School Values Across Three Cultures

A new typology of school-level values is reported in three cultural contexts. School values were assessed by aggregating the scores of 862 students, (ages 15-19) in 32 Jewish and Arab Israeli schools (Study 1), and 1,541 students (ages 11-21) from 8 European schools and 163 teachers from 6 of these schools (Study 2), using Schwartz’s Portrait Values Questionnaire. Six school values emerged in both studies: achievement, autonomy, egalitarianism, harmony, compliance, and dominance. The importance of studying school-level values was demonstrated by relating the values of compliance and dominance to violence, and harmony values to student support measures (Study 1). Strong (minimal r = .64) school-level correlations between students of different ages and teachers supported the validity of the findings (Study 2).

Schools play an important role in children's emotional, physical, and educational development (e.g., Kasen, Berenson, Cohen, & Johnson, 2004;Khoury-Kassabri, Benbenishty, & Astor, 2005;Roeser, Eccles, & Sameroff, 2000). Alongside the family, schools can be considered a primary socialization agent in the development of children. One of their main goals is socialization to specific values, considered by the society to be desirable (Halstead, 1996). It is therefore crucial to gain a better understanding of the values prevalent in schools. Surprisingly, however, no coherent typology exists that outlines the cultural dimensions of schools and allows researchers to study schools by their values. In this article, we address this gap by providing a typology of school values, testing it in three cultural contexts, and looking for meaningful relationships of school values with several additional variables.
Values are abstract trans-situational goals, varying in importance, that pertain to desirable end-states or behaviors, and guide selection or evaluation of behaviors and events (Rokeach, 1973;Schwartz, 1992). Group values are implicitly or explicitly shared among group members. For example, national and organizational cultures are distinguished by their values, the ideas they hold about the right, wrong, and the desirable in the group context (Hofstede, 1991;Schwartz, 1999Schwartz, , 2008. Cultural values are expressed in widely shared norms, symbols, rituals, practices, and ways of thinking. They are the vocabulary of socially approved goals used to motivate action and to express and justify the preferred solutions to societal problems (Schwartz, 1999(Schwartz, , 2008.
We suggest that schools can be characterized by the values prevalent among their members, students, and teachers. Moreover, schools differ in the values they stress. The differential importance placed on values may be manifest through school rules, teaching practices, relations among teachers and students, and more. Describing a school by its values can allow researchers to predict behaviors at schools, and provide focused tools for intervention in case of school malfunctioning.
In Study 1, we propose several potential value dimensions that are likely to differentiate between schools. We test for the existence and replication of these dimensions using school-level value scores derived from the aggregated responses of students in 32 Jewish and Arab schools from Israel. We also relate these dimensions to several schoollevel variables, such as school violence and ethnic school system.
Group-level values are shared to some extent by most members of the group (Schwartz, 1999). Although, for example, there are likely to be average differences between teachers and students, all school members are expected to share their values to a large extent, because they belong to the same social environment. In Study 2, we replicate the value structure, and test the hypothesis that the values of teachers, older students, and younger students across schools are positively correlated in a third, Central European, cultural context.

Theoretical Background: School Values As a Private Case of Group Values
Past research on school characteristics described schools using a number of dimensions in the purpose of predicting school outcomes and especially achievement. Among the dimensions researched, one can find past abilities as well as demographic characteristics, such as mean socioeconomic status (SES) of the students' families and level of education of parents (e.g., Opdenakker & Van Damme, 2001;Raudenbush & Bryk, 1986). Other characteristics measured are school climate, such as the social-emotional climate at school (e.g., whether the school is characterized by enthusiasm or suspicion; Emmons, Efimba, & Hagopian, 1998;Lehr & Christenson, 2002;Lynch, Lerner, & Leventhal, 2012;MacNeil, Prater, & Busch, 2009), communitarian practices and beliefs (Phillips, 1997), cooperation among staff, and the importance assigned to orderly atmosphere (Opdenakker & Van Damme, 2001;Raudenbush & Bryk, 1986).
In the current research, we propose that a characterization of schools based on the values they endorse will be useful for differentiating between schools, and predicting norms and behaviors in them. Research on school values can be undertaken using a number of techniques, each holding advantages and disadvantages. Values can be extracted from artifacts of the schools, values of school principals and teachers, and descriptions of the school values by the students (e.g., Feather, 1972;Meglino & Ravlin, 1998). We take a different approach, relying on students or teachers as reporters of their own values, and using the aggregated responses as indicators of the priority given to each value at the school level (meaningful within-school variations in values notwithstanding). Our approach is based on the notion that school culture comprised the opinions, attitudes, and values of those inhabiting the schools. The aggregation of individuals' values keeps in check the individual variability, revealing the values common in the school.
Past studies found the aggregation of individual characteristics a reliable, valid, and useful technique to measure the cultural values of nations (Schwartz, 1999) as well as organizations (Hofstede, Bond, & Luk, 1993). The values of the individuals in organizations are a product of their shared culture as well as the unique personal experience. The variation between individuals can express the uniqueness, while the average priorities reflect the central thrust of the shared enculturation. The variation between the aggregated scores is largely independent of the variation at the individual level leading to independent analyses and results (Hofstede et al., 1993;Liska, 1990;Schwartz, 1999).
Although little theoretical work has been done regarding the meaning of school values, there is extensive work about values in other groups, particularly national cultures. We rely on this work in defining school values. Similarly to national values (Schwartz, 1999(Schwartz, , 2008, school values are desirable abstract goals that are shared to some extent by members of the school community, such as students, teachers, and other staff. Group values prescribe norms for individual members, thus directing their behavior (Fischer, 2006;Schwartz, 1999). In schools in which prosocial values are considered highly important, for example, school members are likely to be encouraged to engage in mutual aid, and be responsible for the welfare of others, through behaviors such as help in doing school assignments.
Not all school members are likely to have the same impact on school values. For example, school principals may be more influential than the typical student or the typical teacher. However, over time, the school community comes to share values to a certain degree. This process occurs through selection of like-minded members and socialization to value preferences (De Cooman et al., 2009). The value preferences of the school are conveyed to the members through everyday experiences, such as the content and perceived difficulty level of the learning tasks, the dominant patterns of interaction among students and teachers, and the evaluation procedures used (Boekaerts, de Koning, & Vedder, 2006). Thus, school regulations and practices should reflect in part the values of the schools. These regulations serve the purpose of socializing new members to hold these values (e.g., Chatman, 1991).
The existence of shared school values does not preclude individual differences within schools. For example, Ferrari, Kapoor, and Cowman (2005) provided evidence that institutional-level values of universities do not fully overlap with students' personal value systems (see also Astill, Feather, & Keeves, 2002;Owens, 2005). Thus, although normative pressures are expected to increase adherence to school values with time, there is still room for individual differences in the extent these values are accepted by school members due to their individual personalities or backgrounds. These withinschool differences in value priorities are reflected in withinschool individual differences in behavior, school achievement, and well-being (Feather, 1992;Owens, 2005).

