Internet use linked to higher wellbeing, global study suggests
May 14, 2024 7:13 AM Subscribe
Internet use is linked to higher wealth and income and better infrastructure. I think the higher wellbeing comes from that, not being online.
posted by thecjm at 7:21 AM on May 14 [38 favorites]
posted by thecjm at 7:21 AM on May 14 [38 favorites]
the authors could not entirely discount the possibility that increases in incomes, which were also linked to rises in internet access, were behind people feeling better.
That was my first thought.
posted by rmd1023 at 7:22 AM on May 14 [17 favorites]
That was my first thought.
posted by rmd1023 at 7:22 AM on May 14 [17 favorites]
This seems like it might be more correlation than causation. The people I know who are mostly offline are also usually one or more of poor, disabled, rural, or part of very conservative communities or families. It doesn't seem like they controlled for any other factors than gender and age.
posted by The Manwich Horror at 7:22 AM on May 14 [7 favorites]
posted by The Manwich Horror at 7:22 AM on May 14 [7 favorites]
But seriously, thank you for posting this. I look forward to reading the research article when it’s published. I regularly work with students researching this topic (broadly) all the time, and it’s nice to have a big study like this to point them at for consideration.
posted by cupcakeninja at 7:22 AM on May 14 [7 favorites]
posted by cupcakeninja at 7:22 AM on May 14 [7 favorites]
Depressed + terminally online -> now empirically a failure -> even more depressed.
Sensing a pattern here.
posted by Ryvar at 7:28 AM on May 14 [3 favorites]
Sensing a pattern here.
posted by Ryvar at 7:28 AM on May 14 [3 favorites]
I'm all for more data and more data-driven policy, but isn't it trivial to control for income?
posted by signal at 7:36 AM on May 14 [1 favorite]
posted by signal at 7:36 AM on May 14 [1 favorite]
But it was all right, everything was all right, the struggle was finished. He had won the victory over himself. He loved The Internet.
posted by The Card Cheat at 7:42 AM on May 14 [20 favorites]
posted by The Card Cheat at 7:42 AM on May 14 [20 favorites]
The study did not look specifically at social media - which is what much of the most heated debate around online safety is focussed on - but took a broader approach to assessing access to the internet.
This study is ridiculously bad and can tell us nothing. They just slap together a pic of a generic attractive young woman from the Getty Images stock photo site and hope you don't have ad block.
posted by AlSweigart at 7:43 AM on May 14 [1 favorite]
This study is ridiculously bad and can tell us nothing. They just slap together a pic of a generic attractive young woman from the Getty Images stock photo site and hope you don't have ad block.
posted by AlSweigart at 7:43 AM on May 14 [1 favorite]
@dril: *understands tthe full potential of the net all at once and stumbles backward wwhile struggling to breathe* christ,. my god
posted by credulous at 8:00 AM on May 14 [4 favorites]
posted by credulous at 8:00 AM on May 14 [4 favorites]
study suggests dunking on trite 'academic' analyses temporarily improves nerd wellbeing and outlook
posted by lalochezia at 8:00 AM on May 14 [6 favorites]
posted by lalochezia at 8:00 AM on May 14 [6 favorites]
The OII is a research institute based inside of University of Oxford that specifically researches how humans engage with the internet using, from what I can gather, sociological and epidemiological methods. It's not, as what someone had suggested above, a pro-internet shill group funded by uber-evil internet pioneers like... Tim Berners-Lee? lol
They do have a list of their past publications which includes such evil, biased works arguing things like how private interests have unfairly shaped policy around gig work and how bundled software packages like Microsoft Teams are anti-competitive and inefficient.
Perhaps dismissive scoffing and intellectually lazy dismissals of things that we haven't even had a chance to read could be replaced with better faith digging into the topic at large.
posted by paimapi at 8:00 AM on May 14 [20 favorites]
They do have a list of their past publications which includes such evil, biased works arguing things like how private interests have unfairly shaped policy around gig work and how bundled software packages like Microsoft Teams are anti-competitive and inefficient.
