Twitter trolls may be nasty because they don't know names

Why there are so many Twitter trolls: We're more likely to be nasty to someone if we don't know their name

An experiment found only a quarter of people will cooperate with strangers

But they are almost three times as likely to be kind if they know a person's name

Researchers say reducing anonymity online could help tackle trolling

Published: 14:00 EST, 29 March 2017 | Updated: 07:12 EST, 30 March 2017

A new experiment could explain why trolls on Twitter post horrible comments to people they have never met.

Scientists have discovered that people are meaner to others when they have no clues about their identity.

They found that revealing a person's name, even after participants had met and spent time together, was enough to make people act more kindly to each other.

Scroll down for video

A new experiment could explain why trolls on Twitter post horrible comments to people they have never met (stock photo)

WHAT'S IN A NAME?

Scientists have discovered that people are meaner to people when they don't know their name.

They found that revealing a person's name, even after participants had met and spent time together, was enough to make people act more kindly to each other.

Humans have evolved to be unselfish where it is better to work as a group to share in the rewards.

But it is believed this effect goes away when they do not know each other.

Researchers said something as simple as a name could introduce a level of mutual understanding between people.

This could explain why Twitter trolls target strangers and hide behind their anonymity.

People were almost three times as likely to cooperate with each other when they were no longer anonymous strangers.

Scientists led by Northwestern Polytechnical University in China recreated an experiment originally designed by US mathematicians in the 1950s called the Prisoner's Dilemma.

This is a set-up in which two people with no knowledge of each others’ actions are told they are on trial and choose whether to testify against each other.

If one ‘punishes’ the other by testifying, they stand to win a cash reward, but if both behave badly they receive high fines. Both people working together, and refusing to testify, are able to walk free but get no reward.

The study, involving 154 students, found just a quarter cooperated rather than acting out of self interest in the first round of the game.

But, when they were told the name of the people they stood to hurt, 70 per cent of people cooperated.

The students knew their fellow participants, having attended the same classes, so were better able to empathise.

The authors said reducing anyomity online could be used to tackle trolling on sites such as Facebook and Twitter.

Dr Zhen Wang, from Northwestern Polytechnical University, said: ‘Today, it often seems that conflict trumps cooperation, be it on the internet or in national politics - likewise in evolution, Darwinian selection should result in individuals pursuing their own selfish interest.

Scientists have discovered that people are meaner to others when they have no clues about their identity. This could explain why Twitter trolls target strangers (stock photo)

‘Our findings suggest that it is crucial to ask one rather straightforward question - do the prospective cooperators know each other reasonably well?

'If they do, they will more likely not try to win at the expense of each other, but together.’

Humans have evolved to be unselfish where it is better to work as a group to share in the rewards.

However it is believed this effect goes away when they do not know each other.

This could explain why Twitter trolls target strangers and hide behind their anonymity.

THE PRISONER'S DILEMMA

The Prisoner's Dilemma refers to a scenario where two prisoners are taken into custody.

The model shows the consequences of acting selfishly (by confessing to your crime) and by staying loyal to your partner (by staying silent).

If both prisoners act selfishly by confessing to their crimes, they are sent away to spend five years in prison.

But if they both remain loyal to each other and stay silent, they are punished with a smaller one-year sentence.

However if one prisoner confesses while their partner remains silent, they are able to walk through.

Scientists use this model as a basis for predicting when people, animals and machines will choose to cooperate with one another or to act selfishly.

The study allowed people to make choices anonymously or with knowledge of the people involved in the experiment.

Anonymous participants became known to each other simply by giving them each other’s names.

The students had attended the same classes, so had previous knowledge of each other, were the same age and shared interests.

This knowledge increased the cooperation rate from 25 to 70 per cent in the first round and from eight per cent to 23 per cent in a subsequent round of the game.

Professor Marko Jusup, from Hokkaido University in Japan, said: ‘We found that when participants knew each other, this significantly increased the frequency of cooperation.

'This paid out very well for all, so winners play nice.’

Professor Jürgen Kurths, from the Potsdam Institute for Climate Impact Research in Germany, who analysed the results, added: ‘Since the spirit of cooperation that social cohesion is based upon is crumbling away in some places, be it on Facebook or in societies that are about to be torn apart about issues such as immigration, we sought insight into what enhances cooperation.

‘This might also apply to conflicts about environmental resources.

'However, we have to further explore the continuum, the many states between complete anonymity and very well knowing the other person.

‘It will be exciting to learn what kind of information, what degree of mutual recognition is needed to promote cooperation.’

The study was published in the journal Scientific Advances found.

HOW TO SPOT AN ONLINE TROLL

Researchers from Stanford and Cornell University recently studied 40 million posts made by 1.7 million web users to try to identify trolls.

From this data they were able identify so-called Future-Banned Users (FBUs) - and Never-Banned Users (NBUs).

The former are people that are more likely to exhibit antisocial behaviour on online forums, while the latter are everyday users that don't show signs of troll-like behaviour.

The researchers then built an algorithm that scans posts for signs of such antisocial behaviour, and a study shows this algorithm can identify potential trolls in 80% of cases.

For example, posts made by trolls are less readable.

Trolls are more likely to veer off-topic and have fewer similarities in terms of language and content compared to other posts.

Trolls are also less likely to use positive words than other users and they swear more.

In terms of activity, the study found that trolls make more comments each day, and post more times on each thread.

Trolls were also more likely to have had posts deleted than other users.