How were the numbers on this site collected?
Every chart on this site comes from a published survey, government table or legislative count, and each one is listed below with who was asked, how the reading was collected, when it was fielded, when it was published and what it cannot tell you. Nothing is estimated, smoothed or interpolated. Where a source does not publish a detail, that is stated rather than filled in.
Rules that apply to every chart
- A missing year is drawn as a gap, never as a zero and never as an interpolated point.
- Only readings verified against a primary source are plotted, and each one is recorded in the verification ledger.
- Where the question wording changed between waves, the change is named and the affected line is treated as a rough trend.
- Shares asked only of a platform users describe that audience, not the public.
- Every chart can be downloaded as a spreadsheet or an image, and both carry the source line.
Share of U.S. adults who often get news from each platform
- Source organization
- Pew Research Center
- Who was measured
- U.S. adults
- How it was collected
- Nationally representative panel survey, read from the published chart data file.
- Question wording
- Whether the respondent often gets news from each named platform.
- Wording changes across years
- Print was asked as print publications from 2020 to 2024 and as printed newspapers or magazines in 2025, so it is not plotted. Email newsletters and AI chatbots were first asked in 2025.
- Field dates
- August 18 to 24, 2025, annual waves back to 2020
- Published
- September 25, 2025
What these numbers cannot tell you
- A respondent can appear on several lines, so the lines do not sum to 100.
- Often is a self assessment rather than a measured frequency.
See the chart on the facts page · Primary source: Pew Research Center, News Platform Fact Sheet (September 25, 2025)
Share with a lot or some trust in the information from each source
- Source organization
- Pew Research Center
- Who was measured
- U.S. adults
- How it was collected
- Repeated cross sectional surveys, each value combining a lot and some trust.
- Question wording
- How much trust the respondent has in the information that comes from national news organizations, local news organizations and social media sites.
- Wording changes across years
- The social media item named different platforms across waves, moving from Facebook and Twitter to Facebook, X and YouTube. The 2016 social item was asked of web users only. Pew fielded twice in 2019 and the later fielding is plotted.
- Field dates
- March 10 to 16, 2025 for the most recent plotted point, waves back to 2016
- Published
- October 29, 2025
What these numbers cannot tell you
- Years without a published reading are left as gaps rather than filled.
- A later September 2025 wave reported headline figures without the itemized topline, so it is not plotted.
- Wording changes mean the social media line is a rough trend rather than an identical question over time.
See the chart on the facts page · Primary source: Pew Research Center, trust in information from news organizations and social media sites
Would be unhappy if their child married across party lines
- Source organization
- Multiple survey houses, most recently PRRI
- Who was measured
- U.S. adults, by party
- How it was collected
- Separate surveys by different organizations, each point labeled with its house.
- Question wording
- Displeased in 1960, upset in 2008 and 2010, and unhappy in 2019, asked about a child marrying across party lines.
- Wording changes across years
- The verb changes across every fielding, which is why the points are labeled rather than joined as one question.
- Field dates
- 1960, 2008, 2010 and 2019
- Published
- PRRI American Values Survey, 2019
What these numbers cannot tell you
- Party ordering flips between houses, which is a house and wording effect rather than a documented reversal.
- This is a set of comparable readings, not a single tracked series.
See the chart on the facts page · Primary source: PRRI, 2019 American Values Survey
Trust in mass media (great deal / fair amount)
- Source organization
- Gallup
- Who was measured
- U.S. adults
- How it was collected
- Annual telephone survey within the Media Use and Evaluation trend.
- Question wording
- How much trust and confidence the respondent has in the mass media to report the news fully, accurately and fairly, combining a great deal and a fair amount.
- Field dates
- Annual since 1997, with 2006 skipped, plus 1972 and 1976 anchor readings
- Published
- Gallup trend page, updated annually
What these numbers cannot tell you
- Mass media is a single broad object, so the number does not separate outlets a respondent uses from ones they do not.
- The two pre 1997 anchors come from a different fielding era than the annual trend.
See the chart on the facts page · Primary source: Gallup, Media Use and Evaluation trend
Partisans rating the opposing party at 0 (coldest) on the 0-100 feeling thermometer
- Source organization
- American National Election Studies, restated in peer reviewed work
- Who was measured
- U.S. voters in presidential election cycles
- How it was collected
- Feeling thermometer ratings of the opposing party, share giving an exact zero.
- Question wording
- A zero to one hundred feeling thermometer rating of the opposing party.
- Field dates
- 2016 and 2020 cycles
- Published
- Iyengar and Krupenkin 2018, Tyler and Iyengar 2023
What these numbers cannot tell you
- Only cycles restated numerically in published text are plotted, so interior cycles are absent rather than interpolated.
