Keywords

1 Introduction

The emergence of new information and communication technologies (ICTs) has led to a Western society that is more connected than ever before. These days computers, mobile phones and the Internet serve as central means of communication, social interaction and entertainment. As a result, many existing social structures have had to re-formulate, raising questions over the possible impacts that these developments have on modern society.

The notion of online versus offline relationships has been an especially integral part of the debate concerning personal relationships, identification and impacts of digitalization [1, 2]. In the past, individuals based their identities on only a handful of social contexts, for example home, work, school and in the company of close friends. This is no longer the case, since the majority of individuals use the Internet and have hobbies through which they connect to many social networks online. Digitalization has indeed had a significant impact on society, serving as an influential factor in identity formation especially for younger generations. In this era, people have a multiplicity of general or specific groups to identify with through convenient access and effective communication tools online.

The tension between offline and online environments can be better understood by tracking where social ties are born in the first place. As Preece and Maloney-Krichmar have noted, “increasingly it is accepted that online communities rarely exist only online – Either they start as face-to-face communities and then part or all of the community migrates on to digital media, or conversely, members of an online community seek to meet face-to-face” [3]. In this sense, if the members of these communities are feeling that they have formed social ties outside of online spaces in the first place, these platforms are more likely to be categorized as “real life” communities that are utilizing online platforms. On the other hand, if these social ties between community members are realized primarily on online platforms, these platforms are considered online communities. Taking this notion into account, it is crucial to locate where the social action and networks are discursively placed.

While the digitalization of society obviously continues, there are differing views in academic discourse about the relationship between offline and online networks [3, 4]. Central here is the extent to which offline social interaction has actually been replaced by online relationships. Is there a trade-off between online and offline communities or are online communities more likely to be extensions of “real life” communities? With this research we elaborate on this timely question and offer a new perspective for academic discussion on social identifications.

Earlier literature shows that age is one of the most important factors when considering online social action [5,6,7]. Younger generations have been utilizing new technology and new social media applications much earlier and at a much wider scale when compared to older population groups [8, 9]. On the other hand, however, a growing number of ‘silver surfers’ have emerged in recent years. Not only are older people becoming more frequent Internet users, but their activities online have become more versatile [10,11,12]. In this sense we may also ask whether the age differences in general online behaviours have diminished among age groups.

Our interest in this paper revolves around questions regarding online identifications. This is a profound question in the online setting, since social identification signals are not only how we define ourselves, but also show what we consider important in life. Therefore, in addition to examining the traditional quality and usage of ICT services, it is important to examine the social experiences involved in using different digital tools and services. We intend to examine the extent to which different age groups and generations in Finland identify with offline and online communities in 2009 and 2017. The empirical part of this paper focuses on identification with online communities and more traditional social groups in Finland. We are particularly interested in exploring the possible differences between age groups and generations. In addition, we are also interested in how other socio-demographic factors connect with the strength of identification. To summarize, we present the following three research questions:

  • RQ1: Were there temporal changes in identifying with online and offline communities between 2009 and 2017?

  • RQ2: Were there distinct socio-demographic profiles of identification with an online community in 2009 and 2017?

  • RQ3: Was there cohort variation in identifications with an online community between 2009 and 2017?

2 Identifications on the Internet, Platform Evolution and Online Communities

Social scientists have shown the importance of group memberships and social networks in past research, which remain a fundamental building block in society today as well despite fundamental changes due to ICT developments. All manner of social position are arranged through in-groups, including business deals, political influence, and positions of employment, among others. Here, advantages are afforded to individuals who are viewed as members, leaving outsiders who are structurally excluded at a disadvantage [13]. In addition to being a fundamental aspect of society, groups and communities act as a valuable psychological anchoring point, by bringing individuals of various groups a source of security, validation and self-esteem [14].

