Identity and social determinants of perceiving ethnic discrimination of Hungarians from Romania

This paper analyses the subjective ethnic discrimination in case of Transylvanian Hungarians from Romania. We would like to know to what extent the subjective perception about discrimination is determined from a social, national identity and ethnopolitical point of view, based on a representative survey data from 2007 and 2010. During our research, we have pointed out what are the social factors that determine the perception of subjective discrimination. We analyzed the extent to which social and demographic factors, as well as the character of the ethno-national identity and the character of the regional ethnocultural environment probable alter the perception of subjective ethnic discrimination among the members of an ethnonational minority, such as Hungarians from Romania. According to our hypothesis, the perception of ethnic discrimination is primarily determined by the ethno-national minority identity loaded with nationalistic or ethnocentric elements and by the character of relationship between majority and minority, while social-demographic determination or even educational level or knowledge of other languages are less important. A logistic regression model was involved in the analysis in order to explain the social determinants of perceiving ethnic discrimination, in three steps, including independent variables the socialdemographic ones, as well as the variables about education, school socialization and ethnolinguistic environment (variables of regional belonging). Thirdly, the model incorporates elements of national identity variables and the attitudes towards Romanians.


INTRODUCTION
It is not unusual for persons belonging to national minorities to be discriminated in certain situations.In the last few decades, several countries have made significant political efforts to eliminate or at least reduce any kind of discrimination affecting persons belonging to national minorities.They have launched several educational and information programs in this field and they have also adopted legal norms to regulate situations in which discrimination is prohibited.
This paper relies on the concept of ethnic discrimination and we shall analyze to what extent its subjective perception is determined among Hungarians from Romania from a social, national identity and ethnopolitical point of view after Romania's accession to the European Union, based on 2007-2010 data collections.During our research, we have pointed out that the subjective perception of ethnic discrimination is significantly connected to national identity filled with nationalism and to the perception of the minority-majority relationship.Therefore we may indirectly identify a consequence of the social functioning of national ideologies.
Hungarians from Transylvania became a minority in Romania in 1918 when, following the disintegration of the Austrian-Hungarian empire, Transylvania was unified with Romania.In our days, the number of Hungarians from Romania is approximately 1,23 million1 and they represent 19% of the population from Transylvania and 6,5% of the population from Romania, while 99% of them live in Transylvania.From the point of view of the aim of our study it is important to see the position of Hungarians from Transylvania from a territorial and ethnopolitical point of view.The proportion and territorial structure of Hungarians is strongly different from a regional point of view.As shown in Table 1, Transylvanian Hungarians live in local majority in the Eastern Transylvania (historically known as Szeklerland, now are Covasna and Harghita, and mostly, Mures counties).In Northern Transylvania they live in smaller insular blocks, their majority live in towns and in certain towns the number of Hungarian communities is significant (20-40%).Hungarians live in minority in Western area, Crisana (Patium), the border region with Hungary (Bihor, Satu Mare, Salaj), but their proportion is significant: it varies between 20-40% county level, and their majority live along the Hungarian-Romanian border.In Southern Transylvania and in Banat, the Hungarians live in diaspora, therefore their proportion is below 10% in each county and their majority live in towns (in which their proportion is, in most cases, below 10%).
Table 1 The proportion of Ethnic Hungarians in different regions of Transylvania, 2011(Source: Romanian Census from 2011. INS, 2013)

THEORETICAL CONSIDERATIONS: ETHNICITY, NATIONALISM AND CULTURAL DISCRIMINATION
The basis of perceiving one's minority situation in a negative manner is represented by experiencing negative discrimination or its perception.In modern thinking, treating people with equal rights in an unequal manner is considered discrimination.(SCHNAPPER 1999. 138.)Discrimination can be defined as "those actions designed to maintain own-group characteristics and favored position at the expense of the comparison group" (JONES 1972) Although individual prejudice and stereotypes may produce actions, such as political support for laws and public policies that lead to institutional discrimination, which can operate separate from individual discrimination.The institutional discrimination can be supported by laws and policies.Cultural discrimination appears when one group exerts the power to define values for the entire society, privileging their own culture, traditions and values.In this case the dominant group, imposes this culture on other minority groups.This kind of discrimination is deeply embedded in the fiber of a culture's history, customs and norms; it involves not only privileging the culture and values of the dominant group, but also imposing this culture on other non-dominant groups.As a consequence, everyday activities implicitly communicate group-based bias, passing it to new generations.Thus, members of a minority group may develop a 'false consciousness' in which they not only comply with but also endorse cultural values that systematically disadvantage them.(DOVIDIO et. al. 2010. 8-11.)The contemporary legal concept of discrimination may be interpreted in the context of the modern normative order.Differentiations that refer to the origins or to a biological characteristic of a person or which are based on one of these may be considered unjust and therefore sanctioned based on this normative order.However, discrimination does not only consist of acts, but it is also their qualification.Thus, it is also important to know what is considered discriminative and on what basis.(HORVÁTH 2006. 135., 140.).
