brycewang-stanford Contribution Framing (revacc-contribution-framing)
When to trigger
- The results are in but you cannot state the contribution in one sentence
- The intro lists what you did rather than what the field learns
- A referee says "interesting, but what is the takeaway for accounting?"
- An analytical paper has proven propositions but no crisp accounting implication
- The contribution is overclaimed relative to what the design or model actually shows
State the contribution as what accounting learns
A RAST contribution is not "we run X regression" or "we prove Proposition 1." It is a sentence of the form: "We show that [accounting construct/mechanism] [does what] through [friction/channel], which means [implication for reporting, standard-setting, investors, auditors, or theory]." RAST's fast first-round-decision culture means the editor reads the contribution claim hard on the first pass; an underframed or overclaimed contribution is a leading reject reason.
Calibrate the claim to the evidence:
- Empirical archival. The contribution is a credibly identified accounting effect plus the channel evidence. Do not promise causality your design cannot deliver; if the cleanest claim is an association under stated assumptions, say so and make the association informative.
- Analytical. The contribution is the insight the model delivers — an optimal-disclosure result, a rationalized puzzle, a comparative static with a reporting reading — not the algebra. Translate the proposition into the accounting decision it informs.
- Experimental. The contribution is the channel the manipulation isolates that archival data cannot, and what it implies for how the accounting information is processed or produced.
The four contribution registers RAST rewards
| Register | What it offers | Watch out for |
|---|---|---|
| Empirical regularity | A new, credibly identified fact about reporting/markets | Over-claiming causality; thin channel |
| Theory ↔ evidence bridge | A model that explains data, or a test of a disclosure-theory prediction | Loose fit between the model and the test |
| Construct / measurement | A better measure of an accounting construct, validated | A measure with no question behind it |
| Mechanism / channel | Why an established effect occurs | A channel the design cannot distinguish |
The theory↔evidence bridge is RAST's sweet spot — lean on it when the paper supports it.
Sharpen and bound the claim
- Write the one-sentence contribution; cut every clause not supported by a table or a proof.
- State the scope condition — the setting, sample, or assumptions under which it holds.
- Name what the paper does not show, pre-empting the obvious referee objection.
- Connect the contribution to a decision: what should a standard-setter, investor, auditor, or theorist do differently?
Checklist
- The contribution is one sentence: construct → effect → channel → implication
- The claim is calibrated to the evidence (no causality beyond the design; no insight beyond the proof)
- The scope condition is stated explicitly
- What the paper does not show is named
- The contribution names a decision or debate it changes
- The theory↔evidence bridge is used where the paper supports it
Anti-patterns
- Activity for contribution: listing what you did instead of what is learned.
- Causality inflation: claiming a causal effect a DiD/IV cannot support.
- Insight inflation: selling algebra as if the equilibrium itself were the contribution.
- Unbounded claim: no scope condition, so a referee finds the boundary for you.
- Orphan result: a finding connected to no decision, debate, or theory.
Output format
【One-sentence contribution】[construct] → [effect] via [channel] ⇒ [implication]
【Register】regularity / theory-evidence bridge / measurement / mechanism
【Calibration】claim ≤ evidence? (causal / associational / model-insight)
【Scope condition】where it holds
【Does NOT show】the honest limit
【Decision changed】standard-setter / investor / auditor / theorist
【Next skill】revacc-tables-figures