A dimensional investigation of the State Anxiety Inventory (SAI) in an exercise setting: cognitive vs. somatic.
Bixby, Walter R. ; Hatfield, Bradley D.
The anxiolytic effects of acute aerobic exercise have been widely
reported (Bahrke & Morgan, 1978; Buckworth & Dishman, 2002;
Landers & Arent, 2001; Morgan, 1979; Petruzzello, Landers, Hatfield,
Kubitz, & Salazar, 199 l; Raglin, 1997). The primary instrument
employed in such studies has been the state anxiety sub-scale (SAD of
the State-Trait Anxiety Inventory (STAI; Spielberger, 1983). Morgan
(1979) first elaborated on the delayed self-reported anxiolytic effect
of aerobic exercise, which typically occurs between 5 and 30 min
following work, and the magnitude of the response appears similar for
work of varying intensity (Cox, 2000; Ekkekakis, Hall, &
Petruzzello, 1999; Petruzzello et al., 1991). However, the pattern of
change during and immediately following exercise appears to be more
intensity dependent (Ekkekakis et al., 1999; Petruzzello et al., 1991).
As such, self-reported anxiety typically increases during and
immediately following exercise of higher intensity (i.e., ~ 75% of VO2
max) while showing little increase, and in some cases a decrease, during
and immediately following exercise of lower intensity (i.e., < ~ 65%
of VO2 max). The SAI was originally developed to assess anxiety
responses to various psychologically threatening stimuli and to this end
the validity and reliability of the scale is well established. However,
the SAI was not designed specifically to assess exercise-induced state
anxiety changes in which physiological arousal is inherently increased
as a result of ergogenic demand.
Recently, the SAI has drawn criticism as a valid measure of state
anxiety in an exercise setting (for full review see Ekkekakis et al.,
1999). Beyond the points raised previously in the literature, it appears
that the main criticism of the scale is an inability to distinguish
arousal from ergogenic demands and that associated with cognitive
appraisal of threat. Despite these concerns, researchers have continued
to use the SAI in an exercise setting with little to no discussion of
the difference between increased arousal from ergogenic demands and that
associated with a cognitive appraisal of threat (Cox, Thomas, &
Davis, 2000; Focht & Hausenbals, 2004; Kahn, Marlow, & Head,
2008; Kraemer & Marquez, 2009; Masters, LaCaille, & Shearer,
2003; Motl, O'Connor, & Dishman, 2004; Szabo, 2003).
This distinction is not trivial as Spieiberger, Lushene, and McAdoo
(1977) stated that increases in state anxiety occur when a cognitive
appraisal of threat is made with a concomitant increase in the autonomic
system activity (i.e., a physiologic or somatic response; Ekkekakis et
al., 1999, Spielberger et al., 1977). Thus, the SAI incorporates items
that assess perceived autonomic arousal (i.e., tense, calm) in addition
to those that assess cognitive appraisal of threat (i.e., worried,
frightened). In this manner, the self-report nature of the scale may be
insensitive to the distinction between arousal of a psychogenic nature
and that of an ergogenic nature. In this manner the alteration in the
pattern of SAI response during work of varying intensity may be due to
the influence of metabolic demand rather than anxiety per se.
In the majority of experimental settings, this possibility is not a
limitation, as the stimuli employed typically involve arousal of a
psychogenic nature (i.e., perceived threat and autonomic activity).
However, this is likely not the case in an acute exercise setting as the
ergogenic nature of the work stimulus (i.e., exercise) also causes an
increase in self-reported arousal that would be difficult to distinguish
from that caused by a perception of threat with employment of a
self-report instrument. As such, self-report items that assess autonomic
arousal may increase during and immediately following the negotiation of
exercise without a concomitant increase in items that assess cognitive
appraisal of threat and misrepresent the actual state anxiety response.
