Nature Reporting Summary Auto-Fill

The Nature Research Reporting Summary is a mandatory checklist for all papers submitted to Nature-family journals. It asks dozens of questions about sample sizes, statistical methods, replication, randomization, and blinding. Every answer should be derivable from what was done — but most researchers fill it out from memory at submission. Licklider fills it from the record.

STEP 1 —The Pitfall: Checklists Filled from Memory

The Nature Research Reporting Summary is not a formality —it is an accountability document. It asks researchers to confirm that sample sizes are justified, that replicates are defined, that statistical tests are appropriate, and that randomization and blinding were performed. These confirmations are then reviewed by Nature's statistics editors.

The problem is that the checklist is typically filled out at the end of the submission process, weeks or months after the analysis. Researchers answer from memory and from the Methods section — which, as established earlier, may not accurately reflect what was done. The result:

• Inconsistencies between checklist and Methods. The checklist says "n determined by power analysis"; the Methods says "n = 6 per group" with no justification. Statistics editors flag the discrepancy.
• "Yes" answers that should be "No." "Were samples randomized?" is answered Yes by default, when the actual allocation was not randomized.
• Missed items. Researchers unfamiliar with the checklist miss questions about effect size reporting, multiple comparison correction, and independent replication.

STEP 2 —Journal Requirement

The Nature Research Reporting Summary is mandatory for all papers submitted to Nature, Nature Medicine, Nature Methods, Nature Genetics, and all other Nature-family journals. It covers the following categories:

• Study design (sample sizes, randomization, blinding)
• Statistical reporting (tests used, correction methods, effect sizes, CIs)
• Replication (independent experimental replicates)
• Reporting (data availability, code availability)
• Specific method categories (cell biology, animal experiments, human subjects)

Nature's statistics editors routinely return papers at the revision stage with requests to correct checklist entries that are inconsistent with the reported Methods. A pre-filled, internally consistent checklist is a material advantage at submission.

STEP 3 —Licklider's Solution

Input

• The complete Data Contract (observation unit, variable mapping, outcome type, analysis intent).
• The Statistical Audit record (Statistical Validity Score dimensions, power analysis results, multiple comparison decisions).
• The Transparency Trail audit log (preprocessing steps, exclusion criteria, N disclosure).
• The Reproducible Output records (Methods text, version snapshot).

Output

• Pre-filled Reporting Summary. A completed Nature Research Reporting Summary with all auto-fillable questions answered from the session record. Each answer is tagged with the source record it was derived from (e.g., "[Source: Statistical Audit 4.1.4 —power analysis]").
• Flagged items. Questions that cannot be auto-filled (e.g., "Were samples randomized?" —requires the researcher's input) are highlighted with the specific information needed to complete them.
• Consistency check. The filled checklist is cross-referenced with the auto-generated Methods text to verify that answers are consistent. Discrepancies are flagged before export.
• Downloadable Word/PDF. The completed Reporting Summary is exported in Nature's required format, ready for direct upload to the submission system.

Guard

STEP 4 —Draft Output (Draft / Needs review)

Example: Auto-filled checklist entries

NATURE REPORTING SUMMARY —AUTO-FILLED BY LICKLIDER (v2.4.1)

Statistical analysis
Q: Confirm that the following items are present in your manuscript:
  [OK]n for each experimental group as a number and biological replicate type
     [Auto-filled: n = 12 per group (biological replicates, individual mice).
      Source: Data Contract observation unit declaration + N Disclosure panel]

  [OK]Description of statistical methods with sufficient detail to reproduce
     [Auto-filled: Methods paragraph auto-generated by Licklider. Methods text ID: meth-4a7b]

  [OK]Justification for the statistical test(s) chosen
     [Auto-filled: Welch t-test applied to continuous, approximately normal data (Shapiro-Wilk
      W > 0.90 for both groups). Source: Statistical Audit normality check]

  [!]  Whether any adjustments were made for multiple comparisons
     [Requires input: one primary comparison was performed. If this is the only analysis,
      no correction is needed. If part of a larger analysis family, declare here.
      Source: Analysis Family Ledger —family size = 1 confirmatory comparison]

  [OK]Effect sizes and confidence intervals for the primary result
     [Auto-filled: Cohen's d = 0.91, 95% CI [0.35, 1.47]. Source: Statistical Audit 4.1.2]

Related Features