Values and prejudice: Predictors of attitudes towards Australian Aborigines

Abstract

This study related prejudice towards Australian Aborigines to value types assessed by the Schwartz Value Survey (SVS), right-wing authoritarianism (RWA), and social dominance orientation (SDO). One hundred and forty-eight students in Adelaide, South Australia, completed a Modern Racism Scale adapted for Australian Aborigines, the SVS, the RWA Scale, and the SDO Scale. We predicted that prejudice would be positively related to the importance of self-enhancement and conservation values from the SVS such as power and security and negatively related to the importance of self-transcendence values such as universalism and benevolence. Relations between the prejudice measures and RWA and SDO were also expected to reflect their degree of overlap with discrete value types from the SVS. These predictions were supported. Results were discussed in relation to the importance of considering how prejudice relates to a person's specific value priorities as well as to more general value variables such as RWA and SDO.

There is now an extensive literature on prejudice and stereotyping that owes much to the classic analysis by Allport (1954) but that has now gone beyond that analysis in many respects (Dovidio, Glick, & Rudman, 2005; Fiske, 1998). Many of Allport's key insights have been taken up and framed in relation to important new developments from social cognition, group processes, and personality dynamics. Relatively neglected, however, is the analysis of prejudice as it relates to a person's specific value priorities. Allport (1954) drew attention to anti-egalitarianism and conformity pressures in his discussion, and the subsequent development of the concepts of ambivalent racism (Katz & Haas, 1988), symbolic racism (Sears & Kinder, 1971; Sears & Henry, 2005), aversive racism (Dovidio & Gaertner, 2004; Gaertner & Dovidio, 1986), and ambivalent sexism (Glick & Fiske, 2001a, 2001b) referred to general value dimensions such as individualism, Protestant Ethic values, egalitarianism, humanitarianism, and dominance that, depending on the contributors' particular focus, were assumed to underlie modern and more subtle forms of prejudice. Biernat, Vescio, Theno, and Crandall (1996) discussed these various concepts and provided evidence that both individualistic and humanistic/egalitarian values are involved in prejudiced responses as well as other values depending on context and outgroup. Other researchers developed the concepts of right-wing authoritarianism (Altemeyer, 1981, 1998), social dominance orientation (Pratto, Sidanius, Stallworth, & Malle, 1994), and dangerous versus competitive world views (Duckitt, 2001, 2005) and related these variables to prejudice.

These various contributions are important additions to the literature but their focus is more on a limited set of values and on relatively general belief and attitude systems rather than on a broader range of specific value types. They do not set research on values and prejudice within a framework that distinguishes between values in terms of their content and motivational goals and that draws attention to the way values are organised within individuals. The research to be described in the current article takes a step in filling that gap by investigating relations between a wider set of specific value types and a measure of prejudice towards Australian Aborigines. The research also investigates relations between prejudice, right-wing authoritarianism (RWA) and social dominance orientation (SDO), on the assumption that both RWA and SDO partly reflect the value priorities that people hold.

Values

Values may be conceived as general beliefs about desirable ways of behaving or about desirable general goals. They are held by persons and vary in their importance for self. They are abstract in nature, transcend situations, and are linked to the affective and motivational systems. People use them as guiding principles in their lives. They are important influences on the way a person construes situations, affecting the more specific attitudes a person holds and influencing planning and decision-making. They are central to a person's self concept, affecting thought and action in many different ways and forming part of the causal texture of emotional responses when the pursuit of goals relating to important values is either frustrated or fulfilled. These defining characteristics of values are distilled from the seminal contributions of Rokeach (1973) and from later theoretical ideas proposed by Feather (1975, 1999, 2005) and Schwartz (1992, 1996).

