Uncertainty Management and Social Issues: Uncertainty as an Important Determinant of Reactions to Socially Deviating People1

Abstract

In this paper, the social psychology of uncertainty management is used to explain reactions to socially deviating people. In Study 1, we examine how people react to a person communicating negative messages about their home country; in Study 2, how a representative sample of the Dutch society reacts to encounters with a homeless person; and in Study 3, what the behavioral and psychological reactions are of people in an anticipated interaction with a homeless individual. All 3 studies reveal that personal uncertainty—whether made salient in a subtle manner or measured by means of individual differences in the extent to which uncertainty is considered an emotionally threatening experience—is an important determinant of reactions to socially deviating persons.

Full text

Page 1

Uncertainty Management and Social Issues: Uncertainty as
an Important Determinant of Reactions to
Socially Deviating People1
Kees van den Bos,2Martin C. Euwema, P. Marijn Poortvliet, and
Marjolein Maas
Department of Social and Organizational Psychology
Utrecht University
Utrecht, The Netherlands
In this paper, the social psychology of uncertainty management is used to explain
reactions to socially deviating people. In Study 1, we examine how people react to a
person communicating negative messages about their home country; in Study 2, how
a representative sample of the Dutch society reacts to encounters with a homeless
person; and in Study 3, what the behavioral and psychological reactions are of
people in an anticipated interaction with a homeless individual. All 3 studies reveal
that personal uncertainty—whether made salient in a subtle manner or measured by
means of individual differences in the extent to which uncertainty is considered an
emotionally threatening experience—is an important determinant of reactions to
socially deviating persons.
Give me your tired, your poor,
Your huddled masses yearning to breathe free,
The wretched refuse of your teeming shore.
Send these, the homeless, tempest-tossed, to me:
I lift my lamp beside the golden door.
These lines from the poem “The New Colossus” by 19th-century American
poet Emma Lazarus appear on a plaque at the base of the Statue of Liberty.
The poem ends with the statue speaking the quoted words. We refer to these
lines here because they illustrate, among other things, an important problem
of current society, as well as basic and applied social psychology; namely, the
issue of how vulnerable (e.g., “tired,” “poor”) and socially deviating people
(e.g., “homeless”) get treated by society and by other people.
1Marijn Poortvliet is now at the University of Groningen, The Netherlands. The authors
thank Kelly See and Dirk Steiner for their comments on previous versions of the paper; Jurgen
Visser and Ine Voorham for their assistance with Study 2; and Michala Kovakova, Joris Moons,
Yvonne Stomphorst, and Angelique van Drenth for their help with Study 3.
2Correspondence concerning this article should be addressed to Kees van den Bos, Depart-
ment of Social and Organizational Psychology, Utrecht University, Heidelberglaan 1, 3584 CS
Utrecht, The Netherlands. E-mail: k.vandenbos@fss.uu.nl
1726
Journal of Applied Social Psychology, 2007, 37, 8, pp. 1726–1756.
© 2007 Copyright the Authors
Journal compilation © 2007 Blackwell Publishing, Inc.

Page 2

To better understand this important matter, we present three studies in
the current paper that explore how people react to persons whom they think
are socially deviating. In Study 1, we examine how people react to a person
communicating negative (vs. positive) messages about their home country. In
Study 2, we investigate how a representative sample of Dutch society reacts
to encounters with a homeless person. Finally, in Study 3, we examine the
behavioral and psychological reactions of people in an anticipated interac-
tion with a homeless individual.
All three studies will reveal that personal uncertainty—whether made
salient in a rather subtle manner (Studies 1 and 3) or measured by means of
stable individual differences in the extent to which uncertainty is considered
to be a threat or not (Study 2)—is an important determinant of how people
react to socially deviating persons. We base our reasoning on recent social
psychological insights suggesting that uncertainty concerns may have con-
siderable impact on how people react to social transgressions and other
events that they consider aversive and negative. After we introduce this work
on uncertainty management, we will present the research hypotheses for the
current research.
Uncertainty Management
Most work on uncertainty management (for recent overviews, see Hogg &
Mullin, 1999; Lind & Van den Bos, 2002; Sorrentino & Roney, 1999; Van den
Bos & Lind, 2002; Weary, Jacobson, Edwards, & Tobin, 2001) begins with
the observation that the world is an uncertain place. For example, many
people have jobs with indefinite tenure, and success at work often depends on
adaptability and flexibility in the face of an uncertain future (Lord & Hartley,
1998). Rapid change is occurring everywhere, and news of layoffs as well as
national and international conflicts reaches us almost daily. Furthermore,
people are unpredictable, and most people have experienced unanticipated
disappointments and unexpected successes in their personal, work, or politi-
cal worlds.
Various important social psychological theories (e.g., Festinger, 1954;
Fiske & Taylor, 1991; Hogg, 2000; Lopes, 1987; Weary & Edwards, 1996)
have argued that people have a fundamental need to feel certain about their
world and their place within it, that uncertainty can be threatening, and that
people generally feel a need either to eliminate uncertainty or to find some
way to make it tolerable and manageable (but for exceptions to this rule, see
Sorrentino & Roney, 1986). Wilson, Centerbar, Kermer, and Gilbert (2005),
for instance, showed that people even want to avoid uncertainty when it
prolongs pleasurable experiences.
UNCERTAINTY AND SOCIAL ISSUES
1727