In their analysis, BRUBAKER et al. (2006) have shown, based on a research in Transylvanian city, Cluj-Napoca, that ethnic and national consciousness and the ethnicization of space and time is more present in the everyday life of the Hungarian population than among the majority of Romanians who form the majority population6.This type of worldview may increase the subjective perception of ethnic discrimination as minority neurosis, which mean that the actions not necessarily qualify as discrimination in legal terms, it is more likely an oversensitivity in interethnic interactions (see MAGYARI 2000).We also analyze the perception of discrimination related to national identity, ethnocentrism, and the Romanian -Hungarian interethnic relations in Romania.
The national identity is a multidimensional construct with a categorization dimension, that is, "patriotism," and a discrimination dimension, that is, "nationalism."Patriotism is simply defined as a positive emotional attachment to one's own nation-state, whereas nationalism means discrimination against others.Theoretically, national identity may have two main forms, ethnocultural or civic version.This depends on how and based on what principles the nation to which people express, their belonging has been formed (GELLNER 1987).The national identity of the Hungarian minority from Romania is defined mostly ethnoculturally, as demonstrated in several studies, but certain elements of civic identification with Romania, with the Romanian citizenship can also be identified in the collective identity of Transylvanian Hungarians. 2n VERDERY'S (1993. 179-203) view "nationalism is using the symbolic content of nation in an emotionally exalted form for political reasons".Although nationalism is not induced through "nation", but through particular manifestations of the political sphere (BRUBAKER 1996. 13-17.).It may be assumed that influenced by nationalism, the members of an ethnocultural national minority tend to perceive the relationship with the majority as characterised by conflict, especially on a national level and their attitude is more repulsive than that of the members of the majority ethnocultural national community.These attitudes may influence the way in which different everyday situations are perceived, thus the perception of discrimination as well.Naturally, the dynamics of intergroup relationships may not be viewed in an idealistic way.In CALHOUN'S (2003. 531-553.)view, "culture and social relationships are as real as individuals, even if they lack bodies", criticizing the "extreme cosmopolitanism" which promotes the elimination of all loyalties lesser than that of each individual to humanity as a whole.
Inasmuch as the ethno-national identity of minorities is influenced by the phenomenon of affectedness, these people are oversensitive to national belonging, while they are less sensitive to other things (CSEPELI 1997).Ethnic minority neurosis may also be found behind this phenomenon which refers to the oversensitiveness induced exactly by former cultural discrimination, not necessarily experienced personally, but oriented against the community to which the person belongs to.In last decades the Transylvanian Hungarians experienced nationalist political actions which could imply oversensitivity to national belonging.Starting from 1956, and more likely after 1965, the state politics gradually ceased previous linguistic and educational rights (SCHÖPFLIN 1991), with hidden aim to assimilate the Transylvanian Hungarians, on cultural and linguistic sense (BUGAJSKI 1995. 200.).Even in first decade of political transition after 1989, Romania "underwent problematic democratic transitions at the beginning of the 1990s, marked by continued patterns of ethnic minority discrimination, security fears, nationalist rhetoric, and the presence of leaders with extremist views" (MIHAILESCU 1999).
We may speak of subjective discrimination in the case of our respondents, which means that actors who do not have legal or symbolic legitimacy qualify an attitude or act as discriminatory as opposed to qualified discrimination which is established by a socially acknowledged authority, for ex. the National Council for Combating Discrimination in Romania (HORVÁTH 2006. 143.).
The research question of this paper is: what are the social factors that determine the perception of subjective discrimination?Searching for an answer, we analyze the extent to which social and demographic factors, as well as the character of the ethno-national identity and the character of the ethnocultural environment ender probable the perception of subjective ethnic discrimination among the members of an ethnonational minority, such as Hungarians from Romania.
