Women Targeted on Twitter

By Kait Scalisi, MPH

Towards the end of 2014, Women, Action, Media (WAM), a nonprofit dedicated to building a robust, effective, inclusive movement for gender justice in media, became an authorized reporter on Twitter. This status allowed them to monitor, report, and follow-up on instances of harassment experienced on the platform. From over 800 reports of harassment, WAM discovered disturbing patterns in online behavior and Twitter’s response. They published their findings and recommendations this week.

Women experience significantly more and more severe forms of online harassment. WAM sought to quantify this data with a pilot study assessing the types and frequency of harassment. In three weeks, WAM processed over 800 reports and sent over 300 to Twitter. From there, Twitter took action on 161.

Examining the nearly 400 genuine harassment reports led to some interesting findings. These include

Bystanders play a huge role in reporting online harassment. Over 50% of reports received were submitted by someone other than the person being harassed. Most of these reports mentioned only one receiver of harassment.
Ongoing harassment is a problem. Nearly one-third of reports dealt with on-going harassment and over two-thirds mentioned having reported this problem to Twitter in the past. However, most harassing accounts were reported only once, indicating there’s a lot of jerks out there.
Hate speech and doxxing (releasing someone’s private information) were the most common forms of harassment. Together they accounted for nearly half of all reports. An additional 19% of reports did not fall neatly into any category.
Twitter took action in 55% of reported cases. In most cases, accounts were suspended. Only one account was deleted.
Twitter was more likely to take action on reports of hate speech. Twitter was much less likely to take action for other forms of harassment though they showed no favoritism towards longstanding accounts.
The reporting process is complicated. Harassment receivers must provide URLs, not screenshots. This makes it difficult to capture instances of harassment in which the perpetrator “tweets and deletes.” Reviewing reports involves in-depth communication and may be triggering to reviewers.
There are four major challenges of proving harassment. These include: context, interpretation, mode, and format of evidence.

Taken together these conclusions highlight the need for more oversight and accountability for alleged harassers as well as updated policies. As a result of the results, WAM also made six specific suggestions.

More broadly and clearly define what constitutes online harassment and abuse.
Update the abuse reporting interface, using researched and tested trauma-response design methods.
Develop new policies which recognize and address current methods that harassers use to manipulate and evade Twitter’s evidence requirements.
Expand the ability for users to filter out abusive mentions
Hold online abusers accountable for the gravity of their actions.
Diversify Twitter’s leadership.

The report itself may not be super surprising if you’ve paid attention to online harassment incidences over the last few years. REgardless, the report is vital as its the first piece of well-researched evidence quantifiying how online harassment is going down- and what to do about it.