Emerging Trends in The Social and Behavioral Sciences

Herd Behavior

Item

Title

Herd Behavior

Author

Kameda, Tatsuya

Hastie, Reid

Research Area

Class, Status and Power

Topic

Social Movements

Abstract

There are many manifestations of herding in the human species—one of the most socially interdependent species on the earth. Herding here refers to an alignment of thoughts or behaviors of individuals in a group through local interactions among individuals rather through than some purposeful coordination by a central authority in the group. Herding underlies many collective phenomena in the Internet era, ranging from everyday social behavior, consumer choices, economic bubbles, and political movements. Accumulating evidence in various behavioral science disciplines suggests that we humans are equipped with neural, psychological, and behavioral mechanisms that constitute our highly socially sensitive minds. These built‐in mechanisms are evolutionary products that have promoted our survival. Yet, these adaptive tools can cause serious errors in modern environments, in which interconnectivities of individuals are much denser and externalities accruing from individual behaviors are much greater and more far‐reaching, compared to primordial environments in which the human mind evolved. Growing evidence in the behavioral sciences also suggests that the two contrasting collective phenomena in humans, maladaptive herding and the wisdom of crowds, are based on similar underlying mechanisms. In this sense, the two apparently opposite macro phenomena may be seen as twins produced and governed by the social receptivity of our minds. Given this commonality, understanding the neural, psychological, and behavioral mechanisms that could distinguish these twins will be one of the most important challenges for behavioral sciences in the next decade.

Related Essays

Culture and Cognition (Sociology), Karen A. Cerulo

Peers and Adolescent Risk Taking (Psychology), Jason Chein

Adaptation for Culture (Anthropology), Jill M. Church

The Inherence Heuristic: Generating Everyday Explanations (Psychology), Andrei Cimpian

Misinformation and How to Correct It (Psychology), John Cook et al.

Problems Attract Problems: A Network Perspective on Mental Disorders (Psychology), Angálique Cramer and Denny Borsboom

Household Wealth Effects and the US Macroeconomy (Economics), Karen Dynan and Daniel Cooper

Micro-Cultures (Sociology), Gary Alan Fine

State of the Art in Competition Research (Psychology), Márta Fülöp and Gábor Orosz

Economics and Culture (Economics), Gérard Roland

Emotion and Decision Making (Psychology), Jeff R. Huntsinger and Cara Ray

Genetic Foundations of Attitude Formation (Political Science), Christian Kandler et al.

Search and Learning in Markets (Economics), Philipp Kircher

Cultural Neuroscience: Connecting Culture, Brain, and Genes (Psychology), Shinobu Kitayama and Sarah Huff

Niche Construction: Implications for Human Sciences (Anthropology), Kevin N. Laland and Michael O'Brien

From Individual Rationality to Socially Embedded Self-Regulation (Sociology), Siegwart Lindenberg

Two-Systems View of Children's Theory-of-Mind Understanding (Psychology), Jason Low

Implicit Attitude Measures (Psychology), Gregory Mitchell and Philip E. Tetlock

A Bio-Social-Cultural Approach to Early Cognitive Development: Entering the Community of Minds (Psychology), Katherine Nelson

Organizations and the Production of Systemic Risk (Sociology), Charles Perrow

The Sociology of Religious Experience (Sociology), Douglas Porpora

War and Social Movements (Political Science), Sidney Tarrow

The Social Science of Sustainability (Political Science), Johannes Urpelainen

Theory of Mind (Psychology), Henry Wellman

Identifier

etrds0157

extracted text

Herd Behavior
TATSUYA KAMEDA and REID HASTIE

Abstract
There are many manifestations of herding in the human species—one of the most
socially interdependent species on the earth. Herding here refers to an alignment of
thoughts or behaviors of individuals in a group through local interactions among
individuals rather through than some purposeful coordination by a central authority in the group. Herding underlies many collective phenomena in the Internet era,
ranging from everyday social behavior, consumer choices, economic bubbles, and
political movements. Accumulating evidence in various behavioral science disciplines suggests that we humans are equipped with neural, psychological, and behavioral mechanisms that constitute our highly socially sensitive minds. These built-in
mechanisms are evolutionary products that have promoted our survival. Yet, these
adaptive tools can cause serious errors in modern environments, in which interconnectivities of individuals are much denser and externalities accruing from individual
behaviors are much greater and more far-reaching, compared to primordial environments in which the human mind evolved. Growing evidence in the behavioral
sciences also suggests that the two contrasting collective phenomena in humans,
maladaptive herding and the wisdom of crowds, are based on similar underlying
mechanisms. In this sense, the two apparently opposite macro phenomena may be
seen as twins produced and governed by the social receptivity of our minds. Given
this commonality, understanding the neural, psychological, and behavioral mechanisms that could distinguish these twins will be one of the most important challenges
for behavioral sciences in the next decade.

In an increasingly connected world, an event in one place, be it economic,
political, or social, can cause large-scale chain reactions across many other
places. We have abundant examples of this sort, including the recent
global financial crisis, the spread of civil uprisings in the Middle East, the
widespread adoption of technological innovations such as the iPad, and so
on. Until recently, such mass phenomena have been studied sporadically
across social science disciplines without much mutual communication. Yet,
with advances in technology and new theoretical frameworks, these mass
phenomena are becoming a focus of substantial interdisciplinary interests
(Akerlof & Shiller, 2009). An umbrella concept, “herding,” has facilitated
such cross-disciplinary communication over the past 5 years.
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

HERD BEHAVIOR: A DEFINITION AND EXAMPLES
What is meant by herding? Herding refers to an alignment of thoughts or
behaviors of individuals in a group. Most importantly, such convergence
often emerges through local interactions among agents rather than through
purposeful coordination by a central authority or a leading figure in the
group. In other words, the apparent coordination of the herd is an emergent
property of local interactions (Raafat, Chater, & Frith, 2009).
Textbook examples of herding in the social science literature include riots,
panics, fads, mass hysterias, urban legends, economic bubbles, and so on
(Smelser, 1963; Turner & Killian, 1993). However, besides these familiar
examples, recent research suggests that herding may encompass a much
wider range of our social behaviors than had been previously thought.
CRIMES
Proliferation of crimes in a city may be seen as an example of herding. One
of the most striking aspects of crime is that crime rates vary dramatically
across time and space. For example, homicide rates across nations ranged
from 6.1 cases per million in Japan, 12.6 in Sweden, to 98.0 in the United
States in 1990. Within the United States, rates of serious crimes in the year
ranged from 0.008 per capita in Ridgewood Village, New Jersey to 0.384 in
nearby Atlantic City (Glaeser, Bruce Sacerdote, & Scheinkman, 1996). Such
high variances are observed within cities as well, where one street can have
much higher crime rates than streets just a few blocks away.
One obvious explanation for such variety may be that socioeconomic conditions also vary over time and space, creating temporal and geographical
clusters of crime. However, an econometric analysis by Edward Glaeser and
others showed that less than 30% of the variation in cross-city or cross-district
crime rates could be explained by the local socioeconomic differences. These
researchers developed a model in which agents’ decisions about crime were a
function of their own attributes (e.g., socioeconomic as well as psychological
attributes) and of their neighbor’s decisions about criminal activities. Glaeser
and others then estimated impacts of the second element of the model (i.e.,
social influence from neighbors) for a variety of crimes in the United States
in 1985, in 1970, and across New York City in 1985. The results showed that a
positive interaction among agents’ decisions about crime was the only viable
explanation for the large residual variance not explained by the local socioeconomic conditions. More specifically, the local social influence was strong
for larceny and auto theft; moderate for assault, burglary, and robbery; and
weak for arson, murder, and rape. These results suggest that one agent’s decision to commit crimes (especially minor crimes) affects his or her neighbor’s
decisions, which constitutes a positive feedback loop for the collective. The