Our hypothesis is that the perception of ethnic discrimination is primarily determined by the ethno-national minority identity loaded with nationalistic or ethnocentric elements and by the character of relationship between majority and minority, while social-demographic determination or even educational level or knowledge of other languages are less important.

DATA AND METHODS
The empirical data source of the research is represented by the Charpatian Panel project, waves 2010.In some cases, we will refer to data from wawe 2007. 3e carried out the first wave of the Kárpát Panel research from Romania, in Transylvanian counties, in 2007 by means of a representative survey with 900 subjects, employing a questionnaire based method on a Hungarian speaking population sample, using an aleatory multistadial survey method.We took samples from 15 counties of Transylvania in a broader sense (Caras-Severin county was left out due to the small number of Hungarians living there).In this phase, a questionnaire based survey among Hungarians from Slovakia, Serbia and Ukraine, as well a Hungarian control sample survey were carried out, on a total aleatory sample of 2915 cases.The second wave was carried out only in Transylvania during 2010, on a sample consisting of 890 cases.The sampling method used was the following: the population from the 2007 representative sample was contacted again and those who could not be contacted anymore were replaced by using the original supplementary sampling lists (from 2007).
Respondents were selected if they spoke Hungarian.The questionnaire comprised questions on national/ethnic self-identification as well.Respondents were asked to indicate, in order, the communities they felt they belonged to, and we processed the first identifications mentioned.(VERES 2010) We used the following question in order to measure the subjective perception of discrimination: How often were you discriminated against for the following reasons?The answers were recorded on the following scale (1 = never, 2 = rarely, 3 = frequently, 4 = very frequently).We considered that the measuring level of the scale was ordinal.The respondents could provide different types of discriminatory experiences based on ethnic origin, sex, age, material situation, social origin, political views or region of origin.
Secondly, we analysed the associations with different social-demographic variables (sex, age, type of locality, educational level, region or residence) by employing the Chi-square test.
Next, we analysed the indicators used to verify the nature of ethnic and national identity, as well as of the ethnocultural environment.By means of the Chi-square, t test we analysed the bivariate correlations that these variables had with social-demographic variables.
First, we analysed the ethnonym they primarily designated by addressing the following open question: What do you primarily consider yourself?, then we addressed the next open question: What is homeland for you?The respondents could answer anything they wanted and the answers were grouped at a later moment.
For measuring national attitudes toward the ethnonational majority (Romanians), we used a five-level scale of attitudes of like/dislike towards specific groups to capture the distance between the Hungarian minority and other social groups.The question was: How do you see the following groups?
We operationalised general and auto heterostereotype for measuring prejudice towards majority Romanians.The respondents had to estimate to what extent each reference group (one's own minority group, Romanians as majority nation, Hungarians from Hungary) could be characterised by four previously provided positive and and four previously provided negative traits.These percentual answers were used to elaborate two aggregate indicators: one by calculating the mean value of the positive characteristics and one by calculating the mean value of negative characteristics.Then, we analysed their mean and variance values by means of the t test.
The perception of interethnic relations on local or country level (i.e.perceived interethnic conflict or cooperation) were measured by the question: What is the relationship between the Hungarian minority and the Romanian majority like at the country/at local level?The possible answers were the following: 1 = conflict, 2 = neutral, 3 = cooperation.For the purpose of multivariate analysis, we have combined these three categories into a dummy variable, where conflict = 1, else = 0.
We also used other variables in the logistic model but in this paper we did not present the answers to these in detail.They were used to measure the nature of the ethnocultural environment.One was used to measure the degree of knowledge of the official language (Romanian) on a six degree scale.The other question referred to the language used in relationships with public administrative authorities (during office routine) in the locality of residence.Romanian or Hungarian could be opted here.The third referred to the language of general (mandatory) education.Again, Romanian or Hungarian could be opted as languages of education.
Last, the analysis involved a logistic regression model in three steps.
The dependent variable was a dummy variable created on the basis of the following question: How often were you discriminated because of your ethnonational belonging?The value of the Yes, I was (frequently) discriminated and the value of the Rarely answers was 1, while the value of the Never answer was 0.
The model was elaborated in three steps.First, the social-demographic variables, as well as the variables strongly related to them, i.e. school socialisation and language environment were the independent variables.
