1

Running head: STRUCTURAL ORGANIZATION OF HUMAN VALUES

The Structural Organization of Human Values – Evidence from Three Rounds of the

European Social Survey (ESS)

Wolfgang Bilsky and Michael Janik

University of Muenster

Shalom H. Schwartz

The Hebrew University of Jerusalem and the University of Bergen

Keywords: Value Structure, European Social Survey, Societal Development

Corresponding author:

Prof. Dr. Wolfgang Bilsky, University of Muenster, FB 07/Psych. Inst. IV, Fliednerstrasse 21,

48149 Muenster, Germany, Tel.: 0049 251 83 34198, Fax: 0049 251 83 31343

Email: [email protected]
 2

Abstract

Since 1987, a multitude of studies referring to the Schwartz (1992) structural model of human

values have been published. Although most studies support this conceptual approach, few

were based on representative samples. The implementation of the biennial European Social

Survey (ESS) in 2002, made responses from 71 representative national samples from 32

countries to a 21-item version of the Portrait Values Questionnaire available for assessing this

model of human values. We present structural analyses of these data using a theory-based

approach to multidimensional scaling (Bilsky, Gollan & Döring, 2008) that can be applied to

optimally assess the fit of data to diverse theories. The analyses support the circular structure

of basic values across countries and within countries across time. They also replicate two

findings based on other samples, surveys, and methods of analysis (Fontaine et al., 2008):

Deviations from the structure are fewer and the contrast between protection and growth values

is sharper in more developed societies.

 3

Introduction

Universals in the content of human values and their structural organization have been

the focus of research for more than two decades (Schwartz & Bilsky, 1987, 1990; Schwartz,

1992, 2006). In this context, ordinal multidimensional scaling (MDS) has been a central

approach for analyzing value structures. Starting from regional hypotheses, MDS displays the

discriminability of values in an easily accessible geometric representation. Furthermore, this

approach is relatively free of mathematical restrictions and additional assumptions not

relevant to the problem under study (Borg & Shye, 1995).

Studies using the Schwartz Value Survey (SVS) measured values in teacher, university

student, and other adult samples in 67 countries from 1988 to 2007 (Schwartz, 1992, 2007). In

1994, Schwartz developed an alternative instrument, the Portrait Values Questionnaire (PVQ),

that is cognitively less demanding than the SVS and can be applied to younger and less

educated samples (Schwartz, 2006; Schwartz, et al., 2001). Since then the scope of studies has

expanded considerably, including samples of adolescents and children (Bubeck & Bilsky,

2004; Bilsky, Niemann, Schmitz & Rose, 2005; Boehnke & Welzel, 2006; Hofmann–

Towfigh, 2007). Most of these studies corroborated Schwartz (1992) assumptions about ten

basic values (Table 1, columns 1 and 2) and their structural relations (Figure 1) as outlined in

his revised values model (Schwartz, 1992, p. 45). However, few strictly representative

national samples were studied.

With the inclusion of a short form of the PVQ in the European Social Survey (ESS; see

items in Table 1, column 3), data from many representative national samples are now

available to assess the Schwartz theory of basic values. The ESS is a biennial multi–country

survey in which 32 countries have participated at least once, thus far. We analyze the data on

value preferences in 71 samples from ESS rounds 1-3, completed between 2002 and 2007.

Two sources have reported analyses of the structure of values in the first ESS round.

Schwartz (2007) reported complete support for the theorized structure in 15 countries and
 4

substantial support in the remaining 5 countries. In contrast, Mohler, Rammstedt and Wohn

(2006) claimed that the “German ESS data showed, like those of many other cultures, quite

some deviations from the ideal value circle ...” (p. 257). Neither of these sources provided

sufficient information about their multidimensional scaling methods to permit a comparative

evaluation. We therefore undertook to analyze the PVQ data available from all three ESS

rounds, documenting our statistical approach to multidimensional scaling in a way that allows

easy reproduction by others for future ESS rounds and other uses of the short form of the

PVQ. Data from the first three ESS rounds allow not only an extensive assessment of the fit

of the theorized circular structure of ten basic values in representative national samples, but

also examination of the stability of the structure within countries over time and comparisons

of the structure across countries.

Fontaine, Poortinga, Delbeke, and Schwartz (2008) examined the cross-cultural

equivalence of the structure of values, measured with the SVS, in samples of teachers and

students from 38 countries. They found that the average value structure across samples

corresponded very well with the circular structure of values posited by the Schwartz value

theory. Sampling fluctuations accounted for most of the observed deviations from the average

structure. However, there were also systematic patterns of variation across countries. We

examine whether these patterns of variation replicate.

Specifically, Fontaine et al. (2008) reported that the higher the level of societal

development of a country, the less the structure of values in the sample deviated from the

average value structure and the stronger the contrast between two sets of values, ‘growth’

versus ‘protection’. These sets differ based on their relations to anxiety (Schwartz, 2005b,

2006).1 Growth values (self-direction, universalism, benevolence, stimulation, and hedonism)

express anxiety-free self-expansion; protection values (security, power, achievement,

1
Anxiety relations provide a supplementary explanation for the organization of the basic value structure by
contrasting self-transcendence and openness to change values (growth) with self-enhancement and conservation
values (protection).
 5

conformity, and tradition) express anxiety-based self-protection. Fontaine et al. (2008) linked

these patterns to the differences in value forming experiences tied to societal development.

The current study examines whether similar systematic patterns of variation are found when

values are measured with a different instrument, in representative national samples, and in a

set of countries that includes 14 not studied by Fontaine et al. (2008).

Method

Data and Samples

Data from the first three rounds were downloaded from the data archive of the European

Social Survey2 (ESS13, ESS24, and ESS35). Because missing data and failure to discriminate

among value items might distort results, we first cleaned the data following the procedure

described by Schwartz (2005a)6. We excluded persons with more than 5 missing responses

and those who gave the same answer to more than 16 value items.

Next, we computed Pearson correlation coefficient matrices from the unweighted

responses to the 21 PVQ–items, separately for each sample. We applied listwise deletion for

missing cases. These correlation matrices provided the basis for our MDS analyses. Table 2

lists the countries studied in each ESS round and, in columns 2-6, it presents the sample sizes

before and after data cleaning and data processing.

