JOURNAL OF SPORT & EXERCISE PSYCHOLOGY, 1993,15,290-303
O 1993 Human Kinetics Publishers, Inc.
Reducing Warm-up Decrement in the Performance
of the Tennis Serve
Mark H. Anshel
Craig A. Wrisberg
University of Wollongong
The University of Tennessee, Knoxville
,
In the present study an attempt was made to determine the relative effectiveness of various warm-up activities in eliminating postrest warm-up decrement
(WUD) in the tennis serve. Seventy highly-skilled players hit 20 serves,
rested for either 5 or 15 min, and then attempted 4 final serves. During the
last 2 min of the rest period, players continued to rest, ran in place, engaged
in mental imagery, performed practice swings, or repeatedly hit the ball
against the ground and caught it. In addition to estimates of serving accuracy,
measures of somatic and cognitive arousal were obtained at the beginning
and end of the rest interval. Multiple regression procedures revealed that
reductions in WUD were significantly related to the restoration of prerest
arousal levels. Between-group comparisons indicated that practice swings
were the most effective warm-up activity for restoring somatic and cognitive
arousal to prerest levels and for eliminating WUD. Theoretical discussion
centered on possible applications of Nacson and Schmidt's (1971) activityset hypothesis to the tennis serve.
Key words: mental imagery, physical practice, rest, somatic and cognitive
arousal
One of the most pervasive phenomena in the domain of human motor
behavior is the temporary decrement in performance that occurs following an
interval of rest. This phenomenon, termed warm-up decrement (WUD), is usually
brief in duration, lasting for only the first few trials of postrest performance.
Because WUD has been shown to be reduced or eliminated by the practice of
interpolated tasks that presumably contribute little to the niemory strength of the
criterion task (e.g., Wrisberg, Salmoni, & Schmidt, 1975), it has been assumed that
WUD is not due to forgetting the criterion response,but rather to inappropriately
readjusting underlying mechanisms that support criterion-task performance
(Adarns, 1961).
Mark H. Anshel is with the Department of Psychology at the University of Wollongong, Wollongong, New South Wales, Australia 2522. Craig A. Wrisberg is with the
Department of Human Performance and Sport Studies at the University of Tennessee,
344 HPER, Knoxville, TN 37996-2700.
Warm-up Decrement / 291
Of the various theoretical accounts for WUD (for reviews see Adams,
1961; Schmidt, 1972), the one receiving the most empirical support has been the
activity-set hypothesis (Ainscoe & Hardy, 1987; Anshel & Wrisberg, 1988;
Schmidt & Nacson, 1971; Schmidt & Wrisberg, 1971), proposed by Nacson and
Schmidt (1971). According to this hypothesis, motor skill practice serves, among
other things, to adjust relevant systems (e.g., attentional focus, arousal level,
speed-accuracy trade-off requirements) that underlie task performance. A brief
bout of rest, however, usually results in the readjustment of one or more of these
systems. For example, during a change of side in tennis, the attentional focus of
a player might shift from an external orientation required for successful visual
tracking of the ball to an internal one that is conducive to keeping track of firstserve percentage. If appropriate adjustment of attentional focus (or of other
systems comprising the activity set for tennis) does not occur prior to the resumption of play, then WUD may result (e.g., in serving or receiving). In summary,
then, WUD is presumed to be due to an insufficient readiness to respond that
occurs when relevant support systems (i.e., the activity set) are not properly
readjusted following a rest interval.
Tests of the activity-set hypothesis have typically involved a paradigm in
which various warm-up activities are inserted at the end of a rest period prior
to the resumption of performance on the criterion task. If a particular warm-up
activity appropriately adjusts an underlying component of the activity set for the
task, then a reduction in postrest WUD (relative to that of a no warm-up control
condition) should occur.
One underlying system that appears to be a part of the activity set for a
number of tasks is arousal level. Usually associated with the intensity dimension
of behavior, arousal has in recent years been examined as a multidimensional
concept. Somatic (physiological) arousal refers to a "state of activation of the
autonomic nervous system (e.g., heart rate or blood pressure) on a continuum
from deep sleep to high excitement',' (Anshel et al., 1991, p. 113). Cognitive
(psychological) arousal, on the other hand, is the "heightened state of alertness
and readiness on a continuum from deep sleep to extreme excitement [and] may
include positive or negative emotions" (Anshel et al., 1991, p. 121). The results
of several studies have revealed that groups performing warm-up tasks designed
to readjust levels of somatic arousal demonstrate significantly less WUD during
postrest performance relative to a no-warm-up control condition. For example,
Anshel(1985) and Anshel and Wrisberg (1988) found that WUD (for gymnastics
vaulting and baseball batting, respectively) was eliminated when subjects engaged
in a warm-up that involved bicycle ergometer work at 50% maximum heart rate.
