Skip to main content
Media & Representation

The Coverage Record

What the research actually establishes about how American media represents Muslims — twelve studies with their methods, their numbers, and their limits — plus the language and editorial choices that a newsroom can change tomorrow.

This is a history site. It exists because the 250-year record it documents is routinely narrowed to a few decades and a few frames. This page shows where that narrowing is measurable, and what the measurements do and do not support.

Read the evidenceHow we source claims →

What this page is, and is not

It is
  • A cited index of published research on coverage patterns, with each study's method stated.
  • A working reference for reporters, editors, students, and anyone checking a claim they were sent.
  • A record that shows disagreement between studies rather than resolving it by preference.
It is not
  • A rating of individual journalists, outlets, or stories. Nothing here scores a byline.
  • An automated sentiment dashboard. See the limits section for why we did not build one.
  • A claim that coverage causes every outcome it correlates with. Association is labeled as such.
Evidence hub

Twelve studies, with their methods attached

Each entry states what was counted, over what period, by whom. Evidence labels follow the same five-tier vocabulary used across this site — see the source hierarchy. Where a source is a research institute or a monitoring body with a stake in the finding, that is stated on the card rather than left for the reader to discover.

Showing 12 of 12 studies.

  • 2019News coverage

    Why Do Some Terrorist Attacks Receive More Media Attention Than Others?

    Erin M. Kearns, Allison E. Betus & Anthony F. Lemieux — Justice Quarterly 36(6)

    Strong evidenceScholarly work
    What was measured
    All 136 terrorist attacks recorded in the United States between 2006 and 2015, matched against coverage in LexisNexis Academic and CNN.com.
    What it found
    Controlling for target type, fatalities, and whether the perpetrator was arrested, attacks by Muslim perpetrators received on average 357% more coverage than other attacks. Perpetrator religion was the largest single predictor of coverage volume. In the same sample, Muslim perpetrators accounted for 12.4% of attacks and 41.4% of the coverage.
    Why it matters
    Readers estimate how common a threat is from how often they encounter it. A 357% coverage premium teaches a threat profile the underlying incident record does not support.
  • 2019News coverage

    Media Portrayals of Minorities Project — 2018 newspaper coverage report

    Media Portrayals of Minorities Project, Middlebury College

    Strong evidenceInstitutional source
    What was measured
    26,626 articles published during 2018 in The New York Times, The Washington Post, The Wall Street Journal, and USA Today, mentioning African Americans, Asian Americans, Latinos, Jews, or Muslims.
    What it found
    Of the five groups measured, coverage of Muslims was the most negative in tone. Separately: 17% of 2018 articles about Jewish people mentioned antisemitism at least once, while 2% of articles about Muslims mentioned Islamophobia.
    Why it matters
    Anti-Muslim bias is under-named even inside coverage about Muslims. A newsroom can report a hate incident accurately and still leave the category of prejudice unnamed.
  • 2022News coverage

    Covering Muslims: American Newspapers in Comparative Perspective

    Erik Bleich & A. Maurits van der Veen — Oxford University Press

    Strong evidenceScholarly work
    What was measured
    Automated tone analysis of US newspaper articles published between 1996 and 2016, benchmarked against coverage of other religious and ethnic groups and against a representative sample of all articles.
    What it found
    The average article about Muslims is more negative in tone than 84% of the authors' representative sample of US newspaper articles, and more negative than coverage of Catholics, Jews, Hindus, African Americans, Latinos, Mormons, or atheists. The pattern holds across the full two decades.
    Why it matters
    Negativity in Muslim coverage is not an artifact of one outlet, one election, or one news cycle. It is a twenty-year baseline.
  • 2026News coverage

    The State of British Media 2025: Reporting on Muslims and Islam

    Centre for Media Monitoring, a project of the Muslim Council of Britain

    Moderate evidenceInstitutional source
    What was measured
    Roughly 40,000 articles and broadcast items from 30 UK outlets, coded for framing and bias — the largest study of its kind conducted in the UK.
    What it found
    70% of items associated Muslims or Islam with negative aspects or behaviors. Contextual omission — leaving out information a reader needs to understand the story — was the most common single failure, present in 44% of items coded as biased.
    Why it matters
    A different media market, a similar result. It also names a failure mode US newsrooms share: the damage is done more often by what is left out than by what is said outright.

    Read with care. The Centre for Media Monitoring is a monitoring body run by a representative Muslim organization, not an independent academic lab; it selects its own sample and sets its own bias thresholds. It appears here as a monitoring finding alongside the peer-reviewed work above, never as sole support for a claim.

