Stratified Variance

Problem

Return the stratified sampling variance sum(w_h^2 * var_h / n_h) for the given per-stratum variances, sample sizes, and weights. Strata with n_h <= 0 are skipped and empty input returns 0.0.

Hint

Each stratum contributes its weighted squared share.

Test cases

case 1

in: [[4,9],[10,20],[0.5,0.5]]

expected: 0.21250000000000002

case 2

in: [[1],[10],[1]]

expected: 0.1

case 3

in: [[4,9],[10,0],[0.5,0.5]]

expected: 0.1

case 4

in: [[],[],[]]

expected: 0

case 5

in: [[1,1],[5,5],[0.2,0.8]]

expected: 0.13600000000000004

Starter codePython

def stratified_variance(variances, sizes, weights):
    # Your code here
    pass

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