Dangerous and competitive worldviews: A meta-analysis of their associations
with Social Dominance Orientation and Right-Wing Authoritarianism
Ryan Perry, Chris G. Sibley
, John Duckitt
School of Psychology, University of Auckland, New Zealand
Available online 24 October 2012
Social Dominance Orientation
A meta-analysis of 46 studies (N= 12,939) examined the cross-sectional associations between dangerous
worldview and Right-Wing Authoritarianism (RWA), and between competitive worldview and Social
Dominance Orientation (SDO). These dual associations were asymmetric; the dangerous worldview-
RWA partial-correlation was moderate (r= .37), whereas the competitive worldview-SDO partial-
correlation was stronger in size (r= .53). The results support a dual-process model perspective and indi-
cate that RWA and SDO are consistently linked with distinct social schemas of the social world as dan-
gerous and threatening (versus safe and secure), and competitive and cut-throat (versus co-operative
and characterized by mutually beneﬁcial exchange). We present a reﬁned competitive worldview scale
that reduces content overlap and provides a more accurate estimate of the worldview-SDO association.
Ó2012 Elsevier Inc. All rights reserved.
A dual-process model (DPM; Duckitt, 2001) of ideology and pre-
judice identiﬁes dual threat- and competition-based cognitive-
motivational processes that determine individual differences in
prejudice. According to the model, dual dimensions of prejudice
proneness are reliably indexed using measures of Right-Wing
Authoritarianism (RWA; Altemeyer, 1981) and Social Dominance
Orientation (SDO; Pratto, Sidanius, Stallworth, & Malle, 1994).
The DPM further asserts that SDO and RWA are not personality
traits, but rather two dimensions of ideological attitudes that ex-
press relatively independent motivations for group-based domi-
nance and superiority in the case of SDO, and social cohesion and
collective security in the case of RWA. Activation of these motiva-
tional goals is inﬂuenced by differences in personality, and beliefs
about the social world as dangerous and threatening (versus safe
and secure), and as competitive and cut-throat (versus co-
operative and characterized by mutually beneﬁcial exchange).
The DPM is presented in Fig. 1. According to the DPM, differences
in personality are derived from individual socialization experi-
ences. These inﬂuences probably interact with more contemporary
socio-environmental cues that some individuals attend to more or
less so than others depending upon their personality (Duckitt,
2001; Duckitt, Wagner, du Plessis, & Birum, 2002).
The motivational goal process in general, and the DPM world-
view concept in particular, were drawn from an earlier theoretical
construct developed by D’Andrade (1992), Strauss (1992), and Ross
(1993). Social attitudes, these earlier scholars suggested, are an
expression of motivational goals – motivating beliefs that have
been made salient for the individual by the activation of speciﬁc
social schemas. These schemas can be seen as forming coherent so-
cial worldviews – relatively stable interpretations or beliefs about
the social world and other people in that world (Ross, 1993). There
is a variety of evidence consistent with basic premises of the DPM
relating to the role and function of social worldviews, showing (a)
that perceptions of the social world as competitive should predict
SDO, and (b) that perceptions of the social world as dangerous and
threatening should predict RWA. For example, a longitudinal panel
study by Sibley, Wilson, and Duckitt (2007a) found signiﬁcant
cross-lagged effects in support of these hypothesized causal path-
ways between social worldviews at time one and ideological atti-
tudes 5 months later.
Another highly informative, but little known, example of the
worldview socialisation process was provided by Robertson
(2006) in his unpublished doctoral dissertation. Proposing that
fundamentalist religious socialisation contributes to the develop-
ment of a threatening worldview, Robertson (2006) reported that,
in a large community sample of Christian New Zealanders, an
emphasis on religion in childhood was positively related to a dan-
gerous worldview, which in turn predicted RWA. Religious social-
isation thus emphasizes the notion that the world is becoming
more dangerous, with these experiences leading people to form
stable perceptions of the world as threatening and unpredictable.
According to Robertson (2006), this worldview triggers a
0092-6566/$ - see front matter Ó2012 Elsevier Inc. All rights reserved.
Authors’ note: This manuscript is based on part of Ryan Perry’s PhD thesis
supervised by Chris Sibley and John Duckitt.
Corresponding author. Address: School of Psychology, University of Auckland,
Private Bag 92019, Auckland, New Zealand.
E-mail address: email@example.com (C.G. Sibley).
Journal of Research in Personality 47 (2013) 116–127
Contents lists available at SciVerse ScienceDirect
Journal of Research in Personality
journal homepage: www.elsevier.com/locate/jrp
motivation for religious group security and social control, which in
turn leads to prejudice toward groups seen as threatening their
safety and social values.
In this paper we present a meta-analysis of associations be-
tween the worldview and ideology components of the DPM. That
is, the respective associations of dangerous and competitive world-
views with RWA and SDO. A substantial research literature inves-
tigating the relationships between different components of the
DPM has often included these worldview constructs, modeled as
mediating factors linking individual differences in personality to
ideological attitudes captured by SDO and RWA. Relative to other
components of the DPM, however, worldview beliefs tend not to
be the focus of such studies and as such are probably not as sys-
tematically or thoroughly understood as they could otherwise be
(but see Duckitt & Fisher, 2003; Federico, Hunt, & Ergun, 2009;
Perry & Sibley, 2010; Sibley et al., 2007a). An exhaustive meta-
analysis of this relationship in the DPM has remained lacking and
would contribute to the literature by providing a more conclusive
assessment of the role of worldviews (as they relate to ideology
and prejudice) than has been previously available from stand-
alone studies. Here we aim to make such a contribution.
1.1. An asymmetry in the worldview–ideology association
Although never stated explicitly, the DPM implies that compet-
itive worldview should predict change in SDO at a comparable rate
to the extent that dangerous worldview predicts change in RWA.
However, a possible asymmetry in the associations between
worldviews and ideological attitudes in the DPM was ﬁrst high-
lighted by Van Hiel, Cornelis, and Roets (2007). These authors re-
ported that the correlation between competitive worldview and
SDO was markedly stronger than that between dangerous world-
view and RWA across undergraduate and community samples.
On the one hand, this might occur because of unique study or sam-
ple factors that differentially impact the extent to which dangerous
versus competitive worldview beliefs lead to changes in RWA and
SDO. Another possibility, however, is that the asymmetry results
from possible contamination due to differences in item content
overlap between the worldview measures and their respective
ideological attitude indexes.
Van Hiel, Cornelis, and Roets (2007) extracted two separate
dimensions, representing worldviews and ideological attitudes,
for each of the dual pathways. Items loading on the counter-
dimensional factor (e.g., a competitive worldview item loading
on the SDO factor) or with a weak loading on their intended factor
were then discarded from reanalysis. Critically, discarding prob-
lematic items substantially reduced the association between com-
petitive worldview and SDO, but not between dangerous
worldview and RWA. Van Hiel, Cornelis, and Roets (2007) con-
cluded that content overlap between items in the competitive
worldview and SDO scales may therefore have been artiﬁcially
inﬂating the relationship between these two components of the
DPM beyond the more reasonable and valid associations observed
between dangerous worldview and RWA.
Is this asymmetry reliable across samples and contexts? The
current meta-analysis assesses an asymmetry-hypothesis across
all studies that have measured these components of the DPM. If
asymmetry is highly consistent, this would concur with Van Hiel,
Cornelis, and Roets’ (2007) concerns that the worldview and ideo-
logical attitude measures in the DPM may not be sufﬁciently inde-
pendent of one another, and would support their proposal that the
worldview constructs ought to be re-operationalized. If, however,
the association between competitive worldview and SDO is not
consistently stronger across samples, our meta-analysis may be
able to shed light on the study and sample factors that determine
this asymmetrical relationship.
1.2. Overview and aims of the meta-analysis
Despite the substantial research literature assessing the rela-
tionships of dangerous and competitive worldviews with RWA
and SDO as components in a process underlying prejudice, much
of the research is unpublished and, moreover, many of the pub-
lished studies do not report the full correlations between all mea-
sures. A systematic empirical review of this research remains
lacking, and is required in order to provide a coherent and thor-
ough picture of the relationship between worldview beliefs and
ideological attitudes. We aim to address this deﬁcit, presenting a
meta-analysis of the associations between dangerous worldview,
competitive worldview, RWA and SDO. Meta-analysis provides an
indication of the average effect sizes between the DPM worldview
indexes, SDO and RWA observed across all available research in the
area. Further to this, meta-analytic assessment of effect sizes from
multiple independent studies allowed us to examine possible mod-
erating factors (differences between studies) that might inﬂuence
the degree to which worldview beliefs correlate with SDO and
Our primary aim is to determine the average effect sizes be-
tween worldviews and ideological attitudes based on prior re-
search. Consistent with a DPM perspective, we expected that (a)
dangerous worldview would predict RWA and weakly predict
SDO and (b) competitive worldview would predict SDO and weakly
Fig. 1. A dual-process motivational model of the impact of personality, social environment, and social worldview beliefs on the two ideological attitude dimensions of Right-
Wing Authoritarianism (RWA) and Social Dominance Orientation (SDO) and their impact on socio-political behavior and attitudes as mediated through perceived social threat
or competitiveness over group dominance, power, and resources (adapted from Duckitt and Sibley, 2010, p. 1868).
R. Perry et al. / Journal of Research in Personality 47 (2013) 116–127 117
predict RWA. We note that these weaker effects should only reach
signiﬁcance given the power of our very large meta-analytic sam-
ple size. Consistent with Van Hiel, Cornelis, and Roets (2007),we
also hypothesize an asymmetry in which the relationship between
competitive worldview and SDO should appear to be consistently
and substantially stronger than that between dangerous world-
view and RWA across samples. We aim to determine the reliability
of this effect across all available samples to ascertain whether this
is something speciﬁc to certain studies (including Van Hiel, Cornel-
is, & Roets, 2007), or represents a more spurious phenomenon re-
lated to construct design. If this is the case, then a reﬁned
worldview measure would be warranted – an issue we return to la-
ter in the discussion.
1.3. Moderating factors
We examined the extent to which associations between world-
views and ideological attitudes were consistent (or varied) across a
number of sample characteristics: publication status, student ver-
sus adults samples, number of scale items, and cross-national dif-
ferences in income inequality.
1.3.1. Publication bias
Meta-analytic research often documents a publication bias in
which published studies tend to report stronger effects (Rosenthal,
1979). We thus sought to examine (and control for) potential dif-
ferences in effect sizes between worldview beliefs and RWA and
SDO across published and unpublished studies.
1.3.2. Number of scale items
Duckitt et al. (2002) originally included 10 items assessing dan-
gerous worldview beliefs and 20 items assessing competitive
worldview beliefs. Although there is a general consensus regarding
these two dimensions of worldviews, studies routinely use various
shortened versions of these scales. This is also the case for RWA
and SDO. We therefore tested for potential differences in effect
sizes across studies resulting from the number of scale items in-
cluded. Considering the generally high reliability of these dimen-
sions, we expect similar effect sizes regardless of the number of
scale items however. Any differences should be in the magnitude
rather than direction of effects.
1.3.3. Undergraduate students versus community samples
As is often the case in social psychological research, the major-
ity of studies measuring worldview beliefs and ideological atti-
tudes sampled undergraduate student populations, limiting the
ability to generalize results from these studies to non-student pop-
ulations. Although schematic worldview beliefs are presumably
relatively stable over time, the process by which worldviews are
formed and in turn inﬂuence ideological attitudes should also be
cumulative to an extent, with different critical developmental peri-
ods in times of chronic and powerful situational inﬂuence (e.g.,
Newcomb’s Bennington College study, 1943/1967). This suggests
potential for change across the life span. Duckitt (2001) also re-
ported a signiﬁcant effect of age on the association between
RWA and SDO where the two are consistently more closely associ-
ated in older participants. The implication here is that competitive
and dangerous worldviews may become more strongly associated
as people are exposed to consistent socialization experiences.
1.3.4. Income inequality
Our sample enabled us to examine differences (or consistencies)
in the magnitude of associations between worldviews and ideolog-
ical attitudes across nations as well as sample characteristics. The
GINI index is a measure of the inequality of personal incomes in a
nation and ranges from 0 (where everyone has the same income) to
1 (where one person has all the income and everyone else has no
income). The global GINI coefﬁcient is thought to be around .60–
.70 and has been steadily increasing (i.e., becoming more unequal)
in most countries since the 1980s (United Nations Development
Program (UNDP), 2010).
Averaged indexes for the years 2000–2010 were taken from the
Human Development Report of the United Nations (UNDP, 2010)
using data for each of the seven countries included in our meta-
analysis. Income inequality has been shown to have particularly
strong and pervasive effects on a broad range of various social
and psychological indicators (Wilkinson & Pickett, 2009). Typically,
more unequal societies exhibit more prevalent social problems,
abuse, violence and other negative social indicators. Recent
meta-analyses including RWA and SDO do not support a direct link
between these attitude indices and income inequality however. In
Cohrs and Stelzl (2010), the GINI index was not correlated with the
effect of either RWA or SDO on attitudes toward immigrants (when
outlying countries were controlled for). And in Fischer, Hanke, and
Sibley (2012) the GINI index did not predict national levels of SDO
over all. We did not expect that greater income inequality in a soci-
ety (as measured by the GINI index) would moderate associations
between dangerous worldview and RWA, or competitive world-
view and SDO, across nations.
2.1. Literature search
Our literature search was conducted using the PsychINFO and
SCOPUS online databases and the Google Scholar internet search
engine and was ﬁnalized in early 2010. When searching for rele-
vant articles, we combined terms such as social worldviews, sche-
mas, beliefs, dangerous, competitive, jungle, Right-Wing
Authoritarianism, RWA, authoritarianism, Social Dominance Orienta-
tion, SDO, dominance, dual process model and DPM. We also re-
viewed articles citing the original DPM papers by Duckitt (2001)
and Duckitt et al. (2002). In addition, we contacted numerous
researchers who had previously published research on the DPM,
or of whom we knew to be currently conducting research in the
area, and requested unpublished or in-press manuscripts and data.
We also posted a request on the Society for Personality and Social
Psychology ‘SPSP Discuss’ electronic mailing list. All studies in-
cluded in our analysis used a recognizable measure of one or both
(dangerous and competitive) DPM worldview dimensions (Duckitt,
2001) and included (a) a full or shortened version of the SDO scale
developed by Pratto et al. (1994), or (b) a full or shortened set of
items from one of Altemeyer’s (1981, 1988, 1996, 1998) measures
2.2. Study characteristics
As shown in Table 1,k= 46 studies with a total of N= 12,939
participants were identiﬁed. The majority of studies were unpub-
lished (k= 25 [54.3%]), and the remaining studies (k= 21 [45.7%])
were published or in-press (note that this included reanalyses of
published data in which the relevant correlations had not been re-
ported). All 46 studies included both a measure of dangerous
worldview beliefs and Right-Wing Authoritarianism. Thirty-nine
studies (84.8%) contained a measure of competitive worldview be-
liefs and 41 studies (89.1%) contained a measure of Social Domi-
nance Orientation. Of these, 37 studies (80.4%) contained all four
measures of interest. Forty-four studies (95.7%) assessed the
worldview factors using the measures developed for the DPM
(Duckitt, 2001) and only two studies (4.3%) used ad hoc, single-
item indices (in both cases these were reportedly modeled on
118 R. Perry et al. / Journal of Research in Personality 47 (2013) 116–127
Sample details and correlation coefﬁcients for all studies included the meta-analysis of dual-process model worldview measures (DW and CW), RWA and SDO.
Reference Sample details No. items in scales Nation Publication status Sample size Bivariate (and partial) correlations
SDO RWA CW DW SDO–RWA CW–DW CW–SDO CW–RWA DW–RWA DW–SDO
1. Altemeyer (1998) Undergrad 14 30 14 Canada Pub. 354 .22 .49 .00
2. Asbrock (2010) Undergrad 12 11 4 4 Germany Unpub. 42 .39 .09 .56 (.56) .08 (.12) .33 (.34) .04 (.01)
3. Asbrock and Fritsche (2010, Sample 1) Undergrad 12 1 1 Germany Unpub. 144 .53 .22 (.15) .18 (.07)
3. Asbrock and Fritsche (2010, Sample 2) Undergrad 11 4 4 Germany Unpub. 186 .19 .36 (.33) .27 (.23)
5. Cohrs (2005) Undergrad 12 11 1 1 Germany Pub. 133 .30 .44 .11 (.13) .18 (.08) .25 (.19) .01 (.07)
6. Crowson (2009) Undergrad 16 30 14 12 USA Pub. 242 .45 .18 .55 (.53) .23 (.16) .61 (.59) .27 (.21)
7. Dallago, Mirisola, and Roccato (2012) Adult 14 20 Italy Unpub. 483 .52
8. Derrick and Murray (2006) Undergrad 30 12 USA Unpub. 88 .17
9. Dhont and Van Hiel (2009) Adult 14 11 11 13 Belgium Unpub. 90 .47 .47 .32 (.23) .44 (.44) .12 (.11) .26 (.13)
10. Duckitt (2000) Adult 6 18 8 6 NZ Unpub. 179 .45 .32 .60 (.58) .45 (.36) .46 (.37) .17 (.03)
11. Duckitt (2001, Study 2) Undergrad 10 18 12 NZ Pub. 497 .37 .45 .15
12. Duckitt (2001, Study 3) Undergrad 12 18 12 10 NZ Pub. 377 .40 .11 .60 (.59) .11 (.06) .54 (.53) .21 (.18)
13. Duckitt (2001, Study 4) Undergrad 10 10 14 10 South Africa Pub. 215 .21 .19 .58 (.56) .03 (.06) .45 (.45) .29 (.22)
14. Duckitt and Fisher (2003) Undergrad 12 20 14 NZ Pub. 280 .16 .42 .21
15. Duckitt et al. (2002, Study 1) Undergrad 12 20 20 10 USA Pub. 146 .21 .03 .55 (.55) .00 (.01) .43 (.43) .08 (.08)
16. Duckitt et al. (2002, Study 2) Undergrad 10 14 14 10 South Africa Pub. 231 .21 .19 .58 (.56) .03 (.06) .45 (.45) .29 (.22)
17. Federico, Weber, Ergun, and Hunt (2010, Sample 1) Undergrad 16 12 13 10 USA Unpub. 1065 .29 .23 .63 (.62) .17 (.08) .40 (.38) .16 (.02)
18. Federico et al. (in preparation, Sample 2) Undergrad 16 12 13 10 USA Unpub. 434 .25 .04 .51 (.51) .09 (.11) .39 (.39) .08 (.07)
19. Jugert and Duckitt (2009, Study 1) Undergrad 6 10 6 NZ Pub. 218 .09 .61 .09
20. Jugert and Duckitt (2009, Study 2) Undergrad 10 6 NZ Pub. 136 .33
21. Mirisola, Di Stefano, and Falgares (2007) Adult 8 14 20 10 Italy Pub. 177 .34 .11 .36 (.36) .22 (.19) .41 (.40) .06 (.02)
22. Perry et al. (in press) Undergrad 16 30 20 10 NZ Unpub. 345 .33 .15 .63 (.62) .25 (.20) .52 (.51) .19 (.13)
23. Robertson (2006) Undergrad 16 10 20 10 NZ Unpub. 185 .09 .07 .47 (.47) .20 (.26) .45 (.47) .08 (.05)
24. Sibley (2007) Undergrad 6 6 6 6 NZ Unpub. 470 .08 .22 .50 (.50) .02 (.04) .26 (.26) .04 (.08)
25. Sibley (2010) Adult 10 10 6 6 NZ Unpub. 531 .09 .36 .45 (.41) .10 (.03) .21 (.19) .19 (.04)
26. Sibley and Duckitt (2006) Undergrad 10 10 8 8 NZ Unpub. 428 .30 .22 .58 (.56) .06 (.05) .51 (.51) .20 (.09)
27. Sibley and Duckitt (2012) Adult 6 6 4 4 NZ Unpub. 843 .13 .12 .40 (.40) .05 (.02) .28 (.28) .07 (.02)
28. Sibley and Perry (2010) Adult 8 8 6 6 NZ Unpub. 331 .08 .29 .43 (.43) .00 (.15) .45 (.46) .09 (.04)
29. Sibley and Duckitt (2009, Sample 1) Undergrad 8 8 6 6 NZ Pub. 666 .17 .29 .56 (.54) .08 (.03) .36 (.35) .16 (.00)
30. Sibley and Duckitt (2009, Sample 2) Undergrad 8 8 6 6 NZ Pub. 258 .13 .18 .56 (.56) .03 (.07) .21 (.22) .04 (.07)
31. Sibley, Liu, Duckitt, and Khan (2008) Undergrad 6 6 6 6 NZ Pub. 194 .23 .36 .57 (.53) .16 (.02) .40 (.37) .26 (.07)
32. Sibley et al. (2007a) Undergrad 10 10 8 8 NZ Pub. 331 .22 .15 .56 (.55) .14 (.08) .43 (.42) .13 (.06)
33. Sibley, Wilson, and Duckitt (2007b) Undergrad 16 16 8 8 NZ Pub. 213 .35 .23 .65 (.64) .22 (.15) .38 (.35)
34. Sibley, Wilson, and Robertson (2007) Adult 10 10 8 8 NZ Pub. 340 .27 .30 .57 (.54) .12 (.03) .49 (.48) .25 (.10)
35. Van Hiel (2010a, Unpub. data) Adult 14 11 13 11 Belgium Unpub. 80 .54 .37 .58 (.56) .54 (.49) .29 (.12) .18 (.05)
36. Van Hiel (2010b, Unpub. data) Adult 14 11 13 11 Belgium Unpub. 105 .31 .44 .58 (.54) .21 (.07) .35 (.29) .26 (.01)
37. Van Hiel, Cornelis, and Roets (2007, Sample 1) Undergrad 14 24 14 10 Belgium Pub. 183 .60 .27 .55 (.52) .36 (.29) .40 (.34) .24 (.11)
38. Van Hiel, Cornelis, and Roets (2007, Sample 2) Adult 14 24 14 10 Belgium Pub. 276 .47 .39 .58 (.50) .38 (.21) .56 (.48) .43 (.27)
39. Van Hiel, Cornelis, and Roets (2007, Sample 3) Adult 14 24 14 10 Belgium Unpub. 68 .50 .35 .64 (.59) .40 (.25) .61 (.55) .37 (.20)
40. Van Hiel, Duriez, and Kossowska (2006) Undergrad 14 11 13 11 Belgium Unpub. 208 .44 .40 .54 (.44) .36 (.19) .53 (.45) .44 (.29)
41. Weber and Federico (2007) Undergrad 8 12 14 12 USA Pub. 255 .29 .22 .68 (.67) .26 (.20) .37 (.33) .13 (.03)
42. Webster, Whitley, and Miller (2006) Undergrad 16 26 14 9 USA Unpub. 312 .45 .22 .65 (.63) .21 (.17) .23 (.19) .20 (.08)
43. Wilson (2006) Adult 16 30 8 8 NZ Unpub. 135 .46 .33 .53 (.48) .27 (.09) .63 (.59) .28 (.13)
44. Wilson (2010c) Undergrad 16 30 8 8 NZ Unpub. 141 .45 .11 .69 (.69) .29 (.27) .35 (.33) .22 (.19)
45. Wilson (2010a, Unpub. data) Undergrad 16 30 8 8 NZ Unpub. 148 .52 .25 .63 (.60) .41 (.34) .59 (.56) .29 (.18)
46. Wilson (2010b, Unpub. data) Undergrad 16 30 8 8 NZ Unpub. 153 .45 .21 .59 (.58) .35 (.30) .34 (.29) .12 (.00)
R. Perry et al. / Journal of Research in Personality 47 (2013) 116–127 119
the DPM worldview component deﬁnitions). The majority of stud-
ies were conducted in New Zealand (k= 23 [50.0%]), the remainder
were conducted in a further six countries: Belgium (k= 7 [15.2%]),
the USA (k= 7 [15.2%]), Germany (k= 4 [8.7%]), Italy (k= 2 [4.3%]),
South Africa (k= 2 [4.3%]), and Canada (k= 1 [2.2%]). The majority
of studies assessed undergraduate student participants (k=33
[71.7%]), and the remainder (k= 13 [28.3%]) assessed adult (i.e.,
non-undergraduate student) populations.
2.3. Data analysis
We followed the meta-analytic procedures outlined by Hedges
and Olkin (1985) and Lipsey and Wilson (2001). Bivariate correla-
tions between measures of worldview beliefs and SDO and RWA
were transformed to Fisherized z-score effect sizes using the for-
[(1 + r)/(1 r)] and then weighted by their inverse
3, where n
is the number of participants in study i)
and averaged before being converted back to rusing the formula
+ 1). According to Cohen’s (1988) widely accepted
effect size conventions, rvalues less than or approaching .10 repre-
sent weak effects. Only bivariate associations greater than .10 were
therefore treated as noteworthy.
To determine the extent to which associations between each
worldview dimension and RWA and SDO might be due to shared
association with the other worldview dimension, we calculated
partial correlations controlling for the other worldview dimension
in each case. For example, when examining the association be-
tween dangerous worldview beliefs and RWA (also with SDO),
we calculated partial correlations between these two variables that
also controlled for their shared association with competitive
worldview beliefs. Partial correlations were calculated using the
of partial correlations were conducted using the same procedures
as those assessing bivariate correlations (outlined above) but with
the inverse variance adjusted to reﬂect the reduced degrees of free-
dom (e.g., inverse variance = n4).
To examine whether variation in effect sizes was greater than
could be attributed solely to chance, we conducted initial homoge-
neity tests using a ﬁxed-effects model: Q
. Here Q
is distributed as a
statistic with k1
degrees of freedom where kis the number of studies. A signiﬁcant
statistic would thus indicate that there was signiﬁcantly more
variation in an effect size than expected by chance and that this
variation might be attributable to systematic differences across
studies. We examined whether such variation could be attributed
to cross-study differences using a combination of inverse-
variance-weighted one-way analysis of variance (ANOVA) and in-
verse-variance-weighted regression analysis.
For both methods of analysis, we adopted a mixed-effects mod-
el using an iterative method based on maximum likelihood
(Raudenbush, 1994). This approach allowed us to model the
variance in effect sizes beyond that expected by chance or sam-
pling error deriving from two components: variance attributable
to systematic between-study differences and variance deriving
from other (essentially) random sources. Much like a ﬁxed-effects
model, this method of analysis allowed us to examine the extent to
which effect sizes were moderated by between-study differences.
However, the model also includes a random component represent-
ing the effect of unmeasured (and, as Lipsey & Wilson, 2001 noted,
potentially unmeasurable) random effects. Mixed-effects models
thus provide a more conservative test of cross-study differences
than ﬁxed effects models and are generally recommended in cases,
such as ours, in which effect sizes are generally not homogeneous.
Mixed-effects models are more accurate in such cases because they
are less prone to Type I errors (Overton, 1998). We conducted these
analyses using the macros provided by Lipsey and Wilson (2001).
3.1. Bivariate and partial correlations
A summary of the bivariate and partial correlations between
dangerous and competitive worldview measures and SDO and
RWA for each study is presented in Table 1. Coefﬁcients in paren-
theses represent the partial associations between each worldview
dimension and RWA or SDO controlling for the other worldview
dimension. Table 2 presents the average bivariate and partial cor-
relations between the worldview factors and SDO and RWA. Aver-
age weighted rvalues, lower and upper 95% conﬁdence intervals
(CIs), homogeneity statistics (Q
), number of studies (k), and total
number of participants (n) for analyses of the bivariate and partial
associations between dangerous worldview, competitive world-
view, SDO and RWA are presented in Table 2.
