jom-theory-development

Theory Development for Empirical OM (jom-theory-development)

When to trigger

• Your hypotheses describe what happens but not why, operationally
• You are importing a reference theory (TCE, RBV, agency, institutional, behavioral decision theory, contingency, queueing/coordination logic) and must adapt it to an operations setting
• A reviewer says "this is a correlation, not a mechanism" or "the theory could apply to any context"
• You need to specify mediators/moderators that operations data can actually identify

JOM's theory bar

JOM is empirical, but empirical strength without theory reads as a technical report. The argument must give an operations mechanism — a causal logic rooted in how work, capacity, inventory, information, incentives, or human behavior in operations actually function — not a generic management story bolted onto an OM dataset. Because JOM explicitly excludes purely analytical/optimization work, your theory is verbal and falsifiable, developed to be tested against observation, not derived as an optimization proof.

Build the mechanism a priori

1. Name the operations phenomenon and its level (task, shift, line, plant, project, dyad, supply network). Operations theorizing is acutely level-sensitive.
2. State the core mechanism in operational terms — e.g., how variability propagates, how slack absorbs shocks, how a behavioral bias distorts ordering, how coordination cost rises with interdependence.
3. Derive hypotheses before seeing results. Post-hoc hypothesizing (HARKing) is a reject signal. For intervention-based and field work, state the theorized effect of the intervention up front.
4. Adapt, don't import wholesale. Show why the borrowed theory needs modification in the operations context; that modification is often the contribution.

Behavioral vs. operational vs. organizational logic

JOM houses behavioral/empirical OM strongly. Be explicit about which engine drives the effect: a behavioral mechanism (heuristics, bias, fatigue, learning) needs human-decision evidence; an operational mechanism (flow, congestion, buffering) needs process/transaction evidence; an organizational/economic mechanism (governance, incentives, contracts) needs relational/firm evidence. Mismatched theory and evidence is a common rejection.

Mediation / moderation

• Mediation: specify the operational process variable that transmits the effect; plan to measure it, not assume it.
• Moderation: contingency theory is native to OM — state the operational condition (e.g., demand uncertainty, automation level, supplier dependence) that strengthens/weakens the effect, with a directional rationale.

Anti-patterns

• A generic management mechanism that ignores operational structure.
• Optimization-style "the firm chooses to minimize cost" framing instead of an empirically testable behavioral/operational claim.
• HARKing; hypotheses reverse-engineered from the output.
• Importing a theory verbatim with no OM-specific adaptation.

Matching the mechanism engine to the evidence

JOM reviewers expect the theoretical engine and the evidence type to align; the map below is interpretive guidance, confirmed against current Department missions.

| Mechanism engine | Evidence it requires | Mismatch flagged |

|------------------|----------------------|-------------------|

| Behavioral (bias, fatigue, learning) | Human-decision data | Bias inferred from firm-level archival |

| Operational (flow, congestion, buffering) | Process/transaction data | A flow claim from survey perceptions |

| Organizational/economic (governance) | Relational/firm-level data | A governance claim as individual-level |

| Contingency (a condition moderates) | A measured operational moderator | A contingency asserted but unmeasured |

Desk-reject and return triggers on theory

• A generic management mechanism bolted onto an OM dataset, no operational structure doing work.
• Optimization-style "the firm minimizes cost" framing instead of an empirically testable claim.
• HARKing — hypotheses reverse-engineered from the output, not derived a priori.

Worked vignette: building an operations mechanism a priori

A team theorizes that automation reduces operator errors but the benefit reverses past a threshold because automation erodes situational awareness (illustrative). The phenomenon is operational and line-level; the mechanism is behavioral-operational: automation offloads routine cognition (lowering error) yet degrades the operator's mental model (raising error on exceptions). The reference theory is automation-complacency, adapted to a high-variability setting. Hypotheses precede estimation: H1, automation lowers routine error; H2, it raises exception error; H3, the net effect is non-monotonic, moderated by exception frequency. Because the prediction is verbal and falsifiable against observed error logs, it is JOM-shaped — not an optimization proof.

Theory objections and the mechanism-grounded fix

• "This is a correlation, not a mechanism." State the operational logic that generates the effect and name the process variable that transmits it.
• "The theory could apply to any context." Show why the mechanism depends on operational structure (capacity, flow, interdependence) a generic setting lacks.
• "The borrowed theory is used off the shelf." Make the OM-specific adaptation explicit — that adaptation is often the contribution.

Output format

【Phenomenon & level】...
【Core OM mechanism】(operational / behavioral / organizational) ...
【Reference theory + adaptation】...
【Hypotheses】H1 ... (a priori, directional)
【Mediator / moderator】operational variable to be measured ...
【Next step】jom-literature-positioning