Scientific Discussion of the Week--The art of coaching: feedback, modelling and skill failure

[ryancostill]

I’ve been PMing Venom over the last couple of weeks to garner some insight into the area of motor learning and specifically the mechanisms through which motor learning optimal occurs. Over the course of this correspondence he was sent me some fascinating information and I thought it might be worth posting here for other members’ benefit. Feel free to add to the discussion!!

Let me set the scene. I have just begun coaching weightlifting(snatch and clean and jerk) to novices. These are individuals who have no weightlifting experience at all. A number of things I have been concerned with are the issues of feedback, demonstration and failing lifts. For example, how much feedback should the learners be given and can there be such thing as too much feedback/instruction/coaching.

Secondly, I am an ok weightlifter but have some problems with my own technique. Will demonstrating with my (imperfect) technique be of detriment to the learners?

Finally, is it beneficial or detrimental for a weightlifter to fail lifts? For example should we always keep the novice weightlifter lifting weight which we know he or she will be able to successfully lift or should we allow them to increase the weight to a point where they “fail” the lift?

Obviously I am compiling this information with the sport of weightlifting in mind and applying it to that sport. However, the information here is generally applicable to all sports/skills.

Below is edited information he has sent me along with some of my own thoughts and comments as well as other online and text sources I have come across.

All feedback and discussion would be warmly welcomed. How would you apply this information to body building or other sports?? Do you have any personal experiences relating to this discussion?

[ryancostill]

Firstly we will discuss feedback:

It is important to establish a number of definitions:

Intrinsic feedback -information received by the athlete as a direct result of producing a movement through the kinaesthetic senses - feelings from muscles, joints and balance. For example, if an athlete is performing a snatch as misses the lift, he may realise he has missed this lift because he has “felt” that he pulled the bar away from his body, or did not drop underneath the bar fast enough.
Extrinsic feedback - information not inherent in the movement itself but which improves intrinsic feedback. This is also known as AUGMENTED FEEDBACK . There are two main categories:

1. Knowledge of performance (KP) - information about the technique and performance. This can be provided verbally from the coach or visually via video. This enables athletes to establish a kinaesthetic reference for the correct movement. e.g. if we video taped a sprinters running action and watched it back with his coach highlighting technical issues.

2. Knowledge of results (KR) - information with regards the result of the athlete's performance. e.g. the sprinter's 100 metre time.

(adapted from http://www.brianmac.demon.co.uk (http://www.brianmac.demon.co.uk/infofb.htm)))

So how much feedback should we provide to the novices??

This is from: from, Specificity Part VI: The effect of Practice Distribution & Contextual Interference on Performance & Learning

[ QUOTE ]
The guidance hypothesis posits that a high relative frequency of augmented feedback during practice is detrimental to motor learning (Winstein, 1994). This is consistent with studies on the effect of feedback on retention (Sawyer, 2005). To elaborate, there are two types of feedback: intrinsic feedback and augmented feedback (Schmidt and Lee, 1999). Intrinsic feedback is inherent to the execution of movement and provided through various sensory channels. Augmented feedback is supplemental to intrinsic feedback. In this context, Knowledge of Results (KR), supplies augmented feedback about movement outcome. For instance, if a coach tells a basketball player who made a free throw shot, “You made it!”, this would be a form of KR. Evidence suggests that too much feedback appears to degrade learning. For instance, several studies show that reduced relative frequency (the percent of KR trials given) produces more learning (Lee, White, & Carnahan, 1990; Sparrow & Summers, 1992; Vander Linden, Cauraugh, & Greene, 1993.). Salmoni, Schmidt, and Walker (1984) propose that when KR is given on every trial, participants become reliant on the feedback and fail to process the information required to learn the task.

[/ QUOTE ]


[ QUOTE ]
Guadagnoli and Kohl (2001), had participants perform a force replication task in which they were given 100 % knowledge of results. Knowledge of Results refers to a practice variable in which participants are told the outcome of a movement upon its completion. There were two 100 % conditions. The first were given 100 % KR for each trial, while the second condition was first asked to estimate their error and then given 100 % KR. It was found that the 100 % KR condition with a hypothesized difference had greater performance in retention than the other 100 % KR condition. The theoretical rationale is the guidance hypothesis, which suggests that 100 % KR causes participants to become reliant on KR to make adjustments in their errors, instead of the underlying cognitive processes, therefore negating learning effects.