How then might values influence prejudice? All of the contributions to the conceptual analysis of values that were just listed consider values to be important influences on the attitudes that people hold. For example, Rokeach (1973) argued that values are central to the self-concept and key influences on a person's system of beliefs and attitudes. Feather (1975, 1999, 2005) conceived of values as abstract structures that are linked to attitudes and valences via associative networks, sensitising the way a person construes situations in terms of what is attractive and what is aversive. Schwartz (1992) conceived of values as “desirable, transituational goals, varying in importance, that serve as guiding principles in people's lives” (p. 2). Functional approaches to the analysis of attitudes have also recognised value expression as an important function of attitudes among other functions that attitudes may serve (e.g., Katz, 1960; Smith, Bruner, & White, 1956).

The expression of values via specific attitudes may be taken to reflect the motivational base that underlies the values. Thus, Rokeach (1973) argued that the long-range functions of values are to give expression to basic human needs. Schwartz (1992, 1996) proposed different motivational goals associated with different value types. Feather (1995, 1999, 2004) assumed that values are like needs in their ability to motivate goal-directed behaviour by inducing positive or negative valences on objects and events within the immediate situation. Maio and Olson (2001a, 2001b) also argued that different attitude objects are relevant to different values and that the motivational goal of the value has to be taken into account when the functional role of a value is considered.

This analysis implies that prejudiced attitudes may be related to a range of values that differ in their motivational goals. What are the kinds of values that we should explore? In the current study we focused on the analysis by Schwartz because his measurement instrument, the Schwartz Value Survey (SVS), was the main direct measure of value importance that we used. Schwartz (1992, 1996) developed a theory of values that identified ten different value types, each of which was associated with a different motivational goal. He proposed that the value types formed a circular structure ( Figure 1) with adjacent values around the circle being compatible when they are simultaneously pursued (e.g., power and achievement) and value types on opposite sides of the circle coming into conflict when simultaneously pursued (e.g., stimulation and security). The circular structure was supported by analyses of importance ratings for the values listed in the SVS using Guttman's (1968) smallest space analysis. The ratings were obtained from students and teachers in many different countries. According to Schwartz, most value conflicts for individuals involve an opposition between value types that express openness to change (e.g., stimulation, self direction) and value types that express conservation (e.g., conformity, tradition, security), and between value types that express a concern with self-enhancement (e.g., achievement, power) and value types that express a concern with self-transcendence (e.g., benevolence, universalism).

Figure 1. The circular structure of value systems

Using the SVS, Sagiv and Schwartz (1995) showed that the readiness of Israeli Jewish teachers for contact with Israeli Arabs correlated positively with emphasizing universalism and self-direction values and negatively with emphasizing tradition, conformity, and security values. They noted that their findings were generally consistent with results reported by Rokeach (1973) using the Rokeach Value Survey (RVS) in research on attitudes towards Black Americans when the values on the RVS that discriminated between racists and nonracists were classified into the value types identified by Schwartz.

We also included measures of right-wing authoritarianism (RWA) and social dominance orientation (SDO) in the current study on the assumption that subsets of the value types from the SVS feature in measures of these two variables. We expected that these subsets of values may underlie relations between RWA and SDO and prejudiced attitudes. Altemeyer (1981, 1998) conceived of right-wing authoritarianism as a personality variable involving authoritarian submission, authoritarian aggression, and conventionalism. He proposed that authoritarians would be more disposed to submit to established authority, to control the behaviour of others through punishment, and to accept and be committed to the traditional social norms of their society when compared to non-authoritarians. Altemeyer developed the Right-Wing Authoritarianism Scale (RWA Scale) to measure this syndrome of authoritarian characteristics that he assumed were the product of social learning. The measure of social dominance orientation that we used (the SDO Scale) was assumed by Pratto et al. (1994), to tap a “general attitudinal orientation towards intergroup relations, reflecting whether one generally prefers such relations to be equal, versus hierarchical” and the “extent to which one desires that one's ingroup dominate and be superior to outgroups” (p. 742).