Then, the model was extended and the variables of regional belonging were included.Thirdly, we included the elements of national identity and the attitudes towards Romanians in the model, as well as a general variable measuring satisfaction with one's life (but the influence of this latter one was not significant).The variables included in the second and third step also served as control variables to measure the influence of variables included in the previous step.
The independent variables included in logistic regression model were the following: age, education (the number of years of study completed) dummies for gender (women = 1), settlement type (urban = 1), Hungarian language of instruction in secondary school (Hungarian = 1, else = 0).Multivalue variables: the language use in relation of local administration (office routine) (1 = Hungarian, 2 = Romanian, reference value: in both languages), Regional (4 values, reference: Partium) Identity: you primarily consider yourself as (reference value: Transylvanian), Homeland (reference value Romania, knowledge of the Romanian language (1 to 6-grade scale), perception of interethnic relationships with the Romanian majority population at country level (conflict = 1), interethnic relationships at the local level (conflict = 1), attitudes towards Romanians (5-grade scale, see description above), negative stereotypes about Romanians, positive stereotypes about Romanians (values on a scale between 1-100), satisfaction with one's standard of living (1 = very unsatisfied, 6 = very satisfied)

RESULTS
In 2007, 57% of the Transylvanian Hungarian respondents (59% in 2010) responded that they were never discriminated on grounds of ethnic belonging.34% of the respondents were rarely discriminated on grounds of ethnonationality both in 2007 and in 2010.While 6% of the respondents said that he were discriminated frequently and 1,7% very frequently in 2007, their percentage toned down to 4,2 and 1,6% respectively during 2010.However, these differences are within the margin of error, therefore we may assert that there is no significant change in the degree of the perception of ethnic discrimination in time.Comparing these answers of ethnic Hungarians from Romania with the Romanian data on national level regarding discrimination, according to a national survey from year 2004, a similar percent of 6-7% of Romanians consider that in Romania, ethnic Hungarians are discriminated frequently or very frequently, in different situations, because of their ethno-nationality. 4In a recent national survey from 2013 commanded by National Council against Discrimination from Romania 22% of respondents said that they were ever discriminated (on any reason) in Romania, while 12% of Romanians said that the persons belonging to ethnonational minorities are discriminated very frequently and other 20% said frequently.In case of getting a new job, 30% of the Romanian respondents considered that an ethnic Hungarian can find a job more difficult than an ethnic Romanian.Also, 52% of Romanians don't accept a Hungarian in family (as relative) and 47% of them don't accept a Hungarian as a friend. 5n comparison of other types of discrimination, the Transylvanian Hungarian respondents consider that discrimination on grounds of ethnonationality is the most frequent type of discrimination.A significant part of the respondents were rarely discriminated on grounds of their material situation, thus for poverty and their region of origin.This tendency has increased during the last 3 years from 17% to 21% in the case of the material situation and from 11% to 13,6% in the case of the region of origin.Yet, the number of those mentioning discrimination on grounds of sex has decreased from 7% to 3% (see Table 2 and see also PAPP -VERES 2007).This may also be due to the fact that our respondents are oversensitive to national belonging and less sensitive to other issues.It is also possible that minority neurosis is to be found in the background of all this.Minority neurosis refers exactly to the oversensitiveness entailed by one kind of discrimination.We may primarily speak of subjective discrimination in the case of our respondents, which means the qualification of an attitude or action as discriminatory by actors who do not have the legal or symbolic legitimacy as opposed to qualified discrimination which is established by a socially recognised authority, for ex. the National Council for Combating Discrimination of Romania.6 (see Table 2) There are significant association both according to sex and age groups: a higher percentage of men, people from towns/cities and younger people experienced discrimination -although rarelyas compared to other categories.A higher percentage of the latter categories considered that they were not discriminated (see Table 3).Differences related to the level of education could be observed among those who were discriminated on grounds of their ethnonationality.As also shown in Table 4, in 2010, 17% more higher education graduates than basic education graduates mentioned that they were discriminated on grounds of ethnonationality, even if rarely (the difference was 15% in 2007, see Table 4).