[/ QUOTE ]


[ QUOTE ]
Participants with less KR, however, must attempt to detect their own errors, increase cognitive effort, and go through the full learning process .

[/ QUOTE ]


In my opinion the same would hold true for knowledge of performance (KP). If we videotaped every lift an athlete performs in a clean and jerk session and had them watch each lift immediately upon completion pointing out flaws and suggesting improvements the lifters would become dependent on the feedback and fail to process the information required to learn the task. So we need to allow the learners more scope to identify their own mistakes and adjust their technique accordingly. This will create a more “active” learning environment and facilitate a more optimal learning experience.

Think of your own experience. In school (or college), in which environment have you learned more: a lecture where somebody is speaking to you and telling you lots of things but you are just sitting there listening, not asking questions, not answering questions just listening. Or a more hands on class where you can ask questions, experiment with equipment, try and solve problems yourself. The more active environment invariable creates a better learning environment.

So while augmented feedback through knowledge of performance (KP) and knowledge of results (KR) can be a useful tool for the beginner we should not create a learning environment where the learner becomes dependent on this feedback. After all what would then happen if you took away the feedback? During competition in many sports it is often impossible for the coach to provide constant feedback to the athlete, if the athlete has become wholly dependent on the augmented feedback, his/her performance may completely deteriorate without it.

Active learning, where the learner is forced to analyse his/her own technique, will facilitate more successful learning.

[ryancostill]

Next we will discuss Modelling or Demonstration.

My worry as a new coach is that if I demonstrate the lifts myself, as I have some problems with my own technique, will it actually be beneficial or detrimental to the novice lifters?

The following excerpt is from “Observational Learning: The Forgotten Psychological Method in Sport Psychology” by P. McCullagh and M. R. Weiss

A chapter from the book:
Van Raalte, J.L., & Brewer, B.W. (Eds.). Exploring sport and exercise psychology (2nd Ed.). Washington DC: American Psychological Association.


[ QUOTE ]
When choosing someone to demonstrate a physical skill, the first inclination is to select a model that can execute the behavior flawlessly (i.e., correct model). However, an alternative method is to expose observers to a model who is attempting to learn the skill and has not yet achieved exemplary performance (i.e., learning model). McCullagh and Caird (1990) compared these model types with college students attempting to learn a laboratory timing skill, and found that a learning model was more effective than watching someone execute the skill without errors, but only if observers were privy to feedback given to the model. Hebert and Landin (1994) extended this study to a complex sport skill (tennis forehand). Participants either received feedback about their own performance, watched a learning model and heard the instructor’s feedback to the model, received a combination of both treatments, or received no demonstrations or feedback. The combination of treatments led to best performance, while the learning model participants produced performance similar to that achieved by receiving feedback about one’s own skill execution. Collectively these findings imply that, if it is impossible for the coach or teacher to give individualized feedback, participants can learn from peers who are practicing the skill, especially if the coach is providing them with informational or corrective feedback.

[/ QUOTE ]



[ QUOTE ]
Furthermore, higher self-efficacy may be invoked when learners watch similar others persist and master skills for which they have experienced prior difficulty (McCullagh & Weiss, in press).

[/ QUOTE ]

[ QUOTE ]
For the novice, seeing elite athletes perform skills perfectly may have a negative impact on self-efficacy, motivation, and performance because of perceived dissimilarity between the observer and model.

[/ QUOTE ]

Essentially if novice’s are exposed to an expert model (demonstrator) they may be unable to associate with this performance. They may think “I will never reach that level!”, become disheartened and not carry the learning experience of watching the model into their own performance.

What appears to be a more successful teaching/learning method, especially if coaching large groups where individual attention is not always possible, is to utilise a demonstrator with imperfect technique BUT ensure that feedback is provided so that the onlooking novices know why that individual’s technique is imperfect.

Also a useful tool maybe to use a demonstrator who is just slightly ahead of the learners in the skill development process. This way the learners may be able to identify with the model and feel that with some effort they will be able to achieve such a performance level.

