While you are out there looking at data, maybe you can keep an eye out for psychology warfare tactics, designed to destabilize countries. I listened to a really interesting podcast where they discuss how psychological warfare is affecting the Israel/Palestine war but got into the origins of it in the US/Vietnam war as well as Russia’s influence in it during the Cold War.
Not sure if anyone is doing a deep dive into this but it could be we are seeing the fruits of such things in the US and the West at large.
While interesting, I’d be cautious about applying your interpretations of the results in such a manner because there are a lot of assumptions and variables at play, many combinations of which could produce similar results, none of which were controlled for.
Content itself.
An example variable is the fact that YouTube, instagram, and TikTok are quite different ages. There may be substantially different core content bias as the result of simply having more time to accumulate content which creates the profiles you’ve seen. As an example, I would look at age of results. If most negative-China results on YouTube or Instagram are older than TikTok itself consider filtering results so that you only look at the same time window of content and see if you have the same distribution. You may look for total content by bias - negative China content vs positive vs neutral vs redirect etc. If YouTube simply has more negative China content in the same time segment as TikTok then that’s a reasonable explanation of results.
Engagement algorithms.
Then, there is monetization bias. Social media tend to bias attention to the most surprising information (surprising, shocking, dismaying, upsetting) because users engage longer and spread results more with those qualities, as is well understood. TikTok may not be as good at funneling irritating content as YouTube and Instagram. It may produce more randomized (irrelevant) content for new users to profile algorithms over time of what they respond to. Negative China content may provide lower total engagement over time so is deprioritized. There are many possible explanations besides the obvious. I would see if the bias pattern is the same for different ages, different length of usage and click-through by age etc.
Country.
You assume this effect is due to the human rights abuses location in China. What about human rights abuses elsewhere in the world? For instance Myanmar, Afghanistan, Syria, North Korea, Iran, Eritrea, Venezuela, etc. Are their profiles different than China across platforms? Then ask a different question, perhaps look at counties today which may have had past human rights abuses - Germany and ask about Jewish genocide, forced labor and, confiscatory laws… ask about the US and Japanese internment camps - and see how the question about human rights abuses varies by country, and time.
Advertisement.
TikTok may accumulate much more positive Chinese travel posting and advertising for the fairly neutral reason that it is a Chinese company, and advertisers may believe (not irrationally) that positioning China (positioning is the sum of all positive material which may induce sales) on a Chinese site for Chinese travel and goods is more productive than advertising on a US Site for Chinese travel and goods if they have equivalent foreign usage.
I would change the word China to other counties and repeat your sample analysis, and in sample sets see the age of content to assess sizes of similar-age results; and test how profiling is done to sell for advertising - that may yield a much much stronger case for your thesis, or it may show that other variables can account for the effect, and cannot be eliminated (yet) as a cause.
My name is Greg and I'm the executive producer for Paula Wright's (@SexyIsntSexist ) 'Vane Tempest Podcast'
Paula mentioned that you two had spoken previously about having a conversation together on her show. I was hoping that I could help schedule that with you.
Twitter is the number one performing platform in terms of video virality. Basically, a video uploaded to Twitter will likely cross over to appear on other platforms, such as Instagram, whereas a video posted on Instagram will likely only ever appear on Instagram. TikTok is the number two platform for cross-platform virality. You will see more TikTok videos on Instagram than you will ever find Instagram videos on TikTok. This is in fact why the progressive left was so vociferously opposed to congressional action against TikTok. Elon Musk had just purchased their number one platform, so losing their second choice was seen as catastrphic. (Source: I have left the left but still monitor some webinars.) Hope this helps contextualize the issue.
Way to go Lee! With this 'feedback' from reviewer CCP, you have definitely arrived. Now for extra points, try to get this published in a journal and see if you can dig out some ChiCom sympathizers. But of course, be careful where you travel, what documents/gifts you accept, and probably keep a credible chaperone with you at all times.
Danit is sweating bullets to get a peer review-ready version of the report for submission to a journal that is sponsoring a special issue on threats to democracy. Can never tell about reviewers (for any paper!), but the special issue editor studies leftwing authoritarianism, so, no promises, but I am cautiously optimistic. The due date for submissions was ... a year ago. But she has already given us a "get out of being excluded just because you are a year late" card.
Maybe. I am pretty sure folks at NCRI are working on this. Whether I get involved remains to be seen. They can and usually do pure social media analysis stuff without me (I do occasionally collaborate with them even on those when they want some heavy psych explanations). They always bring me in if there is a conventional social science study involved, as here (the survey).
Awesome job!!
While you are out there looking at data, maybe you can keep an eye out for psychology warfare tactics, designed to destabilize countries. I listened to a really interesting podcast where they discuss how psychological warfare is affecting the Israel/Palestine war but got into the origins of it in the US/Vietnam war as well as Russia’s influence in it during the Cold War.
