Two-Proportion Z-Score

Problem

Compute the pooled two-proportion z-score: (p_a - p_b) / sqrt(p_pool * (1 - p_pool) * (1/n_a + 1/n_b)), where p_pool = (success_a + success_b) / (n_a + n_b).

Return 0.0 when a sample size is 0 or the standard error is 0.

Hint

The pooled proportion combines both groups under the null hypothesis of equal rates.

Test cases

case 1

in: [60,1000,50,1000]

expected: 0.980816477227499

case 2

in: [100,500,80,500]

expected: 1.6462159944159829

case 3

in: [50,100,50,100]

expected: 0

case 4

in: [0,100,0,100]

expected: 0

Starter codePython

import math
def two_proportion_z(success_a, n_a, success_b, n_b):
    # Your code here
    pass

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