Herd Behavior

3

large variations in crime rates across time and space seem to emerge as aggregated outcomes of such individual local decisions.
OBESITY
Recent research suggests that obesity may be contagious as well. Using a data
set from a longitudinal survey on cardiovascular disease (the Framingham
Heart Study; see http://www.framinghamheartstudy.org/), Nicholas Christakis and James Fowler (2007) examined how social relations in a community
affect obesity. The original survey traced health states of people residing in
Framingham, Massachusetts, over 32 years. Christakis and Fowler focused
on family and friendship relations among the participants, and applied longitudinal statistical models to examine whether weight gain in one person
was associated with weight gains in his or her friends, siblings, spouse, and
neighbors.
Results of the network analysis revealed that obese people (defined as those
with a body mass index >30) and nonobese people formed different clusters
and that social influences through the network extended up to three degrees
of separation. In other words, the average obese person was more likely to
have obese friends, friends of friends, and friends of friends of friends than
was the average nonobese person. Moreover, a person’s chances of becoming obese increased by 57% if he or she had a friend who became obese in the
time period, by 40% if a sibling became obese, and by 37% if a spouse became
obese. These patterns suggest that obesity spreads through social network
similar to a pathogen. A person’s overeating behavior is affected through the
social network, even if one may not know another overeater directly. Segmentations of obese and nonobese people in a community seem to emerge
as aggregated consequences of local influences (see also Lyons, 2011, for criticisms of the social network analysis employed by Christakis and Fowler).
HAPPINESS
Happy people and unhappy people also seem to inhabit different clusters in a
community. A reanalysis of the Framingham Heart Study data set suggested
that these clusters did not simply reflect a tendency for individuals to associate with similar individuals. Instead, these macro patterns resulted from
spread of happiness and unhappiness through the social network, just as in
the case of obesity. According to the analysis, the probability that one was
happy increased by 25% if a friend who lived within a mile became happy,
and these local influences also extended up to three degrees of separation.
Thus, similar to obesity, happiness also seems to be contagious (Fowler &
Christakis, 2008).

4

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

WHY DOES HERDING OCCUR?—POTENTIAL MECHANISMS
The earlier examples suggest that herding is a robust phenomenon, characterizing a wide range of social behaviors in our life. If so, what are the neural,
psychological, or sociological mechanisms that produce herding?
EMOTIONAL CONTAGION, FACIAL MIMICRY, AND MIRROR NEURONS
As implied by the saying that your smile makes others happy, humans
often reproduce others’ emotions in themselves. This phenomenon, which
is called emotional contagion (Hatfield, Cacioppo, & Rapson, 1994), has
long been known among psychotherapists who treat depressed clients.
Therapists, especially those who are inexperienced, sometimes “catch”
their clients’ emotions expressed during interviews, and feel themselves
depressed afterwards. Elaine Hatfield and her colleagues see emotional
contagion as a primitive, automatic, and unconscious process. It occurs
through a series of steps: when a receiver is interacting with a sender, he or
she first perceives the emotional expressions of the sender. The receiver then
automatically transfers the perceived emotional expressions to his or her
bodily expressions (e.g., facial expressions, postures). Through the process
of afferent feedback, these mimicked bodily expressions are translated into
the receiver feeling the same emotion that the sender experienced, which
leads to emotional convergence among the sender and the receiver.
Indeed, it can easily be demonstrated that we have a tendency to mimic
the facial expressions of others in everyday social interactions. Research suggests that such facial mimicry is an automatic, reflex-like process, in which
the observer’s facial expression matches the observed facial expression (e.g.,
happy, sad, fearful, angry, disgusted faces) rather quickly—typically within
less than a second (Hess & Blairy, 2001). Such automatic mimicry extends to
bodily posture, voice pitch, and so on, and is known to emerge very early in
human development. Even 12- to 21-day-old infants imitate both facial and
manual gestures displayed by an adult model (Meltzoff & Moore, 1977).
Furthermore, recent developments in neuroscience suggest that there may
be a biological system in our brains that helps us to mirror others’ actions.
One of the most intriguing recent findings in brain science is the discovery of
“mirror neurons.” In the late 1980s when Giacomo Rizzolatti and others were
recording electrical activity in the brain of a macaque, these researchers found
neurons that fired both when the animal acted and when the animal observed
the same action performed by another. The same neurons fired when the
monkey grasped something with its hand, and when the monkey observed
the experimenter grasping it. However, these neurons did not discharge in
response to simple presentation of food or other interesting objects. The neuron “mirrored” the motor behavior of the other, as though the observer were

Herd Behavior

5

itself executing the motor act. Although it remains controversial, some recent
data suggest that a similar “mirror neuron system” exists in human brains as
well (Rizzolatti & Craighero, 2004).
Taken together, these psychological, behavioral, and neural findings
strongly suggest that mimicking others may be a fundamentally human
activity.
SOCIAL NORMS, MUTUAL EXPECTATIONS, AND SHARED STORIES
Another mechanism for herding involves more conscious, deliberate, and
controlled psychological processes, that are distinguishable from our automatic “aping” propensities as reviewed earlier. These processes have been
studied mainly by social psychologists.
Classic experimental demonstrations of such herd behaviors in social psychology include the famous line-comparison perception study by Solomon
Asch, where subjects conformed to an erroneous majority view to avoid
potential embarrassment or other social consequences in a group (Asch,
1956); the optical judgment study by Muzfer Sherif demonstrating that
individual perceptions of the autokinetic illusion converged to a shared
social reality through communication (i.e., everybody in the same group
ended up experiencing a similar optical illusion Sherif, 1936), and so on.
A key element underlying these herd behaviors is a fundamental characteristic of our mind, which may be labeled docility or receptivity to social norms
(Kameda & Tindale, 2006; Simon, 1990). Herbert Simon defined this concept
as our tendency to depend on others’ suggestions, recommendation, persuasion, and information obtained through social channels as a major basis of
choice. Compared to other gregarious species, humans are unique in developing social norms and mutually shared expectations, which inform us about
what action is normal, appropriate, or fair in a given social situation. As seen
in the Asch experiment, the human mind is built to be receptive to social
norms, and tends to self-censor actions in order to avoid violating norms.
Notice that the high receptivity to social norms is also fundamental to our
ability to learn culturally. Humans are a cultural species that can take full
advantage of socially acquired knowledge. Without docility by learners to
their “cultural parents,” such cognitive capacities would be highly limited
(Tomasello, 1999).
The human mind is also built to think in terms of, and be influenced by,
narratives or stories (sequences of events with an internal logic and dynamics: Shank & Abelson, 1977). Stories, especially stories shared in a community
or across a whole society, lead us to see, interpret, have feelings about, and
react to experiences from a shared perspective (see Akerlof & Shiller, 2009,
for interesting recent examples of influential political-economic stories).

6

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

This characteristic social receptivity of individual minds can yield effects
that are visible at the societal level. In the aforementioned case of contagious
obesity, for example, one may decide to eat more because the action seems to
be normal given one’s spouse’s or friend’s eating practices, which in turn provides a normative signal for another’s overeating. Our actions have spillover
effects (which economists call externalities) on others, which can lead to spiraling proliferations of action across a whole society (Granovetter, 1978).
RATIONAL CONFORMITY AND INFORMATION CASCADES
Sometimes it is rational to conform to a majority behavior in a group, even
if one would otherwise choose differently. Hans Christian Andersen’s “The
Emperor’s New Clothes” provides a case in point. To recall, an emperor
who cares greatly about his appearance and attire hires two tailors who
promise him the finest suit of clothes made from a fabric invisible to
anyone who is inferior or “just hopelessly stupid.” The Emperor cannot
see the cloth himself, but pretends as if he can for fear of appearing unfit
for his exalted position or stupid, and is joined in this pretense by his
ministers, subordinates, and subjects. Notice that the “spiral of silence”
(Noelle-Neumann, 1993) occurs because it is rational to keep quiet given
another’s silence. Standing up to tell the truth is risky given a possibility
(even if it may be small) that the cloth may be visible to another’s eyes.
This situation is also called pluralistic ignorance in social psychology (Katz &
Allport, 1931) that occurs when a majority of group members privately reject
a norm, but assume (incorrectly) that most others accept it; no one believes,
but everyone thinks that everyone else believes. If such a perception holds
for everybody simultaneously, this constitutes an equilibrium where one’s
unilateral deviations (seem to) work against oneself. A bank run that is
triggered initially by some groundless rumor provides a similar example,
where the (ungrounded) prophecy of bankruptcy can be self-fulfilling
through a positive feedback loop (Merton, 1968).
Information cascades are another example of rational conformity. An
information cascade occurs when it is optimal for an individual, who has
observed the “consensus” prior actions of others, to follow the predecessors’
actions regardless of the private information known to that individual.
Some forms of herding behaviors in financial markets, legal decision making
(Farnsworth, 2007), and other collective endeavors can be understood as
manifestions of cascades. Sushil Bikhchandani and others illustrated this
process with an example of a paper submission to an academic journal
(Bikhchandani, Hirshleifer, & Welch, 1992). A referee in a first journal
reads the submitted paper, assesses its quality, and makes a decision about
whether to accept or reject it. Now suppose that a referee at a second journal

