It is an example I sometimes use in my lectures: suppose you give perfect feedback. Will your student have learned? Dylan Wiliam described eight possible reactions to feedback, but only two of them lead to more learning. The simplest situation in which this does not happen is when the pupil or student ignores the feedback. But even if they do not, they still have to understand, evaluate, and then act on the feedback. All this may sound rather self-evident, but precisely that final step—what people actually do with received feedback, has been remarkably little researched. And that certainly applies to peer feedback.
By now, we know quite a lot about peer feedback, but much of the research presents a fairly simple dichotomy regarding its use: feedback is processed or not processed. A new study by Xingshi Gao and colleagues, published in Contemporary Educational Psychology, attempts to add more nuance to this. The researchers distinguish four possible reactions:
- simply adopt feedback (accept),
- build on it (elaborate),
- solve the identified problem in another way (modify),
- do nothing with it ( no uptake).
What exactly determines the response a fellow student gives?
What did the students do with the feedback?
(Only…) 74 Master’s students in Nutritional Sciences at a Dutch university participated in the study. They wrote an argumentative essay of 600 to 800 words and subsequently provided feedback on two fellow students’ texts. Afterwards, they were given the opportunity to revise their own essay. This allowed the researchers to compare the received feedback with the original and the revised text.
In total, the researchers ultimately analyzed 476 usable feedback comments relating to substantive matters, such as the structure of an essay or the quality of an argument. Feedback regarding, for example, grammar and word choice was not included in further analyses.
Sobering: students did absolutely nothing with 40 percent of the feedback received. Only 8 percent was simply adopted. In 10 percent of cases, students recognized the problem but chose a different solution from the feedback provider. The largest category, however, was elaboration, accounting for 42 percent. Students did not simply adopt the feedback but continued working with it themselves. For example, they added extra arguments, examples, or evidence.
But what feedback is used then?
The researchers also coded the characteristics of the feedback itself. In doing so, they distinguished between cognitive characteristics, such as identifying a problem, explaining why something is a problem, or suggesting a solution, and affective characteristics, such as softening criticism with compliments.
Most of the feedback was quite substantive. In 81 percent, the problem was identified; in 60 percent, there was an explanation; in 39 percent, a general suggestion; and in 36 percent, a concrete solution. Remarkably, but also not illogically, I think, the researchers found not a single example of what they coded as metacognitive feedback. Subsequently, they investigated which characteristics were associated with students’ responses to the feedback.
Two findings stand out in this regard. When feedback included a concrete solution, the likelihood that a student simply adopted it was approximately three times higher. When the feedback clearly identified the problem, the likelihood that students did absolutely nothing with it was approximately halved.
What did they not find? Explaining why something was a problem did not significantly predict how students subsequently used the feedback. The same applies to the affective characteristics of feedback. And none of the characteristics studied significantly predicted whether students would act on the feedback or develop a different solution on their own.
Doing nothing can also mean something.
I don’t necessarily find those results negative, on the contrary. Not even 40% of the feedback is ignored. Does this not necessarily mean a bad choice? We don’t know at all if that is really what is happening.
Suppose a fellow student tells you to discard an argument; you review that argument and subsequently decide that your fellow student is wrong. According to the structure of this study, you have not acted on the feedback. However, cognitively, quite a lot may have happened in the meantime.
The researchers themselves explicitly point this out. No uptake can sometimes result from an active evaluation of the feedback received. Unfortunately, they were unable to investigate why students did or did not process the feedback. Therefore, it’s impossible to tell what portion of that 40 percent stems from misunderstanding, lack of motivation, or trust in the feedback provider, and what portion from a reasoned decision to disregard the feedback.
Limitations
It is a fine study, but also relatively small and quite specific in its target group. It involves 74 graduate students from a single programme at a single Dutch university, 90 percent of whom were female. They received one round of peer feedback on one specific task. We must therefore be careful not to simply translate this to, for example, twelve-year-olds giving each other feedback on a math task. Moreover, it is correlational research. Nevertheless, the study provides a valuable addition to how we can think about feedback. The question is not just whether pupils or students receive feedback. Even the question of whether they use the feedback is perhaps still too simplistic. We also need to ask what they actually do with it. And that brings me back to the example I started this post with: even perfect feedback does not guarantee learning.