Perhaps dismissive scoffing and intellectually lazy dismissals of things that we haven't even had a chance to read could be replaced with better faith digging into the topic at large.
posted by paimapi at 8:00 AM on May 14 [20 favorites]
Here's the actual study.
Looks like having access to the internet was associated with 0.04-0.08 greater life satisfaction on a 1-5 scale, so 1-2%. Funny how they don't mention any numbers in the article.
posted by ssg at 8:03 AM on May 14 [18 favorites]
Looks like having access to the internet was associated with 0.04-0.08 greater life satisfaction on a 1-5 scale, so 1-2%. Funny how they don't mention any numbers in the article.
posted by ssg at 8:03 AM on May 14 [18 favorites]
I just went to send someone a GIF reaction on a text, searched for “fast bird,” and got an endless stream of obese birds, obese children, and pictures of cartoon people overeating. Believe me, none of that was good for me, or would have been good for the recipient, therefore the internet is not linked to wellbeing. Certainly it would not have been if I had sent any of those GIFs and my friend had responded with appropriate force.
posted by cupcakeninja at 8:10 AM on May 14 [1 favorite]
posted by cupcakeninja at 8:10 AM on May 14 [1 favorite]
it looks like they specifically control for both income and educational attainment along with a number of others:
We chose six variables to represent such potentially confounding factors that have been previously considered important in the literature on well-being: the respondent’s income, educational, work, and relationship statuses, their ability to meet basic needs for food and shelter, and whether or not they reported having health problems.
I think the fact that even if the effect size is extremely small, this already disproves Metafilter's favorite anecdotal recounting of how internet use has actively harmed human beings on net since it seems to have improved things for most people (not adolescent women, unfortunately). anecdotally, from my end, I've seen the internet facilitate connections in marginalized communities who otherwise would not have been able to find others like themselves. anecdotally, I've also seen organizers use the internet for great effect, increasing turnout and fundraising for important, regularly ignored causes, and generally utilizing it as one very useful tool in a toolbox for building community with others
I think a more nuanced take of the internet is that it's a neutral-to-good tool for social connection that's unfortunately been vastly over-commodified by private, capital-driven interests. but, like any tool, you can also use it for good - it depends, at the moment, entirely on how you, individually, choose to engage with it and use it. I hope that policy one day controls how much of a shithole dopamine-clicker that apps/games/etc have turned the internet into so it wouldn't take so much work actively choosing how and where to use the internet as it currently exists but the handwavey overt dismissal that Metafilter ironically throws out as a tired canard regarding the internet's impact on individual lives should, at the very least, deserve some critical skepticism and introspection since even with all the enshittification it seems to be making people a little bit happier on the whole
posted by paimapi at 8:14 AM on May 14 [21 favorites]
We chose six variables to represent such potentially confounding factors that have been previously considered important in the literature on well-being: the respondent’s income, educational, work, and relationship statuses, their ability to meet basic needs for food and shelter, and whether or not they reported having health problems.
I think the fact that even if the effect size is extremely small, this already disproves Metafilter's favorite anecdotal recounting of how internet use has actively harmed human beings on net since it seems to have improved things for most people (not adolescent women, unfortunately). anecdotally, from my end, I've seen the internet facilitate connections in marginalized communities who otherwise would not have been able to find others like themselves. anecdotally, I've also seen organizers use the internet for great effect, increasing turnout and fundraising for important, regularly ignored causes, and generally utilizing it as one very useful tool in a toolbox for building community with others
I think a more nuanced take of the internet is that it's a neutral-to-good tool for social connection that's unfortunately been vastly over-commodified by private, capital-driven interests. but, like any tool, you can also use it for good - it depends, at the moment, entirely on how you, individually, choose to engage with it and use it. I hope that policy one day controls how much of a shithole dopamine-clicker that apps/games/etc have turned the internet into so it wouldn't take so much work actively choosing how and where to use the internet as it currently exists but the handwavey overt dismissal that Metafilter ironically throws out as a tired canard regarding the internet's impact on individual lives should, at the very least, deserve some critical skepticism and introspection since even with all the enshittification it seems to be making people a little bit happier on the whole
posted by paimapi at 8:14 AM on May 14 [21 favorites]
"The internet" is such a broad and non-specific thing that trying to glean any insight into its impact on people is pretty much meaningless, as evidenced by the study's results and methodology, which pretty much amounted to asking people if they had access to the internet. Given that everything, including government services, social lives, school, work, etc. is now on the internet and people are mostly required to use it, what is really to be learned here? People who have access to that thing that pretty much everyone has to use to exist in the world are a very tiny degree more satisfied with their lives than people who don't. News at 11.
