When Social Media Influencers Go Political: An Exploratory Analysis on the Emergence of Political Topics Among Finnish Influencers

Social media influencers are often understood as non-political actors who, due to the impact they exert on their followers’ purchase decisions and brand attitudes, play an important role in marketing and branding. In recent years, various influencers have addressed political issues during important political events such as elections. This political aspect of social media influencing has not received much scholarly interest thus far. In one of the first exploratory studies on the topic, we surveyed over one hundred Finnish influencers and investigated the degree to which they engaged with political topics. Our results show that political topics are commonly brought up among Finnish influencers. However, many influencers also deliberately avoid addressing them because they fear what comments and conversations these topics could generate and that they could personally be targeted by aggressive internet commentators. Nevertheless, based on our findings, we argue that the digital spaces maintained by influencers can constitute a new kind of third space where the emergence of political topics can have a greater impact on the political behaviour of influencers’ followers.


Introduction
Social media influencers (henceforth abbreviated as SMIs) have been a topic of scholarly interest for some time now, and they represent one of the latest resemblances of opinion leaders (since Katz and Lazarsfeld 1955). Thus far, SMIs have been primarily studied from a marketing and strategic communication perspective due to their ability to impact their followers' purchase decisions and brand attitudes (Colliander and Dahlén 2011;Djafarova and Rushworth 2017;Lee and Watkins 2016;Munnukka et al. 2019). However, in this article we argue that the possible roles SMIs can take may be significantly broader than to only impact consumer behaviour and that the societal and political roles of SMIs are becoming increasingly relevant. This is exemplified by numerous instances in recent years during which SMIs have commented on societal topics via their channels and drew their followers' attention towards politics. For example, a German gamer YouTuber called Rezo, released a video shortly before the elections to the European Parliament in which he tackled the climate policy of the ruling conservative party (The Guardian, May 22, 2019). In some cases, SMIs collaborate in institutional communication, such as when the Finnish government utilised SMIs in their efforts to raise awareness about the COVID-19 pandemic during the spring of 2020 (The Guardian, April 1, 2020).
As these examples show, while SMIs are not often considered in a political context, some of them are nevertheless engaging with the political sphere. However, this is a research topic that has not received much attention even though it can have important implications for political behaviour. Since SMIs have the capacity to influence their followers' opinion formation and behaviour in various non-political arenas (e.g. Kapitan and Silvera 2016;Lee and Watkins 2016;Reinikainen et al. 2020), it is plausible that their influence can extend to the political sphere when these SMIs so-to-speak go political. Indeed, extensive literature exists on preceding non-political online spaces where the emergence of political talk has been shown to have bearing on political opinion formation and behaviour of those who are exposed to it and engage with it (e.g. Van Zoonen 2007;Wright and Graham 2016).
Hence, we propose that the emergence of political issues initiated by SMIs is an important phenomenon that calls for more research. In this article, we conduct one of the first exploratory study on the topic. Our main objectives are to explore the extent to which SMIs raise political issues, what types of political issues or topics they tend to focus on, and what reasons prevent them from bringing up political topics on their channels.

