6 ms·

From folks making the correlation vs causation argument here, I would like to hear their rebuttal to the points the author already laid out in this article http

From folks making the correlation vs causation argument here, I would like to hear their rebuttal to the points the author already laid out in this article https://jonathanhaidt.substack.com/p/social-media-mental-illness-epidemic https://jonathanhaidt.substack.com/p/social-media-mental-ill..., especially the section on quasi experiments. Do we just think that because it is probably impossible to conduct pure double-blind RCTs in human populations, it is therefore impossible to reasonably attribute causality to anything on a social scale, and our efforts to understand and maybe even shape social outcomes are thus inherently pointless?

cauch 3y ago
My rebuttal would be: it is not impossible to attribute causality, and THIS IS THE PROBLEM. Right now, there are two identically possible situations (well, even more, but you get the jest): A) smartphone -> depression B) depression -> smartphone But the author does not do a good job in explaining why he choose A instead of B, except "it feels logical to me". It feels like a coin toss. (and sure, the author has some arguments, but they are arguably disputable, more than in other cases where the idea that there is causality ends up being adopted as a consensus) It does not mean it is impossible to reasonably attribute causality. It just means that the author does a bad job at the "reasonably" part.
- bhk 3y ago
  Or also.... C) X -> depression AND X -> smartphone First example, X could be "strongly needs to feel included"
  - cauch 3y ago
    yes, exactly
  - illiac786 3y ago
    Why then is there an uptick since 2012, for example? Your theory doesn't explain this, while the author's does. We can't just consider all theories equal because they are all in the realm of the possible. Some evidence, while not proof, strongly points toward social media being the cause or at least a very strong factor. I agree with multiple other commenters here, that banning phones in school would be a high reward chances / low risk decision. We're talking about LOTS of kids getting their mental health damaged, but in many comments I don't hear the sense of urgency, rather a detached a almost snobbish focus on correlation != causality. I think such answers are partially missing the point.
    - cauch 3y ago
      There are TONS of things that changed in the society around 2012. (for example, a specific and brand new phase in globalisation, in climate change crisis, in global loss of trust in the media, in disillusion in the capacity of changing elements that matters, in faster access of overwhelming communication, a more open communication about personal issues (gender, ...), ...) The "smartphone change" is the "what people who don't know sociology think a sociology change looks like". It is still a sociology change, but it's the "nose in the middle of the face". > I agree with multiple other commenters here, that banning phones in school would be a high reward chances / low risk decision. And, surprise, I'm 100% for the ban in school too. But we need to solve the problem, not hide it. The problem can be, for example, social media. Banning smartphone at school will not solve the depression crisis, because people will still have their fix on their laptop after school, and will just maintain the toxic social media relationship dynamic in "real life" in school. If you really care about the kids, you should understand why some people are so unhappy about having a conclusion as naive and simplistic as "let's just ban the smartphone and change nothing in the fundamental ways we are building a society, and I promise you, everything will magically be fine". (and again, it does not mean smartphones are not part of the equation. But we need to find the real cause, and this is why correlation != causality is important. Nobody cares about the theory, people raise this argument because they know this mistake will have bad consequences)
- naasking 3y ago
  > Right now, there are two identically possible situations (well, even more, but you get the jest): This is misleading. They are not "identically possible", which I assume to mean that you think they are "equally plausible". There is more evidence for smartphones and social media causing issues, because we don't see those declines until after the smartphone is introduced, regardless of age. Furthermore, other factors causing both depression and social media use would have to be cross-cultural because the issue is global. Other than social media, what factors do you know of are cross-cultural? Guess what, they eliminated most of them in prior articles. Haidt and co have published multiple articles going over the data and the possibilities, so I recommend reading them all before claiming the situation is so murky.