Accordingly, it would be logically expected that anxiety, as measured by
the SAI, would increase during and immediately following acute exercise
stress of sufficient intensity. And this is indeed the pattern of
response typically described in studies that use the SAI during and
immediately following exercise. However, this often reported observation
might be driven by the nature of the instrument, and not an actual
increase in state anxiety.
The majority of factor-analytic investigations of the SAI have
revealed a two-factor structure (Brown & Duren, 1988; Sherwood &
Westerback, 1983; Spielberger, 1983; Spileberger, Vagg, Barker, Donham,
& Westberry, 1980; Vagg, Spielberger, & O'Hearn, 1980),
which is manifest as anxiety-present (i.e., cognitive appraisal of
threat) and anxiety-absent items (i.e., somatic or autonomic arousal:
Ekkekakis et al., 1999) but the SAI scale is typically scored as a
unidimensional measure. Interestingly, Spielberger, Golzalez, Taylor,
Algaze, and Anton (1978) acknowledged the two-dimensional nature of
anxiety (distinguishing between anxiety present and anxiety absent
items) in an earlier report; however, when the STAI was revised
(Spielberger, 1983), no attempt was made to recognize the scoring of the
two dimensions. As such, the SAI fails to identify the two factors
separately and, in essence, the instrument yields a composite score, or
a one-dimensional measure, comprised of anxiety-present and
anxiety-absent items.
Related to this point, a number of researchers have suggested that
affect is made up of two consistent dimensions, a valence dimension and
an arousal dimension (Russell, 1980; Thayer, 1967, 1986; Watson, Clark,
& Tellegen, 1988; Watson & Tellegen, 1985). Additionally,
numerous instruments have been developed to examine the multidimensional
nature of affect (e.g, AD-ACL; Thayer, 1986; PANAS; Watson &
Tellegen, 1985; Watson et al., 1988). Thus, based on the structure of
affect, these investigators have attempted to account for the
multidimensionality in affective response. No such attempt to account
for the multidimensionality of affect has been made in the SAI.
Ekkekakis et al. (1999) conducted two investigations into the internal
consistency of the SAI in an exercise setting. In the first examination,
using the eight-item version of the SAI, a divergent pattern of response
for individual scale items was found, suggesting poor inter-item
reliability or non-homogeneity between items. In the second
investigation, using the ten-item version of the SAI, similar item
divergence was found. Interestingly, in both investigations, the overall
pattern of response for the SAI to exercise was consistent with that
commonly reported. The fmdings indicate that the two-dimensional nature
of the SAI could misrepresent self-reported anxiety in a physiologically
arousing setting when reported as a single outcome score (Ekkekakis et
al., 1999).
One logical way to examine the construct validity of the SAI in
response to exercise would be to assess the two dimensions of the SAI
(cognitive and somatic) separately. Theoretically, if the SAI is not
confounded by the ergogenic demands of exercise, than both the cognitive
and somatic factors should respond in a similar manner to the exercise
stimulus. Conversely, if the SAI is confounded by the ergogenic demands
of exercise, the cognitive and somatic factors would be expected to
respond differently. As such, if the factors of the SAI do indeed
respond differently, the cognitive factor would mirror the perception of
threat, while the somatic factor would mirror the ergogenic demands of
the exercise stimulus. Thus, in a young healthy population, little
change in the cognitive factor would be expected during the work
stimulus, which is unlikely to be viewed as threatening. A reduction
following exercise would likely occur due to a number of factors such as
the time-out hypothesis or the attendant physiological changes such as
cortical relaxation, thermogenic influences, or alterations in biogenic
amines (Hatfield, 1991; Bahrke & Morgan, 1978). However, the somatic
factor would be expected to increase during the exercise stimulus in
proportion to the intensity of the work (i.e., higher somatic score
during higher intensity exercise). Also the elevation in the somatic
factor would likely persist for a period of time following the exercise,
as the associated physiologic arousal would be maintained as a result of
the work involved. As the recovery period progresses, and physiologic
arousal decreases, a decrease in the somatic factor would then be
expected; however, the time course for this change would be dependent on
the intensity of the exercise. To date, no attempt has been made to
examine the SAI in a two-dimensional manner in an exercise setting,
separating out the cognitive (anxiety present) and somatic (anxiety
absent) factors.