As shown in Table 2, meta-analytic averages indicated that dan-
gerous worldview was positively correlated with RWA (r= .41),
with a moderately strong effect size. Dangerous worldview was
also positively correlated with SDO, although this effect was smal-
ler in magnitude (r= .17). Competitive worldview, in contrast, was
positively and strongly correlated with SDO (r= .55) and more
weakly positively correlated with RWA (r= .19). In addition, dan-
gerous worldview and competitive worldview were positively cor-
related with one another (r= .24), with a moderate effect size. RWA
and SDO were also moderately positively correlated with one an-
other (r= .31).
As outlined in Fig. 2, effects remained comparable when exam-
ining partial correlations between each worldview dimension and
RWA and SDO, controlling for the other worldview dimension in
each case. Meta-analytic averages indicated that dangerous world-
view was moderately positively correlated with RWA (r= .37)
when controlling for competitive worldview. Likewise competitive
worldview was strongly positively correlated with SDO (r= .53)
when controlling for dangerous worldview. Partial correlations
also indicate that dangerous worldview was weakly positively cor-
related with SDO (r= .08), whereas competitive worldview was
weakly correlated with RWA (r= .11).
The ﬁxed-effects homogeneity tests reported in Table 2 indi-
cated that there was signiﬁcantly more variation in the bivariate
associations of both worldview dimensions with SDO and RWA
than expected by chance (assuming that such variation is not ran-
dom). Variation was considerably pronounced in all associations.
As noted earlier, these signiﬁcant Q
statistics suggest that there
were systematic differences across studies that might moderate
the association of worldviews with SDO, RWA and prejudice. We
thus tested mixed-effects models (which included both a ﬁxed
and random component) to assess potential sources of variation.
3.2. Moderating factors
We used inverse-variance-weighted multiple regression to
simultaneously examine the unique effects of a number of moder-
ating factors controlling for shared variation attributable to two or
more moderating factors. We conducted regression analyses test-
ing the extent to which the cross-study differences that we in-
cluded moderated the associations between worldview beliefs,
RWA and SDO.
For analyses of the bivariate and partial (controlling for the
other worldview dimension in each case) associations of dangerous
worldview with RWA (see Table 3) and competitive worldview
with SDO (see Table 4), we ﬁrst entered (and therefore controlled
for) the (contrast-coded) effects of publication status (unpub-
lished = 0.50, published = 0.50). We entered (and controlled for)
separate mutually exclusive variables representing the number of
120 R. Perry et al. / Journal of Research in Personality 47 (2013) 116–127
Meta-analysis of bivariate (and partial) correlations between dual-process model worldview measures, RWA and SDO.
r95% CI z-Test Q
No. of studies N
DW with RWA .412 .377 .446 20.51
DW with SDO .174 .141 .207 10.04
CW with RWA .187 .140 .233 7.68
CW with SDO .552 .521 .582 27.85
DW with CW .242 .206 .278 12.66
RWA with SDO .309 .265 .351 13.15
DW with RWA (cntrl. for CW) .374 .333 .414 16.36
DW with SDO (cntrl. for CW) .077 .045 .108 4.69
CW with RWA (cntrl. for DW) .106 .060 .152 4.44
CW with SDO (cntrl. for DW) .531 .499 .561 26.81
Note. A positive correlation represented a stronger association between worldviews, RWA and SDO. r-values and associated conﬁdence intervals were calculated assuming a
random-effects model. z-tests reﬂect signiﬁcance tests of weighted r-values, also assuming a random-effects model. Q
statistics reﬂect a test of the homogeneity of effect
sizes, calculated assuming a ﬁxed-effects model.
Fig. 2. Average meta-analytical bivariate (and partial) effect sizes between worldviews, RWA and SDO across 46 studies.
Inverse variance mixed-effects weighted regression analyses examining the effects of various study characteristics on the bivariate and partial (controlling for CW) relations
between the dangerous worldview dimension and RWA.
Study factor Model of DW–RWA association (bivariate) Model of DW–RWA association (controlling for CW)
Constant .398 .136 2.94
.151 .153 .99
Published (0.50 = no; 0.50 = yes) .068 .039 1.74 .052 .046 1.12
Number of DW items (centered) .016 .010 1.61 .003 .011 .24
Number of RWA items (centered) .007 .003 2.75
.009 .003 3.03
Sample (.50 = undergraduates; .50 = community/adult) .063 .044 1.44 .017 .049 .34
Income inequality (GINI) .002 .004 .01 .007 .004 1.70
Note. RWA was scored so that a higher value reﬂected a higher level of each dimension. brefers to the unstandardized regression coefﬁcient in Fisherized (z
) effect size units,
se refers to the standard error of b.z-tests reﬂect signiﬁcance tests of b. All effects were calculated using an inverse variance weighted regression analysis assuming a mixed-
Inverse variance mixed-effects weighted regression analyses examining the moderating effects of various study characteristics on the bivariate and partial (controlling for DW)
relations between the competitive worldview dimension and SDO.
Study factor Model of CW–SDO association (bivariate) Model of CW–SDO association (controlling for DW)
Constant .423 .147 2.87
.382 .142 2.70
Published (0.50 = no; 0.50 = yes) .017 .044 .39 .024 .042 .58
Number of CW items (centered) .002 .005 .43 .002 .005 .39
Number of SDO items (centered) .007 .007 .96 .004 .005 .55
Sample (.50 = undergraduates; .50 = community/adult) .086 .048 1.58 .106 .046 2.31
Income inequality (GINI) .005 .004 1.24 .005 .004 1.32
Note. SDO was scored so that a higher value reﬂected a higher level of each dimension. brefers to the unstandardized regression coefﬁcient in Fisherized (z
) effect size units,
se refers to the standard error of b.z-tests reﬂect signiﬁcance tests of b. All effects were calculated using an inverse variance weighted regression analysis assuming a mixed-
R. Perry et al. / Journal of Research in Personality 47 (2013) 116–127 121
items measuring dangerous worldview beliefs and competitive
worldview beliefs in each study. We also controlled for possible
differences between undergraduate student (coded .50) and
adult/community (coded .50) samples. One beneﬁt of using con-
trast codes (.50 and .50) rather than dummy codes (0 and 1) in
this type of analysis is that the intercept or constant for each
regression equation represents an easily interpretable mean effect
size controlling for cross-study differences. Finally, we entered and
controlled for levels of income inequality across the seven coun-
tries included in the meta-analysis using each country’s average
score from 2000 to 2010 on the GINI index. These ﬁve moderating
factors were also assessed for the association between the two
worldview dimensions (see Table 5).
As shown in Table 3, our regression analyses of the association
between dangerous worldview and RWA indicated that studies
employing a greater number of RWA items to assess this construct
appear to demonstrate a stronger relationship between dangerous
worldview and RWA with (b= .01, z= 3.03, p< .01) and without
(b= .01, z= 2.75, p< .01) controlling for competitive worldview.
In Table 4, our regression analysis of the association between com-
petitive worldview and SDO indicates that this effect was margin-
ally stronger in studies that used an undergraduate student sample
relative to an adult/community sample only when controlling for
dangerous worldview (b=.11, z=2.31, p< .05). There were no
further signiﬁcant moderating effects demonstrated by the other
We also tested the extent to which cross-study differences
moderated the association between the two worldview dimen-
sions. As shown in Table 5, our regression analysis of the associa-
tion between dangerous worldview and competitive worldview
indicates that this bivariate effect was modestly, but signiﬁcantly,
stronger in studies that used an adult sample relative to an under-
graduate student sample (b= .09, z= 2.15, p< .05) and marginally
stronger in studies that used fewer dangerous worldview items
to measure this construct (b=.02, z=2.19, p< .05).
Of the components comprising Duckitt’s (2001) dual-process
model, social worldviews have, until recently, garnered the least
attention from researchers (but see Federico et al., 2009; Perry &
Sibley, 2010; Sibley et al., 2007a; Van Hiel, Cornelis, & Roets,
2007). Evident from the number of unpublished data gathered in
the present meta-analysis, worldview associations with RWA and
SDO do not appear to comprise central research hypotheses when
they are included. Social worldviews are, however, a critical facet
of the DPM as the cognitive mechanism through which personality
and situational factors interact to produce the motivational goals
indexed by SDO and RWA. We conducted a meta-analysis to (a)
determine general effect sizes between the worldview and ideolog-
ical attitude components of the DPM, and (b) assess the extent to
which an asymmetry in effect sizes was robust across various sam-
ple and societal factors. We were able to determine that an asym-
metry in effect sizes (see Van Hiel, Cornelis, & Roets, 2007) was
reliable across samples and probably related to both item content
overlap between competitive worldview and SDO, and the tripar-
tite structure of RWA.
Meta-analysis of 46 independent samples (21 published, 25
unpublished) found good support for the DPM. Analysis of average
partial correlations indicated that dangerous and competitive
worldview beliefs were strong predictors of concurrent levels of
RWA and SDO respectively. The dual effects were also distinct, as
the opposing dangerous worldview-SDO and competitive world-
view-RWA associations were small (rs = .08, .11). Although the ex-
pected pattern of correlations was strong overall, the dual
associations were asymmetric – the dangerous worldview-RWA
partial-correlation was moderate in size (r= .37), whereas the
competitive worldview-SDO partial-correlation was stronger in
magnitude (r= .53). Finally, the two worldview dimensions were
also moderately positively correlated with one another, as were
RWA and SDO.
