Fixed vs Random Weights

Problem

Return [fixed_weights, random_weights] with fixed weights 1/se_i^2 and random weights 1/(se_i^2 + tau2) for the supplied between-study variance. Return [[], []] for empty input, a negative tau2, or any se <= 0.

Hint

Random-effects weights are always smaller than fixed-effect weights when tau2 > 0.

Test cases

case 1

in: [[0.1,0.2],0.01]

expected: [[99.99999999999999,24.999999999999996],[49.99999999999999,19.999999999999996]]

case 2

in: [[0.5],0]

expected: [[4],[4]]

case 3

in: [[],0.1]

expected: [[],[]]

case 4

in: [[0,0.1],0.1]

expected: [[],[]]

case 5

in: [[0.2,0.2,0.2],0.04]

expected: [[24.999999999999996,24.999999999999996,24.999999999999996],[12.499999999999998,12.499999999999998,12.499999999999998]]

Starter codePython

def fixed_vs_random_weights(sei, tau2):
    # Your code here
    pass

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