Although warm-up tasks in these two studies were designed to increase arousal
levels relative to those associated with rest conditions, it is possible that, in
some situations, arousal may need to decrease in order to facilitate subsequent
performance. For example, a golfer may need to diminish her arousal level before
attempting a putt following a two-iron shot to the green. Although Nacson and
Schmidt (1971) conceptualize the activity set as a "generalized preparation to
respond" (p. 3), it would appear that studies exploring the importance of arousal
adjustments in reducing WUD need to differentiate somatic and cognitive components. In the present experiment, somatic arousal was estimated by monitoring
heart rate while positive and negative components of cognitive arousal were
292 / Anshel and Wrisberg
operationalized by obtaining subjects' responses on an adult version of the Children's Arousal Scale (CAS-A, Anshel, 1985).
Another possible component of the activity set is mental imagery. The
results of research exploring this possibility indicate that reductions in WUD
following an imagery warm-up occur for some tasks but not for others. Ainscoe
and Hardy (1987) found that the WUD of gymnasts was eliminated in the
performance of double leg circling on the pommel horse after subjects warmed
up by imaging themselves correctly performing the task. However, Anshel(1985)
found no influence of positive imaging on postrest WUD for the handspring
vault. Anshel and Wrisberg (1988) observed reductions in WUD in baseball
batting for subjects who engaged in relevant imagery (e.g., visualizing the release
of the ball from the pitcher's hand and the moment of impact with the bat), but
not for those who practiced irrelevant imagery (e.g., visualizing the trajectory of
an approaching fly ball followed by a successful catch) as a warm up.'
Taken together, the available evidence suggests that some support systems
(e.g., arousal) may be a part of the activity set for a variety of tasks whereas
others (e.g., mental imagery) may have a more limited usefulness. In the present
study the tennis serve was used as the criterion task, and warm-up activities were
selected that were reasoned to be possible candidates for the activity set. Advanced
tennis players performed serves by hitting balls toward a designated target situated
in the ad court. Following initial trials, subjects rested and then, during the last
2 min of this period, continued to rest or performed one of four warm-up activities.
One of the warm-up activities was-designed to increase somatic arousal (i.e.,
running in place), whereas another was intended to create a successful mental picture
of the correct serve (i.e., imagery). The remaining two warm-up activities involved
separate components of the criterion task that could be performed by players within
the rules of competitive tennis. One emphasized the mechanics of the serving motion
(i.e., practice swings without the ball), whereas the other contained hand-eye coordination requirements (i.e., striking the ball toward the ground and then catching it).
It should be noted that the latter two tasks represent a modest departure from the
kinds of warm-up activities employed in most laboratory investigations of the activityset hypothesis. Because the two activities simulate quasi-components of the serve
itself, it might be argued that they contribute memob strength to the serve rather
than adjust underlying nonhabit mechanisms. However, it was reasoned that highperformance athletes such as those used in the present study would be likely to
derive little in the way of increased memory strength for the serve (which they have
probably executed thousands of times previously) from short bouts of isolated practice
swings or ball striking.
A secondary purpose of the present investigation was to determine the
effect of rest interval length on WUD. The majority of previous studies have
involved fixed periods of rest ranging from 10 min (Nacson & Schmidt, 1971;
Schmidt & Nacson, 1971; Schmidt & Wrisberg, 1971) to 20 min (Anshel, 1985)
in length. In a more recent experiment, Wrisberg and Anshel (1993) employed
rest intervals of 5 and 15 min and found no significant differences in the tennis
groundstroke as a function of rest interval length. However, additional research
with a variety of tasks is needed to determine whether postrest decrements in
performance differ as a function of length of the rest period. Therefore, in
the present study subjects rested for either 5 or 15 minutes between criterion
performance periods.
Warm-up Decrement / 293
Method
Subjects
The subjects were male (n = 55) and female (n = 15) tennis players
from New South Wales, Australia, who had competed in at least one state-level
tournament. Subjects ranged in age from 19.2 to 26.7 years (M = 23.8). Prior to
participation in the study, players were subjectively rated by two tennis instructors
on their ability to hit the serve consistently over the net and into a designated
target area using the proper technique. A 5-point Likert scale was employed with
scores ranging from 0 (poor) to 5 (excellent). Only those players receiving a
rating of 4 or higher by both judges (interrater r = .94) were selected as subjects.
The range of scores for the subjects who qualified for the study was 4.12 to
4.90.
Apparatus
Measures of somatic and cognitive arousal were obtained at various times
during testing. Somatic arousal was inferred by using a Sports Tester 3000 Heart
Rate Monitor to record the heart rates of subjects. Estimates of cognitive arousal
were determined by administering an adult version of the Children's Arousal
Scale (CAS-A, Anshel, 1985). The CAS-A was chosen over other sport-related
scales (e.g., CSAI-2; Martens, Burton, Vealey, Bump, & Smith, 1990) because
it could be administered quickly and contained no reference to competitive
conditions. The CAS-A is a 7-point semantic differential scale designed to measure general state arousal based on perceived emotions accompanying experiences
of success and failure. It consists of four items associated with positive arousal
(e.g., happy, relaxed) and six items related to negative arousal (e.g., worried,
nervous). Subjects are asked to indicate the extent to which each item describes
the way they feel at the present time. A Pearson Product-Moment scale intercorrelation revealed a low positive relationship between positive and negative arousal
items (r = .30). Internal consistency of the total scale (Cronbach's alpha = .79)
had been previously established (Anshel, 1985) and was reaffirmed in the present
study (Cronbach's alpha = .81).
