PropertyValue
?:linkContext
  • Participation in civic associations varies among and within countries and within the same country among different social groups. Studies report different rates of immigrants' civic commitment within the same host society depending on their country of origin . Im m igrants' participation in civic associations has been approached mainly through qualitative research methods, but some studies based on quantitative methodology are available as well. underlines the impact of differences in civic and political culture in Surinam and Turkey on Surinamese and Turkish immigrants' civic participation in Amsterdam, while Fennema and noticed a similar effect in the case of Turkish and Antillean migrants in Amsterdam. However, the results are restricted to a specific host city and a limited number of ethnic groups. This article focuses on civic participation of first-generation immigrants in Spain and tries to explain why some immigrants are more active than others. We investigate the effect of the level of civic activism in the home country on civic participation in the host society using quantitative research methods. Using data provided by the Spanish National Immigrant Survey, we employ hierarchical multilevel regression models. The hierarchi cal models allow testing simultaneously for individual-and group-level characteristics that may influence im migrants' civic commitment. Because all immigrants included in the sample live in the same society, we have the opportunity to analyse the 'pure' effect of a country-of-origin participatory culture maintaining constancy in the societal context in the host culture. We define civic participation as involvement in civic associations, and we consider it a specific category of political participation, following . We intend to predict the effect of a home country's civic activism on three different outcome varia bles: participation in all types of civic associations, participation in associations for immigrants and participation in non-immigrant associations. We do not expect different effects of our main predictor on the three dependent variables, but since participation in associations for immigrants has different connections with the control variables, we employed three different analyses for the dependent variables. The results show that immigrants who have lived longer in and are in closer contact w ith highly participatory home countries are more likely to get involved in civic associations in the host country because the immigrants previously acquired civic skills that can be transferred to the new social context. A n alternative explanation may be that they have built a larger social network at home and they retain ties that can be activated when moving to the host society. The first section covers existing literature on im migrants' adaptation to host societies' political cultures as well as on factors influencing immigrants' civic participation. The second part introduces the indicators and the strategy used for analysis, while the third comprises the data analysis. The discussion and conclusions are found in the final section. Three different approaches have been used to explain how immigrants adapt their politi cal beliefs and behaviours to a new political environment: resistance theory, exposure theory and transferability theory. According to the first approach, political attitudes and behaviours are acquired during early socialization and are resistant to change later in life . Consequently, immigrants encounter many difficulties in adapting their political behaviour to the host society because they were socialized in a different context. Time spent in the country of origin is a key variable; younger immigrants adapt more easily to the new environment because they have had little contact with their original society . The last two approaches assume that political attitudes and behaviour are open to change even during adulthood and immigrants can be politically resocialized into the host society. Exposure theory considers the time spent in the host society as a key vari able for resocialization, the likelihood of adaptation increasing with the time spent at the destination . This approach does not pay much attention to the effect of early socialization in the country of origin and considers the immigrant a 'tabula rasa' from the political point of view . The third approach, transferability theory, notes that immigrants acquire political skills and information during early socialization in their country of origin, which pro vides a background for political learning in the host society . Thus, immigrants can transfer political knowledge obtained during early socialization and use it to build new political skills in the new political environment. showed that 'there is no single pathway to political learning: which pathway applies depends upon which orientation, or which political behavior, is under consideration'. They found that the exposure mechanism can better explain voter turnout and political partisanship, while the transfer mechanism provides a reliable explanation of political interest. According to the authors, while voting and political partisanship are more demanding activities requiring judgement and decision-making, political interest is often a byproduct of other common activities involving basic skills that can easily be transferred from one political system to another. Civic participation has some particular characteristics compared with general political participation, and we assume that a different mechanism operates therein. According to , civic participation during adulthood requires three 'participatory facts': motivation, capacity and skills and networks. To participate, indi viduals should be willing to get involved; should have the capabilities to join an associa tion in terms of money, time and skills; and should have connections with others involved in civic activities. In the case of immigrants, the existence of a larger group from the same country of origin seems to facilitate civic participation . Participation in various types of civic associations operates as a mechanism of inte gration into the new society and helps individuals in their relations with the home and host societies . Participation in civic activities brings rewards to immigrants, motivating them to get involved. It can enable the practice of civic and political skills , and improve language skills . Moreover, participation facilitates contacts with the home country and helps immigrants to deal with local institutions . Participation in civic associations requires skills and abilities. According to the soci ological literature, civic skills are acquired early in life, and participation during the formative years and youth increases the likelihood of participation in civic activities during adulthood . Three factors are particularly important for learning the skills required for civic participation: fam ily school and involvement in extracur ricular activities. Parents are usually models for their offspring, and their participation will lead to similar behaviour among the younger generation . A similar effect was reported in the case of neighbourhoods . Schools may boost civic participation through various channels that contribute to the acquisition of fundamental civic skills . Some countries are more successful than others at building civic skills. Citizens from more participative cultures are more likely to transfer civic skills from the old environ ment to the host society when people are exposed to the migration experience and can benefit from civic participation. We assume that a transfer mechanism operates in the case of immigrants' civic participation, i.e. existing skills are used and adapted to the new environment. People who have been involved in civic activities in their home coun try are more likely to participate in the host society. Since the available data do not allow for the effect of previous civic experience to be tested at the individual level, we will focus only on the impact of the level of civic participation in the country of origin on current civic participation. As we have shown, civic participation early in life increases the probability of civic involvement during adulthood . Several studies note the existence of institutional settings that can boost or reduce the chances of civic participation during adult life . The workplace is such a setting. Employed people have more opportunities to practise civic skills compared w ith those outside the labour market. Almond and Verba (1963) find significant intercountry differences in employees' involvement in decision making, which boosts civic participation. Some countries are more successful than oth ers in extending civic-skills learning during adult life, as well. Citizens from societies that encourage training in civic skills during adulthood are more likely to become involved in civic activities. The longer these citizens live in such a social context, the more likely they will have many civic skills that can be transferred and adapted when they migrate to a new country. We assume that a combination of trans fer and resistance mechanisms is likely to occur, which stimulates civic participation by immigrants coming from more civically active societies. Thus, immigrants who lived for a long time in a home country with higher levels of civic involvement and who maintain close contacts w ith it are more likely to participate in the host society because they have better skills for such activities. Studies dedicated to civic participation identify various individual and contextual fac tors influencing people's propensity to engage in civic activities. At the individual level, studies have shown that characteristics such as education, employment status, income, family and gender have a significant impact on civic participation . Employment also has a positive effect on civic participation, because employment integrates people in various social networks and leads to civic involvement and participation, while gender induces differences in the level and type of civic participation due to the different involvement of women and men in family, employment and religion, women being more active compared to men. Moreover, people with higher incomes and more education are more civically active because they have more resources to share, while marriage and parenthood have positive effects. Parenthood boosts civic participation because parents have a larger social net work by meeting other parents, while marriage with an active person increases the level of both spouses' civic activism. Other studies have focused on the national context. found that economic development (measured by GDP/capita) has a positive impact on mem bership in civic associations but no effect on volunteering, whereas the level of democ racy (measured by the Freedom House Gastil Index) has a positive impact on volunteering but no significant impact on membership. Generally speaking, social wealth is associated with more resources required and a higher level of education, both increasing associational membership ). Other individual and contextual factors directly influence immigrants' participation in civic associations. Among individual factors is the length of stay in the host society, which stands for the exposure mechanism , and language skills, which help communication and integration into the host society . The size of the immigrant group from the same country of origin exerts a contex tual influence on the likelihood of immigrant participation; people belonging to larger groups are more likely to be active because they have an extended social network . Starting from the theoretical approaches referred to above and controlling for all the relevant predictors, we propose two hypotheses: H I: Immigrants coming from countries with higher levels of membership in civic associations are more likely to get involved in civic associations in the host countries. H2: Immigrants who have lived longer in and stayed in closer contact with a home country with a higher level of membership in civic associations are more likely