Not sure if anyone is doing a deep dive into this but it could be we are seeing the fruits of such things in the US and the West at large.
Here’s a link to the podcast…
https://youtu.be/dumtmCnrIKU?si=WYyZykJb9d9wHIeo
Now, I am jealous!
While interesting, I’d be cautious about applying your interpretations of the results in such a manner because there are a lot of assumptions and variables at play, many combinations of which could produce similar results, none of which were controlled for.
Content itself.
An example variable is the fact that YouTube, instagram, and TikTok are quite different ages. There may be substantially different core content bias as the result of simply having more time to accumulate content which creates the profiles you’ve seen. As an example, I would look at age of results. If most negative-China results on YouTube or Instagram are older than TikTok itself consider filtering results so that you only look at the same time window of content and see if you have the same distribution. You may look for total content by bias - negative China content vs positive vs neutral vs redirect etc. If YouTube simply has more negative China content in the same time segment as TikTok then that’s a reasonable explanation of results.
Engagement algorithms.
Then, there is monetization bias. Social media tend to bias attention to the most surprising information (surprising, shocking, dismaying, upsetting) because users engage longer and spread results more with those qualities, as is well understood. TikTok may not be as good at funneling irritating content as YouTube and Instagram. It may produce more randomized (irrelevant) content for new users to profile algorithms over time of what they respond to. Negative China content may provide lower total engagement over time so is deprioritized. There are many possible explanations besides the obvious. I would see if the bias pattern is the same for different ages, different length of usage and click-through by age etc.
Country.
You assume this effect is due to the human rights abuses location in China. What about human rights abuses elsewhere in the world? For instance Myanmar, Afghanistan, Syria, North Korea, Iran, Eritrea, Venezuela, etc. Are their profiles different than China across platforms? Then ask a different question, perhaps look at counties today which may have had past human rights abuses - Germany and ask about Jewish genocide, forced labor and, confiscatory laws… ask about the US and Japanese internment camps - and see how the question about human rights abuses varies by country, and time.
Advertisement.
TikTok may accumulate much more positive Chinese travel posting and advertising for the fairly neutral reason that it is a Chinese company, and advertisers may believe (not irrationally) that positioning China (positioning is the sum of all positive material which may induce sales) on a Chinese site for Chinese travel and goods is more productive than advertising on a US Site for Chinese travel and goods if they have equivalent foreign usage.
I would change the word China to other counties and repeat your sample analysis, and in sample sets see the age of content to assess sizes of similar-age results; and test how profiling is done to sell for advertising - that may yield a much much stronger case for your thesis, or it may show that other variables can account for the effect, and cannot be eliminated (yet) as a cause.
Hey Lee ,
My name is Greg and I'm the executive producer for Paula Wright's (@SexyIsntSexist ) 'Vane Tempest Podcast'
Paula mentioned that you two had spoken previously about having a conversation together on her show. I was hoping that I could help schedule that with you.
Email me at thevanetempest@gmail.com and hopefully we can get something arranged. Cheers
Congratulations, this is a badge of honor!
Twitter is the number one performing platform in terms of video virality. Basically, a video uploaded to Twitter will likely cross over to appear on other platforms, such as Instagram, whereas a video posted on Instagram will likely only ever appear on Instagram. TikTok is the number two platform for cross-platform virality. You will see more TikTok videos on Instagram than you will ever find Instagram videos on TikTok. This is in fact why the progressive left was so vociferously opposed to congressional action against TikTok. Elon Musk had just purchased their number one platform, so losing their second choice was seen as catastrphic. (Source: I have left the left but still monitor some webinars.) Hope this helps contextualize the issue.
Way to go Lee! With this 'feedback' from reviewer CCP, you have definitely arrived. Now for extra points, try to get this published in a journal and see if you can dig out some ChiCom sympathizers. But of course, be careful where you travel, what documents/gifts you accept, and probably keep a credible chaperone with you at all times.
Danit is sweating bullets to get a peer review-ready version of the report for submission to a journal that is sponsoring a special issue on threats to democracy. Can never tell about reviewers (for any paper!), but the special issue editor studies leftwing authoritarianism, so, no promises, but I am cautiously optimistic. The due date for submissions was ... a year ago. But she has already given us a "get out of being excluded just because you are a year late" card.
Will you do Russia next? It will probably net you another denouncement to your collection.
Maybe. I am pretty sure folks at NCRI are working on this. Whether I get involved remains to be seen. They can and usually do pure social media analysis stuff without me (I do occasionally collaborate with them even on those when they want some heavy psych explanations). They always bring me in if there is a conventional social science study involved, as here (the survey).