European Journal of Personality Eur. J. Pers. 21: 91–111 (2007) Published online 3 January 2007 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/per.615
Social Axioms and Values: A Cross-Cultural Examination KWOK LEUNG1*, AL AU2, XU HUANG3, JENNY KURMAN4, TOOMAS NIIT 5 and KAISA-KITRI NIIT 5 City University of Hong Kong, Hong Kong 2 University of Hong Kong, Hong Kong 3 Hong Kong Polytechnic University, Hong Kong 4 University of Haifa, Israel 5 Tallinn University, Estonia 1
Abstract The relationships between social axioms, general beliefs that people hold about the social world, and values, defined as desirable goals for life, were examined in five cultural groups. Results show that the correlations between social axioms and Schwartz’s (1992) values are generally low, suggesting that they represent two distinct types of construct. Based on a structural equation modelling approach, results further show that generally speaking, the five axiom dimensions are related to the value types in a meaningful and interpretable manner, and that these relationships are generally similar across the five cultural groups. Implications of these results and directions for future research are discussed. Copyright # 2007 John Wiley & Sons, Ltd. Key words: social axioms; Schwartz’s values; culture INTRODUCTION In the study of culture, two approaches are prominent. Cultural psychology and indigenous approaches focus on culture-specific processes and meaning systems, whereas crosscultural psychology focuses on universal or etic processes across cultures (Greenfield, 2000; Shweder, 2001). Both approaches are useful and complement each other. Perhaps the most obvious example of an etic approach is the identification of cultural dimensions for characterizing diverse cultures in the world. Most dimensions of culture are based on values, such as the now classic dimension of individualism–collectivism (Hofstede, 1980). To provide an alternative conceptualization, Leung, Bond, and their associates (Leung et al., 2002) have turned to general beliefs, or social axioms, for characterizing culture and its psychological manifestations.
*Correspondence to: Kwok Leung, Department of Management, City University of Hong Kong, Hong Kong. E-mail:
[email protected]
Copyright # 2007 John Wiley & Sons, Ltd.
Received 14 October 2005 Revised 8 October 2006 Accepted 9 October 2006
92
K. Leung et al. A FRAMEWORK OF SOCIAL AXIOMS
Social axioms refer to context-free beliefs and assume the following form: A is related to B. A and B can be any constructs, and people’s belief in a belief statement is reflected by their perceived likelihood of the relationship. For instance, for the belief ‘competition leads to progress’, some people may see a strong link between competition and progress, while others may think that the two are unrelated. In contrast, values refer to the importance or desirability that people attach to a construct, such as religion or power. We use the label ‘axiom’ to reflect the axiomatic nature of these general beliefs, because a person assumes their validity without meticulous evaluation and scrutiny. The term ‘social’ is used to refer to the assumption that these axioms are acquired through social experiences. With social axiom items culled from the Western psychological literature as well as from interviews and cultural materials from China and Venezuela, five pan-cultural axiom dimensions were identified across forty cultural groups (Leung & Bond, 2004). Briefly put, Social Cynicism indicates a generally negative view of people and social institutions, and the extent to which actors expect negative outcomes from their engagements with life, especially with more powerful others. Social Complexity indicates an actor’s judgements about the variability of individual behaviour and the number of influences involved in determining social outcomes. Reward for Application indicates how strongly a person believes that challenges and difficulties will succumb to persistent inputs, such as relevant knowledge, exertion of effort or careful planning. Religiosity indicates an assessment about the positive, personal and social consequences of religious practice, along with the belief in the existence of a supreme being. Note that religiosity is sometimes defined as a personal value, but our definition is based purely on general beliefs about religions and has nothing to do with one’s own religiosity values. In the present study, we measured religiosity with belief items based the social axioms framework. Fate Control indicates the degree to which important outcomes in life are believed to be fated and under the control of impersonal forces, but are predictable and alterable. These five dimensions are identifiable in a wide range of cultural groups (Leung & Bond, 2004), even with the adoption of stringent statistical procedures (Cheung, Leung, & Au, 2006). Although cultures can be rank-ordered according to these five axiom dimensions, we do not assume that social axioms are highly consensual in a given society. This feature differentiates social axioms from mathematical axioms, which are endorsed by all mathematicians. To validate the meaning of these five axiom dimensions, Leung and Bond (2004) correlated the average axiom scores of a cultural group, its citizen scores, with a diverse range of variables at the culture level. Many meaningful relationships are found, supporting the validity of these axiom dimensions. For instance, Social Cynicism is related to lower life satisfaction; Social Complexity is related to higher voter turnout and stronger interest in politics; Reward for Application is related to more working hours per week; Religiosity is related to higher agreeableness and more church attendance; and Fate Control is related to a higher heart disease death rate. At the individual level, a number of studies have also found meaningful relationships with a wide range of variables (see Leung & Bond, 2004 for a review). For instance, in a study involving Hong Kong Chinese, Bond, Leung, Au, Tong, and Chemonges-Nielsen (2004a) found that Reward for Application was related to the conflict resolution style of accommodation; Social Cynicism was related to a lower tendency to use collaborative and compromising styles; Social Complexity was related to the use of compromise and of collaboration; and Religiosity was related to both accommodation and competition Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
Social axioms and values
93
positively. Fate Control was significantly related to distancing, a coping style characterized by being passive and the avoidance of thinking about difficulties. Singelis, Hubbard, Her, and An (2003) found that, among American college students, Social Cynicism correlated negatively with interpersonal trust and cognitive flexibility; Social Complexity correlated positively with cognitive flexibility, and feeling comfortable with talking to strangers and speaking one’s mind even if it may hurt others’ feelings; Reward for Application was related to trying harder the next time when unsuccessful; Religiosity correlated positively with traditional Christian beliefs, and with seeking advice from a spiritual adviser, praying and reading scriptures; Fate Control correlated positively with external locus of control, negatively with traditional Christian beliefs, and positively with spiritual beliefs, supernatural beliefs and belief in precognition.
SCHWARTZ’S THEORY OF VALUES Schwartz (1992) has developed a theory of values that consists of 10 value types organized in the form of a circumplex. Schwartz used smallest space analysis to examine the spatial relationships among a broad range of values, and found that the optimal solution is a two-dimensional representation, with the values organized in the form of a circle: Stimulation (e.g. exciting life), Self-Direction (e.g. independent), Universalism (e.g. world of beauty), Benevolence (e.g. forgiving), Conformity (e.g. self-discipline), Tradition (e.g. respect for tradition), Security (e.g. social order), Power (e.g. authority), Achievement (e.g. successful) and Hedonism (e.g. enjoying life). For a schematic representation, see Figure 1.
Figure 1. A schematic representation of Schwartz’s theory of values (based on Schwartz, 1994b). Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
94
K. Leung et al.
Schwartz (1992) further proposed that his 10 value types can be organized into two pairs of opposing higher-order categories. In one contrast, some values reflect the motivation to enhance the self (power and achievement), and other values reflect the motivation to transcend the self (universalism and benevolence). In the other contrast, some values reflect the motivation for conservation (tradition, conformity and security), and other values reflect the motivation for openness to change (self-direction and stimulation). Hedonism is the only value type that lies between two higher-order categories: openness to change and self-enhancement. Schwartz’s theory has been validated in a wide range of cultures (Schwartz, 1992; Schwartz & Boehnke, 2004). Many meaningful relationships between his value types and other variables have been reported, such as those with worries (Schwartz, Sagiv, & Boehnke, 2000) and with voting behaviour (Barnea & Schwartz, 1998).
