BJMB
Brazilian Journal of Motor Behavior
Research Article
!
Barros, Mora,
Becker
2020
VOL.14
N.3
121 of 133
The effects of instruction on feedback requests
JOAO A. C. BARROS
1
| ERIKA G. MORA
2
| ANDREA BECKER
3
1
California State University, Fullerton,!800 N. State College Blvd. Fullerton, California 92831, USA.
2
Louisiana State University, Baton Rouge, Louisiana 70803. USA.
3
California State University, Sacramento, 6000 J Street, Sacramento, CA 95819, USA.
Correspondence to:!Joao Barros, KHS 230. Department of Kinesiology, California State University, Fullerton, 800 N. State College Blvd, Fullerton, California 92831, USA.
email: jbarros@fullerton.edu
https://doi.org/10.20338/bjmb.v14i3.190
HIGHLIGHTS
• Autonomy supportive language led to greater
amount of feedback requests.
• Motor skill acquisition was similar regardless
of instructional language.
• Instructions can be used to increase or
decrease number of feedback requests.
ABBREVIATIONS
AE Absolute constant error
ANOVA Analysis of variance
B1-5 Blocks 1-5
CE Constant error
η2 Partial eta-squared
NEED Do you NEED feedback?
N No
VE Variable error
WANT Do you WANT feedback?
Y Yes
PUBLICATION DATA
Received 28 07 2020
Accepted 28 08 2020
Published 01 09 2020
BACKGROUND: Self-controlled feedback enhances skill acquisition. Notably, the pattern and frequency of
feedback requested varies and impacts the amount of learning. Self-controlled feedback benefits have been
attributed to an increase in learner’s feelings of autonomy. It is possible that autonomy supportive instructional
language modulates feedback requests and consequently skill acquisition.
AIM: We investigated if autonomy-supportive language leads to different pattern and increases frequency of
feedback requests and skill acquisition.
METHOD: Forty-two participants (22 women and 20 men) were assigned to a controlling or autonomy supportive
instructions group. After each trial, participants were asked “Do you NEED feedback?” or “Do you WANT
feedback?”, respectively. The task consisted of pressing a specific sequence of 5 computer keys in 1200ms.
Then, participants completed 24hrs retention/transfer tests without feedback. During transfer participants
performed the same sequence in 1500ms.
RESULTS: Repeated measure ANOVAs indicated participants in the WANT group requested more feedback
than participants in the NEED group. Both groups distributed feedback evenly throughout acquisition. No
differences in performance in acquisition or in retention/transfer tests were identified.
CONCLUSION: Autonomy supportive instructional language increased feedback requests but not learning.
Including measures of feelings of autonomy is encouraged to clarify the mechanisms underlying these findings.
KEYWORDS: Self-controlled feedback | KR | Frequency | Skill acquisition | Motor learning | Autonomy
INTRODUCTION
Allowing learners to decide when to receive feedback has been shown to enhance
the acquisition of motor skills compared to externally determined feedback regimens.
1,2
The benefits of self-controlled feedback have been explained from two perspectives.
1,3
One explanation suggests that self-controlled feedback leads to greater engagement in
cognitive processes, such as the identification of performance errors,
3,4,5,6
which in turn
promotes motor skill acquisition. A more prevalent explanation suggests self-controlled
feedback fulfills the learner’s psychological needs, like the need for autonomy,
7,8,9,10
which
ultimately enhances motor skill acquisition.
Although the benefits of self-controlled feedback are consistently reported, the
pattern and frequency of feedback requests that lead to these benefits is varied.
1,11,12
Some studies report a decrease in feedback requests as practice progresses
11,13,14
while
others report participants’ feedback requests are evenly distributed throughout practice
trials.
15,16,17
Additionally, in some studies
14,18
participants request feedback after a relatively
low percentage of trials (ranging from an average of 7% to 11% of practice trials) while in
BJMB! ! ! ! ! ! ! ! Research Article!
Brazilian(Journal(of(Motor(Behavior!
Barros, Mora,
Becker
2020
VOL.14
N.3
122 of 133
others
15,19,20
participants request a relatively high frequency of feedback (ranging from an
average of 31.3% to 97% of practice trials).
The frequency of feedback chosen by participants appears to influence skill
acquisition. For example, Chiviacowsky, de Medeiros, Kaefer, Wally, and Wulf
21
compared
children who requested relatively high frequencies of feedback (i.e., 39.3%) to those who
requested relatively low frequencies of feedback (i.e., 8.4%) while learning a beanbag
tossing task. The results indicated that participants who requested relatively more
feedback performed better in the retention test compared to participants who requested
relatively less feedback. The authors argued that the benefits of self-controlled feedback
are likely mediated by the feedback frequency requested by the learners. Further, the
authors suggest that identifying instructions that impact the frequency of feedback
requests could increase the observed benefits of self-controlled feedback.
Examining the effects of instructions on skill acquisition can also provide additional
insight into the mechanisms underlying the benefits of self-controlled learning feedback.
