THE EFFECTS OF AUGMENTED FEEDBACK ON SKILL LEARNING

THE EFFECTS OF
AUGMENTED FEEDBACK
ON SKILL LEARNING
1
THIS CHAPTER’S CONCEPT
Augmented feedback provides
information that can facilitate skill
learning
2
PART I: INTRODUCTION
 
 
Giving augmented feedback is a significant
role as a physical therapist, teacher, coach &
athletic trainer
Augmented feedback provides answers that
save the learner’s time and energy in
learning skills
3
Reinforcement or Feedback
 
 
Reinforcement is usually verbal and is direct to
the learner to motivate or maintain persistency
in learning a task.
Feedback can come in many form but directly
relates to the performance to the intend
outcome and primary guides the performer to
the intended outcome.
4
Wallace & Hagler’s Basketball
Set shot Study
1.  KP group
exceeded verbal
encouragement
initially and
later in practice.
2.  Verbal
encouragement
group showed
no further
improvement.
5
FEEDBACK
 
Two types of performance information
  Task-intrinsic
feedback
  Sensory-perceptual
information that is a natural
part of performing the skill
  Augmented
  Add-on
feedback
to task intrinsic feedback
•  Adds to information detected
•  Adds to information one cannot detect
6
Types of Augmented Feedback
 
Knowledge of Results  Externally
presented information about the
outcome of performing a skill or achieving the
goal
  KR does not describe the outcome, only tells
the performer if they achieved the goal
 
Knowledge of performance  Information
about movement characteristics
that led to the performance outcome
7
Roles of Augmented Feedback
 
 
 
Helps the learner achieve the goal more
quickly (guidance)
Influences the person’s perception of his/
her own ability in a skill (motivation)
Improves the chance that the performer
will repeat the performance
(reinforcement)
8
How essential is augmented
feedback depends on:
 
Depends on the skill being learned.
 
Depends on the stage of learning
9
Augmented Feedback Affect:
Essential
A benefit
Not needed
Hinders
10
ANSWER 1: AUGMENTED
FEEDBACK IS ESSENTIAL
WHEN:
 
 
 
Some situations where sensory information is not
available (e.g., learner cannot see the target)
When learner’s sensory pathways are impaired due to
injury, accident , age, or disease.
Situation where the learner sensory system is
capable of detecting feedback but they cannot use
the feedback because of:
 
 
Lack of experience
It’s not meaningful to the learner
11
ANSWER 2: AUGMENTED
FEEDBACK IS NOT NEEDED
WHEN.
 
 
 
Skill that inherently provides task-intrinsic feedback
Skill that has a detectable external reference in the
environment that the learner can use to determine the
appropriateness of the action (i.e. targets, basket,
goal)
In observational learning situations
 
Observational learning situations seems to precludes
the need for augmentation.
12
ANSWER 3:AUGMENTED
FEEDBACK CAN ENHANCES
SKILL ACQUISITION WHEN
  In
about any complex, multi limb skill where a
person is required to attain a certain degree of
success simply by making repeated attempts to
achieve a performance goals.
  Wallace
& Hagler basketball shooting study
13
ANSWER 4: AUGMENTED
FEEDBACK HINDERS SKILL
LEARNING
When the beginning learner become dependent on
augmented feedback that will not be available in a test
situation.
When the learner is given erroneous feedback
When the learner is given concurrent feedback but!!!
14
SUMMARY
 
 
 
 
 
 
 
Task intrinsic feedback comes from sensory perceptual
information
Augmented feedback adds on to task intrinsic feedback
KR is externally presented information
KP information is about movement characteristics
Feedback motivates, reinforces and speeds learning
Feedback’s effect depends on the skill and the performer
We need to know what situations require feedback and
situations where feedback is not necessary or may hinder
performance.
15
Part II :CONTENT OF FEEDBACK
16
The Content of Feedback
 
 
 
 
 
