5 ms·

> It didn't just impact the individuals, it reduced the behavior site-wide. And? Even saying it reduced the behavior site-wide is not an effective measurement

> It didn't just impact the individuals, it reduced the behavior site-wide. And? Even saying it reduced the behavior site-wide is not an effective measurement to conclude that deplatforming works in reducing the prevalence of those views in society. Again, "Deplatforming X views from platform Y resulted in less of X on platform Y" is effectively stating the obvious. To demonstrate the effectiveness of deplatforming, one would have to determine whether deplatforming actually results in fewer people believing in the views that are being deplatformed. I have not encountered any instance of this occurring. Ask yourself this: when you are banned from a forum for views you believe in, or you witness someone banned for views you agree with do you tend to turn around agree with the censor? Or do you become more enthusiastic for that view and lose respect for the censor > This is a pretty clear correlation/causality confusion. If factor A is on the rise and triggers reaction B, you cannot use the rise of B to prove that B caused A. Now, your narrative may be correct, but the narrative story you've provided does not demonstrate it. We're seeing trust in media and tech companies plummet. While the fact that a rise in extremist views is correlated with increases in deplatforming is not hard evidence of causation, it's extremely difficult to claim that deplatforming works to reduce said views in the face of that positive correlation between the two. That's trying to claim a causal relationship in the face of evidence of the opposite correlation. In other words, if we see A rise alongside B it is indeed jumping the gun to say that A certainly causes B. But it's even more dubious to say that A reduces B in the face of that correlation.

darawk 8y ago
> I have not encountered any instance of this occurring. How could you witness such a thing occurring? This seems like an unreasonable evidentiary standard. > We're seeing trust in media and tech companies plummet. While the fact that a rise in extremist views is correlated with increases in deplatforming is not hard evidence of causation, it's extremely difficult to claim that deplatforming works to reduce said views in the face of that positive correlation between the two. That's trying to claim a causal relationship in the face of evidence of the opposite correlation. Do HIV drugs cause HIV? Do civil rights movements cause racism? De-platforming is a treatment. Of course it's going to co-occur with the thing it's attempting to treat. This is evidence of nothing at all. What you need to do, and what the studies I reference did do, is examine individual communities pre and post treatment. That is how you start to get at causality. The analysis is imperfect, to be sure, but it's a lot better than looking at simple correlation.
- manfredo 8y ago
  > Do HIV drugs cause HIV? How is immunodeficiency treatment at all related to deplatforming? Viruses aren't thinking human beings. > Do civil rights movements cause racism? The Civil Rights movement did not enagage in deplatforming. Many of them explicitly acknowledged that their opponents also deserve the ability to speak. It was often the civil Rights movement itself that was subject to deplatforming. > Of course it's going to co-occur with the thing it's attempting to treat. This is evidence of nothing at all. The rise of deplatforming preceded the rise of the Alt Right by about a year or two. They didn't always co-occur, one preceded the other. Their continued co-occurrence suggests that the implementation of deplatforming either 1. has no effect on that brand of extremist, or 2. maybe even causes it. This is not consistent with treating a disease. Usually, a disease is present sometime before treatment is administered. Then as treatment is administered the symptoms are reduced, if the treatment is successful. This is not what we are witnessing with the relationship between deplatforming and the brand of right wing extremism we've been seeing lately. > What you need to do, and what the studies I reference did do, is examine individual communities pre and post treatment. Limiting measurement to individual communities is not a good way to measur it's overall effect. Again pointing out the fact that when a community deplatforms a certain view that view is no longer present is pointing out an obvious consequence. Of course the platform sees a reduction in the view that was deplatformed. That's basically just restating the definition of deplatforming: kicking a person or group off the platform. If you want to measure the effect of deplatforming on society, then the analysis has to be society-wide. Otherwise one is effectively just building a bubble of the communities that do engage in deplatforming, and burying their head with respect to it's impact on the rest of society.
  - darawk 8y ago
    > How is immunodeficiency treatment at all related to deplatforming? Viruses aren't thinking human beings. > The Civil Rights movement did not enagage in deplatforming. Many of them explicitly acknowledged that their opponents also deserve the ability to speak. It was often the civil Rights movement itself that was subject to deplatforming. My point is that the type of reasoning you used here would lead you to draw both of those conclusions. > The rise of deplatforming preceded the rise of the Alt Right by about a year or two. They didn't always co-occur, one preceded the other. Their continued co-occurrence suggests that the implementation of deplatforming either 1. has no effect on that brand of extremist, or 2. maybe even causes it. That's an a-factual statement. When did the alt right "rise"? Was it when Mencius Moldbug started writing Unqualified Reservations in 2007? When Richard Spencer joined the National Policy Institute in 2011? During Gamergate in 2014? Similarly, when did de-platforming 'start'? Was it when people first started protesting The Bell Curve when it was published in 1994? Was it when British National Union of Students adopted a no-platform policy? The point is, neither of these events have a well-defined starting point, so any claim of one preceding the other is silly, and has no basis in fact. > Limiting measurement to individual communities is not a good way to measur it's overall effect. Again pointing out the fact that when a community deplatforms a certain view that view is no longer present is pointing out an obvious consequence. Of course the platform sees a reduction in the view that was deplatformed. That's basically just restating the definition of deplatforming: kicking a person or group off the platform. Your objection suggests a specific causal model, though. You're right that kicking users off of the platform will tautologically reduce the content. However, what if you didn't kick people off the platform? What if instead, as in the example I linked, you banned the sub-communities dedicated to advocacy of the proscribed topics? The people stay, the community goes. Then, you look at the level of the material in other sub-communities on the site. That is what those studies did, and that is why they demonstrate causality.

