Featured
The pattern of pro-Palestinian posts is consistent with a prolonged social movement, the research suggests, while the pattern of pro-Israeli posts is typical of what follows a major news event.
With millions of users across the globe, including the Middle East, TikTok has become a popular source of information and commentary on the Israel-Hamas war.
But pro-Palestinian posts to the social media app significantly outnumber pro-Israeli posts and follow a very different pattern, new research from Northeastern University reveals.
The number of pro-Israeli content posted to TikTok has steadily declined since Hamas’ Oct. 7 attack on Israel, the research shows.
“There’s a lot of posting activity initially, and then there’s a gradual decline over time,” says Laura Edelson, an assistant professor at Northeastern’s Khoury College of Computer Sciences.
Sign up for NGN’s daily newsletter for news, discovery and analysis from around the world.
The opposite has happened with pro-Palestinian content.
“What we see with pro-Palestine posting activity — it grows organically over time, it culminates and then it has a symmetric decline,” says Edelson, a computer scientist who studies large online platforms. “That’s the kind of pattern that is more commonly associated with social movements.”
Edelson was granted access to the TikTok Research API, which allows independent and academic researchers who work for nonprofit institutions to access certain data to support their work.
She collected data on more than 280,000 TikTok posts from the United States that had specific hashtags related to the Israel-Hamas war. Examples include political statements such as #IStandWithIsrael or #SavePalestine, as well as more general apolitical tags like #Gaza or #Israel.
The data was gathered from 12 three-day windows — beginning Oct. 7 to Oct. 9, 2023, and ending Jan. 26 to Jan. 29, 2024.
Edelson examined the data in three major ways.
First was the number of posts. There were 170,430 pro-Palestinian posts, 8,843 pro-Israeli posts and 101,706 neutral or general posts.
Edelson also looked at the posts’ page views — there were 236 million views for pro-Palestinian posts, 14 million for pro-Israeli posts and 492 million views for neutral or general posts.
Finally, Edelson compared whether the number of posts and the post views were proportionate — this enabled Edelson to conclude whether TikTok was amplifying certain types of posts.
“It’s not enough to look at content,” Edelson says. “Big differences in how people experience content come from differences in amplification too.”
Content was amplified on both sides, the research suggests.
“There’s periods of time when TikTok is disproportionately amplifying pro-Palestine content, and there’s times when it’s disproportionately amplifying pro-Israel content,” Edelson says. “When you sum up everything over the entire study period, they amplify those two things equally, but it changes over time, initially.”
Edelson thus divides the data into three “phases.”In the first phase, the 3½ weeks following the Oct. 7 Hamas attack, neutral or general content is both most posted and most seen.
Edelson says the high number of general posts and views per general post is “fairly similar” to what happens on the platform following any major news event. She notes that TikTok is unique among social media platforms. Rather than amplifying the most extreme voices, it is “majoritarian,” she says.
“TikTok wants to find the most popular thing and then show that most popular thing as widely as possible,” Edelson says.
She also notes that neutral or general content is predominantly higher in quality — in terms of production value — which means it is more likely to be amplified by TikTok.
Pro-Israeli content follows the same pattern as general news: it is highest in this phase as well, but it steadily declines after the first week.
Pro-Palestinian content, on the other hand, jumps significantly in the second week and continues to grow steadily.
But things begin to change on Oct. 27 when the number page views on pro-Israel posts skyrockets — 2,555 views per post compared to 336 views per post previously.
Edelson says she doesn’t know why this occurred. But it lasts through the return of some Israeli hostages during a ceasefire that began Nov. 24 and into the first half of December.
The final phase begins on Dec. 15 when views per post for all categories fall dramatically.
“Interest in topics declines over time, that’s very normal,” Edelson says. “But the speed of the fall off is striking and not well explained by other events.”