posted by grumpybear69 at 8:14 AM on May 14 [3 favorites]
posted by grumpybear69 at 8:14 AM on May 14 [3 favorites]
Among other things, the internet helps connect people who are marginalised or Othered in some way with other people who have the same life experiences
eg connecting Disabled people to other Disabled people
connecting Chronically Ill people to other Chronically ill people
connecting LGBT people to other LGBT people
connecting Women of Colour to other Women of Colour.
posted by chariot pulled by cassowaries at 8:20 AM on May 14 [18 favorites]
eg connecting Disabled people to other Disabled people
connecting Chronically Ill people to other Chronically ill people
connecting LGBT people to other LGBT people
connecting Women of Colour to other Women of Colour.
posted by chariot pulled by cassowaries at 8:20 AM on May 14 [18 favorites]
Yes. Connecting minorities of all sorts is what the internet does best. You may be the only gay in the village, but the world is a click away.
posted by pracowity at 8:38 AM on May 14 [7 favorites]
posted by pracowity at 8:38 AM on May 14 [7 favorites]
Unfortunately it also connects the religious fanatics, racists, and conspiracy theorists who might otherwise be lonely in their different villages.
Maybe an interesting way to look at the effect of the internet is that it has removed shunning as an effective social punishment. Shunned people can find each other and connect with each other now, instead of being isolated. And that's true whether they were shunned for good reasons or bad ones...
posted by OnceUponATime at 8:59 AM on May 14 [5 favorites]
Maybe an interesting way to look at the effect of the internet is that it has removed shunning as an effective social punishment. Shunned people can find each other and connect with each other now, instead of being isolated. And that's true whether they were shunned for good reasons or bad ones...
posted by OnceUponATime at 8:59 AM on May 14 [5 favorites]
racists and religious fanatics were already given space spaces aplenty by our modern institutions. compare, for example, the police reaction to alt-right 'protests' and even January 6th versus how aggressive they've been with campus protests. look, too, at the 'non-profit' statuses of many a religious organization with incredibly harmful beliefs
marginalized folks are marginalized not just because of social shunning - institutions work to silence and discredit them too. at least online there are services that facilitate connection outside of these institutions for once
posted by paimapi at 9:03 AM on May 14 [4 favorites]
marginalized folks are marginalized not just because of social shunning - institutions work to silence and discredit them too. at least online there are services that facilitate connection outside of these institutions for once
posted by paimapi at 9:03 AM on May 14 [4 favorites]
I know for me personally, the internet has been a net good. My three best friends in the whole world, I met through the internet. I've found new hobbies, expanded old ones, found ways to support my mental and physical health, and have been able to learn a ton of new things.
The internet has taught me so much about how to be a better person, better professional, better friend, and more conscious in the world. Plus, all those cat videos are a lovely distraction on days that just epically suck.
posted by teleri025 at 9:05 AM on May 14 [12 favorites]
The internet has taught me so much about how to be a better person, better professional, better friend, and more conscious in the world. Plus, all those cat videos are a lovely distraction on days that just epically suck.
posted by teleri025 at 9:05 AM on May 14 [12 favorites]
> I think the fact that even if the effect size is extremely small, this already disproves Metafilter's favorite anecdotal recounting of how internet use has actively harmed human beings
No, it doesn't.
"Attempt to control for" does not mean "actually eliminate influence of".