For complete novices, using an elite or expert demonstrator may actually be detrimental to the motor learning process.

[ryancostill]

Finally, should the learner be allowed to progress to a point where he/she is making mistakes? In the context of weightlifting should I allow my athletes to add on weight which may cause them to fail the lift or should I keep them in a comfort zone where they can continually make successful performance?

Again it is important to give a basic definition:

Schmidt's theory (1975) was based on the fact that actions are not stored rather we refer to abstract relationships or rules about movement. Schmidt's schema is based on the theory that that every time a movement is conducted four pieces of information are gathered:
· the initial conditions - starting point
· certain aspects of the motor action - how fast, how high
· the results of the action - success or failure
· the sensory consequences of the action - how it felt
Relationships between these items of information are used to construct a recall schema and a recognition schema. The Recall schema is based on initial conditions and the results and is used to generates a motor program to address a new goal. The recognition schema is based on sensory actions and the outcome.

(adapted from: www.brianmac.demon.co.uk (http://www.brianmac.demon.co.uk))


[ QUOTE ]
Sawyer (2005) suggests that in order to learn, the learner must make mistakes

[/ QUOTE ]

[ QUOTE ]
Hagman (1983) and Winstein, Pohl, and Lewthwaite (1994) revealed that guided (i.e., errorless) practice was detrimental to retention and transfer, especially if presented on every acquisition trial. This was also a proposed mechanism for the findings of Lee, Wishart, Cunningham, and Carnahan (1997) in their study on modeling. Four mechanisms may facilitate this process. The first two being that errors may increase cognitive effort and elaborative processing, enhancing learning (Sherwood & Lee, 2003; Shea and Morgan, 1979). Third, it may help develop a schema, as evidence suggests that there are positive benefits from correct and incorrect movements for schema learning, which is based on a relationship among all stored elements (Schmidt and Lee, 1999). And fourth, 100% relative feedback impedes the intrinsic processes, such as problem solving, in the inter-trial interval, known to be important for learning (Sawyer, 2005; Bjork, 1988; Landauer & Bjork, 1978; Schmidt, 1991).

[/ QUOTE ]

The take home point? Allow learners to fail. They will learn from their mistakes. With respect to weightlifting, one important consideration to be made is that learners can fail safely (ie drop the bar safely and effectively).

[book]

Interesting stuff. Best of luck with the training as well

[President Wilson]

What an awesome post! Its two in the morning, but I have a lot of things I want to say on this one tomorrow!

Thanks for posting your insight Ryan!

[aquariyums]

I've played football from junior high to college. I've coached at the junior high and high school level. I've assisted the college football team in the gym.............just thoguht I'd tell you. LOL

[Venom]

Ryan,

I am simply amazed at how sharp you are. Every concept you discussed was absolutely perfectly portrayed, and very articulate.

[ QUOTE ]

Think of your own experience. In school (or college), in which environment have you learned more: a lecture where somebody is speaking to you and telling you lots of things but you are just sitting there listening, not asking questions, not answering questions just listening. Or a more hands on class where you can ask questions, experiment with equipment, try and solve problems yourself. The more active environment invariable creates a better learning environment.

[/ QUOTE ]

I never considered that this would increase active learning, great point! The president and I submitted an article that may get published on the importance of active learning. It is absolutely invaluable. Excellent point.

I am going to absorb all that you discussed on feedback. This made me think of many more ideas. Prez has actually been researching feedback intensively for years now—especially the last 3 months. He will definitely write an article for JHR in the future. I look forward to hearing his comments on this. I am going to study the rest tonight. Thanks so much, bro!

[President Wilson]

[ QUOTE ]
e.g. if we video taped a sprinters running action and watched it back with his coach highlighting technical issues.

[/ QUOTE ]

This is critical, especially the part about providing cues for the athlete. Kernoodle and Carlton (1992) had participants either watch a video or watch a video with cues on specific aspects to look for, and found that the video with cues was far greater for learning. This is because a straight up video appears to have too great an information load to process

[ QUOTE ]
In my opinion the same would hold true for knowledge of performance (KP).

