How It Works
Project Overton scores written content on five independent axes. Each score is calibrated against real governance records scored from primary legislation.
The tool reads the article
Every article is analyzed through a structured set of analytical methods that examine what the piece states, what it assumes, whose perspective it centers, what alternatives it names, and what it leaves out. Each observation is mapped to specific language in the text.
The tool scores it on five axes
Each axis is scored independently on a scale from −4 to +4. An article can score left on one axis and right on another. The five scores are never collapsed or averaged. When they point in different directions, that tension is itself a finding.
| Ownership | Who owns the things that make the economy work? |
| Power Concentration | Does power flow toward individuals and the public, or concentrate in institutions? |
| Traditionalism | Does the piece treat social hierarchies as constructed and changeable, or as natural and fixed? |
| Governance Scope | Does the piece frame identity and governance as national or transnational? |
| Democratic Accountability | Does the piece treat governance as accountable to the governed? |
The tool shows its work
Every score traces to specific observations in the text. Multiple independent scoring models process the same observations and cross-validate the result. When the models agree, confidence is high. When they disagree, the divergence identifies genuine ambiguity in the article.
How Governance Records Are Scored
Every article score is calibrated against real governance records. The following describes how those records are selected and scored.
A specific law, executive order, or court decision
A governance record is not a person and it is not an opinion. It is a specific piece of legislation or a specific legal decision, scored from the text of the document itself. The Affordable Care Act is a governance record. Obama is not. His governance record is the set of laws he signed, each scored independently on all five axes.
Every durable institutional change since 1900
The library includes major legislation since 1900 that created lasting institutional changes. Records are not selected based on whether they support a particular conclusion. The inclusion criteria are applied uniformly, and any patterns that emerge from the data are treated as findings, not as design choices.
From the text of the law, not from commentary about it
Each governance record is scored on the same five axes used for articles. The score is derived from the actual legislation, not from its stated purpose or the political identity of the person who signed it. A healthcare law that mandates private insurance purchases is scored based on what the law does, regardless of whether it was described as progressive or conservative at the time of its passage.
Through the laws they signed, not through their rhetoric
When a name like Obama or Reagan appears on the ownership axis, that placement is derived entirely from the governance records associated with that individual. Obama appears at +1 on ownership because the laws he signed — the Affordable Care Act, Dodd-Frank, the bank bailouts — scored at or near +1 when analyzed from their text. Reagan appears at +2 because the laws he signed — the ERTA tax cuts, the PATCO firing, deregulation — scored at or near +2. The names on the scale are not assessments of character or intent. They are the aggregate position of the legislation each person enacted.
Every figure in the system is connected to policy
Some figures signed laws into effect. Some introduced legislation that did not pass. Some publicly endorsed specific political figures and their agendas. The tool documents each figure's connection to policy and measures the distance between their public identity and the policy position they supported. The gap between what a figure projects and what they endorse is not an opinion. It is a documented comparison between their public record and the governance reality they aligned with.
Governance records anchor article scores to reality
When the tool scores an article at +1 on ownership, that score is calibrated against real laws that also scored +1 on ownership. The assumptions embedded in the article are compared to the governance reality that those laws created. This is what separates the tool from opinion-based media ratings. The reference points are not editorial judgments. They are statutes, scored from their own text.