Citizenship and Socio-Political Integration: A Person-Oriented Analysis Among Russian-Speaking Minorities in Estonia, Finland and Norway

This survey study utilized a person-oriented approach to explore the patterns of socio-political integration among Russian-speaking minority group members in three neighboring countries in the Baltic area: Estonia (n = 482), Finland (n = 252), and Norway (n = 215). Three profiles were obtained in all countries: critical integration, separation, and assimilation. In the whole sample, critical integration was the most common acculturation profile. After the profiles were established, they were examined vis-à-vis citizenship and integration context to see, whether and to what extent, the objective (i.e., citizenship) and subjective (i.e., perceived social status and sense of belonging) socio-political integration of Russian-speakers corresponded with each other. Critical integration and separation were the most common profiles among participants holding national citizenship of the country of residence, while foreign citizenship was not related to any specific profile. Separation was rare among participants holding dual citizenship, but it was the most common profile among participants with undetermined citizenship. Also, intergroup context was associated with socio-political integration: critical integration and separation were the most common profiles of Russian-speakers in Estonia, critical integration and assimilation profiles in Finland, and assimilation profile in Norway. The results are discussed in relation to previous variable-oriented research and official integration policies of the countries studied.

argued for the centrality and the constructive role of dual (i.e., ethnic and national) identification in immigrants' social integration (Fischer-Neumann, 2014). Previous research from one of the present research contexts, Finland, echoes this idea. Larja (2017) studied 15-29 year old youth with foreign background and examined their social integration from the viewpoint of ethnic and national identification (cf., Hutnik, 1991; see also Dovidio, Gaertner, & Saguy, 2009, p. 9-10). The results indicated that most youth with Finnish citizenship were integrated (35%; high national and ethnic identification) or assimilated (32%; high national and low ethnic identification), while these profiles were not that common among youth with foreign citizenship (21% and 11%, respectively). Among foreign citizens, separation orientation (51%; low national and high ethnic identification) was more common than among Finnish citizens (20%).
When interpreting previous results related to national identification, it should be noted that in some studies (e.g. Ersanilli & Koopmans, 2010;Fick, 2016), civic identification with the country or society has been measured (e.g., "I feel that I am a member of the German society"), while other studies (e.g. Larja, 2017) have focused on identification with the ethnic majority group (e.g., "I feel that I am Finnish"). In European nation states, national identification is often very ethnically laden, but society as a whole creates more inclusive grounds for common civic identification (for discussion, see e.g., Brylka, Mähönen, & Jasinskaja-Lahti, 2015). In this study, we focus on civic national identification with society.
Importantly, when studying socio-political integration from the viewpoint of social identities, also perceptions of acceptance and social status need to be acknowledged. For example, Simon and Grabow (2010) studied Russian immigrants in Germany and showed the importance of perceived maltreatment for politicized dual identification to develop. When facing inequality and/or pressure to assimilate, ethnic minorities often transform their identities into more defensive or reactive forms (e.g., Jasinskaja-Lahti, Mähönen, & Liebkind, 2012;Mähönen, Jasinskaja-Lahti, & Liebkind, 2011;Verkuyten & Martinovic, 2012;Vetik & Helemäe, 2011). In other words, strong ethnic identification can be associated both with integrationist and separatist orientation in relation to engaging in contact with the national majority -depending on how immigrants perceive the prevailing intergroup relations in society (Barrette, Bourhis, Personnaz, & Personnaz, 2004; see also Verkuyten, 2006). Another reaction to perceived inequality among immigrants is national dis-identification, which compromises positive intergroup relations with the majority and socio-political engagement (Jasinskaja-Lahti, Celikkol, Renvik, Eskelinen, Vetik, & Sam, 2018;Jasinskaja-Lahti, Liebkind, & Solheim, 2009). Therefore, following previous research discussed above (e.g., Simon & Ruhs, 2008;Verkuyten, 2006;Verkuyten & Martinovic, 2012;Vetik 2018;Wright & Bloemraad, 2012), we approach Russian-speaking minority group members' socio-political integration through their sense of belonging in the ethnic ingroup and the society at large, accompanied by perceptions of group status (perceived discrimination, status legitimacy and permeability of group boundaries) and stance towards the national majority (attitudes towards and trust in majority). Finally, it should be acknowledged that the way immigrants see their possibilities for equal membership and belonging cannot be understood in isolation from legal recognition. Thus, as discussed in more detail below, in this study, we focus on both objective and subjective socio-political integration -recognition in terms of granted citizenship and subjective sense of membership (cf., Wright & Bloemraad, 2012) -and their interplay in different integration contexts.

