Far-Right Politicians, Alternative media & Twitter
A explanatory study on far-right politicians use of alternative media
on twitter
By: Hugo Peterson
Words: 8588
Hugo Peterson
Bachelor's thesis in political science at Gothenburg University
SK1523
3-1-2023
Supervisor: Gefjon Off
Abstract
As alternative media and its dangers have become increasingly established in the last few
years and got to the forefront of Swedish politics because of the planned murder of Sweden
politician Annie Lööf, this study aims to look into if far-right politicians use alternative
media in order to spread their message using the research question: Do politicians of the
Swedish Democrats spread more alternative media on their social media accounts than
politicians of the mainstream conservative Moderates? The result of this research question
was a positive causality where Sweden democrats to a statistically significant degree do post
more links to alternative media than moderates politicians meaning they most certainly do.
The most interesting thing about this study may be that politicians with a higher position
within the party, and thus should be most in the know with the potentially damaging aspects
of alternative media, post the most. Suggesting that this posting very well can be a planned
strategy.
Innehållsförteckning
Introduction 1
Purpose 2
Literature review 2
Alternative media 2
Radicalization 3
Alternative media and radicalization 4
Party sympathy for radical parties & “dangerous worldview” 6
The relationship between politicians and voters 7
Social media and reputation 8
Theory 9
Hypothesis 10
Method 11
Collecting the data 13
Data analysis 14
Dataset 15
Results 16
Conclusions 21
References 24
Appendix 26
Introduction
Alternative media has become more and more at the forefront of Swedish politics. This is due
to, in part, the right-wing party Sweden Democrats gaining more and more prominence but
more importantly the planned murder of the Swedish politician Annie Lööf during the
political ‘Almedals’ week, by an alleged Nazi sympathizer (Expressen). Although the
planned murder never got past the planning stage, the to-be murderer instead murdered
another participant whom he did not agree with, Ing-Marie Wieselgren. Even though the
information that Lööf was the planned target only surfaced a while after the murder of
Wieselgren, it still sparked a big debate about alternative media, and radicalization, in
Swedish politics, which this study aims to look further into.
The party who has received maybe the most criticism due to this scandal is the Sweden
Democrats, the far-right party in Sweden, and, more specifically, the alternative media that
share the core principles with the party. This is because the Sweden Democrat-tied alternative
media were very critical towards Lööf specifically. Alternative media, with anti-immigration
tendencies, have in turn been linked to radicalization in studies by both Kaisler et al (2020)
and Nygaard (2019). This, in turn, lays the groundwork for this article as I will look further
into if far-right politicians spread these articles to a higher degree than other parties, such as
mainstream conservative politicians. We already know that far-right politicians use twitter to
spread their message. We also know that Politicians have an effect on their voters, and not
only the other way around, and we also think, with some evidence, that alternative media
have an impact on its readers' political beliefs. But we do not know if politicians use
alternative media as a sort of tool for spreading their ideology in their twitter usage, despite
its dangers, thus motivating this study.
In summary I will in this paper do an explanatory study where I look into the following: Do
politicians of the Swedish Democrats spread more alternative media on their social media
accounts than politicians of the mainstream conservative Moderates? The choice of the
Moderates as the comparison party is based on the fact that they are the established party that
lays closest to the Sweden Democrats concerning immigration policy thus is most likely to
1
consume the same alternative media as the Sweden Democrats based on the definition of
alternative media used in this study.
Purpose
By studying if the Sweden Democrats do indeed link more to alternative media than the
Moderates in Sweden, I aim to further add to the literature in the political communication
field, as I add information on whether or not far-right politicians actually use alternative
media, despite its dangers, to spread their message. If far-right politicians do spread
alternative media more than mainstream conservative politicians this could be a warrant for
further research on the effect alternative media has on its readers, for example, as it is a
subject not thoroughly researched and an area that could be a large actor in the increasingly
more polarized political spectrum in Sweden, and other Western countries.
If far-right politicians are proven to be a large distributor of alternative media, it would be
important for that to spark a debate in, mainly, Sweden about what role politicians should
have in our information society, should they be a ‘guiding light’? Or should they act merely
in the best interest of themselves and their party?
Literature review
This literature review will look into earlier research concerning alternative media and its
dangers, it will also look into how different political elites act on social media.
Alternative media
Firstly, we must establish what alternative media is, which is a lot more complicated than it
may seem. The direct and wide definition of alternative media is as follows:
2
“... media that challenge the established channels and put forward alternative approaches and
perspectives that contradict or diverge from an experienced dominant discourse in the mainstream
media” (Holt 2018)
Although it is when you try to specify it further that it gets a lot harder, because there are a
few different ways to go. But in this paper the definition that I am going to use is Holt’s
(2018) definition that is adapted to the Swedish context, as it is the best definition to capture
the alternative media that have ties to the Sweden Democrats or have an ideology that
matches best with the Sweden Democrats. This definition goes as follows:
“A self-assumed term that signals an opposition to traditional media (“old media”), which many of
the writers in this field regard as failing to report properly on important societal issues, for example,
by avoiding reporting on social problems related to immigration” (Holt 2018)
In this paper I am purposefully looking at alternative media that are just immigration critical,
as they are more linked to the right of Swedish politics. Holt (2018) mentions a few
alternative media that fit the example and will be used in this paper. These alternative media
are: Fria Tider, Samhällsnytt, Nya Tider, Nyheter Idag och Ledarsidorna.
