Our Methodology

The Challenge: Quantifying Values

Values are deeply personal and inherently subjective. People don't naturally think about ethics in numerical terms. Simple rating scales (“rate fairness 1–10”) produce unreliable data because individuals anchor differently—one person's “7” is another's “4.” Forced ranking creates artificial hierarchies. The challenge is to surface genuine trade-off preferences in a way that is mathematically rigorous but feels natural to the participant.

Step 1Discovering Your Values Through Pairwise Comparisons

We use the Analytic Hierarchy Process (AHP), developed by Thomas Saaty in 1980, because it is one of the most rigorous and well-studied methods for quantifying subjective trade-offs. Rather than asking you to rate values on an abstract scale, AHP presents you with direct comparisons: “Which matters more to you, and by how much?”

Each question pits two objectives against each other. You choose which matters more and indicate your conviction strength on a scale from 1 (slightly more) to 9 (overwhelmingly more). The winner receives the intensity value as points; the other receives the reciprocal (1/intensity). This reciprocal scoring is what makes AHP mathematically consistent—if you say A is 7× more important than B, then B is exactly 1/7th as important as A.

Code reference: app/questionnaire/questionnaire-client.tsx — the calculateWeights() function

Step 2Mapping Values to Real-World Criteria

Knowing that someone values equity over stability doesn't directly tell you their view on automating warehouse jobs. We need a bridge—the relevance matrix.

Each theme (AI Labor, Education, Economic Policy) defines a set of measurable criteria. The relevance matrix defines how much each ethical objective cares about each criterion, using values from −1.0 to +1.0:

Code reference: lib/theme-config.ts — relevance matrices and criteria definitions for all themes

Step 3Bringing It All Together: The Scoring Formula

With your ethical weights and the relevance matrix, we compute a personalized score for every item (occupation, policy, or education approach). The formula combines your priorities with each item's characteristics:

Code reference: lib/theme-config.ts — the calculateItemScore() function

Step 4Your Ethical Profile

Your weight distribution maps to a named ethical profile based on your primary and secondary objectives. This helps create a recognizable identity for your ethical perspective, making it easier to understand and compare with others.

Code reference: lib/relevance-matrix.ts — getProfileName() and getEducationPerspective()

A Starting Point, Not a Final Answer

This methodology is a starting point for continuous refinement. The relevance matrices, the scoring formula, the normalization parameters, and the question design all benefit from community feedback. We are searching for a more accurate model—one that faithfully reflects the many different views that come together to form our collective consciousness.

Feedback from users about the evaluated weight of themes, predicted views, and analysis results is critical to this process. If your results don't match your self-understanding, that is valuable data. It might mean:

• •A relevance matrix value needs recalibration
• •The scoring normalization needs adjustment for this theme
• •A criterion is missing or weighted incorrectly
• •The pairwise questions aren't capturing a nuance in your values

Every piece of feedback moves us closer to a model that accurately predicts how different value systems lead to different perspectives—which is the entire purpose of this tool.

Open Source: See the Code, Shape the Future

Views to Values is fully open source. Every formula, every relevance matrix, every scoring decision is visible in the code. We believe that a tool asking people to examine their values must itself be open to examination.

We welcome contributions of all kinds: bug reports, relevance matrix refinements, new theme proposals, academic review of the methodology, and ideas for improving the user experience. If you have expertise in decision science, ethics, survey design, or data analysis, your perspective is especially valuable.