Did you ever feel offended at something you’ve seen in Twitter? According to SimpleTexting’s new survey, Twitter tops the “most harmful apps” list. Twitter was the platform with most trolling, according to 38.1% of those surveyed. This compares to 26.9 percent, who stated that trolling is a problem on Facebook and 14.8 per cent on Reddit.
Instead of bringing us together, social media apparently is tearing us apart – so much so that six in 10 people in the SimpleTexting survey said that they were afraid to post about certain topics for fear of negative feedback. 90% of respondents said they had seen or heard racist comments from others in their social network.
Additionally, 86% of respondents said they have seen inappropriate content about sexual orientation or gender shared by others in their networks. Notable is the fact that 46% of respondents admitted to having had negative interactions with people online, while 87% said they would unfollow or block someone they didn’t like on social media.
Spam is a big problem on Twitter
A second study found that Twitter has more than toxicity problems. It also suffers from serious spam issues. GlobalData, an international analytics company released its findings. It found that up to 10 percent of Twitter users are spammers.
Given that Elon Musk is unable to take control of Twitter, his bid for the company is currently on hold because there is a dispute over the percentage of spam accounts. Twitter says bot/spam accounts make up less than 5 percent of Twitter accounts, while GlobalData’s senior data scientist Sidharth K Kumar said otherwise.
Kumar stated that “the exact proportion of spam account is hard to compute” as it’s almost impossible to identify the person behind the tweet handle. The definition of spam accounts may be different for every person. Spam can also be defined as persistent tweeting non-original content, although some people may view it as an active user who shares articles/opinions.
GlobalData developed a mathematical model that estimated spam accounts using several parameters. It then calculated a weighted scoring, which was used to classify the account as “spam”, or “nonspam.” GlobalData determined these parameters by looking at the activity differences between spam accounts and average Twitter users.
Kumar stated that “there were research papers published earlier in media which looked at certain followers to estimate spam/bot proportions.” The best approach was to examine live streams because that’s more representative of Twitter activity. Because we wanted to ensure that spam was not being identified, our estimation is somewhat conservative. This estimate is only an estimation. It is impossible to determine if an account is spam or bot.