Cumulative disinformation through citation: Investigating the longitudinal construction of Sweden as the ‘rape capital’ of the world

Abstract

Sweden has increasingly come into international far right focus, and digital media sites in the United States have proven especially active in spreading mis- and disinformative narratives about the country. Using social network analysis and descriptive statistics, we trace 776 links over a 27 year-period to understand how U.S. far-right media sites construct the idea of Sweden as ‘the rape capital’ of the world. The article makes several contributions as it pertains both to the technological and ideological functions of citing for digital far-right media. We highlight the (lack of) timeliness of sourcing, the role of mainstream sources for repurposing and amplification, and practices of exaggeration and pseudo-legitimisation. Against the backdrop of our findings, we suggest the need for approaching digital far-right media information practices as more complex and subtle than they have previously been conceived, and highlight the necessity to problematise misinformation as a label for misrepresentation in digital far-right media.

Introduction

There is an increasing prevalence of false and misleading information in contemporary digital media environments, which can be exemplified with misinformation and disinformation related to the side-effects of vaccines, the Covid-pandemic, and the Russian war on Ukraine (Benkler et al., 2018; Hameleers, 2021; Strömbäck et al., 2022a). No matter if it is produced and disseminated intentionally or unintentionally, such information may have far-reaching consequences in terms of the (re)production of misperceptions and misbeliefs (Glüer and Wikforss, 2022; Nyhan, 2020; Strömbäck et al., 2022b), which has in turn been shown to have a distorting influence on people’s attitudes as well as behaviours (Ahler and Sood, 2018; Kavanagh and Rich, 2018; Nyhan, 2020).

In this context, digital far-right ¹ or so-called ‘alternative’ media plays a key role in the dissemination of misinformation online (Åkerlund, 2023; Benkler et al., 2018; Hameleers, 2021; Heft et al., 2021; Jamieson and Cappella, 2008). More specifically, research shows that they tend to compete by providing partisan-consistent and attitude-confirming information, and prioritise political goals higher than the truthfulness and relevance of information (Benkler et al., 2018; Hameleers, 2021; Strömbäck et al., 2022a). Hence, although far-right media proponents may advocate their importance as watchdogs of mainstream media (Holt et al., 2019; Mayerhöffer and Heft, 2022), these have repeatedly been shown to represent ideologically driven, and even Islamophobically centred and conspiratorial reporting, rather than functioning as corrective sources of information (Ekman and Widholm, 2022; Haller and Holt, 2019; Nygaard, 2019; Palmgren et al., 2023).

As suggested by previous research, a key strategy that digital far-right media use to spread false and misleading information, increase its persuasiveness, and delegitimise its targets, is by constructing specific strategic narratives and repeating these across both time and space (Åkerlund, 2022b, 2023; De Fina, 2018; Hameleers, 2021; Van Leeuwen, 2007; Wikforss, 2023). Online, hyperlinked citations are a central means of getting these narratives across. Yet, there is only limited understanding of how far-right narratives are constructed through hyperlink citation practices over time and across digital media settings. To remedy this, in this study we investigate the far-right narrative that Sweden would be especially exposed to Muslim-perpetrated rape (Åkerlund, 2023; Lindgren and Lundström, 2010; Thorleifsson, 2019). Specifically, by tracing how the idea of Muslim-perpetrated rape in Sweden ends up in extremist far-right contexts in the United States, this article aims to explore (1) how far-right narratives are constructed over time and across digital media settings through citing practices, (2) how misinformation contributes to this, and (3) the key digital sources referenced to form this narrative.

Defining and understanding mis- and disinformation

Because of digitalisation, media settings across the world have transformed into high-choice media environments. Where traditional news media have lost influence both in terms of gatekeeping and as a source of information, digital and social media have become increasingly important (Benkler et al., 2018; Strömbäck et al., 2022a). This development has made it easier for governments, various political actors, as well as individual citizens to bypass traditional gatekeepers and disseminate what is variously labelled disinformation, misinformation, and fake news (Broda and Strömbäck, 2024). The common denominator for these concepts is that the information that is spread is false or misleading, but they differ with respect to whether the information is disseminated with intention to mislead (disinformation and fake news) or without (misinformation) (Benkler et al., 2018; Kavanagh and Rich, 2018; Lecheler and Egelhofer, 2022; O’Connor and Weatherall, 2019; Wardle and Derakshan, 2017). A problem though, is that true beliefs and intent is most often hard to prove. After all, even those who claim that the world is governed by reptiles disguised as humans might believe in it. Given this, we will follow in the footsteps of others and use misinformation as an umbrella term (e.g. Törnberg and Chueri, 2025).