Both right-wing authoritarianism and social dominance orientation have been shown to predict prejudice (e.g., Altemeyer, 1981, 1998; Duckitt, 2005, 2006; Duckitt & Sibley, 2007; Guimond, Dambrun, Michinov, & Duarte, 2003; Heaven & St. Quintin, 2003; McFarland & Adelson, 1997; Sibley, Wilson, & Duckitt, 2007; Whitley, 1999; Whitley & Kite, 2006). They have also been shown to be related to ideological beliefs and values (e.g., Heaven, Organ, Supavadeeprasit, & Leeson, 2006). If we focus on the value types from the SVS, Feather (1996, 1998, 1999, 2005) consistently found in different studies that RWA scores were positively correlated with the importance assigned to tradition, conformity, security, and power values from the SVS, and negatively correlated with hedonism, stimulation, self-direction, and universalism values from the SVS (see also Altemeyer, 1998; Cohrs, Moschner, Maes, & Kielmann, 2005; Duriez & Van Hiel, 2002; Rohan & Zanna, 1996). Thus the evidence shows that right-wing authoritarians are more tuned in to conservation values rather than to values associated with openness to change.

The pattern of values relating to social dominance orientation overlaps with these values but also shows some differences. Research shows that people high in SDO tend to assign more importance to power values from the SVS, reflecting more concern with self-enhancement values (Altemeyer, 1998; Cohrs et al., 2005). However, like the high RWA scorers, they also show less concern with universalism and benevolence values from the SVS. As Duckitt (2005) concluded, “people high in RWA value security, order, and control whereas people high in SDO value power, dominance, and group enhancement” (p. 404).

Hypotheses

We investigated prejudice towards Australian Aborigines in the current study, measured in a subtle form by an adaptation by Augoustinos, Ahrens, and Innes (1994) of the Modern Racism Scale (McConahay & Hough, 1976). As Augoustinos et al. point out, Aborigines are a highly salient group in Australian society and they are often the targets of discrimination and prejudice (e.g., Larsen, 1981; Marjoribanks & Jordan, 1986). They occupy positions of low status and are disadvantaged when compared with the rest of the Australian community. Using the adaptation of the Modern Racism Scale, Pedersen and Walker (1997) found that more than half of their Perth sample scored above the midpoint on this measure while only approximately one quarter of their sample scored above the midpoint on more blatant, old-fashioned prejudice. Pedersen, Griffiths, Contos, Bishop, and Walker (2000) also found stronger evidence for modern prejudice when compared with blatant prejudice in their Perth and Kalgoorlie samples. Qualitative research with South Australian university students has also revealed subtle and covert forms of prejudice in statements made about Australian Aborigines (e.g, Augoustinos, Rapley & Tuffin, 1999; Sanson, Augoustinos, Gridley, Kyrios, Reser, & Turner, 1998).

In accord with our previous discussion, we assumed that prejudiced attitudes would express in part the values that are engaged in a situation, consistent with the motivational base that underlies the values. Prejudiced attitudes towards Australian Aborigines were expected to be positively related to both conservation value types from the SVS such as security, and to self-enhancement value types from the SVS such as power, and negatively related to self-transcendence values from the SVS, such as universalism. For prejudiced people, Australian Aborigines would more likely be perceived as a threat to their security and their power or status. On the other hand, those who are relatively unprejudiced would be more likely to perceive Australian Aborigines as having rights to fair treatment without discrimination, reflecting stronger universalism values held by those making the judgements.