Social Media Influencers and Opinion-Leadership
Ever since Katz and Lazarsfeld popularised the concepts of an opinion leader and the two-step flow of communication (1955), the notion that ordinary individuals can have a large-scale impact on the opinion and behaviour of others through their personal influence has become commonplace. In this, SMIs can be understood similarly as opinion leaders, who occupy a central place within a communication network by maintaining connections to many other individuals and to have the ability to influence them by decoding messages that are disseminated by media or other organisations (cf. Casaló, Carlos Flavián, and Ibáñez-Sánchez 2018). Moreover, like opinion leaders (Katz and Lazarsfeld 1955), the influence of SMIs often stems from the expertise they have obtained on a specific topic such as fashion, fitness, travelling, or gaming.
Opinion leadership provides a helpful starting point to understand the SMI phenomenon. However, there exists many different types of opinion leaders that share conceptual similarities with SMIs, such as bloggers, micro-celebrities, and citizen journalists (Brundidge et al. 2014;Khamis, Ang, and Welling 2017). This conceptual similarity requires us to define SMIs more specifically, and we do this by using theoretical work conducted by Freberg et al. (2011), andEnke andBorchers (2019). An early definition of SMIs brings forth two specific characteristics of SMIs by stating that they are "third party [i.e. independent] endorsers who shape audience attitudes through blogs, tweets, and the use of other social media" (Freberg et al. 2011, 90). This definition underscores that the SMI phenomenon is based on the heavy use of social media platforms that are based on Web 2.0 technologies. Next, the characterisation of SMIs as "endorsers" implies that SMIs are willing to monetise their status as opinion leaders and establish collaborations with corporations in which they endorse brands and products.
Two further characteristics of SMIs are relevant for the purpose of this article. First, Enke and Borchers (2019, 267) define SMIs as "third-party actors that have established a significant number of relevant relationships with a specific quality to and influence on organisational stakeholders through content production, content distribution, interaction, and personal appearance on the social web". Accordingly, SMIs' status depends on the relationships that they have established with their audiences. This status is highly dependent upon how many individuals become their followers, how regularly these followers engage with the content the SMIs create, and how much they interact with the SMIs (Gräve and Annika 2018). Due to their importance, SMIs commonly establish two-way interactions with their followers and relate with them in intimate ways (Berryman and Kavka 2017;Raun 2018). Furthermore, these interactions are not singular, non-recurring events, but rather they are constantly ongoing where users return to interact with SMIs more or less regularly (Enke and Borchers 2019, 265). This way, SMIs foster parasocial relationships with their followers (Colliander and Dahlén 2011).The second SMI characteristic that can be derived from Enke's and Borchers's definition is the practices of self-branding. SMIs need a wellcurated public persona that allows them to maintain consistency and to be perceived as authentic by their followers (Khamis, Ang, and Welling 2017). The successful curation of a public persona can increase the attachment and size of the SMIs follower base and furthermore improve the SMIs endorsement potential for commercial products (Abidin 2016).
In summary, in this article we conceive of SMIs as opinion leaders who 1) intensively use social media in their communication practices, 2) collaborate with corporations to monetise their opinion leader status, 3) establish regular two-way interactions with their followers that often lead to parasocial relationships, and 4) engage in self-branding strategies to curate a consistent public persona. These characteristics distinguish SMIs from other types of opinion leaders sciences, such as traditional bloggers (Brundidge et al. 2014) and social media wielding activists like Greta Thunberg (Olesen 2020) who do not, for example, establish commercial collaborations or systematically pursue self-branding strategies.

Defining Political Topics
To investigate to what extent SMIs bring up political topics to their followers we must first have a way to identify them. As the nature of politics has become increasingly "elusive" (Papacharissi 2011), there exists no single easy answer to this question. One could use a somewhat formal and "rigid" definition for a political topic where they refer to formal political processes and institutions or when institutional politics itself is the topic (Graham, Jackson, and Wright 2015, 653). The problem with such an approach is that it can be largely unrepresentative of the full spectrum of what constitutes politics. Indeed, for most people, the everyday ebb and flow of politics is connected to their personal, subjective experiences (Van Zoonen 2005), where what is political is not always understood in connection with formal political actors or institutions but also through individual lifestyle values and personal narratives (Graham 2008).
Due to these two distinct ways of understanding what constitutes a political topic, we opt to treat them separately and to distinguish between formal and lifestyle-based political topics. Here formal topics refers to issues that relate to political processes, institutions, or actors such as politicians, parties, or policies. Lifestyle-based topics, on the other hand, are broader in their nature and we understand them to occur when SMIs make a reference to any issue that they think is a collective concern and which has an impact on the wider society, such as when they make references to topics such as health, travel, family, or housing.

Research Questions
Due to the novelty of our research topic, it is difficult to formulate precise expectations about the phenomenon. Hence, the main research questions of this article are explorative in nature and aim to provide an initial understanding on whether bringing up political topics is a type of behaviour that is an anomaly among SMIs, and to establish whether certain patterns exists as to which SMIs are more likely to bring up political topics on their channels. However, while we argue that political topics can be prevalent among SMIs, we recognise that political topics are contentious and that there can be several reasons to actively avoid them, especially in non-political digital spaces. For example, in her influential work on why American citizens avoid politics, Eliasoph (1998) shows that avoidance is often predominant because the context of the public setting deliberately discourages political talk. Similarly, the setting and the context of the digital spaces that have formed around SMIs can develop in a direction where political talk is not considered appropriate and thus is actively avoided. Moreover, harassment, verbal abuse, and even threats of violence have been found to be common within social media, discussion forums and other online arenas that include political discussions (Cicchirillo, Hmielowski, and Hutchens 2015), which surely act as a deterrent for SMIs to start engaging with politics. Indeed, online discussions about leisure and hobbies tend to be precisely the type of online chat and message spaces where the likelihood of encountering political disagreement is high (Wojcieszak and Mutz 2009).
These issues make it highly plausible that SMIs might actively avoid political topics even when they would have something to say or would like to use their channels to highlight some political issue that is important to them. Thus, two further research questions emanate: RQ 2.1: To what extent do SMIs deliberately avoid bringing up formal or lifestyle-based political topics on their channels? RQ 2.2: What are the most common reasons why such topics are avoided?