Criterion Task
The criterion task involved serving a tennis ball from a location 2 ft (.61
m) to the left of the center mark into a designated target area located in the ad
service court. The target consisted of a rubber mat covering a rectangular area
bounded by the service line, the center line, and two additional lines, one located
3 ft (.91 m) from the service line and the other 2 ft (.61 m) from the center line.
The objective on each shot was to hit the ball into the ad court as close to the
target area as possible.
Interpolated Tasks
Subjects performed one of the following tasks during the last 2 min of
the rest interval. The 2-min interval was monitored using a stopwatch. Verbal
commands (e.g., "go") and heart rate feedback were provided by an experimenter
at predetermined moments.
.saseqd pqyl pw lSr!j ayl 8 u y p pau!eiu!em SBM IBNalUr
puuaiu! s-01 v .SIBUI a~wuuopadpug rnoj pw '(isar ~euo~i~ppe
10 diyn!i3e
dn-WM y l ! ~palIg araM y 3 ; [ yjo~ U!W i s e ~ayi) pouad pal e 's~euia3wm
-ropad p!i!ur 02 :saseyd a a q jo paislsuo3 iuama8w.r.m ~eluamuadxaa u
-oraz e st! pa.103~
araM earn ia8rni ayi uy %u!pw[sahIaS 'sa.w ia8.wi aqi jo raiawuad aqi uo lured
lsamau ayi pw iurod ley1 uaawaq m3 u! a~ut?ls!p
ayi palnseaur uaql pw 8u!pwl
Ilaq jo uo!ie3o-[ayi payrnm sraiuamuadxa o ~joi m a i e pui y3ea-rod -a~q!ssod
st! earn ia8m ayi 01 asop se uno:, pe ayi oiu! 1189aqi i!y 01 pay3ruisu! araM
siaafqng -1anb3er sluuai U M O ray ro s!y 8ysn dlpnp!n!pu! paisa1 8ulaq pafqns
y3ea r l , .~y
~ ~ A~a~eur!xo~dde
jo popad a ~ o palsq
j
uorssas leluampadxa u v
-sdnor8 11sroj a m s ayl aq ( E = u) salemaj put? ( I 1 = u) salem jo uo!uodord ayi
d1uropw.1araM pw siuamn3op
yeyi uoli3gsa.1ayl y i ! ~S U O ! I ~ 01
~ Upau%!sse
O~
iuasuo3 pauuoju! pau8ls sparqns 'sa~npa3oldLmuymqard jo uo!ia~dwos8uy
-MOIIO+J - V - S Vp
~w roiruour ale1 m a y ayi y i ! ~
pazpqymej a1aM pw iuaurpadxa
aqi jo sluaurar!nbar ayi jo uoyyewldxa w u a ~ l 8a1aM si3arqns ' d ~ p p ! u ~
.Ai!~!i3es ~ y 8uunp
i
pasn SBM ~leqou 'ranaMoH
-sleg pz jo pi01 e 103 s s Lana parmmo ysa s!yi jo lepi y3ea 'ayuis dnorf) jo
asm ayi ul s v 'amas ayi jo sz1!uey3au1%U!MS
ayi pa~!i3wdsi3arqns 'a~!i3afqo
s!ql a~ayy3eo~ qsei uouaiw aqi q l ! ~
paley30sse Iaaj a~pda3oudordpm uqd
~
SF%
'(8u.2~
S ~~ U
) ~ Ma3!13w~d
S
roiour aw aie~nm!s01 papuaiu! s e Aqnyi3e
'qeui p z jo ploy a roj s s L a ~ apaun330 ysei sryi jo ppi auo .pwy ai!soddo
r'ayi y i ! ~~leq8uypunoqa~ayi 1y8ne3 uayi pue ' l a ~ ay s~! e ~moqe ie pun018
aqi prv~oi11eq ayi ym.11~
'puey pa.uaja1d qayi u! ianb3er ayi play si~arqns' s ~ y i
ys!lduro33e o~ .ysei uouaiu:, aq? jo iuauodmo3 uo~ieu!proo3ada-pwy ayi azrs
-eydma 01 SBM Ai!n!i3e dn-u~rems!yijo aso&nd a u ' ( a ~ p i ys )3 i w p ~ y ~ u ~
'(EL' = as ' 6 s ' =
~ N ) s s a u p ! ~ put?
