English · Русский · Українська

Status

Research design / not validated and not ready for individual verdicts. The instrument reports evidence about proposed operations and position modes. It does not reveal a true type.

Construct frame

The target is the BWB hypothesis in Socionics as a Hypothesis of Latent Reality Modeling: an aspect pairs a class of distinctions with an operation for selecting, compressing, organizing, inferring from, and updating a partial representation of one shared world. Aspect evidence and Model A position evidence are separate constructs and must not be inferred from the same thematic preference.

Four measurement facets

Facet Evidence question
Selection What is noticed without an aspect cue?
Compression Which distinctions survive summarization, delay, or limited capacity?
Inference What novel conclusion, prediction, or alternative is generated?
Updating What error signal triggers revision, and how well is the representation corrected?

Record accuracy where a task has an answer, coverage, error pattern, response time, confidence calibration, transfer, and revision after feedback. No single metric is an aspect score by itself.

Eight provisional operation scales

Aspect Proposed operation
Ne latent properties and unrealized possibilities
Se present force, form, boundaries, resistance, and direct impact
Te action/method links to observable results and feedback
Fe changing activation and state through expressed reactions
Ni timing, trajectory, and transition among sequential processes
Si concurrent conditions and their effects on state and quality
Ti structural relations, classifications, constraints, and invariants
Fi subjective significance, attraction, need, distance, and acceptability

These scales are candidate constructs, not established eight-factor structure.

Position-mode evidence

Test position claims independently across repeated tasks. Candidate evidence concerns foreground versus background use, accepting versus producing, stable versus situational use, response to pressure, reliance on external scaffolding, and valuation. The Model A labels leading, creative, role, vulnerable, suggestive, mobilizing, ignoring, and demonstrative remain hypotheses. Positions 1/2/7/8 are not ability scores; 3/4/5/6 are not deficits.

Four parallel microtask families

  1. Condition state ↔ direct impact (Si/Se). Present the same changing physical or resource scene. Measure detection and updating of condition-quality effects separately from capacity, boundary, resistance, and leverage estimates.
  2. Temporal trajectory ↔ possibility space (Ni/Ne). Present an event sequence with hidden alternatives. Measure timing/direction inference separately from discovery of viable latent properties and branches.
  3. Structural rule ↔ effective method (Ti/Te). Present a novel system with rules and outcome feedback. Measure invariant/constraint recovery separately from choosing and revising a result-producing procedure.
  4. Stable relation ↔ expressed-state dynamics (Fi/Fe). Present interaction traces over time. Measure inference about significance and distance separately from detection of expressed-state transition and propagation.

Each family needs matched difficulty, equivalent information access, counterbalanced order, neutral wording, and tasks where either operation can help without one being socially superior.

Auxiliary self-report

Self-report may ask about perceived effort, preference, support, and familiarity after behavioral tasks. It cannot serve as the primary aspect measure because identity, prestige, prior typing knowledge, role, and memory can imitate the target. Observer reports and interviews are also auxiliary.

Public output contract

The report must contain four separate blocks:

  1. Aspect-operation evidence: selection, compression, inference, and updating observations for all eight candidate operations.
  2. Position-mode evidence: separate repeated-task evidence, including contradictions and dependence on support or pressure.
  3. Context and rivals: expertise, language, health, sensory access, role, stress, familiarity, motivation, missingness, and measurement error.
  4. Ranked Model A hypotheses: only when preregistered minimum evidence exists for both aspect and position structures; otherwise report that no code inference is warranted.

Never report an overall ability score, innateness score, compatibility score, calibrated-looking percentage without calibration data, or a type inferred only from preferred topics.

Analysis sketch

  • Preserve task-level observations before aggregation.
  • Estimate facet profiles with uncertainty and report missing data.
  • Test whether aspect operation patterns replicate across task variants.
  • Estimate position-mode evidence separately; do not backfill it from the expected stack.
  • Compare candidate profiles with simpler baselines such as expertise, general task performance, response style, and broad traits.
  • Attempt type-like classification last, on held-out data, and permit “insufficient evidence” as the default outcome.

Validation requirements

Use cognitive pretesting, blinded coding, inter-rater reliability where relevant, test-retest designs, factor/model comparison, measurement invariance across EN/RU/UK and relevant groups, convergent and discriminant measures, held-out prediction, external replication, calibration, harm review, and preregistered failure thresholds.

Caveats

  • Task success may reflect knowledge, intelligence, reading ability, motor or sensory access, or familiarity rather than the proposed operation.
  • A stable aspect-like pattern does not establish Model A positions or biological innateness.
  • All people may use all operations; the target is a recurring differential pattern, not ownership of a channel.
  • No result authorizes moral, clinical, employment, relationship, or safety decisions.

Next reading

See Typology Test Design Protocol, Compatibility Measurement Methods, and Socionics as a Hypothesis of Latent Reality Modeling.