The Relationship Between Naturalization and Integration
concluded that citizenship fosters immigrants' socio-political inclusion and political engagement. Hainmueller et al. (2017) found in Switzerland that naturalization also promoted immigrants' long-term integration. The design of Hainmueller et al. (2017Hainmueller et al. ( , see also 2015 was unique in its focus on close naturalization referendums, in which the naturalization decision was as good as random, enabling them to study the causal relationship between citizenship and integration outcomes. There is also valuable longitudinal evidence from Germany (Fick, 2016;Maehler et al., 2019) supporting the idea that citizenship acquisition promotes identification with receiving society and might also support the compatibility of ethnic and civic national identities.
However, not all immigration contexts are equally welcoming: there are substantial differences in the ease of naturalization processes, as well as in integration policies. In their study among Turkish immigrants in the Netherlands, France and Germany, Ersanilli and Koopmans (2010) studied whether naturalization is positively associated with socio-cultural integration measured as civic national identification, proficiency and use of the national language, and interethnic social contacts. They found this pattern of results only in France and Germany, where a certain degree of cultural assimilation has traditionally been required from new citizens. Context also played a role in a Simonsen's (2017) study in 14 Western countries: While citizenship appeared as an important marker of national belonging for national majority members in all countries studied, the largest difference was between Norway and Finland. The importance of citizenship was the least pronounced in Finland, while in Norway, the role of citizenship was stressed more than in any of the other European countries (Simonsen, 2017).
One way towards inclusiveness and formal recognition of dual identities is to allow immigrants' dual citizenship of country of residence and country of origin/emigration. Even within Europe, countries differ greatly in their views on dual citizenship. This has evoked also relevant research on whether allowing dual citizenship is a risk or an aid when it comes to integration. Ersanilli and Koopmans (2010) compared the situation of Turkish immigrants in the Netherlands, France, and Germany. In France, where dual citizenship is unconditionally allowed, the positive association between naturalization and socio-cultural integration was the strongest, compared to the other two contexts studied. In contrast, the strict restrictions regarding dual citizenship in Germany did not make German Turks any better integrated than their fellows in France. Ronkainen (2011) has pointed out that the key reason for nation states to restrict dual citizenship has been the expectation of undivided loyalty from the citizens toward their state. As recently shown in a study in Finland, Germany, and the Netherlands (Jasinskaja-Lahti, Renvik, Van der Noll, Eskelinen, Rohmann, & Verkuyten, 2019), immigrants' dual citizenship is indeed associated with majority group members' reserved attitudes, mistrust, and reluctance to assist immigrants in their integration process (see also Kunst, Thomsen, & Dovidio, 2019). However, as noted by Ronkainen (2011), dual citizenship has different meanings among the dual citizens themselves. In the Finnish context, he found that the functions of dual citizenship were mainly described in terms of transnationalism, rather than in terms of political affiliations or state loyalties.
Being questioned can feel unfair by immigrants themselves, leading to identity and attitudinal reactions (see Verkuyten & Martinovic, 2012). From this viewpoint, more research is clearly needed also on the less studied group of dual citizens.
is the only country that is not a member of the NATO, and Norway is the only country that is not a member of the EU. Second, as discussed in more detail below, the numbers of Russian-speakers and the inclusiveness of citizenship and integration policies vary between the countries (MIPEX, 2015). Third, the migration reasons of Russianspeakers in the countries studied vary: while family reunification and ethnic remigration have mostly characterized Russian immigration to Finland, work and family reunification have been the primary motivations of Russians to come to Norway (Sandnes, 2018;Sutela & Larja, 2015;Utlendingsdirektoratet, 2018). In Estonia, it was rather the state border than the people that moved: Russians in Estonia are internal migrants from the times when the country was a part of the Soviet Union. As a result, Russians in Estonia have lived there for generations and do not regard themselves as immigrants at all (see Brubaker, 1996).
As regards the situation of Russian-speakers in these countries, in Estonia, the share of foreign-origin population was 27 percent at the beginning of 2016 (Tammur, 2017). Most of these people are Russian-speakers, who constitute around 25 percent of the total population according to the 2011 census. After Estonia gained independence, ethnic Russians have lost their previously privileged status and have become a relatively deprived group with devalued group identity and problems with integrating into the society (Kus-Harbord & Ward, 2015). The strict, ethnically laden citizenship policies have had a negative impact on the integration of Russian speakers in the labor market as well as society in general, and as a reaction, this minority has shown strong defensive identities (Lauristin et al., 2012;Vetik & Helemäe, 2011).
In Finland, around seven percent of the whole population have an immigrant background, and around 20 percent of them speak Russian as their mother tongue (Statistics Finland, 2017). The intergroup relations between Russian immigrants and the Finnish majority are characterized by discrimination owing to historical conflicts (e.g., the wars between Finland and the Soviet Union in 1939-1940and 1941-1944, and Russians have been considered one of the least welcome immigrant groups (Jaakkola, 2009). Prior research among Russian-speaking immigrants has attested that ethnic discrimination is common (e.g., Larja et al., 2012) and it has detrimental effects on their identification with Finnish society (Jasinskaja-Lahti, Mähönen, & Ketokivi, 2012).
