308). New York, NY: Oxford University
Press.
Todd, P. M., & Minard, S. L. (2014). Simple heuristics for deciding what to eat. In
S. D. Preston, B. Knutson & M. Kringelbach (Eds.), The interdisciplinary science of
consumption. Cambridge, MA: MIT Press.
Tversky, A. (1972). Elimination by aspects: A theory of choice. Psychological Review,
79(4), 281–299.
Von Neumann, J., & Morgenstern, O. (2007). Theory of games and economic behavior
(Commemorative ed.). Princeton, NJ: Princeton University Press.
Wardle, J., Guthrie, C., Sanderson, S., Birch, L., & Plomin, R. (2001). Food and activity
preferences in children of lean and obese parents. International Journal of Obesity &
Related Metabolic Disorders, 25(7), 971–977.
FURTHER READING
Agent-Based Modeling
Epstein, J. (2006). Generative social science: Studies in agent-based computational modeling.
Princeton, NJ: Princeton University Press.
14
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Macy, M., & Willer, R. (2002). From factors to actors: Computational sociology and
agent-based modeling. Annual Review of Sociology, 28, 143–166.
Miller, J., & Page, S. (2009). Complex adaptive systems: An introduction to computational
models of social life: An introduction to computational models of social life. Princeton,
NJ: Princeton University Press.
Systems Science and Public Policy
Luke, D. A., & Stamatakis, K. A. (2012). Systems science methods in public health:
Dynamics, networks, and agents. Annual Review of Public Health, 33, 357–376.
Mabry, P. L., Marcus, S. E., Clark, P. I., Leischow, S. J., & Mendez, D. (2010). Systems
science: a revolution in public health policy research. American Journal of Public
Health, 100, 1161–1163.
Mabry, P. L., Olster, D. H., Morgan, G. D., & Abrams, D. B. (2008). Interdiciplinarity
and systems science to improve population health: a view from the NIH Office of
Behavioral and Social Sciences Research. American Journal of Preventive Medicine,
35(S2), 11–24.
Heuristic Decision-Making
Gigerenzer, G., Hertwig, R., & Pachur, T. (2011). Heuristics: The foundations of Adaptive
Behavior. New York, NY: Oxford University Press.
Gigerenzer, G., Todd, P. M., & The ABC Research Group (1999). Simple heuristics that
make us smart. New York, NY: Oxford University Press.
Kahneman, D., Slovic, P., & Tversky, A. (Eds.) (1982). Judgment under uncertainty:
heuristics and biases. Cambridge, England: Cambridge University Press.
Payne, J., Bettman, J., & Johnson, E. (1993). The adaptive decision maker. Cambridge,
England: Cambridge University Press.
Co-Evolution of Heuristic Strategies and Social Environments
Hammond, R. & Ornstein, J. (2014). A model of social influence on body weight.
Annals of the New York Academy of Science. Forthcoming.
Hutchinson, J. M. C., Fanselow, C., & Todd, P. M. (2012). Car parking as a game
between simple heuristics. In P. M. Todd, G. Gigerenzer & The ABC Research
Group (Eds.), Ecological rationality: Intelligence in the world (pp. 454–484). New York,
NY: Oxford University Press.
Todd, P. M., Billari, F. C., & Simão, J. (2005). Aggregate age-at-marriage patterns from
individual mate-search heuristics. Demography, 42(3), 559–574.
ELIZABETH BRUCH SHORT BIOGRAPHY
Elizabeth Bruch, PhD is an Assistant Professor in sociology and complex
systems, and Affiliate of the Population Studies Center at the Institute for
Coevolution of Decision-Making and Social Environments
15
Social Research. Much of her work blends statistical and agent-based methods to examine the relationship between individuals’ decisions about where
to live and patterns of residential segregation. She is also working on a project
exploring how individuals’ mate search strategies and willingness to settle
intersect with demographic constraints to generate dating or marriage market dynamics in US metro areas. She has served as a consulting editor for the
American Journal of Sociology, and on the editorial board for Sociological
Methodology. Her work has received the Gould Prize from the American
Journal of Sociology, as well as best article prizes from the Mathematical
and Urban Sociology sections of American Sociological Association. She is
a member of the National Institute of Health’s Network on Inequality, Complexity, and Health.
