The PRB of each component is estimated using the code below. Notice that PRBs are relatively small, with absolute values below 1%, with the largest bias in the GEI index component.
# repeat true population values
(theta.pop)
# estimate the expected values of the components estimators
# using the average of the estimates
(theta.exp <- rowMeans(sapply(estimate.list , coef)))
# estimate the percentage relative bias
(percentage_relative_bias <- 100 * (theta.exp / theta.pop - 1))
# stopifnot( percentage_relative_bias_variance < 1 )
ajdamico
commented
1 year ago
The PRB of each component is estimated using the code below. Notice that PRBs are relatively small, with absolute values below 1%, with the largest bias in the GEI index component.