“Teen suicide” has been falling since 2017. You never heard about it because anti-social-media and other interests profit from spreading fear. Now, they’ll start grabbing undeserved credit.
Absolutely, and there's an even bigger bias: Blue states are substantially richer and more racially diverse than Red states. Suicide rates are higher among Whites, especially poorer and rural ones. My larger point is that for Haidt to blame liberal girls' depression on social media and then elsewhere state that social-media-caused depression must be driving suicide is both illogical and academic malpractice. Thanks for comment.
Don't you think the conservative/liberal analysis suffers a bit of omitted variable bias? Might the difference be explain in part by education of parents, household income, criminality, etc?
I take your point Haidt may inflate his narrative. But this just seems to be inflating an opposing narrative. Regardless, I do appreciate you balancing the scales a bit.
Yes, red-blue differences are very likely explained by omitted variables. However, things like education, income, criminality, public health, anti-LGBTQ, etc., deficits are much worse in Republican states and localities, due in large part to their policies. I analyze some of these in my YES article, https://www.yesmagazine.org/economy/2022/03/21/republican-conservative-america-angry . The history of teen suicide is one of such grotesque exploitation, by overbuilt psychiatric hospitals and sensational professionals in the 1980s and now by social-media blamers.
I accuse Haidt of deliberate deception on this dangerous topic because he bases huge parts of his Anxious Generation book on the CDC's 2021 YRBS survey. So, either he is a very poor researcher, or he surely noticed vital facts like these: suicide attempts among teens under age 16 (who he wants to ban from social media): 11% for those who use social media less than daily, versus 8% for those who use social media daily or more. Self-harm: 3% for those who rarely use social media, vs 1% for those who frequently use social media.
This pattern is repeated across the 2021 CDC survey: teens who use social media more are more likely to be depressed, but LESS likely to attempt suicide or self harm. That is a gigantic finding with profound implications for addressing suicide. Yet, Haidt completely ignores (if he is a bad researcher) or suppresses (if he is a dishonest researcher) these important results. I choose to accuse him of being dishonest in his presentation rather than being so incompetent that he is dishonest to accept accolades as an expert. Maybe he simply is inexcusably ignorant of the data he claims are so important. Either way, flat-out lies are being perpetrated.
Agreed, focusing on "liberal areas" and "conservative areas" seems to be arguing about something different from Haidt. Maybe liberals in conservative areas are doing much worse than liberals in liberal areas, but that doesn't tell you anything about the impact of social media on mental health.
All of this information is very interesting and should be part of the discussion, but was way too aggressively portrayed. I think Haidt holds genuine beliefs, I don't think this is just about book sales. And the underlying motivation has nothing to do with whether or not he is correct about the issue.
It’s weird how you approach this as a need to expose the researchers you disagree with for lying. Haidt may be empirically wrong, and his errors may mutually reinforce political conclusions you disagree with (though I think your model of his politics is distorted). But it’s a naive, unproductive worldview to assume people you disagree with are bad intentioned liars. I suppose it drives engagement.
The difference between legitimate disagreement and lying is the refusal to correct clearly documented mistakes. Prior to a couple of years ago, I saw social-media blamers like Haidt as honest players. However, their abject deterioration into demagoguery wildly hyping social media harm to teens (which Haidt admitted in the past is "very small"), blaming social media for teen suicide trends (which they know they have no evidence for), and deceptively dismissing the effects of parental violence and abuse in teen depression and suicide (which Haidt's own analysis of the 2021 CDC survey would show as crucial) places them in the category of liars. Suicide is too important an issue for Haidt et al to play with to promote books and politics even after being informed multiple times of their gross errors. Their argument has degenerated into lying.
I don't fully disagree. But I think it undermines your own credibility if you devote too much time to personal attacks. Ideally we would all be capable of analyzing the data dispassionately in order to come to the correct understanding. If it sounds like your biggest goal is to discredit a single person, you risk deteriorating into demagoguery just as much as Haidt. I can't say if that has affected your analysis at all, but the aggressive nature of your tone will make it seem as though it has.
None of your data here actually disproves what Haidt is claiming, suicide rates still went up at the same time that phones were becoming common. Haidt certainly isn't including the level of context that you are providing and I think it's really great that you are countering the narrative and providing additional context. Keep it up!
