untitled 199 Abstract The paper investigates the relationship between ethnocultural diversity and political culture. The first part of the study explores the literature on ethnocultural diversity as a sociologically relevant variable, its impact on other structural variables, and the way in which ethnicity became relevant to the study of political culture. In the second part, the secondary analysis of data provided by a research conducted in Romania in 2003 shows that the relationship which can be established between the explored variables is highly inconsistent with relevant studies from the international literature. The findings are in agreement, however, with the contact theory which asserts that the exposure to diversity has a positive impact on the minority- majority relations. Broadly speaking, this research proves that ethnocultural diversity is not necessarily conducive to an environment permeable to extremist views. Due to methodological limitations, the conclusions of the study need to be dealt with precaution. ETHNOCULTURAL DIVERSITY AND ASPECTS OF POLITICAL CULTURE Levente SALAT Monica CĂLUŞER Amelia GORCEA Olimpia MOŞTEANU Levente SALAT Associate Professor, Department of Political Sciences, Faculty of Political, Administrative and Communication Sciences, Babeş-Bolyai University, Cluj-Napoca, Romania Tel.: 0040-264-431505 Email: levente.salat@polito.ubbcluj.ro Monica CĂLUŞER Program Coordinator, The Ethnocultural Diversity Resource Center, Cluj-Napoca, Romania Tel.: 0040- 0264-420490 Email: mcaluser@edrc.ro Amelia GORCEA Program Coordinator, The Ethnocultural Diversity Resource Center, Cluj-Napoca, Romania Tel.: 0040- 0264-420490 Email: agorcea@edrc.ro Olimpia MOŞTEANU Program Coordinator Assistant, The Ethnocultural Diversity Resource Center, Cluj-Napoca, Romania Tel.: 0040- 0264-420490 Email: omosteanu@edrc.ro Transylvanian Review of Administrative Sciences, 27E/2009 pp. 199-219 200 1. Introduction1 The study Intolerance, discrimination and authoritarianism of public opinion in Romania claims that the Romanian society is permeable to extremism not only because of the instability of the democratic regime, but also due to Romanian political parties’ inclinations toward populism (IPP, 2003). The research report also argues that not only the extremist parties, but also any other type of political organization is to some extent inclined to get involved in debate over controversial issues looking to enhance their political capital. It is averred that these views take form as ideological programs of marginal groups. The report asserts that even if the marginal groups lack the power to impose their views on public opinion, if the political milieu presents some particular characteristics then some of these extremist views enter the public sphere. It is the aim of the IPP research report to identify the characteristics that make possible the spread of extremist ideas into public sphere. The present approach links the aforementioned assumptions to studies regarding the detrimental impact of ethnocultural diversity on economic growth, social cohesion, trust, institutional performance and social and political stability. It also considers research studies which regard ethnocultural diversity as a predictive factor for the emergence of conflict. Bearing in mind that both the characteristics of the political milieu and the “content” of public sphere refer to the political culture, this analysis will place the issues of intolerance, discrimination, extremism in politicians’ or media discourse and the negative social representations in the larger framework of political culture. Broadly speaking, the research attempts to investigate the extent to which ethnocultural diversity is conducive or not to an environment permeable to extremist views. 2. The study of ethnocultural diversity and political culture The first attempt to use ethnocultural diversity as a predictor for certain structural variables of societies dates back to 1967, when Marie R. Haug tried to classify societies according to varying levels of pluralism and to establish systematic differences among the ranked societies in critical structural variables. With the help of an index of pluralism she ranked 114 independent polities of the world (based on data from 1963) and found that the degree of pluralism is in a linear relationship with a series of demographic, economic and political variables. Based on her findings, Haug concluded that pluralism (ethnocultural diversity in our terms) is a factor which cannot be ignored in social system analysis (Haug, 1967).Based on data included in the Atlas Narodov Mira (1964), Taylor and Hudson (1972) constructed an index of ethnolinguistic fractionalization2, measuring the probability that two randomly selected individuals in a country will belong to different ethnolinguistic groups. 1 This article has been elaborated as part of a research project entitled The Impact of Diversity on Political Culture in Romania, financed through UEFISCSU, Contract no.419/1.10.2007. 2 1 1 2n i i pF = −= , where p1, p2, p3...pn are the share of the various thnic groups. 201 Fearon (2003) provides a cross-national database with values of ethnic fractionalization calculated with the help of available data, warning however that there are significant inconsistencies among the sources he used. In addition to the fractionalization index, Fearon constructs a measure of cultural diversity3 calculated with the help of the estimated linguistic and religious distance among groups. Alesina et al. (2003) distinguish among the ethnic, linguistic and religious dimensions of diversity, calculating for each of these dimensions the value of the fractionalization index, characteristic for 215 states and dependent territories. In spite of the different sources they used, the results of Fearon and Alesina et al. are largely similar. Montalvo and Reynal-Querol (2005) construct an index of polarization4 which measures the normalized distance of a particular distribution of ethnic and religious groups from a bimodal distribution. The Q index is used to assess the relationship between social heterogeneity and the likelihood of conflict, offering a more nuanced picture of the widespread belief that increase in diversity increases the probability of conflict. Since evidence show that there are more conflicts in societies where a large ethic minority faces a majority, the polarization index is a more reliable predictor of ethnic conflict than fractionalization. Montalvo and Reynal-Querol provide a cross-national database with the values of the four variables: ethnic polarization, ethnic fractionalization, religious polarization and religious fractionalization, calculated for the world countries. While the fractionalization index grows with the number of the groups, the polarization index has a maximum at two groups. For low levels of fractionalization the correlation between ethnic fractionalization and ethnic polarization is positive and close to linear, for the medium range of the fractionalization values the correlation is zero, while for high values of fractionalization the correlation is negative (Montalvo and