New study: Feedback can undermine learning for some students

Here is an interesting release from Vanderbilt in a new study that reaffirms for me the complexities of the teaching profession:

From Vanderbilt:

A child solves a math problem and gets the wrong answer. A teacher or parent tells her, “Good try, but your answer is incorrect.”

This kind of instructional input is helpful to the child’s learning, right? Not necessarily.

A new Vanderbilt study finds that sometimes providing verbal feedback (positive or negative) actually causes more harm than good. Developing ways to improve problem solving and early mathematics understanding is the research focus of Emily Fyfe, a doctoral student at Vanderbilt’s Peabody College of education and human development.

A new study suggests kids may do better solving math on their own with a lot of teacher feedback.  (AP Photo/The Winchester Star, Jeff Taylor)

A new study suggests kids may do better solving math on their own without a lot of teacher feedback. (AP Photo/The Winchester Star, Jeff Taylor)

A new paper published in the Journal of Educational Psychology by Fyfe and Bethany Rittle-Johnson, associate professor in the Department of Psychology and Human Development at Peabody, continues their examination of how feedback provided to elementary school students affects their mathematics learning.

Fyfe’s recent study included 108 children from second- and third-grade classrooms in Nashville-area schools. One group of students received instruction on how to solve the problems. The other group did not receive instruction; they were true novices with low prior knowledge of the problems.

All of the students were then asked to solve a set of math problems. Some students were assigned to the no-feedback condition, meaning they solved all the problems without any input from the tutor. The other students were assigned to the feedback condition — after each problem, the tutor told them if their answer was right or wrong.

“Most people assume that giving children feedback after they solve a math problem is helpful because it allows them to see their errors and adjust their approach,” Fyfe said. “But we found that feedback only had positive effects for children who didn’t know much about the problems. For children who were already taught how to solve the problems, giving them feedback during problem solving actually led to lower performance on subsequent math problems than giving them no feedback at all.”

One possible reason for the negative effects is that the feedback may have distracted the higher-knowledge learners by drawing attention to their self-image and performance. The children who already knew how to solve the problems likely had some expectation of performing well, she said. Feedback may have fixated their attention on whether they were right or wrong, and how getting a wrong answer reflected on their self-image, rather than on how to approach solving the problems.

“This study shows that children with really low prior knowledge of the task benefited from feedback,” Fyfe said. “But, the higher-knowledge children performed best when they were allowed to practice a set of relevant problems at their own pace without feedback.”

The study results challenge the notion that feedback is always a good thing. It also reinforces that small changes to teaching approaches may significantly improve a child’s learning experience.

“More and more research shows that micro-level changes to instructional input can affect how children learn and understand mathematics,” Fyfe said. “This means that small, purposeful changes can help shape children’s understanding and performance in a positive way.”

Another takeaway, Fyfe said is that there is value in allowing a child to have time for unguided problem solving.

“Although our natural inclination may be to step in and guide children through the difficult tasks, it may actually undermine their efforts and deprive them of the opportunity to make sense of the math problem on their own,” she said. “Under some conditions, we may need to refrain from ‘rescuing’ children by providing them with feedback, and instead let them struggle, engage and learn on their own.”

Read Fyfe and Rittle-Johnson’s paper, “Feedback both helps and hinders learning: The causal role of prior knowledge, in the Journal of Educational Psychology. Fyfe is supported by a graduate research fellowship from the National Science Foundation.

 

 

Reader Comments 0

26 comments
living-in-outdated-ed
living-in-outdated-ed

@maureen, it looks like you have to purchase the PDF - not a free study based on your link - fyi.

living-in-outdated-ed
living-in-outdated-ed

I plan on reviewing this study in more detail.  I believe the headline is misleading and can lead to misinterpretation.


Isn't this kind of obvious?  Of course, feedback is critical to a child's learning.  Math is a process - the answer is less important than how you got to the answer.   It's building the foundation of a house.  Every concept builds on the foundation.  If the foundation is weak, the house will crumple.  But feedback in an of itself is a misnomer.   It's HOW you deliver feedback.    If you take a negative approach, then you will damage a child's confidence.   So on the surface, is this research really surprising??