Errors versus Correct
KR versus KP
Qualitative versus Quantitative information
Performance Bandwidths
Erroneous Augmented Feedback
17
INFORMATION ABOUT ERRORS
VERSUS CORRECT ASPECTS
 
Should one provide mistakes he or she made
or those aspects of the performance that are
correct?
1. 
2. 
Research consistently has shown that error
information is more effective for skill
improvement.
Information about correct performance serves to
motivate the person to continue.
18
KNOWLEDGE OF RESULTS
VERSUS KNOWLEDGE OF
PERFORMANCE
 
Two question are asked about KR and KP in skill
learning situations:
  Do
professionals use one form over the other?
  Does KR and KP have the same influence on skill
learning?
19
The benefits in using each:
 
 
 
 
 
KR
Leaner Use it to
confirm their own
assessment of task
intrinsic feedback
Learner use it when
intrinsic is not present
Learner uses it to
motivate themselves.
Learn uses it to
promotes active
learning
 
 
 
 
 
KP
Learner uses when the
skill requires a specific
movement pattern.
Helps when multi-limb
movement need to be
corrected
Helps when the out come
goal is kinematic, kinetic,
or requires a specific
muscle activity
When KR is redundant.
20
QUALITATIVE VERSUS
QUANTITATIVE INFORMATION
 
 
How precise should feedback be???
Should we provide feedback that is numerical in
value or feedback that identifies the quality of
performance?
  In
terms of Gentile’s model:
  Early
learner needs qualitative feedback;
  Once learned quantitative feedback
21
AUGMENTED FEEDBACK BASED
ON PERFORMANCE BANDWIDTHS
 
How large an error should occur before one gives
augmented feedback?
  Performances outside the performance bandwidth
augmented feedback is given
  Research support the use of bandwidths
  One does not have to reduce the size of bandwidth
limits in relation to the stage of learning
  Participants need to know in advanced that they will
not receiving KR if inside the performance bandwidth
22
REDUNDANT AUGMENTED
FEEDBACK
 
When augmented feedback is redundant will the learner
will use it or ignore it?
  If learner ignores the feedback but task-intrinsic
feedback is present, performance is not effected
  If learner used the feedback but the feedback is
erroneous information, it will bias the performance
  Early stages of learning the beginner will use
augmented feedback whether it is correct or not
  Helps
them deal with uncertainty about what the task
intrinsic feedback is telling them
23
Descriptive and Prescriptive
 
Descriptive and Prescriptive.
  Descriptive
simply describe the movement.
  Prescriptive tells a person what to do
correct it!
  Which form is best?
  Prescriptive
in beginning
  Descriptive once the skill is learned
24
PART III: DIFFERENT TYPES OF
KNOWLEDGE OF
PERFORMANCE
25
DIFFERENT TYPES OF
KNOWLEDGE OF
PERFORMANCES
 
 
 
 
Verbal Knowledge of performance
Videotape as augmented feedback
Movement kinematics as augmented feedback
Biofeedback as augmented feedback
26
VERBAL KNOWLEDGE OF
PERFORMANCE
 
Verbal knowledge of performance is useful
when:
  The
content of what you tell the person practicing
is critical.
  How do you know if is critical information?
  Perform
a skill analysis of skill being practiced.
  Prioritize each part by listing the most critical parts
first and so on.
  Which part of skill must be done properly for the
entire skill to be performed correctly
27
VIDEOTAPE AS AUGMENTED
FEEDBACK
 
Is the use of videotape an effective form of
augmented KP feedback?
  Videotape
replays is a common practice that
many assume aids in skill acquisition
  Effectiveness of video tape replays depends on
one’s stage of learning
  Beginners
need assistants to point out critical
information
  Advanced player benefit form attention-directing
instructions such as verbal cues or checklists
Continued
28
VIDEOTAPE AS AUGMENTED
FEEDBACK
 
Moderately skilled golfers were used in study where three
groups were formed:
Video play was used as source of KP
  Just verbal KP
  NO KP
 
 
Groups were involved in 90 minute practice session for 4
days
PGA pro gave verbal feedback to one group
  Video group viewed video replay of their swing
  Accuracy measures across different distance were
recorded.
 