Reynal-Querol, 2005). As far as the impact of the different diversity measures on various structural variables is concerned, the opinions of different authors vary on a large scale. Haug found that geographic size and thinly scattered, illiterate population regularly correlates with high values of the pluralism index, as dependent variable, while pluralism, as independent variable in this case, goes hand in hand with low per capita GDP and political instability (Haug, 1964). According to Mauro (1995), between ethnolinguistic fractionalization and institutional efficiency a strong negative correlation can be established, while fractionalization correlates positively with corruption. Furthermore, high level of ethnocultural diversity implies generally a lower level of investment. Mauro sees fractionalization as a proxy for ethnic conflict, which may lead to political instability or even civil war. Easterly and Levine (1997) argue that ethnic diversity helps in explaining cross-country differences as far as public policy choices and other economic indicators 3 ,1 1 1 −= = = ijj n i n j i rppCD where: pi – ethnic group i; pj – ethnic group j; rji – resemblance factor. 4 w,45.0 5.0 1 j1 1 2 2 1 πππ π = ≠ = = − −= N i j ii N i iQ where πs are the sizes of the group in the population 202 are concerned. Focusing on Sub-Saharan Africa, they found that the region’s high ethnic fragmentation can explain a significant part of the indicators which reflect low schooling, political instability, underdeveloped financial systems, high government deficits and poor infrastructure. In their view, the findings are not characteristic exclusively to Africa: it can be generally stated that high ethnic diversity is regularly associated with interest group polarization which leads to rent-seeking behavior and reduces the consensus on public goods (Easterly and Levine, 1997, p. 1241). La Porta et al. (1999) found, as well, that ethnolinguistic fractionalization is in a significant degree negatively correlated with latitude and per capita income, and has a consistent adverse effect on government performance: higher fractionalization is associated with more interventionism (worse property rights and regulation), lower government efficiency (more corruption, longer delays and lower tax compliance), inferior provision of public goods (higher infant mortality and illiteracy, lower school attainment and infrastructure quality), more state enterprises and less political freedom, due mainly to the predisposition of ethnic groups in power to redistribute income (La Porta et al., 1999, p. 245). Fearon (2003), on his turn, warns that the effect of ethnic fractionalization on structural variables like economic growth, political conflict, party structure etc. might depend on the specific measure of diversity the particular researches deploy. In a more nuanced analysis of the different dimensions of diversity as possible predictor of certain structural variables, Alesina et al. (2003) conclude that high values of ethnic and linguistic fractionalization regularly correlate with low GDP output and poor economic performances, democratic deficits, bad governance (reflected in the quality of public policies, literacy rate and infant mortality), poor performance of public institutions (measured by the extent of corruption, indicators of fundamental and political freedom etc). However, the study emphasizes that it is difficult to assess precisely the size of such effects, due to the strong correlation of ethnolinguistic fractionalization variables with a number of other potentially explanatory variables, especially geographic ones (latitude, for instance). Alesina et al. found on the other hand that religious fractionalization correlates positively with economic growth, public policy quality, corruption control, literacy rate and civic and political liberties. It is also observed that ethnic fractionalization has a stronger predictive potential than linguistic fractionalization, which seems to generate effects less significant statistically (Alesina et al., 2003, pp. 182-183). Montalvo and Reynal-Querol (2005) go further in refining the analysis by assessing separately the impact of ethnolinguistic fractionalizations (F) and ethnolinguistic polarization (Q). They confirm the significant negative direct effect of ethnolinguistic fractionalization on economic growth, reported by many other authors, while the ethnolinguistic polarization exerts, in their view, its negative impact through indirect channels: investment, government consumption and the incidence of civil war. As far as the religious dimension is concerned, according to Montalvo and Reynal-Querol, religious polarization exerts the expected negative effect on the above mentioned indirect channels, while the effect of religious fractionalization is the opposite: more 203 diversity increases the investment rate, decreases the level of government consumption over the GDP and reduces the probability of civil war, which is largely consistent with the findings of Alesina et al. (2003). The relationship between ethnocultural diversity and the probability of civil war is addressed by Fearon and Laitin (2003). The authors argue that conditions which favor insurgency - weak states marked by poverty, large population and political instability - are better predictors of the likeliness of ethnic violence than the indicators of ethnic and religious diversity or even measures of various grievances, like economic inequality, lack of civic and political liberties, state discrimination against minority religions and languages. In addition to the cross-national perspective detailed in the above, Alesina, Baqir and Easterly (1999) observe that more ethnically diverse jurisdictions in the United States have higher spending and higher deficits/debt per capita, and yet devote lower shares of spending to core public goods like education, roads, sewer and trash pickup. These findings confirm theories holding that heterogeneous and polarized societies value less public goods, patronage more and are collectively careless about fiscal discipline. Putnam (2007) distinguishes between the immediate and the long lasting effects of ethnic diversity on social solidarity and social capital. Stemming from a quantitative analysis conducted in United States, his findings suggest that in ethnically diverse neighborhoods, residents of all races are likely to present lower levels of trust. As Putnam argues, “trust (even of one’s own race) is lower, altruism and community cooperation rarer, friends fewer; but in the long run, however, successful immigrant societies have overcome such fragmentation by creating new, cross-cutting forms of social solidarity and more encompassing identities” (Putnam, 2007, p. 137). Other studies focusing on the constituents of social capital and their macro or micro effects on society claim that social trust presupposes homogeneous societies, therefore researchers should keep in mind that this form of social capital is apparently less characteristic in the case of more diverse societies (Hooghe, 2005). Subsequently, it is argued that, when studying the relationship between