With respect, Mike, I'm sure we can agree that evidence is not self-interpreting. Haidt may have been informed multiple times of his gross errors, but if he declines to change his analysis in response to these corrections, it is likely that he simply disagrees with your interpretation of the evidence, not that he is lying.
I'm not trying to attack your argument. I truly believe that this area of public health and policy is so morally, politically, ethically fraught that we need to work really hard to turn down the emotional volume. Even if I am wrong, and Haidt is a demogogue and liar, consider the possibility that you will be more persuasive to people who are currently under his influence if you treat him and other opponents as esteemed colleagues who are mistaken rather than as bad actors.
I agree generally that attributing bad motives is disagreeable in itself. However, I've been at this for years, I've tried the conciliatory approach, and it does not work -- and worse, the soft approach was not honest on my part. The "teen suicide" crusade is so thoroughly dishonest, inflammatory, and potentially damaging that I think we have to state facts available from our own research.
Haidt bases a large part of his book on the 2021 CDC survey which he claims to have analyzed extensively -- yet that survey clearly shows that teens who never or rarely use social media are MORE likely to attempt suicide and harm themselves than teens who use social media several hours a day. Surely he saw that -- or he did not analyze the survey as he claimed.
At the least, Haidt should have mentioned this startling finding and called for more study of what it might mean for such a crucial topic as suicide. Yet, he never mentioned it at all, nor did Twenge or anyone else, which reveals a level of callous dishonesty I term lying. The 2023 CDC survey shows a similar pattern for teens under age 16 -- less social media use is associated with MORE suicide attempts and self harm, especially for girls, the very groups Haidt et al wax incendiary over protecting. I don't see this as a mere mistake or difference in interpretation -- and, I repeat, it is dangerous for those teens and adults who do need help.
I'm guessing it's real, since Americans of all ages (not just teens) check the lonely box on surveys. There are a rising number of good studies showing social media is not contributing to loneliness or social maladaptation; just the opposite, those who do well with social media also do well in real-life socialization. I don't get this. Unlike personal interactions among family, associates, etc., technology use is controllable. I have never had a computer force me to use it.
It seems likely that loneliness going up is simply a product of modern lifestyles regardless of technology. Bowling Alone can explain it, while adding the fact that our work lives expose us to fewer and fewer people and kids don't run around in groups anymore, etc. But also who knows if we just have different expectations today? Like maybe people used to not mind being alone as much. It wouldn't surprise me if media (whether movies, tv, or social media) changed our expectations for how much social interaction is desirable. I'm skeptical of making any kind of judgement about a thing that is as subjective as self-described loneliness.
I think that's largely right. I would suggest watching "Social Media" on Fx, because of its bizarre disconnect between claiming teens today are isolated and lonely due to social media -- then, 99% of the series' scenes showed teens actively socializing in person, often in groups, talking vigorously even though their phones are out. I'm wondering if more teens are saying they're lonely because an avalanche of daily media commentaries tell them they are the most mentally disturbed generations ever and how idyllically happy past generations of youth were -- despicable lies, of course, but teens have no way of knowing what past generations were really like.
I hate I’m just discovering your Substack at 12:54 am! I have a lot of reading to do. I actually just published a podcast episode about Heidi’s book yesterday. Very interested to read all of your work.
Thanks, Mike, for your thorough analytical work on these demographics.
However, I have a question of the validity of the data coming from the CDC that we know especially through Covid1984 lied about nearly everything, so can we trust them on this data?
Is there another data set on this topic we can go to to compare?
The main topic I think we are in agreement on is what I term the Elephant in the Womb—the taken for granted “natural” right for so-called adults to rule so-called children and keep them in a category akin to pets and slave.
My neologism, Parentarchy, along with Childism/Adultism (Holt, Farson, Gordon) is trying to make visible this Abuse of Power that is the real cause of "Adults" causing their "Children" to rebel with self-and-other harm.
First the Voluntaryist insight:
“If we cannot trust people with freedom how can we trust them with power?”
Then John who put it so well:
“To trust children we must first learn to trust ourselves; and most of us were taught as children that we could not be trusted.” John Holt
Thanks for comment. I don't think CDC lied about COVID. A misleading private report citing CDC figures that only 6% of COVID deaths involved COVID alone has been debunked. 92% of COVID deaths involved COVID as the underlying cause along with comorbidities, and 8% another underlying cause. As for suicide, that determination is made by local coroners requiring evidence of intent, otherwise the default is accident. There are flaws in such a death certification system, particularly for rural areas where the coroner is the sheriff (!), but I think the figures can be trusted.
