497064

2013
17110.1177/1368430213497064Group Processes & Intergroup RelationsDhont et al.

G
Group Processes & P
Intergroup Relations I
Article R

Group Processes & Intergroup Relations

Changing the ideological roots of

2014, Vol 17(1) 27­–44
© The Author(s) 2013
Reprints and permissions:
prejudice: Longitudinal effects of sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1368430213497064
ethnic intergroup contact on social gpir.sagepub.com

dominance orientation

Kristof Dhont,1 Alain Van Hiel,1 and Miles Hewstone2

Abstract
Social dominance orientation (SDO) has been reported to be strongly related to a multitude of
intergroup phenomena, but little is known about situational experiences that may influence SDO.
Drawing from research on intergroup contact theory, we argue that positive intergroup contact is able
to reduce SDO-levels. The results of an intergroup contact intervention study among high school
students (Study 1, N = 71) demonstrated that SDO levels were indeed attenuated after the intervention.
Furthermore, this intervention effect on SDO was especially pronounced among students reporting
a higher quality of contact. A cross-lagged longitudinal survey among adults (Study 2, N = 363)
extended these findings by demonstrating that positive intergroup contact is able to decrease SDO
over time. Moreover, we did not obtain evidence for the idea that people high in SDO would engage
less in intergroup contact. These findings indicate that intergroup contact erodes one of the important
socio-ideological bases of generalized prejudice and discrimination.

Keywords
cross-lagged design, intergroup contact, longitudinal, prejudice, social dominance orientation

Paper received 07 June 2012; revised version accepted 22 May 2013.

Bringing members of different groups together, as affect people’s socio-ideological beliefs about
originally proposed by Allport’s (1954) intergroup intergroup relations, which have been shown to
contact hypothesis, is one of the most effective underlie this diverse set of dimensions.
methods of reducing prejudice (R. Brown &
Hewstone, 2005; Hewstone, 2009; Pettigrew & 1
Ghent University, Belgium
Tropp, 2011). Intergroup contact has, however, 2
University of Oxford, UK
been reported to have a more widespread impact
than merely reducing prejudice as it affects a vari- Corresponding author:
Kristof Dhont, Department of Developmental, Personality
ety of dimensions related to generalized preju- and Social Psychology, Ghent University, Henri Dunantlaan
dice (Hewstone, 2009; Pettigrew & Tropp, 2011). 2, B-9000 Ghent, Belgium.
We therefore argue that contact may, in addition, Email: [email protected]

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

28 Group Processes & Intergroup Relations 17(1)

In the current research, we focused on the Nevertheless, some studies testing the reverse
potential impact of positive intergroup contact on causal direction also obtained evidence that atti-
social dominance orientation (SDO; Pratto, tudes and feelings towards a particular group can
Sidanius, Stallworth, & Malle, 1994), which is con- longitudinally predict SDO (Kteily et al., 2011;
ceptualized as a broad social attitude expressing an Matthews, Levin, & Sidanius, 2009; Sibley & Liu,
individual’s preference for hierarchically structured 2010), which complements research showing that
group relations and inequality among social groups SDO is sensitive to situational and social influ-
(Pratto et al., 1994; Sidanius & Pratto, 1999). ences. Experimental studies have shown, for
Previous cross-sectional studies have reported a instance, that SDO is enhanced with higher levels
negative relationship between positive intergroup of societal resource scarcity and competition, or
contact and SDO (e.g., Asbrock, Christ, Duckitt, & when membership of dominant social groups is
Sibley, 2012; Dhont & van Hiel, 2009; Hodson, made salient (Guimond, Dambrun, Michinov, &
2008), which seems to support the hypothesis that Duarte, 2003; Huang & Liu, 2005; Morrison &
people high in SDO (“social dominators”) tend to Ybarra, 2008; Schmitt, Branscombe, & Kappen,
avoid intergroup contact. However, because these 2003). Together, these studies corroborate recent
previous studies based on cross-sectional data do theorizing of Duckitt (2001) arguing that, despite
not allow us to draw inferences about the direction being a strong predictor of prejudice, SDO is not
of the relationships, we used both an intervention an unchangeable or deeply rooted personality
study and a cross-lagged longitudinal study to trait, but varies as a function of changing percep-
investigate the relationship between positive inter- tions of the social world as a competitive place.
group contact and SDO. These competitive worldviews are, in turn,
derived from individual socialization experiences
and exposure to particular social contexts charac-
SDO: Cause or Effect of Intergroup terized by high levels of inequality and competi-
Attitudes, or Both? tion (Perry, Sibley, & Duckitt, 2013; Sibley,
SDO has been reported to be a strong and unique Wilson, & Duckitt, 2007b).
predictor of a multitude of intergroup phenom- Furthermore, social dominance theorists
ena and types of prejudice across different (Pratto et al., 2006; Sidanius & Pratto, 1999) have
domains, for example, racial and ethnic prejudice repeatedly acknowledged that individual SDO
(Hodson & Esses, 2005; Pratto et al., 1994; van levels are sensitive to life and socialization expe-
Hiel & Mervielde, 2005) and sexism (Roets, van riences such as, among others, education and
Hiel, & Dhont, 2012; Sibley, Wilson, & Duckitt, multicultural experiences (Pratto et al., 2006).
2007a). According to social dominance theory However, due to the main research focus on
(Pratto, Sidanius, & Levin, 2006; Sidanius & SDO as a predictor of various measures of inter-
Pratto, 1999), the rejection of particularly low- group bias (e.g., Kteilly, Ho, & Sidanius, 2012) or
status groups by people high in SDO should be on (perceptions of) contextual factors that may
considered as an endorsement of hierarchy- increase SDO (e.g., Huang & Liu, 2005; Sibley
enhancing legitimizing myths that serve the pres- et al., 2007b), the potential role of individual
ervation of group-based hierarchies in society. contextual experiences like intergroup contact in
SDO is thus primarily conceived as a cause of decreasing SDO has remained an understudied
prejudice and group-based attitudes. This causal topic.
perspective has been supported by several recent
studies showing that SDO does indeed have a The Relationship Between
longitudinal influence on prejudice (Asbrock,
Sibley, & Duckitt, 2010; Kteily, Sidanius, & Levin,
Intergroup Contact and SDO
2011), as well as on attitudes towards inequality In his seminal book, Gordon Allport (1954)
policies for specific groups (Sibley & Duckitt, proposed that contact between members of dif-
2010; Sibley & Liu, 2010). ferent groups reduces mutual prejudice when

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

Dhont et al. 29

contact occurs under conditions of equal status, 2012; Dhont & van Hiel, 2009, 2011; Hodson,
cooperation, common goals, and institutional 2008, 2011). These studies reported that SDO is
support. The prejudice-reducing effect of posi- negatively related to intergroup contact which
tive intergroup contact has been confirmed by points, according to Asbrock et al. (2012), to an
Pettigrew and Tropp’s (2006) meta-analysis of important barrier for social dominators to ben-
515 studies, revealing a negative relationship efit from intergroup contact. Perhaps people
between intergroup contact and prejudice with a who do not want to have intergroup contact
mean effect size of r = −.21. This impact is typi- might be thought to be unlikely to benefit from
cally stronger when the contact situation meets it, although contact effects are actually stronger
Allport’s necessary conditions, but these condi- where there is no choice (see Pettigrew & Tropp,
tions are not essential to achieve the effects of 2006). Asbrock et al. (2012) were therefore
intergroup contact. rather pessimistic about the usefulness of inter-
While contact studies typically used cross- group contact among people high in SDO. It
sectional designs in the past, the literature has should also be noted that Asbrock et al. (2012)
recently been enriched by a number of longitudi- reported that intergroup contact was still signifi-
nal studies, demonstrating the effects of inter- cantly associated with reduced prejudice among
group contact on prejudice across time (e.g., those high in SDO in their second study, whereas
Binder et al., 2009; Dhont, van Hiel, De Bolle, & a nonsignificant relationship was found in their
Roets, 2012; Swart, Hewstone, Christ, & Voci, first study.
2011). Some of these studies also found evidence However, the cross-sectional nature of the
for a bidirectional relationship between inter- available data on the relationship between inter-
group contact and prejudice, indicating that posi- group contact and SDO (Asbrock et al., 2012;
tive intergroup contact reduces prejudice, but Dhont & van Hiel, 2009; Hodson, 2008) does not
also that prejudiced people tend to avoid inter- allow for inferences concerning causality.
group contact (e.g., Binder et al., 2009; Swart Therefore, the pessimistic conclusion of Asbrock
et al., 2011). Furthermore, research has shown et al. (2012) needs further investigation. Indeed,
that both frequency (“more contact”) and quality just like the relationship between intergroup con-
(“better contact”) of intergroup contact are asso- tact and prejudice (e.g., Binder et al., 2009; Swart
ciated with less prejudice. When measured sepa- et al., 2011), the contact–SDO relationship may
rately, however, contact quality is typically the also reflect a bidirectional effect. People high in
stronger predictor of the two (e.g., Binder et al., SDO may engage less in intergroup contact, but
2009; Eller & Abrams, 2003). Nevertheless, an the negative correlation between contact and
optimal combination of both frequent and posi- SDO might also indicate that intergroup contact
tive intergroup contact, as expressed for instance has the potential to decrease SDO. Although
in cross-group friendships or a high frequency of both effects may operate simultaneously, we
positive contact, is generally considered the most argue that the latter is more plausible than the
effective way to reduce prejudice (e.g., Davies, former.
Tropp, Aron, Pettigrew, & Wright, 2011; Dhont On the one hand, people tend to select envi-
et al., 2012; Pettigrew, 1997; Swart et al., 2011; ronments that fit their attitudes and values (e.g.,
Voci & Hewstone, 2003). Bretz & Judge, 1994). Along similar lines, social
Despite the overwhelming body of research dominance theorists (e.g., Haley & Sidanius,
on intergroup contact, the role of social-ideologi- 2005), have argued that people endorsing anti-
cal attitudes in intergroup contact has been egalitarian values are likely to be attracted by
ignored for a long time. Researchers have only environments and institutions that support hier-
recently included measures of ideological atti- archically structured group relationships,
tudes in their research designs while investigat- whereas those with egalitarian values are more
ing the effects of contact (e.g., Asbrock et al., attracted by environments and institutions that