social capital and ethnocultural diversity, researchers need to pay attention to contextual information about the environment in which diverse social interactions occur or to the attitudinal aspects of social cohesion (Hooghe, 2005; Hooghe et al., 2006). If the study of ethnocultural diversity and the research dedicated to its impact on various structural variables of the polity are relatively recent phenomena, political culture has been in the focus of the attention of several influential authors since the seminal work of Almond and Verba (1963), The Civic Culture. Based on a comparative study conducted between 1957-1962, aiming to assess public attitudes towards politics and their role in shaping the political system in five countries (US, Great Britain, Germany, Italy and Mexico), Almond and Verba concluded that a nation’s political culture reflects its citizens orientations and evaluations towards three levels: the political system, the political and policymaking process and the policy outcomes and outputs. They found that political cultures are defined essentially by the combination of three patterns of orientation toward involvement in the political process, each the result of different circumstances of political socialization: parochialism (characterized by total 204 lack of awareness of the ingredients of the system and process), subject-mentality (passive obedience and lack of active involvement) and participatory behavior (active involvement in the process). The share of the three patterns in the overall political culture of a nation, respectively the dominance of one or the other define the nature of the political system as authoritarian or democratic,traditional/preindustrial or industrial, and combinations of the former categories. The concept of civic culture was coined by Almond and Verba to describe a particular type of political culture, found to be dominant for Western democratic countries, which combines the allegiant subject mentality with the constraints and filtering effects of the participant mentality. The concept itself emphasizes the key ingredients of political stability and equilibrium in a modern society: the willingness of the citizens to be governed, accepting the legitimacy of power and authority, as well as the citizen’s willingness to participate in power and decisions (Almond, 1996). Another important element of the theory of Almond and Verba is the cultural congruence thesis which holds that political structures and political culture are mutually reinforcing in the sense that sustaining democratic institutions in a country lacking a sufficient proportion of subjects ready and prepared to participate, or to impose authoritarianism over a society with a significant share of people willing to participate, are equally impossible. Due to its heuristic and interpretative power, the concept of political culture has become over time the target of serious academic scrutiny. The congruence thesis was confirmed by Inglehart (1990), while Putnam (1993) offered consistent arguments in the support of the political socialization theory. The balance inclines, however, in favor of those who have found the concept vulnerable. The most frequent critical arguments target the concept’s weaknesses as an explanatory variable: instead of being empirically measurable, some see it as a vague, stereotypic description of the unit of analysis (Kaase, 1983 apud Dalton, 2000), while others highlight the weak predictive power of the concept, with reference to the uneven evidence of its causal effects (Laitin, 1995). Indeed, Inglehart (1977, 1990, 1997), Barnes et al. (1979), Dalton et al. (1994), Putnam (1995), Norris (1999), and Dalton (2000) provided evidence of political transformations in the world which the concept itself could not forecast5. The equilibrium thesis of the civic culture has been challenged by the emergence of the postmaterial value change followed by participatory populism, decline of the public trust in government, voter dealignment and realignment, growing political spontaneity and disrespect for traditional political authority. The congruence thesis has also come under pressure during the 1990s, in the context of the democratic transformation in Eastern Europe, where the role of the political culture has been less evident (Dalton, 2000; Armingeon and Careja, 2008), the 5 Bad publicity to the forecasting potential of the concept has been provided especially by Walter A. Rosenbaum, who reached the conclusion, in a study dedicated to the cohesive capacity of political culture over language differences, regionally dispersed resources and parochial loyalties, that in authoritarian or totalitarian systems are more likely to dominate among the world’s political cultures (Rosenbaum, 1975). 205 developments sustaining Eckstein’s (1992) thesis about the gradual change of political culture and the frequent syncretism between the institutions of new political system and elements of the previous regime’s political culture. In 1996, under the pressure of mounting evidence, Almond saw himself obliged to conclude that “the balanced mix of the Civic Culture of loyal subjects and consensual participatory elements celebrated in our book of 1963, begins to give way to alienated subject combined with a form of participation weakened and demoralized by populism, extremism and apathy” (Almond, 1996, p. 10). Political culture remains nevertheless in the focus of scholarly research in political science. The growing sophistication of political behavior attracts much attention, researchers aiming to understand the new complexities of public-issue interests’ increasing heterogeneity, the refusal of more and more citizens to depend on party elites or on conventional, institutionalized methods of political participation, the shift from involvement in electoral politics to more direct forms of political participation, or the effort of several democratizing nations across the world to seek variants of democracy which are more congruent with their history and cultural traditions (Dalton, 2000). A special chapter of these more recent research endeavors is dedicated to the study of the new forms of political culture and emerging new social movements, amongst which identity politics, politics of difference and politics of recognition. Several of the studies published under these headings (Nelson, 1979; Bickford, 1999; Scarritt and Mozzafar, 1999; Alesina and La Ferrara, 2000; Deleon and Naff, 2004; Posner, 2004; Segura and Rodrigues, 2006; Birnir, 2007 etc) discover the relevance of ethnocultural diversity to politics in general, and political culture in particular, exploring the various ways in which race, ethnicity, gender, identity and group membership influence the different forms of political participation. 