Therefore, the purpose of this investigation was to examine the
two-dimensional nature of the SAI before, during, and following acute
bouts of exercise. Specifically, participants completed two 30-min bouts
of aerobic exercise at low- and high-intensity. The SAI was administered
at numerous points throughout the protocol allowing for thorough
temporal resolution of self-reported anxiety. Additionally, the two
dimensions (i.e., anxiety-present and anxiety-absent items) were
examined separately to determine if the cognitive and somatic factors
would exhibit different response patterns. We predicted that the overall
response pattern for both low- and high-intensity exercise would be
similar to that typically reported in the literature. For low-intensity
exercise we predicted little to no change during and a significant
reduction following the work, while for high-intensity exercise we
predicted a significant increase during and a similar reduction
following the work. Examination of the two factors separately was
expected to reveal a divergence in response pattern. Specifically, the
cognitive factor was expected to show little to no change during
exercise accompanied by a reduction following both low- and
high-intensity work. Alternately, the somatic factor was expected to
significantly increase during both low- and high-intensity exercise,
with higher elevation for the latter. Following exercise, a return to
baseline in the somatic factor was expected for both intensities but
that the time course would be delayed for higher intensity work.
Methods
Participants
Participants were 32 (20 female) healthy, nonsmoking volunteers
recruited from kinesiology classes at a large university on the east
coast of the United States. Participant characteristics are presented in
Table 1. Participants were informed of the requirements of the study and
provided written consent on a form approved by the institutional review
board.
Psychological Instruments
Anxiety was assessed with the State Anxiety Inventory (SAI) of the
State-Trait Anxiety Inventory (STAI: Spielberger, 1983). The SAI was the
10-item version that could yield a score between 10 and 40. A higher
score was representative of greater anxiety. Speilberger (1983) found
that alpha coefficients for the State Inventory were above .90
suggesting strong internal reliability and validity of the scale was
established by showing that scores increased when subjects were put in
anxiety provoking situations. The 10-item short form was used to
decrease test reactivity as multiple measures were employed. The SAI was
recorded 10-min before the start of exercise, at the initiation of
exercise, 10-, 20-, and 30-min during exercise, and 10-, 20-, and 30-min
during recovery for a total of eight recording times.
Procedures
The experiment involved testing on three separate days. Prior to
each visit to the laboratory participants were instructed to refrain
from eating within 2 hr of testing, to avoid exercising and the
consumption of alcoholic and caffeinated beverages on the day of
testing, and to be well rested (i.e., to obtain 8 hr of sleep the night
before testing).