  - cauch 3y ago
    ("identically possible" as: as legitimately grounded. But of course, if you are convinced that people who are not convinced by your argument can only be dishonest, of course, for you, they are not identically possible) > because we don't see those declines until after the smartphone is introduced, regardless of age. Again, there are a lot of identically possible explanation: the whole society has changed and became globally more anxiogenic, the measurement of the decline may be biased as people are nowadays more incline to be honest with their mental health or less incline to "shake it off", or ... > Furthermore, other factors causing both depression and social media use would have to be cross-cultural because the issue is global. Other than social media, what factors do you know of are cross-cultural? Guess what, they eliminated most of them in prior articles. What do you mean "cross-cultural"? Isn't the study showing that it's mainly a US problem? And how can you have a cross-cultural effect when the usage of the smartphone amongst young people is so linked to cultural trends that are very different from country to country? I have the impression the argumentation jumps from one to the others according to what benefits the authors: if it happens at two places, they say "see, it proves there is no other sources", if it does not happen in other places, they say "sure, but it's just because the smartphone usage is socially different in these countries". > Haidt and co have published multiple articles going over the data and the possibilities, so I recommend reading them all before claiming the situation is so murky. And other experts, with similar credentials than Haidt, have been critical of the work. My position is that we should not jump to the conclusion: the jury is still out. Your conclusion seems to be "I choose to trust Haidt and to distrust equivalent scientists, just because Haidt's explanation seems more obvious to me".
- cpsempek 3y ago
  I commented this elsewhere, but it's worth repeating because actually reading the author's arguments (granted, in other writings, not OP) will allow you to see he's considered this point and is will to accept that causation works the other way around. It's a fair point. But what explains the uptick in depression and mental health issues starting around 2012, disproportionately impacting pre-teen girls and not contained to any geo. The author's entire point is that social media is the only explanation that that has been proposed [1]. Moreover, it's not a far-fetched explanation. He very aware that correlation does not provide causation, and that it could be the other way around. However, no one (according to him) has offered up a theory which explains the data like the social media theory does. [1] https://jonathanhaidt.substack.com/p/social-media-mental-illness-epidemic https://jonathanhaidt.substack.com/p/social-media-mental-ill...
  - cauch 3y ago
    And these arguments have been heard, but they are still not convincing. I don't know what is the reality, maybe indeed the smartphone is causing the problem. But the author is just jumping to the conclusion without proper scientifically caution. There are a lot of other hypothesis that explain exactly the same observations, and that are not far-fetched (as you say, there is none _according to the author_, who then list all of these hypotheses in his rebutal). In his article, he goes through a lot of them but don't address them properly. For example, the more recent social anxiety about climate change is just "I don't think so". Personally, if I have to bet, I think it is social media. But then, 1) I will not claim "it is social media", just "it may be social media", 2) it would be stupid of me to then isolate smartphone: if it is social media, then forbidding smartphone but not the laptops would be totally useless (and if the argument is turned into "no but the problem is the constant usage all the time, which you cannot do with a laptop", then, again, it's just pure conjecture: it may end up being true, but it is still scientifically incorrect to present these conclusions as scientific)
    - jemmyw 3y ago
      While I agree with you on the scientific front, I agree with the author on the risk/action cost front. There's not much cost to preventing the use of phones during school hours, it seems to be a net benefit to education and teacher/student relations even if the hypothesis is wrong. Banning platforms from allowing under 16s to have accounts is harder because kids will work around it very easily, but I would be happy to see it happen anyway, and made Facebook's problem to deal with.