ROSS A. HAMMOND SHORT BIOGRAPHY
Ross A. Hammond, PhD is a Senior Fellow in Economic Studies at the
Brookings Institution, where he is also the Director of the Center on Social
Dynamics & Policy. His primary area of expertise is modeling complex
dynamics in economic, social, and public health systems using methods
from complexity science. His current research topics include obesity etiology
and prevention, food systems, tobacco control, behavioral epidemiology,
crime, corruption, segregation, and decision-making. Hammond received
his BA from Williams College and his PhD from the University of Michigan. He has authored numerous scientific articles, and his work has been
featured in New Scientist, Salon, The Atlantic Monthly, Scientific American,
and major news media. Hammond was recently appointed to the Institute
of Medicine/National Research Council committee Framework for Assessing
the Health, Environmental, and Social Effects of the Food System, and is both a
Public Health Advisor at the National Cancer Institute and an Advisory
Special Government Employee at the FDA Center for Tobacco Products. He
serves on the editorial boards of the journals Behavioral Science & Policy and
Childhood Obesity, and has been a member of the NIH-funded research networks MIDAS (Models of Infectious Disease Agent Study), ENVISION (part
of the National Collaborative on Childhood Obesity Research), and NICH
(Network on Inequality, Complexity, and Health). Hammond currently
holds appointments at the Harvard School of Public Health, Washington
University, and University of Michigan. He has been a consultant to the
World Bank, the Asian Development Bank, the Food and Drug Administration, the Institute of Medicine, and the National Institutes of Health. He
has taught computational modeling at Harvard, the University of Michigan,
Washington University, the National Cancer Institute, and the NIH/CDC
Institute on Systems Science and Health.
16
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
PETER M. TODD SHORT BIOGRAPHY
Peter M. Todd grew up in Silicon Valley, studied mathematics and electronic
music at Oberlin College, received an MPhil in computer speech and
language processing from Cambridge University, and developed neural
network models of the evolution of learning for his PhD in psychology at
Stanford University. In 1995, he moved to Germany to help found the Center
for Adaptive Behavior and Cognition (ABC), based at the Max Planck Institute for Human Development in Berlin. The Center’s work was captured in
the book Simple Heuristics That Make Us Smart (Gigerenzer, Todd, and The
ABC Research Group; Oxford, 1999); the sequel, Ecological Rationality: Intelligence in the World, covering information-environment structures and their
impact on decision-making, came out in 2012, along with a book on search
behavior, Cognitive Search: Evolution, Algorithms, and the Brain (Todd, Hills,
and Robbins, eds.; MIT Press). Todd moved to Indiana University in Bloomington as Professor of Cognitive Science, Psychology, and Informatics in 2005
and set up the ABC-West lab there (http://www.indiana.edu/∼abcwest/).
His ongoing research interests cover the interactions between and coevolution of decision-making and decision environments, focusing on the
ways that people and other animals search for resources—including mates,
information, and food—in space and time.
RELATED ESSAYS
Models of Revealed Preference (Economics), Abi Adams and Ian Crawford
Choice Architecture (Psychology), Adrian R. Camilleri and Rick P. Larrick
Behavioral Economics (Sociology), Guy Hochman and Dan Ariely
Emotion and Decision Making (Psychology), Jeff R. Huntsinger and Cara Ray
Against Game Theory (Political Science), Gale M. Lucas et al.
From Individual Rationality to Socially Embedded Self-Regulation (Sociology), Siegwart Lindenberg
Event Processing as an Executive Enterprise (Psychology), Robbie A. Ross and
Dare A. Baldwin