Page 3

Furthermore, consider the threats that can accompany uncertainty:
Uncertainty deprives one of confidence in how to behave and what to expect
from the physical and social environment within which one finds oneself.
Uncertainty about one’s attitudes, beliefs, feelings, and perceptions—as well
as about one’s relationship to other people—is generally aversive (e.g., Fiske
& Taylor, 1991; Hogg & Mullin, 1999; Lopes, 1987). Therefore, uncertainty
often motivates behavior that reduces subjective uncertainty (Van den Bos,
Heuven, Burger, & Fernández Van Veldhuizen, 2006). Consequently, people
will react very strongly and negatively to anything that does not help them to
cope with uncertainty once it is salient (Van den Bos, Poortvliet, Maas,
Miedema, & Van den Ham, 2005). Moreover, epistemic motives related
to uncertainty are important social psychological principles. Festinger
(1954), for example, based social comparison theory on the proposition
that knowing that one is correct is a critical human motivation that drives
people to make interpersonal social comparisons when nonsocial means are
unavailable.
Based on these theories and notions, Van den Bos and Lind (2002; Lind
& Van den Bos, 2002) recently proposed their uncertainty management
model. Following the previously reviewed work, this model argues that
uncertainty and managing uncertainty play an important role in human life.
This is not to say that people want to reduce uncertainty all the time or that
all uncertainties are the same. Of course, being completely certain about all or
many aspects of one’s life may make one’s life rather dull, and there are
clearly instances in which people strive for uncertainty rather than seek to
reduce it. After all, sometimes people want to experience new, uncertain
events; and on occasion, they even seek the thrill of possible danger, such as
bungee jumping or skydiving. Furthermore, the uncertainty involved in a fair
gamble is stimulating, at least for some people. But the model proposes that
even when uncertainty is sought, it usually is still managed, at least to some
extent.
Personal Uncertainty and Worldview Defense
There are different types of uncertainty that people can encounter (Van
den Bos & Lind, 2002), but the uncertainty management model focuses
especially on the experience of personal uncertainty, which is the result of
people being uncertain about themselves (e.g., see Van den Bos, 2001; Van
den Bos, Poortvliet et al., 2005; see also De Cremer & Sedikides, 2005; Hogg,
2005; McGregor, Zanna, Holmes, & Spencer, 2001). The model proposes
that people want to protect themselves from being in or thinking of situations
in which they are uncertain about themselves. One way in which people can
1728
VAN DEN BOS ET AL.