  • 2021Entertainment

    Missing & Maligned: The Reality of Muslims in Popular Global Movies

    Annenberg Inclusion Initiative, USC Annenberg School for Communication and Journalism

    Strong evidenceInstitutional source
    What was measured
    200 popular films released in the United States, United Kingdom, Australia, and New Zealand between 2017 and 2019, with every speaking character coded.
    What it found
    Fewer than 2% of the more than 8,500 speaking characters were Muslim — 1.1% in the US and UK films. Not one of the 23 animated films examined included a Muslim character.
    Why it matters
    Erasure and stereotype are one problem seen from two angles. When a group appears on screen almost only inside one kind of story, that story becomes the whole picture.
  • 2022Entertainment

    Erased or Extremists: The Stereotypical View of Muslims in Popular Episodic Series

    Annenberg Inclusion Initiative, USC Annenberg School for Communication and Journalism

    Strong evidenceInstitutional source
    What was measured
    200 top-rated television series from 2018 and 2019 across the United States, United Kingdom, Australia, and New Zealand.
    What it found
    1.1% of speaking characters were Muslim, and 87% of the series included no Muslim character at all.
    Why it matters
    Television is where most Americans meet people they do not know. An almost empty set is itself an editorial decision, made repeatedly.
  • 2018Entertainment

    Entertainment-education effectively reduces prejudice

    Sohad Murrar & Markus Brauer — Group Processes & Intergroup Relations 21(7), 1053–1077

    Strong evidenceScholarly work
    What was measured
    Randomized experiments in which participants watched either an entertainment-education sitcom featuring Arab and Muslim characters (Little Mosque on the Prairie) or a control sitcom with an all-White cast.
    What it found
    Participants who watched the sitcom with Arab and Muslim characters scored lower on both implicit and explicit measures of prejudice, and the reduction was still measurable four weeks later. The more strongly viewers identified with the characters, the larger the effect.
    Why it matters
    This is the strongest available evidence that representation is not merely symbolic. Ordinary, relatable characters move measured prejudice, and the movement lasts.
  • 2025Entertainment

    Stereotypes on Screen: The Effects of Muslim Portrayals in Entertainment Media

    Institute for Social Policy and Understanding (ISPU)

    Moderate evidenceInstitutional source
    What was measured
    Experiment: participants viewed either a positive depiction of Muslim characters (an episode of 9-1-1: Lone Star) or a negative one (an episode of Criminal Minds), then answered survey items on policy and social attitudes.
    What it found
    Viewers of the positive depiction were more opposed both to anti-Muslim policies and to broadly anti-democratic ones, and more willing to be in close social proximity to Muslims, than viewers of the negative depiction.
    Why it matters
    Portrayal effects do not stop at attitudes toward Muslims. They register in support for democratic norms generally.

    Read with care. ISPU is a research institute that also advocates for American Muslim communities. It is paired here with Murrar and Brauer's peer-reviewed experiments, which point in the same direction; on its own it is not sole support for any claim on this page.

  • 2022Platforms & policy

    Hate, amplified? Social media news consumption and support for anti-Muslim policies

    Nazita Lajevardi, Kassra A. R. Oskooii & Hannah L. Walker — Journal of Public Policy 42(4), 656–683

    Moderate evidenceScholarly work
    What was measured
    Survey analysis testing whether self-reported reliance on social media as a primary news source predicts support for anti-Muslim policies.
    What it found
    A strong and consistent association between relying on social media for news and support for a range of anti-Muslim policies.
    Why it matters
    Where people get their news is itself a variable in policy attitudes, alongside what that news says.

    Read with care. This is an observational association, not an experiment. It does not establish that social-media use causes these policy attitudes; people who already hold them may also select into those feeds.

  • 2018Law & consequence

    Equal Treatment? Measuring the Legal and Media Responses to Ideologically Motivated Violence

    Institute for Social Policy and Understanding (ISPU)

    Debated / contestedInstitutional source
    What was measured
    52 randomly selected US cases of ideologically motivated violence between 2002 and 2015, with coverage measured in The New York Times and The Washington Post.
    What it found
    Perpetrators perceived as Muslim received sentences averaging four times longer than other perpetrators of comparable plots, and prosecutors sought charges carrying 302% longer sentences (230 months versus 76 months). Foiled or uncompleted plots by Muslim-perceived suspects drew 770% more coverage.
    Why it matters
    This is the clearest attempt to measure coverage volume and legal outcomes inside a single sample of cases.

