Effect of Audience On Music Performance Anxiety
Effect of Audience On Music Performance Anxiety
Effect of Audience On Music Performance Anxiety
Author(s): Albert LeBlanc, Young Chang Jin, Mary Obert and Carolyn Siivola
Source: Journal of Research in Music Education , Autumn, 1997, Vol. 45, No. 3 (Autumn,
1997), pp. 480-496
Published by: Sage Publications, Inc. on behalf of MENC: The National Association for
Music Education
REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/3345541?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Sage Publications, Inc. and are collaborating with JSTOR to digitize, preserve and extend access
to Journal of Research in Music Education
We tested 27 male and female high school band members performing solos under three
levels of audience presence. Participants performed alone in a practice room, in a prac-
tice room with one researcher present, and in the rehearsal room with all researchers, a
peer group, and a tape recording being made. Dependent measures were an analog scale
self-report of perceived anxiety, heart rate recorded during performance, judges' rating of
the final performance, and an exit interview. Self-reported anxiety rose with each suc-
ceeding performance condition, and each reported increase was significant. Heart rate
was steady across the first two performance conditions, but rose significantly at the third.
Female participants presented better performances, reported significantly higher anxiety
levels than did males in the third performance condition, and attained significantly
higher heart rates than did males in the first and third conditions. Gender emerged as a
significant predictor of heart rate during performance, with femaleperformers attaining
higher heart rates.
Performance Anxiet
Performance anxiety is a problem that has long hindered s
musicians from reaching their full potentials as performers. Affl
performers as disparate as concert pianist Vladimir Horowitz and
singer Carly Simon, performance anxiety has also caused discomfo
generations of music students, their teachers, and their parents. I
been documented that musicians as successful as Leopold Godo
(Nicholas, 1989), Enrico Caruso (Greenfeld, 1983), Pablo Casals (Kirk,
1974), and Maria Callas (Stassinopoulos, 1981) had to cope with nota-
When playing with an orchestra he would hide in the wings until the opening tutti
was over, rush out and literally pounce on the piano. On one occasion he forgot
to put aside the cigar he was nervously chomping-this was in Russia-and played
the concerto cigar in mouth, smoking away, much to the amusement of the Czar.
The mere thought of giving a concert made him physically ill. He gave very few
throughout his career-far fewer than any of the great pianists, including Alkan-
and in the last thirty-three years of his life apparently gave no more than three. (p.
199)
Lund (1972) conducted the only study we found in which music per-
formance anxiety with subjects younger than college age was examined.
Lund's study was much like Wardle's, in which a small audience was
used as part of the procedure, but the experimental focus was on con-
trasting methods of treating performance anxiety. Unlike Wardle,
Lund took no physiological data.
Leglar (1978) worked with 30 organists who ranged in experience
from undergraduate organ majors to postgraduate professionals. Her
participants performed alone, in the presence of a critic, and in the
presence of a critic and an audience of about 15 peers. Participant
heart rates increased in the presence of a critic and again in the pres-
ence of a critic and a peer-group audience.
Abrams and Manstead (1981) studied 80 college students (40 men
and 40 women) who performed under four conditions: alone, with two
people present in an adjacent room, in front of a mirror, and for a tape
recorder (with the tape to be evaluated later by a professor). Female
participants made fewer performance errors, reported significantly
more anxiety than did male participants, and attained significantly
higher pulse rates than did the male subjects. Hamann (1982) worked
with 90 (42 male and 48 female) college-level music majors who played
alone with a tape recorder as a reduced-anxiety condition and for a
peer-group and instructor audience as an enhanced-anxiety condition.
Those performing for the peer-group and instructor audience report-
ed significantly greater anxiety, and those with more years of formal
study were judged to have presented better performances.
Hamann and Sobaje (1983) worked with 60 (25 male and 35 female)
college music students who performed under conditions of enhanced
anxiety (jury) and reduced anxiety (no jury). Participants reported sig-
nificantly higher anxiety under jury conditions, and those with more
years of formal study were judged to have presented better perfor-
mances.
METHOD
Participants
Materials
The band director had advised us that lack of a solo would be one of
the main obstacles to participation in our study by students who did n
already intend to enter the solo and ensemble festival. We therefore
decided to furnish solos for participants who did not already have one
and we searched for a published collection of wind instrument solos
that would be relatively easy and available for a wide variety of instru
ments. We had earlier considered standardizing the music to be per-
formed by requiring a single selection to be played by all participant
in the study, but we rejected this idea because we thought it would b
unattractive to our participants, and it would require the peer-group
audience to listen to the same selection many times.