Aerobic Capacity and Ventilatory Breakpoint Testing
On the first day of testing participants were given a brief
description of the study and provided informed consent. Height and
weight were obtained. Participants then completed a graded exercise test
on a cycle ergometer (Monark Exercise AB, model 818, Sweden) to
determine aerobic capacity. The exercise test began with a warmup period
during which the participant cycled for 3 min with no load. Thereafter,
the load (determined by output wattage) was increased every min until
voluntary exhaustion was reached. The increase in load was based on the
participant's exercise history and weight (Wasserman, Hansen, Sue,
Whipp, & Casaburi, 1994). ECG was monitored throughout via a
Sensormedics VMAX 229 metabolic cart (Sensormedics, Inc., Yorba Linda,
CA) from pregelled, disposable Ag/AgCL electrodes (Marquette Medical
Systems, #900703-230, Jupiter, FL) attached to the participant in a V5
configuration. Ratings of perceived exertion (RPE; Borg, 1982) were
obtained every minute during the test. Expired gases were analyzed with
a calibrated Sensormedics VMAX 229 metabolic cart (Sensormedics, Inc.,
Yorba Linda, CA) to obtain breath-by-breath averages of minute volume
and fractional gas concentrations of oxygen and carbon dioxide. An
estimate of V[O.sub.2] max was deemed valid if two of the following
three criteria were met: 1) heart rate equaled the agepredicted maximum
+/- 10 bpm, 2) the increase in oxygen consumption was less than 150 ml
with an increase in workload, and 3) the respiratory quotient exceeded
1.10 (Taylor, Buskirk, & Henschel, 1955). All participants met these
criteria. The calculation of V[O.sub.2] max (ml*[kg.sup.-1] *
[min.sup.-1]) was based on the highest oxygen consumption value
obtained. The ventilatory breakpoint was defined as the percentage of
aerobic capacity associated with an upward deflection in VE/ V[O.sub.2]
without a change in VE/VC[O.sub.2] (Wilmore & Costill, 1994) and was
chosen with the use of VMAX229 software (Sensormedics, Inc., Yorba
Linda, CA). The calculated values were also compared to values derived
from visual inspection of the data for accuracy.
Exercise Testing
On the second and third visits to the laboratory participants
completed 30 min of steady state exercise at either a low or high
intensity. The order of exercise intensities was randomly assigned and
counterbalanced across participants. Except for exercise intensity, the
procedures on the second and third days were the same. The participant
was seated on a recumbent cycle (Lifecycle, Inc., model 9500R, Franklin
Park, IL) that was equipped with toe straps to secure the
participant's feet on the pedals and the seat position was adjusted
to maximize the efficiency and comfort of pedaling (i.e., such that the
knee was bent slightly when the foot pedal was fully depressed), after
which the experiment was initiated. The protocol involved three
contiguous periods: a 15-min baseline, a 30-min bout of exercise, and 30
min of recovery. Throughout the experiment the participant was seated on
the cycle with the feet secured on the pedals. Participants were
instructed to sit quietly during the baseline. After the baseline ended
the participant began pedaling at 75 rpm and 67 watts. The load (i.e.,
wattage) was progressively increased during the first 5 min of the
exercise period to bring the participant's heart rate to the
appropriate level. In the high intensity condition participants worked
just below a target heart rate that was associated with the ventilatory
breakpoint determined during the exercise capacity test (heart rate M=
151.5 bpm, SD = 8.4; RPEM= 15.1, SD = 1.7). In the low intensity
condition participants maintained heart rate at a level corresponding to
75% of that observed at ventilatory breakpoint (heart rate M= 111.6 bpm,
SD = 6.9; RPEM= 10.4, SD = 2.1). Heart rate was derived from continuous
ECG recordings (Hewlett Packard, model 78352C) using a V5 configuration
and workload was adjusted accordingly to ensure that the participant
worked at the targeted intensity. Upon completion of the exercise bout
the participant was allowed 2 5 min of active recovery (i.e., pedaling
< 60 rpm with no load), after which he or she sat quietly until the
end of recovery. SAI, heart rate, and RPE measures were taken 5 min into
baseline and every 10 min thereafter for a total of 8 recordings during
each exercise session. The recordings corresponded to the 5-min baseline
measure, initiation of exercise, 10-min into exercise, 20-min into
exercise, 30-min into exercise, 10-min into recovery, 20-min into
recovery, and 30-min into recovery.
Results
Statistical Analyses
An initial analysis was conducted to verify that the participants
exercised at the targeted intensities. Heart rate and RPE were subjected
to separate 2 (intensity: low vs. high) x 8 (time) multivariate analyses
of variance (MANOVA: Wilks' Lambda) with repeated measures on both
factors. To examine changes in anxiety, SAI scores were subjected to a 2
(intensity: low vs. high) x 8 (time) MANOVA (Wilks' Lambda) with
repeated measures on both factors. The factor scores were then subjected
to a 2 (dimension: cognitive and somatic) x 2 (intensity: low vs. high)
x 8 (time) MANOVA (Wilks' Lambda) with repeated measures on all
three factors. Where appropriate, post-hoc analysis was performed using
a Student Neuman-Keuls procedure and effect sizes were calculated for
the two dimensions of the SAI to determine the magnitude of any
significant differences.
