1[2 {"id":"C1","text":"11 (procedure, cell) pairs have coverage below 0.90: CR0 at all G=5 and G=10 cells, WCB at all G=5 cells, CR1 at G=5 with ICC 0.2 and 0.5; none at G=30.","results_keys":["failure_cells","n_failure_cells","coverage"],"status":"supported","caveat":"cr1_G5_icc0.0 and cr0_G10_icc0.0 lie within 2 MC SE of 0.90, so their classification is fragile."},3 {"id":"C2","text":"CR1 has the highest average coverage over the G=5 cells (0.891 vs 0.853 WCB and 0.740 CR0); the CR1-WCB gap is many paired MC SEs.","results_keys":["best_procedure_small_G","avg_coverage_G5","avg_coverage_G5_paired_diff"],"status":"supported","caveat":"Specific to Rademacher unrestricted WCB with B=399 and Gaussian DGP."},4 {"id":"C3","text":"Only 5 pairs are consistent with nominal 95% coverage (CR1 at G=10 and G=30 with ICC=0; WCB at all G=30 cells); CR0 never is.","results_keys":["nominal_cells","coverage","mc_se_coverage"],"status":"supported","caveat":"2-SE screen without multiplicity control."},5 {"id":"C4","text":"CR0 and CR1 variance estimators are biased downward, most at small G and high ICC (CR0 ratio 0.48-0.66 at G=5, 0.93-0.94 at G=30).","results_keys":["rel_bias_var","mc_se_rel_bias_var"],"status":"supported","caveat":"MC estimate; agrees with closed-form expectation within 2 SE in 8 of 9 cells (2.46 SE in G5_icc0.2)."},6 {"id":"C5","text":"The naive i.i.d. baseline interval covers only 0.29-0.49 when ICC>0, so all cluster-robust procedures improve on it in clustered cells.","results_keys":["baseline_iid_coverage","coverage"],"status":"supported","caveat":"Baseline is supplementary, not a required estimand."},7 {"id":"C6","text":"At G=10 with ICC>0, WCB covers better than CR1.","results_keys":["coverage","mc_se_coverage"],"status":"supported","caveat":"Differences (0.014 and 0.026) are of the order of 2-4 MC SE of individual coverages; not formally tested."}8]9