What is important to note is that disinformation, misinformation, and fake news does not necessarily have to be outright false. The information can include some snippets of truth, but still be low in facticity (Tandoc, 2019) and highly misleading (Kavanagh and Rich, 2018). It is also important to distinguish between individual pieces of information and the overall narrative that is created through a dissemination of multiple, individually more or less truthful, accounts of reality. Although there are plenty of definitions of narratives (Shaffer et al., 2018), what is most important here is that narratives ‘are ways to connect events in a non-random way, and therefore they are powerful means to provide meaning to the political world we live in’ (Bottici, 2010, cited in De Fina, 2018: 235). An example might be an immigration-critical site only reporting when immigrants commit crimes but not when natives do. Hence, individual stories might theoretically be (at least partly) factually correct, but on an aggregated level still create an overall misleading and misinformative narrative. Research also suggest that it might be more effective to disseminate a highly selective, cherry-picked and non-random, but not necessarily outright false, set of information to promote misinformation (Cho et al., 2014).

To understand misinformative narratives then, it is important to not only investigate the actual facticity of individual pieces of information at one given point in time and source. Instead, the broader pattern of selected pieces of information cited also need to be considered. This holds not least when seeking to understand the role of digital far-right media in producing and disseminating mis- and disinformation.

Digital far-right media and citing practices

A distinguishing feature of political alternative media is that their content is influenced and shaped by specific political goals and values. Central to this process is the selection of what they cover in their reporting and how they cover it (Holt et al., 2019; Mayerhöffer, 2021). This holds in particular for digital far-right media, which tend to be more ‘alternative’ than left-wing alternative media (Åkerlund, 2022b; Benkler et al., 2018). While the decisions of what to cover and how to cover it are ideologically motivated, they are also very much dependent on, and enabled through, digital technology (Leidig, 2021).

Although the digital far-right media landscape is varied and includes a range of different actors and audiences (Holt, 2020; Staender et al., 2024), they have been shown to often use technological means to create an appearance of objectivity by borrowing formats and aesthetics from credible sources, like legacy news media, to appear more trustworthy (Benkler et al., 2018; Ekman and Widholm, 2022; Nygaard, 2019). Furthermore, in-text citations, not the least clickable ones, are a key technological feature for far-right media in attempts to increase their appearance of legitimacy. They serve the purpose of connecting different pieces of digital information. By interlinking far-right sites, content, as well as users, they have previously been shown to contribute to the formation of far-right socio-technical networks (Åkerlund, 2022a; Burris et al., 2000; Caiani and Wagemann, 2009; Heft et al., 2021; Klein, 2012).

While some research suggests that far-right media may not outright reframe the sources they cite (Mayerhöffer and Heft, 2022), others argue that they do, for instance by putting cited content out of context to make specific points (Roberts and Wahl-Jorgensen, 2022). Such recontextualisations, which might at times only slightly alter the content of referenced sources, have furthermore been shown to augment racially coded, conspiratorial dimensions which were not in the original news items (Haanshuus and Ihlebæk, 2021). Digital far-right media have also been shown to rely – to different extents – on copying or adding context to existing material published elsewhere online (Holt, 2016; Mayerhöffer, 2021; Mayerhöffer and Heft, 2022).

Overall, digital far-right media have been shown to reference content from a variety of sources. Contrary to previously held assumptions about ‘echo chambers’ (Sunstein, 2007) and ‘filter bubbles’ (Pariser, 2011), which suggest that digital media would only draw from sources similar to their own, more recent studies have found that mainstream media, including public service outlets, are highly important for far-right media sourcing (Haller and Holt, 2019; Holt, 2016; Mayerhöffer and Heft, 2022; Palau-Sampio, 2022). By citing traceable mainstream news, far-right media sometimes try to piggy-back on their legitimacy (Haanshuus and Ihlebæk, 2021). Yet, there is often a discrepancy between the claims made by far-right media and the references used to make them.

More specifically, far-right media have been shown to cherry-pick citations to news items that help make specific points, which support already held beliefs (Haller and Holt, 2019). Beyond this, as Palau-Sampio (2022) has shown, these mismatches can vary greatly in scope, ranging from cited texts not supporting the points made by the citing far-right article, to cited texts being in inaccessible languages. Thus, instead of providing verified and impartial information – as is expected of legitimate news media organisations – they seek to compete by providing attitude-confirming content, while the veracity is typically treated as less important than what is politically expedient (Benkler et al., 2018; Strömbäck, 2023; Waisbord, 2022).