to partici pate in civic associations. The analysis is based on the data set from the Spanish National Immigrant Survey (ENI) carried out in 2007 by the Spanish National Institute of Statistics on a representative sample of the immigrant population living in Spain.1 The questionnaire provides infor mation about respondents' participation in various associations as well as information regarding social background and migration history. The data set offers information about 15,465 immigrants from 94 countries living in Spain. From this sample, we selected a subsample of 8575 immigrants from 34 countries of origin. We included only immi grants who came directly from their country of origin and did not report previous migra tion to other countries. We did so because we assumed the effect of intermediary migration would be difficult to interpret. We selected only people who had resided in their home country at least until the age of 10 because we wanted to capture the effect of early socialization, and migrants who leave their home country early do not have enough opportunity to interact with societal institutions influencing civic participation. We have excluded immigrants from countries of origin with fewer than 25 cases in the ENI database. The 34 countries of origin included in the analysis are shown in . The total number of cases by country of origin ranges from 25 for Norway to 1850 for Morocco, with an average of 252 cases. The macro-level indicators used in the analyses come from the 2005 World Values Survey and the 2000 European Values Survey/World Values Survey (EVS/WVS) for the level of civic activism in the country of origin, from the 2008 KILM for the GDP (data reported for 2005) and from the 2007 ENI for the total number of immigrants by country of origin ( our own computations).2 We produced several m ultilevel regression m odels using the HLM 6.04 program for hierarchical linear m odelling. We estim ated hierarchical logistic regression models, because our dependent variables are dichotomous ones, predicting individuals' prob ability o f participating in civic associations. The m ultilevel regression allows testing of the com bined effect of individual independent variables and contextual-level variables. The ENI provides separate information about immigrants' participation in associa tions dedicated to foreigners and associations not directly dedicated to foreigners. We do not expect different effects of the participatory culture from the country of origin on civic participation in associations addressed to foreigners versus to non-foreigners. However, previous studies suggest different effects of some control variables on the probability of joining migrant (ethnic) civic associations versus non-migrant associations. Thus, on arrival, immigrants are much more bound to the cultures of their countries of origin and are more likely to adhere to migrant associations. Consequently, length of time spent in a country exerts a different influence on membership in migrant versus non-migrant associations. Moreover, since migrant associations provide legal advice and help in solving legal status problems ), there is an increased likelihood that migrants who are not Spanish citizens will participate more in migrant associations. We estimated three separate series of models for predicting the general probability of joining an association either for migrants or for non-migrants, the probability of joining an association for migrants and the likelihood of participating in associations not especially directed to immigrants. Separate models were run for the three target variables. Each set comprises three different regression models. The first model has no explanatory variables and indicates the amount of variance allotted to the individual traits and the variance existing at the group level. The second regression includes only the individual-level predictors. The third model contains both individual and country-independent variables, while the fourth tests the effect of cross-level interac tions as well. For each dependent variable, we have employed the same set of predictors, and we have added them stepwise. In all models, we employed multiple imputation of missing values. Since the country samples have different sizes, we used a weight variable in all individual-level analyses so that all samples have the same size, the average of the ethnic groups included in the sample (N = 252). We tested the multicollinearity among the independent variables, both individual and contextual, using tolerance (estimated in SPSS 17). For all indicators, the level of tolerance was above 0.6. The ENI questionnaire determined im migrants' civic participation with two series of questions: one assessed involvement in associations for immigrants and the other the membership in associations for the general population. Respondents were asked to indi cate whether they participated in associations dedicated to immigrants such as organiza tions that provide assistance to migrants, sport associations and clubs, educational and cultural clubs, religious entities and groups and other social groups. In addition, respond ents had to specify whether they joined groups and associations that were not specifically dedicated to immigrants, such as non-governmental organizations, political, trade union and neighbourhood activities, religious entities and groups, educational and cultural groups, sporting associations and clubs and other social groups. General participation in associations is determined by a dummy variable that takes the value 1 if the respondent joined at least one civic association either for immigrants or for the general population and 0 if he/she did not join such an association. We chose to use a dichotomous dependent variable because we are interested in the likelihood of participat ing in civic associations. Moreover, similar to we consider that a count variable does not precisely reflect different levels of civic involvement. Participation in associations for immigrants is captured by a dummy variable that scores 1 if the respondent joined at least one civic association for immigrants and 0 if he/ she did not join such an association.3 Participation in regular associations (not for immigrants) is measured with a dichoto mous variable with a score of 1 if the respondent joined at least one civic association not specifically targeting immigrants and 0 if he/she did not join such an association. Three indicators ascertain exposure and integration in the host society: length of time since the person arrived in Spain (Time in Spain), whether the respondent has Spanish nationality (Spanish nationality) and language skills (Speaking Spanish). The first vari able is measured in years and ranges from 0 to 79. The second is a dummy variable indicating if the respondent has Spanish citizenship or not. The third is a subjective evaluation of the respondent's own language abilities, measured on a four-point scale and ranging from 'speaks Spanish very w ell' to 'needs improvement'. Higher values indicate better language skills. Two different indicators point out the connection w ith home country: length of time spent in the home country (Age of immigration) and frequency of contact with home society (Contacts home country). The first variable is measured in years and ranges from 11 to 88. The second indicates how often the respondent communicates by phone with his/her country of origin. The answers range from 1 to 6, from daily to less than once a year. Higher values indicate more frequent contacts. Socioeconomic status is indicated by three variables: income, work hours and educa tion. Income is denoted by monthly income in euros. Weekly work hours are shown by how much time the respondent spends in the labour market. Education is estimated by years spent in school for the highest level of studies and ranges from 1 ('no formal education') to 7 ( 'second level of tertiary studies'), no matter where the education was acquired, in Spain or in the country of origin. The effect of family context is captured with two indicators. The first indicates whether the respondent is married ( dummy variable), while the second shows parental status, tak ing a value of 1 if the respondent has children and 0 for childless persons. Gender is tapped as a dummy variable, which takes the value of 1 for female respondents. Age is a relevant predictor of civic participation, as we showed in the previous sec tion. However, we do not include this predictor in our models because of the identifica tion problem: age = time spent in the home country + time spent in Spain. Since time spent in the home country and time spent in Spain are key variables for testing our hypotheses, we left age out of our models. Country level of civic participation is revealed by the percentage of respondents who said they belonged to at least one voluntary association. The figures were computed from the WVS/EVS data collected in 2005 (or 2000 if the data for a particular country were unavailable). We used an aggregate variable that captures the level of civic participation at one point in time. According to , the associational activity is influenced by 'historical cultural fram es', which are quite stable in time. The cited authors rank countries included in the WVS in 1981 and 1991 according to the percentage of individuals reporting membership in any associations. The results show that the countries' hierarchy is very similar after a period of 10 years, even if the percent age of people reporting membership changed over time. Therefore, we consider the per centage of individuals reporting membership in an association indicator quite reliable, even if individuals were socialized at different points in time. Country-level economic development is shown with GDP p e r capita; the data come from the World Bank's World Development Indicators . Figures for each ethnic group in Spain were computed based on the ENI data, because no official data are avail able for each ethnic group. The GDP and an ethnic group's dimension values were divided by 1000 to facilitate interpretation of the output. In , we provide descrip tive statistics for the dependent and explanatory variables used in the analyses. Data provided by the ENI show that in Spain there is a high variability in immigrants' participation in civic associations, depending on the country of origin. The results shown in indicate greater participation in all types of associations for people born and socialized in countries with higher levels of civic involvement such as Sweden, the United States, the United Kingdom and Austria, as reported by and . The results confirm the findings of previous research, revealing a very low level of civic involvement among immigrants coming from post-communist countries such as M oldova, Romania and Bulgaria . The lack of civic participation in these countries is documented by previous research . At the descriptive level, the ENI data confirm our expectation, showing a relationship between im migrants' civic activism and the participatory culture of their home country. Comparing participation in associations for immigrants involved in non-immigrant associations produces different findings. The overall level of civic commitment is lower for the first type of associations, 8% compared with 15% in regular associations (see ). In addition, people socialized in more-participatory cultures are more involved in regular associations. Thus, early socialization in a participatory culture increases the propensity for civic participation, as these immigrants participate mainly in regular associations. The results of the multilevel regression models are shown in and 5. The value of the interclass correlation coefficient for the three dependent variables is 0.11 for participation in any type of association, 0.12 for participation in non-immigrant associa tions and 0.09 for participation in associations for immigrants, showing that 9-12% of the total variation is among countries of origin. For the three dependent variables, x2 is statistically significant (for all types of associations x2 = 186.982, p < .000 for d.f. = 