E-values for Observed Treatment Effects
Based on primary model: ‘Unweighted with Lin adjustment’. Minimum strength of association an unmeasured confounder needs with treatment and outcome to explain away the effect.
Outcome	Model	Estimate	p-value1	E-value (Point Est.)	E-value (Lower CI)2
Wave 1
Feeling Thermometer (W1)	Unweighted with Lin adjustment	4.09	0.000	1.50	1.36
Behavioral Intentions (W1)	Unweighted with Lin adjustment	0.24	0.000	1.65	1.49
Attitudes (W1)	Unweighted with Lin adjustment	0.09	0.006	1.38	1.18
Latent Support (W1)	Unweighted with Lin adjustment	0.23	0.000	1.57	1.43
Wave 2
Feeling Thermometer (W2)	Unweighted with Lin adjustment	1.62	0.214	1.27	1.00
Behavioral Intentions (W2)	Unweighted with Lin adjustment	0.06	0.292	1.26	1.00
Attitudes (W2)	Unweighted with Lin adjustment	0.12	0.021	1.44	1.14
Latent Support (W2)	Unweighted with Lin adjustment	0.12	0.045	1.36	1.04
1 Bold p-values indicate significance at p < 0.05.
2 E-value (Lower CI) is 1 if the 95% CI for the treatment effect includes the null.