This bolstering of individuality and group belonging together meets deep needs in socialising individuals who find agreeable social connections. According to social identity theory [14], individuals make sense of their social environment by categorizing themselves and others into groups that can be compared with one another. Here, comparison of group behavioural norms and identifying characteristics help to determine the individual member’s place in society. Notably, these social identities exist simultaneously and evolve over time; being a father, son, sports fan and member of a political party are all social identities of an individual, the strengths of which can change over time. Notably, certain social identities through experiences of identification are more likely for some population groups than others; age, gender, level of education and numerous additional factors can influence the likelihood of becoming attracted to identifying with a certain group [1]. As such, connection with others is central to this theme of membership and the social value ascribed to it.

The rapid transformation of Western societies through ICT developments has, at its core, been one of connectivity. Social networks and online communities have been key in this new level of social identification online where interacting partners are far more easily reachable regardless of geographic limitations. This feeds into a core need, namely for socialisation, which helps to explain the massive growth of social platforms online. This extension of the offline world has brokered countless new memberships and experiences of connection globally.

To understand what has happened to identification with online communities, it is important to understand changes over the years during which much online evolution took place. This study focuses on the years between 2009 and 2017, which can be described as the social media era, particularly in Western countries. When considering online communities’ development, social media plays a crucial role, despite there already being online communities before the advent of social media. In the era of “Web 1.0”, online communities were typically based on text-based platforms, such as various kinds of discussion forums, chat sites, and messaging services [15, 16]. These platforms exist today as well, but their popularity and importance is significantly diminished when comparing to modern social networking sites.

Social networking platforms began to emerge after in the early 2000s, enabling unparalleled methods of social interaction in the online environment. These new sites – such the first wave of platforms including Friendster, LinkedIn, and Myspace – had novel features of social interaction and featured new visual characteristics. After the boom of social media, several pioneering social networking sites were overridden by emerging social media giants, especially Facebook. Subsequently, Facebook grew to be by far the most popular social networking site and the most important media company in the world. Nowadays there are roughly two billion monthly active Facebook users worldwide [17]. In addition to Facebook, there are also many other popular social networking sites utilized today as platforms for online communities, such Youtube, WhatsApp, Facebook Messenger, Twitter, and Instagram.

When comparing “Web 1.0” online communities to modern day online communities, namely social networking sites, there are major differences between them. First, today’s social networking sites are highly visual and packed with various features, far surpassing the user experience of older online community platforms. Secondly, during the first phase of online community platforms, people were gathered around various topics and issues, which they were interested in [16]. These “old school” online communities were more likely to be based on small niches of people sharing the same interest, than large-scale social networks extending users’ offline experiences and social circles. Thirdly, today it is common, especially in platforms like Facebook, Twitter and LinkedIn, for users to interact and navigate using their own names and identities, while platforms such as discussion forums and chat rooms more prominent in the past allowed for use of a pseudonym or in some cases were fully anonymous [15, 18]. These differences in characteristics play a crucial role in the kinds of communities that these platforms are enabling.

Notably, the development of new emerging social networking sites can be seen as taking steps back to these first phase online community platforms. For example, platforms based on more private interactions and conversation, such as WhatsApp and Facebook Messenger, have gained popularity during this past decade. Also, platforms based on full anonymity, such as Jodel, have emerged. These new platforms mentioned are more concentrated in terms of features instead of being multi-media and multi-featured online platforms. When considering this reversion in social media services, it seems that demand remains for smaller groups and communities without the attention of the wider public and large-scale networks.

In the time before the spread of social media, online communities were typically made up of a group of people interacting based on a shared interests or purpose, who were guided by some form of protocols, norms, policies or rules, with the interaction supported or mediated by technology [3, 19]. Compared to the past, online environments are generally based on social ties from offline surroundings. In this sense, the line between online and offline space have become even blurred.

3 The Changing Social Media Landscape in Finland

Overall, meanings, settings and surroundings for online communities evolved over the past decade. Online social networks have become more important, commonly shared and more personal after the wide scale spread of social media. This transition of “web 1.0” platforms to social networking sites can be observed in how people are utilizing various platforms. For instance, at the starting point of our research in year the 2009, only 30% of the Finns have a registered profile on social media [6], while approximately the same amount of people were also participating in conversations on various discussion forums and news groups [20]. In the year 2017, the proportion of people who utilize discussion forums or news groups has diminished to less than 10% [21]. In turn, in the year 2017 roughly 65% of Finns were utilizing social networking sites. Figure 1 highlights these recent changes in the shares of those Finns who have registered for at least one SNS site.

Fig. 1.
figure 1

Social media penetration rates in Finland, 2010–2017 (%)

Full size image

As the figure shows, younger age groups are still more likely to be registered as SNS users than their older counterparts. In 2017, over 90% of under 35-year-olds were registered users, whereas less than 50% of over 55 year-olds belonged to this category. However, in 2010 the differences between the youngest and oldest age groups were even stronger. In addition, the middle-aged, namely 45- to 54-year-olds, have become clearly more active over this time period.

These observations indicate that there have been changes in the Finnish SNS landscape. Therefore, it would be necessary to address whether there have been certain qualitative differences in online identifications over time. If we assume that individuals’ urge to define themselves through social media communities has generally increased, it would be especially important to know how people were identifying with different online communities earlier compared to now. Here, it is also important to compare online community identification with identification with different and more traditional communities formed primarily in offline surroundings.

4 Data and Methods

Our empirical data are derived from national-level surveys collected in Finland in 2009 (n = 1,202) and 2017 (n = 1,648). These nationally representative samples consist of respondents aged 18 to 74 years, thus providing an extensive look at the phenomenon in question. The surveys used simple random sampling as the respondents’ home addresses were drawn from the Finnish population register database. Final samples are corrected with sample weights by balancing variation in sample sizes and response bias in terms of age and gender.

We use subjective measures of identifications as dependent variables. The variables were elicited with the question: “How strongly do you feel part of the following groups?” A total of eight items were displayed in the questionnaire for evaluation: city or town, church or religious community, a hobby group, an online community, residential community, colleagues at work or at school, friends, and family. Respondents gave their answers using a five-point Likert-type scale (ranging from 1 = “Not at all” to 5 = “Very much”). While no restrictions were given in the questionnaire, it is perhaps reasonable to assume that most of the respondents answered on the basis of their views relating to the circumstances of their daily life. In this way the interpretations of an online community or a hobby group, for instance, may vary between respondents. Nevertheless, possible ambiguities are taken into account in the interpretation of the results. In our analyses the dependent variables are treated as dichotomous measures as we estimate likelihood for identifying with different communities. In order to do this, we dichotomized original variable values (1 or 2 = “Not identifying”; 3, 4 or 5 = “Identifying”).

Our main independent variable is age. In the first section of analysis, age was specified in the questionnaire as the year of birth, thus providing a continuous measure. In order to allow parallel comparisons with the other independent measures, age was categorized into seven groups: 18–24, 25–31, 32–38, 39–45, 46–52, 53–59 and 60–74.

In the second stage of analysis, we examine cohort effects while acknowledging problems related to the analysis of age, such as whether the phenomena discovered are related to certain life-cycle stage or whether they are actually typical of broader groupings, such as generations. However, we must bear in mind that our analysis is based on the utilization of cross-sectional data sets, which do not necessarily allow causal interpretations. In order to have observations from each age group in both survey periods, we had to re-categorize the data by establishing the following cohorts: 1943–49, 1950–56, 1975–63, 1964–1970, 1971–1977, 1978–84 and 1985–1991. As a result, the oldest age group, namely cohort 1935–1942, observed in 2009 and the youngest group, namely cohort 1992–1999, observed in 2017 were excluded from the cohort analysis.

