{
  "slug": "statistical-power",
  "term": "Statistical power",
  "definition": "Statistical power is the probability that a specified test rejects its null hypothesis when a specified alternative is true.",
  "note": "Power equals 1 minus the Type II error probability for a specified alternative. It depends on the test, significance level, effect size, sample size and sampling assumptions; it is not a probability that a hypothesis is true. Plan against a practically important alternative, not an effect selected after seeing the results. [hoenig-heisey]\n\nA fixed-horizon binomial design counts each independent, equal-risk device once as failed or not failed after complete follow-up. With a known 1% probability benchmark and nominal 5% upper-tail test, the synthetic example detects a true 3% probability with 35.3% power for 100 devices and 79.7% for 300. Discrete thresholds keep actual null rejection below 5%. These are horizon-specific probabilities, not AFR values. [nist-binomial-definition] [nist-oc] [hesela-power]\n\nNon-rejection does not establish equivalence. Equivalence requires prespecified practical bounds and a suitable test; observed-effect post-hoc power does not resolve an inconclusive result. Shared failure causes, unequal device risks or incomplete follow-up require a design that models those features. [lakens-equivalence] [hoenig-heisey] [nist-binomial-definition]\n\nExact-test power need not rise at every larger sample size. Raising the integer rejection threshold can lower it. In the same 1% versus 3% example, n=301 first reaches 80% power, n=330..339 falls below, and every prespecified n>=340 reaches it. This fixed-design guarantee does not permit outcome-dependent stopping. [stata-exact-power] [hesela-power-floor]",
  "related": [
    "binomial-proportion",
    "confidence-interval",
    "prediction-horizon"
  ],
  "citations": [
    {
      "id": "nist-oc",
      "title": "NIST e-Handbook 6.2.3.2: Binomial sampling plans and operating-characteristic probabilities",
      "url": "https://www.itl.nist.gov/div898/handbook/pmc/section2/pmc232.htm",
      "accessed": "2026-11-19"
    },
    {
      "id": "nist-binomial-definition",
      "title": "NIST: Binomial distribution, independent Bernoulli experiments with common probability",
      "url": "https://csrc.nist.gov/glossary/term/Binomial_Distribution",
      "accessed": "2026-11-19"
    },
    {
      "id": "hoenig-heisey",
      "title": "Hoenig and Heisey (2001): The Abuse of Power, The American Statistician 55(1), 19-24",
      "url": "https://sciences.ucf.edu/biology/d4lab/wp-content/uploads/sites/24/2016/11/HoenigHeisey2001.pdf",
      "accessed": "2026-11-19"
    },
    {
      "id": "lakens-equivalence",
      "title": "Lakens (2017): Equivalence Tests: A Practical Primer, DOI 10.1177/1948550617697177",
      "url": "https://pure.tue.nl/ws/portalfiles/portal/80918653/lakeequi2017.pdf",
      "accessed": "2026-11-19"
    },
    {
      "id": "hesela-power",
      "title": "Hesela: Fixed-horizon statistical-power example, exact arithmetic and assumptions",
      "url": "https://hesela.com/analyses/afr-rate-not-probability/#statistical-power",
      "accessed": "2026-11-19"
    },
    {
      "id": "stata-exact-power",
      "title": "Stata: power oneproportion, Example 7, discrete thresholds and nonmonotonic power",
      "url": "https://www.stata.com/manuals/pss-2poweroneproportion.pdf",
      "accessed": "2026-11-19"
    },
    {
      "id": "hesela-power-floor",
      "title": "Hesela: Exact fixed-design power floor, finite enumeration and rational tail certificate",
      "url": "https://hesela.dev/research/statistical-power/power-floor.json",
      "accessed": "2026-11-19"
    }
  ],
  "aliases": [
    "Test power",
    "Power analysis"
  ],
  "status": "shipping",
  "advantages": [],
  "limitations": [],
  "provenance": [
    "The 100- and 300-device examples are synthetic exact-binomial calculations, not observations of a storage fleet. [hesela-power]",
    "The equivalence source develops general inference principles and t-test examples; it is not a validated storage-specific equivalence protocol. [lakens-equivalence]",
    "For this synthetic design, exact enumeration covers n=1..849 and a rational exponential-Markov bound covers every n>=850. The proof cutoff is not the required sample size. [hesela-power-floor]"
  ],
  "supersedes": [],
  "supersededBy": [],
  "firstSeen": null,
  "lastSeen": null,
  "updated": "2026-11-19"
}