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

30 Group Processes & Intergroup Relations 17(1)

attenuate group-based hierarchy. Sidanius, van against outgroups uninvolved in the contact situ-
Laar, Levin, and Sinclair (2003), for instance, ation (e.g., Schmid, Hewstone, Küpper, Zick, &
found that high-SDO students are more Wagner, 2012; Tausch et al., 2010).
attracted to “hierarchy enhancing” careers such The finding that positive intergroup contact
as national security officer or military personnel, affects a multitude of intergroup variables seems
compared to low-SDO students. Moreover, to suggest that intergroup contact erodes the
such environments are unlikely to be frequented socio-ideological basis of generalized prejudice
by members of minority or subordinate groups and discrimination, as represented by SDO
and therefore high-SDO people may have less (Duckitt, 2001; Pratto et al., 1994). Indeed, during
(positive) contact with them. Furthermore, if positive intergroup encounters, people are pro-
we conceive positive intergroup contact itself as vided with a social context that is characterized
a hierarchy-attenuating environment, then we by equality and cooperation rather than inequality
may expect that high-SDO people will not be and competition. The frequent experience of a
eager to engage in such intergroup situations cooperative intergroup environment challenges
(Haley & Sidanius, 2005). people’s view of the world as a competitive jun-
Yet, on the other hand, in modern multicul- gle, which is a typical psychological basis of SDO
tural societies most people are likely to come into (Duckitt, 2001; Perry et al., 2013; Sibley et al.,
some kind of contact with outgroup members, 2007b). Moreover, in terms of social dominance
although to what extent will depend on various theory, positive intergroup contact may be con-
factors ranging from community segregation to sidered as a small-scale, hierarchy-attenuating sit-
motivation. The question of interest to us is why uation which promotes egalitarianism. Frequent
people high in SDO would try actively to avoid exposure to such an egalitarian microenviron-
intergroup contact? According to Duckitt (2001), ment may gradually lead to the internalization of
SDO is, unlike right-wing authoritarianism, less egalitarian norms by the interaction partners,
fueled by threat-driven motives and negative which attenuates SDO levels (Haley & Sidanius,
emotions of fear and anxiety which are psycho- 2005). In sum, combining insights from inter-
logical processes that are typically related to group contact research (e.g., Hewstone, 2009)
avoidance tendencies (e.g., Mackie, Devos, & with theorizing on SDO (e.g., Duckitt, 2001;
Smith, 2000). Instead, SDO is primarily related to Haley & Sidanius, 2005; Pratto et al., 2006), it can
dominance-driven motives, coldness, and a lack be expected that positive intergroup contact
of positive emotions and empathy (Duckitt, decreases SDO. Furthermore, a combination of
2001; Lippa & Arad, 1999; Pratto et al., 1994), frequent and high-quality contact is likely to have
psychological processes which do not seem to be the strongest effect on SDO.
associated with avoidant behavior in a straight-
forward manner.
Moreover, the relative stability of SDO over The Current Research
time does not preclude that SDO is influenced by Van Laar, Levin, Sinclair, and Sidanius (2005)
social experiences (Pratto et al., 2006; Sidanius & reported the first direct evidence that positive
Pratto, 1999). Positive intergroup contact likely intergroup contact has an effect on SDO. Based
represents such an influential social experience on a large cohort sample of college students,
that might impact upon SDO, given its beneficial these authors investigated the effect of ethnic
effects on implicit group associations, attitude heterogeneity of students’ roommates during
strength, outgroup trust, forgiveness, and sup- college years on a range of outcome variables in
port for positive outgroup-targeted policies (see the last year, including intergroup affect, sym-
Hewstone, 2009; Pettigrew & Tropp, 2011). bolic racism, antimiscegenation attitudes, inter-
Furthermore, the impact of positive intergroup group unease and competence, and SDO (see
contact also generalizes to reduced prejudice Sidanius, Levin, van Laar, & Sears, 2009). Ethnic

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

Dhont et al. 31

roommate heterogeneity was found to have a 2009; Pedersen, 2009; Pizam, Fleischer, &
beneficial effect on several of the outcome vari- Mansfeld, 2002; Schild, 1962), although not in
ables, including SDO (see van Laar et al., 2005, the context of intractable conflict (e.g., Milman,
Table 3). Given the focus of these authors on the Reichel, & Pizam, 1990). In line with these previ-
reduction of prejudice, this initial finding of a ous findings, the contact-based intervention pro-
contact effect on SDO was not discussed in its gram investigated in the current study has been
own right and has therefore remained relatively demonstrated to successfully reduce negative
unnoted by researchers. From these initial results, attitudes towards Moroccan immigrants (Dhont,
however, it can thus be inferred that SDO levels Roets, & van Hiel, 2011, Study 3). In the current
can indeed be reduced, at least by high-quality, study, based on newly collected data, we investi-
long-term contact. gated whether contact with Moroccans during a
The aim of the present studies was to further 1-week school trip abroad would also affect ado-
investigate the potential of intergroup contact to lescents’ levels of SDO.
decrease SDO. Such research is needed because Our second study, a longitudinal survey, aimed
the characteristics of the intergroup context to cross-validate the findings of Study 1 in a het-
studied by van Laar et al. (2005) were highly spe- erogeneous sample of adults who reported on
cific and by no means representative for inter- their levels of SDO and daily contact experiences
group contact as it generally occurs. Indeed, given with Muslim immigrants of Moroccan or Turkish
the room-sharing context in van Laar et al.’s origin. Moreover, by using a cross-lagged panel
(2005) study, the members of different ethnic design in Study 2, we were able to simultaneously
groups were closely connected to each other investigate the longitudinal effect of intergroup
because they needed, as the authors described, to contact on SDO and the longitudinal effect of
maintain a mutually satisfactory home environ- SDO on intergroup contact, while controlling for
ment, were likely to share a wide range of activi- the stability of both variables over time and the
ties, and were supported by a university setting cross-sectional associations between the variables
which promotes egalitarian norms (van Laar (see, Christ & Wagner, 2013).
et al., 2005). These unique conditions may, how-
ever, rarely occur in other intergroup contexts.
The question thus remains whether relatively Study 1
short-term contact interventions or daily experi-
ences of positive intergroup contact in other Method
contexts would yield similar results. To address Participants and procedure. We recruited 71 students
this question, we first conducted a short-term without migration background (79% female,
contact intervention study (Study 1) among Mage = 16.92, SDage = .91) from three Belgian
Belgian high school students who went on a high schools. All three schools are located in
1-week school trip to Morocco. Along with peo- Flanders, Belgium. Two schools provide general
ple of Turkish descent, the Moroccan commu- educational tracks, whereas one school provides
nity constitutes the largest Muslim minority technical and vocational educational tracks.
group in Belgium. Together, the two groups rep- Because the school principals and class teachers
resent 11.5% of the immigrant population had chosen to participate in the program and all
(approximately 10% of the total Belgian popula- students of these classes were involved, self-
tion have an immigrant background). selection effects were eliminated.
Short-term visits abroad that stimulate posi- The goal of the school trip was to become
tive intergroup contact have been shown to acquainted with Moroccan students as well as
increase favorable outgroup attitudes, to reduce with their school, religion, and way of life. The
ethnocentrism, and to improve intercultural sen- visiting Belgian and host Moroccan students
sitivity (e.g., Amir, 1969; Church, 1982; Jackson, spent much time together, being involved in joint

