Skip to toolbar

50% of online abuse is authored by women – Evaluation

By Jun.01, 2016




In the past week, considerable attention has been given to the launch of the #reclaimtheinternet campaign, a cross-party initiative aimed at opening a debate concerning strategies and interventions to address issues connected to online abuse.




The homepage of the initiative states: ‘misogyny, sexism, racism, homophobia, transphobia, intimidation or abuse online mean that some voices are silenced. We’ve heard from teenagers bullied online, women targeted by rape threats, teachers subjected to sexist abuse by pupils     and parents, people enduring racist attacks, smears and slurs’.


The criticism


Looking at the #reclaimtheinternet hashtag associated with the campaign, several tweets referred to the findings of a study conducted by Demos which were released at the same time as the launch: the study reported that in the UK alone, 65.000 women were targeted by abuse on Twitter in the three week period of the study. The interesting twist, and one which was often emphasised on the Twitterspere as well as on mainstream media, was that 50% of the abuse came from women’s accounts.

Critics therefore took this as proof that there is no need to protect women online, as they are the main perpetrators of abuse.

It is important therefore to consider what the study actually teaches us.








The tweets were selected over a (non-specified) three-week period, using two search keywords ‘slut’ and ‘whore’: these words were identified in a previous study (May, 2014) as terms which men and women almost equally use on Twitter to attack women. The 2014 report was concerned with three keywords: ‘rape’, ‘slut’ and ‘whore’.


The missing Keyword




It is interesting that the 2016 study excluded the third keyword featured in the original study, ‘rape’. The original study identified over 100,000 Tweets mentioning ‘rape’ between 26th December 2013 and 9th February 2014. More than 1 in 10 appeared to be threatening in nature.Tweets containing the keyword ‘rape’ were classified on the basis of their imputed function as Serious/News, Metaphore/Casual, Threat/Abusive, or Other.

The majority of Tweets (40% of the overall total) mentioning rape were connected to discussion of news or current affairs, or in the context of political debates and campaigns.

The second most frequent category was defined by the researcher ‘casual’/metaphor and featured uses of the word rape that the coders (or the algorithm) did not deem threatening: while one example provided (@^^^^ this is my rape face LOL) is not a simply ‘casual’ mention and actually could be quite disturbing for recipients, I convene that the use of the word ‘rape’ in this case does not imply a deliberate attack on the individual.

Threat and Abusive Tweets were 12% of the sampled Tweets, and the remaining 27% was classified as ‘Other’.

Although men seemed to be using this word more frequently, the difference did not reach significance. This is almost certainly due to the fact that this result is based on the overall sample, rather than focusing on when the term was used as a threat. I would expect many women to mention rape in the context of news, political debates or campaigning. But I wonder how many of those 12% of Threatening or abusive Tweets were by women?


The two keywords selected



The result concerning the equal use of ‘slut’ and ‘whore’ by men and women to attack women is in line with Preston and Stanley’s 1987 study, in which these terms appeared in the top 10 insults directed towards women from both genders. In this sense, the study replicates previous research by identifying these terms as part of a dominant repertoire across genders.



Morality and Competence


Insults such as the ones considered above have a strong moral component: by addressing someone with those terms, one is challenging their moral integrity and their moral standing (Rubini et al., 2016). The repertoire of insults directed to women is nonetheless not exclusively focussed on morality, but also on competence (Semin and Rubini, 1990; Heflick & Goldenberg, 2009): thus, insults like ‘dumb’, ‘idiot’, ‘stupid’, ‘dim’, etc… serve the function of undermining women by denying them intellectual and/or professional competence (ibid.).

The study conducted therefore focuses on a very limited subset of the repertoire of insults available and in doing so it is likely to underestimate the amount of verbal abuse received by women, as well as differences of repertoires across genders.


Identifying Users



It is not clear in the report what were the criteria used to identify the gender of the users whose tweets were analysed: some people on social media adopt false identities (Caspi and Gorsky, 2006), and although this is not a particularly widespread phenomenon, it is likely that a proportion of those using the internet to abuse others would be doing so under false pretences. This begs the question as to how many of the self-professed women in the sample were actually women.

This is not to say that women do not insult each other online using those derogatory terms (as argued above, these are likely to be the most common insults in general), but whether and how the researchers managed to accurately estimate users’ gender remains unclear.


Does it matter?




A more general point on this study is that it highlighted the scale of abuse, especially when considering the extremely limited definition of abuse adopted (based on only two keywords). Whether coming from men or women, such abuse can have detrimental effects on the recipients (e.g. Ortega et al. 2012) and it is important to open up a dialogue on how to best address this social issue and how to support the victims of abuse online.



Image 1 Image 2  Image 3: n/a  Image 4 Image 5
Leave a comment

Leave a Reply