MultiCausalData(df=(20000, 12), treatment_names=['d_0', 'd_1', 'd_2'], control_treatment='d_0')outcome='y', confounders=['tenure_months', 'avg_sessions_week', 'spend_last_month', 'premium_user', 'urban_resident', 'support_tickets_q', 'discount_eligible', 'credit_utilization'], user_id=None,

{'overall_flag': 'GREEN', 'all_comparisons_pass': True}

{'overall_flag': 'RED', 'flags': {'psi_tail_ratio': 'YELLOW', 'psi_kurtosis': 'RED', 'ortho_max_|t|': 'RED', 'oos_moment': 'GREEN', 'ortho_max_|t|gk': 'RED', 'ortho_max|t|g0': 'RED', 'ortho_max|t|mk': 'GREEN', 'ortho_max|t|m0': 'GREEN'}, 'flags_by_comparison': comparison psi_tail_ratio psi_kurtosis ortho_max|t| oos_moment
0 d_1 vs d_0 YELLOW RED RED GREEN
1 d_2 vs d_0 YELLOW RED RED GREEN

overall_flag
0 RED
1 RED }

1.) Are your clients independent (i). Outcome of ones do not depend on others? 2.) Are all clients have full window to measure metrics? 3.) Do you measure confounders before treatment and outcome after? 4.) Do you have a consistent label of treatment, such as if a person does not receive a treatment, he has a label 0?

avg_sessions_week 0.057140 tenure_months 0.046042 urban_resident 0.028845 premium_user 0.024638 support_tickets_q 0.016798 spend_last_month 0.013420 discount_eligible 0.007124 credit_utilization 0.004336 dtype: float64

{'overall_flag': 'GREEN', 'flags': {'balance_max_smd': 'GREEN', 'balance_violations': 'GREEN'}, 'overall_balance_pass': True}

================== Bias-aware Interval ==================

------------------ Scenario ------------------ Significance Level: alpha=0.05 Null Hypothesis: H0=0.0 Sensitivity parameters: cf_y=0.010101010101010102; r2_d=[0.01 0.01], rho=[1. 1.], use_signed_rr=False

theta se max_bias max_bias_base bound_width sigma2 nu2 sampling_ci_l sampling_ci_u theta_l theta_u bias_aware_ci_l bias_aware_ci_u rv rva d_1 vs d_0 -1.225401 0.053151 0.071358 7.064466 0.071358 11.726175 4.256007 -1.329575 -1.121226 -1.296759 -1.154043 -1.400830 -1.048932 0.748664 0.713779 d_2 vs d_0 2.570377 0.109287 0.075791 7.503273 0.075791 11.726175 4.801148 2.356178 2.784576 2.494587 2.646168 2.281413 2.861468 0.920747 0.907083

{'theta': array([-1.22540079, 2.57037715]), 'se': array([0.05315128, 0.10928717]), 'alpha': 0.05, 'z': 1.959963984540054, 'H0': 0.0, 'sampling_ci': array([[-1.32957538, -1.1212262 ], [ 2.35617823, 2.78457606]]), 'theta_bounds_cofounding': array([[-1.29675903, -1.15404255], [ 2.49458652, 2.64616778]]), 'bias_aware_ci': array([[-1.40082993, -1.0489322 ], [ 2.28141319, 2.86146838]]), 'max_bias_base': array([7.06446594, 7.50327258]), 'max_bias': array([0.07135824, 0.07579063]), 'bound_width': array([0.07135824, 0.07579063]), 'sigma2': 11.726175012924882, 'nu2': array([4.25600667, 4.8011478 ]), 'rv': array([0.74866425, 0.92074748]), 'rva': array([0.71377934, 0.90708265]), 'contrast_labels': ['d_1 vs d_0', 'd_2 vs d_0'], 'params': {'cf_y': 0.010101010101010102, 'r2_d': array([0.01, 0.01]), 'rho': array([1., 1.]), 'use_signed_rr': False}}