34; for non-immigrant associations x2 = 178.494, p < .000 for d.f. = 34; for associations for immigrants x2 = 131.970, p < .000 for d.f. = 34). Therefore, there is a significant variance in immigrants' participation in various types of associations depending on the country of origin, which can be explained by individual as w ell as contextual characteristics. In and 5, the first model includes only individual-level predictors, the second adds the contextual variables (the country of origin's characteristics) and the third checked for the effect of cross-level interactions on the dependent variables. The propor tion of explained variance varies fromAE2didlo = 0.123 for M odel 1 for the dependent vari able participation in associations for immigrants (see ) to R2 didl0 = 0.207 for Model 3 for the dependent variable participation in non-immigrant associations (see R 2 dkho is the proportion of explained variance (for computing details see . ). In presenting the results, we will focus more on hierarchical models because our hypothesis assumes second-level effects as well as cross-level interactions. For participation in all types o f associations, the final hierarchical model with cross level interaction supports only the second hypothesis. Both cross-level interactions prove to have a statistically significant impact on the target variable. Thus, those who stay in closer contact with a home country that has a higher level of civic participation and those living for a longer period in a home country with a higher level of civic participation are more likely to participate in the country of destination. However, the M odel 2 results in show a statistically significant effect of membership in civic associations in the home country on immigrants' civic participation, but this effect disappears when con trolled for the cross-level interactions. The results show that it is not the effect of country of origin per se which matters, but a longer and closer contact with a higher participative culture that makes the difference. In the case of participation in all types of associations, the first hypothesis is not supported by the data, but the second one is. ft2 dkho is the proportion of explained variance (for computing details see ). In the case of participation in non-immigrant associations, if one considers the inter action between the level of civic participation in the country of origin and the length of stay in this country combined with the frequency of contacts with the home country, then the level of civic participation in the country of origin has no effect per se, and the first hypothesis is rejected. However, the second hypothesis is supported by the data. Migrants who have longer and more frequent contact w ith the home country with a higher level of civic activism are more likely to be members of non-immigrant associations in the host society. The results for participation in associations fo r immigran ts do not support our first hypothesis in any of the hierarchical models. The level of civic participation in the home country has no significant effect on this outcome variable. However, if we take into account the interaction between the level of civic participation in the country of origin and the length and frequency of contact w ith the home country, then the second hypoth esis is supported by our results. R1 dkho is the proportion of explained variance (for computing details see . In the final models, the effect of the home country's level of civic activism per se is non-significant. The second hypothesis is validated by the data, for all three dependent variables. According to our results, longer exposure to and better connection with an active civic culture increase the likelihood of participation in civic associations in the host country, no matter which type of association is under investigation. The data support the idea that people from active societies have had the opportunity to acquire civic skills that they can easily transfer to a new social context. Thus, being born in an active civic society is not important per se, but exposure to and integration in such a participatory culture matter. presents a summary of the hypotheses and the outcomes. The control variables exert similar effects on all three target variables and partially support the findings of previous research. We comment on the results for the three differ ent models together, because they are similar. Individual-level variables in the third model support the resistance theory, showing that the longer an individual has lived in his/her home country the lower the likelihood of becoming involved in any type of civic associations or in non-immigrant associations. However, this might be an age effect, too. In addition, a closer connection with the home country decreases the probability of being active in associations for immigrants in Spain. This outcome is not surprising since eth nic associations provide opportunities to get in touch with the country of origin, and those who get in touch on their own do not need the extra support provided by associa tions. Better integration in the home society reduces civic participation in the host soci ety, except in the case of people who come from highly participative cultures. The results do not support exposure theory. Time spent in the host society and citizen ship do not have a significant impact on any of the dependent variables. Language skills have a strong and significant effect on participation in associations for immigrants and in all types of associations. In the case of non-immigrant associations, the effect is positive, but significant only in the full model ( see M odel 3 in ). The data indicate that income has a positive effect on civic participation; wealthier people are more inclined to join all types of associations. Time spent weekly in the labour market has the opposite effect. Our results support role overload theory (see Markham and that states employed people have less time to spend in civic associations and contradicts our first expectation. Contrary to other research on civic participation, we found women are more likely to join both types of associations. Family context, marriage and parenthood seem to play no role in influencing immigrants' civic participation. The same result is valid for the individual level of education, no matter where it was acquired. Considering the second-level indicators, those related to the immigrants' country of origin, the results indicate a significant negative impact of economic development on participation in immigrants' associations in the model without control for cross-level interactions. This outcome indicates that the likelihood of participating in associations for immigrants is higher for people coming from poorer countries and lower for those from wealthier countries. However, the effect disappeared when controlling for cross level interactions. For the other dependent variables, the level of economic development in the country of origin has no impact. Our results show that the difference in the immi grants' level of civic participation is not produced by the different level of economic development of the home country, but by the level of civic participation in the home country moderated by the length of contact with the home versus the host society. The data show that the intercountry differences are not due to economic development, but to an intersection between micro-and macro-level factors. The results shown in indicate that the larger the ethnic group, the more likely immigrants are involved in associations. Larger networks provide individuals with many connections that help them get involved in the host country's civic life. According to theory, we expected a sim ilar effect for ethnic associations, but not for the non immigrant associations. Our results may be explained by findings, which state that the relationship between the size of an ethnic group and individual involvement in ethnic associations is not linear. There is a lower propensity for participation among very small and very large groups of immigrants. In the first case, there are too few people to set up an organization. In groups that are too large, however, ethnic attachment is diluted. Moreover, our dependent variable does not measure involve ment in ethnic associations; our variable measures participation in associations for immi grants, no matter which country they come from. The article focuses on the effect of civic commitment in the home country on immi grants' participation in civic associations in the host society. Using data from the Spanish National Immigrant Survey carried out in 2007, the article tests whether immigrants transfer their civic skills from the home to the host society. The article investigates whether people socialized in more-participatory contexts are more likely to get involved in civic activities after migrating because they acquired civic skills during their early socialization and use them to become integrated into the new society. Multilevel logistic regressions were employed to test the hypotheses. The results of multilevel logistic regressions partially confirm our hypotheses. The outcome reveals a significant positive effect of a home country's civic activism on cur rent involvement in associations for immigrants or non-immigrants, mediated by longer exposure to and closer contact with the country of origin. Thus, immigrants who have spent more time in a more-participatory context and who are in closer contact with these societies are more likely to get involved in both types of civic associations in their host country. The data support our second hypothesis but do not confirm the first one. The current results support the transferability theory, but they show that transferability operates in certain circumstances. Therefore, the culture of the home country is important for the participation in the host society, if the migrant has spent enough time in a country with a higher level of civic participation and she or he has had enough time to acquire the skills needed for civic participation. Moreover, a closer contact with a participatory cul ture boosts civic skills even after migration. Basically, these results point out that in the case of civic skills a combination of transferability and resistance explains better how the culture of the home country influences participation at destination. Thus, civic skills are transferred in the host society only if they are sufficiently consolidated by socialization in the home country and if contact with this culture is persistent even after migration. However, our results are limited by the available data. The ENI data do not allow checking for previous participatory experiences in the home country. It is quite likely that previous experiences will increase the likelihood of participation in the host society. Therefore, future research should consider the impact of previous participatory experi ences and the participatory culture of the home country. Cette étude s'intéresse au niveau d 'adhésion à des associations de la population immigrée en Espagne. Trois types d 'implication civique sont pris en considération : la participation dans des associations de tous types, dans des associations d 'immigrés et dans des associations de nonimmigrés. L'article étudie si les immigrés originaires de pays ayant des niveaux de participation civique plus élevés sont plus susceptibles d 'être actifs dans des associations civiques et si les immigrés ayant vécu plus longtemps ou ayant maintenu un contact plus étroit avec un pays d 'origine à plus fort niveau de participation civique sont plus susceptibles de devenir membres d 'associations civiques. Les données utilisées proviennent de l 'Encuesta Española N acional de Inm igración . Les résultats des régressions logis tiques multi-niveaux montrent que les immigrés qui ont passé plus de temps dans un environne ment participatif dans leur pays d 'origine et qui sont en contact plus étroit avec ces sociétés sont plus susceptibles de s'impliquer dans des associations civiques dans le pays de destination. Mots-clés: Espagne, participation civique