Despite that the appearances and marketing strategies of many far-right media outlets have grown increasingly professional over time (Ekman and Widholm, 2022; Figenschou and Ihlebæk, 2019), which may have helped increase their perceived legitimacy for audiences, there remains little knowledge of how far-right misinformation takes shape over time, through citing practices specifically.

Case and context

To provide context, Sweden has long enjoyed high international regard for its relative equality, high quality of life, and social welfare (Ericson, 2018; Schierup and Ålund, 2011; Strömbäck, 2022; Swedish Institute, 2022). Sweden has a progressive sexual assault and gendered harms legislation (Carroll, 2021), which also includes what counts as rape. For example, lack of consent is sufficient for a sexual act to be considered a rape, while in most other countries, some kind of violence is required. In Sweden, it is also sufficient with some kind of penetration, while in most other countries, coition is required. Furthermore, in cases of repeated rapes by one man against a woman, each rape is counted, while in many other countries, they count only once. Altogether then, Sweden has the most restrictive legislation in Europe and stricter than in the United States and virtually all other countries (BRÅ, 2020). International rankings also show that Sweden is one of the world’s leading countries with respect to gender equality and the safety of women (Directorate-General for Research and Innovation, 2025; Strömbäck, 2022).

Despite this, there has in recent years been intense and polarised debates in Sweden about the connections between Muslim immigration and crime, not the least regarding sexual assault (Skilbrei, 2020). The sexual threat that Muslim men as a group supposedly pose to (White) women has deep historical roots (Said, 1978; Venäläinen and Virkki, 2019). But in recent years, not the least following the so-called ‘refugee crisis’ in 2015, this idea has been invigorated and the far right has sought to cast immigration as a threat to women’s rights a (Farris, 2017; Horsti, 2017; Sager and Mulinari, 2018). In the construction of Muslim immigrant men as rapists, Western ideals of progressiveness and equality are positioned in irreconcilable contrast to the alleged dogmatism and backwardness of Islam (Hafez, 2014; Keskinen, 2013).

Alongside national debates, Sweden has increasingly also become a main target for international far-right mis- and disinformation campaigns. More specifically, the international far right has for years sought to position Sweden as a nation that is in spiralling decline – plagued by shootings, gang violence, parallel societies, and riots – and where sexual assault, as a consequence, is rampant (Cassinger et al., 2018; Colliver et al., 2018; Ericson, 2018). So much so, in fact, that Sweden is increasingly being constructed by the far right as the rape capital of the world (Åkerlund, 2023; Thorleifsson, 2019; Titley, 2019).

Moreover, research has shown that U.S. far-right media constitute the centre of ‘an emerging alternative news ecology on the right [that] bring audiences into the broader community of an interconnected right’ (Heft et al., 2021: 495). With the United States in the centre, this begs the question of how far-right narratives like that about immigrant-perpetrated rape in Sweden may be spread through in-text hyperlinks. Against this backdrop, and given that previous research has shown how the U.S. far-right has been especially active in promoting Sweden’s supposed demise (Åkerlund, 2023; Colliver et al., 2018; Robinson and Enli, 2022), we take the U.S. far-right as our point of departure for this analysis.

Data and methodological approach

Citations and data collection

The study’s focus on citation practices is based on the notion of intertextuality. According to this approach, ‘texts are constructed through other texts being articulated in particular ways, ways which depend upon and change with social circumstances’ (Fairclough, 1992: 9). Thus, meaning is not static but changeable over time as texts draw upon others, and texts are not self-contained entities produced in a social vacuum. Instead, texts will in different ways relate to other texts: responding to, cementing, contradicting, repurposing, and reinterpreting them (Fairclough, 2010; Kristeva, 1980). In its most apparent form, intertextuality refers to quotations (Fairclough, 1992), and analysing citing practices can thus provide insight into how depictions of Sweden and rape, over time, might get warped as they pass through digital media.

We trace the uses of in-text hyperlinks over time to see how the most extremist far-right sites end up constructing their idea(s) of Swedish immigrant-perpetrated rape. To do this, we work backwords, starting with the most extremist sources. We do this against the backdrop that citations tend to be unidirectional, where more extremist far-right sources reference those that are right-wing or mainstream, but seldom the other way around (e.g. Kaiser et al., 2020; Mayerhöffer and Heft, 2022) By starting (and therein ending up) with extremist far-right sources, we can trace the links and narrative to its most overt, unmoderated, and unfiltered forms of the U.S. far-right construction of the Sweden rape narrative.

