,
9 tweets,
1 min read
Read on Twitter
In light of reporting on Facebook/Russia, worth remembering what we know about causal effect of targeted FB ads 1/
Assumption (among both commentators and Parscale) seems to be that FB ads effectively demobilized key pro-Clinton demographic groups 2/
To my knowledge (correct me if wrong), no public study has looked at that precise Q. But a couple have looked at related questions 3/
.@dbroockman and Green find null effect of real-world Cong. campaign FB ads on candidate knowledge, favorability, and vote intent 4/
... and that's with people in study exposed to the ads an average of 38 times! If these move the needle, imperceptibly small effects 5/
Ryan (2012) also did FB ad experiments and found that anger can generate clicks. Given past year, not surprising. 6/
But real question is whether ads can boost (or deflate) turnout. Classic Bond et al. FB experiment found non-social treatment = 0.0 7/
So I'm suggesting that our priors should be adjusted accordingly. Caveat is possibility of heterogeneity ... 8/
i.e., some of the targeted subgroups are especially susceptible to whatever appeals they were getting.