Herd Behavior

7

learns that the paper was rejected by the first journal. Assuming that the
referee cannot evaluate the paper’s quality perfectly, knowledge of the
previous rejection should (rationally) make the referee lean toward rejection.
If the paper is rejected at the second journal, this process can continue
at other journals, yielding a chain of rejections. Economists proposed a
model that showed that, at some stage in a sequential-choice task, a rational
decision maker should ignore his or her private information and act only
on the public information obtained from previous decisions. Once this
stage is reached, all decision makers thereafter in the sequence should do
the same, producing an information cascade. And if the earlier decisions in
the sequence happen to be erroneous (e.g., rejecting a high-quality paper),
the cascade leads to undesirable outcomes (Anderson & Holt, 2008; Banerjee,
1992).
HERDING AND THE WISDOM OF CROWDS
The mechanisms reviewed, ranging from unconscious, automatic mimicry to
reasoned, deliberate conformity to rational herding, are fundamental building blocks of collective phenomena. The robustness of these mechanisms,
raises a question about nature of herding: Is herding always problematic, as
is implied by some popular images (e.g., mass hysterias, mobs, panics, fads,
economic bubbles, and groupthink)? What about “the wisdom of crowds,” a
contrasting image of collective behavior, popularized by James Surowiecki?
While herding in humans often refers to defective social processes that
degrade toward suboptimal performance, the “wisdom of crowds” implies
intelligent group processes that can have collective benefits (Surowiecki,
2004). How can we reconcile the two contrasting images of collective action?
GROUP DECISION MAKING BY HONEYBEES
It seems instructive to extend our scope to include herd behavior by nonhuman animals that also live collective lives. Although humans are a gregarious
species, we are arguably not the most gregarious species of all. Our rivals
in this respect are eusocial animals, including, for example, bees, ants, termites, and naked mole rats. Eusocial species are colonial animals that live in
multigenerational genetically related groups, in which the vast majority of
individuals cooperate to aid a relatively few reproductive group members.
They often exhibit extreme task specialization, which makes colonies efficient
in gathering resources.
The puzzle of these species is how they can achieve such high efficiencies
collectively, despite the fact that they have relatively much smaller brains as
compared to humans. More specifically, how do they avoid defective social

8

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

processes leading to problematic herd behavior? We examine group decision
making by honeybees to address these questions.
In late spring or early summer, as a large hive outgrows its nest, a colony
of honey bees often divides itself. The queen leaves with about two-third of
the worker bees to create a new colony, and a daughter queen stays in the old
nest with the rest of the worker bees. The swarm leaving the colony must find
a new home in a short time, which is essential to their survival. The moving
swarm, which is composed of 10,000 or so bees, clusters on a tree branch,
while several hundred scout bees search the neighborhood for a new home.
These scout bees fly out to inspect potential nest sites, and upon returning
to the swarm, perform waggle dances to advertise any good sites they have
discovered. The duration of the dance depends on a bee’s perception of the
site’s quality. Other scout bees that have not flown out yet, as well as those
that have stopped dancing, observe these dances and decide where to visit.
In these decisions, the bees are likely to visit and inspect the sites that have
been advertised strongly by many predecessors. This process constitutes a
positive feedback loop. Thomas Seeley and others, who conducted a series of
experiments with honeybees in natural settings, found that the bees usually
choose the best nest site. Even though none of the bees visit all the potential
nest sites individually, they can aggregate partial individual information to
form a collective wisdom that enables optimal decisions (Seeley, 2010).
Although the bees’ performance is impressive, the puzzle still remains.
How do the bees solve the problem of interdependency? As we have seen,
the bees communicate their findings via waggle dances that are performed
sequentially by scout bees. This could create statistical dependencies among
decision makers, in which initial errors committed by earlier scouts can carry
over and be amplified in the sequence. In this sense, the honeybee group
decision-making system may be susceptible to the erroneous information
cascade (Kameda, Wisdom, Toyokawa, & Inukai, 2012).
A recent paper has addressed this question theoretically with a computer
simulation model (List, Elsholtz, & Seeley, 2009). In line with the previous
empirical observations, the model assumes that scout bees are dependent
on other bees in that they give more attention to nest sites strongly advertised by their predecessors. The bees essentially conform to a majority view
in their decisions about where to visit. However, simultaneously, the model
assumes that the bees are independent in assessing the quality of the visited site. The duration of the scout’s dance, which indexes the strength of the
bee’s preference for the site, is not affected by others’ waggle dances, but is
determined solely by the scout’s own perception of the site’s quality. The computer simulation results showed that, when a suitable mixture of conformity
and independence exists, the honeybee group decision-making process works

Herd Behavior

9

well. Of course, this particular mix of conformity and independence solves
the rational information cascade problem.
COLLECTIVE WISDOM ON THE INTERNET?
Honeybee nest search provides an impressive example of how animals that
have only limited cognitive capacity as individuals can make “wise” decisions collectively as a swarm. It is also important to note that the “swarm
intelligence” (Krause, Ruxton, & Krause, 2009) in honeybees emerges not
from some purposeful coordination by a central authority (e.g., the queen)
but through local interactions among the bees—a key element in the definition of herding, as discussed earlier in this essay. Interestingly, the honeybee
nest-search situation seems to have counterparts in modern human societies,
where individuals can use public information as well as private information
to make a well-informed decision. Examples include information search on
the Internet when buying books or music, choosing a restaurant for dinner,
deciding which hotel to stay at, and so on. Potential options are quite large
in number, yet our time budget for private information search is limited. In
these occasions, we often visit relevant websites (e.g., Amazon, Yelp) to see
how others have decided. Do these social information-pooling systems on the
Internet, in which individuals informed by predecessors’ experiences report
their own new experiences to share with others, yield collective wisdom as
in the honeybee case?
A recent experiment on a “cultural market” by Matthew Salganik and others examines this question (Salganik, Dodds, & Watts, 2006). In cultural markets, sales volumes of hit songs, books, and movies are many times greater
than sales for a typical product. This might imply that the hits are qualitatively different from “the rest,” yet experts can rarely predict which cultural
products will succeed. Why is predicting hits so difficult?
Intrigued by the unpredictability of cultural markets, these researchers
created an experimental music market, where a total of 14,341 participants
downloaded previously unknown songs under one of two conditions—the
“social influence” condition or the “independent” condition. In both conditions, participants could listen to any song they were interested in to
have a direct experience of the product. On top of the individual learning
opportunity, participants in the “social influence” condition were provided
information about how many times each song had been downloaded by
other participants. Notice that there is a structural similarity between the
social influence condition and the honeybee nest search situation. In both
situations, agents had to make choices between unfamiliar options that
could differ in quality. Also, when making individual decisions, social

10

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

frequency information (predecessors’ behaviors) was available, in addition
to the opportunities for individual information search.
The experiment revealed that inequality in overall download counts
among songs was much greater in the social influence condition, as compared to the independent condition in which participants could not access
the social-frequency information. Obviously, participants in the social influence condition copied predecessors’ choices, which produced a “rich get
richer” outcome. Thus, the experiment replicated the robust phenomenon
in cultural markets that hit songs are many times more successful than
average.
Furthermore, the most popular songs (with the highest download frequencies) in the independent condition did not necessarily match the most popular songs in the social influence condition. Mapping of the songs in terms
of popularity ranking between the two conditions was moderate—the most
popular songs in the independent condition never did badly in the social
influence condition, and the least popular songs never did extremely well
either. However, almost any other result could happen. The success of a song
in the social influence condition was path-dependent and susceptible to random fluctuations. This may explain why it is difficult for even experts to
predict which products will succeed in cultural markets.
Overall, how did the human performance in the experimental music
market compare to the honeybee performance in nest search? A tentative
answer does not seem to be flattering to humans. Honeybees mix dependence and independence in nest search. They conform to predecessors
to decide which sites to visit, but assess the quality of the visited sites
independently from predecessors’ evaluations. This leads to the typical
swarm’s high performance. On the other hand, human participants in the
experimental music market seemed to fail to separate the two aspects and
relied too much on others’ choices. Of course, the inherent subjectivity of
music preferences means that the quality of experimental cultural market
outcomes cannot be assessed objectively (as can the nest choice decisions).
Yet, the lack of correspondence in song popularity between the independent
and the social influence conditions suggests that such subjective preferences
are unstable and nonoptimal. In this sense, the hypersusceptibility of
mass behavior to social influence is problematic not only for marketers
of cultural products but also in other sociopolitical domains where no
demonstrably correct answer exists (see Hastie & Kameda, 2005; Kameda,
Tsukasaki, Hastie, & Berg, 2011; Toyokawa, Kim, & Kameda, 2014, for assessments of the “wisdom of crowds” when decision quality can be assessed
objectively).