The biggest differences may be observed according to region: 52,5% of the respondents from Northern Transylvania (where the Hungarian diaspora and insular blocks of Hungarians live) and almost 40% of the respondents from Southern Transylvania and Banat answered that they were rarely discriminated on grounds of nationality, while their percentage is 33,5% in the Partium region and 27,8% in Eastern Transylvania.Almost 6% of the Hungarians from Northern Transylvania experienced discrimination frequently.This percentage is smaller in other parts (see Table 5).
In the following, we shall briefly present the values of those national identity and attitude variables based on which we will try to explain the evaluation of ethnic discrimination in the logistic model.
We used several indicators for identity variables.One refers to what they primarily consider themselves.Based on their answers, a strengthening of regionality may be observed by 2010: while 16,6% of our respondents considered themselves Transylvanian without an ethnonym in 2007, the percentage of those who declared themselves Transylvanian was 29,6 in 2010.This is followed by the option with a regional attribute, Transylvanian Hungarian with 27% in 2007 and already 24,4% by 2010.Thus, the two regional "Transylvanian" options make up the majority of the answers.
Another 24,4% of the respondents declared themselves Hungarian without a regional attribute in 2010.Almost 22% of the respondents from Transylvania declared themselves to be something else.
The most characteristic is Szekler: 9% of the respondents declared themselves Szekler in 2010, while their percentage was 7% in 2007, which also reflects the strengthening of regionality.
The other question was used to analyse what people considered as their homeland.In 2010, 40% of the respondents considered Romania as their homeland and 45,5% (as compared to 46,6% in 2007) indicated Transylvania as their homeland (as compared to 31% in 2007).
Hungary was only indicated by 3% as their homeland and there were several other distributed answers (more often: locality of residence, locality of birth, other answers).We used a 5-level scale for measuring attitudes towards ethnic-national groups, where 1 denoted very likeable and 5 very antipathetic.The question referred to the respondents' attitude towards different ethnic groups, to what extent they are likeable: in 2010 the average value of Hungarians from Hungary was 2,3, while that of Romanians was 2,4.There is only a small, not significant difference between these two values.Average values are 0,1 smaller than for the group analysed in 2007.Looking at the distribution of answers, it may be seen that the attitude towards Romanians is rather indifferent, while somewhat more respondents feel that Hungarians from Hungary are more likeable or antipathetic than Romanians.The majority of the respondents feel antipathy towards Roma and the average value is considerably higher than for other groups: 3,1 in 2010 (3,2 in 2007).We have also analysed the averages for the answers according to locality types and we have noticed that there are significant differences in the perception of Hungarians from Hungary: urban people feel more antipathy towards them than village people.The answers indicate a new situation as compared to research carried out in previous years.While in 1997 and 2000, for example, Germans were the most accepted ethnic group, by 2007 and 2010 this had changed: Hungarians from Transylvania also consider Romanians to be closer than Germans (see Figure 1).
While measuring preconceptions, our respondents had to estimate to what extent the analysed groups (one's own minority group, Hungarians from Hungary, Romanians as the majority nation) could be characterised as described above (Method) by four positive and four negative traits.The answers show how minority ethnocentrism functions: positive traits are mainly attributed to Hungarians from Transylvania, they are followed by Hungarians and then by Romanians.The hierarchy is reversed in the case of negative traits: negative traits are attributed approximately to the same extent to Hungarians than to Romanians and to a smaller extent to their own minority group.In the case of positive traits the biggest differences may be noticed for "helpfulness" in 2007: the respondents consider that 56% of their own group, 45% of the Romanians and only 40% of the Hungarians are helpful on average.In 2010, the difference decreased for both Hungarians (45%) and Romanians (48%), while the value for Hungarians from Transylvania did not change.As for the other traits, the changes are not significant, except that, in 2010, 3% less respondents consider that Hungarians are selfish than in 2007.No significant correlations between the different characteristics attributed to Hungarians or Romanians and the main socialdemographic variables could be revealed (VERES -PAPP 2012).
We used the answers to each trait to elaborate aggregate indicators taking into consideration if they were positive or negative traits and we analysed the average and deviation values of the new variables for each region in relation to Romanians and Hungarians from Hungary (see the annex for the regional averages calculated according to each trait).43% of the entire analysed population attribute negative traits to Romanians and 40,3% to Hungarians from Hungary, which also shows the extent of the difference between the relationship with these two groups.Regionally, the results clearly outline that the average values for the negative traits that characterise Romanians are higher in Eastern7 and Northern Transylvania (around 45%) and they are lower in the Western area (Crisana-Partium)8 region, in Southern Transylvania and Banat (around 39%).Hungarians from  Hungary are attributed negative traits to the greatest extent in Eastern Transylvania (43%) and this decreases in correlation with the proportion of Hungarians (it is only 32,7% in Southern Transylvania and Banat).Positive traits do not significantly differ for Romanians.There are also small differences for Hungarians from Hungary.The average value is the highest in Crisana (50%): they are the ones who consider them the closest as we have already seen it at the group definition or at the concept of homeland, while it is the smallest in Northern Transylvania (45%).