Alternative media and radicalization
As we now well know there are different definitions of alternative media but, as already
explained, the far-right definition that Holt (2018) uses is the one that this study will use. The
thing that these alternative medias have in common are that they are more critical towards
immigration, aswell as islam, than mainstream media, according to studies from the US
aswell as Scandinavia and central Europe (Nygaard, S. 2019. Kaisler, J et al. 2020. Holt, K.
2018).
In Nygaards article from 2019 she reviews the far-right alternative media in Sweden, Norway
and Denmarks whereas she finds two different types of alternative media, named descriptive
and normative. These two types of far-right alternative media are opposing extremes, in terms
of how they convey their messaging, whereas the normative is mainly found in Denmark and
the descriptive one is the main one in Sweden, Norway's alternative media can be found ‘in
between’ these two extremes.
3
Out of the two types of alternative media presented in Nygaards (2019) study the most
interesting for this study is the descriptive one as that is the one found in Sweden. When
Nygaards (2019) explains the descriptive style she refers to a quote from former ‘Chicago
daily news’ editor Paul Mowrer where he highlights the importance of “the appearance of
objectivity”:
“Thus it is better to scrap an editorial calling the mayor a liar and a crook, and to write another
which, by reciting facts without using adjectives and without calling names, makes it obvious that the
mayor is a liar and a crook" (Mott and Casey 1937 in Nygaard 2019, p.1159)
In this quote Mowrer rather explicitly, depicts how Sweden's alternative media works
according to Nygaard where they try very hard not to explicitly say what they think about an
event and instead try to ‘guide’ the reader to the same conclusion. This ‘false objectivity’
Nygaard argues is the same thing as political propaganda.
Nygaards (2019) findings is something that is backed up in Odağ et al. (2019) literary review
with the name of “Reviewing the role of the internet in radicalization processes”. In this
review, where they are generally critical towards most research on radicalization on the
internet, they do however find earlier research that point out that right-wing sites do their best
to cloak their connection to the extreme right-wing communities. They do this by portraying
democratic values such as education and freedom of speech but beneath this they spread
white supremacy propaganda by using rhetorical devices in order to legitimize these
right-wing groups.
The third study looked at in the context of alternative media and its link to alternative media
is a study by Kaisler et al.(2020) in which the do a topic modeling study, which is a kind of
statistical model that looks for topics in a large text base, and hyperlink analysis, which looks
for information through hyperlink networks, from several alternative media in the United
States. In this study they see a clear linkage between size and professionalization of
alternative media, something that has been seen as something good for democracy for a long
time. But in this study Kaisler et al. (2020) problematize this as they do see that this
professionalization does not lead to social inclusion and is not progressive, in turn they look
at the effect this professionalization has on the mainstream media in America. They argue
4
that when right-wing media becomes more professional and thus becomes more mainstream
that in turn pulls the mainstream media more towards them meaning that the whole media
landscape radicalizes to some extent (Kaisler, J et al. 2020).
The last study looked at in this section of the literary review is a study by Leung & Lee
(2014) which is not perfect for this study as they looked at alternative media in Hong Kong
and thereby ‘left’ alternative media. Why this Leung & Lee study is still relevant to this study
is because they are the only ones, I could find, who have even remotely looked at the
empirical political impact from alternative media. The precious studies presented have rather
looked at alternative media in itself and then speculated, based on their research, that its
aspects should have an impact. In Leungs & Lees (2014) study they set out to see if internet
alternative media do in fact have an impact on the readers, and found support for their
hypothesis: “Internet alternative media usage relates positively to protest participation.” and
“Internet alternative media usage relates positively to support for the planned civil
disobedience campaign for universal suffrage.”(Leung & Lee 2014; 8). Although they did
find support for their hypothesis they did concede that they could not determine which way
the causality went.
To summarize these studies we can look at three aspects, firstly they find that alternative
media, as defined in this study, often present themselves as objective, an objectivity that turns
out to be a sort of political propaganda called ‘false objectivity’. Secondly, alternative media
may also have an effect on mainstream media by pulling them closer to themselves in the
process of their own professionalization, meaning that the whole political spectrum in the
country may be affected. Lastly, alternative media seem to have a political effect on its
readers.
The relationship between politicians and voters
Two separate studies, made by Bischof and Wagner (2019) and Silva (2018) have both looked
at the correlation between radical parties and political polarization, and thus the effect
politicians have on voters, in the Netherlands since the breakthrough of radical right parties in
their 2002 election. Both of their studies did indeed find that the political polarization of the
voters has increased since the emergence of these parties and that this increase is higher than
5
it would have been without them. This is then further established in Castanho Silvas (2018)
article when he tests the increase in polarization in relation to another 13 Western European
countries, in order to get a general polarization to compare with, where the Dutch case indeed
has a larger polarization. These findings then showed a larger effect on polarization of the
radical right as Bischof and Wagner (2019) find that this polarization could not be seen in
relation to the entry of radical green parties, which in turn suggest that this polarization is
maybe more driven by the radical right.
What makes these studies interesting in the case of this study is that they show that politicians
have an effect on the voters' views, and not just the other way around. This then begs the
question as to what politicians do with this influence and maybe more importantly why
far-right politicians seem to have an even larger impact on voters than other politicians. This
inturn is a motivation to this study as it is looking at whether or not there is a difference
between far-right politicians and mainstream conservative ones concerning the spread of
alternative media, which in turn could be part of an explanation to this question.
Politicians use of Twitter
In a study from 2016 that van Kessel and Castelein named “Shifting the blame. Populist
politicians’ use of Twitter as an oppositional tool” they look into how two political leaders of
populist parties in the Netherlands used Twitter. By looking at and comparing the twitter
feeds of Geerts Wilders, Party for freedom(PVV), and Emile Roemer, former leader of the
Socialist party(SP) they tried to answer the following question: Did their Twitter usage reflect
well the anti-establishment strategies of populist parties, and can the medium therefore be
considered suitable to study the construction of populist injustice frames?
