A recent report by US-based Atlantic Council’s Digital Forensic Research Lab (DFRLab) has shed some light on just how much Twitter traffic can be manipulated in political scenarios.
Looking at how Twitter traffic was influenced during Prime Minister Narendra Modi’s visit to Tamil Nadu in February, the report found that while both hashtags that trended at the time - #TNWelcomesModi and #GoBackModi – were pushed by bot accounts, the bot activity for pushing #TNWelcomesModi was much higher and more aggressive.
The report, published on Medium on Tuesday, says that an 'eyeball test' of accounts that posted #TNWelcomesModi hashtag revealed that they were mostly bots that have since been suspended.
DFRLab, that has worked with Facebook in the past to weed out pages and profiles with “inauthentic behaviour” in the run up to the Lok Sabha elections, found that pro-Modi traffic on Twitter “was far more heavily manipulated than any large-scale traffic flow the DFRLab has analyzed as of yet.”
To calculate the level of manipulation on February 9 and 10, DFRLab used the Coefficient of Traffic Manipulation (CTM) method, which “allows researchers to compare a given Twitter flow with known organic traffic, and traffic that was heavily gamed by small groups.”
The hashtag ‘TNWelcomesModi’ had an astonishingly high CTM score. “The first 49,727 tweets in #TNwelcomesModi scored a CTM of 123.98, the highest the DFRLab has ever recorded, indicating that it was very heavily manipulated by a very small group,” the report said. This is alarming, because in earlier studies, organic traffic was found to score a CTM of 12 or lower.
#GoBackModi trend
The report says that #GoBackModi - that trended during his visit to Tamil Nadu was driven in part by bots and by “a small number of high-volume accounts that posted hundreds of times an hour”.
While calculating the CTM score for #GoBackModi, that trended on February 10, DFRLab kept the sample size close to 50,000 tweets so as to draw a fair comparison. While it trended faster with 49,538 tweets in just over three hours, researchers found that it peaked at a lower rate and had 4,47,000 posts on February 9 and 10, compared to #TNWelcomesModi being mentioned 7,77,000 times in those two days.
The CTM score for #GoBackModi was 46.81. While it was “far above the usual range for organic traffic and comparable with some of the most heavily gamed hashtag campaigns the DFRLab has encountered hitherto, it paled in comparison to the pro-Modi effort,” the report said.
“Unlike #TNwelcomesModi, these accounts (tweeting with #GoBackModi) were still not suspended at the time of drafting,” it adds.
How a few accounts can manipulate traffic
To give an idea of just how these high-volume accounts are and how they can manipulate traffic, DFRLab gave some examples. For #TNWelcomesModi, the accounts @SasiMaha6 was found to be posting most frequently. It tweeted 1,803 times with the hashtag, averaging a tweet every 15 seconds. Similarly, another high-volume account @priyamanaval6 posted with the hashtag 1,677 times which translates to a tweet every 17 seconds.
For #GoBackModi, the top accounts were even more hyperactive. For instance, @PhillyTdp tweeted with the hashtag a whopping 2,179 times or a staggering rate of a tweet every 5.3 seconds, for over three hours. While it tweeted 2,453 times on February 10, 98% of the its tweets the previous day were retweets. Another account, @nritdpusa interestingly posted with #GoBackModi 2,459 times on February 10 – almost exactly the same as @PhillyTdp.
“These individual bots, if anything, were even more prolific than their pro-BJP rivals. Overall, however, the scan of #GoBackModi revealed a larger number of lower-activity accounts,” the report noted. “Overall, the nearly 50,000 tweets in the #TNwelcomesModi scan were posted by just 891 accounts, while the nearly 50,000 tweets in the #GoBackModi scan were posted by 7,394 accounts,” it added.
The research concluded that, “By any measure, #TNwelcomesModi saw a much more aggressive attempt to make the hashtag trend from a much smaller user base”.