Digital microaggressions and everyday othering: an analysis of tweets sent to women members of Parliament in the UK

ABSTRACT Online abuse directed at female politicians has the potential to present barriers to the political representation of women in the digital age. Previous research has produced mixed results, with some finding little difference between the extent of abuse received by male and female MPs, while others found that women and minority groups are subjected to specific kinds of abuse. While straightforward examples of abuse and discrimination are easy to identify, categorise and quantify, tweets which include more subtle microaggressions that position women and minority MPs as unqualified and unwelcome in politics deserve attention. This research therefore employs a qualitative thematic analysis of 11,543 tweets sent to UK members of Parliament. The analysis identified four themes: 1. ‘overt online abuse’, 2. ‘everyday sexism and othering’, 3. ‘dismissing discrimination and victim blaming’, and 4. ‘claiming reverse discrimination’. We argue that these digital microaggressions serve as constant reminders of the marginalised status of female representatives, and women of colour specifically, and should be conceptualised as forms of psychological and semiotic violence that reconstitute online political spaces as a hostile environment for women and may discourage women from seeking political office or compel women representatives to leave. Digital microaggressions may therefore be as damaging to women’s democratic representation as outright abuse.


Introduction
In the UK, there has been a high-profile debate surrounding online abuse directed at female MPs, and its potential to present barriers to the political representation of women in the digital age. Studies of this phenomenon have produced mixed results, for example, Ward and McLoughlin (2020) found very little difference in the extent of abuse received by male and female MPs, while others have found that the women receive particular kinds of abuse (Southern & Harmer, 2019). Black women MPs have been found to attract high levels of abuse compared to their colleagues (Stambolieva, 2017). Many of these studies have used some form of automated analysis, searching a Twitter corpus for keywords.
While this approach has illuminated the scale of the problem, this deductive approach misses the more subtle forms of sexism, racism and other forms of discrimination people experience online. In our previous work we deployed a manual content analysis of 117,802 tweets sent to UK MPs, and found that female MPs were more likely than men to receive generally uncivil tweets, tweets which stereotyped their identity, and tweets which questioned their position or ability as politicians (Southern & Harmer, 2021). While straightforward examples of abuse and discrimination were easy to identify and categorise, tweets which included more subtle forms of othering which positioned women and minority MPs as unqualified and unwelcome in politics were more challenging to quantify. These microaggressions that are not overtly abusive but nevertheless undermine women and other underrepresented groups, potentially discouraging their participation in politics, are the focus of this article.
Since most academic studies in this area employ large scale quantitative analyses of Twitter data, we argue that more qualitative analysis is needed to identify and make sense of more everyday forms of othering and discrimination which undermines politicians online. This article builds on our previous work by conducting a qualitative thematic analysis of 11,543 tweets. The analysis in this article identifies digital microaggressions which are often too subtle or nuanced to easily quantify. The research is guided by two main aims: to establish what kinds of microaggressions MPs experience and also whether these microaggressions target multiple dimensions of their identity, such as race and class, in addition to their gender. Four themes were identified, 1. overt online abuse (which we include to demonstrate its presence), 2. 'everyday sexism and othering', 3. 'dismissing discrimination and victim blaming', and 4. 'claiming reverse discrimination'. We argue that these microaggressions reinforce politics as a hierarchical space where certain politicians are stigmatised or unwelcome. The article will first set out the existing literature and key concepts which inform this research before explaining the data collection and analysis. The article will then present our findings and will finally discuss the potential implications of our findings for the participation of women in politics.

