U.S. high-school survey · 2023 evidence
Social media and mental health: what the survey shows.
A national school survey found an association between frequent social-media use and persistent sadness or hopelessness. It cannot establish which came first or whether one caused the other.
of students using social media at least several times a day reported persistent sadness or hopelessness, compared with 31.9% among those using it less often, including nonusers.
People · U.S. public and private schools · Grades 9–12 · 2023 YRBS · Outcome in the previous 12 months · Known-data responses, weighted
Source: Young et al., CDC MMWR, corrected Table 4. Source 95% confidence intervals: frequent 40.4–44.8%; less frequent 25.3–39.3%. This is a reported experience, not a diagnosis of depression.
Source: CDC. Original materials are available free of charge on CDC.gov. Use by Social Usage does not imply endorsement by CDC, HHS or the U.S. Government.
Two groups, one survey question.
The sadness-or-hopelessness item asks about a period of at least two consecutive weeks when feelings were present almost every day and led the student to stop some usual activities. Its recall window is the past 12 months.
Definition: CDC Table 2, persistent sadness or hopelessness.
Source: CDC, corrected Table 4, sadness row. Original Social Usage graphic. Estimates and intervals are CDC's; the bars do not show a causal effect.
The displayed difference is 10.7 percentage points, calculated by Social Usage as 42.6% − 31.9%. This descriptive subtraction uses rounded source percentages. It has no separately calculated confidence interval and is not the study's adjusted comparison.
View percentages and confidence intervals for all four outcomes
| Reported outcome | Frequent use % (95% CI) | Less frequent / nonuse % (95% CI) |
|---|---|---|
| Persistent feelings of sadness or hopelessness | 42.6 (40.4–44.8) | 31.9 (25.3–39.3) |
| Seriously considered attempting suicide | 20.2 (18.8–21.8) | 18.7 (12.8–26.6) |
| Made a suicide plan | 16.6 (15.1–18.2) | 17.5 (10.3–27.9) |
| Attempted suicide | 9.5 (8.4–10.8) | 9.5 (6.6–13.5) |
Frequent means at least several times a day. The comparison combines less-frequent users and nonusers. Denominators use known data and vary by item; exact outcome-specific sample sizes are not supplied. Students can report more than one outcome, so these percentages must not be added.
Source: CDC Table 4 and its denominator footnote. All percentages and intervals above are source-reported.
What does the adjusted comparison add?
CDC's adjusted prevalence ratio for persistent sadness or hopelessness is 1.35 (95% CI 1.23–1.47). The model compares the two frequency groups while adjusting for age, race and ethnicity, sex, and sexual identity.
A prevalence ratio compares how common an outcome is in one group relative to another. A value of 1 means equal prevalence. This ratio is a different measure from a percentage or a percentage-point difference. Adjustment for these recorded characteristics does not make the association causal.
Source: CDC Table 4, aPR column and adjustment footnote; analysis method.
Inspect unadjusted and adjusted ratios, including source significance flags
| Outcome | PR (95% CI) | Adjusted PR (95% CI) | CDC adjusted flag |
|---|---|---|---|
| Persistent sadness or hopelessness | 1.33 (1.07–1.65) * | 1.35 (1.23–1.47) * | Significant * |
| Seriously considered attempting suicide | 1.08 (0.75–1.55) | 1.21 (1.06–1.37) * | Significant * |
| Made a suicide plan | 0.95 (0.58–1.55) | 1.16 (1.00–1.35) * | Significant * |
| Attempted suicide | 1.00 (0.70–1.43) | 1.11 (0.89–1.39) | Not flagged |
* Preserves CDC's significance marker. The plan row is flagged even though its displayed lower bound rounds to 1.00; CDC explicitly explains this rounding issue. We do not recompute significance from rounded intervals. PR is unadjusted; adjusted PR uses the four covariates listed above. Neither is an odds ratio or an estimate of future individual risk.
Source: CDC corrected Table 4, PR/aPR columns and significance footnote.