! ~ (92- = as
'6~'=
p j q ) a~w!ldmo3jo slanaf y8!y d~an!$e~ar
palexpuy sqnsal a% .sa8emy
st A8aie.q~La8em! ayi pasn Aayi
rlayi jo ssaup!n!A jo aar8ap ayi p w pai~nrisu~
) 01 payse araM si3arqns
y 3 ~ 01
y ~iuaixa ayi (isay8zy = s 01 lsaMol = 0 m o ~ jaler
'uoyied!3!md%U!MOIIO+J
.auoz ~a8rniaql jo ralua:,aql ur p a p q 1vq1sauas 13alro3
LIIB~!UB
8u!mopad
~ ~ ~ U I sanIasurayi pau!8em! pw sada r!ayi pasop sparqns
'(p861)SWH pue s p e dq
~ pasodord saugap!n8 ayl %u!sn.ysai uouai!r3 ayi jo
a8em! an!i!sod e aloisal 01 pau%!sap seM Aqn!13e sly& .(a8vzq) cCtaOvw~
'lanai
Ay!suaiu! parpap ayi jo a3ueuaiu!ew amsse 01 rapro u! s S I Lana aier m a y lrayi
jo pauuoju! araM si3afqns -amas s!uuay ayy se y3ns ysey e jo a3wuuogad ayi
roj aiepdo~ddeaq plnoM IanaI 1esno.w w y3ns ley1 pamnsse s e 11~ 'ranoaropu
'an8!1ej p31sAyd inoyiy pw A~ddnspoo~qaienbape w dq lam a n sanssri ayi
jo spwurap ua8Axo q 3 1 y U~I aleis Apeays e %u!ysgqeisaapyM sassa3ord xqorae
ayohu! oi u ~ o y uaaq
s
say i! asnwaq uasoy3 SBM Al!suaiu! jo IaAal s ! u '(9861
' p u a ~ s vS.) samg a8e s'i3afqns ayi s n u y 022 se paieln3p3 'mnm!xem pa13!pa~d
qayi jo %OSjo ale1 m a y e paqibar ivy1 di!suaiu! w @ a q d u! uru 01 svarqns
8uurnba.I dq pays!1dmo33e s e sryL
~ .a3afqns jo lesnorn 3!iemos ayi asear3~
01 seM 113!~!13edn-WM sly1 jo aso&nd ayL '(uny) aavld u.2 Bupiunx
.(sarnp
-a3oJd aas) au!ze8em sFuuay e morj apym w pear si3afqnsasayL .lo~iuo3
Warm-up Decrement 1 295
During the first 10 s of the rest period, subjects' heart rates were monitored.
Immediately after this, they completed the CAS-A. During the remainder of the
rest period, subjects were seated and read an article from a tennis magazine that
detailed competitive strategies used by well-known professional players. This
activity was designed to prevent mental rehearsal of the criterion task. To assure
that subjects read the material, they were told that several questions would be
asked about the contents of the article at the completion of testing. All subjects
were asked between two and four questions, depending on the length of their
answers. During the last 2 min of the rest period, treatment subjects engaged in
their respective warm-up activities, while control subjects continued to read.
Following the rest period, heart rate was again monitored and the CAS-A was
completed. Subjects then performed four final trials of the serve.
The entire experimental protocol was completed twice by each subject,
once for each length of rest period (5 and 15 min). The order of rest periods was
counterbalanced among subjects in each condition with 7 subjects receiving a
5-min rest during the first round and a 15-min rest during the second while the
remaining 7 received the opposite order.
Results
Performance Accuracy
Since preliminary analyses revealed no significant effect of rest interval
length during prerest performance, a 5 x 20 (Groups x Trials) ANOVA with
repeated measures on the second factor was performed on mean prerest performance accuracy scores (Figure 1). Where appropriate, omega squared (0') was
calculated to determine the meaningfulness of significant differences (Tolson,
1980). The results revealed significant effects for groups, F(4, 65) = 5.10, p <
.01, m2 = .01; trials, F(19, 1235) = 338.69, p < .01, m2 = .76; and the Groups x
Trials interaction, F(76, 1235) = 2.80, p <.01, m2 = .02. Simple main effects
analysis of the Groups x Trials interaction revealed that Group Strike was significantly ( p < .05) more accurate than Groups Swing, Run, and Control on the first
two prerest trials. However, no significant between-group differences existed on
any other trials. Thus, it was assumed that the performance of all groups was
equivalent immediately prior to the rest interval.
Postrest performance accuracy was evaluated by a 5 x 2 x 5 (Groups x Rest
Intervals x Trials) ANOVA with repeated measures on the last two factors. Trial
20 of prerest performance was included in the analysis to determine the magnitude
and pattern of postrest WUD in the various groups. Follow-up tests using simple
main effects analysis and Tukey's procedure were conducted to locate the source
of any significant differences. Significant effects were obtained for groups, F(4,
65) = 47.39, p < .01, d = .17; trials, F(4, 520) = 415.30, p < .01, 03 = .42; the
Groups x Trials interaction, F(16, 520) = 44.38, p < .01, d = .17; and the Groups
x Rest Intervals interaction, F(4, 65) = 10.31, p < .01, (3 = .01.
