Don't read this post: Be a Language Log reader!

« previous post | next post »

The big deal in a new paper "Motivating voter turnout by invoking the self" (see also the official PNAS site, or e.g. this Discover magazine article "The power of nouns….") is that people can be manipulated into voting simply by clever use of nouns instead of verbs in a questionnaire. In each of several studies, potential voters were split into two groups and given (amongst other questions which didn't vary by group) one of two questions to answer:

Group 1 question: How important is it to you to be a voter in the upcoming election?

Group 2 question: How important is it to you to vote in the upcoming election?

Turned out that Group 1 turned out. Really. In one of the studies an amazing 95.5% of them actually turned out to vote, whereas only 81.8% of Group 2 voted. That's obviously a huge effect on voting behavior. And it appears to be caused by the use of a construction with the nominal "voter" instead of the verb "vote".

As hinted at by the paper's title, the authors (Christopher Bryana, Gregory Walton, Todd Rogers, and Carol Dweck) don't think the difference is due to the use of nouns versus verbs per se. Rather, they give an explanation that has a sociolinguistic personal identity feel: when people are asked about "being a voter", they reflect on who they really are, and whether they are the sort of person that does the right thing, and this is very motivating. But when people are asked about whether they will vote, the idea is that people are less likely to consider what sort of person they really are, and instead will just consider the mundane and practical question of what they will do on election day. Cute!

Readers are probably thinking about how such manipulations could be used in other ways to influence their loved ones, customers, students, or professors. And they're probably also thinking: is this really about nouns versus verbs? Here's my 2 cents worth: no. I'd guess that the important property that the nominal constructions have is that they are, to introduce some semanticist's jargon originated by Greg Carlson, individual level, whereas the Group 2 verbal construction is stage level. Put simply, individual level predications tell you about general properties of things, whereas stage level predications tell you what those things are doing, or what state they are in, at a particular time. Thus to say "Barack will read this post" is to make a stage level claim about Barack, whereas to say "Barack is a Language Log reader" is to make an individual level claim about a general property Barack has.

And now it should be clear why this presumably isn't a matter of nouns or verbs. To say "How important is it to you to be one of the voters tomorrow?" would still involve a nominal use of (the plural of) "voter", and yet would be stage level. I'd guess it wouldn't have a huge motivating effect, despite the use of the nominal. On the other hand, to ask "How important is it to you that you vote in elections?" would be to ask about a generic property of the addressee, the individual level tendency to vote. I'd guess this might have a relatively higher motivating effect than a stage level question. Someone please do the study!


  1. Ramsay said,

    July 22, 2011 @ 5:02 am

    Does it also matter what the negative answer implies? To not vote might not be that important. To be a "non-voter" sounds like you have lost the right to self-determination…

  2. Adam said,

    July 22, 2011 @ 5:25 am

    For some reason this reminds me of two random things (which I'm not endorsing, just mentioning).

    Strunk & White said something like "write with nouns and verbs"; maybe nouns are best of all?

    One chapter of Freakonomics says there's a strong correlation between what parents are (educated, for example) but not what they do (take kids to museums, keep them from watching TV) and the kids' educational achievements.

  3. Keith M Ellis said,

    July 22, 2011 @ 5:44 am

    My not-very-deep-or-rigorous reaction to this is that the first question with its "to be a voter in the upcoming election" sort of encourages the reader to more concretely imagine themselves, as themselves, being someone who is voting in the upcoming election. Whereas the second question just asks the reader to query themselves about a value in a mostly abstracted context.

    The first question, in the specific way in which it asks the reader to imagine themselves as themselves actually voting in the upcoming election, does so with a strong but hard-to-pin-down emphasis on "themselves as themselves a a voter" in specifically the social identity context. Which is where our desire to have our positive self-image be reflected in how others view us comes powerfully into play in our motivations.

    Is this related to your individual level and stage level distinctions in some way that I'm missing?