Method and Data
This study was carried out using an online-survey conducted in Finland in October-December 2019. Finland is a useful country to conduct studies on SMIs particularly due to the high internet and social media use rates among its population. Finns are among the most active internet users in the EU with a high social media adoption rate (Internet World Stats 2020). However, it is notable that the Finnish social SMI sphere is small compared to, for example, the US, Germany or Sweden. While the most influential Finnish YouTubers have around 400,000-500,000 followers (Niemi 2019), media intelligence company Meltwater Finland estimates that based on their Social Influencers Tool, which lists the profiles of SMIs globally, the majority of Finnish SMIs have either between 500-5,000 followers or 5,000-30,000 followers on their social media channels.
In order to reach Finnish SMIs, we established collaborations with fifteen Finnish communication agencies and networks that manage and/or represent SMIs who agreed to share the survey link to individuals within their SMI networks. Collectively, these agencies and networks have the potential to reach approximately 1,500-1,600 SMIs on different platforms. This initial survey dissemination was further enhanced by snowball sampling, whereby the SMIs who responded to the survey were encouraged to share it in their own personal networks to those who they perceive to be SMIs.

Operationalisations
In order to distinguish between formal political topics and lifestyle-based political topics, we define the concepts by adopting prior work on political third spaces. Following Graham, Jackson, and Wright (2015) and Munson and Resnick (2011), we defined formal political topics to occur when a SMI makes mentions to formal political processes, institutions or actors such as politicians, parties or policies. In respect to lifestyle-based definition for political topics, we adopt the first criteria proposed by Graham for a political discussion, in which lifestyle-based political topics occur when an individual first refers to a personal experience, interest, issue, or topic in general to society, and then secondly connects it to the wider society. As Graham puts it, a connection to the wider society is established when such topics are not only seen as important to single individuals, but also a collective concern that should be likewise discussed collectively (2008,(22)(23).
We operationalised these two types of political topics through two survey questions. Formal political topics were captured by asking the respondents "Have you ever brought up political topics or controversial public issues in any of your main platforms?" and further explaining that this refers to times when they mentioned topics related to formal political areas such as to political actors, political processes, public policy, political institutions, and other public issues. Lifestyle-based political topics were operationalised through the following survey question: "Have you ever brought up some interest, issue, topic or personal experience in any of your main platforms because you felt that it is societally important and that there is a need to have a public discussion about it?" Again, for clarifying purposes we further explained that this could refer to non-formal political issues, such as to sexuality and gender, animal rights, health and eating, climate change, religion, consumer behaviour or social equality. The survey also included two sets of more detailed questions about what kinds of political topics the SMIs had specifically brought up within the past year, whether they had ever deliberately avoided such topics, and what were the most common reasons why they avoided them (the complete survey is included in online supplementary material 1).

Sample Description
The survey was answered by 114 individuals, but due to question specific nonresponse, the sample size varies from 90 to 114. The respondents were 72 percent THE EMERGENCE OF POLITICAL TOPICS AMONG FINNISH INFLUENCERS female, 26 percent male. The age of the respondents varied from 18 to 62 where the average age was 35. The most common number of followers the respondents had ranged around 1,000-5,000 followers, which is typical for most Finnish SMIs. Finally, most respondents had around 1-10 ongoing collaborations with brands or companies. These background variables are summarised in Table 1.
The survey allowed respondents to make a multiple choice from 21 different items to indicate what topics they usually create content on for their followers. Among these, lifestyle (48%) and travel (28%) were the most popular ones. Fewer than 20 percent of the sample created content on remaining topics such as health or well-being, and fashion or clothing. Politics was mentioned by only 6 per cent of the respondents. Over 30 percent of the sample indicated that they created content on other topics that were not present in the survey items, which most commonly were related to cooking, food, and eating.