When you have to apply a large number of control effects and you get a small effect size, this is evidence you don't have evidence.
There are approaches that will highlight this problem, but the base issue is that your ability to eliminate the impact of those issues is imperfect. If their effects are large, then what you end up measuring is the error in your elimination attempt; or, the degree to which the error in your elimination attempt correlates with what you are measuring.
Imagine measuring athletic skill vs height. You attempt to control for athletic experience - have they been a pro player in a league, do they self-report regular athletic activities. This control isn't going to be perfect. So what you may conclude - that height leads to natural athletic skill after controlling for athletic experience - what you are actually seeing may just be that height leads to athletic experience that isn't reported, which results in athletic skill.
And here, the effects that also cause more internet use have insanely huge impacts on well being; what more, the effects that prevent internet use do as well: if you are a 13 year old girl who has never been on the internet, this would require a very draconian level of parenting today. That level of control could cause significant harm.
Sociology is insanely hard, and you don't get paid or promoted to publish "this is hard, I have no conclusions". So we get people publishing these small effect papers with huge numbers of controls on confounding factors that prove nothing whatsoever.
posted by NotAYakk at 9:21 AM on May 14 [6 favorites]
No, it doesn't.
"Attempt to control for" does not mean "actually eliminate influence of".
When you have to apply a large number of control effects and you get a small effect size, this is evidence you don't have evidence.
There are approaches that will highlight this problem, but the base issue is that your ability to eliminate the impact of those issues is imperfect. If their effects are large, then what you end up measuring is the error in your elimination attempt; or, the degree to which the error in your elimination attempt correlates with what you are measuring.
Imagine measuring athletic skill vs height. You attempt to control for athletic experience - have they been a pro player in a league, do they self-report regular athletic activities. This control isn't going to be perfect. So what you may conclude - that height leads to natural athletic skill after controlling for athletic experience - what you are actually seeing may just be that height leads to athletic experience that isn't reported, which results in athletic skill.
And here, the effects that also cause more internet use have insanely huge impacts on well being; what more, the effects that prevent internet use do as well: if you are a 13 year old girl who has never been on the internet, this would require a very draconian level of parenting today. That level of control could cause significant harm.
Sociology is insanely hard, and you don't get paid or promoted to publish "this is hard, I have no conclusions". So we get people publishing these small effect papers with huge numbers of controls on confounding factors that prove nothing whatsoever.
posted by NotAYakk at 9:21 AM on May 14 [6 favorites]
No, it doesn't.
ah perfect, so from this we can then conclude that we should go back to dismissively calling the internet a net evil then and not a complicated thing deserving of nuanced discussion, right? because I feel like the thesis of my commenting is pointing to that but you picked one specific bone to contend with and ignored the rest of it
posted by paimapi at 9:29 AM on May 14 [1 favorite]
ah perfect, so from this we can then conclude that we should go back to dismissively calling the internet a net evil then and not a complicated thing deserving of nuanced discussion, right? because I feel like the thesis of my commenting is pointing to that but you picked one specific bone to contend with and ignored the rest of it
posted by paimapi at 9:29 AM on May 14 [1 favorite]
internet-hosted indy 3d model designs + resin 3d printer = my much reduced Games Workshop "plastic crack" addiction
So that's a plus.
Also, learning pretty much my entire career skillset via the internet over the last 30 years. Paying eyewatering amounts, waiting for, and trying to search giant paper brick manuals to learn anything about your business IT equipment or setup absolutely *sucked*
posted by Absolutely No You-Know-What at 9:30 AM on May 14 [1 favorite]
So that's a plus.