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

32 Group Processes & Intergroup Relations 17(1)

activities, such as hiking, sightseeing, and visiting SD = 0.92, α = .77) were: “How often did you
the families of the Moroccan students. They also have contact with Moroccans during the trip?”,
needed to work together on assigned school tasks “How often did you have a conversation with
(e.g., preparing and conducting an interview of a Moroccans during the trip?”, and “How often did
local resident) and organized activities during lei- you work with Moroccans?” (1, never; 7, very fre-
sure time. One week prior to the intervention, quently). The contact quality measure (M = 5.54,
participants completed a pretest questionnaire in SD = 0.95, α = .75) started with the question
their classrooms during school hours. A posttest “How would you characterize your contact with
questionnaire was completed within 2 weeks of the Moroccan population?” followed by four
the students’ returning from Morocco. adjectives: (a) pleasant, (b) superficial (reverse
Both the pre- and posttest questionnaires coded), (c) annoying (reverse coded), and (d)
included measures of SDO and prejudice. Because enjoyable (1, certainly not; 7, very certainly).
we did not have access to a comparable group of
students who could have reasonably served as a
control group, the posttest questionnaire included Results and Discussion
measures of quantity and quality of intergroup First, we investigated whether there were signifi-
contact during the intervention. As such, we were cant differences in SDO and anti-immigrant prej-
able to check whether changes in SDO and preju- udice before and after the contact intervention.
dice could be attributed to the contact experiences We therefore performed a one-way, within-
during the intervention. In other words, we could subjects (pretest vs. posttest) multivariate analysis
investigate whether students reporting more or of variance (MANOVA) with SDO and prejudice
better contact experiences would show greater as the dependent variables. In line with our expec-
effects of the intervention. tations, this analysis showed a significant multi-
variate difference between the scores on the
Measures. In the pretest and posttest, we admin- pretest and the posttest, F(2, 69) = 9.59, p < .001,
istered the 14-item SDO scale (Pratto et al., partial η2 = .22. The univariate test results demon-
1994, translated into Dutch by van Hiel & strated that both SDO and prejudice posttest
Duriez, 2002) on a 5-point Likert scale (1, strongly scores were significantly lower (MSDO = 2.21 and
disagree; 5, strongly agree). Sample items are “Some Mprejudice = 2.81) than the pretest scores (MSDO =
groups of people are simply not the equals of 2.39 and Mprejudice = 2.94), F(1, 70) = 13.56, p <
others” and “It’s sometimes necessary to step on .001, partial η2 = .16, and F(1, 70) = 9.03, p < .005,
others to get ahead in life” (pretest, α = .82; post- partial η2 = .11, respectively. These results thus
test, α = .88). demonstrate an overall decrease in SDO and prej-
At both pretest and posttest, we also measured udice against Moroccans after the intervention.
prejudice towards Moroccans with an adapted We hypothesized that this intervention
10-item modern racism scale (McConahay, 1986; effect would be attributable to the students’
translated into Dutch and adapted by Dhont, contact experiences with Moroccans during the
Cornelis, & van Hiel, 2010) rated on a 5-point trip. Or, in other words, we expected that the
Likert scale (1, strongly disagree; 5, strongly agree). A quality and quantity of intergroup contact
sample item is “Moroccans are getting too would moderate the intervention effect. To test
demanding in their push for equal rights” (pretest, this moderation effect, we followed the analyti-
α = .72; posttest, α = .71). cal recommendations of Judd, Kenny, and
Furthermore, the posttest questionnaire McClelland (2001) for testing moderation in
included scales measuring the quantity and quality within-subject designs using ordinary least
of contact with Moroccans during the school trip squares (OLS) regression analyses. In the cur-
(based on Dhont & van Hiel, 2011; Voci & rent design, moderation is demonstrated when
Hewstone, 2003). The quantity items (M = 5.80, contact quantity or quality significantly

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

Dhont et al. 33

predicts the difference between the posttest quality relative to quantity in the prediction of
and pretest scores of SDO and prejudice, that intergroup attitudes (see R. Brown & Hewstone,
is, the intervention effect. Accordingly, we cal- 2005). The limited role of contact quantity in the
culated a difference score for both dependent present study might, however, also be attributed
variables by subtracting the pretest scores from to the fact that all respondents had experienced a
the posttest scores. Hence, a higher difference considerable amount of intergroup contact
score of SDO or prejudice indicated a stronger because they were all involved in a contact inter-
intervention effect and thus a stronger decrease vention program. Such regulated context charac-
in SDO or prejudice, respectively. terized by a high prevalence of intergroup contact
In a first regression analysis, we regressed the may have reduced the predictive value of inter-
SDO difference scores on both contact quantity group contact quantity.
and quality. The results showed that contact
quality but not contact quantity significantly pre-
Study 2
dicted the SDO difference score, β = .31, p = .01
and β = .05, p = .70, respectively, R² = .11. Although the findings of Study 1 indicated that
Similarly, a second regression analysis with the intergroup contact is able to reduce SDO, these
difference scores of prejudice also revealed that findings do not clarify whether the negative asso-
contact quality, but not contact quantity, signifi- ciation between intergroup contact and SDO
cantly predicted the prejudice difference scores, reported in previous cross-sectional studies
β = .28, p = .03 and β = −.10, p = .43, respec- (Asbrock et al., 2012; Dhont & van Hiel, 2009;
tively, R² = .07. Hodson, 2008) is solely unidirectional or bidirec-
To summarize, the results of Study 1 demon- tional. Moreover, Pettigrew and Tropp’s (2006)
strated that a short-term period of intensive meta-analysis has shown that intergroup contact
intergroup contact reduced the levels of both effects are strongest in confined settings where
SDO and prejudice against Moroccans. However, participants cannot avoid intergroup contact.
an important limitation of the design of Study 1 Hence, the question arises whether the effects of
is that we were not able to compare the interven- positive contact on SDO would still emerge in
tion group with a control group (e.g., students of the context of daily intergroup interactions,
the same age from the same school who did not where people can avoid intergroup contact (see
travel to Morocco). As such, we cannot rule out also Dhont & van Hiel, 2009; Hodson, Harry, &
completely that a testing effect influenced our Mitchell, 2009).
results (see Shadish, Cook, & Campbell, 2001). To address this issue, we conducted a cross-
For instance, respondents may have talked about lagged longitudinal survey study in a heterogene-
the measures during the intervention, potentially ous sample of adults. In this second study, we
resulting in more socially desirable responses in focused on the frequency of positive contact
the posttest.1 experiences with immigrants of Turkish and
Despite this limitation, we were able to dem- Moroccan descent. Thus, rather than making a
onstrate that the intergroup contact experiences distinction between contact quantity and quality,
during the intervention were indeed significantly we focused on the effects of a combined index
related to changes in SDO, which increases our of both contact quantity and quality. It was rea-
confidence in the validity of our findings. In par- sonable to assume that the nonsignificant role of
ticular, the higher the contact quality, but not the mere quantity of intergroup contact in Study
quantity, was, the stronger the effect of the inter- 1 was caused by the highly regulated intergroup
vention was on both SDO and prejudice. The context. In less or nonregulated intergroup con-
superior role of contact quality compared to con- texts, however, having both frequent and positive
tact quantity is consistent with several studies that intergroup contact has been considered to result
have demonstrated the stronger value of contact in the most powerful effects on intergroup

