prior-study-replication

name: prior-study-replication description: "Companion to analysis-report. Provides report.md structure for analyses that replicate or follow a prior study's methodology. Enforces: prior study methods/results/conclusions → methodological adaptation rationale → user data results → systematic comparison with prior findings."

Prior Study Replication Report

Trigger

• Analysis that replicates, follows, or extends a specific published study's methodology
• Requests like "follow the paper's method", "replicate study X", "apply method from Y to our data"
• Analysis where results should be compared against published findings

Non-Trigger

• Exploratory analyses with no specific prior study to follow
• Descriptive statistics only (no methodological replication involved)
• Novel analyses where the user defines the method from scratch

Relationship to Other Skills

• Always use together with analysis-report — that skill handles folder structure, code rules, environment notes, color conventions, script boilerplate
• This skill only defines the report.md structure
• Use analysis-report/templates/script_boilerplate.py for the Python script
• Use this skill's templates/report_template.md for the report

Report Structure (mandatory sections, strict order)

Section-Specific Rules

Prior Studies

• One subsection per referenced paper: ### Author et al. (Year) — short title
• Each subsection MUST include all four items:
  • Methodology: Model, variables, analysis approach
  • Experimental design: N subjects, equipment, protocol
  • Key results: Specific numbers (e.g., "r ≈ 0.54", "AUC = 0.71", "predicted 11% of stepping")
  • Conclusions: Authors' main takeaway in 1-2 sentences
• Include reference paper PDFs in the analysis folder when available

Methodological Adaptation

• MUST contain a comparison table with columns: Prior Method | Current Implementation | Deviation Rationale
• Every row where the method was NOT replicated 1:1 MUST have a rationale (e.g., equipment constraints, missing parameters, coordinate system incompatibility)
• End with a summary sentence: "This analysis adopts [aspect X] from [study] but modifies [aspect Y] because [reason]."

Comparison with Prior Studies

• MUST contain a comparison table with columns: Comparison Item | Prior Study Result | Current Result | Verdict
• Verdict values: Consistent / Partially consistent / Inconsistent / Not tested
• Any Inconsistent or Not tested item MUST include a brief explanation

Results Reporting Convention (default)

• User-facing Results tables should hide numeric p values and df unless the user explicitly requests them.
• Report significance with Sig only: *, **, ***, n.s. at alpha=0.05.
• If multiple comparison correction is used, compute Sig from corrected p-values (e.g., BH-FDR) and name the method in Analysis Methodology.
• For metrics named *_peak, define the term explicitly as absolute peak: abs_peak = max(|x|) within the analysis window.

Validation

• report.md contains all 11 mandatory sections in order
• Prior Studies: ≥1 study with all four items (methodology, design, results, conclusions) with specific numbers
• Methodological Adaptation: comparison table present; all deviation rows have rationale
• Comparison with Prior Studies: comparison table present with verdicts; inconsistencies explained
• All numbers in Comparison table match Results section and Prior Studies section
• Results tables follow Sig-only convention by default (no numeric p/df columns unless requested)

Completion Checklist

• report.md follows this skill's template (not analysis-report template)
• Prior Studies section has concrete numbers from each referenced paper
• Methodological Adaptation table explains all deviations
• Comparison table has verdicts for all items
• Reference PDFs included in analysis folder (if available)