New media-hyped studies show again why nearly all “research” on teenagers and social media is junk

Journal editors and reviewers need to crack down on authors who wildly exaggerate near-nothing social “science” findings that omit crucial factors

The unconscionable public relations crusade by popular social scientists who tout the barest “statistical significance” created by leaving out key factors entirely, ignoring others, and hyping their pet findings has manufactured a national panic over teenagers and social media at the expense of endangering young people by burying the truly important family and school issues.

A few examples among dozens

“Young people show addictive behavior with phones, social media, video games,” announced the Washington Post, citing a study in JAMA Pediatrics of 4,235 US teenagers that claimed to find: “31.3% had increasing addictive use trajectories for social media and 24.6% for mobile phones” associated with “elevated risks of suicidal behaviors or ideation compared with low addictive use.” Oddly, this study also found more screen time was not associated with worse problems in children – a finding that got downplayed.

The authors’ broadly simplistic definition of “addictive use” included things like, “I feel the need to use social media apps more and more;” “The thought of being without my phone makes me feel distressed;” “I play video games so I can forget about my problems;” etc.

These definitions proved severely problematic because authors admitted they “did not include psychosocial and behavioral factors such as… adverse childhood experiences.” That is, no effort was made to factor in parental and household-adult abuses, parents’ addiction, mental health issues, criminality, etc., “nor neighborhood or school” stresses.

The authors, like 99.9% of those studying social media and teens, were strikingly uncurious about what might really be causing grade-school “problems.” They simply assumed social media was the only factor in children’s lives.

Big mistake. The omitted “other factors” are exactly the ones the U.S. Centers for Disease Control found account for two-thirds of teens’ depression and 89% of suicidal thinking. They also pivotally affect why a grade schooler would use apps more often, feel distressed at not having their phone, and escape into entertainment to forget about problems. Where were the JAMA editors and reviewers?

More nothin’

That omission, as well as trivial findings, also demolish another media-hyped “meta-analysis of 117 studies of nearly 300,000 kids younger than 10½ when the research began” that sought to answer the question, “Does screen use lead to socioemotional problems, and do socioemotional problems lead children to use screens more often?” The study’s real answer: hardly at all for either.

Instead of saying that, the authors blamed “screen time” for children’s “anxiety, depression, hyperactivity and aggression… The association was small but significant, especially for girls.”

“Small” vastly overstates what this study found. Their barely-significant association (β = 0.06, which equals a ‘way-below-“small” Cohen’s d of 0.12, for the wonks) amounts to nothing, especially for such a huge sample, which tends to make a ton of factors theoretically “significant.” For one example of many, the 2021 Centers for Disease Control survey associated eating vegetables with substantially higher levels of teens’ poor mental health than this study found for screen time.

Further, “high screen use isn’t just a cause of problems — sometimes, it’s a symptom,” the study’s lead author said of their second tiny non-finding. “Children who are already struggling emotionally turn to screens, especially video games, as a way to cope or escape.”

What on earth could be causing children to already be “struggling emotionally”? Authors apparently didn’t want to know. They admitted their meta-analysis “couldn’t account for factors such as parenting style or socioeconomic status” or even “pinpoint the effects of social media use on kids’ mental health.” As in scores of previous studies, they just assumed the only thing in teens’ lives are screens, and those screens must be damaging.

When the article got to the press, the shouting began: “Does your child not listen to you? Do they kick and scream when they get angry? You may need to rethink their screen time,” starts off a CNN article titled: “Screen time is both a cause and symptom of kids’ bad behavior, according to new research.”

Worse overhype ensued. A study author quoted in Medical Express exploited their study’s gross omission of family and parental abuses to blame EVERYTHING on social media; "Children are spending more and more time on screens, for everything from entertainment to homework to messaging friends,” said Michael Noetel, Ph.D., an associate professor in the School of Psychology at Queensland University. "We found that increased screen time can lead to emotional and behavioral problems, and kids with those problems often turn to screens to cope."

Enough of this crap!

That is not at all what their zero-finding study “found.” Psych prof Noetel must be unaware that in the real Australia teens live in, social service agencies investigate 1,200 cases of child abuse every day and have removed 45,000 abused children and youths from parents. The pretense that all teen mental health problems are caused by teens’ social media use is professional, official, and media malpractice.

Another press-featured study asking, “could the positive association between screen time and depression be explained by effects of screen time on sleep and white [brain] matter integrity among adolescents?” found just about no screen-time association (β = 0.06, nothing) between white-matter cerebral bundling.

However, they did find a small association between more screen time and loss of sleep (β = 0.23, d = 0.47). While this study did control for income, race, and parents’ education, it failed to control for parental abusiveness and personal troubles, which have been strongly associated with sleep problems and slower brain development. CDC survey numbers showed emotionally and/or violently abused teens are 2.6 times more likely to get 5 or fewer hours of sleep a night compared to non-abused teens.

No matter. The media flooded itself with “social media causing 'an epidemic of sleep problems' in kids” headlines. Much as I personally suspect that grossly excessive time spent in front of a TV, video, phone, or computer screen does have some bad effects on both adolescent and adult brains, authors did not show screen time was the real culprit.

Omitting crucial variables invalidates entire studies

The elementary reason researchers don’t loudly tout to the press that, say, listening to country music causes the high rates of lung cancer in rural White men is that the real cause (high rates of smoking) occurs in this population, one also incidentally likely to patronize Waylon, Willie, and the boys.

A comparison of how important various factors are to teens’ mental health shows the splashes surrounding these new studies – and, even more, previous studies – of social media effects are just silly. They don’t measure or explain anything. I included being sunburned as a mathematically significant factor causing teens’ poor mental health (certainly sunburn can be depressing; wouldn’t that call into question efforts to get teens away from screens and outdoors more?) to show how many factors become technically “significant” with large sample sizes.

Distractions aside, six factors from our only multi-factorial surveys can so far be shown to have real effects on teen mental health and suicidality: parents’ emotional abuse (the umbrella factor for other family dysfunctions), parents’ own poor mental health, parents’ violent abuses, schools viewed as unfriendly places, parents’ drug/alcohol abuse, and violent homes (Table 1). Beyond those, a decade of science shows researchers are wasting our time. If they can’t deal with parental/adult abuses, they should fold up shop and stop distracting us from real issues.

Table 1. What factors really affect teens’ mental health?

Source: CDC 2024.

Of course, useful science examines multiple factors in “multivariate” analysis, in which major factors (i.e., parental emotional and violent abuse) subsume others like violent abuses and pinpoint the best possible explanations for teen mental health. Social media use doesn’t even show up as a minor factor in such multivariate analyses, which consistently show parental, household, and school factors explain nearly all depression/anxiety, suicidal inclinations, self-harm, opiate abuse, etc.

Sure, blaming social media and popular culture is popular and fun. It’s much more entertaining to amuse oneself with psychologist Jonathan Haidt’s and the attorneys general’s salacious regalings of sordid pornography sites, lurking predators, evil media moguls’ confessions, and indignant “save our children!” moralizings, compared to perusing dismal child-abuse reports, statistics, and poverty studies.

But as grownups – such as we are – our first loyalty in our oft-boasted-of mature, executive-functioning cerebral cortexes should be to support the vulnerable and future young even when doing so might make us uncomfortable. Take it from a study author, me, who’s had to report nothing findings (many with considerably higher but still too-small-to-bother-with effect sizes than found in popular studies claiming to associate social media with teen troubles): null findings are not satisfying, but it’s vitally important to report what isn’t causing the problem. Social media researchers need to back off their addiction to zero-zip-zilch hyping.