  4. Stuart said,

    July 22, 2011 @ 6:06 am

    Interesting. I work in market research,so the results of this study are interesting professionally. Never voting is also a fundamental principle in my personal life, so I am not sure that the changed phrasing would have the desired impact. The LL article says"when people are asked about "being a voter", they reflect on who they really are, and whether they are the sort of person that does the right thing, and this is very motivating" – but that presupposes that the reader of the question believes (as it seems the LL author does) that "being a voter" ="being the sort of person that does the right thing". The idea of being a voter is anathema to me, so the motivation I get from both the survey question and the LL article is perhaps non-standard.

  5. GeorgeW said,

    July 22, 2011 @ 6:09 am

    Did they say what the total turnout was for this particular election? The turnout seems higher for both groups than elections generally. If the general turnout was much less, this might suggest that just asking either question prompted voting.

    [(myl) From the paper:

    In addition to the condition difference, the overall turnout rate in our sample merits consideration. The turnout rate in the verb condition of 81.8%, although not meaningfully higher than the statewide turnout rate of 79.4% among registered voters in this election, was higher than would be expected given the demographic makeup of the sample. Using US Census data on turnout rates in this election by age, gender, and level of education (19), the expected premanipulation turnout rate of participants in our sample was 63.9% (95% confidence interval: 61.9–66.0%).

    We suggest two possible reasons why the turnout rate, even in the verb condition, was higher than this expected baseline rate. First, people who choose to participate in a survey related to an upcoming election—or even those who have the time and inclination to participate in any survey shortly before an election— may be more likely to vote than the general population. Second, some research suggests that merely responding to a pre-election survey can increase voter turnout (16, 17, 20, 21; but see ref. 18). These findings highlight the rigorous nature of the control condition in the present research, which used questions that were nearly identical to those in the treatment condition. In any case, the critical finding from experiment 2 is that a subtle change in the phrasing of survey items, which cast voting as a reflection of the kind of person one is rather than as merely a behavior, significantly increased voter turnout.


    FWIW, my initial reaction was similar to the authors', that it is a mater of personal identity.

  6. F said,

    July 22, 2011 @ 10:44 am

    Transit blogs (e.g. have sometimes discussed the opposite problem: when you identify people by nouns (cyclists, transit users, drivers, pedestrians), many people will disassociate themselves from all but one of the categories, even though, say, everyone is sometimes a pedestrian. When you frame the discussion in terms of "drivers do X, pedestrians do Y", you risk losing support from people who drive a lot of the time and will bin themselves with "drivers".

    Here, I feel like when you mention "being a voter", people think "if I don't vote, I will be a nonvoter, part of a large block of people who are not represented in the political system." But maybe I'm overthinking it.

  7. Julie Sedivy said,

    July 22, 2011 @ 11:55 am

    Interesting post. The individual-level strategy is one that is used all the time in marketing, whether implicitly or explicitly, and with considerable success. At one point, friends of mine bought a Subaru despite having ascertained that a Passat would be the best vehicle for their needs. They justified their decision by saying "We're just not Passat people."

    The strategy was exploited to the hilt in those long-running "Mac vs. PC" commercials. As you may recall, just about every one of these ads began with the words "Hello, I'm a Mac. And I'm a PC." And the ads of course left no doubt as to what it meant to be either a Mac or a PC.

  8. McLemore said,

    July 22, 2011 @ 12:04 pm

    "social identity" means identity with a group. "to be a voter" evokes one's identity with people who vote, and also therefore participation in a communal activity with shared goals (e.g. sustaining a democracy). The important thing is that it emphasizes participating in a social activity.

    "to vote" puts the emphasis on a single instance of individual activity, like something from a list of errands. You can miss voting in one election and still think of yourself as a voter…. unless you're called on it by the survey, I suppose.

    the increase in turnout by people who respond to such surveys would also suggest an evocation of communal activity and goals.