We have also compared the answers related to discrimination with the answers to certain aspects of the ethno-cultural national identity, thus with self-identification, with the answers to the concept of homeland and with the attitude towards and preconceptions about Romanians.We have found significant correlations with the perception of ethnic discrimination and the attitude towards the majority both in 2007 and 2010: on average, those who assert that they have never been affected by ethno-national discrimination assume closer relationships with the majority population than those who assert that they have been affected by discrimination, but we have also carried out a more complex analysis of this by means of a multivariate logistic regression model (see below, Table 8).
In order to evaluate the minority situation, we inquired how they perceive the relationship between the Hungarian minority and the Romanian majority.In many cases, ethnic groups in contact have a different representation of the commonly used geographical space and common past in their collective consciousness, which may give birth to conflicts (SMITH 1991).
There are significant differences between the evaluation of the country-level relationships shaped by power ideologies and the intergroup relationships directly perceived by our respondents on local level, which may also be attributed to the conflictive Hungarian-Romanian relationships during the course of history.
While in 2010, 61% considered that locality level Hungarian-Romanian relationships from Transylvania were characterized by cooperation (as compared to 68% in 2007), 32-33% of Hungarians from Transylvania considered that country-level Hungarian-Romanian relationships were characterized by cooperation.However, a significant change may be noticed here: while in 2007, 33.5% considered that Hungarian-Romanian relationships were conflictive, their proportion toned down to 26% by 2010.Therefore, in 2010, more respondents, i.e. 35.9% viewed country-level Hungarian-Romanian relationships with indifference.
The perception of conflict is significantly different according to regions.Country-level Hungarian-Romanian relationship are perceived as conflictive by Hungarians living in blocks and Hungarians living in significant proportion in Eastern Transylvania and in Western Area (Partium) (31%), while in Northern Transylvania only 24,4%, in Southern Transylvania and Banat 15,5% perceive that there is a conflict.Accordingly, the majority of diaspora Hungarians consider that this relationship is characterized by cooperation, while in the other two regions not even 30% consider it as such!However, the situation is different on locality level: an above average proportion of Hungarians from Northern Transylvania perceive that there is a conflict (more than 17% as compared to the average 8,3%).80,4% of Hungarians living in Southern Transylvania and Banat, almost 70% of Hungarians from Western Area (Partium) consider that Hungarian-Romanian relationships are characterized by cooperation at the level of their respective locality, while in the other two regions only 55% consider that this is the case (Table 7).
Rather surprisingly, there are no significant correlations between the perception of Hungarian-Romanian relationships and different social-demographic variables, neither at country level nor at the level of localities, except a weak correlation with the type of locality (a few percent more consider that there is a conflict in villages than in towns/cities).

THE MULTIVARIATE MODEL OF THE PERCEPTION OF ETHNIC DISCRIMINATION
In the following, we analysed the variables that could explain those who answered that they had been discriminated on grounds of their nationality in Romania.The analysis was carried out by means of logistic regression.The dependent variable was a double-value dummy variable which was 1 if the person in question had ever experienced subjective ethnic discrimination and 0 if never.The model has three levels.First, we analysed the social-demographic variables, as well as the influence of the closely related school socialisation and language environment.Secondly, we extended the model by including the variables of regional belonging and, thirdly, we included the elements of national identity and the variables related to attitudes towards/ relationships with Romanians, as well as a general satisfaction variable.The variables included during the second and third step also serve as control variables for measuring the effects of the variable included in the previous step.