Some of the values just mentioned may have direct effects on prejudice. We also assumed, however, that there may also be indirect effects that are mediated by right-wing authoritarianism (RWA) and by social dominance orientation (SDO). As we noted previously, research has shown that there are value differences associated with RWA and SDO. These differences can be related to the way RWA and SDO are conceptualised in the dual process model developed by Duckitt (2001, 2005). He proposed that high levels of RWA expressed a motivational goal for social cohesion and collective security and that low levels of RWA expressed more concern with independence and autonomy. This distinction would map into the distinction made by Schwartz (1992, 1996) between the conservation value dimension (incorporating conformity, tradition, and security value types) and the openness to change value dimension (incorporating self-direction and stimulation value types). Duckitt (2001, 2005) also proposed that high levels of SDO expressed a motivational goal for group-based dominance and superiority and that low levels of SDO expressed more concern with egalitarianism and prosocial concern. This distinction would map into Schwartz's (1992, 1996) distinction between the self-enhancement value dimension (incorporating power and achievement value types) and the self-transcendence value dimension (incorporating benevolence and universalism value types).

Thus, RWA and SDO were expected to be associated with different value profiles. Conservation values were expected to be positive predictors of RWA and both self-transcendence and openness to change values to be negative predictors. Self-enhancement values were expected to positively predict SDO and self-transcendence values to negatively predict this variable.

But how might RWA and SDO predict prejudice towards Australian Aborigines? Increasing attention is now being paid to differences in the targets of prejudice. For example, Fiske, Cuddy, Glick, and Xu (2002) distinguished between different targets of prejudice depending on where they are located on dimensions of competence and warmth (see also Glick & Fiske, 2001a, 2001b). In their analysis the combination of low competence and low warmth would be associated with contemptuous prejudice that, for example, might be directed towards low status welfare recipients and poor people. Other types of combination would lead to paternalistic prejudice (low competence, high warmth) that might be directed towards elderly people and disabled people; admiration (high competence, high warmth) that might be directed towards the in-group and close allies; and envious prejudice (high competence, low warmth) that might be directed towards high status and competitive people (they give Asians, Jews, rich people, and feminists as examples).

Duckitt and Sibley (2007) also distinguished different targets of prejudice in terms of whether they were derogated, perceived as dangerous, or groups that dissented from, challenged, or opposed mainstream norms and values. Their research supported this classification and also showed that ratings of warmth correlated negatively with SDO for groups in the “derogated” classification, negatively with RWA for groups in the “dangerous” classification, and negatively with both SDO and RWA for groups in the “dissident” classification. Duckitt and Sibley also found positive correlations between both RWA and SDO and negative attitudes towards New Zealand minority groups (Anti-Asian, Anti-Pacific Islanders, Anti-Maori).

We expected that relations between prejudiced attitudes towards Australian Aborigines and RWA and SDO would mirror those just noted for the New Zealand minorities. Those with high prejudice scores might perceive Australian Aborigines as a dissident group that threatens existing social cohesion, order and stability and that also challenges social inequality and relative dominance. Thus, both RWA and SDO were expected to predict more prejudice towards this Australian target.

Finally, given the ideological basis of RWA and SDO in relation to the values each reflected, we expected that these two variables would partially mediate the effects of different value types on prejudice. Thus, an important aim of our study was to investigate the extent to which RWA and SDO mediate the effects of discrete value types on prejudice towards Australian Aborigines, thereby extending the type of analysis provided by Duckitt and his colleagues.

Method

Participants and procedure

The participants in the study were 148 students (42 men, 105 women, and 1 individual who did not specify gender) who were sampled from the introductory classes in psychology at Flinders University, Adelaide, Australia in 2005. The mean age of the sample was 22.05 years (SD = 7.49). Participants completed a questionnaire that contained the 57-item Schwartz Value Survey (SVS; Schwartz, 1992), the 30-item Right-Wing Authoritarianism Scale (RWA; Altemeyer & Hunsberger, 1992); the 14-item social dominance scale (SDO; Pratto et al., 1994); and the 7-item Modern Racism Scale adapted for Australian Aborigines (Augoustinos et al., 1994). Questionnaires were completed anonymously and participants were subsequently provided with feedback about the study.

