Understanding “Good” and “Bad” Twitter Practices in Alternative Media: An Analysis of Online Political Media in the UK (2015–2018)

ABSTRACT Alternative Online Political Media (AOPM) have become increasingly important within international news landscapes, but their social media practices have received limited academic attention. Our large-scale content analysis (N = 14807) offers the first comprehensive study of how APOM in the UK use Twitter. Drawing on a pertinent model of social media use that enhances notions of “good” and “bad” journalism, and through our own sentiment analysis, we find Twitter norms closely aligned with those of legacy media, including a relatively limited online interaction with audiences. We conclude that while AOPM follow many social media logics consistent with mainstream news sites and add to the wider realm of political analysis, their highly partisan content means that their Twitter use cannot be considered balanced, neutral or objective.


Introduction
Lying on a continuum between "well-funded, digital-native start-ups" and the "lone individual able to blog or tweet or comment" (Carlson and Lewis 2020, 123) are a range of online news outlets noticeably different to mainstream media (MSM) or "legacy news".The labels used to define them variously highlight their mainly political (and often opinionated) content, their online shareability, the often "clickbait" headlines and political partisanship (see McDowell-Naylor, Thomas, and Cushion 2021a, 171 for summary).Given the wide diversity within these outlets (McDowell-Naylor, Cushion, and Thomas 2021b) and their range of approaches, organisational structures and content focus, applying homogenising labels is problematic, and while recognising this fundamental imperfection, for expediency, we nonetheless describe a specific group of these new organisations as "alternative online political media" (hereafter "AOPM").This label reflects, we feel, the alternative or "corrective" (Holt, Figenschou, and Frischlich 2019, 3) they claim to offer to MSM and the independent governance/ ownership models generally distinguishing them from legacy media brands often under the control of large corporations.While these traits, we argue, are fairly consistent throughout, there the similarities generally end.Indeed, we argue later that when using Twitter to disseminate their work, even the description of "alternative" might actually be challenged.
Many such outlets rely on Twitter to share content and reach their audience.Accordingly, this study breaks new ground in its analysis of how AOPM use Twitter, contrasting this with how legacy news brands use Twitter, and asking whether AOPM's Twitter strategies exhibit traits consistent with "good" or "bad" journalism.We are contextualising "good" and "bad" as indicative of professional practices largely adhering to traditional understandings of the normative aims of journalism ("good") or apparently contravening such aims ("bad").This in turn, enables us to make a contemporary intervention into debates surrounding editorial practices, how AOPM serve audiences and the wider democratic value of media.
Drawing on 14,807 manually coded Tweets and an analytical model developed by Orellana-Rodriguez and Keane ( 2018) that provides a taxonomy of identifiable themes within news organisations' use of Twitter, we find merit in the ways AOPM use Twitter, but also that there are conventions that might undermine public understandings of politics.Above all, our sentiment analysis reveals how AOPM use Twitter to amplify partisanship, and their reluctance to engage directly with social media users.
If alternative media sites operating on the fringes of journalism both confirm and challenge what we recognise as conventional news reporting (Carlson 2016), our study can provide some insight into how "alternative" media are distinctive from the mainstream.Indeed, within Eldridge's (2018, 857-858) notion of "interloper" journalism, AOPM nonetheless perform similar "socio-informative functions, identities, and roles".They are frequently critical of the MSM (Cushion, McDowell-Naylor, and Thomas 2021;Figenschou and Ihlebaek 2019;Holt, Figenschou, and Frischlich 2019) and establishment politics (Holt 2018;Rae 2020).Most pertinent here, they are reliant on social media to disseminate content (Kalsnes and Larsson 2019;Haller and Holt 2019;Manthorpe 2018).Research has considered the formation of "digital news infrastructures" (Heft et al. 2019) which have been presented as highly successful (Conte 2016) despite contrasting evidence that they have limited audience reach (Klawier, Prochazka, and Schweiger 2021;Fletcher, Newman, and Schulz 2020, 8).
We draw on an analytical taxonomy of themes to identify "good" and "bad" journalistic use of Twitter developed by Orellana-Rodriguez and Keane (2018, 12).This model also enables us to determine how AOPM's use of Twitter might differ from legacy media, the degree to which the audience is engaged with, and indeed, whether it is the "good" or the "bad" that prevails.In sum, we examine the extent to which AOPM's Twitter use is consistent with the wider normative aims of journalism.