Content of School Values
In proposing a typology of school values, one may consider the values of individuals (Schwartz, 1992) and the values of national cultures (Schwartz, 2008) as possible reference points. We see cultural values as a more relevant value typology, as schools, like nations, are an assembly of individuals, linked by common goals, practices, and norms. School values, like national values, can be seen as ideas about what is right and wrong, desirable and undesirable, which are implicitly or explicitly shared among members of the group.
The cultural dimensions of values reflect the basic issues or problems that societies confront to regulate their members' actions. Schools face somewhat similar issues or problems to the ones confronted by nations, such as regulating the nature of relations between the individual and the group. We therefore choose to base the typology of school values on the values of national cultures and not on the values of individuals. Nevertheless, schools also face issues and problems relevant to their unique social role of teaching and socializing children and adolescents, and we should therefore also consider school-specific issues that may lead to value priorities. Such issues are the questions of evaluating students and emphasis on cognitive functioning. Hereinafter, we will describe the school value dimensions hypothesized to be similar to nations' value dimensions and the value dimensions expected to be unique to schools.

Embeddedness Versus Autonomy
An important issue confronting societies is the nature of the relations between the individual and the group. Particularly, social entities differ markedly in the extent to which individuals are considered autonomous or embedded in their groups (Hofstede, 1980;Markus & Kitayama, 1991;Schwartz, 1999Schwartz, , 2008. Cultures differ in the importance given to embeddedness (emphasizing group identification and conforming to group norms and rules and the traditional order) versus autonomy (viewing each person as an autonomous, bounded entity encouraged to pursue and express his or her own uniqueness; Schwartz, 1999Schwartz, , 2008. These values are also relevant to organizations, which differ in whether they put an emphasis on interdependence, rules, and codes, on one hand, or on independence and change, on the other hand (Sagiv & Schwartz, 2000;Victor & Cullen, 1988). Applied to schools, valuing rules at school has been found relevant to variables such as parental school involvement (Rosenblatt & Peled, 2002). Autonomy has also been described as an important ethical principle distinguishing between schools (Schulte et al., 2002). For example, some schools may vary in their views about students' personal responsibility for their academic success or failure. In sum, we expected schools to differ in the importance given to embeddedness and autonomy values.

Hierarchy Versus Egalitarianism
This second issue is especially relevant to schools because of their hierarchical structures. Groups must find a way to guarantee responsible behavior that preserves the social fabric, including cooperation and mutual consideration. In their solution to this problem, societies are differentially positioned on the spectrum of values going from hierarchy (legitimizing an unequal distribution of power, roles, and resources), on one end, and egalitarianism (an emphasis on equality pursued through unselfish pursuit of others' welfare; Schwartz, 1999), on the other. Hierarchical organizations emphasize the chain of authority, in contrast to egalitarian organizations, which assign flexible roles and enable negotiation as means of motivating members (Sagiv & Schwartz, 2000). To ensure the smooth activity of the school, decisions have to be made on whether to establish a hierarchical nature of school organization, with teachers as more powerful and students as compliant (and perhaps also with power differences between different students, such as older and younger students), or whether students are considered equal to teachers when important decisions are made. Indeed, Pang (1996) found school differences in how much each school has a formalized, centralized, and well-established system of superordinate-subordinate relationships.
Hierarchy values at schools were expected to focus on humility and compliance, on one hand, and on dominance and social power, on the other hand, because in a hierarchical society it is important for members to acknowledge their inferiority to some as well as their superiority to others (Schwartz, 1999). For example, in hierarchical schools, older students can give high importance to their higher status as compared with younger students, but accept the higher status ascribed to teachers.
Previous studies have also indicated that schools vary in two aspects of egalitarianism as described by Schwartz (1999Schwartz ( , 2008. Differences in the importance ascribed to the participation of teachers (Pang, 1996) and students (Harber & Trafford, 1999;Vieno, Perkins, Smith, & Santinello, 2005) reflect the egalitarian aspect of this value. The caring aspect of egalitarianism is reflected in past studies indicating school differences in the importance of collegiality and students' mutual support (Doyle Lynch et al., 2012;Higgins-D'Alessandro & Sadh, 1997;Pang, 1996;Rosenblatt & Peled, 2002). In sum, we expected a dimension of school values contrasting egalitarianism (and care) and hierarchy at schools.

Mastery Versus Harmony
A third basic issue that cultures encounter concerns our relation their relations to the natural and social world. According to Schwartz (1999), harmony values describe acceptance of the world as it is, trying to fit in rather than change or exploit it. Harmony values emphasize fitting harmoniously into the environment (Schwartz, 1999). In organizations, this issue is reflected in the level of competitiveness versus cooperation (Sagiv & Schwartz, 2000). Harmony values may be important to school functioning through the regulation of relationships inside schools, and between the schools and their communities. For example, in high-harmony schools, decisions may be made through agreements rather than through a voting procedure or simply principal-initiated regulations.
The opposite solution to the issue, mastery values, emphasizes getting ahead through active self-assertion, by responding to challenges, mastering, and changing the world (note that despite the similar label, this concept is very different than that of mastery goals, for example, Elliot & Church, 1997). Mastery values have two major aspects. One aspect describes an emphasis on facing challenge and adventure. A value representative of this aspect will be the value "daring." In the school context, this aspect may be of secondary importance. A second aspect of mastery values, the aspect of striving ahead and achieving, may be of supreme importance to schools. Values representing this aspect may focus on ambition and success. Due to this differential relevance, we decided to emphasize achievement as the defining feature of mastery values in the school context.