[/ QUOTE ]

Absolutely. A recent study found the same results in KP for a golf shot

[ QUOTE ]
If we videotaped every lift an athlete performs in a clean and jerk session and had them watch each lift immediately upon completion pointing out flaws and suggesting improvements the lifters would become dependent on the feedback and fail to process the information required to learn the task.

[/ QUOTE ]

Yes. It may also degrade learning because the modeling between each set will mainatin the current action plan in working memory. The action plan would consist of the motor program and the paramaters selected.

It is suggested that an individual will enhance learning if they have to initiate an action plan before each trial. This would be why random practice works so well, it causes a participant to abandon their action plan and recontruct a new one on each trial.

[ QUOTE ]
The more active environment invariable creates a better learning environment.

[/ QUOTE ]

Amen! Gabriel and I are submitting a paper on this concept as a paradigmn for success for kinesiology students. William James (1890) suggested that “A curious peculiarity of our memory is that things are impressed better by active than by passive repetition (p. 686).”

Evidence since then has only enhanced this statement

I have some other theories as well for constant KR.

1. The maladaptive correction Hypothesis - if KR is given on each trial, it may be given at a time in which the KR provides information about an error that is too small for the participants motor system to be able to correct for. They then attempt to make the correction anyway, and end out causing another error.

2. Enhanced variability - The aquisition of skill appears to be
a factor of

A. Gaining adaptability - here adjustment of the pattern is important

B. Stabilization of the pattern. - Each time KR is delivered the participant may make an adjustment, this while necessary enhances variability and makes it more difficult to stabilize a movement pattern. So reduced relative frequency can serve as a method to stabilize the pattern (i.e. they make less corrections).

[ryancostill]

[ QUOTE ]
Prez has actually been researching feedback intensively for years now—especially the last 3 months. He will definitely write an article for JHR in the future.

[/ QUOTE ]

It would be fascinating to see an article in JHR on this topic. If you need a proof reader let me know! /forum/images/graemlins/grin.gif

[Ratcat]

G'day. This is very interesting. In our dance classes the teacher demonstrates a move and everyone has to copy that move. Some times the move is complex and she only shows us a few times and then leaves everyone to their own interpretation. It makes for an interesting ten minutes while we all sort it out in our heads. Then adapt to fit in with all the other dancers. The attitude of the people is interesting. Some find it all to hard, some find it silly and want to complain till they get it. Some just accept the mistakes and just get on with it till the move finaly works. Sometimes you'll go so well for a while and then for no reason you'll make mistake after mistake and then a big effort is required to get going again.
The teacher is interesting as she has to deal (or not deal) with everyones personalities and ideas of what is going on. She does it well.
Cheers. /forum/images/graemlins/smile.gif

[Adam Knowlden]

Nice discussion guys! I'm learning a lot! /forum/images/graemlins/cool.gif

[ryancostill]

[ QUOTE ]
[ QUOTE ]
e.g. if we video taped a sprinters running action and watched it back with his coach highlighting technical issues.

[/ QUOTE ]

This is critical, especially the part about providing cues for the athlete. Kernoodle and Carlton (1992) had participants either watch a video or watch a video with cues on specific aspects to look for, and found that the video with cues was far greater for learning. This is because a straight up video appears to have too great an information load to process

[/ QUOTE ]

I agree that feedback + video is far more effective than video alone. This method of technical analysis can be a very useful tool. However would you agree that it should not be used too frequently so that the athlete becomes engaged in more active problem solving themselves and as to not develop a dependence on such tools?

[William Ustav]

Okay, I haven't read any of the comments so far, so I'll just give my input and then try to comment on the other's. /forum/images/graemlins/smile.gif

[ QUOTE ]
Secondly, I am an ok weightlifter but have some problems with my own technique. Will demonstrating with my (imperfect) technique be of detriment to the learners?

[/ QUOTE ]

That depends on how imperfect your technique is - if we're not talking minimal differences, then perhaps it is better to not demonstrate. Most important though is that you know where your technique is imperfect, and then highlight these things in your own feedback to the students.

[ QUOTE ]
Finally, is it beneficial or detrimental for a weightlifter to fail lifts? For example should we always keep the novice weightlifter lifting weight which we know he or she will be able to successfully lift or should we allow them to increase the weight to a point where they “fail” the lift?