The answer to this question was generally yes, meaning that the twitter feeds of both these
political leaders, to a certain degree, matched the party's political manifesto in criticizing their
relevant issues. Meaning that they very well used twitter as a tool of political opposition. But
the main difference between the two party leaders was that the far-right leader of the PVV,
Geert Wilders, was much less consistent with its messaging than Emile Roemer. This
basically means that Roemer to a larger degree kept tweeting about the same views and
opinions while Wilders used to switch between different political actors during the time
6
period depending on the relevant issues at hand and the current political landscape. The result
of this is that the main issues of Wilders ideology, being opposed to progressive elites,
immigration and islam, sometimes took a backseat to more adverserviable tweets depending
on the issues at hand. The target of the criticism by the ideology also changed during the time
period depending on, as said, the issues at hand and the political landscape, as when PVV
chose not to support the centre-right VVD-CDA minority government anymore(van Kessel &
Castelein 2016).
This result is further backed up in a study from 2017 by Engesser et.al in which they also
studied populist politicians' use of twitter, but in this case also facebook. In this study they
looked at a few different European countries, Austria, Switzerland, United Kingdom and
Italy, which gives the result further legitimacy and generability. The result of this study also
showed that populist parties used social media as a tool for the spread of its ideas, although
the major finding of this study showed that these ideas only were spread in a fragmented
form.
Lastly, to put this into comparison to more established politicians, Jacobs et.al (2020) found
in a study that populist politicians were more likely to be “...naming and shaming on Twitter
and anger activation on Facebook”(Jacobs et.al 2020; 628). This result is interesting in the
context of this study as it shows a clear divide in the usage of social media between populist
and non-populist politicians, something that this study also looks at.
These findings are interesting for this study as we now know that populist politicians to a
higher degree use social media to push their narrative, or ideology, than more established
politicians. We also know, referring to the last chapter, that politicians, especially far-right
politicians, have an effect on voters. Lastly, we think that alternative media does have an
impact on its readers. These three findings point to a potential answer to the question of this
study, Do politicians of the Swedish Democrats spread more alternative media on their social
media accounts than politicians of the mainstream conservative Moderates? These findings
suggest that they should post more alternative media links than the Moderates as, for one,
Sweden Democrat politicians share the same basis of ideology as the alternative media and
thus can use them to spread their message. Secondly, we know that politicians have an effect
on their voters, and not only the other way around, and we also think that alternative media
7
has an effect on its readers meaning that exposure to the narrative of alternative media
should, theoretically, lead to a voter gain for the Sweden Democrats.
Social media and reputation
Although there is evidence to suggest that far-right politicians should spread links to
alternative media on their twitter there is another aspect that also should be looked at. In a
study by Kalsnes (2016) with the name of “The social paradox explained: Comparing
political parties’ facebook strategy versus practice” Kalsnes looks into which strategies
political parties in Norway have with social media and then how the strategies played out.
Quickly summarizing it, the conclusion was that parties generally have grander strategies for
social media than what they then actually turn out doing during the campaign. What is
interesting from this in the context of this study is that party leaders are generally bad at
prioritizing social media, something that is based on a couple of arguments. The most
interesting argument in terms of this study is that there is a lack of control of online
reputation and negative media attention when a politician/party is active on social media
(Kalsnes 2016). An argument that could be used in this study aswell when looking at
politicians. This is because the same motivation could be used as to why someone with a high
position within the party, as a party leader, could be afraid to post links to alternative media,
as alternative media generally does not have a good reputation in Sweden meaning that they
do not want that bad reputation ‘rubbing’ of upon them. This could be a reason as to why we
would not see more links to alternative media from Sweden democrats in relation to
Moderates.
Theory
According to the earlier research far-right politicians do use social media, and twitter
specifically, in order to spread their ideologies. This in itself would suggest that far-right
politicians should post more links to alternative media as they share the same core principles
of their ideology meaning that alternative media could be a way of further spreading that
message. To substantiate this theory even further we also know that politicians have an effect
on voters. This combined with the small amount of evidence that suggest alternative media
8
have an effect on voters should motivate far-right politicians into posting links to them
regularly, as it could be used as a tool to spread political propaganda as Nygaard (2019)
theorizes. But in order to actually know if they see a benefit to sharing alternative media to
their followers, we need to know if they in fact share it regularly. This in turn leads us to the
causal model of this study, see figure 1.
Figure 1. The causal model of this study
By testing this model, I hope to establish if far-right politicians link more to alternative media
than other politicians. If this causality turns out to be positive it would suggest that these
politicians indeed use alternative media to spread their ideology, even though they should be
aware of the negative aspect of alternative media. It can also suggest that they use Alternative
media to spread political propaganda in a more hidden way, as the media are not directly
linked to the party. If the causality turns out to be negative that would start to ask questions as
to why they do not use alternative media to help them spread their ideology, is it because they
do not see any benefit to it and is that then related to them not wanting to harm their
reputation.
The politicians position within the party will also be looked at as an independent variable
because of Kalsnes(2016) research as it may have a rather large impact on the result of the
study. If it is so that politicians with a higher position within the party, and thus have a larger
reputation to uphold, are less likely to be controversial on social media because of the
hardship in controlling the narrative it could have a negative effect on the causality, thus the
need of having it as an independent variable. This variable could be looked at as a risk versus
reward kind of scenario where someone with a higher position, and thus reputation, should
have the largest positive effect of spreading alternative media on social media; but they are
also the ones with a reputation to lose. Another potential scenario could be that this only
affects the Moderates, their party leader Ulf Kristersson does not have a twitter for example,
9
as they are more established and seen as non populist leading them to be more careful in not
being controversial, something that could have a positive effect on the causality. The amount
of followers will also be looked at as a control variable as followers potentially could have
the same effect, as those who have a larger following on social media also could want to be
more careful.