Also, learning pretty much my entire career skillset via the internet over the last 30 years. Paying eyewatering amounts, waiting for, and trying to search giant paper brick manuals to learn anything about your business IT equipment or setup absolutely *sucked*
posted by Absolutely No You-Know-What at 9:30 AM on May 14 [1 favorite]
There's some irony in paimapi complaining that someone has picked one line out of their comment to respond to, by themselves responding only to the first line of NotAYakk's comment.
posted by subdee at 9:51 AM on May 14
posted by subdee at 9:51 AM on May 14
it was certainly an excellent comment about statistics and I quoted the crux of it's point - that the point was not disproven. it's not a difficult thing to understand.
also your third-person passive voice critique as if you were addressing a crowd of everyone but me is simultaneously annoying, rude, and somehow cowardly - if you have a problem with an assertion, feel free to engage with me directly
posted by paimapi at 9:59 AM on May 14 [1 favorite]
also your third-person passive voice critique as if you were addressing a crowd of everyone but me is simultaneously annoying, rude, and somehow cowardly - if you have a problem with an assertion, feel free to engage with me directly
posted by paimapi at 9:59 AM on May 14 [1 favorite]
Hey folks! Why hasn’t this study which didn’t change my opinion changed your opinion?
posted by Wood at 10:05 AM on May 14 [3 favorites]
posted by Wood at 10:05 AM on May 14 [3 favorites]
I was in the middle of editing my comment when the comment window closed. I agree that it was an excellent comment about statistics (I have a stats degree).
I'm trying to think of how to word this reply to stay within community guidelines, because I don't want to derail the discussion. But basically, pulling one line out of a comment to reply only to that part is a bog-standard way to reply to someone else online and it doesn't necessarily mean anyone is ignoring the rest of your comment.
posted by subdee at 10:25 AM on May 14 [2 favorites]
I'm trying to think of how to word this reply to stay within community guidelines, because I don't want to derail the discussion. But basically, pulling one line out of a comment to reply only to that part is a bog-standard way to reply to someone else online and it doesn't necessarily mean anyone is ignoring the rest of your comment.
posted by subdee at 10:25 AM on May 14 [2 favorites]
I am generally in agreement with paimapi here -- in my quick skim of the paper (and I am not a statistician so the methods are a bit much for me), the crux from the analysis seems to be that of the many, many factors that influence well-being (as the data are represented in this Gallup dataset), the internet is ONE of them, and it more frequently has a very very small, but still net positive association.
I think the fact that they tested so many covariates is a strength, in fact, because they are clearly acknowledging the complexity of the issue. And they are very clear to point out the permutations where internet use was NOT a net positive. The point is not that the effect of internet use is small -- the authors acknowledge that it would be.
They did this analysis because: a. this study used a global dataset whereas previous research was biased towards data from western/northern hemisphere countries and b. policy about the internet is frequently made based not on data but on knee-jerk reactions and impressions. It is important to provide evidence that the world is more complicated than INTERNET=BAD or INTERNET=GOOD.
Honestly the bigger issue is once again that reporting on the paper elides the very careful way the authors describe their results. "It's Complicated" doesn't make for very sexy headlines.
posted by DiscourseMarker at 10:25 AM on May 14 [8 favorites]
I think the fact that they tested so many covariates is a strength, in fact, because they are clearly acknowledging the complexity of the issue. And they are very clear to point out the permutations where internet use was NOT a net positive. The point is not that the effect of internet use is small -- the authors acknowledge that it would be.
They did this analysis because: a. this study used a global dataset whereas previous research was biased towards data from western/northern hemisphere countries and b. policy about the internet is frequently made based not on data but on knee-jerk reactions and impressions. It is important to provide evidence that the world is more complicated than INTERNET=BAD or INTERNET=GOOD.
Honestly the bigger issue is once again that reporting on the paper elides the very careful way the authors describe their results. "It's Complicated" doesn't make for very sexy headlines.
posted by DiscourseMarker at 10:25 AM on May 14 [8 favorites]
But finding a very small effect when looking at a bunch of different outcomes and trying to control for a handful of different (and clearly very important) variables is basically the same as finding nothing. It doesn't tell us anything, because it's pretty much impossible to tell the difference between a very small, but real, difference and residual confounding.
posted by ssg at 11:02 AM on May 14 [1 favorite]
posted by ssg at 11:02 AM on May 14 [1 favorite]
"The internet" is such a broad and non-specific thing
This is why I never look at the Internet anymore, and only use my smartphone.
posted by chavenet at 11:07 AM on May 14
This is why I never look at the Internet anymore, and only use my smartphone.
posted by chavenet at 11:07 AM on May 14
But finding a very small effect when looking at a bunch of different outcomes and trying to control for a handful of different (and clearly very important) variables is basically the same as finding nothing.
