The multi-part series is one of the most ambitious reporting efforts on the status of nuclear weapons in decades.
On March 4, 2024
The New York Times
published a new series on
the threat of nuclear weapons
Graphic depictions of an esclating nuclear exchange were portrayed on The Times website and social media.
Especially on Instagram
Drawing hopeless responses online...
THIS IS NOT @laurenlevatocoyne
Who at @nytimes thought it was a good idea to share this? @vasu_sojitra
The Audience Team wanted to answer those readers asking "What should we do with this information?" and invited Emma Pike, a nuclear disarmament campaigner with a large following on TikTok to create a short video highlighting reasons to feel hopeful despite the nuclear threat.
The Video garnered a largely positive response on TikTok and Instagram.
Then the team posted the video to Twitter/X where the comments quickly took a different turn...
A research project examining the intense reaction to nuclear weapons media and discourse.
This is Voices at the Brink
500
Comments
Why did one seemingly innocuous video drive millions of views and unleash an online firestorm with thousands of angry and vicious comments?
NYTOpinion
10M
Views
2.2k
Shares
1.5k
likes
1.2k
Bookmarks
100 Twitter/X Posts Manually
We Examined
Three Questions Guided Our Research
1.
2.
3.
300 top responses and their profiles using analytic tools
4 levels of responses and shares of top posts
A close look at the top replies and threads reveals a set of disturbing patterns.
36% Attacked her experience and expertise.
22% Contained explicit sexism.
28% Called Gen Z naive including one Swedish post suggesting
"We must retroatcitvely abort all of Gen Z."
Many responses however did contain substanative though condescending commentary that expressed strongly held opinions on the topic centering on a set of key themes.
Analysis of top threads including direct replies to the original post and repost threads shared five common areas of argument with the concept that nuclear disarmament is achievable and desirable.
Content analysis showed that many posts not only shared a tone that can be described as "snark," but also were connected by a common set of arguments that were dismissive of international law and treaty-based approach, and also viewed nuclear weapons as a source of stability. Many brought up Ukraine as an example of how non-nuclear weapons states are at risk of attack.
The majority of comments with substantive arguments suggested Ukraine was made vulnerable by dismantling and transferring nuclear weapons to Russia between 1992 and 2001. The commentary was connected to the dogma that nuclear weapons provide protection from invasion and that deterrence works to limit or block conflict.
This argument ignores that the potential of proliferation closer to Russia’s borders is a stated cause for the invasion of Ukraine. Russia perceives the eastward movement of NATO as a threat that brings nuclear encroachment through NATO’s nuclear umbrella. Nuclear armed states sharing borders does not increase security, it increases the urgency for one side to control the other and potential for escalation from diplomatic disagreements or low-intensity conflict to nuclear war.
The suggestion that the majority of nations in the world “choose” not to possess nuclear weapons was a major trigger for many respondents. This was seen as an example of naïveté and a lack of understanding of great power politics. These commenters simply could not accept that nations would choose not to develop and deploy nuclear arms if they could.
The short length of the video did not allow for expanding on the different forms these “choices” take. How does a nation choose to eschew nuclear weapons? Who actually makes the choice? A head of state, parliament, the citizenry?
More stories detailing the various ways nuclear weapons have been rejected are needed to illuminate the role the diplomatic process and civil society play in securing disarmament and non-proliferation. Illuminating the ways nuclear weapons have been limited to a fraction of nations also highlights that treaties have value and disarmament is possible — responding to two of the other major critiques.
It is important to note that Twitter/X now favors users with premium subscriptions Those willing to pay receive a bigger boost for their posts. This creates an atmosphere where users who most align with Elon Musk's vision self-select on the platform and are overrepresented due to quirks of the algorithm.
Twitter/X, like most social networks, gives each user a unique experience. Premium users may get the highest "boost" when replying to a post, but other factors influence what a user can see including who they follow and which posts they interact with. Conducting analysis and recreating the timeline is constrained by the user-specific nature of Twitter/X and the minimal API. For example, only 250-300 replies to a post will load rendering viewing all replies impossible. However, we looked for common themes in replies and highly engaged posts in order to reconstruct a timeline of key momnets. From this we gained insight into why the post went viral and the profile of users involved in "dogpiling" the post with negative replies and "Quote Tweets."
Quote posts by those working in the foreign policy and nuclear weapons space pushed the original post into communities that follow "experts" on military-related issues.