To get an initial set of seed texts from which to begin sampling hyperlinks we accessed all available U.S. extremist far-right websites as identified by the Southern Poverty Law Centre’s Designated Hate Group list. ² Data were sampled through manual keyword searches for texts containing words related to Swed*, rape*, sex*, and molest* on all these sites, from when they were first launched until data collection in August 2023. Given the oftentimes coded and covert language use of the far right (Berlet, 2014; Haney López, 2014), the texts were then closely read to determine whether or not they were in fact discussing rape in Sweden, and if they did so while also focusing on immigrants or immigration. ³ This sampling resulted in an initial dataset of 144 texts dating back as far as September 2004, primarily from VDARE (n = 78) and National Alliance (n = 41), but also from American Renaissance (n = 12), Centre for Security Policy (n = 5), Council of Conservative Citizens (n = 3), Federation for American Immigration Reform (n = 2), Centre for Immigration Studies (n = 1), Media2Rise (n = 1), and Social Contract Press (n = 1).

All hyperlinks found in the 144 seed texts were documented, accessed, and assessed regarding their fit with the above-mentioned sampling criteria. Where needed, we accessed archived versions of webpages. If they fit the sampling criteria, any URL references in the texts were used to snowball sample further texts. These new texts, in turn, were then matched with the above-mentioned criteria, and if relevant, URLs were extracted from these texts too. This process was repeated with new texts and respective URLs until no more new texts could be identified, a total of nine rounds. By approaching the data collection in this way, the sampled texts capture the actual U.S. far-right argumentation regarding immigrant-perpetrated rape in Sweden. Simultaneously, those texts which were not used to make this case are automatically left out of the sample. Consequently, the most important texts and sources are not predetermined by the researchers, but instead emerge because of actual far-right citations.

For the sake of mapping the Swedish rape narrative, where texts were direct copies of other texts, only the hyperlink to the original source was noted. Furthermore, we only sampled sources referred to in a given text, not any URLs of recommended or related content on the same webpage. In texts on multiple topics, only those references used when discussing rape were sampled. In total, after cleaning the dataset from entries with incomplete information (e.g. inaccessible or dead links), we had collected 776 in-text citations, from 653 texts, published between 1996 and 2023 on 166 different websites.

Method of analysis

To investigate how the idea of Sweden’s supposed immigrant rape problem has come to take shape over time, and to map the key digital sources used in constructing this narrative, we primarily rely on Social Network Analysis (SNA) and descriptive statistics. SNA is a framework for investigating the patterns of interaction and relationships between different parts of a network, both from the perspective of the network as a whole and from that of individual actors or entities within it (Freeman, 2004). SNA posits that entities and their actions within networks are interdependent with other entities and their actions, that the connections between these represent paths for resources to flow, and that individual entities’ positions within a network impacts their opportunities for action (Wasserman and Faust, 1994).

To prepare the data for SNA, we collected meta information and categorised the dataset on different levels of abstraction. To do this, on a text-level, we identified the directionality (Wasserman and Faust, 1994) of the links, that is documented all citations in terms of the flow of hyperlinks from citing to cited text, and made note of the languages and publication dates of all these. On a site level, we identified the country of origin for each of the sources. We also coded the sources broadly in terms of their content, as either far-right media (including also far-right blogs), mainstream news media, social media, governmental sources, conservative and/or right-wing media, and academic sources (see Table 1.). Appendix 1 shows an excerpt of the top 20 sites, which constitute 60% of the sources in the dataset. The full list can be made available upon request.

Table 1.

Distribution of the texts across the sources.

	Source distribution		Text distribution
	count	%	count	%
far-right media	73	44	387	59,2
mainstream news media	45	27,1	162	24,8
conservative or right-wing media	15	9	21	3,2
governmental sources	9	5,4	29	4,4
academic sources	7	4,2	10	1,5
social media	4	2,4	28	4,2

The remaining thirteen sites and sixteen texts could not confidently be categorised to either category or lacked information to assess.

From the categorisations of citing practices, dates, languages, countries, and content types we mapped how this intertextual network formed over time. Practically, we used Python code to process and clean the data, and to calculate and visualise descriptive statistics, but primarily we relied on SNA through Gephi (Bastian et al., 2009) to calculate and visualise relationships between, and centrality measures for, individual texts, sites, and source types.

To provide context, we complement the network analysis with qualitative exploration on the text level to comprehend how texts draw upon others as means to (re)interpret and (re)contextualise them. We paid specific attention to the most highly cited texts throughout the dataset and identified how these were referenced by other sources.