Twitter, Journalism, and Alternative Online Political Media
Journalism is accessible via different routes (Kavanagh et al. 2019).Twitter especially is widely used by journalists (Canter 2015;Willnat and Weaver 2018) and is normalised within mainstream journalism (Molyneux and Mourão 2019).Its use by political journalists in particular is central to a hybrid media system (Chadwick 2017;Oschatz, Stier, and Maier 2021).
Twitter enables news organisations and journalists to directly engage with audiences, by "following many users, providing links, soliciting information and disclosing personal information" (Hanusch and Bruns 2017, 39).Journalists also use Twitter to develop their individual branding branding (Molyneux 2019) and for adding personal elements (Hanusch and Bruns 2017).More specifically, Twitter presents public opinion (McGregor 2019;Ross and Dumitrescu 2019), helps with newsgathering (Enli andSimonsen 2018: McGregor andMolyneux 2020) and provides embedded content (Oschatz, Stier, and Maier 2021).It is well suited to journalism's imperative to break news quickly (Canter 2015;Chadwick 2017), the holding of politicians to account (Chadwick 2017, 186), and provides testimony from "actual eyewitnesses on the ground" and "second-hand live discussion of unfolding events" (Bruns and Burgess 2012, 801).
But journalism practice has been transformed by social media (Chadwick 2017).The journalistic use of Twitter has been scrutininsed (Hermida and Mellado 2020;Parmelee 2013;Barnard 2016;Bentivegna and Marchetti 2018), and generally, journalists blur traditional and online norms when using it (Chadwick 2017).Accordingly, research tends to focus on how traditional norms and practices are challenged by Twitter (Bentivegna and Marchetti 2018;Mellado and Hermida 2021) and indeed, our own study follows this same general trajectory.But it also breaks new ground in that AOPM Twitter strategies can now be compared with those used by MSM.
What represents journalism's "best practice" on Twitter remains contested (Molyneux and Mourão 2019, 250).The notion of journalistic objectivity is both debated and contextual, and we recognise the ongoing debate as to whether journalists achieve "objectivity" by being detached from the story or by being involved within it (see Skovsgaard et al. 2013).Our own conceptualising of objectivity is based on the notion that "any dialogical relationship will damage the journalist's outsider and unbiased position" (Soffer 2009, 474) even though such interactions are facilitated by a busy platform like Twitter (Singer 2008).We define objectivity, therefore, more along the lines of journalists and news organisations being distanced observers with a neutral overview (Merritt 1995) rather than the alternative "Journalist-centred role" where "news professionals have a voice in the story" (Hartley and Askanius 2021, 864).
Amid a wide range of potential measures of "objectivity" for this study, we evaluate the degree to which AOPM insert themselves into the story on Twitter with comments and responses to the audience.The blurring of traditional objectivity can lead to more "opinionated" journalism (Lawrence et al. 2014) and journalists typically exhibit wariness about how to behave (Hanusch and Bruns 2017, 40;Parmelee 2013, 303).
The use of Retweets also challenges norms of objectivity and independence, as news content and opinion are easily shared (Molyneux 2015).Accordingly, ideological bias and partisanship are often clearly apparent (Mills, Mullan, and Fooks 2020), as is Twitter's one-directional "broadcast" approach of not engaging with users (Hanusch and Bruns 2017, 40).This broadcast model has been described as a "one-to-many diffusion" capable of generating "large information cascades" (Liang et al. 2019, 7-8), where Tweets do not obviously invite interaction beyond their simple consumption.Our study considersthrough analysing the types of Tweets preferred by AOPMwhether such a model prevails, or if a more unstructured pattern of interaction is preferred.