These average effects were robust across a number of study char-
acteristics including publication status, undergraduate versus com-
munity/adult samples, scale length, and national level of income
inequality. However, highly signiﬁcant heterogeneity in all cases
suggested that these relationships were still affected by a variety
of factors. The number of scale items used to measure the four con-
structs mattered in some cases, but was not a factor in others. The
type of sample (undergraduates or adults) also mattered, but only
to a limited extent. Overall, the robustness of the relationships be-
tween worldview beliefs and ideological attitudes argues for the rel-
ative stability of the relationships between these DPM components,
and therefore that the observed asymmetry is due to measurement
issues – most likely content overlap. The few instances of signiﬁcant
cross-study variability do however indicate factors that subtly affect
the relationships between social worldviews and SDO and RWA.
We identiﬁed a consistent and substantial asymmetry in the
relationships between worldviews and ideological attitudes. In line
with Van Hiel, Cornelis, and Roets (2007), we found a moderate
association of dangerous worldview with RWA, but a stronger
association between competitive worldview and SDO. This differ-
ence was reliable across samples, suggesting that it is a general
and robust effect and not idiosyncratic or limited to a particular
study context. The robustness of the asymmetry effect also sug-
gests it may be at least partially a spurious effect due to measure-
ment design, and we agree with Van Hiel, Cornelis, and Roets
(2007) that there is a need to purify or re-operationalize the exist-
ing DPM worldview construct to eschew this asymmetry.
In the following sections we review in detail the results of our
analyses of cross-study moderating effects, and present a reﬁned
and shortened DPM worldview measure that omits the competi-
tive worldview items that most appear to overlap with SDO.
Inverse variance mixed-effects weighted regression analyses examining the effects of various study characteristics on the relation between DW and CW.
Study factor Model of DW–CW association
Constant .514 .130 3.95
Published (0.50 = no; 0.50 = yes) .018 .039 .47
Number of DW items (centered) .016 .007 2.19
Number of CW items (centered) .028 .015 1.93
Sample (.50 = undergraduates; .50 = community/adult) .091 .042 2.15
Income inequality (GINI) .007 .004 1.92
Note. b refers to the unstandardized regression coefﬁcient in Fisherized (z
) effect size units, se refers to the standard error of b.z-tests reﬂect signiﬁcance tests of b. All effects
were calculated using an inverse variance weighted regression analysis assuming a mixed-effects model.
122 R. Perry et al. / Journal of Research in Personality 47 (2013) 116–127
4.1. Cross-study moderating effects
4.1.1. Publication status
Although there is a well-known publication bias in which pub-
lished studies show consistently stronger effects sizes than unpub-
lished studies (Rosenthal, 1979), this did not appear to be the case
in studies concerning relationships between DPM social world-
views and ideological attitudes, as publication status did not mod-
erate these associations. In an earlier meta-analysis of associations
between Big-Five personality and ideological attitudes (Sibley &
Duckitt, 2008), the authors suggested that successful publication
of correlational studies (typical of those included in our meta-
analysis) may be less contingent upon the detection of speciﬁc
hypothesized signiﬁcant differences relative to more experimental
study designs. The speciﬁc associations between worldviews and
RWA and SDO were also not typically the focus of studies included
in the present meta-analysis meaning these associations may have
been of limited relevance when being considered for publication.
4.1.2. Sample composition
In Van Hiel, Cornelis, and Roets (2007), dangerous worldview
was a more powerful mediator between personality and RWA in
an adult sample relative to an undergraduate student sample.
There was no such difference between their samples for competi-
tive worldview and SDO. On the contrary, in our meta-analysis of
bivariate associations, we did not observe a moderating effect of
sample composition (students or adults) on the relationship be-
tween dangerous worldview and RWA, and only marginal effects
of this moderator on other associations (discussed below). It there-
fore appears as if the effect observed by Van Hiel, Cornelis, and
Roets (2007) may be an exception, or situation-dependent –
generally the effect of dangerous worldview on RWA is consistent
across student and adult samples. It may be simply that there is
only a difference between student and adult samples in the medi-
ating effect of dangerous worldview however, and not the direct
bivariate effect of this component on RWA.
Furthermore, if we can consider the consistency between stu-
dent and adult samples proximal to an age-effect, our overall ﬁnd-
ings support the proposition that schematic worldview beliefs,
once formed, tend to be relatively stable over time. However, lon-
gitudinal research designs are needed to formally make inferences
about stability over time. Beliefs about the social world are subject
to situational changes throughout an individual’s life, but are only
substantially altered if such changes are sufﬁciently chronic and
powerful. Historiometric evidence, for instance, generally suggests
that worldview-related attitudes vary systematically over the
years depending upon levels of societal threat rather than eliciting
permanent change (Doty, Peterson, & Winter, 1991; Sales, 1972,
1973), even after major social event such as the 2001 World Trade
Centre attack (e.g., Perrin & Smolek, 2009). It seems safe to assume
that experiences within versus outside of university would not dif-
fer in perceived levels of threat anywhere near the extent that
would signiﬁcantly alter social worldviews and, moreover, this
sample characteristic should only moderate the mean levels of
dangerous worldview and RWA, not associations between them.
It is becoming more widely acknowledged that in some circum-
stances, however, the use of undergraduate student samples in
psychology (and other ﬁelds) may often be less generalizable than
researchers want to assume (Henrich, Heine, & Norenzayan, 2010).
In our study the associations between competitive worldview and
SDO were marginally signiﬁcantly stronger when undergraduate
student samples were used relative to adult/community samples.
There is no immediate explanation for this ﬁnding however, and
we again argue that any sample speciﬁc characteristics would most
likely lead to differences in mean levels of worldview and ideolog-
ical attitude scores rather than in the associations between these
There was also an indication that adult samples demonstrated a
slightly stronger association between the two worldview dimen-
sions compared with undergraduates, a ﬁnding consistent with
Duckitt (2001) who showed that SDO and RWA are more closely
associated in older participants. Though also only marginally sig-
niﬁcant, this difference may indicate that schematic worldview be-
liefs become slightly more convergent with age as well. As shown
in Table 2, there was a moderate correlation between the two
dimensions on average across samples. This is further evidence
that dangerous and competitive worldview beliefs, once formed,
are relatively stable and, therefore, as they stabilize over time their
mutual inﬂuence should become more consistent. Hence, we saw a
stronger association in adults relative to (typically younger) stu-
dents as worldview schemas should be less stable in this younger
population. Younger persons may be more amenable to change be-
cause their circumstances are changing more rapidly and differen-
tially in terms of exposure to socio-structural aspects relating to
competition versus danger. Indeed, Newcomb’s Bennington Col-
lege study beginning in 1935 revealed marked changes in socio-
political attitudes amongst conservative afﬂuent female college
students whereas a later 25 year follow up (and again after
50 years) showed remarkable stability of these attitudes from
graduation through adulthood (Newcomb, 1943; Newcomb,
Koenig, Flacks, & Warwick, 1967).
This is also consistent with Guimond, Dambrun, Michinov, and
Duarte (2003) and others’ research that showed socialization of
different university majors differentially affects levels of SDO
(Dambrun, Guimond, & Duarte, 2002; Dambrun, Kamiejski,
Haddadi, & Duarte, 2009; Sinclair, Sidanius, & Levin, 1998; see also
Roberts, Caspi, & Mofﬁtt, 2003). In addition to socialization there is
a complimentary process of self- or institutional-selection in which
people choose careers, university majors and so on that are norma-
tively consistent with their existing beliefs (e.g., Gatto, Dambrun,
Kerbrat, & De Oliveira, 2010; Sidanius, Van Laar, Levin, & Sinclair,
2004). In a reciprocal pattern, worldviews will therefore increas-
ingly stabilize with age as people self-select environments that
are consistent with those they have been formerly socialized by.
Roberts et al. (2003) describe this formally in their ‘‘corresponsive
principle’’ showing that certain personality traits predict work
experiences over time (measured at ages 18 and 26), and these
same personality traits were affected, in turn, by work experiences
over that time.
4.1.3. Number of scale items
Both the RWA and SDO scales show high levels of stability over
time when assessing test–retest correlations (Altemeyer, 1996;
Pratto et al., 1994; Sibley et al., 2007a), even when very short (4-
item) versions of the scales are employed (e.g., Cohrs, Moschner,
Maes, & Kielmann, 2005). In the present study, however, the stron-
gest variation in effect size arose from methodological differences
in the number of items used to assess RWA. Studies employing a
greater number of RWA items reported stronger associations be-
tween dangerous worldview and RWA. A greater number of SDO
items did not strengthen the association with competitive world-
view. For the worldview measures, the only signiﬁcant effect of
scale length was that a greater number of dangerous worldview
items appeared to decrease the effect size between the two world-
view dimensions – this was only marginally signiﬁcant however.
The moderating effect of RWA scale-length is consistent with a
growing body of evidence that suggests RWA comprises three
underlying dimensions (Duckitt, Bizumic, Krauss, & Heled, 2010;
Duckitt & Fisher, 2003; Funke, 2005; Mavor, Louis, & Sibley,
2010; Smith & Winter, 2002; Stellmacher & Petzel, 2005; Van Hiel,
Cornelis, Roets, & De Clercq, 2007b). Duckitt and Bizumic (in press)
R. Perry et al. / Journal of Research in Personality 47 (2013) 116–127 123
and Duckitt et al. (2010) presented a tripartite authoritarianism–
conservatism–traditionalism (ACT) scale of RWA, reconceptualis-
ing the scale items to overcome an acquiescence-driven dual factor
structure. Mavor et al. (2010) also proposed their authoritarian-
ism–conventionalism–submission (ACS) scale applying a novel ap-
proach to control for acquiescence bias in the original RWA items.