2
http://www.europeansocialsurvey.org/; data archive and distributor of the ESS data: Norwegian Social Science
Data Services (NSD), retrieved 31.10.08.
3
ESS1: http://ess.nsd.uib.no/index.jsp?year=2003&country=&module=download, retrieved 11.01.08.; R Jowell
and the Central Co-ordinating Team, European Social Survey 2002/2003: Technical Report, London: Centre for
Comparative Social Surveys, City University (2003).
4
ESS2: http://ess.nsd.uib.no/index.jsp?year=2005&country=&module=download, retrieved 17.09.08; data for
Italy were added from ESS2IT: http://ess.nsd.uib.no/index.jsp?year=2005&country=&module=download,
retrieved: 29.10.08; R Jowell and the Central Co-ordinating Team, European Social Survey 2004/2005:
Technical Report, London: Centre for Comparative Social Surveys, City University (2005).
5
ESS3: http://ess.nsd.uib.no/index.jsp?year=2007&country=&module=download, retrieved 07.07.08; data for
Latvia and Romania were added from ESS3LVRO: http://ess.nsd.uib.no/index.jsp?
year=2007&country=&module=download, retrieved 07.07.08; R Jowell and the Central Co-ordinating Team,
European Social Survey 2006/2007: Technical Report, London: Centre for Comparative Social Surveys, City
University (2007)
6
SPSS-Syntax used for data cleaning, see: http://essedunet.nsd.uib.no/cms/topics/1/4/all.html; retrieved
31.10.08.
 6

In order to assess relations of societal development to deviations from the theorized

value structure and to the strength of the contrast between protection and growth values, we

constructed an index of societal development in a manner parallel to that used by Fontaine et

al. (2008).7 We performed a principal components analysis of seven indicators of societal

development: the Human Development Index, country wealth, occupational diversification,

level of education, life expectancy, level of democracy, and democratic participation. A single

factor accounted for 74.52% of the total variance; the remaining factors had eigen values

below .80. Countries with higher scores on this factor were higher on each of the indicators.

Norway scored highest on this social development index (SDI) in our data set and Ukraine

scored lowest. We report relations of this index to the patterns of variation because it captures

relations of each of the component context characteristics quite well.

A Theory-based MDS Approach

Numerous factors may affect the results of multidimensional scaling (Borg & Groenen,

2005; Borg & Staufenbiel, 2007; Shye, 1985). The measures of proximity used, the loss

function, and the respective starting configuration may all make a difference. Many statistical

programs are available, such as FSSA, HUDAP–SSA, KYST, SPSS–ALSCAL and –

PROXSCAL, and SYSTAT–MDS (see the detailed overview in Borg & Groenen, 2005), that

have different default options and choices. Hence, failure to report the many statistical

decisions and computational steps taken prevents exact replication and may result in

confusion at best and misinterpretation of findings at worst.

We chose a theory-based MDS approach for our analyses. Central to this approach is a

starting configuration which assigns every variable (i.e., every value item) its place within the

hypothesized structure of values (this is also called a “weak confirmatory” MDS by Borg &

Staufenbiel, 2007; cf. Bilsky, 2008). Of course, a strict confirmatory MDS with regional

restrictions would be an optimal choice. However, its routine application is impractical

7
This index of societal development correlated .86 (p<.001) with the index used by Fontaine et al. (2008), across
the 17 countries in both datasets.
 7

because of far-reaching statistical implications and computational demands which cannot be

handled by standard software packages (for special applications, see Groenen & van der Lans,

2006; Borg, Groenen, Jehn & Bilsky, 2009).

Multidimensional scaling using a theory-based starting configuration is appropriate for

our purposes because Schwartz (1992) offers an explicit, theoretically grounded hypothesis

about the structure of the values. More general methodological considerations in Borg and

Groenen (2005) also support this choice which can be used to optimally assess diverse

theories:

The MDS program optimizes Stress, which is substantively blind: that is, it is not

tailored to the particular questions that are being asked. ... Minimizing Stress gives a

solution that is locally optimal. Yet, other local minimum solutions may exist with a

similar Stress, or possibly even with lower Stress ... The question is which solution

should be preferred. If a hypothesis for the data is available, then, of course, we would

be particularly interested in the solution that most directly speaks to this hypothesis.

This is obviously the solution that most closely satisfies the hypothesis…. (p. 228;

italics added by authors)

The value items represent the 10 more general basic values in the values theory. The

hypothesized overall structure of relations among the 10 values determines the hypothesized

locations of the 21 items relative to one another. We therefore deduced a design matrix of

values from the Schwartz model (Figure 1) as a first step (Bilsky, Gollan & Döring, 2008).

Design matrix. In this model (Schwartz, 1992, p. 45), the 10 values are represented by

nine sectors (Figure 1). One sector is further divided into an inner and an outer segment, each

of which represents a different value (conformity and tradition). Equal spacing of the nine

sectors at 40° angles is not a defining feature of the Schwartz model. Nonetheless, such a

simple and regular structure is functional. In the absence of evidence for a more specific

structure, it may be adequate.

 8

The nine sectors serve as the basis for specifying the prototypical location of each value

by corresponding coordinates (Bilsky, Gollan & Döring, 2008). We determine the coordinates

trigonometrically by referring to the unit circle and summarize them in the design matrix.

Nine of the 10 values are represented by points on the periphery of this circle; their

coordinates derive from the centre of that circular arc which is marked by the respective

(value) sector. The coordinates of the tenth value (conformity) are determined in the same

way, though with a radius of 0.5 instead of 1.0. Table 3 shows the design matrix and Figure 2

shows the prototypical locations of the ten values in a two–dimensional space.

Starting configuration. As a second step, the starting configuration for all the value

items is defined. In the PVQ21, two items operationalize each of nine values and three items

operationalize the tenth (universalism). The starting configuration, like the design matrix,

should represent the prototypical structure of values. Therefore, all items that index the same

value receive the coordinates specified for that value in the design matrix.

Data analysis

Our structural analyses of the ESS values data were accomplished with PROXSCAL, an

MDS program in SPSS. This program offers many options for multidimensional scaling. The

user can choose between different proximities, proximity transformations, restrictions on

common space, and initial (starting) configurations. We analyzed the matrices of Pearson

correlation coefficients between the 21 PVQ–items with ordinal MDS (defaults for ties and

iteration criteria: keepties; stress convergence=.0001, minimum stress=.0001, maximum

iterations=100), using the starting configuration described before.

Results

Given the many samples in our study, we first introduce the general form of the two–

dimensional graphical MDS–splits of value items by referring to the three German ESS

samples. Table 4 summarizes findings of the analyses in each sample and ESS round.8
8
A comprehensive research report including all MDS plots per country and ESS round is available from the first
author ([email protected]).
 9

Exemplary Analysis: Values Structure in Germany

Figure 3a, 3b, and 3c plot results of the MDS–analyses in the three German samples. To

provide a simple, clear overview, the value items have been replaced by the numbers of the

values they represent as shown in Figure 2. The sequence of the numbers corresponds to the

circular structure of Schwartz original model (1992, p. 14) in which all ten values were

represented by separate sectors. The plots present two alternative splits for tradition (3) and

conformity (4), one referring to the original model (dashed line) and the other to the revised

model (solid line; see Figure 1).