Of the three countries, Norway inhabits the smallest community of Russian-speakers. At the time of the data collection, immigrants accounted for ca. 13 percent of the total population in Norway, with only two percent of immigrants speaking Russian as their mother tongue (Statistics Norway, 2016). However, Russians were the fifth largest immigrant group and the largest non-EU group in the country before the so-called refugee crisis in 2015 (Olsen, 2015). Similar to Finland, a larger increase in Russian-speaking happened after the collapse of the Soviet Union in 1990. Systematic research on the integration of this group is limited, but as marriages are one of the main reasons to migrate to Norway from Russia, there is a body of research on intercultural marriages, transnational family ties, and the labour market situation of Russian women in Norway (e.g., Flemmen, 2008;Heyse, 2010;Munkejord, 2017). Based on these studies, it is clear that also in Norway, Russian-speakers face prejudice, and that their social status can be considered as relatively low.
As regards citizenship policies, in all these countries, extended residence and language ability are the basic requirements for citizenship. Dual citizenship of Russia is allowed only in Finland, with approximately one third of the Russian-speaking minority possessing dual citizenship . Until recently, Norway did not welcome or encourage dual citizenship, but this is changing. Estonia belongs to the countries least tolerant of dual citizenship in the European Union (Bauböck et al., 2013). Also more generally speaking, citizenship acquisition rates among the foreign-born are among the lowest in Estonia when compared to other European countries (Bauböck et al., 2013, see also MIPEX, 2015. In 2016, one third of the first generation Russian-spekers living in Estonia had Estonian citizenship, another one third were Russian citizens, one fifth had so-called undetermined citizenship (i.e., were not recognized by any state as its citizen), and the rest were citizens of other countries (often from the former Soviet Union) (Tammur, 2017). However, the share of Estonian citizens is increasing, and the share of undetermined citizenship is decreasing (Tammur, 2017;Estonian Ministry of Internal Affairs, 2018).
As regards the background of undetermined citizenship, the adoption of the Citizenship Act in 1993 (see Barrington, 1995) forced most Estonian-Russians to go through of naturalization process that made language proficiency as a criterion for acquiring citizenship. Many Russian-speakers regard this process as unjust and discriminatory, and seemingly well-intended efforts to increase Estonian language proficiency in schools is often seen as assimilationist (Vetik & Helemäe, 2011, see also Kruusvall et al., 2009). As a result, many Russian-speakers have not acquired Estonian citizenship. Estonia does grant social protection to all legal residents regardless of their citizenship status, and residents with undetermined citizenship can vote in local elections. Also at the level of practicalities, undetermined citizenship does not hinder everyday living in Estonia and makes it easier to travel to Russia (Nimmerfeldt, Schulze, & Taru, 2011). Nevertheless, undetermined citizens are not structurally integrated as they do not have equal political rights in terms of voting or standing for in national elections, or equal access to the major institutions such as education and labour market (Nimmerfeldt et al., 2011). Thus, in the present research, the special case of Estonian Russians with undetermined citizenship is particularly acknowledged.

Aim and Research Questions
Recently, Goodman (2015) evaluated the current state of research on the relationship between citizenship and immigrant integration and called for studies that would not necessarily be theory-confirming but theory-developing.
One possibility for theoretical development is to strengthen the line of research that acknowledges the mutuality of socio-political integration (see Berry & Sam, 2013), instead of approaching it merely though the strategies or skills of an immigrant individual. In this study, we are able to study the relationship between objective socio-political integration (i.e. "Am I legally 'one of us'?") and subjective socio-political integration by taking into account also perceived opportunities for integration (i.e. "Am I considered as 'one of us'? Do I feel that I am a full member of this society?"). Furthermore, one empirical possibility for specifying the existing theories further is to utilize a personoriented analysis. Namely, previous research on immigrant integration could be criticized for over-reliance on sample homogeneity: most studies have assumed that patterns at the level of the sample apply to all (or most) participants (for research examples, see Sam & Berry, 2016). Acculturation psychological research has increasingly started to acknowledge the value of a person-oriented approach and complemented the predominantly As pointed out by Verkuyten (2006), integration policies in one socio-political context can have a different meaning in another, and something that works in one country does not necessarily work in another. Moreover, there is research evidence showing that due to within-country and within-group variation, the integration profiles of immigrants rarely parallel the state's official integration policy (Berry et al., 2006; see also Horenczyk et al., 2013). Thus, we start the analysis by utilizing a person-oriented approach and exploring the characteristics of socio-political integration of Russian-speaking minority group members separately and jointly in the three countries studied (RQ1).
