Methods
From observation to a bounded claim
A readable methods guide to claim classes, independent decisions, null values, chronological replay, limitations and the implementation appendix.
Methods begin with a boundary
The method must say what kind of claim it is making.
A captured moment, a policy decision, a deterministic calculation and an interpretation are not interchangeable. Keeping them visibly separate is the first safeguard against false precision.
The reading order
From observation to a bounded claim.
- 01
Observe
Record a bounded event, its time, context, evidence and provenance.
- 02
Decide
Keep source settlement, human impact and valuation as explicit decisions.
- 03
Calculate and interpret
Derive only permitted outputs, then state what they support and what remains unproven.
Claim discipline
Seven claim classes
Every mathematical statement on this layer must declare what kind of statement it is. A public explanation cannot silently turn an implementation detail, individual observation or hypothesis into ratified policy.
01
Ratified policy
A rule adopted by its named authority. It applies only within the scope and version that authority declares.
02
Implementation behaviour
What a particular system currently stores, computes, blocks or displays.
03
Synthetic fixture
A purpose-built teaching example. It tests a branch but contains no personal observation and proves no real-world effect.
04
Observed redacted case
An authorised, minimised observation used to explain a route. It cannot establish a general norm or causal claim.
05
Interpretation
A bounded explanation of what a result may mean. It is not itself a policy rule or research finding.
06
Hypothesis
A testable proposal that remains unvalidated.
07
Open research question
A named uncertainty requiring evidence, methods and governance before stronger claims are justified.
Independent questions
Economic value and human movement are separately governed.
Performance valuation
Did this contribution acquire typed value?
Vcom_event = validated rate × validated duration
VCperformance_event = (Wbefore / 1000) × Vcom_event
A rate must be explicitly validated or deterministically resolved from an approved, versioned policy referenced by the event. If it is unavailable, typed economic fields remain null—not zero.
Human-state movement
Did the represented state change?
decision = applied | no_transition | pending
Wafter = Wbefore + eligible weighted movement
The impact review may move the represented state, declare no transition or remain pending. Economic valuation and wellbeing impact are separately governed and not mechanically coupled.
Do not collapse uncertainty
Five answers to five different questions.
- 01 · Source
- Who settled what happened?
- 02 · Impact
- Was human movement applied, absent or pending?
- 03 · Replay
- May this record enter the active chronology?
- 04 · Valuation
- Is a typed economic output available?
- 05 · Projection
- What may be treated as canonical or preview-only?
Chronology and integrity
A correction changes the future without erasing the past.
01
Earlier active record
An eligible event produces the prior state read by later events.
02
Correction or supersession
The active input changes while the earlier revision remains in the audit chain.
03
Deterministic replay
Downstream derived outputs are recomputed in effective-time order.
Each eligible event reads the exact prior state. One logical occurrence produces at most one transition in a branch. Later canonical inputs and decisions are not silently edited.
Evidence from before the declared starting boundary remains outside the constituted branch. Replay does not convert historical records into new approvals.
What the method supports
Deterministic boundary testing
- Checking typed branches, statuses, arithmetic and chronological replay.
- Showing why value and human movement can diverge.
- Making provenance, null values, corrections and limitations inspectable.
What remains unproven
Validity beyond the fixture
- No population norms, clinical validity, causal effect or cross-person comparability has been established.
- The score is not human worth, eligibility, compensation, diagnosis or a universal ranking.
- A validated contribution does not prove positive wellbeing impact; a positive impact does not prove economic value.
- Synthetic cases test implementation boundaries; they are not empirical evidence.
Falsifiability
The open research programme
- 01Content and construct validity of any declared component model.
- 02Within-person stability, sensitivity and meaningful-change thresholds.
- 03Reviewer agreement for direction, intensity and no-transition decisions.
- 04Component discrimination where one event plausibly affects several nodes.
- 05Missing-data patterns, rate provenance and selection effects.
- 06Cultural and linguistic adaptation, gaming resistance and downstream harms.
- 07Whether any typed value relation predicts outcomes without becoming a coercive target.
Implementation evidence—not validationIssuer-policy appendixOpen the current authority map, implementation-specific statuses, precision rules and observed-evidence limitations.