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

34 Group Processes & Intergroup Relations 17(1)

variables (e.g., Davies et al., 2011; Dhont et al., 7, very frequently). A sample item is “How often do
2012; Swart et al., 2011). you have pleasant contact with immigrants?”
An effect of positive intergroup contact on (Time 1, α = .94; Time 2, α = .96). A higher score
SDO can be demonstrated if contact at Time 1 thus indicated more frequent positive intergroup
affects SDO at Time 2 after controlling for (a) the contact.
stability of both these variables over time, that is, Participants also completed the same 14-item
the autoregressive paths, and (b) the indirect SDO scale as administered in Study 1 and a nine-
effects from the covariation of the two variables item anti-immigrant prejudice scale (based on
and the autoregressive path (Christ & Wagner, Billiet & De Witte, 1991) rated on a 5-point Likert
2013; Dhont et al., 2012). Such a design allows scale (1, strongly disagree; 5, strongly agree). A sample
for simultaneously testing the cross-lagged effect item of the prejudice scale is “Belgium should
of contact at Time 1 on SDO at Time 2 and the never have allowed immigrants into the country.”
effect of SDO at Time 1 on contact at Time 2. Both scales were internally consistent at Time 1
(SDO, α = .86; prejudice, α = .90) and Time 2
(SDO, α = .90; prejudice, α = .91).
Method
Participants and procedure. A heterogeneous sam-
Results and Discussion
ple of 363 Belgian adults was recruited at the
start of the academic year (Time 1) by under- Preliminary analyses. We first conducted a MANOVA
graduate psychology students. The students on the three variables under study (intergroup con-
were asked to recruit one non-Muslim adult tact, SDO, and prejudice) to determine whether the
without migration background from their par- respondents who completed the questionnaire at
ents’ social network or from their neighbor- both time points differed significantly from the
hood. Participants received an envelope that respondents who dropped out after Time 1. The
included a questionnaire and a letter of informed results of this analysis revealed neither a significant
consent explaining the survey procedure and the multivariate difference between both groups, F(3,
participants’ rights. The study was introduced as 359) = .97, p = .41, nor a significant univariate dif-
an investigation of attitudes and beliefs about ference between the groups for any of these varia-
societal topics. To assure confidentiality, ques- bles, all Fs < 2.34, ps > .12. Based on these results,
tionnaires were returned in a closed envelope. we can conclude that selective attrition did not play
All respondents from Time 1 were contacted a significant role in subsequent findings and miss-
again approximately three months later (Time 2) ing data are missing at random. All respondents
with the request to complete a second question- who participated at Time 1 were therefore included
naire. A total of 92% of the initial sample (NTime 2 in the subsequent analyses, and missing values were
= 333) participated at Time 2. estimated using maximum likelihood estimation
Participants (71% female) were between 19 with the expectation maximization algorithm.
and 64 years old, with a mean age of 48.29 years Applying maximum likelihood procedures has
(SD = 5.09). Of this sample, 18% had attended been shown to produce more reliable parameter
university, 46% had completed nonuniversity estimates and standard errors as compared to con-
higher education, 26% had completed secondary ventional methods of dealing with missing data, for
school, and 10% had completed only basic example, pair-wise or list-wise deletion (Schafer,
education. 1997; Schafer & Graham, 2002; see also Al Ramiah
& Hewstone, 2012; Swart et al., 2011).
Measures. Intergroup contact was assessed with Table 1 presents the descriptive statistics and
four questions about the frequency of positive the correlations among the measures within and
contact with immigrants (see also Dhont et al., across time points. The cross-sectional and cross-
2011), answered on a 7-point Likert scale (1, never; lagged correlations showed that positive

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

Dhont et al. 35

Table 1. Descriptive statistics and correlations between the variables in Study 2 at Time 1 (T1) and Time 2 (T2).

Mean SD Contact SDO Prejudice

T1 T2 T1 T2 T1 T2
*** ** *** ***
Contact T1 3.23 1.58 .63 −.14 −.26 −.38 −.40***
T2 3.01 1.72 −.12* −.23*** −.35*** −.37***
SDO T1 2.17 0.65 .68*** .42*** .43***
T2 2.20 0.64 .49*** .58***
Prejudice T1 2.17 0.78 .80***
T2 2.23 0.78
Note. *p < .05; **p < .01; ***p < .001.

intergroup contact was negatively related to SDO on the relationships between intergroup contact
and prejudice within and across time points, and SDO (without including prejudice). How-
whereas SDO and prejudice were positively inter- ever, before testing these longitudinal relation-
related within and across time points. ships, it was necessary (a) to test the fit of
the longitudinal measurement model to investi-
Longitudinal analyses with latent constructs. To test gate the factorial validity and construct independ-
the hypothesized longitudinal effects, we used ence of the latent constructs and (b) to investigate
structural equation modeling (SEM) with latent whether the measurement properties of the con-
variables using Lisrel (Jöreskog & Sorbom, 2004). tact and SDO measures could be considered
To smooth measurement error and to maintain invariant over time (Byrne, Shavelson, & Muthén,
an adequate ratio of cases to parameters, we aver- 1989; Little, Preacher, Selig, & Card, 2007; Mere-
aged subsets of items into indicator parcels (Lit- dith, 1993; see also Swart et al., 2011).
tle, Cunningham, Shahar, & Widaman, 2002), Therefore, we first tested a model including
which were held constant over time. Two parcels the latent factors and accompanying indicators
were created for the latent factor of the frequency of SDO and contact from each time point with
of positive intergroup contact, and three parcels freely estimated parameters. This longitudinal
were created for the latent factors of SDO and measurement model showed good model fit,
prejudice.2 The first factor loading of each latent χ²(24) = 36.04, p = .05; CFI = 1.00; RMSEA =
variable was set to unity in order to scale the fac- .037; SRMR = .026.3 Next, we compared this
tors and the residual errors of parallel indicators unrestrictive longitudinal model with a second
were allowed to correlate in all analyses, reflecting model in which factor loadings of correspond-
stability in systematic error over time. The chi- ing indicators across time were constrained to be
square test statistic (χ²), the comparative fit index invariant, that is, a longitudinal metric invariance
(CFI), the root-mean-square error of approxima- model (cf. T. A. Brown, 2006; Christ & Wagner,
tion (RMSEA), and the standardized root- 2013; Swart et al., 2011) to establish longitudinal
mean-square residual (SRMR) were used to measurement invariance (MI). The restrictions
evaluate the goodness-of-fit of the tested models. imposed in this second model did not result in a
A satisfactory fit is indicated by a CFI value significantly worse fit compared to the less
greater than .95, an RMSEA value close to .06, an restricted model (with freely estimated parame-
SRMR close to .08 (Hu & Bentler, 1999), and a ters), ∆χ²(3) = 2.10, p = .55, confirming metric
χ²/df ratio smaller than 3 (Kline, 2010). MI over time. Longitudinal metric invariance
(i.e., partial MI) is considered sufficient for pur-
Longitudinal relationships between intergroup contact and poses of testing and comparing latent models
SDO. In a first series of model tests, we focused (Byrne et al., 1989).

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

36 Group Processes & Intergroup Relations 17(1)

Time 1 Time 2

Intergroup .70*** Intergroup

Contact Contact
−.17***
−.05
−.15***

−.04
SDO .64*** SDO

Figure 1. Cross-lagged model testing the longitudinal relationships between intergroup contact and SDO in
Study 2.
Note. Presented values are standardized coefficients. The error term for each indicator at Time 1 was allowed to correlate with
the error term for the same indicator at Time 2.
**
p < .01; ***p < .001.

Having established satisfactory measurement that people scoring high on SDO would avoid
invariance for the latent factors of contact and intergroup contact.5
SDO, we simultaneously investigated the effects
of contact and SDO at Time 1 on contact and Longitudinal relationships between intergroup contact,
SDO at Time 2. This full cross-lagged model thus SDO, and anti-immigrant prejudice. In a second series
included all paths from contact and SDO at Time of model tests, we also included the latent factors
1 to contact and SDO at Time 2 (i.e., the autore- of prejudice at Time 1 and Time 2 in addition to
gressive and cross-lagged paths) while controlling the Time 1 and Time 2 factors of intergroup con-
for the within-time associations between the two tact and SDO. As such, we wanted to rule out the
variables, that is, the latent variables were allowed possibility that the obtained longitudinal effect of
to be correlated at Time 1, whereas the latent contact on SDO emerged solely because contact
variable residuals (the disturbance terms) were with immigrants and SDO are both related to anti-
allowed to be correlated at Time 2. Figure 1 immigrant prejudice. A longitudinal measurement
depicts the results (i.e., standardized estimates) of model including the Time 1 and Time 2 factors of
this analysis, χ²(27) = 38.14, p = .08; CFI = 1.00; intergroup contact, SDO, and prejudice with freely
RMSEA = .034; SRMR = .026.4 In line with our estimated parameters showed a good model fit,
hypotheses, intergroup contact had a significant χ²(81) = 128.26, p < .001; CFI = .99; RMSEA =
negative longitudinal effect on SDO, β = −.17, p .040; SRMR = .038.6 Again, restricting this model
< .001, whereas no significant longitudinal effect by constraining the factor loadings of correspond-
of SDO on contact was found, β = −.04, ns. ing indicators to be invariant across time did not
Moreover, constraining the path from intergroup result in a significantly worse fit, ∆χ²(5) = 6.84, p =
contact at Time 1 to SDO at Time 2 to 0, resulted .23, confirming metric MI over time.
in a significantly worse model fit, ∆χ²(1) = 16.54, Then, we tested a full cross-lagged model that
p < .001. In contrast, constraining the path from enabled us to simultaneously investigate the lon-
SDO at Time 1 to intergroup contact at Time 2 gitudinal effects of contact, SDO, and prejudice
had no significant impact on the model fit, ∆χ²(1) at Time 1 on contact, SDO, and prejudice at Time
= 1.08, p = .30. Hence, the results of Study 2 2. As with the previous cross-lagged model with-
demonstrated that intergroup contact decreases out prejudice, we controlled for the stability
SDO, whereas no support was found for the idea effects of the three factors over time (i.e., the