cauch 3y ago

My rebuttal would be: it is not impossible to attribute causality, and THIS IS THE PROBLEM. Right now, there are two identically possible situations (well, even more, but you get the jest): A) smartphone -> depression B) depression -> smartphone But the author does not do a good job in explaining why he choose A instead of B, except "it feels logical to me". It feels like a coin toss. (and sure, the author has some arguments, but they are arguably disputable, more than in other cases where the idea that there is causality ends up being adopted as a consensus) It does not mean it is impossible to reasonably attribute causality. It just means that the author does a bad job at the "reasonably" part.

bhk 3y ago
Or also.... C) X -> depression AND X -> smartphone First example, X could be "strongly needs to feel included"
- cauch 3y ago
  yes, exactly
- illiac786 3y ago
  Why then is there an uptick since 2012, for example? Your theory doesn't explain this, while the author's does. We can't just consider all theories equal because they are all in the realm of the possible. Some evidence, while not proof, strongly points toward social media being the cause or at least a very strong factor. I agree with multiple other commenters here, that banning phones in school would be a high reward chances / low risk decision. We're talking about LOTS of kids getting their mental health damaged, but in many comments I don't hear the sense of urgency, rather a detached a almost snobbish focus on correlation != causality. I think such answers are partially missing the point.
  - cauch 3y ago
    There are TONS of things that changed in the society around 2012. (for example, a specific and brand new phase in globalisation, in climate change crisis, in global loss of trust in the media, in disillusion in the capacity of changing elements that matters, in faster access of overwhelming communication, a more open communication about personal issues (gender, ...), ...) The "smartphone change" is the "what people who don't know sociology think a sociology change looks like". It is still a sociology change, but it's the "nose in the middle of the face". > I agree with multiple other commenters here, that banning phones in school would be a high reward chances / low risk decision. And, surprise, I'm 100% for the ban in school too. But we need to solve the problem, not hide it. The problem can be, for example, social media. Banning smartphone at school will not solve the depression crisis, because people will still have their fix on their laptop after school, and will just maintain the toxic social media relationship dynamic in "real life" in school. If you really care about the kids, you should understand why some people are so unhappy about having a conclusion as naive and simplistic as "let's just ban the smartphone and change nothing in the fundamental ways we are building a society, and I promise you, everything will magically be fine". (and again, it does not mean smartphones are not part of the equation. But we need to find the real cause, and this is why correlation != causality is important. Nobody cares about the theory, people raise this argument because they know this mistake will have bad consequences)
naasking 3y ago
> Right now, there are two identically possible situations (well, even more, but you get the jest): This is misleading. They are not "identically possible", which I assume to mean that you think they are "equally plausible". There is more evidence for smartphones and social media causing issues, because we don't see those declines until after the smartphone is introduced, regardless of age. Furthermore, other factors causing both depression and social media use would have to be cross-cultural because the issue is global. Other than social media, what factors do you know of are cross-cultural? Guess what, they eliminated most of them in prior articles. Haidt and co have published multiple articles going over the data and the possibilities, so I recommend reading them all before claiming the situation is so murky.
- cauch 3y ago
  ("identically possible" as: as legitimately grounded. But of course, if you are convinced that people who are not convinced by your argument can only be dishonest, of course, for you, they are not identically possible) > because we don't see those declines until after the smartphone is introduced, regardless of age. Again, there are a lot of identically possible explanation: the whole society has changed and became globally more anxiogenic, the measurement of the decline may be biased as people are nowadays more incline to be honest with their mental health or less incline to "shake it off", or ... > Furthermore, other factors causing both depression and social media use would have to be cross-cultural because the issue is global. Other than social media, what factors do you know of are cross-cultural? Guess what, they eliminated most of them in prior articles. What do you mean "cross-cultural"? Isn't the study showing that it's mainly a US problem? And how can you have a cross-cultural effect when the usage of the smartphone amongst young people is so linked to cultural trends that are very different from country to country? I have the impression the argumentation jumps from one to the others according to what benefits the authors: if it happens at two places, they say "see, it proves there is no other sources", if it does not happen in other places, they say "sure, but it's just because the smartphone usage is socially different in these countries". > Haidt and co have published multiple articles going over the data and the possibilities, so I recommend reading them all before claiming the situation is so murky. And other experts, with similar credentials than Haidt, have been critical of the work. My position is that we should not jump to the conclusion: the jury is still out. Your conclusion seems to be "I choose to trust Haidt and to distrust equivalent scientists, just because Haidt's explanation seems more obvious to me".
cpsempek 3y ago
I commented this elsewhere, but it's worth repeating because actually reading the author's arguments (granted, in other writings, not OP) will allow you to see he's considered this point and is will to accept that causation works the other way around. It's a fair point. But what explains the uptick in depression and mental health issues starting around 2012, disproportionately impacting pre-teen girls and not contained to any geo. The author's entire point is that social media is the only explanation that that has been proposed [1]. Moreover, it's not a far-fetched explanation. He very aware that correlation does not provide causation, and that it could be the other way around. However, no one (according to him) has offered up a theory which explains the data like the social media theory does. [1] https://jonathanhaidt.substack.com/p/social-media-mental-illness-epidemic https://jonathanhaidt.substack.com/p/social-media-mental-ill...
- cauch 3y ago
  And these arguments have been heard, but they are still not convincing. I don't know what is the reality, maybe indeed the smartphone is causing the problem. But the author is just jumping to the conclusion without proper scientifically caution. There are a lot of other hypothesis that explain exactly the same observations, and that are not far-fetched (as you say, there is none _according to the author_, who then list all of these hypotheses in his rebutal). In his article, he goes through a lot of them but don't address them properly. For example, the more recent social anxiety about climate change is just "I don't think so". Personally, if I have to bet, I think it is social media. But then, 1) I will not claim "it is social media", just "it may be social media", 2) it would be stupid of me to then isolate smartphone: if it is social media, then forbidding smartphone but not the laptops would be totally useless (and if the argument is turned into "no but the problem is the constant usage all the time, which you cannot do with a laptop", then, again, it's just pure conjecture: it may end up being true, but it is still scientifically incorrect to present these conclusions as scientific)
  - jemmyw 3y ago
    While I agree with you on the scientific front, I agree with the author on the risk/action cost front. There's not much cost to preventing the use of phones during school hours, it seems to be a net benefit to education and teacher/student relations even if the hypothesis is wrong. Banning platforms from allowing under 16s to have accounts is harder because kids will work around it very easily, but I would be happy to see it happen anyway, and made Facebook's problem to deal with.