Figure 3 reveals that the MDS–splits of all German samples correspond perfectly to the

Schwartz (1992) structural theory. This holds for both the original (p. 14) and the revised

model (p. 45) of the value structure. Note that in this context bent lines pose no problems with

respect to interpretation as long as a particular value region does not include value items of a

different basic value (Borg & Shye, 1995; Shye, Elizur & Hoffman, 1994).

Structural Analyses in All Samples

Table 4 summarizes results of the MDS structural analyses. It includes information

about all observed deviations from the hypothesized circular structure. As can be seen, in

42/71 samples, the theorized circular order of values was perfectly reproduced. Moreover,

every one of the deviations, whether reversals of the order around the circle or mixing of two

values, involved values that are adjacent in the circle.

One may ask whether providing a theory-based starting configuration biases the

resulting MDS structure in favor of the theory at the expense of the adequacy of the fit with

the data. To address this concern we also performed an MDS in each sample using the three

other options for starting configurations in the PROXSCAL program, Simplex (the default),

Torgerson, and Random (we ran 1000 random starts). Columns 7-10 of Table 2 report the

stress-1 values obtained with each type of starting configuration. As can be seen, these stress

values are virtually the same. The mean stress values across all samples are: Simplex=0.121,
 10

Torgerson=0.117, Random=0.116, and Custom (the theory-based starting configuration)

=0.118 (additional analyses with stress convergence = .000001 and maximum number of

iterations = 1000 did not result in reduced stress values or different MDS structures). Thus, by

using a theory-based starting configuration one can obtain an optimal assessment of the match

between the data and the theory at no loss in accuracy of representing the data. In contrast,

using any other starting configuration with the same data may reach an alternative local

minimum solution that is less similar to the hypothesized structure.

Figure 4 illustrates the dependence of the MDS solution on the starting configuration.

It compares the solutions obtained in the Czech values data of ESS round 2 with the theory-

based starting configuration and the default starting configuration (Simplex). With the theory-

based starting configuration, this sample showed two deviations from the hypothesized

circular structure (UN peripheral to BE, HE and ST mixed). With the Simplex starting

configuration, there were additional deviations--mixing of AC with HE and ST and location

of one UN item in the SE region.

Relations of Societal Development to Value Structure

Extent of Deviations.9 In order to examine relations of societal development to the extent

of deviations from the theorized circular structure, we counted the number of deviations per

sample within each country. For example there was one deviation in the three samples in

Austria, yielding a count of .33, and 3 deviations in the one sample in Cyprus, yielding a

count of 3.00. The correlation of the number of deviations with SDI across the 32 countries

was -.649 (p<.001). Thus, the higher the societal development in a country, the less the

observed value structure deviated from the theorized circular structure. SDI accounted for a

substantial 42% of the variance in the number of deviations.

9
We assessed deviations from the theory-based structure whereas Fontaine et al. (2008) focused on deviations
from the overall average structure. This reflects our primary interest in the fit of the theory to the data in all
countries in contrast to their interest in the stability of the structure across countries. Our approach also sheds
light on the stability of the hypothesized structure.
 11

Contrast between Protection Values and Growth Values. The method we used to assess

the strength of the contrast between protection and growth values followed the approach

Fontaine et al. (2008) used with their raw SVS data.10 They computed two scales, one

including the items that shifted toward the protection pole of the protection-growth value

dimension with increased SDI and one including the items that shifted toward the growth

pole. The 21 PVQ value items parallel many of the items in the SVS. We identified seven

PVQ items (items 2,4,5,7,13,14 and 17) whose contents parallel those of the Fontaine et al.

(2008) SVS protection-shift items and five PVQ items (1,6,8,18,19) whose contents parallel

those of their SVS growth-shift items. We then computed a scale for each based on each

individual’s centered responses. Individuals from all ESS rounds constituted the sample for

each country.

For the 32 countries, the within-country correlation between the two scales ranged from

-.669 to -.418. The more negative this correlation, the stronger the contrast between protection

and growth values in the country. To examine relations of the strength of the protection-

growth contrast with societal development, we correlated the SDI with the correlation

between the two scales (Fisher’s z transformed) across the 32 countries. The resulting

correlation of -.852 (p<.001) indicates that the higher the level of societal development of a

country, the stronger the contrast between protection and growth values.

Discussion

When evaluating the empirical structures, it is important to consider the form of any

deviations. Deviations from the Schwartz model may take several forms. A first type of

deviation is failure to find a distinct region for each value. Table 4 indicates such deviations

by [x+y], with x and y standing for the values that mix together (e.g., Finland, round 2, TR +

CO). If these values are adjacent in the hypothesized circle (see Figure 1), the deviation is of

minor importance. Schwartz and Sagiv (1995) demonstrated that such deviations are likely to
10
Because the strength of this contrast depends on correlations among raw responses, our approach to assessing
it does not use the MDS solutions we obtained.
 12

be chance variations. They may well be ignored, especially if the mixed values form distinct

regions in the country in other ESS rounds, as they did in Finland.

A second form of deviation is a peripheral rather than side-by-side split between values.

The values form distinct regions but one emerges behind the other rather than in wedge–like

regions around the circle. The value theory hypothesizes such a split for tradition (3) and

conformity values (4) but not for other values. Table 4 indicates such deviations by [x/y], with

x standing for the peripheral and y for the central value (e.g., Austria, round 3, CO and SE). If

the two values are adjacent in the hypothesized value structure, this type of deviation too is

likely to be a chance variation. It too, therefore, is of minor importance, especially if the

hypothesized order is found in other ESS rounds.

A third form of deviation is a reversal of the order of values. Table 4 indicates such

deviations by a reversal in the order of the numbers that stand for the values (e.g., Turkey,

round 2, CO and SE). Reversals of adjacent values are less important than other reversals,

especially if they do not recur across ESS rounds. If reversals recur, it is worth considering

possible conceptual or methodological reasons for this shift.

Finally, a fourth form of deviation is the mislocation of single items to the spatial region

of another value. Table 4 indicates such deviations in the ‘deviations’ column by listing the

item and its location (e.g., CO_7 in Hungary, round 3, between SE and PO). Such

mislocations are more serious if they entail location in a region distant from the hypothesized

region of the item and if they recur across ESS rounds. The only two such mislocations in the

71 samples analyzed here were neither distant from their hypothesized region nor did they

recur across ESS rounds. Should more serious deviations appear in a sample, however, it

would suggest either problems of translation or a difference between the meaning of the value

item in the sample and its intended meaning that is confirmed in most other samples.

Five countries exhibited recurrent deviations from the hypothesized value structure. In

Estonia, hedonism and stimulation values were mixed together rather than adjacent in both
 13

rounds. In Hungary, universalism was peripheral to benevolence rather than adjacent to it in

all three rounds. In Poland as in Hungary, universalism was peripheral to benevolence in all

three rounds. In Portugal, security reversed with tradition and conformity in all three rounds.

