U.S. adult survey · 2024 evidence
Who watches health videos on social media?
Reported viewing differs across adult age groups, even when occasional viewing counts.
82.2% of adults aged 18–34 reported watching a health-related video on a social-media site at some frequency in the past 12 months, compared with 35.5% of those aged 75+, in our analysis of the 2024 HINTS survey.
United States · Civilian noninstitutionalized adults 18+ · Within each recorded age band · Includes less-than-monthly viewing
Social Usage analysis of NCI HINTS 7 public-use data, B12e / SocMed_WatchedVid and AgeGrpB. Fieldwork March 25–September 16, 2024. Sampling SEs: 1.96 and 2.32 percentage points respectively. NCI does not endorse this presentation.
Viewing across five adult age groups
Source: NCI HINTS 7, B12e and AgeGrpB. Original weighted estimates and chart. Each group includes missing video answers in its denominator. Accessible table and sampling errors.
Age is unclassified for 7.3% of the weighted sample; those participants are excluded here, without additional age-nonresponse adjustment.
The point estimates are lower in each successive older age band. That describes different groups surveyed in the same wave. It does not show that individuals reduce their viewing as they get older, and we have not tested whether the differences are statistically significant.
These are shares within each age group, not the age composition of the health-video audience. The bars cannot be added or averaged to recover the national estimate. The separate national health-information guide uses all survey participants, including those whose age is unclassified, and compares watching with sharing and interacting.
View the five estimates, sampling errors and sample sizes
| Age band | Weighted % | SE (pp) | Relative SE (%) | Viewing (n) | All in band (n) |
|---|---|---|---|---|---|
| 18–34 | 82.2 | 1.96 | 2.39 | 890 | 1,080 |
| 35–49 | 75.3 | 1.69 | 2.24 | 1,047 | 1,355 |
| 50–64 | 65.5 | 2.22 | 3.39 | 1,145 | 1,728 |
| 65–74 | 49.7 | 2.04 | 4.10 | 762 | 1,466 |
| 75+ | 35.5 | 2.32 | 6.52 | 378 | 1,036 |
n is an unweighted participant count. Percentages and errors use person weights. Standard errors (SEs) are in percentage points; they are not 95% margins of error. Relative SE divides the SE by the estimate and multiplies by 100. No confidence intervals or significance tests are reported.
Public codebook: B12e PDF p84 (printed p83), AgeGrpB PDF p524 (printed p523). NCI Methodology Report §5.5; subgroup percentages and sampling errors calculated by Social Usage.
What counts as watching a health video?
The questionnaire asks about watching a health-related video on a social-media site during the past 12 months. YouTube appears as an example. This does not isolate YouTube viewers, specific channels, short-form video, or every health video available elsewhere online.
Our numerator combines four mutually exclusive answers: Almost everyday; At least once a week; A few times a month; Less than once a month. “Never” and missing answers stay outside the numerator but inside each recorded age group’s denominator. Occasional viewing counts; the result is not a weekly-use or daily-use rate.
Questionnaire: HINTS 7 B12e, PDF p5 (printed p4); exact category labels in the public codebook.
Does viewing mean seeking advice or trusting it?
The item does not establish intentional health-information seeking, trust, belief, treatment decisions or clinical outcomes. It also does not assess whether a video was accurate. The separate health-misinformation question measures perceptions and cannot be inferred from these viewing estimates.
Can this identify a platform or a medical-condition audience?
No. This analysis reads one age recode and one activity field. It does not identify platforms, creators, diagnoses or intersections with other demographic characteristics. The age bands come from NCI’s AgeGrpB variable, not generational labels.
Methods, missing age and dates
Age is unclassified for 613 participants, 7.3% of the weighted sample. They are excluded from these five age domains, with no imputation or additional age-nonresponse adjustment. That limitation matters when interpreting the age pattern.
Sampling and reproducible calculations
HINTS uses an address-based probability sample of U.S. civilian noninstitutionalized adults, with one adult selected by the next-birthday method and paper responses with a web option. Final person weights account for selection, nonresponse and population calibration.
Within each age domain: percentage = 100 × sum of final weights for B12e response codes 1–4 ÷ sum of final weights for all participants in that domain. We retain missing video answers; we do not calculate 100 minus “Never” or add rounded category percentages.
For each of 50 JK1 replicate weights, we recompute both numerator and denominator within the same domain. SE = √[0.98 × sum of (replicate percentage − full-weight percentage)²]. The analysis preserves all 50 aggregate replicate inputs. Sampling errors do not account for every possible recall, coverage or nonresponse bias.
All five domains pass our publication screen: numerator and complementary cell n ≥ 50, domain n ≥ 100 and relative SE ≤ 30%. These are Social Usage safeguards, not thresholds issued by NCI. No respondent rows or small missing-code breakdowns are published.
HINTS 7 Methodology Report, chapter 1 and §5.5; Analysis Recommendations PDF pp4–9 and History Document PDF pp2,8,14 in the original package.
Source version, review and reuse
- Data collection
- March 25–September 16, 2024
- Reference window
- Past 12 months at each response
- Public package
- HINTS7_STATA_20250731; updated August 2025, exact publication day unknown
- Original source retrieval
- October 1, 2026
- Retained rights evidence
- October 7, 2026
- Assisted analysis review
- 2026-10-08
Both the age estimates and their standard errors are original Social Usage calculations. The age domains reproduce NCI’s codebook counts and rounded weighted totals; the national video calculation reproduces our previously reviewed analysis. This report does not refresh the survey or replace the original national release.
Assisted factual and technical review is not independent human editorial approval. HINTS public-use guidance and the NCI reuse policy support this attributed aggregate analysis. This report provides no data download, image export or embed. Respondent records, source graphics and agency logos are not republished.