Achievement
The role of individual students' achievement values has been investigated substantially (Wigfield & Eccles, 1992. Students' achievement goals were demonstrated to be important for predicting success (Harackiewicz, Baron, Tauer, Carter, & Elliot, 2000). Moving on to the school level, all modern schools share the task of helping students reach a certain standard in their studies. However, schools differ in the relative importance given to achievement of this standard. This emphasis is expressed by multiple characteristics, such as reinforcement and incentives for high academic standards (Pang, 1996), as well as physical plant, and teaching methodology (Roeser, Urdan, & Stephens, 2009). Not only do schools differ in the level of achievement value but they also frame achievement differently. Schools can emphasize learning, thus supporting effort and self-improvement. They can also emphasize performance, thus encouraging demonstration of ability, comparison among students, and striving for high grades as a main goal of learning (Anderman & Midgley, 1997;Wang & Holcombe, 2010).
School culture of achievement has important contributions to student-level characteristics. School achievement culture was found related to the motivation of students, their identification with the school, and their engagement in the school; to students' self-esteem; to self-regulation strategies; and finally, to the achievement of students (Church, Elliot, & Gable, 2001;Wang & Holcombe, 2010). Being such an important characteristic shaping the everyday lives of students at schools, we expected to find achievement values to be a defining feature of mastery values, differentiating among schools in a substantive manner in our study.

The Structure of Value Types
The value types suggested above are hypothesized to form dimensions that express the conflicts between alternative resolutions to issues faced by schools. Because of these contradictions, we expect an emphasis on a value to go along with a de-emphasis of the opposing type. At the same time, compatibilities are possible between values from different dimensions. That is, certain value emphases are likely to be accompanied by other value emphases, to the extent they share similar assumptions and motivations. These relations of conflicts and compatibilities are summarized in the circular structure formed by the school value system. For example, egalitarianism values and autonomy values share the assumption that people can and should be held responsible for their actions and decisions. Embeddedness and hierarchy values share the assumption that each individual has group obligations and roles that may override personal inclinations (Schwartz, 2008). We expected school values to follow the pattern of conflicts and compatibilities described at the cultural level.
In conclusion, we hypothesized that schools differ across three value dimensions relevant to groups in general: hierarchy versus egalitarianism, embeddedness versus autonomy, and harmony versus mastery. We hypothesized that mastery values will take the form of achievement values. Moreover, we expected the structure of the values at the school level to resemble the structure of group values. Last, we hypothesized that the school values found will be able to predict school-level characteristics. We addressed these issues in Study 1.

Study 1
Study 1 investigated the structure of values at the school level in Israeli Jewish and Arab high schools. In Israel, Jewish and Arab students study in different school systems, both under the supervision of the Ministry of Education. The structure of the curriculum in both school system is nearly identical. It differs mainly in the cultural and religious subjects, and in the language of instruction (Benavot & Resh, 2003). The students in the Jewish schools are members of the Jewish majority in Israel. This group comprises 79% of the Israeli population. The students in the Arab schools are Arab citizens of Israel. Their families are likely to have lived in the area that is now Israel since before its foundation. They are approximately 18% of the Israeli population, and belong to two main religions, Islam and Christianity. They live mostly in homogeneous Arab villages or segregated neighborhoods in mixed cities (Rabinowitz, 2001). Israeli Jews were previously found to report higher importance assigned to values of autonomy and mastery than Israeli Arabs, while Israeli Arabs reported higher importance assigned to values of hierarchy and embeddedness than Israeli Jews (Schwartz, 2008).
Value structure was analyzed using multidimensional scaling (MDS) as recommended by Schwartz (1999;Schwartz & Sagiv, 1995). Schwartz (1992) considered value types as fuzzy sets arrayed on a continuum of related emphases. Value items are grouped into value domains based on theoretical considerations and the empirical location of items, where items emerging adjacent in the value structure are considered to share similar meanings and to point to a shared emphasis (Schwartz, 1992). Replication of the structure across sample types and cultures provides the best evidence for the validity of a value theory (e.g., Schwartz & Sagiv, 1995). We measured school values by aggregating responses across individuals attending the same school, as explained earlier.
Our first question was whether meaningful value dimensions could be found that may distinguish between schools. We addressed this issue with data from a substantial number of schools (N = 32). The second question addressed the issue of replicability of the value dimensions found. We compared the school-level value structure in two ethnic cultures by comparing Arab and Jewish schools in Israel. Next, we addressed the relevance of the value dimensions by asking whether the value dimensions were meaningful for comparing schools of different ethnic backgrounds in their value priorities, and for studying the relationships between schoollevel values and several social characteristics of schools (violence levels and school treatment of students). Meaningful relationships would exemplify the usefulness of the value dimensions in understanding between-school differences in other variables.