Herd Behavior

11

CONCLUSION
In this essay, we have reviewed various manifestations of herding in
humans. As we have seen, humans are a highly socially receptive species,
as compared to other gregarious animals. Accumulating evidence from
various behavioral science disciplines strongly suggests that we humans
are equipped with neural, psychological, and behavioral mechanisms that
support this receptiveness—our abilities to learn from and be influenced
by others. It is no doubt that these capacities are mechanisms, selected by
evolution, that have served our survival and contributed to our adaptive
success on the earth. Yet, these adaptive tools can cause serious errors in
modern environments, in which interconnectivities of individuals are much
denser and externalities accruing from individual behaviors are greater
and more far-reaching, as compared to ancient environments in which the
human mind evolved.
Interestingly, growing evidence in the behavioral sciences also suggests
that the two contrasting collective phenomena in humans, maladaptive herding and the wisdom of crowds, are both produced by similar basic mechanisms (Kameda et al., 2011, 2012). In this sense, the two apparently opposite
macro phenomena may be seen as twins produced and governed by our basic
human social receptivity. Given this commonality, understanding the neural,
psychological, and behavioral mechanisms that could help distinguish these
twins will be one of the most important challenges for behavioral sciences in
the next decade.

REFERENCES
Akerlof, G. A., & Shiller, R. J. (2009). Animal spirits: How human psychology drives the
economy, and why it matters for global capitalism. Princeton, NJ: Princeton University
Press.
Anderson, L. R., & Holt, C. A. (2008). Information cascade experiments. In C. R.
Plott & V. L. Smith (Eds.), The handbook of experimental economics results (Vol. 1,
pp. 335–343). Amsterdam, The Netherlands: North Holland.
Asch, S. E. (1956). Studies of independence and conformity: A minority of one against
a unanimous majority. Psychological Monographs, 70(416).
Banerjee, A. V. (1992). A simple model of herd behavior. Quarterly Journal of Economics, 107, 797–818.
Bikhchandani, S., Hirshleifer, D., & Welch, I. (1992). A theory of fads, fashion, custom,
and cultural change as informational cascades. Journal of Political Economy, 100,
992–1026.
Christakis, N. A., & Fowler, J. H. (2007). The spread of obesity in a large social network over 32 years. New England Journal of Medicine, 357, 370–379.

12

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Fowler, J. H., & Christakis, N. A. (2008). Dynamic spread of happiness in a large social
network: Longitudinal analysis over 20 years in the Framingham Heart Study.
British Medical Journal, 337(a2338), 1–9.
Farnsworth, W. (2007). The legal analyst: A toolkit for thinking about the law. Chicago,
IL: University of Chicago Press.
Glaeser, E. L., Bruce Sacerdote, B., & Scheinkman, J. A. (1996). Crime and social interactions. Quarterly Journal of Economics, 111, 507–548.
Granovetter, M. (1978). Threshold models of collective behavior. American Journal of
Sociology, 83, 1420–1443.
Hastie, R., & Kameda, T. (2005). The robust beauty of majority rules in group decisions. Psychological Review, 112, 494–508.
Hatfield, E., Cacioppo, J. T., & Rapson, R. L. (1994). Emotional contagion. New York,
NY: Cambridge University Press.
Hess, U., & Blairy, S. (2001). Facial mimicry and emotional contagion to dynamic
emotional facial expressions and their influence on decoding accuracy. International Journal of Psychophysiology, 40, 129–141.
Kameda, T., & Tindale, R. S. (2006). Groups as adaptive devices: Human docility
and group aggregation mechanisms in evolutionary context. In M. Schaller, J. A.
Simpson & D. T. Kenrick (Eds.), Evolution and social psychology (pp. 317–341). New
York, NY: Psychology Press.
Kameda, T., Tsukasaki, T., Hastie, R., & Berg, N. (2011). Democracy under uncertainty: The wisdom of crowds and the free-rider problem in group decision making. Psychological Review, 118, 76–96.
Kameda, T., Wisdom, T., Toyokawa, W., & Inukai, K. (2012). Is consensus-seeking
unique to humans? A selective review of animal group decision-making and its
implications for (human) social psychology. Group Processes and Intergroup Relations, 15, 673–689.
Katz, D., & Allport, F. H. (1931). Students’ attitudes: A report of the Syracuse University
reaction study. Syracuse, NY: Craftsman Press.
Krause, J., Ruxton, G. D., & Krause, S. (2009). Swarm intelligence in animals and
humans. Trends in Ecology and Evolution, 25, 28–34.
List, C., Elsholtz, C., & Seeley, T. D. (2009). Independence and interdependence in
collective decision making: an agent-based model. Philosophical Transactions of the
Royal Society B, 364, 755–762.
Lyons, R. (2011). The spread of evidence-poor medicine via flawed social-network
analysis. Statistics, Politics, and Policy, 2(1), Article 2.
Meltzoff, A. N., & Moore, K. (1977). Imitation of facial and manual gestures by human
neonates. Science, 198, 75–78.
Merton, R. K. (1968). Social theory and social structure. New York, NY: Free Press.
Noelle-Neumann, E. (1993). The spiral of silence: Public opinion—our social skin.
Chicago, IL: University of Chicago Press.
Raafat, R. M., Chater, N., & Frith, C. (2009). Herding in Humans’ (2009). Trends in
Cognitive Sciences, 13, 420–428.
Rizzolatti, G., & Craighero, L. (2004). The mirror-neuron system. Annual Review of
Neuroscience, 27, 169–192.

Herd Behavior

13

Salganik, M. J., Dodds, P. S., & Watts, D. (2006). Experimental study of inequality and
cultural market. Science, 311, 854–856.
Seeley, T. D. (2010). Honeybee democracy. Princeton, NJ: Princeton University Press.
Shank, R. C., & Abelson, R. P. (1977). Scripts, plans, goals and understanding. New York,
NY: John Wiley & Sons, Inc.
Sherif, M. (1936). The psychology of social norms. New York, NY: Harper Collins.
Simon, H. A. (1990). A mechanism for social selection and successful altruism. Science, 21, 1665–1668.
Smelser, N. J. (1963). Theory of collective behavior. Glencoe, IL: Free Press.
Surowiecki, J. (2004). The wisdom of crowds: Why the many are smarter than the few and
how collective wisdom shapes business, economies, societies and nations. New York, NY:
Doubleday.
Tomasello, M. (1999). The cultural origins of human cognition. Cambridge, MA: Harvard University Press.
Toyokawa, W., Kim, H., & Kameda, T. (2014). Human collective intelligence under
dual exploration-exploitation dilemmas. PLoS One, 9(4), e95789.
Turner, R. H., & Killian, L. M. (1993). Collective behavior (4th ed.). Englewood Cliffs,
NJ: Prentice-Hall.

TATSUYA KAMEDA SHORT BIOGRAPHY
Tatsuya Kameda is currently a professor of Social Psychology at the University of Tokyo after spending 20 years as a professor of behavioral science at
Hokkaido University. He led the Global COE (Centers of Excellence) project
funded by the Japanese Government (2007–2012), as a director of the Center for Experimental Research in Social Sciences at Hokkaido University. He
studies social behavior from the adaptationist perspective, through combining evolutionary games and agent-based simulations with behavioral, cognitive, and fMRI experiments (http://lynx.let.hokudai.ac.jp/∼kameda/). He
was a Fulbright Research Fellow at the University of Colorado at Boulder
(1997–1998), a Deutscher Akademischer Austausch Dienst Research Fellow
at Max Planck Institute for Human Development in Berlin (2001), and a Residential Fellow at the Center for Advanced Study in the Behavioral Sciences
at Stanford University (2008–2009).
REID HASTIE SHORT BIOGRAPHY
Reid Hastie is the Ralph and Dorothy Keller Distinguished Service Professor of Behavioral Science at the University of Chicago. His primary research
interests are concern individual and group judgment and decision making.
He is the author of Rational Choice in an Uncertain World (with Robyn Dawes)
and Wiser: Getting beyond groupthink to make groups smarter (with Cass Sunstein).