SOCIAL AXIOMS AND VALUE TYPES As described before, some research has examined the joint effects of social axioms and the Schwartz value types on various dependent variables. Bond et al. (2004a) found that social axioms and the Schwartz values do not correlate highly among Hong Kong Chinese college students. Kurman and Ronen-Eilon (2004) found that social axioms were more predictive of the adaptation of immigrants in Israel than were values. Nonetheless, a systematic investigation of their relationships has not been conducted, which is the major objective of the present paper. An exploration of the relationships between social axioms and the Schwartz value types is important for two reasons. First, social axioms and values are different constructs, and a systematic investigation of their inter-relationships will shed light on how to integrate them for a better understanding of preferences and behaviours. For instance, in the expectancy model, valence and expectancy are jointly used to predict specific behaviours (Vroom, 1964). Values may be regarded as generalized valences, and social axioms, generalized expectancies. The examination of how values are related to axioms can help us better understand and predict general patterns of social behaviours (Leung & Bond, 2004). Second, while the cross-cultural validity of the Schwartz values is well established, social axioms as relatively novel constructs require more evidence for their cross-cultural generality. If some meaningful relationships between social axioms and values can be established, the validity and universality of social axioms can be further strengthened. How may social axioms relate to the Schwartz value types? Before we delve into specific predictions, two issues should be addressed. First, values signal one’s priorities in life, and are often ascribed a motivational function in guiding people to focus their effort on goals deemed as important (e.g. Rokeach, 1972; Schwartz, 1996). In contrast, axioms are judgements about the social world, which are not based on self-description and self-perception. Although axioms and values both serve as general guidelines for choices and behaviours, these two types of construct differ in the way they operate. Values provide the ‘what’ answer, in a sense that they define what one should pursue, be it wealth or social justice. Axioms provide the ‘how’ answer, because how one construes the social world bear on the strategies and actions adopted for goal achievement. Viewed in this perspective, social axioms should be distinct from and independent of values. However, motivational and cognitive processes are not isolated, and they exert mutual influence upon each other (e.g. Jost, Glaser, Kruglanski, & Sulloway, 2003; Strack & Deutsch, 2004). It is well known Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
Social axioms and values
95
that values provide a lens through which people perceive and interpret social reality, and that people who endorse certain values are likely to see the world in a specific way. For instance, Goodwin, Nizharadze, Luu, Kosa, and Emelyanova (2001) found that values predicted people’s perceived availability of social support in three post-Communist nations. Reciprocally, cognitions also influence the goals that people pursue. For instance, self-efficacy, the general assessment of one’s ability for goal attainment, has been shown to influence goals, such as academic goals (Zimmerman, Bandura, & Martinez-Pons, 1992) and career goals (Kahn & Scott, 1997). The reciprocal influence between values and beliefs suggests that although social axioms and values are distinct constructs, some meaningful linkages between these two types of construct should be detectable. The second issue is that values are organized in a circumplex structure, which implies that if the data fit the model perfectly, a variable should relate to the 10 value types in the form of a sinusoid function (Schwartz, 1992). Most adjacent values would show similar relationships with a variable, but occasionally some adjacent values may show opposite relationships with a variable. This pattern is possible because if the data are perfect, the sinusoid curve should pass through the X-axis twice, and a predictor would show opposite relationships with two adjacent values if the curve passes through the X-axis between them.
PREDICTIONS ABOUT THE RELATIONSHIPS BETWEEN VALUES AND AXIOMS We now discuss the expected relationships between axioms and values. In the absence of previous research, our analysis is by necessity speculative. Obviously, the ideal way to generate predictions is to rely on a coherent theoretical framework, but unfortunately such a framework has yet to be developed. We instead adopt an empirical approach and rely on findings based on constructs that are related to social axioms to generate some predictions. We note that a number of studies examined specific values rather than value types, and we therefore have to rely on these specific values to generate predictions for the value types to which they belong. In a nutshell, the current research should be viewed as the first step to develop a coherent theoretical framework for linking social axioms and values. Social Cynicism involves a negative view of people and social institutions, and it is likely to correlate negatively with universalism, and positively with power. Mistrust is a core component of Social Cynicism, and we draw upon the research on mistrust to support these speculations. Mistrust in co-workers is related negatively to a universalistic as opposed to a particularistic orientation in personnel practices (Pearce, Branyiczki, & Bigley, 2000), suggesting a negative relationship between Social Cynicism and Universalism. People high in Social Cynicism should also see Power as important because it can be used to protect their well being against the potential exploitation by others. Consistent with this argument, perceived powerlessness is related to mistrust (Ross, Mirowsky, & Pribesh, 2001). People high in Social Cynicism may also have a deficiency need with regard to power (Bilsky & Schwartz, 1994). Reward for Application involves the belief in the effectiveness of effort and hard work, and it should correlate positively with Achievement, and negatively with Hedonism. Consistent with these speculations, Protestant work ethic, which involves the belief in hard work, is related to achievement motivation positively (Furnham, Kirkcaldy, & Lynn, 1994), the provision of good service to customers (Tang & Weatherford, 1998), and to an emphasis Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
96
K. Leung et al.
on leisure negatively (Hirschfeld & Feild, 2000). We note that Achievement and Hedonism are adjacent values types in Schwartz’s model, and we predict that Reward for Application will show opposite relationships with them. In other words, we expect that the sinusoid curve will pass through the X-axis between these two value types. In addition, we expect Reward for Application to correlate with Conformity positively, since conformity involves the restraints of actions and impulses that violate social norms and expectations, which seem conducive to perseverance. In line with these arguments, Feather (1984) found that Protestant Ethic value is correlated positively with obedience and self-control, which are definers of Conformity. Social Complexity is associated with a complex view of people and social events, as well as openness to diverse views and pluralism. We therefore expect it to correlate with Tradition and Conformity negatively, both of which emphasize adherence to social and moral norms. We also expect it to correlate positively with Self-Direction, which emphasizes independence, and with Universalism, which emphasizes understanding, tolerance and protection for the welfare of people and nature. Our predictions are based on results associated with the openness dimension of the five-factor model of personality, which also involves beliefs that emphasize openness to experience and imaginativeness. Openness has been found to correlate negatively with Tradition and Conformity, and positively with Self-Direction and Universalism (Roccas, Sagiv, Schwartz, & Knafo, 2002). Using a lexical approach based on the German language, Renner (2003) reported that openness correlated negatively with values associated with conservation (e.g. national identity and tradition), but positively with values associated with balance (e.g. human kindness, human rights and human dignity). The finding with regard to balance seems to suggest that Social Complexity may correlate with Benevolence. It seems possible that the pluralistic view espoused in Social Complexity may encourage a benign relationship with other people. Fate Control is associated with the belief in fate and the possibility of altering fate to one’s benefit. Fate Control is likely to relate to Tradition positively, as traditional values often involve fatalistic elements, and to Self-Direction negatively. Consistent with this argument, Unger et al. (2002) found that fatalism, which is characterized by perceived uncertainty and uselessness of planning, correlated positively with such traditional values as filial piety and machismo (male dominance). Their findings show that fatalism also correlated positively with collectivism, which is characterized by high interdependence with friends, and seems opposite to Self-Direction. Finally, Religiosity is likely to relate to Tradition and Conformity positively, as religions are part of the tradition of most cultures and they require conformity to religious practices. Religiosity should also relate to Benevolence positively, as religions are typically associated with the advocacy of selfless love and sacrifice, and to Hedonism, Stimulation and Self-Direction negatively, as religions typically de-emphasize enjoyment, pleasure and excitement, and instead emphasize restraint, other-worldliness and subjugation to a divine being. These speculations are confirmed by a meta-analysis of the relationships between the Schwartz values and religiosity, defined as a general personal orientation, with 21 independent samples from 15 different countries. The findings generally show that religious people tend to endorse Tradition, Conformity and Benevolence, and reject Hedonism, Stimulation and Self-Direction (Saroglou, Delpierre, & Dernelle, 2004). Table 1 summarizes our predictions. We believe that the relationships between social axioms and values are universal and are relatively unaffected by cultural context. To test this general hypothesis, the relationships Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
Social axioms and values
97
Table 1. Predictions of the relationships between social axioms and values Social Axioms
Hypothesized relation with Schwartz’s values
Social Cynicism
Power (þ) Universalism () Achievement (þ) Conformity (þ) Hedonism () Self-direction (þ) Universalism (þ) Benevolence (þ) Conformity () Tradition () Tradition (þ) Self-direction () Tradition (þ) Conformity (þ) Benevolence (þ) Hedonism () Stimulation () Self-Direction ()
Reward for Application Social Complexity
Fate Control Religiosity
between axioms and values were examined in a diverse group of cultures: Hong Kong, mainland China, the Netherlands, Israel and Estonia.
