Differential visibility and narrative framing of LGBT+ identities in a decade of U.S. online news
Although scholarly interest in LGBT+ media representation has grown, most published work relies on qualitative content analysis, single-identity coverage patterns, or focuses on non-U.S. contexts. Large-scale, quantitative investigation of how American online news distributes attention across the four distinct LGBT+ identity subgroups — and what that asymmetry produces — remains sparse.
This gap matters because media visibility is not a neutral measure. Drawing on Agenda-Setting Theory (McCombs & Shaw, 1972), Framing Theory (Entman, 1993), and the Minority Stress Model (Meyer, 2003), this project treats differential coverage as a mechanism: one that shapes public cognition, informs legislative agendas, and is associated with documented disparities in mental health and physical safety among the communities least served by that coverage.
What follows is an analysis of 954,433 English-language online news articles retrieved from Media Cloud, spanning January 2016 through December 2025 — four identity labels, tracked week by week, across a decade of American political life.
Each row shows two things: media coverage share (solid bar) and actual LGBT+ population share (faded bar below it). In a world where coverage matched demographics, both bars would be the same length.
Gay identity accounts for 51% of coverage while representing 17% of the LGBT+ population. Bisexual identity represents 56% of the population and receives 6% of coverage.
Divide each identity's media share by its population share. A ratio of 1.0 means perfect parity. Gay coverage runs at 3.07×. Bisexual coverage runs at 0.11× — nearly invisible relative to its demographic size.
The share of U.S. adults identifying as LGBT+ doubled between 2012 and 2025, driven primarily by bisexual identification among younger generations. Bisexual adults grew from 1.8% to 5.3% of the total U.S. adult population.
Media coverage did not follow. The gap between who exists and who gets reported on widened as the community grew.
Four charts, one shared Y-axis. All four identities measured on the same scale. Gay coverage is so dominant that the other three identities barely register at this scale. Trans coverage spikes sharply around legislative attacks, then recedes. Lesbian and bisexual coverage is near-flat throughout.
Gay coverage maintains the highest baseline across the full decade. The two dominant peaks — Obergefell v. Hodges (June 2015) and the Pulse nightclub shooting (June 2016) — are large enough to compress the other three rows into near-invisibility at the shared scale.
Trans coverage is sparse outside major headlines, but surges around legislative attacks: HB2 (2016), the school guidance rescission (2017), the military ban (2017), and the accelerating state-level care and sports bans from 2022 onward. Coverage of trans identity is contingent almost entirely on policy conflict.
Lesbian coverage rises only in cultural contexts — Pride Month, film releases, representation stories. It almost never drives coverage in political or policy debates. The absence from legislative cycles, which so consistently produce trans spikes, is itself a pattern worth noting.
Bisexual coverage reads as a near-flat line throughout. Small spikes correspond to narrow contexts: the 2014 FDA blood donation debate, Ryan Russell's NFL coming-out (2019), monkeypox vaccine equity (2022), the final FDA blood ban repeal (January 2023). None produced coverage visible alongside the Gay row.
From 5,000 sampled articles per identity, these are the terms that appeared most consistently alongside each identity label. The percentage shows how many of those articles contained each word. Read across all four columns and you see not four coverage patterns but four distinct kinds of stories.
Each rectangle is one word. Its height is proportional to the share of sampled articles in which it appeared. Stacked by thematic category, the chart makes visible what the collocate table only implies: what kind of conversation each identity is embedded in.
For Trans identity, the Legislative Attack category alone accounts for more than 80% of the cumulative word signal — dwarfing every other identity's total. Hover any segment for the specific word and its frequency.
| Identity | Primary Frame | Representative Collocates | What this means |
|---|---|---|---|
| Gay | Civil rights and political identity | rights · pride · marriage · court florida · anti · buttigieg · trump |
Gay coverage sits inside the power structure. Named politicians appear. Legal milestones anchor major spikes. The framing positions gay identity as a legitimate political subject with standing in institutional debates. |
| Lesbian | Cultural representation and relationships | women · queer · couple · court film · tv · show · marriage |
Lesbian coverage is culturally legible but politically secondary. Film, television, and relationship contexts dominate. Legal contexts appear, but rarely as a driver of coverage. Lesbian identity is present in stories about art and intimacy; largely absent from stories about power. |
| Bisexual | Health policy and celebrity disclosure | blood · hiv · health · star celebrity · survey · men · women |
Bisexual identity appears mainly in two contexts: public health policy debates (blood donation eligibility, HIV prevention discourse) and celebrity coming-out stories. The community itself almost never drives a news cycle. Bisexual people are covered when they are useful to someone else's story. |
| Trans | Legislative opposition and institutional conflict | trump · ban · military · sports athletes · court · supreme · care |
The most striking collocate profile of the four. Trans coverage is defined almost entirely by adversarial institutional terms. The word "trump" appears in 17.4% of all sampled trans articles. Affirmative or community-centered language is nearly absent. Trans identity in U.S. news is framed predominantly as a political problem to be legislated and adjudicated, not a community to be reported on. |
Media visibility is not a neutral measure. The Minority Stress Model (Meyer, 2003) posits that chronic social exclusion and stigma — including invisibility in public discourse — produce measurable health disparities. The data below show that the identities least served by media coverage face the most severe mental health outcomes, and that crisis-driven media framing of transgender identity corresponds with a documented rise in anti-trans hate crimes.