If this factor structure is not taken into account when determining
shortened versions of the RWA scale, there is an obvious risk of
inﬂating variation between studies that may inadvertently be tap-
ping distinct sub-dimensions.
Although scale shortening procedures are usually arbitrary,
there may still be systematic variation in the DW–RWA association
due to shortened RWA scales inadvertently tapping one or another
of the RWA subscales. Van Hiel, Cornelis, and Roets (2007; see also
Onraet, Van Hiel, Roets, & Cornelis, 2011; Van Hiel & De Clercq,
2009), for example, used a translated RWA scale that consisted of
10 pro-trait and 1 con-trait items. Considering item valences are
now known to assess different subscales of RWA (see Mavor
et al., 2010), these studies were assessing only part of the tripartite
Unfortunately, we did not obtain the actual worldview and
ideological attitude items used in studies; this would have allowed
us to examine differences across RWA subscales. A subsequent
meta-analysis comparing the content of shortened scales, as well
as their reliabilities (i.e., Cronbach’s alphas), would clarify the
implications of assessing different subscales on the association be-
tween RWA and other variables. The moderation effect of RWA
scale length observed here may account for the observed asymme-
try to an extent. This would potentially account for asymmetry in
Van Hiel, Cornelis, and Roets (2007), for example, considering their
shortened RWA scale did not assess the full tripartite construct.
4.1.4. Income inequality
As predicted, there was no signiﬁcant moderating effect of na-
tion-level income inequality on any of the associations that we re-
ported. The typically negative effect of high levels of income
inequality on various social and psychological indicators suggests
that this phenomenon should increase the salience of competition
in a society. As demonstrated in past meta-analyses (Cohrs & Stelzl,
2010; Fischer et al., 2012) however, this does not tend to be evi-
dent in mean levels of SDO. Moreover, we did not expect income
inequality to affect the relationship between competitive world-
view and SDO (or between dangerous worldview and RWA for that
matter). Increasing perceptions of competition through the sal-
ience of social inequality should only serve to increase mean levels
of competitive worldview, not the ability of this construct to pre-
dict SDO. It is unfortunate that we did not obtain mean level data
for the purposes of the present meta-analysis. Future investiga-
tions using between-nation constructs such as the GINI index
would be prudent to consider the effects on competitive worldview
in addition to SDO. Worldviews operate as a mechanism through
which both personality and social conditions lead to prejudice re-
lated ideology, so the true effect of social inequality may be more
accurately reﬂected at this stage of the cognitive process underly-
4.2. Dual path asymmetry
Our meta-analysis indicated that the associations between the
dual dimensions of worldviews and ideological attitudes are strong
and robust across a number of potential moderating variables.
However, there was an asymmetry in the strength of these rela-
tionships; the dangerous worldview association with RWA (partial
r= .37) was moderate whereas the competitive worldview associ-
ation with SDO was comparably strong (partial r = .53). Van Hiel,
Cornelis, and Roets (2007) suggested that this asymmetry may be
spurious. These authors tested the extent to which item content
overlap between the worldview and ideological attitude measures
artiﬁcially inﬂated associations between these DPM components
by removing items that cross-loaded between them (or loaded
on the counter-theoretical factor). Discarding the problematic
items reduced the correlation between competitive worldview
and SDO to a level comparable with that observed between dan-
gerous worldview and RWA. Of note, the latter association was
unaffected by the removal of overlapping items. Van Hiel, Cornelis,
and Roets (2007) concluded that the worldview construct there-
fore ought to be re-operationalized in order to eschew content
overlap with ideological attitude measures.
4.3. Removing spurious asymmetry
Our meta-analysis generally supported Van Hiel, Cornelis, and
Roets’s (2007) conclusion – the asymmetry was robust across sam-
ples and thus seems likely due to content overlap between the
competitive worldview and SDO measures, rather than sample-
or societal-level factors. Following Van Hiel, Cornelis, and Roets
(2007), we factor analyzed our previous sample containing the full
worldview scales to develop a reﬁned worldview measure that
should attenuate possible content overlap. We append a reﬁned
and balanced DPM social worldviews scale comprising 10-item
competitive and dangerous worldview measures in Table 6. A con-
ﬁrmatory factor analysis of data re-analyzed from Perry, Sibley and
Duckitt (2012) supported our reﬁned measure. The model demon-
strated good ﬁt to the data when modeling competitive worldview
as a latent factor predicting SDO (
(298) = 843.077; sRMR = .068).
Bivariate associations in these data were particularly high (for CW–
SDO, r= .63), though when employing the reﬁned 10-item measure
of competitive worldview, this association was reduced (r= .55) in
line with that between dangerous worldview and RWA (r= .52).
As originally proposed in the DPM (Duckitt, 2001; Duckitt et al.,
2002), social worldviews should assess descriptive beliefs about
the social world – as in, what the social world and people in it
are generally like. Conversely, ideological attitudes should assess
prescriptive beliefs – what the social world should be like, or
how we should respond to it. The ‘extra’ variance predicted by
the original competitive worldview measure in our meta-analysis
thus arguably represents mostly prescriptive ideological item-
content that overlaps with SDO, artiﬁcially inﬂating the relation-
ship (see also Perry, Sibley, & Duckitt, in press). We also note that
the association between dangerous worldview and RWA in these
data was substantially stronger than the meta-analytic average.
Since the full RWA measure was used here, this supports our con-
tention that the use of shortened RWA measures across studies
may contribute in part to the observed asymmetry. Unlike Van
Hiel, Cornelis, and Roets (2007), our full RWA was, presumably,
not contributing to the asymmetry between pathways here (i.e.,
by suppressing the DW–RWA relationship). We see no reason to
predict that shortened versions of the SDO scale would affect the
association with competitive worldview.
Another way forward in terms of addressing this issue is to cre-
ate puriﬁed measures of dangerous and competitive social world-
views that do not share excessive content overlap with SDO and
RWA. We have recently begun to address this issue by exploring
new ‘ideology-free’ measures of social worldviews (Perry & Sibley,
2010; Perry et al., in press).
4.4. Summary and conclusion
Our meta-analysis offers two main implications about the rela-
tionship between worldview beliefs and ideological attitudes in
the DPM. First, our analyses conﬁrmed that dangerous and com-
petitive worldviews are strong and independent predictors of
124 R. Perry et al. / Journal of Research in Personality 47 (2013) 116–127
RWA and SDO respectively, consistent with the DPM model. Our
analyses generally indicated that these relationships are consistent
across sample and societal characteristics. This is in line with a ba-
sic premise of the DPM – that the cognitive processes underlying
prejudice attitudes should respond to day-to-day situational
changes, but retain a stable baseline of association with one an-
other in the face of all but the most severe and chronic social envi-
ronmental change (e.g., Perrin & Smolek, 2009; see also Perry &
Sibley, 2011). Second, the spurious asymmetry in worldview-
ideological attitude associations provides an important impetus
for re-operationalizing the DPM worldview constructs to distin-
guish the variance explained by this component of the DPM from
the subsequent ideological attitude components indexed by RWA
and SDO. To this end, we present a reﬁned 20-item social world-
view measure that eschews those competitive worldview items
found to overlap with SDO (see Table 6).
There are almost certainly contextual and dispositional factors
other than social worldviews and personality that contribute to
variance in RWA and SDO. These may include shortened RWA
scales inadvertently tapping one or another of three underlying
sub-dimensions (e.g., Duckitt et al., 2010; Mavor et al., 2010). An-
other candidate emerging in recent literature is that of intelligence
(Heaven, Ciarrochi, & Leeson, 2011; Hodson & Busseri, 2012; Van
Hiel, Onraet, & De Pauw, 2010). Heaven et al. (2011) reported a
cross-lagged negative effect of generalized intelligence and verbal
ability on RWA. SDO was also predicted by low verbal ability.
Hodson and Busseri (2012) showed that childhood levels of general
intelligence were signiﬁcantly related to racism in adulthood and
this relationship was mediated by socially conservative ideological
beliefs. In a second experimental study, they showed that abstract
reasoning (a form of cognitive ability) was negatively related to
RWA, which in turn predicted prejudice against homosexuals.
These ﬁndings suggest that there may be information processing
differences underlying RWA and SDO – those high in RWA, for
example, appear to be ‘‘cognitive misers’’ that favor simpler (i.e.,
stereotype-conﬁrming) attributions (Heaven et al., 2011; Van Hiel,
Pandelaere, & Duriez, 2004).
In conclusion, this meta-analysis contributed to recent efforts to
systematically review research examining different aspects of the
DPM (Cohrs & Stelzl, 2010; Fischer et al., 2012; Sibley & Duckitt,
2008). Compiling a total sample of 46 studies with 12,939 partici-
pants across seven nations, we conﬁrmed that dangerous and com-
petitive worldview beliefs are strong and independent predictors
of respective concurrent levels of RWA and SDO (Duckitt et al.,
2002). These dual associations were asymmetric – the association
of dangerous worldview with RWA was moderate in size (partial
r= .37), whereas the association of competitive worldview with
SDO was stronger in magnitude (partial r = .53). Moreover, we con-
ﬁrmed that this asymmetry was spurious – probably resulting
from item content overlap between competitive worldview and
SDO, and possibly also from the multidimensional nature of
RWA. These dual patterns were also distinct – the opposing dan-
gerous worldview-SDO and competitive worldview-RWA associa-
tions were small in size (partial rs = .08, .11). All effects were
robust across published and unpublished undergraduate and com-
munity samples using different scales and across seven nations
varying in income inequality. These results provided good support
for the DPM, and indicate that, as expected, RWA and SDO are con-
sistently linked with distinct social schemas of society as danger-
ous and threatening (versus safe and secure), and competitive
and cut-throat (versus co-operative and characterized by mutually
Denotes article included in the meta-analysis.