More specifically, we will focus on how Russian-speaking minority group members perceive the status hierarchies Citizenship and Socio-Political Integration 58 and intergroup boundaries in the societies they are living in, what are their identification patterns, and how do they evaluate national majority group members (for a similar approach to socio-political integration, see e.g., Verkuyten & Martinovic, 2012). Then, with pooled data, we will examine the extent to which different integration profiles are represented in Estonia, Finland and Norway (RQ2) and among participants holding national, foreign, dual, or undetermined citizenship (RQ3).

Method Participants and Procedure
This study utilizes Estonian, Finnish, and Norwegian survey data collected for a broader cross-cultural research project MIRIPS i between 2012 and 2016. The Estonian data was collected among a random sample of Russianspeaking adult population in May 2015. Out of the participants, 81% were born in Estonia and represent second or third generation Russians. More than 90% reported having lived in Estonia for more than 25 years, i.e., for the whole post-Soviet period. The questionnaires were completed during face-to-face interviews. In Finland, data was collected between June and November 2012, by approaching a representative sample of Russian-speaking adults by mail. All participants had been born in the Soviet Union/the Russian Federation and had moved to Finland before 2008. The Norwegian data was collected between November 2015 and January 2016. From all legal residents of Norway, adults who identified themselves as Russians were randomly selected and invited by phone to take part in the study. Upon agreeing, a link to an electronic questionnaire was sent. In all countries, participation in the study was fully voluntary and anonymous.
The response rate was 65% for Estonia, 39% for Finland and 33% for Norway. Due to data attrition, the final sample used in the analyses was not representative of the populations. However, the response rates can be considered as typical for survey studies conducted in these countries. In the final sample there were 955 participants (n = 482 for Estonia, n = 254 for Finland, and n = 219 for Norway).
The socio-demographic characteristics of the sample are presented in Table 1. The majority of participants in all countries were female and, on average, middle-aged. As regards citizenship, there were differences between the countries, χ 2 (6) = 502.157, p < .001: there were more national citizens in the Estonian sample compared to Finland and Norway, and the number of foreign citizens differed between all three countries (Norway > Estonia > Finland).
Dual citizenship is not allowed in Estonia, and in Norway it is only a special case: thus, also in our data there were more dual citizens in Finland than in Norway. In Estonia, the citizenship of approximately every fifth participant was undetermined.

Measures
In this study, we utilized measures used in previous social and acculturation psychological research and backtranslated them from English to Russian. The descriptive statistics of each scale by country are presented below.
Perceived discrimination was measured with five items (Berry, Phinney, Sam, & Vedder, 2006)  Perceived social status of the ingroup in relation to that of the national majority group was measured using a single item adapted from Louis, Duck, Terry, Schuller, and Lalonde (2007)  In addition, a set of socio-demographic background variables (i.e., sex, age, marital status, education and employment status) were used in the following analyses (see Table 1). For correspondence and cluster analyses, age was recoded into six categories (18-24, 25-34, 35-44, 45-54, 55-64, and 65-74). Educational background was transformed into three groups: Basic, Secondary and Higher Education. In Estonia and Norway, data on education was gathered by specific educational categories (see Table 1). Data on education was gathered in Finland by years of schooling, and so in order to match Estonian and Norwegian data as well as possible, less than 11 years of schooling was considered as Basic Education, between 11 and 15 years of schooling was considered as Secondary Education, and over 15 years was considered as Higher Education. In the transformed variable used in the analyses, less than secondary education was considered as Basic, and university or college level education was considered as Higher Education. To compare this classification to the International Standard Classification of Education (ISCED, 2011), Basic Education is comprised of school levels up to ISCED Level 2, Secondary Education of ISCED Levels 3 to 4, and Higher Education of ISCED Levels 6 to 8 (5 being virtually non-existent). ISCED classification of lower secondary education is considered as Basic Education, and higher secondary education is included as Secondary Education.

Data Analysis
First, a confirmatory factor analysis was conducted to test the four-factor structure of ethnic identification, national identification, perceived discrimination and outgroup trust items. We also tested this four-factor model against a one-factor baseline model, to show the distinctiveness of the four latent constructs. After this, we performed multivariate measurement invariance tests to see if the four-factor model holds across the three countries.
To analyze the general pattern of relationships between the eight variables tapping socio-political integration and socio-demographic characteristics, we performed a multiple correspondence analysis (MCA) with principle variable normalization (SPSS 25) separately for each country. Next, to identify integration profiles, the object score variables obtained from MCA were utilized in a two-step cluster analysis with log-likelihood distance measure and Schwarz's Bayesian Criterion Clustering (Schwarz, 1978). Finally, based on the preliminary country-specific analyses, we ran correspondence and cluster analyses in pooled data, and used cross-tabulation and chi-square tests to examine the distribution of participants to different profiles within and between countries and within each citizenship category.

Confirmatory Factor Analysis
Confirmatory factor analysis of items used to measure ethnic identification, national identification, perceived discrimination and outgroup trust was run in Mplus 8.1 with robust maximum likelihood estimator. Goodness of fit indices indicated that a four-factor model had a relatively good fit with the data, χ 2 (98) = 285.28, p < .001; RMSEA = .05; CFI = .96; TLI = .96; SRMR = .04, unlike a single-factor model which did not fit the data, χ 2 (106) = 4226.95, p < .001; RMSEA = .20; CFI = .20; TLI = .09; SRMR = .25. Thus, the factors were distinctive at the level of the whole sample. Next, we tested the factorial equivalence of these four measures across the three countries. When testing configural invariance, the model fit was adequate, χ 2 (294) = 635.10, p < .001; RMSEA = .06; CFI = .95; TLI = .93; SRMR = .05, and we also found support for partial metric invariance, χ 2 (315) = 566.01, p < .001; RMSEA = .05; CFI = .96; TLI = .95; SRMR = .06, after setting two items measuring discrimination to covary (i.e. <national majority group members> have teased or insulted me because I'm a Russian immigrant; <national majority group members> have threatened or attacked me because I'm a Russian immigrant). Thus, we concluded that the factorial structure works well enough in all three countries for the purpose of using sum scores of ethnic identification, national identification, perceived discrimination and outgroup trust in the following analyses.    While also socio-demographic characteristics were included in the analysis as background variables, the theoretical interpretation ii of the profiles was based on the variables tapping socio-political integration from the viewpoint of perceived status and sense of belonging. In all clusters, i.e. profiles in all countries, ethnic identification was quite high, but there were some interesting differences in the other variables forming the profiles iii . For example, regardless of the integration profile, Russian-speakers in Estonia saw their ingroup's status and its legitimacy in a more negative light than their co-ethnics in Finland and Norway. However, the most notable and systematic finding was that in all countries, the pattern of the profiles found was very similar.

Integration Profiles: A Country-Specific Analysis
Namely, the first profile was characterized by highest levels of perceived discrimination, lowest levels of perceived status, status legitimacy and permeability of group boundaries, high ethnic identification, lowest national identification, lowest outgroup trust and most critical attitudes towards the national majority. In other words, participants in this profile identified mainly with their ethnic ingroup and took a negative stance towards the majority that was perceived to treat the ingroup in an unjust manner. In contrast, the second profile was characterized by lowest levels of perceived discrimination, highest levels of perceived status, status legitimacy and permeability of group boundaries, high ethnic identification, highest national identification, highest outgroup trust and most positive attitudes towards the national majority. Thus, these participants identified relatively highly with both ethnic and national groups and took a very positive stance towards the majority that was seen to grant access to the society.
In the third profile, ethnic identification was again high, and the mean values of the rest of the variables fell in between the mean values present in the two previous clusters. In other words, this profile was characterized by high ethnic but moderate national identification, paired with critical views about the majority group and ingroup's possibilities to be considered as an equal member of the society.
Based on previous acculturation and social psychological research (e.g., Berry et al., 2006;Berry & Sam, 2013;Bourhis et al., 1997;Horenczyk et al., 2013), these profiles were named as Separation, Assimilation, and Critical Integration, respectively. The third profile was characterized by orientation towards both ethnic and national groups, thus resembling the integration orientation defined in previous research (cf., Berry et al., 2006). However, this profile was also characterized by critical views of the ingroup's position: participants in this profile perceived higher levels of discrimination and lower social standing of their ingroup than those in the assimilation profile, but they were also more concerned about the legitimacy of this group's status position than assimilationists. Following similar observations from Estonia by Kruusvall and colleagues (2009), we decided to call this profile critical integration. It should be noted that ethnic identification was high in all the profiles: even within assimilation profile, ethnic identification was equally high to national identification. This speaks for the salience of ethnic identity among Russian participants in all three countries. Given that the pattern of the results was so similar in the three countries studied, we moved to the final analyses where the pooled data set was utilized.

Integration Profiles vis-à-vis Citizenship and Country of Residence
With pooled data, MCA of socio-demographic characteristics and the eight variables tapping socio-political integration produced two dimensions explaining 42.9 percent of the total variance, out of which 26.9 percent was explained by the first dimension and 16 percent by the second dimension. Again, a two-step cluster analysis of the object scores obtained from MCA identified three clusters with an adequate silhouette measure of cohesion and separation (0.6). The interpretation of the clusters, i.e., socio-political integration profiles, was the same as in the country-specific analyses presented above, and the descriptive statistics of the variables forming the three profiles are presented in Table 2. To validate the meaningfulness of the profiles with another method, we tested whether the mean values of each variable differed statistically significantly between the profiles in the pooled data. According to post-hoc comparisons, F(2, 952) = 9.22, p < .001, η 2 = .019, the level of ethnic identification was higher in separation profile (M = 4.45, SD = .04) than in assimilation (M = 4.18, SD = .05) and critical integration profiles (M = 4.27, SD =.03) at p < .001 level, but there was no difference in the mean level of ethnic identification between assimilation and critical integration profiles (p = .468). More generally speaking and confirming the results of cluster analysis, the other variables of interest differed statistically significantly (p < .001) between acculturation profiles in a post-hoc ANOVA with Bonferroni correction. At the level of the whole sample, critical integration was the most common profile (49.5%), followed by separation (28.1%) and assimilation (22.4%) profiles.
