Social network site usage and personal relations of migrants

: In this study, we examine the relation between social network site (SNS) usage and the personal networks of immigrants, using a unique dataset composed of a representative sample of immigrants living in the Netherlands. In theory, SNSs can be a helpful tool for immigrants, because they may help establish social ties in the destination country and help maintain ties with people in the country of origin. We examine whether this is also true in practice by analyzing whether the frequency of using two SNSs—Facebook and Hyves (a Dutch SNS)—is associated with the number of ingroup and outgroup ties, as well as the quality of social relations. In addition, we test whether general emotional disclosure boosts the effect of SNS usage on the quality of relationships. We find that SNS usage is associated with more outgroup ties, but not with more ingroup ties. Our analyses also show that SNS usage is associated with greater quality social relationships among migrants. Contrary to our expectations, we found no interaction between general emotional disclosure and SNS usage on satisfaction with social relations. The implications of these findings are discussed.


Introduction
Social Internet applications, and especially social network sites (SNSs), are inherently connected to people's social relationships and networks.However, there is considerable debate about how the Internet affects social networks exactly (and vice versa).Several studies have looked at whether the rise of the Internet has decreased social relations, increased them or left them unaffected.Earlier studies resulted in ambivalent empirical evidence: some found a negative effect of Internet usage on offline social networks [1,2]; others found a positive effect [3].Recent studies have been more consistent.The results differ according to the measure of social networks or relationships under study, but most studies show positive effects [4][5][6][7] and some show non-significant effects [8].According to data from the World Internet Project [9], where respondents are asked directly about the effects of their Internet usage on contact with others, most individuals indicate that their contacts with other people remained the same, and there are more individuals who report increased contact than there are individuals who report decreased contact due to their Internet usage; this is consistent across different types of contacts.Wang and Wellman [10] found that heavy Internet users had more offline friends than light users in 2007, but the two groups showed a similar (positive) change in the number of offline friends between 2002 and 2007.
There are many questions about the relation between Internet usage and offline social networks that remain unresolved.For example, it is fairly unclear how Internet usage is associated with different types of ties in social networks.Therefore, a few recent studies have started distinguishing between types of ties, such as strong and weak ties [7,11] and bridging and bonding ties [5].Furthermore, it remains unclear whether Internet usage is related to the (perceived) quality of relationships.These questions have become even more relevant since the rapid increase of the popularity of social network sites.
In the present paper, we contribute to this literature by analyzing representative data of first generation immigrants in the Netherlands.These data are unique in the sense that they include detailed measures of Internet usage, as well as coverage of all immigrant groups in the Netherlands.We believe that migrants are a very instructive group for those interested in studying the consequences of the Internet on interpersonal relationships and communication.After migration, people face a double challenge: finding ways to maintain their existing relationships (family members and friends in the country of origin) and establishing new social ties in the destination country.We will argue that the usage of social network sites may play a role in dealing with this situation.
In this contribution, we answer the following research questions: (1) Is SNS usage related to the number of ingroup and outgroup ties in the core discussion networks of migrants?(2) Is SNS usage related to the perceived quality of social interactions among migrants?Additionally, (3) do migrants with high levels of emotional disclosure profit more from SNS usage than those with low levels of emotional disclosure?These topics have only been studied scarcely, while they may have important implications for acculturation processes, especially given the sharp rise of computer-mediated social interactions in the past two decades and given the fact that successful acculturation is predicted by the degree of contact with the native population [12].