METHOD Participants The five cultural groups studied vary on many cultural dimensions. Dutch, Israelis and Estonians are individualistic and low in power distance, whereas the two Chinese groups (Hong Kong and mainland China) are collectivistic and high in power distance (Hofstede, 2001). Schwartz (1994a) has analysed his values at the culture level, with culture as the unit of analysis, and obtained dimensions that are different from the individual-level value types described before. With regard to his culture-level value framework, Hong Kong, mainland China, Israel and Estonia (urban) are in the top half regarding conservatism, while the Netherlands is in the lower half. The five cultural groups also span across the dimension of hierarchy with mainland China near the top and Estonia (urban) near the bottom. The Netherlands is very high in egalitarian commitment, Estonia (urban) in the middle, while Hong Kong, Israel and mainland China are in the bottom half. As for the culture-level dimensions of social axioms (Bond et al., 2004b), the Netherlands and Israel are relatively lower in societal cynicism and dynamic externality, which refers to a belief constellation defined primarily by an emphasis on effort and the positive consequences of religions, while mainland China, Hong Kong and Estonia are relatively higher in these two dimensions. In addition, these cultural groups have different dominant religions (e.g. Protestantism, Judaism and Buddhism). Different socio-political-economic systems are represented. Hong Kong, the Netherlands and Israel are capitalistic, mainland China is Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
98 Table 2.
K. Leung et al. Sample information
Sample size M F Mean age
Hong Kong
Mainland China
The Netherlands
Israel
Estonia
180 90 90 20.2
229 93 134 26.1
170 65 105 23.1
89 41 48 35.1
244 80 164 20.5
Note: Two respondents in the mainland Chinese sample did not report their gender.
characterized by market-oriented socialism, and Estonia is a former communist society. Hong Kong, Israel and the Netherlands are relatively wealthy, and mainland China and Estonia have relatively lower GDP per capita. In summary, if a consistent pattern emerges across these diverse cultural groups, the generality of the relationships is compelling. There were a total of 912 participants from the five cultural groups, and Table 2 presents the sample characteristics. Participants either received course credit for the participation or participated voluntarily. The mainland Chinese sample was collected specifically for this paper, but the social axiom data for the Netherlands, Israel and Estonia have been reported in both Leung and Bond (2004) and Bond et al. (2004b), and the social axiom data for Hong Kong were reported in Bond et al. (2004a). The value data have not been reported before, except for the value data of Hong Kong, which were reported in Bond et al. (2004a). The data set was comprised of predominately college students with the exception of mainland China and Israel, in which adults were also included. However, the heterogeneity in age should not be a problem because both the axiom and value constructs had been validated with students and adults (Leung & Bond, 2004; Schwartz, 1992). To be certain, we calculated partial correlations between axioms and values controlling for age in these two cultures to assess the importance of age. Among the 100 axiom-value pairs (50 for each culture), the average absolute difference between simple and partial correlations was 0.006 for mainland China and 0.01 for Israel. We therefore conclude that age was not important and did not consider its impact in subsequent analyses. Measures and procedure For all questionnaires, local languages were used, and back translation was used to assure fidelity in translation. Participants completed the 82-item or 60-item version of the Social Axioms Survey (Leung et al., 2002). The response scale used five points, with ‘Strongly disbelieve’ and ‘Strongly believe’ as labels for the two end points. Sample items included ‘Powerful people tend to exploit others’ (Social Cynicism), ‘There are various ways to achieve a goal’ (Social Complexity), ‘One will succeed if he or she really tries’ (Reward for Application), ‘Good luck follows if one survives a disaster’ (Fate Control), and ‘Religious faith contributes to good mental health’ (Religiosity). Except for the Israeli sample, all participants completed the 57-item version of the Schwartz Value Survey (SVS, Schwartz, 1994b), which consisted of the 10 value types. Participants responded to value phrases such as ‘Equality [equal opportunity for all]’ (Universalism) and ‘Indulge myself [self-indulgence]’ (Hedonism) on a 9-point scale ranging from ‘opposed to my values’ (-1), ‘not important’ (0), to ‘of supreme importance’ (7). The Israeli participants completed the 40-item Portrait Value Questionnaire (PVQ), which is an alternate and less abstract form of the original value survey and was validated with participants from four countries: South Africa, Uganda, Italy and Israel (Schwartz Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
Social axioms and values
99
et al., 2001). PVQ can be regarded as an indirect measure of respondents’ endorsement of various value types because they are asked to compare themselves to a specific person possessing a particular value described in each item. PVQ consists of the same 10 value types, and sample items include ‘He thinks it is important that every person in the world be treated equally. He believes everyone should have equal opportunities in life’ (Universalism) and ‘Enjoying life’s pleasures is important to him. He likes to ‘‘spoil’’ himself’’’ (Hedonism). Participants responded to a 6-point scale ranging from ‘not like me at all’ to ‘very much like me’. In validating the PVQ, Schwartz et al. conducted a multitrait-multimethod analysis by asking the same group of respondents to complete both the SVS and the PVQ in a counterbalanced order. Results showed that the single-trait-multimethod correlations were all significant and sizeable, and were larger than the multitrait-multimethod correlations. Thus, we regard the two methods as equivalent and did not differentiate them in our data analysis. Participants received the anonymous survey instrument from an administrator in groups. They were briefed about the content of the survey and the confidentiality of the data collected. The time required to complete the survey was about 15 to 25 minutes. Analyses The pan-cultural structures of the axiom and value constructs have been validated with a wide range of cultures, and these structures were used to ensure valid cross-cultural comparisons. As a result, only 39 items from the 60 axiom items and 46 from the 57 value items were included in the analyses (see Leung & Bond, 2004; Schwartz, 1992; Schwartz & Boehnke, 2004). For the PVQ, all 40 items were included. Scale composites in the form of averages (sum of all items in a scale divided by the number of items) were used in the main analysis. Schwartz (1992) suggested that his value survey should reveal the value priorities of a person and individual differences in response styles may confound the value profile of a person (see also Leung & Bond, 1989). To overcome this problem, value scores were within-subject centred for each respondent by subtracting the overall mean of all his or her value items from each value item. As a result of this within-subject centring procedure, the sum of all the value scores of a respondent is zero, thus allowing the individual value scores to reveal the value priorities of the person. No within-subject centring was conducted for axioms because it is unclear whether a comprehensive set of axioms has been identified.