Follow-up tests of the Groups x Trials interaction (see Figure 2) revealed
significant differences between groups on Trials 21,22, and 23, F(4,65) = 89.75,
p < .01, m2 = .77; F(4, 65) = 19.92, p < .01, 0' = .40; and F(4, 65) = 12.63,
p < .01, m2 = .28, respectively, but not on Trials 20 (last prerest trial) or 24, F(4,
65) = .65, p > .05, and F(4, 65) = 1.88, p > .05, respectively. For Trial 21, all
296 / Anshel and Wrisberg
TRIAL
Figure 1-Error
scores for prerest trials of the tennis serve.
warm-up groups demonstrated significantly higher performance accuracy (i.e.,
less error) than Group Control (M = 54.89, SD = 12.87). In addition, Group
Swing (M = 8.61, SD = 4.46) was significantly more accurate than Groups Strike
(M = 25.75, SD = 5.67), Run (M = 32.21, SD = 8.05), and Image (M = 35.18,
SD = 7.30), and Group Strike was more accurate than Group Image.
This pattern persisted on Trial 22 as Group Control (M = 30.39, SD =
11.07) was significantly less accurate than all warm-up groups, whereas Group
Swing (M = 10.25, SD = 3.34) exhibited significantly more accuracy than Groups
Run (M = 19.86, SD = 8.24), Strike (M = 17.43, SD = 5.87), and Image (M =
22.75, SD = 8.07). By Trial 23, the only significant differences were between
Group Swing (M = 6.18, SD = 3.39) and the other groups (Strike, M = 12.96,
SD = 4.49; Run, M = 13.57, SD = 5.86; Image, M = 14.57, SD = 7.81; Control,
M = 16.82, SD = 4.97).
Within-group analyses of performance accuracy across the 5 trials revealed
a classic pattern of WUD for Groups Control, Image, and Strike, F(4, 104) =
176.29, p < .01, a2= 3 1 ; F(4, 104) = 1 4 7 . 8 6 , ~< .01,& = .68; and F(4, 104) =
78.01, p < .01, 02 = .63, respectively. For all three groups, performance was
significantly Cp < .05) less accurate on Trials 21, 22, and 23 than on Trials 20
and 24. A significant trials effect was also found for Group Run, F(4, 104) =
Warm-up Decrement / 297
*
----.---.
!WING
---0-
\
IMAGE
U STRIKE
----
f
.--.
FUN
TRIAL
Figure 2 -Error
serve.
scores for the last prerest trial and all postrest trials of the tennis
70.36, p < .01, a2= .61. For this group, performance accuracy on Trials 21 and
22 was significantly lower than that on Trials 20, 23, and 24. Perhaps most
impressive, however, was the absence of a significant trials effect for Group
Swing.
Post hoc inspection of the Groups x Rest Intervals interaction revealed a
significant effect of rest interval length only for Group Image, F(1, 26) = 1'8.39,
p < .01, m2 = .07. The performance accuracy of this group was significantly ( p
< .05) lower after the 15-min interval (M = 21.63, SD = 12.49) than after the 5min interval (M = 15.41, SD = 9.65).
In summary, the analysis of performance accuracy scores indicated that practice
swings was the only warm-up activity that eliminated WUD. Although the amount
of WUD in the postrest performance of the other warm-up groups was less than that
of the control condition, patterns of recovery from WUD were similar.
298 / Anshel and Wrisberg
Table 1
Prerest and Postrest Scores for Heart Rate,
CAS-A Positive, and CAS-A Negative
Heart rate
Group
Control
Swing
Image
Strike
Run
Rest
interval
5
15
5
15
5
15
5
15
5
15
Prerest
M SD
107
110
103
105
104
104
102
101
111
110
11
9
10
9
9
9
12
11
10
10
GAS-A positive
Postrest
M SD
73
62
96
100
86
72
88
78
107
104
9
7
10
8
8
6
9
8
7
9
Prerest
Postrest
M
SD
M
SD
4.5
4.3
5.0
4.9
4.5
4.5
4.9
4.8
5.2
4.8
.72
.66
.57
.51
.46
.49
.59
.52
.56
.51
3.8
3.7
5.8
5.5
4.4
4.6
5.0
4.8
5.4
4.7
.64
.58
.58
.51
.70
.78
.65
.50
.56
.43
CAS-A negative
Prerest
M SD
3.0
3.2
2.8
2.6
2.5
2.5
2.8
2.8
3.2
3.3
.77
.76
.38
.39
.57
.50
.62
.59
.68
.74
Postrest
M
SD
3.8
3.9
2.4
2.4
2.5
2.5
2.7
2.8
3.0
3.5
.71
.76
.37
.39
.69
.70
.65
.58
.70
.80
Somatic and Cognitive Arousal
In order to determine the influence of the various warm-up activities on
somatic and cognitive arousal, separate 5 x 2 (Groups x Rest Intervals) ANOVAs
with repeated measures on the second factor were performed on difference scores
calculated for heart rate, CAS-A positive, and CAS-A negative measures. Each
score was determined by subtracting the value obtained at the end of the rest
interval from that obtained at the beginning. Descriptive statistics for these three
measures as a function of warm-up condition and rest interval length are presented
in Table 1. For heart rate, significant effects were found for groups, F(4, 65) =
86.76, p < .01, w2 = .73; rest intervals, F(l, 65) = 108.35,p < .01, w2 = .05; and
the Groups x Rest Intervals interaction, F(4,65) = 24.98, p < .01, w2 = .05. Post
hoc analysis of the groups effect using Tukey's procedure revealed that (a) the
pre-post difference for Group Control was significantly (p < .05) greater than
that of all warm-up conditions, (b) the difference for Group Image was significantly greater than that of the other warm-up conditions, and (c) the difference
for Group Strike was significantly greater than that for Groups Run and Swing.