des immigrés, pays d 'origine, socialisation Este estudio se centra en el nivel de pertenencia a asociaciones de la población inmigrante en España. Se analizan tres tipos de implicación cívica: participación en asociaciones cívicas de todos los tipos, en asociaciones para inmigrantes y en asociaciones de no inmigrantes. El artículo inves tiga si es más probable que los inmigrantes que vienen de países con mayores niveles de partici pación cívica participen en asociaciones y si es más probable que los inmigrantes que han vivido durante más tiempo y han estado más en contacto con un país de origen con mayores niveles de participación cívica entren en asociaciones. Los datos usados provienen de la Encuesta Española N acional de Inmigración (2007) y la Encuesta Mundial de . Los resultados de las regresiones logísticas multinivel m uestran que es más probable que los inmigrantes que han pasado más tiempo en un ambiente más participativo en origen y que han estado más en contacto con esas sociedades se impliquen en asociaciones cívicas en el país de destino. Palabras clave: España, país de origen, participación cívica de inmigrantes, socialización Immigrant political adaptation in Canada: Some tentative findings Canadian Journal o f Political Science The practice o f politics in two settings: Political transferability among recent im mi grants to Canada Canadian Journal o f Political Science Age, resistance, and political learning in a new environment. The case o f Canadian immigrants Comparative Politics The Private Roots o f Public Action: Gender, Equality, and Political Participation Integration, participation, identity: Immigrant associations in the province o f Milan International M igration Comm unity-based organizations and migration in N ew Y ork City. Journal o f Ethnic and Migration Studies Values and the social sciences: A global w orld o f global values? Easter P, Braun M and M ohler P (eds) Globalization, Values Change, and Generation: A Cross-National Perspective Local Integration Policies fo r Migrants in Europe. Dublin: European Foundation for the Improvement o f Living and W orking Conditions Political participation and political trust in Amsterdam: Civic communities and ethnic networks Journal o f Ethnic and Migration Studies Party identification and political adaptation o f A m erican migrants in Australia Journal o f Politics The role o f social networks in immigrant w om en's political incor poration International Migration Review Im migrant volunteering: A stepping stone to integration? N onprofit and Voluntary Sector Quarterly Volunteering in global perspective The Values o f Volunteering. N ew Y ork and Dordrecht The stability o f political attitudes and behaviors across ado lescence and early adulthood: A comparison o f survey data on adolescents and young adults in eight countries Journal o f Youth and Adolescence Political participation and associational life o f Turkish residents in the capital o f Europe Turkish Studies Pathways to voluntarism: Fam ily socialization and status transm is sion models Social Forces Developing citizens: The impact o f civic learning opportunities on students commitment to civic participation American Educational Research Journal Civic skills building: The missing component in service programs? Political Science and Politics M otivation fo r civic participation o f immigrants: The role ofp erso n a l resources, social iden tities, and personal traits. POLITIS-W orking Paper No Persistence and change in political confidence over the life-span: Embedding life-cycle socialization context British Journal o f Political Science Employment status and the attitudes and behavior o f higher status w om en volunteers Sex Roles Immigrants and associations: A global and historical perspective Journal o f Ethnic and Migration Studies Bridge over troubled water? M igration and social capital Journal o f Ethnic and Migration Studies Political socialization A nnual Sociological Review Civil society, migrant organizations and political parties: Theoretical linkages and applications to the Swedish context Journal o f Ethnic and M igration Studies N ational context, religiosity and volunteering: Results from 53 countries American Sociological Review The structural contexts o f civic engagement: Voluntary association membership in comparative perspective American Sociological Review Im migrant organizations Journal o f Ethnic and Migration Studies Evidence o f the long-term persistence o f adults' political participation Journal o f Politics Politics matters: Political events as catalysts for preadult sociali zation American Political Science Review Political incorporation among immigrants from ten areas o f origin: The persistence o f source country effects International Migration Review M ultilevel Analysis: An Introduction to Basic and Advanced M ultilevel Modeling Naturalization, socialization, participation: Immigrants and (non-) voting The Journal o f Politics Social capital o f organizations and their members: Explaining the political integra tion o f immigrants in Amsterdam Journal o f Ethnic and M igration Studies The school's role in developing civic engagement: A study o f adolescents in twenty-eight countries A pplied D evelopm ental Science M embership in voluntary organizations and democratic performance: European Post-Comm unist countries in comparative perspective Communist and Post-Com m unist Studies Voice and Equality: Civic Voluntarism in American Politics Organizational patterns: Surinamese and Turkish associations in Amsterdam Journal o f Ethnic and Migration Studies Civic activation o f immigrants: An introduction to concep tual and theoretical issues. POLITIS-W orking Paper No Im migrant involvement in voluntary association Journal o f Ethnic and Migration Studies H om eland interests, hostland politics: Politicized ethnic identity among Middle Eastern heritage groups in United States International M igration Review The political resocialization o f immigrants: Resistance or lifelong learning? Political Research Quarterly The effects o f age and political exposure on the development o f party identification among A sia A m erica and Latino immigrants in the U nited States Political Behavior Im m igrants' membership in civic associations: W h y are some im m igrants more active than others? GESIS Malina Voicu malina.voicu@gesis.org GESIS Leibniz Institute for the Social Sciences Germany Leibniz Institute for the Social Sciences Unter Sachsenhausen 6-8 50667 Cologne Germany Alexandra Loana University of Bucharest Romania Rusu University of Bucharest Romania Im m igrants' membership in civic associations: W h y are some im m igrants more active than others? 10.1177/0268580912452172 GROBID - A machine learning software for extracting information from scholarly documents Country of origin immigrant civic participation socialization Spain This study focuses on the level of membership in associations of the migrant population in Spain. Three types of civic engagement are considered: participation in all types of civic associations, in associations for immigrants and in non-immigrant associations. The article investigates whether immigrants coming from countries with higher levels of civic participation are m ore likely to participate in civic associations and if immigrants who have lived longer in and stayed in closer contact with a home country with a higher level of civic participation are m ore likely to join civic associations. Data used come from the Spanish National Immigrant Survey (2007) and the W o rld Values Survey (2000, 2005). The results of multilevel logistic regressions show that immigrants who have spent more time in a more participatory context at origin and who are in closer contact with these societies are more likely to get involved in civic associations at destination. Introduction Participation in civic associations varies among and within countries and within the same country among different social groups. Studies report different rates of immigrants' civic commitment within the same host society depending on their country of origin (Nannestad et a l, 2008; Tillie, 2004; Vermeulen, 2005) . Im m igrants' participation in civic associations has been approached mainly through qualitative research methods, but some studies based on quantitative methodology are available as well. Vermeulen (2005) underlines the impact of differences in civic and political culture in Surinam and Turkey on Surinamese and Turkish immigrants' civic participation in Amsterdam, while Fennema and Tillie (1999) noticed a similar effect in the case of Turkish and Antillean migrants in Amsterdam. However, the results are restricted to a specific host city and a limited number of ethnic groups. This article focuses on civic participation of first-generation immigrants in Spain and tries to explain why some immigrants are more active than others. We investigate the effect of the level of civic activism in the home country on civic participation in the host society using quantitative research methods. Using data provided by the Spanish National Immigrant Survey, we employ hierarchical multilevel regression models. The hierarchi cal models allow testing simultaneously for individual-and group-level characteristics that may influence im migrants' civic commitment. Because all immigrants included in the sample live in the same society, we have the opportunity to analyse the 'pure' effect of a country-of-origin participatory culture maintaining constancy in the societal context in the host culture. We define civic participation as involvement in civic associations, and we consider it a specific category of political participation, following Verba et al. (1995) . We intend to predict the effect of a home country's civic activism on three different outcome varia bles: participation in all types of civic associations, participation in associations for immigrants and participation in non-immigrant associations. We do not expect different effects of our main predictor on the three dependent variables, but since participation in associations for immigrants has different connections with the control variables, we employed three different analyses for the dependent variables. The results show that immigrants who have lived longer in and are in closer contact w ith highly participatory home countries are more likely to get involved in civic associations in the host country because the immigrants previously acquired civic skills that can be transferred to the new social context. A n alternative explanation may be that they have built a larger social network at home and they retain ties that can be activated when moving to the host society. The first section covers existing literature on im migrants' adaptation to host societies' political cultures as well as on factors influencing immigrants' civic participation. The second part introduces the indicators and the strategy used for analysis, while the third comprises the data analysis. The discussion and conclusions are found in the final section. Review of literature and hypotheses Three different approaches have been used to explain how immigrants adapt their politi cal beliefs and behaviours to a new political environment: resistance theory, exposure theory and transferability theory. According to the first approach, political attitudes and behaviours are acquired during early socialization and are resistant to change later in life (Easter et al., 2006; Sears and Funk, 1999; Sears and Valentino, 1997) . Consequently, immigrants encounter many difficulties in adapting their political behaviour to the host society because they were socialized in a different context. Time spent in the country of origin is a key variable; younger immigrants adapt more easily to the new environment because they have had little contact with their original society ( White et al., 2008) . The last two approaches assume that political attitudes and behaviour are open to change even during adulthood and immigrants can be politically resocialized into the host society. Exposure theory considers the time spent in the host society as a key vari able for