Page 4

do this, the model argues, is by adhering to their cultural norms and values
(Van den Bos, Poortvliet et al., 2005). Experiences that are supportive of
people’s cultural worldviews lead people to be less uncertain about them-
selves or to be able to better tolerate the uncertainty (Van den Bos et al.,
2006). As a result, the model predicts, people who are uncertain about
themselves or who have been reminded about their personal uncertainties will
react very positively toward worldview-supportive experiences (Van den Bos,
2001).
In contrast, experiences that threaten or impinge on people’s worldviews
do not help people at all to cope with their uncertainties; hence, people will
respond very negatively toward these worldview-threatening experiences
(Van den Bos, Poortvliet et al., 2005). In this way, the model hypothesizes
that under conditions of uncertainty, people will react especially positively
toward the occurrence of events or toward persons that uphold their cultural
norms and values and particularly negatively toward transgressions of these
concepts (Van den Bos & Lind, 2002; Van den Bos, Poortvliet et al., 2005).
Although an elaborate overview of the empirical work on uncertainty
management is beyond the scope of the present paper (for more complete
descriptions, see Van den Bos, 2001; Van den Bos, Poortvliet et al., 2005),
the results are in accordance with predictions derived from the uncertainty
management model. For example, research by Van den Bos (2001) was
founded on the observation that in most (if not all) societies, being treated
in a fair manner is in accordance with cultural norms and values, whereas
being treated in an unfair way is a violation of these norms and values (e.g.,
Folger, 1984; Folger & Cropanzano, 1998; Tyler & Smith, 1998). Integrat-
ing this observation with the previously reviewed uncertainty management
model, Van den Bos argued that this should imply that asking (as opposed
to not asking) people to think about their uncertainties should lead them to
react more positively toward fair events and more negatively toward unfair
events. The findings reported by Van den Bos were supportive of this
hypothesis.
Furthermore, in addition to reactions to fair and unfair events (Van den
Bos, 2001), salience of personal uncertainty may also moderate reactions to
other experiences that bolster or violate people’s cultural norms and values.
For example, Van den Bos, Poortvliet et al. (2005) built their third and
fourth experiments on the observation that because of social identity con-
cerns (e.g., Tajfel & Turner, 1979, 1986) and belongingness needs (e.g.,
Baumeister & Leary, 1995), praise of students’ own university constitutes a
bolstering of their cultural worldviews, whereas criticism of the university
represents a violation of participants’ worldviews (Dechesne, Janssen, & Van
Knippenberg, 2000). Following this line of research, Van den Bos, Poortvliet
et al. hypothesized and showed that when university students were reminded
UNCERTAINTY AND SOCIAL ISSUES
1729

Page 5

of their personal uncertainties, they reacted more positively toward infor-
mation that was favorable about their university and more negatively
toward information that was unfavorable about their university (Experi-
ments 3 and 4). Moreover, the five experiments presented in Van den Bos,
Poortvliet et al.’s paper all suggest that, at least sometimes, the models of
uncertainty management (e.g., Van den Bos & Lind, 2002; see also Martin,
1999; McGregor et al., 2001) may better explain people’s reactions to cul-
tural worldview defense reactions than a viable alternative account (i.e.,
terror management theory; e.g., see Greenberg, Solomon, & Pyszczynski,
1997; Pyszczynski, Greenberg, & Solomon, 1999; Solomon, Greenberg, &
Pyszczynski, 1991).
Other studies have provided support for predictions by related uncer-
tainty management models. Hofstede (2001), for example, showed that com-
pared to people low in uncertainty avoidance, those high in uncertainty
avoidance are more conservative, less tolerant of diversity, and less open to
new experiences and alternative lifestyles. In addition, they want immigrants
to be sent back to their countries of origin, and they reject people from other
races as their neighbors.
McGregor and colleagues (McGregor, 2004; McGregor & Marigold,
2003; McGregor et al., 2001) revealed that people who are made uncertain
about themselves react more defensively toward events that threaten their
cultural worldview, and that people do so because in this way they want to
restore their sense of self (namely, being persons who can be certain about
themselves; see also Martin, 1999). Related to this, Hogg (2000, 2005)
showed that extreme self-uncertainty motivates people to believe more in
ideological belief systems related to orthodoxy, hierarchy, and extremism (see
Towler, 1984).
The Current Research
Because of space considerations, we refer readers to the publications
reviewed here for more extensive introduction to the uncertainty manage-
ment models that have been reported in the literature (e.g., Hogg, 2000;
Martin, 1999; Sorrentino & Roney, 1999; Towler, 1984; Van den Bos & Lind,
2002; Weary & Edwards, 1996) and the empirical studies that support impor-
tant components of these models (e.g., De Cremer & Sedikides, 2005; Hogg,
2005; Sorrentino & Roney, 1986; Van den Bos, 2001; Van den Bos, Poortvliet
et al., 2005; Weary et al., 2001; Wilson et al., 2005). Here, we would like to
use the insights we can derive from the previously reviewed uncertainty
management work to better understand how people react to others whom
they think are socially deviating from themselves.
1730
VAN DEN BOS ET AL.

