NTIA research · United States
Watching videos online in the United States
How common is watching video online?
November 2023 · Internet users aged 15+ · Prior six months
of U.S. internet users aged 15+ were reported as streaming or downloading videos at least occasionally during the prior six months in NTIA’s November 2023 survey.
People in the civilian noninstitutionalized population · Within the internet-user population, not all U.S. residents · Source SE: 0.2853 percentage points
Source: NTIA Analyze Table, Nov 2023 / videoUser / adultInternetUser, row 772. Latest actual aggregate wave in the source checked 2026-10-08; no 2025 or 2027 estimate is inferred from a collection notice.
Watching videos online by age
The reported participation estimates are higher in the two younger age bands than in the two older bands. These are within-group shares: the chart does not show how the total video audience is divided by age.
Internet users aged 15+ · November 2023 · Prior six months
Source: NTIA, videoUser, age-group Prop columns. Original Social Usage chart. Read source sampling errors below; apparent differences do not establish causation or statistical significance.
Read exact age estimates and source standard errors
| Age group | Share (%) | Source SE (pp) | Source column |
|---|---|---|---|
| 15–24 | 86.1836 | 0.7952 | age1524Prop |
| 25–44 | 82.6596 | 0.4412 | age2544Prop |
| 45–64 | 68.9616 | 0.4631 | age4564Prop |
| 65+ | 50.5776 | 0.6217 | age65pProp |
An SE is a standard error, not a 95% confidence interval or margin of error. The aggregate table does not supply the selected cell sample sizes or separate confidence intervals. We do not calculate a difference test.
Explore other published single-dimension breakdowns, including income, education and metropolitan status, in the digital-behavior comparison explorer. Each comparison retains its own group denominator; age-by-income or local estimates are not available here.
Broad video participation is not a YouTube audience
NTIA asks whether someone watches videos over the Internet and gives YouTube and Netflix as examples. A Yes answer can cover either service, other services, or more than one. The estimate does not isolate a video platform, a paid subscription or short-form video.
The question captures at least occasional participation over six months. Someone who watches one video and someone who streams every day can both answer Yes. It therefore cannot establish watch time, daily reach, a count of subscriptions or the share of viewing devoted to any platform.
This distinction matters when planning video research. Use the survey to understand the broad activity within its internet-user population; use separately defined company or platform surveys for a particular service. Dividing a platform account total by this percentage would not reveal the number of video viewers.
The original question
What about streaming or downloading videos? [Do you/Does NAME] watch videos over the Internet? Examples include YouTube and Netflix.
The preceding USEINT instruction asks about activities done online at least occasionally during the past six months, at home or elsewhere. The answer options are Yes and No. A household proxy may answer on behalf of the randomly selected internet user.
Source: NTIA November 2023 questionnaire, PDF page 148, VIDEO. The question’s examples and combined activities cannot be separated using the published marginal cells.
Historical snapshots with question changes visible
The 2019 question used “streaming or watching”; the 2021 and 2023 questionnaires use “streaming or downloading.” That wording change is visible below so a lower or higher bar is not automatically read as a like-for-like change in viewing.
U.S. internet users aged 15+ · Not a harmonized growth series
2021 · Video wording changed by 2021
2019 asked about streaming or watching videos; 2021 and 2023 ask about streaming or downloading videos. YouTube and Netflix remain examples, not separate estimates. Original methodology
2022 · CPS population controls changed in 2022
CPS introduced a blended 2020 population base in January 2022. No NTIA-specific adjustment is supplied. Separate survey snapshots are shown; cross-wave growth, interpolation and significance claims are disabled. This is our comparability caution, not an NTIA break flag. Original methodology
Source: NTIA aggregate CSV, national videoUser cells. Separate bars do not interpolate between survey waves. No growth rate, cross-wave difference or causal effect is calculated.
Read historical values, questions and source standard errors
| Wave | Share (%) | Source SE (pp) | CSV row |
|---|---|---|---|
| Nov 2019 | 74.3430 | 0.2426 | 598 |
| Nov 2021 | 70.1169 | 0.2900 | 687 |
| Nov 2023 | 72.8196 | 0.2853 | 772 |
- 2019 questionnaire
- What about streaming or watching videos? (Do you/Does NAME) watch videos over the Internet? Examples include YouTube and Netflix. (1) Yes (2) No Blind (D) Don’t Know Blind (R) Refused Original PDF page 155
- 2021 questionnaire
- Random respondents What about streaming or downloading videos? [Do you/Does NAME] watch videos over the Internet? Examples include YouTube and Netflix. (1) Yes (2) No Original PDF page 149
- 2023 questionnaire
- What about streaming or downloading videos? [Do you/Does NAME] watch videos over the Internet? Examples include YouTube and Netflix. Original PDF page 148
The CSV contains earlier rows too. Those are not imported into this history because earlier questionnaire and sampling changes have not been harmonized. A source variable with the same name is not sufficient proof of comparability.
Methods, dates and reuse
Population, weights and uncertainty
NTIA estimates from one randomly selected internet user aged 15+ per household, using PWPRMWGT; proxy responses allowed. Source standard errors use successive difference replication with 160 respondent replicate weights.
The November 2023 fieldwork ran from 12–18 November. The selected population is internet users aged 15+ in the civilian noninstitutionalized population of the 50 states and District of Columbia. It is neither every resident aged 15+ nor all adults aged 18+.
Social Usage reads published aggregate proportions and source proportion standard errors, then multiplies each by 100. This produces percentages and standard errors in percentage points. We process no individual responses, multiply no survey shares by unrelated population totals, and create no small-area estimates.
Source standard errors describe sampling uncertainty. Reporting, nonresponse, imputation and coverage errors can also affect these survey estimates. The five online activities in this research collection overlap and cannot be summed into a unique audience.
Dates, source selection and permitted downloads
- Latest data period
- November 2023
- Aggregate update published
- 2024-06-06 · This is not the original publication date of every older wave.
- Source access checked
- 2026-10-08
- Assisted release review
- 2026-10-08 · Independent human review is not claimed.
Source: NTIA Internet Use Survey, U.S. Census Bureau Current Population Survey. Selection and presentation by Social Usage; no agency endorsement implied.
Selected NTIA federal aggregates and our original attributed charts and tables use the recorded federal-work basis. Agency logos, third-party graphics, respondent records and a blanket licence over other sources are excluded. Download the reviewed aggregate data with source credit and scope.
Data Commons aids source discovery and auditing. These percentages and standard errors come directly from the original NTIA Analyze Table because its survey universe and uncertainty are explicit. They are not computed from Data Commons counts divided by an unrelated population.