The final analytical step needed to answer our research questions was to examine the integration profiles vis-àvis country and citizenship status. Table 3 presents the cross-tabulation between country and integration profile, χ 2 (4) = 264.07, p < .001. Numerically, critical integration was the most typical profile in Estonia and Finland. In Estonia, separation profile was slightly rarer than critical integration. In Finland, there was no statistically significant difference between the prevalence of critical integration and assimilation profiles. In Norway, assimilation was the most common profile, followed by critical integration. The main differences between the countries, χ 2 (4) = 264.035, p < .001, as presented in Table 4, boil down to separation and assimilation profiles. Separation was relatively more common in Estonia compared to Finland and Norway, and assimilation was relatively more common in Norway compared to Estonia and to a lesser degree, Finland. Note. Each subscript letter denotes a subset of two-step cluster number categories whose column proportions do not differ significantly from each other within each country at the .05 level. Table 5 presents the cross-tabulation between citizenship and integration profile, χ 2 (4) = 115.351, p < .001. Critical integration and separation were the most common profiles among participants holding national citizenship. Foreign citizenship was not clearly associated to any specific profile, but only differentiated between critical integration and assimilation profiles. Among participants holding dual citizenship, critical integration was the most common profile, followed by assimilation. While separation was rare among participants holding dual citizenship, it was the most common profile among participants with undetermined citizenship. However, also a considerable number of participants with undetermined citizenship represented critical integration. When interpreting the results related to citizenship, it should be remembered that in our data, all participants with undetermined citizenship resided in Estonia, and dual citizens mainly came from Finland (with some from Norway). Statistical analysis of acculturation profiles simultaneously by citizenship and country was not meaningful due to limited sample size. Note. Each subscript letter denotes a subset of two-step cluster number categories whose column proportions do not differ significantly from each other between each country at the .05 level. Note. Each subscript letter denotes a subset of two-step cluster number categories whose column proportions do not differ significantly from each other within each citizenship category at the .05 level. Dual citizenship is not allowed in Estonia, and in Norway, it is only an aberration.
Undetermined citizenship, in turn, only concerns Russian-speakers in Estonia.

Discussion
This was one of the very few studies that utilized a person-oriented approach to socio-political integration and examined it vis-à-vis with citizenship (however, see Lauristin et al., 2012) and societal context. With this approach, it was possible to depict meaningful subgroup differences within the socio-political integration of Russian-speaking minority group members in the Baltic area. Before turning to the role of citizenship and our country-specific findings, a few words regarding the overall picture are needed, as the general pattern of socio-political integration profiles was similar between the three countries studied.
In previous studies, integration is often found to be the most common acculturation orientation among immigrants (Berry et al., 2006), at least when conceptualizing it as relatively high involvement in both ethnic and national cultures (cf. Snauwaert et al., 2003). This is what we also found when approaching socio-political integration from the viewpoint of mutual intergroup relations: through perceived status and sense of belonging among Russianspeakers in the Baltic region at the level of the whole sample. More specifically, our data spoke for critical integration, as orientation towards both ethnic and national cultures was paired with a critical view on ingroup's social standing. This interpretation echoes the findings of another cluster-analytic study conducted among Russian-speakers in Estonia. Namely, Kruusvall and colleagues (2009) found participants representing an integration profile to be quite well adapted but still not fully oriented towards participation in the larger society, as they did not regard civic engagement to be meaningful in the context of strong assimilation pressure from the majority group. Thus, Russianspeaking minority members in the Baltic region seem to take a critical stance towards larger society, even when they are quite well integrated in terms of involvement in both ethnic and national cultures.
Also in line with most previous research, we did not find a marginalization profile, in which people would be oriented towards neither ethnic nor national cultures (see e.g., Berry et al., 2006; but for an exception among Russianspeakers in Estonia, see Kruusvall et al., 2009). However, speaking for the previously attested context-dependency of acculturation orientations (e.g., Jasinskaja-Lahti, Horenczyk, & Kinunen, 2011;Phinney et al., 2001), there were some important differences between the countries studied. While critical integration and separation were the most common profiles in Estonia, in Finland, critical integration and assimilation profiles were the most typical ones, and in Norway, assimilation was found most often.