Hypothesis
Based on the above literature and theoretical framework, I hypothesize that:
H1: The far-right politicians in Sweden, the Sweden Democrats, post more links to
alternative media articles than the mainstream conservative politicians in Sweden.
Hypothesis 1 is based on earlier research. Considering that we know that far-right politicians
to a high degree use social media, and most importantly for this study, twitter in order to
spread their message and that combined with the findings that Alternative media and
politicians has an impact on their readers and voters respectively it i thus then highly likely
that far-right politicians would use alternative media as a tool in spreading their ideology.
This is because it seems, from earlier research, that it would have a positive effect for their
party and thus individual politicians. This is in spite of the fact that it may have a negative
effect on their reputation as it is difficult to have control of the narrative on social media and
thus running the risk of the bad reputation of alternative media spilling over on you. This is as
the potential rewards outways the risk.
H2: People with a higher position within the party post fewer links to alternative media than
their political colleagues with lower positions
Although the potential reward should outway the risk for the average politician that may not
necessarily be true for the politicians with a higher position within the party, as reputation and
legitimacy should be much more important for a party leader than a ‘regular’ parliament
member for example. This could make for an interesting scenario where primarily the people
with a lower position within the party post more links to alternative media as they do not
have the same risk hanging to it. This could however be looked at in a completely different
10
way as they, the politicians with a lower position within the party, probably won't have the
same reach and thus same reward. It could also be so that this only affects the Moderates as
they are seen as more established, and thus legitimate, leading them to be more careful. But
for the interest of this study a path would have to be chosen and my, personal, first instinct
was that people with a higher position would post less links due to the risk of a hampered
reputation and thus hypothesis 2 was based on that argument.
Method
Firstly, I have to establish which country this study is going to be based in. The country
chosen is Sweden because of the fact that the motivation for this study is based on an event in
Swedish politics, the attempted murder of politician Annie Lööf. Next up is the choice of the
first party studied. The criterion of this party is that it needs to be far-right. The only party in
the Swedish parliament that can be considered far-right is the Sweden Democrats, which
made their selection straightforward. The party needs to be in the parliament because, as we
can see in the earlier research, it is mainly when a far-right party enters parliament that it
starts to affect the voters (Bischof, D. Wagner, M. 2019; Silva, C. 2018).
The next step will be to decide where we will find the potential posts where the Sweden
Democrats’ politicians link to the alternative media articles. Here, only Twitter is chosen due
to the time limit and its high quality search engine. This is of course not perfect as Twitter is
not the largest social media available to me, Facebook and Instagram are both larger. But due
to Twitter's ‘advanced search’ engine, where you can be very specific when you search, it
was the obvious choice. The alternative media that will be searched for is the ones mentioned
in Holt’s (2018) article: Fria Tider, Samhällsnytt, Nya Tider, Nyheter Idag och Ledarsidorna.
The next step is to narrow the study down in scope. To do so, I will only look at politicians
that have been elected to the Swedish parliament in the latest election, as they are the ones
who have an effect on voters (Bischof, D. Wagner, M. 2019; Silva, C. 2018). This does
however come with the negative effect that they may have ‘cleaned’ up their twitter account
when they ran for this seat, because alternative media still generally have a bad reputation in
Sweden. The number of their Twitter followers will be looked at as a control variable
whereas position within the party will also be coded as an independent variable. Position
11
within the party will be looked at as it will give us the answer to H2, people with a higher
position within the party post fewer links to alternative media than their political colleagues
with lower positions. The amount of followers will also be looked at as it could have the
same effect, as politicians with more followers are more likely to be less controversial on
twitter in order to not upset people and thus keeping their high number of followers.
The third part is to decide which comparison party I should look more into, in order to
establish whether or not the Sweden Democrat politicians post more alternative media than
other politicians. The main criterion for the comparison party is that they are the closest to the
Sweden Democrats politically. If they are too far apart politically, I run the risk that they still
consume alternative media but just not the Swedish, immigration critical, version that I am
looking at. In other words, a party whose politicians, and voters, would read the same
alternative media as Sweden Democrats if they were to consume alternative media is desired.
The Moderates is the party best aligned with the Sweden Democrats concerning the
immigration policy according to CHES (2019). Meaning they are the party most likely to be
best aligned ideologically to the alternative media chosen in this study.
Another part that would need to be decided is which media sources are going to represent
‘mainstream’ media, in order to be able to use the search engine on Twitter which in turn
saves a lot of time. As this study will be conducted using “advanced search” on twitter I need
to decide which links should be considered mainstream media. The media chosen were
mainly based on a survey conducted by SIFO where 45 000 random people of the Swedish
public asked them about their media consumption. Out of the papers mentioned in the SIFO
study I chose the first five, as the sixth one was a local paper. The first five were: Dagens
Nyheter, Aftonbladet, Expressen, Svenska Dagbladet and Dagens Industri. A few other media
was also added on an inductive basis from me looking at various media often linked from
politicians. These were: TT nyhetsbyrå, Omni and SVT. SVT can be considered mainstream
as they are the public service media for Sweden and thus are very highly used, TT are a news
agency who also can be considered mainstream as Aftonbladet, and many other newspapers,
often buy articles and have much collaboration with them. Omni can also be defined as
mainstream as they are a so-called “news aggregator” which basically means they mainly
summarize the reporting of other media.
12
The last aspect that needs to be established is the time aspect. This is about how long I will go
back in time on the politician’s Twitter profile while searching for the posts. This is mainly
because of how long it will take to gather all the information. So, considering this, I will only
go back to when they first entered the Parliament plus one year, as that would have roughly
been when they got put on the list to enter parliament, and that would be the first time they
would benefit from spreading alternative media to their potential voters. If they have had gap
years during their time in political office, I will only go back to the most recent time in order
to not make it too difficult.
Collecting the data
As already mentioned, Twitter is the social media chosen due to its very high quality search
engine called “advanced search”. Advanced search is basically a system to help you navigate
through all the different commands that can be used in the search bar and to help you
formulate them, this means when you fill in the parts you want the engine to search for you
will get a special ‘code’ that will run that search. As for this study I got “(samnytt.se OR
nyheteridag.se OR nyatider.nu OR friatider.se OR ledarsidorna.se) (from:) since:-12-31
filter:links -filter:replies” . The OR in the command means that the engine will show every hit
it gets from either of the keywords, the sites, searched for. Meanwhile, “from:” entails which
account the tweet shall come from, meaning the different politicians’ accounts. “Since:” is
then the time period, meaning in this case one year before the politicians entered parliament
most recently. To determine the time period, I searched for the politician on “Riksdagen.se”
to see when they entered, and then went back in time by one year. In order to save time, I did
not look at the exact date, only the year, meaning 12-31; month/year was exactly the same for
every politician. “filter:links” means that only posts where links were shared came through
the search engine. “filter-replies” meant that replies to other people were not shown, as the
study wanted to only look at the articles the politician actively wanted to distribute.
This search basically searches for all links shared that contain the alternative media filled in
from a certain date. Then when looking into the different politicians from each party I put
their username after “from:” and the year they entered the year they ran for parliament and
got in after “since:”, this was then done for every politician looked at as well as the two
respective parties twitter, but then using the date they joined twitter as the start of the time
13
period. This practically means that if two different people entered, let's say, 2018 and 2020
respectively they both would get 2017 as they both were put on the list to the parliament
roughly one year before the election; depending on party rules. This means that politicians
who replaced other politicians on leave and then got elected were counted from the time they
stood on the list as replacement for the year they got the chance to act in the parliament for
the first time. The exact same thing was then done when searching for the mainstream media
articles but this time instead with the mainstream media websites, giving me the following
code: “(aftonbladet.se OR dn.se OR di.se OR expressen.se OR svd.se OR tt.se OR omni.se
OR svt.se) (from:) since:-12-31 filter:links -filter:replies”.
Data analysis
With the material collected for this study there are three variables that I look further into,
based on the above-developed hypotheses. The main one would be to see if Sweden
Democrat politicians do post links to alternative media more often than the Moderate
politicians. The main way of doing this is by firstly adding all links to alternative media that
Sweden Democrats politicians have shared together. Then doing the same with the Moderates
politician, then dividing this sum by the number of investigated politicians from each party.
With this method I will get the arithmetic mean of the average number of alternative media
links per politician of the respective parties. By comparing these two, I will get the average
number of posts to alternative media made during the investigated time period from the
Sweden Democrats and the Moderates.
The main problem with arithmetic means though is that they are very sensitive to extreme
cases. If one politician posts a lot of links this would of course make the average go up,
something that can build a skewed picture. Because of this I am going to use another measure
of central tendency called the median. Calculating the median works by basically lining up all
the number of posts per politician, ranked by size, before counting each and every one until
you get up to half the number of respondents, you calculate the 50th percentile statistically
speaking (Esaiasson et al., 2017, pp. 364-366; Teorell & Svensson, 2007, pp.116-117). By
doing this, I can in a better way get rid of the most extreme linkers, the ones that post
abnormally many links, and instead see what number of links most of the politicians post per
party.
14
When this data collection and descriptive analysis has been made and a mean and median
value have been revealed it is time for further analysis. The next part of the analysis will be a
so-called bivariate regression analysis, in which I will be able to further establish whether
there is a positive or negative relationship between the party and the number of alternative
media links posted. In this analysis, being a parliamentarian of the Sweden Democrats or the
Moderates constitutes the independent variable and the number/share of posts linking to
alternative media constitutes the dependent variable. The Moderates will be given the number
0 and the Sweden Democrats the number 1. This would mean that, if the hypothesis is
correct, that we would get a positive correlation between the party and the number of
alternative media links posted. This would be seen with the help of bivariate regression
analysis and then further determined with the use of the p-value in order to establish whether
or not the result is statistically significant. Then I will add the independent variable and the
control variable in order to see if they have any effect on the result. The same bivariate
regression analysis will then be run but this time looking at the share of alternative media
links posted in relation to total links posted to both alternative and mainstream media. Lastly,
both of these bivariate regression analysis will be run with followers and position in party as
a control and independent variable respectively in order to see which effect they have and if
H2 thus is correct. Position in party will be coded with a 1 given to them with a high position,
party leadership, committee chairman and cabinet members, and a 0 given to the rest of the
politicians.
Dataset
The data collection has left me with 82 politicians representing the Moderates and 74
politicians representing the Sweden Democrats, meaning 156 politicians altogether. In
addition to politicians’ Twitter feeds, this number does not include the official Twitter
accounts of the parties. Although the Sweden Democrats received more votes in the last
election, and thus has more politicians in parliament, the study looked at more Moderates as
13 of the original parliament politicians ended up in the cabinet1. Because of this I decided to
1 If a politician from the parliament ends up chosen into the cabinet by the prime minister, he/she will
have a replacement in parliament as they can no longer serve in it. Ministers and their replacements
in parliament have both been included in this data collection.