In Sweden, security and conformity mixed in two of three rounds. Each of these deviations is

minor, but their recurrence suggests that they merit special attention. They may reflect

possible artifacts (e.g., problematic translation of one or two items) or, more interestingly,

culture-specific organizations of values. The fact that all of these deviations were due to

mixing or shifts in location between values adjacent in the theorized structure suggests that

whatever culture-specific variation there is in the significance of the values is not great.

The fact that these country deviations replicated across samples from the different

rounds points to the stability of the value structure over time when using the same instrument

to measure values. In nine countries that participated in all three ESS rounds, the theorized

structure replicated across all three rounds, further supporting the stability of the structure

across time.

In six of the 10 countries that participated in only one ESS round, minor deviations of

the sort Schwartz and Sagiv (1995) demonstrated as likely to be due to chance were found

(Bulgaria, Cyprus, Italy, Latvia, Romania, and Turkey). It is probably too early to attribute

significance to these deviations in the absence of evidence that they recur.

One specific deviation occurred many more times than any other. This was the location

of universalism peripheral to benevolence (15 times). According to the values theory, this pair

of values shares the motivation to transcend self-interest and to seek to benefit others.

Analyses of values data with earlier questionnaires also revealed that universalism values

emerged peripheral to benevolence in a substantial minority of cases (Schwartz, 1992, 2006).

The peripheral location was attributed to the greater abstractness of universalism compared

with benevolence. Universalism values concern the welfare of those in the wider society and

world outside the ingroup rather than the welfare of close others. Similar reasoning explained
 14

the location of tradition values peripheral to conformity values (Schwartz, 1992, 2006). Both

values express the motivation to submit to external expectations and avoid conflict. Tradition

values refer to submission to more abstract sources of expectation such as religion and

accepted customs, conformity refers to submission to the expectations of those with whom

one interacts in everyday life.

In sum, the vast majority of our analyses corroborate the Schwartz model (see Table 4,

and the individual plots available from the authors for all samples). This finding suggests that

this theoretical approach is a conceptually and empirically appropriate and sound basis for

cross–cultural research. Regarding the contradiction between the Schwartz (2007) and Mohler

et al. (2006) findings cited at the outset, the current analyses clearly support the former. By

explicitly specifying the detailed procedures in carrying out the structural analyses, such

controversies can be avoided in the future. The current article demonstrates that one can avoid

methodological artifacts and arrive at unequivocal empirical results by using a well–

documented MDS with a theory-based starting configuration.

Our findings about the circular structure of human values are in line with other research

using confirmatory factor analysis (e.g., Schwartz & Boehnke, 2004) and support the general

stability of the value model. Nonetheless, we too observed two systematic patterns of

variation in the findings that parallel variations reported by Fontaine et al. (2008). We

measured values with a different instrument, in representative national samples rather than in

teacher and student samples, in a set of countries that overlapped only partly with the set they

studied, and we used somewhat different measures of deviation, of contrast in values, and of

societal development. Despite these differences, both patterns of variation replicated in this

study. Specifically, the higher the level of societal development, the fewer the deviations from

the theoretical circular structure of ten basic values and the stronger the contrast between

protection and growth values.11

11
Related findings are reported by Fischer, Milfont & Gouveia (2009) who compared the value structure of
physician samples across the Brazilian states using another value instrument.
 15

Why might societal development reduce deviations from the theorized structure of

values? Fontaine et al. (2008) offered three explanations that imply that the observed

association is a methodological artifact. First, they noted that high SDI countries were

overrepresented in their dataset. This influenced the overall average structural configuration

to which they compared each country configuration. The greater influence of high SDI

countries on the target of comparison could create a spurious impression of more structural

deviation in low SDI countries. This explanation cannot apply to our findings, however. We

compared each country configuration to the theory-based structural configuration that is

unaffected by the countries in the dataset rather than to an average configuration.

Second, Fontaine et al. (2008) noted that shifts in respondents’ understandings of items

due to translation or context of administration might be greater in the less developed societies

whose languages were more distant from the languages of the majority of samples that

contributed to the overall average configuration. In the current study, the context of

administration was quite similar in all samples; respondents completed the PVQ during an

individual face-to-face interview. Moreover, despite use of 27 languages in the ESS samples,

the rate of deviation from the theory-based configuration (which privileges no particular

language) differed little across the major language groups as well as the non-Indo-European

languages (Turkish and Hebrew). Hence translation and context differences apparently do not

explain the association between SDI and the extent of deviation from the hypothesized value

structure.

Third, Fontaine et al. (2008) noted that people in less developed countries may have less

experience using numerical scales such as those in the SVS. This could lead to less accurate

communication of their value priorities by people from low SDI countries and hence to more

deviant value structures. Although a lack of ‘test wiseness’ might also apply in the current

study, it is less likely to have been a problem for three reasons: The PVQ does not use a

numerical response scale, its items are more concrete than those of the SVS, and a response
 16

format that involves comparisons between others and self draws upon people’s everyday

experience in all societies.

It therefore seems likely that the observed differences in the rate of deviation from the

hypothesized value structure may reflect real differences in how well-articulated people’s

value systems are in societies at different levels of development. Fontaine et al. (2008) cited

the findings of Schwartz (2004) indicating that cultures in more developed countries

encourage individual autonomy more and cultivate people’s unique personal ideas and

preferences. The stronger cultural autonomy and weaker embeddedness in developed societies

mean that prevailing norms and customs impose fewer constraints on individual decision

making and behavior. This confronts people with more opportunities and obligations to make

choices, the main vehicle for developing articulated value systems (Schwartz, 2006). This, in

turn, would lead to fewer deviations from the theory-based value structure in higher SDI

countries.

The explanations Fontaine et al. (2008, p. 362) offer for the greater contrast between

protection and growth values in more developed societies also pertain to our replication of

this finding. We briefly recapitulate their arguments applied to our setting.

Increasing SDI in the set of ESS countries largely reflects a shift in their labor markets

from jobs in agriculture, construction, heavy manufacturing and other primary and secondary

occupations to jobs in service and information industries. The latter demand more innovation,

creativity, critical thinking, complex interpersonal skills, and less routine activity. The nature

of jobs, socialization in smaller families, exposure to higher education, and prevailing cultural

autonomy in more developed societies all encourage an interest in what is unique, novel,

challenging, and different. In other words, they encourage people to pursue growth values.

But such pursuit entails forgoing to some extent the security of the status quo, of conformity

to norms and traditions. That is, it entails sacrificing protection values, the primary default

option of people according to Maslow (1959) and Higgins (1997). As a result, people in more
 17

developed countries are likely to experience conflict between protection and growth values

more frequently than are people in less developed countries. This gives rise to the stronger

contrast between protection and growth values in more developed societies.