Value measures

Schwartz Value Survey

Participants completed the SVS with the usual instructions (Schwartz, 1992). Each of the 57 values is accompanied by a short descriptive phrase. Participants rated how important each value was as “a guiding principle in your life” using a 9-point rating scale ranging from −1 to 7 (−1 = opposed to my values, 0 = not important, 3 = important, 6 = very important, 7 = of supreme importance).

The value types with their motivational goals, constituent values, and internal reliabilities were as follows:

•	Power: social status and prestige, control or dominance over people and resources (social power, wealth, authority, preserving my public image; α = .75);
•	Achievement: personal success through demonstrating competence according to social standards (ambitious, influential, capable, successful; α = .53);
•	Hedonism: pleasure and sensuous gratification (pleasure, enjoying life, self-indulgent; α = .65);
•	Stimulation: excitement, novelty, and challenge in life (an exciting life, a varied life, daring; α = .70);
•	Self-Direction: independent thought and action, choosing, creating, exploring (freedom, creativity, independent, choosing own goals, curious; α = .61);
•	Universalism: understanding, appreciation, tolerance and protection for the welfare of all (equality, a world at peace, unity with nature, wisdom, a world of beauty, social justice, broadminded, protecting the environment; α = .77);
•	Benevolence: preservation and enhancement of the welfare of people with whom one is in frequent personal contact (loyal, honest, helpful, responsible, forgiving; α = .68);
•	Tradition: respect, commitment, and acceptance of the customs and ideas provided by traditional culture or religion (respect for tradition, moderate, humble, accepting my portion in life, devout; α = .60);
•	Conformity: restraint in actions, inclinations, and impulses likely to upset or harm others and violate social expectations or norms (politeness, self-discipline, honouring of parents and elders, obedient; α = .70);
•	Security: safety, harmony, and stability of society, of relationships, and of self (social order, national security, reciprocation of favours, family security, clean; α = .62).

To obtain a score for each value type, we averaged the importance ratings for the values that comprised the value type.

Right-Wing Authoritarianism (RWA)

The items in the 30-item RWA Scale (Altemeyer & Hunsberger, 1992) cover authoritarian submission, authoritarian aggression, and conventionalism and are balanced so as to control for acquiescence response set. Participants answered each item by using the following 8-point scale: +1, I slightly agree; +2, I moderately agree; +3, I strongly agree; +4, I very strongly agree; −1, I slightly disagree; −2, I moderately disagree; −3, I strongly disagree; −4, I very strongly disagree. Responses were recoded by adding a constant of 5 and negative items were reverse coded. Total scores could then range from 30 to 270 with a midpoint of 150 (α = .90).

Social Dominance Orientation (SDO)

Examples of items in the 14-item SDO Scale (Pratto et al., 1994) are “Some people are just more worthy than others,” and “Social equality should be a major goal for everyone”. Items in the scale are balanced so as to control for acquiescence response set. Participants answered each item by using a 7-point scale anchored by the following labels: Disagree a lot (1) and Agree a lot (7). After recoding the seven negatively worded items, total scores could range from 14 to 98 with a midpoint of 56 (α = .85).

Prejudice measure

Modern Racism Scale

The 7-item measure of prejudice towards Australian Aborigines (Augoustinos et al., 1994) contains items such as the following: “Indigenous people have more influence upon government policy than they ought to have”, “Indigenous people are getting too demanding in their push for land rights”, and “Discrimination against indigenous people is no longer a problem in Australia”. Participants answered each item by using a 7-point scale anchored by the following labels: Disagree (1) and Agree (7). After recoding one negatively worded item total scores could range from 7 to 49 with a midpoint of 28 (α = .89).

Results

Means and standard deviations

Table I presents the means and standard deviations (SDs) for the value measures and the prejudice measure for male and female participants as well as the results of t tests that were used to test gender differences.