Heart Rate and Ratings of Perceived Exertion
MANOVA revealed a significant intensity x time interaction for both
heart rate, F (7, 25) = 122.76,p < .001; and RPE, F(7, 25) = 17.34,p
< .001. Post hoc analysis showed that there was no difference in
heart rate or RPE at baseline between the low-intensity and high
intensity. Heart rate and RPE were significantly elevated from baseline
during both exercise sessions; however, a larger elevation occurred
during the high-intensity exercise, relative to the low intensity for
both heart rate and RPE. Following exercise, heart rate decreased in
both the low-and high-intensity exercise; however, the decrease occurred
more rapidly in the low-intensity exercise relative to the
high-intensity (see figure 1A). RPE following exercise decreased below
exercise levels for both the low- and high-intensity conditions and the
decrease was not different between the conditions (see figure 1B).
State Anxiety Inventory scores
MANOVA revealed a significant intensity x time interaction, F (7,
25) = 6.23, p < 0.001 for the overall SAI scores. Post hoc inspection
revealed no difference between the low- and high-intensity scores at
both baseline and recovery. During exercise, scores were significantly
higher in the high-intensity relative to the low-intensity condition.
For the low-intensity condition, SAI scores did not change from baseline
to exercise and then significantly decreased during recovery. For the
high-intensity condition, SAI scores significantly increased from
baseline to exercise and significantly decreased below baseline levels
during recovery. See figure 2 for details.
Cognitive and Somatic Factors of the SAI
To examine the dimensions of the SAI, factor scores were calculated
by summing the 5 items that make up the cognitive (4, 5, 6, 7, & 9)
and somatic (1, 2, 3, 8, & 10) factors. This yielded possible scores
of between 5 and 20 for each factor. Following calculation of factor
scores, a 2 (dimension) x 2 (intensity) x 8 (time) MANOVA (Wilks'
Lambda) was performed, which revealed a Dimension x Intensity x Time
interaction, F(7, 25) = 4.26, p = .003. Post hoc analysis showed that
for both the low- and high-intensity conditions, the somatic factor was
significantly higher than the cognitive factor. For the somatic factor,
baseline and recovery scores were not different between the low- and
high-intensity conditions; however, scores during exercise were
significantly higher in the high-intensity condition, relative to the
low-intensity condition. For low-intensity, somatic factor scores
increased relative to baseline at the start and 10-min into exercise,
returned to baseline level at 20- and 30-min into exercise and 10- and
20-min into recovery and then decreased below baseline level at 30-min
into recovery. For high-intensity, somatic factor scores remained at
baseline level at the start of exercise, increased relative to baseline
at 10-min, 20-min, and 30-min into exercise, and then decreased below
baseline at 10-min, 20-min, and 30-min into recovery. For the cognitive
factor, scores were not different at all time points between the low-
and high-intensity conditions. For the low-intensity condition,
cognitive scores remained at baseline level throughout the exercise and
recovery, except for 30-min into recovery, which was lower than
baseline. For the high-intensity condition, cognitive factor scores
remained at baseline level at the start of exercise,
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
Discussion
Employment of the overall State Anxiety Inventory (SAI:
Spielberger, 1983) revealed an increase in scores during higher
intensity exercise followed by a decrease below baseline levels during
recovery. For lower intensity work, little to no change was observed
during exercise followed by a decrease below baseline levels during
recovery. This pattern of change for the overall SAI is in line with
other investigations, which have shown similar response patterns
(Landers & Arent, 2001; Morgan, 1979; Petruzzello, Landers,
Hatfield, Kubitz, & Salazar, 1991; Raglin, 1997). However, the
examination of the SAI as a two-dimensional measure revealed divergent
response patterns for the cognitive and somatic factors. This result
brings into question the use of the SAI as a composite measure of
anxiety in an exercise setting as it appears that the anxiety-absent
items fail to distinguish perceived arousal of a psychogenic nature and
that resulting from ergogenic demand. In this manner it would seem
desirable to score the anxiety-present (cognitive) and anxiety-absent
(somatic) items separately to assess the anxiolytic effects of exercise
in an unambiguous manner.