Towards a Normative Understanding of AOPM Twitter use
As "digital-born organisations" (Nicholls et al. 2018), AOPM are well-acquainted with social media.They might even be viewed as having evolved from bloggers or "semi-professional or semi-amateur journalist-activist-experts", who "occasionally intervene to break important political news" (Chadwick 2017, 214) perhaps even achieving "viral distribution" (Lasorsa, Lewis, and Holton 2012, 20).Accordingly, social media platforms enable AOPM to reach significant audiences (Conte 2016;Manthorpe 2018).
Twitter use provides insight about best practice and appropriate journalistic norms since "any new technology has the potential to change journalism" (Parmelee 2013, 291).Orellana-Rodriguez and Keane (2018, 76) argue that social media has brought both positive and negative changes, the more negative aspects probing whether increased interaction with audiences adversely impacts perceptions of objectivity.Their analytical framework not only enables the quantification of Twitter habits and whether these are different to those practiced by legacy media, but also provides a mechanism to determine positive or negative operational social media traits and whether such behaviours reflect notions of "good journalism".
"Good" journalism has long been debated and a notional "normative theory" for journalism includes elements such as truthfulness, "watchdog" journalism embracing critical and objective reporting (McQuail 1987) and holding political and commercial power to account.Evaluating journalism "quality" is important, particularly "in an era of 'fake news' and ubiquitous information overload" (Achtenhagen, Melesko, and Ots 2018, 129).But "quality" within journalism is increasingly nebulous.Fundamental principles such as "impartiality"once clear indicators of "quality journalism" are increasingly complex (Thomas 2021, 160).Moreover, news organisations face growing financial pressure leading to the cutting of resources.The concentration of media ownership, for example, results is a loss of variety (Curran 2011), and as the lines between journalism, PR and politics become increasingly blurred within the so-called "symbiotic relationship" (Louw 2010), the profession is under significant pressure.Indeed, as commercial interests eclipse democratic ideals there is a tangible sense that "journalism's best days are over" (Iggers 1998, 3).Amid the notion that legacy media have "seized the opportunity to assert their own role as guardians of quality journalism" (Kalsnes, Falasca, and Kammer 2021, 300), searching for quality within new forms of news reporting is particularly relevant in the age of "disintermediated, distributed, on-demand journalism that is blended into our individual, AI-driven news streams or feeds and recommendation engines" (Beckett 2014).
Some of journalism's central pillarsobjectivity, providing multiple viewpoints and audience engagement -are incorporated within our analytical model, represent our evaluative measures, and enable wider empirically-driven conclusions about the habits, operations and quality of AOPM.By using this particular framework, we can draw some conclusions about "good" or "bad" journalism, where "good" is aligned with a normative model embracing traditional objectives associated with "quality" operational behaviours on behalf of journalists and the organisations that employ them.The model provided by Orellana-Rodriguez and Keane (2018, 77-79) and our adaptation of it is outlined in Figure 1.
Informed by this model, our specific research questions are: The left-wing outlets within our study were The Canary, The Skwawkbox, Evolve Politics, Another Angry Voice and Novara Media.The right-wing outlets were Guido Fawkes, Breitbart London, Westmonster and The Conservative Woman.This sample was selected as being representative of the right-and left-leaning outlets with the greatest reach in terms of their audience on social media.As empirical research into AOPM is still relatively new, organising or grouping sites is provisional.But while these sites are seemingly united in their wish to offer something different to MSM, they have differing political affiliations.Figure 2 shows how these sites self-identify, notionally reflecting either a "left" or "right" approach to reporting politics.
Our manual content analysis of 14,807 Tweets concentrated on the main Twitter accounts run by these nine AOPM outlets between 2015 and 2018.Our sample was drawn from four periods: 6-25 October 2015; 9-29 October 2016; 30 April-7 June 2017 (the UK general election); and 8-28 October 2018.We stress that while our data offers a retrospective analysis, it nonetheless focuses on a pivotal period in terms of understanding these social media during a time when many sites launched (in either 2015 and 2016) and tracks their development over successive years.We chose October as a "routine" period in the political calendar, avoiding the party conference season and parliamentary recesses, and the likelihood of elections which habitually (though not exclusively) run in late spring.In sum, we wanted to see how AOPM used Twitter during what we determined as period of "regular" political activity.
We chose 2015-2018 as this would provide some insight into the patterns of Twitter use and behaviour either side of a general election (in June 2017).We selected this particular general election since it coincided with public acknowledgement that "a small group of hyperpartisan British media outlets have quietly built enormous audiences on Facebook" (Waterson 2017a).Further, while "the Sun and the Daily Mail sell 3 million copies a day between them", nonetheless "their decades-long claim to dominating public opinion may have finally come to an end" (Waterson 2017b) as AOPM apparently began to wield considerable influence of their own.Tweets were collected using Twitter's Full Archive Search AP, which we accessed using Twurl to collect JSON files, and subsequently converted into Excel files ready for manual coding.Our sample, therefore, represents the "full" content from each account, excluding deleted content, but including all Tweet types.In total, there were 9,284 standard Tweets, 634 quote Tweets, 1443 reply Tweets, and 3446 Retweets.Following analysis carried out by Molyneux and Mourão (2019, 254; see also Hermida and Mellado 2020, 865), we treated each class of Tweet as discrete subsamples within our analysis (see Table 1) in order to understand activity across the affordances of Twitter.Besides quantifying Tweets, we coded each type and share metrics (as of August 2019).The interpretive analysis in terms of "purpose" focused on five key variables:

Tweet Purpose
. to share content e.g., links to articles, videos, images produced by the outlet who is Tweeting; . to share content from other media publications; . to share opinion, conjecture, speculation, viewpoints, hypothesis, predictions; . to share information e.g., a fact, figure, report, announcement, event; . to share hominem, dismissive, inflammatory, sarcastic, insulting content aimed at others; .Other purposes, including promoting individuals or organisations, appeals for subscribers, running polls, etc.
In practice, coding was straightforward, since the limit of characters naturally restricts Tweets from performing many functions simultaneously.Accordingly, there was no double coding, and where there was a decision to make, the more dominant Tweet function was chosen.If an opinion, for example, was in any way inflammatory and specifically directed, then this was coded as "attack", rather than the sharing of less contentious and less targeted opinions.Most often, Tweets simply share online content, and this was straightforward to code.
Political reference and sentiment.Where a Tweet referred to a UK political party, politician, representative or general references to the "left", "right" or "the government".We determined "positive" as anything supportive of a party or associated ideology, including the validity of its policies or the behaviours of those representing it.We coded "Negative" for anything interpreted as critical, such as a policy failing or suggestions of poor practice, or corruption.Whenever there was no evaluative judgement, we coded "neutral".Manual coding enabled us to capture nuanced versions of these categories, including sarcasm or more oblique references that nonetheless could clearly be assigned as either "positive" or "negative".
Media Reference and sentiment.Where a Tweet referred to the BBC, a UK media outlet or journalist, the "mainstream media", or other alternative media outlets.As before, we coded "positive" for anything supportive of legacy media, such as the quality of their journalism and whether they were performing well, for example."Negative" was anything interpreted as critical of legacy media bands, perhaps pointing to "biased" coverage or no coverage of a particular issue at all.As before, "neutral" was coded in the absence of any evaluation.
The data was analysed by three coders and intercoder reliability tests were performed on 1,485 Tweets (10% of the sample).Levels of agreement for all variables were between 89.7% and 94.3%, and the more intuitive Krippendorf Alpha scores ranged between 0.83 and 0.91, indicating a robust and repeatable framework and a reliable coding process.

