GUEST POST: When feedback is forgettable
By Robert Nash & Naomi Winstone
Which type of feedback do your students prefer to receive: past-oriented feedback about how they performed, or future-oriented feedback about how they could improve? Education studies, and our own informal surveys, consistently tell us that students prefer future-oriented feedback, and judge it as more beneficial for their learning (1). Likewise, a quick web-search for educational resources using the buzzword “feed-forward” illustrates that many teachers and education researchers share the same views.
But here’s another question: Do your students actually remember any of your feed-forward? This question has received scant attention in prior research, but it seems an important one to answer, because people who quickly forget helpful advice might as a result be unequipped to improve their skills and performance. In a paper recently published at the Journal of Experimental Psychology: Learning, Memory, & Cognition, we tested the prediction that people should be especially good at remembering future-oriented advice, relative to past-oriented advice (2). And the outcome was… surprising.
Our participants signed up for a two-part study supposedly investigating persuasive writing skills. The first part of the study was straightforward – participants simply chose some controversial topics from a list (e.g., “Should Valentine’s Day be abolished?”), and they wrote short persuasive arguments about each topic. We told participants that when they returned a couple of days later, we would provide personalized written feedback on their writing, which they should take on board in a second persuasive writing task.
This instruction was only partly true. When participants returned, they did receive feedback, but in fact everyone received the same feedback, with each individual comment written in one of two different styles. Half of the comments in each participant’s written feedback were evaluative: past-oriented advice that focused on aspects of how the participant had supposedly performed (e.g., some people were told “you sometimes said in multiple sentences what you could potentially have said in just one”). The other half were directive, appearing to offer suggestions for future assignments (e.g., other people were told “avoid saying in multiple sentences what you could potentially say in just one”). As these examples illustrate, the past- and future-oriented feedback comments were virtually identical except for minor tweaks to the wording.
There was also a surprise in our procedure. Instead of us next asking participants to complete a second persuasive writing task, we actually gave them a memory test, asking them to write out as much of the feedback as they could remember. We then coded which of the original feedback comments each person had remembered. As we explained above, our prediction was that people would be more likely to remember those feedback comments that had been presented in a directive style.
What we actually found, though, was the exact opposite. In fact, across six experiments, participants were a huge 46% more likely to recall evaluative comments compared with directive comments. Over and over in each experiment—some of them using different feedback materials and different participant populations—we got the same result. And even when they did remember directive feedback, more often than not, they remembered it as evaluative anyway.
Why might future-oriented feedback be so forgettable? To cut a long story short, we still don’t know! But several of the more intuitive post-hoc explanations are not supported by our data. For example, the memory advantage for evaluative feedback doesn’t seem to be a consequence of evaluative comments feeling more concrete or self-relevant; in fact, we find the same bias even when participants try to remember another person’s feedback. Nor is the memory advantage apparently driven by the fact that evaluative comments were seen as more negative. At this time we are still testing out other possible explanations—with the generous financial support of the Leverhulme Trust—to better understand what is going on.
While we try to figure out this puzzle, what do our findings mean for educators? Should we all forget about telling students how/what to improve, and instead give them only evaluative feedback? No. Our data do not warrant that kind of response. After all, we haven’t yet any evidence that this memory bias carries over to students’ behavior. That is to say, we don’t know whether people would be more likely to actually implement evaluative feedback in their next assignment (although we are working on answering that question right now).
Moreover, directive feedback could still be more beneficial for learning than is evaluative feedback, even if the latter is easier to remember. If so, then we should focus our efforts on improving the likelihood of feedback “sticking”, rather than on cutting out particular kinds of feedback entirely. Interventions that encourage students to refer back to their feedback more frequently, or to reflect on it more thoroughly, might perhaps be straightforward ways of achieving this goal.
In fact, we would argue that any feedback intervention is doomed to failure if its focus is on making cynical, quick-fix tweaks to the wording of feedback comments, rather than on creating opportunities for dialogue and for implementing the advice (3). In this vein, creating true “feed-forward” clearly involves more than merely re-writing our criticisms in the future tense. Yet we have heard many educators describe exactly this kind of verbal reframing as an intervention they have followed—and been advised to follow—for the purpose of making their feedback more effective. It is therefore striking to consider that by doing so, with all good intentions, educators’ efforts might ultimately make their feedback more forgettable.
We have so many questions still to answer about these surprising memory effects, but at least one thing is crystal clear: Just because we give students feed-forward, doesn’t mean they hear feed-forward.
(1) Nash, R. A., Winstone, N. E., Gregory, S. E. A., & Papps, E. (in press). A memory advantage for past-oriented over future-oriented performance feedback. Journal of Experimental Psychology: Learning, Memory & Cognition. http://psycnet.apa.org/record/2018-09034-001
(2) Winstone, N. E., Nash, R. A., Rowntree, J., & Menezes, R. (2016). What do students want most from written feedback information? Exploring necessities and luxuries using a budgeting methodology. Assessment & Evaluation in Higher Education, 41, 1237-1253.
(3) Nash, R. A., & Winstone, N. E. (2017). Responsibility-sharing in the giving and receiving of assessment feedback. Frontiers in Psychology, 8, 1519.