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

Dhont et al. 37

Time 1 Time 2

Intergroup .59*** Intergroup

contact *
contact
-.08
-.07*
-.01
-.15**

.03 -.02
-.42*** SDO .59*** SDO
.09*
**
.45 .11**

-.15**
.22***
Prejudice Prejudice
.78***

Figure 2. Cross-lagged model testing the longitudinal relationships between intergroup contact, SDO, and
prejudice in Study 2.
Note. Presented values are standardized coefficients. The error term for each indicator at Time 1 was allowed to correlate with
the error term for the same indicator at Time 2.
*
p < .05; **p < .01; ***p < .001.

autoregressive effects) as well as all cross- stronger, β = .22, p < .001, than the effect of
sectional associations between the three variables SDO on prejudice, β = .09, p < .05.8
within each wave. The results of this model test The results of this second study, using a het-
showed that the model had adequate fit, χ²(86) = erogeneous adult sample, extended the findings
135.10, p < .001; CFI = .99; RMSEA = .040; of the short-term contact intervention study
SRMR = .038.7 The results (i.e., standardized esti- (Study 1) and indicated that intergroup contact is
mates) of this model test are presented in able to reduce SDO over a 3-month period. More
Figure 2. The longitudinal negative effect of con- specifically, more frequent positive intergroup
tact on SDO remained significant, β = −.08, p < contact at Time 1 was significantly associated
.05, even after including, and thus controlling for, with lower SDO levels at Time 2. We did not,
the variance explained by the factors representing however, obtain evidence for the idea that people
prejudice at Time 1 and Time 2. The longitudinal high in SDO would engage less in intergroup
effect of SDO on contact was still nonsignificant, contact, since SDO at Time 1 was not signifi-
β = .03, ns. Furthermore, the results showed that cantly related to contact levels at Time 2. This
contact had a significant longitudinal negative nonsignificant effect of SDO on contact cannot
effect on prejudice, β = −.07, p < .05, but also be attributed to the general absence, in this par-
that prejudice had a significant longitudinal nega- ticular sample, of a tendency to avoid contact.
tive effect on contact, β = −.15, p < .01. A final Indeed, despite the absence of a bidirectional
noteworthy result is that SDO and prejudice effect between SDO and contact, the relationship
mutually influenced each other over time, between prejudice and intergroup contact was
although the effect of prejudice on SDO was found to be bidirectional. The results showed

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

38 Group Processes & Intergroup Relations 17(1)

that intergroup contact leads to lower prejudice the school context and in applied settings (see
levels, but also that higher prejudice levels are also Church, 1982; Stephan & Vogt, 2004), and
related to less intergroup contact over time. In therefore clarify the value of intergroup contact
fact, the latter path, that is, from prejudice to as an intervention strategy to counter group
intergroup contact, was even stronger than the dominance and inequality beliefs.
former, that is, from intergroup contact to Nevertheless, both the high school students in
prejudice. the contact intervention of Study 1 as well as the
university students with close outgroup room-
mates (van Laar et al., 2005) were located in a set-
General Discussion ting characterized by strong situational
Despite almost two decades of research on social constraints, likely maximizing the effects of inter-
dominance orientation, relatively little is known group contact (Pettigrew & Tropp, 2006), and
about situational factors that may affect SDO. particularly the effects of quality of intergroup
Some authors have focused on the influence of contact. Such intergroup contexts might not cor-
changes in contextual factors and individuals’ respond to every-day life, where people are free
competitive world views that may increase SDO to avoid or to engage in intergroup contact.
(e.g., Huang & Liu, 2005; Sibley et al., 2007b) but Therefore, in Study 2, we adopted a cross-
the question whether individual contextual expe- lagged panel design and tested the longitudinal
riences may decrease SDO has not yet been well effects of self-reported levels of frequency of
researched. The current research addressed this positive intergroup contact as they occur in
lacuna by focusing on the attenuating effect of respondents’ daily life. This cross-lagged approach
positive intergroup contact on SDO. Consistent and the focus on daily intergroup experiences ena-
with our expectations, a pretest–posttest inter- bled us to simultaneously test whether positive
vention study among high school students and a intergroup contact reduces SDO, as well as
cross-lagged longitudinal survey study using a whether people high in SDO tend to avoid inter-
heterogeneous adult sample revealed that posi- group contact, while controlling for the stability
tive contact with ethnic outgroup members effects of contact and SDO over time and the
reduces SDO. cross-sectional associations between contact and
Study 1 demonstrated that a contact interven- SDO. This rigorous longitudinal test provided
tion lowered school students’ SDO levels. In the further evidence for the effect of intergroup
absence of a no-contact control condition, we contact on SDO, an effect that remained signifi-
could show that a marked decrease in SDO was cant after controlling for participants’ prejudice
obtained especially among school students scores.
reporting higher quality contact during the inter- The suggestion that people high in SDO
vention. These results corroborated the findings would avoid intergroup contact was not con-
obtained by van Laar et al. (2005) who reported firmed, as demonstrated by the nonsignificant
that ethnic roommate heterogeneity at a college longitudinal path from SDO to intergroup con-
campus longitudinally predicted reduced SDO tact. A self-selection effect did, however, emerge
over several years. The particular type of inter- for prejudice, indicating that highly prejudiced
group contact studied by van Laar et al. (2005), people tend to avoid intergroup contact (see also
roommate contact at college, is expected to be of Binder et al., 2009; Swart et al., 2011; for an
high quality, yet is likely to be difficult to engineer exception, see Dhont et al., 2012). This does not,
in other situations. The present results clarified however, indicate that contact is somehow less
that a short-term contact intervention of 1 week important than prejudice. The significant effect
is also able to lower SDO levels. Short-term con- from contact to reduced prejudice over time rein-
tact interventions like the one studied in the cur- forces the view that contact can be a part of
rent research are relatively easy to implement in interventions aimed at reducing prejudice. The

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

Dhont et al. 39

present results added the demonstration that the are highly salient groups that are debated in the
tendency for prejudiced people to avoid contact media and among politicians (see also Billiet &
is not driven by their SDO levels. De Witte, 2008; Coenders, Lubbers, Scheepers, &
Overall, the results of the current studies are Verkuyten, 2008). It can therefore reasonably be
encouraging with respect to the potential of argued that positive personal contact with mem-
intergroup contact to change people’s attitudes bers of such a highly salient group will not only
towards group dominance and inequality. lead to more positive attitudes towards this par-
Nevertheless, it should be noted that the opera- ticular outgroup, but also represents a relevant
tionalization of intergroup contact employed in and influential experience on which people rely to
Study 2 is more common in intergroup contact shape or adjust their attitudes about intergroup
research than the methods used in Study 1. relations and inequality “in general.”
Indeed, in Study 2, we focused on participants’ The effect of intergroup contact on SDO
contact experiences with a minority group in reminds us of the secondary transfer effect of
their own country, whereas in Study 1 partici- intergroup contact, which refers to the finding
pants went to the country of the target outgroup that contact with members of one outgroup also
where they were the numerical minority during leads to more positive attitudes towards other,
their stay abroad. The different perspectives in “secondary” outgroups that were not involved in
these two contexts raise the question of whether the contact situation (Pettigrew, 2009; Schmid
similar processes were operating in the two con- et al., 2012; Tausch et al., 2010). Instead of focus-
texts. At the same time, these different contexts ing on attitudes towards specific secondary out-
provide a cross-validation of the effects, attesting groups, however, we focused on a generalized
to their robustness. Indeed, we obtained consist- orientation that represent an important, social
ent evidence for the effect of positive intergroup attitudinal basis of attitudes towards all kinds of
contact on SDO using different methods and secondary outgroups (see Kteily et al., 2012;
research designs (i.e., a pretest–posttest interven- Sibley & Liu, 2010).
tion abroad vs. a longitudinal survey study), which Studies by both Schmid et al. (2012) and
were applied in different samples (high school Tausch et al. (2010) further indicated that the
students vs. a community sample). The consistent generalized effects of contact on attitudes
findings increase the confidence in the validity of towards secondary outgroups operate through an
our findings. attitude generalization process. That is, attitudes
We now elaborate on how these findings can towards the primary outgroup mediate the effect
be integrated within social dominance theory and of contact with that outgroup on attitudes
relate them to some other recent findings from towards secondary outgroups. Based on these
intergroup contact research. We also point to findings, we may expect that intergroup contact
promising pathways for future research and high- first has an effect on prejudice and, then, further
light the implications of our findings. generalizes to SDO. This explanation is consist-
ent with the present findings showing cross-
Bridging Social Dominance Theory lagged effects of contact on prejudice and SDO
as well as a cross-lagged effect of prejudice on
and Intergroup Contact Theory SDO. The design of the present studies did not,
Sidanius and Pratto (1999) argued that striving however, allow for an appropriate test of longitu-
for dominance is likely to be strongest with dinal mediation; future studies collecting longitu-
respect to groups that are most salient in society, dinal data across three waves (e.g., Swart et al.,
while such striving is weaker with respect to 2011) are needed to clarify this issue.
groups that do not elicit much attention. In The work of Duckitt (2001) offers another
Western European countries like Belgium, immi- plausible explanation worthy of further investiga-
grants with a Moroccan or Turkish background tion. According to Duckitt’s model (Duckitt,

