[check] Table dataset/descriptives vs metric_descriptives.csv: ok
[dataset] Claude 708/1000 (70.8%), Gemini 604/1000 (60.4%), GPT 602/1000 (60.2%) | total 1914/3000 (63.8%)
[descriptives] Claude: CC 1.77 (med 0), CYC 2.31, LOC 5.2, smells 0.40, zero 56.1% | Gemini: CC 2.06 (med 1), CYC 2.52, LOC 6.1, smells 0.37, zero 45.0% | GPT: CC 2.61 (med 1), CYC 2.87, LOC 6.4, smells 0.36, zero 43.7%
[check] pooled distribution diagnostics vs 09_cc_distribution_diagnostics.csv: ok
[distribution] mean 2.13, median 1, Q1 0, Q3 3, max 25, zero 48.7%, skewness 2.42 (adjusted G1), Shapiro W 0.685 p 8.9e-51
[check] FF1 Friedman, W, mean ranks, block means vs ff1_friedman.csv / ff1_rank_sums.csv: ok
[check] agreement by minimum level (Fig. 3b) vs 12_agreement_by_level.csv: ok
[FF1] chi2 84.48, p 4.51e-19, W 0.0791 | mean ranks 1.795/2.043/2.162 | block means 1.547/1.918/2.453 | identical 46.8%, all zero 37.5%, max spread 17 (task 129 standard: 8/10/25)
[Fig3b] min=0: n 336, identical 59.5%, spread 1.17 | min=1: n 74, identical 31.1%, spread 2.22 | min=2-3: n 61, identical 26.2%, spread 2.02 | min>3: n 63, identical 17.5%, spread 3.32
[check] output length medians and token-complexity correlations vs 21_verbosity*.csv: ok
[output length] standard medians Claude 47, Gemini 110, GPT 40 | readability-standard median diff 0/28/0, p_holm 4.38e-02 / 4.78e-22 / 1.55e-06 (no continuity correction) | rho(tokens, CC) 0.791/0.605/0.859
[check] robustness table vs 21_robustness_*.csv: ok
[robustness] cc              534 blocks | W 0.0791 | p 4.51e-19 | order Claude < Gemini < GPT | mean ranks 1.79/2.04/2.16 | means 1.55/1.92/2.45
[robustness] cyc             534 blocks | W 0.0775 | p 1.05e-18 | order Claude < Gemini < GPT | mean ranks 1.81/2.02/2.17 | means 2.18/2.44/2.77
[robustness] loc             534 blocks | W 0.1098 | p 3.35e-26 | order Claude < GPT < Gemini | mean ranks 1.72/2.07/2.21 | means 4.77/6.13/5.96
[robustness] code_smells     534 blocks | W 0.0054 | p 0.0561 | order Gemini < GPT < Claude | mean ranks 1.99/1.99/2.02 | means 0.36/0.36/0.38
[robustness] cc standard     270 blocks | W 0.0302 | p 0.000288 | order Claude < Gemini < GPT | mean ranks 1.89/1.99/2.11 | means 1.66/1.76/2.27
[robustness] cc readability  264 blocks | W 0.1512 | p 4.66e-18 | order Claude < Gemini < GPT | mean ranks 1.70/2.09/2.21 | means 1.43/2.08/2.64