Thus, it appeared that all warm-up activities increased somatic arousal
when compared to the no-warm-up control condition and that only running in
place and practice swings restored heart rate to prerest levels. With respect to
the significant rest intervals effect, heart rate decreased more over the longer
rest period than over the shorter one. However, simple main effects analysis of
the Groups x Rest Intervals interaction revealed that only Group Control and
Group Image had significantly (p < .05) greater decreases in heart rate over the
15-min interval than over the 5-min interval. These results indicate that, of the
various warm-up activities, imagery was the least effective in restoring somatic
arousal.
Warm-up Decrement / 299
A significant effect of groups was found for both CAS-A positive and
CAS-A negative scores, F(4, 65) = 18.31, p < .01, 0' = .46, and F(4, 65) =
36.65, p < .01, a2= .57, respectively. For the former measure, only Group Control
experienced a reduction in positive arousal. The results of the post hoc test
revealed significant (p < .05) differences between this group and all warm-up
conditions and between Group Swing (which had the largest increase in positive
arousal) and the other warm-up conditions. For CAS-A negative scores, Group
Control was the only one to report an increase. Post hoc results indicated that
the score of this group was significantly (p < .05) different from those of all
warm-up conditions and that the score of Group Swing (which had the largest
decrease in negative arousal) was significantly different from those of the other
warm-up conditions, with the exception of Group Strike. Taken together, these
findings suggest that all warm-up activities restored cognitive arousal to levels
found near the end of initial performance trials, with the most prominent influence
resulting from practice swings.
To determine the extent of relationship between arousal level and immediate
postrest performance, multiple regression analyses were performed for each rest
interval condition using the difference scores for the three arousal measures as
predictor variables and using the performance accuracy score on Trial 21 as the
dependent ~ a r i a b l eThe
. ~ R-square selection method (Freund & Littell, 1991) and
Mallows's (1973) C(p) statistic were used to determine the most useful subset
model for each rest interval. For the 5-min interval, a three-term model was
identified in which the interaction of heart rate and CAS-A positive scores was
the best predictor of postrest performance (R2= .46) followed by heart rate (R2 =
.53) and finally by the interaction of CAS-A positive and CAS-A negative scores
(R2 = .59). For the 15-min interval, a two-term model was selected in which
heart rate was the best predictor of postrest performance (R2= .49) followed by
CAS-A negative scores (R2 = .64). Tests for multicollinearity revealed small
variable inflation factor values (VIFs < 2) for both models. Taken together, these
results suggest that, for both the shorter and longer rest intervals, postrest serving
accuracy was higher when postrest somatic arousal approached or surpassed
prerest arousal levels. To a lesser degree, maintenance of positive cognitive
arousal was related to higher performance following the shorter interval, whereas
suppression of increased negative cognitive arousal was associated with serving
accuracy following the longer interval.
Discussion
In the present study an attempt was made to determine the relative effectiveness of various warm-up activities in reducing postrest performance decrement
(referred to as warm-up decrement, or WUD) in the tennis serve. In addition, an
effort was made to determine (a) the influence of different warm-up tasks on
somatic and cognitive arousal, and (b) whether the length of the rest interval
interpolated between performance bouts influences WUD. Experienced players
performed 20 serves aimed at a designated target area in the ad court. They then
rested for 5 or 15 min, the last two of which involved additional rest or the
performance of one of four warm-up activities. Each warm-up task (a) emphasized
a possible support system (Nacson & Schmidt, 1971) underlying performance
300 / Anshel and Wrisberg
of the serve and (b) was considered practical for use by players under normal
competitive conditions.