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

40 Group Processes & Intergroup Relations 17(1)

2001; Perry et al., 2013), SDO is driven by sche- From these latter findings, it may be expected
matic beliefs and perceptions about the social that positive intergroup contact may be positively,
world as a competitive and cut-throat place ver- rather than negatively, related to SDO among
sus a cooperative place characterized by mutually minority group members. In sum, the extension
beneficial exchange. These schematic social of the present findings to the minority perspec-
worldview beliefs, in turn, vary as a function of tive surely represents a challenging and exciting
changes in the social environment and socializa- field of future inquiry.
tion experiences. In particular, exposure to social
situations characterized by high levels of inequal-
ity and competition increases competitive world- Conclusion
views, whereas exposure to an egalitarian social The main focus on SDO as a generalized orienta-
context characterized by cooperation—as in the tion towards group-based inequality and hierarchy
case of positive intergroup contact—should (Kteily et al., 2012; Pratto et al., 1994) and as the
decrease these worldviews and lead to the inter- basis of intergroup attitudes diverted research
nalization of egalitarian norms. As a result, a attention from the questions of which situational
decrease in SDO can be expected (see Haley & variables may influence SDO, and ultimately how
Sidanius, 2005). However, because we did not SDO can be changed. Nevertheless, the identifica-
measure either perceived intergroup norms or tion of situational factors that impact on SDO is
competitive worldviews in the current studies, an important topic, given that such findings may
future studies may further clarify their role as lead to the development of techniques to attenu-
intermediary processes. ate SDO. Intergroup contact (see Hewstone,
Future research may also extend the current 2009; Pettigrew & Tropp, 2011) holds promise as
findings by studying the effects of positive inter- a situational influence on SDO, as shown by our
group contact on SDO from a minority perspec- two studies. These findings have potentially far-
tive. It is unlikely that the current results will reaching implications because intergroup contact
generalize in a straightforward manner to mem- may attack some of the ideological roots of preju-
bers of minority status groups. Indeed, being well dice and thus indirectly affect the range of varia-
aware of their lower status, minority members bles that are predicted by SDO.
tend to approach and experience an intergroup
contact situation in a different way than members Funding
of majority groups because they likely anticipate This research was supported by a postdoctoral research
prejudice and discrimination against them from grant from the Research Foundation – Flanders (FWO,
dominant group members (Saguy, Tropp, & Belgium) awarded to Kristof Dhont.
Hawi, 2013; Tropp, 2006). As such, the effect of
positive intergroup contact on reduced prejudice Notes
is typically weaker for minority than for majority 1. If the students completed the SDO and preju-
group members. Allport’s optimal conditions dice measures differently after the intervention
(e.g., equal status and cooperation) have also been because they had discussed the measures during
found not to facilitate the effects of intergroup the intervention, this should also be reflected in
contact among minority members (Tropp & a higher intraclass correlation (ICC) between the
SDO and prejudice scores at Time 2 compared to
Pettigrew, 2005). Furthermore, recent research
Time 1. This was, however, not the case, ICC =
has indicated that positive contact with majority .54 at Time 1 and ICC = .51 at Time 2.
members may also attract minority members’ 2. Details of the items included in each parcel can
attention away from group inequality and decrease be obtained from the first author on request.
their efforts to challenge the status quo of group- 3. Tests of the cross-sectional measurement models at
based inequality and injustice (Dixon, Tropp, Time 1 and Time 2 with the latent factors of contact
Durrheim, & Tredoux, 2010; Saguy et al., 2013). and SDO also yielded satisfactory model fits, χ²(4) =

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

Dhont et al. 41

10.50, p = .03; CFI = .99; RMSEA = .067; SRMR = Asbrock, F., Christ, O., Duckitt, J., & Sibley, C. G.
.028 for Time 1 and χ²(4) = 1.24, p = .87; CFI = 1.00; (2012). Differential effects of intergroup con-
RMSEA = .00; SRMR = .006 for Time 2. tact for authoritarians and social dominators: A
4. The model fit of this same model but without dual process model perspective. Personality and So-
correlating the latent variable residuals at Time 2 cial Psychology Bulletin, 38, 477–490. doi:10.1177/
was, χ²(28) = 40.04, p = .07; CFI = 1.00; RMSEA 0146167211429747
= .034; SRMR = .028. Asbrock, F., Sibley, C. G., & Duckitt, J. (2010). Right-
5. We also tested our model for the participants who wing authoritarianism and social dominance orien-
provided data at both Time 1 and Time 2. These tation and the dimensions of generalized prejudice:
analyses yielded analogous results. The model A longitudinal test. European Journal of Personality,
had a very good fit, χ²(27) = 30.84, p = .28; CFI 24, 324–340. doi:10.1002/per.746
= 1.00; RMSEA = .021; SRMR = .024, and we Billiet, J., & De Witte, H. (1991). Naar racisme neigende
obtained virtually identical estimates of all paths houdingen in Vlaanderen: Typologie en maatschappelijke
to the ones presented in Figure 1, that is, with achtergronden [Attitudes towards racism in Flanders:
cross-lagged paths of contact on SDO, β = −.17, Typology and societal backgrounds]. Leuven, Bel-
and of SDO on contact, β = −.05. gium: Sociologisch Onderzoeksinstituut (SOI).
6. The tests of the cross-sectional measurement Billiet, J., & De Witte, H. (2008). Everyday racism as
models at Time 1 and Time 2 with the inclusion predictor of political racism in Flemish Belgium.
of the latent factor of prejudice also yielded sat- Journal of Social Issues, 64, 253–567. doi:10.1111/
isfactory model fits, χ²(17) = 36.59, p = .004; CFI j.1540-4560.2008.00560.x
= .99; RMSEA = .056; SRMR = .041 for Time 1 Binder, J., Zagefka, H., Brown, R., Funke, F., Kes-
and χ²(17) = 32.13, p = .014; CFI = 1.00; RMSEA sler, T., Mummendey, A., & …Leyens, J.-P. (2009).
= .00; SRMR = .026 for Time 2. Does contact reduce prejudice or does prejudice
7. The model fit of this same model but without reduce contact? A longitudinal test of the contact
correlating the latent variable residuals at Time hypothesis among majority and minority groups
2 was, χ²(89) = 161.94, p < .001; CFI = .99; in three European countries. Journal of Personal-
RMSEA = .048; SRMR = .045. ity and Social Psychology, 96, 843–856. doi:10.1037/
8. Again, testing this model exclusively on the par- a0013470
ticipants who provided data on the two measure- Bretz, R. D., & Judge, T. A. (1994). Person–organiza-
ment points yielded a good fit, χ²(86) = 118.32, tion fit and the theory of work adjustment: Impli-
p = .01; CFI = .99; RMSEA = .034; SRMR = cations for satisfaction, tenure, and career success.
.037, and analogous estimates of the paths as Journal of Vocational Behavior, 44, 32–54. ). http://
those presented in Figure 2, that is, for the effect dx.doi.org/10.1006/jvbe.1994.1003
of contact T1 on SDO T2 and prejudice T2, β = Brown, R., & Hewstone, M. (2005). An integrative the-
−.08 and β = −.07, respectively, for the effect of ory of intergroup contact. Advances in Experimental
SDO T1 on contact T2 and prejudice T2, β = .02 Social Psychology, 37, 255–343. doi:10.1016/S0065-
and β = −.08, respectively, and for the effect of 2601(05)37005-5
prejudice T1 on SDO T2 and contact T2, β = .22 Brown, T. A. (2006). Confirmatory factor analysis for applied
and β = −.15, respectively. research. New York, NY: Guilford Press.
Byrne, B. M., Shavelson, R. J., & Muthén, B. (1989).
References Testing for the equivalence of factor covariance
and mean structures: The issue of partial measure-
Allport, G. W. (1954). The nature of prejudice. Reading, ment invariance. Psychological Bulletin, 105, 456–466.
MA: Addison-Wesley. doi:10.1037/0033-2909.105.3.456
Al Ramiah, A., & Hewstone, M. (2012). “Rallying Christ, O., & Wagner, U. (2013). Methodological issues
around the flag”: Can an intergroup contact inter- in the study of intergroup contact: Towards a new
vention promote national unity? British Journal of wave of research. In G. Hodson & M. Hewstone
Social Psychology, 51, 239–256. doi:10.1111/j.2044- (Eds.), Advances in intergroup contact (pp. 233–261).
8309.2011.02041.x London, UK: Psychology Press.
Amir, Y. (1969). Contact hypothesis in ethnic relations. Church, A. T. (1982). Sojourner adjustment. Psycho-
Psychological Bulletin, 71, 319–342. doi:10.1037/ logical Bulletin, 91, 540–572. doi:10.1037//0033-
h0027352 2909.91.3.540