RESULTS Cronbach alphas of both axiom and value scales for the five cultural groups are shown in Table 3, and the reliability of quite a number of the scales is below 0.70. For Social Complexity and Fate Control, reliabilities are relatively low across the samples, but these reliabilities obtained are comparable to those reported in Leung and Bond (2004). For values, reliabilities for Self-Direction, Tradition and Security are relatively lower, but they are also similar to the values obtained in Schwartz et al. (2001), where they speculated that it was due to the small number of items used for each value type as well as the broad coverage of concepts by each value type. To assure whether subsequent analysis was meaningful in the face of some low reliabilities, confirmatory factor analysis (CFA) was conducted separately for axioms as well as values to check whether our data fitted the Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
Copyright # 2007 John Wiley & Sons, Ltd.
(3) (3) (4) (6) (4) (4) (4) (5) (3) (4)
4.14 3.40 4.70 4.07 4.56 2.36 3.81 4.03 2.77 4.22
(4.07) (3.82) (4.94) (4.57) (4.64) (2.99) (3.95) (4.43) (3.43) (4.32)
2.99 3.64 4.16 2.79 2.95
Raw mean
1.48 1.62 0.90 1.00 0.97 1.24 1.14 1.10 1.51 1.19
(1.20) (1.22) (0.77) (0.82) (0.76) (1.16) (1.10) (0.92) (1.11) (1.10)
0.51 0.51 0.37 0.56 0.66
SD
0.76 0.78 0.54 0.71 0.76 0.59 0.67 0.64 0.76 0.68
0.72 0.64 0.59 0.48 0.77
Hong Kong
0.61 0.70 0.55 0.74 0.74 0.53 0.58 0.57 0.67 0.60
0.72 0.67 0.49 0.42 0.69
Mainland China
Note: Numbers in parentheses refer to the items in the portrait value questionnaire answered by the Israeli respondents.
3 3 5 8 5 5 4 5 4 4
11 9 6 6 7
No. of item
Descriptive statistics and reliability of the measures
Social Axioms Social Cynicism Reward for Application Social Complexity Fate Control Religiosity Schwartz values Hedonism Stimulation Self-direction Universalism Benevolence Tradition Conformity Security Power Achievement
Scale
Table 3.
0.71 0.75 0.58 0.73 0.70 0.77 0.59 0.64 0.72 0.73
0.69 0.49 0.60 0.46 0.73
The Netherlands
Alpha
0.86 0.75 0.62 0.75 0.67 0.75 0.77 0.68 0.65 0.80
0.75 0.66 0.39 0.72 0.73
Israel
0.64 0.85 0.59 0.75 0.74 0.65 0.69 0.63 0.71 0.62
0.62 0.66 0.46 0.51 0.65
Estonia
100 K. Leung et al.
Eur. J. Pers. 21: 91–111 (2007)
DOI: 10.1002/per
Social axioms and values
101
prescribed structures reported in previous studies. We applied the technique of item parcelling to reduce the model complexity because many items were involved. We also centred the responses within each culture prior to the analysis to control for possible cultural differences in response sets. This procedure also allowed us to perform a pan-cultural analysis, in which the five cultural groups were merged into a single data set (Leung & Bond, 1989). In particular, for axioms, three parcels were formed for each of the axiom dimensions by randomly assigning items to each parcel. The number of items in each parcel ranged from two to four. A five-factor structure was fitted, and the CFA results suggested an acceptable fit: x2 (df ¼ 125) ¼ 479.8, p < 0.001, with x2/df ¼ 3.84, CFI ¼ 0.934, IFI ¼ 0.935, SRMR ¼ 0.041, RMSEA ¼ 0.059. For values, two parcels were formed due to the smaller number of items in some of the value types. For Hedonism and Stimulation, which consist of only three items, a two-item parcel and the remaining item were used as the two indicators for each value type. The number of items in each parcel ranged from two to four, and the CFA results also suggested a good fit for a 10-factor structure: x2(df ¼ 80) ¼ 244.4, p < 0.001, with x2/df ¼ 3.06, CFI ¼ 0.921, IFI ¼ 0.922, SRMR ¼ 0.046, RMSEA ¼ 0.047. In sum, the universal structures of axioms and values reported before were confirmed with the present data set, suggesting that the low reliabilities of some scales were not too problematic. The unreliability issue is also addressed in subsequent analyses. Table 4 shows the correlation coefficients between axioms and values for the five cultural groups. As expected, most of the correlations are low and within the range of 0.25 and 0.25, and the highest correlation scored is 0.54, which is between Religiosity and Tradition in the Israeli sample. In fact, Religiosity yields most of the larger correlations. We may conclude from these correlations that the overlap between axioms and values is indeed small. Generally speaking, we expect sinusoid relationships between an axiom dimension and values because the values are based on a circumplex model (Schwartz, 1992). Social Cynicism is correlated most positively with Power, and most negatively with Self-Direction. For Reward for Application, the most positive correlation is with Achievement, and the most negative correlations are with Hedonism and Tradition. For Social Complexity, the most positive correlation is with Self-Direction, and the most negative correlation is with Tradition. For Fate Control, the most positive correlations are with Tradition and Power, and the most negative correlation is with Benevolence. For Religiosity, the most positive correlations are with Tradition, Conformity and Benevolence, and the most negative correlation is with Hedonism. The sinusoidal shape is quite clear for Religiosity, and moderately clear for Social Cynicism, Social Complexity and Reward for Application. There is no obvious pattern for Fate Control in its correlations with values. We conclude that the pattern of correlations generally supports our expectation. A formal test was conducted to evaluate whether the correlations between axioms and values are equivalent across the five cultural groups. Given that the structure of axioms and values has been validated across diverse cultures, our focus is on testing the equality of the axiom-value correlation coefficients across the five cultural groups. Procedures like those proposed by Olkin and Finn (1995) for testing correlation coefficients across samples or by Hunter and Schmidt (1990) for meta-analysis can be used, but these procedures are designed to assess a pair of correlations at a time across different samples. For our purpose, we need to assess 50 correlations (5 axioms 10 values) across 5 samples, and these procedures become tedious and inefficient. Recently, Cheung and Chan (2004) propose that the equality of a set of correlations across several samples can be tested efficiently by Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
Copyright # 2007 John Wiley & Sons, Ltd.
