Research Sandbox Environment
All experiments use synthetic data only. No production data is accessed or modified.
Experiment Lab
Double-blind experiments with statistical validation
Synthetic Data
Active Experiments
2
Currently running
Completed
1
This week
Significant Results
2
p < 0.05
Avg Effect Size
0.42
Cohen's d
Active Experiments
Resonance Threshold CalibrationSIGNIFICANT
Lowering resonance detection threshold from 0.7 to 0.5 improves early detection of trust degradation
Progress72%
Control (0.7)7.2
n = 1,245
Treatment (0.5)7.8
n = 1,189
p-value: 0.0023Effect: 0.42
Started: 2/10/2026Bedau Index Window Size
Extending the temporal window for Bedau Index calculation from 10 to 25 interactions improves emergence detection
Progress45%
10 interactions6.8
n = 678
25 interactions7.1
n = 645
50 interactions6.5
n = 612
p-value: 0.0890Effect: 0.18
Started: 2/10/2026Multi-Modal Coherence WeightsSIGNIFICANT
Equal weighting across modalities outperforms text-heavy weighting for trust scoring
Completed
Progress100%
Text-Heavy (70/30)7.4
n = 2,341
Equal (50/50)8.1
n = 2,287
p-value: 0.0001Effect: 0.67
Started: 2/8/2026