The scale of abuse directed at MPs has revealed that in a sample of nearly three million tweets mentioning MPs over a six-week period, more than 130,000 were considered toxic.

The BBC Shared Data Unit analysed 2.9 million tweets aimed at UK politicians across six weeks between March 15 and April 23 2022, as part of a large-scale investigation.

Around 20,000 of those tweets analysed were deemed as severely toxic.

Which MPs received the most 'toxic tweets'?

Boris Johnson, the prime minister at the time of the study, received 18,907 tweets deemed toxic, more than double that of the nearest MP.

However, he was not in the threshold for the top 20 MPs to receive the most abuse proportionally, which instead was made up of backbenchers and no cabinet or shadow cabinet members.

Within the top 20 was MP for North West Cambridgeshire Shailesh Vara, with seven per cent of the tweets towards him measured as being 'toxic'.

Fellow Cambridgeshire MPs Jonathan Djanogly, Lucy Frazer and Stephen Barclay all ranked inside the top 155 of MPs to receive the most abuse proportionally from a total pool of almost 600 MPs.

The high levels of proportional abuse directed at Cambirdgehsire MPs was a reflected trend for MPs in the East of England, where a total of 28,167 toxic-rated tweets were sent to MPs in the region, second only to London MPs (35,679). 

What is a toxic tweet?

The BBC Shared Data Unit used Jigsaw, a research unit within Google, to use Artificial Intelligence to spot toxic comments online.

The software utilised was Perspective AI, which rated comments against five main qualities: toxicity, identity attack, insult, profanity and threat.

The algorithm defines a toxic tweet as a "rude, disrespectful or unreasonable comment that is likely to make someone leave a conversation" and ranks the probability text contains an insult or is toxic.

A piece of text would have to break a threshold of 0.7 to be considered a toxic comment.

For instance, the report found that "stupid" and "pathetic" were two of the most commonly used adjectives in what was classified as toxic tweets.

The software also found that female MPs were more likely to be called "thick" and "ignorant" and be subject to sexualised language, while their male counterparts were more likely to be called "liars".

Lucy Frazer

South East Cambridgeshire MP Lucy Frazer received 633 tweets in the time frame, of which 19 were deemed toxic.

Therefore three per cent of the tweets she received were toxic, the joint 153rd highest proportion amongst all MPs.

Jonathan Djanogly

Huntingdon MP Jonathan Djanogly received 588 tweets during the investigation, and 34 were classified as toxic.

The proportion of toxic tweets received by Mr Djanogly was 3.4 per cent, joint 119th for all MPs. 

Stephen Barclay

North East Cambridgeshire MP Stephen Barclay received 6893 tweets, and 295 were measured as toxic.

The proportion of toxic tweets he received was 4.3 per cent, which was a higher percentage than Boris Johnson, and placed Mr Barclay ninth in the East of England for MPs who received more than 50 tweets and 61st amongst all MPs.

Shailesh Vara

When filtering out the MPs who received less than 1,000 tweets during the study, Mr Vara was the seventh most abused MP proportionally when measured in regards to toxicity.

Of the 5627 tweets sent to Mr Vara, 393 were classified as toxic (seven per cent).

Mr Vara, categorised in the report as a self-identified ethnic minority, was the highest MP categorised as so to have received the most abuse proportionally out of MPs who received more than 1,000 tweets.

In response to the findings, Mr Vara said: “Social media abuse is a sad fact of public life, and it affects all politicians, irrespective of their party.

"Online abuse continues to increase and despite all their rhetoric, social media companies are not doing anything near enough to police and stop it.

"There needs to be much tougher regulation and enforcement, as well as high levels of fines that genuinely impact on the companies’ behaviour as they are profit driven.

"There is much to be said for the phrase that if you wouldn’t say something face-to-face then don’t type it.

"This sort of abusive behaviour only serves to discourage good people from putting themselves forward to serve in public office.”