15
also look through these politicians' feeds as well, as they are some of the most high-profile
politicians in the moderates, as their replacements in parliament. Three of the Moderates in
the cabinet were not elected to the parliament, despite this they were also studied as they are
seen as people with a high position in the party. Out of the combined 156 politicians looked
at, 43 did either not have a Twitter account or have not tweeted from their account, and are
thus excluded from the sample. 5 politicians did not tweet more than 3 tweets and were
therefore excluded from the sample, as these tweets were just replies. In total, the data
collection thus comprised posts from the Twitter accounts of 113 politicians.
Results
The combined number of links to alternative media from Sweden Democrat politicians is
301. 20 of them were posted by the official Twitter account of the Sweden Democrats and 9
were posted by their party leader Jimmie Åkesson. The biggest alternative media posters
were Jonas Andersson, with 69 links, and Tobias Andersson and Björn Söder with 51 links,
respectively, all three of these have a place within the party leadership and thus are given a 1
in regards to position in party. 301 divided by 74, which is the number of Sweden Democrats
politicians in parliament, gives us an arithmetic mean of circa 4.1 (counting all politicians of
the Sweden Democrats, even those without Twitter). If you exclude the 21 Sweden
Democrats politicians without Twitter accounts, the arithmetic mean goes up to circa 5.7
posted links per Sweden Democrat politician. The mean number of shared posts by Sweden
Democrat politicians is in stark contrast to the respective number for Moderates politicians:
0.15. Not counting the Moderates politicians without Twitter accounts, the average Moderates
politician shares 0.2 alternative media links throughout the investigated time period. In figure
2 below you can see the amount of posts to alternative media from every politician who
posted a minimum of 1 link to alternative media, both parties' politicians can be seen.
16
Figure 2. The box plot of alternative media posts by politician(of them who did post)
As mentioned, in this box plot you can see the number of posts to alternative media by each
politician who actually posted. The y-axis shows the politician and the X-axis shows the
number of posts, only Beckman, Ericson and Tolgfors represent the Moderates. As you can
see here, it is mainly five heavy posters who impact the average number of posted alternative
media links of the Sweden Democrats. Meanwhile there is mainly one of the Moderates
politicians who posts alternative media links, committee member Lars Beckman, although on
this plot he merely represents an average poster in contrast to the Sweden democrat posters.
The results further reveal that there are some heavy posters of alternative media links in both
of the parties. Here is when the median link posted per politician becomes interesting,
without counting the politicians without Twitter accounts. The median number of alternative
media posts per Moderate politician is 0, considering that only 3 Moderates actually posted
links to alternative media. For the Sweden Democrats, the median number of alternative
media posts by politicians is only 1 compared to the arithmetic mean of 5.7. This shows that
the heavy posters have had a rather large impact on the arithmetic mean for the Sweden
Democrats.
17
Table 1. The bivariate regression analysis
In table 1, you can see the results of a bivariate regression analysis between being a Sweden
Democrat (vs. Moderate) politician and the number of shared alternative media links. Here
we can see that the party has a statistically significant impact on the number of alternative
media links shared on Twitter per politician (p-value < 0.01). The coefficient shows that a
politician of the Sweden Democrats would post 5.2 more links to alternative media during his
or her time in parliament and a year before than a Moderates politician. If you look at the
constant, you can also see that a Moderates politician would post 0.188 links to alternative
media during their time in parliament and a year before.
Table 2. Multivariate regression
18
Table 2 shows the positive relationship between being a Sweden Democrats politician and the
number of shared alternative media links. We can also see that a politician’s position in a
party has a large effect concerning the result, even though this is not as statistically
significant. This means that people with a higher position in the respective parties, as defined
in the collection of data, in general are more likely to post links to alternative media. In
contrast, the number of followers does not have an effect on the number of links to alternative
media that a politician posts.
Now when we start to look at the share, rather than the absolute number, of alternative media
that's been linked in contrast to mainstream media by the politicians it gives us another
perspective on the results. But firstly, in order to work the data, we needed not only to remove
all the people who did not have a Twitter account but also all of those who have not linked to
either alternative or mainstream media, as you cannot divide 0 by 0. After calculating the
share of alternative media from all politicians who have linked to either mainstream or
alternative media, the same bivariate regression analysis with the same independent variables
was run, giving us figure 3 and table 3.
Figure 3. The box plot analysis looking at the share
19
Table 3. The independent variables in relation to the share
Here, when looking at figure 3, we see a much more even distribution. This is because this
figure is looking at the share or in other words the percent of alternative media shared per
politician of those who posted links to alternative media. The main thing to mention here is
that two out of the Moderate politicians who posted links to alternative media, Ericson and
Tolgfors, had a very low percentage, and thus share. When looking at the share of alternative
media and mainstream media in a bivariate regression analysis, table 3, we see an even more
statistically significant effect of being a Sweden Democrat politician vs. a Moderates
politician (p-value < 0.001). This is likely because only one of the politicians in the
Moderates, Beckman, scored a competitive share in contrast to the Sweden Democrat
politicians. On another hand, a politician's position in the party and number of followers is
not significantly related to the share of alternative media sources shared by a politician
anymore.
Some more interesting findings in the collected data that should be mentioned is that Jimmie
Åkesson, the party leader of the Sweden Democrats and the person with the most followers
on social media, did in fact post to alternative media regularly, 9 times, which is a share of
circa 41%, since his re-entry to parliament in 2015. Another finding that goes against H2 was
that the official Sweden Democrat page on Twitter posted even more links to alternative
media than Jimmie Åkesson, posting 20 links since its creation in 2010, although it is only a
share of 33.33%.