0.02 0.26
0.10
0.02
Power
Achievement
0.11
0.13
0.27
0.11
0.04
0.02
0.05
0.10
0.03
0.21
0.30
0.27 0.09
0.08
0.12
0.19 0.34
0.17
0.04
0.08 0.00
0.04
0.03
0.08 0.05
0.09
0.09
0.10
0.07
0.06
0.10
0.06
0.02 0.15
0.20
ES
0.12
0.09
0.06
CN
0.08 0.13
0.00 0.04
0.14
0.03
0.02
HK
0.07
0.00
0.09
0.02
0.02
NE
0.17
0.12
0.05
0.15
0.15
IS
Social Complexity
0.30
0.13
0.03
0.09
0.17 0.02
0.10
0.08
0.09
0.13 0.03
0.05
0.02
0.16 0.09 0.03 0.02
0.06
0.18 0.02 0.12
.01
0.03
CN
0.16 0.09
0.13
0.11
HK
0.18
0.00
0.07
0.10 0.22
0.16 0.16
0.02 0.03
0.13
0.10 0.09 0.14
0.21
0.08
0.02
0.05
0.05
NE
0.18
0.16
0.01
0.03
IS
0.07
0.10
0.08 0.00
IS
ES
0.25
0.54
0.24
0.18
0.01
0.07
0.00 0.00 0.20 0.15 0.0.04 0.14
0.06
0.14 0.15
0.19 0.05
0.41
0.26
0.10
0.20
0.07
0.10
0.14
0.03
0.11
0.22
0.10
0.14
0.03
0.28
0.34 0.22 0.17
NE
Religiosity
0.31
0.03
0.05
0.31 0.16
0.45
0.25
0.15
0.20 0.04 0.27 0.04
0.09 0.03
0.02
0.26 0.02
0.21 0.13
0.14
0.06
0.01
0.05
CN 0.05 0.35
HK
0.09 0.19
ES
0.13 0.09 0.25
0.11
0.09
Fate Control
0.23 0.05
0.04 0.21 0.04 0.09
0.05
0.04 0.06
0.19 0.07 0.11
0.05
0.02
ES
0.06 0.25 0.10 0.11 0.26 0.25 0.24
0.02 0.13
0.21 0.04 0.05
0.05
0.07
0.24
0.08
IS
0.17 0.26
0.04 0.21 0.09
0.07
0.07 0.08
0.04
0.00
NE
0.00
0.02
0.07
CN
0.16 0.08
HK
Reward for Application
HK ¼ Hong Kong (n ¼ 180); CN ¼ mainland China (n ¼ 229); NE ¼ the Netherlands (n ¼ 170); IS ¼ Israel (n ¼ 89); ES ¼ Estonia (n ¼ 244). p < 0.05. p < 0.01.
0.05
0.00
0.02 0.03
Conformity
Security
0.04
0.06 0.07
Tradition
0.14 0.08
0.01 0.22 0.07 0.12
Benevolence 0.10 0.18
Universalism
Self-direction 0.02 0.09
0.08
0.05
0.03 0.19
0.03
0.14 0.01 0.20
ES
IS
0.09
NE
0.09
CN
Stimulation
HK
Social Cynicism
Correlations between axioms and values
Hedonism
Table 4.
102 K. Leung et al.
Eur. J. Pers. 21: 91–111 (2007)
DOI: 10.1002/per
Social axioms and values
103
structural equation modelling (SEM). Specifically, a multiple group SEM model is set up, in which the relevant covariance paths that correspond to the correlations under examination are constrained to be equal across groups. If the fit statistics are acceptable, we then conclude that the correlations are equivalent across groups (see Cheung & Chan, 2004 for technical details). Their approach allows us to test the cross-cultural equivalence of all axiom-value correlations in a single omnibus test. We extended Cheung and Chan’s (2004) approach to 50 correlations across five samples. We first fitted a model in which each of the 10 axiom-value correlations was set to be equivalent across the five cultural groups with EQS 6.1 (Bentler & Wu, 2004). This package is preferable because it provides a Lagrange Multiplier test function for improving model fit by relaxing the equality constraints placed on the 50-covariance paths using a data-driven approach. In this model, there are 15 latent variables for each cultural group, representing the five axiom dimensions and the 10 value types. For each latent variable, the composite score of the corresponding axiom or value dimension was regarded as an observed variable. Following Cheung and Chan (2004), the factor variance of the latent variables was set as 1, and the error variance of the observed variables was set as zero. A full matrix of the covariance paths was constructed based on the 15 latent variables, but correlation matrices of the five samples were actually input as covariance matrices. This ‘trick’ led to the display of correlation coefficients in the covariance paths, and the end result was a saturated SEM model with zero degree of freedom. To test for cross-cultural equivalence, we imposed an equality constraint on each of the 10 axiom-value pairs across the five cultural groups. Each set of constraints had 4 degrees of freedom, and the final model had 200 degrees of freedom because we had 50 axiom-value correlations. An omnibus test was then performed on this model, which assumes all axiom-value correlations to be congruent across the five cultural groups. The SEM test statistics were: x2 (df ¼ 200) ¼ 380.07, p < 0.001, with x2/df ¼ 1.90, which rejected the cross-cultural equivalence of all axiom-value pairs. This result is not surprising, because the chi-square statistic is very sensitive in a large sample. Thus, we also examined various fit indexes to evaluate the fit of the model: CFI ¼ 0.975, IFI ¼ 0.976, SRMR ¼ 0.072 and RMSEA ¼ 0.032. Using the criteria recommended by Hu and Bentler (1998), the model achieved a very good fit except for SRMR, which is slightly larger than the criterion of 0.06. Table 5 shows the SEM estimates of the pooled correlation coefficient estimates. All Table 5. Pooled correlation coefficient estimates from the SEM results Social Cynicism Hedonism Stimulation Self-direction Universalism Benevolence Tradition Conformity Security Power Achievement
p < 0.05;
p < 0.01;
0.03 0.02 0.09 0.04 0.04 0.01 0.08 0.02 0.08 0.02
Reward for Application
0.05 0.02 0.01 0.04 0.02 0.13 0.07 0.02 0.07 0.15
Social Complexity
Fate Control
Religiosity
0.02 0.05 0.11 0.04 0.09 0.19 0.05 0.01 0.08 0.05
0.03 0.04 0.07 0.02 0.05 0.07 0.02 0.02 0.04 0.01
0.14 0.10 0.15 0.01 0.14 0.33 0.10 0.05 0.10 0.11
p < 0.001.
Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
104
K. Leung et al.
correlations fall within the range of 0.2 to 0.2, with the exception of the correlation between Tradition and Religiosity (0.33). Out of the 18 predictions presented in Table 1, 15 are supported. The three unsupported predictions include: Social Cynicism with Universality, and Social Complexity with Universalism and Conformity. A total of seven unpredicted correlations are significant, but only three are from a value type that is not the neighbour of a predicted value type: Power with Social Complexity and Religiosity, and Conformity with Social Cynicism. In addition, the correlation between Reward for Application and Tradition was negative. Although Tradition is a neighbour of a predicted value type, the sign was opposite to our expectation. In other words, a total of four significant correlations deviated drastically from our original predictions. The unsupported predictions and the major deviations are explored in the Discussion. Finally, the sinusoidal curve pattern is clear for Religiosity, and vaguely discernable for Social Cynicism, Social Complexity and Reward for Application. No obvious pattern was found for Fate Control (see Figure 2). Despite the acceptable fit provided by the equality model, we used the Lagrange Multiplier (LM) test to improve the model fit. This procedure first identified the most problematic case among all the equality constraints, and relaxed this constraint. The next most problematic constraint was then identified, and the procedure repeated until all the 200 cases were being evaluated. We first examined the results of the univariate LM test to assess the improvement of relaxing each of the 200 equality constraints. Ten cases reached the 0.01 level of significance, and were all concerning Religiosity and one of the value types across Hong Kong and Estonia. In other words, relaxing each of these 10 equality constraints would result in a significantly better model fit. We then conducted the
Figure 2. Pooled correlation coefficient estimates based on the SEM results. Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
Social axioms and values
105
cumulative multivariate LM test to identify the next constraint that would yield the largest improvement after the most problematic constraint was relaxed. It should be noted that results of the univariate and multivariate LM tests may not be identical, and we used both tests to identify problematic constraints. After relaxing the equality constraint associated with the largest improvement in model fit (the religiosity-stimulation pair), Dx2 ¼ 20.96 (df ¼ 1), p < 0.01, none of the remaining nine cases remained significant in the cumulative multivariate LM test. We therefore concluded that in general the axiom-value correlations were similar across these five cultural groups. As mentioned before, the reliability of some scales is low, so that the estimated correlations from the model may have been attenuated and do not reflect the true population correlations accurately. To explore this possibility, we followed a modelling procedure that takes into account scale unreliability by the use of latent variables with only one indicator (Netemeyer, Johnston, & Burton, 1990). The model configuration was identical to the model described before, but latent variables were used instead, with one indicator for each latent variable. The use of single indicators is particularly useful in complex models when the focus is on structural paths (e.g. Carlson & Perrewe, 1999; Zacharatos, Barling, & Iversion, 2005). The error variance was fixed at (1alpha) variance of an indicator, and the inclusion of the error variance in the model is equivalent to a correction for unreliability. The factor variance of each of the latent variable was fixed at 1, which was for model identification purposes as well as for setting the estimated covariance paths to be equal to correlations after correction for attenuation. Covariance matrices rather than correlation matrices were used as data input. As before, equality constraints on the 200 axiom-value covariance paths across the five groups were set. The SEM results obtained for this model were comparable to those of the model reported before, x2 (df ¼ 200) ¼ 373.38, p < 0.001, x2/df ¼ 1.87, CFI ¼ 0.976, IFI ¼ 0.977, SRMR ¼ 0.071, and RMSEA ¼ 0.031. As expected, due to correction for attenuation, the average of the absolute magnitude of the 50 correlations was 0.105, which was larger than the average based on uncorrected correlations (0.066). To assess whether the pattern of the correlations between axioms and values was similar across the two sets of SEM results, Spearman rank correlation coefficients were computed based on the two sets of axiom-value correlations for each axiom, and was found to have an average value of 0.98, with Cynicism and Reward for Application being the highest (1.0), and Fate Control being the lowest (0.95). We therefore conclude that the low reliability of some scales somewhat attenuated the axiom-value correlations, but it did not have much effect on the pattern of the correlations between axioms and values. DISCUSSION Relationships between axioms and values Despite the inclusion of five diverse cultural groups and both college students and adults as respondents, and the use of two different instruments to measure Schwartz’s values, the results show considerable convergence. The results support two major predictions we made. First, consistent with previous results (Bond et al., 2004a), the overlap between social axioms and Schwartz’s values is small, even when the unreliability of the scales is taken into account. It is well known that values are related to a wide range of preferences and behaviours, but the observed relationships are usually not strong (e.g. Bardi & Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
106
K. Leung et al.
Schwartz, 2003). Social axioms, which are distinct from values, provide a different perspective that adds to our understanding and prediction of preferences, judgements and behaviours. Second, social axioms show some meaningful relationships with values, which are similar across the cultural groups studied. We made 18 predictions, and 15 were confirmed. These meaningful correlations with values support the validity and cross-cultural generality of the axiom dimensions. The rationale for the confirmed predictions is described in the introduction, and we briefly review it in the following. A total of seven unpredicted correlations are significant, but only three are from a value type that is not the neighbour of a predicted value type, and these significant deviations are also discussed below. We note again that in the absence of previous research, our account is obviously speculative. We also note that we use some specific values of a value type to illustrate our arguments, and it does not mean that we regard these specific values as key drivers of the value-axiom relationships observed. It should be emphasized again that among the significant axiom-value pairs, some relationships are weak (e.g. reward for application with hedonism). We regard these significant but weak relationships as indicative of associations between distinct concepts, because the relationships between distinct concepts may vary from weak to strong. Social Cynicism is correlated positively with Power, and a likely explanation is that people with a cynical view desire power as a countervailing resource to avoid being abused and used. The predicted negative correlation between Social Cynicism and Universalism was not found, and people high in Social Cynicism do not reject such universalistic values as wisdom and broadmindedness. It seems that the belief in a malevolent social world is not related to the rejection of self-transcendence values. Not predicted, Social Cynicism was also correlated with Conformity positively and Self-Direction negatively. These correlations are interesting because people high in Social Cynicism view social institutions and authority figures as untrustworthy, but they endorse Conformity and de-emphasize Self-Direction. Perhaps they see Conformity and low Self-Direction as instrumental for avoiding rejection and punishment from other people. Reward for Application is correlated positively with Achievement and Conformity, and negatively with Hedonism. It is easy to understand that the belief in hard work and effort is related to the desire for success and the de-emphasis of pleasure. The relationship with Conformity, however, is less straightforward. Conformity is defined by such values as self-discipline, politeness and obedience. Self-discipline is obviously needed for working hard, and it is interesting that people who believe in hard work also value politeness and obedience. Perhaps hard work is often associated with taking instructions from authority figures and working within rules, which explains the importance of politeness and obedience. Not predicted, Reward for Application correlated negatively with tradition (e.g. humble, moderate and accepting my portion in life) and positively with power. The negative correlation with Tradition is interesting, and one likely explanation is that people high in Reward for Application strive to achieve their goals and break away from their current status position, a tendency that is opposite to traditional values. In other words, people high in Reward for Application work within rules and obey authorities (high endorsement of conformity), but they would like to go up the social ladder (low endorsement of tradition). Future research is needed to explore the psychological processes associated with this paradoxical finding. Finally, the correlation with power is straightforward, because people high in Reward for Application pursue not only achievement, but also power. Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
Social axioms and values
107
Social Complexity is correlated positively with Self-Direction and Benevolence, and negatively with Tradition. People who subscribe to a pluralistic view of the world endorse independence, curiosity, helpfulness and honesty, which are major definers of Self-Direction and Benevolence. People high in Social Complexity seem to emphasize a caring world that allows individuality, and they resist the constraints of Tradition (e.g. humble and accepting my portion in life). Contrary to our prediction, Social Complexity did not correlate with Universalism and Conformity. People high in Social Complexity do not particularly emphasize such universalistic values as wisdom and broadmindedness, nor do they reject Conformity (e.g. polite and obedient). Not predicted, Social Complexity was correlated negatively with Power. People high in Social Complexity seem to reject power and dominance over others. Taken as a whole, these findings suggest that people high in Social Complexity place much emphasis on the individual self and endorse a benign social world in which people help each other and are not dominated by others. Fate Control correlates with Tradition positively and Self-Direction negatively as predicted. It is noteworthy that for Fate Control, the usual sinusoidal relationship with values is not evident, which suggests that Fate Control is related to values in a somewhat haphazard way. Fate Control seems more distant from values, which requires further exploration in future research. As predicted, Religiosity correlates positively with Tradition, Conformity and Benevolence, and negatively with Hedonism, Stimulation, and Self-Direction. Not predicted, it also correlates negatively with power and achievement. This pattern of results is quite similar to the meta-analytic results of Saroglou et al. (2004), who reported some relationships between religiosity as a personal orientation and Achievement and Power, and weak relationships between religiosity as a personal orientation and Security and Universalism. Taken as a whole, this pattern suggests that people high in the axiom dimension of Religiosity are low in openness to change and self-enhancement values, which are expected given that Religiosity is associated with the observance of religious teachings and the subjugation of the self to a divine being. These people also endorse conservation values except for Security, and self-transcendence values except for Universalism. These two exceptions are noteworthy, and one way to explain them is that people high in Religiosity are indifferent to Security (e.g. family and national security) and Universalism (e.g. wisdom, social justice and equality) because of their belief that a benign divine being will protect their well being and guide them with wisdom and justice. It is interesting to note that Religiosity is the only axiom dimension that shows sizable correlations with values. As described before, Religiosity as an axiom dimension is based on general beliefs about religions and its operational definition has nothing to do with one’s religiosity values. Nonetheless, previous research finds religiosity and values to be intertwined, and most if not all religions have strong value components and proffer prototypes of human values (e.g. Rokeach, 1969; Schwartz & Huismans, 1995). Our results echo these earlier conclusions despite the fact that Religiosity is entirely based on general beliefs about religions in the social axioms framework. Different functions of axioms and values As discussed before, social axioms are about social truths and values are concerned with goals. Values define what people strive for, and axioms shed light on how to achieve important goals. Leung and Bond (2004) argued that people need a general understanding of the social world to decide on effective courses of action, giving rise to the universality of Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