 
Video KP group performed better than the other two
groups on a retention test given after a 2-week no
practice period.
29
VIDEOTAPE AS AUGMENTED
FEEDBACK
Girls learned a balance beam routine by:
1) using verbal KP with a checklist
2) using video tape replay with a checklist.
At the end of 6 weeks, the videotaped group scored
significantly higher on precision, execution, amplitude
and direction.
In the end, video replay facilitates features the performers
that can or cannot be readily observed.
30
MOVEMENT KINEMATIC AS KP
AUGMENTED FEEDBACK
 
Graphic computer simulations of performances
Lindahl study enabled machine operators to achieve
performance levels in cutting tungsten discs
  Wood et al. graphically displayed golf shot improved
golfers swing
  Physical therapist use of kinematic computer simulation
of the patient’s knee improved the patient’s range of
motion of flexion and extension
 
 
Take into account the performer stage of learning
  Beginner need help to interpret it
  Skilled can use more complex kinematic info
31
BIOFEEDBACK
 
 
Use of task-intrinsic feedback related to physiological
processes (heart rate, muscle activity)
Most common form in EMG
Brucker & Bulaeva ‘s study of EMG feedback study
found it to be effective for cervical spinal cord patients.
  Intiso et al. found recovery with stroke victims were
better when using EMG Biofeedback
  Using swimming cap with an audible signal to maintain
velocity and arm cycle was effective.
  Heart rate presented audibly to competitive rifle
shooters was found to be effective.
 
 
Again, beginners needs help to interpret it
32
Early in rehab or learning
 
Patient or student should be given:
  Error information
  Both KR & KP
  General (qualitative feedback)
  Prescriptive feedback
  Verbal instructions should focus on few critical
movement errors
  If you use self video tapes, kinematic computer
simulation, or biofeedback devices, the beginner will
need help in interpreting it!
33
At the latter stages….
 
Patient or student should be given
  Both
error and corrective feedback
  Both KP and KR feedback
  Descriptive Feedback
  More quantitative feedback
  Verbal feedback can be more complex.
  Use self-video tapes with a checklist
  Use computer simulation and EMG devices
34
The bottomline is
 
The content of feedback needs to change across
the different stages of rehabilitation,
development, or learning.
35
PART IV: Timing of Feedback
36
Timing of Feedback
 
Concurrent & Terminal Augmented Feedback
  Should
one give feedback during or after one
has performed?
 
The KR-Delay & Post-KR Intervals
  What
affect does feedback have just before or
after one has performed?
 
Frequency of Presenting Augmented
Feedback
  How
frequently should I give feedback?
37
PRESENTING FEEDBACK
DURING OR AFTER?
 
Feedback given during performing is concurrent
feedback
  I.E.
Steering a car through traffic; knee extension
device that measures ROM; walking to class
 
 
Feedback at the end is terminal feedback
Which form is better???
  Terminal
feedback is effective in any skill learning
situation but concurrent feedback depends on
level of task-intrinsic feedback in performing.
38
EFFECTS OF GIVING CONCURRENT
AUGMENTED FEEDBACK
 
A negative learning effect occurs when:
  Feedback
directs the learner attention from critical
task intrinsic feedback and towards the
augmented feedback (feedback becomes the
important information).
  Coach
provides feedback to player....pass the
ball…which causes the performer attention to focus
on passing instead of taking an easy short shot to
make a basket in basketball.
Continued
39
EFFECTS OF GIVING
CONCURRENT AUGMENTED
FEEDBACK
 
Enhances skill learning when
  Concurrent
feedback enhances intrinsic feedback.
  Involved skills where task intrinsic feedback back
were difficult to process.
  Training
pilot using a Flight simulator
  Activation of a muscle group in physical therapy
  Learning bimanual tasks
  Information
is directed to the movement outcome
not the movement itself**
40
PREDICTING AFFECTS OF
CONCURRENT AUGMENTED
FEEDBACK
 
 
 
 
When information value of task intrinsic feedback is low,
augmented feedback information value high learner will
depend on concurrent feedback(Annett)
Concurrent feedback must facilitate the learning of critical
features in the task (Lintern & others)
Concurrent feedback directed to movement outcome is
benetifical
Negative effects of feedback occur when it distracts
attention away from these features
41
Give feedback before or after?
 