Findings
Promoting "GOOD" Twitter Behaviours Our first analytical element is to determine whether AOPM's Twitter use supported the dissemination of real-time information.Orellana-Rodriguez and Keane (2018) identify election campaigns as indicative of such "real-time" delivery of news, and Twitter is important for the coverage of political campaigns (Lawrence et al. 2014;Parmelee 2013).While our 2015, 2016 and 2018 sample periods cover "routine" political weeks, our 2017 sample embraced a week within the UK general election campaign.
Table 1 shows -with the exception of The Conservative Woman -that the volume of Tweets for all sites increased significantly in 2017, only to decrease the following year.
By virtue of their intensified Twitter activity during the election, AOPM satisfy our first analytical element of "positive" Twitter use.Indeed, they are recognised as capable of generating social media "traffic" many legacy news providers would envy (Waterson, 2017b).Table 2 shows that standard Tweets dominated (62.7%) for all but three outlets (AAV, The Skwawkbox and Evolve Politics).Importantly, scholarship suggests that standard Tweets and Retweets tended to conform to "traditional journalistic values", whereas quote Tweets and reply Tweets tended to follow a "looser set of values" (Molyneux and Mourão 2019, 262), aligning with our measure of "objectivity".Table 3 considers standard Tweets in closer detail and reveals that all outlets overwhelmingly focused on providing links to their own content."Other" contains a series of minor objectives/purposes, with no single purpose comprising of more than 1.1% of the total 14,807 Tweets.We found little uniformity in the use of Twitter across the nine sites.For example, while AAV regularly offers opinions, other sites such as Evolve Politics were more prone to attack within their Tweets.However, 7 out of 9 sites emphatically preferred to generate Tweets linking to their own content.Accordingly, we conclude that this next criterion within our analytical model of "good" or "regular" Twitter practice is emphatically met, as AOPM clearly provide a conduit to news provision.
By sharing their own Twitter content, AOPM strongly adhere to the established "sharing" logic of Twitter.Twitter, therefore, functions as a gateway, but importantly, there is no distinctive difference between the way that AOPM and MSM brands used Twitter.In sum, it might be concluded that despite AOPM's "alternative" and radical raison d'etre, their use of Twitter emerges as something altogether more traditional.
The next "positive" indicator of "good journalism" in our model is to provide news and commentary to specific audiencesin other wordsto provide content and links to content that might reasonably be expected by those following AOPM sites on Twitter.At face value, we conclude that audiences might reasonably anticipate what sort of content they might expect via Tweets and the linked material they are directed to.AOPMs develop a politically motivated news agenda and all that entails, and as we develop later, this is closely connected to the emergence of filter bubbles.
The final "positive" metric is to determine whether Twitter use facilitated self-promotion and branding.By examining AOPM Twitter activity we can determine the extent to which they are developing and growing their audience.Given that online content relies on its "shareability" to establish a foothold in a frenzied Twittersphere, we can establish how successful AOPM sites were in building a distinctive presence within it.On the basis that Retweeting indicates interest, trust and agreement (Metaxas et al. 2015), we analysed Retweet rates across our 4-year sample.We calculated an average Tweet rate by dividing the total number of times AOPMs were retweeted by the total number of Tweets.As Table 4 shows, 2017 was a watershed for all sites as their content was retweeted more than in 2016.But only 5 of the AOPM sites (Westmonster, Breitbart, Evolve, AAC and Novara) were able to sustain and build on this rate of Retweets.For others, Retweet rates fell in 2018.
The Canary, Guido Fawkes and Skwawkbox -3 of the 4 busiest Tweeters across our samplewere unable to sustain their "shareability" after 2017.Perhaps the appropriate conclusion herealbeit only measuring their performance on Twitter -is that it is inconclusive as to whether these sites have been able to develop any wider momentum as distinctive and growing entities within the wider landscape of journalism.But overall, when regarding these "positive" indicators, we can reasonably conclude that these sites generally conform to a model of practice that might be considered as enhancing journalism with "good" behaviours.