108
K. Leung et al.
social axioms. For instance, Social Cynicism provides the basis for choosing between trusting and self-protective behaviours in dealing with others. Reward for Application provides the basis for deciding how much effort should be exerted for task accomplishments. Social Complexity provides the basis for deciding whether or not to search for the single best solution or asserting that there is only one solution. Fate Control provides the basis for deciding whether one should adapt to events that cannot be easily altered or engaging in practices judged effective in counteracting fateful outcomes. Finally, Religiosity provides the basis for accepting a divine being and acting in a religious manner. The major difference in how axioms and values shape behaviour can be illustrated by conflict-handling behaviour. As described before, Bond et al. (2004a) found that Social Cynicism was related to a lower preference for collaborative and compromising styles. Social Cynicism is associated with a negative view of people, and because collaboration and compromising are based on mutual trust, it is obvious why these two strategies are not preferred by people high in Social Cynicism. In other words, one’s standing on Social Cynicism shapes one’s views with regard to how conflict can be resolved effectively. In contrast, values are appropriate for explaining why people prefer some goals over others in handling a conflict. For instance, Ohbuchi, Fukushima, and Tedeschi (1999) provided evidence to show that in conflict situations, the collectivistic values of Japanese led them to emphasize relationship goals (to maintain a positive relationship with others), whereas the individualistic values of Americans led them to emphasize justice goals (to restore justice). In other words, values provide a motivational account of why some goals are salient in a conflict situation for a particular cultural group. Although values and axioms influence people’s preferences and behaviours through different processes, because of the mutual influence of values and cognitions on each other, our findings point to some overlap between them. To continue with our conflict example, Self-Direction, a key marker of individualism, shows a negative correlation with Social Cynicism. Thus, people who endorse Self-Direction are more likely to emphasize justice goals in their conflict handling behaviour. Because they tend to have lower Social Cynicism, they are also more likely to use compromising and collaborative strategies to resolve a conflict. Obviously, we are not in a position yet to theorize about the complex interplay of axioms and values because of the nascent nature of this line of work. Nonetheless, this paper helps stimulate more research on how to integrate these two constructs into a coherent framework. Limitations and directions for future research The present study has several limitations, and they are discussed in the context of identifying some directions for future research. First, we claim that our results are universal, but the generality of our results needs to be evaluated with more cultural groups. A related issue is that our samples are based mostly on college students, and it is useful to evaluate the relationships between axioms and values with adults from diverse walks of life. Second, we adopt a cross-cultural approach in our research, which emphasizes universal relationships across diverse cultures. On the other hand, culture-specific processes are definitely important, and are best examined by cultural and indigenous approaches (Greenfield, 2000; Shweder, 2001). For instance, there is some culture-specific research on values in Estonia (Aavik & Allik, 2002), and future research should integrate our findings with results based on cultural and indigenous approaches. Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
Social axioms and values
109
Third, we adopt a liberal definition of cross-cultural equivalence, which does not demand perfect congruence in a statistical sense, but instead is based on a high degree of similarity across cultures. This strategy is commonly adopted in large-scale cross-cultural projects because the range of these projects makes it almost impossible to obtain perfect cross-cultural congruence (Leung & Bond, 2004). Nonetheless, when we accumulate more knowledge about the specific relationships between axioms and values, fine-grained, meaningful cultural differences in the relationships between axioms and values may be identified and meaningfully interpreted. Forth, despite the low reliability of Social Complexity and Fate Control, these two dimensions are clearly identifiable across diverse cultural groups, and with college students as well as with adults (Leung & Bond, 2004). The lower reliability of these two dimensions does not seem to be too problematic for the present study for three reasons. First, the confirmatory factor analysis supports the proposed structures of both axioms and values, suggesting that the low reliability of some scales does not lead to a high level of noise. Second, the structural equation modelling approach adopted corrects for unreliability, and compensates for the low reliability of some scales. We also note that for axiom dimensions with an alpha 0.65, the relationships with values are not much stronger. For instance, social cynicism generally has good reliability, but it does not show strong relationships with values. This observation suggests that the negative impact of the low reliability of some scales is perhaps small. Third, the alphas we obtain are comparable to those reported in previous studies, in which meaningful results were obtained (e.g. Bond et al., 2004a). In sum, the totality of the results suggests that the issue of low reliability should not be a serious problem, but it is desirable to develop more reliable measures for evaluating the axiom-value relationships in future research. Finally, we have provided some speculations to explain the significant relationships between social axioms and values, and at this point it is uncertain whether axioms affect values, or values affect axioms, or they exert mutual influence on each other. We are only beginning to understand the relationships between axioms and values, and these complex issues await future evaluation. ‘‘In truths dependent on our personal action, then, faith based on desire is certainly a lawful and possibly an indispensable thing.’’ William James, The will to believe
REFERENCES Aavik, T., & Allik, J. (2002). The structure of Estonian personal values: A lexical approach. European Journal of Personality, 16, 221–235. Bardi, A., & Schwartz, S. H. (2003). Values and behavior: Strength and structure of relations. Personality and Social Psychology Bulletin, 29, 1207–1220. Barnea, M., & Schwartz, S. H. (1998). Values and voting. Political Psychology, 19, 17–40. Bentler, P. M., & Wu, E. J. C. (2004). EQS 6.1 for Windows [Computer software]. Encino, CA: Multivariate Software. Bilsky, W., & Schwartz, S. H. (1994). Values and personality. European Journal of Personality, 8, 163–181. Bond, M. H., Leung, K., Au, A., Tong, K.-K., & Chemonges-Nielson, Z. (2004a). Combining social axioms with values in predicting social behaviors. European Journal of Personality, 18, 177–191. Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