Among cisgender LGBT+ identities, bisexual youth report the highest rates of suicidal ideation and depression — higher than gay youth, despite receiving far less media coverage and community recognition.
Transgender youth face the most severe outcomes overall, corresponding with their exposure to crisis-driven, adversarial media framing and sustained legislative attacks.
The trans media spike pattern closely mirrors the legislative calendar. Total anti-LGBT+ bills tracked by the ACLU grew from roughly 50 in 2016 to 510 in 2023, 533 in 2024, and 616 in 2025 — a twelvefold increase over the decade. FBI-reported anti-transgender hate crimes rose in parallel, from 111 in 2016 to a peak of 492 in 2023.
This suggests a feedback loop: adversarial legislation generates media coverage that frames trans identity as a contested political problem, which may normalize anti-trans hostility and contribute to real-world violence.
Healthcare restrictions, school curricula, sports participation, identity documents, public facilities — each category has its own escalating trajectory. The breadth of categories means trans people face simultaneous restrictions in every aspect of daily life.
Despite bisexual people outnumbering gay people in the U.S. adult population, anti-bisexual hate crimes register at roughly one-tenth the rate of anti-gay crimes in FBI data.
This likely reflects a reporting and classification problem: when bisexual individuals are targeted, crimes tend to be categorized as anti-gay or anti-lesbian, mirroring the same erasure mechanism at work in media coverage.
The total bisexual count sounds like a number. But it obscures a direction. Coverage has been falling year after year since 2016. The New York Times published 706 bisexual-mentioning articles in 2016. By 2025, that number was 34 — a 95% decline at the country's most influential newspaper.
This is not about bisexual people becoming less newsworthy. It is about bisexual people never having been treated as newsworthy in the first place.
The pattern extends beyond news. GLAAD's Where We Are on TV 2024–25 found bisexual-plus characters declined for three consecutive years, reaching 20% of all LGBT+ characters — down from 30% in 2016–17. GLAAD's Studio Responsibility Index 2025 found bisexual-plus film characters fell to 10% of all LGBT+ film characters.
News coverage falls. Television representation falls. Film representation falls. The community grows. The stories do not follow.
Bisexual-plus characters fell from 30% of all LGBT+ TV characters in 2016–17 to 20% in 2024–25, mirroring the news coverage decline almost exactly. Total LGBT+ character counts peaked in 2021–22 and have contracted since, particularly on broadcast television.
At 98,036 articles, lesbian coverage is larger in absolute terms than bisexual coverage — but it follows the same downward trajectory. In 2016, the dataset recorded 20,041 articles. By 2025, that number had fallen to 4,425: a 78% decline over the decade.
The New York Times published 1,089 lesbian-mentioning articles in 2016. By 2025, that number was 60 — a 94% decline, nearly identical to the bisexual drop at the country's most influential newspaper.
The decline pattern differs from bisexual in one key way: lesbian coverage had a higher baseline to begin with, and its drop is more gradual. Bisexual coverage fell sharply between 2016 and 2018, then plateaued near zero. Lesbian coverage declined more steadily, without a single dramatic collapse — suggesting a slow editorial drift rather than an abrupt decision to stop covering the community.
The top outlets are identical to the bisexual list: NYT, Guardian, LAT, Washington Post, Reuters. The mainstream chose both communities less and less, at nearly the same rate.
A corpus of 954,433 English-language online news articles was retrieved from Media Cloud (a non-commercial academic media monitoring platform), spanning January 2016 through December 2025. Articles were identified through Boolean keyword search isolating each of four target identity terms: gay, lesbian, bisexual/bisexuality, and transgender/transsexual. The 954,433 figure is the sum of four separate searches: gay (489,001) + transgender/transsexual (310,418) + lesbian (98,036) + bisexual/bisexuality (56,978). A single article mentioning multiple identity terms may appear in more than one count.
Weekly mention counts were aggregated from daily article totals to smooth short-term noise. For each identity, the top 100 most frequent collocating terms were extracted from a stratified sample of 5,000 articles per identity and analyzed using document frequency ratios (doc_ratio). Population benchmarks are from the Gallup 2025 national survey of LGBT+ adult identity, with four categories normalized to sum to 100%. Thematic labelling follows the methodology of Ng, Yang et al. (2024), PLOS ONE e0300385. Mental health data: The Trevor Project National Survey 2023–2024. Hate crime data: FBI Uniform Crime Reports 2016–2024. Legislative data: ACLU Legislative Tracker. Entertainment media data: GLAAD Where We Are on TV 2024–25 and Studio Responsibility Index 2025.
Note on "gay" as umbrella term: The term "gay" functions both as a specific marker for gay men and as a generic synonym for the LGBT+ community in U.S. news discourse. This dual usage is itself analytically significant — the conflation of "gay" with "LGBT+" in editorial practice is precisely the phenomenon this study documents: gay identity as the normative default through which the entire community is represented.
All sources accessed March–April 2026. Sources marked PRIMARY are government data or original survey data. Sources marked INSTITUTIONAL are major advocacy or professional organizations. Sources marked NOTE carry credibility caveats.