Altemeyer, B. (1981). Right-wing authoritarianism. Winnipeg, Canada: University of
Altemeyer, B. (1988). Enemies of freedom: Understanding right-wing authoritarianism.
San Francisco: Jossey-Bass.
Construct deﬁnitions and items of the two social worldview scales – reﬁned (SWS-R).
Belief that the social world is a competitive jungle characterized by a ruthless, amoral struggle for resources and power in which might is right and winning is
everything versus belief that the social world is a place of cooperative harmony in which people care for, help, and share with one another
1. It’s a dog-eat-dog world where you have to be ruthless at times
2. There is really no such thing as ‘‘right’’ and ‘‘wrong.’’ It all boils down to what you can get away with
3. One of the most useful skills a person should develop is how to look someone straight in the eye and lie convincingly
4. My knowledge and experience tells me that the social world we live in is basically a competitive ‘‘jungle’’ in which the ﬁttest survive and succeed, in which power,
wealth, and winning are everything, and might is right
5. Basically people are objects to be quietly and coolly manipulated for one’s own beneﬁt
6. Life is not governed by the ‘‘survival of the ﬁttest.’’ We should let compassion and moral laws be our guide
7. It is better to he loved than to be feared
8. Do unto to others as you would have them do unto you, and never do anything unfair to someone else
9. Honesty is the best policy in all cases
10. One should give others the beneﬁt of the doubt. Most people are trustworthy if you have faith in them
Belief that the social world is a dangerous and threatening place in which good, decent people’s values and way of life are threatened by bad people versus belief that
the social world is a safe, secure and stable place in which almost all people are fundamentally good
1. My knowledge and experience tells me that the social world we live in is basically a safe, stable and secure place in which most people are fundamentally good
2. It seems that every year there are fewer and fewer truly respectable people, and more and more persons with no morals at all who threaten everyone else
3. Although it may appear that things are constantly getting more dangerous and chaotic, it really isn’t so. Every era has its problems, and a person’s chances of living a
safe, untroubled life are better today than ever before
4. Any day now chaos and anarchy could erupt around us. All the signs are pointing to it
5. There are many dangerous people in our society who will attack someone out of pure meanness, for no reason at all
6. The ‘‘end’’ is not near. People who think that earthquakes, wars, and famines mean God might be about to destroy the world are being foolish
7. My knowledge and experience tells me that the social world we live in is basically a dangerous and unpredictable place, in which good, decent and moral people’s
values and way of life are threatened and disrupted by bad people
8. Despite what one hears about ‘‘crime in the street,’’ there probably isn’t any more now than there ever has been
9. If a person takes a few sensible precautions, nothing bad is likely to happen to him or her; we do not live in a dangerous world
10. Every day as society become more lawless and bestial, a person’s chances of being robbed, assaulted, and even murdered go up and up
Note. Items reﬁned from the scale originally developed by Duckitt et al. (2002). Construct deﬁnitions are from Duckitt et al. (2002).
R. Perry et al. / Journal of Research in Personality 47 (2013) 116–127 125
Altemeyer, B. (1996). The authoritarian spectre. London: Harvard University Press.
Altemeyer, B. (1998). The other ‘‘authoritarian personality’’. In M. Zanna (Ed.).
Advances in experimental social psychology (Vol. 30, pp. 47–92). San Diego:
Asbrock, F. (2010). Unpublished data. Philipps-University of Marburg.
Asbrock, F., & Fritsche, I. (2010). Unpublished data. Philipps-University of Marburg.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.).
Lawrence Hillsdale, NJ: Lawrence Erlbaum.
Cohrs, J. C. (2005). Interindividuelle Unterschiede in Vorurteilen: Zur
Konstruktvalidität und Erklärungskraft von rechtsgerichtetem Autoritarismus
und sozialer Dominanzorientierung. Zeitschrift für Politische Psychologie, 13,
Cohrs, J. C., Moschner, B., Maes, J., & Kielmann, S. (2005). The motivational bases of
right-wing authoritarianism and social dominance orientation: Relation to
values and attitudes in the aftermath of September 11, 2001. Personality and
Social Psychology Bulletin, 31, 1–10.
Cohrs, J. C., & Stelzl, M. (2010). How ideological attitudes predict host society
members’ attitudes toward immigrants: Exploring cross-national differences.
Journal of Social Issues, 66, 673–694.
Crowson, H. M. (2009). Right-wing authoritarianism and social dominance
orientation as mediators of worldview beliefs on attitudes related to the war
on terror. Social Psychology, 40, 93–103.
D’Andrade, R. (1992). Schemas and motivation. In R. D’Andrade & C. Strauss (Eds.),
Human motives and cultural models (pp. 23–44). Cambridge, England:
Cambridge University Press.
Dallago, F., Mirisola, A., & Roccato, M. (2012). Predicting right-wing
authoritarianism via personality and dangerous worldview beliefs: Direct,
indirect, and interactive effects. The Journal of Social Psychology, 152, 112–127.
Dambrun, M., Guimond, S., & Duarte, S. (2002). The impact of hierarchy-enhancing
vs. attenuating academic major on stereotyping: The mediating role of
perceived social norm. Current Research in Social Psychology, 7, 114–136.
Dambrun, M., Kamiejski, R., Haddadi, N., & Duarte, S. (2009). Why does social
dominance orientation decrease with university exposure to the social
sciences? The impact of institutional socialization and the mediating role of
‘‘geneticism’’. European Journal of Social Psychology, 39, 88–100.
Derrick, J. L., & Murray, S. L. (2006). Unpublished data. University at Buffalo, The
State University of New York.
Dhont, K., & Van Hiel, A. (2009). We must not be enemies: Interracial contact and
the reduction of prejudice among authoritarians. Personality and Individual
Differences, 46, 172–177.
Doty, R., Peterson, B., & Winter, D. (1991). Threat and authoritarianism in the United
States, 1978–1987. Journal of Personality and Social Psychology, 61, 629–640.
Duckitt, J. (2000). Unpublished data. The University of Auckland.
Duckitt, J. (2001). A dual-process cognitive-motivational theory of ideology and
prejudice. In M. P. Zanna (Ed.). Advances in Experimental Social Psychology (Vol.
33, pp. 41–113). New York: Academic Press.
Duckitt, J., & Bizumic, B. (in press). Authoritarianism–conservatism–traditionalism:
Evidence for the multidimensionality of threat-based ideological attitudes.
Duckitt, J., Bizumic, B., Krauss, S. W., & Heled, E. (2010). A tripartite approach to
right-wing authoritarianism: The authoritarianism–conservatism–
traditionalism model. Political Psychology, 31, 685–715.
Duckitt, J., & Fisher, K. (2003). The impact of social threat on worldview and
ideological attitudes. Political Psychology, 24, 199–222.
Duckitt, J., & Sibley, C. G. (2010). Personality, ideology, prejudice, and politics: A
dual process motivational model. Journal of Personality, 78, 1861–1894.
Duckitt, J., Wagner, C., du Plessis, I., & Birum, I. (2002). The psychological bases of
ideology and prejudice. Testing a dual process model. Journal of Personality and
Social Psychology, 83, 75–93.
Federico, C. M., Hunt, C. V., & Ergun, D. (2009). Political expertise, social worldviews,
and ideology: Translating ‘‘competitive jungles’’ and ‘‘dangerous worlds’’ into
ideological reality. Social Justice Research, 22, 259–279.
Federico, C. M., Weber, C. R., Ergun, D., & Hunt, C. V. (in preparation). Reconsidering
the ideological structure of moral foundations. Journal of Experimental Social
Fischer, R., Hanke, K., & Sibley, C. G. (2012). Cultural and institutional determinants
of social dominance orientation: A cross-cultural meta-analysis of 27 societies.
Political Psychology, 33, 437–467.
Funke, F. (2005). The dimensionality of right-wing authoritarianism: Lessons from
the dilemma between theory and measurement. Political Psychology, 26,
Gatto, J., Dambrun, M., Kerbrat, C., & De Oliveira, P. (2010). Prejudice in the police.
On the processes underlying the effects of selection and group socialisation.
European Journal of Social Psychology, 40, 252–269.
Guimond, S., Dambrun, M., Michinov, N., & Duarte, S. (2003). Does social dominance
generate prejudice? Integrating individual and contextual determinants of
intergroup cognitions. Journal of Personality and Social Psychology, 84, 697–721.
Heaven, P. C. L., Ciarrochi, J., & Leeson, P. (2011). Cognitive ability, right-wing
authoritarianism, and social dominance orientation: A ﬁve-year longitudinal
study amongst adolescents. Intelligence, 39, 15–21.
Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. Orlando, FL:
Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world?
Behavioral and Brain Sciences, 33, 61–135.