For the cognitive factor, there was little to no change during and
a decrease following both the low- and high-intensity exercise, thus the
negotiation of the exercise stimulus did not cause an increase in the
cognitive appraisal of anxiety while a significant decrease in the
cognitive appraisal of anxiety occurred following the exercise session.
Therefore, the changes in the cognitive factor of the SAI lend support
to the widely held belief of an anxiolytic effect following exercise.
The magnitude of this change was considered large based on effect size
calculations (.77 to .9, see table 2 for details). This finding appears
logical, as exercise would not be thought to cause an increase in the
cognitive appraisal of threat, which is necessary for an increase in
anxiety, in such a young healthy population. Additionally, the reduction
in the cognitive factor (i.e., the "feel better" effect;
Bahrke & Morgan, 1978) following exercise may be a result of a
"time-out" from the stresses of the daily routine (Bahrke
& Morgan, 1978) or possibly a result of the thermogenic effect of
the exercise stimulus (Hatfield, 1991). Thus, it appears that a
cognitive anxiolytic effect occurs following exercise; however, the
pattern of anxiolytic response may be different than that commonly
reported in the exercise psychology literature for the composite score,
in which anxiety increases during and immediately following exercise and
then decreases following the exercise session (Landers & Arent,
2001; Morgan, 1979; Petruzzello, et al., 1991; Raglin, 1997).
For the somatic factor, a significant increase occurred during both
the low- and high-intensity exercise followed by a significant decrease
during recovery. The elevation during exercise was greater for the
high-intensity exercise, relative to the low-intensity (see figure 3).
Interestingly, the pattern of change for the somatic factor is
strikingly similar to the pattern of change for both the heart rate and
RPE (Borg, 1982) response (see figure 4). The change in the somatic
factor occurred without a corresponding change in the cognitive factor
which suggests that the somatic factor is sensitive to physiologic
arousal not just from threatening situations, but also from
non-threatening situations that require an increase in ergogenic demands
(i.e., exercise). Therefore, the somatic factor appears to be influenced
by the ergogenic demands of the exercise, causing a misrepresentation of
the anxiety response. Thus, when using the SAI as an undifferentiated
measure of anxiety in an exercise setting, care must be taken to
determine if the changes in anxiety are actual concomitant changes in
the cognitive and somatic factors, or the result of a divergence in
response pattern for the two factors. Through examining the cognitive
and somatic factors separately, a clearer picture of the anxiety
response during and following exercise will develop.
In summary, this investigation showed the typical anxiety response
reported in the literature using the SAI to low- and high-intensity
exercise, with both intensities leading to a reduction in anxiety
following the exercise session. However, by examining the SAI as a
two-dimensional measure, it was further shown that the exercise stimulus
caused a divergence in response pattern for the two factors which make
up the SAI (i.e., cognitive and somatic). This divergence brings into
question the nature of anxiety change commonly reported in the exercise
psychology literature, as the majority of investigations have employed
the SAI.
However, this investigation did show an anxiolytic effect of
exercise, as the cognitive factor decreased significantly following both
low- and high-intensity exercise. This finding indicates that
individuals felt better (i.e., less cognitive anxiety) following the
exercise session and the magnitude of this finding was large. This is
important, as the anxiolytic effects of exercise have been widely touted
in both the scientific and popular literature as a benefit of
participation in exercise. The present investigation does not refute
this finding; however, the magnitude and pattern of the anxiety changes
during and following exercise of varying intensities needs to be
examined more closely to determine the influence of the increased
ergogenic demands of exercise. The present findings point to the need
for closer examination of the selfreport measures used in exercise
psychology, as the physiological requirements of exercise could confound
the affective response both during and following the exercise session.
