Module statistics 

Statistical Analysis Module (PMAT-024)

Implements statistical analysis per F221 for 95% nonparametric confidence intervals, effect size calculation, and bootstrap sampling.

Components

Component	Formula	Use Case
Bootstrap CI	Resampling with replacement	Nonparametric 95% CI
Cohen’s d	(M1-M2) / pooled_std	Effect size magnitude
Welch’s t-test	t-statistic with unequal variances	A/B comparison
Mann-Whitney U	Nonparametric rank test	Non-normal distributions
IQR Outlier Filter	Q1 - 1.5IQR to Q3 + 1.5IQR	Robust statistics

Citations

• [Efron & Tibshirani 1993] “An Introduction to the Bootstrap”
• [Cohen 1988] “Statistical Power Analysis for Behavioral Sciences”
• [Hoefler & Belli 2015] “Scientific Benchmarking of Parallel Computing Systems”

Structs

ComparisonResult

Result of statistical comparison between two samples

EffectSize

Effect size calculation result

MannWhitneyResult

Mann-Whitney U test result (nonparametric)

OutlierFilter

IQR-based outlier filter

StatisticalAnalysis

Statistical analysis result with confidence interval

Enums

EffectCategory

Effect size category per Cohen’s conventions

Functions

bootstrap_ci

Bootstrap confidence interval calculation

percentile

Calculate percentile (0.0 to 1.0)

trimmed_mean

Calculate robust mean (trimmed mean, removing top/bottom 10%)