brycewang-stanford Conceptual Figures & Propositions Tables (amr-tables-figures)
AMR exhibits do theoretical work. There are no scatterplots, bar charts, or regression tables — there is no data. Every figure must be readable directly off the theory: each box is a defined construct, each arrow is a stated proposition or mechanism.
When to trigger
- The theory is built but has no visual model, or the model is decorative
- Your figure is boxes and arrows with no mechanism behind the arrows
- You have many propositions and the reader cannot see their structure
- A typology is asserted in prose but never shown as a grid
The exhibit toolkit
| Exhibit | Use it when | It must show |
|---|---|---|
| Process model | The theory is a sequence / becoming | Stages, triggers, feedback loops, timing — and the engine that drives movement |
| 2x2 (or NxM) typology | The phenomenon has theoretically distinct types | Two theory-derived dimensions; mutually exclusive, jointly exhaustive cells; ideally theorized transitions between cells (cf. Oliver's antecedent-conditions × responses logic, AMR 1991, DOI 10.5465/amr.1991.4279002) |
| Multi-level framework | The action spans individual ↔ collective | Levels as horizontal bands; emergence (bottom-up) and top-down arrows labeled |
| Variance/relational model | Core constructs and their specified links | Constructs as nodes; arrows labeled with proposition numbers and form (causal/moderating/mediating) |
| Propositions table | There are many propositions | Pn |
Design rules (what reviewers check)
- Every box is a defined construct. No box should appear that was not defined in the text.
- Every arrow is earned. Each arrow corresponds to a proposition or a stated mechanism; label it (e.g., "P3") so the figure and text are in lock-step.
- The mechanism is visible or cited. A bare arrow with no associated mechanism is decoration. If the space is tight, label the arrow with the proposition number whose mechanism the text supplies.
- The figure adds, not repeats. It should let a reader grasp the structure of the argument at a glance — not merely re-list the propositions.
- Levels and time are explicit when the theory is multi-level or processual.
- Consistency: construct labels in the figure match the text exactly; numbering matches.
The empirical-exhibit reframe (what each AMR exhibit replaces)
| Empirical sibling exhibit (AMJ/ASQ/SMJ) | AMR conceptual replacement |
|---|---|
| Table 1: descriptive statistics / correlations | Construct-definition table (definitions + scope conditions) |
| Regression / coefficient table | Propositions table (P1...Pn) |
| Marginal-effects / interaction plot | Conceptual model figure with moderator arrows |
| Event-study / time-series plot | Process model with stages and feedback loops |
If any exhibit reports a number estimated from data, the manuscript is a misfiled empirical paper. A construct-definition table also supports construct clarity (Suddaby, AMR 2010, DOI 10.5465/amr.2010.0419).
Propositions table template
| # | Proposition (constructs + form) | Mechanism (why) | Boundary condition |
|----|--------------------------------------------------|-------------------------|--------------------|
| P1 | The stronger A, the more likely B | M1: ... | holds when C |
| P2 | A's effect on B is amplified by D | M2: ... | weakens when ... |
Checklist
- Every box/node is a construct defined in the text
- Every arrow maps to a proposition or a stated mechanism, and is labeled
- The figure shows argument structure, not just a restatement
- Typologies use theory-derived dimensions; cells are MECE; transitions theorized if relevant
- Process models show stages, triggers, feedback, timing, and the driving engine
- Multi-level frameworks label emergence and top-down effects and the levels
- A propositions table accompanies any paper with more than ~4 propositions
- Figure labels and numbering match the text exactly
- Resolution / format meet AOM submission requirements (verify current guidelines)
Anti-patterns
- A "model" figure of boxes and arrows with no theoretical mechanism behind the arrows
- A figure introducing constructs that never appear in the text
- A typology whose dimensions were chosen for tidiness, not derived from theory
- Arrows with no direction or no form (is it causal? moderating? nobody can tell)
- A figure that simply re-renders the propositions list, adding nothing
- Decorative clip-art / gradients that obscure the logical structure
Output format
【Exhibit type】process / typology / multi-level / relational + propositions table
【Constructs shown】all defined in text? yes/no
【Arrows】each maps to Pn or mechanism? yes/no
【Adds beyond text?】shows structure / merely repeats (fix)
【Format check】meets AOM figure guidelines (verify)
【Next step】amr-writing-style