Communities with specific interest in the military with a focus on weapons and technology next picked up the post.
Followers of that account, many of whom have the Ukranian flag in their profiles, then Retweeted the post gaining visibility in their communities.
This last push brought the original post into the English-speaking pro-Ukraine online community who Quote Tweeted the post with their own commentary at an exceptionally high rate.
These connections lead us back to our original questions
1.
2.
3.
The "curtain-raiser" post for the series has been pinned on the Times Opinion profile since March and has 165K views and 78 replies.
Posts promoting other articles in the series also have low engagement.
The Times does not make any readership data available to third-parties, but we can compare other posts on the @nytopinion Twitter/X account to answer this question.
Was the engagement significantly higher than usual?
If we broaden to look at all posts for the period where data is available (2019), we find that in the past five years there have been 21 posts with 2K comments or more.
The most replied to post was a video essay questioning American greatness.
It is clear that the reason to hope post saw an unusually high level of engagement in terms of views, replies and Retweets.
Was the high engagement due to manipulation?
We looked for key indicators to understand if the post was boosted through means other than the standard Twitter/X algorithm and if there were directed attacks or coordination. We consulted with Junkipedia and gained access to their analysis tools.
Junkipedia is a tool created by the Algorithmic Transparency Institute, a project of the National Conference on Citizenship, to provide a collaborative platform for collecting, monitoring, analyzing, and responding to problematic content online.
Through discussions with the Junkipedia team and in utilizing the platform, we found no clear evidence of manipulation. No indications of bots were found. However, artificial manipulation cannot be completely ruled out due to Twitter/X limitations. Researchers cannot view comments and Retweets beyond 200-300 posts, which means bots could be operating at the level of 500-2800 replies without notice. Even so, it is our opinion, shared by Junkipedia, that bot usage is highly unlikely due to the fact that we saw no indication of bots in posts reviewed, and there is a plausible explanation for the high level of engagement.
1.
2.
This leaves the final question: who exactly are these users and why did they engage so fully with this particular post on TwitterX about nuclear weapons?
While others see their role as online warriors
Some have gone so far as to join the Ukranian military to fight Russia.
The post attracted a group of Twitter/X users who see the defense of Ukraine as their "Greatest Generation" moment.
Meet the Man on a Mission
They see themselves not as a primrily online group, but as Warriors
Participating in the battle to save democracy, which makes many fiercely anti-Trump.
And even strongly loyal to Joe Biden and the Democrats.
NAFO members join through making donations to approved Ukraine defense foundations. The group traces its founding to a 2022 Twitter/X "dogpile" on a Russian diplomat where they posted over 60,000 comments on the platform and caused the diplomat to suspend his account.
North Atlantic Fella Oganization
Some commenters are part of an organization that sprung up following the invasion of Ukraine and describes their mission as fundraising for "Ukraine's defenders" and combatting disinformation on social media spaces.
Many promote the Come Back Alive foundation in their profiles, a crowdfunding effort providing military hardware and training for Ukranian forces.
A pro-nuclear disarament post presented a stark to challenge to this group's worldview -- one where disarament is weakness and military superiority is the force that can stop the Russian menace. Without realizing it, The Times was issuing a challenge to these "Warriors" by posting the video on Twitter/X.
This cohort is highly motivated and a force online. It is tempting to label this instance a fluke. Emma Pike's content is calibrated for a TikTok audience and performs well on there and Instagram. There is certainly a large segment of the population who is open to positive and accesible messages. We could write this moment off as an instance where an audience was served content that was not intended for them and responded harshly.
However, this group's intense committment to causes could make them vital allies for disaramemnt. We must simply understand that the type of content that moves them to action is often counter to the content the movement often is encouraged to produce. This unique cohort is overly active online, overly educated, and yes, mostly male.
A large percentage of searches about nuclear weapons online comes from this group and at least 65% of searches about nuclear weapons come from those who identify as male (BuzzSumo).
The majority are aged 36-50 with disposable income and are politically active.
This cohort is incredibly overrepresented in education with 52% possessing a post-graduate degree. In the U.S. only approximately 5% of the population has attained a professional or doctoral degree.
Conclusion
It is time to revisit stories of disarmament success and frame those as tales of epic victories against impossible odds. We should also be aware that some cohorts may not be swayed by disarmament messaging, but there can be a valuable effect on other observers and telling stories of anti-nuclear victories is powerful on its merits.
- Show that disarmament has worked and treaties have made a difference.
