Comparison of Standard Estimates and Lee Bounds
Assessing treatment effect robustness to differential attrition
	Standard Estimate1		Lee Bounds2
	Point Est.	95% CI	Lower	Upper	Width
Wave 1
Feeling Thermometer (W1)	4.09	[2.49, 5.69]	−2.01	10.12	12.14
Behavioral Intentions (W1)	0.24	[0.16, 0.31]	0.07	0.54	0.46
Attitudes (W1)	0.09	[0.03, 0.16]	−0.14	0.28	0.42
Latent Support (W1)	0.23	[0.16, 0.30]	−0.05	0.56	0.60
Wave 2
Feeling Thermometer (W2)	1.62	[-0.93, 4.16]	0.92	5.60	4.67
Feeling Thermometer (W2)	1.62	[-0.93, 4.16]	−5.84	10.56	16.40
Behavioral Intentions (W2)	0.06	[-0.05, 0.18]	0.07	0.26	0.19
Behavioral Intentions (W2)	0.06	[-0.05, 0.18]	−0.13	0.51	0.64
Attitudes (W2)	0.12	[0.02, 0.22]	0.06	0.25	0.19
Attitudes (W2)	0.12	[0.02, 0.22]	−0.17	0.41	0.58
Latent Support (W2)	0.12	[0.00, 0.24]	0.08	0.34	0.26
Latent Support (W2)	0.12	[0.00, 0.24]	−0.21	0.61	0.82
1 Standard model: Unweighted with Lin adjustment used for Standard Estimate & 95% CI.
2 Bounds assume monotonicity of selection effect.