Page 6

More specifically, we argue that the uncertainty findings reviewed here tell
us something that is fundamental to the issues we are attempting to investi-
gate. On the basis of the previously reviewed uncertainty management
models and empirical studies pertaining to these models (De Cremer &
Sedikides, 2005; Hogg, 2000, 2005; Martin, 1999; Sorrentino & Roney, 1986,
1999; Towler, 1984; Van den Bos, 2001; Van den Bos & Lind, 2002; Van den
Bos, Poortvliet et al., 2005; Weary & Edwards, 1996; Weary et al., 2001;
Wilson et al., 2005), as well as the notion that persons who are socially
deviant can be threatening to people’s cultural worldviews (Greenberg et al.,
1997; Pyszczynski et al., 1999; Solomon et al., 1991), we hypothesize that
people may be especially likely to react negatively to people they consider to
be social deviants when they have been thinking about what makes them
uncertain (Studies 1 and 3) or when they consider uncertainty to be a threat
rather than a challenge (Study 2). This is presumed to occur because such
experiences that threaten or impinge on people’s worldviews do not help
people to cope with the uncertainty at all. This general hypothesis will be
tested in three studies, after which we will discuss the theoretical and applied
implications.
Study 1
In Study 1, participants participated in ostensibly two unrelated studies.
In the first study, we used the method developed by Van den Bos (2001) to
make personal uncertainty salient or not salient to our Dutch participants. In
the second study, building on the method used Van den Bos, Poortvliet et al.
(2005; Experiments 3 and 4), participants were asked to read a newspaper
article in which a person was interviewed who was either very positive or very
negative about The Netherlands. After this, we assessed participants’ incli-
nations to protest against the ideas of the person in the article. We did this
because from the social psychological literature on social identity (e.g., Tajfel
& Turner, 1979, 1986) and belongingness (e.g., Baumeister & Leary, 1995), it
is interesting to explore people’s reactions to positive and negative commu-
nications about their in-group in general (e.g., Tajfel & Turner, 1979) and
their home country in particular (e.g., Greenberg et al., 1997). Investigating
this is even more important following the events of 9/11/2001 and the chal-
lenges that this imposes on communications between members of different
civilizations (Pyszczynski, Solomon, & Greenberg, 2003; Skitka, Bauman, &
Mullen, 2004).
In The Netherlands, a particularly important social issue in this respect is
how people from Muslim-oriented cultures communicate about the Dutch
society. Therefore, in Study 1, we asked Dutch participants to respond to
UNCERTAINTY AND SOCIAL ISSUES
1731

Page 7

communications that were apparently coming either from a Muslim-oriented
out-group source that is negative about The Netherlands or from another
source that is positive about Dutch society.
Method
Participants and Design
Participants were 80 students (19 men, 61 women)3at Utrecht University
who were paid for their participation. They were randomly assigned to one of
the conditions of the 2 (Salience: uncertainty vs. television) ¥ 2 (Article:
positive vs. negative) between-participants design.
Procedure
Students at Utrecht University were invited to the laboratory to partici-
pate in a study on human judgment. Upon their arrival at the laboratory,
participants were led to separate cubicles, each of which contained a com-
puter with a monitor and a keyboard. Next to the monitor, participants
found pieces of paper and a pencil. The computers were used to present the
stimulus information and to collect data on the dependent variables. Partici-
pants answered the questions that constituted the dependent variables and
manipulation checks before or after participating in other, unrelated studies.
The studies lasted a total of 40 min, and participants were paid 5 Euros for
their participation (1 Euro was equal to approximately $1.20 U.S. at the time
the studies were conducted).
The study was presented to participants as two separate studies. In the
first study, the salience manipulation was induced. Following Van den Bos
(2001), participants in the uncertainty-salient condition were asked questions
about their thoughts and feelings of being uncertain: Participants were asked
to write down on the pieces of paper next to the computer their responses to
the following: “Please briefly describe the emotions that the thought of your
being uncertain arouses in you,” and “Please write down, as specifically as
you can, what you think physically will happen to you as you feel uncertain.”
Participants in the television-salient condition were asked two questions
that were similar in format and that did not remind participants about their
3Gender and other demographic variables did not influence the results of the three studies
reported here and, hence, were dropped from the analyses that are reported. Study 1 participants
indicated, when asked the extent to which they identify themselves as Dutch persons (1 = very
weak to 7 = very strong), that they identify strongly as Dutch persons (M = 5.46, SD = 1.21).
1732
VAN DEN BOS ET AL.