This pattern of results can be seen to reflect the different situations of the Russian-speaking minority in the countries studied. Even though there is value in looking at the general picture of relationships between citizenship and socio-political integration, from an acculturation and social psychological viewpoint it is also interesting to discuss the intergroup contexts in which specific relationships occur (cf., Verkuyten, 2006). It should be noted Finland and Norway shared the fourth place, while Estonia came 22nd of the 38 countries compared. Also in our data, there were similarities between Finland and Norway: assimilation profile was found often in both countries (cf. Larja, 2017, for Finnish context), even though it was more pronounced in Norway. Based on MIPEX indicators, it is difficult to make strong claims about the differences between Finland and Norway as integration contexts.
However, regarding the factors central to the present study, the state of affairs related to anti-discrimination and access to (especially dual) nationality is slightly better in Finland than in Norway. Also, there might be somewhat higher pressure to assimilate in Norway (cf. Simonsen, 2017). In the official integration program of Norway, immigrants are expected to show high Norwegian proficiency and good understanding of the society (MIPEX, 2015), while in Finland, integration is officially portrayed as a mutual, two-way process (Government of .
There might be also more practical reasons for the differences found between Norway and Finland. Namely, it might be easier to maintain Russian culture by being in touch with family and friends when living among a bigger group of Russian-speakers in Finland, closer to Russia. Thus, in the light of all this information, it is understandable that assimilation profile was the most typical acculturation profile among the small group of Russian-speaking immigrants in Norway.
In Finland and Norway, Russian-speakers represent relatively small and recent immigrant groups, with possibly positive expectations about their future prospects in the new home country. In contrast, Russian-speakers in Estonia represent a big, established minority group. Despite having lived in the country for generations, the position of Russian-speakers continues to be disadvantaged and segregated (MIPEX, 2015;Vetik & Helemäe, 2011). For example, Kus-Harbord and Ward (2015) reported ethnic identification and perceived ingroup devaluation to be negatively associated with adaptation and intergroup relations with majority Estonians. Further, in another personoriented study by Kruusvall and colleagues (2009), every third Russian-speaker in Estonia represented a separation profile, integration profile was characterized by a critical stance towards societal participation, and no clear-cut assimilation profile was found. Thus, our country-specific as well as comparative findings corroborate previous variable-and person-oriented research that has brought forward the difficult situation of Russian-speakers in Estonia. From the viewpoint of mutual integration, the core of the problem is in the discordance between majority and minority groups: Russian-speakers feel pressure to assimilate and see their social position as unjust, and react to this with distancing themselves from the Estonian society (cf. Bourhis et al., 1997;Horenczyk et al., 2013).
As regards citizenship, critical integration and separation were the most common profiles among participants holding national citizenship of the country of residence. Recent studies utilizing a variable-oriented approach (e.g., Fick, 2016;Hainmueller et al., 2015Hainmueller et al., , 2017Verkuyten & Martinovic, 2012) have suggested national citizenship to promote immigrant integration, and also our person-oriented results pointed to this direction. However, in our data, national citizenship was also associated with separation profile, and both ethnic and national identities were relatively pronounced. The high number of those with national citizenship in the separation profile shows that objective and subjective socio-political integration do not always go hand in hand. Also importantly, gaining national citizenship does not need to require assimilation or even a non-critical integrationist orientation. Instead, we argue that our results speak for the politicization of the identities of Russian-speakers in the three countries studied: even among naturalized immigrants or minority group members, strong ethnic and national identities and a critical stance towards the ingroup's position are needed for challenging inequalities effectively (see, e.g., Simon & Grabow, 2010;Simon & Klandermans, 2001). Our analysis may have identified the potential agents of social change especially in Estonia, where separation profile and national citizenship were relatively more common than in the other two countries.
It was somewhat surprising to find that foreign citizenship was not strongly related to any specific profile, as could have been expected particularly for separation, based on previous research (e.g., Larja, 2017). One explanation for this may be found from research on psychological citizenship (i.e., the subjective sense of being a citizen) (Sindic, 2011). Namely, immigrants without a legal recognition of their countries of residence can develop different forms of affiliation with society, and this way form alternatives to recognized national membership (Sindic, 2011).
Thus, foreign citizenship does not necessarily support nor prevent subjective socio-political integration. Alternatively, our finding may support the notion by Ronkainen (2011) that foreign citizenship may serve a rather pragmatic function for many immigrants, while being unrelated to perceived status and feeling of belonging. Yet another, more technical reason is that the share of participants with foreign citizenship differed substantially in different countries. This made it impossible to fully disentangle context-and citizenship-related patterns in integration profiles.
When it comes to the special case of Russian-speakers with undetermined citizenship in Estonia, they mostly represented separation profile. Also Nimmerfeldt and colleagues (2011) found Russian-speakers with undetermined citizenship to be the least structurally integrated, when looking at their rights and access to education and labour market. Lauristin and colleagues (2012), in turn, found people with undetermined citizenship to be often linguistically, culturally and politically poorly integrated. In his integration report, Vetik (2015) found that people with undetermined citizenship in Estonia had low trust in local institutions, but the majority of them would have still liked to obtain Estonian citizenship. One could argue that there is a will but no way to full societal membership, without meeting the strict language requirements for Estonian citizenship (see also Nimmerfeldt et al., 2011). The situation is problematic particularly in North-East of Estonia, where opportunities to practice Estonian language are limited, as Estonian Russians form about 80% of the population.