In summary, the vast majority of our analyses corroborate the circular model of the

structure of human values. These analyses also provide detailed guidance on how best to

assess the fit of new data to this model by using MDS with a theory-based starting

configuration. This approach can also be applied to other instruments used to measure basic

values such as the SVS and to instruments that operationalize other theories.

Even when this optimal starting configuration is used with the PVQ, however, there are at

least two systematic variations in the pattern of value relations. On the one hand, the circular

model fits somewhat less well in less developed societies, perhaps, as we have argued,

because the culture and organization of experience in these societies confronts people with

fewer opportunities and demands to make choices independently among alternatives, the

vehicle through which value systems take shape. In addition, the contrast between protection

and growth values is sharper in more developed societies. This may also reflect greater

opportunities and demands to make choices in more developed societies, specifically between

maintaining the status quo and innovating. Future studies may uncover other variations within

the basic structure of values that arise as a consequence of cross-national differences in

culture and social structure and consequently in the everyday experiences and choices through

which individuals crystallize their value systems.

 18

References

Bilsky, W. (2008). Value structure – On the use of weakly constrained confirmatory MDS.

Paper presented at the 19th International Congress of the International Association for

Cross–Cultural Psychology July 27–31, 2008, Bremen, Germany. (Berichte aus dem

Psychologischen Institut IV, 32/2008. Münster: Westfälische Wilhelms–Universität).

Bilsky, W., Gollan, T., & Döring, A. (2008). Análise confirmátoria de escalonamento

multidimensional (EMD) de valores baseda em uma matriz de desenho. In M.L. Mendes

Teixeira (Ed.), Valores humanos e gestão (pp. 213–221). São Paolo: Senac.

Bilsky, W., Niemann, F., Schmitz, J., & Rose, I. (2005). Value structure at an early age:

Cross–cultural replications. In W. Bilsky & D. Elizur (Eds.), Facet Theory: Design,

analysis and applications. Proceedings of the 10th International Facet Theory

Conference in Rome, 10–13 July 2005 (pp. 241–248). Prague: Agentura Action M.

(http://miami.uni–muenster.de/servlets/DerivateServlet/Derivate–4360/

bilsky_elizur_2005.pdf)

Boehnke, K., & Welzel, C. (2006). Wertetransmission und Wertwandel: Eine explorative

Drei–Generationen–Studie. Zeitschrift für Soziologie der Erziehung und Sozialisation,

26, 341–360.

Bubeck, M., & Bilsky, W. (2004). Value structures at an early age. Swiss Journal of Psychology,

63, 31–41.

Borg, I., & Groenen, P. (2005). Modern Multidimensional Scaling. Berlin: Springer.

Borg, I., Groenen, P., Jehn, K.A., & Bilsky, W. (2009). Embedding the organizational culture

profile into Schwartz`s universal value theory using multidimensional scaling with

regional restrictions. Manuscript submitted for publication.

Borg, I., & Shye, S. (1995). Facet theory: form and content. Newbury Park, CA: Sage.

Borg, I., & Staufenbiel, T. (2007). Theorien und Methoden der Skalierung. Huber: Bern.
 19

Fischer, R., Milfont, T.L., & Gouveia, V.V. (2009). Does social context affect value

structures? Testing the within-country stability of value structures with a functional

theory of values. Journal of Cross-Cultural Psychology (submitted).

Fontaine, J.R.J., Poortinga, Y.H., Delbeke, L., & Schwartz, S.H. (2008). Structural

equivalence of the values domain across cultures. Journal of Cross-Cultural

Psychology, 39, 345-365.

Groenen, P. & van der Lans, I. (2006). Multidimensional scaling with regional restrictions for

Facet Theory: An application to Levy’s political protest data. In M. Braun & P.Ph.

Mohler (Eds.), Beyond the horizon of measurement (ZUMA-Nachrichten Spezial, Vol.

10, pp. 41-64). Mannheim: ZUMA.

Higgins, E.T. (1997). Beyond pleasure and pain. American Psychologist, 52, 1280-1300.

Hofmann–Towfigh, N. (2007). Do students‘ values change in different types of schools?

Journal of Moral Education, 36, 453–473.

Maslow, A. H. (Ed.) (1959). New knowledge in human values. New York: Harper.

Mohler, P., Rammstedt, B., & Wohn, K. (2006). The great value divide – testing Hofstede’s

convergence assumption. In P. Ester, M. Braun & P. Mohler (Eds.), Globalization, value

change, and generations (pp. 251–272). Leiden: Brill.

Schwartz, S. H. (1992). Universals in the content and structure of values: Theoretical

advances and empirical tests in 20 countries. In M. Zanna (Ed.), Advances in

experimental social psychology (Vol. 25, pp. 1–65). New York: Academic Press.

Schwartz, S. H. (2004). Mapping and interpreting cultural differences around the world. In H.

Vinken, J. Soeters, & P Ester (Eds.), Comparing cultures, Dimensions of culture in a

comparative perspective (pp.43-73). Leiden, The Netherlands: Brill.

Schwartz, S. H. (2005a). Human values. European Social Survey Education Net. Retrieved

from: http://essedunet.nsd.uib.no/cms/topics/1/, 31.10.2008.

 20

Schwartz, S. H. (2005b, July). The structure and implications of individuals’ value systems.

Keynote address at the 1st European Association for Survey Research Conference,

Barcelona, Spain.

Schwartz, S. H. (2006). Les valeurs de base de la personne: Théorie, mesures et applications.

[Basic personal values: Theory, measurement and application.] Revue française de

sociologie, 42, 249-288.

Schwartz, S. H. (2007). Value orientations: measurement, antecedents and consequences

across nations. In R. Jowell, C. Roberts & R. Fitzgerald (Eds.), Measuring attitudes

cross-nationally - lessons from the European Social Survey (pp. 161-193). London:

Sage.

Schwartz, S. H., & Bilsky, W. (1987). Toward a universal psychological structure of human

values. Journal of Personality and Social Psychology, 53, 550–562.

Schwartz, S. H., & Bilsky, W. (1990). Toward a theory of the universal content and structure

of values: Extensions and cross–cultural replications. Journal of Personality and Social

Psychology, 58, 878–891.

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., 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. (1995). Identifying culture–specifics in the content and structure

of values. Journal of Cross–Cultural Psychology, 26, 92–116.

Shye, S. (1985). Multiple Scaling. Amsterdam: North–Holland.