For example, Hooyman, Wulf, and Lewthwaite
22
examined if instructions that increased
learners’ feelings of autonomy would enhance the acquisition of cricket pitching to a target
compared to controlling or neutral instructions. As mentioned earlier, increased feelings of
autonomy as a result of self-controlled feedback have been associated with enhanced
learning.
23
Three groups were exposed to either autonomy supportive, controlling, or
neutral instructions before practice begin. The results indicated that participants who
received autonomy-supportive instructions demonstrated enhanced skill acquisition
compared to participants who received controlling or neutral instructions. The autonomy-
supportive instructions also led to greater feelings of autonomy, higher self-efficacy and
positive affect. The authors argued that the instructions increased learners’ feelings of
autonomy and in turn enhanced learning. That increase in feelings of autonomy did not,
however, lead to statistically significant changes in self-controlled behavior (i.e. pacing)
during acquisition. That is, participants did not take advantage of the autonomy afforded to
them. This might have happened because participants had limited opportunity to exert
control over the learning environment (i.e., participants were only allowed to control
pacing), differences in the content of the instructions themselves (i.e., participants in the
autonomy supportive group were told “feel free to go at a pace you are comfortable with”
while participants in the controlling group were told that “you must maintain a consistent
pace”), or limited exposure to the instructions (i.e., participants were only given the
instructions once at the beginning of practice).
In the present study, we investigated if autonomy-supportive language leads to
different pattern and frequency of feedback requests and skill acquisition. Primarily, we
expected that participants exposed to language that promoted autonomy would request
more feedback than those who were exposed to more controlling instructions and that the
differences between groups in feedback requests would increase during the acquisition
phase. Secondarily, we expected that language promoting autonomy would lead to
increased retention of a motor skill compared to instructions containing more controlling
language.
METHODS
Participants and experimental groups
BJMB! ! ! ! ! ! ! ! Research Article!
Brazilian(Journal(of(Motor(Behavior!
Barros, Mora,
Becker
2020
VOL.14
N.3
123 of 133
Forty-two right-handed college-aged volunteers (22 women and 20 men) free from
cognitive or sensory-motor impairments participated in the experiment. Sample size was
based on Lohse, Buchanan, and Miller
24
who noted that most motor learning experiments
are underpowered citing a selected group of experiments with a median n/group = 11.
Therefore, we selected a n/group = 21. In-class announcements were used to recruit
participants at a large public university in southern California. Participants were free from
musculoskeletal injuries and sensory-motor or cognitive impairments. Participants were
inexperienced with the task and naïve to the purpose of the experiment. They were quasi-
randomly assigned to one of two experimental groups: NEED (M
age
=21.14, SD
age
= 1.74
years), and WANT (M
age
= 20.10, SD
age
= 2.17 years). Participants in the NEED group
were presented with a screen that read “Do you NEED feedback? Y/N” and participants in
the WANT group were presented with a screen that read “Do you WANT feedback? Y/N”
after every trial. All other instructions were identical. Campbell
25
indicates that although
individuals sometimes use the words “need” and “want” interchangeably in casual
conversations, in the context of consumption (i.e., using resources available to them), they
are aware of the critical differences between the two terms and use them differently in
achieving their goals. Additionally, Wilensky
26
argues that the words “need” and “want” are
perceived differently. The author indicates that “need” is used when it refers to something
that is a necessary precondition or action for the subject’s goal. “Want” is used when it
refers to something that is itself the subject’s goal, given that the goal is not one of
preservation or obligation. Indeed, O’Boyle
27
defines a need as something indispensable
while a want is something desired. Further, avoiding instructions that contain certain words,
like “have to” or “must”, can increase learners’ sense of choice and flexibility.
28
In the
present context then, asking participants if they need feedback suggests feedback is
indispensable, or something that they have to have, and as such restricts participants to
only request it when it is absolutely necessary. On the other hand, asking participants if
they want feedback allows learners more freedom in deciding if feedback would or not be
appropriate for that trial. Quasi-random group assignment matched the number of women
and men in each group. All procedures were approved by the university’s Institutional
Review Board.
Task and material
The task consisted of pressing five computer keys sequentially (3-6-5-8-4) in
exactly 1200ms (Figure 1). Previous literature has demonstrated advantages of self-
controlled feedback in the acquisition of this type of task.
6,15,16,17,19,29
The use of this task
then allows for the comparison of the frequency of feedback requests in this experiment to
others in the self-controlled feedback literature. The task was performed on the numeric
pad of a standard keyboard on a desktop computer. Custom software (E-Prime 3.0,
Psychology Software Tools, Inc., Pittsburg, PA, USA) was used to provide general
information about the task, specific instructions regarding feedback requests, feedback,
and to collect data. Information about the task, specific instructions regarding feedback
requests and feedback were displayed on a standard computer monitor.
BJMB! ! ! ! ! ! ! ! Research Article!
Brazilian(Journal(of(Motor(Behavior!