Table I. Means and standard deviations for study variables in relation to gender

Variable	Male		Female		t
	M	SD	M	SD
Value measures
Value types
Power	2.91	1.50	1.94	1.21	4.03***
Achievement	4.47	.93	4.50	.97	−.14
Hedonism	4.37	1.18	4.28	1.17	.43
Stimulation	3.86	1.20	3.75	1.39	.47
Self-direction	4.74	.85	4.70	.92	.26
Universalism	4.16	1.19	4.57	.94	−2.18*
Benevolence	4.73	1.13	5.00	.80	−1.61
Tradition	3.42	1.28	2.85	1.15	2.62**
Conformity	4.35	1.30	4.19	1.05	.78
Security	4.38	1.04	4.08	.95	1.67
Right-Wing Auth. (RWA)	142.63	30.41	127.45	32.79	2.50*
Social Dom. Orient. (SDO)	45.34	12.09	36.00	13.02	3.97***
Prejudice measure
Modern racism	24.33	7.73	20.56	8.99	2.58*
Note. Ns for male participants ranged from 40 to 42; Ns for female participants ranged from 93 to 105. Values were rated for importance on a −1 to 7 scale as described. Scale midpoints were 150 for RWA, 56 for SDO and 28 for Modern Racism.
*p < .05; **p < .01; ***p < .001, all two-tailed.

These results show that men assigned significantly more importance than women to power and tradition values. They also scored higher on right-wing authoritarianism (RWA) and on social dominance orientation (SDO). Women rated universalism values as more important for themselves.

Men scored significantly higher than women on prejudice towards Australian Aborigines (Modern Racism), although mean scores were below the midpoint of the scale (28) denoting relatively low prejudice.

Correlations for value measures and prejudice measures

Table II presents the correlations between the value measures and the prejudice measures, after using partial correlation to control for the effects of the gender differences that we just noted. The correlations involving the 10 value types were corrected for individual differences in scale use by computing each participant's mean importance rating for the set of 57 values and then centering the rating for each item around that mean rating. This procedure provides an ipsatised measure of value importance for each individual, where value importance is relative to a person's mean rating for the total set of values in the SVS.

Table II. Partial correlations between value and prejudice variables controlling for gender differences

Variable	RWA	SDO	Modern racism
Value measures
Value types
Power	.09	.39***	.38***
Achievement	−.15	.23*	−.00
Hedonism	−.20*	.39***	.11
Stimulation	−.12	.14	.11
Self-direction	−.42***	.09	−.15
Universalism	−.34***	−.50***	−.25**
Benevolence	−.10	−.36***	−.37***
Tradition	.39***	−.23*	−.07
Conformity	.43***	−.01	.05
Security	.25**	.29**	.23*
Right-Wing Authoritarianism	–	.18	.48***
Social Dominance Orientation		–	.50***
Note. N = 113 for partial correlations.
*p < .05; **p < .01; ***p < .001, all two-tailed.

The partial correlations in Table II show that the Modern Racism measure of prejudice towards Australian Aborigines was positively related to the importance assigned by participants to power and security values, and to their scores on the RWA and SDO scales. In contrast, this prejudice measure was negatively related to the importance of universalism and benevolence values. All of these correlations were statistically significant.

The partial correlations in Table II also show that there were different patterns of relations between RWA, and SDO and the importance ratings for each value type. Consistent with previous findings (Feather, 1996, 1998, 2005), authoritarianism was linked to a pattern of values in which tradition, conformity, and security values were emphasised and pleasure, self-direction and universalism values were de-emphasised. Social dominance scores were positively related to the importance of power, achievement, hedonism, and security values but participants scoring higher on social dominance tended to de-emphasize universalism, benevolence, and tradition values. All of these correlations were statistically significant.

In general the correlations reported in Table II held up for the separate male and female samples. An exception was the Stimulation/Modern Racism correlation. The importance of stimulation values was a positive predictor of Modern Racism for the female participants but a negative predictor for the male participants (rs of .24 and−.25 respectively, Fisher's z = 2.62, p < .01) This result requires replication with a larger sample of male participants.