Method
Participants. High school students (N = 907) attending 33 different schools participated in the study. Of the participants, 3% could not be assigned to a school or attended a school with less than 15 participants in the sample, and were therefore not included in the analysis. The final sample included 15 Jewish and 17 Arab schools with 862 students or 28 students on average in each school. Adolescents' ages ranged from 15 to 19 with 95% between 16 and 18 (M Jews = 16.72, SD = 0.80; M Arabs = 16.91, SD = 0.70). About half (49%) of the respondents were female (49% among Jewish adolescents, 50% among Arab adolescents). Arab adolescents were 51% of the sample (20% of the age-relevant population in the country; Israel Central Bureau of Statistics, 2005) to ensure an adequate sample size for cross-cultural comparisons.
Procedure. Schools were randomly sampled in each sector across different levels of SES, as classified by the Israeli Ministry of Education. An attempt was made to reach about equal numbers of schools from the lower, middle, and upper range of socioeconomic development. This resulted in 34% of the schools being from the top three deciles of socioeconomic development relative to schools of the same ethnic group, 36% from the middle three deciles, and 27% from the lower four deciles. This distribution did not differ significantly between Arab and Jewish schools, χ 2 (2) = 1.14, ns, meaning that schools similarly represented socioeconomic variation within each ethnic group. Nevertheless, it is important to note there were between-group differences in SES. For example, the mean parental education level in Jewish schools was 13.13 years (SD = 1.44) as compared with Arab schools (M = 11.55, SD = 1.27), t(30) = 3.30, p < .01. Only schools in which at least 15 students participated were retained for school-level analysis.
Adolescents were recruited by telephone during 2005. Each ethnic group was reached in a procedure that was more likely to yield a relatively high participation rate within that group. Adolescents attending Jewish schools were approached through phone numbers taken from student directories for high schools. Classes were randomly sampled from within the school, and all students who were listed for the specific class were approached, and requested to participate. Adolescents attending Arab schools, for which no equivalent phonebooks existed, were approached with the use of a snowball technique.
A university student of the relevant ethnic group administered the questionnaire in the adolescents' homes to the adolescents who agreed to participate (46%) and received parental permission to do so. A small (7%) proportion of the sample (n = 62 Jewish adolescents) answered the questionnaire on a secure website. No significant differences on any of the study variables were found between adolescents who answered a paper-pencil questionnaire and those who responded on the Internet. Adolescents were assured that their responses would be anonymous and confidential. They were instructed to answer the questionnaires alone, without parental intervention or the presence of peers, and were offered to hand the questionnaires back in sealed envelopes.
Materials. The questionnaires contained several parts, regarding adolescents' values, perception of school characteristics, violent victimization, and violent behavior. Each respondent completed the questionnaire in his or her native language (Hebrew or Arabic). Back-translation procedures ensured the comparability of the Arabic and Hebrew questionnaires.
Values. Values were assessed with the Portrait Values Questionnaire (PVQ; Schwartz et al., 2001;Schwartz, Lehmann, & Roccas, 1999). We used a version tested by Schwartz and Rubel (2005) in a multicultural study. The PVQ has been shown to be suitable for use with adolescents (Bubeck & Bilsky, 2004;Knafo & Schwartz, 2003;Knafo & Spinath, 2011;Schwartz et al., 2001). It includes short verbal portraits of 40 people (matched to the respondent's gender). Each portrait describes the person's goals, aspirations, or wishes, pointing implicitly to the importance of a single broad value. Two exemplary portraits are as follows: "It is important to her to be rich. She wants to have a lot of money and expensive things" and "It is important to him always to behave properly. He wants to avoid doing anything people would say is wrong." Respondents indicated "How much like you is this person?" for each portrait, using a 6-point scale ranging from "very much like me" to "not like me at all." Thus, respondents' own values were inferred from their self-reported similarity to people who are described in terms of particular values.
As a preliminary analysis, we ensured that Jewish and Arab adolescents understood the value items in a similar way. We conducted smallest space analyses (SSA; Guttman, 1968) in which the values of Jewish and Arab adolescents yielded structures very similar to the prototypical, circular structure of values described by Schwartz (1992). This analysis indicated that Jewish and Arab adolescents construed their values at the individual level very similarly. Three items were located in different areas of the structure for both groups (country safety, stable government, fit into nature). These items were omitted from the school-level analyses to preserve the cross-cultural equivalence of value meaning.
To control for scale use differences across cultures that are independent of value structures (Fischer, 2004;Schwartz, 1992), we standardized each value item within individual, so that each individual student's responses reflected the importance he or she gave to that item relatively to all value items. We then averaged the standardized scores of all students available from each school to form an aggregated schoollevel item score. We performed all further analyses on these transformed data.
School violence. This measure was adapted from Benbenishty, Zeira, and Astor (2000) by selecting eight items (from a larger pool) that measured direct violent behavior, in different levels of severity, and by asking adolescents not only to report on violent behaviors they were victims of, as in the original version, but also on their own behavior and the behavior of their best two friends. Every student was asked about the frequency of his or her own violent behavior (e.g., picked up a stone or another object to hurt another student; took things from another student by use of force), about the frequency in which he or she was a victim of another student's violent behavior, and finally about the frequency of his or her two best friends' violent behavior over the last 12 months. Answers were given on a 3-point scale (0 = never, 1 = once or twice, 2 = 3 times or more). A total score was calculated by summing responses across all items of each scale, with higher scores indicating more frequent violence. We then aggregated the scores on each of the three variables across all respondents from each school, creating three scores (perpetration, victimization, and friends' behavior) for the violence level in that school. The violence variables, like the school values, were therefore created as characteristics of schools, and not of individuals. They describe the level of violence prevalent in this school and encountered by the students studying in it, whether they take part in it themselves or not.
School climate. We assessed school climate with a scale of nine items taken from scales by Benbenishty et al. (2000). In the present study, the scale consisted of two subscales. One subscale focused on students' participation in school decision-making processes (e.g., "Our school staff makes an effort to enable students' participation in making important decisions"). The other one assessed teachers' support (e.g., "When I have a problem, I feel comfortable discussing it with my teachers"). Respondents indicated how much they agreed with each of eight statements on a scale, ranging from 1 = disagree to 5 = agree. Items were then aggregated within schools. Again, school climate was a characteristic of the school as all respondents reported of it, and not of the subjective perspective of any individual student.