Results
We first present the response distribution on whether the respondents had brought up formal political topics or lifestyle-based political topics on their social media channels. The analyses show that both formal and lifestyle-based political topics had been brought up at least once by the vast majority of the respondents: over 60 percent having brought up topics related to formal politics and approximately 94 percent related to lifestyle-based politics. While we cannot claim that these results are generalisable beyond our sample, the fact that most of the surveyed SMIs mainly create content about non-political topics (such as about lifestyle in general, travel, and entertainment) suggests that this result is not only due to self-selection whereby only highly politically active and interested SMIs responded to the survey.
We further asked the respondents whether they had brought up five more specific types of formal political topics and seven lifestyle-based political topics on their channels (see Q4 and Q7 in Online supplementary material 1). Figure 1 shows how frequently the respondents had mentioned these topics on their channels within the past year, where the frequency was measured through an ordinal variable that ranged from 0 to 3 (0 = not at all, 1 = once or twice, 2 = three to five times, 3 = more than five times).  Figure 1 highlights again that various lifestyle-based political topics are generally more commonly mentioned by SMIs. Topics such as consumer behaviour, health and eating, and climate change are by far the three most popular, all of which had been mentioned at least once or twice at the minimum by three-quarters of the respondents (shown in Very Common). Moreover, almost a quarter of the respondents were highly active in mentioning all three topics: at least 24 percent of them had brought up these topics more than five times within the past year.
Following these three most common topics, we find that six topics were mentioned by over 50 percent of the respondents at least once or twice (shown in Common). The most intensely brought up topics are again related to lifestyle-based political topics, such as sexuality and gender, and social equality, but formal political topics such as public issues, public policy, and political processes are now also represented.
However, while lifestyle-based political topics tend to be more intensely brought up within these six items, the formal political topic labelled as public issues is a notable exception. Public issues were mentioned at least three to five times by 35 percent of the respondents whereas only 28 percent had brought up sexuality and gender, and only 20 percent had brought animal rights at least three or five times. It is worth noting that immigration was given as an example of a public issue to the respondents and that immigration was a very visible topic within the Finnish media during the survey dissemination. Thus, it is likely that the salience of the example had a significant impact on how the respondents respond to this question item. Finally, the rarest topics mentioned by the SMIs are related to political figures, political institutions and to religion (shown in Uncommon), all of which are not mentioned at all by the vast majority of the respondents (60 percent or more).

Respondents on Two Scales: The Intensity of Formal and Lifestyle-based Political Topics
To further investigate what types of SMIs are more or less likely to bring up formal or lifestyle-based political topics, we reduced the question items presented in Figure 1 into two standardised activity scales that represent how actively the respondents brought up our two types of political topics. The values of the scales ranged from 0 to 3, and we categorised each respondent to belong in one of four different activity levels based on the following cut-off points: 0 = not active, 0 < somewhat active < 1, 1 ≤ active < 2, 2 ≤ highly active. The distribution of each scale and the number of participants belonging to each activity category is presented in Figure 2.
As the figure shows, the distribution of the lifestyle scale is much more even when compared to the formal scale, and that most of the respondents are at least somewhat active. In contrast to this, the distribution of the formal political topic scale is highly skewed because a significant portion of the respondents are not at all active. About 15 percent of the respondents, however, are unusually active in bringing up formal political

FIGURE 2
The distribution of topic scales. (Circles in the figure represent individual respondents. Those who indicated that they had never brought up formal or lifestyle-based topics were coded 0 on the relevant question items.) topics, out of which two respondents had brought up all types of formal political topics more than five times during the past year.
We used the non-parametric Mann-Whitney u (for two groups) and Kruskal-Wallis (for groups of three or more) tests to investigate whether different groups of SMIs were prone to score higher or lower on these two scales. These groups were created by using 10 different survey questions, which are summarised in Appendix 1. The following section presents differences between groups that the tests found to be statistically significant at a .05 level or relatively near it.

Major Differences on the Lifestyle Topic Scale
For the lifestyle topic scale, we found four variables that displayed statistically significant differences between the scores of different respondents. First, SMIs who had more followers tended to score higher on the scale, but only up to 50,000 followers. Table 2 shows that the mean score of SMIs tends to rise as they have more followers, and those with 10,001-50,000 followers are the most active in bringing up lifestyle-based political topics. However, the trend seems to change after this.
Similarly, respondents who have more ongoing collaborations with brands or companies tend to score higher on the lifestyle scale. Whereas SMIs who had ten or less collaborations tend to be only somewhat active on the scale, those with more than ten collaborations tend to fall into the active category.
Furthermore, SMIs whose followers interact more with them, or react positively towards them when they mention topics related to lifestyle-based politics, tend to score higher on the scale. In both cases, the SMIs who reported that their followers either interact more with them or react positively towards them tend to score higher on the scale. However, it is worth noting that it is extremely rare for the respondents to report that their followers would interact less with them or to react negatively towards them when they brought up lifestyle-based political topics.