20
Conclusions
The main conclusion to be drawn from this study is that H1 “The far-right politicians in
Sweden, the Sweden Democrats, post more links to alternative media articles than the
mainstream conservative politicians in Sweden.” is confirmed and that the answer to the
question: “Do politicians of the Swedish Democrats link to more alternative media on their
social media than politicians of the mainstream conservative Moderates?” is yes. As
presented in the results, there is a statistically significant positive relationship, meaning that
politicians of the Sweden Democrats do in fact post more links to alternative media, during
their time in parliament and a year before, than the Moderates. The relationship is even
stronger when looking at the share of alternative media posted in relation to mainstream
media giving further evidence to support H1. This result shows what is at best carelessness
from Sweden democrat politicians regarding the dangers of alternative media, who politicians
at parliamentary level which are supported by a team of communicators and consultants
should know, and at worst a calculated effort in trying to gain party support with the help of
the ‘false objectivity’ that alternative media uses as Nygaard(2019) mentioned. The result
also starts to highlight a possible answer, or part of an answer, as to why far-right politicians
have a larger impact on political polarization than other ideologies, something that should be
researched further.
A result that could suggest that it is the latter is the answer to H2 “People with a higher
position within the party post fewer links to alternative media than their political colleagues
with lower positions” was not confirmed. This means that politicians with a higher position
within the party, and thus are working closer to the communicators hired by the party, are not
less likely to post links to alternative media sources. This is even more emphasized when
looking at the parties’ own official twitter page, which presumably is run by the
communicators hired to handle the communication for the party, that in turn is the fourth
highest poster of links to alternative media of all accounts looked at in this study. This
suggests that even those who should have the most knowledge about the fallacies of
alternative media, the communicators, still choose to regularly post links to it, meaning they
probably think it is a net gain for the party. It also suggests that the bad reputation they
eventually could get from it, based on the arguments presented by Kalsnes (2016), either does
not exist, as alternative media maybe does not have a bad reputation, or that the politicians
21
and their communicators make the call that it does not matter in regard to the net gain of
spreading alternative media.
If we now move on to look at some limitations of this study, one is that it does not actually
prove if alternative media spread actually have an effect on voters and ‘only’ gives us an
suggestion as to whether or not the communicators and politicians of the Sweden Democrats
think that the spread of these alternative medias, with all of the faults mentioned mainly by
Nygaard (2019), may lead to them getting more support in the future. But this is maybe also
the best thing about this study because what this study does in fact show is that there
definitely is something here that needs to be further studied by the political communication
field, something that I suggest the field should do. One good place to start may be to research
more about if alternative media has an effect on its readers' political beliefs.
If we continue to look at deficiencies with the study two that probably did have an impact is
that the study did not look at what the individual politicians wrote when linking the articles,
they could have supported it as well as criticized it. This should have, too a small part, been
mitigated by looking at the share of alternative media links, as this measure takes into
account both mainstream and alternative media.
The last criticism that I have of my own study is that twitter’s code did not include retweets,
meaning that retweets were not accounted for in the study. However, this should only have
had an impact on the number of links behind the study as it should have had a similar effect
on both mainstream and alternative media. This is because all of the media looked at does
have individual twitter sites, meaning they did not have to post links in order to show the
article, they should have had the option to retweet for both alternative as well as mainstream
media.
This result should also call into question which role politicians in general have in today's
information society. Should they act as a guiding light? Or only post what best fits their
narrative and thus have a positive impact on their individual careers/their party's successes? If
it is the first option then both mainstream media and the general public should to a greater
degree call out politicians who do not act according to the norm of being a ‘guiding light’.
22
References
Bischof, D. Wagner, M. (2019). “Do Voters Polarize When Radical Parties Enter
Parliament?”. American Journal of Political Science.
Engesser, S. Ernst, N. Esser, F. Büchel, F. (2017). “Populism and social media: how
politicians spread a fragmented ideology”. Information, Communication & Society.
Esaiasson, P. Gilljam, M. Oscarsson, H. Towns, A. Wängerud, L. (2017). “Metodpraktikan:
Konsten att studera samhälle, individ och marknad”. Fifth edition. Wolters Kluwer.
Göthlin, E. and Ronge, J. (2022). “Annie Lööf (c) Tilldelad Målsägandebiträde Efter Mordet
I almedalen”. Available at:
https://www.expressen.se/nyheter/theodor-engstrom-planerade-att-morda-annie-loof/
(Accessed: January 3, 2023).
Holt, K. (2018). “Alternative Media and the Notion of Anti-Systemness: Towards an
Analytical Framework”. Media and communication.
Jolly, Seth, Ryan Bakker, Liesbet Hooghe, Gary Marks, Jonathan Polk, Jan Rovny, Marco
Steenbergen, and Milada Anna Vachudova. 2022. “Chapel Hill Expert Survey Trend File,
1999-2019.” Electoral Studies 75 (February). https://doi.org/10.1016/j.electstud.2021.102420
Jacobs, K. Sandberg, L. Spierings, N. (2020). “Twitter and Facebook: Populists’
double-barred gun?”. New media & Society.
Kaisler, J. Rauchfleisch, A. Bourassa, N. (2016) “Connecting the (Far-)Right Dots: A Topic
Modeling and Hyperlink Analysis of (Far-)Right Media Coverage during the US Elections”.
Routledge.!
Kalsnes, B. (2016). “The social media paradox explained: Comparing political parties’
facebook strategy versus practice”. Social media + Society.