Results and analysis

Longitude and political expedience

Turning to the results, the first far-right texts where immigrants are discussed in the context of rape in Sweden date back to 2004. Although the actual term ‘rape capital’ is not used in the dataset until 2013 (Frontpage Magazine, 2013; CBN News, 2013) ⁴ early texts nevertheless clearly focus on ‘Muslim rape’ as a societal threat with titles like:

Muslims rule major Swedish city (Jihad Watch, 2004)

Muslim Rape Epidemic in Sweden and Norway–Authorities Look the Other Way (Fjordman, 2005a)

Rape: Nothing to do with Islam? (Fjordman, 2005b)

Muslim rape wave in Sweden (Majority Rights, 2005).

These titles not only share a high resemblance with contemporary far-right framings of immigrant-perpetrated rape (Leidig, 2021; Saresma, 2019; Thorleifsson, 2019; Titley, 2019), but the presence of these articles a decade before the so-called ‘refugee crisis’ showcases the longitude of this far-right trope; a trope which is thus perhaps better situated as an Islamophobic far-right response to 9/11 than directly to current political circumstances, as the far right often claims. With this said, there is frequent monthly publication activity from 2010 and onwards, and especially from the years immediately following 2015, indicating the increasing focus on Sweden following the refugee crisis and the rise of the Trump-led Make America Great Again movement in the United States (see Figure 1) (see also Åkerlund, 2023; Cassinger et al., 2018; Titley, 2019).

Figure 1.

Monthly counts of citing, cited, and overall published texts.

Figure 1 also shows that while the timing of citing and cited sources seems to correspond, there are some cited texts used to make far-right claims which date back several years earlier than the first far-right reports on immigrant rape. The earliest cited source is the ‘Immigrants’ and immigrant children’s criminality’ ⁵ (Ahlberg, 1996) report by The Swedish National Council for Crime Prevention (BRÅ) from 1996. This report, along with one published by BRÅ in 2005 called ‘Criminality among persons born in Sweden and abroad’ (Martens and Holmberg, 2005), will over time become among the most referenced texts in constructing the far-right immigrant rape narrative. The appropriation of the 1996 BRÅ report is however not instant; it is not until 2012 that far-right sources begin using it to make their immigrant rape case. In fact, cited sources are on average 453 days old when they are used, and one reference dates back as far as 19 years. This clearly contradicts far-right media’s self-perception as watchdogs of mainstream media (Holt et al., 2019; Mayerhöffer and Heft, 2022). Instead, it shows how far-right media reiterate and repurpose information when it fits their pre-constructed narrative and worldview, and when it is politically expedient for them (Giusti and Piras, 2020; see also (Strömbäck, 2023; Waisbord, 2022). Technologically, this is a practice which is entirely enabled through the continuous access to these materials over time (Åkerlund, 2022a, 2022b; Merrill and Åkerlund, 2018).

One particularly well-referenced and reiterated text is entitled ‘Immigrant Rape Wave in Sweden’ and was published in 2005 by the Norwegian far-right blogger Fjordman (see Figure 2). It was cited by other far-right blogs around the time when it was first posted, but it was also used to make a case about the threat of Muslim rape in Sweden in 2011, 2012, and 2017 (e.g. Majority Rights, 2005; VDARE, 2017a, 2017b). Primarily, its related imagery portraying a bloody, blond, White, Swedish woman has been identified as central in legitimising an imagined White global community around the idea of ‘Muslim rape’ (Horsti, 2017). In the dataset, there are also explicit attributions to said blogger for informing about the ‘immigrant rape wave’ in Sweden.

Figure 2.

Blurred screenshot from Fjordman’s blog post ‘Immigrant Rape Wave in Sweden’.

The image of the woman was taken at the hospital after she had been beaten and raped by immigrants on New Year’s Eve 2004 and was originally used by a mainstream Swedish news media outlet that interviewed her about the assault. In the interview, she talked about her experience and sought to make a broader point about men’s sexual violence towards women. Because the perpetrators were non-White; however, she soon became an involuntary symbol of immigrant-perpetrated rape in Sweden. So much so, that the blogpost was later also referenced in far-right terrorist Anders Behring Breivik’s manifesto (Horsti, 2017).

Although it might not be intended by the original source, this showcases how far-right actors might appropriate mainstream news and reiterate them to create powerful stories about the threat of Muslim rape. This kind of pattern is repeated throughout the dataset, through the broad-scale repurposing and amplification of mainstream news.

Mainstream amplification

Unsurprisingly, given the sampling method, far-right sites constitute the bulk of data. Overall, 59% of texts come from far-right media and blogs. Even more prominently, as shown on the left-hand side of Figure 3, the texts which most commonly cite others are almost exclusively far-rightist. In fact, almost nine out of 10 of all citations made come from far-right media. As shown on the right-hand side of Figure 3, the source types of the texts cited are more varied. These results show how the texts that dominate in constructing the rape narrative are not the same as those being used to make these arguments.

