Project Results
contributions
This effort helped us analyse a representative sample of tweets sent to 100 women politicians in India over the General Elections 2019.
involved
That’s enough Decoders to fill India’s Parliament three times over!
Most Decoders did come from India - with Canada, Nigeria, United Kingdom, Bangladesh, and US next.
Forum visits
Pageviews of the Amnesty Decoders discussion forum
contributed
That’s the equivalent of someone working full-time for a year!
Number of posts to the Amnesty Decoders discussion forum during the project
Within India, we had over 26 States or Territories represented. Most Decoders who gave their location came from Delhi, Uttar Pradesh, Chandigarh, Karnataka and Maharashtra.
The Decoders decoded in languages including English, Hindi, Bengali, Gujarati, Kannada, Malayalam, Marathi, Tamil and Telegu. Many Hindi tweets were in combination with English - in Hinglish.
What’s next?
With the help of hundreds of Decoders, we analysed a representative sample of tweets that were sent to 100 women politicians in India over a three-month period around the 2019 General Elections in India. This will help us conduct an unprecedented large scale analysis of abuse against Indian women politicians on Twitter.
We are working with a team of data scientists to validate and analyse all the data submitted by Decoders. We are also experimenting with the latest machine learning technologies to scale up the sample tweet analysis to more than 8 million (80 lac) tweets.
Watch this space for an upcoming report and interactive visualisation of the results.
REGISTER FOR UPDATES AND JOIN THE DISCUSSIONAbout the project
Women in India have reported facing severe online violence on Twitter. However, the exact nature and scale of abuse in the Indian context is under-researched. Research shows that online abuse against women has an adverse effect on their political participation and impacts their public life. Online abuse is silencing - it leads women to self-censoring, limiting what they post, anonymizing their accounts, or leaving Twitter altogether. It also has a strong negative impact on mental well-being and perceptions of offline safety. Women who are from religious and ethnic minorities, transwomen, queer women, and Dalit women are especially affected.
Although Amnesty International, among many others, has undertaken extensive research on the extent of abuse on women politicians and journalists in the Global North, there is a dearth of research on forms of online abuse in the Global South. This project will aim to build that knowledge, specifically with a focus on #ToxicTwitter in India.
How it worked
Amnesty collected tens of thousands of tweets sent to women politicians in India. For each task, Decoders saw a random anonymised tweet and were asked to identify whether it contained content that according to their opinion was 'abusive' or 'problematic' or 'not'. If identified as 'abusive' or 'problematic' they had to mention the nature of abuse. The majority of tweets were in Hindi followed by English. The Indian regional languages comprised over 16% of tweets.
Anyone with a mobile phone or computer was able to contribute. The Discussion forum enabled the Decoders to raise queries, concerns and engage with each other. The Forum was actively co-moderated by Volunteers themselves.
Why this campaign?
Women in India face gendered abuse from both individual stand-alone accounts and organised “troll armies”. The abuse faced by women politicians is aimed at silencing them and at discouraging their political participation. In the run-up to the elections, women leaders were called a number of hateful, sexist abuses including “Chudail (witch)”, “Rakhel” (concubine), and “Randi” (prostitute). They were subjected to racist, ethnic and religious slurs and have even been even threatened with rape and murder.
The timing of this project was selected to study tweets to women politicians in the run-up and during the 2019 General Elections of India.
The Troll Patrol India project is helping to identify the volume and type of abusive and problematic tweets directed at these women politicians. It is part of a larger research project where the data collected are being analysed to shed light on how these tweets may deter women from freely posting their views on Twitter and how they instil fear in women contesting elections.
Impact of previous research
In 2018, Amnesty has led extensive research into online abuse against women politicians and journalists from the UK and USA. Troll Patrol was a joint effort by human rights researchers, technical experts and thousands of online volunteers to build the world’s largest crowd-sourced dataset of online abuse against women.
The findings revealed for the first time the sheer scale and nature of online abuse faced by women. We found that 7.1% of tweets sent to the women in the study were problematic or abusive. This amounts to 1.1 million problematic or abusive mentions of 778 women across the year, or one every 30 seconds on average. Women of colour, (black, Asian, Latinx and mixed-race women) were 34% more likely to be mentioned in abusive or problematic tweets than white women. Black women were disproportionately targeted, being 84% more likely than white women to be mentioned in abusive or problematic tweets.
Acknowledgements
Amnesty Decoders: Troll Patrol India is a joint initiative by Amnesty International and Amnesty India. Troll Patrol India was the seventh project for the Amnesty Decoders - a global network of digital volunteers for human rights research. Already there are more than 50,000 volunteers from more than 150 countries.
This project was developed in collaboration with Focal Labs. Powered by Hive, a modular, flexible, open-source platform developed by the New York Times and Discourse, an open source discussion platform.
Amnesty would like to thank all the volunteers who have helped so far with this project. Shout out to our most active Decoders who decoded thousands of tweets. They were - @Junaid, @Akriti, @Pallavi, @Gargii, @shreya_07, @Tannia, @Lavanya, @xyzzz, @Rohit, @mthassan, @Azeemmian, among others.
A special mention to our star volunteer @alisaleem who along with 19 more volunteers further engaged other volunteers from across India. They were - @rabiya, @Nihalkaur, @Ipshita7, @Siddharth, @Kopal, @womensaviour2000, @Muskan, @Zaara, @mohitsimalti, @aayushi_30, @Akan26, @Prashant, @Mishti, @jatin5652, @bibanshu_amnestyindia, @Bhoomika, @Shweta_attri, @Primzie and @Charvi.
A special thank you to our Volunteer moderators @indiatrollpatrol (Natasha and Madhulika) who diligently and passionately managed the Discussion Forum and would engage with Decoders on a daily basis. Our most active community members @Akash, @Bayzid, @gmk87, @Katyayani, @snehal_dhote, @Tannia and many others actively participated in the discussions.
Decoders are people power at its best and we truly couldn't have done it without you.
Get involved
Become a Decoder and be notified of project updates and new project launches. Thank you for your support.
BECOME A DECODER