Teachers need to be better trained as communicators - positive reinforcement.  I have seen how a teacher can take a child who loved math and make them hate it.   And that has to be an educator's (and parent's ) worst nightmare.

gactzn2
gactzn2

This research supports the development of "grit" in students.  It is not emphasized as much anymore because learning has become to "fun" oriented.  Part of the learning experience requires that they put some skin in the game.  Struggling with a problem for a little while is part of the learning process.

Lee_CPA2
Lee_CPA2

Ahhhh, so that's why some teachers just hand out worksheets and then go sit at their desk and surf the internet.  All this time, I through they were just goofing off.   Little did I realize they were employing a RESEARCHED BASED teaching methodology.   ROFLMAO

AvgGeorgian
AvgGeorgian

@Lee_CPA2 Ahhhh, now I am understanding, same reason CPAs bill hours while surfing net, drinking coffee, eating lunch etc., 

popcornular
popcornular

Paralysis by Analysis. Most educators are by now completely unable to move anything but their tongues. Stop yapping and DO something.

Cover your classroom, all 360 degrees, with whiteboards/chalkboards. Send a bunch of kids up to work problems. Scan their work and use a laser to point out/discuss the trouble spots. Erase, Repeat. Erase, Repeat. Over and over, until the darlings can solve as fast as they can write. It works. Kids love being out of their seats actually doing something. Please, teachers, stop talking, stop ‘researching’, and stop thinking. Put on your Nike shoes and just do it. Erase, Repeat. 

AvgGeorgian
AvgGeorgian

@popcornular Ahhhhh, Pop - Please be so good as to become principal of worst school in state and bring your mightiness to bear on problems, thereby propelling all children to Gwinett STEM Academy levels. Please!!!!!

Wascatlady
Wascatlady

I wonder if the WAY feedback is done may be the salient point.

Wascatlady
Wascatlady

For example, have the student explain to YOU how they worked the problem, so you can identify where the error (whether in plain old computation or in the math concept) lies.

Wascatlady
Wascatlady

And when you are teaching and scaffolding, use "teacher talk" to demonstrate to the student the questions and processes they should be asking themselves as they work through the problem.  This also works well in reading comprehension.


One cannot stress, also, the important role of experiential background, another reason many of our poorest students struggle.  So much that is part of our examples, or reading selections, presupposes middle class, suburban experiences of our children.  Easy example:  We were reading something that took place in an alley.  Now, being country kids, none of them knew what an alley was; there are none even in the small town.  So a lot of the story did not make much sense.


Even my own children were not immune to this. They knew what an alley was--there was one behind their grandparents' house.  But at age 5, my son came running in to tell his Pawpaw that some men were out there in the alley, STEALING HIS GARBAGE!  Looking to forestall this with his younger sister, when she was 4, I explained that some men would be coming by in a big truck to pick up Pawpaw's garbage.  Wide-eyed, she looked at me and asked, "But WHY?"  In the country we don't have garbage service, and it had not occurred to me that I would need to explain it.


How much we assume that others have the same experiences that we do!

class80olddog
class80olddog

I wonder if they have done a study to verify the positive or negative effects of studies.

For example "studies have shown that kids who are retained have a higher dropout rate"

(Reminds me of that government statistical study that found that 100% of all people who drink water die!)

For the record, kids who fail courses, yes, are more prone to drop out.

The secondary ramifications of the blind acceptance of these studies leads to children in classrooms where they are hopelessly behind, fall further behind, and then drop out! (Or,worse, they get promoted all through school, get a diploma, can't read or do simple math, get hired, and the employer no longer has any faith in a HS diploma)

class80olddog
class80olddog

I have read your articles, and while I agree with the concept, I believe the execution would be extremely difficult

MaryElizabethSings
MaryElizabethSings

@class80olddog


From my experience of having implemented well, for a decade, continuous progress throughout a school, grades 1 - 7, it would not be difficult.  In fact, once teachers understand how it is accomplished logistically, it is not difficult, at all, to accomplish.

AvgGeorgian
AvgGeorgian

@class80olddog Please post your experience in managing a work group made up of random people from the neighborhoods surrounding your workplace. Many do not want to be there. Your rules - you cannot fire them, they must remain part of your team, and you are evaluated on how well they meet performance standards for regular, motivated, employees. Oh yeah - you expect them to do 1-3 hours of practice at home to stay up to date.  

redweather
redweather

"One possible reason for the negative effects is that the feedback may have distracted the higher-knowledge learners by drawing attention to their self-image and performance. The children who already knew how to solve the problems likely had some expectation of performing well, she said. Feedback may have fixated their attention on whether they were right or wrong, and how getting a wrong answer reflected on their self-image, rather than on how to approach solving the problems."