Although the prerest performance accuracy of all groups was comparable,
WUD was evidenced in the immediate postrest performance of the no-warm-up
control condition as well as that of the warm-up groups that mentally imaged
successful performance of the serve, hit the ball against the ground and caught
it, and ran in place. The only warm-up task that eliminated postrest WUD was that
involving practice swings of the serve. Interestingly, this activity also produced
significantly greater restorations of positive cognitive arousal and, except for
running in place, somatic arousal and a significantly greater reduction of negative
cognitive arousal than did all other rest-interval activities.
The most plausible explanation for the superiority of practice swings as a
warm up was that it involved an activity that was more similar to the criterion
task, even though no ball was used during the performance of this task. For
closed skills (Poulton, 1957) like the tennis serve, in which a high degree of
environmental stability exists, the performer is able to devote more attention to
response selection and execution. Therefore, it should not be surprising that the
best warm up is one that simulates the conditions of the criterion movement as
closely as possible.
Recent models dealing with strategies for closed-skill preparation (e.g.,
Boutcher & Rotella, 1987; Singer, 1986) emphasize the importance of creating
a positive feeling state about the movement that is about to be performed. In
particular, Boutcher and Rotella (1987) recommend a "kinesthetic coupling"
(p. 134) of the look and feel of the upcoming movement. One explanation, then,
for the superiority of practice swings in the present study is that, among other
things, it may have energized the sensation associated with effective serve performance. It is also possible that this activity facilitated the adjustment of a greater
number of support systems comprising the activity set (Nacson & Schmidt, 1971)
for the serve. For example, taking practice swings produced restorations (i.e.,
increases) in somatic arousal and positive cognitive arousal, which likely contributed to the reinstatement of appropriate (i.e., prerest levels) of physiological and
psychological activation. It is also possible that subjects taking practice swings
may have engaged in some incidental imagery, although this appears less likely
given the brief time (i.e., 5 s) between practice attempts.
Of the remaining warm-up activities, striking and catching the ball
seemed to exert the most influence on postrest performance, suggesting that
hand-eye coordination may be one support system of the activity set for
serving. Running in place successfully restored prerest levels of somatic
arousal but had only a modest influence on postrest performance accuracy.
This result conflicts with the findings of earlier research in which somatically
arousing activity was found to eliminate WUD in gymnastics vaulting (Anshel,
1985), thereby suggesting that the performance of some skills may be more
dependent on the maintenance of somatic arousal levels than is that of others
(see Martens & Landers, 1970, for a discussion). Although somatic arousal
was found to be the best predictor (of the three arousal measures) of immediate
postrest performance accuracy, it appears that a warm-up designed solely to
increase somatic arousal levels (e.g., running in place) does not always guarantee reductions in postrest WUD.
Least effective of the warm-up activities was the imaging of successful
Warm-up Decrement / 301
execution of the task. That an imagery warm-up failed to diminish WUD is
consistent with the results of Anshel(1985), who found no influence of imaging
on postrest WUD for the handspring vault. However, the present findings conflict
with those of Ainscoe and Hardy (1987) and Anshel and Wrisberg (1988) who
reported reductions in WUD following positive imaging of double leg circling
on the pommel horse and baseball batting, respectively. The inconsistency of
results across these studies may be due to differences in the imagery ability of
subjects. Although all of the subjects in the mental imagery group in the present
study indicated that they (a) regularly incorporated imagery in their training, (b)
complied with imagery instructions, and (c) were able to produce relatively vivid
images during their warm up, no measure of imagery ability was obtained.
unfortunately, this seems to be the case in most investigations examining the
role of imagery in reducing WUD.
Recent research by Goss, Hall, Buckolz, and Fishbume (1986) suggests
that the imagery ability of subjects should not be taken for granted. In that study,
it was discovered that mental rehearsal of a movement was more advantageous
for individuals who had high levels of imagery ability than for those with low
imaging ability. Thus, to more accurately determine the relative efficacy of
imagery as a warm-up activity, future research should include assessments of
the imagery ability or imagery experience of subjects (see Smith, 1987, for a
review of sport imagery issues).
Except for the findings that heart rate declined more for control and imagery
subjects over 15 rnin of rest than over 5 rnin and that the performance accuracy
of the imagery group was lower after the longer rest interval than after the shorter
one, rest interval length did not appear to have much influence on the results of
the present investigation. This finding is consistent with the results of Wrisberg
and Anshel(1993) who compared 5- and 15-min intervals in an investigation of
WUD in the performance of tennis groundstrokes. In other previous research,
the existence of WUD has been demonstrated after rest periods as short as 10
min and as long as 20 rnin (Ainscoe & Hardy, 1987; Anshel, 1985; Anshel &
Wrisberg, 1988; Nacson & Schmidt, 1971; Schmidt & Nacson, 1971; Schmidt &
Wrisberg, 1971). Anshel and Wrisberg (1988) reported postrest performance
decrements after both a 15-min rest and a 6-month lay-off. However, in the latter
instance. it was concluded that the decrement in ~erformancewas due to losses
in memiry strength rather than to the insufficient Bdjustment of support systems.