    Read with care. Small sample (52 cases), produced by a research institute that advocates for American Muslim communities, and not reproduced by the larger peer-reviewed analysis below. Treat the direction as a hypothesis and the magnitude as unsettled.

  • 2024Law & consequence

    Sentencing Disparities in US Terrorism Prosecutions, 2001–2018

    Jesse J. Norris & Hanna Grol-Prokopczyk — Criminology, Criminal Justice, Law & Society 25(2), 18–40

    Debated / contestedScholarly work
    What was measured
    All identified US terrorism cases from 2001 to 2018 (n = 825), state and federal, testing sentence length against race and ethnicity, gender, citizenship, court, and type of terrorism.
    What it found
    Some disparities appear — foreign nationals and White supremacist defendants received longer sentences, foreign non-jihadi defendants shorter ones — with only limited evidence of racial disparity overall. The analysis does not find a general sentencing premium for Muslim or jihadi-classified defendants.
    Why it matters
    This is where the record disagrees with itself. A larger, peer-reviewed sample complicates the widely repeated four-times-longer figure. We show both rather than publishing the more quotable one.
  • 2026Public opinion

    25 Years After 9/11, a Portrait of Muslim Americans

    Pew Research Center

    Strong evidenceInstitutional source
    What was measured
    Nationally representative surveys, including an American Trends Panel survey fielded May 4–17, 2026, read against Pew's earlier trend data.
    What it found
    51% of US adults say Islam is more likely than other religions to encourage violence, up from 25% in March 2002. 43% say Muslims are less patriotic than other Americans, and 42% say Muslim Americans have a negative impact on the country. The split by party is wide: 76% of Republicans and Republican leaners said Islam encourages violence, against 29% of Democrats and Democratic leaners.
    Why it matters
    Opinion is more negative now than in the months immediately after the attacks. Whatever twenty-five years of coverage accomplished, it did not settle into familiarity.
Language & framing

Six words and frames worth a second look

None of these is a banned word. Each is a place where a common phrase carries a claim the reporting has not made — and where a named alternative is available and more accurate.

  • “Islamist” used to mean “militant”

    The problem

    The Associated Press Stylebook defines an Islamist as an advocate or supporter of a political movement that favors reordering government and society in accordance with laws prescribed by Islam. That describes a political position, not an act of violence.

    Better practice

    Follow AP's own instruction: do not use it as a synonym for Islamic fighters, militants, extremists, or radicals, who may or may not be Islamists. Where possible, name the affiliation — al-Qaida-linked, Hezbollah, Taliban.

    AP Stylebook entry revised April 2013.

  • “No-go zones”

    The problem

    In January 2015 Fox News aired claims that European cities including Birmingham, England were areas non-Muslims simply do not enter. The network broadcast four on-air corrections and apologies inside a week; the then-most-recent UK census recorded roughly 22% of Birmingham residents as Muslim. The mayor of Paris announced plans to sue.

    Better practice

    Treat a demographic claim as a factual claim and check it against census or administrative data before it airs. A guest asserting it is not a source.

    Fox News on-air corrections, January 17–18, 2015.

  • “The Muslim world”

    The problem

    The phrase implies a single geopolitical bloc with one interest and one voice. Historian Cemil Aydin traces the idea to late-nineteenth-century imperial politics and the racial ordering that came with it, not to any unified polity that ever existed. Muslims live under every kind of government, on every continent.

    Better practice

    Name the actual subject: Indonesia, the Gulf states, the Egyptian government, Muslim-majority countries, or simply Muslims.

    Cemil Aydin, The Idea of the Muslim World: A Global Intellectual History (Harvard University Press, 2017).

  • Anti-Muslim bias that goes unnamed

    The problem

    In 2018 US newspaper coverage, 17% of articles about Jewish people named antisemitism at least once. Two percent of articles about Muslims named Islamophobia.

    Better practice

    Name anti-Muslim bias as directly, and as high in the story, as other forms of prejudice are named — including in coverage of incidents that are about nothing else.

    Media Portrayals of Minorities Project, Middlebury College, 2019.

  • Religious labeling applied to one perpetrator and not another

    The problem

    Across all 136 US terrorist attacks from 2006 to 2015, perpetrator religion was the largest single predictor of how much coverage an attack received — larger than fatalities or target type.