Promoting "BAD" Twitter Behaviours
The first more negative trait within our adapted model is to determine whether Twitter use facilitates less objective behaviour by journalists, which here we operationalise as "unprofessional" behaviour.It is commonly opined, for example, that journalists should report the news and not be part of it (see Smith and Grabowski 2010).
Twitter has long been seen by media organisations as a way of building consumer relationships and raising brand awareness (Hermida, 2013, 299).Consistent with Twitter being used to provide insight into audience engagement (Conte 2016; Manthorpe 2018), the analytical model we have adopted does, to a reasonable degree, facilitate such insight.Accordingly, the first "negative" indicator is audience interaction.We can draw a conclusion as to whether, by interacting directly with their audience, AOPM might be trending towards more subjective behaviour that conflicts with the fundamental tenet of journalistic objectivity.
As we have mentioned, for the limited context of this study, we conceptualise "objectivity" as maintaining an appropriate distance from stories and particularly audiences.On a platform that facilities interaction, news organisations have a choice whether to further interact with those commenting on their Tweets, or to ignore them.If Twitter encourages journalists "to converse with their readers and express themselves more freely", then they might be breaching any sense of "detachment" as they do so (Orellana-Rodriguez and Keane 2018, 77).This element gets to the heart of whether personal involvement makes for "good journalism" (Beckett 2014).The debate is further complicated since if AOPM aims to provide alternative news narratives challenging the MSM and more accurately representing audience interests and needs, do they achieve this by interacting and "mixing"in virtual spaces at leastwith those they claim to be serving with better journalism?While replying to Tweets or adding quotes to Retweets is admittedly a rather crude metric to determine "interaction" more broadly, we nonetheless judge this to be a strong indicator as to the degree to which these AOPM are prepared to directly connect with their audiences.
Table 5 shows that AAV are the outlier here, in that they were more willing to engage with their audience via replies or by adding commentary (good, bad or neutral) to Tweets posted by others.Otherwise, such interaction was much less across all sites (only 14.0% of all Tweets) and audience interaction is relatively unusual within AOPM Twitter behaviour.We conclude that the more traditional "Broadcast model" of "one to many" is preferred.
Turning now to the final element of less positive Twitter behaviour that might point to "bad" journalism, our starting position is that -albeit loosely -AOPM sites can be categorised as either politically left-or right-leaning.As such, and as was made clear in Figure 2, they are considered as transparently partisan in their objectives and Table 6 shows that even the more ambiguously self-identifying Guido Fawkes, for example, can be fairly clearly assigned as supporting a right-sided ideology.
Of course, we understand that reducing political choice to a simple Labour (left-wing)/ Conservative (right-wing) binary is not always helpful, but it is justified here, since 86.9% (7647) of all political sentiments expressed within our sample (8803) were about these two parties.Table 6 shows the sentiment expressed when Tweets were commenting about each party.
From Table 6 we conclude that while there is no simple nor consistent pattern, most often the left-wing sites were positive about Labour.Right-wing sites were almost never positive about Labour and were most often negative.Left-wing sites were overwhelmingly negative about the right-wing Conservatives, but the right-wing sites' support for the Conservatives is more tentative.Amid a generally predictable pattern of positive and negatives, the much less certain support for the Conservative party from the right leaning sites within our sample is an interesting nuance that merits further, future scrutiny.Right wing editors and journalists, for example, might be able to shed some light as to the reasons for this apparently less certain support for the Conservatives.In sum, though, it can be reasonably concluded that within their Twitter activity at least, followers of left and right AOPM could expect a predictable ideological thrust.Accordingly, the "self-selected, target audience" (Orellana-Rodriguez and Keane 2018, 77) will expect to find content aligned with a particular political viewpoint.Besides the specific political direction, AOPM Twitter followers are also exposed to regular criticismparticular and specificof MSM (see also Cushion, McDowell-Naylor, and Thomas 2021).Table 7 shows the sentiments expressed when AOPM Tweets refer to the MSM either via a specific journalist, media organisation or more generic references to something like "the UK media".
Most often (in 59.1% of Tweets), MSM were reported neither positively nor negatively, but in 37.9% of cases, references to them were critical.AAV (72.6%) and The Canary (59.3%) are particularly critical, while Guido Fawkes (22.4% of cases) and Evolve Politics (35.2%) are less so.What can be concluded more emphatically is that there was almost no positive support for mainstream media, with only 2.0% of all Tweets being supportive/complimentary.This is perhaps understandable, given that an associate of Evolve Politics suggested that a top "traffic driver" for their site is criticising the BBC's "political coverage".
We conclude that this particular metric of positiveor "good" -journalistic practice is met; audiences can largely self-select sites that are transparent about their ideological / editorial direction; they are generally affiliated to either the left or right but are seemingly united by their dissatisfaction with the MSM.Perhaps more alarmingly, however, the final measure within our model is also met, and we conclude that it is only a short step between the provision of content with a predictable and partisan editorial thrust and the development of so called "echo chambers" or "filter bubbles" (Grossetti, du Mouza, and Travers 2019, 212) which "restrain the diversity of opinions".