I've never heard of a case where the CDC was found to alter data to suit a narrative. That doesn't mean they won't push a narrative based on what data they have, but that's a horse of a different color. Also, it's their job to figure out how to reduce disease spread, it's not their job to figure out what interventions are "worth it". That's for us and our politicians to decide. The CDC will always advocate for the largest possible intervention, that's what they should be doing. It's just our job to tell them no sometimes.
Makes sense for the phone thing. You cant profit off of someone dead. So the profit incentive for social media naturally encourages not killing yourself.
Additionally social media is very, very good at pandering to human desire. Depressed? Heres a community of people who love your favourite game or tv show! Or heres some cool podcast that can let you engage in gossip without ruining real life relationships. Very effective circuses.
Would you share the effect sizes the effect of conservative vs liberal areas on suicide rates? And what you controlled for in estimating the effect sizes?
I think it's good, better than the last post. I assume you noticed, but my problem with the last one was that you took Haight et al. to task for not using effect sizes, but then you offered evidence that didn't use effect sizes.
Here you do, and you drive your point home by offering a comparison, basically "if you want to say that phones cause suicide, then (by the same standard of evidence) be prepared to believe other things you don't want to!" Make sense.
I tried to analyze the YRBSS survey myself using multivariate regression, but I use Stata and the data isn't offered in that format. Drat. Anyway, what you wrote is in line with what I've read in other studies: there isn't strong evidence showing social media causes harm. So I'll just take your word on this and the political data.
I like that you emphasize that this is all correlational, since your point is that Haight and co. are relying on correlations. One interpretation of the political data is that conservative partisans take liberal leadership in stride, whereas liberal partisans are more likely to despair under GOP leadership--and maybe it would help if we turned down the heat on our political rhetoric. But who knows?
Absolutely, and there's an even bigger bias: Blue states are substantially richer and more racially diverse than Red states. Suicide rates are higher among Whites, especially poorer and rural ones. My larger point is that for Haidt to blame liberal girls' depression on social media and then elsewhere state that social-media-caused depression must be driving suicide is both illogical and academic malpractice. Thanks for comment.
Absolutely!
Don't you think the conservative/liberal analysis suffers a bit of omitted variable bias? Might the difference be explain in part by education of parents, household income, criminality, etc?
I take your point Haidt may inflate his narrative. But this just seems to be inflating an opposing narrative. Regardless, I do appreciate you balancing the scales a bit.
Yes, red-blue differences are very likely explained by omitted variables. However, things like education, income, criminality, public health, anti-LGBTQ, etc., deficits are much worse in Republican states and localities, due in large part to their policies. I analyze some of these in my YES article, https://www.yesmagazine.org/economy/2022/03/21/republican-conservative-america-angry . The history of teen suicide is one of such grotesque exploitation, by overbuilt psychiatric hospitals and sensational professionals in the 1980s and now by social-media blamers.
I accuse Haidt of deliberate deception on this dangerous topic because he bases huge parts of his Anxious Generation book on the CDC's 2021 YRBS survey. So, either he is a very poor researcher, or he surely noticed vital facts like these: suicide attempts among teens under age 16 (who he wants to ban from social media): 11% for those who use social media less than daily, versus 8% for those who use social media daily or more. Self-harm: 3% for those who rarely use social media, vs 1% for those who frequently use social media.
This pattern is repeated across the 2021 CDC survey: teens who use social media more are more likely to be depressed, but LESS likely to attempt suicide or self harm. That is a gigantic finding with profound implications for addressing suicide. Yet, Haidt completely ignores (if he is a bad researcher) or suppresses (if he is a dishonest researcher) these important results. I choose to accuse him of being dishonest in his presentation rather than being so incompetent that he is dishonest to accept accolades as an expert. Maybe he simply is inexcusably ignorant of the data he claims are so important. Either way, flat-out lies are being perpetrated.
Agreed, focusing on "liberal areas" and "conservative areas" seems to be arguing about something different from Haidt. Maybe liberals in conservative areas are doing much worse than liberals in liberal areas, but that doesn't tell you anything about the impact of social media on mental health.
All of this information is very interesting and should be part of the discussion, but was way too aggressively portrayed. I think Haidt holds genuine beliefs, I don't think this is just about book sales. And the underlying motivation has nothing to do with whether or not he is correct about the issue.