 
KR-delay interval is between the end of one
practice attempt and the augmented feedback
Post-KR internal is the interval between the
augmented feedback and the beginning of the
next practice attempt
42
T1
Time Intervals
T2
KR
Post-Kr
Delay
KR-Delay
Trial 1
T = trial
43
RELATIONSHIP BETWEEN
INTERVALS AND SKILL
LEARNING
 
To understand the relationship we need to
understand the influence of two variables:
-Length of the intervals (Time)
-Activity during the intervals
44
LENGTH OF KR-DELAY
INTERVAL
 
 
Augmented feedback has informational value for
performers to solve problems.
Seems to be a minimum amount of time that
must pass before giving feedback (Swinnen, et.
al. 1990)
  Too
soon seems to intrefer with task intrinsic
feedback processing
  But if one waits just a few seconds, task intrinsic
feedback processing is facilitated.
45
ACTIVITY DURING THE KRDELAY INTERVAL
 
Three Outcomes have been cited in research that indicates that it
might hinder, benefit, or have not have an affect
  The most common effect is no influence
  It hinders learning when the learner is engaged in error
estimation activities.
 
 
 
E.g., I am going to show you want you did wrong?
E.g., Show me what you did wrong!
It is of benefit if you under go some type of subjective
performance strategy.
 
 
E.g., If you ask the learner to evaluate their performance by
comparing two different strategies in performing or engaging
E.g.,Some type of activity that enhances task-intrinsic feedback
such as watching their performance via a self-tape.
46
LENGTH OF POST- KR
INTERVAL
 
 
 
Very important interval because most believe
this is the period where we develop a plan of
action
Learner uses this time to develop task intrinsic
feedback to determine future plan of action
There seems to be a minimum length of time
for this interval but no upper length of time limit
has not been established
47
ACTIVITY DURING POST –KR
INTERVAL
 
Again the engaging activity has similar results
to that of KR-delay, that is, no effect, hinder,
or benefits.
  Most
common results is no affect
  Much of what goes on during this period is
cognitive so any type of attentional problem
solving activity during this interval will hinder
learning.
  E.g.
I would give you checklist that points out
your mistakes and have you resolve your own
problems.
48
PART V: HOW OFTEN SHOULD
ONE GIVE AUGMENTED
FEEDBACK?
 
 
Giving augmented feedback after every trail is
not practical nor optimal for learning
Some type of relative or reduced frequency
feedback schedule is appropriate
49
DO YOU SUPPORT THE VIEW
THAT “MORE IS BETTER?”
 
Winstein and Schmidt study indicates that
“more is not better.”
  Reduced
frequency of feedback is better for
learning skills
  Optimal relative frequency of feedback is
specific to the skill learned
50
WHY GIVING AUGMENTED
FEEDBACK LESS THAN 100%
BENEFITS LEARNING?
 
Related to attention-capacity overload
100% feedback leads to different type of learning
processes
 
Guidance hypothesis (Winstein & Schmidt)
 
  100%
feedback guides the learner and they
become dependent on augmented feedback
  Relative frequency does not develop dependency
51
Summary
 
 
 
 
 
Terminal feedback is the most common
Concurrent feedback if attention gaining may
affect one’s performance
KR should be delayed and not immediate.
Once you give KR try to minimize problem solving
activities.
Absolute feedback causes dependence where as
relative feedback results in better retention and
transfer.
52
Part VI: Strategies in Giving
Feedback
Strategies in Giving Feedback
12
Absolute Error
10
8
6
4
2
0
Faded
Self-Select
Summary
Feedback Conditions
53
CONCEPT VI: TECHNIQUES
THAT REDUCE AUGMENTED
FEEDBACK FREQUENCY
 