After examining a number of possibilities, we selected a collection o
songs from the musical The Sound of Music (Rodgers & Hammerstein
1959) and identified a number of songs of highly comparable difficu
ty that we would allow participants to select if they did not already ha
a solo they wished to perform in our study. These songs were of a rel
tively easy difficulty level, and the band director assured us that they
were within the reach of any student in the band. We bought enoug
copies of the collection so that every participant had printed music t
take home and practice.
We designed an analog self-report scale, the Personal Performance
Anxiety Report, which participants used immediately after each per-
formance to tell us how much performance anxiety they had felt in th
performance just completed. Instructions were printed on the scale,
which was drawn in the shape of a mercury thermometer. By using th
"thermometer," participants reported their perceived anxiety on a sca
of 0 to 10, with higher numbers associated with greater anxiety.
We used two Polar Vantage XL heart-rate monitors to sample an
accumulate each participant's heart rate during performance. Thi
instrument is widely available in stores that cater to athletes and is pr
marily marketed as a device to facilitate the monitoring of exercise an
athletic training programs. Its predecessor, the CIC Heartwatch, was
field-tested by LeBlanc, Campbell, and Codding (1993) and found
be well accepted by participants in music research. The Polar Vantage
XL was recently used in major studies of music performance anxiety b
Brotons (1994) and Tartalone (1992). We recorded student perfor-
mances on a Sony TC-FX6C stereo cassette recorder using Realistic 33
992 microphones.
Procedure
In our first visit to the band to explain the study, we described the
heart-rate monitor that we would use to record participant heart rates.
We also described this instrument in the letter requesting informed
consent that we sent home to parents so participants had little appre-
hension about the monitor when it was time to begin the study.
Participants were allowed to select their own music to perform. The
band director verified that no solo being used was too difficult for the
ability level of the student performing it. Participants who did not have
their own solos were allowed to select one of the songs we had
approved from The Sound of Music. We noted that the participants
who supplied their own music generally selected more difficult music.
We made every effort to ensure that participants prepared their solos
in conditions that were ordinary for them. The band director made
himself available to help any student who desired his help. Students
were free to devote as much or as little practice to the solos as they
wished. No student was required to perform from memory, and no stu-
dent chose to do so.
All testing was done during band rehearsal in the choral rehe
room adjoining the band room. Although normal clothing could
worn over it, it was necessary for the sensor belt of the heart rate
itor to directly contact the skin of a participant's chest in ord
record an accurate heart rate. To protect the personal privacy o
participants, we arranged for them to use a private and windo
room to put on the sensor belt. We commissioned a professional gr
ic artist to draw illustrations showing correct wearing of the mon
sensor belt and transmitter unit, and a copy of these illustra
together with a simplified set of instructions was provided to part
pants as they entered the dressing room. Participants had no diffi
putting on the sensor belt correctly.
The receiver unit of the Polar Vantage XL looks like a large w
watch, and it is normally worn on the wrist, but we needed to touch
controls of the receiver to enter an event marker immediately bef
each participant began to play, and we were afraid this would creat
distraction if the unit were strapped to the participant's wrist. We t
fore bought two military-style adjustable belts and hung a receiver
from each belt. Participants wore the belts and it was easy for one o
researchers to check the monitor and enter event signals without d
turbing the participant, who would have both hands on the instru
in playing position at the time the signal needed to be entered
field-tested this procedure before the actual study to verify that t
receiver would still be within practical range of the transmitter u
mounted on the sensor belt.
In our first performance condition, participants wore the heart-
monitor while performing the first 2 minutes of their solos alone
practice room. In this and in every performance condition, the mo
tor was set to sample and record heart rate at 5-second intervals. T
yielded 24 measurements of each participant's heart rate under
performance condition. Immediately after the 2-minute performan
each participant was brought to an adjoining practice room whe
or she filled out a Personal Performance Anxiety Report. This p
dure was repeated in every performance condition.
Our second performance condition required participants to play t
first 2 minutes of their solo in a practice room with one of t
researchers present. This performance took place 2 weeks after the f
one.
RESULTS
Reliability
Table 1
1 2.04 1.60
2 3.33 1.80
3 5.11 2.33
Heart-rate monitor
1 109.19 21.34
2 109.04 18.95
3 116.72 22.15
Heart-Rate Monitor
The mean heart rate of our participants was virtually identical in the
first two performance conditions, but there was a distinct rise in heart
rate in the third condition, as shown in Table 1 and Figure 2. Again, we
computed orthogonal difference contrasts. The contrast between the
first two performance conditions was not significant, F (1, 26) = .0029,
p = .9575. The second contrast, comparing the mean of the first and
second conditions with the mean of the third condition, was signifi
Table 2
Orthogonal Contrasts for Personal Performance Anxiety Report
GO "
C) C
Ia
cc>
0 0
LM
L LM
F mCa
"o la
(1)
LM
Ei
a1) LM
.W Lw
V)Q
Q) a
Performance Conditions
o.