Digital politics: incivility, intolerance, and violence against women in politics
Online participation in politics is fraught with difficulties for women, and many have argued that the intemperate and uncivil tone of much online discussion represents a threat to democratic values (Harmer & Lumsden, 2019;Papacharissi, 2004). The experiences of women political representatives are further complicated by the tension that arises from the increased importance of digital media in their campaigning and everyday political work. Social media use is becoming an important method of communication with constituents which makes it difficult to opt-out. Not having a social media presence may at the very least be a political disadvantage to women, or they may risk being seen as out of touch by voters who have a right to contact their elected representatives (Southern & Harmer, 2021).
Scholars have defined a variety of behaviours as uncivil, including ad hominem attacks, vulgarity, derogatory language, direct insults, and name-calling (Kenski et al., 2020;Rossini, 2020). It has also been claimed that incivility in online debates may discourage talented candidates from seeking elected office (Maisel, 2012). Rossini (2020) argues that it is not incivility that threatens democratic norms because rudeness and swearing are not always used to offend. Instead, Rossini argues that more extreme forms of incivility, such as racial slurs and threats to violence are better conceived of as political intolerance, which does pose a specific threat to democratic pluralism and participation.
Political intolerance can also be considered an expression of violence. In her foundational book on violence against women in politics, Krook (2020) outlines a typology of violence against women in politics which covers traditional understandings of violence (physical, sexual, and economic) but also advocates that our understanding of violence should be extended beyond these narrow parameters to include psychological and, what she calls 'semiotic' violence. Krook argues that the ultimate goal of any form of violence against women is to deter them from participating in politics. The concepts of psychological and semiotic violence are of particular relevance to this research.
Psychological abuse can be defined as attempts 'to disempower targets by degrading, demoralising or shaming themoften through efforts to instil fear, cause stress or harm their credibility' (Krook, 2020, p. 139). The effects of psychological violence can be severe for female representatives, and according to previous research, it is the most common form of violence against women politicians (Bardell, 2011;Herrick et al., 2019). Furthermore, and perhaps counter-intuitively, some recipients of psychological and emotional violence have reported greater trauma in some instances than if they received physical abuse (Follingstad et al., 1990;Stark, 2007).
'Semiotic violence' is 'using words and images … to injure, discipline and subjugate members of marginalised groups' (Krook, 2020, p. 188). It can take many different forms. Semiotic violence draws on earlier concepts such Tuchman's 'symbolic annihilation' (2000) whereby women (and members of other marginalised groups) are attacked figuratively, rather than physically, by marginalising, trivialising, and demonising them. Such treatment has already been thoroughly established in traditional media coverage (see, for example, Ross, 2002). More recently, and in the online context, there has been a spate of 'cheapfakes' or 'deepfakes' produced where images of female politicians' faces are digitally placed on other images or videos, sometimes pornographic, without their consent (Southern & Harmer, 2019;Maddocks, 2020). These practices serve to undermine the contribution that women make to the political sphere. We argue that online discrimination and othering of women MPs on Twitter constitutes a form of semiotic and psychological violence designed to undermine their participation.

Online abuse of politicians and public figures
Research into the issue of online discrimination and abuse aimed at female political representatives and other public figures is still relatively scarce. A study that used data from the 2017 UK general election candidate survey found that 32.7% of candidates had experienced at least one type of harassment (Collignon & Rüdig, 2020). This research further suggests that women and young candidates are more likely to be harassed or intimidated. The Inter-Parliamentary Union (2016) identified online abuse as the most common type of psychological abuse experienced by women in politics. Wagner (2020) conducted 101 in-depth interviews with political candidates or representatives in Canada. While she found no evidence that online abuse curbs women's ambition to go into politics, she did find that women politicians were more concerned about online harassment than male candidates. Women also reported that it impaired their ability to do their job and that it led some women to consider leaving their posts.
Women in the public eye have spoken about developing ways of coping with online abuse, such as blocking and deleting abusive messages. However, these actions increasingly take valuable time away from their duties as representatives (Mason, 2014). Research by Sobieraj (2020) about women who had been subjected to impersonal abuse from strangers online, some of whom were public figures, showed that many of the women interviewed reported that they had been intimidated, shamed, and discredited. Some reported that these experiences had a 'chilling' effect on what they said, although some were also defiant and resolved to continue speaking out. Women also reported having to engage in 'preventative' or 'ameliorative' labour in the face of online abuse, taking time and energy away from other tasks. Women politicians also recognise that they risk becoming the target of online abuse for addressing specific topics, particularly feminist issues (Rawlinson, 2018). Female politicians claim that, combined with 'real-world' violent incidents (such as the assassination of MP Jo Cox during the EU Referendum), online threats and abusive messages produce a hostile environment towards women (Carter & Sneesby, 2017).
Other studies have analysed the content of abusive messages. Ward and McLoughlin (2020) analysed tweets send to MPs over three months and found that while male MPs received most of the casual insults, female MPs received the vast majority (86%) of hateful speech. Our own previous work analysing replies sent to MPs found that women politicians were subject to higher levels of general incivility, messages stereotyping them and messages telling them they do not belong in politics (Southern & Harmer, 2021). Stambolieva (2017) analysed online abuse sent to female MPs during the 2017 election in the UK. The results showed that 60% of all abuse during the election had been sent to one MP -Diane Abbottthe first Black woman to be elected to parliament in 1987. Evidence from a machine learning analysis of 2.2 million tweets sent to US and Canadian politicians by found a higher level of incivility toward female representatives but only at the highest levels of name-recognition (Rheault et al., 2019). These studies all found an association between having a higher profile and receiving more incivility and abuse. However, in the first three studies, this was particularly pronounced for women representatives. The evidence suggests that women are censured more for gaining prominence or success in politics, which may deter women from reaching their full potential, thus creating a barrier to women aiming for the highest office. Soberaj (2020) posits that there are three facets shared by those women who experience the worst online abusebeing a member of one or multiple marginalised groups, speaking up about feminist issues or those that otherwise challenge the status quo, and speaking in a male-dominated space.
The evidence from these empirical studies suggests that women are more likely to experience and are more concerned about online abuse and harassment. It also suggests that minority groups are more likely to experience the worst kinds of online abuse. These findings demonstrate that online abuse and harassment often target multiple sites of one's identity (Harmer & Lumsden, 2019). Hackworth (2018) similarly argues that it is difficult to separate attacks on one's identity and calls for more intersectional approaches to online harassment research. This article responds to that call by seeking to identify whether digital microaggressions target multiple aspects of the recipients' identities.
Empirical work that analyses the ways that politicians experience harassment based on multiple aspects of their identity remains rare. Previous work by the authors is a notable exception (Southern & Harmer, 2019). We analysed tweets sent to female MPs who were also people of colour, disabled or identified as LGBT (or some combination of these identities). The results showed that Black and Asian women MPs received abusive tweets which were simultaneously racist and sexist, demonstrating the need for more research of this kind. Crucially, this work also identified several more subtle ways in which tweets sent to women MPs reinforced sexist assumptions, such as attempting to silence them and questioning their intelligence. These results show that it is essential to adopt an intersectional approach when trying to understand the way that women MPs experience online harassment.
The research also indicates that while research into online abuse is vital for understanding how online messages can be harmful to politicians from marginalised groups, focusing on these more 'extreme' types of harassment (such as death and rape threats) ignores the fact that low-level forms of othering and discrimination can be equally harmful to democratic participation (Harmer & Lumsden, 2019). Our study will contribute further evidence that tweets sent to MPs can be characterised by a constant 'dripping tap' or 'background noise' of sexism, racism, and other discrimination that complicates the everyday working lives of women political representatives. The research aims to establish what kinds of microaggressions MPs experience and also whether these microaggressions target multiple dimensions of their identity, such as race and class, in addition to their gender. To this end we draw on the concepts of intersectionality and microaggressions to demonstrate the discriminatory potential of subtle forms of othering and stereotyping directed at women politicians.