The control variables include three demographic variables: gender, residential area and education. An earlier study has found that these variables have a strong association with identifying with an online community [5]. In the present study, residential area was measured simply by asking participants to choose their type of residential area, urban or non-urban. It can be argued that this variable reports unambiguously whether the respondent’s residence is located in an urban or a non-urban setting. Education was measured in the data as vocational education. Here, the classification used consisted of four categories on the basis of ISCED classification [22]: (1) “Primary” (including all without at least secondary education), (2) “Secondary” (including lower, upper and post-secondary), (3) “Tertiary” (including tertiary and bachelor) and (4) “Master” (including master and doctoral degree).

The analytic techniques include logit models on the basis of which we post-estimate citizens’ likelihood for identifying with offline and online communities. In the cohort analysis, we also equate effects of gender, education and residential area and show average year effects for each cohort. Before that we present unadjusted effects of applied background variables on identifying with online communities. All results are illustrated in figures by utilizing coefplots developed by Jann and figures are improved with schemes developed by Bischoff [23, 24]. Statistical tests are shown in the appendix in the Tables A1, A2 and A3.

5 Results

We begin our examination from the first research question (RQ1): Were there temporal changes in identifying with online and offline communities between 2009 and 2017?

Fig. 2.
figure 2

Changes in the probabilities of identifying with different communities between 2009 and 2017

Full size image

Figure 2 shows estimations for identifications with various social communities. The values indicate probabilities of expressing identification experience on a scale from 0 to 1 for both years separately. The figure shows us that there have been changes over the time period examined. Social belongingness with online communities especially has grown stronger. In fact, the magnitude of increase is as much as 20 percentage points between the years 2009 and 2017 (from the likelihood of 0.29 to 0.49; logit coefficient = 0.84). This effect is statistically very significant (at the level of p < 0.001). Otherwise, the observed changes are relatively marginal. In general, identification experiences with most traditional communities have either increased or stayed at the same level. However, identification with a church or a religious community has weakened slightly. School and work colleagues as well as hobby groups have become more important sources of variation over time (the differences are not otherwise statistically significant). Statistical tests for temporal changes by item are given in Table A1.

The findings above indicate that Finns identified with an online community more intensively in 2017 when compared to 2009. This interpretation naturally applies only to the whole population on average. This is why we need to continue our investigation by looking at the possible differences between different population groups. This leads us to our second research question (RQ2): Were there distinct socio-demographic profiles of identification with an online community in 2009 and 2017?

Fig. 3.
figure 3

Unadjusted probabilities of identifying with an online community by socio-demographic characteristics in 2009 and 2017

Full size image

Figure 3 shows how different age and educational groups, men and women, and urban and non-urban dwellers identified with an online community in 2009 and 2017. The results are represented as unadjusted likelihoods (on a scale from 0 to 1), which make it possible to compare the effect of each independent variable between the years (the estimates are based on logit model shown in the Table A2). The most notable finding has to do with age. It appears that the disparities between the age groups have diminished. This is true especially between the youngest age group and others. In 2009, under 25-year-old Finns were a distinct category expressing a clearly stronger level of identification compared to all other age cohorts. In 2017 this is no longer the case. Moreover, what is noteworthy is that in the youngest age group the likelihood of identification has dropped by 13% points (from 0.80 to 0.67) between 2009 and 2017. All other age groups report stronger identifications in 2017 when compared to 2009.

The figure also shows that the differences between educational categories have become significant in 2017, while the differences were not statistically significant in 2009. Those with primary education identified less with an online community than others. The differences between males and females, as well as with urban and non-urban residents, were significant both years.

Together, these results indicate that certain temporal changes have taken place in online identifications between the years. The important changes relate to age. On one hand, the differences between all age groups have generally weakened. On the other hand, the youngest age group tend to be less attached to online communities than earlier. Given the fact that our data also allow us for comparing the responses within the age cohorts at different points in time, a further analysis was performed. This leads us to the final research question (RQ3): Was there cohort variation in identifications with an online community between 2009 and 2017?

