Did New York Times manipulate data to create #FakeNews about @elonmusk's Twitter takeover? A thread.
On Dec 2, New York Times led with the headline "Hate Speech’s Rise on Twitter Is Unprecedented, Researchers Find."
nytimes.com
On Dec 2, New York Times led with the headline "Hate Speech’s Rise on Twitter Is Unprecedented, Researchers Find."
nytimes.com
In the data it cites, it claims "slurs against Black Americans" jumped from 1,282 times a day.. to 3,876 times a day.
It makes similar claims about anti-Semitic and other hate categories.
Buried in the text however is a small detail...
It makes similar claims about anti-Semitic and other hate categories.
Buried in the text however is a small detail...
What is the period that is being measured?
The period, it turns out, is *two weeks* after Elon acquired the site.
If the claim is that the change happened due to Elon's *policies* the effect would take some time to show up. Was 2 weeks the right measure? Read on.
The period, it turns out, is *two weeks* after Elon acquired the site.
If the claim is that the change happened due to Elon's *policies* the effect would take some time to show up. Was 2 weeks the right measure? Read on.
Elon himself had tweeted that there was a spike of hate speech soon after he took over. Clearly this was not due to policy, but due to unknown actors (1500 per his account) with motivations we could speculate.
Now, here's @nytimes' manipulation ...
Now, here's @nytimes' manipulation ...
Elon had put out a tweet that showed the hate speech spike.
This doesn't take any "researchers." A high school student could clip the chart and claim the average of hate speech was higher. This is exactly what NYT does.
It cherry picks the 2-week window that include the spike.
This doesn't take any "researchers." A high school student could clip the chart and claim the average of hate speech was higher. This is exactly what NYT does.
It cherry picks the 2-week window that include the spike.
This is manipulative "research"
1. The spike right after Elon's takeover can hardly be attributed to his policies. Even at startup speed policies take time for an impact.
2. Unknown actors cause the spike.
3. The average after 2 weeks is *lower* than before the acquisition.
1. The spike right after Elon's takeover can hardly be attributed to his policies. Even at startup speed policies take time for an impact.
2. Unknown actors cause the spike.
3. The average after 2 weeks is *lower* than before the acquisition.
However, NY Times steam rolls on with its hatchet job.
Firstly, by citing and dismissing Elon's claim of hate speech being lower but negating it by saying "he did not provide underlying numbers or details of how he was measuring hate speech."
Well, neither did NYT Researchers.
Firstly, by citing and dismissing Elon's claim of hate speech being lower but negating it by saying "he did not provide underlying numbers or details of how he was measuring hate speech."
Well, neither did NYT Researchers.
They haven’t. But they do give the time period which includes the spike.
Loading suggestions...