If this is true (and I don't have the stats background to argue one way or the other) then I think the research is even more interesting, because it still points to public policies that are based on thinking internet = bad are wrong headed and ineffective. But I am also somebody who has long been an advocate for publishing null results. I think sometimes it's even more fascinating when something you think should have a big effect, turns out not to.
posted by DiscourseMarker at 11:40 AM on May 14 [3 favorites]
If this is true (and I don't have the stats background to argue one way or the other) then I think the research is even more interesting, because it still points to public policies that are based on thinking internet = bad are wrong headed and ineffective. But I am also somebody who has long been an advocate for publishing null results. I think sometimes it's even more fascinating when something you think should have a big effect, turns out not to.
posted by DiscourseMarker at 11:40 AM on May 14 [3 favorites]
One thing that occurs to me, whether internet access makes people happier than lack of internet access doesn't really say if the internet is a good thing or not. If the internet has had a corrosive effect on offline social spaces, for example, the downside would effect everyone, but people with internet access would have those downsides mitigated in a way people without access wouldn't, even if they would all be happier with no internet existing at all. I'm not arguing this is the case. I have no idea. I just think that any result is going to be less illuminating than it might seem.
In reality, anything as huge as the internet cannot be reduced to just good or bad. People still argue about the net effect of the industrial revolution today. I doubt we'll have any clear notion of how the internet really effected society in our lifetimes.
posted by The Manwich Horror at 12:49 PM on May 14 [3 favorites]
In reality, anything as huge as the internet cannot be reduced to just good or bad. People still argue about the net effect of the industrial revolution today. I doubt we'll have any clear notion of how the internet really effected society in our lifetimes.
posted by The Manwich Horror at 12:49 PM on May 14 [3 favorites]
People still argue about the net effect of the industrial revolution today
It’s been over 24 hours since the last AI thread so technically we argued about the net effect of the industrial revolution yesterday.
posted by Ryvar at 3:51 PM on May 14 [4 favorites]
It’s been over 24 hours since the last AI thread so technically we argued about the net effect of the industrial revolution yesterday.
posted by Ryvar at 3:51 PM on May 14 [4 favorites]
MetaFilter: dismissive scoffing and intellectually lazy dismissals of things that we haven't even had a chance to read
posted by biogeo at 3:55 PM on May 14 [5 favorites]
posted by biogeo at 3:55 PM on May 14 [5 favorites]
The OII says some of its findings are "consistent" with reported links between social media use and depressive symptoms among young women.
Canaries in the coal mine.
posted by betweenthebars at 4:06 PM on May 14 [1 favorite]
Canaries in the coal mine.
posted by betweenthebars at 4:06 PM on May 14 [1 favorite]
"Attempt to control for" does not mean "actually eliminate influence of".
posted by NotAYakk
This.
The point of control in a scientific sense is not to remove variables, per se, it is to quantify the degree of influence of each variable so it can be properly accounted for in the analysis phase of a study.
Removing a variable completely is just one way to do that. Basically it sets the range of possible influence of that variable to zero. It becomes a known fixed value for the analysis.
posted by Pouteria at 1:50 AM on May 15 [2 favorites]
posted by NotAYakk
This.
The point of control in a scientific sense is not to remove variables, per se, it is to quantify the degree of influence of each variable so it can be properly accounted for in the analysis phase of a study.
Removing a variable completely is just one way to do that. Basically it sets the range of possible influence of that variable to zero. It becomes a known fixed value for the analysis.
posted by Pouteria at 1:50 AM on May 15 [2 favorites]
« Older The Last Pre-Raphaelite | User Inyer Face Newer »
This thread has been archived and is closed to new comments
posted by cupcakeninja at 7:20 AM on May 14 [13 favorites]