Page 8

uncertainties (Van den Bos, 2001). These participants were asked to write
down on a piece of paper their responses to the following: “Please briefly
describe the emotions that the thought of you watching TV arouses in you,”
and “Please write down, as specifically as you can, what you think physically
will happen to you as you watch TV.”
After this, all participants completed the Positive and Negative Affect
Schedule (PANAS; Watson, Clark, & Tellegen, 1988) on which they reported
on 20 items how they felt at the moment. Following previous uncertainty
salience studies (Van den Bos, 2001), the PANAS was included as a filler task
and to determine if the salience manipulation engendered positive and nega-
tive affect. The PANAS consists of two 10-item subsets (Watson et al., 1988),
one measuring positive affect (PA) and one measuring negative affect
(NA). Both subsets were averaged to form reliable scales (as = .77 and .78,
respectively).
The second study then began. In this study, participants were asked to
read an article that had appeared in a newspaper recently. In the article,
extending Arndt and Greenberg (1999), people had been interviewed about
their opinions on The Netherlands. In the positive article condition, partici-
pants read the following:
That is an easy question to answer. I think this is a great
country: beautiful and very friendly. There is a great amount of
freedom here: Everything goes and people can do whatever they
like as long as they respect others. I can imagine well that a lot
of people admire us. After all, there are lots of things to be
proud of: beautiful canals, people on bikes everywhere, cozy
winters, nice bread, not a lot of bureaucratic procedures, a lot of
tolerance, and you can count on each other when needed.
People are really open-minded and interested here. Tourists
often tell you how great they find it to be here, and quite rightly
so. We should be proud of living in The Netherlands!
The negative article read as follows:
That is an easy question to answer. I think this is a filthy
country: ugly and hostile. There is no real freedom here: Those
Dutch people think they have sorted it all out, but they cannot
even make the trains run on time. And this bureaucratic gov-
ernment is interfering with everybody’s lives. They are talking
about freedom of religion and tolerance, but we cannot really
discuss the possibility of something such as circumcision of
women. Well, I will circumcise my daughters, even if I have to
do so by illegal means. They cannot stop me from being a good
UNCERTAINTY AND SOCIAL ISSUES
1733

Page 9

Muslim. I want to be proud of my children, so I do not allow
them to mingle with the Dutch. After all, Christianity is an
inferior religion. People think The Netherlands is great, but that
is not true at all. It is a spoiled and cold country, the weather is
terrible, and the people are interfering with other people’s lives.
I am suffocating here. There is plenty of reason why so many
people in this world hate The Netherlands. People who are
satisfied with this country are just stupid!
Following the presentation of the article, the dependent variables were
solicited. Dependent variables of Study 1 asked participants the extent to
which they would like to protest against the ideas of the person in the article,
andtheextenttowhichtheywouldliketocriticizetheideasofthepersoninthe
article. Both questions were answered on a 7-point Likert-type scale ranging
from1(veryweak)to7(verystrong).Participants’responsestothesetwoitems
were averaged to form a reliable scale of their protest intentions (a = .84).
The salience manipulation was checked by asking participants whether
they had been thinking about uncertainty when they were writing down their
answers. Responses were rated on a 7-point scale ranging from 1 (definitely
did not) to 7 (definitely did). They were then asked the extent to which they
had been thinking about uncertainty when they were writing down their
answers (a = .98). Responses were rated on a 7-point scale ranging from 1
(very weak) to 7 (very strong). Participants were also asked whether they had
been thinking about watching television when they were writing down their
answers. Responses were rated on a 7-point scale ranging from 1 (definitely
did not) to 7 (definitely did). Finally, participants were asked the extent to
which they had been thinking about watching television when they were
writing down their answers (a = .98). Responses were rated on a 7-point scale
ranging from 1 (very weak) to 7 (very strong).
After they completed the study, participants were paid for their partici-
pation and were thoroughly debriefed (for the complete debriefing proce-
dure, see Van den Bos, 2003). During the debriefing procedure, participants
indicated that they did not experience a direct relationship between the
salience manipulation and their reactions to the article they read. None of the
participants objected to the procedures used in Study 1.
Results
Manipulation Check
A 2 ¥ 2 MANOVA on the uncertainty and television salience scales indi-
cates only a main effect of the salience manipulation at both the multivariate
1734
VAN DEN BOS ET AL.

