How confident can we be that rankings drawn from point estimates are accurate?
Because of margins of errors, we cannot make precise rankings. However, we can still make inferences based on confidence intervals. If we for instance divide the distribution of countries’ estimates into two categories (low-high governance rating), we can calculate the probability that any given country might indeed belong to the opposite side of the distribution. A more detailed discussion of confidence intervals, standard errors and rankings can be found in the Governance Matters III paper (pages 11-14, see also figure 1 on pages 45-47) and the Governance Matters VII paper.
Because of margins of errors, we cannot make precise rankings. However, we can still make inferences based on confidence intervals. If we for instance divide the distribution of countries’ estimates into two categories (low-high governance rating), we can calculate the probability that any given country might indeed belong to the opposite side of the distribution. A more detailed discussion of confidence intervals, standard errors and rankings can be found in the Governance Matters III paper (pages 11-14, see also figure 1 on pages 45-47) and the Governance Matters VIII paper.