Intersectionality and microaggressions
Intersectionality is an approach which derives from the Black feminist thought (Crenshaw, 1989) and ultimately critiques the idea of taking 'one group at a time' to solve discrimination (Collins & Bilge, 2020, p. 7). Instead, its advocates argue that it is crucial to assess 'the mutually constitutive relations among social identities' that impact people at 'individual, interpersonal, and social structural levels' (Shields, 2008, p. 302). While intersectional insights and frameworks are highly contested and complex (Cho et al., 2013), they offer an analytic tool for understanding multiple levels of oppression and discrimination. This approach helps us to understand how women and minority groups navigate political institutions which were created by and for white men. Puwar (2004) highlights the pernicious and subtle ways which certain bodies are entitled to certain spaces, while others are deemed 'space invaders'. Women and people are colour are subjected to invisible or indistinct forms of discrimination which positions them as interlopers.
These subtle forms of othering and discrimination are called microaggressions (Pierce et al., 1978). Initially, this concept was used to theorise a broader understanding of racism by accounting for the more subtle and insidious aspects of racial discrimination. Later research expanded this to include types of behaviour directed towards any marginalised group (Ikuenobe, 2011;Sue et al., 2007). Sue et al. (2007, p. 1) describe microaggressions as 'the everyday verbal, non-verbal and environmental slights, snubs or insults, whether intentional or non-intentional, that communicate hostile, derogatory or negative messages to target persons based solely upon their marginalised group membership'. These constant reminders about the marginalised status of disadvantaged groups make them aware that even 'well meaning' people have internalised negative stereotypes about them. Experiencing constant low-level slights, or as Nadal (2008, p. 23) describe them 'brief and commonplace daily … indignities' can be harmful because they increase the risk of psychological stressors and can even lead to low self-esteem and depression (Swim et al., 2001).
Evidence also suggests that because these daily indignities are thought of as 'perceived minimal harms' (Sue, 2010, p. 14), this can exacerbate the damage, as recipients feel pressured into dismissing or ignoring them. However, the cumulative and everyday nature of such experiences are far from trivial. They can have severe detrimental cognitive effects, such as expending cognitive energy on deciding whether a particular incident was intentional or not or having to decide to respond or not; thus, taking time and energy away from other tasks. Microaggressions can also lead to behavioural changes such as hyper-vigilance in social settings in case an incident should occur or increased scepticism towards all members of a dominant group (Sue, 2010).
Responding to microaggressions is also risky because it may unwittingly play into existing stereotypes. For example, women can be described as 'hysterical' for taking a microaggression seriously, or Black people are labelled as 'angry' for calling out discrimination (Cadinu et al., 2005). Respondents are left in a no-win situation. They can either choose to let it go, and then feel a loss of integrity that they did not stand up for themselves or their group, or they are left feeling as though they created conflict over a 'minor' point which might undermine perceptions of them or their group. Considering all of this, we argue, therefore, that experiencing microaggressions in online spaces constitutes a form of psychological and semiotic violence against women in politics which poses a risk to women's participation in democratic politics.
Existing studies have demonstrated that women politicians are more likely to receive extreme forms of abuse online, especially if they also belong to another minority group. While overt forms of harassment are easy to identify, there is also evidence that online communication is awash with more subtle microaggressions which undermine women and which impact on the way women MPs can use social media for their political work which are relatively unexplored in the literature. Our study addressed this gap by identifying some of the ways that these microaggressions manifest themselves in everyday communication and identifying the extent to which they target multiple aspects of the recipients' identity.