resocialization, the likelihood of adaptation increasing with the time spent at the destination (Black, 1982; White et al., 2008; Wong, 2000) . This approach does not pay much attention to the effect of early socialization in the country of origin and considers the immigrant a 'tabula rasa' from the political point of view ( White et al., 2008) . The third approach, transferability theory, notes that immigrants acquire political skills and information during early socialization in their country of origin, which pro vides a background for political learning in the host society (Black, 1987; Black et al., 1987; Finifter and Finifter, 1989; Simpson Bucker, 2005; White et al., 2008) . Thus, immigrants can transfer political knowledge obtained during early socialization and use it to build new political skills in the new political environment. White et al. (2008: 277) showed that 'there is no single pathway to political learning: which pathway applies depends upon which orientation, or which political behavior, is under consideration'. They found that the exposure mechanism can better explain voter turnout and political partisanship, while the transfer mechanism provides a reliable explanation of political interest. According to the authors, while voting and political partisanship are more demanding activities requiring judgement and decision-making, political interest is often a byproduct of other common activities involving basic skills that can easily be transferred from one political system to another. Civic participation has some particular characteristics compared with general political participation, and we assume that a different mechanism operates therein. According to Verba et al. (1995) , civic participation during adulthood requires three 'participatory facts': motivation, capacity and skills and networks. To participate, indi viduals should be willing to get involved; should have the capabilities to join an associa tion in terms of money, time and skills; and should have connections with others involved in civic activities. In the case of immigrants, the existence of a larger group from the same country of origin seems to facilitate civic participation (Cordero-Guzman, 2005; Moya, 2005; Vogel and Triandafyllidou, 2007) . Participation in various types of civic associations operates as a mechanism of inte gration into the new society and helps individuals in their relations with the home and host societies (Caselli, 2010; Cordero-Guzman, 2005) . Participation in civic activities brings rewards to immigrants, motivating them to get involved. It can enable the practice of civic and political skills (Caselli, 2010; Cordero-Guzman, 2005; Fennema and Tillie, 1999; Kosic, 2007; Tillie, 2004) , and improve language skills (Handy and Greenspan, 2009) . Moreover, participation facilitates contacts with the home country (Cordero-Guzman, 2005; Odmalm, 2004) and helps immigrants to deal with local institutions (Caselli, 2010; Cordero-Guzman, 2005; Odmalm, 2004) . Participation in civic associations requires skills and abilities. According to the soci ological literature, civic skills are acquired early in life, and participation during the formative years and youth increases the likelihood of participation in civic activities during adulthood (Kirlin, 2002) . Three factors are particularly important for learning the skills required for civic participation: fam ily school and involvement in extracur ricular activities. Parents are usually models for their offspring, and their participation will lead to similar behaviour among the younger generation ( Janoski and Wilson, 1995; Khane and Sporte, 2008; Langton, 1984; Niemi and Sobieszek, 1977) . A similar effect was reported in the case of neighbourhoods (Khane and Sporte, 2008) . Schools may boost civic participation through various channels that contribute to the acquisition of fundamental civic skills (Khane and Sporte, 2008; Niemi and Sobieszek, 1977; Torney-Purta, 2002) . Some countries are more successful than others at building civic skills. Citizens from more participative cultures are more likely to transfer civic skills from the old environ ment to the host society when people are exposed to the migration experience and can benefit from civic participation. We assume that a transfer mechanism operates in the case of immigrants' civic participation, i.e. existing skills are used and adapted to the new environment. People who have been involved in civic activities in their home coun try are more likely to participate in the host society. Since the available data do not allow for the effect of previous civic experience to be tested at the individual level, we will focus only on the impact of the level of civic participation in the country of origin on current civic participation. As we have shown, civic participation early in life increases the probability of civic involvement during adulthood (Hooghe and Wilkenfeld, 2008; Kirlin, 2002) . Several studies note the existence of institutional settings that can boost or reduce the chances of civic participation during adult life (Almond and Verba, 1963; Bums et al., 2001; Verba et al., 1995) . The workplace is such a setting. Employed people have more opportunities to practise civic skills compared w ith those outside the labour market. Almond and Verba (1963) find significant intercountry differences in employees' involvement in decision making, which boosts civic participation. Some countries are more successful than oth ers in extending civic-skills learning during adult life, as well. Citizens from societies that encourage training in civic skills during adulthood are more likely to become involved in civic activities. The longer these citizens live in such a social context, the more likely they will have many civic skills that can be transferred and adapted when they migrate to a new country. We assume that a combination of trans fer and resistance mechanisms is likely to occur, which stimulates civic participation by immigrants coming from more civically active societies. Thus, immigrants who lived for a long time in a home country with higher levels of civic involvement and who maintain close contacts w ith it are more likely to participate in the host society because they have better skills for such activities. Studies dedicated to civic participation identify various individual and contextual fac tors influencing people's propensity to engage in civic activities. At the individual level, studies have shown that characteristics such as education, employment status, income, family and gender have a significant impact on civic participation (Ruiter and De Graaf, 2006; Schofer and Fourcade-Gourinchas, 2001) . Employment also has a positive effect on civic participation, because employment integrates people in various social networks and leads to civic involvement and participation, while gender induces differences in the level and type of civic participation due to the different involvement of women and men in family, employment and religion, women being more active compared to men. Moreover, people with higher incomes and more education are more civically active because they have more resources to share, while marriage and parenthood have positive effects. Parenthood boosts civic participation because parents have a larger social net work by meeting other parents, while marriage with an active person increases the level of both spouses' civic activism. Other studies have focused on the national context. Ruiter and De Graaf (2006) found that economic development (measured by GDP/capita) has a positive impact on mem bership in civic associations but no effect on volunteering, whereas the level of democ racy (measured by the Freedom House Gastil Index) has a positive impact on volunteering but no significant impact on membership. Generally speaking, social wealth is associated with more resources required and a higher level of education, both increasing associational membership (Schofer and Fourcade-Gourinchas, 2001 ). Other individual and contextual factors directly influence immigrants' participation in civic associations. Among individual factors is the length of stay in the host society, which stands for the exposure mechanism (Gidegil and Stolle, 2009; Wald, 2008; White et al., 2008) , and language skills, which help communication and integration into the host society (Jacobs et al., 2006; Kosic, 2007; Tam Cho, 1999; Vogel and Triandafyllidou, 2007) . The size of the immigrant group from the same country of origin exerts a contex tual influence on the likelihood of immigrant participation; people belonging to larger groups are more likely to be active because they have an extended social network (Black, 1982; Cordero-Guzman, 2005; Moya, 2005; Vogel and Triandafyllidou, 2007) . Starting from the theoretical approaches referred to above and controlling for all the relevant predictors, we propose two hypotheses: H I: Immigrants coming from countries with higher levels of membership in civic associations are more likely to get involved in civic associations in the host countries. H2: Immigrants who have lived longer in and stayed in closer contact with a home country with a higher level of membership in civic associations are more likely to partici pate in civic associations. Data, methods and measurement Data and methods The analysis is based on the data set from the Spanish National Immigrant Survey (ENI) carried out in 2007 by the Spanish National Institute of Statistics on a representative sample of the immigrant population living in Spain.1 The questionnaire provides infor mation about respondents' participation in various associations as well as information regarding social background and migration history. The data set offers information about 15,465 immigrants from 94 countries living in Spain. From this sample, we selected a subsample of 8575 immigrants from 34 countries of origin. We included only immi grants who came directly from their country of origin and did not report previous migra tion to other countries. We did so because we assumed the effect of intermediary migration would be difficult to interpret. We selected only people who had resided in their home country at least until the age of 10 because we wanted to capture the effect of early socialization, and migrants who leave their home country early do not have enough opportunity to interact with societal institutions influencing civic participation. We have excluded immigrants from countries of origin with fewer than 25 cases in the ENI database. The 34 countries of origin included in the analysis are shown in Table 1 . The total number of cases by country of origin ranges from 25 for Norway to 1850 for Morocco, with an average of 252 cases. The macro-level indicators used in the analyses come from the 2005 World Values Survey and the 2000 European Values Survey/World Values Survey (EVS/WVS) for the level of civic activism in the country of origin, from the 2008 KILM for the GDP (data reported for 2005) and from the 2007 ENI for the total number of immigrants by country of origin ( our own computations).2 We produced several m ultilevel regression m odels using the HLM 6.04 program for hierarchical linear m odelling. We estim ated hierarchical logistic regression models, because our dependent variables are dichotomous ones, predicting individuals' prob ability o f participating in civic associations. The m ultilevel regression allows testing of the com bined effect of individual independent variables and contextual-level variables. The ENI provides separate information about immigrants' participation in associa tions dedicated to foreigners and associations not directly dedicated to foreigners. We do not expect different effects of the participatory culture from the country of origin on civic participation in associations addressed to foreigners versus to non-foreigners. However, previous studies suggest different effects of some control variables on the probability of joining migrant (ethnic) civic associations versus non-migrant associations. Thus, on arrival, immigrants are much more bound to the cultures of their countries of origin ( Wald, 2008) and are more likely to adhere to migrant associations. Consequently, length of time spent in a country exerts a different influence on membership in migrant versus non-migrant associations. Moreover, since migrant associations provide legal advice and help in solving legal status problems (Caselli, 2010; Cordero-Guzman, 2005 ), there is an increased likelihood that migrants who are not Spanish citizens will participate more in migrant associations. We estimated three separate series of models for predicting the general probability of joining an association either for migrants or for non-migrants, the probability of joining an association for migrants and the likelihood of participating in associations not especially directed to immigrants. Separate models were run for the three target variables. Each set comprises three different regression models. The first model has no explanatory variables and indicates the amount of variance allotted to the individual traits and the variance existing at the group level. The second regression includes only the individual-level predictors. The third model contains both individual and country-independent variables, while the fourth tests the effect of cross-level interac tions as well. For each dependent variable, we have employed the same set of predictors, and we have added them stepwise. In all models, we employed multiple imputation of missing values. Since the country samples have different sizes, we used a weight variable in all individual-level analyses so that all samples have the same size, the average of the ethnic groups included in the sample (N = 252). We tested the multicollinearity among the independent variables, both individual and contextual, using tolerance (estimated in SPSS 17). For all indicators, the level of tolerance was above 0.6. Dependent variables The ENI questionnaire determined im migrants' civic participation with two series of questions: one assessed involvement in associations for immigrants and the other the membership in associations for the general population. Respondents were asked to indi cate whether they participated in associations dedicated to immigrants such as organiza tions that provide assistance to migrants, sport associations and clubs, educational and cultural clubs, religious entities and groups and other social groups. In addition, respond ents had to specify whether they joined groups and associations that were not specifically dedicated to immigrants, such as non-governmental organizations, political, trade union and neighbourhood activities, religious entities and groups, educational and cultural groups, sporting associations and clubs and other social groups. General participation in associations is determined by a dummy variable that takes the value 1 if the respondent joined at least one civic association either for immigrants or for the general population and 0 if he/she did not join such an association. We chose to use a dichotomous dependent variable because we are interested in the likelihood of participat ing in civic associations. Moreover, similar to Ruiter and De Graaf (2006) we consider that a count variable does not precisely reflect different levels of civic involvement. Participation in associations for immigrants is captured by a dummy variable that scores 1 if the respondent joined at least one civic association for immigrants and 0 if he/ she did not join such an association.3 Participation in regular associations (not for immigrants) is measured with a dichoto mous variable with a score of 1 if the respondent joined at least one civic association not specifically targeting immigrants and 0 if he/she did not join such an association. Independent variables -individual level Three indicators ascertain exposure and integration in the host society: length of time since the person arrived in Spain (Time in Spain), whether the respondent has Spanish nationality (Spanish nationality) and language skills (Speaking Spanish). The first vari able is measured in years and ranges from 0 to 79. The second is a dummy variable indicating if the respondent has Spanish citizenship or not. The third is a subjective evaluation of the respondent's own language abilities, measured on a four-point scale and ranging from 'speaks Spanish very w ell' to 'needs improvement'. Higher values indicate better language skills. Two different indicators point out the connection w ith home country: length of time spent in the home country (Age of immigration) and frequency of contact with home society (Contacts home country). The first variable is measured in years and ranges from 11 to 88. The second indicates how often the respondent communicates by phone with his/her country of origin. The answers range from 1 to 6, from daily to less than once a year. Higher values indicate more frequent contacts. Socioeconomic status is indicated by three variables: income, work hours and educa tion. Income is denoted by monthly income in euros. Weekly work hours are shown by how much time the respondent spends in the labour market. Education is estimated by years spent in school for the highest level of studies and ranges from 1 ('no formal education') to 7 ( 'second level of tertiary studies'), no matter where the education was acquired, in Spain or in the country of origin. The effect of family context is captured with two indicators. The first indicates whether the respondent is married ( dummy variable), while the second shows parental status, tak ing a value of 1 if the respondent has children and 0 for childless persons. Gender is tapped as a dummy variable, which takes the value of 1 for female respondents. Age is a relevant predictor of civic participation, as we showed in the previous sec tion. However, we do not include this predictor in our models because of the identifica tion problem: age = time spent in the home country + time spent in Spain. Since time spent in the home country and time spent in Spain are key variables for testing our hypotheses, we left age out of our models. Independent variables -country-of-origin level Country level of civic participation is revealed by the percentage of respondents who said they belonged to at least one voluntary association. The figures were computed from the WVS/EVS data collected in 2005 (or 2000 if the data for a particular country were unavailable). We used an aggregate variable that captures the level of civic participation at one point in time. According to Schofer and Fourcade-Gourinchas (2001) , the associational activity is influenced by 'historical cultural fram es', which are quite stable in time. The cited authors rank countries included in the WVS in 1981 and 1991 according to the percentage of individuals reporting membership in any associations. The results show that the countries' hierarchy is very similar after a period of 10 years, even if the percent age of people reporting membership changed over time. Therefore, we consider the per centage of individuals reporting membership in an association indicator quite reliable, even if individuals were socialized at different points in time. Country-level economic development is shown with GDP p e r capita; the data come from the World Bank's World Development Indicators (2005) . Figures for each ethnic group in Spain were computed based on the ENI data, because no official data are avail able for each ethnic group. The GDP and an ethnic group's dimension values were divided by 1000 to facilitate interpretation of the output. In Table 2 , we provide descrip tive statistics for the dependent and explanatory variables used in the analyses. Results Data provided by the ENI show that in Spain there is a high variability in immigrants' participation in civic associations, depending on the country of origin. The results shown in Table 1 indicate greater participation in all types of associations for people born and socialized in countries with higher levels of civic involvement such as Sweden, the United States, the United Kingdom and Austria, as reported by Schofer and Fourcade-Gourinchas (2001) and Hodgkinson (2003) . The results confirm the findings of previous research, revealing a very low level of civic involvement among immigrants coming from post-communist countries such as M oldova, Romania and Bulgaria (Kosic, 2007) . The lack of civic participation in these countries is documented by previous research ( Valkov, 2009) . At the descriptive level, the ENI data confirm our expectation, showing a relationship between im migrants' civic activism and the participatory culture of their home country. Comparing participation in associations for immigrants involved in non-immigrant associations produces different findings. The overall level of civic commitment is lower for the first type of associations, 8% compared with 15% in regular associations (see Table 1 ). In addition, people socialized in more-participatory cultures are more involved in regular associations. Thus, early socialization in a participatory culture increases the propensity for civic participation, as these immigrants participate mainly in regular associations. The results of the multilevel regression models are shown in Tables 3, 4 and 5. The value of the interclass correlation coefficient for the three dependent variables is 0.11 for participation in any type of association, 0.12 for participation in non-immigrant associa tions and 0.09 for participation in associations for immigrants, showing that 9-12% of the total variation is among countries of origin. For the three dependent variables, x2 is statistically significant (for all types of associations x2 = 186.982, p < .000 for d.f. = 34; for non-immigrant associations x2 = 178.494, p < .000 for d.f. = 34; for associations for immigrants x2 = 131.970, p < .000 for d.f. = 34). Therefore, there is a significant variance in immigrants' participation in various types of associations depending on the country of origin, which can be explained by individual as w ell as contextual characteristics. In Tables 3, 4 and 5, the first model includes only individual-level predictors, the second adds the contextual variables (the country of origin's characteristics) and the third checked for the effect of cross-level interactions on the dependent variables. The propor tion of explained variance varies fromAE2didlo = 0.123 for M odel 1 for the dependent vari able participation in associations for immigrants (see Table 5 ) to R2 didl0 = 0.207 for Model 3 for the dependent variable participation in non-immigrant associations (see R 2 dkho is the proportion of explained variance (for computing details see Snijders and Bosker, 2002: 224) . Table 4 ). In presenting the results, we will focus more on hierarchical models because our hypothesis assumes second-level effects as well as cross-level interactions. For participation in all types o f associations, the final hierarchical model with cross level interaction supports only the second hypothesis. Both cross-level interactions prove to have a statistically significant impact on the target variable. Thus, those who stay in closer contact with a home country that has a higher level of civic participation and those living for a longer period in a home country with a higher level of civic participation are more likely to participate in the country of destination. However, the M odel 2 results in Table 3 show a statistically significant effect of membership in civic associations in the home country on immigrants' civic participation, but this effect disappears when con trolled for the cross-level interactions. The results show that it is not the effect of country of origin per se which matters, but a longer and closer contact with a higher participative culture that makes the difference. In the case of participation in all types of associations, the first hypothesis is not supported by the data, but the second one is. ft2 dkho is the proportion of explained variance (for computing details see Snijders and Bosker, 2002: 224 ). In the case of participation in non-immigrant associations, if one considers the inter action between the level of civic participation in the country