Major Differences on the Formal Topic Scale
For the formal topic scale, the background variables did not produce statistically significant differences between the SMIs and neither did the respondents' evaluation of how their followers reacted to them when they brought up topics related to formal politics. Out of the tested variables only one produced significant differences: SMIs who disagreed with the statement that they sometimes avoid bringing up formal political topics because they want to keep their social media channels as non-political as possible tended to score higher on the formal topic scale than those who agreed (mean score 1.47 compared to 0.39). Moreover, those who disagreed with the statement that they had nothing to say about formal political topics were also prone to score higher on this scale (mean 1.03 vs. 0.70), although the difference is significant only at .068 level.

The Most Common Reasons to Avoid Formal or Lifestyle-based Political Topics
In regards to whether the respondents had ever deliberately avoided bringing up either formal or lifestyle-based political topics, we find that both topics were avoided by over a half of the respondents and that formal politics is more commonly avoided. Approximately 77 percent had deliberately avoided formal political topics, whereas 52 percent had deliberately avoided lifestyle-based political topics. For the respondents who admitted avoiding either topic type, Table 3 summarises the level of agreement among respondents with a list of reasons why they had avoided either topic type.
As can be seen from Table 3, the most common reason why the respondents avoid formal or lifestyle-based political topics was that they did not want to be targeted by aggressive internet-commentators. General fear of comments and conversations was likewise a common reason why the respondents avoided both lifestyle-based and formal political topics, although the reason seems to be somewhat more common within lifestylebased political topics than in formal political topics (mean of 0.51 and 0.38 respectively). A significant portion of the respondents also indicated that they explicitly want to keep their social media channels as non-political as possible, which is why they tend to avoid bringing up topics that relate to formal politics, but this reason is not as prevalent within lifestyle-based political topics.

Discussion and Conclusions
This article is one of the first studies that investigated the political activities of SMIs. As the results show, addressing political topics is prevalent among SMIs since over 90 percent of our respondents have at some point of their careers brought up a political topic on their channels. This is particularly true for lifestyle-based political topics, which had been mentioned multiple times within a year by the majority of the SMIs included in our sample.
Our results also suggest that no single factor is connected to a higher frequency of mentioning both formal and lifestyle-based political topics. Rather, for lifestyle-based political topics, an important factor is the quantity and nature of the followers that a SMI has. In general, SMIs who have more followers tend to bring up lifestyle-based political topics more frequently. Similarly, SMIs whose followers interact with them more or react positively when they bring up lifestyle-based political topics also tend to bring these topics up more frequently. This suggests a positive feedback loop, whereby some SMIs get positive reinforcement from their followers, which encourages them to keep engaging with lifestyle-based political topics. Another important factor was the number of on-going collaborations that an SMI had; those with more collaborations were associated with a higher frequency of bringing up lifestyle-based political topics. This could be because SMIs who have participated in a larger number of collaborations are usually firmly established in their field. Their status gives them the opportunity to address political issues with more confidence and be less concerned about potential risks. They can also expect stronger support from their followers in the case of a crisis. However, it is likely that these two factors (number of followers and collaborations) are interlinked with one another. Our data showed a strong positive correlation between the number of followers and the number of collaborations (Spearman's ρ = .636, p < .001) which makes it difficult to determine what their independent impact is. The tendency to bring up formal political topics, on the other hand, was not influenced by the SMIs' followers. Rather, only those respondents who did not deliberately want to keep their social media channels as non-political as possible, and who did not believe that they had nothing to say about formal political topics were more prone to bring up formal political topics on their channel. This suggests that the tendency to bring up formal political topics is mainly driven by whether or not an SMI has made a conscious choice to opt-out from formal politics altogether.
Lastly, while it is rather common among SMIs to address political topics, avoidance of these is also occurring since the majority of the respondents reported that they have deliberately avoided them in the past. Overall, it seems that there exist two major reasons for this avoidance. First, the respondents did not want to be targeted by aggressive internet commentators. Second, the respondents generally feared what kinds of comments and conversations these topics could generate. Although, due to our relatively small sample size, we cannot investigate whether gender has an impact on the prevalence of these reasons, it is possible that female SMIs are more responsive to them since especially women who blog about political topics have been found to face a great risk of online abuse, through aggressive comments and even threats (Eckert 2018).
Before discussing the implications of these findings, a brief note on limitations is due. One obvious limitation of the current study is its small sample size and the non-random sampling method used. These issues challenge the representativeness of the results and prevented us from using multivariate statistical methods and lowered the statistical power of the analyses. However, this is one of the very few studies that uses a quantitative approach in studying SMIs. Most previous studies that were interested in SMIs have adopted qualitative designs for data collection and analysis (see e.g. Abidin 2016; Berryman and Kavka 2017; Raun 2018). Even non-optimal quantitative samples, we argue, can make a valuable addition to the existing body of research as it provides a new perspective on the scope of the phenomenon and advances our understanding of SMI communication. A second limitation is that our country context limits how generisable our results are, as Finnish SMIs are very likely to have their own unique characteristics. For instance, the Finnish SMI sphere is small in comparison to bigger European countries, which makes it less likely that Finnish SMIs are as professionalised as their global colleagues are.