Fig. 4.
figure 4

Effects of year 2017 in predicting probability for identifying with an online community. Cohort analysis among 7-year age groups between 2009 and 2017 (average marginal effects adjusted for gender, education and residential area).

Full size image

Figure 4 shows the average marginal effects for identification within each age cohort between 2009 and 2017. These estimates report the average change from 2009 to 2017 when taking into account respondents’ level of education, gender and residential area (the estimates are based on logit model shown in the Table A3). The values are reported on a scale from −1 to 1, in which a positive value indicate an increase in the strength of identification and a negative one a decrease. The findings are clear: identifications have generally grown stronger within cohort groups over time. More importantly, however, the youngest age cohort (those born between 1985 and 1991) makes an exception here. Therefore, the findings are perfectly in line with our earlier interpretation regarding the youngest age group. It is indeed the youngest Finns who report weakened identification with an online community. What implications do these findings indicate?

6 Discussion and Conclusion

Our findings indicate that identification with online communities has become more prominent in Finland between 2009 and 2017. It is also noteworthy that demographic differences have also diminished between observed years especially in terms of age. Accordingly, the popularity of online groups has increased among middle-aged citizens especially. Also, women and people living in urban districts were more likely to be embedded in online communities. Interestingly, we did not find variation in terms of education level and online community identification in 2009, but in 2017 people with only primary level education were less likely to identify with online communities.

In the more detailed analysis, we found an interesting interaction between age cohort and observed year. Younger generations have apparently experienced a sort of communal backlash in which identification with online communities has decreased, whereas the probability of identifying with traditional social groups has not changed between the years 2009–2017.

According to the diffusion of innovations tradition, it is expected that the growth in the use and attachment of social media should begin to slow down within a period of time after a boom in popularity. In view of our findings, it seems that younger generations have tired of with the wide-scale online networks and multi-featured online platforms. This finding is in line with earlier statistics suggesting that younger users have started to abandon far matured and multi-featured social media platforms, such as Facebook [25]. This can be seen also in the adoption of nascent and recently growing social media platforms such as Snapchat and Instagram, which are more concentrated with a narrower set of features [26, 27].

However, these generational differences may also be explained by life stage transition. Even though the youngest generation still has the highest probability of identifying with online communities, online communities seem to be the first social groups they neglect after reaching the hurried thirties. This finding related to life transitions also underlines the order of importance between offline and online communities. As our results suggest, online communities are the first to be abandoned when prioritizing time consumption and when assessing what is essential and what is not.

On the basis of these findings, we argue that identifications with different communities can be seen as a good indicator of the impact that digitalization has had on society, especially on a generational level. Despite the fact that the growth rate of social media attachment has declined among the youngest groups, the growth is still strong among older population segments. However, it is also noteworthy here that traditional communities and social groups are still holding strong, even though the popularity of online groups has increased at the population level.

Notably, there are limitations in this study. Our findings represent the views of one European country alone, which do not allow us to make broader generalizations. It is also noteworthy to acknowledge that our data come from quantitative population surveys and not from qualitative in-depth interviews. However, the most difficult interpretations of the study deal with theoretical research implications.

Our findings partly support interpretations that online and offline communities can still be to be separated from each other. Namely, while online communities have become a much more essential part of people’s lives, traditional offline communities, groups and networks have not been replaced by them. This suggests that the blurred line between online and offline in people’s understanding is still possible to determine. However, it is evident that various offline social communities have embedded themselves in online environments, resulting in the blurring of past boundaries. As such, it has become even harder to make a distinction between online and offline communities. The question of what we are referring here is how users understand the term “online community.” For instance, is a WhatsApp group of close friends an online community or just an extension of a particular offline community? Do family members or colleagues at work cease to be offline community members on Facebook or ResearchGate?

Such questions lead us to wonder whether it is even possible to separate online and offline communities in the same way as before the era of social media. It is very likely that responses from different age cohorts would be contradictory with each other. Together, these notions indicate that more research is needed to assess this conceptualization of a possible boundary among social media users.