Path analysis

We used simultaneous equations path analysis to test a model in which both RWA and SDO were included as mediators between the value types and prejudice towards Australian Aborigines (Modern Racism). In this model the conservation value types (security, conformity, and tradition) were positive predictors of RWA, and the openness to change value types (self-direction and stimulation) were negative predictors. The self-enhancement value types (power, achievement and hedonism) were positive predictors of SDO and the self-transcendence value types (benevolence and universalism) were negative predictors. The analysis used the maximum likelihood procedure available in LISREL, version 8.3 (Jöreskog & Sörbom, 1999).

The initial model test suggested that the model did not fit the data particularly well (χ² (21) = 41.76, p = .004 (CFI = .96, NNFI = .86, RMSEA = .097, RMR = .053). Subsequent modifications were made to the model by dropping the non-significant paths involving stimulation and benevolence and adding in a path from universalism to RWA as well as direct paths from power and tradition to Modern Racism. This resulted in the model in Figure 2, with LISREL estimates of χ² (14) = 12.02, p = .61 (CFI = 1.00, NNFI = 1.01, RMSEA = 0.0, RMR = .022), all of which were at acceptable levels.

Figure 2. Final path diagram presenting significant paths that link value types, SDO, and RWA to Modern Racism. Path coefficients are standardised betas. *p < .05; **p < .01; ***p < .001

The statistically significant path coefficients presented in Figure 2 show that social dominance orientation (SDO) was positively predicted by the importance of achievement, hedonism, and power values and negatively predicted by the importance of universalism values. SDO was a positive predictor of prejudice towards Australian Aborigines (Modern Racism). There was also a direct positive path linking the importance of power values to Modern Racism.

Right-wing authoritarianism (RWA) was positively predicted by the importance of security, tradition, and conformity values and negatively predicted by the importance of universalism and self-direction values. RWA was a positive predictor of prejudice towards Australian Aborigines (Modern Racism). There was also a direct negative path linking the importance of tradition values to Modern Racism.

These results therefore provide evidence that SDO and RWA mediated some of the effects of values on prejudice and they also showed that power and tradition values had direct effects on prejudice.

Discussion

Our results help to fill a gap in the analysis of prejudice by providing new information about how prejudiced attitudes towards Australian Aborigines are related to specific value types. We argued that past discussions of values and prejudice have tended to focus on a limited number of general value dimensions such as egalitarianism and individualism that are not integrated into a formal theory of values and that are presented without recognition of the existence of other value types that differ in their content and that also have the potential to influence attitudes. The correlations clearly show that power and security values positively predicted prejudice relating to racial attitudes, and that universalism values that include equality and social justice, and benevolence values that include honesty and forgiveness, were negative predictors of these forms of prejudice. Thus, participants who valued social status, prestige, and dominance over others, as well as safety and stability for self and society, were more likely to express prejudiced attitudes. Those who valued tolerance and protection for the welfare of society as a whole, and who also valued preserving and enhancing the welfare of people with whom they were in personal contact, were less likely to express prejudice.

Note that it was individual value types that predicted prejudice – power and security were positive predictors, universalism and benevolence were negative predictors ( Table II). Other value types from Schwartz's higher order dimensions (self-enhancement, self-transcendence, conservation, openness to change) were not predictors of prejudice. However, as we hypothesised, value types from these higher order dimensions did predict right-wing authoritarianism (RWA) and social dominance orientation (SDO). The correlations and path analysis showed that RWA scores were more reflective of the conservation dimension of the value types (security, tradition, conformity) and less reflective of both the openness to change dimension (self-direction, stimulation, hedonism) and the self-transcendence dimension (benevolence, universalism). SDO scores were more reflective of the self-enhancement dimension (power, achievement and hedonism) and less reflective of the self-transcendence dimension. Our results also showed that both variables were positive predictors of prejudice towards Australian Aborigines and that they were partial mediators of the effects of values on prejudice. When RWA and SDO were included in the path analysis, security, universalism, and benevolence values no longer had direct links to prejudice but power values continued to have a direct effect.