It’s weird how you approach this as a need to expose the researchers you disagree with for lying. Haidt may be empirically wrong, and his errors may mutually reinforce political conclusions you disagree with (though I think your model of his politics is distorted). But it’s a naive, unproductive worldview to assume people you disagree with are bad intentioned liars. I suppose it drives engagement.
The difference between legitimate disagreement and lying is the refusal to correct clearly documented mistakes. Prior to a couple of years ago, I saw social-media blamers like Haidt as honest players. However, their abject deterioration into demagoguery wildly hyping social media harm to teens (which Haidt admitted in the past is "very small"), blaming social media for teen suicide trends (which they know they have no evidence for), and deceptively dismissing the effects of parental violence and abuse in teen depression and suicide (which Haidt's own analysis of the 2021 CDC survey would show as crucial) places them in the category of liars. Suicide is too important an issue for Haidt et al to play with to promote books and politics even after being informed multiple times of their gross errors. Their argument has degenerated into lying.
I don't fully disagree. But I think it undermines your own credibility if you devote too much time to personal attacks. Ideally we would all be capable of analyzing the data dispassionately in order to come to the correct understanding. If it sounds like your biggest goal is to discredit a single person, you risk deteriorating into demagoguery just as much as Haidt. I can't say if that has affected your analysis at all, but the aggressive nature of your tone will make it seem as though it has.
None of your data here actually disproves what Haidt is claiming, suicide rates still went up at the same time that phones were becoming common. Haidt certainly isn't including the level of context that you are providing and I think it's really great that you are countering the narrative and providing additional context. Keep it up!
With respect, Mike, I'm sure we can agree that evidence is not self-interpreting. Haidt may have been informed multiple times of his gross errors, but if he declines to change his analysis in response to these corrections, it is likely that he simply disagrees with your interpretation of the evidence, not that he is lying.
I'm not trying to attack your argument. I truly believe that this area of public health and policy is so morally, politically, ethically fraught that we need to work really hard to turn down the emotional volume. Even if I am wrong, and Haidt is a demogogue and liar, consider the possibility that you will be more persuasive to people who are currently under his influence if you treat him and other opponents as esteemed colleagues who are mistaken rather than as bad actors.
I agree generally that attributing bad motives is disagreeable in itself. However, I've been at this for years, I've tried the conciliatory approach, and it does not work -- and worse, the soft approach was not honest on my part. The "teen suicide" crusade is so thoroughly dishonest, inflammatory, and potentially damaging that I think we have to state facts available from our own research.
Haidt bases a large part of his book on the 2021 CDC survey which he claims to have analyzed extensively -- yet that survey clearly shows that teens who never or rarely use social media are MORE likely to attempt suicide and harm themselves than teens who use social media several hours a day. Surely he saw that -- or he did not analyze the survey as he claimed.
At the least, Haidt should have mentioned this startling finding and called for more study of what it might mean for such a crucial topic as suicide. Yet, he never mentioned it at all, nor did Twenge or anyone else, which reveals a level of callous dishonesty I term lying. The 2023 CDC survey shows a similar pattern for teens under age 16 -- less social media use is associated with MORE suicide attempts and self harm, especially for girls, the very groups Haidt et al wax incendiary over protecting. I don't see this as a mere mistake or difference in interpretation -- and, I repeat, it is dangerous for those teens and adults who do need help.
What do you think about the loneliness epidemic? Is it real? Do smartphones contribute? Are zoomer socially awkard because too much technology?
I'm guessing it's real, since Americans of all ages (not just teens) check the lonely box on surveys. There are a rising number of good studies showing social media is not contributing to loneliness or social maladaptation; just the opposite, those who do well with social media also do well in real-life socialization. I don't get this. Unlike personal interactions among family, associates, etc., technology use is controllable. I have never had a computer force me to use it.
I was skeptic about technology and loneliness myself. From personal experience usually people who use a lot social are very...social.
That's my experience as well. Social media simply offers another way to connect with others, not a barrier to connection.
It seems likely that loneliness going up is simply a product of modern lifestyles regardless of technology. Bowling Alone can explain it, while adding the fact that our work lives expose us to fewer and fewer people and kids don't run around in groups anymore, etc. But also who knows if we just have different expectations today? Like maybe people used to not mind being alone as much. It wouldn't surprise me if media (whether movies, tv, or social media) changed our expectations for how much social interaction is desirable. I'm skeptical of making any kind of judgement about a thing that is as subjective as self-described loneliness.