14

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

RELATED ESSAYS
Culture and Cognition (Sociology), Karen A. Cerulo
Peers and Adolescent Risk Taking (Psychology), Jason Chein
Adaptation for Culture (Anthropology), Jill M. Church
The Inherence Heuristic: Generating Everyday Explanations (Psychology),
Andrei Cimpian
Misinformation and How to Correct It (Psychology), John Cook et al.
Problems Attract Problems: A Network Perspective on Mental Disorders
(Psychology), Angélique Cramer and Denny Borsboom
Household Wealth Effects and the US Macroeconomy (Economics), Karen
Dynan and Daniel Cooper
Micro-Cultures (Sociology), Gary Alan Fine
State of the Art in Competition Research (Psychology), Márta Fülöp and
Gábor Orosz
Economics and Culture (Economics), Gérard Roland
Emotion and Decision Making (Psychology), Jeff R. Huntsinger and Cara Ray
Genetic Foundations of Attitude Formation (Political Science), Christian
Kandler et al.
Search and Learning in Markets (Economics), Philipp Kircher
Cultural Neuroscience: Connecting Culture, Brain, and Genes (Psychology),
Shinobu Kitayama and Sarah Huff
Niche Construction: Implications for Human Sciences (Anthropology), Kevin
N. Laland and Michael O’Brien
From Individual Rationality to Socially Embedded Self-Regulation (Sociology), Siegwart Lindenberg
Two-Systems View of Children’s Theory-of-Mind Understanding (Psychology), Jason Low
Implicit Attitude Measures (Psychology), Gregory Mitchell and Philip E.
Tetlock
A Bio-Social-Cultural Approach to Early Cognitive Development: Entering
the Community of Minds (Psychology), Katherine Nelson
Organizations and the Production of Systemic Risk (Sociology), Charles
Perrow
The Sociology of Religious Experience (Sociology), Douglas Porpora
War and Social Movements (Political Science), Sidney Tarrow
The Social Science of Sustainability (Political Science), Johannes Urpelainen
Theory of Mind (Psychology), Henry Wellman



Herd Behavior
TATSUYA KAMEDA and REID HASTIE

Abstract
There are many manifestations of herding in the human species—one of the most
socially interdependent species on the earth. Herding here refers to an alignment of
thoughts or behaviors of individuals in a group through local interactions among
individuals rather through than some purposeful coordination by a central authority in the group. Herding underlies many collective phenomena in the Internet era,
ranging from everyday social behavior, consumer choices, economic bubbles, and
political movements. Accumulating evidence in various behavioral science disciplines suggests that we humans are equipped with neural, psychological, and behavioral mechanisms that constitute our highly socially sensitive minds. These built-in
mechanisms are evolutionary products that have promoted our survival. Yet, these
adaptive tools can cause serious errors in modern environments, in which interconnectivities of individuals are much denser and externalities accruing from individual
behaviors are much greater and more far-reaching, compared to primordial environments in which the human mind evolved. Growing evidence in the behavioral
sciences also suggests that the two contrasting collective phenomena in humans,
maladaptive herding and the wisdom of crowds, are based on similar underlying
mechanisms. In this sense, the two apparently opposite macro phenomena may be
seen as twins produced and governed by the social receptivity of our minds. Given
this commonality, understanding the neural, psychological, and behavioral mechanisms that could distinguish these twins will be one of the most important challenges
for behavioral sciences in the next decade.

In an increasingly connected world, an event in one place, be it economic,
political, or social, can cause large-scale chain reactions across many other
places. We have abundant examples of this sort, including the recent
global financial crisis, the spread of civil uprisings in the Middle East, the
widespread adoption of technological innovations such as the iPad, and so
on. Until recently, such mass phenomena have been studied sporadically
across social science disciplines without much mutual communication. Yet,
with advances in technology and new theoretical frameworks, these mass
phenomena are becoming a focus of substantial interdisciplinary interests
(Akerlof & Shiller, 2009). An umbrella concept, “herding,” has facilitated
such cross-disciplinary communication over the past 5 years.
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

HERD BEHAVIOR: A DEFINITION AND EXAMPLES
What is meant by herding? Herding refers to an alignment of thoughts or
behaviors of individuals in a group. Most importantly, such convergence
often emerges through local interactions among agents rather than through
purposeful coordination by a central authority or a leading figure in the
group. In other words, the apparent coordination of the herd is an emergent
property of local interactions (Raafat, Chater, & Frith, 2009).
Textbook examples of herding in the social science literature include riots,
panics, fads, mass hysterias, urban legends, economic bubbles, and so on
(Smelser, 1963; Turner & Killian, 1993). However, besides these familiar
examples, recent research suggests that herding may encompass a much
wider range of our social behaviors than had been previously thought.
CRIMES
Proliferation of crimes in a city may be seen as an example of herding. One
of the most striking aspects of crime is that crime rates vary dramatically
across time and space. For example, homicide rates across nations ranged
from 6.1 cases per million in Japan, 12.6 in Sweden, to 98.0 in the United
States in 1990. Within the United States, rates of serious crimes in the year
ranged from 0.008 per capita in Ridgewood Village, New Jersey to 0.384 in
nearby Atlantic City (Glaeser, Bruce Sacerdote, & Scheinkman, 1996). Such
high variances are observed within cities as well, where one street can have
much higher crime rates than streets just a few blocks away.
One obvious explanation for such variety may be that socioeconomic conditions also vary over time and space, creating temporal and geographical
clusters of crime. However, an econometric analysis by Edward Glaeser and
others showed that less than 30% of the variation in cross-city or cross-district
crime rates could be explained by the local socioeconomic differences. These
researchers developed a model in which agents’ decisions about crime were a
function of their own attributes (e.g., socioeconomic as well as psychological
attributes) and of their neighbor’s decisions about criminal activities. Glaeser
and others then estimated impacts of the second element of the model (i.e.,
social influence from neighbors) for a variety of crimes in the United States
in 1985, in 1970, and across New York City in 1985. The results showed that a
positive interaction among agents’ decisions about crime was the only viable
explanation for the large residual variance not explained by the local socioeconomic conditions. More specifically, the local social influence was strong
for larceny and auto theft; moderate for assault, burglary, and robbery; and
weak for arson, murder, and rape. These results suggest that one agent’s decision to commit crimes (especially minor crimes) affects his or her neighbor’s
decisions, which constitutes a positive feedback loop for the collective. The

Herd Behavior

3

large variations in crime rates across time and space seem to emerge as aggregated outcomes of such individual local decisions.
OBESITY
Recent research suggests that obesity may be contagious as well. Using a data
set from a longitudinal survey on cardiovascular disease (the Framingham
Heart Study; see http://www.framinghamheartstudy.org/), Nicholas Christakis and James Fowler (2007) examined how social relations in a community
affect obesity. The original survey traced health states of people residing in
Framingham, Massachusetts, over 32 years. Christakis and Fowler focused
on family and friendship relations among the participants, and applied longitudinal statistical models to examine whether weight gain in one person
was associated with weight gains in his or her friends, siblings, spouse, and
neighbors.
Results of the network analysis revealed that obese people (defined as those
with a body mass index >30) and nonobese people formed different clusters
and that social influences through the network extended up to three degrees
of separation. In other words, the average obese person was more likely to
have obese friends, friends of friends, and friends of friends of friends than
was the average nonobese person. Moreover, a person’s chances of becoming obese increased by 57% if he or she had a friend who became obese in the
time period, by 40% if a sibling became obese, and by 37% if a spouse became
obese. These patterns suggest that obesity spreads through social network
similar to a pathogen. A person’s overeating behavior is affected through the
social network, even if one may not know another overeater directly. Segmentations of obese and nonobese people in a community seem to emerge
as aggregated consequences of local influences (see also Lyons, 2011, for criticisms of the social network analysis employed by Christakis and Fowler).
HAPPINESS
Happy people and unhappy people also seem to inhabit different clusters in a
community. A reanalysis of the Framingham Heart Study data set suggested
that these clusters did not simply reflect a tendency for individuals to associate with similar individuals. Instead, these macro patterns resulted from
spread of happiness and unhappiness through the social network, just as in
the case of obesity. According to the analysis, the probability that one was
happy increased by 25% if a friend who lived within a mile became happy,
and these local influences also extended up to three degrees of separation.
Thus, similar to obesity, happiness also seems to be contagious (Fowler &
Christakis, 2008).

