By Gus Skorburg, Duke University
Across a series of papers, Christopher Bryan and colleagues have reported that invoking the self can be a powerful way to motivate political, social, and moral behavior. But what does it mean to ‘invoke the self’? And what kinds of behaviors can be motivated?
In one study (Bryan, Walter, Rogers, & Dweck, 2011), the researchers reported that subtle linguistic cues can change voting behavior. Across three experiments, the framing of survey questions was varied to either evoke identity (“How important is it to you to be a voter in the upcoming election?”) or an action (“How important is it to you to vote in the upcoming election?”). The researchers hypothesized that participants in the noun phrase condition, which made the self salient, would be more likely to vote than participants in the verb condition, which did not appeal to the self.
Based on decades of research in social psychology which has shown that people have a desire to see themselves as competent, morally good, and well-regarded socially, their idea is that “being the kind of person who votes may be seen as a way to build and maintain a positive image of the self” and that “people may be more likely to vote when voting is represented as an expression of self—as symbolic of a person’s fundamental character—rather than as simply a behavior” (Bryan, Walter, Rogers, & Dweck, 2011, p. 12653).
In the first experiment, Bryan and colleagues gave people who were eligible, but not yet registered, to vote in the 2008 presidential election in California either the noun-based or the verb-based survey. After taking the survey, participants were then asked to indicate their level of interest in registering to vote. Those who took the self-evoking noun survey (“to be a voter”) were significantly more likely to be interested to register to vote (87.5%) than those who took the verb-based (“to vote”) survey (55.6%).
The second experiment looked at voter turnout in the 2008 presidential election in California. After taking either the noun-based or verb-based surveys, the researchers used official state records to check whether participants actually voted. After controlling for the prior probability of turnout, they found that those who took the self-evoking noun survey (“to be a voter”) were significantly more likely to have voted (95.5%) than those who took the verb-based (“to vote”) survey (81.8%).
In a third experiment, the researchers extended the results from the second experiment to a larger and more diverse sample in the 2009 New Jersey gubernatorial election. And once again, participants in the noun condition voted at a higher rate (89.9%) than participants in the verb condition (79.0%).
Can a such a subtle linguistic manipulation – the difference between “a voter” and “to vote” – really have such a significant impact on political behavior? A follow-up study (Gerber, Huber, Biggers, & Hendry, 2015) suggests that perhaps the results are too good to be true.
Gerber et al. used the same noun and verb-based survey (though administered over the phone, rather than the internet) and added a Get Out the Vote (GOTV) condition which used a standard phone script to encourage people to vote in an upcoming election. Their participants were voters in 2014 primary elections in Michigan, Missouri, and Tennessee. In contrast to the large effects in the Bryan et al. study, Gerber et al. found no significant difference in turnout between the noun and verb conditions. Moreover, they found that both the noun and verb-based surveys were less effective than the shorter, basic GOTV message.
In a reply, Bryan, Walton, & Dweck (2016) brush off the results of this follow-up study by noting that their original experiments were conducted in the context of “high-profile elections that received substantial public attention,” while the Gerber et al. study was conducted in the context of lower-profile congressional primaries which received much less attention.
While it is often difficult to understand the generalizability of psychological findings, or the status of replication studies, it should be noted that “self as motivator” effects have been observed in other domains.
Bryan, Master, & Walton (2014) report that children helped with significantly more tasks in a noun condition (“being a helper”) than in a verb condition (“helping”). Similar results obtained in a study by Bryan, Adams, & Monin (2013): people were less likely to cheat in a noun condition (“don’t be a cheater”) than a verb condition (“don’t cheat”).
Across all of the studies reviewed here, Bryan and colleagues claim that invoking the self can motivate political behavior, social behavior, and moral behavior. Our team is interested in whether and how these effects are related to political polarization. Are there ways to invoke self and identity to promote dialogue across political differences? How important is political identity to the self? Are there certain aspects of political identity that are more important than others?