The explaining power of the model is 0.052 on the first level based on the Nagelkerke R-square, 0.101 on the second level and it has increased to 0.244 on the third level by including the elements of national identity and the variable measuring the attitude towards Romanians, which counts as considerable based on the peculiarities of this method.Based on odds ratios Exp(B),9 of the variables included in the first round, living in an urban environment, age (0,988) and the nature of language use in administration (1,5) display significant roles in the increase of the chances of discrimination, according to the following: living in urban areas significantly increases the chances of discrimination by an odds ratio of 1,669.To a small degree, a younger age also increases the odds of discrimination (0,988), while the chances of being discriminated on grounds of nationality are 1,58 higher for those who mainly use Hungarian in their relationships with administrative authorities (office routine) than for those who use both languages and 1,3 higher than for those who use only Romanian.Sex, educational level, the language of the last school graduated or the level of proficiency in Romanian do not have a significant influence on discrimination based on ethnic origin [we omitted this latter one from the extended version of the model because the value of Exp(B) was 1,023].
If we include regional belonging into the model, the situation slightly changes.The value of Northern Transylvania will have the most significant influence from the point of view of the regional variable with an odds ratio of 2.22, while Eastern Transylvania has an odds ratio of 0,56.Thus, the chances of ethnic discrimination are significantly increased by residence in Northern Transylvania, while residence in Eastern Transylvania decreases the chances of ethnic discrimination.Of the previously analysed variables, the influence of age and locality of residence have barely changed.On the contrary, in the case of administrative language use the influence of the use of Hungarian has become even stronger (2.25), while it has practically disappeared in the case of Romanian (1.09).Thus, if a respondent is from Northern Transylvania and he/she lives in a locality where the language of office routine is mainly Hungarian, the chances of ethnic discrimination are higher than for others.Moreover, the chances of people from Northern Transylvania of being ethnically discriminated are on average higher even if they use Romanian in their relationships with administrative authorities.
We may notice smaller changes due to the influence of the control variables measuring identity and attitude introduced during the third step: the influence of urban residence increases (1.868), the influence of the use of Hungarian in office routine remains the same, but it decreases to the level of the first version (1.519), the influence of the Northern Transylvanian residence decreases, but it is still the most significant (1.775), while residence in Eastern Transylvania contributes to an even smaller degree to the chances of being discriminated (0.407).
Of the control variables, the attitude towards Romanians (to what extent they are likeable) has the most significant influence with an odd ratio of 1.633, therefore those who relate to Romanians in a more negative way are more likely to experience discrimination.The homeland variable is another variable with a significant influence: i.e., those who named Transylvania as their homeland (with a value of 1.646) as compared to those who named Romania as their homeland, the influence of those choosing another option is not significant due to the number of cases, although odds ratios are rather high (see Table 8).We have included several other variables into the model, but they do not have a significant influence, except the respondents who use negative stereotypes to characterise Romanians.The experience of ethnic discrimination is more characteristic among them, although the odds ratio is barely above average (1.025), yet the significance level is smaller than 0.001.The results lead to the conclusion that subjective ethnic discrimination is partially explained by the spatial variables that may objectively increase the chances of ethnic discrimination given that the modification of ethnic proportions in Northern Transylvanian towns/cities, the symbolic space loss and the sometimes ethnically charged political fights may otherwise also increase the chances of Hungarians living in these towns/cities to ethnicise their life events as also observed by Brubaker (2006) and other scholars.10Cluj-Napoca.The fact that in some cases mentioning discrimination is the result of minority neurosis increases the chances of subjective discrimination not being explained by weak proficiency in Romanian besides regional and locality variables, but rather by the rejection of Romanians or by certain identity elements.

DISCUSSION
Research on interethnic relations and national identity carried out in Romania shows that in the first two decades following the political and economic regime change from 1989/90, the relationships between the Romanian majority and the Hungarian minority had improved, the perception of interethnic conflict had decreased and the attitudes of minority Hungarians towards the Romanian majority had also improved (see VERES 2010VERES , 2012)).As a result, the social atmosphere obstructing ethnic discrimination had also intensified, but a significant part of the Hungarian population from Romania (almost 40%) subjectively still perceived ethnic discrimination even in 2010, even if most of them rarely.
Based on the interpretation of the multivariate logistic regression model, the chances to perceive ethnic discrimination are increased by the regional location of residence, especially Northern Transylvanian, mainly urban residence and this does not significantly change even if several control variables are introduced.The chances of perceiving national discrimination are higher among those who rather use Hungarian language during their relationships with local administrative authorities (office routine).Yet, the use of Hungarian language in administration is not explained by a low proficiency in Romanian language, because this does not influence the perception of discrimination, but rather by negative preconceptions about Romanians.Thus, those for whom the Romanian majority population is antipathetic are more likely to perceive ethnic discrimination.