4

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

WHY DOES HERDING OCCUR?—POTENTIAL MECHANISMS
The earlier examples suggest that herding is a robust phenomenon, characterizing a wide range of social behaviors in our life. If so, what are the neural,
psychological, or sociological mechanisms that produce herding?
EMOTIONAL CONTAGION, FACIAL MIMICRY, AND MIRROR NEURONS
As implied by the saying that your smile makes others happy, humans
often reproduce others’ emotions in themselves. This phenomenon, which
is called emotional contagion (Hatfield, Cacioppo, & Rapson, 1994), has
long been known among psychotherapists who treat depressed clients.
Therapists, especially those who are inexperienced, sometimes “catch”
their clients’ emotions expressed during interviews, and feel themselves
depressed afterwards. Elaine Hatfield and her colleagues see emotional
contagion as a primitive, automatic, and unconscious process. It occurs
through a series of steps: when a receiver is interacting with a sender, he or
she first perceives the emotional expressions of the sender. The receiver then
automatically transfers the perceived emotional expressions to his or her
bodily expressions (e.g., facial expressions, postures). Through the process
of afferent feedback, these mimicked bodily expressions are translated into
the receiver feeling the same emotion that the sender experienced, which
leads to emotional convergence among the sender and the receiver.
Indeed, it can easily be demonstrated that we have a tendency to mimic
the facial expressions of others in everyday social interactions. Research suggests that such facial mimicry is an automatic, reflex-like process, in which
the observer’s facial expression matches the observed facial expression (e.g.,
happy, sad, fearful, angry, disgusted faces) rather quickly—typically within
less than a second (Hess & Blairy, 2001). Such automatic mimicry extends to
bodily posture, voice pitch, and so on, and is known to emerge very early in
human development. Even 12- to 21-day-old infants imitate both facial and
manual gestures displayed by an adult model (Meltzoff & Moore, 1977).
Furthermore, recent developments in neuroscience suggest that there may
be a biological system in our brains that helps us to mirror others’ actions.
One of the most intriguing recent findings in brain science is the discovery of
“mirror neurons.” In the late 1980s when Giacomo Rizzolatti and others were
recording electrical activity in the brain of a macaque, these researchers found
neurons that fired both when the animal acted and when the animal observed
the same action performed by another. The same neurons fired when the
monkey grasped something with its hand, and when the monkey observed
the experimenter grasping it. However, these neurons did not discharge in
response to simple presentation of food or other interesting objects. The neuron “mirrored” the motor behavior of the other, as though the observer were

Herd Behavior

5

itself executing the motor act. Although it remains controversial, some recent
data suggest that a similar “mirror neuron system” exists in human brains as
well (Rizzolatti & Craighero, 2004).
Taken together, these psychological, behavioral, and neural findings
strongly suggest that mimicking others may be a fundamentally human
activity.
SOCIAL NORMS, MUTUAL EXPECTATIONS, AND SHARED STORIES
Another mechanism for herding involves more conscious, deliberate, and
controlled psychological processes, that are distinguishable from our automatic “aping” propensities as reviewed earlier. These processes have been
studied mainly by social psychologists.
Classic experimental demonstrations of such herd behaviors in social psychology include the famous line-comparison perception study by Solomon
Asch, where subjects conformed to an erroneous majority view to avoid
potential embarrassment or other social consequences in a group (Asch,
1956); the optical judgment study by Muzfer Sherif demonstrating that
individual perceptions of the autokinetic illusion converged to a shared
social reality through communication (i.e., everybody in the same group
ended up experiencing a similar optical illusion Sherif, 1936), and so on.
A key element underlying these herd behaviors is a fundamental characteristic of our mind, which may be labeled docility or receptivity to social norms
(Kameda & Tindale, 2006; Simon, 1990). Herbert Simon defined this concept
as our tendency to depend on others’ suggestions, recommendation, persuasion, and information obtained through social channels as a major basis of
choice. Compared to other gregarious species, humans are unique in developing social norms and mutually shared expectations, which inform us about
what action is normal, appropriate, or fair in a given social situation. As seen
in the Asch experiment, the human mind is built to be receptive to social
norms, and tends to self-censor actions in order to avoid violating norms.
Notice that the high receptivity to social norms is also fundamental to our
ability to learn culturally. Humans are a cultural species that can take full
advantage of socially acquired knowledge. Without docility by learners to
their “cultural parents,” such cognitive capacities would be highly limited
(Tomasello, 1999).
The human mind is also built to think in terms of, and be influenced by,
narratives or stories (sequences of events with an internal logic and dynamics: Shank & Abelson, 1977). Stories, especially stories shared in a community
or across a whole society, lead us to see, interpret, have feelings about, and
react to experiences from a shared perspective (see Akerlof & Shiller, 2009,
for interesting recent examples of influential political-economic stories).

6

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

This characteristic social receptivity of individual minds can yield effects
that are visible at the societal level. In the aforementioned case of contagious
obesity, for example, one may decide to eat more because the action seems to
be normal given one’s spouse’s or friend’s eating practices, which in turn provides a normative signal for another’s overeating. Our actions have spillover
effects (which economists call externalities) on others, which can lead to spiraling proliferations of action across a whole society (Granovetter, 1978).
RATIONAL CONFORMITY AND INFORMATION CASCADES
Sometimes it is rational to conform to a majority behavior in a group, even
if one would otherwise choose differently. Hans Christian Andersen’s “The
Emperor’s New Clothes” provides a case in point. To recall, an emperor
who cares greatly about his appearance and attire hires two tailors who
promise him the finest suit of clothes made from a fabric invisible to
anyone who is inferior or “just hopelessly stupid.” The Emperor cannot
see the cloth himself, but pretends as if he can for fear of appearing unfit
for his exalted position or stupid, and is joined in this pretense by his
ministers, subordinates, and subjects. Notice that the “spiral of silence”
(Noelle-Neumann, 1993) occurs because it is rational to keep quiet given
another’s silence. Standing up to tell the truth is risky given a possibility
(even if it may be small) that the cloth may be visible to another’s eyes.
This situation is also called pluralistic ignorance in social psychology (Katz &
Allport, 1931) that occurs when a majority of group members privately reject
a norm, but assume (incorrectly) that most others accept it; no one believes,
but everyone thinks that everyone else believes. If such a perception holds
for everybody simultaneously, this constitutes an equilibrium where one’s
unilateral deviations (seem to) work against oneself. A bank run that is
triggered initially by some groundless rumor provides a similar example,
where the (ungrounded) prophecy of bankruptcy can be self-fulfilling
through a positive feedback loop (Merton, 1968).
Information cascades are another example of rational conformity. An
information cascade occurs when it is optimal for an individual, who has
observed the “consensus” prior actions of others, to follow the predecessors’
actions regardless of the private information known to that individual.
Some forms of herding behaviors in financial markets, legal decision making
(Farnsworth, 2007), and other collective endeavors can be understood as
manifestions of cascades. Sushil Bikhchandani and others illustrated this
process with an example of a paper submission to an academic journal
(Bikhchandani, Hirshleifer, & Welch, 1992). A referee in a first journal
reads the submitted paper, assesses its quality, and makes a decision about
whether to accept or reject it. Now suppose that a referee at a second journal

Herd Behavior

7

learns that the paper was rejected by the first journal. Assuming that the
referee cannot evaluate the paper’s quality perfectly, knowledge of the
previous rejection should (rationally) make the referee lean toward rejection.
If the paper is rejected at the second journal, this process can continue
at other journals, yielding a chain of rejections. Economists proposed a
model that showed that, at some stage in a sequential-choice task, a rational
decision maker should ignore his or her private information and act only
on the public information obtained from previous decisions. Once this
stage is reached, all decision makers thereafter in the sequence should do
the same, producing an information cascade. And if the earlier decisions in
the sequence happen to be erroneous (e.g., rejecting a high-quality paper),
the cascade leads to undesirable outcomes (Anderson & Holt, 2008; Banerjee,
1992).
HERDING AND THE WISDOM OF CROWDS
The mechanisms reviewed, ranging from unconscious, automatic mimicry to
reasoned, deliberate conformity to rational herding, are fundamental building blocks of collective phenomena. The robustness of these mechanisms,
raises a question about nature of herding: Is herding always problematic, as
is implied by some popular images (e.g., mass hysterias, mobs, panics, fads,
economic bubbles, and groupthink)? What about “the wisdom of crowds,” a
contrasting image of collective behavior, popularized by James Surowiecki?
While herding in humans often refers to defective social processes that
degrade toward suboptimal performance, the “wisdom of crowds” implies
intelligent group processes that can have collective benefits (Surowiecki,
2004). How can we reconcile the two contrasting images of collective action?
GROUP DECISION MAKING BY HONEYBEES
It seems instructive to extend our scope to include herd behavior by nonhuman animals that also live collective lives. Although humans are a gregarious
species, we are arguably not the most gregarious species of all. Our rivals
in this respect are eusocial animals, including, for example, bees, ants, termites, and naked mole rats. Eusocial species are colonial animals that live in
multigenerational genetically related groups, in which the vast majority of
individuals cooperate to aid a relatively few reproductive group members.
They often exhibit extreme task specialization, which makes colonies efficient
in gathering resources.
The puzzle of these species is how they can achieve such high efficiencies
collectively, despite the fact that they have relatively much smaller brains as
compared to humans. More specifically, how do they avoid defective social

