Skip to main content

Documentation Index

Fetch the complete documentation index at: https://www.rhetoricaudit.com/docs/llms.txt

Use this file to discover all available pages before exploring further.

The Bias Spectrum is Rhetoric Audit’s approach to measuring political lean without collapsing complex argumentation into a binary label. Instead of stamping an article “left” or “right,” the analysis positions it on a continuous gradient, decomposes it across 10 distinct ideological dimensions, and tells you how confident the analysis is — and when an article is too multi-positional to classify cleanly. The result is a political profile you can reason about rather than a verdict you either accept or reject.

Why a continuous gradient matters

Binary left/right labels are common because they are easy to communicate, but they destroy the information that matters most. A single label cannot tell you whether an article’s political lean comes from its economic framing, its position on state authority, its treatment of environmental policy, or its stance on global governance. All of those can point in different directions in the same piece. The Bias Spectrum addresses this by measuring lean as a continuous numeric valuepolitical_inclination — on a gradient from Far Left to Far Right. A position near the center does not necessarily mean the article is balanced; it can mean the article blends positions from multiple ends of the spectrum in ways that average to the center. The ambiguity_flag captures exactly this case.

The political inclination score

The political_inclination score is the primary output of the Bias Spectrum. It places the article on a seven-position gradient:
PositionDescription
Far LeftStrong progressive or revolutionary framing across most dimensions
LeftConsistent progressive framing with institutional engagement
Center-LeftPredominantly progressive with pragmatic accommodation
CenterMixed or balanced ideological signals
Center-RightPredominantly conservative with pragmatic accommodation
RightConsistent conservative framing
Far RightNationalist, authoritarian, or reactionary framing across most dimensions
The score is derived from the weighted aggregate of the 10 ideology dimension scores below — it is not a human-assigned label.

The philosophical frame

Alongside the numeric position, the analysis assigns a Philosophical Frame label that characterizes the worldview the article is reasoning from. This goes deeper than political lean — it captures the underlying assumptions about institutions, society, and human nature. Examples include:
  • Progressive Institutionalism — change through existing democratic structures
  • Market Libertarianism — individual economic freedom as the organizing principle
  • Anti-Imperialist Internationalism — skepticism of Western-led global institutions
  • Nationalist Conservatism — cultural and national identity as primary values
  • Technocratic Governance — evidence-based, expert-led policy as the ideal
The Philosophical Frame often clarifies why an article lands where it does on the spectrum, not just where it lands.

Ten ideology dimension scores

The Bias Spectrum breaks political lean into 10 independent ideological dimensions, each scored from 0 to 100. A high score on any dimension means the article’s rhetoric is strongly consistent with that ideology; a low score means it is inconsistent or absent.
Measures alignment with market-led economic governance: free trade, deregulation, privatization, fiscal austerity, and the primacy of market mechanisms over state intervention. High scores appear in articles that treat market outcomes as legitimate by default and frame state intervention as a distortion.
Measures alignment with social-democratic and democratic socialist frameworks: public ownership of key industries, universal social programs, redistribution through progressive taxation, and worker rights. High scores appear in articles that foreground economic inequality and institutional remedies for it.
Measures alignment with civil and economic libertarian principles: minimal state interference in personal choices, strong individual rights, skepticism of surveillance and regulation, and anti-interventionist foreign policy. Note that libertarianism scores independently from both Neoliberal Capitalism and Nationalist Conservatism — an article can score high here while scoring low on both of those.
Measures alignment with national-identity-centered conservatism: emphasis on cultural heritage, border sovereignty, skepticism of supranational institutions, and social traditionalism. High scores frequently co-occur with high Populism scores.
Measures alignment with expert-led, evidence-based governance frameworks: deference to scientific consensus, institutional technocracy, and skepticism of populist or democratic override of expert opinion. High scores appear in articles that frame policy debates primarily in terms of efficiency and evidence rather than values or representation.
Measures alignment with populist rhetoric patterns: framing politics as a conflict between “the people” and a corrupt elite, appeals to common-sense wisdom over institutional expertise, and distrust of media, academia, or government. Populism is ideologically non-directional — this dimension captures the rhetorical structure regardless of whether the article is left-populist or right-populist.
Measures alignment with ecosocialist and radical environmental frameworks: framing ecological crisis as inseparable from capitalist production, support for systemic economic transformation as a prerequisite for environmental sustainability, and centering frontline and indigenous communities in climate discourse.
Measures alignment with authoritarian governance frameworks: centralized state power, securitization of social problems, law-and-order framing, and justification of constraints on civil liberties in the name of stability or national security. This dimension scores independently of the left/right axis — authoritarian statism can appear in both left-wing and right-wing contexts.
Measures alignment with decentralist and federalist frameworks: devolution of power to local communities or regions, subsidiarity principles, skepticism of centralized national or supranational authority, and preference for bottom-up over top-down governance.
Measures alignment with conspiratorial narrative frameworks: attribution of events to hidden coordinated actors, framing of official accounts as deliberate deception, and use of evidential standards that treat absence of disproof as confirmation. High scores here are a strong signal to cross-reference with the External Claim Grounding parameter.

Classification confidence and the ambiguity flag

Two meta-signals accompany every Bias Spectrum output:

classification_confidence

A percentage from 0 to 100 reflecting how clearly the article’s rhetoric aligns with a position on the spectrum. High confidence (above 75) means the article sends consistent ideological signals. Low confidence (below 40) means the signals are weak, mixed, or contradictory.

ambiguity_flag

A boolean flag that activates when the article scores significantly above the midpoint on two or more ideologically opposed dimensions simultaneously. An article that scores high on both Neoliberal Capitalism and Democratic Socialism, for example, is genuinely multi-positional — and forcing it onto a single point on the spectrum would be misleading.
When ambiguity_flag is active, treat the political_inclination score as an average rather than a position. The ideological dimension scores will tell you more than the summary gradient in this case. Articles written for broad audiences, or articles covering ideological debates from multiple perspectives, commonly trigger this flag.

What the Bias Spectrum does not do

The Bias Spectrum measures the rhetorical posture of the article — the ideological assumptions embedded in its framing, language, and argument structure — not the biographical politics of the author or the editorial line of the publication. An article can score Center-Left even if it was written by a conservative author, if the rhetorical choices made in that specific piece align with center-left framing. The analysis is text-first, not author-first.