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

42 Group Processes & Intergroup Relations 17(1)

Coenders, M., Lubbers, M., Scheepers, P., & Verkuyten, Guimond, S., Dambrun, M., Michinov, N., & Duarte, S.
M. (2008). More than two decades of chang- (2003). Does social dominance generate prejudice?
ing ethnic attitudes in the Netherlands. Journal Integrating individual and contextual determinants
of Social Issues, 64, 269–285. doi:10.1111/j.1540- of intergroup cognitions. Journal of Personality and
4560.2008.00561.x Social Psychology, 84, 697–721. doi:10.1037/0022-
Davies, K., Tropp, L. R., Aron, A., Pettigrew, T. F., 3514.84.4.697
& Wright, S. C. (2011). Cross-group friendships Haley, H., & Sidanius, J. (2005). Person–organization
and intergroup attitudes: A meta-analytic review. congruence and the maintenance of group-based
Personality and Social Psychology Review, 15, 332–351. social hierarchy: A social dominance perspective.
doi:10.1177/1088868311411103 Group Processes & Intergroup Relations, 8, 187–203.
Dhont, K., Cornelis, I., & van Hiel, A. (2010). Interra- doi:10.1177/1368430205051067
cial public–police contact: Relationships with police Hewstone, M. (2009). Living apart, living together?
officers’ racial and work-related attitudes and be- The role of intergroup contact in social integra-
haviors. International Journal of Intercultural Relations, tion. Proceedings of the British Academy, 162, 243–300.
34, 551–560. doi:10.1016/j.ijintrel.2010.07.004 doi:10.5871/bacad/9780197264584.003.0009
Dhont, K., Roets, A., & van Hiel, A. (2011). Open- Hodson, G. (2008). Interracial prison contact: The
ing closed minds: The combined effects of inter- pros for (social dominant) cons. British Journal of
group contact and need for closure on prejudice. Social Psychology, 47, 325–351. doi:10.1348/014466
Personality and Social Psychology Bulletin, 37, 514–528. 607X231109
doi:10.1177/0146167211399101 Hodson, G. (2011). Do ideologically intolerant
Dhont, K., & van Hiel, A. (2009). We must not be people benefit from intergroup contact? Cur-
enemies: Interracial contact and the reduction rent Directions in Psychological Science, 20, 154–159.
of prejudice among authoritarians. Personality and doi:10.1177/0963721411409025
Individual Differences, 46, 172–177. doi:10.1016/j. Hodson, G., & Esses, V. M. (2005). Lay perceptions of
paid.2008.09.022 ethnic prejudice: Causes, solutions, and individual
Dhont, K., & van Hiel, A. (2011). Direct contact differences. European Journal of Social Psychology, 35,
and authoritarianism as moderators between 329–344. doi:10.1002/ejsp.251
extended contact and reduced prejudice: Low- Hodson, G., Harry, H., & Mitchell, A. (2009). Independ-
er threat and greater trust as mediators. Group ent benefits of contact and friendship on attitudes to-
Processes and Intergroup Relations, 14, 223–237. ward homosexuals among authoritarians and highly
doi:10.1177/1368430210391121 identified heterosexuals. European Journal of Social Psy-
Dhont, K., van Hiel, A., De Bolle, M., & Roets, A. chology, 39, 509–525. doi:10.1002/ejsp.558
(2012). Longitudinal intergroup contact effects on Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit in-
prejudice using self- and observer-reports. British dexes in covariance structure analysis: Convention-
Journal of Social Psychology, 51, 221–238. doi:10.1111/ al criteria versus new alternatives. Structural Equation
j.2044-8309.2011.02039.x Modeling, 6, 1–55. doi:10.1080/10705519909540118
Dixon, J., Tropp, L. R., Durrheim, K., & Tredoux, Huang, L. L., & Liu, J. H. (2005). Personality and so-
C. (2010). “Let them eat harmony”: Prejudice- cial structural implications of the situational prim-
reduction strategies and attitudes of historically ing of social dominance orientation. Personality and
disadvantaged groups. Current Directions in Psy- Individual Differences, 38, 267–276. doi:10.1016/j.
chological Science, 19, 76–80. doi:10.1177/096372 paid.2004.04.006
1410363366 Jackson, J. (2009). Intercultural learning on short-
Duckitt, J. (2001). A dual-process cognitive-motivation- term sojourns. Intercultural Education, 20, 59–71.
al theory of ideology and prejudice. Advances in Ex- doi:10.1080/14675980903370870
perimental Social Psychology, 33, 41–113. doi:10.1016/ Jöreskog, K. G., & Sorbom, D. (2004). LISREL 8.7 for
S0065-2601(01)80004-6 Windows [Computer software]. Lincolnwood, IL:
Eller, A., & Abrams, D. (2003). “Gringos” in Mexico: Scientific Software International, Inc.
Cross-sectional and longitudinal effects of lan- Judd, C. M., Kenny, D. A., & McClelland, G. H. (2001).
guage school-promoted contact on intergroup bias. Estimating and testing mediation and moderation
Group Processes and Intergroup Relations, 6, 55–75. in within-participant designs. Psychological Methods, 6,
doi:10.1177/1368430203006001012 115–134. doi:10.1037/1082-989X.6.2.115

