The Methodology

Measuring Match

The framework claims that environmental alignment predicts human thriving. Here's how we measure it — domain by domain, with documented ancestral ranges, pre-registered weights, and a formula designed to catch cults.

If Match doesn't predict outcomes, the theory fails. That's what makes it science.

The Problem of Tautology

A framework risks circularity if "Match" is defined by its results. "A matched community is one where people are happy" tells you nothing. To be useful, we must:

Define Match strictly by environmental inputs (independent variables)
Measure outcomes (pathology/thriving) separately
Test whether the inputs predict the outcomes

This is what the Match Score system does.

Plausible Ancestral Ranges

We reject invented benchmarks. Scoring is grounded in documented ranges from extant forager populations — Hadza, !Kung, Ache, Tsimane, Agta.

Domain	Metric	Documented Range (PAR)	Primary Sources
Social	Face-to-face hours/day	4–9 hours	Marlowe (2010), Konner (2005)
Movement	Active hours/day	4–6 hours	Pontzer et al. (2012)
Light	Daylight exposure (>1000 lux)	6–10 hours	de la Iglesia et al. (2015)
Group Size	Daily contact group	20–50 people	Hill & Dunbar (2003)
Care	Alloparents per child	4–20 adults	Hrdy (2009)

Scoring principle: Values within PAR = 100 points. Values diverging from PAR suffer proportional reduction.

The Seven Domains

Seven positive domains are measured and combined using a geometric mean. This enforces Liebig's Law of the Minimum — a deficiency in one essential domain limits overall thriving, regardless of abundance in others.

Domain 1: Social Density & Depth (Weight: 0.25)

Measurement is strictly behavioral/structural, not self-reported feelings.

Band Layer Co-presence (0-40):Hours/day in physical proximity with stable core group (n=5–50). Verified via Bluetooth/audio. Target: 4+ hours.
Bond Infrastructure (0-30):Count of individuals meeting criteria: tenure >5 years OR kinship OR weekly resource exchange. Target: 5+ individuals.
Stranger Ratio (0-30):Percentage of daily interactions involving unknown individuals. Target: <10%.

Domain 2: Agency & Closed Loops (Weight: 0.20)

Replaces subjective "sense of purpose" with the Jurisdiction Test.

For the subject's top 5 active life concerns, we audit:

Means: Do they possess the resources to act now?
Authority: Do they require permission to act?
Physics: Is the outcome determined by their action or external probability?

Metric: Control-to-Responsibility Ratio (CRR). High responsibility matched by high jurisdiction = high score. High responsibility with low jurisdiction (middle management, poverty) = near zero.

Domain 3: Circadian & Environmental Alignment (Weight: 0.15)

Solar Synchrony: Mid-sleep point deviation from solar midnight
Lux Contrast: Ratio of daytime to post-sunset light exposure. Target: >10:1
Sleep Integrity: Duration and fragmentation index via actigraphy

Domain 4: Movement Patterns (Weight: 0.10)

Active Volume: Hours above resting metabolic rate. Target: 4–6 hours
Diversity Index: Count of distinct movement types daily (walk, carry, squat, climb, sprint)
Terrain Complexity: Gait variability measuring surface irregularity

Domain 5: Nature Contact (Weight: 0.10)

Immersion Hours: Time outdoors in >1 hour blocks
Acoustic Ecology: % of day with anthropogenic noise <40dB
Fractal Exposure: Visual analysis of environment (natural vs. rectilinear geometry)

Domain 6: Resource Interdependence (Weight: 0.10)

Convenience Tier (0-30): Can borrow daily necessities without debt/ledger?
Safety Net Tier (0-30): Can access 1 month of resources within 24 hours via informal request?
Existential Tier (0-40): Does the system need you AND do you need the system?

Domain 7: Governance & Exit (Weight: 0.10)

Voice-to-Decision Ratio: Probability that a stated objection modifies a group decision
Exit Cost Index: Inverse of financial/social penalty for leaving. High penalty = Low score.
Information Symmetry: Audit of transparency available to average member

The Exit Cost Index is what differentiates tribes from cults.

The Interference Domain

This domain measures active harms — supernormal stimuli that hijack evolved cognition. These are subtracted from the final score. There is no cap. A thoroughly hijacked individual can score negative.

Interference Type	Metric	Penalty
Parasocial Load	Hours/day in unidirectional social bonding (consume without interact)	2 points per hour
Scope Mismatch Index	Ratio of global news to local/actionable news	5 pts (>2:1), 10 pts (>5:1), 20 pts (>100:1)
Algorithm Exposure	Hours/day with variable-ratio reinforcement (infinite scroll, gacha)	3 points per hour

The Formula

We use a geometric mean to enforce the "limiting factor" principle. A high-scoring social environment with zero agency must fail the test. A cult with great community but no exit rights must fail the test.