3. Data sources, variables and method of analysis The data analyzed in the research were provided by a survey conducted in 2003, within the frameworks of a project initiated by the Institute for Public Policy, aiming to identify indicators of intolerance, discrimination and authoritarianism in the country’s public opinion.6 The project was meant to raise the awareness of political and civil society actors upon extremist movements and also to identify policies and programs that promote tolerance, strengthen democratic values and diminish the impact of radical views on public discourse. The Institute for Public Policy averred that extremist views were not marginal phenomena, but an alarming indicator of intolerance sustaining social deprivations that had characterized the Romanian society for a long period of time (IPP Report, 2003, p. 6). Since the project explored extremist attitudes and behaviors such as 6 The survey Intolerance, Discrimination and Authoritarianism of Public Opinion in Romania was part of the Extremism in Romania program run by IPP. The data were collected by the Gallup Organization Romania in September 2003. The sample, including 1500 respondents, was representative for the adult, non-institutionalized Romanian population, with an accepted error of ± 2.7 %. The survey book can be accessed on the www.ipp.ro website. The analyzed data set has been kindly made available to us by the Institute for Public Policy. 206 discrimination and intolerance of ethnic, religious and sexual minorities, authoritarian inclinations and religious fundamentalism, it seemed to provide a suitable research target for our intent to assess the relationship between ethnocultural diversity and political culture. Given that the data were collected “after the facts”, and since at the moment we initiated the research we had no control over possible extraneous variables, the research was carried out with the help of an ex post facto research design. Taking into consideration the research question, the sampling method, the degree of control over the variables, as well as the time factor, we reached the conclusion that a correlational research design is most appropriate. With regard to the nature of the analysis, a secondary data analysis has been employed, consisting in a number of analyses of pairs of values for each respondent, in order to assess the corresponding relations between the explored variables. The major conceptual issues that should be addressed when carrying on a secondary data analysis include the problem of theoretical congruence and variable operationalization (Orsi et al., 1999). The issue of theoretical congruence refers to the extent to which the questions used in secondary analysis are different from the research questions in the original study. Fawcett et al. (1992) argue that once the questions and the corresponding concepts have been identified, the researcher must ascertain whether the operational definitions of the concepts in the original data set are consistent with the concepts that are to be used in the secondary analysis. When the operationalization of the concepts in the secondary analysis research questions does not fit the theoretical definitions in the original data sets, then the concept operationalization may limit the ability to use the original empirical indicators (Kiecolt et al., 1985). Moreover, the use of those empirical indicators in the secondary analysis may be detrimental for the validity of the research design. Creating empirical indicators from original data sets can be employed, but the issues of the operational definitions and item comparability should be considered. Our hypothesis was generated with the help of the existing theories and the previously evoked research findings regarding the impact of various indicators of ethnocultural diversity upon political culture. Based on the findings reflected in the literature, we expect that the level of ethnocultural diversity to be related, in away or other, to the political culture, i.e. there are statistically significant correlations between the level of ethnocultural diversity and the political culture. There is a high level of theoretical congruence between the data set resulted from the Intolerance, discrimination and authoritarianism of public opinion in Romania survey and the scope of this research project. Therefore, the concept of political culture will be operationalized in accordance with the original empirical indicators, as follows7. The 7 An inclusive, yet hard to use, operationalization of political culture devised by Elazar comprises the following dimensions: 1) Orientation to political organization: “how have particular people or groups of people addressed the organization of political life?” 2) Orientation to civil society: elitist, participatory elitist, leader-centered populist and participatory populist. 3) Orientation to polity: civic, corporatist, statist, subject, and tribal.4) Orientation to political action: represents the summation of persistent patterns of underlying political attitudes and 207 measurement of the political culture consists of eight ordinal variables, each measured on 4-point scales. The first 15-item variable reflects the feelings one has about being Romanian. Among others, the question explores the pride of being Romanian, the preference of Romanians over the members of other ethnocultural groups, and the presence or absence of tolerance for minorities. The second question is based on the Bogardus social distance scale and investigates individuals’ willingness to participate in social contacts of varying degrees of closeness with members of diverse social groups (in our study we look at Roma, Hungarian, Arab, Chinese, Black people, Jehovah’s Witnesses, Muslim, homosexual, lesbian, citizen of Republic of Moldova). The third variable deals with views pertaining to the Roma people. The forth one considers a more general topic, namely the opinion about the number of rights held by minorities in Romania. The variable explores the extent to which Romanians feel that minority groups in Romania have “too many rights”, “too few rights” or “just as needed”. The fifth variable is a 6-item question concerning the Romanian-Hungarian relations ranging from education to autonomy issues. The last three variables deal with several aspects of people’s orientation to political organization focusing on elements as different as “the man should decide in a family”, “it is good to have a market economy” or “books and ideas that distrust state authority should be forbidden”8. Ethnocultural diversity is measured using the ethnic and linguistic fractionalization indices. The attractiveness of these indices in empirical applications can be understood if one looks at two different aspects. On the one hand, the computation of these indices can be performed both from micro and from aggregate data: all that is required is the value of shares of the various groups in the population. On the other hand, these indices benefit from an intuitive interpretation: they measure the probability that two randomly drawn individuals from the overall population belong to different ethnic groups. 