8

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

processes leading to problematic herd behavior? We examine group decision
making by honeybees to address these questions.
In late spring or early summer, as a large hive outgrows its nest, a colony
of honey bees often divides itself. The queen leaves with about two-third of
the worker bees to create a new colony, and a daughter queen stays in the old
nest with the rest of the worker bees. The swarm leaving the colony must find
a new home in a short time, which is essential to their survival. The moving
swarm, which is composed of 10,000 or so bees, clusters on a tree branch,
while several hundred scout bees search the neighborhood for a new home.
These scout bees fly out to inspect potential nest sites, and upon returning
to the swarm, perform waggle dances to advertise any good sites they have
discovered. The duration of the dance depends on a bee’s perception of the
site’s quality. Other scout bees that have not flown out yet, as well as those
that have stopped dancing, observe these dances and decide where to visit.
In these decisions, the bees are likely to visit and inspect the sites that have
been advertised strongly by many predecessors. This process constitutes a
positive feedback loop. Thomas Seeley and others, who conducted a series of
experiments with honeybees in natural settings, found that the bees usually
choose the best nest site. Even though none of the bees visit all the potential
nest sites individually, they can aggregate partial individual information to
form a collective wisdom that enables optimal decisions (Seeley, 2010).
Although the bees’ performance is impressive, the puzzle still remains.
How do the bees solve the problem of interdependency? As we have seen,
the bees communicate their findings via waggle dances that are performed
sequentially by scout bees. This could create statistical dependencies among
decision makers, in which initial errors committed by earlier scouts can carry
over and be amplified in the sequence. In this sense, the honeybee group
decision-making system may be susceptible to the erroneous information
cascade (Kameda, Wisdom, Toyokawa, & Inukai, 2012).
A recent paper has addressed this question theoretically with a computer
simulation model (List, Elsholtz, & Seeley, 2009). In line with the previous
empirical observations, the model assumes that scout bees are dependent
on other bees in that they give more attention to nest sites strongly advertised by their predecessors. The bees essentially conform to a majority view
in their decisions about where to visit. However, simultaneously, the model
assumes that the bees are independent in assessing the quality of the visited site. The duration of the scout’s dance, which indexes the strength of the
bee’s preference for the site, is not affected by others’ waggle dances, but is
determined solely by the scout’s own perception of the site’s quality. The computer simulation results showed that, when a suitable mixture of conformity
and independence exists, the honeybee group decision-making process works

Herd Behavior

9

well. Of course, this particular mix of conformity and independence solves
the rational information cascade problem.
COLLECTIVE WISDOM ON THE INTERNET?
Honeybee nest search provides an impressive example of how animals that
have only limited cognitive capacity as individuals can make “wise” decisions collectively as a swarm. It is also important to note that the “swarm
intelligence” (Krause, Ruxton, & Krause, 2009) in honeybees emerges not
from some purposeful coordination by a central authority (e.g., the queen)
but through local interactions among the bees—a key element in the definition of herding, as discussed earlier in this essay. Interestingly, the honeybee
nest-search situation seems to have counterparts in modern human societies,
where individuals can use public information as well as private information
to make a well-informed decision. Examples include information search on
the Internet when buying books or music, choosing a restaurant for dinner,
deciding which hotel to stay at, and so on. Potential options are quite large
in number, yet our time budget for private information search is limited. In
these occasions, we often visit relevant websites (e.g., Amazon, Yelp) to see
how others have decided. Do these social information-pooling systems on the
Internet, in which individuals informed by predecessors’ experiences report
their own new experiences to share with others, yield collective wisdom as
in the honeybee case?
A recent experiment on a “cultural market” by Matthew Salganik and others examines this question (Salganik, Dodds, & Watts, 2006). In cultural markets, sales volumes of hit songs, books, and movies are many times greater
than sales for a typical product. This might imply that the hits are qualitatively different from “the rest,” yet experts can rarely predict which cultural
products will succeed. Why is predicting hits so difficult?
Intrigued by the unpredictability of cultural markets, these researchers
created an experimental music market, where a total of 14,341 participants
downloaded previously unknown songs under one of two conditions—the
“social influence” condition or the “independent” condition. In both conditions, participants could listen to any song they were interested in to
have a direct experience of the product. On top of the individual learning
opportunity, participants in the “social influence” condition were provided
information about how many times each song had been downloaded by
other participants. Notice that there is a structural similarity between the
social influence condition and the honeybee nest search situation. In both
situations, agents had to make choices between unfamiliar options that
could differ in quality. Also, when making individual decisions, social

10

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

frequency information (predecessors’ behaviors) was available, in addition
to the opportunities for individual information search.
The experiment revealed that inequality in overall download counts
among songs was much greater in the social influence condition, as compared to the independent condition in which participants could not access
the social-frequency information. Obviously, participants in the social influence condition copied predecessors’ choices, which produced a “rich get
richer” outcome. Thus, the experiment replicated the robust phenomenon
in cultural markets that hit songs are many times more successful than
average.
Furthermore, the most popular songs (with the highest download frequencies) in the independent condition did not necessarily match the most popular songs in the social influence condition. Mapping of the songs in terms
of popularity ranking between the two conditions was moderate—the most
popular songs in the independent condition never did badly in the social
influence condition, and the least popular songs never did extremely well
either. However, almost any other result could happen. The success of a song
in the social influence condition was path-dependent and susceptible to random fluctuations. This may explain why it is difficult for even experts to
predict which products will succeed in cultural markets.
Overall, how did the human performance in the experimental music
market compare to the honeybee performance in nest search? A tentative
answer does not seem to be flattering to humans. Honeybees mix dependence and independence in nest search. They conform to predecessors
to decide which sites to visit, but assess the quality of the visited sites
independently from predecessors’ evaluations. This leads to the typical
swarm’s high performance. On the other hand, human participants in the
experimental music market seemed to fail to separate the two aspects and
relied too much on others’ choices. Of course, the inherent subjectivity of
music preferences means that the quality of experimental cultural market
outcomes cannot be assessed objectively (as can the nest choice decisions).
Yet, the lack of correspondence in song popularity between the independent
and the social influence conditions suggests that such subjective preferences
are unstable and nonoptimal. In this sense, the hypersusceptibility of
mass behavior to social influence is problematic not only for marketers
of cultural products but also in other sociopolitical domains where no
demonstrably correct answer exists (see Hastie & Kameda, 2005; Kameda,
Tsukasaki, Hastie, & Berg, 2011; Toyokawa, Kim, & Kameda, 2014, for assessments of the “wisdom of crowds” when decision quality can be assessed
objectively).

Herd Behavior

11

CONCLUSION
In this essay, we have reviewed various manifestations of herding in
humans. As we have seen, humans are a highly socially receptive species,
as compared to other gregarious animals. Accumulating evidence from
various behavioral science disciplines strongly suggests that we humans
are equipped with neural, psychological, and behavioral mechanisms that
support this receptiveness—our abilities to learn from and be influenced
by others. It is no doubt that these capacities are mechanisms, selected by
evolution, that have served our survival and contributed to our adaptive
success on the earth. Yet, these adaptive tools can cause serious errors in
modern environments, in which interconnectivities of individuals are much
denser and externalities accruing from individual behaviors are greater
and more far-reaching, as compared to ancient environments in which the
human mind evolved.
Interestingly, growing evidence in the behavioral sciences also suggests
that the two contrasting collective phenomena in humans, maladaptive herding and the wisdom of crowds, are both produced by similar basic mechanisms (Kameda et al., 2011, 2012). In this sense, the two apparently opposite
macro phenomena may be seen as twins produced and governed by our basic
human social receptivity. Given this commonality, understanding the neural,
psychological, and behavioral mechanisms that could help distinguish these
twins will be one of the most important challenges for behavioral sciences in
the next decade.