Methods
This study draws on a subset of messages from a much larger data set of 117,802 tweets. 1 The original study identified 11,543 uncivil tweets from the data set. The previous study used content analysed to identify quantitative patterns of abusive or more subtle forms of othering, such as stereotyping and questioning intelligence (see Southern & Harmer, 2021). The content analysis identified broad trends in the kinds of uncivil messages MPs received (the results have been discussed in the previous section). However, there were also instances where MPs were othered in more subtle and nuanced ways which are difficult to quantify. This study builds on that previous work to capture and reveal these microaggressions more effectively. In this article we subjected the 11,543 tweets, which were deemed uncivil in the previous study, to a qualitative thematic analysis. While the sample contains tweets directed at men and women MPs, those cited in the analysis section focus specifically on the microaggressions targeted at women.

Data collection
At the time of the original data collection, there were 578 MPs on Twitter (MPs on Twitter, 2018). Eighteen were deemed inactive and were removed from the sample. The authors also removed the 50 most followed MPs. Although analysing messages sent to these more high-profile figures is, of course, important, the authors felt that the experiences of more 'ordinary' representatives were the key population of interest here. This left a sample of 500 MPs. We then captured every public tweet aimed at each of these 500 MPs (both direct replies and @-mentions). We did so across a sampling period which ran from 19 March until 2 April 2018 inclusive, during standard Parliamentary business (that is, not recess or vacation), using NodeXL software (version one).

Data analysis
The 11,543 uncivil tweets were analysed using thematic analysis. We followed Braun and Clarke's (2006) approach for conducting the thematic analysis. This analysis was undertaken by the two authors. Thematic analysis is an inductive method which identifies themes within the data, rather than imposing a predetermined coding scheme. To familiarise ourselves with the data, we read and re-read the tweets carefully. Each tweet was qualitatively coded independently by the authors. This means we produced informal notes highlighting instances where the tweet contained specific forms of othering or microaggressions. For illustration, some of our initial codes included: 'implying recipient is disloyal to wider community for religious reasons'; 'recipient has used sex for career progression', and 'fat shaming'. To ensure that the analysis was well founded and rigorous, the authors held frequent discussions where we adjudicated any disagreements, and discussed what the qualitative codes meant, and how we could categorise them into broader themes. Areas of significant overlap were identified and developed into the broader themes which are elaborated in our analysis. Although we present them singularly for clarity, each form of microaggression should be considered as interwoven and mutually reinforcing, resulting in an online environment which is hostile to women and minority groups in politics.
The aim of this approach is not to quantify the presence of these microaggressions or make any claims about their prevalence. These tweets mostly evade discrete categorisation because they contained multiple slights or indignities aimed at marginalised groups, as our analysis will demonstrate. Moreover, the analysis does not claim to be representative of all tweets in the sample. We have focused on those microaggressions which highlight sexist or intersectional forms of othering, to that end some MPs feature more prominently in the analysis than others to highlight how Twitter can be a particularly hostile space for women of colour. The point of the analysis is to render these microaggressions visible to enable a more nuanced understanding of how women politicians are othered and discriminated against in online spaces. Following Jane (2014), we believe that to capture the full impact of online harassment, it is essential to quote these sentiments directly rather than couch them in euphemism. We, therefore, include all tweets, unedited (including spelling and grammatical errors). All Twitter handles have been removed apart from the recipient MP.

