# Humrel Theory Development > Use when building the theoretical engine for a Human Relations (HR) manuscript — mechanisms grounded in social theory, process vs. variance logic, constructs, and boundary conditions. Builds the argument; it does not choose methods or run analysis. - Skill: `thedixitjain/humrel-theory-development` (Agent Skill) - Install (CLI): `npx skillmds add thedixitjain/humrel-theory-development` - Raw SKILL.md: https://api.skillmd.com/api/skills/thedixitjain/humrel-theory-development/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: thedixitjain (https://skillmd.com/u/thedixitjain) - Updated: 2026-09-08 - Page: https://skillmd.com/skills/thedixitjain/humrel-theory-development --- # Theory Development (humrel-theory-development) ## When to trigger - You have a work-and-society puzzle (from `humrel-topic-selection`) but no theoretical machinery to explain it - Your "theory" is a box-and-arrow diagram with labelled constructs but no *social mechanism* - A reviewer says the theory is "thin," "post hoc," "atheoretical," or "a relabelling of the finding" - You borrowed a fashionable construct (resilience, agility, the algorithmic) as a veneer over an under-theorized result ## The HR standard: a substantive theoretical contribution drawing on social theory HR's gatekeeping criterion is **"a unique and substantive theoretical contribution"** to understanding work and organizations as social phenomena. Unlike a hypothesis-confirmation register, HR rewards theory that *engages social theory* — power, identity, institutions, practice, sensemaking, emotion, discourse, the labour process — and uses it to illuminate the relational dynamics of work. The deliverable is a change in how readers across organization studies, sociology, and psychology *understand* something, not a coefficient or a managerial recommendation. ## Mechanism first, grounded in a social process A construct is not a mechanism. State the **generative social mechanism** — the relational/organizational process that produces the pattern or sequence: - Name the **actors** and what they are doing relationally (negotiating identity, exercising or resisting power, making sense, categorizing, controlling, caring, contesting). - Name the **social structure** in which they act (occupation, hierarchy, field, institution, network, gender/class/race order, technological regime). - Specify the **processual or causal logic**: why does this actor, in this structure, produce this outcome or this next move? - Specify **boundary conditions**: where does the mechanism hold, weaken, or reverse? HR readers expect scope, not universal laws. A good test: can a reader *predict a new, non-obvious implication* for working life from your mechanism? If not, you have a label, not a theory. ## Process theory vs. variance theory — match the form to the phenomenon - **Process theory** explains *how and why* something unfolds over time: sequences, events, becoming, contestation, feedback. Natural for ethnographic, historical, and critical HR papers. - **Variance theory** explains *how much* an outcome co-varies with antecedents, with moderators/mediators. Natural for survey/quantitative HR papers. HR accommodates both, but punishes a mismatch — variance framing forced onto a processual phenomenon ("control *increased* resistance") flattens exactly the relational dynamics HR exists to study. Let the form of the theory match the form of the phenomenon (this decision feeds `humrel-methods`). ## Critical and reflexive theorizing HR's Tavistock lineage makes critical theorizing first-class: theory that problematizes power, ideology, and taken-for-granted arrangements is welcome — but it must be **generative as well as deconstructive**. A critique that only unmasks, without offering a new way to understand or a constructive reframing, reads as polemic. Pair the critical move with a positive theoretical claim that other scholars can build on. ## Inductive vs. deductive - **Inductive/grounded:** theory is the *output*, presented after the data structure (see `humrel-data-analysis`). A grounded process model can be the *entire* HR contribution. Avoid "theory by adjective" (calling a pattern "dynamic" without specifying the process). - **Deductive:** hypotheses must be *derived* with an explicit causal-logic paragraph each, tied to one core mechanism — not a bag of hypotheses. Moderators must be theoretically motivated, not data-mined. ## Checklist - [ ] A named generative social mechanism (actors + structure + logic), not just constructs - [ ] Theory engages a recognizable social-theoretical conversation, not a borrowed buzzword - [ ] Process vs. variance form matches the phenomenon - [ ] Boundary conditions stated (where it holds / weakens / reverses) - [ ] Mechanism yields at least one non-obvious implication for working life - [ ] Critical papers are generative, not only deconstructive - [ ] Deductive: each hypothesis derived from the core mechanism; inductive: a process model planned ## Anti-patterns - Box-and-arrow models with no mechanism ("constructs in search of a story") - Variance framing imposed on a processual/relational phenomenon - A fashionable construct used as a veneer over a thin finding - Critique that only unmasks, with no constructive theoretical payoff - "Theory by adjective"; hypotheses asserted rather than derived; a moderator zoo ## Output format ```text 【Journal】Human Relations 【Skill】humrel-theory-development 【Theory form】process / variance + why it matches the phenomenon 【Core mechanism】actors + social structure + processual/causal logic 【Social-theory anchor】the conversation it engages (power/identity/institutions/…) 【Boundary conditions】where it holds / weakens / reverses 【Non-obvious implication】the surprise for working life 【Next skill】humrel-literature-positioning ``` --- **Source:** [`brycewang-stanford/Awesome-Journal-Skills`](https://github.com/brycewang-stanford/Awesome-Journal-Skills) → `Human-Relations-Skills/skills/humrel-theory-development/SKILL.md`