Theory Development (orgsci-theory-development)

When to trigger

• Your hypotheses or propositions are descriptive ("A relates to B") with no organizational mechanism
• You have rich inductive data and need to build, not test, theory
• A reviewer says "what is the theoretical engine?" or "the levels are conflated"
• You need to bridge micro behavior and macro structure explicitly

Two legitimate modes — pick deliberately

Organization Science is theory-driven but methodologically eclectic, so theory can be built two ways and the journal respects both:

• Deductive / a priori. State the organizational construct relationships, derive the mechanism from organization theory, and articulate testable propositions or hypotheses before the data. Make the mediating mechanism and the boundary conditions explicit.
• Inductive / grounded. When the contribution is a new process or construct, build theory up from data. Aim for an abstracted process model or a grounded set of propositions — not a description. The journal's signature openness to qualitative and inductive work means this is a first-class path, not a fallback.

Build the mechanism, not just the relationship

• Name the organizational mechanism. Routines, sensemaking, identity work, learning, search, selection, legitimation, coordination, power — say which engine drives the effect and why it is organizational, not merely psychological or economic.
• Specify the level(s) and any cross-level bridge. Be explicit about whether the construct lives at the individual/team (micro), organization (meso), or field/population (macro) level. If you cross levels, state the bridging logic (emergence, composition, top-down constraint) — conflating levels is a common reject signal.
• State boundary conditions. When does the mechanism operate, and when does it break? Boundary conditions are part of the theory, not a limitations afterthought.
• Theorize, do not gap-spot. A "no one has studied X" framing is not theory. The contribution can come from a new mechanism, a new setting that tests generalizability, or a reframing — novelty alone is "neither necessary nor sufficient."

Anti-patterns

• Hypotheses with no mechanism; mediation labeled but never theorized.
• Inductive findings written as thick description with no abstracted model.
• Level conflation (individual data, organizational claims) with no composition logic.
• Borrowing a discipline's theory wholesale without an organizational translation.

Theory pass for Organization Science

Treat this skill as an executable review pass, not a prose hint. First lock a level map, a mechanism paragraph, and the cover-letter contribution statement; then judge whether the current manuscript answers the venue's real reader: interdisciplinary organization reviewers who ask whether the mechanism travels across levels of analysis.

• Do the pass: Name the construct, mechanism, boundary condition, and falsifiable implication separately; do not let a literature summary masquerade as theory.
• Return a ledger: give claim / evidence / risk / manuscript location rows, so the next agent can edit rather than rediscover the issue.
• Sibling guard: compare against AMJ for empirical management framing, ASQ for organization-theory depth, Management Science for formal/quantitative operations; if a sibling owns the contribution, recommend re-routing before polishing format.
• Submission-ready gate: do not give final advice until the pack's resources/official-source-map.md has been checked for upload-week rules and the manuscript has one concrete fix for the largest venue-specific risk.

Output format

【Mode】deductive a priori / inductive grounded
【Mechanism】the organizational engine driving the effect
【Level(s)】micro / meso / macro / cross-level bridge logic
【Propositions/Hypotheses】or grounded process model
【Boundary conditions】when it holds / breaks
【Next step】orgsci-literature-positioning or orgsci-methods