Theory Development (asq-theory-development)

When to trigger

• You have a puzzle (from asq-topic-selection) but no theoretical machinery to explain it
• Your "theory" is a box-and-arrow diagram with labeled constructs but no mechanism
• You are unsure whether your phenomenon needs a process theory or a variance theory
• Reviewers say the theory is "thin," "post hoc," or "a relabeling of the finding"

Process theory vs. variance theory

This is the central choice at ASQ. Get it right before drafting.

• Variance theory explains how much of an outcome varies with antecedents: X → Y, with moderators/mediators. Appropriate for "whether and under what conditions" questions, typically quantitative.
• Process theory explains how and why something unfolds over time: sequences, events, feedback, transformation. Appropriate for emergence, becoming, contestation, and dynamics — typically qualitative/inductive.

ASQ punishes variance framing forced onto a process question: if your data show a sequence of events transforming an organization but you write it as "factor A increases outcome B," reviewers will say the theory and phenomenon are mismatched. Match the form of the theory to the form of the phenomenon. ASQ's lineage makes process and interpretive theorizing fully legitimate as the main contribution — this is a real difference from journals where the variance/hypothesis template is the implicit standard. Barley's (1986) ASQ study of CT scanners, for instance, theorized structuring as a process rather than estimating an effect of technology on structure.

Mechanism first

A construct is not a mechanism. State the generative mechanism — the social/organizational process that produces the relationship or sequence:

• Name the actors and what they are doing (sensemaking, competing, imitating, categorizing, coordinating, resisting). ASQ's sociology-of-organizations roots mean collective and field-level mechanisms (mimetic, normative, and coercive isomorphism in the DiMaggio & Powell 1983 sense) are as welcome as micro ones.
• Name the social structure in which they act (field, hierarchy, network, category system, status order)
• Specify the causal/processual logic: why does this actor, in this structure, produce this outcome or this next step?
• Specify boundary conditions: where does the mechanism hold, weaken, or reverse?

A good test: could a reader predict a new, non-obvious implication from your mechanism? If not, you have a label, not a theory.

For inductive papers

If theory is emergent (grounded), this skill still applies — but in reverse order:

• Theory is the output, presented after the data structure (see asq-data-analysis)
• Still required: a named mechanism, a process model (often a figure with phases/feedback), and a clear statement of what is novel relative to existing organizational theory
• A grounded, inductively-built theory can be the entire contribution at ASQ — it need not be a stepping stone to a later deductive test, as it might be framed elsewhere
• Avoid "theory by adjective": calling a pattern "dynamic" or "recursive" without specifying the process

For deductive papers

• Hypotheses must be derived, each with an explicit causal logic paragraph — not asserted
• Each hypothesis should connect back to the same core mechanism (avoid a "bag of hypotheses")
• Moderators must be theoretically motivated (why this boundary?), not data-driven fishing

Checklist

• Process vs. variance form matches the phenomenon
• A named generative mechanism (actors + structure + logic), not just constructs
• Boundary conditions stated (where it holds / weakens / reverses)
• Mechanism yields at least one non-obvious, testable/observable implication
• Constructs are defined precisely and consistently with the literature
• For deductive: each hypothesis has its own derivation paragraph tied to the mechanism
• For inductive: a process/theory model figure is planned, with novelty stated

Anti-patterns

• Box-and-arrow models with no mechanism ("constructs in search of a story")
• Variance framing for a process phenomenon (or vice versa)
• Hypotheses asserted rather than derived; a moderator zoo with no logic
• "Theory by adjective" — naming a pattern instead of explaining it
• Borrowing a fashionable construct as a veneer over an under-theorized finding

Output format

【Theory form】process / variance + why it matches the phenomenon
【Core mechanism】actors + structure + causal/processual logic
【Boundary conditions】where it holds / weakens / reverses
【Novel implication】the non-obvious prediction or insight
【Deductive only】hypotheses + derivation status
【Next step】asq-literature-positioning