Conceptual & Analytical Framework (worlddev-theory-model)

When to trigger

• The paper has a "theory" or "conceptual framework" section that is decorative — cited and then never used
• Reviewers say the empirics are fine but the paper "doesn't add up to an argument"
• A formal model is included by reflex, when a clear conceptual framework would serve WD's readers better
• The mechanism connecting cause and outcome is asserted, not theorized
• A qualitative paper's themes are described but not organized by any analytical lens

What "theory" means at WD

WD is not a journal that demands a formal model. For most WD papers, "theory" means a conceptual or analytical framework drawn from development scholarship — a structured account of the mechanism that tells the reader what to look for, why the variables matter, and how the pieces connect. A formal model is welcome when it earns its place (it disciplines an argument that prose cannot), but a model bolted on for prestige is a liability: WD's multidisciplinary referees punish "theory theater" — equations that do no work the empirics depend on.

The test of a good WD framework is load-bearing: remove it and the paper's interpretation collapses. It should generate the hypotheses (quant) or sensitizing concepts (qual), define the scope conditions, and frame what the findings mean for development.

Choosing the right framework vehicle

Pick one primary vehicle. Stacking a formal model, a theory-of-change diagram, and three analytical lenses signals that none is doing the work.

Making the framework load-bearing — a sequence

1. State the mechanism in one sentence before any formalism: "X improves Y because it shifts the bargaining power of Z."
2. Derive what we should observe if the mechanism holds — and crucially what we should not observe (a discriminating prediction).
3. Map each prediction to a specific piece of evidence later in the paper (a coefficient, a case contrast, an interview pattern).
4. State scope conditions — where the mechanism should and should not operate. WD readers care intensely about context-dependence; a framework that claims universality is suspect.
5. Return to the framework in the discussion — show what the evidence taught you about the theory, not just about the world.

A worked example (illustrative)

A paper argues microcredit raises women's autonomy. A decorative version cites empowerment theory and runs regressions. A load-bearing WD version specifies the mechanism (credit relaxes a liquidity constraint that previously forced dependence on a spouse), derives a discriminating prediction (effects concentrate where women lacked independent income, not everywhere), maps it to a heterogeneity test, and states the scope condition (no effect where social norms bar women from controlling the loan). Now the theory predicts the pattern of effects, not merely their existence.

Checklist

• One primary framework vehicle chosen, fit to the contribution
• The mechanism is stated in plain language before any formalism
• The framework generates a discriminating prediction (or sensitizing concepts) used in the empirics
• Each prediction maps to a specific exhibit, case, or pattern
• Scope conditions / context-dependence are explicit
• The discussion revisits the framework in light of the evidence
• No formal model included purely for prestige; no lens merely name-checked

Anti-patterns

• "Theory theater": equations or a diagram that the empirical strategy never uses
• Citing a development-theory tradition (capabilities, political settlements) and then ignoring it in analysis
• Asserting the mechanism instead of theorizing it, leaving the reader to infer the logic
• Claiming a universal mechanism with no scope conditions — WD distrusts context-free claims
• A qualitative paper whose themes float free of any analytical organizing principle

Output format

【Journal】World Development (WD)
【Skill】worlddev-theory-model
【Vehicle】formal model / conceptual framework / analytical lens / grounded
【Mechanism】one plain-language sentence
【Discriminating prediction】what we should and should NOT observe
【Prediction→evidence map】which exhibit/case tests each
【Scope conditions】where the mechanism does / does not operate
【Source status】verified URL / 待核实 / not asserted
【Next skill】worlddev-robustness