Scope and authority
The current implementation is a founder-specific, longitudinal HouseOfBros instrument. HouseOfBros owns its six components, weights, transition matrix, Genesis boundary and issuer-specific score. The IoV Spec owns shared envelopes, typed fields, replay, correction and interoperability. Value Log implements capture and deterministic replay. Word Wallet controls private projection and disclosure.
The Spec claims matrix is an ownership and status index. It points to the normative owner; it does not override HouseOfBros policy or another repository’s authority.
Current settlement behaviour
- Source settlement
- Final, provisional, pending or not applicable. GFE founder self-review remains provisional here.
- Impact review
- Applied, no_transition or pending under HouseOfBros authority.
- Replay eligibility
- Eligible, blocked by a pending authority decision, or excluded_pre_genesis. Exclusion is not pending.
- Valuation
- Valued, not_yet_valued, inputs ready but noncanonical, or not applicable to the current Non-Performance branch.
- Projection
- Canonical, pending, provisional preview only, or excluded_pre_genesis.
Current Non-Performance behaviour
For Physiology, Emotion, Feeling, Thought and Habit, the current implementation applies the relevant HouseOfBros transition and records Wafter and deltaW. Both are WellbeingScore units.
They are not economic VC. They are not money, credits or a price for personal life. The current branch is not an operational Non-Performance implementation of VC = W × Vcom. Any future Non-Performance value relation remains a research question unless separately specified and ratified.
Precision and presentation
Value Log and the synthetic casebook use ECMAScript Number arithmetic (IEEE-754 binary64). Governed calculation steps use Math.round(value × 10¹²) / 10¹². This is the verified production implementation and no stronger decimal rounding semantics are asserted.
Public Wellbeing display uses toFixed(3) for presentation only. Twelve stored decimal places support deterministic comparison; they do not claim twelve-decimal measurement accuracy.
Implementation limitations
- Founder self-review is not independent external review. GFE founder review remains provisional until an eligible independent reviewer settles it.
- The economic rate-bearing founder branch has not yet been demonstrated through an observed post-starting-point founder event. Every rate-bearing casebook test example is synthetic.
- The pinned policy is authoritative only for its declared issuer-specific implementation. It is not a universal human model.
Evidence map
Academic bibliography and claim-to-source map
Cronbach & Meehl (1955). Construct validity in psychological tests.
Use here: Frames validation as an evidence programme rather than a label attached to a score.
Shrout & Fleiss (1979). Intraclass correlations: uses in assessing rater reliability.
Use here: Supports explicit selection and reporting of reliability models when independent raters are introduced.
Stone & Shiffman (1994). Ecological momentary assessment in behavioral medicine.
Use here: Provides a primary methodological basis for collecting observations close to lived events while limiting retrospective reconstruction.
Shiffman, Stone & Hufford (2008). Ecological momentary assessment.
Use here: Reviews event-contingent and time-contingent within-person observation, burden and reactivity limitations.
Molenaar (2004). A manifesto on psychology as idiographic science.
Use here: Supports separating within-person longitudinal inference from population-level or cross-person claims.
Vohra et al. (2015). CENT 2015: reporting guideline for N-of-1 trials.
Use here: Original CENT 2015 article. Used only as a methodological reporting reference; the founder implementation is not asserted to be a clinical N-of-1 trial.
BMJ correction (2016). Correction: CENT 2015.
Use here: Correction notice for the CENT 2015 article; it is not the original guideline DOI.
Beaton et al. (2000). Cross-cultural adaptation of self-report measures.
Use here: Supports adaptation and equivalence checks before claims across language or culture.
Campbell (1979). Assessing the impact of planned social change.
Use here: Motivates safeguards against turning an indicator into a target or human ranking.
Terwee et al. (2018). COSMIN content-validity methodology.
Use here: Used only as methodological vocabulary for future content-validity work; it does not make the founder implementation a PROM.
Mokkink et al. (2010). COSMIN checklist.
Use here: Provides methodological vocabulary for future measurement-property evaluation, not evidence that such properties are established here.
External sources inform evaluation methods. They do not ratify IoV equations, HouseOfBros parameters or a clinical interpretation.