SMR Writing Style

Use this to make the prose match a methods journal. SMR readers are methodologists who want the contribution stated early, the assumptions visible, and the math legible — and they want ASA style and the abstract rules honored exactly, because conformance signals seriousness.

The methods-paper introduction arc

Open in this order (model it on resources/worked-examples/01-introduction.md):

1. The methodological problem in an applied setting — what goes wrong when researchers use the current method, stated concretely, not "X has received much attention."
2. Why existing methods fail — the gap, named, not a literature tour.
3. The contribution — develop / evaluate / correct, in one proportional sentence (from smr-method-contribution).
4. The properties and evidence — what you prove, what the simulation shows, what the illustration demonstrates, each labeled by evidentiary status.
5. The applied payoff — what researchers can now do, and that usable software is released.

A compact roadmap closes the introduction; do not over-signpost.

The abstract (a hard, checkable rule)

• ≤150 words and no parenthetical citations (检索于 2026-06；以官网为准). SAGE allows non-parenthetical citations ("Smith (2015) showed…") but not "(Smith 2015)". Rewrite any parenthetical cite or drop it.
• The abstract must name the method, the problem it solves, and the evidence type — not just the topic. A methods abstract that reads like a substantive abstract signals scope confusion.

Exposition discipline for math-heavy prose

• State the estimand before the estimator; state assumptions near the result they support.
• Short interpretation over algebra restatement: after a theorem, say what it means for a user, do not re-narrate the derivation.
• Label evidentiary status in the prose: "we prove," "the simulation shows," "we conjecture." A methods reader tracks these precisely.
• Define notation once, use it consistently; a methods paper cannot afford symbol drift.

ASA style and house conventions

• In-text citations and references in ASA style; dataset references in DataCite style (检索于 2026-06；以官网为准). Use a reference manager configured for ASA rather than hand-formatting.
• Keep the title page separate from the main document for double-anonymized review; remove self-identifying phrasing ("in our previous paper") from the body.
• Disclose generative-AI tool use in the back matter when applicable; if none was used, no acknowledgement is needed.

Before/after micro-rewrites in the SMR register

• Before: "Latent class analysis has received considerable attention in recent years." → After: "When classes are weakly separated, standard LCA enumeration selects the wrong number of classes in a majority of realistic samples." (opens on the failure, not the fashion)
• Before: "Equation (7) follows from substituting (5) into (6) and rearranging." → After: "Equation (7) says the bias grows with the imbalance ratio: applied users with rare exposure groups are hit hardest." (interprets, does not re-derive)
• Before: "Our results suggest the estimator performs well." → After: "We prove consistency under A1–A3; the simulation shows the nominal 95% interval covers between 93.1% and 95.4% across designs." (labels evidentiary status and quantifies)

Checklist

• The introduction follows problem → why methods fail → contribution → evidence → payoff.
• The contribution sentence is proportional and appears early.
• The abstract is ≤150 words with zero parenthetical citations.
• The abstract names method, problem, and evidence type.
• Estimand precedes estimator; assumptions sit near their results.
• Evidentiary status (proved/simulated/conjectured) is labeled in prose.
• ASA citation style and DataCite dataset references are used.
• No self-identifying language; AI-use disclosed if applicable.

Anti-patterns

• Substantive-paper opening: "much attention has been paid to…" instead of the methodological gap.
• Parenthetical citations in the abstract: a direct violation of the stated rule.
• Algebra narration: restating the derivation instead of interpreting the result.
• Status blur: simulation regularities written in the language of proof.
• Hand-formatted citations: ASA inconsistencies that signal carelessness at a methods venue.
• Deanonymizing phrasing left in the body under double-anonymized review.

Output format

[Prose status] ready / needs revision / not ready
[Intro arc] complete / missing step (which)
[Abstract] words=<n>, parenthetical cites=<n>, names method+problem+evidence? yes/no
[Style] ASA + DataCite applied? yes/no
[Anonymization] clean / risk where
[Next SMR skill] smr-software-and-reproducibility