Wouldn't the "one possible reason for the negative effects" noted above apply to all subjects?  Isn't problem solving a feature of all learning? 

hssped
hssped

I like the  scaffolding/modeling (whatever it is called) method  where I do it, step-by-step, three or four times and then we do it together three or four times and then the student does it alone.  And I give feedback, but my feedback is way more than "right" or "wrong."  I like to help the kid find where they went off the trail.  I can totally see how ONLY telling a kid the answer is wrong would be frustrating.


However, if the Vanderbilt study is correct.....maybe I should try the no feedback method?  But what if they are wrong and I waste precious time????


I will start my 22nd year of teaching  in the fall.  I tend to go with tried and true, but I'm game to try something new.  This, however, doesn't sound like it will work.   I work with mostly LD students. 

Wascatlady
Wascatlady

@hssped No, if the authors are merely saying "correct" or "incorrect", well, that really isn't much feedback is it?  What are these researchers thinking?!  Maybe they should swat the kid with a ruler instead, and see how helpful THAT is!

MaryElizabethSings
MaryElizabethSings

Very interesting.  I can believe this is true.  Every student's mind will process somewhat differently.  That is why I always incorporated questioning by students within my lectures, not after I had finished.  The student must learn by being true to his mind's own mental processes, not the mental processes of the teacher, per se.


Here is a reading truism which, I think, will help readers to understand the depth of my point, above.  There is a saying in the reading field that the reader understands what the writer has written only to the extent of the experiences that the reader, himself, brings to the content to be absorbed on the written page before he even begins reading.  For example, if the reader has a degree in psychology, he will understand a passage from "The Brother's Karamazov" with greater insight than if he has little understanding of human nature to begin with.  


One other point:  It is not simply saying to the student, "You have the right or wrong answer," that matters as much as explaining to the student(s) HOW his/her mental processes were malfunctioning in getting the wrong answer and why the answer chosen was incorrect.  For example, in my SAT class, in private at home, I would itemize the particular questions on the practice SAT test which many of the students in my class missed.  Then, I would reteach how to get that particular question's correct answer, by a process of elimination.  If I noticed that most of those who missed that SAT question answer B, not the correct answer of D, then we would analyze why that was so as a class.  We would go back to the passage itself and look for key word, such as pronoun referents which may have been misleading to the students, as well as other types of comprehension skills, tying the ideas of sentence #2 in the paragraph to sentence #5 in that paragraph, or comparing and contrasting the ideas within paragraph 3 with the ideas in paragraph 4 to understand why we had to eliminate certain question items, which can back the students into the correct answer.  Thinking must be detailed and complex, and having an expansive experiential background brought to the subject matter helps any student perform better.

MiltonMan
MiltonMan

"A child solves a math problem and gets the wrong answer. A teacher or parent tells her, “Good try, but your answer is incorrect.”"


The very first sentence is factually incorrect.  A student does not "solve" a math problem if the answer is incorrect.  The student merely "worked on" the math problem.  A math problem is not solved unless the answer is correct.

AvgGeorgian
AvgGeorgian

@MiltonMan That is what you got from the article? Interesting. Please take time to look at various usages of solve and solution when it comes to problems, It is perceived as more of a process as in" solve for x". Can there be a wrong solution? Perhaps seeing things in less of a black and white manner and allowing a little gray could be the solution to your problem.

DawgDadII
DawgDadII

@AvgGeorgian @MiltonMan  If we are educating people who will cut my paycheck I'd prefer someone who sees things very "black and white", who can learn a process and apply it consistently and reliably to get the "correct" result. If we are educating someone to erase answers on tests, I suppose it doesn't matter so much, though we hope the judge and jury can ultimately get it right.


Not being in the profession of education my initial reaction to this article is the study seems to make the case for tracking kids within a grade level, as was the case when I was in school. The nerdier kids wind up in the classrooms where they read tomes of imperceptible Brit Lit and learn to calculate and rank their GPA in multiple planes of time-space, the average kids wind up in the classrooms where they snap and share selfies and tweet about where to hook up after class.