Base Match Score

M_Base = ∏⁷_i=1 (S_i + ε)^w_i

SᵢScore of Domain i (0-100)

wᵢWeight of Domain i (sum = 1.0)

ε1.0 (prevents mathematical errors at zero while keeping score functionally zero)

Final Match Score

M_Total = M_Base − I_Interference

Why Geometric, Not Additive

Consider a "Golden Cage" cult:

Domain	Score
Social	95
Governance/Exit	5
Agency	5
Other domains	~50

Additive Model

Score ≈ 50/100

"Moderately Matched"

Geometric Model

Score ≈ 18/100

"Severely Mismatched"

The geometric mean correctly identifies that a high-control environment is not matched, no matter how good the social density looks.

Pre-Registered Weights

To prevent adjusting weights until they fit the data, we fix the theoretical weights before any empirical testing.

Domain	Weight
Social Density & Depth	0.25
Agency & Closed Loops	0.20
Circadian Alignment	0.15
Movement Patterns	0.10
Nature Contact	0.10
Resource Interdependence	0.10
Governance & Exit	0.10

These weights are fixed based on evolutionary priors. The primary hypothesis test uses only these weights.

As secondary analysis, we run Lasso regression to determine empirically observed weights. If empirical weights diverge significantly from theoretical weights — for instance, if Nature explains 40% of variance instead of 10% — this constitutes a finding about human biology, not a license to retrofit the theory.

Accounting for Dropout

Measuring depression prevalence in a community is flawed if depressed people leave. A toxic environment might show 0% depression simply because everyone who struggles gets pushed out.

Solution: Total Prevalence Load (TPL)

TPL = (N_current × P_current) + (N_exited × P_exited) / N_total

P_currentPrevalence of High Pathology (PHQ-9 >10) among current members

P_exitedPrevalence among those who left in last 12 months, measured 3 months post-exit

Protocol

All participants agree to 6-month post-exit follow-up as condition of study entry. If >20% of exiters are lost to follow-up, Maximum Bias Assumption is applied — we assume lost exiters are high-pathology, which penalizes communities that can't retain contact with former members.

Study Design

The Subjects

Group	Examples	Purpose
High Match	Twin Oaks, Ache, Hadza (where ethics permit)	Test upper bound of environmental alignment
Transitional	Cohousing, "pod" living arrangements	Test intermediate conditions
Standard Control	Urban apartment dwellers	Baseline modern mismatch
Negative Control	High-control/Low-agency groups (prisons, strict sects)	Validate geometric mean identifies low agency

Addressing Selection Bias

We cannot randomly assign people to tribes. We use:

Waitlist Controls: Individuals accepted to communities but waiting for openings. They share the "seeking" trait but lack the environment.
Inverse Propensity Weighting: Controlling for baseline mental health, childhood trauma (ACE scores), and socioeconomic status.

Timeline

Phase	Description	Sample
Phase 1	Validate "Jurisdiction Test" and "Bond Infrastructure" metrics against cortisol/HRV markers	N=50
Phase 2	Assess 3 distinct communities using Match Score v7.0. Check for mathematical anomalies	3 sites
Phase 3	24-month longitudinal tracking across all groups	N=500

Falsification Criteria

The framework makes specific predictions. Here are the conditions under which we would conclude it's wrong.

Condition 1: High-Match / High-Pathology

If communities scoring ≥80 on the Match Score show retention-adjusted depression/anxiety prevalence ≥15% (Western baseline), with dropout properly accounted for, the theory fails.

Condition 2: The Prison/Cult Paradox

If environments with high Social Density (>90) but near-zero Agency/Governance (<10) produce high wellbeing, the theory fails. The framework predicts agency is a biological necessity, not a preference.

Condition 3: Null Dose-Response

If an increase in Match Score from 30→70 shows no correlation (r < 0.15) with outcome improvements across N > 1000, the theory fails.

Applications

For Researchers

The weights are pre-registered. The falsification criteria are defined. The study design is specified. Run the study. Replicate, extend, or refute.

For Practitioners

The Match Score gives you a diagnostic tool. Instead of asking "what's wrong with this person," ask "which domains are bottlenecked?" The geometric mean tells you where the limiting factor is.

For System Builders

Before you build, you can score your design. Does your community structure meet the Agency threshold? Does your platform create or reduce Interference? The spec sheet has numbers.

For Communities

Self-assess. Which domain is dragging your score down? The geometric mean makes it visible. Fix the bottleneck before optimizing what's already working.

Participate

The Demismatch framework is open. If you're a researcher interested in running these protocols, a community willing to be assessed, or a funder interested in supporting empirically grounded mental health research — we want to hear from you.

Read the Full Framework See the Evidence Base Contact

Match is measurable. Outcomes are measurable. The prediction is clear: alignment predicts thriving. Test it.