Digital microaggressions
Four themes were identified in our analysis, firstly overt online abuse (which we include to demonstrate its presence), 'everyday sexism and othering', 'dismissing discrimination and victim blaming', and finally 'claiming reverse discrimination'. While the exclusionary nature of the first theme is evident, the other types of online othering work in much more subtle ways to undermine women MPs and portray them as inadequate or unworthy political representatives.

Overt online abuse
The analysis identified examples of tweets which were openly abusive and misogynistic: @Dr_PhilippaW Smug Tory biatch @EstherMcVey1 More lies from the vile bitch! @claireperrymp I am glad I no longer live in Devizes and have such a sanctimonious stuck up know all bitch as my MP.
@little_pengelly Why are you poking your nose in? Nosy bitch These examples often used gender-specific insults and were supposed to offend their intended recipient. In some cases, women also received abuse based on religious or racial identity, demonstrating the intersectional nature of online othering: @joanryanEnfield Still taking the shekels from Israel are we dear? … still trying to undermine UK democracy with your Mossad associates? @RuthSmeeth You are a spy! You are evil, satanic! Leave! @rushanaraali defending @SadiqKhan record as @MayorofLondon. I can't for the life of me think why. Can you? @ShabanaMahmood @GiselaStuart What's keeping you here … ?
These examples demonstrate that MPs can be exposed to antisemitism and Islamophobia in their everyday online interactions. In addition to displaying overt racism, these tweets also question the loyalties of women MPs and seem to argue that these women do not belong in politics. Moreover, some imply that they do not belong in the UK either. These messages demonstrate that online othering often incorporates attacks on multiple aspects of women's identities, in addition to their gender (Harmer & Lumsden, 2019). These overtly abusive tweets are obvious examples of harassment and discrimination, and they mirror the kinds of content reported in our previous work, as well as other studies which have sought to measure the volume and nature of online abuse directed at political representatives in large data sets (Southern & Harmer, 2021;Stambolieva, 2017;Ward & McLoughlin, 2020). The remaining themes that were identified in this analysis reveal the more subtle and indirect ways that women are harassed online, which are central to our argument.

Everyday sexism and othering
The tweets sometimes refer to the gender of women politicians in ways which undermine or 'other' them, but which were not overtly abusive. Some tweets deployed gendered stereotypes against women MPs which reinforce the idea that they are unsuited to or are unwelcome in public life. @DawnButlerBrent shaking her head, worried if she left the iron on in 1 of her 2 London houses. #pmqs @SarahChampionMP @ITV Gosh how sensitive you are?

@EstherMcVey1 WHAT ABOUT DISABLED PEOPLE WHO ARE TOO ILL AND DIS-ABLED TO WORK AND YOURE TAKING AWAY THEIR DLA AND PIP YOU EVIL HEARTLESS WOMAN
These tweets contain references to well-worn gendered stereotypes. The first example accuses a woman MPs of being too distracted by domestic matters to fulfil her political role effectively. Other tweets accused the recipient of being too sensitive to take part in political discussions. The final tweet is more complicated because it attempts to hold the recipient to account for their political record but ends with a stereotypical insult which undermines her femininity by calling her uncaring. Although this tweet is uncivil, it also contains a subtle microaggression which shows how ingrained sexism can often become a default basis for insulting women.
Sexist tweets also drew attention to the physical appearance of women MPs to criticise them. @Sandbach Lay off the cookies fatty @HeatherWheeler When someone obese votes to take food from poor childrens mouths @annietrev She doesn't give a damn Colin, she's a self-promoting upper class hag, end of.
While the first example insults the MP by calling her unattractive, the second two tweets both go further by directly equating the recipients' perceived attractiveness with their political failings. These tweets invoke their unattractive bodies as evidence for their greed or self-interest. A similar phenomenon was observed in tweets which referred to the private lives and relationships of women MPs. @EstherMcVey1 What an utterly evil person you are. Ester, care to enlighten us all about the "special" relationship you had with Ian Duncan Smith as his junior minister???? Don't remember the press releasing the gory details? @annietrev Social mobility annie style eh Find a rich guy, marry him The first example directly accuses the recipient of having a sexual relationship with a male superior for political gain, while the second critiques the MPs political beliefs by implying that she has married for money. These examples demonstrate, once again, the convenience with which tweeters deploy sexist insults against MPs whose political decisions they dislike, or which directly harm constituents. The focus on their appearance and private lives mirror the ways that traditional media often represent women in politics (see Ross, 2002;Trimble, 2017).
In addition to these overtly sexist tweets, the analysis also identified more subtle forms of online othering of women MPs which sought to undermine and patronise them. These examples deploy gendered terms of endearment to patronise the recipient and call into question their capabilities as political representatives. Some of these tweets also use infantilising language such as 'wee lassie' to undermine their political authority. These examples show that women are often told to resign and shut up, or their capabilities are questioned. This supports findings from our previous work which showed that women were more likely to have their position in politics questioned than male MPs (Southern & Harmer, 2021). Crucially what this qualitative analysis reveals is that tweets make use of gendered naming practices and stereotypes to undermine the contribution that women make to the political sphere.
Several examples implied that the recipients were stupid and invoked gendered terms at the same time: @angelasmithmp Wishing for the moon could never be a "principled position" you daft bat.
@annietrev Get a grip woman, nobody will buy it because of the tariffs it will attract because of brexit! These examples either call attention to the fact that the recipient is female or use gendered language. The phrase 'daft bat' is a mild insult but in the UK it is unlikely to be used against a man. These more subtle microaggressions may not be overtly abusive or harassing. However, they nevertheless underscore the idea that women do not make good political representatives or do not belong in public life.
These findings demonstrate that political Twitter is pervaded by ambient sexism (Fox et al., 2015) which positions women as poor political representatives by repeatedly emphasising their gender identity. Their femininity or domesticity is invoked in different ways to criticising their political decisions, performance, or motivations. In many ways, these messages are striking similar to how mainstream media coverage has historically stereotyped and undermined women politicians (Ross, 2002).

