Do you think of yourself as a beer-drinker, or do you merely drink beer? Are you a Volkswagen driver or a Democrat, or do you just drive a Volkswagen or vote for Democratic candidates? The difference in how these notions are worded involves more than just splitting linguistic hairs. Phrases like drive a Volkswagen or vote for Democrats simply describe behaviors that you take part in. But if you assert that you're a Volkswagen driver or a Democrat, you're expressing something fundamental about who you are. Wielding the right language to tap into people's sense of identity, as it turns out, can make for potent persuasion.
Every now and then, though, someone makes it clear that covert techniques of persuasion can be used for the public good and not just for nefarious purposes. For example, one recent study shows that tweaking language to focus on identity can drive people to the voting booths. Christopher Bryan and several colleagues from Stanford and Harvard Universities recruited citizens who were registered to vote in the 2008 presidential election and gave them a short "election survey" less than 24 hours before the election. Some people saw questions such as "How important is it to you to vote in the upcoming election?" Others instead answered questions like "How important is it to you to be a voter in the upcoming election?" Subtle difference in wording, no? A mere flick of the linguistic wrist. But guess what? It had a massive effect on actual voter turnout. About 82% of the people who were asked the first question later turned up at the polls, just slightly above the national average for registered voters (79%). In contrast, the voter rate after the second question (which prompted people to think of themselves as either voters or non-voters) was a staggering 95.5%. A second study of the 2009 gubernatorial election in New Jersey yielded an effect that was just slightly smaller—a 10% difference due to questionnaire wording.
According to the authors, these findings are "among the largest experimental effects ever observed on objectively measured voter turnout." To put them into perspective, the researchers also measured the impact of age, gender, ethnicity, level of education, political knowledge, and even interest in politics—none of which showed any statistically reliable effects on voter turnout for their sample of registered voters. (Though perhaps they should also have studied the impact of supplying donuts at the polling stations.)
As the study by Bryan and colleagues suggests, we humans appear to resonate like plucked strings to messages that appeal to our self-concept. On the more commercial side of things, marketers have of late been putting more emphasis on telling us about who are and less on what their products can do for us. Techniques of identity marketing range from linking brands with certain symbols or spokespersons to choosing regional linguistic dialects in commercials. But no one has raised the practice to quite the same explicit level as Apple with its Mac vs. PC ad campaign, in which stereotyped characters were offered up as the very embodiment of competing computer brands. Remember how each ad began with the words "I'm a Mac" and "I'm a PC"? You don't get identity marketing much clearer than that. (And the central message was equally clear: to avoid being fatally dorky, get a Mac.)
Consumers, incidentally, jump at the opportunity to use products as a way of defining and broadcasting their personal identities, as documented by Rob Walker in his interesting book Buying in: What We Buy and Who We Are. These days, to drink Pabst beer rather than Heineken or to use French's yellow mustard over Dijon-style is more than just a matter of personal taste. It says something about who we are, and we thank the makers of beer and mustard for giving us the tools with which to say it.
Maybe I shouldn't have been as startled as I was by the whopping size of the effect of identity-based language on voter turnout. After all, I've recently witnessed an example of how compelling the need can be to align consumer behavior with a sense of self, and how this need can override even the most careful deliberation. Some close friends of mine were in the market for a new car. They're thorough, thoughtful types, and did a lot of research before ever setting foot in a dealership. They narrowed the field down to two station wagons they were interested in: the Volkswagen Passat and the Subaru Outback. Over dinner one evening, they reported that they'd made a list of the features they wanted in a car and had pored over consumer reports. They were going to test-drive both cars, but their research already suggested that the Passat had the greatest number of their desired features at a better price. Two weeks later, they pulled up in their shiny new Subaru. "What happened to the Passat?" I asked. The husband looked sheepish. "When it comes down to it," he explained, "We're just not Passat people."
Follow Sold on Language on Twitter
I first learned of this interesting paper in a blog post by linguist David Beaver. In his post on the Language Log blog, Beaver addresses some more subtle linguistic points than I have talked about here.
To read the paper on identity-based language and voter turnout:
Bryan, C. J., Walton, G. M., Rogers, T. & Dweck, C.S. 2011. Motivating voter turnout by invoking the self. Proceedings of the National Academy of Science, 108, 12653-12656
Source for comic: http://www.phdcomics.com/comics.php
Copyright © 2011 Julie Sedivy, All Rights Reserved