I think that's largely right. I would suggest watching "Social Media" on Fx, because of its bizarre disconnect between claiming teens today are isolated and lonely due to social media -- then, 99% of the series' scenes showed teens actively socializing in person, often in groups, talking vigorously even though their phones are out. I'm wondering if more teens are saying they're lonely because an avalanche of daily media commentaries tell them they are the most mentally disturbed generations ever and how idyllically happy past generations of youth were -- despicable lies, of course, but teens have no way of knowing what past generations were really like.
I hate I’m just discovering your Substack at 12:54 am! I have a lot of reading to do. I actually just published a podcast episode about Heidi’s book yesterday. Very interested to read all of your work.
Hi Ann -- welcome to the Substack. Please send link to your podcast. Here's my article on Haidt, published yesterday: https://www.laprogressive.com/law-and-justice/global-nanny
Thanks, Mike, for your thorough analytical work on these demographics.
However, I have a question of the validity of the data coming from the CDC that we know especially through Covid1984 lied about nearly everything, so can we trust them on this data?
Is there another data set on this topic we can go to to compare?
The main topic I think we are in agreement on is what I term the Elephant in the Womb—the taken for granted “natural” right for so-called adults to rule so-called children and keep them in a category akin to pets and slave.
My neologism, Parentarchy, along with Childism/Adultism (Holt, Farson, Gordon) is trying to make visible this Abuse of Power that is the real cause of "Adults" causing their "Children" to rebel with self-and-other harm.
First the Voluntaryist insight:
“If we cannot trust people with freedom how can we trust them with power?”
Then John who put it so well:
“To trust children we must first learn to trust ourselves; and most of us were taught as children that we could not be trusted.” John Holt
Thanks for comment. I don't think CDC lied about COVID. A misleading private report citing CDC figures that only 6% of COVID deaths involved COVID alone has been debunked. 92% of COVID deaths involved COVID as the underlying cause along with comorbidities, and 8% another underlying cause. As for suicide, that determination is made by local coroners requiring evidence of intent, otherwise the default is accident. There are flaws in such a death certification system, particularly for rural areas where the coroner is the sheriff (!), but I think the figures can be trusted.
I've never heard of a case where the CDC was found to alter data to suit a narrative. That doesn't mean they won't push a narrative based on what data they have, but that's a horse of a different color. Also, it's their job to figure out how to reduce disease spread, it's not their job to figure out what interventions are "worth it". That's for us and our politicians to decide. The CDC will always advocate for the largest possible intervention, that's what they should be doing. It's just our job to tell them no sometimes.
Makes sense for the phone thing. You cant profit off of someone dead. So the profit incentive for social media naturally encourages not killing yourself.
Additionally social media is very, very good at pandering to human desire. Depressed? Heres a community of people who love your favourite game or tv show! Or heres some cool podcast that can let you engage in gossip without ruining real life relationships. Very effective circuses.
Would you share the effect sizes the effect of conservative vs liberal areas on suicide rates? And what you controlled for in estimating the effect sizes?
See https://mikemales.substack.com/p/social-media-isnt-driving-teen-suicide . Any critiques?
I'm flattered that you asked. I'll read it quickly and reply, since I know you probably want to push it out.
I think it's good, better than the last post. I assume you noticed, but my problem with the last one was that you took Haight et al. to task for not using effect sizes, but then you offered evidence that didn't use effect sizes.
Here you do, and you drive your point home by offering a comparison, basically "if you want to say that phones cause suicide, then (by the same standard of evidence) be prepared to believe other things you don't want to!" Make sense.
I tried to analyze the YRBSS survey myself using multivariate regression, but I use Stata and the data isn't offered in that format. Drat. Anyway, what you wrote is in line with what I've read in other studies: there isn't strong evidence showing social media causes harm. So I'll just take your word on this and the political data.
I like that you emphasize that this is all correlational, since your point is that Haight and co. are relying on correlations. One interpretation of the political data is that conservative partisans take liberal leadership in stride, whereas liberal partisans are more likely to despair under GOP leadership--and maybe it would help if we turned down the heat on our political rhetoric. But who knows?
Thanks for looping me in.
Ack, I wrote "make sense" but I meant to write "makes sense." That letter S really changes the meaning of the sentence, doesn't it?
Great suggestion. I'm going to do a short blog post on this. Give me a day or two and I'll send the link. -Mike
Well said, Mike.