peopleanalyst

research / principia / reports

Construct-family roadmap

17 construct families across 4 tiers (foundational · derivative · composite · outcomes), 8 queued for sequencing, 3 parallel threads (surveys · infrastructure · book draft).

Principia·Reports·source: people-analyst/principia/docs/research/PROGRAM.md

Principia — research program

The construct-family roadmap, sequenced by literature density and measurement-model maturity. This is what gets surveyed, in what order, and why.

Three parallel threads

The work runs in three threads concurrently:

  1. Construct-family surveys — the bulk of the work. One construct family at a time, end-to-end, before moving to the next.
  2. Infrastructure — the queryable database (Postgres + API), the @measurement/core consumer wiring, the verification log surface, the book renderer.
  3. Book draft — assembled from completed surveys. Each completed survey is a chapter draft; the introductory and synthesis chapters are written last.

The surveys lead. Infrastructure follows enough to support the surveys. The book trails — it is the last rendering, not the first deliverable.

Construct-family roadmap

Sequenced by literature density and measurement-model maturity (see methodology.md §4 for the rationale).

Tier 1 — foundational constructs (densest literatures)

#FamilyWhy firstAnchor instruments
1EngagementDensest accumulated literature; methodology proof-of-method. Three measurement traditions in tension (Schaufeli's UWES, Kahn's qualitative tradition, the practitioner-survey tradition) — surveying engagement is itself a methodology contribution.UWES, Gallup Q12, MEI, JES
2Job satisfactionLong history; classic measurement-model debates (global vs facet; affective vs cognitive).JDI, MSQ, JSS, Brief Index
3Organizational commitmentTripartite measurement model (Allen & Meyer affective/continuance/normative) is canonical and well-tested.OCQ, TCM (Allen & Meyer)
4Organizational climateMultiple traditions (climate vs culture; specific climates: safety, service, innovation, justice). Survey clarifies the ontology.OCDQ, Litwin & Stringer, climate-for-X variants

Tier 2 — derivative constructs (rest on tier 1 scaffolding)

#FamilyWhy second
5Organizational justiceDistributive / procedural / interactional / informational measurement model is dense and well-validated.
6Psychological safetyEdmondson 1999 onwards. Construct is dense; instruments cluster around a core.
7Perceived organizational supportSPOS family. Tight literature; clean validation work.
8Work-related stress, strain, burnoutMBI, OLBI, BAT. Construct boundaries against engagement matter — the survey adjudicates.

Tier 3 — composite constructs (heterogeneous, require disambiguation)

#FamilyWhy third
9LeadershipTransformational, transactional, authentic, servant, ethical, charismatic — overlapping instruments, contested measurement-model assumptions. The survey provides the disambiguation.
10PerformanceTask / contextual / counterproductive / adaptive. Multi-source rating, behaviorally-anchored vs trait-rated debates.
11Organizational citizenship behavior (OCB)Heterogeneous taxonomy; overlap with contextual performance; survey clarifies.

Tier 4 — outcomes (covered after antecedents land)

#FamilyWhy fourth
12Turnover (intent and actual)Surveyed after antecedents are in place so the predictor-outcome rows have a target schema.
13Withdrawal behaviorsLateness, absenteeism, work-effort reduction.
14Counterproductive work behavior (CWB)Adjacent to performance and OCB; surveyed alongside them.

Tier 5 — adjacent / methodology

#FamilyWhy surveyed at all
15Psychological capital (PsyCap)Composite of hope, efficacy, resilience, optimism. Surveyed because the composite measurement model is itself a methodology case study.
16Person-environment fit familyPerson-job, person-organization, person-supervisor, person-group fit. The measurement model varies more than the literature acknowledges; surveying clarifies.
17Work design / job characteristicsJDS, WDQ. Foundation for predictor variables in many of the above families.

Open / queued for sequencing decision

  • Innovation, creativity at work
  • Identification (organizational, occupational, team)
  • Trust (organizational, supervisor, team)
  • Voice and silence
  • Workplace incivility, harassment, mistreatment
  • Diversity climate and related workplace-experience constructs
  • Compensation justice and pay-related attitudes (intersects directly with AnyComp on the People Analytics Platform)

Per-family work-product

Every construct-family survey ships the same artifact:

  1. Construct definition + history — what it is, where it came from, the measurement-model debates that shaped how it gets measured today
  2. Instrument inventory — every instrument that has measured the construct, with item-level schemas (item text, response options, scale anchors, reverse-coding, scoring rules)
  3. Reliability / validity evidence — per instrument, cited from the validation literature, source-graded
  4. Effect-size table — the construct's predictors and outcomes, with study-quality grades, replication status, and population notes
  5. Cross-cultural notes — where the construct travels, where it does not
  6. Operational notes for builders — implementation pitfalls, common errors, instrument-choice guidance
  7. Threats and limitations — coverage bias, extraction-error risk, author-position notes

The first survey (engagement) doubles as the methodology proof-of-method. Its threats-and-limitations section will be more elaborate than later surveys' because it documents the methodology's own first-pass mistakes.

Sequencing constraints

Several construct families wait on @measurement/core extraction (vela ASN-1013) because their effect-size tables cannot land in a typed registry that does not yet exist. Until the package extracts, surveys produce drafts in markdown and Construct / Instrument / Item surrogates that match the meta-factory schema names; once @measurement/core lands, a single import-rewrite sweep migrates all surveys to canonical types.

The construct-family ordering is also subject to external pull. If the People Analytics Platform's Calculus or AnyComp work needs a specific construct surveyed before the foundational order would otherwise reach it, the construct-family ordering re-sequences. The roadmap is durable; it is not rigid.

What this is not a roadmap for

  • Empirical research on these constructs. Original research happens in Vela, the People Analytics Platform, and Fourth & Two. Principia surveys; the apps test.
  • Instrument creation. Principia indexes existing instruments; it does not develop new ones. Where coverage gaps exist (Mike's adaptive-measurement work in Vela's Reincarnation engine is one example), Principia will register the new instrument once it has primary validation evidence.
  • Replication studies. Principia surfaces replication status as a column in the effect-size table; it does not run replications. Replications are research-app work.

Status

Tier 1 — engagement queued as the first survey, blocked on @measurement/core extraction. See PIPELINE_STATUS.md.