Hodson, G., & Busseri, M. A. (2012). Bright minds and dark attitudes: Lower
cognitive ability predicts greater prejudice through right-wing ideology and
low intergroup contact. Psychological Science, 23, 187–195.
Jugert, P., & Duckitt, J. (2009). A motivational model of authoritarianism:
Integrating personal and situational determinants. Political Psychology, 30,
Lipsey, M. W., & Wilson, D. B. (2001). Practical meta-analysis. London: Sage.
Mavor, K. I., Louis, W. R., & Sibley, C. G. (2010). A bias-corrected exploratory and
conﬁrmatory factor analysis of right-wing authoritarianism: Support for a
three-factor structure. Personality and Individual Differences, 48, 28–33.
Mirisola, A., Di Stefano, G., & Falgares, G. (2007). I processi duali ideologici nel
contesto italiano. In F. Di Maria, G. Di Stefano, & G. Falgares (Eds.), Psiche e
società: La polis siciliana tra conservazione e trasformazione (pp. 29–435). Milan:
Newcomb, T. (1943). Personality and social change: Attitude formation in a student
community. New York: Dryden.
Newcomb, T., Koenig, K. E., Flacks, R., & Warwick, D. P. (1967). Persistence and
change: Bennington College and its students after 25 years. New York: Wiley.
Overton, R. C. (1998). A comparison of ﬁxed-effects and random (mixed-effects)
models for meta-analysis tests of moderator variable effects. Psychological
Methods, 3, 354–379.
Onraet, E., Van Hiel, A., Roets, A., & Cornelis, I. (2011). The closed mind: ‘Experience’
and ‘cognition’ aspects of openness to experience and need for closure as
psychological bases for right-wing attitudes. European Journal of Personality, 25,
Perrin, A. J., & Smolek, S. J. (2009). Who trusts? Race, gender, and the September 11
rally effect among young adults. Social Science Research, 38, 134–145.
Perry, R., & Sibley, C. G. (2010). Dangerous and competitive schemas: A new
frequency estimation index of the dual process model’s social worldviews
component. Personality and Individual Differences, 49, 983–988.
Perry, R., & Sibley, C. G. (2011). Social dominance orientation: Mapping a baseline
individual difference component across self-categorizations. Journal of
Individual Differences, 32, 110–116.
Perry, R., Sibley, C. G., & Duckitt, J. (in press). A comparison of broad-bandwidth and
frequency-speciﬁc measures of competitive and dangerous worldview.
Personality and Individual Differences.
Pratto, F., Sidanius, J., Stallworth, L., & Malle, B. (1994). Social dominance
orientation: A personality variable predicting social and political attitudes.
Journal of Personality and Social Psychology, 67, 741–763.
Raudenbush, S. W. (1994). Random effects models. In H. Cooper & L. V. Hedges
(Eds.), The handbook of research synthesis (pp. 301–321). New York: Russell Sage.
Roberts, B. W., Caspi, A., & Mofﬁtt, T. E. (2003). Work experiences and personality
development in young adulthood. Journal of Personality and Social Psychology,
Robertson, A. (2006). In search of a theoretical explanation for the relationship
between religiosity and prejudice among self-identiﬁed Christians. Unpublished
doctoral dissertation, Victoria University of Wellington.
Rosenthal, R. (1979). The ‘‘ﬁle drawer problem’’ and tolerance for null results.
Psychological Bulletin, 86, 638–641.
Ross, M. (1993). The culture of conﬂict. New Haven, CT: Yale University Press.
Sales, S. M. (1972). Economic threat as a determinant of conversion rates in
authoritarian and nonauthoritarian churches. Journal of Personality and Social
Psychology, 23, 420–428.
Sales, S. M. (1973). Threat as a factor in authoritarianism: An analysis of archival
data. Journal of Personality and Social Psychology, 28, 44–57.
Sibley, C. G. (2007). Unpublished data. The University of Auckland.
Sibley, C. G. (2010). The dark duo of post-colonial ideology: A model of symbolic
exclusion and historical negation. International Journal of Conﬂict and Violence, 4,
Sibley, C. G., & Duckitt, J. (2006). Unpublished data. The University of Auckland.
Sibley, C. G., & Duckitt, J. (2008). Personality and prejudice. A meta-analysis and
theoretical review. Personality and Social Psychology Review, 12, 248–279.
Sibley, C. G., & Duckitt, J. (2009). Big-ﬁve personality, social worldviews, and
ideological attitudes: Further tests of a dual process cognitive-motivational
model. The Journal of Social Psychology, 149, 545–561.
Sibley, C. G., & Duckitt, J. (2012). Personality geneses of authoritarianism: The form
and function of openness to experience. In F. Funke, T. Petzel, J. C. Cohrs, & J.
Duckitt (Eds.), Perspectives on Authoritarianism (pp. 169–199). Weisbaden,
Sibley, C. G., Liu, J. H., Duckitt, J., & Khan, S. H. (2008). Social representations of
history and the legitimation of social inequality: The form and function of
historical negation. European Journal of Social Psychology, 38, 542–565.
Sibley, C. G., & Perry, R. (2010). An opposing process model of benevolent sexism.
Sex Roles, 62, 438–452.
Sibley, C. G., Wilson, M. S., & Duckitt, J. (2007a). Effects of dangerous and
competitive worldviews on right-wing authoritarianism and social dominance
orientation over a ﬁve-month period. Political Psychology, 28, 357–371.
Sibley, C. G., Wilson, M. S., & Duckitt, J. (2007b). Antecedents of men’s hostile and
benevolent sexism: The dual roles of social dominance orientation and right-
wing authoritarianism. Personality and Social Psychology Bulletin, 33, 160–172.
Sibley, C. G., Wilson, M. S., & Robertson, A. (2007). Differentiating the motivations
and justiﬁcations underlying individual differences in Pakeha opposition to
bicultural policy: Replication and extension of a predictive model. New Zealand
Journal of Psychology, 36, 25–33.
126 R. Perry et al. / Journal of Research in Personality 47 (2013) 116–127
Sidanius, J., Van Laar, C., Levin, S., & Sinclair, S. (2004). Ethnic enclaves and the
dynamics of social identity on the college campus: The good, the bad, and the
ugly. Journal of Personality and Social Psychology, 87, 96–110.
Sinclair, S., Sidanius, J., & Levin, S. (1998). The interface between ethnic and social
system attachment: The differential effects of hierarchy-enhancing and
hierarchy-attenuating environments. Journal of Social Issues, 54, 741–757.
Smith, A. G., & Winter, D. G. (2002). Right-wing authoritarianism, party
identiﬁcation, and attitudes toward feminism in student evaluations of the
Clinton–Lewinsky story. Political Psychology, 23, 355–383.
Stellmacher, J., & Petzel, T. (2005). Authoritarianism as a group phenomenon.
Political Psychology, 26, 245–274.
Strauss, C. (1992). Models and motives. In R. D’Andrade & C. Strauss (Eds.), Human
motives and cultural models (pp. 1–20). Cambridge, England: Cambridge
United Nations Development Program (UNDP) (2010). Statistics of the human
development report.<http://hdr.undp.org/en/statistics/>. Retrieved June, 2011.
Van Hiel, A. (2010a). Unpublished data. Ghent University.
Van Hiel, A. (2010b). Unpublished data. Ghent University.
Van Hiel, A., Cornelis, I., & Roets, A. (2007). The intervening role of social
worldviews in the relationship between the ﬁve-factor model of personality
and social attitudes. European Journal of Personality, 21, 131–148.
Van Hiel, A., Cornelis, I., Roets, A., & De Clercq, B. (2007). A comparison of various
authoritarianism scales in Belgian Flanders. European Journal of Personality, 21,
Van Hiel, A., & De Clercq, B. (2009). Authoritarianism is good for you: Right-wing
authoritarianism as a buffering factor for mental distress. European Journal of
Personality, 23, 33–50.
Van Hiel, A., Duriez, B., & Kossowska, M. (2006). The presence of left-wing
authoritarianism in Western Europe and its relationship with conservative
ideology. Political Psychology, 27, 769–793.
Van Hiel, A., Onraet, E., & De Pauw, S. (2010). The relationship between social–
cultural attitudes and behavioral measures of cognitive style: A meta-analytic
integration of studies. Journal of Personality, 78, 1765–1800.
Van Hiel, A., Pandelaere, M., & Duriez, B. (2004). The impact of need for closure on
conservative beliefs and racism: Differential mediation by authoritarian
submission and authoritarian dominance. Personality and Social Psychology
Bulletin, 30, 824–837.
Weber, C., & Federico, C. M. (2007). Interpersonal attachment and patterns of
ideological belief. Political Psychology, 28, 389–416.
Webster, R. J., Whitley, B. E. Jr., & Miller, K. A. (2006). Adult attachment style and
prejudice. In Poster session presented at the annual meeting of the Midwestern
Psychological Association, Chicago, IL.
Wilkinson, R. G., & Pickett, K. E. (2009). Income inequality and social dysfunction.
Annual Review of Sociology, 35, 493–511.
Wilson, M. S. (2006). Unpublished data. Victoria University of Wellington.
Wilson, M. S. (2010a). Unpublished data. Victoria University of Wellington.
Wilson, M. S. (2010b). Unpublished data. Victoria University of Wellington.
Wilson, M. S. (2010c). Unpublished data. Victoria University of Wellington.
R. Perry et al. / Journal of Research in Personality 47 (2013) 116–127 127