darawk 8y ago

> I have not encountered any instance of this occurring. How could you witness such a thing occurring? This seems like an unreasonable evidentiary standard. > We're seeing trust in media and tech companies plummet. While the fact that a rise in extremist views is correlated with increases in deplatforming is not hard evidence of causation, it's extremely difficult to claim that deplatforming works to reduce said views in the face of that positive correlation between the two. That's trying to claim a causal relationship in the face of evidence of the opposite correlation. Do HIV drugs cause HIV? Do civil rights movements cause racism? De-platforming is a treatment. Of course it's going to co-occur with the thing it's attempting to treat. This is evidence of nothing at all. What you need to do, and what the studies I reference did do, is examine individual communities pre and post treatment. That is how you start to get at causality. The analysis is imperfect, to be sure, but it's a lot better than looking at simple correlation.

manfredo 8y ago
> Do HIV drugs cause HIV? How is immunodeficiency treatment at all related to deplatforming? Viruses aren't thinking human beings. > Do civil rights movements cause racism? The Civil Rights movement did not enagage in deplatforming. Many of them explicitly acknowledged that their opponents also deserve the ability to speak. It was often the civil Rights movement itself that was subject to deplatforming. > Of course it's going to co-occur with the thing it's attempting to treat. This is evidence of nothing at all. The rise of deplatforming preceded the rise of the Alt Right by about a year or two. They didn't always co-occur, one preceded the other. Their continued co-occurrence suggests that the implementation of deplatforming either 1. has no effect on that brand of extremist, or 2. maybe even causes it. This is not consistent with treating a disease. Usually, a disease is present sometime before treatment is administered. Then as treatment is administered the symptoms are reduced, if the treatment is successful. This is not what we are witnessing with the relationship between deplatforming and the brand of right wing extremism we've been seeing lately. > What you need to do, and what the studies I reference did do, is examine individual communities pre and post treatment. Limiting measurement to individual communities is not a good way to measur it's overall effect. Again pointing out the fact that when a community deplatforms a certain view that view is no longer present is pointing out an obvious consequence. Of course the platform sees a reduction in the view that was deplatformed. That's basically just restating the definition of deplatforming: kicking a person or group off the platform. If you want to measure the effect of deplatforming on society, then the analysis has to be society-wide. Otherwise one is effectively just building a bubble of the communities that do engage in deplatforming, and burying their head with respect to it's impact on the rest of society.
- darawk 8y ago
  > How is immunodeficiency treatment at all related to deplatforming? Viruses aren't thinking human beings. > The Civil Rights movement did not enagage in deplatforming. Many of them explicitly acknowledged that their opponents also deserve the ability to speak. It was often the civil Rights movement itself that was subject to deplatforming. My point is that the type of reasoning you used here would lead you to draw both of those conclusions. > The rise of deplatforming preceded the rise of the Alt Right by about a year or two. They didn't always co-occur, one preceded the other. Their continued co-occurrence suggests that the implementation of deplatforming either 1. has no effect on that brand of extremist, or 2. maybe even causes it. That's an a-factual statement. When did the alt right "rise"? Was it when Mencius Moldbug started writing Unqualified Reservations in 2007? When Richard Spencer joined the National Policy Institute in 2011? During Gamergate in 2014? Similarly, when did de-platforming 'start'? Was it when people first started protesting The Bell Curve when it was published in 1994? Was it when British National Union of Students adopted a no-platform policy? The point is, neither of these events have a well-defined starting point, so any claim of one preceding the other is silly, and has no basis in fact. > Limiting measurement to individual communities is not a good way to measur it's overall effect. Again pointing out the fact that when a community deplatforms a certain view that view is no longer present is pointing out an obvious consequence. Of course the platform sees a reduction in the view that was deplatformed. That's basically just restating the definition of deplatforming: kicking a person or group off the platform. Your objection suggests a specific causal model, though. You're right that kicking users off of the platform will tautologically reduce the content. However, what if you didn't kick people off the platform? What if instead, as in the example I linked, you banned the sub-communities dedicated to advocacy of the proscribed topics? The people stay, the community goes. Then, you look at the level of the material in other sub-communities on the site. That is what those studies did, and that is why they demonstrate causality.