Research over the last 20 years has shown that at least part of the academic problems children show in school are not wholly academic; some difficulties are rooted in self-image. Students may be hobbled by their perception of themselves as not fitting an academic context.
More pernicious, that threat to social identity can be cast on the student. That’s the heart of stereotype threat (Steele & Aronson, 1995). When a student who belongs to a group stereotyped as low-performing is asked to accomplish an academic task, he or she feels anxious, and his or her thoughts are preoccupied with concern about fulfilling the stereotype. Those thoughts occupy working memory, and the student is, indeed, less able to perform the task. The student doesn’t need to endorse the stereotype to feel its effects; worrying about fulfilling the stereotype that others hold is enough.
Geoff Cohen (Cohen et al, 2006) came up with a strategy to combat the problem. Students complete brief writing exercises that require them to think about personal values, things that are important in their lives. The subject might be friends, family, music, sports, politics, whatever. The idea is that stereotype threat situations narrow the sense of self—suddenly, an African-American boy sees himself as nothing but “kid who is not supposed to succeed in math.” The writing exercises are supposed to remind him that he’s much more than that, and so feel less threatened.
Some early experiments showed that several such writing exercises over the course of a school year (each taking just 20 minutes) had a substantial impact on academic achievement for 7th grade African-American and Hispanic students, presumably through the reduction of stereotype threat, and perhaps through other mechanisms.
Despite early successes, recent attempts to replicate the intervention are failing.
The most recent (Hanselman et al 2016) tested nearly 500 middle school students, and attempted to replicate closely the original methods used by Cohen. This graph summarizes effect sizes of the intervention on GPA from different studies. Note that the larger effect sizes come from studies with smaller Ns.
Error bars are 95% confidence intervals. Dotted line is weighted mean.
What are we to make of this?
Hanselman and colleagues suggest that the intervention may be sensitive to various modifying variables and presumably, it will be difficult identify and measure them all, casting doubt on the utility of the writing exercises in schools. (They were able to test the impact of some moderators that current theory would predict was important--they seemed not to matter.)
For my part, I’ve always found it difficult to understand why the intervention worked in the first place. Why would a writing intervention influence sense of self months later? Sense of self Is surely the product of many experiences over a long time, and there’s no reason to suspect that the stereotype-threat situations would trigger a memory of the writing exercise and thus influence sense of self at that moment.
The writing sounds sort of similar to one I blogged about recently—having college freshman watch videos of older student describing their experiences from freshman year, and then writing about the videos. The important difference is that the target was not students’ sense of self (which I’m suggesting is robust and hard to change) but their sense of what college is like, which, I would expect, would be less closely held and more easily influenced.
The door is not closed on the values affirmation intervention, but much work is to be done if it is to prove useful in schools.
Hanselman and colleagues suggest that the intervention may be sensitive to various modifying variables and presumably, it will be difficult identify and measure them all, casting doubt on the utility of the writing exercises in schools. (They were able to test the impact of some moderators that current theory would predict was important--they seemed not to matter.)
For my part, I’ve always found it difficult to understand why the intervention worked in the first place. Why would a writing intervention influence sense of self months later? Sense of self Is surely the product of many experiences over a long time, and there’s no reason to suspect that the stereotype-threat situations would trigger a memory of the writing exercise and thus influence sense of self at that moment.
The writing sounds sort of similar to one I blogged about recently—having college freshman watch videos of older student describing their experiences from freshman year, and then writing about the videos. The important difference is that the target was not students’ sense of self (which I’m suggesting is robust and hard to change) but their sense of what college is like, which, I would expect, would be less closely held and more easily influenced.
The door is not closed on the values affirmation intervention, but much work is to be done if it is to prove useful in schools.