As exercise psychologists, an essential goal of the field is the
discovery of the psychological changes associated with exercise, while
reducing the artifact associated with assessment instruments. Only
through a better understanding of the measures employed in exercise
psychology will we be able to more clearly understand the affective
response pattern during and following exercise.
[FIGURE 4 OMITTED]
References
Bahrke, M. S., & Morgan, W. P. (1978). Anxiety reduction
following exercise and meditation. Cognitive Therapy and Research, 2,
323-333.
Borg, G. A. V. (1982). Psychological bases of perceived exertion.
Medicine and Science in Sports and Exercise, 14, 377-381.
Buckworth, J. & Dishman, R. K. (2001). Anxiety. In Exercise
Psychology (pp. 115-130). Champaign, IL: Human Kinetics
Brown, M. T. & Duren, P. S. (1988). Construct validity for
Blacks of the State-Trait Anxiety Inventory. Measurement and Evaluation
in Counseling and Development, 21 25-33.
Cox, R. H., Thomas, T. R., & Davis, J. E. (2000). Delayed
anxiolytic effect associated with an acute bout of aerobic exercise.
Journal of Exercise Physiology (online), 3(4), 59-66.
Ekkekakis, P., Hall, E. E., Petruzzello, S. J. (1999). Measuring
state anxiety in the context of acute exercise using the State Anxiety
Inventory: An attempt to resolve the brouhaha. Journal of Sport and
Exercise Psychology, 21,205-229.
Focht, B. C. & Hausenblas, H. A. (2004). Perceived evaluative
threat and state anxiety during exercise in women with social physique
anxiety. Journal of Applied Sport Psychology, 16(4), 361-368.
Hatfield, B. H., (1991). Exercise and Mental Health: The mechanisms
of exercise-induced psychological states. In L. Diamant (Ed.),
Psychology of sports, exercise and fitness, social and personal issues
(pp. 17-50). Washington DC: Hemisphere.
Khan, R. S., Marlow, C., & Head, A. (2008). Physiological and
psychological response to a 12-week bodybalance training programme.
Journal of Science and Medicine is Sport, 11,299-307.
Kraemer, J. M .& Marquez, D. X. (2009). Psychological
correlates and outcomes of yoga or walking among older adults. The
Journal of Psychology, 143(4), 390-404.
Landers, D. M. & Arent, S. (2001). Physical activity and mental
health. In R. Singer, H. A. Hausnblas, & C. M. Janelle (Eds.),
Handbook of sport psychology (pp. 740-765). New York: John Wiley &
Sons, Inc.
Masters, K. S., LaCaille, R. A., & Shearer, D. S. (2003). The
acute affective response of type A behavior pattern individuals to
competitive and noncompetitive exercise. Canadian Journal of Behavioral
Science, 35(1), 25-34.
Morgan, W. P. (1979). Anxiety reduction following acute physical
activity. Psychiatric Annals, 9(3), 36-45.
Motl, R. W., O'Connor, P. J., & Dishman, R. K. (2004).
Effects of cycling exercise on the soleus H-reflex and anxiety among men
with low or high trait anxiety. Psychophysiology, 41, 96-105.
Petruzzello, S. J., Landers, D. M., Hatfield, B. D., Kubitz, K.A.,
& Salazar, W. (1991).A MetaAnalysis on the anxiety reducing effects
of acute and chronic exercise. Sports Medicine, 11(3), 143182.
Raglin, J. S. (1997). Anxiolytic effects of physical activity. In
Physical activity and mental health, ed. W. P. Morgan. Washington, DC:
Taylor & Francis.