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

Dhont et al. 43

Kline, R. B. (2010). Principles and practice of structural equa- Social Psychology, 44, 156–163. doi:10.1016/j.jesp.
tion modeling. New York, NY: Guilford Press. 2006.12.006
Kteily, N. S., Ho, A., & Sidanius, J. (2012). Hierarchy Pedersen, P. J. (2009).Teaching toward an ethnorela-
in the mind: The predictive power of social domi- tive worldview through psychology study abroad.
nance orientation across social contexts and do- Intercultural Education, 20, 73–86. doi:10.1080/
mains. Journal of Experimental Social Psychology, 48, 14675980903370896
543–549. doi:10.1016/j.jesp.2011.11.007 Perry, R., Sibley, C. G., & Duckitt, J. (2013). Danger-
Kteily, N. S., Sidanius, J., & Levin, S. (2011). Social ous and competitive worldviews: A meta-analysis
dominance orientation: Cause or “mere effect”?: of their associations with social dominance orien-
Evidence for SDO as a causal predictor of preju- tation and right-wing authoritarianism. Journal of
dice and discrimination against ethnic and racial Research in Personality, 47, 116–127. doi:10.1016/j.
outgroups. Journal of Experimental Social Psychology, jrp.2012.10.004
47, 208–214. doi:10.1016/j.jesp.2010.09.009 Pettigrew, T. F. (1997). Generalized intergroup con-
Lippa, R., & Arad, S. (1999). Gender, personality, and tact effects on prejudice. Personality and Social
prejudice: The display of authoritarianism and so- Psychological Bulletin, 23, 173–185. doi:10.1177/
cial dominance in interviews with college men and 0146167297232006
women. Journal of Research in Personality, 33, 463– Pettigrew, T. F. (2009). Secondary transfer effect of
493. doi:10.1006/jrpe.1999.2266 contact: Do intergroup contact effects spread to
Little, T. D., Cunningham, W. A., Shahar, G., & Wida- noncontacted outgroups? Social Psychology, 40, 55–
man, K. F. (2002). To parcel or not to parcel: Ex- 65. doi:10.1027/1864-9335.40.2.55
ploring the question and weighing the merits. Struc- Pettigrew, T. F., & Tropp, L. R. (2006). A meta-ana-
tural Equation Modeling, 9, 151–173. doi:10.1207/ lytic test of intergroup contact theory. Journal
S15328007SEM0902_1 of Personality and Social Psychology, 90, 751–783.
Little, T. D., Preacher, K. J., Selig, J. P., & Card, N. doi:10.1037/0022-3514.90.5.751
A. (2007). New developments in latent vari- Pettigrew, T. F., & Tropp, L. R. (2011). When groups meet:
able panel analyses of longitudinal data. Jour- The dynamics of intergroup contact. Philadelphia, PA:
nal of Behavioral Development, 31, 357–365. Psychology Press.
doi:10.1177/0165025407077757 Pizam, A., Fleischer, A., & Mansfeld, Y. (2002). Tour-
Mackie, D. M., Devos, T., & Smith, E. R. (2000). ism and social change: The case of Israeli ecotour-
Intergroup emotions: Explaining offensive ac- ists visiting Jordan. Journal of Travel Research, 41,
tion tendencies in an intergroup context. Jour- 177–184. doi:10.1177/004728702237423
nal of Personality and Social Psychology, 79, 602–616. Pratto, F., Sidanius, J., & Levin, S. (2006). Social domi-
doi:10.1037//0022-3514.79.4.602 nance theory and the dynamics of intergroup
Matthews, M., Levin, S., & Sidanius, J. (2009). A longi- relations: Taking stock and looking forward.
tudinal test of the model of political conservatism European Review of Social Psychology, 17, 271–320.
as motivated social cognition. Political Psychology, 30, doi:10.1080/10463280601055772
921–936. doi:10.1111/j.1467-9221.2009.00733.x Pratto, F., Sidanius, J., Stallworth, L. M., & Malle, B. F.
McConahay, J. B. (1986). Modern racism, ambivalence, (1994). Social dominance orientation: A personal-
and the modern racism scale. In J. F. Dovidio & S. ity variable predicting social and political attitudes.
L. Gaertner (Eds.), Prejudice, discrimination and racism Journal of Personality and Social Psychology, 67, 741–
(pp. 91–126). New York, NY: Academic Press. 763. doi:10.1037/0022-3514.67.4.741
Meredith, W. (1993). Measurement invariance, factor Roets, A., van Hiel, A., & Dhont, K. (2012). Is sex-
analysis and factorial invariance. Psychometrika, 58, ism a gender issue? A motivated social cogni-
525–543. doi:10.1007/BF02294825 tion perspective on men’s and women’s sexist
Milman, A., Reichel, A., & Pizam, A. (1990). The im- attitudes toward the own and other gender. Euro-
pact of tourism on ethnic attitudes: The Israeli– pean Journal of Personality, 26, 350–359. doi:10.1002/
Egyptian case. Journal of Travel Research, 29, 45–49. per.843
doi:10.1177/004728759002900207 Saguy, T., Tropp, L., & Hawi, D. (2013). The role of
Morrison, K., & Ybarra, O. (2008). The effects of group power in intergroup contact. In G. Hodson
realistic threat and group identification on social & M. Hewstone (Eds.), Advances in intergroup contact
dominance orientation. Journal of Experimental (pp. 113–132). London, UK: Psychology Press.

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015

44 Group Processes & Intergroup Relations 17(1)

Schafer, J. L. (1997). Analysis of incomplete multivari- Sidanius, J., van Laar, C., Levin, S., & Sinclair, S. (2003).
ate data. London, UK: Chapman & Hall. Social hierarchy maintenance and assortment
Schafer, J. L., & Graham, J. W. (2002). Missing data: into social roles: A social dominance perspective.
Our view of the state of the art. Psychological Meth- Group Processes & Intergroup Relations, 6, 333–352.
ods, 7, 147–177. doi:10.1037/1082-989X.7.2.147 doi:10.1177/13684302030064002
Schild, E. O. (1962). The foreign student as a stranger Stephan, W. G., & Vogt, W. P. (Eds.). (2004). Education
learning the norms of the host culture. Journal of So- programs for improving intergroup relations: Theory, research
cial Issues, 75, 41–54. doi:10.1111/j.1540-4560.1962. and practice. New York, NY: Teachers College Press.
tb02570.x Swart, H., Hewstone, M., Christ, O., & Voci, A. (2011).
Schmid, K., Hewstone, M., Küpper, B., Zick, A., & Affective mediators of intergroup contact: A longi-
Wagner, U. (2012). Secondary transfer effects of tudinal analysis in South Africa. Journal of Personality
intergroup contact: A cross-national comparison and Social Psychology, 101, 1221–1238. doi:10.1037/
in Europe. Social Psychology Quarterly, 75, 28–51. a0024450
doi:10.1177/0190272511430235 Tausch, N., Hewstone, M., Kenworthy, J. B., Psaltis,
Schmitt, M., Branscombe, N., & Kappen, D. (2003). C., Schmid, K., Popan, J. R., & Hughes, J. (2010).
Attitudes toward group-based inequality: So- Secondary transfer effects of intergroup contact:
cial dominance or social identity? British Journal Alternative accounts and underlying processes. Jour-
of Social Psychology, 42, 161–186. doi:10.1348/ nal of Personality and Social Psychology, 99, 282–302.
014466603322127166 doi:10.1037/a0018553
Shadish, W. R., Cook, T. D., & Campbell, D. T. (2001). Tropp, L. R. (2006). Stigma and intergroup contact
Experimental and quasi-experimental designs for gen- among members of minority and majority status
eralized causal inference. Boston, MA: Houghton groups. In S. Levin & C. van Laar (Eds.), Stigma
Mifflin. and group inequality: Social psychological perspectives (pp.
Sibley, C. G., & Duckitt, J. (2010). The ideological le- 171–191). Mahwah, NJ: Erlbaum.
gitimation of the status quo: Longitudinal tests of Tropp, L. R., & Pettigrew, T. F. (2005). Relation-
a social dominance model. Political Psychology, 31, ships between intergroup contact and prejudice
109–137. doi:10.1111/j.1467-9221.2009.00747.x among minority and majority status groups. Psy-
Sibley, C. G., & Liu, J. H. (2010). Social dominance chological Science, 16, 951–957. doi:10.1111/j.1467-
orientation: Testing a global individual differ- 9280.2005.01643.x
ence perspective. Political Psychology, 31, 175–207. Van Hiel, A., & Duriez, B. (2002). Een meetinstrument
doi:10.1111/j.1467-9221.2009.00748.x voor individuele verschillen in Sociale Dominantie
Sibley, C. G., Wilson, M. S., & Duckitt, J. (2007a). Oriëntatie [A scale to measure individual differ-
Antecedents of men’s hostile and benevolent ences in social dominance orientation]. Nederlands
sexism: The dual roles of social dominance ori- Tijdschrift voor de Psycholgie en haar Grensgebieden, 57,
entation and right-wing authoritarianism. Per- 114–116.
sonality and Social Psychology Bulletin, 33, 160–172. Van Hiel, A., & Mervielde, I. (2005). Authoritari-
doi:10.1177/0146167206294745 anism and social dominance orientation: Rela-
Sibley, C. G., Wilson, M. S., & Duckitt, J. (2007b). Ef- tionships with various forms of racism. Jour-
fects of dangerous and competitive worldviews nal of Applied Social Psychology, 35, 2323–2344.
on right-wing authoritarianism and social domi- doi:10.1111/j.1559-1816.2005.tb02105.x
nance orientation over a five-month period. Po- Van Laar, C., Levin, S., Sinclair, S., & Sidanius, J.
litical Psychology, 28, 357–371. doi:10.1111/j.1467- (2005). The effect of university roommate contact
9221.2007.00572.x on ethic attitudes and behavior. Journal of Experi-
Sidanius, J., Levin, S., van Laar, C., & Sears, D.O. (2009). mental Social Psychology, 41, 329–345. doi:10.1016/j.
The diversity challenge: Social identity and intergroup rela- jesp.2004.08.002
tions on the college campus. New York, NY: The Russell Voci, A., & Hewstone, M. (2003). Intergroup contact
Sage Foundation. and prejudice towards immigrants in Italy: The me-
Sidanius, J., & Pratto, F. (1999). Social dominance: An in- diational role of anxiety and the moderational role of
tergroup theory of social hierarchy and oppression. Cam- group salience. Group Processes & Intergroup Relations,
bridge, UK: Cambridge University Press. 6, 37–54. doi:10.1177/1368430203006001011

Downloaded from gpi.sagepub.com at UNIV OF PITTSBURGH on March 14, 2015