4. Data analysis Alesina et al. (2003, p. 188) found the following fractionalization indices for Romania: the index of ethnic fractionalization is 0.3069; the index of linguistic fractionalization is 0.1723; the index of religious fractionalization is 0.2373. As for the present construction of indices we used the 2002 Romanian Census in order to come up with the values of the shares of the various groups in the Romanian population. These indices were constructed to match the cases comprised in the Institute for Public Policy’s data set; therefore, from the total of 2952 Romanian localities we used the 106 localities included in the IPP sample, in order to properly link the available data on political culture to the ethnolinguistic distribution of Romanians. Hence, unlike Alesina et al. (2003), the values of the fractionalization indices we use are not illustrative for the entire characteristic responses to political concerns that is manifest in a particular political order. 5) Orientation to political economy: market, corporatist, commonwealth, state involved, and state controlled political economy (The Daniel Elazar On-Line Library, available at http:// www.jcpa.org/index.htm). 8 For the complete list of variables see Appendix, List of the variables used to measure political culture. 208 Romanian population, but they reflect the level of diversity for those 106 localities our sample has individuals from. The indices at the level of the respective localities are as follows: 0.1104 the index of ethnic fractionalization; 0.0890 the index of linguistic fractionalization; 0.1608 the index of religious fractionalization. These variables are measured on an interval scale. Table 1: Indices of ethnic and linguistic fractionalization Ethnic fractionalization index Linguistic fractionalization index N Valid 1500 1500 Missing 0 0 Mean .110473 .089071 Std. Error of Mean .0036842 .0034244 Median .058302 .030132 Mode .0583 .0301 Std. Deviation .1426873 .1326257 Variance .020 .018 Range .6191 .5494 Minimum .0000 .0000 Maximum .6191 .5494 Given the interpretation of the fractionalization indices (a “.0000” value indicates that there is only one ethnic or linguistic group; a “.5000” value signifies an equal distribution between two groups; a “1.0000” value corresponds to a great number of small groups) we can state that, with regard to the cases investigated, in the most frequently met situations we have one large group and a small number of tiny groups. It is important to note that the “large group” can represent (1) a group that is both a majority at the national level and in a certain locality; or (2) a group that is a minority at the national level but a majority in a certain locality. Table 2: Correlations between the indices of ethnic, linguistic and religious fractionalization Ethnic fractionalization index Linguistic fractionalization index Religious fractionalization index Ethnic fractionalization index Pearson Correlation 1 .987(**) .804(**) Sig. (2-tailed) .000 .000 N 1500 1500 1500 Linguistic fractionalization index Pearson Correlation .987(**) 1 .812(**) Sig. (2-tailed) .000 .000 N 1500 1500 1500 Religious fractionalization index Pearson Correlation .804(**) .812(**) 1 Sig. (2-tailed) .000 .000 N 1500 1500 1500 ** Correlation is significant at the 0.01 level (2-tailed). 209 The second table shows the positive high correlations between these three constructed indices. As one can see, the correlation between the index of ethnic fractionalization and the index of linguistic fractionalization is very strong: 0.987, which means that ethnic and linguistic identity overlap to a significant degree in Romania. As far as the correlation coefficients of the operationalized variables are concerned, the following can be stated. Throughout this analysis we used Spearman’s Rho nonparametric coefficient because the studied relationships entail both interval (ethnic fractionalization index) and ordinal data (all the variables that measure political culture). In other cases we constructed indices to discover if some multiple dimensional questions were explained by latent factors and to ascertain how these factors correlate with the ethnolinguistic indices. A 0.903 Cronbach’s Alpha coefficient showed the internal consistency of the 15 items of the first variable. A factor analysis was employed for this variable and it rendered a 0.922 statistically significant Kaiser-Meyer-Olkin coefficient, proving the appropriateness of conducting such an investigation. A two-dimensional factor was found: the first dimension referred to “the pride of being Romanian” and the second one to “the preference of Romanians over other people”. The factors explain the 56.84% of the existent variance. The factor scores estimates were saved as variables using the regression method (the synthetic variable was created by applying weights to the all observed variables). These were used afterward in a correlation analysis that revealed that “the pride of being Romanian” dimension negatively correlates with the ethnic and linguistic indices and “the preference of Romanians over others” positively correlates with the ethnolinguistic fractionalization indices. In other words, the correlational analysis showed that, when the pride of being Romanian is assessed in more diverse localities (diversity being illustrated by the increase in the fractionalization index), it has lower values, but when it is measured in rather homogeneous localities (characterized by an index of fractionalization almost equal to.0000), it records higher values. The positive Pearson’s coefficients explained that individuals from rather homogeneous communities are less likely to share “the feeling of preferring Romanians over other people” compared to those from heterogeneous communities. The not so high values of Pearson’s correlation coefficients could stem from the difficulty to clearly assess both “the pride of being Romanian” and “the preference of Romanians over others” when the whole Romanian population is taken into account, namely because at the national level there are various levels of experienced ethnocultural diversity. 210 Table 3: Correlations between the regression factor scores and the ethnolinguistic indices Ethnic fractionalization index Linguistic fractionalization index Pride of being Romanian Preference of Romanians over others Ethnic fractionalization index Pearson Correlation 1 .987(**) -.166(**) .096(**) Sig. (2-tailed) .000 .000 .001 N 1500 1500 1147 1147 Linguistic fractionalization index Pearson Correlation .987(**) 1 -.157(**) .108(**) Sig. (2-tailed) .000 .000 .000 N 1500 1500 1147 1147 Pride of being Romanian Pearson Correlation -.166(**) -.157(**) 1 .023 Sig. (2-tailed) .000 .000 .443 N 1147 1147 1147 1147 Preference of Romanians over others Pearson Correlation .096(**) .108(**) .023 1 Sig. (2-tailed) .001 .000 .443 N 1147 1147 1147 1147 ** Correlation is significant at the 0.01 level (2-tailed). With regard to the indicators of social distance9, the analysis revealed that it is significantly correlated with the indices of ethnolinguistic fractionalization. We extracted two factors which explained 41.467% of the common variance. The first one refers to the sexual diversity and the other one comprises all the other types of ethnic, racial and religious diversity. We found small significant positive Pearson’s correlation coefficients between the first factor and the ethnolinguistic indices and moderate significant negative correlations between the second factor and the indices