2023 CDC Youth Risk Behavior Survey · Students' Social Media Addiction Dataset
A data story in two chapters: what the CDC's 2023 national survey reveals about the pressures converging on teen mental health, and what a global social-media survey adds about the digital layer on top.
Scroll to begin1. Meet the Cohort
20,103 U.S. high school students, weighted to be nationally representative. Before any finding, the denominator matters: who is in this survey, and in what proportions?
2. The Pressure Map — Chapter One: Broad Health Pressure
Poor mental health sits at the centre of a web. Each node is a pressure; each link is how much the poor-mental-health rate rises when that pressure is present. The web is not symmetric — some pressures are far more predictive than others.
Sleep loss, bullying, and low school connectedness have the strongest associations with poor mental health. But national averages hide who carries which pressures — that comes next.
3. Not Everyone Carries the Same Load
The national average dissolves when you separate by sex, grade, age, or race. Some groups carry heavier profiles on almost every axis; others diverge sharply on one or two. Each cell shows how far above or below the national baseline that subgroup sits.
Female students show the most extreme deviation from the national baseline — highest on nearly every pressure. That pattern repeats across grade levels, which raises a more specific question: how does the landscape shift as students move through high school?
4. High School Changes the Equation
Not all pressures stay equally prominent across high school. Substance use rises; school connectedness falls; sleep loss stays stubbornly high. Each line tracks how one pressure's rank shifts from 9th to 12th grade — the same cohort navigating a changing landscape.
Some pressures worsen with time; some ease. But this is still a rank picture — it doesn't say which individual factors have the sharpest impact on mental health. That comes next.
5. The Steepest Slopes
Each line connects the poor-mental-health rate when a risk is absent (left) to the rate when it is present (right). The steepest slopes are the factors that most reliably separate students in poor mental health from those who are not. Click a factor to select or deselect it for the simulator below.
No single risk is a switch — but bullying and insufficient sleep produce the largest gaps. The real question is what happens when these don't arrive alone.
6. Pressure Stacks
Students rarely carry just one pressure. When two or three arrive together, the poor-mental-health rate doesn't add — it compounds. Toggle pressures on or off to see how the rate climbs as they accumulate.
Each row is a pressure bucket (0 to N selected pressures). Row width is proportional to the bucket's share of the student population; the red portion shows what fraction of those students reports poor current mental health. The rate indicator on the right tracks how that fraction rises as pressures accumulate.
Students with zero of the selected pressures report poor mental health at roughly 8%. Students carrying all five report it at roughly 72%. The pressures work together, not independently — accumulation is the mechanism.
7. One More Dimension: Relationship Safety
Safety, consent, and agency in relationships are part of the same health system. Violence and coercion produce mental-health gaps comparable to the largest in the dataset. Among sexually active students, prevention and agency rates reveal where support is — and isn't — reaching.
Sexual dating violence produces one of the largest single mental-health gaps in the entire dataset. The pressure story is incomplete without it. That closes Chapter One — but one pressure thread runs into the next chapter: the phone.
8. Before We Leave the YRBS: The Phone Signal
The YRBS includes questions about phone use, social media, and screen time. These correlate with sleep loss and mental health — but the survey doesn't go deep enough to say which platforms, or how habits differ globally. For that, we need a different dataset.
Phone use and sleep loss are tightly correlated in the YRBS — and both connect to poor mental health. Chapter Two examines that link with a survey built specifically to measure it.
Chapter Two: The Social Media Layer
Chapter One showed that phone use correlates with the pressures already in play. Now we switch to a dataset built specifically to measure it: 705 university-age students from 110 countries, answering a focused questionnaire on social-media habits, sleep, mental health, conflicts, and academic impact. The goal is not to replace the national picture — it is to inspect one modern pressure in much greater detail.
9. Where the Habit Lives
Different countries use different platforms as their primary channel, and usage intensity varies widely. Country fill shows overall exposure; platform badges show the local leader. Click a platform pill to switch into country-by-country percentage mode for that platform.
10. Platform Fingerprints
Each platform's typical user has a different shape across six pressure axes: daily usage, addiction score, sleep loss, conflicts, mental health, and academic impact. Toggle platform pills to compare fingerprints directly.
The platform you use is a stronger predictor of your pressure profile than simply how many hours you spend online. TikTok and Instagram users stretch outward on nearly every axis at once; LinkedIn traces a much smaller figure near the centre. The contrasts hold across six independent metrics — which platform attracts which kind of user and which platform shapes behavior are harder to separate, but the differences are real. This raises the next question: does the pressure ease as students get older?
11. Does It Ease With Age?
Six indicators are converted onto the same 0–100 pressure scale, where higher always means worse. Follow the colored currents across age; dot size shows how many students sit behind each point — the largest dots carry the most evidence.
The data is cautiously optimistic: addiction, daily usage, and academic impact all drift downward with age. But the signal is clearest in the 19–22 band where most of the sample sits. The youngest and oldest ages should be read as directional. One question remains: where exactly does the risk concentrate?
12. The Risk Zone
Each point is one student. X-position is daily usage, y-position is sleep hours, color is addiction score, and a dark outline marks students reporting academic impact. The shaded terrain shows where high-addiction cases concentrate — a landscape of compounding pressure.
Heavy use alone and short sleep alone each carry risk. But the steepest drop in wellbeing appears where both arrive together — the same compounding logic that ran through Chapter One. Single pressures are manageable. Multiple pressures arriving at the same time are not. That is the through-line of this entire story.