Shye, S., Elizur, D., & Hoffman, M. (1994). Facet theory. Newbury Park, CA: Sage.
 21

Table 1

Basic Values, Core Goals, and PVQ–Items in the European Social Survey (ESS)

Basic Value Core Motivational Goal PVQ–Items as numbered & labeled in the ESS
3. Ipeqopt: He thinks it is important that every person
UN Understanding, appreciation, in the world should be treated equally. He believes everyone
should have equal opportunities in life.
Universalism tolerance and protection for the 8. Ipudrst: It is important to him to listen to people
welfare of all people and for who are different from him. Even when he disagrees with them,
nature. he still wants to understand them.
19. Impenv: He strongly believes that people should
care for nature. Looking after the environment is important to
him.

BE Preservation and enhancement 12. Iphlppl: It is very important to him to help the
people around him. He wants to care for their well–being.
Benevolence of the welfare of people with 18. Iplylfr: It is important to him to be loyal to his friends. He
whom one is in frequent wants to devote himself to people close to him.
personal contact.

TR Respect, commitment and 9. Ipmodst: It is important to him to be humble and

modest. He tries not to draw attention to himself.
Tradition acceptance of the customs and
20. Imptrad: Tradition is important to him. He tries
ideas that one's culture or to follow the custom handed down by his religion or his family.
religion impose on the
individual.
7. Ipfrule: He believes that people should do what
CO Restraint of actions, they are told. He thinks people should follow rules at all times,
even when no–one is watching.
Conformity inclinations, and impulses 16. Ipbhprp: It is important to him always to behave
likely to upset or harm others properly. He wants to avoid doing anything people would say is
and violate social expectations wrong.
or norms.
5. Impsafe: It is important to him to live in secure
SE Safety, harmony and stability surroundings. He avoids anything that might endanger his
safety.
Security of society, of relationships, and 14. Ipstrgv: It is important to him that the
of self. government ensures his safety against all threats. He wants the
state to be strong so it can defend its citizens.
2. Imprich: It is important to him to be rich. He
PO Social status and prestige, wants to have a lot of money and expensive things.
17. Iprspot: It is important to him to get respect from
Power control or dominance over others. He wants people to do what he says.
people and resources.
4. Ipshabt: It is important to him to show his abilities.
AC Personal success through He wants people to admire what he does.
13. Ipsuces: Being very successful is important to
Achievement demonstrating competence him. He hopes people will recognize his achievements.
according to social standards.
10. Ipgdtim: Having a good time is important to him.
HE Pleasure and sensuous He likes to "spoil" himself.
21. Impfun: He seeks every chance he can to have
Hedonism gratification for oneself. fun. It is important to him to do things that give him pleasure.

6. Impdiff: He likes surprises and is always looking

ST Excitement, novelty, and for new things to do. He thinks it is important to do lots of
different things in life.
Stimulation challenge in life. 15. Ipadvnt: He looks for adventures and likes to take
risks. He wants to have an exciting life.
1. Ipcrtiv: Thinking up new ideas and being creative
SD Independent thought and action is important to him. He likes to do things in his own original
way.
Self choosing, creating, exploring. 11.Impfree: It is important to him to make his own
Direction decisions about what he does. He likes to be free and not
depend on others.
 22

Table 2

Participating countries, ESS round, sample size (N), and Stress 1 for different starting configurations

N after computing Stress 1 Stress 1

N N
ESS- Pearson Deleted Custom Stress 1 Stress 1 1000
Country raw after data
round correlations cases Schwartz Simplex Torgerson Random
data cleaning
listwise deletion 1992 starts

Austria 1 2257 2203 2189 14 .081 .083 .082 .081

2 2256 2179 1991 188 .086 .086 .086 .086
3 2405 2326 2110 216 .095 .096 .095 .095

Belgium 1 1899 1819 1692 127 .116 .117 .116 .116

2 1778 1734 1671 63 .115 .114 .115 .114
3 1798 1767 1743 24 .121 .121 .121 .121

Bulgaria 3 1400 1248 981 267 .110 .112 .112 .110

Cyprus 3 995 933 809 124 .117 .116 .115 .115

Czech 1 1360 1213 1064 149 .108 .100 .096 .095

Republic 2 3026 2445 2118 327 .102 .108 .103 .101

Denmark 1 1506 1457 1363 94 .124 .130 .118 .118

2 1487 1457 1331 126 .120 .121 .120 .120
3 1505 1451 1376 75 .125 .126 .126 .124

Estonia 2 1989 1931 1768 163 .123 .121 .116 .108

3 1517 1420 1265 155 .123 .125 .122 .122

Finland 1 2000 1758 1705 53 .113 .117 .115 .113

2 2022 1692 1556 136 .113 .126 .114 .112
3 1896 1645 1077 568 .130 .142 .129 .128

France 1 1503 1312 1232 80 .118 .118 .118 .118

2 1806 1661 1575 86 .122 .122 .122 .121
3 1986 1948 1880 68 .143 .143 .143 .143

Germany 1 2919 2785 2685 100 .096 .095 .094 .094

2 2870 2800 2640 160 .109 .109 .108 .108
3 2916 2828 2706 122 .110 .110 .110 .110

Greece 1 2566 2453 2413 40 .108 .137 .108 .107

2 2406 2293 2239 54 .118 .111 .112 .109

Hungary 1 1685 1564 1467 97 .141 .143 .141 .130

2 1498 1407 1332 75 .135 .140 .136 .134
3 1518 1409 1327 82 .157 .156 .163 .149

Iceland 2 579 525 474 51 .140 .209 .138 .137

Ireland 1 2046 1838 1679 159 .115 .115 .115 .115

2 2286 1139 1050 89 .133 .134 .134 .133
3 1800 1582 1453 129 .126 .126 .126 .126

Israel 1 2499 2167 1982 185 .140 .144 .140 .140

Italy 2 1529 1430 1366 64 .113 .112 .113 .112

 23

Table 2 (continued)

N after computing Stress 1 Stress 1

N N
ESS- Pearson Deleted Custom Stress 1 Stress 1 1000
Country raw after data
round correlations cases Schwartz Simplex Torgerson Random
data cleaning
listwise deletion 1992 starts

Latvia 3 1960 1825 1789 36 .149 .163 .157 .146

Luxembourg 2 1635 1549 1410 139 .122 .122 .122 .121

Netherlands 1 2364 2301 2210 91 .127 .131 .127 .126

2 1881 1824 1759 65 .129 .135 .128 .128
3 1889 1814 1772 42 .120 .127 .120 .119

Norway 1 2036 1806 1753 53 .112 .110 .112 .110

2 1760 1543 1488 55 .134 .138 .134 .130
3 1750 1533 1447 86 .110 .110 .110 .108

Poland 1 2110 1982 1826 156 .108 .110 .110 .108

2 1716 1621 1445 176 .110 .110 .108 .107
3 1721 1629 1478 151 .110 .111 .110 .109