Barros, Mora,
Becker
2020
VOL.14
N.3
124 of 133
Figure 1. Key-pressing sequence and goal movement time. Fig. 1 depicts the sequence used for both
experimental phases and the goal movement time for acquisition. For the transfer test, goal movement time
was 1500ms. Once this diagram was displayed, participants were allowed to start. The diagram was visible
until the last key was in the sequence was pressed.
Procedures
Data collection was conducted individually in a secluded room. The experiment
included two phases held on two consecutive days: acquisition and retention/transfer.
During the acquisition phase, participants were given a brief overview of the experimental
task and asked to read and sign the informed consent form. After providing consent,
participants were seated in front of a desktop computer and asked to read the instructions
on the screen, then explain the instructions back to the experimenter. During the
acquisition phase participants completed 50 trials of the experimental task. A trial began
with “Ready?” display on the screen for 2 seconds, then a diagram with the sequence and
goal movement time (Figure 1) was displayed. Participants could then start to press the
keys when they were ready. Once the participant pressed the first key, time started to be
recorded (i.e. movement time). The diagram remained on the screen throughout the
response. Once the last key of the sequence was pressed a blank screen appeared for
one second. Participants in the NEED group were then prompted with the question “do you
NEED feedback? Y / N” after every trial. In the WANT group, participants were prompted
with the question “do you WANT feedback? Y / N” after every trial. After the prompt,
participants would press “Y” to receive feedback or “N” to move on to the next trial without
receiving the feedback. Feedback consisted of sequence errors (i.e., if the sequence was
completed correctly or not), and constant error (i.e., the difference between movement time
and goal movement time). No sequence errors were observed. Feedback, if requested,
was presented on the screen for 2 seconds. If participants did not request feedback, a
blank screen was displayed for 2 seconds.
Approximately 24 hours later, participants completed a 10-trial retention test and a
10-trial transfer test, in which the goal was to perform the same sequence in 1500ms. No
feedback was provided during retention or transfer tests. The retention/transfer phase
lasted approximately 10 minutes.
Measures
Participants’ choice regarding feedback was recorded for every trial. Additionally,
the time elapsed from the pressing of the first key to the pressing of the final key in the
sequence (i.e. movement time) was recorded. Constant error (CE) was calculated as the
BJMB! ! ! ! ! ! ! ! Research Article!
Brazilian(Journal(of(Motor(Behavior!
Barros, Mora,
Becker
2020
VOL.14
N.3
125 of 133
difference between movement time and goal movement time for each trial. The dependent
variable used to investigate the effects of instruction on pattern and frequency of feedback
requests was number of feedback requests in each 25-trial block. The dependent variables
used to assess performance and learning of the task were: 1) mean CE, calculated by
averaging the CE obtained in each 10-trial block; 2) mean absolute constant error (AE),
obtained by calculating the mean of absolute CE in each 10-trial block; 3) mean variable
error (VE), the standard deviation of the CE obtained in each 10-trial block. Schmidt and
colleagues
30
indicates that these measures of performance are typically used in motor
learning research and provide different insight into the performers’ capabilities.
Analyses
To test the effects of instruction on the pattern and frequency of feedback requests,
a 2 (groups: NEED; WANT) by 2 (blocks of 25 trials: first; last) analysis of variance
(ANOVA) with repeated measures on the last factor was conducted. Changes in
performance and differences between groups during acquisition were tested via separate 2
(groups: NEED; WANT) by 5 (blocks of 10 trials: B1, B2, B3, B4, B5) repeated measures
ANOVAs for CE, AE, and VE. Performance and differences between groups on the
retention and transfer test were tested via separate 2 (groups: NEED; WANT) by 2 (tests:
Retention; Transfer) repeated measures ANOVAs for CE, AE and VE. These statistical
analyses are fairly typical in the literature on autonomy support
22
and self-control
21,22
given
the number of groups, nature of the measurements, and repetitive nature of the
performance. As an additional measure for skill retention, separate 2 (groups: NEED;
WANT) by 2 (blocks of 10 trials: B1, RET) repeated measures ANOVAs for CE, AE, and
VE were included as secondary analyses. A visual examination the data suggested no
outliers and approximate normal distribution for the vast majority of the dependent
variables. Where appropriate Levene’s test of homogeneity of variances was used. The
vast majority of the dependent variables met the assumptions for the chosen analyses.
When the sphericity assumption was violated the Greenhouse-Geisser adjusted values are
reported. Partial eta-squared (η
2
) is reported as an estimate for effect size. Sidak post hoc
procedures were used when appropriate. For all analyses, alpha level was set at 0.05, and
SPSS
®
V21 software was used.
RESULTS
Feedback Requests
Table 1 depicts feedback requests during acquisition for participants in the NEED
and WANT groups. Participants in the WANT group requested more feedback than
participants in the NEED group. It also appeared that participants requested slightly more
feedback in the second half of the acquisition phase than in the first half. The ANOVA
partially confirmed these observations indicating a group effect (F(1, 40) = 4.175, p = .048,
η
2
= .095) but no block effect (F(1, 40) = 1.525, p = .224, η
2
= .037) or block by group
interaction (F(1, 40) = .731, p = .398, η
2
= .018).