Bootstrap F1 CI Sample Size Guidance Method

A Bootstrap F1 CI Sample Size Guidance Method is a bootstrap parameter selection method that provides evidence-based recommendations for resample count, sample size thresholds, and method variants when constructing bootstrap F1 confidence intervals.

• AKA: Bootstrap B Selection Method, Resample Size Guidance Method, Bootstrap F1 Parameter Guide, Small-n Bootstrap Warning Method.
• Context:
  • It can typically recommend B ≥ 2000 bootstrap resamples for stable percentiles.
  • It can typically warn about degenerate intervals (e.g., [0, 100%]) for n < 12.
  • It can typically prefer BCa over basic percentile ("should not be used generally").
  • It can often acknowledge bootstrap correctly reflects high uncertainty with wide intervals.
  • It can often suggest sufficient B to ensure percentile stability (B = 5000 for publication).
  • It can often document trade-off between computation time and interval accuracy.
  • It can range from being a Minimal Bootstrap F1 CI Sample Size Guidance Method to being a Comprehensive Bootstrap F1 CI Sample Size Guidance Method, depending on its recommendation detail.
  • It can range from being a Conservative Bootstrap F1 CI Sample Size Guidance Method to being a Efficient Bootstrap F1 CI Sample Size Guidance Method, depending on its resource priority.
  • It can range from being a Fixed-B Bootstrap F1 CI Sample Size Guidance Method to being a Adaptive-B Bootstrap F1 CI Sample Size Guidance Method, depending on its resample strategy.
  • It can range from being a Warning-Only Bootstrap F1 CI Sample Size Guidance Method to being a Alternative-Suggesting Bootstrap F1 CI Sample Size Guidance Method, depending on its response type.
  • ...
• Example(s):
  • Resample Count Recommendations, such as:
    • Quick exploration: B = 1000 (adequate for rough estimate).
    • Standard analysis: B = 2000 (recommended minimum).
    • Publication/production: B = 5000-10000 (ensure stability).
  • Small Sample Warnings, such as:
    • n=12 example: Precision CI was [0%, 100%] reflecting true uncertainty.
    • n < 10: Bootstrap may produce uninformative full-range intervals.
    • n < 20: Consider Wilson+CC as analytical alternative.
  • Method Variant Selections, such as:
    • Basic percentile: Not recommended without adjustment.
    • BCa: Preferred for bias and skewness correction.
    • ABC: Similar coverage to BCa but 10x faster computation.
  • ...
• Counter-Example(s):
  • Fixed Bootstrap Protocol, using same B regardless of context.
  • No-Warning Bootstrap Method, applying bootstrap blindly.
  • Percentile-Only Method, ignoring BCa improvements.
• See: Bootstrap Method, BCa Bootstrap F1 Confidence Interval Method, Sample Size Planning, Resample Size Selection, Degenerate Interval, Computational Trade-off, Small Sample Problem, Bootstrap Stability, F1 Interval Selection Guide Method, Monte Carlo Convergence.

---