CAN A STUDY LOCATE A PROBLEM IN THE INDIVIDUAL SIMPLY BECAUSE IT CANNOT ACCESS EVIDENCE TO THE CONTRARY?

Oct. 2026

Published in 2026 in the International Journal of Mental Health and Addiction, Dorrestein, Nutley, and Thorell’s study of 1,820 Swedish high school and university students finds that motives—escaping worries, comparing oneself with others, compensating for unmet social needs—and content focused on appearance or negativity, rather than screen time, account for 33–40% of the variance in psychosocial problems. Its stated conclusion: move beyond screen time and target users’ individual motivations.

The problem lies in what the study takes for granted: that motives and content are psychological variables belonging to each adolescent, measurable and amenable to targeted intervention. Yet neither the content nor the motive it activates comes from nowhere. Both are the visible products of an invisible infrastructure. Platforms do not simply distribute neutral material for users to interpret freely. They continuously run millions of A/B tests on population segments to discover which content—anxiety-provoking, comparative, flattering—maximizes time spent and return visits. The motive “escaping worries” has already been selected and reinforced by a platform that has statistically established its value for engagement. In 2014, Kramer, Guillory, and Hancock manipulated the feeds of almost 700,000 users without consent, showing that a social network could experimentally shift a population’s mood: evidence that architecture, as well as visible content, shapes affect.

This inversion has a name: opaque machinic epistemology. Knowledge about users—their motives, vulnerabilities, and what keeps them engaged—is inseparable from the system that has shaped them to keep them there. Psychology continues to treat these traces as stable traits, without access to the platforms’ proprietary experiment logs.

As long as researchers cannot audit the A/B tests themselves, the literature will continue, in good faith, to treat effects engineered at scale as individual problems. Each study recommending that we “target motivations” will unwittingly strengthen the platform that shaped those motivations, rather than challenge its role.

Liviu Poenaru

References

Dorrestein, Melissa, Sissela B. Nutley, and Lisa B. Thorell. 2026. “Screen Time, Addictive Use of Social Media, Motives for Social Media Use and Social Media Content.” International Journal of Mental Health and Addiction 24 (3): 2051–2067. https://doi.org/10.1007/s11469-025-01491-5.

Kramer, Adam D. I., Jamie E. Guillory, and Jeffrey T. Hancock. 2014. “Experimental Evidence of Massive-Scale Emotional Contagion through Social Networks.” Proceedings of the National Academy of Sciences 111 (24): 8788–8790.