New preprint!

Why are people’s policy opinions influenced by cues from political parties?

We studied the theory that people use party cues as heuristics, stand-ins for their lack of policy information and motivation for effortful thinking.

https://psyarxiv.com/tdk3y 

Highlights:

The theory that party cues are used as heuristics for policy information and cognitive effort has been interpreted as encouraging for democracy, because it ostensibly offers a part-solution to the “democratic dilemma” of widespread political ignorance & disinterest.

A key prediction that follows from the heuristic theory is that the influence of party cues would diminish if only people were to possess more information about policy, a greater propensity for effortful thinking, or both.

Does existing evidence bear this prediction out?

In the paper, we argue that it’s hard to tell: the designs of many previous studies face severe challenges to both causal inference (due to plausible unobserved confounders) and empirical generalizability (due to small and idiosyncratic samples of policy issues).

We designed two large experiments in an effort to overcome these challenges. We randomized exposure to party cues, policy information, and inducement to effortful thinking, and we measured US adults’ opinions over 20 contemporary US policies. Treatment compliance was high.

Three highlights:

1. Exposure to policy information causally attenuated the influence of party cues, but engagement in effortful thinking per se did not—neither by itself, nor when paired with exposure to policy information. The average attenuation under information was ~35%.

2a. There was large variation between issues in the magnitude of attenuation of party cue influence under policy information: attenuations ranged from ~20% to ~100%, and everything in between. This illustrates the importance of studying more than a handful of issues at once.

2b. This point is hard to overstate: had we only studied 1-4 issues in the previous figure, we could have drawn vastly different conclusions about the magnitude of attenuation. For more on this point:

https://academic.oup.com/isq/article-abstract/61/2/425/4065443

https://psyarxiv.com/jqw35  https://psyarxiv.com/p48zb/ 

3. On aggregate, people exposed to party cues behaved qualitatively similarly to those exposed only to policy information: the ATEs are both > 0. This is generally consistent with the notion that party cues offer a valid informational basis for people’s opinion formation.

The broader implications of these results and others are in the paper.

The work was done with Ryan McKay (recently Twitter-less, but trying to tempt him back).

Thanks for listening; we would appreciate feedback of any kind.