Page 10

and univariate levels: multivariate F(2, 75) = 67.04, p < .001; uncertainty
salience scale, F(1, 76) = 84.85, p < .001; television salience scale, F(1,
76) = 83.32, p < .001. As expected, uncertainty was more salient in the uncer-
tainty condition (M = 4.25, SD = 0.68) than in the television condition
(M = 2.76, SD = 0.76). Similarly, thoughts of watching television were more
salient in the television condition (M = 4.50, SD = 1.94) than in the uncer-
tainty condition (M = 1.41, SD = 0.83). This shows that the salience manipu-
lation was successful in affecting the relative strength of participants’
thoughts in ways that were intended with this manipulation.
PANAS
The PANAS (Watson et al., 1988) was administered following the
salience manipulation to serve as a filler task and to determine whether
unintended effects of the screen manipulation on the positive and negative
subsets would be found. In correspondence with earlier uncertainty salience
studies (Van den Bos, 2001; Van den Bos, Poortvliet et al., 2005), a 2 ¥ 2
MANOVA on the positive and negative subsets of the PANAS yielded no
effects at both the multivariate and univariate levels. This suggests
that affect cannot explain the findings reported here. Overall means of the
positive and negative subsets were 2.99 (SD = 0.50) and 1.34 (SD = 0.36),
respectively.4
Dependent Variables
Means and standard deviations of participants’ protest intentions are
presented in Table 1. A 2 ¥ 2 ANOVA on this scale yields a significant effect
of the article manipulation, F(1, 76) = 5.54, p < .03; and a significant interac-
tion effect between the salience and article manipulations, F(1, 76) = 5.94,
p < .02. The main effect of the salience manipulation was not significant.
These effects show that the negative article yielded more protest than did the
positive article. Furthermore, in accordance with our expectations, partici-
pants’ protest intentions were more strongly influenced by the article manipu-
4Participants did not experience PA differently in the uncertainty-salient condition
(M = 2.96, SD = 0.50) than in the television-salient condition (M = 3.02, SD = 0.50), F(1,
78) = 0.37, p = .55. NA did not differ when comparing the uncertainty-salient condition
(M = 1.35, SD = 0.39) with the television salient condition (M = 1.33, SD = 0.33), F(1,
78) = 0.03, p = .85. Furthermore, PA (r = -.15, p = .18) and NA (r = -.06, p = .63) did not cor-
relate significantly with participants’ protest intentions. Moreover, controlling for both PA and
NA still yielded the predicted interaction effect on participants’ protest intentions, F(1,
74) = 6.61, p < .02.
UNCERTAINTY AND SOCIAL ISSUES
1735

Page 11

lation in the uncertainty condition than in the television condition. In fact,
the results show a significant effect of the article manipulation within the
uncertainty condition, F(1, 76) = 11.48, p < .01; and a nonsignificant effect
within the television condition, F(1, 76) = 0.00, ns.
As an aside, it can be noted here that the results indicate nonsignificant
effects of the salience manipulation within both the positive and negative
article conditions. We will come back to this observation in the General
Discussion.
Study 2
As predicted, the findings of Study 1 reveal that when uncertainty has
been made salient to people, they react more strongly toward positive versus
negative communications about their home country. This suggests that a
subtle manipulation of salience of uncertainty concerns can have reliable
effects on people’s reactions; particularly on how likely they are to protest
against communications about their in-group. However, before strong con-
clusions are drawn on the basis of these findings, it is important to provide
more empirical evidence pertaining to the hypothesis that is studied in the
present paper.
In Study 2, we attempt to broaden the evidence regarding our research
hypothesis in several ways. First, we want to obtain evidence for our predic-
tions not only in a sample of university students, but also in a more repre-
sentative sample. Therefore, a sample of the Dutch society serves as the
respondents in Study 2.
Second, research thus far on uncertainty management has largely
neglected other important variables that may moderate the relationship
between uncertainty concerns and people’s reactions to other people (e.g.,
Van den Bos, 2001; Van den Bos & Lind, 2002; Van den Bos, Poortvliet et al.,
Table 1
Means of Protest Intentions as a Function of Salience and Article: Study 1
Uncertainty salientTelevision salient
M SDM SD
Positive article
Negative article
3.73
5.15
1.46
1.36
4.35
4.33
1.11
1.36
1736
VAN DEN BOS ET AL.