12_e0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
C
00
0
Pe 4o..0. . . . . . . .
20
Performance Conditions
Table 3
p .957 .044
Eta squared .00 .15
Exit Interview
In the exit interview, 17 students (63% of the group) said that play-
ing for the researchers and the peer group was the most stressful per-
formance condition, while 8 (30%) said playing for one researcher in
a practice room was most stressful, and 2 (7%) said playing alone in a
practice room was most stressful.
At the conclusion of the exit interview, participants were asked if
they had any general comments to offer. The most prominent com-
ment, made by 5 participants (18% of the group), was that they were
most anxious when playing for a group of classmates and friends. One
said "It is nerve-racking to play for your friends, but if you can play for
your friends, you can play for anybody." Three (11%) said they were
not worried because the music was easy, three said it was fun to partici-
pate, two (7%) said they would be more nervous if their band director
were present, and two said they did not realize they were being record-
ed in the third performance condition. One said he was not nervous
because he was not playing for a festival rating, and another said he was
not nervous playing for his classmates because it was the third time he
was playing his solo for this study.
Results by Gender
Table 4
Mean Scores on Personal Performance Anxiety Report, Heart-Rate Monitor and Performance
Rating by Gender
Heart-rate monitor
1** M 99.83 18.50
1** F 122.81 18.06
2 M 103.84 14.61
2 F 116.60 22.52
3* M 109.32 21.56
3* F 127.50 19.04
Performance rating
3** M 5.70 2.46
3** F 8.20 1.07
heart-rate monito
noted that 16 of
Responses by th
Performance An
when female part
ference was signif
p = .0283. We use
nificance tests be
Female participan
performance con
10.25, p= .0037. F
dition, but this d
p = .0854. Female
formance conditi
Female participan
rating than did
mance ratings wa
Table 5
.58**
Gender .47**
Grade .56**
.53**
Grade .53**
Music selection .39*
Grade and music selection .41*
.35
Gender .36
Music selection .31
Gender and music selection .33
Correlation Analysis
Regression Analysis
We pooled the heart rate scores from all three performance condi-
tions and entered music selection, participant grade, participant ge
der, and performance score into a simultaneous multiple-regressio
equation to predict heart rate. A squared multiple correlation of .
indicated that these four predictors accounted for 27% of heart-ra
variance. Of these four predictors, gender was by far the strongest, wit
a beta weight of .53, and it was the only significant predictor, t (22)
2.41, p = .0249 (see Table 6). If performance anxiety is operationaliz
as obtained heart rate, gender was the best predictor, with female par
ticipants attaining higher heart rates during performance. To further
check the effect of music selection on heart rate, we entered it as the
lone predictor of heart rate in a regression equation. Music selection
accounted for less than 1% of the variance in heart rate, attained a beta
weight of .09, and was not a significant predictor, t (25) = 0.45, p =
.6550.
We pooled the scores on the Personal Performance Anxiety Report
across all three performances and entered performance score, heart
rate, participant grade, music selection, and participant gender into a
simultaneous multiple-regression equation to predict the anxiety score.
These variables accounted for 24% of the variance in the anxiety
scores, and heart rate emerged as the strongest predictor with a beta
weight of .39 (see Table 7). However, it failed to attain significance,
t (21) = 1.76, p = .0933. The fact that obtained heart rate was the
strongest predictor of the Personal Performance Anxiety Report speaks
Table 6
Summary of Simultaneous Regression Analysis for Variables Predicting Heart Rate (N = 27)
Variable B SE B Beta
*p > .05.
Note. B = partial regression coefficient; SE B = standard error of the partial coefficient; Beta =
standardized beta coefficient.
Table 7
Summary of Simultaneous Regression Analysis for Variables Predicting Personal Performance
Anxiety Report (N = 27)
Variable B SE B Beta
well for the validity of the self-report measure. We removed heart rate
from the regression equation to evaluate the predictive power of the
remaining variables and noted that the remaining variables accounted
for only 12% of the variance in the anxiety score. Grade emerged as the
strongest predictor with a beta weight of -.27. It was not significant, t
(22) = -1.25, p = .2234. The negative beta weight indicated that partic-
ipants in higher grades reported lower anxiety.
DISCUSSION
likely to have been taking private music lessons and to have been plan-
ning to enter solo festival. This would have made them more motivat-
ed to excel, and it also would have made their performances in our
study more important to them.
REFERENCES