Figure 3.

Network visualisation with node sizes by most citing (left), and by most cited (right).

After far-right sites, mainstream media are the second most common source type throughout the dataset, constituting a quarter of all texts. This is primarily because mainstream media are cited by far-right sites, as seen through the higher prominence of mainstream media texts on the right-hand side, compared to on the left-hand side of Figure 3.

More than this, mainstream media sources are seldom present in the dataset because they discuss rape and Sweden and immigrants together. In fact, mainstream news media rarely mention all these dimensions. Important to note here is that the ethical guideline for Swedish journalists’ states ‘Do not highlight individuals’ ethnic origin, sex, nationality, profession, political affiliation, religious views or sexual orientation if it lacks importance in the context and is disrespectful’. ⁶ Interestingly then, given the antagonistic stance towards mainstream news, far-right media remain highly reliant on their reporting to build their arguments and produce their own content, linking to them although there is seldom legitimate reason to do so.

Only 39 of the mainstream media texts – a mere 24% of news media articles – discussed rape, Sweden, and immigrants simultaneously. Most often, these articles cite other mainstream news articles, and at times governmental texts. Sometimes, these mainstream news texts cite others as a means to question or mitigate claims of immigrant rape. One example is the BBC (2017) article entitled ‘Reality Check: Is Malmo the “rape capital” of Europe?’, which for instance notes that ‘Sweden does not publish the ethnicity or national background of perpetrators of any crime, including sexual offences’, or the Swedish Public Service (2016) article correcting wrongful accusations, stating that ‘[t]he police retract their initial information that sexual assault suspects during [festival] are unaccompanied children’. There are however some mainstream media articles that do in fact centre on immigrants in negative terms, as illustrated for instance by the following titles, the last of which, from Swedish Public Service (2021), is among the most cited texts in the dataset overall:

Swedish music festivals hit by reports of rapes by ‘migrants’ (The Telegraph, 2016)

Swedish police accused of covering up sex attacks by refugees at music festival (The Guardian, 2016)

Alleged gang rape shown on Facebook shocks Sweden (USA Today, 2017)

New study: Nearly half of all convicted rapists are born abroad (SVT, 2021)

The foci of many of these news items coincide with the most prolific events and periods of interest throughout the sample period. Thus, in the rare instances when mainstream news media do mention ethnicity in their coverage on rape in Sweden, this becomes powerful fuel for far-right reporting, which in turn showcases how mainstream news media content – no matter their intent – can be amplified through citations for far-right purposes (Åkerlund, 2022a).

The results furthermore show that far-right sources often cite mainstream news articles directly, even when these are in other languages. Across the dataset far-right media sources are primarily composed by U.S. far-right media texts in English (66%), while Swedish news sources constitute the bulk of mainstream media texts (74%). Interestingly, in almost half the instances where Swedish mainstream media is referenced, it is done directly by English language U.S. far-right media texts. This not only highlights, as previously shown, that there are mismatches between the language of far-right media citing and consequent cited texts (Palau-Sampio, 2022), but also the breadth in information scoping by far-right sites, and the absence of mediation through ‘gateway’ sources (cf. Åkerlund, 2022a). Whether this is due to technological savvy in finding and translating relevant texts, or whether this is due to access to Swedish speakers to do this work, remains unclear. Nevertheless, this finding points to the interconnectedness of the international far right through citation practices, and the consequent need to explore far-right communication practices beyond individual national contexts.

Exaggeration and pseudo-legitimacy

When focusing on the peaks in publication numbers, our findings show that these are primarily centred around a few events that drew extensive media attention, as mentioned above. This coverage includes sexual assaults perpetrated by unaccompanied boys at music festivals around Sweden and alleged surrounding cover-ups by police of these events in 2016; then the Facebook livestreaming of a gangrape perpetrated by immigrants; the now infamous ‘Last Night in Sweden’ remarks by former president Trump; and finally, an anti-immigrant post made by a police officer on his Facebook page (e.g. Ekman, 2019; Merrill and Åkerlund, 2018), all in 2017. The last of which is also among the most cited individual texts in the dataset.

