Skip to main content



Want to fight online misogyny? Look to the bystanders

Are our offhand likes and shares on social media as innocuous as we might think? Perhaps not, if we are bystanders to online misogyny and violence.

Want to fight online misogyny? Look to the bystanders

For one thing, bystanders might believe that by liking and retweeting misogynistic content is “not as bad” as originating it themselves. Yet the practical effect of liking and retweeting is reinforcement and amplification, which can be devastating to victims.

Are our offhand likes and shares on social media as innocuous as we might think? Perhaps not, if we are bystanders to online misogyny and violence.

A bystander is someone who witnesses an act of abuse, including violence against women, and has the choice whether or not to intervene (for example, to stop the violence or help the victim in some way).

Bystanders might know the parties involved –– friends, family, colleagues –– or they could be strangers, meaning they have no personal stake in the situation.

The traditional concept of a bystander has been limited by geographic circumstance, for example, commuters at MRT stations who alert security, the victim or both about sexual voyeurism.

However, as the sites of abuse have changed –– with our lives increasingly led in online spaces  –– our idea of bystanders has also evolved. And rightfully so.

Given the chance, we’d all like to think that we would be the good bystander, heroically coming to a victim’s defence.

But is that actually what’s happening online? A new study by Aware and technology firm Quilt.AI suggests otherwise.

By analysing publicly available data, our research team tracked and categorised different types of misogynistic statements across select platforms, to get a better understanding of what misogyny looks like and the kind of support it enjoys.

First, we sourced sexist and misogynistic content from local Twitter, Facebook, Instagram, YouTube accounts and the forums Reddit, HardwareZone and Sammyboy.

We found comments such as “women deserve to be enslaved”, “you’re a man-hating gargoyle” and “doctor, the sandwich won't make itself”.

Then we trained Quilt.AI’s proprietary machine-learning model to recognise similar content on its own.

The model picked up a random sample of around 700 misogynistic tweets between 2016 and this year, and found that most fell into the misogyny category of “belittling and objectifying”.

The second most prevalent category was “flipping the narrative” (that is, implying that men are the oppressed gender, not women), followed by “rape myths”.

A closer look at how users were interacting with this misogynistic content revealed that, in terms of engagement, misogynistic tweets handily outperform non-misogynistic tweets.

They are twice as likely to be liked and 4.5 times more likely to be retweeted when compared to non-misogynistic tweets.

In short, users are not only failing to call out misogynistic content, or simply ignoring it –– they are actively engaging with and promoting it, thus perpetuating misogynistic behaviour themselves.


One might have expected to see more positive bystander intervention on Twitter than on other social media platforms, given that  –– unlike Facebook, Instagram or TikTok  –– it is designed expressly to facilitate discourse between individuals.

Nevertheless, we have a few theories as to why misogynistic content seems to enjoy such popularity with bystanders on Twitter.

For one thing, bystanders might believe that by liking and retweeting misogynistic content is “not as bad” as originating it themselves. (Think of this as the “he started it” argument for evading culpability.)

Yet the practical effect of liking and retweeting is reinforcement and amplification, which can be devastating to victims.

After all, when a piece of content  –– say, a sexist statement like “the gender pay gap has been debunked by economists”  –– picks up traction on social media, it racks up eyeballs exponentially, gaining cultural significance and influence in spite of being blatantly disprovable.

Secondly, these users may simply not recognise this content to be harmful or problematic in the first place.

Research indicates that understanding of gender inequality in its various forms is not particularly strong in Singapore.

For example, a 2019 Ipsos survey of 1,019 Singapore citizens and permanent residents found that only 32 per cent believed that there was a gender pay gap in the country, while 45 per cent believed that women who wear revealing clothes “should not complain” if men make comments about their appearance.

The latter, in particular, falls under the category of ideas known as rape myths  –– ideas commonly used to justify sexual assault, often by displacing blame from perpetrator to victim  –– which comprised 11 per cent of the misogynistic tweets captured by our model.

Another reason why bystanders might be hesitant to intervene in online misogyny is how difficult it can be to breach the social norms of masculine solidarity. This may be especially salient if the originator of the content is a peer or a user they admire.

It’s hard for friends to call each other out, a process that typically leads to some combination of embarrassment, frustration and anger.

A power dynamic might also be at play, based on “influence”. A user with thousands of followers might be more difficult to publicly rebuke than someone with a negligible following.

Lastly, bystanders might not feel compelled to intervene in violence because of the way an online space such as Twitter compounds diffusion of responsibility.

By this principle, the more bystanders are present in a situation, the less personal responsibility falls upon each individual bystander to take action –– in other words, everybody expects somebody else to step up.

Such diffusion can dissuade even individuals in small groups to take action –– all the more so in a virtual space like Twitter, where thousands, even millions of potential bystanders are in the “room”.

Though our study focused on Twitter, we can expect these factors to play a part in bystander behaviour on other online platforms, too.


Maybe you’re thinking: What’s the big deal? It’s just social media.

If so, consider that globally, there are twice as many male Twitter users as female users, a statistic that many have linked to the rampant misogyny on the platform –– which, as Amnesty International pointed out, “leads women to self-censor what they post, limit their interactions, and even drives women off Twitter completely”.

For many professional women (from artists to academics to politicians), having a presence on Twitter is a crucial part of their work.

Consider too that the tendency of online hostility to veer into real-world violence is palpable. A study published in February, for example, found that geolocated misogynistic tweets in the United States correlated with domestic- and family-violence incidents in those areas.

The engagement rates of misogynistic content speak to the need for policymakers to prioritise bystander education.

It’s not just the loudest voices we need to watch out for — the online equivalent of that one guy who insults his female colleagues behind their backs.

We also need to reach those who would nod and laugh at that guy’s jokes, or even repeat them to others later on.

Instead of turning the other cheek, or contributing to a “pile-on”, a bystander can take real action against online abuse by posting something supportive about the victim or sending them a sympathetic private message.

They could also directly call out the abuser by responding critically to their original post.

Comprehensive sexuality education that instils gender-equal values and the skills to critique violent, misinformed ideas is ever more essential in the digital age.

Beyond that foundational education, though, bystander intervention programmes that encourage more active prevention of misogyny and violence should be developed in Singapore.

A systematic review of bystander programmes in the US, Canada and India observed that their most pronounced benefit was changing attitudes towards rape myths.

Students’ rejection of rape myths was sustained both immediately after the programme, and months afterwards.

The Government can work with non-profits and sex educators to develop such programmes in Singapore — and design them to cover online situations, in addition to offline ones — as part of the solution to the spread of misogyny.

Over a three-hour online session, a successfully designed programme can help bystanders identify the problem, assume responsibility and teach them how to intervene while keeping themselves and the victims safe.

Ultimately, we should think of these efforts as not only a means to stop potential perpetrators of misogyny and abuse, but measures that strengthen the resilience of our entire society against inequality.



Shailey Hingorani is head of research and advocacy at the Association of Women for Action and Research (Aware) and Michelle Gay is director of operations at Quilt.AI.

Related topics

misogyny violence online bystanders gender women

Read more of the latest in




Stay in the know. Anytime. Anywhere.

Subscribe to get daily news updates, insights and must reads delivered straight to your inbox.

By clicking subscribe, I agree for my personal data to be used to send me TODAY newsletters, promotional offers and for research and analysis.