Kvíz spojený s průvodcem · Střední úroveň

Vícenásobná srovnání opravný kvíz

Procvičte si, proč opakované testování zvyšuje falešně pozitivní výsledky, a odlište rodinnou kontrolu chyb od postupů s falešnou mírou odhalení.

Související vodicí cestyOprava více srovnání
Otázka 1 z 8

With 20 independent true nulls tested at 0.05, what is the approximate chance of at least one false positive?

Náhled každé otázky8 otázky

Odpovědi a vysvětlení se zobrazují, jak budete v kvízu.

  1. With 20 independent true nulls tested at 0.05, what is the approximate chance of at least one false positive?
    • 0,64
    • 0.05
    • 0,20
    • 0.95
  2. For 10 tests and family-wise alpha 0.05, what per-test cutoff does Bonferroni use?
    • 0.05
    • 0.005
    • 0,5
    • 0.0005
  3. Which error criterion does Bonferroni target?
    • Average prediction error
    • Expected fraction of false discoveries only
    • Family-wise error rate
    • False omission rate
  4. What does FDR concern among rejected hypotheses?
    • The variance of all test statistics
    • The chance every rejection is false
    • The probability of any single miss
    • The expected proportion that are false
  5. Why might analysts prefer an FDR procedure for a large discovery search?
    • It guarantees each finding is true
    • It can offer more power while controlling a discovery proportion criterion
    • It avoids defining the test family
    • It removes the need for replication
  6. Why define the hypothesis family before examining results?
    • It guarantees independence
    • It makes all p-values equal
    • The family determines which comparisons need joint interpretation
    • It ensures no data are missing
  7. What does a corrected p-value not fix?
    • The chosen alpha threshold
    • Multiplicity itself
    • The stated family size
    • Confounding or invalid measurement
  8. After an adjusted exploratory discovery, what is a useful next step?
    • Seek independent validation or replication
    • Declare a confirmed mechanism immediately
    • Remove all non-significant outcomes from the report
    • Retest until the p-value falls further