 
 
 
Fading Technique
Self-selected frequency
Performance Based Bandwidths
Summary Feedback
54
FADING TECHNIQUE
 
Form of relative feedback in which one
systematically reduces KR frequency.
Example
• Give augmented feedback for 50% of the trials
• Give augmented feedback for 25% of the trials
• Give augmented feedback for 12% of the trials
• Give augmented feedback when requested or not at
all.
55
SELF-SELECTED FREQUENCY
 
 
The practitioner gives the learner augmented feedback
only when they ask for it
Self-regulates the presentation of augmented feedback
 
 
Janelle, et al. overhand throwing study showed that only
11% of feedback in self-controlled condition was needed.
Why do we ask for help?
It is usually after a good trial not bad
  Use it to confirm our self-evaluation of the trial
(Chiviakowsky & Wulf)
 
56
Lab Results
Feedback %
Good Versus Bad
30%
Percent of FB
70
25%
60
20%
50
40
15%
30
10%
20
5%
10
0
0%
Good
Faded
Self-select
Bad
Summary
Conditions
57
PERFORMANCE BASED
BANDWIDTH FEEDBACK
 
 
Where one gives feedback on a small band of correct
trials or standard set by the teacher, coach, or therapist.
When ever the learner is outside this band
 
 
When the learner is within this band
 
 
Feedback is given frequently
Feedback is given sparingly.
Provides a useful means to individualize the frequency
of augmented feedback
Continued
58
SUMMARY AUGMENTED
FEEDBACK
 
 
Defined as listing performance related augmented
feedback after a certain number of trials
What is the optimal number of performance trials to
be included in summary feedback?
No specific number of trials
  But longer summaries are better for simple skills where
as shorter summaries are better for complex skills
(Guadagnoli, et al., 1996).
  One can give a summary of each trial or an average
across trials (Young & Schmidt, 1996)
 
59
Guidelines about Feedback
The following slides summarize how we should
use the tool of feedback to guide the client or
learner the the desired outcome.
60
GUIDELINE FOR GIVING
AUGMENTED FEEDBACK
1. The person must be capable of using the
information.
- Beginners need “ballpark” info (qualitative)
- Skilled needs more specific info(quantitative)
2. Error-based feedback in the beginning and then
combine performance information of what is
done correctly with error
continued
61
GUIDELINES FOR GIVING
AUGMENTED FEEDBACK
3. Verbal KP should be based on the most critical
errors made during practice
- Analysis of skill
- Prioritize list of components of skill
- Use checklists
4. Prescriptive KP is better for novice;descriptive
KP is appropriate for advanced
Continued
62
GUIDELINES FOR GIVING
AUGMENTED FEEDBACK
5. Videotape replays can be effective with beginners
- Practitioners need to provide direction to help them detect
6. Computer generated displays is effective for advanced
performers
7. Biofeedback needs to give performer’s info that can be
use to alter their movements
63
GUIDELINES FOR GIVING
AUGMENTED FEEDBACK
8. Do not give feedback after every trial and
develop a faded feedback schedule using
bandwidth standards.
9. Wait a few seconds before giving feedback.
10. Have the student or client engage in selfevaluation of their performance then give
feedback. Once given do not have them
engaged in any cognitive problem solving.
64
GUIDELINES FOR GIVING
AUGMENTED FEEDBACK
11. 
12. 
13. 
14. 
Provide a systematic schedule of frequency of
feedback that starts with absolute then
becomes faded
Summary feedback should be provided over
trials. More trials for easy skill and less trial for
more complex tasks.
Give the performer a choice in when they want
feedback.
When using bandwidth feedback inform the
learner of the strategy before giving it.
65
The End
66