[/ QUOTE ]

An airplane always takes off against the wind /forum/images/graemlins/smile.gif And I'm a firm believer of using that in philosophy in training as well. Carrying a student to always complete lifts will not let him see the failure (when it will happen) from a good perspective. He has to get accustomed to being able to fail, but still not wanting to fail. I believe it would be best to always let the student push their limits, and if they fail, they get a chance to learn how to pick themselves up and strengthen themselves again.

[ QUOTE ]
1. Knowledge of performance (KP) - information about the technique and performance. This can be provided verbally from the coach or visually via video. This enables athletes to establish a kinaesthetic reference for the correct movement. e.g. if we video taped a sprinters running action and watched it back with his coach highlighting technical issues.

2. Knowledge of results (KR) - information with regards the result of the athlete's performance. e.g. the sprinter's 100 metre time.

(adapted from http://www.brianmac.demon.co.uk)

So how much feedback should we provide to the novices??

[/ QUOTE ]

A combination is definitely necessary, however feedback regarding technique and performance is really important. When you learn something new it's like a puzzle, and every time you get feedback (that you understand) you get a new piece to the puzzle. Eventually, you will have most of the pieces (if not all) and you'll understand the full picture. Without the pieces of the puzzle it's impossible to see the picture.

[ QUOTE ]
In my opinion the same would hold true for knowledge of performance (KP). If we videotaped every lift an athlete performs in a clean and jerk session and had them watch each lift immediately upon completion pointing out flaws and suggesting improvements the lifters would become dependent on the feedback and fail to process the information required to learn the task. So we need to allow the learners more scope to identify their own mistakes and adjust their technique accordingly. This will create a more “active” learning environment and facilitate a more optimal learning experience.

[/ QUOTE ]

I would opt for the instructor to know which flaws to stress. At first, when the athlete is completely new to the lift, some things may be important, while others aren't necessary to comment on. If we take the puzzle example again, it's always easier to start with corners and edges, and then you can continue building on that.

Also, it is imperative that the student understands the critique.

[ QUOTE ]
Think of your own experience. In school (or college), in which environment have you learned more: a lecture where somebody is speaking to you and telling you lots of things but you are just sitting there listening, not asking questions, not answering questions just listening. Or a more hands on class where you can ask questions, experiment with equipment, try and solve problems yourself. The more active environment invariable creates a better learning environment.

[/ QUOTE ]

I think there are different kinds of people that respond to different kinds of teaching methods. Some are listeners and arguably learn the best from hearing things. Others are seers, and really have to see everything with their own eyes to register it. And others can be touchers, where they need to feel everything physically. If possible, one should try to have a mix of these three methods in instruction.

[ QUOTE ]
Active learning, where the learner is forced to analyse his/her own technique, will facilitate more successful learning.

[/ QUOTE ]

Definitely - but the learner needs to have the tools to work with at first. Often the student is blind to his own faults in the beginning.

[ QUOTE ]
Essentially if novice’s are exposed to an expert model (demonstrator) they may be unable to associate with this performance. They may think “I will never reach that level!”, become disheartened and not carry the learning experience of watching the model into their own performance.

What appears to be a more successful teaching/learning method, especially if coaching large groups where individual attention is not always possible, is to utilise a demonstrator with imperfect technique BUT ensure that feedback is provided so that the onlooking novices know why that individual’s technique is imperfect.

[/ QUOTE ]

I must disagree there. Or no, I agree with that IF there is an imperfect technique being shown, feedback MUST be provided as to why it's imperfect. BUT, I don't think it's better than to have a perfect technique shown. As a student, I feel one should know there is a level one can attain where there is perfect technique, and one should see what it looks like, and get to know why it is perfect. The imperfect parts can be shown from what the student does (wrong). I don't know if I made my point accurately here...

My 2 cents /forum/images/graemlins/smile.gif

[William Ustav]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
e.g. if we video taped a sprinters running action and watched it back with his coach highlighting technical issues.

[/ QUOTE ]

This is critical, especially the part about providing cues for the athlete. Kernoodle and Carlton (1992) had participants either watch a video or watch a video with cues on specific aspects to look for, and found that the video with cues was far greater for learning. This is because a straight up video appears to have too great an information load to process

[/ QUOTE ]

I agree that feedback + video is far more effective than video alone. This method of technical analysis can be a very useful tool. However would you agree that it should not be used too frequently so that the athlete becomes engaged in more active problem solving themselves and as to not develop a dependence on such tools?