Results
Identifying School-Level Value Dimensions. We ran a series of consecutive MDS analyses on the school-level value scores. We used the alternating least-squares algorithm (ALSCAL) procedure in SPSS 20, adding no initial configuration. ALS-CAL differs from other MDS procedures in minimizing S-Stress rather than Stress, thereby fitting squared distances to squared similarities (Borg & Groenen, 2010). Because of the large number of items and the relatively small number of schools, we were interested in reducing the number of items to include only those that could robustly and meaningfully address variation across schools.
Our first criterion was cross-cultural similarity in meaning. We dropped PVQ Items 14, 35, 40, 1 which were not shown to have a cross-culturally similar meaning at the individual level in preliminary analyses (Knafo, Daniel, & Khoury-Kassabri, 2008). We also dropped Item 20 that pertains to the importance of religion, because the educational system in Israel is highly segregated based on religiosity level (e.g., there are Jewish religious and Jewish nonreligious schools), and the meaning of this item may be different across these school systems.
Our second criterion was face validity and content meaningfulness of the groups of items. We attempted to find meaningful regions in the school-level MDS analysis. A number of items were positioned in the proximity of items that were not conceptually related to them. For example, the three individual-level Schwartz (1992) hedonism values were dispersed in the school-level value structure and did not relate in an internally valid way to the items in their vicinity. Possibly, hedonism has little meaning as a school-level value. Six more items from diverse Schwartz (1992) value types (4, 15, 25, 28, 31, 36) did not show theoretically meaningful relations to the items in the regions they appeared in. Such items were therefore dropped from further analysis. We ended up with 27 items that had cross-cultural meaning equivalence and formed meaningful clusters of at least 3 items each, with Stress 1 value of .03, which represents an excellent representation (Kruskal & Wish, 1978). Figure 1 presents the final MDS structure.
The first dimension we described in our theoretical background referred to embeddedness versus autonomy. A clear autonomy region emerged (Figure 1, middle right) with six items (α = .70), emphasizing openness to intellectual experiences and behavioral choice. Potential embeddedness items such as secure surroundings, cleanliness, and behaving properly appeared at the opposite end of the figure, but did not form a clear embeddedness region.
As expected, a region emerged that included both aspects of hierarchy values, namely, humility and compliance, on one hand, and dominance and social power, on the other hand ( Figure 1, top left). However, in contrast to the expectations, a clear distinction also emerged between these two aspects, in this study as well as in Study 2. The values of compliance formed a distinct group, relating to issues of obedience and acceptance of one's place in the social world. The values of dominance formed another group, relating to issues of authority and command. We therefore retained two separate values. Dominance values (Figure 1, top left) included three items (α = .68), emphasizing being at the top of the social hierarchy: being rich, being in charge, and leading. Compliance values (Figure 1, middle left) included three items (α = .50), emphasizing accepting one's position when at the lower end of the social hierarchy (being humble and modest, behaving properly, doing as told, and following rules).
We also expected to find a region of values emphasizing the importance of egalitarianism. This region was found (Figure 1, top right). This value included seven items (α = .80), emphasizing broadmindedness and equality, protection of the weak, forgiveness, loyalty, and supporting and helping others.
A harmony region was also found (Figure 1, bottom left) with five value items (α = .77). As suggested by Schwartz (1999) regarding national cultures, harmony values emphasize both harmony with the surrounding physical environment (cleanliness, caring for nature) and social harmony (secure surroundings, being satisfied with less, and world peace).
Finally, as was theoretically postulated, a clear achievement region emerged (Figure 1, bottom right) with three items (α = .69), emphasizing ambition, getting ahead, and succeeding and impressing others. This value is very similar to the individual-level achievement value described by Schwartz (1992). Table 1 presents the definitions of the six values we identified.
The order of the values emerging from the analysis was as follows: dominance, compliance, harmony, achievement, autonomy, and egalitarianism. Thus, harmony values were located in a position different than hypothesized, alongside the opposing achievement values, resulting in shifts in the structure.
Cross-Culturally Replicating the School-Level Value Structure. We next asked whether the structure obtained would replicate in the two cultural groups. We averaged the items belonging to each value domain to scores representing the six school-level values and ran the analysis separately for Jewish and Arab schools. (We used average scores to reduce the number of items in an MDS based on less than 20 schools.) The ALS-CAL procedure yielded Stress 1 values smaller than .01 for Jews and Arabs. As Figure 2 shows, the structure was similar for the two groups, with the values similarly ordered (going from the bottom left, clockwise): dominance, compliance, harmony, egalitarianism, autonomy, and achievement. This cross-cultural replication supports the validity of our findings, and enables a mean comparison of the importance of the six school values in each group. Moreover, when separated across cultures, harmony values were located in the hypothesized position, between the values of compliance and egalitarianism, and in opposition to the values of achievement.
Cross-Cultural Differences in Value Importance. The value dimensions that characterize Jewish and Arab schools were found similar in structure. Thus, schools from both cultures can be distinguished based on the same value dimensions. We tried to examine the relevance of the six values derived in this study by demonstrating their ability to distinguish between schools coming from the separate ethnic cultures. If  Accepting one's position and obligations toward the social system and toward the upper end of the social hierarchy; being humble, modest, and obedient. Harmony Aspiring for harmony with the surrounding physical and social environments; avoiding conflict, danger, and disorder. Egalitarianism The importance of caring for others and respect for them; broadmindedness and equality, protection of the weak, loyalty, and helping. Autonomy Openness to intellectual experiences and the importance of behavioral choice. Achievement Motivation for striving ambitiously to get ahead, succeeding and impressing others.
ethnic culture has an effect on the values prevalent in the schools, then the schools originating from different cultures will support different values. We compared the mean importance given to values in Arab and Jewish schools. Large and significant differences were found in all values (Table 2). Even the smallest effect size (D = 0.77 for dominance) is considered large in Cohen's (1992) terms. In Jewish schools, there was higher importance to dominance, egalitarianism, and autonomy values, whereas in Arab schools higher importance was given to compliance, harmony, and achievement. Thus, school values were successful in differentiating between schools coming from different cultures.

