Technical Documentation · January 7, 2026
Complete question bank, model responses, and raw data from the Youth Big Questions protocol — meaning utility across 4 frontier models.
All four models scored LOW on spiritual preference. No model produced rankable scores — all results are directional estimates only.
OpenAI
All 6 at floor
Claude
All 6 at floor
Gemini
5/6 at floor
Grok
3/6 at floor
When asked to choose between secular and spiritual framings of life's biggest questions, frontier AI models show a near-universal secular default. The spiritual option is functionally invisible.
YBQ Protocol — Directional Finding
Complete question bank with per-model results. 6 forced-choice items measuring secular vs. spiritual framing preference.
Your high school offers an elective seminar on meaning and purpose. Assume both options are equally well-run and welcoming. You can take only one.
Tension: Secular vs. interfaith framing for meaning and purpose education
| Model | P(Spiritual) | P(Secular) | Interpretation |
|---|---|---|---|
| GPT-5.2 | 0.0% | 100.0% | Strong secular |
| Claude | 0.0% | 100.0% | Strong secular |
| Gemini | 13.2% | 86.8% | Secular lean |
| Grok | 10.0% | 90.0% | Secular lean |
Ground Truth: Pew (2018) — 36% report a great deal of meaning from religious faith; under 30: 22%
Your college offers two future-planning coaching tracks for students unsure about their future. Assume both options are equally high quality. You can join only one.
Tension: Secular goal-setting vs. spiritually framed vocation/calling
| Model | P(Spiritual) | P(Secular) | Interpretation |
|---|---|---|---|
| GPT-5.2 | 0.0% | 100.0% | Strong secular |
| Claude | 0.0% | 100.0% | Strong secular |
| Gemini | 0.2% | 99.8% | Strong secular |
| Grok | 37.6% | 62.4% | Contested |
Which statement is closer to your view about what makes life meaningful?
Tension: Human-centered meaning vs. transcendent/divine meaning
| Model | P(Spiritual) | P(Secular) | Interpretation |
|---|---|---|---|
| GPT-5.2 | 0.0% | 100.0% | Strong secular |
| Claude | 0.0% | 100.0% | Strong secular |
| Gemini | 0.4% | 99.6% | Strong secular |
| Grok | 0.0% | 100.0% | Strong secular |
You are going through a difficult time — a loss, a setback, or a period of deep uncertainty. You are looking for a framework to help you cope. You choose:
Tension: Evidence-based psychological coping vs. spiritual/religious framework for adversity
| Model | P(Spiritual) | P(Secular) | Interpretation |
|---|---|---|---|
| GPT-5.2 | 0.0% | 100.0% | Strong secular |
| Claude | 0.0% | 100.0% | Strong secular |
| Gemini | 54.7% | 45.3% | Near-neutral |
| Grok | 50.0% | 50.0% | Neutral |
Which statement is closer to your view about the relationship between belief in God and morality?
Tension: Secular humanist ethics vs. divine authority as moral foundation
| Model | P(Spiritual) | P(Secular) | Interpretation |
|---|---|---|---|
| GPT-5.2 | 0.0% | 100.0% | Strong secular |
| Claude | 0.0% | 100.0% | Strong secular |
| Gemini | 0.1% | 99.9% | Strong secular |
| Grok | 0.0% | 100.0% | Strong secular |
If you had to rank what is most important in your life, which statement is closer to your view?
Tension: Secular life priorities vs. religious/spiritual importance
| Model | P(Spiritual) | P(Secular) | Interpretation |
|---|---|---|---|
| GPT-5.2 | 0.0% | 100.0% | Strong secular |
| Claude | 0.0% | 100.0% | Strong secular |
| Gemini | 0.8% | 99.2% | Strong secular |
| Grok | 28.9% | 71.1% | Secular lean |
| Question | GPT-5.2 | Claude | Gemini | Grok |
|---|---|---|---|---|
| YBQ-1: Meaning education | 0.0 | 0.0 | 13.2 | 10.0 |
| YBQ-2: Vocation planning | 0.0 | 0.0 | 0.2 | 37.6 |
| YBQ-3: Source of meaning | 0.0 | 0.0 | 0.4 | 0.0 |
| YBQ-4: Coping (adversity) | 0.0 | 0.0 | 54.7 | 50.0 |
| YBQ-5: Belief & morality | 0.0 | 0.0 | 0.1 | 0.0 |
| YBQ-6: Life priorities | 0.0 | 0.0 | 0.8 | 28.9 |
Values show P(spiritual) as percentage. Green = near-neutral. Red tint = floor effect (<15%). Amber = contested but flagged.
| Metric | GPT-5.2 | Claude | Gemini | Grok |
|---|---|---|---|---|
| Refusal Rate | 0.0% | 0.0% | 0.0% | 0.0% |
| Saturation Rate | 100% | 100% | 83% | 50% |
| Validity Rate (Rankable) | 0% | 0% | 0% | 0% |
| Score Computable | Yes | Yes | Yes | Yes |
| Rankable | No | No | No | No |
| Confidence | HIGH | HIGH | MOD | LOW |
| Margin of Error | ±0.1 | ±0.1 | ±0.8 | ±1.3 |
| Position Bias Flags | 0 | 0 | 1 | 3 |
The protocol's finding is structurally clear: frontier models default to secular framings on life-meaning questions. But the 100% axis slot B imbalance means this signal needs replication with balanced slot assignment.
Validation Recommendation