- An exact zero is one point on a scale, not a measure of intended behavior.
See the chart on the facts page · Primary source: Iyengar & Krupenkin (2018) / Tyler & Iyengar (2023), ANES exact-zero out-party thermometer share (text-verified endpoints only)
U.S. adults who regularly get news on TikTok
- Source organization
- Pew Research Center
- Who was measured
- U.S. adults
- How it was collected
- Annual fact sheet and report readings, each year verified against a Pew publication.
- Question wording
- Whether the respondent regularly gets news on TikTok.
- Field dates
- Annual waves 2020 through 2025
- Published
- September 25, 2025 for the most recent reading
What these numbers cannot tell you
- Growth reflects both platform adoption and news use on the platform, which the single number does not separate.
See the chart on the facts page · Primary source: Pew Research Center, Social Media and News Fact Sheet (2025)
"Most people can be trusted", General Social Survey
- Source organization
- NORC at the University of Chicago, General Social Survey
- Who was measured
- U.S. adults
- How it was collected
- In person and mixed mode survey waves, weighted with WTSSCOMP.
- Question wording
- Whether most people can be trusted or you cannot be too careful in dealing with people.
- Field dates
- Selected waves across the GSS series
- Published
- NORC GSS Data Explorer
What these numbers cannot tell you
- Intermediate waves are omitted where a primary extraction was not on hand.
- The 2020 and 2022 waves are not yet verified against a primary source.
- The 2020 wave shifted collection mode, which can move a trend independently of opinion.
See the chart on the facts page · Primary source: NORC, General Social Survey Data Explorer, "trust" variable
"Very unfavorable" view of the other party (Pew Research)
- Source organization
- Pew Research Center
- Who was measured
- U.S. partisans
- How it was collected
- Cross year comparison published inside a partisan hostility report.
- Question wording
- Whether the respondent has a very unfavorable view of the opposing party.
- Field dates
- Years restated numerically in the 2022 report
- Published
- August 9, 2022
What these numbers cannot tell you
- Intermediate 2016 and 2019 waves exist in other Pew reports and are not yet verified in this ledger.
See the chart on the facts page · Primary source: Pew Research Center, Rising Partisan Antipathy (2022)
Split-ticket U.S. House districts (President vs. House winner)
- Source organization
- Brookings Institution, Vital Statistics on Congress
- Who was measured
- Congressional districts in presidential election years
- How it was collected
- Count of districts carried by a presidential candidate of one party and a House candidate of the other.
- Question wording
- Not a survey. This is a tabulation of certified election results.
- Field dates
- Presidential years 1972 through 2020
- Published
- Table 2-16, updated November 2024
What these numbers cannot tell you
- The 2024 cycle is not yet published in this table.
- District boundaries change with redistricting, so the denominator is not identical across decades.
See the chart on the facts page · Primary source: Brookings, Vital Statistics on Congress, Table 2-16 (Nov 2024)
Public laws enacted per Congress
- Source organization
- GovTrack.us
- Who was measured
- Public and private laws enacted by each Congress
- How it was collected
- Count of enacted legislation per Congress, plotted on the first year of each Congress.
- Question wording
- Not a survey. This is a legislative count.
- Field dates
- By Congress, across the GovTrack series
- Published
- GovTrack statistics page, updated continuously
What these numbers cannot tell you
- Fewer and larger bills mean a lower count does not necessarily indicate less legislative activity.
- The chart shows raw law counts and does not weight a law by its scope.
See the chart on the facts page · Primary source: GovTrack.us, Statistics and Historical Comparison
Trust the government in Washington to do what is right (just about always / most of the time)
- Source organization
- Pew Research Center, pooling several survey houses
- Who was measured
- U.S. adults
- How it was collected
- Pooled trend drawing on Pew, National Election Studies, Gallup, ABC and Washington Post, CBS and New York Times, and CNN surveys, with every point labeled by house.
- Question wording
- How often the respondent trusts the government in Washington to do what is right, combining just about always and most of the time.
- Field dates
- 1958 through 2025
- Published
- December 4, 2025
What these numbers cannot tell you
- Values are individual polls rather than the smoothed moving average Pew draws as a line.
- One reading per year is selected by a fixed house order, so a different selection rule would move some points.
See the chart on the facts page · Primary source: Pew Research Center, Public Trust in Government: 1958-2025 (Dec 2025)
Party identification, Gallup annual averages
- Source organization
- Gallup
- Who was measured
- U.S. adults
- How it was collected
- Annual averages built from every Gallup telephone poll fielded that year, more than 13,000 interviews in 2025.