Portugal 1 1511 1417 1327 90 .121 .124 .119 .118

2 2052 1987 1889 98 .116 .117 .118 .113
3 2222 2117 1937 180 .091 .090 .091 .090

Romania 3 2139 2003 1900 103 .137 .137 .138 .128

Russia 3 2437 2306 1903 403 .114 .101 .100 .099

Slovakia 2 1512 1420 1281 139 .111 .126 .110 .110

3 1766 1670 1567 103 .118 .131 .117 .111

Slovenia 1 1519 1390 1342 48 .114 .115 .114 .114

2 1442 1297 1241 56 .127 .124 .124 .123
3 1476 1329 1328 1 .121 .121 .121 .120

Spain 1 1729 1638 1585 53 .084 .084 .084 .083

2 1663 1544 1427 117 .095 .095 .095 .095
3 1876 1802 1735 67 .080 .080 .080 .079

Sweden 1 1999 1677 1608 69 .130 .130 .130 .129

2 1948 1663 1604 59 .119 .120 .119 .119
3 1927 1585 1534 51 .130 .129 .129 .128

Switzerland 1 2040 2009 1884 125 .118 .116 .117 .116

2 2141 2084 1902 182 .123 .122 .121 .121
3 1804 1758 1630 128 .126 .126 .126 .126

Turkey 2 1856 1591 1424 167 .135 .147 .134 .133

Ukraine 2 2031 1882 1446 436 .108 .102 .103 .094

3 2002 1877 1451 426 .124 .112 .111 .109

United 1 2052 1748 1645 103 .121 .123 .122 .121

Kingdom 2 1897 1806 1719 87 .139 .144 .138 .138
3 2394 2301 2188 113 .119 .119 .119 .119
 24

Table 3

Prototypical specification of value structure: Design matrix based on the revised Schwartz model
(1992, p. 45)

Value Value Sequence Dimension 1 Dimension 2 Angle

Number
UN 1 .34 .94 70
BE 2 .87 .50 30
TR 3 .98 –.17 350
CO 4 .49 –.09 350
SE 5 .64 –.77 310
PO 6 .00 –1.00 270
AC 7 –.64 –.77 230
HE 8 –.98 –.17 190
ST 9 –.87 .50 150
SD 0 –.34 .94 110

Notes: 1=Universalism(UN), 2=Benevolence(BE), 3=Tradition(TR), 4=Conformity(CO),

5=Security(SE), 6=Power(PO), 7=Achievement(AC), 8=Hedonism(HE),
9=Stimulation(ST), 0=Self–direction(SD)
 25

Table 4

Synopsis: Results of the Structural Analyses (ESS1–ESS3)

Country ESS- Stress1 Distinct Sequence of Values a,b Deviations

round Regions

Austria 1 .08 10 1,2,3,4,5,6,7,8,9,0

2 .09 10 1,2,3/4,5,6,7,8,9,0
3 .10 10 1,2,3,4/5,6,7,8,9,0 CO peripheral to SE

Belgium 1 .12 10 1,2,3/4,5,6,7,8,9,0

2 .11 10 1,2,3/4,5,6,7,8,9,0
3 .12 10 1,2,3/4,5,6,7,8,9,0

Bulgaria 3 .11 10 1/2,3/4,5,6,7,8,9,0 UN peripheral to BE

Cyprus 3 .12 10 1/2,3,4/5,6,7,9,8,0 UN peripheral to BE; CO peripheral to SE;

HE&ST reversed

Czech Republic 1 .11 8 [1+2],3/4,5,6,7,8/9,0 UN+BE mixed; HE peripheral to ST

2 .10 8 1/2,3/4,5,6,7,[8+9],0 UN peripheral to BE; HE+ST mixed

Denmark 1 .12 10 1,2,3,4,5,6,7,8,9,0

2 .12 10 1,2,3/4,5,6,7,8,9,0
3 .12 10 1,2,3/4,5,6,7,8,9,0

Estonia 2 .12 6 2,1,[3+4],5,6,7,[8+9],0 UN&BE reversed; TR+CO mixed;

HE+ST mixed
3 .12 6 [1+2],3,4,5,6,7,[8+9],0 UN+BE mixed; HE+ST mixed

Finland 1 .11 10 1,2,3/4,5,6,7,8,9,0

2 .11 8 1,2,[3+4],5,6,7,8,9,0 TR+CO mixed
3 .13 10 1,2,3,4,5,6,7,8,9,0

France 1 .12 10 1,2,3,4,5,6,7,8,9,0

2 .12 10 1,2,3/4,5,6,7,8,9,0
3 .14 10 1,2,3,4,5,6,7,8,9,0

Germany 1 .10 10 1,2,3/4,5,6,7,8,9,0

2 .11 10 1,2,3/4,5,6,7,8,9,0
3 .11 10 1,2,3/4,5,6,7,8,9,0

Greece 1 .11 8 1,2,3,4,5,6,7,[8+9],0 HE+ST mixed; HE_10 between PO&AC

2 .12 10 1,2,3/4,5,6,7,8,9,0

Hungary 1 .14 10 1/2,5,3/4,6,7,8,9,0 UN peripheral to BE; SE&TR/CO reversed

2 .14 8 1/2,5,[3+4],6,8,7,9,0 UN peripheral to BE; TR+CO mixed;
SE&TR+CO reversed; AC&HE reversed
3 .16 8 1/2,3/4,5,6,7/8,9,0 UN peripheral to BE; AC peripheral to HE;
CO_7 between SE&PO

Iceland 2 .14 10 1,2,3,4,5,6,7,8,9,0

Ireland 1 .12 10 1,2,3/4,5,6,7,8,9,0

2 .13 10 1,2,3/4,5,6,7,8,9,0
3 .13 10 1,2,3/4,5,6,7,8,9,0

Israel 1 .14 10 1,2,3/4,5,6,7,8,9,0

Italy 2 .11 10 1/2,3,4,5,6,7,8,9,0 UN peripheral to BE

 26

Table 4 (continued)

Country ESS- Stress1 Distinct Sequence of Values a,b Deviations

round Regions

Latvia 3 .15 8 1/2,3/4,5,6,7,[8+9],0 UN peripheral to BE; HE+ST mixed

Luxembourg 2 .12 10 1,2,3/4,5,6,7,8,9,0

Netherlands 1 .13 10 1,2,3/4,5,6,7,8,9,0

2 .13 10 1,2,3/4,5,6,7,8,9,0
3 .12 10 1,2,3/4,5,6,7,8,9,0

Norway 1 .11 10 1,2,3/4,5,6,7,8,9,0

2 .13 8 1,2,[3+4],5,6,7,8,9,0 TR+CO mixed
3 .11 10 1,2,3,4,5,6,7,8,9,0

Poland 1 .11 10 1/2,3/4,5,6,7,8,9,0 UN peripheral to BE

2 .11 10 1/2,3/4,5,6,7,8,9,0 UN peripheral to BE
3 .11 10 1/2,3/4,5,6,7,8,9,0 UN peripheral to BE