Relating School-Level Values to Other School Characteristics.
The school-level value scores of each school in each value were correlated with school-level characteristics of these schools, including the climate in the school and the level of violence prevalent in it. Harmony values, that emphasize smooth interpersonal relations, correlated meaningfully with the ratings of students of the support they received from teachers, r = .50, p < .01, and with the ratings of students' participation in decision making, r = .43, p < .01, controlling for ethnic differences.
Next, we related school-level values to school violence. An unexpected finding emerged with achievement values correlating positively (controlling for ethnic differences) with school violence as indicated by adolescents' reports of perpetration, r = .52, p < .01; the violence of their two best friends, r = .52, p < .01; and reports of victimization, r = .40, p < .05.
We were interested in the potential predictive power of compliance and dominance. Although neither of these values correlated with school violence, we reasoned that their combination may predict violence. Specifically, a school in which compliance values are not important and at the same time dominance is very important is likely to have the highest degree of violence, because individuals in this context will strive to obtain power and control over others (high dominance) with little regard for rules, norms, and laws (low compliance). We ranked schools as either high or low in these two values, based on them being, respectively, above or below the median of schools of the same ethnic sector. We then ran a MANCOVA with ethnic school system as the covariate. Figure 3 presents the mean violence levels (using the three indices of violence) for schools high or low in compliance and dominance. The multivariate test showed that neither compliance, F(3, 25) = 0.53, ns, nor dominance, F(3, 25) = 1.25, ns, had a significant main effect on school violence. However, the two values interacted in predicting violence levels, F(3, 25) = 5.66, p < .01. Schools in which dominance values were high in importance but compliance values low stood out as highest in violence, regardless of whether it was indexed by adolescents' reports of perpetration, F(1, 27) = 4.74, p < .05, η 2 (variance accounted for) = .10; by reports about the violence of two best friends, F(1, 27) = 11.98, p < .01, η 2 = .25; or by reports of victimization, F(1, 27) = 4.76, p < .05, η 2 = .14.

Discussion
The study found meaningful clusters of values at the school level. These values were mostly similar to the nation-level  values identified in previous research (Schwartz, 1999(Schwartz, , 2008: dominance, compliance, harmony, egalitarianism, autonomy, and achievement. The values of embeddedness were not distinguished as a meaningful value at the school level. The order of the values at the culture level, but not at the overall level, resembled the hypothesized structure. The difference may stem from the high importance assigned by Arab adolescents, relative to Jewish adolescents, to both harmony and achievement values, resulting in high correlation between the values at the school level. The cross-cultural similarity in the value structure supports the validity of the results. The meaningful relationships with other variables such as teacher support and school violence suggest that these values may be important variables to be pursued in further research. However, the first study had two main weaknesses. First, we relied on data from a limited number of students within each school to characterize the whole school. Moreover, the sampling of the adolescents within each school was not random, particularly in the Arab sample. Although more than 800 students participated and every data point is based on at least 15 students, it would be preferable to see whether the results could be replicated with larger samples.
Second, although the Arab and Jewish school systems in Israel are different in many respects, they are still nested within the same context, and managed by the same Ministry of Education. A replication in a different cultural and educational context is therefore needed.
Most importantly, we have assumed that aggregated students' values could represent the values of the school. However, teachers influence school culture to a large extent, and they may have substantially different values. A study with data from students and teachers is needed to see whether the value priorities of the schools are shared by teachers and students, and to check for mean-level differences between these two school subpopulations.

Study 2
In Study 2, our first question was whether the value typology and structure found would be replicated within schools nested in a European context. Our second question dealt with whether Study 1 results were unique to student reports. We addressed it with data from students and teachers, using a large number of students in each school. A replication of the value structure provided by student and teacher data would provide support for the proposed value dimensions.
The third question dealt with a different kind of replication. Not only did we look at the structure of values for students and teachers but we also looked at whether the values of each specific school were similar, regardless of whether they were measured with student data or with teacher data. In other words, we expected the meaningful school value dimensions to be reflected in substantial positive correlations between the school-level values of students and teachers, reflecting the unique culture of the school.
Similarly, we wanted to compare the value priorities of younger and older students, and to see whether the same values are shared by school members of different ages. High positive correlations across schools between the importance given to each value by younger and older students would support the validity of our findings.

Method
Participants. The study was conducted with students (5th-13th grade, N = 1,541: 695 girls, 845 boys, and 1 unknown) from eight different high schools in central Europe (six German schools, one international school in Germany, and one international school in the Czech Republic). Students' age range was between 11 and 21 years (M = 15.61, SD = 1.99) with 93% between the ages of 12 and 18. The majority (72%) of the students were Germans, 28% were of other nationalities. In addition, teachers from six of these schools (n = 163: 82 females, 78 males, and 3 unknown) participated as well. Teachers' ages ranged from 21 to 71 years (M = 42.48, SD = 10.12) with 95% between the ages of 26 and 65.
Procedure. The eight schools were located by sending letters to approximately 35 high schools asking the principals whether they would be willing to allow their students to participate in the study. Depending on the requirements of each school, in some of the schools, we administered the questionnaires class by class; in others, they were handed out and filled in during school assembly time. Teachers' questionnaires were administered when permission was granted (six schools). Materials. As in Study 1, all participants filled in a questionnaire which, in addition to basic variables such as sex and age, measured their values using the PVQ (Schwartz et al., , 2001. An individual-level MDS showed that the value items measured the 10 Schwartz (1992) values, enabling a cross-cultural replication for the purposes of this study. The same standardization procedures as in Study 1 were used. We then averaged the scores of all students available from each school to form an aggregated school-level score. Similarly, in the six schools in which teacher data were available, we aggregated teachers' scores. We performed all analyses on these transformed and aggregated data.

Results
Replication of the School-Level Value Structure Using Student Data. We ran an MDS on the school-level scores, using the ALSCAL procedure, with students' data across eight schools, and the same 27 items identified in Study 1. Figure 4 presents the structure obtained in this MDS. The structure from the European schools replicated almost perfectly the Israeli structure, and the stress was .03 that represents an excellent representation (Kruskal & Wish, 1978). For five of the six values, all items clustered together as in Study 1 (Figure 1). One minor exception was the location of harmony values, which were located in front of compliance values, instead of beside it, as in the separate cultures in Israel. In spite of this relatively small deviation, and considering the small number of schools (n = 8), Study 2 provides a very good replication of Study 1's results. This assertion is based on the standards provided by Schwartz and Sagiv (1995) for cross-cultural comparison of value structures.
Replicating the Value Structure with Teacher Data. We next asked whether the structure obtained would replicate using teacher and student samples. We averaged the items belonging to each value domain to scores representing the six values and ran an ALSCAL analysis, as described before, separately for students across eight schools and for teachers in six schools (Stress 1 = .03 for students and .05 for teachers). Plotting the resulting structure in Figure 5, the structure was very similar for the two groups, with the values similarly ordered (going from the middle left, clockwise): dominance, compliance, harmony, egalitarianism, autonomy, and achievement. The similar structure for teachers and students provides further support to the school value classification we propose. In addition, the structure is very similar to the one found in the analysis of the two separate Israeli samples, which provides a second cross-cultural replication.