The findings are not uniform across outcomes. For reported suicide attempts, the adjusted ratio is 1.11 (95% CI 0.89–1.39), and CDC does not flag it as statistically significant. That does not establish that there is no association. For making a plan, CDC reports 1.16 (1.00–1.35) and retains its significance flag despite the rounded lower bound.
Keep the question with the statistic.
Read the exact outcome questions
Persistent sadness or hopelessness
During the past 12 months, did you ever feel so sad or hopeless almost every day for 2 weeks or more in a row that you stopped doing some usual activities?
Analysis coding: Yes versus no.
Seriously considered attempting suicide
During the past 12 months, did you ever seriously consider attempting suicide?
Analysis coding: Yes versus no.
Made a suicide plan
During the past 12 months, did you make a plan about how you would attempt suicide?
Analysis coding: Yes versus no.
Attempted suicide
During the past 12 months, how many times did you actually attempt suicide?
Analysis coding: One or more times versus zero times.
Source: CDC Table 2, printed p25 / PDF p3. These are self-reported experiences; the questions do not establish a clinical diagnosis.
Is “less frequent” the same as no social media?
No. It combines nonusers with students reporting use below several times a day. The comparison therefore cannot isolate abstaining from social media, quitting, a particular platform or a particular number of hours. The frequency question has no explicit recall window; the outcomes ask about the preceding 12 months.
Source: CDC frequency categories and limitations.
Does this measure depression, anxiety or perceived harm?
The mental-health item measures persistent sadness or hopelessness under the stated duration and activity limits. It is not a diagnostic assessment of depression or anxiety. The selected analysis also does not ask students whether they believe social media harmed their mental health. A perceived-impact survey would answer a separate question.
Question boundaries: CDC Table 2.
Can we compare this with the 2025 hourly-use figures?
The separate high-school frequency guide uses an hourly threshold. This analysis uses at least several times a day and outcomes from 2023. Combining their values would change both the question and time period. Adult visits and teen duration surveys have different populations and definitions too.
Methods, uncertainty and source history.
Sample, denominator and statistical method
The national 2023 YRBS is a cross-sectional survey of U.S. public- and private-school students in grades 9–12 across the 50 states and D.C. There were 20,103 respondents overall and 15,203 answers to the social-media item. Neither count is the exact sample size of every outcome or adjusted model; those counts are not supplied in the selected table.
CDC uses known data for each percentage. Missing answers are not No answers. Source sample weights account for the complex design and nonresponse; Taylor series linearization supplies uncertainty estimates. PRs come from logistic regression with predicted marginals, with the adjusted models adding the stated demographic variables. We retain the published confidence intervals and source flags.
The design cannot determine causal direction. Self-report, missing data and the school-based population limit interpretation; findings do not describe all U.S. teenagers, adults or clinical patients. Confidence intervals do not capture every potential source of bias.
Source: CDC methods, Table 4 footnotes and limitations.
Data dates, correction and editorial review
- Data through
- 2023 survey; exact fieldwork boundaries are not established in this selection.
- Publication
- MMWR issue: October 10, 2024. Page metadata first-online date: October 8, 2024.
- Source correction
- The current report is marked corrected and last reviewed November 8, 2024. An exact correction day is not supplied. We preserve the corrected table and significance markers.
- Retrieved / reviewed
- 2026-10-07 / 2026-10-07. Assisted evidence and calculation review; no independent human editorial approval is claimed.
This is a dated analysis, not a current-year prevalence estimate or an exhaustive review of mental-health research. The selected source measures associations and reported experiences; this page does not provide a treatment recommendation.
Source credit and reuse
Selected CDC statistical facts support this original guide. The agency-materials policy, retained from its October 6 check, has exceptions for protected third-party material. We use no source artwork, logo, respondent records or blanket licence over the article. This record has no CSV, image-export or embed distribution.
Source: CDC. Original materials are available free of charge on CDC.gov. Use by Social Usage does not imply endorsement by CDC, HHS or the U.S. Government.