Statistical Power

Problem

Compute the power of a one-sided z-test: 1 - Phi(z_alpha - delta * sqrt(n) / sigma), where Phi is the standard normal CDF.

Return 0.0 when n <= 0 or sigma <= 0.

Hint

Power is the probability of rejecting the null when the true effect is delta.

case 1

in: [100,0.2,1,1.645]

expected: 0.6387052043836872

case 2

in: [400,0.1,1,1.96]

expected: 0.5159534368528308

case 3

in: [50,0.5,2,1.28]

expected: 0.6871425391678323

case 4

in: [0,0.2,1,1.645]

expected: 0

Starter codePython

import math
def statistical_power(n, delta, sigma, z_alpha):
    # Your code here
    pass

4 test cases · run your code against them in the editor