- Question wording
- Whether the respondent identifies as a Republican, a Democrat or an independent.
- Field dates
- Annual averages across the Gallup party affiliation series
- Published
- January 2026 for the most recent reading
What these numbers cannot tell you
- Only years Gallup restates numerically in published text are plotted, and missing lines stay missing rather than interpolated.
- Leaned party preference is a separate Gallup measure and is not plotted here.
See the chart on the facts page · Primary source: Gallup, New High of 45% in U.S. Identify as Political Independents (Jan 2026)
Average confidence in 14 major U.S. institutions
- Source organization
- Gallup
- Who was measured
- U.S. adults
- How it was collected
- June battery, reported as the average across the 14 institutions tracked since 1993.
- Question wording
- How much confidence the respondent has in each institution, combining a great deal and quite a lot.
- Field dates
- Each June
- Published
- July 2026 for the most recent reading
What these numbers cannot tell you
- Only readings Gallup restates numerically are plotted, so interior years of the trend are absent rather than interpolated.
- A 14 institution average hides movement inside individual institutions.
See the chart on the facts page · Primary source: Gallup, Confidence in U.S. Institutions Remains Near All-Time Low (July 2026)
Partisan gap in average confidence in major U.S. institutions
- Source organization
- Gallup
- Who was measured
- U.S. adults, by party
- How it was collected
- Spread between Republican and Democratic average confidence, published alongside the overall average each July.
- Question wording
- The same confidence battery, split by party identification.
- Field dates
- Each July
- Published
- July 2026 for the most recent reading
What these numbers cannot tell you
- The release carries a margin of error of about 4 points on a sample of roughly 1,001 adults, so a two year move of this size should be read as flat.
- The direction of the gap flips with control of the presidency, so the sign carries less meaning than the size.
See the chart on the facts page · Primary source: Gallup, Confidence in U.S. Institutions Remains Near All-Time Low (July 2026)
U.S. adults who regularly get news on each social media site
- Source organization
- Pew Research Center
- Who was measured
- Users of each named site
- How it was collected
- Annual fact sheet, values taken from the published underlying data file.
- Question wording
- Whether the respondent regularly gets news on a site they already use.
- Field dates
- Annual waves since 2020
- Published
- September 2025 for the most recent reading
What these numbers cannot tell you
- The chart is capped at the four highest platforms in the 2025 reading, for legibility rather than sourcing.
- Only people who use a site were asked, so the lines are not comparable across audience sizes.
See the chart on the facts page · Primary source: Pew Research Center, Social Media and News Fact Sheet (Sept 2025)
Reported voting rate, November elections
- Source organization
- U.S. Census Bureau
- Who was measured
- The voting age population
- How it was collected
- Current Population Survey November Voting and Registration Supplement, Table A-1.
- Question wording
- Whether the respondent voted in the November election.
- Field dates
- November 1964 through November 2024
- Published
- 2024 historical time series release
What these numbers cannot tell you
- Voting is self reported, which runs above validated turnout.
- Values use the total voting age population base for comparability rather than the citizen base the Census reports from 1978 onward.
- Presidential and midterm years alternate, so the sawtooth is the election calendar rather than a trend.
See the chart on the facts page · Primary source: U.S. Census Bureau, Historical Reported Voting Rates, Table A-1 (2024)
Reported registration versus reported voting, November elections
- Source organization
- U.S. Census Bureau
- Who was measured
- The voting age population
- How it was collected
- The same Table A-1, plotted as two published lines rather than a computed difference.
- Question wording
- Whether the respondent registered and whether the respondent voted.
- Field dates
- November 1964 through November 2024
- Published
- 2024 historical time series release
What these numbers cannot tell you
- Registration was not collected on the same basis in 1964, so that year carries a voting value only.
- Both lines are self reported.
See the chart on the facts page · Primary source: U.S. Census Bureau, Historical Reported Voting Rates, Table A-1 (2024)
What partisans believe about the other side, next to what the other side actually says
- Source organization
- More in Common, fielded with YouGov
- Who was measured
- U.S. partisans after the 2018 midterms
- How it was collected
- Single study pairing the share of one party holding a view with the share the other party estimates.
- Question wording
- The full item wording is carried in the primary source rather than restated here.
- Field dates
- After the 2018 midterm election
- Published
- June 2019
What these numbers cannot tell you
- This is a single study rather than a tracked trend, so it carries no year over year reading.
- Later replications exist and are not merged into these figures.
See the chart on the facts page · Primary source: More in Common, The Perception Gap (June 2019)
Sourced facts · Where Americans get their news · Bibliography