110
K. Leung et al.
Bond, M. H., Leung, K., Au, A., Tong, K.-K., Reimel De Carrasquel, S., & Murakami, F, et al. (2004b). Cultural-level dimensions of social axioms and their correlates across 41 cultures. Journal of Cross-Cultural Psychology, 35, 548–570. Carlson, D. D., & Perrewe, P. L. (1999). The role of social support in the stressor-strain relationship: An examination of work-family conflict. Journal of Management, 25, 513–540. Cheung, M. W. L., & Chan, W. (2004). Testing dependent correlation coefficients via structural equation modeling. Organizational Research Methods, 7, 206–223. Cheung, M. W. L., Leung, K., & Au, K. (2006). Evaluating multilevel models in cross-cultural research: An illustration with social axioms. Journal of Cross-Cultural Psychology, 37, 522–541. Feather, N. T. (1984). Protestant ethic, conservatism, and values. Journal of Personality and Social Psychology, 46, 1132–1141. Furnham, A., Kirkcaldy, B. D., & Lynn, R. (1994). National attitudes to competitiveness, money, and work among young people: First, second, and third world differences. Human Relations, 47, 119–132. Greenfield, P. M. (2000). Three approaches to the psychology of culture: Where do they come from? Where can they go? Asian Journal of Social Psychology, 3, 223–240. Goodwin, R., Nizharadze, G., Luu, L. A. N., Kosa, E., & Emelyanova, T. (2001). Social support in a changing Europe: An analysis of three post-Communist nations. European Journal of Social Psychology, 31, 379–393. Hirschfeld, R. R., & Feild, H. S. (2000). Work centrality and work alienation: Distinct aspects of a general commitment to work. Journal of Organizational Behavior, 21, 789–800. Hofstede, G. (1980). Culture’s consequences: International differences in work-related values. Beverly Hills, CA: Sage. Hofstede, G. (2001). Culture’s consequences: Comparing values, behaviors, institutions, and organizations across nations (2nd ed.). Thousand Oaks, CA: Sage. Hu, L., & Bentler, P. M. (1998). Fit indices in covariance structure modeling: Sensitivity to underparameterized model misspecification. Psychological Methods, 3, 424–453. Hunter, J. E., & Schmidt, F. L. (1990). Methods of meta-analysis. Orlando, FL: Academic Press. Jost, J. T., Glaser, J., Kruglanski, A. W., & Sulloway, F. J. (2003). Political conservatism as motivated social cognition. Psychological Bulletin, 129, 339–375. Kahn, J. H., & Scott, N. A. (1997). Predictors of research productivity and science-related career goals among counseling psychology doctoral students. Counseling Psychologist, 25, 38–67. Kurman, J., & Ronen-Eilon, C. (2004). Lack of knowledge of a culture’s social axioms and adaptation difficulties among immigrants. Journal of Cross Cultural Psychology, 35, 192–208. Leung, K., & Bond, M. H. (1989). On the empirical identification of dimensions for cross-cultural comparisons. Journal of Cross-Cultural Psychology, 20, 133–151. Leung, K., & Bond, M. H. (2004). Social axioms: A model of social beliefs in multi-cultural perspective. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 36, pp. 119–197). San Diego, CA: Academic Press. Leung, K., Bond, M. H., Reimel de Carrasquel, S., Mun˜oz, C., Herna´ndez, M., Murakami, F., Yamaguchi, S., Bierbrauer, G., & Singelis, T. M. (2002). Social axioms: The search for universal dimensions of general beliefs about how the world functions. Journal of Cross Cultural Psychology, 33, 286–302. Netemeyer, R. G., Johnston, M. W., & Burton, S. (1990). Analysis of role conflict and role ambiguity in s structural equations framework. Journal of Applied Psychology, 75, 145–157. Ohbuchi, K.-I., Fukushima, O., & Tedeschi, J. T. (1999). Cultural values in conflict management: Goal orientation, goal attainment, and tactical decision. Journal of Cross-Cultural Psychology, 30, 51–71. Olkin, I., & Finn, J. D. (1995). Correlation redux. Psychological Bulletin, 118, 155–164. Pearce, J. L., Branyiczki, I., & Bigley, G. A. (2000). Insufficient bureaucracy: Trust and commitment in particularistic organizations. Organization Science, 11, 148–162. Renner, W. (2003). A German value questionnaire developed on a lexical basis: Construction and steps toward a validation. Review of Psychology, 10, 107–123. Roccas, S., Sagiv, L., Schwartz, S. H., & Knafo, A. (2002). The Big Five personality factors and personal values. Personality and Social Psychology Bulletin, 28, 789–801. Rokeach, M. (1969). Value systems and religion. Review of Religious Research, 11, 2–23.
Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
Social axioms and values
111
Rokeach, M. (1972). Beliefs, attitudes, and values: A theory of organization and change. San Francisco, CA: Jossey-Bass. Ross, C. E., Mirowsky, J., & Pribesh, S. (2001). Powerlessness and the amplification of threat: Neighborhood disadvantage, disorder, and mistrust. American Sociological Review, 66, 568–591. Saroglou, V., Delpierre, V., & Dernelle, R. (2004). Values and religiosity: A meta-analysis of studies using Schwartz’s model. Personality and Individual Differences, 37, 721–734. Schwartz, S. H. (1992). Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 25, pp. 1–65). New York: Academic Press. Schwartz, S. H. (1994a). Beyond individualism/collectivism: New cultural dimensions of values. In U. Kim, H. C. Triandis, C. Kagitcibasi, S.-C. Choi, G. Yoon (Eds.), Individualism and collectivism: Theory, method, and applications (pp. 85–119). Thousand Oaks, CA: Sage. Schwartz, S. H. (1994b). Are there universal aspects in the content and structure of values? Journal of Social Issues, 50(4), 19–45. Schwartz, S. H. (1996). Value priorities and behavior: Applying a theory of integrated value systems. In C. Seligman, J. M. Olson, Mark P. Zanna (Eds.), The psychology of values (pp. 1–24). Mahwah, NJ: Lawrence Erlbaum. Schwartz, S. H., & Boehnke, K. (2004). Evaluating the structure of human values with confirmatory factor analysis. Journal of Research in Personality, 38, 230–255. Schwartz, S. H., & Huismans, S. (1995). Value priorities and religiosity in four Western religions. Social Psychology Quarterly, 58, 88–107. Schwartz, S. H., Melech, G., Lehmann, A., Burgess, S., Harris, M., & Owens, V. (2001). Extending the cross-cultural validity of the theory of basic human values with a different method of measurement. Journal of Cross-Cultural Psychology, 32, 519–542. Schwartz, S. H., Sagiv, L., & Boehnke, K. (2000). Worries and values. Journal of Personality, 68, 309–346. Shweder, R. A. (2001). Thinking through cultures: Expeditions in cultural psychology. Cambridge, MA: Harvard University Press. Singelis, T. M., Hubbard, C., Her, P., & An, S. (2003). Convergent validation of the social axioms survey. Personality and Individual Differences, 34, 269–282. Strack, F., & Deutsch, R. (2004). Reflective and impulsive determinants of social behavior. Personality and Social Psychology Review, 8, 220–247. Tang, T. L., & Weatherford, E. J. (1998). Perception of enhancing self-worth through service: The development of a service ethic scale. Journal of Social Psychology, 138, 734–743. Unger, J. B., Ritt-Olson, A., Teran, L., Huang, T., Hoffman, B. R., & Palmer, P. (2002). Cultural values and substance use in a multiethnic sample of California adolescents. Addiction Research and Theory, 10, 257–280. Vroom, V. H. (1964). Work and motivation. New York: Wiley. Zacharatos, A., Barling, J., & Iversion, R. D. (2005). High-performance work systems and occupational safety. Journal of Applied Psychology, 90, 77–93. Zimmerman, B. J., Bandura, A., & Martinez-Pons, M. (1992). Self-motivation for academic attainment: The role of self-efficacy beliefs and personal goal setting. American Educational Research Journal, 29, 663–676.
Copyright # 2007 John Wiley & Sons, Ltd.
Eur. J. Pers. 21: 91–111 (2007) DOI: 10.1002/per
