Thread by @davidlazer

See our latest report, (survey) experiments with vaccine messaging. In first experiment, we varied the messenger-the question, what impact did this have on vaccine resistance? Treatment effects below (higher = more vaccine resistance).1/7
COVIDstates.org

Take away: lots of potential for a backfire here. Celebrities and athletes have null effects, and anyone with a political hue risks driving away anti-partisans more than they persuade same partisans. (See this figure for partisan breakdowns). 2/7

Fauci is an illuminating case, because he has become a partisan figure, and evoking him makes Democrats a bit less resistant, but Republicans and Independents substantially more resistant, with a net effect of increasing resistance. 3/7

But, if we vary message type, we find substantial treatment effects to reduce resistance, especially if own doctor or "scientists" recommend. Interestingly, appeals from people you know and to preventing harm more effective in counties with more cases. 4/7

Not shocking, and it does highlight, especially, the role that community embedded professionals will play in persuading people to be vaccinated. 5/7

All the usual caveats re the generalizability of survey experiments/etc. 6/7

COVIDstates.org
w/
@_Jon_Green

Alexi Quintana
@Ognyanova

@hanyuchwe

@matthewabaum

@royperlis

@mattysimonson

@Atauslu

@AdinaGitomer

@MauSantillana

@jennifer_lin16

@Northeastern

@Harvard

@NorthwesternU

@NUlabTMN

@NUnetsi

@IQSS

@Kennedy_School
7/7