An efficient method of constructing confidence intervals that correspond to randomization tests is described. The method is based on the Robbins-Monro search process and conducts a separate search for each endpoint of the confidence interval. Each search requires only slightly more permutations than the number needed for a randomization test. Diagnostics to monitor searches are suggested and the method is illustrated with examples.
No takes yet. Share an insight, caveat, or question.
Paul H. Garthwaite (1996) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: