A shared vocabulary for better comparisons

The same word can hide a different measurement. Use the source’s definition, not just the label.

U.S. social-network use · NTIA socialNetworkUser

Reported use of online social networks at least occasionally during the previous six months. The denominator is internet users aged 15+ within the selected group or state, not all U.S. adults. Fieldwork: November 12–18, 2023. Source, wording and uncertainty.

Adult visit frequency · NCI HINTS SocMed_Visited

Frequency of visiting a social-media site during the previous 12 months among U.S. civilian noninstitutionalized adults aged 18+. The source codebook retains item nonresponse in its denominator. “Almost everyday” is a response category, not an exact daily count. Definition, categories and calculation.

High-school frequency · CDC YRBS hourly use

Share of students in grades 9–12 reporting social-media use about once an hour or more often, among valid item responses. This differs from the broader several-times-daily threshold. A frequency response cannot be converted into hours spent. Source question and matching history.

Reported time · Gallup selected-platform mean

Mean reported typical-day hours across seven tested platforms among U.S. teenagers aged 13–19 in summer 2023. This is a dated survey estimate, not a device log or an all-adult average. Scope, source and limitations.

Social-network participation · Eurostat I_IUSNET

Use of the internet for private purposes to participate in social networks in the last three months. This release uses PC_IND: percent of all individuals in the selected age band. Source table.

Survey adoption · Pew “ever use”

Share of all U.S. adults reporting ever using a named website or app. It does not imply daily or monthly activity. Source and question context.

Active users · DAU / MAU

An operating measure with a daily or monthly window. Definitions can vary by company, activity, counted entity, deduplication method and averaging convention. Check the original disclosure before comparing.

Denominator

The population against which a percentage is calculated. “All adults” and “adult internet users” are different denominators.

Standard error

A measure of sampling uncertainty for an estimate. It is not itself a 95% margin of error and does not capture every source of survey error. NTIA supplies standard errors in its tables; Gallup supplies one for its mean time estimate. HINTS standard errors here are SocialUsage analysis using the source’s replicate weights. Percent estimates use percentage-point errors; the time estimate uses hours. HINTS calculation and uncertainty.

Confidence interval

An interval produced by a statistical method to express sampling uncertainty. The teen-frequency page preserves CDC’s published 95% intervals. A standard error alone is not a confidence interval, and overlapping intervals are not a substitute for a survey-design-aware comparison test.

Percentage point

The arithmetic difference between two percentages: 60% minus 40% is 20 percentage points. A relative change would be 50%; they are different calculations.

Reference period

When the underlying observation applies. For a survey it can include both a fieldwork period and a recall window.

Source publication / retrieval / evidence-check dates

When the source was released, when we accessed it and when its evidence was checked. None of these changes the reference period.

Break in series · b

A source flag indicating a discontinuity, potentially from a changed method or definition. Do not calculate continuous growth across it.

Confidential · c; low reliability · u

A confidential value is unavailable for this extract. A low-reliability value is retained with its warning. Neither should silently become an ordinary unflagged observation.

Eurostat flag meanings follow the dataset status metadata in the source response. See dataset scope and lineage.