408 Social Influence and Group Dynamics
Two sources of data provide empirical support for this
perspective on social transition: the development of the pri-
vate sector of the Polish economy and the emergence of vot-
ing preferences in the Polish parliamentary elections during
the transition from socialism to private enterprise in the late
1980s and early 1990s (Nowak, Urbaniak, & Zienkowski,
1994). For a description of these data, as well as a compre-
hensive depiction of the cellular automata model and its im-
plications for societal transition, the reader is referred to
Nowak and Vallacher (2001).
Implications for Cultural Differences
The cellular automata model is useful in understanding and
predicting differences among cultures in the dynamics of
social influence and societal organization. A primary theme
in cross-cultural comparisons centers on collectivism versus
individualism (cf. Markus & Kitayama, 1991). In so-called
collectivist cultures—China and Japan, for example—
interdependence among individuals is stressed at the expense
of personal independence, so that individuals are readily in-
fluenced by the beliefs, attitudes, and expectations of other
people. In so-called individualistic cultures—the United
States, for example—greater emphasis is placed on inde-
pendence, with individuals maintaining a relatively strong
degree of autonomy in their self-concept, attitudes, and
lifestyle. This dimension of cultural variation maps directly
onto the variable of self-influence in the cellular automata
model. In a society that values independence in decision-
making and judgment, the magnitude of self-influence is cor-
respondingly strong and operates at the expense of the
opinions and expectations of others. Computer simulations
have revealed that as self-influence increases in magnitude,
the number of individuals changing their opinion on a given
issue decreases, there is less polarization and clustering, and
the average cluster is smaller in size (Latané & Nowak, 1997;
Lewenstein et al., 1993).
Societies also differ in their relative stability. In less mod-
ernized societies, which are predominantly rural and agrarian
rather than industrial in nature, the social context for individ-
uals is relatively stable over time. In contrast, relatively mod-
ernized and industrial societies tend to be characterized by
greater social mobility (e.g., travel, permanent relocation)
and greater frequency of communication over large distances
(by means of phone, e-mail, and fax). These features disturb
the stability of social influence exerted by the social context
on the individual. At different times, in other words, the indi-
vidual is exposed to a broad range of opinions that go beyond
those expressed in the immediate social context. This aspect
of modernized society can be represented in the model as
noise, which reflects the sum of influences (e.g., exposure to
mass media, contact with people from other cultures) not ac-
counted for by local influence. The greater the magnitude of
noise in a society, the weaker the relative role played by the
individual’s local context. The opinions of someone in a dif-
ferent part of the country, for example, may have a greater
impact on an individual’s opinions than do the opinions of his
or her immediate neighbors. This is clearly not the case in a
stable society, in which everyone is exposed to the same local
contacts throughout much of his or her life.
Computer simulations of the model have demonstrated a
nonlinear relationship between noise and the distribution of
opinions in a society (Latané & Nowak, 1997; Lewenstein
et al., 1993). Small values of noise tend to destabilize weak
clusters (e.g., Nowak, Vallacher, Tesser, & Borkowski,
2000). Because weak clusters tend also to be small, low-level
noise has the effect of increasing the average size of clusters
in the society, which is reflected in higher overall clustering
and polarization. Higher values of noise, however, can desta-
bilize all minority clusters and thus promote unification of
opinions in the society. At very high levels of noise, however,
individuals are likely to adopt opinions that are independent
of their immediate social context. This not only disrupts clus-
ters, but it also prevents unification of opinions in the society.
In effect, everyone switches his or her opinions in a more-or-
less random fashion.
In a stable society characterized by low levels of noise,
then, a stable pattern of relatively small clusters is to be ex-
pected, whereas in a somewhat less stable society character-
ized by moderate levels of noise, larger clusters and greater
opinion polarization is to be expected. With further increases
in societal instability, one might expect a breakdown in mi-
nority opinion clusters and a tendency toward societal uni-
fication in opinion. Finally, in a highly modernized and
unstable society, one would expect the pattern of opinions to
be largely independent of the pattern of social ties (e.g.,
neighborhood influence), demonstrating instead the influence
of other factors, such as selective exposure to the media and
contact with other cultures.
Cultures also differ in their respective values and prefer-
ences regarding everything from clothing to religion. This
feature is represented in the model as bias. If a new idea res-
onates well with a culture’s prevailing values and prefer-
ences, it will take somewhat less social influence for the idea
to take hold in the society. But if the idea runs counter to cul-
tural values, it is likely to be resisted even if it is supported by
considerable influence. Communist ideology was never fully
embraced in Poland, for example, despite the considerable
influence exerted by the government, because communist
values ran counter to strong Polish traditions of independence