Purpose
This protocol defines how Before We Build should design, review, validate, and interpret typology-based instruments for Socionics, Psychosophy, and Temporistics.
The target “type” is a model hypothesis about perception and experience organization, not a personality type. A proposed natural predisposition is a separate unvalidated hypothesis. See Reconceiving Typologies.
It exists because typology tests can easily become overconfident: a questionnaire may look precise while measuring social desirability, community stereotypes, verbal style, mood, or prior self-typing rather than the intended construct.
The goal is to build instruments that translate typological hypotheses into measurable constructs while preserving empirical caution.
Scope
This protocol applies to:
- self-report questionnaires;
- observer-rating forms;
- interview coding schemes;
- scenario-based tests;
- behavioral task proposals;
- hybrid typing instruments.
It does not assume that traditional type categories are already valid. Validation must occur at the level of constructs, scales, profiles, and decision rules.
A score produced by an individual typology instrument is not a compatibility score for a pair. It may summarize response evidence for a construct or type hypothesis only; it cannot be converted into a pair verdict without a separately defined and validated outcome model, which Before We Build does not currently have.
Before We Build Construct Frame
All test design should preserve the project's three-system framing:
- Socionics → latent processes of selecting, compressing, inferring from, and updating partial models of one shared reality;
- Psychosophy → latent processes of synthesis and analysis in action;
- Temporistics → latent processes of abduction, induction, and deduction in temporal/existential experience.
These mappings are working hypotheses, not established psychological facts.
Socionics instruments must keep aspect-operation evidence separate from Model A position-mode evidence, cover selection, compression, inference, and updating, and treat self-report as auxiliary. The system-specific design is defined in Socionics Test Specification and Socionics as a Hypothesis of Latent Reality Modeling.
Recommended Expert Routing
Before We Build should use one shared psychometric pipeline rather than separate psychometric experts for each typology.
| Role | Responsibility |
|---|---|
| Domain expert | Defines the typology-specific construct and checks theory fidelity |
| Psychometrics methodologist | Converts constructs into measurable variables and reviews item quality |
| Statistical validation agent | Tests reliability, dimensionality, calibration, uncertainty, and invariance |
| Empirical claims caveat reviewer | Prevents overclaiming and deterministic interpretation |
| Ethics reviewer | Required when real participant data, sensitive use, or recommendations are involved |
Domain experts include:
- Socionics Overview / Socionics researcher for information-modeling constructs;
- Psychosophy Overview / Psychosophy researcher for action-priority constructs;
- Temporistics Model / Temporistics researcher for temporal/existential constructs.
Why Not Separate Test Experts Per Typology
Separate test experts for each system are not the first choice because psychometric quality standards are shared across systems:
- construct definition;
- item clarity;
- reliability;
- validity;
- invariance;
- calibration;
- bias control;
- uncertainty reporting.
The better structure is:
- centralized psychometric standards;
- distributed domain review for Socionics, Psychosophy, and Temporistics.
For recurring test and item-bank review, Before We Build uses typology-test-evaluation-expert as a coordinating safety layer. It does not replace the psychometrics, statistics, caveats, ethics, or domain-specific experts. A future psychometric-item-bank-steward may still be proposed if item-bank maintenance becomes a large recurring workload.
Core Design Principles
- Measure latent constructs, not lore labels.
- Prefer observable tendencies over identity statements.
- Separate theory-informed hypotheses from validated findings.
- Avoid deterministic claims about personality, compatibility, role fit, or life outcomes.
- Treat type assignment as model-based inference, not direct observation.
- Preserve dimensional scores before deriving categories.
- Avoid exact confidence percentages unless calibrated against validation data.
Construct Definition Template
Each proposed scale or subscale should document:
- Construct being measured - plain-language definition; - typology source system; - expected manifestations; - plausible opposites or neighboring constructs.
- Why it matters - theoretical role in Before We Build; - expected relation to profiles, behavior, compatibility, or role-fit hypotheses.
- Observable indicators - recurring choices; - attentional patterns; - decision habits; - narrative style; - interaction patterns.
- Exclusions - what the construct is not; - adjacent constructs likely to contaminate measurement.
Item Writing Rules
Items should:
- describe concrete tendencies or recurring patterns;
- avoid typology jargon where possible;
- avoid prestige-loaded wording;
- avoid forcing identification with a type narrative;
- avoid double-barreled statements;
- avoid metaphysical certainty claims;
- avoid obvious “I am type X” cueing;
- be understandable across languages and educational backgrounds.
