One of the double-edged swords of working in the humanities is that the news media doesn’t often report on the findings we publish in journals. Double-edged because on the one hand, we don’t get much popular exposure for the work we do, and on the other hand, that work doesn’t get grossly mischaracterized as tends to happen when the media reports on the findings of a particular study. So it’s not often that I see sloppy reports about study findings that actually relate to my academic interests.
But a recent article in the Journal of Personality and Social Psychology about reading and identification (though that’s not how the authors phrase it) has been getting (typically hyperbolic) headlines: Do Your Favorite Book Characters Change Your Life?, You are what you read, study suggests, ‘Losing Yourself’ in a Fictional Character Can Affect Your Real Life, etc. Each of those articles was more unhelpful than the next in figuring out what this study was actually claiming about reading and (implicitly) fiction, so I had to take a look at the study itself.
The results summary in the abstract gives a nicely succinct synopsis of what the authors are claiming about their research:
“Results from Studies 1–3 showed that being in a reduced state of self-concept accessibility while reading a brief fictional work increased—and being in a heightened state of self-concept accessibility decreased—participants’ levels of experience-taking and subsequent incorporation of a character’s personality trait into their self-concepts. Study 4 revealed that a first-person narrative depicting an ingroup character elicited the highest levels of experience-taking and produced the greatest change in participants’ behavior, compared with versions of the narrative written in 3rd-person voice and/or depicting an outgroup protagonist. The final 2 studies demonstrated that whereas revealing a character’s outgroup membership as a homosexual or African American early in a narrative inhibited experience- taking, delaying the revelation of the character’s outgroup identity until later in the story produced higher levels of experience-taking, lower levels of stereotype application in participants’ evaluation of the character, and more favorable attitudes toward the character’s group.”
There are a couple of things about this study I find interesting and potentially useful, and a great many more I have reservations about. First of all, this being social psychology, the terminology is significantly different than what I would tend to use. The authors focus on “experience-taking,” which they define as follows: “We propose that when experience-taking occurs, readers simulate the events of a narrative as though they were a particular character in the story world, adopting the character’s mindset and perspective as the story progresses rather than orienting themselves as an observer or evaluator of the character.” That definition has some implicit genre-limitations, since any number of fictional modes attempt to intentionally disrupt the reader’s ability to adopt a character’s mindset. But that aside, “experience-taking” seems to be very similar to Kenneth Burke’s idea of identification or consubstantiality, a definition I lean heavily on when I talk about fiction, identification, and reading here and here. According to Burke, “In being identified with B, A is ‘substantially one’ with a person other than himself. Yet at the same time he remains unique, and individual locus of motives. Thus he is both joined and separate, at once a distinct substance and consubstantial with another.”
Burke also notes, though, that “to begin with ‘identification’ is, by the same token, though roundabout, to confront the implications of division,” since “Identification is compensatory to division.” That’s what seems to be happening in studies 4-6, which reveals that identification works through associations with a shared ingroup, and is disrupted when the person meant to be identified with is a member of an outgroup. No surprise there, though I’d argue that the findings of the last two studies, which claim that deferring the revelation that the identifiable character is a member of an outgroup increases the likelihood of identification, are at least partially contradicted by the sort of readings evident in responses to the Hunger Games.
And then there is the fairly significant problem that the study seems to conflate “stories” or “narratives” with “fiction” or “novels.” The procedure for each of the individual studies was to assemble a group of undergraduates, coerced by the offer of partial course credit, have them read different variations of a similar narrative (main character is black, main character is white, main character goes to same university as students, main character goes to different university, etc.), and then have them fill out a questionnaire about their responses. The first two studies also attempted to manipulate the students sense or awareness of themselves to either heighten or dampen identification (“experience-taking”).
But near as I can tell, the material the students were given to read was dead boring, akin to the kind of dreck one gets on standardized tests in grade school. This is the description of the material used in study 4:
“Pages 2–5 of the booklet contained the narrative; Participants were randomly assigned to read one of four versions of a short story that resulted from varying both the voice of the narrative (first-person versus third-person) and the university affiliation of the story’s protagonist, who was identified in the first sentence as a student at either Ohio State University (the same university that all participants attended) or Denison University (another university in the same state) when he was described as reading a voter’s guide provided by his particular university. The narrative depicted the character enduring several obstacles on the morning of Election Day (e.g., car problems, rainy skies, and long lines at the polling location) before ultimately entering the booth to cast a vote.”
Throughout the article, the authors use well-known works of fiction as anecdotal examples to imply the sort identifying reading experience they’re trying to simulate. They name-check The Catcher in the Rye, The Color Purple, To Kill a Mockingbird, and Brokeback Mountain, among others. But these studies really only tell us about readers’ response to narrative, and narrative alone does not a novel make. The implication that college students’ responses to some very dry-sounding reading material can be compared to readers’ necessarily more complex responses to texts that use a host of narrative and formal techniques is questionable at best.
Obviously, I think identification plays an important role in the most common reading modes available to readers today (I don’t think that’s historically been the case, but that’s a whole different question). But reading for identification is a mode that can often get pre-empted or short-circuited by other reading modes, particularly close reading. Dry stories about a college student failing to vote because his car broke down don’t require much of their readers. More to the point, the idea that we can make any generalizable claims about the effects of narrative based entirely on the reading experiences of a bunch of Ohio State freshmen poses some significant problems.
Unfortunately, social psychology is always going to get more popular attention, thanks to the veneer of scientific objectivity. But for the record, Kenneth Burke totally got there first.
Edited to add citations, so as not to be completely sloppy:
Burke, Kenneth. A Rhetoric of Motives. Berkeley: U of California Press, 1950.
Kaufman, G. F., & Libby, L. K. “Changing Beliefs and Behavior Through Experience-Taking.” Journal of Personality and Social Psychology. March 2012. Advance online publication.