Dismissing discrimination and victim-blaming
In addition to the everyday sexist othering, another form of microaggression identified was how some replies openly disparaged any discussion of discrimination in politics. Several women MPs received tweets which attempted to deny their experiences of inequality. @ThangamMP Proof of this volumes of abuse please? Your alleging criminal offences and should report this all to the police … people in your position should know all this … @RuthSmeeth claimed antisemitism when no one knew or cared whether she was Jewish. We thought she was a tranny.
The first example asks the MP to provide evidence that she and others had been discriminated against and abused, by saying that if she could corroborate her accusations, then she would have reported them to the police. The second example shows the complexity of microaggressions. The tweet attempts to undermine the recipient's claim that she had been discriminated against because of her Jewish identity by suggesting that instead it was because she is trans. The MP in question does not publicly identify as trans, nevertheless the tweet deploys transphobia to discredit her claims about antisemitism. This example underlines that online othering of women MPs is often intersectional (Harmer & Lumsden, 2019).
Other examples went much further by claiming that women MPs either deserve or invite abuse and discrimination in social media posts: @DawnButlerBrent From what I've seen of you Dawn youre reaping what you sow. @little_pengelly I'm starting to think that you want to be trolled.
@little_pengelly Oh, it's what you signed up for. You know the old saying. Can't stand the heat … The first two tweets explicitly state that the recipients are inviting or deserve to be discriminated against or abused. The final tweet frames harassment as part and parcel of politics which MPs should expect. One tweeter attempted to downplay the harassment experienced by a specific woman MP by contrasting it unfavourably to other women: @little_pengelly Maybe if you were more critical and less bias it may help. If you would like to see how a female politician gets abused by mainly DUP supporters and also going back to the leaflets of the flag protests which your party helped pay for then have a look at Naomi Long Twitter feed. This example suggests that the recipient was complicit in discriminatory treatment of a political opponent and therefore cannot complain about her own experiences. These messages downplay or dismiss concerns about discrimination by victim-blaming, accusing women of fabricating or exaggerating the harassment or discrimination that they experienced, or explicitly endorsing the idea that they deserved it.
Previous research has shown that victim-blaming narratives are pervasive on Twitter about violence against women (Stubbs-Richardson et al., 2018). Harmer and Lewis (2020) argue that denying the lived experiences of women who experience harassment in online spaces is problematic because it frames women as untrustworthy and unreliable sources for understanding these phenomena, which means their concerns are often downplayed or ignored. It is also a way of silencing women. Moreover, Krook (2020) argues that violence against women in politics is often dismissed as being the cost of doing business. However, it is a threat to democracy and the political representation of women and minority groups. These microaggressions work to prevent women from speaking out about their experiences and discipline them into accepting discrimination as just part of the job, which could drive them out of politics altogether.