of fractionalization. This implies that, to some extent, ethnic diversity could be conducive to a more tolerant feeling towards all a various ranges of groups, except sexually diverse minorities10. The small values of the Pearson’s coefficients support the previous remarks concerning the difficulty of assessing the experienced diversity at the national level, and this becomes even more noticeable when one peruses the correlation coefficients corresponding to the relations between the sexual diversity and the indices of ethnocultural diversity. The investigation of the third and fourth questions referring to the Romanian political culture, namely to the opinion about the amount of rights held by the Romanian national minorities and regarding the Roma population, lead to the conclusion that there are no statistically significant relationships between the ethnolinguistic fractionalization indices and these aspects of political culture. Therefore, it is hard to articulate any conclusions about how ethnocultural diversity relates either to the rights held by the Romanian minorities or to the life and needs of the Roma population11. 9 This is an 11-item question measured on a 6-point scale, for details see Appendix, Q2. It has a 0.947 Cronbach’s Alpha coefficient which proves its internal consistency. 10 For more information on these correlations see Appendix, Table 5. 11 For further statistical data on this issue see Appendix, Table 6. 211 The analysis dedicated to the fifth variable, targeting the extent to which people agree/disagree with several statements about the Hungarians in Romania12, revealed that the more diverse a community is the more openness people show towards the needs of minorities. The results also ascertained that in homogeneous communities individuals are less inclined to be tolerant. For the sake of exemplification, when the ethnic fractionalization index has a lower value the subjects tended to agree with the statement “Hungarians should use only the Romanian language in their relation with the public administration”. Conversely, when the ethnolinguistic indices have higher values the subjects were inclined to agree with the statements “Romanian state should provide Hungarian children education in Hungarian” and “The counties where the Hungarians represent majority should be offered increased autonomy”13. The analysis of the last three questions that operationalize political culture illustrated a set of interconnected aspects with regard to the Romanian political culture. The results revealed that ethnic diversity is less likely to be conducive to extremist or radical views. As a case in point, there is a medium negative correlation coefficient between the “it is good to have only one political party” variable and the index of ethnic fractionalization, meaning that to a higher degree of diversity corresponds a lower degree of agreement on the issue and that to a lower degree of diversity corresponds a higher degree of agreement. 5. Conclusions and limitations This research study placed the concept of political culture within the framework of the ethnic diversity and tried to investigate their relationships. Given the constraints discussed in the methodological part, the focus was on the correlations between a set of variables measuring the Romanian political culture and the index of ethnic fractionalization. In the light of the presented data we can state that there are, indeed, statistically significant correlations between the level of ethnocultural diversity and political culture, therefore the hypothesis is confirmed. As far as the nature of the established relationship is concerned, the findings provide a somewhat contradictory picture about the investigated phenomena. One simple answer to our research question would be that the relationship found is highly inconsistent with other international studies which reflect the detrimental effect of ethnocultural diversity on a large variety of structural variables, both in cross-national and in national contexts. Nevertheless, our findings are in agreement with the contact theory which asserts that exposure to diversity has a positive impact on the minority-majority relations. Broadly speaking, this research revealed that ethnocultural diversity is not necessarily conducive to an environment permeable to extremist views. One should keep in mind, however, that this statement is based on correlational analyses and this aspect, together with other limitations, asks for a precautious interpretation of the results. 12 Measured on a 4-point scale, using a 7-item question, for details see Appendix, Q5. 13 For further statistical data on this issue see Appendix, Table 7. 212 The most significant limitations of the present research paper are the following: (1) the selection of the units of analysis was influenced by the available data set, together with the employed variables and indicators; (2) the operationalization of the political culture might not be satisfactory, the deployed indicators failing to reflect the complexity of the concept; (3) should have we used other sets of indicators, the outcome might have been different. References 1. Alesina, A. and La Ferrara, E., ‘Participation in Heterogeneous Communities’, 2000, The Quarterly Journal of Economics, vol. 115, no. 3, pp. 847-904. 2. Alesina, A., Baqir, R. and Easterly, W., ‘Public Goods and Ethnic Divisions’, 1999, The Quarterly Journal of Economics, vol. 114, no. 4, pp. 1243-1284. 3. Alesina, A., Devleeschauwer, A., Easterly, W., Kurlat, S. and Wacziarg, R., ‘Fractionalization’, 2003, Journal of Economic Growth, vol. 8, no. 2, pp. 155-194. 4. Almond, G. and Verba, S., The Civic Culture, Princeton, NJ: Princeton University Press, 1963. 5. Almond, G., ‘Comparative Political Systems’, 1956, Journal of Politics, vol. 18, no. 3, pp. 391-409. 6. Almond, G., ‘The Civic Culture: Prehistory, Retrospect, and Prospect’, Paper 96-01, Center for the Study of Democracy, University of California, Irvine, [Online] available at http://repositiries.cdlib.org/csd/96-01. 7. Armingeon, K. and Careja, R., ‘Institutional Change and Stability in Post-communist Countries, 1990-2002’, 2008, European Journal of Political Research, vol. 47, no. 4, pp. 436-466. 8. Barnes, S. and Kaase, M., (eds.), Political Action: Mass Participation in 5 Western Democracies, Beverly Hills: SAGE Publications, 1979. 9. Bickford, S., ‘Reconfiguring Pluralism: Identity and Institutions in the Inegalitarian Polity’, 1999, American Journal of Political Science, vol. 43, no. 1, pp. 86-108. 10. Birnir, J.K., Ethnicity and Electoral Politics, Cambridge: Cambridge University Press, 2007. 11. Bourdieu, P., “The Forms of Capital” in Richardson, J.G., (ed.), Handbook of Theory and Research for the Sociology of Education, Connecticut: Greenwood Press, 1986. 12. Bruck, S.I. and Apachenko, V.S., (eds.), Atlas Narodov Mira, Moscow: Glavnoe Upravlenie Geodezii i Kartografii, 1964. 13. Dalton, R.J., ‘Citizen Attitudes and Political Behavior’, 2000, Comparative Political Studies, vol. 33, no. 6-7, pp. 912-940. 14. Dalton, R.J., Flanagan S. and Beck, P.A., (eds.), Electoral Change in Advanced Industrial Democracies, Princeton, NJ: Princeton University Press, 1984. 