It is not difficult to imagine why far-right actors might find the policeman’s post appealing given its insider perspective on Swedish criminal activity and its alleged association to (a particular group of) immigrants (see Appendix 3). The most manifest pattern in proceeding citations of the Facebook post is the exaggeration of the policeman’s authority and status, as well as the information that the post supposedly conveyed. For instance, as seen in the following international far-right uses: a Breitbart (2017) article wrongfully claimed the officer had stated that ‘almost all of the serious criminal activity in Sweden is committed by foreigners and migrants’, and likewise did an Express (2017) text note that ‘Migrants ARE to blame for most serious crimes in Sweden’. Finally, when Summit News (2021) references RT’s (2017) article about the Facebook post 4 years after the fact, it was warped into being a question the officer answered after he ‘was asked about the demographics of those responsible for violent crimes’. While the officer speaks from his own personal experience, the intensification of the statements and officer’s authority slowly warps the meaning of the original text. Through slightly altered interpretations in following citations, the police officer is transformed into a (pseudo-)expert who speaks for the whole of Swedish law enforcement, and about all criminal activity, in the entire county. Yet, it is potentially a practice which the far right might easily get away with since fact-checking would not only require visiting the original source, but also to translate it and identify possibly subtle linguistic inconsistencies.

The pursuit for trustworthiness and legitimacy through citing is also seen in the use of statistics. Beyond citations of selected parts of Swedish governmental reports and academic literature, pseudo-statistics originating from far-right sources are also found among the most referenced texts (Cockbain and Tufail, 2020). Prominently, among the most cited texts is a report seeking to ‘provide an overall picture of the origins of convicted sex offenders in Sweden [. . .] and confirm if what many spontaneously believe about foreigners and sexual crime is really true’ (Jonasson, 2017). The 80-page long text, self-labelled as a ‘citizen study’, is written in Swedish by a private individual who identifies as supporting the authoritarian right-wing populist party, the Sweden Democrats. Through close readings of links to this source, it becomes clear that the text gets increasingly warped the further away citations get from the original source. This holds true in particular regarding the share of non-European immigrants committing aggravated rape.

In the report, the author mentions two numbers, wherein the lower share is identified as the final, and official statistic:

five out of six aggravated rapes were committed by perpetrators of non-European origin (84%). That is a lower figure than in the preliminary study, which then showed that 92% of those convicted of that crime classification were non-European (Jonasson, 2017).

Among the citations of the report is a summary written in Swedish by the author himself for the Swedish far-right media outlet Avpixlat (2017). This overview, however, does not go into sufficient detail to mention neither the 84% nor 92 % figure. Yet, when the Avpixlat article in turn is cited by far-right site 10News.one, not only does it mention the 92% pilot study finding, but it entirely leaves out the 84% figure. The ‘citizen study’ has by this stage also warped into seemingly legitimate ‘data from Sweden’:

New data from Sweden shows how [. . .] 92 percent of all severe rapes (violent rapes) are committed by migrants and refugees. (10News.one, 2017)

As U.S. far-right site VDARE cites 10News.one the following year, ‘Swedish data’ turns into the slightly more official label ‘Swedish statistics’ and the ‘92 percent of all severe rapes’ becomes distorted once more to encompass all rape:

According to Swedish statistics, 92 percent of rapes are perpetrated by migrants. (VDARE, 2018)

This brief example among many, illustrates the subtle practices of far-right misinformation regarding the employment of pseudo-statistics, which at times can work similarly to a ‘game of telephone’– ever so slightly changing meaning and emphasis over the course of several iterations of interpretation. Another (which there is no space to discuss in depth here) is how immigrants and refugees in more or less subtle ways are transformed into Muslims. Again, these are enabled through the continuity and ease-of-access to far-right source material online, and on the reliance that audience will not fact-check the seemingly minor adjustments made to the meaning of the text. Yet, as the original information and ideas get warped, we find that it is persistently in a direction that enhances the supposed threat and seriousness of Muslim-perpetrated rape.

Discussion

In this study we have investigated the far-right narrative that Sweden is especially exposed to Muslim-perpetrated rape (Åkerlund, 2023; Lindgren and Lundström, 2010; Thorleifsson, 2019). Specifically, by tracing how the idea of immigration-perpetrated rape in Sweden ends up in extremist far-right contexts in the United States, this article explored (1) how far-right narratives are constructed over time and across digital media settings through citing practices, (2) how false and misleading information contributes to this, and (3) the key digital sources referenced when forming this narrative. Using descriptive statistics and SNA, we analysed 776 in-text citations from 653 texts on 166 different websites, spanning 27 years.

The article makes several empirical contributions. While much focus has been on Sweden’s portrayal after the so-called refugee crisis (Åkerlund, 2023; Ericson, 2018; Titley, 2019), our study shows the longevity of the ‘Muslim rape’ narrative, thus contributing new knowledge into how deep-rooted the international far right’s interest in Sweden is. We furthermore identify the outdatedness of references used to construct the idea of Muslim rape in Sweden, and therein strengthen the findings from extant literature that political goals often outweigh truth and timeliness for the far right (Strömbäck, 2023; Törnberg and Chueri, 2025; Waisbord, 2022). Consequently, digital media – alternative and mainstream alike – provide technological means for the far right to reiterate powerful stories through citations, no matter their novelty. In turn, they afford the far right with an ever-growing arsenal of ideologically driven, easily-accessible material which can be drawn upon when deemed appropriate (Åkerlund, 2023; see also Horsti, 2017), also after quite extensive periods of time.