Discussion and Conclusion
Social media such as Twitter are complex platforms that can potentially disturb journalism's traditional boundaries.Their use redefines the wider power dynamics of journalism and audiences because journalists are not necessarily the dominant participants (Carlson and Lewis 2020).Seeking to examine how this might apply to AOPM Twitter use, our large-scale content analysis offered the first ever comprehensive study, examining how these UK sites used Twitter between 2015 and 2018.Our analysis reveals a patchy synergy between them.So, in answer to RQ1, which asks about the ways AOPM use Twitter and the trends and patterns therein, our conclusion is that these AOPM are a heterogenous group that employ seemingly disparate Twitter strategies.AOPM have "core similarities" but otherwise vary in terms of "content, appearance, audience, and reach" (McDowell-Naylor et al. 2021a, 170).On a simple level, such heterogeneity can be amply shown by the considerable variance in the volume of Tweets across different AOPM; AAV for example, Tweeted 492 times across our sample, while Guido Fawkes Tweeted 3580 times -over 7 times as often.Our longitudinal data reveals that during the general election in 2017, there was a climax of Twitter activity (see also Thomas and McDowell-Naylor 2019) accounting for almost 50% of the whole sample of 14,807.Activity dropped markedly in 2018.
Further, while standard Tweets account for 99% of activity by Breitbart and Conservative Woman and dominate overall, they account for less than 20% of AAV's activity.There was more uniformity in Tweets linking to an external source (usually content produced by the AOPMs themselves), but both AAV and Evolve were significant outliers when compared to almost all other sites, who link to their own content over 80% of the time.AOPM's political affinity and sentiment towards the mainstream were roughly, rather than precisely, aligned, with right wing sites less supportive of the Conservative party than the left-wing sites were for Labour.There is more variance in patterns of audience interaction (via Reply or Quote Tweets) with AAV as an outlier once again.
So, while AAV in particular, had a distinctive way of using Twitter, we more generally conclude that Twitter strategies appear contextual depending on the political calendar, and that there is no standard set of consistent Twitter logics that AOPM seem to follow.Indeed, their Twitter strategies might be described as variable, random and perhaps even haphazard.This we argue, underlines the still emerging nature of AOPM, and indeed, the need to continue to analyse their operations as these news organisations mature.
Whether this lack of homogeneity could be reasonably expected in this context is moot.On one hand, the existence of organising models and frameworks called "social media logics" (van Dijck and Poell 2013), "Twitter logics" (Olausson 2017) and even the "broadcast model" (Liang et al. 2019) does, at face value, point to some anticipated degree of consistency in the use of Twitter.But perhaps more compellingly, it is clear that while these sites are similar in some respects, there are significant differences, for example, in the way they report, how they organise and manage themselves, and the scale of their commercial aspirations (Declan McDowell-Naylor et al. 2021b).It seems logical that sites managed by small teams or single individuals (for example AAV) will operate differently to those who are better resourced (for example, The Canary).But at this stage, since empirical research into AOPMor whatever different label might emerge in the futureis relatively new, we would argue that one of our important contributions here is to determine that little can be taken for granted in terms of a "pack mentality" or a "standard" mode of operation within new online media.
However, by adopting the traditional "broadcast" model of Twitter use, AOPM collectively mimic legacy news brands (see Holcomb and Mitchell 2011) rather than offering something "alternative" to MSM.More widely, we confirm that while these digital news outlets can "challenge journalistic orthodoxy" they can also "borrow from it" (Carlson and Lewis 2020, 123).In our view, this contrasts with narratives claiming that such new media are "rebellious" (Harlow and Harp 2013, 42) and "innovative" (Lee 2005, 12) and suggests at least in the way that they use Twitter -that even new alternative news providers fall back on more traditional modes of journalistic operation.
For RQ2, we were able to go beyond the simple quantification of AOPM's operational behaviour when using Twitter.Our analytical framework enabled us to shine light on the positive/good and negative/bad traits of Twitter behaviour, which in turn enables some conclusions as to whether such behaviours enhance the democratic function or inhibit it.Dividing such behaviour characteristics/ traits into those considered "good" or "bad" is perhaps too binary a concept within such a complex site of study, and perhaps instead might be reconceptualised as traits that are more or less threatening to journalism's wider democratic objectives.Taking the more positive first, the summary in Figure 3 shows that there were some clearly identifiable traits that are "good"according to Orellana-Rodriguez and Keane's (2018) model of Twitter use.
Firstly, we found AOPM provided information about real-time current affairs and by linking to external content, Twitter acts as a gateway to wider sources of information beyond its 140-character model.Content can be reasonably predicted, both in terms of political affiliation and criticism of MSM, and so audiences are specific and self-selected.Given that these AOPM actually run and manage Twitter accounts at allalbeit in different ways -does intrinsically satisfy the next analytical element in that if they wished to do so, the audience could become involved in the news narratives by interacting with journalists, and each other.
