Table 11.

Effects of concordant message topics.

	Ordered logit
	Intent to vaccinate self	Intent to vaccinate child
Concordant messages by topic
Population tested	−0.013	−0.005
	$[-0.228,0.201]$	$[-0.416,0.405]$
Community impact	0.059	0.174
	$[-0.073,0.191]$	$[-0.067,0.414]$
Children affected by COVID-19	0.169	0.189
	$[-0.059,0.397]$	$[-0.038,0.417]$
Protecting the elderly	−0.026	0.018
	$[-0.198,0.145]$	$[-0.223,0.259]$
Protection from vaccine	−0.133	−0.095
	$[-0.342,0.075]$	$[-0.446,0.256]$
Gatherings made possible	0.062	0.058
	$[-0.090,0.214]$	$[-0.156,0.272]$
Vaccination locations	−0.026	−0.128
	$[-0.223,0.172]$	$[-0.356,0.100]$
Recommendation	0.020	−0.054
	$[-0.147,0.186]$	$[-0.351,0.244]$
$\mathrm{\mathbb{P}}(\mathrm{all}\mathrm{topics}=0)$	0.715	0.510
Subsegments	24	11
Observations	2621	1032

Notes: 95% confidence intervals in brackets using heteroskedasticity-robust standard errors. Regressions include subsegment fixed effects. Outcome ranges from 1 (highly unlikely) to 7 (highly likely). Concordant messages are the number of message components customized for that respondent’s segment memberships. Impact–Churches and Gatherings–Freedom were each a separate randomization, but analyzed as part of the Impact and Gatherings bundles, respectively. For example, everyone received one of the Impact messages, and half the Religious segment also received the Impact–Churches message.

⁺$p<0.10$, * $p<0.05$, ** $p<0.01$, *** $p<0.001$.