% Author: Simon-Pierre Boucher — simon-pierre.boucher@uqo.ca
% generated by scripts/20_evaluate.py — do not edit
\begin{tabular}{llrrrrrrrrrr}
\toprule
Arm & {Committee} & {Turns} & {Valid ids} & {Adj. $\pm$40\%} & {In hull} & {Quantiles ok} & {Words} & {IQR$_0$} & {IQR$_R$} & {Anchoring} & {To median} \\
\midrule
E (single llm) & mixed\_random & \num{80} & \num{1.000} & \num{0.900} & \num{1.000} & \num{1.000} & \num{77} & {--} & {--} & {--} & {--} \\
F (independent) & mixed\_random & \num{960} & \num{1.000} & \num{0.709} & \num{0.994} & \num{0.991} & \num{78} & \num{0.086} & \num{0.086} & {--} & {--} \\
G (debate r2) & mixed\_random & \num{2856} & \num{1.000} & \num{0.721} & \num{0.998} & \num{0.988} & \num{78} & \num{0.088} & \num{0.038} & \num{0.713} & \num{0.451} \\
H (debate skeptic) & mixed\_random & \num{2654} & \num{1.000} & \num{0.717} & \num{0.998} & \num{0.985} & \num{78} & \num{0.092} & {--} & {--} & {--} \\
I (placebo) & mixed\_random & \num{2832} & \num{1.000} & \num{0.711} & \num{0.995} & \num{0.988} & \num{72} & \num{0.087} & \num{0.041} & \num{0.751} & \num{0.429} \\
J (cost matched) & mixed\_random & \num{1040} & \num{1.000} & \num{0.917} & \num{0.997} & \num{0.996} & \num{76} & \num{0.014} & {--} & {--} & {--} \\
E (single llm) & mixed\_random & \num{300} & \num{1.000} & \num{0.977} & \num{0.993} & \num{0.997} & \num{77} & {--} & {--} & {--} & {--} \\
F (independent) & mixed\_random & \num{3600} & \num{1.000} & \num{0.716} & \num{0.996} & \num{0.998} & \num{78} & \num{0.088} & \num{0.088} & {--} & {--} \\
G (debate r2) & mixed\_random & \num{10740} & \num{1.000} & \num{0.702} & \num{0.996} & \num{0.995} & \num{80} & \num{0.087} & \num{0.034} & \num{0.677} & \num{0.521} \\
E (single llm) & mixed\_random & \num{60} & \num{1.000} & \num{0.983} & \num{1.000} & \num{1.000} & \num{74} & {--} & {--} & {--} & {--} \\
F (independent) & mixed\_random & \num{720} & \num{1.000} & \num{0.732} & \num{0.998} & \num{0.994} & \num{78} & \num{0.075} & \num{0.075} & {--} & {--} \\
G (debate r2) & mixed\_random & \num{2147} & \num{1.000} & \num{0.728} & \num{1.000} & \num{0.992} & \num{77} & \num{0.073} & \num{0.037} & \num{0.700} & \num{0.465} \\
E (single llm) & mixed\_random & \num{40} & \num{1.000} & \num{0.975} & \num{0.975} & \num{1.000} & \num{71} & {--} & {--} & {--} & {--} \\
F (independent) & mixed\_random & \num{479} & \num{1.000} & \num{0.704} & \num{0.992} & \num{1.000} & \num{77} & \num{0.077} & \num{0.077} & {--} & {--} \\
G (debate r2) & mixed\_random & \num{1404} & \num{1.000} & \num{0.707} & \num{0.988} & \num{0.999} & \num{77} & \num{0.082} & \num{0.041} & \num{0.757} & \num{0.452} \\
E (single llm) & all\_anthropic & \num{40} & \num{1.000} & \num{1.000} & \num{1.000} & \num{1.000} & \num{63} & {--} & {--} & {--} & {--} \\
E (single llm) & all\_openai & \num{40} & \num{1.000} & \num{0.950} & \num{0.975} & \num{1.000} & \num{83} & {--} & {--} & {--} & {--} \\
E (single llm) & mixed\_random & \num{40} & \num{1.000} & \num{0.900} & \num{1.000} & \num{1.000} & \num{76} & {--} & {--} & {--} & {--} \\
F (independent) & all\_anthropic & \num{480} & \num{1.000} & \num{0.738} & \num{1.000} & \num{0.992} & \num{68} & \num{0.058} & \num{0.058} & {--} & {--} \\
F (independent) & all\_openai & \num{480} & \num{1.000} & \num{0.710} & \num{0.994} & \num{0.998} & \num{86} & \num{0.041} & \num{0.041} & {--} & {--} \\
F (independent) & mixed\_random & \num{480} & \num{1.000} & \num{0.731} & \num{0.994} & \num{1.000} & \num{77} & \num{0.064} & \num{0.064} & {--} & {--} \\
G (debate r1) & mixed\_random & \num{960} & \num{1.000} & \num{0.727} & \num{0.999} & \num{0.996} & \num{78} & \num{0.069} & \num{0.043} & \num{0.672} & \num{0.427} \\
G (debate r2) & all\_anthropic & \num{1428} & \num{1.000} & \num{0.748} & \num{1.000} & \num{0.993} & \num{69} & \num{0.056} & \num{0.018} & \num{0.970} & \num{0.525} \\
G (debate r2) & all\_openai & \num{1104} & \num{1.000} & \num{0.719} & \num{0.991} & \num{0.998} & \num{87} & \num{0.042} & \num{0.038} & \num{0.384} & \num{0.075} \\
G (debate r2) & mixed\_random & \num{1440} & \num{1.000} & \num{0.713} & \num{0.999} & \num{1.000} & \num{78} & \num{0.070} & \num{0.034} & \num{0.673} & \num{0.442} \\
G (debate r4) & mixed\_random & \num{1020} & \num{1.000} & \num{0.726} & \num{0.999} & \num{1.000} & \num{78} & \num{0.057} & \num{0.024} & \num{0.781} & \num{0.429} \\
G (fullview) & mixed\_random & \num{1416} & \num{1.000} & \num{0.883} & \num{0.997} & \num{0.993} & \num{78} & \num{0.072} & \num{0.028} & \num{0.651} & \num{0.517} \\
H (debate skeptic) & mixed\_random & \num{1360} & \num{1.000} & \num{0.717} & \num{0.999} & \num{0.998} & \num{78} & \num{0.074} & {--} & {--} & {--} \\
I (placebo) & mixed\_random & \num{1404} & \num{1.000} & \num{0.734} & \num{0.999} & \num{0.997} & \num{73} & \num{0.073} & \num{0.035} & \num{0.776} & \num{0.435} \\
J (cost matched) & mixed\_random & \num{520} & \num{1.000} & \num{0.935} & \num{0.979} & \num{1.000} & \num{76} & \num{0.015} & {--} & {--} & {--} \\
\bottomrule
\end{tabular}