Page 12

2005). In Study 2, we attempt to fill this void by taking into account the
possible moderating role of the valence of respondents’ prior attitudes
toward vulnerable people in society.
Third, the psychology of uncertainty management involves more than just
making uncertainty salient to people or not doing so. Greco and Roger
(2001), in particular, have shown that there are important stable individual
differences in how people emotionally cope with uncertainty. Therefore, in
Study 2 we explore the extent to which these individual differences may serve
as an important moderator of how people react to socially deviating others.
In the pilot studies we found that of the uncertainty scales that have been
reported in the social psychological literature, Greco and Roger’s (2001)
Emotional Uncertainty Scale was the one that had the best chance of reli-
ably predicting people’s reactions to socially deviating others. Specifically,
our pilot studies suggested that Greco and Roger’s Cognitive Uncertainty
Scale and their Desire for Change Scale, as well as self-constructed items
assessing experienced personal uncertainty in modern society, and scales
measuring uncertainty avoidance (e.g., Hofstede, 2001), uncertainty orien-
tation (e.g., Sorrentino, Holmes, Hanna, & Sharp, 1995), need for closure
(e.g., Kruglanski & Webster, 1996), and need for cognition (e.g., Cacioppo,
Petty, & Kao, 1984) would yield weaker effects on our participants’ reac-
tions than would the Emotional Uncertainty Scale. We think that our pilot
studies suggested this result because the Emotional Uncertainty Scale mea-
sures how people respond emotionally to uncertainty and uncertainty-
related events, and that it is precisely this emotional component of
uncertainty that is important in predicting people’s reactions to persons and
events that threaten their worldviews (see also Van den Bos, Van Ameijde,
& Van Gorp, 2006).
The 15-item Emotional Uncertainty Scale (Greco & Roger, 2001) mea-
sures the extent to which people consider uncertainty an emotionally threat-
ening experience. Items of the scale include “I feel anxious when things are
changing,” and “I get worried when a situation is uncertain.” The scale has
been shown to have high internal and test–retest reliabilities and has been
validated using physiological and psychological variables (Greco & Roger,
2001). Therefore, we measured the scale to assess the extent to which Study
2 respondents considered personal uncertainty an emotionally threatening
experience.
Fourth, we want to obtain evidence for our predictions on people’s reac-
tions to other types of socially deviating people who perhaps are less extreme
than the stimulus person that was used in Study 1. More specifically, because
the homeless are socially deviating persons in Dutch and other societies, and
because how citizens react to homeless people is an important social issue,
reactions to a homeless person is the focus of attention in Studies 2 and 3. In
UNCERTAINTY AND SOCIAL ISSUES
1737

Page 13

the Study 2 questionnaire, we interviewed respondents about how they would
interact with a homeless person; and in Study 3, we assessed participants’
behavioral and other reactions toward an anticipated real social interaction
with a homeless person.
A fifth and final extension of Study 2 is to measure respondents’ negative
affective reactions to the homeless individual. Previous uncertainty manage-
ment research has shown that negative affect may be the dependent variable
that is most sensitive to tapping the effects of uncertainty manipulations (e.g.,
Van den Bos, 2001; Van den Bos, Poortvliet et al., 2005), suggesting that this
may well be one of the constructs that is predicted best by people’s uncer-
tainty concerns. Following the uncertainty management work outlined
earlier, we predict that on this dependent variable, respondents will react
strongly to an encounter with a homeless individual when their attitudes
toward vulnerable people are relatively negative and they consider uncer-
tainty to be a threatening emotional experience.
Method
Participants
In September 2003, a representative sample (N = 1,650) of Dutch society
of 18 years and older received our questionnaire by means of an e-mail sent
by TNS NIPO Consult.5Respondents were drawn from a panel of respon-
dents maintained by TNS NIPO Consult. The characteristics of the panel and
the sample closely match those of the Dutch population.
A total of 1,277 respondents (621 men, 656 women) e-mailed the ques-
tionnaire back to TNS NIPO Consult, yielding a response rate of 77%. There
were no significant differences in the demographic profile (gender, age,
marital status, income, and education) of those who responded versus those
who did not respond to the survey. Respondents’ ages varied between 18 and
85 years, and the mean age was 46.33 years (SD = 15.91). Of the respondents,
985 were married or living together. Most respondents earned an annual
gross income between 34,000 (approx. $42,500) and 45,000 Euros (approx.
$56,250; M = 55,580 Euros, approx. $69,475; SD = 975 Euros, approx.
$1,218.75). Of the respondents, 78% who had completed either college (22%)
or a lower level of education (56%), while 6% were students, and 15% had not
completed any level of education.
5For more information on TNS NIPO Consult, which is a consultancy company that
conducts surveys and other studies among the Dutch population, see http://www.tns-nipo-
consult.com/
1738
VAN DEN BOS ET AL.

Page 14

Measures
The questionnaire asked respondents about their attitudes toward the
vulnerable in society, respondents’ personal uncertainty, and their negative
affective reactions toward a homeless person. All responses in Study 2 were
rated on a 5-point scale ranging from 1 (disagree) to 5 (agree).
Respondents’ attitudes toward the vulnerable in society were measured by
asking them to respond to the following questions: “I care about people who
because of being old cannot participate any more in our society,” “I am
concerned about lonely people,” “I care about children who suffer from their
parents’ problems,” and “I am concerned about people who have been sexu-
ally abused.” Respondents’ answers to these four items were averaged to
form a reliable scale of their attitudes toward the vulnerable in society
(a = .77).6
Respondents’ emotional reactions to experiencing uncertainty in their
personal lives were measured by having respondents complete Greco and
Roger’s (2001) Emotional Uncertainty Scale. Respondents’ answers to the 15
items of the scale were averaged to form a reliable index of their emotional
reactions to personal uncertainty (a = .91).
Negative affective reactions toward a homeless person were assessed by
asking respondents to respond to the following situation described in the
questionnaire:
At the entrance of your home, there is this untidy man of about
50 years. He is drinking from a beer can and is smoking a
cigarette. Next to him there is a big plastic shopping bag.7
Respondents’ negative affective reactions were measured by asking
respondents how angry, hostile, furious, and infuriated they felt about this
situation. Respondents’ answers to these four items were averaged to form a
reliable scale of their negative affective reactions (a = .86). Following Study
1, participants’ protest reactions toward the presence of the homeless person
were assessed by asking respondents the extent to which they would like to
6Cross-cultural research (Hofstede, 2001) has suggested that the Dutch are more oriented
toward helping others who are in need than are Americans (also see Van den Bos, Steiner, Van
Yperen, & Dekker, 2005) and that they may consider both those who are not responsible for
their situation (e.g., the elderly, children who suffer because of their parents) and those who are
somewhat responsible for their situation (e.g., the homeless) as vulnerable people. The attitude
scale of Study 3 will be more strongly related to homeless attitudes, thereby making it a
potentially stronger predictor of reactions to homeless individuals.
7Drinking and smoking in public places typically are permitted in The Netherlands. Fur-
thermore, “at the entrance of one’s home” in The Netherlands does not imply that the person is
on the private property of respondents, as this implies still being in public territory in The
Netherlands.
UNCERTAINTY AND SOCIAL ISSUES
1739

Page 15

protest against the presence of this person and the extent to which they would
like to criticize the presence of this person at the entrance of their homes,
again yielding a reliable scale of their protest reactions (a = .80).
Results
Table 2 reports the overall means, standard deviations, and intercorrela-
tions of Study 2 independent and dependent variables. We performed regres-
sion analyses on respondents’ negative affective reactions and their protest
reactions to test for the predicted effects between the centered (Aiken & West,
1991) attitude and uncertainty variables.
On respondents’ negative affective reactions, regression analysis yields a
significant main effect of the uncertainty variable (Beta = .20, p < .001) and a
significant interaction effect between the attitude and uncertainty variables
(Beta = -.10, p < .01). These effects are depicted in Figure 1. The main effect
of the attitude variable was not significant. As can be seen in Figure 1,
respondents with a relatively negative attitude toward vulnerable people (1
SD below the mean; see Aiken & West, 1991) were influenced more strongly
by their emotional uncertainty style than were those with a relatively positive
attitude toward vulnerable people (1 SD above the mean). In fact, simple
slopes were significant among the relatively negative attitude respondents
( p < .001) and were not statistically significant among the relatively positive
attitude respondents (p > .60).
In further correspondence with our predictions, respondents’ protest
reactions show a significant main effect of the uncertainty variable
(Beta = .11, p < .001) and an interaction between the attitude and uncer-
Table 2
Overall Means and Intercorrelations of Study Variables: Study 2
M SD
123
1. Attitude toward
vulnerable people
2. Emotional
uncertainty
3. Negative affective
reactions
4. Protest reactions
1.840.65—
3.050.58.15*—
3.170.70
-.01 .20*—
2.640.90
-.02.11*.60*
*p < .001.
1740
VAN DEN BOS ET AL.