Inferential Statisticshard
0:00.0

Given a sample of size nn from a distribution f(xθ)f(x|\theta), the Score function U(θ)U(\theta) and the Fisher Information I(θ)I(\theta) are used to test H0:θ=θ0H_0: \theta = \theta_0. If we define the Score test statistic as S=U(θ0)2/I(θ0)S = U(\theta_0)^2 / I(\theta_0), what is the fundamental advantage of this test over the Wald test in terms of the likelihood function?