Thus, some Hungarian individuals could have had discriminatory experiences for example before 1989 or in the following years, which led to certain disadvantages on the individual level (the absence of professional promotion at work, verbal warning for school performances) due to their Hungarian ethnonationality.Other respondents, on the other hand, experienced discrimination at the collective level.For example, during the 1980-ies, the Hungarian language had been gradually eliminated from the administration, from signs with the name of public institutions and locality name signs, between 1987 and 1989 even the use of Hungarian locality names in Hungarian newspapers/periodicals was prohibited (BUGAJSKI 1995).
As for the peculiarities of the Transylvanian Hungarian identity, it may be asserted that the most important differences in their national attitudes are to be noticed on a regional level.The regional differences in the proportion of Hungarians and the existence or absence, as well as the extent of the system of Hungarian cultural, educational institutions which are related to regional differences, as well as the frequency of everyday interactions with Romanians, the language use in administration are all decisive: a higher proportion of Hungarians living in the Eastern Transylvanian counties (Covasna-Harghita) and along the Romanian-Hungarian border (Partium) consider that their Hungarianness is something natural.In other parts of Transylvania, Hungarians living in diaspora perceive this as a special resource, although -even if rarely -more people have experienced ethnic discrimination in Northern Transylvania than in other regions, where ethnic Hungarians live in a high proportion, or in very low proportion, like in Southern Transylvania and Banat.
Social distances between Hungarians and Romanians are, on average, greater among Hungarians living in a county-level majority.They have less frequent everyday interactions with Romanians than Hungarians living in diaspora or insular communities in Northern Transylvania.Southern Transylvania and Banat are characterised by an even stronger diaspora situation.People living in Hungarian diaspora areas perceive the Romanian-Hungarian relationships more collaborative, not conflicted, both at the country level or the local level as the Transylvanian Hungarians living in a local or county-level majority.

CONCLUSION
We may assume that a certain discriminatory experience from the recent or distant past -which could have happened at the individual or collective level -has contributed to the evolution of the present minority neurosis, measured by the subjective perception of ethnic discrimination.The state nationalism -institutionalized in the Ceausescu's dictatorship, and also in the first years after the 1989's transition period -has been manifested as the cultural discrimination against ethno-national minorities, especially against Transylvanian Hungarians.However, the proportion of ethnic Hungarians, that often (or very often) perceived ethnic discrimination in their lives, is similar to general perception of Romanian population about the frequent perception of discrimination of Ethnic Hungarians from Romania.
The results of the multivariate model showed us that the subjective perception of the ethnic discrimination significantly differentiate according to local ethnopolitical context and the ethnocentrically loaded ethnonational minority identity, in Verdery's sense, and not according to personal or social status characteristics (as age, educational level etc.).The local ethnopolitical context implies the ethnic composition of the population, and the quality and history of interethnic relations.The local experiences on interethnic relations implied the affectedness in minority identity of the Transylvanian Hungarians living in regions as cities in Northern Transylvania, like in Cluj-Napoca, the period of mayor Funar .In this region, the Hungarian-speaking population have a collective memory about a more equilibrated presence of Romanian and Hungarian language and culture in public sphere, especially in the 1940's, 1950's and even 1960's (BRUBAKER et al 2006)).
According to our multivariate analysis the subjective perception of ethnic discrimination, among Transylvanian Hungarians, is explained, next to the local ethno-demographic context reflected by the territorial location of respondents, by the minority ethnocentrism, measured by the negative attitudes toward the Romanians, associated with the manifestation of certain elements of the local Hungarian minority identity.Consequently the Transylvanian Hungarians from the urban Northern Transylvanian areas, tend to perceive the local ethnonational majorityminority relationship as conflicted in a higher degree than in other regions.

FIGURE 2
FIGURE 2 How could you describe country-level relationships between Hungarians and Romanians today?And at the level of the locality you live in? (%) N 2007 = 894, N 2010 = 846 (Source: Charpatian Panel Survey, Transylvanian Hungarians from Romania, 2007 and 2010 [data calculation by the author])

TABLE 6
differences[Anova-test]; m = mean, s = standard deviation; a.Hungarians in this case mean Hungarians from Hungary)