of origin and the length of stay in this country combined with the frequency of contacts with the home country, then the level of civic participation in the country of origin has no effect per se, and the first hypothesis is rejected. However, the second hypothesis is supported by the data. Migrants who have longer and more frequent contact w ith the home country with a higher level of civic activism are more likely to be members of non-immigrant associations in the host society. The results for participation in associations fo r immigran ts do not support our first hypothesis in any of the hierarchical models. The level of civic participation in the home country has no significant effect on this outcome variable. However, if we take into account the interaction between the level of civic participation in the country of origin and the length and frequency of contact w ith the home country, then the second hypoth esis is supported by our results. R1 dkho is the proportion of explained variance (for computing details see Snijders and Bosker, 2002: 224) . In the final models, the effect of the home country's level of civic activism per se is non-significant. The second hypothesis is validated by the data, for all three dependent variables. According to our results, longer exposure to and better connection with an active civic culture increase the likelihood of participation in civic associations in the host country, no matter which type of association is under investigation. The data support the idea that people from active societies have had the opportunity to acquire civic skills that they can easily transfer to a new social context. Thus, being born in an active civic society is not important per se, but exposure to and integration in such a participatory culture matter. Table 6 presents a summary of the hypotheses and the outcomes. The control variables exert similar effects on all three target variables and partially support the findings of previous research. We comment on the results for the three differ ent models together, because they are similar. Individual-level variables in the third model support the resistance theory, showing that the longer an individual has lived in his/her home country the lower the likelihood of becoming involved in any type of civic associations or in non-immigrant associations. However, this might be an age effect, too. In addition, a closer connection with the home country decreases the probability of being active in associations for immigrants in Spain. This outcome is not surprising since eth nic associations provide opportunities to get in touch with the country of origin, and those who get in touch on their own do not need the extra support provided by associa tions. Better integration in the home society reduces civic participation in the host soci ety, except in the case of people who come from highly participative cultures. The results do not support exposure theory. Time spent in the host society and citizen ship do not have a significant impact on any of the dependent variables. Language skills have a strong and significant effect on participation in associations for immigrants and in all types of associations. In the case of non-immigrant associations, the effect is positive, but significant only in the full model ( see M odel 3 in Table 4 ). The data indicate that income has a positive effect on civic participation; wealthier people are more inclined to join all types of associations. Time spent weekly in the labour market has the opposite effect. Our results support role overload theory (see Markham and Bonjean, 1996) that states employed people have less time to spend in civic associations and contradicts our first expectation. Contrary to other research on civic participation, we found women are more likely to join both types of associations. Family context, marriage and parenthood seem to play no role in influencing immigrants' civic participation. The same result is valid for the individual level of education, no matter where it was acquired. Considering the second-level indicators, those related to the immigrants' country of origin, the results indicate a significant negative impact of economic development on participation in immigrants' associations in the model without control for cross-level interactions. This outcome indicates that the likelihood of participating in associations for immigrants is higher for people coming from poorer countries and lower for those from wealthier countries. However, the effect disappeared when controlling for cross level interactions. For the other dependent variables, the level of economic development in the country of origin has no impact. Our results show that the difference in the immi grants' level of civic participation is not produced by the different level of economic development of the home country, but by the level of civic participation in the home country moderated by the length of contact with the home versus the host society. The data show that the intercountry differences are not due to economic development, but to an intersection between micro-and macro-level factors. The results shown in Table 5 indicate that the larger the ethnic group, the more likely immigrants are involved in associations. Larger networks provide individuals with many connections that help them get involved in the host country's civic life. According to theory, we expected a sim ilar effect for ethnic associations, but not for the non immigrant associations. Our results may be explained by Schrover and Vermeulen's (2005) findings, which state that the relationship between the size of an ethnic group and individual involvement in ethnic associations is not linear. There is a lower propensity for participation among very small and very large groups of immigrants. In the first case, there are too few people to set up an organization. In groups that are too large, however, ethnic attachment is diluted. Moreover, our dependent variable does not measure involve ment in ethnic associations; our variable measures participation in associations for immi grants, no matter which country they come from. Conclusions The article focuses on the effect of civic commitment in the home country on immi grants' participation in civic associations in the host society. Using data from the Spanish National Immigrant Survey carried out in 2007, the article tests whether immigrants transfer their civic skills from the home to the host society. The article investigates whether people socialized in more-participatory contexts are more likely to get involved in civic activities after migrating because they acquired civic skills during their early socialization and use them to become integrated into the new society. Multilevel logistic regressions were employed to test the hypotheses. The results of multilevel logistic regressions partially confirm our hypotheses. The outcome reveals a significant positive effect of a home country's civic activism on cur rent involvement in associations for immigrants or non-immigrants, mediated by longer exposure to and closer contact with the country of origin. Thus, immigrants who have spent more time in a more-participatory context and who are in closer contact with these societies are more likely to get involved in both types of civic associations in their host country. The data support our second hypothesis but do not confirm the first one. The current results support the transferability theory, but they show that transferability operates in certain circumstances. Therefore, the culture of the home country is important for the participation in the host society, if the migrant has spent enough time in a country with a higher level of civic participation and she or he has had enough time to acquire the skills needed for civic participation. Moreover, a closer contact with a participatory cul ture boosts civic skills even after migration. Basically, these results point out that in the case of civic skills a combination of transferability and resistance explains better how the culture of the home country influences participation at destination. Thus, civic skills are transferred in the host society only if they are sufficiently consolidated by socialization in the home country and if contact with this culture is persistent even after migration. However, our results are limited by the available data. The ENI data do not allow checking for previous participatory experiences in the home country. It is quite likely that previous experiences will increase the likelihood of participation in the host society. Therefore, future research should consider the impact of previous participatory experi ences and the participatory culture of the home country. Résumé Cette étude s'intéresse au niveau d 'adhésion à des associations de la population immigrée en Espagne. Trois types d 'implication civique sont pris en considération : la participation dans des associations de tous types, dans des associations d 'immigrés et dans des associations de nonimmigrés. L'article étudie si les immigrés originaires de pays ayant des niveaux de participation civique plus élevés sont plus susceptibles d 'être actifs dans des associations civiques et si les immigrés ayant vécu plus longtemps ou ayant maintenu un contact plus étroit avec un pays d 'origine à plus fort niveau de participation civique sont plus susceptibles de devenir membres d 'associations civiques. Les données utilisées proviennent de l 'Encuesta Española N acional de Inm igración ) et du W orld Values Survey (2000 . Les résultats des régressions logis tiques multi-niveaux montrent que les immigrés qui ont passé plus de temps dans un environne ment participatif dans leur pays d 'origine et qui sont en contact plus étroit avec ces sociétés sont plus susceptibles de s'impliquer dans des associations civiques dans le pays de destination. Mots-clés: Espagne, participation civique des immigrés, pays d 'origine, socialisation Resumen Este estudio se centra en el nivel de pertenencia a asociaciones de la población inmigrante en España. Se analizan tres tipos de implicación cívica: participación en asociaciones cívicas de todos los tipos, en asociaciones para inmigrantes y en asociaciones de no inmigrantes. El artículo inves tiga si es más probable que los inmigrantes que vienen de países con mayores niveles de partici pación cívica participen en asociaciones y si es más probable que los inmigrantes que han vivido durante más tiempo y han estado más en contacto con un país de origen con mayores niveles de participación cívica entren en asociaciones. Los datos usados provienen de la Encuesta Española N acional de Inmigración (2007) y la Encuesta Mundial de Valores (2000 Valores ( , 2005 . Los resultados de las regresiones logísticas multinivel m uestran que es más probable que los inmigrantes que han pasado más tiempo en un ambiente más participativo en origen y que han estado más en contacto con esas sociedades se impliquen en asociaciones cívicas en el país de destino. Palabras clave: España, país de origen, participación cívica de inmigrantes, socialización T a b le 3. Multilevel logistic model -dependent variable: participation in all types of associations. T a b le 6. Summary of the hypotheses, variables used to test the hypotheses and the outcomes. T a b le I. Immigrants' participation in civic associations in Spain by country of origin. Participation in all Participation in associations Participation in types of associations for immigrants regular associations Sweden 48% 10% 45% Austria 40% 27% 27% U SA 37% 5% 34% U K 35% 19% 25% Belgium 32% 15% 25% Denmark 32% 8% 28% Philippines 32% 20% 22% Ireland 31% 8% 31% Uruguay 29% 12% 22% Mexico 26% 2% 25% Argentina 26% 8% 23% Switzerland 25% 7% 20% Chile 25% 6% 21% Germany 24% 1 1% 21% Venezuela 24% 10% 22% France 23% 5% 21% Italy 23% 8% 19% The Netherlands 23% 12% 18% Brazil 23% 10% 17% Algeria 21% 12% 13% Norw ay 20% 13% 13% Peru 20% 7% 16% Colombia 19% 7% 15% Poland 18% 12% 12% Ukraine 15% 9% 8% M orocco 15% 9% 10% Portugal 14% 2% 13% India 14% 14% 5% Lithuania 13% 13% 6% Russia 12% 4% 10% China 12% 9% 7% Romania 1 1% 6% 7% Bulgaria 9% 3% 7% Moldova 