Previous Research and Hypotheses
Social networks do not form and develop randomly.Studies have shown that individuals tend to initiate relationships with those who are similar to them, which is known as the homophily principle [13][14][15][16].This tendency to connect to similar others has been found with regard to values, attitudes, beliefs, socio-demographic characteristics (e.g., age, ethnicity) and religion, among others.In the present paper, we focus on one aspect of homophily: national origin.Race and ethnicity are known to play important roles in social network formation across different types of relationships; a review by McPherson et al. [13] showed that, in the United States, there is strong race and ethnicity homophily in marriages, close-confidant relations, schoolmate friendships and work relations.In other words, the formation of both strong and weak ties depends, among other things, on the ethnicity of a person.
Homophily tendencies are not equally strong in all contexts.Research by Mollenhorst, Völker and Flap [17] showed that voluntariness is a parameter of social contexts that affects homophily.The greater the extent to which participation in a certain social context is chosen voluntarily, the greater the likelihood that those interactions result in homophilous ties in one's personal network.This is intuitive: the preference for similar others cannot manifest itself in situations with little freedom of choice.If this is true, then homophily should be especially visible online, because the Internet puts very few restrictions on choosing one's interaction partners (unlike schools, work places, neighborhoods, etc.).In the case of online dating, Skopek et al. [18] showed indeed that (educational) homophily plays a major role in the selection of potential partners.
The Internet and specifically social media have the potential to affect ingroup relationships, here defined as ties to others with the same national origin, among immigrants, because they facilitate contact with those who are geographically distant [4,5,[19][20][21].For instance, Celene and Esperanza [19] reported that the Internet allowed immigrants to create virtual communities and maintain interactions with people from their places of origin.Brandtzaeg [4] concluded that the Internet "offers free and easy communication with family, friends and acquaintances regardless of time and place."Thus, SNS usage among immigrants should enhance contact with close ones who remained in the country of origin or, in the case of remigration, close ones who remained in the destination country [21].In the case of older migrants, these close ones may also include children who stayed in the country of origin or who moved to other countries [22].
Ingroup ties are not necessarily just to others who remained in the home country.The Internet provides opportunities to meet fellow immigrants from the same country of origin, as well as a means to develop ties with these fellow immigrants.A qualitative study of teenage immigrants in Israel [23] found that online interactions with co-ethnic peers provided several helpful resources to these teenagers, mainly because they were facing many similar problems.
This does not imply a negative association between SNS usage and outgroup contacts (i.e., with individuals with a different national origin).Several scholars argued that the Internet offers opportunities to establish new contacts in the host country, which should lead to easier adaptation [19,20,[24][25][26][27].For instance, Ye [25] found that immigrants who communicate more frequently with locals through the Internet adapt more successfully.Similarly, Chen [26] found that those who communicate more frequently with natives through the Internet showed greater social, cultural and psychological adaptation.Grasmuck et al. [27] provide a theoretical reason for this effect, arguing that SNSs promote the formation of inter-ethnic relationships, because they allow people to keep their personal information hidden, which should, in turn, increase the chances of inter-ethnic relationships being formed.In addition, Elias and Lemish [23] found that teenage migrants experimented with online friendships with locals prior to establishing offline ties.Thus, we formulate the following hypotheses: H1: SNS usage among immigrants is associated with having more ingroup ties in one's personal network.H2: SNS usage among immigrants is associated with having more outgroup ties in one's personal network.
