| name | prl-methods |
| description | Use when deciding what methodological detail belongs in a Physical Review Letters body versus Supplemental Material, so a physicist can trust the result without the Letter becoming a long paper. Partitions methods; does not design figures or run analysis. |
PRL Methods (prl-methods)
When to trigger
- Your Methods text is several paragraphs of apparatus, sample, or derivation detail
- You cannot tell what a reader needs versus what is reassuring completeness
- Reviewers of a prior draft asked for more rigor, and you are tempted to add bulk
- The Letter is over length and methods are a prime trimming target
- You are unsure which derivation steps to keep inline versus move to SM
The trust-minimum principle
In a Letter, methods exist to let a competent physicist believe the central claim — not to enable full replication inline. Full replication detail lives in Supplemental Material. Keep in the body only what is load-bearing for trust:
- The essential experimental configuration or theoretical setup (one compact description).
- The key control(s) that rule out the obvious alternative explanation.
- The decisive statistical/systematic statement (uncertainties, significance) for the headline number.
- The one or two equations that define the quantity being claimed.
Everything else — calibration procedures, sample growth, full Hamiltonians, lengthy derivations, parameter sweeps, additional checks — goes to SM and is cited inline ("see Supplemental Material [ref]").
Partition table
| Detail type | Body (trust-minimum) | Supplemental Material |
|---|
| Core setup / model definition | one compact paragraph | full apparatus / full Hamiltonian |
| Defining equations | 1–2 that define the claim | full derivation chain |
| Key control ruling out the obvious artifact | yes, stated explicitly | secondary controls |
| Uncertainty on the headline result | yes (stat + dominant syst) | full error budget |
| Sample / material preparation | one-line provenance | growth, characterization data |
| Numerical convergence / mesh | one-line assurance | convergence study |
| Reproducibility info (code/data location) | pointer | data-availability detail |
Rigor without bulk
PRL referees in physics are sharp about systematics and alternative explanations. Pre-empt the obvious objection in the body, briefly, and point to the SM for the exhaustive version. A single sentence ruling out the leading artifact is worth more than a paragraph of generic robustness language.
- State uncertainties as numbers, separating statistical and dominant systematic.
- Name the alternative explanation and the control that excludes it.
- For theory: state assumptions explicitly; relegate algebra to SM but keep the physical logic inline.
Worked micro-example: compressing an apparatus paragraph
Before (regular-article register):
"The cryostat was cooled over 48 hours. Thermometry used a calibrated sensor. Wiring consisted of filtered twisted pairs. Measurements used a lock-in amplifier after verifying frequency independence."
After (trust-minimum, one sentence + SM pointer):
"Measurements were performed in a dilution refrigerator with filtered wiring and calibrated thermometry; checks bounding self-heating below the quoted uncertainty are given in the Supplemental Material [ref]."
The rewrite keeps what a skeptical physicist needs — the environment, the leading artifact class, where the exhaustive checks live — and returns a paragraph of budget to the physics.
Sentence patterns that buy trust cheaply
- Control-in-a-clause: "The signal is absent at zero applied field, ruling out pickup as its origin."
- Split uncertainty: "We obtain X = value ± (stat) ± (syst), the systematic dominated by calibration drift."
- Assumption flag (theory): "Throughout we assume [condition]; relaxing it modifies only the prefactor (SM Sec. S2)."
- Convergence one-liner: "Results are converged in system size to within the plotted symbol size."
Each pattern replaces a defensive paragraph with one load-bearing sentence.
Checklist
Anti-patterns
- Treating the Letter like a full PR article and inlining the whole method
- Generic robustness boilerplate instead of the one decisive control
- Quoting a headline number with no uncertainty, or merged stat+syst
- Equations with undefined symbols (define on first use)
- "Details available on request" instead of a data-availability statement
Output format
【Trust-minimum kept inline】setup / key control / uncertainty / defining eqs
【Moved to SM】list
【Leading alternative explanation】named + control inline? yes / fix
【Headline uncertainty】stat + syst stated? yes / fix
【Data availability】pointer present? yes / fix
【Next】prl-figures (lead figure) or prl-supplementary (SM partition)
Data-availability and reproducibility expectations evolve — verify current APS policy on the official PRL author page.