REFERENCES
Akerlof, G. A., & Shiller, R. J. (2009). Animal spirits: How human psychology drives the
economy, and why it matters for global capitalism. Princeton, NJ: Princeton University
Press.
Anderson, L. R., & Holt, C. A. (2008). Information cascade experiments. In C. R.
Plott & V. L. Smith (Eds.), The handbook of experimental economics results (Vol. 1,
pp. 335–343). Amsterdam, The Netherlands: North Holland.
Asch, S. E. (1956). Studies of independence and conformity: A minority of one against
a unanimous majority. Psychological Monographs, 70(416).
Banerjee, A. V. (1992). A simple model of herd behavior. Quarterly Journal of Economics, 107, 797–818.
Bikhchandani, S., Hirshleifer, D., & Welch, I. (1992). A theory of fads, fashion, custom,
and cultural change as informational cascades. Journal of Political Economy, 100,
992–1026.
Christakis, N. A., & Fowler, J. H. (2007). The spread of obesity in a large social network over 32 years. New England Journal of Medicine, 357, 370–379.

12

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Fowler, J. H., & Christakis, N. A. (2008). Dynamic spread of happiness in a large social
network: Longitudinal analysis over 20 years in the Framingham Heart Study.
British Medical Journal, 337(a2338), 1–9.
Farnsworth, W. (2007). The legal analyst: A toolkit for thinking about the law. Chicago,
IL: University of Chicago Press.
Glaeser, E. L., Bruce Sacerdote, B., & Scheinkman, J. A. (1996). Crime and social interactions. Quarterly Journal of Economics, 111, 507–548.
Granovetter, M. (1978). Threshold models of collective behavior. American Journal of
Sociology, 83, 1420–1443.
Hastie, R., & Kameda, T. (2005). The robust beauty of majority rules in group decisions. Psychological Review, 112, 494–508.
Hatfield, E., Cacioppo, J. T., & Rapson, R. L. (1994). Emotional contagion. New York,
NY: Cambridge University Press.
Hess, U., & Blairy, S. (2001). Facial mimicry and emotional contagion to dynamic
emotional facial expressions and their influence on decoding accuracy. International Journal of Psychophysiology, 40, 129–141.
Kameda, T., & Tindale, R. S. (2006). Groups as adaptive devices: Human docility
and group aggregation mechanisms in evolutionary context. In M. Schaller, J. A.
Simpson & D. T. Kenrick (Eds.), Evolution and social psychology (pp. 317–341). New
York, NY: Psychology Press.
Kameda, T., Tsukasaki, T., Hastie, R., & Berg, N. (2011). Democracy under uncertainty: The wisdom of crowds and the free-rider problem in group decision making. Psychological Review, 118, 76–96.
Kameda, T., Wisdom, T., Toyokawa, W., & Inukai, K. (2012). Is consensus-seeking
unique to humans? A selective review of animal group decision-making and its
implications for (human) social psychology. Group Processes and Intergroup Relations, 15, 673–689.
Katz, D., & Allport, F. H. (1931). Students’ attitudes: A report of the Syracuse University
reaction study. Syracuse, NY: Craftsman Press.
Krause, J., Ruxton, G. D., & Krause, S. (2009). Swarm intelligence in animals and
humans. Trends in Ecology and Evolution, 25, 28–34.
List, C., Elsholtz, C., & Seeley, T. D. (2009). Independence and interdependence in
collective decision making: an agent-based model. Philosophical Transactions of the
Royal Society B, 364, 755–762.
Lyons, R. (2011). The spread of evidence-poor medicine via flawed social-network
analysis. Statistics, Politics, and Policy, 2(1), Article 2.
Meltzoff, A. N., & Moore, K. (1977). Imitation of facial and manual gestures by human
neonates. Science, 198, 75–78.
Merton, R. K. (1968). Social theory and social structure. New York, NY: Free Press.
Noelle-Neumann, E. (1993). The spiral of silence: Public opinion—our social skin.
Chicago, IL: University of Chicago Press.
Raafat, R. M., Chater, N., & Frith, C. (2009). Herding in Humans’ (2009). Trends in
Cognitive Sciences, 13, 420–428.
Rizzolatti, G., & Craighero, L. (2004). The mirror-neuron system. Annual Review of
Neuroscience, 27, 169–192.

Herd Behavior

13

Salganik, M. J., Dodds, P. S., & Watts, D. (2006). Experimental study of inequality and
cultural market. Science, 311, 854–856.
Seeley, T. D. (2010). Honeybee democracy. Princeton, NJ: Princeton University Press.
Shank, R. C., & Abelson, R. P. (1977). Scripts, plans, goals and understanding. New York,
NY: John Wiley & Sons, Inc.
Sherif, M. (1936). The psychology of social norms. New York, NY: Harper Collins.
Simon, H. A. (1990). A mechanism for social selection and successful altruism. Science, 21, 1665–1668.
Smelser, N. J. (1963). Theory of collective behavior. Glencoe, IL: Free Press.
Surowiecki, J. (2004). The wisdom of crowds: Why the many are smarter than the few and
how collective wisdom shapes business, economies, societies and nations. New York, NY:
Doubleday.
Tomasello, M. (1999). The cultural origins of human cognition. Cambridge, MA: Harvard University Press.
Toyokawa, W., Kim, H., & Kameda, T. (2014). Human collective intelligence under
dual exploration-exploitation dilemmas. PLoS One, 9(4), e95789.
Turner, R. H., & Killian, L. M. (1993). Collective behavior (4th ed.). Englewood Cliffs,
NJ: Prentice-Hall.

TATSUYA KAMEDA SHORT BIOGRAPHY
Tatsuya Kameda is currently a professor of Social Psychology at the University of Tokyo after spending 20 years as a professor of behavioral science at
Hokkaido University. He led the Global COE (Centers of Excellence) project
funded by the Japanese Government (2007–2012), as a director of the Center for Experimental Research in Social Sciences at Hokkaido University. He
studies social behavior from the adaptationist perspective, through combining evolutionary games and agent-based simulations with behavioral, cognitive, and fMRI experiments (http://lynx.let.hokudai.ac.jp/∼kameda/). He
was a Fulbright Research Fellow at the University of Colorado at Boulder
(1997–1998), a Deutscher Akademischer Austausch Dienst Research Fellow
at Max Planck Institute for Human Development in Berlin (2001), and a Residential Fellow at the Center for Advanced Study in the Behavioral Sciences
at Stanford University (2008–2009).
REID HASTIE SHORT BIOGRAPHY
Reid Hastie is the Ralph and Dorothy Keller Distinguished Service Professor of Behavioral Science at the University of Chicago. His primary research
interests are concern individual and group judgment and decision making.
He is the author of Rational Choice in an Uncertain World (with Robyn Dawes)
and Wiser: Getting beyond groupthink to make groups smarter (with Cass Sunstein).

14

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

RELATED ESSAYS
Culture and Cognition (Sociology), Karen A. Cerulo
Peers and Adolescent Risk Taking (Psychology), Jason Chein
Adaptation for Culture (Anthropology), Jill M. Church
The Inherence Heuristic: Generating Everyday Explanations (Psychology),
Andrei Cimpian
Misinformation and How to Correct It (Psychology), John Cook et al.
Problems Attract Problems: A Network Perspective on Mental Disorders
(Psychology), Angélique Cramer and Denny Borsboom
Household Wealth Effects and the US Macroeconomy (Economics), Karen
Dynan and Daniel Cooper
Micro-Cultures (Sociology), Gary Alan Fine
State of the Art in Competition Research (Psychology), Márta Fülöp and
Gábor Orosz
Economics and Culture (Economics), Gérard Roland
Emotion and Decision Making (Psychology), Jeff R. Huntsinger and Cara Ray
Genetic Foundations of Attitude Formation (Political Science), Christian
Kandler et al.
Search and Learning in Markets (Economics), Philipp Kircher
Cultural Neuroscience: Connecting Culture, Brain, and Genes (Psychology),
Shinobu Kitayama and Sarah Huff
Niche Construction: Implications for Human Sciences (Anthropology), Kevin
N. Laland and Michael O’Brien
From Individual Rationality to Socially Embedded Self-Regulation (Sociology), Siegwart Lindenberg
Two-Systems View of Children’s Theory-of-Mind Understanding (Psychology), Jason Low
Implicit Attitude Measures (Psychology), Gregory Mitchell and Philip E.
Tetlock
A Bio-Social-Cultural Approach to Early Cognitive Development: Entering
the Community of Minds (Psychology), Katherine Nelson
Organizations and the Production of Systemic Risk (Sociology), Charles
Perrow
The Sociology of Religious Experience (Sociology), Douglas Porpora
War and Social Movements (Political Science), Sidney Tarrow
The Social Science of Sustainability (Political Science), Johannes Urpelainen
Theory of Mind (Psychology), Henry Wellman



Media

Herd Behavior