Skip to content

Hypothesis Testing Lab

The black curve is the sampling distribution of t if H₀ is true. Drag the critical value: the red tail is α. The violet curve is where t would live if the true effect were δ; the dashed region is β, and power = 1 − β. Drag the observed t to read p as a tail area.

H₀ and H₁: α = .050, β = 0.950, power = 0.050, p = 1.000

H₁ is drawn as a t curve shifted by δ√n (a visual approximation); α, β and power in the readouts are computed from the noncentral t distribution (Lenth AS 243, checked against SciPy nct). The null simulation uses seeded normal samples with n − 1 df.

p = 1.000, power = 0.050, critical t = 2.09

Formula

α = P(|T| > t_crit | H₀) β = P(|T′| ≤ t_crit | δ) power = 1 − β, T′ ~ t(n−1, ncp = δ√n)

Moving the critical value left raises power and α together — you cannot buy one without paying the other at fixed n.

Method: noncentral t (Lenth AS 243), both tails for two-sided tests · p: Student t(19) · Synthetic, seeded data. Illustrative, not an empirical finding.

Validation: Numerically validated against reference fixtures · numerically validated