Claiming reverse discrimination
In addition to dismissing women's experiences of discrimination or victim-blaming, some tweets targeted MPs for specifically tweeting about equality and diversity issues. This reinforces Sobieraj's (2020) findings that those who speak out on progressive or social justice issues are more prone to online harassment. The responses in our analysis were dismissive of the need for diversity policies and initiatives, deliberately mischaracterised their aims, or accused MPs of supporting reverse discrimination.
Women MPs were asked to justify and explain their decision to support measures such as equal pay and diversity initiatives in the workplace.
@joswinson "Equal pay is necessary, but not sufficient" Please clarify. Are you suggesting that women should be paid more? @DawnButlerBrent You could always just employ people on merit?
These tweets both argue that there is a hidden agenda behind measures designed to track inequalities and that these policies intend to provide privileged access to opportunities for women and people of colour, rather than levelling the playing field for historically disadvantaged groups. These messages suggest that women MPs are favouring one set of constituents over another, positioning them as discriminating against majority groups like white men. These sentiments were occasionally expressed explicitly: @joswinson *Except for men, they can get stuffed. Right? @DawnButlerBrent White men go to the back of the queue … @CPJElmore Why should it happen? Because he was black? Your colour does not make you more deserving of a military cross.
Here white men are positioned as the real victims of discrimination. In the second example, the inclusion of a racial identifier intensifies the alleged discrimination by suggesting that, in this case, a Black woman MP, is trying to disadvantage white men. This accusation that Black MPs are incapable or unwilling to represent the interests of their white constituents was remarkably pervasive amongst the tweets sent to Black women MPs. Dawn Butler, in particular, received lots of these messages: @Dawn ButlerBrent How far out of your way do you have to go to completely avoid the inclusion of white British people. Labour no longer stands for the working-class majority.
@Dawn ButlerBrent Not very diverse, not one white person on the panel! @DawnButlerBrent @UKLabour Will ur policies help white women too Dawn?
Tweets which accuse politicians of colour, and Black women specifically, of being racists have been identified in other analyses of Twitter (see Southern & Harmer, 2019). The first example even seems to suggest that 'British people' and 'the working-class majority' are white and implies Black women MPs are therefore incapable or unwilling to represent them. Implying that Butler is unable to represent the working classes also makes assumptions about her class background. This is a pernicious form of othering which attempts to downplay or deny the existence of structural inequalities and therefore reassert the right of already advantaged groups to space and privilege.
Other tweets marked out Black and Asian MPs as the problem, accusing them of being divisive: @DawnButlerBrent Why don't you have an integration summit instead?
This example was tweeted in response to the MP sharing photographs from a diversity and inclusion event. It accuses a woman of colour of failing to integrate simply for attending an event for women of colour to discuss their experiences of discrimination and receive support. This is a clear microaggression which asserts that a Black woman MP trying to represent and improve things for people of colour is problematic because it somehow means that she is not representing white people. Once again, these messages frame the rights of the imagined 'majority' as more important than those of under-represented groups.
Some tweets go even further and overtly accused Black MPs of being 'the real' racists: @DawnButlerBrent @UKLabour get rid of Dianne Abbott one of the biggest racists in politics @SeemaMalhotra1 @DavidLammy Yes, your proud to support becos hes black no other reason. What a racist you trully are @DawnButlerBrent trust you to quote a racist, very clear you didn't get to your position on merit.
These tweets assert that the MPs in question are racist. The first two tweets also mention other prominent Black MPs, Diane Abbot and David Lammy. These tweets accuse Black MPs of trying to divide people and exploit racial tensions for political gain. The final example not only accuses the recipient of racism but also questions her position in parliament and suggests that she was only elected based on her ethnicity. This finding corroborates those from our previous intersectional analysis of women MPs which showed that women of colour were repeatedly accused of racism by members of the public (Southern & Harmer, 2019). These microaggressions are particularly insidious because they position MPs of Black and Asian heritage as illegitimate political representatives who are dangerous and divisive. Moreover, such tweets deny the existence of structural inequalities and discrimination while simultaneously perpetuating them. Several of these hostile tweets also advanced the idea that women and Black MPs actively exploit divisive politics for their political gain: @DawnButlerBrent Is that going to designed to create more division and hatred. Your friend, the intellectual colossus @DavidLammy is constantly looking at the world through a black and white lens of division. The pair of you are doing our country a great disservice by stoking up hatred.
Oh bollox -Muslim/Pakistani gangs have raped their way across the countryyour gov't response totally inadequate, @jessphillips @UKLabour turned a blind eye "in order 2 accommodate a community expected 2 vote Lab" @CatSmithMP You still forcing Lancashire non-muslim children to eat halal school lunches for votes?
The first example openly accuses two prominent Black MPs of sewing division on purpose, although it stops short of identifying specific motives or goals. The second two examples accuse two white women MPs of trying to make political capital from favouring Muslim constituents over others. These tweets express overtly Islamophobic sentiments which homogenise and demonise Muslims by associating them with criminal behaviour. They also frame Muslims as a threat to the community. While these racist stereotypes are not directly targeted at the recipients of these tweets, such messages are nevertheless harmful and discriminatory and work to create a communicative environment which is hostile to difference.
Tweets claiming reverse discrimination demonstrate Puwar's (2004) argument that within political institutions, minority groups are often only granted recognition within very specific parameters. They are therefore seen as political representatives for other women and/or ethnic minorities, rather than viewed as representing the universal citizen (which is by default white and male). Therefore, their ability to be effective MPs is called into question. Tweets which claim reverse racism are also a pernicious form of microaggression which attempts to dismiss the material effects of racism and sexism by accusing MPs who attempt to address these issues of being divisive or obsessed with identity politics. Lentin (2020) demonstrates that such tactics are an increasingly commonplace means of seeking to control what counts as racism to emphasise the idea of racism as a personal moral failing rather than a political project that requires structural changes to defeat. MPs who speak up about structural inequalities are then reimagined as problematic for pointing out that inequities exist, rather than the persistence of sexism and racism being recognised as undemocratic.