    Better practice

    Apply one disclosure test to every perpetrator. If the faith of a non-Muslim perpetrator would not appear in the lede, a Muslim perpetrator's should not either, unless the reporting establishes an actual causal role and explains it.

    Kearns, Betus & Lemieux, Justice Quarterly, 2019.

  • “The Muslim community,” singular

    The problem

    No racial or ethnic group forms a majority among US Muslims. Pew's figures are roughly 30% White, 30% Asian, 20% Black, 11% Hispanic, and 8% another race or more than one. There is no single community to quote.

    Better practice

    Name who you actually spoke to — a mosque board in Dearborn, Somali American parents in Minneapolis, a Bangladeshi American student group. One “community spokesperson” for a group with no majority is a sampling error, not a shortcut.

    Pew Research Center, How US Muslims compare with other Americans, June 18, 2025.

Newsroom checklist

Eight questions before you publish

Short enough to run on deadline. Every item traces to something on this page or to published guidance for media professionals.

  1. Does religion explain anything in this story?

    If the honest answer is “it identifies him,” that is not an explanation. Say what the evidence actually supports about motive, or leave it out.

  2. Run the substitution test on the headline.

    Swap the perpetrator's faith for another one. If the sentence stops reading like news, the original was carrying the weight the facts were not.

  3. Has every population or geography claim been checked against data?

    The 2015 “no-go zones” corrections turned on a census figure that took minutes to look up and a week of apologies to undo.

  4. Is “Islamist” being used per the AP definition, or as a stand-in for “militant”?

    AP's own entry forbids the second use and tells reporters to name the group instead.

  5. If the story is about anti-Muslim bias, does it use the word?

    Coverage of Muslims names Islamophobia at roughly an eighth the rate coverage of Jewish people names antisemitism.

  6. Are Muslims quoted in this outlet outside of crisis coverage?

    A source list that reaches Muslims only for terrorism, foreign conflict, and civil-liberties stories produces a portrait made of emergencies.

  7. What context would a careful reader need that this draft leaves out?

    Contextual omission — a missing base rate, denominator, or prior — was the most common failure in the largest UK audit of Muslim coverage.

  8. Are the victims and survivors in this story named and centered?

    Coverage that centers the perpetrator's identity and not the harm reproduces the framing that made the harm reportable in the first place.

Adapted from the Institute for Social Policy and Understanding's guide for media professionals, the Associated Press Stylebook's own entry on “Islamist,” and the Shorenstein Center's Journalist's Resource reporting tips, together with the studies above.

Guidance from people who do this full time

What we do not claim

The limits of this page

  • Association is not causation. The social-media finding is a survey association. It does not establish that platform use produces anti-Muslim policy views; the causal arrow could run either way, or through a third factor.
  • The sentencing question is unsettled. A widely repeated figure holds that Muslim-perceived defendants receive sentences four times longer; it comes from a 52-case think-tank study. A peer-reviewed analysis of 825 cases from 2001 to 2018 does not reproduce a general disparity of that kind. Both are on this page. Anyone quoting the first number without the second is quoting half the record.
  • Monitoring bodies are not neutral labs. Two sources here — the Centre for Media Monitoring and ISPU — are organizations with a stake in the finding. Their work is labeled as such on the card and paired with independent research rather than standing alone.
  • One widely circulated claim was left out. A figure holding that more than three-quarters of social posts referencing Muslims or Islam carry a strongly Islamophobic tone circulates widely. We could not trace it to a study whose sample and coding scheme we could read, so it does not appear here. If you can point us to the underlying research, we will add it or explain why it did not qualify.
  • We did not build an automated sentiment dashboard. A language model scoring daily headlines would produce a chart we could not defend claim by claim, which is the standard the rest of this site is held to. Curated, method-stated studies first; live tooling only if it can meet the same bar.
  • This page audits patterns, not people. Nothing here scores an individual journalist or story. Aggregate findings about a corpus do not license a verdict about any single byline in it.

The rest of this site is the other half of the argument: the 250-year timeline, the primary sources, and the classroom toolkit exist so that a reporter looking for something other than crisis coverage has somewhere to start.

Corrections & sources

Suggest a fix or a better source

Every submission is checked against the cited record before anything changes on the site. Prefer primary documents, peer-reviewed work, or institutional pages. Broken links are replaced only with independently verified URLs.

Email hello@usmuslim.org

Free to reproduce with attribution to Threads of Connection. Verify every figure against its linked source before republishing.