Over the past several years, Soviet immigrants in the U.S. were divided over Donald Trump. Scarred by forced collectivism under the USSR, many Soviets saw democratic values as an existential threat to their rights and individualism. Others saw Trump as a triggering reminder of totalitarianism under Soviet rule. Leading up to Trump’s run for reelection, this chasm was particularly evident on Facebook, where groups as vastly different as “Anti Trump Soviet Immigrants” and “Russian Speaking Americans for Trump” thrived. This investigation tracked these ideologies and how they played out on the platform.
Most of the right-wing groups I tracked for the project were taken down by Facebook as I continued reporting this story (I sent Facebook a no-surprises letter detailing my findings, but did not receive a response). The left-wing groups remain, as they do not violate the platform’s policies. The story was shared across the groups and based on feedback I received at least from the groups still running, members felt seen for the first time amidst this complicated, nuanced culture war.
I relied on my Russian language skills (I am bilingual) to join groups and connect with members and adminstrators across the political spectrum. I used CrowdTangle to analyze key words in posts and the types of external links being shared in these groups and any associated pages ahead of and after the election. (For private groups, I did this by hand.) I wanted to track whether there was any evidence of coordinated inauthentic behavior on the platform among these groups, so I looked for connections between the webpages (mostly news sites) being shared. To do that, I used digital verification tools like WhoIs to track domain registrations, as well as the source code to track potential shared Google codes across sites. I created spreadsheets of the most active users within each group and what links they posted and when, and where else on Facebook those links were also shared. I created pivot tables to see things like what user shared the most links and in what groups. Through that, I could plot out connections, if they existed, between the posters, webpages and groups. I also had to verify the profile of each group member I reached out to, so I ran their profile images through TinEye and examined their profiles for clues as to when the account may have been set up. I also ran my own survey to gather data on the community in terms of political leanings, social media use and news preferences.
What was the hardest part of this project?
One challenge I faced was getting sources in right-wing groups to speak with me. I was ultimately able to get a few on the record by approaching them in an op and listening to them intently. I did my best not to form conclusions or assumptions before I had done my reporting, and I believe that all my sources felt heard through that. Another challenge was weighing whether to include the comments of one source who unexpectedly passed away during the project (in the end, I included his story). It was tricky to figure out how to structure the story. Should I include more sources? Should the piece be cultural criticism or a data-driven feature? In the end, it was a bit of both. I had to rewrite the draft for weeks until I got at what was really important and worth including. After my reporting, it was difficult mentally to deal with several trolls who called me a “communist” and “ungrateful, spoiled leftie.” I had to disable comments on my Facebook page. What was most challenging was writing the story post-election after Trump had lost and pitching it to other publications. In hindsight, I would have started the project sooner and published it sooner, so that it could be more timely and have maximum impact. I wish that another publication would have accepted the story to give it a bigger platform, but after Trump lost, lots of publications expressed fatigue around publishing election-related stories, particularly ones that centered around Trump. I was nevertheless able to make the story relevant by showing that for most Soviet immigrants, because of their lived experiences under the Soviet Union, Trump was more than a public figure; he represented a way of life. That is not going to disappear because he is gone from
What can others learn from this project?
Start these types of digital forensic investigations early, and track them over time. Findings are very powerful that way. Even after the election passed, the groups remained active for months as concerns about the election being stolen swept through the Right. I was able to include that in my story as additional evidence of the groups’ ideologies. I recommend taking lots of screenshots, as you never know when a group might get taken down or frozen (as did happen). Take precautions to protect your identity from those who are inherently distrustful of media or women. That could mean setting up a separate social media profile for this project only, using an alternative phone number, not meeting in-person in a private environment, etc.
I think the story itself could serve as a powerful learning tool for other reporters looking to figure out the best way to tell this type of story. By mixing polarizing political commentary of these groups with hard data, the story comes across as both fair to all perspectives and committed to rigorous fact-checking and research standards. I believe all reporting should begin with an open mind, and the final product should reflect that you took that approach while also being accompanied by context that shows you got the bigger picture.