23
Leung, K. Lee, F. (2014). “Cultivating an Active Online Counterpublic: Examining Usage
and Political Impact of Internet Alternative Media”. The international journal of
press/politics.!
Nygaard, S. (2019). “The Appearance of Objectivity: How Immigration Critical Alternative
Media Report the News”. Routledge.!
Odağ, Ö. Leiser, A. Boehnke, K.(2019). “Reviewing the Role of the Internet in
Radicalization Processes”. Journal for Deradicalization.!
Silva, Bruno Castanho. (2018). “Populist Radical Right Parties and Mass Polarization in the
Netherlands”. European Political Science Review.!
Teorell, J. Svensson, T. (2007). “Att fråga och att svara: Samhällvetenskaplig metod”. Liber.
Topplista Sveriges största tidningar. Ocast. Available at:
https://ocast.com/se/toplists/orvesto-konsument-api (Accessed: December 16, 2022).
Van Kessel, S. and Castelein, R. (2016). “Shifting the blame. Populist politicians’ use of
Twitter as a tool of opposition”. Journal of Contemporary European Research.
24
Appendix
All the politicians looked at:
Parti Namn
SD Anders Alftberg
SD Jonas Andersson
SD Lars Andersson
SD Pontus Andersson
SD Tobias Andersson
SD Clara Aranda
SD Mats Arkhem
SD Ludvig Aspling
SD Angelika Bengtsson
SD Sara-Lena Bjälkö
SD Anna-Lena Blomkvist
SD Carita Boulwén
SD Bo Broman
SD Mattias Bäckström
Johansson
SD Alexander
Christiansson
SD Dennis Dioukarev
SD Staffan Eklöf
SD Aron Emilsson
SD Matheus Enholm
SD Mattias Eriksson Falk
25
SD Yasmine Eriksson
SD Mikael
Eskilandersson
SD Rashid Farivar
SD Runar Filper
SD Ann-Christine From
Utterstedt
SD Nima Gholam Ali
Pour
SD Rasmus Giertz
SD Jörgen Grubb
SD Roger Hedlund
SD Erik Hellsborn
SD Richard Jomshof
SD Patik Jönsson
SD Mattias Karlsson
SD Martin Kinnunen
SD Julia Kronlid
SD Fredrik Lindahl
SD Linda Lindberg
SD Ulf Linholm
SD Angelica Lundberg
SD David Lång
SD Adam Marttinen
SD Thomas Morell
SD Mats Nordberg
SD Katja Nyberg
26
SD Mona Olin
SD Eric Palmqvist
SD David Perez
SD Daniel Persson
SD Magnus Persson
SD Charlotte Quensel
SD Anette Rangdal
SD Patrick Reslow
SD Michael Rubbestad
SD Oscar Sjöstedt
SD Jessika Stegrud
SD Robert Stenqvist
SD Carina Ståhl Herrstedt
SD Jimmy Ståhl
SD Johnny Svedin
SD Sven-Olof Sällström
SD Björn Söder
SD Per Söderlund
SD Victoria Tiblom
SD Björn Tidland
SD Beatrice Timgren
SD Henrik Vinge
SD Martin Westmont
SD Eric Westroth
SD Elsa Widding
SD Marcus Wiechel
27
SD Lars Wistedt
SD Leonid Yurkovskiy
SD Jimmy Åkesson
SD Sverigedemokraterna
M Fredrik Ahlstedt
M Emma Ahlström
Köster
M Ann-Sofie Alm
M Alexandra Anstrell
M Kristina Axén Olin
M Lars Beckman
M Sten Bergheden
M Jörgen Berglund
M Helena Bouveng
M Camilla Brunsberg
M Crister Carlsson
M Margareta Cederfelt
M Mikael Damsgaard
M Ida Drougge
M Lars Engsund
M Karin Enström
M Jan Ericson
M Mats Green
M Gustaf Göthberg
M Ann-Charlotte
Hammar Johnsson
M Ulrika Heindorff
28
M Johanna Hornberger
M Johan Hultberg
M Marie-Louise Hänel
Sandström
M Malin Höglund
M Caroline Högström
M Kjell Jansson
M Lars Johnsson
M David Josefsson
M Ellen Juntti
M Arin Karapet
M Mattias Karlsson
M Fredrik Kärrholm
M Ann-Sofie Lifvenhage
M Merléne Lund
Kopparklint
M Josefine Malmqvist
M Noria Manouchi
M Louise Meijer
M Marie Nicholson
M Ulrik Nilsson
M Carl Nordblom
M Charlotte Nordström
M Susanne Nordström
M Peter Ollén
M Stefan Olsson
29
M Erik Ottoson
M Lars Püss
M Saila Quicklund
M Thomas Ragnarsson
M Johanna Rantsi
M Magnus Resare
M Adam Reuterskiöld
M Edward Riedl
M Jessica Rosencrantz
M Oliver Rosengren
M Anna af Sillén
M Jesper Skalberg
Karlsson
M Maria Stockhaus
M Helena Storckenfeldt
M Oskar Svärd
M Magdalena Thuresson
M Katarina Tolgfors
M Hans Wallmark
M John E Weinerhall
M Jennie Wernäng
M John Widergren
M Viktor Wärnick
M Boraina Åberg
M Moderaterna
M Ulf Kristersson
30
M Jessika Roswall
M Elisabeth Svantesson
M Niklas Wykman
M Pål Jonson
M Carl-Oskar Bohlin
M Gunnar Strömmer
M Maria Malmer
Stenegard
M Parisa Liljestrand
M Anna Tenje
M Camilla Waltersson
Grönvall
M Tobias Billström
M Johan Forssell
31