An important theoretical mechanism that helps understand the relation between SNSs and personal relationships is self-disclosure, defined as verbal and non-verbal communication revealing information about an individual [28].Prior studies have shown that online self-disclosure has positive effects on both the quality of existing relationships and the formation of new contacts [6,29,30].Before the rise of SNSs, Wellman et al. [3] found that online communication represented a useful way to keep in contact with others between face-to-face meetings: "network members become more aware of each other's needs and stimulate their relationships through more frequent contact; exchange songs, pictures, and other files; and make online arrangements to meet in person and by telephone."This positive effect of Internet usage on the quality of ties should even be stronger in the context of SNSs, because they are designed to accommodate self-disclosure (e.g., through "status updates"), and self-disclosure is considered essential for relationship development and establishing feelings of intimacy [31,32].Some Internet applications stimulate self-disclosure because of reduced visual, audio and contextual cues.As a result, people "can deliberately edit their utterances, control their audiences, and allocate cognitive resources to the information they are sharing selectively" [28], which should make people more comfortable with sharing intimate information [33].
A related, but distinct, reason why SNSs enhance the quality of relations is that they facilitate easy relationship maintenance and, hence, cause strengthening of ties.Ellison et al. [5] found that "Facebook intensity" was related to "bonding social capital" (emotionally close relationships), and they hint at tie maintenance as a possible explanatory mechanism.For instance, the birthday alert on Facebook makes it easier to remember friends' birthdays and should generally increase the number of happy birthday wishes received.This can contribute to perceiving one's social network as attentive and caring and, subsequently, to greater perceived strength of ties.Qualitative studies of migrants have provided support for this idea.Khvorostianov et al. [22] found that, through its function of maintaining and extending social networks, the Internet solved problems of loneliness and isolation in a sample of older immigrants in Israel.Ogan and Ozakca [21], who studied individuals who remigrated to Turkey, concluded that the Internet was used to "keep personal relations warm".
A third reason why SNSs enhance relational quality is that this type of Internet usage (along with other online activities, like searching for information and visiting online fora) facilitates the exchange of resources and the mobilization of social support.In others words, SNSs usage can make ties more resourceful, thereby increasing their usefulness for a certain actor.Studies of computer-mediated communication have shown how social support is provided through online support groups [34,35].More recently, studies found that also other types of applications, notably SNSs, can provide people with social resources [36][37][38].Hence, we expect that: H3: SNS usage among immigrants is associated with greater satisfaction with social relationships.
Finally, we are interested in whether the relation between SNSs usage and the reported quality of close ties depends on the level of emotional disclosure towards others.Past research revealed a positive relation between general emotional disclosure and the quality of relationships [39][40][41].For instance, Finkenauer and colleagues [41] found that emotional disclosure has a positive effect on the satisfaction with both vertical and horizontal family relationships (the former being the strongest effect).Moreover, Finkenauer and Hazam [40] showed that emotional disclosure has a positive effect on marital satisfaction.Thus, we would also expect those individuals who generally show high levels of emotional disclosure to profit more from their SNS usage.
This is an extension of the research that examines whether online emotional disclosure enhances relational quality.We reason that, since there is a large variation in general emotional disclosure between individuals and since both individuals with little and with much general emotional disclosure use SNSs, the rewards of SNS usage in terms of the quality and quantity of personal relationships should be greater for the latter; or, as Trepte et al. [28] put it: "… a basic willingness for self-disclosure appears to be a precondition for more intense and active ways to use social media."In other words, we propose that those with a general tendency to disclose personal feelings and information are likely to engage in different activities and likely to produce different content on social media compared to those without this disposition.Hence, their time on social media should be more socially productive.Recent studies have shown that the outcomes of Facebook usage are highly dependent on the type of activities performed; especially supportive (online) interactions [42] and directed communication activities [11] are associated with improved socio-psychological outcomes.
We measure the quality of social relationships as satisfaction with social relationships.Furthermore, our data do not include measures of online self-disclosure.We assume that those with a general tendency to disclose their feelings and emotions also do so online.This is supported by studies that simultaneously examined on-and off-line self-disclosure; for example Schouten et al. [43] reported a correlation as high as r = 0.71.Thus, our final hypothesis is the following: H4: The effect of SNS usage on satisfaction with social relationships among immigrants increases with the level of general emotional disclosure.