Unsurprisingly, we identified a unidirectional relationship between far right and mainstream media, with a large-scale discrepancy in the texts promoting the Muslim rape narrative – which are almost exclusively far-rightist, and those drawn upon to legitimise it – which are often mainstream news articles. We also show that when mainstream media is cited, there is seldom information to fully justify these sourcing decisions (Mayerhöffer and Heft, 2022; Palau-Sampio, 2022).

In the few instances where mainstream news media do report on ethnicity in the context of rape in Sweden, the analysis showed how far right were quick to make use of this legitimacy to further pre-held perceptions. Interestingly, and contrary to previous research (Åkerlund, 2022a), we identified that digital far-right media often cite mainstream news without mediation through ‘gateway’ sources, even in cases when there are language barriers. This, in turn, points to the far right’s skill and reach – albeit social or technological – in identifying suitable stories. Overall, these findings showcase a contradiction between the well-established outward attempts to undermine and vilify mainstream news media (Figenschou and Ihlebaek, 2021), and far-right media’s continuous reliance on its content in their daily practice. These findings, furthermore, highlight the importance in approaching far-right media as international phenomenon, no matter their national context or intended audience (Heft et al., 2021).

Finally, beyond identifying that content is selectively cited in stories by the far right (Haller and Holt, 2019; Mayerhöffer and Heft, 2022), we find evidence of augmentations of racially coded, conspiratorial dimensions which were not in the original news item (Haanshuus and Ihlebæk, 2021). Specifically, our analysis shows exceedingly subtle practices of exaggerating the legitimacy of pseudo-experts and the severity of information about Muslim rape and rapists. More than this, beyond the direct recontextualisation of information, facts and news (Haanshuus and Ihlebæk, 2021; Haller and Holt, 2019; Nygaard, 2019; Roberts and Wahl-Jorgensen, 2022), we showcase how meanings ever so slightly morph and change meaning as they travel further from the original source, also several years later and over the course of several iterations of citation. We thus show not only that digital far-right media rely heavily on misinformation in their narratives but also how these practices can at times be remarkably subtle and reliant on an absense of detailed scrutiny by their audiences.

Taken together, these findings provide new theoretical insight into the strategies of digital far-right media generally, and about the obscured ways through which far-right propaganda can work, specifically. We argue, following Benkler et al. (2018), that far right media cannot be conceived of as a mere right-ward extension of the mainstream media ecosystem, at least not in terms of the construction and spread of digital false and misleading information. This brings us back to the distinction between misinformation and disinformation. While we cannot with certainty prove intent, we identify systematic, continuous, and at times veiled practices of inaccurately and skewedly portraying information through citations that are better likened to ‘disinformation’. This suggests that the umbrella term ‘misinformation’ is not only too vague a concept to describe the workings of the far right, in particular over a longer term, but also that its underlying invertedness exempts the far right from agency. Yet, through our exploration of the case of Sweden as the ‘rape capital’ of the world, we identify that far-right practices might often be less manifest and tangible than the disinformation label presupposes. We therefore argue that the construction of far-right narratives through citations online might be more accurately described as cumulative disinformation, a seemingly systematic and intentional practice, enabled through digital technology; and which can be exceedingly subtle, multifaceted, and temporally complex.

Accordingly, these far-right practices cannot solely be measured through numbers of citations or be qualitatively interpreted through exploration of a small set of texts. Instead, they need to be approached through flexible methodological toolkits that are adapted to the intentional covert and deceptive ways of the digital far right (see also Åkerlund, 2022b).

In the case of this study, we found these mechanisms and strategies at work with respect to the misinformative narrative that Sweden has become the ‘rape capital’ of the world due to Muslim rape and rapists, but the theoretical implications are more general. Thus, a key task for future research would be to investigate how other cases of strategic and disinformative narratives are formed through citation practices that span both space, time, and theme. Finally, while it is not digital technology itself that drives these ideological narratives, it should be noted that the technological opportunities to reach a variety of attitude-conforming information, translate it, archive and at a later time reiterate it to make certain points, are all important dimensions for enabling these far-right efforts. Future research would do well to better understand how other aspects of digital technology, beyond citations, interacts with far-right (media) norms.