When Internet Convos Go Toxic: Science to the Rescue

Toxic Online Conversations


“Let me repeat, You are wrong. And an idiot.”  If you have been on the receiving end of this or a similar online reaction, you have a lot of company.  The reason is because online conversations far too often degenerate into personal attacks. <Think: online chats with Customer Service.>

Science to the Rescue: In hopes that those attacks can be averted, researchers have created a model to predict which civil conversations might take a turn and derail.

The study: The study analyzed 1,270 conversations that began civilly but degenerated into personal attacks, culled from 50 million conversations across 16 million Wikipedia “talk” pages, where editors discuss articles or other issues. They examined exchanges in pairs, comparing each conversation that ended badly with one that succeeded on the same topic, so the results weren’t skewed by sensitive subject matter such as politics.

After analyzing the exchanges, the researchers developed a computer program that scans for warning signs in the language used by participants at the start of a conversation.

Tip-Offs for Problematic Convos: Some warning signs for a troubled conversation ahead:  Repeated, direct questioning or use of the word “you” — are among the signs that predict which initially civil conversations would go awry.

Tip-Offs for Smooth Convos: Some signs the convo will not degenerate: Early exchanges that included greetings, expressions of gratitude, hedges such as “it seems,” and the words “I” and “we” were more likely to remain civil.

How good are you at detecting which online convos will go south?

People can test their own ability to guess which conversations will derail at an online quiz.


Research: “Conversations Gone Awry: Detecting Early Signs of Conversational Failure.” 2018, Cornell University.  Authors: Cristian Danescu-Niculescu-Mizil, assistant professor of information science, Cornell Ph.D. information science student Justine Zhang; Ph.D. computer science students Jonathan P. Chang, and Yiqing Hua; Lucas Dixon and Nithum Thain of Jigsaw; and Dario Taraborelli of the Wikimedia Foundation.