Item Review Checklist
Every item should be reviewed for:
- construct contamination;
- acquiescence bias;
- social desirability bias;
- reading complexity;
- translation difficulty;
- self-typing cueing;
- excessive abstraction;
- cultural loading;
- emotional valence imbalance.
Recommended Item Mix
Use multiple formats where possible:
- direct self-report items;
- scenario-based items;
- forced-tradeoff items where justified;
- observer-report variants for externally visible tendencies;
- interview prompts;
- optional behavioral indicators if feasible.
No single item format should be treated as sufficient for final type inference.
Item Lifecycle
- Theory extraction — convert typological claims into candidate latent constructs.
- Operational definition — specify what would count as evidence for the construct.
- Initial item pool — create broad, redundant item coverage.
- Expert review — domain + psychometric review for wording, theory alignment, and contamination.
- Cognitive pretesting — interview participants about comprehension and response process.
- Pilot study — estimate distributions, missingness, variance, and basic dimensionality.
- Item reduction — remove weak, redundant, unstable, or ambiguous items.
- Validation study — test reliability, structure, convergent/discriminant validity, and decision rules.
- Replication — re-test in new samples, contexts, languages, and modes.
- Ongoing monitoring — track drift, subgroup bias, calibration decay, and misuse.
Recommended Measurement Architecture
Preferred order:
- Build construct scales first.
- Derive profiles second.
- Infer type-like classifications last, if justified.
This prevents the system from forcing weak dimensional evidence into rigid type categories too early.
Validation Framework
Reliability Checks
- internal consistency where appropriate;
- test-retest stability;
- inter-rater reliability for observer/coder instruments;
- alternate-form checks if parallel item sets are used.
Validity Checks
- Content validity: expert mapping of items to constructs;
- Convergent validity: relation to nearby measures where theoretically expected;
- Discriminant validity: separation from mood, status, verbal style, social desirability, broad traits, and ideology;
- Criterion validity: relation to relevant behaviors or judgments;
- Predictive validity: cautious testing of future outcomes where justified.
Structural Checks
- dimensionality / factor structure;
- item discrimination;
- local dependence;
- subgroup functioning;
- calibration stability across samples.
Invariance Checks
Test across:
- language;
- culture;
- gender;
- age;
- testing context;
- self-report vs observer-report mode.
Threats to Validity
Common threats include:
- circular validation against prior type assignments;
- construct contamination from Big Five-like traits, intelligence, status, mood, or ideology;
- self-typing bias from participant familiarity with typology systems;
- demand characteristics in typology communities;
- overfitting small enthusiast samples;
- translation drift across languages;
- category inflation from weak dimensional evidence.
Interpretation Rules
- Report findings as evidence for measured tendencies, not proof of fixed essence.
- Do not describe classifications as clinical diagnoses.
- Do not infer compatibility outcomes without direct criterion studies.
- Do not collapse low-validity scales into high-confidence type labels.
- Do not use test scores as the sole basis for high-stakes recommendations.
- Label evidence status: source-backed, expert hypothesis, pilot finding, validation finding, or speculation.
Minimum Evidence Threshold Before Public Use
Before operational deployment, require:
- documented construct definitions;
- pilot-tested item pool;
- acceptable reliability;
- at least one independent validation sample;
- caveat-reviewed interpretation language;
- ethics review if personal data are collected or if results affect recommendations.
Ethics Requirements
Ethics review is required when a test collects real participant data or is used for matching, career, military, dating, team, or other opportunity-affecting recommendations.
Minimum requirements:
- informed consent;
- data minimization;
- deletion and retention policy;
- separation of raw and derived data;
- no hidden profiling;
- no sensitive inference without explicit review;
- no deterministic or punitive use of typology results.
Recommended Next Steps
- Define construct maps for Socionics, Psychosophy, and Temporistics.
- Build broad item banks without typology jargon.
- Run expert domain review.
- Run psychometric item review.
- Conduct cognitive interviews.
- Pilot with heterogeneous samples.
- Test dimensional structure and reliability.
- Evaluate convergent/discriminant validity.
- Test invariance across languages and groups.
- Only then evaluate classification usefulness.
See Also
- Validation Program — Overall validation framework
- Compatibility Measurement Roadmap — Prerequisites for any future pair measure
- Psychosophy Typing Methods — Existing psychosophy typing methods note
- Socionics Model A — Socionics Model A hub
- Afanasyev's Psychosophy Model — Psychosophy model structure
- Temporistics Model — Temporistics model structure
- Epistemic Status and Inference Limits — Research positioning and inference limits