Taken together, then, the available evidence suggeststhat WUD may be evidenced
after as little as 5 min of rest and as much as 20 min, although the minimum
and maximum duration of rest intervals that produces WUD remains to be
determined.
In summary, the results of the present study suggest that an effective warmup for the tennis serve should include rehearsal of the mechanical properties of
the criterion task. However, it must also be noted that all four warm-up activities
(when compared to the control condition) produced significant increases in somatic arousal and positive cognitive arousal, significant decreases in negative
cognitive arousal, and significantly higher immediate postrest performance.
Therefore, it appears that support systems (Nacson & Schmidt, 1971) for the
tennis serve may also include hand-eye coordination, arousal level, and, to a
lesser extent, mental imagery. Finally, it is acknowledged that, although the
warm-up tasks used in the present study were chosen because they could be
302 / Anshel and Wrisberg
performed by tennis players in applied settings, it remains to be determined
whether such activities, alone or in combination, reduce postrest WUD (e.g.,
following a change of sides) during actual competition.
References
Adams, J.A. (1961). The second facet of forgetting: A review of warm-up decrement.
Psychological Bulletin, 58, 257-273.
Ainscoe, M., & Hardy, L. (1987). Cognitive warm-up in a cyclical gymnastics skill.
International Journal of Sport Psychology, 18, 269-275.
Anshel, M.H. (1985). The effect of arousal on warm-up decrement. Research Quarterly
for Exercise and Sport, 56, 1-9.
Anshel, M.H., Freedson, P., Hamill, J., Haywood, K., Horvat, M., & Plowman, S.A.
(1991). Dictionary of the sport and exercise sciences. Champaign, IL: Human
Kinetics.
Anshel, M.H., & Wrisberg, C.A. (1988). The effect of arousal and focused attention on
warm-up decrement. Journal of Sport Behavior, 11, 18-31.
Astrand, P.O. (1986). Textbook of work physiology (3rd ed.). New York: McGraw-Hill.
Boutcher, S.H., & Rotella, R.J. (1987). A psychological skills educational program for
closed-skill performance enhancement. The Sport Psychologist, 1, 127-137.
Freund, R.J., & Littell, R.C. (1991). SAS system for regression. Cary, NC: SAS Institute.
Goss, S., Hall, C., Buckolz, E., & Fishbume, G. (1986). Imagery ability and the acquisition
and retention of motor skills. Memory and Cognition, 14, 469-477.
Harris, D.V., & Harris, B.L. (1984). The athlete's guide to sports psychology: Mental
skills for physical people. Champaign, IL: Human Kinetics.
Mallows, C.P. (1973). Some comments on C(p). Technometrics, 15, 661-675.
Martens, R., & Landers, D.M. (1970). Motor performance under stress: A test of the
inverted-U hypothesis. Journal of Personality and Social Psychology, 16, 29-37.
Martens, R., Burton, D., Vealey, R.,Bump, L., & Smith, D. (1990). The development of
the Competitive State Anxiety Inventory-2 (CSAI-2). In R. Martens, R.S. Vealey, &
D. Burton, Competitive anxiety in sport (pp. 117-190). Champaign, IL: Human
Kinetics.
Nacson, J., & Schmidt, R.A. (1971). The activity-set hypothesis for warm-up decrement.
Journal of Motor Behavior, 3, 1-15.
Poulton, E.C. (1957). On prediction in skilled movements. Psychological Bulletin, 54,
467-478.
Schmidt, R.A. (1972). Experimental psychology. In R.N. Singer (Ed.), The psychomotor
domain: Movement behavior (pp. 18-55). Philadelphia: Lea and Febiger.
Schmidt, R.A., & Nacson, J. (1971). Further tests of the activity-set hypothesis for warmup decrement. Journal of Experimental Psychology, 90, 56-64.
Schmidt, R.A., & Wrisberg, C.A. (1971). The activity-set hypothesis for warm-up
decrement in a movement speed task. Journal of Motor Behavior, 3, 318-325.
Singer, R.N. (1986). Sports performance: A five-step mental approach. Journal of Physical
Education and Recreation, 57, 82-84.
Smith, D. (1987). Conditions that facilitate the development of sport imagery training.
The Sport Psychologist, 1, 237-247.
Tolson, H. (1980). An adjunct to statistical significance: d.Research Quarterly for
Exercise and Sport, 51, 580-584.
'OZE-I 1s '6s '273.f8010~31fid
U13V 'SIIqS lOlOUI alaIas!p $0
~
-v-x'lp!urq3~18 "M-V '!uourtes "v-3'%Iaqs!lM
Su!ua[ ayl u! slaaga d n - m '(5LfjI)
-59-6s 'V9
'gods puv aspdaxg JOYXlnamrma y3dvasay - I I ! ~ Suado ue u! luauraI3ap dn-WM
s!saqlodllq $ ~ S - ~ ! AaYv~30E lSal pIa!J '7j '(£661) -H'W '1qSUV ''v-3' 8 1 a q s ~ ~
IOJ