Clustered MDE Inflation

Problem

Return the clustered minimum detectable effect, mde_individual * sqrt(1 + (cluster_size - 1)*icc). Clustering inflates the MDE by the square root of the design effect. Return 0.0 for non-positive cluster size or negative icc.

Hint

Multiply the individual MDE by the square root of the design effect.

Test cases

case 1

in: [0.2,10,0.05]

expected: 0.24083189157584595

case 2

in: [0.2,1,0.5]

expected: 0.2

case 3

in: [0.5,4,0.1]

expected: 0.570087712549569

case 4

in: [0.3,0,0.1]

expected: 0

case 5

in: [0.3,5,-0.1]

expected: 0

Starter codePython

def clustered_mde(mde_individual, cluster_size, icc):
    # Your code here
    pass

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