Discussion: digital microaggressions as psychological and semiotic violence
The research aimed to establish what kinds of microaggressions MPs experience and also whether these microaggressions target multiple dimensions of their identity, such as race and class, in addition to their gender. The analysis showed the myriad ways in which women MPs experience outright abuse, everyday forms of othering, and digital microaggressions. These findings revealed instances of abusive tweets, messages which contained everyday forms of othering, tweets which dismissed their experiences of discrimination and those which accused them of reverse discrimination. Although it is not always possible to assign motive to the individuals who post tweets containing microaggressions, what prompts members of the public to send such tweets is beside the point. What matters is that digital microaggressions are detrimental to the status of women as equal political representatives because they underscore that their acceptance in the political sphere is conditional and precarious.
Digital microaggressions serve as constant reminders of the marginalised status of female representatives, and women of colour in particular, which make them aware that even 'well meaning' people have internalised negative stereotypes about them (Nadal, 2008;Sue et al., 2007). This is especially visible in those examples where tweets aimed to criticise the politics and performance of MPs while explicitly invoking their gender and or racial identity. We, therefore, argue that the experience of repeatedly being subjected to tweets which question one's authority and capabilities as representatives is a form of psychological and semiotic violence because they reconstitute online spaces as a hostile environment for women. It is particularly hostile for those who are also from a minority group and may ultimately discourage other women from seeking political office, or compel women representatives to leave (Wagner, 2020). At the very least it can make their daily lives in the public sphere more difficult to negotiate (Sobieraj, 2020).
The research offers important insights into the presence of digital microaggressions in online spaces, however there are some limitations. The sample includes two weeks of tweets meaning that we were unable to capture the full extent to which microaggressions accumulate over time. Digital microaggressions can also be very specific so removing tweets from their original context, as happens when collecting large data sets, makes interpreting them more difficult, or risks missing important details. While the research gives insight into what microaggressions look like on Twitter, we also recognise that there is a need for more work which addresses how women experience and deal with digital microaggressions to assess the impact on their participation.
The findings however are important because they demonstrate that online harassment and digital microaggressions often target multiple aspects of women's identities. There were examples of misogyny coupled with antisemitism, Islamophobia, and racism, in addition to more subtle microaggressions. This research, therefore, underscores the importance of taking an intersectional approach to analysing online othering and harassment to understand better its potential consequences for political engagement and participation (Hackworth, 2018;Harmer & Lumsden, 2019). Moreover, the analysis shows the importance of qualitative approaches to the study of online harassment and abuse. While large-scale automated content or sentiment analyses are crucial for understanding the scale of the problem and identifying key trends, they are ill-equipped to deal with the range of more subtle and context-specific microaggressions that we identify here.
Given that the demands of an increasingly digital politics show no sign of going anywhere, women political representatives are almost required to be on social media platforms such as Twitter for campaigning as well as their everyday political work (Southern & Harmer, 2021). This means that women MPs must spend time having to engage in preventative measures to ameliorate the worst effects of online harassment and othering which inevitably takes time and energy away from their work (Sobieraj, 2020). Digital microaggressions of this nature should therefore be considered a form of political intolerance which is damaging for democracy because it represents a threat to the participation of women and minorities in formal political institutions (Krook, 2020;Rossini, 2020). Notes 1. A full description of the sample can be found in Southern and Harmer (2021) see reference list.

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

Notes on contributors
Dr Emily Harmer researches the relationship between gender, media and politics. She is the coeditor of Online Othering: Exploring Digital Violence and Discrimination on the Web.
Dr Rosalynd Southern is an expert in online political communication. She has published research on social media campaigning at the last four UK general elections. She also studies online incivility against women politicians.