- Focus on the ways it could have been worse rather than the failures of nuclear-armed states to do better.
- Highlight stories of complexity and hidden connections.
- Celebrate disarmament heroes.
- Use humor.
- Invite them to participate.
Though they posses much general knowledge, and a unique depth of military information, there is much that this group and others do not know about nuclear disarmament. For those who came of age after the Cold War, the success of disarament campaigns to halt the march towards nuclear annihilation remain largely uknown or misunderstood. They see the world shaped in power and by powerful individuals, not by collective action. We need heroic tales of disarmament that frame the current nuclear order as a victory against impossible odds. The existence of humanity is not thanks to Kennedy or Krushev, Reagan or Gorbachev, but to disarmament campaigners. They faced off against the great empires and won.
Along with the grand victories comes more complexity. Campaigners and communicators in disarament and other disciplines are encouraged to simplify messaging and the pathway to success. We are not asking for something difficult and contentious, but something obvious. This is sensible given the absurd complexity of deterrence theory and is likely the best way to reach broad audiences. However, this movement, like most today, must create connections of depth and meaning. To do so requires asking more of supporters, not less and highlighting the obstacles to success. The North Atlantic Fella Organization requires members to donate to approved charities and then send a screenshot of the donation to the groups leadership through email or a Twitter/X direct message. The multiple steps and donation requirement are in stark contrast to campaigning best practices that create the lowest bar possible to action and streamlined user-experiences. But asking more creates deeper connections and a stronger bond of belonging.
The final recommendation may seem counterintuitive, but there is truth that individuals will act and speak in a manner online that they would not in-person, and that online groups forge incidiary environments. Inviting individuals to in-person events where they can participate -- feel heard, but also hear out others could make inroads with these types of groups. They seek legitimacy and belonging...ultimately, to be told that they have value and what they do has impact. Providing new venues to achieve that may pay off in unique ways.
These people need a mission; give them a better one.
Various data points were taken into account by our team and we used multiple tools to analyze social media posts, search trends, and articles. Though we sought specific data on the series, The Times does not provide reading and viewing data publicly or privately, excepting when required by advertising agreements.
Since our focus was the response rather than the reach of these articles, we used available social media analytics for the pieces in the series.
Twitter/X data was analyzed using several tools and through manual input and analysis.
Notes on Methodology
Twitter/X
Twitter analytics and user profiles.
Twitter WebsiteKeyhole
Twitter analytics and user profiles, tweet reach.
Keyhole.coMeltwater
Twitter analytics, tweet reach, sentiment analysis, engagement rankings, media monitoring and historical trends.
Meltwater.comBuzzSumo
Search trends and user profiles, article social shares and engagement, TikTok analytics.
BuzzSumo.comNewsWhip
Article social shares and engagement.
Newswhip.comSparkToro
Search trends and user profiles.
SparkToro.comGoogle Trends
Search Trends
Trends.Google.comJunkipedia
Monitoring and analytics across multiple social networks, misinformation and coordination analysis.
Read more hereTools
Social media data is notoriously inscrutable and Twitter/X is even more difficult to catalogue since Elon Musk decided to cut the public API in May 2023. Analytic products use a combination of user sharing (often achieved through browser extensions), user direct input (such as authorizing the product's access to your own account), frontend data scraping, and proprietary algorithms to ascertain audience reach.
This research is further complicated by the fact the key accounts and websites being analyzed were not known until after the March 21 tweet. Most systems do not start monitoring until a paying user of the product requests a certain profile, site or topic is recorded.
In short, we cannot gain a perfect set of data related to our key questions. For example, the key post received over 3,000 comments, but Twitter will only show a maximum of 250 replies on its website at any one time. Facebook posts about a particular article could be subject to a multitude of privacy restriction settings.
The tools used do not give us a perfect view of the data and content, but a clear one.
When two analytic tools diverged on engagement data (likes, shares, etc) we attempted to view the post directly when possible. If not, we defaulted to the higher number as engagement statistics rarely suffer from false positives. If BuzzSumo listed a Times article as having 25 shares on Facebook and NewsWhip listed it as having 20, we defaulted to the higher number, 25. We assume that one service was able to discover more shares than the other, rather than that the data is inflated.
However, we took the opposite approach with Reach data and generally only used this statistic for benchmarking. Reach does not tell us the actual views or reads, but attempts to quantify the potential reach and uses the highest number possible. That an article could reach 50 million people, does not mean it did reach them.