Portugal 1 .12 10 1,2,5,3/4,6,7,8/9,0 SE&TR/CO reversed; HE peripheral to ST

2 .12 10 1,2,5,3,4,6,7,9/8,0 SE&TR, CO reversed; ST peripheral to HE
3 .09 8 1,2,5,3/4,6,7,[8+9],0 SE and TR/CO reversed; HE+ST mixed

Romania 3 .14 10 1/2,3/4/5,6,7,8,9,0 UN peripheral to BE; TR/CO peripheral to

SE

Russia 3 .11 10 1,2,3/4,5,6,7,8,9,0

Slovakia 2 .11 10 1,2,3/4,5,6,7,8,9,0

3 .12 8 [1+2],3/4,5,6,7,8,9,0 UN+BE mixed

Slovenia 1 .11 10 1,2,3/4,5,6,7,8,9,0

2 .13 6 1,2,[3+4],5,6,7,[8+9],0 TR+CO mixed; HE+ST mixed
3 .12 10 1,2,3/4,5,6,7,8,9,0

Spain 1 .08 10 1,2,3,4,5,6,7,8,9,0

2 .10 10 1,2,3/4,5,6,7,8,9,0
3 .08 10 1,2,3/4,5,6,7,8,9,0

Sweden 1 .13 8 1,2,3,[4+5],6,7,8,9,0 CO+SE mixed

2 .12 8 1,2,3,[4+5],6,7,8,9,0 CO+SE mixed
3 .13 10 1,2,3/4,5,6,7,8,9,0

Switzerland 1 .12 10 1,2,3,4,5,6,7,8,9,0

2 .12 10 1,2,3/4,5,6,7,8,9,0
3 .13 10 1,2,3/4,5,6,7,8,9,0

Turkey 2 .14 10 1/2,3,5,4,6,7,8,9,0 UN peripheral to BE; CO&SE reversed

Ukraine 2 .11 10 1/2,3,4,5,6,7,8,9,0 UN peripheral to BE

3 .12 10 1/2,3,4,5,6,7,8,9,0 UN peripheral to BE

United 1 .12 10 1,2,3/4,5,6,7,8,9,0

Kingdom 2 .14 10 1,2,3/4,5,6,7,8,9,0
3 .12 10 1,2,3/4,5,6,7,8,9,0
 27

Table 4 (continued)

Notes: 1=Universalism(UN), 2=Benevolence(BE), 3=Tradition(TR), 4=Conformity(CO),

5=Security(SE), 6=Power(PO), 7=Achievement(AC), 8=Hedonism(HE), 9=Stimulation(ST),
0=Self-direction(SD)
a x/y: x = peripheral position, y = central position
b[x+y]: x and y mixed
 28

Growth

Protection

Figure 1. Structural Relations among Basic Values (cf. Schwartz, 1992, 2006).
 29

Figure 2. Location of Values according to the Design Matrix: 2–dimensional MDS.

Notes: 1=Universalism(UN), 2=Benevolence(BE), 3=Tradition(TR), 4=Conformity(CO),

5=Security(SE), 6=Power(PO), 7=Achievement(AC), 8=Hedonism(HE),
9=Stimulation(ST), 0=Self–direction(SD)
 30

Figure 3a. Round 1 (N=2,685); Stress 1=.10 Figure 3b. Round 2 (N=2,640); Stress 1=.11
 31

Figure 3c. Round 3 (N=2,706); Stress 1=.11

 32

0.75

0 1
2 1
0.25 9
8 0 2
DIM_2

1 3
9 8
7 4

7 4
-0.25 6
5
6
5 3

-0.75
-1.1 -0.6 0.0 0.6 1.1
DIM_1
Figure 4a. Theory Based Custom
Starting Configuration, Stress 1=.10

0.75

3
1

0.25 8 9
2
0
9 4
DIM_2

8 4
6 7 1
7 2 3
0
-0.25 6 5
5
1

-0.75
-1.1 -0.6 0.0 0.6 1.1
DIM_1
Figure 4b. Simplex (default) Starting
Configuration; Stress 1=.11

Figure 4. Two-dimensional ordinal MDS (PROXSCAL); Czech values data

of ESS round 2 (N=2.118)
 33

Address for Reprints

Name: Wolfgang Bilsky

Address: E-mail: [email protected]

Author #1

Name: Wolfgang Bilsky

Address: Psychologisches Institut IV

Westfälische Wilhelms-Universität

Fliednerstr. 21

48149 Münster, Germany

Affiliation (University, Institute, etc.):

WWU Münster, FB07, Psychologisches Institut IV

Telephone: +49 251 8334121

Fax: +49 251 8331343

E-mail: [email protected]

Biographical Sketch:

Wolfgang Bilsky is Professor of Personality and Differential Psychology at the Westfälische

Wilhelms-Universität in Münster, Germany. He received his Dr.rer.nat. from the TU
Braunschweig and his Dr.habil. from the University of Freiburg. His research interests include
differential psychology, psychology and law, and cross-cultural studies. He has published on
helping behavior, conflict management and crisis negotiation, facet theory, fear of crime and
victimization, lay theories, and on values and motives.
 34

Author #2

Name: Michael Janik

Address: Psychologisches Institut IV

Westfälische Wilhelms-Universität

Fliednerstr. 21

48149 Münster, Germany

Affiliation (University, Institute, etc.):

WWU Münster, FB 07, Psychologisches Institut IV

Telephone: +49 251 8334198

Fax: +49 251 8331343

E-mail: [email protected]

Biographical Sketch:

Michael Janik is a graduate student and a tutor at the Westfälische Wilhelms-Universität in

Münster, Germany. He has assisted the first author in several studies on values and recently
finished his diploma thesis on testing value structures by means of weak confirmatory MDS.
 35

Author #3

Name: Shalom H. Schwartz

Address: Department of Psychology

The Hebrew University of Jerusalem

Jerusalem 91905, Israel

Affiliation (University, Institute, etc.):

The Hebrew University of Jerusalem

Telephone: +972 25817892

Fax: +972 25817892

E-mail: [email protected]

Biographical Sketch:

Shalom H. Schwartz is the Leon and Clara Sznajderman Emeritus Professor of Psychology at
the Hebrew University in Jerusalem. He received his Ph.D. in social psychology from the
University of Michigan-Ann Arbor in 1967 and taught at the University of Wisconsin before
moving to Israel. He coordinates an international project applying his individual and culture
level values theories and methods in over 75 countries. His recent work concerns two broad
topics: the nature and sources of basic human values and their role as bases of attitudes and
behavior; the nature and sources of cultural value orientations as expressions of and
influences on the institutional structures, policies, and prevailing norms and practices in
different societies.