    I wonder how the Facebook practice of reporting one's having voted in a highly visible community count affected turnout in the '08 presidential election? Guess there's no way to tease out that effect from all the other factors that converged to produce such a large turnout…

  9. blahedo said,

    July 22, 2011 @ 12:56 pm

    When I first read about this, and again today, I wondered about the relationship of this effect to the ser/estar distinction in Spanish. To really tease out the linguistic facts, you could run a study (in a Spanish-speaking country) that has three groups, the verb-form group and one group that gets the noun with "ser" and one that gets the noun with "estar"—like David I suspect it has nothing to do with the part of speech (qua POS) of the vot- form, and everything to do with deeper semantic implications about habituality and group membership.

  10. Karl Weber said,

    July 22, 2011 @ 3:55 pm

    Interesting. Reminds me of the old observation that many more people want to "be writers" than to actually "write." The former names an identity that carries some cachet, glamor, prestige, etc., while the latter denotes an activity that is basically solitary intellectual drudgery.

  11. Lemuel Pitkin said,

    July 22, 2011 @ 4:01 pm

    People who've ridden the New York City subways recently will have noticed the somewhat confusing ads for the newly-paywalled New York Times, with the slogan, "Become a Digital Subscriber." I had thought this was just an awkward phrase by a copywriter who couldn't think of any better way of saying, "Subscribe to the Times online." But maybe they're using cutting-edge psychological research?

    (What makes the ads confusing is that the words "Digital Subscriber" are much larger than "Become a", and usually accompany a picture of someone the Times might write about — a young Bedouin girl, say — but who seems unlikely to subscribe.)

  12. Graham said,

    July 23, 2011 @ 1:26 am

    But 'voter in the … election' is not an individual-level predicate. Only a pattern of voting behavior would make one an (individual-level) voter, and the reliable diagnostic 'used to' shows the contrast:

    We used to be voters.
    ??We used to be voters in the last election.
    (cf. We were voters in the last election)

    'Voter in the election' is like the event-related nouns 'passenger' and 'batter' which Carlson and Manfred Krifka discuss.

    I do wonder whether aspect isn't to blame, and whether "How important is it to you to have voted in the upcoming election?" would also be less encouraging than the more direct "to vote" question.

    [Excellent point! If the stage level/individual level distinction is to be invoked, the claim has to be not that the full VP "be a voter in the upcoming election" expresses an individual level predication, but that it makes salient a certain individual level property, i.e. being a voter, by virtue of containing the phrase "be a voter". Your suggestion that aspect is important is also a good one: "be a voter in the upcoming election" is obviously stative, whereas "vote in the upcoming election" is telic. But it's not clear to me why (non-)telicity should matter beyond how it relates to generality of properties of individuals.

    Incidentally, I suspect that another very important factor is the slight oddity of "be a voter in the upcoming election", as witnessed by (a) the fact that it sounds odd to me, and (b) the fact that related google searches show that phrasing to be very rare. As witness to (b), note e.g. that Google finds only a single hit for "be a voter in the municipal election", but a large number for "vote in the municipal election". I suspect that the unexpectedness of "be a voter in the upcoming election" caused subjects to take more notice of it. But again, someone needs to do the study!

    — David Beaver]

  13. Barbara Partee said,

    July 25, 2011 @ 3:12 pm

    I experienced this effect very vividly in 1986, when I had just quit smoking for the final time, and was looking for all the assistance I could get. I attended a group quit-smoking series run by a local hypnotist; I don't know if I was ever hypnotized (don't think so), but he had lots of very useful suggestions apart from any hypnosis stuff. The one that I found most helpful was directed at the common experience of "relapsing" at a party or something — you have a cigarette, and you feel like you've failed once again. His advice: think: when you quit, you have stopped being a smoker and have become a non-smoker. Having one cigarette at a party is smoking, it's not "being a smoker"; just keep identifying yourself as a non-smoker, even if you had a cigarette or two. I could really feel the difference, and that seemed immediately like a really good idea. (And I'm still a non-smoker.)

RSS feed for comments on this post