[/ QUOTE ]

The question that arises in my head here, is does the student have the tools necessary to be able to solve the problems themselves?

[Venom]

Thanks for your incite, Will! /forum/images/graemlins/smile.gif

[ QUOTE ]


I agree that feedback + video is far more effective than video alone. This method of technical analysis can be a very useful tool. However would you agree that it should not be used too frequently so that the athlete becomes engaged in more active problem solving themselves and as to not develop a dependence on such tools?


[/ QUOTE ]

Yes, that is definitely the case, Ryan. You want to limit the KR/KP as with anything else.

[ QUOTE ]

I would opt for the instructor to know which flaws to stress. At first, when the athlete is completely new to the lift, some things may be important, while others aren't necessary to comment on. If we take the puzzle example again, it's always easier to start with corners and edges, and then you can continue building on that.

[/ QUOTE ]

This is excellent advice, Will.

It is commonly advised to critique starting with big errors, then later going to small errors. This is to avoid information overload.

[ QUOTE ]


Definitely - but the learner needs to have the tools to work with at first. Often the student is blind to his own faults in the beginning.

[/ QUOTE ]

You again make a great point. Evidence suggests you want to fade the feedback. So novice athletes should receive more feedback than advanced athletes. But the more they can learn through intrinsic processes, the better. So feedback should be kept at bay, still.

[ QUOTE ]


I must disagree there. Or no, I agree with that IF there is an imperfect technique being shown, feedback MUST be provided as to why it's imperfect. BUT, I don't think it's better than to have a perfect technique shown. As a student, I feel one should know there is a level one can attain where there is perfect technique, and one should see what it looks like, and get to know why it is perfect. The imperfect parts can be shown from what the student does (wrong). I don't know if I made my point accurately here...

[/ QUOTE ]

Wow, that is an interesting point you bring up.

Well, studies are very clear that a learning model (one that makes errors) is superior to a correct model (one that does not make errors). However, I am not aware of a study that combines both types of models. I’ll ask Dr. McCullough about that. /forum/images/graemlins/grin.gif


Good post, William. /forum/images/graemlins/smile.gif

[ryancostill]

[ QUOTE ]

[ QUOTE ]

I would opt for the instructor to know which flaws to stress. At first, when the athlete is completely new to the lift, some things may be important, while others aren't necessary to comment on. If we take the puzzle example again, it's always easier to start with corners and edges, and then you can continue building on that.

[/ QUOTE ]

This is excellent advice, Will.

It is commonly advised to critique starting with big errors, then later going to small errors. This is to avoid information overload.


[/ QUOTE ]

Excellent insights. It should also be pointed out that correcting the larger errors may indirectly correct the smaller errors. Also there can often be a hierarchy of errors in performance where one error will lead to another will lead to another.

For example, in weightlifting lets say our learner has three major errors: he is rounding his back (not maintaining a strong arch), he is keeping his head down and he is not reaching full extension.

The coach needs to decide which of these errors is most important in the hierarchy. Sometimes correcting just one of the errors can indirectly cause correction of the ther errors.

Here we could stress to the learner the importance of keeping the head up during the lift. If they correct this problem, it will induce a more upright upper body position which may place the back in a stronger arch position. This in turn will give the learner a stronger base of support, a more upright position through the pull of the lifts and resultantly help them fully extend in the pull phases of the lifts.

So in this example (and I myself have seen this happen in a practical setting) we have only instructed the learner in correction of one error, but this correction has actually aided in the correction of three major erros!

[ryancostill]

I should also mention that this identification of a hierarchy of errors is what makes some coaches "great coaches". A great coach will have an ability to identify one key error, correction of which will lead to other errors being indirectly corrected.

[William Ustav]

[ QUOTE ]
I should also mention that this identification of a hierarchy of errors is what makes some coaches "great coaches". A great coach will have an ability to identify one key error, correction of which will lead to other errors being indirectly corrected.

[/ QUOTE ]

Very true!