Broader Implications for Social Media Influencers
Bearing the limitations of this study in mind, we believe that its results show that the lens through which SMIs are seen and understood within the current academic literature is overly narrow. As the results suggest, SMIs can have more diverse roles than mere "brand endorsers". Colliander and Dahlén (2011) have referred to fashion bloggers as "fashionable friends" as they can be perceived as important peer-endorsers by their followers regarding fashion choices. The results of this study show that the role of SMIs can similarly reach beyond "outfits-of-the-day" to "societal-topics-of-the-day" and thus they could be seen as "political friends" as well.
However, the way in which our respondents engaged with political topics varied significantly, and we argue that this variation is likely to be influenced by constraints and obligations that limit the SMIs' agency (Pang et al. 2016). We suggest that the decision to tackle political topics can be explained by the need for SMIs to practice self-branding strategies, which affect the decision to bring up politics in two interconnected ways. First, SMIs are under the obligation to appear authentic (Duffy and Hund 2019). They have to decide whether they want to endow their public persona with a political dimension that subsequently has to be performed in a credible way. Also, they have to evaluate if tackling political topics in general and formal political topics in particular is consistent with the public persona they have already created. Second, SMIs' success depends on their accumulation of person-brand capital (Delisle and Parmentier 2016). They therefore have to evaluate how this capital is affected by tackling political topics.

Broader Implications for the Political Sphere
Finally, although we cannot claim representativeness to the entire social media sphere with our current sample, the fact that over 90 percent of our respondents had brought up political topics suggests that online channels maintained by SMIs represent another third space where non-political online environments can become political. Moreover, there is good reason to suspect that when these third spaces emerge it can have a significant impact on their followers and through them on the broader political sphere. SMIs are largely defined by the strong relationships they have created among their followers, who tend to listen closely to what the SMI has to say. Thus, when SMIs raise both formal and lifestyle-based political topics, their followers are likely to pay attention and will have the chance to encounter politics in their day-to-day lives where such chances might otherwise be in short supply. If these topics then result in extended conversations between the SMIs and their followers or among the followers themselves, prior research has shown that it can lead to positive developments among individuals' political competence and behaviour Wright 2015, 2016).
Such best-case scenarios, however, might be hindered if the SMIs are more responsive to the potential threats that they see in discussions surrounding politics. Indeed, the results of this study show that the fear of aggressive internet commentators and general fear of comments and conversations is a strong motivator that makes SMIs hesitant to bring up political topics at all. Thus, the emergence of extended political discussions might be the very reason which prevents these third spaces from realising their political potential. If we are worried about the current state of political participation among the general public, as many political scientists are (Grasso 2016), it is worthwhile to explore THE EMERGENCE OF POLITICAL TOPICS AMONG FINNISH INFLUENCERS this dilemma further and to investigate what happens after SMIs have brought up political topics on their social media channels. Do these topics, for example, truly lead to extended political discussions that have positive impact on political behaviour, or will they attract aggressive internet commentators that the SMIs want to avoid?
An essential task of future research is to explore the context and style in which SMIs bring up political topics. We have little idea of the ways in which SMIs bring up, for example, environmental or social issues and whether these resulted in debates among the followers or between SMIs and followers. In-depth interviews with SMIs might make an important contribution to our understanding of these issues. Furthermore, while the empirical study presented in this article focused exclusively on SMIs, further studies would do well to investigate followers more closely and to examine how the political topics raised by SMIs impact them. Can engagement with these topics lead to, for example, increased interest in politics among followers and to political participation such as consumerism and even voting?

DISCLOSURE STATEMENT
No potential conflict of interest was reported by the authors.