Methodology

Render scores are created according to Catholic Social Teaching (CST), a tradition dealing with moral absolutes and prudential questions about how societies ought to be organized. It is built through the Catechism, papal encyclicals, and conciliar documents.

Render scores the actions taken by politicians in office according to CST. This means votes cast, legislation enacted, executive actions taken, and official positions. Rhetoric does not directly impact scoring, meaning a politician's statements, campaign branding, or reputation do not directly change scores. The exception is for candidates who have not held office, in which case rhetoric is the only available evidence. Still extensively sourced, dated, and labeled.

Scores are based on sourced evidence, are auditable, and are applied impartially regardless of party affiliation.

Rhetoric does not raise or lower a score, only official, consequence-bearing acts do: votes, enacted legislation, executive orders. However, recent statements are used to modify how quickly previous actions decay. If recent statements are aligned with previous actions, those actions will remain relevant to their score for longer. This is called the rhetoric modifier.

The exception is candidates who have never held office. With no official record to score, Render must use campaign platforms, questionnaire responses, and policy statements as primary evidence. This is done transparently, with the candidate page having it disclosed, labeled as rhetoric/platform evidence, and capped at T3. Once an official record does exist, it supersedes the existing score.

The pipeline for scoring a politician on a given issue follows these steps in order:

Every piece of evidence is classified into a tier based on evidentiary strength. Higher tiers are weighted more heavily. Tiers are impartially assigned based on the nature of the action, as defined below:

Eligible actions are assigned a CST alignment score on a scale from 1 to 5. The score is based on how closely that action is aligned with CST, and is applied to actions individually. This is not an overall evaluation of a politician's position, as a politician may have high scoring and low scoring actions on the same issue. Then, those individual scores are averaged to form the overall score for a politician.

Legislative records age. Render applies exponential decay so recent actions carry more weight without erasing the historical record.

The constant 0.1733 derives from a four-year half-life: ln(2) / 4 ≈ 0.1733. An action taken four years ago carries half the weight of a current one.

Once eligible actions (T1–T3) are scored, tier-weighted, and decay-weighted, the issue score is their normalized weighted average:

Each issue gets a single score between 1 and 5. To get to that score, Render weights every action in that issue by its tier and the recency of that action, then averages them out. A Tier 1 action (direct legislative activity) pulls much harder than a Tier 3 action (rhetoric or association). Further, the formula's denominator prevents politicians with larger dossiers from scoring higher through volume alone. Tier 4 entries are excluded from this math entirely for officeholders.

Note: The /10 conversion happens only at the composite level (§10–11). Issue scores in Render right now were assigned directly by human researchers. See §15 for current data status

An action goes into an issue's scoring pool if it directly changes the policy, law, or funding governing that issue. Related actions can also be included, but only when a specific CST source explicitly connects them to that issue.

Not every office can act on every issue. Render handles this with five classifications:

Composite A measures overall alignment using doctrinal weights only. It asks: how aligned is this politician across the moral universe Render tracks, weighted by doctrinal seriousness?

Issues with actionability = 0 for the office are excluded from both numerator and denominator.

Composite B measures alignment adjusted by what the office can actually do. It preserves the intrinsic-evil multiplier while accounting for each office's practical authority over each issue. It does not say an issue matters less morally — it says this office has more or less power to move it.

• Uses doctrinal × actionability as the combined weight — not actionability alone.
• If actionability = 0 for an issue, that issue drops from the denominator.
• Both A and B are shown together. A reflects principle; B reflects impact.

Each cell is the actionability weight for that office. Intrinsic evils are marked ⊕. The 2× intrinsic-evil multiplier is a separate weight independent of these values. 0.00 = no jurisdiction: the issue is excluded from both denominators for politicians in that office.

Pres = President · Sen = Senator · Rep = U.S. Representative · Gov = Governor · StLeg = State Legislator · Mayor = Mayor

Every dossier entry should carry: a tier classification, a one-sentence action description, a date, a source label, a source URL, and eventually a per-action CST alignment score. Adjacent actions must include the CST citation justifying their inclusion.

The current codebase computes composites from human-assigned issue-level doctrinal scores. Full action-level recomputation requires per-action CST alignment scores in each dossier entry. That field exists in the data model but is not yet populated.