Method
We test our hypotheses using data from the LISS (Longitudinal Internet Studies for the Social Sciences) immigrant panel (defining immigrants as persons who have at least one parent who was born outside the Netherlands [44]), which contains a representative sample of the Dutch immigrant population.Households that were otherwise unable to participate have been provided with a computer and Internet connection.We combined several modules of the panel, administered in 2010 and 2011.The response rates of all separate modules were above 70%.Our analyses only include "first generation" immigrants (not children of immigrants born in the Netherlands), comprising 749 respondents.Furthermore, the study sample is composed of immigrants between 16 and 85 years old, with diverging educational qualifications, many different national origins (48% non-Western) and considerable variation in the length of stay in the Netherlands (ranging between 1 and 85 years).

Variables
In-and out-group ties.Respondents were asked "Most people discuss important things with people they know.If you look back on the last six months, with whom did you discuss important things?"The resulting network is known as the core discussion network [45].The variable outgroup ties is a count of the number of mentioned contacts (a.k.a.alters or confidants) from a different national-origin group (maximum 5 persons), and in a similar vein, ingroup ties is a count of the number of contacts from the same national-origin group; contacts with individuals that may or may not live in the same country as the respondent (maximum 5 persons).The national origin question contained the following categories (representing the largest immigrant groups): Dutch, Turkish, Moroccan, Dutch Antillean, Surinamese, Indonesian, and other.Because of this latter, unspecified category, there were a few cases in which we were unable to detect whether a relationship represented an ingroup or outgroup tie.These undecided ties were removed from the analyses.
Satisfaction with social relationships.Respondents were asked "How satisfied are you with your social contacts?"The answer categories ranged from 0 (not at all satisfied) to 10 (completely satisfied).Even though this item is often used as part of a larger scale [46][47][48][49], research has shown that it is highly correlated with the scales used to measure global relationship quality: Hendrick [48] found a correlation of 0.76 between this item and the Relationship Assessment Scale (RAS).Moreover, this item showed the highest correlation with the scale.The scale has been used not only to measure romantic relationship quality, but also other types of close relationships (e.g., relationships with close friends [49]).
SNS usage.Respondents were asked how often they used SNSs, which included Facebook and Hyves (a Dutch SNS).The original categorical variables were recorded into quasi-continuous variables with the following values: 0, 6, 32, 208, 365 and 730 or more (times a year).We computed a new variable that represents the highest value of the two items.Table 1 shows the descriptive statistics.
General emotional disclosure.Emotional disclosure was measured using the "Expression of Emotion within Socio-Cultural Context Scale" [50], an adapted version of the Display Rule Assessment Inventory (DRAI) developed by Matsumotto and colleagues [51,52].Respondents were asked how they reacted when experiencing eight positive and negative emotions (joy, contempt, guilt, anger, happiness, warmth, fear and sadness) in two situations: (1) in interactions with familiar people; and (2) in interactions with people that the participant does not know very well or not at all.To give an example of an item: "Think about a conversation with someone that you know very well where you felt joy.What did you do with this feeling?"Response categories varied from 1 (I expressed my feelings, but with more intensity than my true feelings) to 5 (I smiled only, with no trace of anything else, and hid my true feelings).For our study, we created a summative score for all 8 emotions in both contexts; hence, the scale includes all 16 items.Higher values on the final scale represent a higher level of emotional expressivity (reversing the original coding).The internal consistency of the scale is 0.84.In the analyses, we used a standardized version of the variable.
We include the following control variables in our models: level of education, gender (women = 1), age, Western/non-Western origin and length of stay in the host country.Previous research has shown that these factors affect both (offline) strong tie networks and SNS usage.For instance, women have been found to use the Internet for communication purposes more often than men do, while men use it more often for instrumental purposes (e.g., checking the news, job opportunities) [53,54].Women are also often found to have more strong ties offline (e.g., Warr, D.J. [55]).DiMaggio and Hargittai [56] state that access to new technologies is associated with higher social status, such as a higher educational level.Furthermore, with regard to SNSs, the more highly educated have been found to participate more often than do the less educated [57].Likewise, the more highly educated show more extensive offline social interaction across many different indicators [58].Age is a well-known determinant of Internet usage and SNS usage [59].In a similar vein, younger people maintain larger offline social networks than do older people [60].We expect that these factors explain part of the relation between SNS usage and in-and out-group ties.However, if our abovementioned reasoning is correct, an effect should remain after adding these controls to our models.

Analytical Strategy
We used structural equation models to test the hypotheses.The estimation method used was maximum likelihood with missing values (Stata 13).Standard errors were corrected for clustering in households by applying Stata's vce(cluster) option, which takes into account that observation within the clusters (households) are not independent of each other [61].

Results
Figure 1 shows the effects of SNS (Facebook and Hyves) usage on the number of ingroup and outgroup ties in the personal networks of immigrants.Interestingly, SNS usage was not associated with the number of ingroup ties (H1 is unsupported), but it did show a positive effect on outgroup ties (in line with H2).In other words, the more frequent the usage of SNSs, the greater the number of ties to others from a different national origin group (most likely to the native Dutch).The coefficients are standardized; in other words, an increase in SNS usage by one standard deviation is associated with a 0.158 standard deviation increase in outgroup ties, which is a small effect.The coefficient of determination (equivalent to R 2 in OLS regression) of the entire SEM was 0.276.
As far as the control variables are concerned, three factors turned out to have significant effects.The higher the education level of an immigrant, the smaller the number of ingroup ties and the greater the number of outgroup ties.Furthermore, the longer the stay in the host country, the greater the number of both in-and out-group ties.Finally, non-Western immigrants have more ingroup ties in their core discussion network than Western immigrants.Table 2 presents the analyses of satisfaction with social relationships.From Model 1, we learn that the more frequent the usage of Facebook and Hyves, the greater the satisfaction with social relationships among immigrants, which supports our third hypothesis.Age is the only control variable that had a significant effect on the quality of relationships.Older immigrants reported greater satisfaction with personal relationships than younger immigrants.
In Model 2, we included an additional factor (general emotional disclosure) and tested whether the relation between SNS usage and the quality of social relationships is stronger for those who easily disclose their personal feelings to others.The results show that the effect of emotional disclosure on relationship satisfaction among those who do not use SNS is non-significant, as well as the interaction effect between emotional disclosure and the frequency of SNS usage (Hypothesis 4 is not supported).