5% 0% 5% Total 19% 8% 15% T a b le 2. Descriptive statistics of variables. Minimum Maximum Mean SD Individual-level indicators Age of migration 1 1 88 30 12 Contacts home country 6 2 Time in Spain (years) 0 79 10 1 1 Speaking Spanish 4 2 Spanish nationality 0 Income 25 9000 1032 572 W o rk in g hours 168 40 13 Education 7 5 Male 0 Parenthood 0 1 1 Married 0 Country-level indicators Ethnic group dimension 25 1850 340 418 Membership 2000 10 96 48 22 G D P 2000 1541 40,000 18,193 1 3,457 T a b le 4. Multilevel logistic model -dependent variable: participation in non-immigrant associations. Model 1 Model 2 Model 3 B Odds ratio B Odds ratio B Odds ratio Intercept -2.145*** 0.1 17 -3.224*** 0.040 _ l 907*** 0.149 Country-level indicators Membership 2000(*I00) 0.020** 1.020 -0.008 0.992 Ethnic group dimension 0.875** 2.399 0.890** 2.435 G D P 2000 -0.003 0.997 0.000 1.000 Individual-level indicators Age of migration -0.036 0.965 -0.071 0.932 -0.346*** 0.707 Contacts home country 0.045 1.046 0.053 1.054 -0.166 0.847 Time in Spain 0.096 1.101 0.074 1.077 0.097 1.102 Spanish nationality -0.028 0.973 -0.014 0.986 -0.012 0.988 Speaking Spanish 0.104 l.l 10 0.1 14 1.121 0.128+ 1.136 W o rk in g hours -0.335*** 0.715 -0.315*** 0.729 -0.318*** 0.728 Income 0.125*** 1.133 0.122*** 1.129 0.120*** 1.127 Education 0.024 1.024 0.01 1 1.012 0.030 1.030 Married -0.029 0.971 -0.009 0.991 0.008 1.008 Parenthood 0.020 1.020 0.036 1.036 0.046 1.048 Male -0.391* 0.676 -0.387* 0.679 -0.379* 0.684 Interactions Contact home country x 0.420** 1.522 Membership 2000 (*100) Age of immigration x 0.485*** 1.624 Membership 2000 (*100) R 2 dich o 0.202 0.202 0.207 Notes: *** p < .001;** p < .01; * p < .05; + p < .1. T a b le 5. Multilevel logistic model -dependent variable: participation in association for immigrants. Model 1 Model 2 Model 3 B Odds ratio B Odds ratio B Odds ratio Intercept -1.663*** 0.190 -1.485*** 0.227 0.293 1.341 Country-level indicators Membership 2000 (*100) 0.007 1.007 -0.033+ 0.968 Ethnic group dimension -0.702 0.496 -0.758 0.469 G D P 2000 -0.032+ 0.969 -0.028 0.972 Individual-level indicators Age of migration -0.039 0.962 -0.010 0.990 -0.430 0.650 Contacts home country -0.096 0.908 -0.105 0.901 -0.310*** 0.733 Time in Spain 0.006 1.006 0.024 1.024 0.056 1.058 Spanish nationality -0.3 17 0.728 -0.366 0.694 -0.378 0.685 Speaking Spanish 0.248** 1.282 0.249+ 1.283 0.271** 1.312 W o rk in g hours -0.340*** 0.712 -0.358*** 0.699 -0.365*** 0.694 Income 0.066*** 1.068 0.069*** 1.072 0.068*** 1.070 Education 0.079 1.083 0.103 1.108 0.122 1.129 Married 0.088 1.092 0.044 1.045 0.082 1.086 Parenthood -0.01 3 0.987 -0.025 0.975 -0.014 0.986 Male -0.55 1 *** 0.576 -0.587*** 0.556 -0.586*** 0.556 Interactions Contact home country x 0.425* 1.529 Membership 2000 (*100) Age of immigration x 0.786** 2.195 Membership 2000 (*100) K 2 d ich o 0.123 0.174 0.169 Notes:*-1* p < .001; ** p < .01 ;* p < .05; + p < . 1. Acknowledgements We are indebted to Bogdan Voicu, H erm ann D ulm er and to several anonymous reviewers for help ful comments and suggestions on the article. Notes 1. D ata downloaded from the website: http://www.ine.es/prodyser/micro_inmigra.litm (accessed January 2010). World Values Survey 1981-2008 Official A ggregate v.20090901, 2009. W orld Values Survey Association (www.worldvaluessurvey.org). Aggregate File Producer: ASEP/JDS, Madrid. 3. Unfortunately, the data do not allow to distinguish between participation in immigrant asso ciations and associations for immigrants. Therefore, we have included membership in both types o f associations in the same category because the data provide valuable information about the propensity o f migrants to jo in associations. On the other hand, our analysis is limited to the immigrant population in Spain. However, the current results are relevant for other countries w ith similar opportunity structures regarding immigrants' civic participation. Political participation by immi grants depends on the community norms of participation ( Tillie, 2004) , on the number of immigrants and on national integration policies ( Voicu and Serban, 2012) . Spain shares with other countries in Southern Europe, such as Italy, Portugal and France, a generally low level of civic participation (see Hodgkinson, 2003: 40; Schofer and Fourcade-Gourinchas, 2001: 808) . Moreover, these countries are currently con fronted w ith a larger inflow of more, often illegal, immigrants, which makes the coun tries more similar regarding the numbers of immigrants. On the other hand, migration is a new phenomenon in the area (except France), and until recently, these countries lacked consistent national integration policies (Eurofound, 2007) . One strategy for coping with this issue was to offer support to NGOs that provide services to immigrants (Eurofound, 2007) . Therefore, there are many similarities between the Spanish case and other coun tries in Southern Europe, in terms of contextual factors with an impact on immigrants' participation. Thus, the current results are theoretically relevant not only for the Spanish case but also can be extrapolated to countries in Southern Europe as a whole. However, further research should not only focus on differences in civic participation among home countries but also analyse different host societies to check the interaction effect between the host and the home society on immigrants' civic participation. Funding This w ork was partially supported by the Romanian National Research Council (grant number PN -II-ID-PCE-2011 -3 -0210). A uthor biographies Malina Voicu is post-doctoral researcher with GESIS Leibniz Institute for the Social Sciences, D ata A rchive for the Social Sciences. H er main fields o f interests and current w ork are in social values, attitudes change and civic participation. Immigrant political adaptation in Canada: Some tentative findings Canadian Journal o f Political Science 15 1 Princeton U niversity Press The Civic Culture The practice o f politics in two settings: Political transferability among recent im mi grants to Canada J B Lack Canadian Journal o f Political Science 20 4 Age, resistance, and political learning in a new environment. The case o f Canadian immigrants J B Lack Bingham Powell G Jr Comparative Politics 20 1 The Private Roots o f Public Action: Gender, Equality, and Political Participation Scholzm B Um Sn Harvard University Press Cambridge, MA Integration, participation, identity: Immigrant associations in the province o f Milan M Caselli International M igration 48 2 Comm unity-based organizations and migration in N ew Y ork H Cordero-Guzm An City. Journal o f Ethnic and Migration Studies 31 5 Values and the social sciences: A global w orld o f global values? P Easter Easter P, Braun M and M ohler P (eds) Globalization, Values Change, and Generation: A Cross-National Perspective Leiden Brill Local Integration Policies fo r Migrants in Europe. Dublin: European Foundation for the Improvement o f Living and W orking Conditions Euro Political participation and political trust in Amsterdam: Civic communities and ethnic networks Tillie J Fennem A M Journal o f Ethnic and Migration Studies 25 4 Party identification and political adaptation o f A m erican migrants in Australia A Finifter Journal o f Politics 51 3 The role o f social networks in immigrant w om en's political incor poration E Gidegil D Stolle International Migration Review 43 4 Im migrant volunteering: A stepping stone to integration? N onprofit and Voluntary Sector Quarterly F Handy 38 Volunteering in global perspective V Hodgkinson The Values o f Volunteering. N ew Y ork and Dordrecht D ekker P and H alm an L Kluwer Academic/Plenum Publishers The stability o f political attitudes and behaviors across ado lescence and early adulthood: A comparison o f survey data on adolescents and young adults in eight countries M Hooghe Journal o f Youth and Adolescence 37 2 Political participation and associational life o f Turkish residents in the capital o f Europe D Jacobs K Phalet M Swyngedouw Turkish Studies 7 1 Pathways to voluntarism: Fam ily socialization and status transm is sion models T Janoski Social Forces 74 1 Developing citizens: The impact o f civic learning opportunities on students commitment to civic participation J Kliane S Sporte American Educational Research Journal 45 3 Civic skills building: The missing component in service programs? M Kirlin Political Science and Politics 35 3 M otivation fo r civic participation o f immigrants: The role ofp erso n a l resources, social iden tities, and personal traits. POLITIS-W orking Paper No A Kosic Oldenburg, Germany U niversity o f Oldenburg Persistence and change in political confidence over the life-span: Embedding life-cycle socialization context K Langton British Journal o f Political Science 14 4 Employment status and the attitudes and behavior o f higher status w om en volunteers C Bonjean Sex Roles 34 Immigrants and associations: A global and historical perspective J M Oya Journal o f Ethnic and Migration Studies 31 5 Bridge over troubled water? M igration and social capital P N Annestad Glh Svendsen G T Svendsen Journal o f Ethnic and Migration Studies 34 4 Political socialization R Niemi B Sobieszek A nnual Sociological Review 3 Civil society, migrant organizations and political parties: Theoretical linkages and applications to the Swedish context P Odmalm Journal o f Ethnic and M igration Studies 30 3 N ational context, religiosity and volunteering: Results from 53 countries S Ruiter De G Raaf N D American Sociological Review 1 2 The structural contexts o f civic engagement: Voluntary association membership in comparative perspective E Schofer M Fourcade-Gourinchas American Sociological Review 66 4 Im migrant organizations M Schrover Journal o f Ethnic and Migration Studies 31 5 Evidence o f the long-term persistence o f adults' political participation D Sears C Funk Journal o f Politics 61 1 Politics matters: Political events as catalysts for preadult sociali zation D Sears N Valentino American Political Science Review 91 Political incorporation among immigrants from ten areas o f origin: The persistence o f source country effects Simpson Bucker C International Migration Review 39 1 T Snijders M ultilevel Analysis: An Introduction to Basic and Advanced M ultilevel Modeling London Sage Naturalization, socialization, participation: Immigrants and (non-) voting Tam Clio W The Journal o f Politics 62 4 Social capital o f organizations and their members: Explaining the political integra tion o f immigrants in Amsterdam J Tillie Journal o f Ethnic and M igration Studies 30 3 The school's role in developing civic engagement: A study o f adolescents in twenty-eight countries J Torney-Purta A pplied D evelopm ental Science 6 4 M embership in voluntary organizations and democratic performance: European Post-Comm unist countries in comparative perspective N Valkov Communist and Post-Com m unist Studies 42 1 Voice and Equality: Civic Voluntarism in American Politics S V Erba Schlozman Kl H arvard U niversity Press Cambridge, MA Organizational patterns: Surinamese and Turkish associations in Amsterdam V Journal o f Ethnic and Migration Studies 31 5 Civic activation o f immigrants: An introduction to concep tual and theoretical issues. POLITIS-W orking Paper No A Triandafyllidou Oldenburg, Germany University o f Oldenburg Im migrant involvement in voluntary association B V Oicu M Serban Journal o f Ethnic and Migration Studies forthcoming H om eland interests, hostland politics: Politicized ethnic identity among Middle Eastern heritage groups in United States International M igration Review 42 2 The political resocialization o f immigrants: Resistance or lifelong learning? S W Hite Political Research Quarterly 61 2 The effects o f age and political exposure on the development o f party identification among A sia A m erica and Latino immigrants in the U nited States J W Ong Political Behavior 22 4 (xsd:string)
?:linkScore
  • 0.0 (xsd:float)
?:linkSource
  • GEFUREX pipeline (xsd:string)
?:linkingMethod
  • search (xsd:string)
?:mainEntity
?:name
  • WVS (xsd:string)
is ?:referenceMetadata of
?:toSource
  • gesis (xsd:string)
rdf:type