The self-promotion and branding elements are much less clear.When measured in terms of Retweet rates, and the degree to which the audience were sharing AOPM-originated content, it seems that for some, Twitter "momentum" -in terms of how often their content was passed on by its followers -has been difficult to develop.The busiest Tweeters -The Canary, Guido Fawkes and The Skwawkbox -could not repeat 2017 Retweet levels in 2018, and perhaps missed the chance to build some critical mass, post-election.
The notion of distinctive branding is also problematised since some AOPMmost notably Evolve Politics, The Skwawkbox, and Guido Fawkes -are more associated with individuals than other outlets.The Skwawkbox, for example, is individually run by Steve Walker, with whom the outlet is almost synonymous.Similarly, Guido Fawkes is closely associated with journalist Paul Staines.Even as these AOPM become established within the wider media system (Thomas and McDowell-Naylor 2019) the lines between ongoing news brands and personal brands are blurred.
Turning to the components characterised as "bad" journalism, our starting point is that "news organizations increasingly regard social media as not only a place for research and distribution of content but also as an important platform for audience participation" (Hedman 2016).However, we found generally low levels of interaction, and the general adoption of the broadcast model of one directional transfer of information.Having found similarly low levels of audience interaction, Molyneux and Mourão (2019, 262) suggest that this might be journalists and editors marginalising an online public "too often full of trolls and harassment" or "simply feeling uncomfortable engaging their audiences".Either case perhaps, seems at odds with a more contemporary news organisation's objective to engage and interact with audiences (Holton and Molyneux 2017).
This low level of interaction, therefore, is difficult to understand.With little or no evidence of journalists inserting themselves into stories within these Tweets at least, their objectivity -such as it is and as we define it -is not obviously violated.However, we are re-igniting the familiar debate concerning whether news providers engaging or interacting on Twitter might compromise perceptions of objectivity.Most pertinently here, however, if AOPM are claiming to give a voice to the voiceless (Atton 2007) or are closely identifying with populist news agendas (Heft et al. 2019), then remaining at arm's length from their audience seems an underdeveloped Twitter strategy and one not obviously designed to bring these news organisations and their regular social media users any closer.It also seems to develop a distance between these emerging news brands and the core followers that they might rely on to promote their brands more informally.
This engages with a much wider debate as to the degree to which audiences "should or should not take part in newsmaking" (Carlson and Lewis 2020, 126).Whether this is part of a purposeful mimicking of mainstream media social media practice or is less intentional is, of course, not revealed by a content analysis.Nonetheless, future research might usefully engage with AOPM editors and managers to discover the degree to which the wider social media strategy is part of a developed plan or something more ad hoc and unstructured, how these Twitter accounts are curated, and the working practices that are followed.
In the final element of our analysis, we found that none of the AOPM sites -whether on the right or left of politics -provided a balanced view of the world.Given their self-identified political leanings of course, this is no surprise, but does not seemingly satisfy the conclusive wish of a majority of digital audiences who want a more neutral approach, and "prefer news that reflects a range of views and lets them decide what to think" (Newman et al. 2021, 9).Instead, such obvious partisanship is likely to solidify the establishing of echo chambersor "information cocoons" that might eventually be "exacerbating extremism" (Guess et al. 2018, 4).
However, while we began by grouping the sites in our study as either "left" or "right", we must add a note of caution for future research.Table 6 shows quite clearly that partisanship is not simply a case of the left supporting the left and the right supporting the right.Indeed, the right-sided sites often seem as frustrated by the Conservative party as anyone much further to the left, while the left-sided sites themselves seem ambivalent about the issue of party leadership in particular.They often seem capable of unleashing invective towards various factions within the Labour party as fierce as anything directed against their usual Conservative opponents (McDowell-Naylor and Thomas 2019).Future research, therefore, should incorporate a nuanced understanding of partisanship.
In summary, our analysis sheds new light on the editorial practices of AOPM by examining their Twitter use over time.We showed how AOPM pursued a unidirectional "broadcast" approach to Twitter, with limited audience engagement.In this vein, the enactment of traditional journalistic norms by AOPM on Twitter calls into question not just the boundaries of mainstream/alternative journalism (Nygaard 2019) but also its fundamental definition amid enormous transformations (Deuze and Witschge 2018).In other words, since some alternative media sites use Twitter in broadly the same way as legacy news outlets, the lines between what are considered "alternative" and "mainstream" media may be blurring.

Disclosure Statement
No potential conflict of interest was reported by the author(s).

RQ1:
How did AOPM use Twitter between 2015 and 2018 and what trends and patterns can be identified?RQ2: To what degree, through their Twitter use, do AOPM fulfil what our analytical model determine as "good" of "bad" journalism?

Figure 2 .
Figure 2. Statements of partisanship on the websites of alt-media sites.

Table 2 .
Types of tweet between 2015 and 2018.

Table 4 .
Retweets across the full sample period.

Table 5 .
Degree of audience interaction.

Table 7 .
Tweets expressing a sentiment about the MSM.