Consistency of School Values across Student Age Groups and between Students and Teachers.
The large student samples obtained within each school enabled us to test the issue of replicability of the findings with different age groups attending the same school. If school values, as argued, are common to some extent to most school members (allowing for individual differences), then the value priorities of teachers as well as younger (ages 11-15) and older (ages 16-20) students within the same school should correlate positively and significantly. In other words, when aggregated school-level scores taken from teachers and students of different ages correlate positively, this provides evidence for meaningful cross-schools variability that is shared by school members, be they older students, younger students, or teachers. We tested this hypothesis using cross-schools correlations for each value between the priorities given to it by younger and older students, aggregated within each school. The hypothesis was supported by the school-level  correlations presented in Table 3. Positive correlations were found for all six values (although the correlation for harmony was not significant, r = .32).
A further test of our findings was whether teachers and students at each school have similar values. In other words, if the value priorities of teachers and students in the same school correlate highly positively, this would indicate that school values as conceptualized and measured here do represent the values of each school, regardless of whether they have been obtained with teacher or student data. Moreover, it would support the idea that school values are a construct relevant to school members of different ranks. This is the case, as demonstrated by school-level correlations of each value across the six schools between students and teachers (Table  3). Across schools, teachers' and students' aggregated values correlated strongly (minimal r = .76, maximal r = .96). The correlations between teachers' scores and students' scores were somewhat higher for the older students (average r = .85) than for the younger students (average r = .75). Interestingly, although students' and teachers' compliance values correlated very highly (r = .96, p < .01) across schools, teachers still valued compliance (M = −0.30, SD = 0.48) more than students did (M = −0.62, SD = 0.26), t(5) = 3.23, p < .05, D = 0.86.

Discussion
Although the number of schools in Study 2 was relatively small, the number of individual students in each school was large (192.6 on average), and thus school values could be measured reliably. The results of Study 2 replicated the value typology and structure found in Study 1 and extended them to a third cultural context. Furthermore, the results showed that school values as assessed here are common to different school subpopulations: teachers, older students, and younger students.

General Discussion
This study proposes and tests a typology of school-level values. Results from three cultural contexts all point to the existence of the same six values. MDS analyses of the value scores find that the order of the values is consistent across value domains: dominance, compliance, harmony, egalitarianism, autonomy, and achievement. In Germany, the substantial correlations between students of different ages and teachers in Study 2 supported the validity of the findings, and the behavioral correlates of values (school violence, support for students, and participation in decision making) demonstrated the importance of studying school-level values as a potential precursor of other school characteristics.

School-Level Value Content
Schools are groups of individuals that work together to promote group and individual goals. We therefore expected schools to differ across the three value dimensions that define the solutions which groups in general give to three issues: embeddedness versus autonomy, hierarchy versus egalitarianism, and harmony versus mastery (Schwartz, 1999(Schwartz, , 2008. This expectation was partly supported, as each of the three issues was represented in the school-level values we identified-autonomy, dominance and compliance (representing hierarchy), egalitarianism and harmony. However, the value of embeddedness, which distinguishes among national cultures, did not emerge as important in distinguishing among schools; and mastery values took a unique form of achievement values when distinguishing among schools. We will next discuss the similarities and differences between the national-level and the school-level value dimensions.
The embeddedness-autonomy dimension is represented by autonomy values in this study, promoting openness to intellectual experiences and the importance of behavioral choice. Thus, schools differ in how much they consider it important that individuals are autonomous (see also Schulte et al., 2002). One way further investigations of this value could go is to study the tightness versus looseness (Gelfand et al., 2011) of rules and codes at schools. For example, it is possible that in high-autonomy schools, the dress code of students (and teachers) will be more flexible than in low-autonomy schools. In these schools, parental school involvement (Rosenblatt & Peled, 2002)  be higher than in schools in which autonomy is not highly important. These schools can also stress students' curiosity and personal responsibility for studies more than other schools (Cheng, 2011). It is interesting to note that a clear embeddedness region had not emerged in any of the studies. Embedded cultures are those that expect meaning in life to come largely through social relationships, through identifying with the group, participating in its shared way of life, and striving toward its shared goals. It is possible that the importance of embeddedness values is largely derived from the national or ethnic culture, or from the family system, and schools do not create their unique atmosphere of embeddedness. Thus, schools cannot be distinguished by their embeddedness values.

may also
The hierarchy-egalitarianism dimension is represented by dominance values, that emphasize striving to reach the top of the social ladder, and compliance values, that emphasize acceptance of one's role in the social hierarchy system. In high-dominance schools, a more differentiated status system is expected as compared with low-dominance schools. Schools that have a relatively formalized, centralized, and well-established system of superordinate-subordinate relationships (Pang, 1996), for example, between teachers and students or between older and younger students, are likely to be those in which dominance values are relatively important. In high-compliance schools, acceptance of the rules and following them are likely to be emphasized. The interaction between compliance and dominance values in predicting school violence emphasizes also the importance of studying the joint contributions of different values to schoollevel behaviors.
It is interesting to note that dominance and compliance emerged as distinct in the present studies, in contrast to national cultures, in which the values were intermixed. This may point to the highly hierarchical system of schools in which the roles of students and teachers are defined and differentiated.
Egalitarianism values were also found to form a clear region in the present study. The role of cooperation and amity goals in achievement setting such as school is being increasingly acknowledged (Doyle Lynch et al., 2012;Levontin & Bardi, 2012). Future studies should examine the relationships between the importance given in schools to collegiality, mutual support (Higgins-D' Alessandro & Sadh, 1997;Pang, 1996;Rosenblatt & Peled, 2002) and egalitarianism values, and school climate and success. The role of schoollevel egalitarianism values in predicting out-of-school prosocial behaviors such as volunteering should also be investigated.
Harmony values represent one end of the masteryharmony dimension. In the school context, they promote interpersonal consideration and the smooth interaction between members of the social group. Accordingly, they were found to relate meaningfully to teacher support and to student participation in decision making in Israeli high schools. The involvement of students and teachers in highharmony schools is also likely to relate to attitudes toward out-group members, because harmony values emphasize a peaceful resolution of conflicts.
Because schools give different emphases to reaching high academic standards (Pang, 1996), we proposed a schoollevel value dimension of achievement. In support of this hypothesis, a clear achievement value domain emerged in both studies. The notion that the amount of reinforcement and incentives for high academic standards (Pang, 1996) is predicted by schools' achievement values should be investigated in further research. Achievement values at the school level may provide a clue as to why some schools are competitive and others less so. It can also point to the source of differences between schools in the motivation of students and their achievements.

