Simpson's Paradox in Risk Tables

Problem

Given risk counts for two strata (a_j cases and b_j non-cases in the exposed group, c_j and d_j in the unexposed group), return True when both stratum risk differences have one sign and the pooled risk difference has the opposite sign. Return False otherwise.

Hint

Pooled counts can reverse a trend that holds in every subgroup.

Test cases

case 1

in: [81,6,234,36,192,71,55,25]

expected: true

case 2

in: [30,70,10,90,30,70,10,90]

expected: false

case 3

in: [234,36,81,6,55,25,192,71]

expected: true

case 4

in: [30,70,10,90,10,90,30,70]

expected: false

case 5

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

expected: false

Starter codePython

def simpsons_paradox(a1, b1, c1, d1, a2, b2, c2, d2):
    # Your code here
    pass

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