Finally, our findings regarding dual citizenship shed new light on this much debated but rarely studied group of immigrants. Namely, dual citizenship was most closely associated with critical integration profile, followed by assimilation profile. As we found no evidence of Russian-speakers with dual citizenship to isolate from broader society, the concerns related to the allegedly divided loyalties of immigrants with dual citizenship especially in the Finnish context (e.g., Jasinskaja-Lahti et al., 2019;Kananen, Ronkainen, & Saari, 2018;Ronkainen, 2011) seem exaggerated. Also recent research from Germany has suggested that tolerating dual citizenship does not jeopardize immigrants' identification with their country of residence (Maehler et al., 2019).
Also the caveats of this study should be discussed. In her evaluation of studies focusing on citizenship and integration policies, Goodman (2015Goodman ( , p. 1905) crudely states that "what scholars know about the effects of immigration and citizenship policy is subject to data and sample selection". Indeed, limitations related to causality and generalizability are central also for the present study. Using observational data, our aim was not to test if citizenship affects socio-political integration. Instead, we leaned on causally robust studies that give a reason to expect naturalization to promote immigrant integration (Hainmueller et al., 2015(Hainmueller et al., , 2017, and linked our research to previous studies on the integration of Russian minority in Baltic region. However, future research should also keep the other causal direction in mind: in their study with longitudinal and representative data, Maehler and colleagues (2019) found immigrants who were already naturalized as well as immigrants aiming to become naturalized identify more strongly with Germany than non-naturalized respondents.
While one of the strengths of our study was to utilize data on the same minority group in three countries, with nonrepresentative samples and only three countries, it is not possible to make broad generalizations or robust crosscultural comparisons. Also, due to limited and varying sample sizes and the different citizenship policies in the three countries studied, the research design was not optimal for between-country comparisons. Thus, we could not simultaneously examine the integration profiles by citizenship category and by country. Even though this research was designed in close collaboration between our teams in Estonia, Finland and Norway, due to practical reasons, the data collection could not be executed simultaneously. However, during the time of data collection, there were no such large-scale changes in political power relations or integration policies that would significantly limit the interpretation of our results. Finally, it should be noted that we were not able to control for the length of stay or the immigrant generation of our participants, as we did not have equivalent indicators in all countries. Besides avoiding these limitations when designing future research, we would like to call for studies with an even broader scope of variables tapping integration. For example, language proficiency and social contacts with national majority group members and co-ethnics could not be included in the present study, although being known as central indicators of integration (e.g., Berry et al., 2006;Ersanilli & Koopmans, 2010).
When thinking about the broader message of this study, it is noteworthy that ethnic identification was high in all three acculturation profiles, while there were substantial differences in the other variables tapping perceived status and sense of belonging. This highlights the importance of including a variety of indicators in studies of socio-political integration. The picture painted by the results would have been much different if we would have focused, for example, only on the identification conceptualization of integration (cf. Hutnik, 1991;Maehler et al., 2019). Including perceived social status in the analysis of citizenship is important because opportunities for equal membership in society are not just perceptions in the heads of immigrants, but reflect societal hierarchies and mutual intergroup relations (cf., Berry & Sam, 2013). For example, in their comparative study, Wright and Bloemraad (2012) concluded that immigrants living in countries characterized by high access to citizenship and recognition of multicultural values felt included by and trusting of their fellow residents and government institutions. Naturalization policies can be considered to be "at the heart of a state's integration policy" (Huddleston & Vink, 2015, p. 1), but in the light of our findings, efforts still need to be made for Russian-speakers in the Baltic region to be able to feel that they are in the heart of the society.

Notes
i) There are two previous publications in which Estonian and Finnish (Renvik, Brylka, Konttinen, Vetik, & Jasinskaja-Lahti, 2018) and Estonian, Finnish and Norwegian (Jasinskaja-Lahti et al., 2018) data from the project was used. The commonalities of these studies are related to a focus on perceived discrimination and national identification, but the present study is the first to focus on citizenship and a broad array of indicators of socio-political integration, and to utilize a person-oriented approach.
ii) Speaking for the robustness of the findings in each country and in the whole sample, the interpretation of the profiles remained similar also when leaving socio-demographic background variables out from the analysis.
iii) Except for ethnic identification, differences in variable means between the profiles were significant at p < .01 level.

Funding
The Finnish part of this research was supported by the KONE Foundation (grant numbers 31-219 and 4704917). The Estonian and Norwegian parts of the research were supported by the EEA funding through the Research Council of Norway (grant EMP 138).