The Interactive Relationships among School Values
The values found to distinguish among schools can form a system of preferences characterized by the contradictions and compatibilities among them. Egalitarianism and autonomy are related positively because they share a view of the social actor as an autonomous entity. Achievement values relate positively to autonomy values, because they both presume that changing the status quo is legitimate and desired, and they both emphasize activity and assuming control over one's fate. Achievement values are also linked to dominance values because the aspiration to be first and get ahead is often at the expense of others, and can result in unequal allocation of roles and resources (Schwartz, 1999). Dominance and compliance values are compatible as they both legitimize the unequal distribution of power within a society that can be expressed by either assuming a dominant or a subordinate role. Compliance values are compatible with harmony values, as they share the emphasis on not disturbing the social or general status quo. Last, harmony values stand in line with egalitarianism values as they both emphasize fit into the social environment rather than controlling it and benefiting from it (Schwartz, 1999).
Harmony values, which emphasize the fit within the social world, stand in contrast to achievement values that emphasize getting ahead and succeeding more than others. Egalitarianism values, stressing equality among individuals, stand in contrast to dominance and compliance values that emphasize legitimization of the social hierarchy (Schwartz, 1999).
The structure of the school-level value system showed one difference between the studies. In each culture independently in Study 1, and in Study 2, the order of the values was as follows: dominance, compliance, harmony, egalitarianism, autonomy, and achievement. In Study 1, when collapsing both cultures, the order was different, with harmony values adjacent to dominance and achievement values. This order is less theoretically reasonable, and may have resulted from cultural differences in value importance, with Arab schools emphasizing harmony and achievement values more than Jewish schools, creating a spurious correlation between the values.

School Values and their Relevance to Different School Members
The finding of substantial school-level correlations in the importance given to the values in Study 2 points to school values as a construct that is relevant to individuals of different ranks and positions within the school system. This does not preclude, however, the possibility of average differences between different groups within the same school. The higher importance given by teachers as compared with students to compliance may well reflect the different roles and statuses that teachers and students have at schools, or the different ages and stages in life of teachers and students. Such mean differences are an important path for future research, preferably with a larger number of schools.
The teacher-student correlations were positive for all values but extremely high for compliance (r = .96) and egalitarianism (r = .93) values. One possibility is that school policies that affect students and teachers affect the importance of these values similarly for both groups. Alternatively, values highly relevant to interpersonal behavior, such as being humble, modest, and obedient (compliance), and caring, helpful, and loyal (egalitarianism), may have a mirroring effect between students and teachers. In schools in which these values are highly important to teachers, relevant behaviors are modeled to students and affect the values of the latter.
Although the current design cannot address directly the issue of teacher influence, the comparison between age groups is telling. The correlations between teachers' and students' values tended to be stronger for the older students. This may imply a socialization effect, in which as students remain at school, their values become increasingly similar to those of their teachers. Another suggestion may be of an age effect, as older students acquire adult-like value priorities and are thus becoming increasingly similar to their teachers.

Study Strengths and Limitations
The study has several methodological strengths. First, the student samples were large. Second, three very different cultural contexts were studied, providing a replication of the value structure in a predominantly Christian context (Europe), as well as predominantly Muslim (Israeli Arabs) and Jewish (Israeli Jews) contexts. The replication of the value structure in these three contexts suggests that the school values identified in this study may apply across a wide range of cultural contexts. The similar meaning of values across cultures is a precondition for meaningful cross-cultural comparisons of value importance (see Schwartz, 1992). A replication of the results in additional cultural contexts, such as North America, South America, and Africa, can add to the understanding of school values.
Although using a variety of schools, the study investigates a limited number of schools. In particular, running an MDS with a large number of variables and a relatively small number of schools is a limitation, although each school-level data point represents groups of 15 individuals or more. Replicating the results using a wider school sample will strengthen the robustness of the results.
Another strength of the study lies in the comparison of teachers and students of different ages in Study 2. The large positive correlations across schools in the importance given to these values by different subpopulations of the schools increase our confidence in the findings. However, the large school samples reached in Study 2 limited the number of schools in which students of different ages and teachers could be asked to participate. Study 1 had a much larger number of schools but with a limited number of students in each. Although each study's advantages compensate for the other's limitation, there is still need for a study with a large number of students and teachers, nested in a substantial number of schools.
The present study extends beyond identifying school value dimensions to show the relevance of the values to important school characteristics. We have demonstrated that such meaningful relationships exist with teacher support and school violence. Future studies can pursue many other possible relations, such as the relations between school values, school performance, and drop-out rates.
Another possible limitation is the reliance mainly on data from students. This is the reason for the inclusion of teacher data in Study 2. Across six schools, student-teacher correlations were positive and large. In addition, the value structure obtained from teachers' data was very similar to that of students ( Figure 5). Nevertheless, future studies should seek to include other members of the school community, such as principals and nonteaching staff, in addition to students and teachers.
The measure we used in this study for assessing values, the PVQ (Schwartz & Rubel, 2005), has been designed specifically for comparing samples of a broad range of educational levels and ages. This is one reason we relied on it when studying the values of adolescents and preadolescents, and comparing them with those of adults. However, it should be noted that this questionnaire was devised for studying the values of individuals, not of schools. It would be beneficial in future research to test the typology we obtained from the aggregated individual responses, this time with a specifically designed measure that directly addresses the values of schools.

Conclusion
Studies in three cultures pointed out the replicability of the six value types, largely shared by school members of different groups. The typology reported in this article should guide future research into the phenomenon of school values. The relevance of the values to variables such as ethnicity and school violence demonstrates the importance of studying school-level values.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.