The results that link SDO and RWA to specific value types from the SVS and to prejudice are generally consistent with theoretical discussions of right-wing authoritarianism and social dominance and also with previous findings (Altemeyer, 1998; Cohrs et al., 2005; Duckitt, 2001, 2005, 2006; Duckitt & Sibley, 2007; Duriez & Van Hiel, 2002; Feather, 1996, 1998, 1999, 2005; Sidanius & Pratto, 1999). In particular, our results resonate with Duckitt's (2001, 2005, 2006) proposal that prejudiced attitudes have an ideological basis. In his view RWA and SDO are “two basic ideological dimensions of social or ideological attitudes with each expressing motivational goals or values made chronically salient for individuals by their social world views and their personalities” (Duckitt, 2005, p. 405). The effects of authoritarianism are assumed to become more prevalent when there are perceived threats from outgroups; the effects of social dominance are assumed to become more prevalent when there is inequality and competition with lower status outgroups. The link between authoritarianism and threat has been recognised by others (e.g., Altemeyer, 1998; Doty, Peterson, & Winter, 1991; Feather 1999, 2005; Sales, 1973), and as we noted earlier, Fiske et al. (2002) have proposed that perceived status and competition are basic determinants of different types of prejudice in relation to dimensions of competence and warmth respectively. So there is some consensus that separates threats to stability and security from threats to status and privilege.

Consistent with these analyses, our results showed that prejudice towards Australian Aborigines was predicted by security values that would be activated under conditions of perceived threats to safety and harmony and by power values that would be activated under conditions of competitive threats to status and dominance. However, our results implicate a wider range of values that were associated with negative attitudes towards indigenous Australians. Social values relating to social justice and equality also played a part, as did more interpersonal values relating to the care and protection of others.

As we noted, the path analysis ( Figure 2) showed that there was a direct positive path linking power values to prejudice and that power values continued to exert an effect on prejudice even after RWA and SDO were included in the analysis. There was also a direct negative path linking tradition values to prejudice. This negative path may reflect a tendency for participants with strong tradition values to show respect for the ideas and customs of the indigenous culture as practiced by Aboriginal communities. Note that the zero-order correlation between tradition values and Modern Racism was negative but nonsignificant ( Table II). The negative relation only became significant when RWA and SDO were included as mediators in the path analysis. It is an interesting finding that requires replication.

Our study used a sample of university students from Flinders University. Clearly there is a need to test the generalisability of our findings by extending the research to wider samples from the Australian population. Future research should also investigate other targets in order to determine whether prejudice towards different targets (e.g., gays, feminists, immigrants, criminals) are predicted by value types that are similar or different from those found in the current study. Does the value base of prejudice generalize across different target groups (Allport, 1954) or does it vary depending on which target group is the focus of study, as the analyses discussed previously might suggest (Duckitt & Sibley, 2007; Glick & Fiske, 2001a; Fiske et al., 2002)?

Finally, we are not arguing that values are the sole determinants of prejudice. Prejudice has deep roots in social learning, family and group dynamics, self-interest, social identification, and in structural variables within a society. But prejudice is also linked to the value systems that people develop in the course of their lives as they adjust to universal issues of satisfying biological needs, relating to others, maintaining some stability and security in their social roles and group memberships, and developing a meaningful and integrated sense of self. Our results show that people with prejudiced attitudes towards Australian Aborigines were those who endorsed stronger power and security values that reflect self-enhancement and conservation and who rated universalism and benevolence values that reflect self-transcendence as less important for self. We are firm in our view that a value analysis holds much promise for enabling an improved understanding of the dynamics of prejudice.

Acknowledgement

The research reported in this article was supported by a grant from the Australian Research Council.