Discussion and Conclusions
This article focused on the relationship between the use of two SNSs-Facebook and Hyves-and the personal networks of immigrants in the Netherlands.Our first conclusion is that SNS usage is positively related to the number of outgroup ties (whereas no relation with ingroup ties was found).This positive association between Internet usage and social relationships contradicts the displacement hypothesis from early Internet studies (e.g., Kraut et al. [1]) and is in line with more recent findings (e.g., Valkenburg [6]).
The lack of support for our first hypothesis is remarkable given the evidence from some previous studies, as discussed in the previous research section of this paper.However, it resembles findings by Burke et al. [11] in a general population: in their study, Facebook usage (or rather directed communication activities on Facebook) was associated with enhanced bridging social capital, but not with enhanced bonding social capital.One explanation for the diverging findings may be the type of Internet usage researchers have looked at: most of the mentioned studies (except Burke et al. [11]) have looked at individuals' total Internet usage.This means that the positive effects of Internet usage on ingroup ties may have been due to migrants' usage of applications other than SNSs.For instance, for maintaining close ties in the country of origin, Skype may be more important, and for the purpose of meeting fellow-migrants (ethnic) forums, it may be more important.Collecting more fine-grained data on different kinds of Internet applications and different types of ties could be an important step in future research to come to a better understanding of how this works.
Our finding is interesting in light of the debates about the acculturation of immigrant groups.They imply that online activities may help establish relations with natives, which is known to be an important success factor in acculturation processes.Our findings strengthen conclusions from previous studies in this regard [19,20,26].One possible mechanism that could explain this finding is that SNSs help strengthen ties.For instance, after having met someone for the first time, one connection strategy (cf.Ellison et al. [62]) may be to search for information about that person using Facebook (possibly after adding that person as a Facebook friend).This search may reveal things both persons have in common (reducing the perceived in-versus out-group difference), which, in turn, encourages further interaction, which, in turn, strengthens the tie.This could work for any person, but among (first-generation) migrants, this may be even more important, since they are in the process of establishing new personal networks.
However, some precaution is needed when drawing conclusions.SNSs can never be autonomous causes of enhanced acculturation: it is very likely that those who are using them to connect to outgroup members already had a desire to connect to others from a different national origin in the first place.In other words, SNSs can be tools that help establish and maintain outgroup contacts, but only in the hands of someone who has the desire to use them that way.Nonetheless, it is plausible that the rise of social Internet applications has made life a bit easier in this regard.
Our second conclusion is that the use of SNSs has a positive effect on the quality of social relationships.Our findings extend previous studies in this regard, because as far as we know, this has never been tested among migrants before.Again, the effects among migrants may even be greater than among non-migrants, since they are still establishing a personal network with resourceful contacts.Although we were unable to test what mechanism causes this effect, one option might be enhanced online self-disclosure [30,63].Especially, SNSs seem to accommodate online self-disclosure well, by providing their members with opportunities to constantly communicate their experiences and feelings to their online friends.
The exact way in which online self-disclosure works and how it interacts with offline self-disclosure remains a topic for debate and further study.We argued that individuals with a general tendency to disclose personal information should profit more from their SNS usage in terms of the quality of their social ties than individuals who do not have this tendency, but our analyses did not support this.Future studies may want to test this directly with measures of online self-disclosure, too, but for now, our preliminary conclusion is that SNS usage enhances self-disclosure (which, in turn, enhances the quality of relations) rather than that a predisposition of self-disclosure changes people's online activities and communication (making them more socially productive).
A few limitations of our study deserve to be discussed.First, our measures of social ties came from a name generator that asks about contacts with whom one discusses important matters and a maximum of five names could be provided.These name generators are known to generate strong ties [64].Thus, our findings pertain to close, intimate contacts, not to weak ties and social networks at large.This may be a reason why immigrant SNS users did not have more ties to others from the same national origin as non-users: they may have more ingroup ties among their weak(er) ties, which were not included in our data.Second, Internet and social media applications, such as Facebook, change rapidly, which means that (like all findings from these type of studies) our findings probably have limited generalizability over time.Third, our measure of general self-disclosure only referred to emotional aspects of self-disclosure.Although self-disclosure is often measured with items that are related to emotions (e.g., Schouten, A.P. et al. [43]) the concept also contains non-emotional aspects.Perhaps the results are different for non-emotional self-disclosure.Finally, another important limitation is that our data are cross-sectional, which means we cannot draw any conclusions about the causal order of the variables.Studying the causal order of these variables is difficult in any research design, because on-and off-line social interactions are extremely entangled nowadays.However, we do think that measuring the acculturation process longitudinally would be an important step forward.This would enable researchers to assess whether SNS users (or, more generally, users of social Internet applications) go through different acculturation trajectories.

Figure 1 .
Figure 1.SEM results of frequency of SNS usage on in-and out-group ties.

Table 2 .
Path model predicting satisfaction with social relationships (standardized coefficients and standard errors).