1[2 {"id":"C1","text":"11 (procedure, cell) pairs have coverage below 0.90: CR0 at all G<=10 cells, CR1 at G=5 with ICC 0.2 and 0.5, WCB at all G=5 cells; none at G=30.","results_keys":["failure_cells","n_failure_cells","coverage"],"status":"supported","caveat":"cr1_G5 (icc0.2/0.5) and cr0_G10_icc0.0 lie only ~2.3-2.8 MC SE below 0.90."},3 {"id":"C2","text":"Only 6 of 27 pairs are consistent with nominal 95% coverage (CR1 at G10/ICC0, G30/ICC0, G30/ICC0.2; WCB at all G=30 cells).","results_keys":["nominal_cells","coverage","mc_se_coverage"],"status":"supported","caveat":"2-SE rule applied to 27 pairs without multiplicity correction."},4 {"id":"C3","text":"With G=5, CR1 has the highest average coverage (0.893) ahead of WCB (0.857) and CR0 (0.746).","results_keys":["best_procedure_small_G","mean_coverage_small_G"],"status":"supported","caveat":"Specific to unrestricted Rademacher WCB with symmetric bootstrap-t and B=399."},5 {"id":"C4","text":"At G=10 with ICC>0 and at G=30, WCB is closer to nominal than CR1.","results_keys":["coverage","mc_se_coverage"],"status":"supported","caveat":"Differences at G=30 are of the order of 1-3 MC SE; paired SEs not computed."},6 {"id":"C5","text":"CR0 and CR1 variance estimators are biased downward in every cell, strongly at G=5 (CR0 0.51-0.65, CR1 0.63-0.82), and Monte Carlo bias estimates agree with analytic expectations.","results_keys":["rel_bias_var","mc_se_rel_bias_var","exact_rel_bias_var","z_rel_bias_mc_vs_exact"],"status":"supported","caveat":"Bias depends on the unbalanced cluster-size pattern."},7 {"id":"C6","text":"At G=5 WCB intervals are wider on average than CR1 intervals yet cover less.","results_keys":["mean_ci_halfwidth","coverage"],"status":"supported","caveat":"Explanation (variable, discrete bootstrap critical value) is interpretive, not separately tested."},8 {"id":"C7","text":"The naive i.i.d. interval covers ~0.95 at ICC=0 but only 0.30-0.48 when ICC>0.","results_keys":["baseline_coverage_iid"],"status":"supported","caveat":"Baseline not required by the task; included for context."}9]10