15. Deleon, R.E. and Naff, K.C., ‘Identity Politics and Local Political Culture: Some Comparative Results from the Social Capital Benchmark Survey’, 2004, Urban Affairs Review, vol. 39, no. 6, pp. 689-719. 16. Elazar, D., “Globalization Meets the World’s Political Cultures”, The Daniel Elazar On- Line Library. [Online] available at http://www.jcpa.org/dje/articles3/polcult.htm 17. Fawcett, J. and Downs, F., The Relationship of Theory and Research, Philadelphia: F.A. Davis, 1992. 18. Fearon, J.D. and Laitin, D.D, ‘Ethnicity, Insurgency, and Civil War’, 2003, The American Political Science Review, vol. 97, no. 1, pp.75-90. 213 19. Fearon, J.D., ‘Ethnic and Cultural Diversity by Country’, 2003, Journal of Economic Growth, vol. 8, no. 2, pp. 195-222. 20. Haug, M.R, ‘Social and Cultural Pluralism as a Concept in Social System Analysis’, 1967, The American Journal of Sociology, no. 73, pp.294-304. 21. Hooghe, M. and Stolle, D., Generating Social Capital: Civil Society and Institutions in Comparative Perspective, New York: Palgrave Macmillan, 2003. 22. Hooghe, M., ‘Social Capital and Diversity. Generalized Trust, Social Cohesion and Regimes of Diversity’, 2007, Canadian Journal of Political Science, vol. 40, no. 3, pp. 709-733. 23. Hooghe, M., Reeskens, T., Stolle, D. and Trappers, A., “Ethnic Diversity, Trust and Ethnocentrism in Europe. A Multilevel Analysis of 21 European Countries”, Paper presented at the annual meeting of the American Political Science Association, Philadelphia, USA, August 31-September 3, 2006. 24. Inglehart, R., Culture Shi in Advanced Industrial Societies. Princeton, NJ: Princeton University Press, 1990. 25. Inglehart, R., Modernization and Post-modernization, Ann Arbor: University of Michigan, 1997. 26. Inglehart, R., The Silent Revolution, Princeton, NJ: Princeton University Press, 1977. 27. Institute for Public Policy, Intolerance, discrimination and authoritarianism of public opinion in Romania. Report, Bucharest: The Institute for Public Policy, 2003. 28. Kaase, M., “Sinn order unsinn des konzepts politische kultur für die vergleichende politikforschung” in Kaase, M. and Klingemann, H.D., (eds.), Wahlen und Politisches system, Oplanden: Westdeutsher Verlag, 1983, pp. 44-171. 29. Kiecolt, K.J. and Nathan, L., Secondary Analysis of Survey Data. Beverly Hills, CA: Sage, 1985. 30. La Porta, R., Lopez-de-Silanes, F., Shleifer, A. and Vishny, R., ‘The Quality of Government’, 2003, Journal of Law, Economics & Organization, vol. 15, no. 1, pp. 222-279. 31. Laitin, D., ‘The “Civic Culture” at Thirty’, 1995, The American Political Science Review vol. 89, no. 1, pp. 168-173. 32. Mauro, P., ‘Corruption and Growth’, 1995, Quarterly Journal of Economics, vol. 110, no. 3, pp. 681-712. 33. Montalvo, J.G. and Reynal-Querol, M., ‘Ethnic Diversity and Economic Development’, 2005, Journal of Development Economics, vol. 76, no. 2, pp. 293-323. 34. Nelson, D.C., ‘Ethnicity and Socioeconomic Status as Sources of Participation: The Case for Ethnic Political Culture’, 1979, The American Political Science Review, vol. 73, no. 4, pp. 1024-1038. 35. Norris, P., (ed.), Critical Citizens: Global Support for Democratic Government. Oxford: Oxford University Press, 1999. 36. Orsi, A., Grey, M., Mahon, M., Moriarty, H., Shepard, M. and Caroll, R., ‘Conceptual and Technical Considerations When Combining Large Data Sets’, 1999, Western Journal of Nursing Research, vol. 21, no. 2, pp. 130-142. 37. Posner, D.N., ‘The Political Salience of Cultural Difference: Why Chewas and Tumbukas are allies in Zambia and adversaries in Malawi’, 2004, The American Political Science Review, vol. 98, no. 4, pp. 529-545. 38. Putnam, R., ‘Bowling Alone’, 1995, Journal of Democracy, vol. 6, no. 1, pp. 65-78. 39. Putnam, R., ‘E Pluribus Unum: Diversity and Community in the Twenty-First Century’, 2007, Scandinavian Political Studies, vol. 30, no. 2, pp. 137-174. 214 40. Putnam, R., Making Democracy Work. Princeton, NJ: Princeton University Press, 1993. 41. Rosenbaum, W.A., Political Culture, New York: Praeger, 1975. 42. Scarritt, J.R. and Mozaffar, S., ‘The Specification of Ethnic Cleavages and Ethnopolitical Groups for the Analysis of Democratic Competition in Contemporary Africa’, 1999, Nationalism and Ethnic Politics, vol. 5, no. 1, pp. 82-117. 43. Segura, G.M. and Alves Rodrigues, H., ‘Comparative Ethnic Politics in the United States: Beyond Black and White’, 2006, Annual Review of Political Science, vol. 9, pp. 375-395. 44. Taylor, C. and Hudson, M.C., The World Handbook of Political and Social Indicators, New Haven, CT: Yale University Press, 1972. Appendix List of the variables used to measure political culture ____________________________________________________________________________ Q1. (This question corresponds to the B1A-O variable from IPP data set) To what extent do you agree with the following items? a. I feel that I am Romanian b. I feel that Romania is my country c. I like the Romanian language d. I like Romania e. It is good to be Romanian f. Generally speaking, I like Romanians g. I am proud of Romania h. I am proud that I am Romanian i. I am proud of what the Romanians accomplished j. Generally speaking, I prefer to get in touch with the Romanian people than with people from other countries k. Generally speaking, I like Romanians more than people from other countries l. I feel that all Romanians are members of a large family which also includes me m. I believe that all Romanians should live in Romania n. Romanians should not mix with other nations o. People of other nationality than Romanian should leave Romania This is an ordinal variable which takes the following values: 1- Disagree; 2 - Rather Disagree; 3 - Rather Agree; 4 - Totally agree; 9 - No response/Don’t know. Q2. (This question corresponds to the B3A-K variable from IPP data set) What is the closest relationship that you would establish with someone who is…? a. Jew? b. Roma c. Hungarian d. Arab 215 e. Chinese f. Black people g. Jehovah’s Witnesses h. Muslim i. Homosexual j. Lesbian k. Citizen of the Republic of Moldova This is an ordinal variable which takes the following values: 1- As a family member; 2 - As a friend; 3 - As a colleague; 4 - To live in my locality; 5 - To live in Romania; 6 - They should not be allowed to live in Romania; 9 - No Response/Don’t know. Q3. (This question corresponds to the C2A-H variable from IPP data set) I will read a list of statements and I would like you to tell me to what extent do you agree with them or not…? a. Roma people should be forced to live separately from the rest of society, because they can not integrate. b. The state should provide more subsidies and assistance to Roma population. c. It is good to have special places for Roma in state universities. d. There should be pubs (bars, restaurants, discos) and shops where the access of Roma to be restricted. e. The inhabitants of localities in Romania should be allowed to decide whether Roma people could live there or not. f. Most of the Roma population breaks the rules. g. The state should take legal action in order to stop the increase of Roma population. h. Roma people shouldn’t be allowed to travel abroad, because they embarrass us and cause troubles. This is an ordinal variable which takes the following values: 1- Disagree; 2 - Rather Disagree; 3 - Rather Agree; 4 - Totally agree; 9 - No response/Don’t know. Q4. (This question corresponds to the C1 variable from IPP data set) Nowadays do you think that, in Romania, minorities have … rights? This is an ordinal variable which takes the following values: 1 - Too many rights; 2 - Too few rights; 3 - Just as needed; 9 - No response/Don’t know. Q5. (This question corresponds to the C5A-G variable from IPP data set) To what extent do you agree with the following statements about Hungarians in Romania? a. Romanian state should provide Hungarian children education in Hungarian b. The interests of Hungarian population in Romania are most often distinctive of other citizens’ interests in our country c. It would be better if the Hungarians went to live in Hungary 216 d. Hungarians in Romania should use Romanian in their relation with public administration, even though, in some areas, their number is higher than the Romanians’ e. The counties where the Hungarians represent majority should be offered increased autonomy f. There are many Hungarians who avoid speaking Romanian, despite knowing it g. Hungarians would never abandon their aspiration to link Transylvania up to Hungary This is an ordinal variable which takes the following values: 1- Disagree; 2 - Rather Disagree; 3 - Rather Agree; 4 - Totally agree; 9 - No response/Don’t know. Q6. (This question corresponds to the D1A-J variable from IPP data set) I will read a list of statements and I would like you to tell me to what extent do you agree with them or not…? a. Things would go better if everybody would always listen to the authorities’ orders b. Keeping public order is more important than individual liberties c. The man should decide in a family d. It is natural that everyone chooses its own style life, even though it differs from majority’s e. Children should be punished violently (be hit) in order to learn how to behave f. Books and ideas that distrust state authority should be forbidden g. Obeying and respecting authorities are the most important things children should learn h. Romania needs a powerful leader to organize the country i. Those who do not pay respect to the hymn, flag or national heroes should be punished j. The use of violence is often more efficient that long discussions This is an ordinal variable which takes the following values: 1- Disagree; 2 - Rather Disagree; 3 - Rather Agree; 4 - Totally agree; 9 - No response/Don’t know. Q7. (This question corresponds to the D2A-B variable from IPP data set) How well you think that would be for Romania...? a. Have a military leadership b. Have a democratic political system This is an ordinal variable which takes the following values: 1- Disagree; 2 - Rather Disagree; 3 - Rather Agree; 4 - Totally agree; 9 - No response/Don’t know. Q8. (This question corresponds to the D3A-B question from IPP data set) From your point of view...? a. It is better to have only one political party b. It is good to have a market economy 217 This is an ordinal variable which takes the following values: 1- Disagree; 2 - Rather Disagree; 3 - Rather Agree; 4 - Totally agree; 9 - No response/Don’t know. Table 4: Values of the ethnic, linguistic and religious fractionalization indices and at the regional level Region Ethnic fractionalization index Linguistic fractionalization index Religious fractionalization index Crişana-Maramureş Mean .298542 .261006 .457765 N 133 133 133 Std. Deviation .1791549 .1624101 .1688854 Banat Mean .228128 .198654 .334767 N 70 70 70 Std. Deviation .0880319 .0773426 .1269062 Oltenia Mean .051592 .041330 .016627 N 161 161 161 Std. Deviation .0513686 .0500022 .0119421 Muntenia Mean .035053 .019908 .028926 N 296 296 296 Std. Deviation .0487527 .0456502 .0196585 Dobrogea Mean .099092 .071412 .102997 N 67 67 67 Std. Deviation .0636889 .0501168 .0544408 Transilvania Mean .209599 .182289 .309152 N 295 295 295 Std. Deviation .1651049 .1642004 .1950130 Moldova Mean .044915 .029660 .113279 N 324 324 324 Std. Deviation .1057274 .0878739 .1219643 Bucureşti Mean .054089 .027723 .070666 N 154 154 154 Std. Deviation .0133253 .0072922 .0171337 Total Mean .110473 .089071 .160827 N 1500 1500 1500 Std. Deviation .1426873 .1326257 .1867526 218 Table 5: Correlations between the second variable of political culture and the ethnolinguistic indices REGR factor score 1 for analysis 1 REGR factor score 2 for analysis 1 Ethnic fractionalization index Linguistic fractionalization index REGR factor score 1 for analysis 1 Pearson Correlation 1 .023 -.166(**) -.157(**) Sig. (2-tailed) .443 .000 .000 N 1147 1147 1147 1147 REGR factor score 2 for analysis 1 Pearson Correlation .023 1 .096(**) .108(**) Sig. (2-tailed) .443 .001 .000 N 1147 1147 1147 1147 Ethnic fractionalization index Pearson Correlation -.166(**) .096(**) 1 .987(**) Sig. (2-tailed) .000 .001 .000 N 1147 1147 1500 1500 Linguistic fractionalization index Pearson Correlation -.157(**) .108(**) .987(**) 1 Sig. (2-tailed) .000 .000 .000 N 1147 1147 1500 1500 ** Correlation is significant at the 0.01 level (2-tailed). Table 6: Correlations between the fourth variable of political culture and the ethnolinguistic indices14 Ethnic fractionalization index Linguistic fractionalization index Rejection of the Roma people Integration of the Roma people Ethnic fractionalization index Pearson Correlation 1 .987(**) .020 .018 Sig. (2-tailed) .000 .519 .551 N 1500 1500 1064 1064 Linguistic fractionalization index Pearson Correlation .987(**) 1 .021 .028 Sig. (2-tailed) .000 .491 .356 N 1500 1500 1064 1064 Rejection of the Roma people Pearson Correlation .020 .021 1 .000 Sig. (2-tailed) .519 .491 .994 N 1064 1064 1064 1064 Integration of the Roma people Pearson Correlation .018 .028 .000 1 Sig. (2-tailed) .551 .356 .994 N 1064 1064 1064 1064 ** Correlation is significant at the 0.01 level (2-tailed). 14 Two regression factor scores were constructed based on the 8-item question we had from the IPP data set. One factor was referring to the rejection of the Roma people and the other one 219 Table 7: Correlations between two of the fifth variable items and the ethic fractionalization index Ethnic fractionalization index Romanian state should provide Hungarian children education in Hungarian The counties where the Hungarians represent majority should be offered increased autonomy Spearman’s rho Ethnic fractionalization index Correlation Coefficient 1.000 .131(**) .152(**) Sig. (2- tailed) . .000 .000 N 1500 1368 1323 Romanian state should provide Hungarian children education in Hungarian Correlation Coefficient .131(**) 1.000 .318(**) Sig. (2- tailed) .000 . .000 N 1368 1368 1280 The counties where the Hungarians represent majority should be offered increased autonomy Correlation Coefficient .152(**) .318(**) 1.000 Sig. (2- tailed) .000 .000 . N 1323 1280 1323 ** Correlation is significant at the 0.01 level (2-tailed). to the integration of the Roma population and together explained 41.467 of the common variance. The KMO coefficient for this factor analysis was 0.846.