Russell, J. A. (1980). A circumplex model of affect. Journal of
Personality and Social Psychology, 39, 1161-1178.
Sherwood, R. D. & Westerback, M. E. (1983). A factor analytic
study of the State-Trait Anxiety Inventory utilized with preservice
elementary teachers. Journal of Research in Teaching, 20, 225-229.
Spielberger, C. D. (1983). Manual for the State- Trait Anxiety
Inventory (form y). Palo Alto, CA: Consulting Psychologists Press.
Spielberger, C. D., Golzalez, H. P., Taylor, C. J., Algaze, B.,
& Anton, W. D. (1978). Examination stress and test anxiety. In C. D.
Spielberger & I. G. Sarason (Eds). Stress and Anxiety (Vol. 5, pp
167-191). Washington, DC: Hemisphere.
Spielberger, C. D., Lushene, R. E., & McAdoo, W. G. (1977).
Theory and measurement of anxiety states. In R. B. Cattell & R. M.
Dreger (Eds.), Handbook of modern personality theory (pp. 239-253). New
York: John Wiley & Sons.
Spielberger, C. D., Vagg, P. R., Barker, L. R., Donham, G. W.,
& Westberry, L. G. (1980). The factor structure of the State-Trait
Anxiety Inventory. In I. G. Sarason & C. D. Spieiberger (Eds.)
Stress and anxiety (Vol. 7, pp. 95-109). Washington, DC: Hemisphere.
Szabo, A. (2003). The acute effects of humor and exercise on mood
and anxiety. Journal of Leisure Research, 35(2), 152-162.
Taylor, H. L., Buskirk, E., & Henschel, A. (1955). Maximal
oxygen intake as an objective measure of cardiorespiratory performance.
Journal of Applied Physiology, 8, 73-80.
Thayer, R. E. (1967). Measurement of activiation through
self-report. Psychological Reports, 20, 663-678.
Thayer, R. E. (1986). Activation-Deactivation Adjective Checklist:
Current overview and structural analysis. Psychological Reports, 42,
747-756.
Vagg, P. R., Spielberger, C. D., & O' Hearn, T. P. Jr.
(1980). Is the State-Trait Anxiety Inventory multidimensional?
Personality and Individual Differences, 1,207-214.
Wasserman, K., Hansen, J. E., Sue, D. Y., Whipp, B. J., Casaburi,
R. (1994). Protocols for exercise testing. In Principles of Exercise
Testing and Interpretation (2nd ed.) (pp. 95 - 111). Malvern, PA: Lea
& Febiger.
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development
and validation of brief measures of positive and negative affect: The
PANAS scales. Journal of Personality and Social Psychology, 54 (6),
1063-1070.
Watson, D. & Tellegen, A. (1985). Toward a consensual structure
of mood. Psychological Bulletin, 98(2), 219-235.
Willmore, J. H. & Costill, D. L. (I 994). Physiology of sport
and exercise. Champaign, IL: Human Kinetics.
Address Correspondence to: Walter R. Bixby, Department of Exercise
Science, Elon University, CB 2525, Elon, NC 27244. E-mail:
wbixby@elon.edu
Walter R. Bixby
Elon University
Bradley D. Hatfield
University of Maryland
Table 1. Mean (SD) Participant Characteristics
Men (n = 12) Women (n = 20)
Age (years) 24.3 (2.9) 23.1 (3.3)
Height (cm) 182.6 (6.4) 167.4 (9.0)
Weight (kg) 77.7 (10.1) 61.6 (10.8)
V[O.sub.2]max 48.3 (8.6) 37.3 (7.5)
(m 1 * kg.sup.-1] * [min.sup.-1])
Table 2. Effect Size Calculations for the Somatic and Cognitive
Factors of the SAI across Time within Low and High Intensity
SAI Somatic to intensity Baseline Start 10 ex 20 ex 30 ex
Baseline * 0.42 0.64 0.31 0.32
Start of exercise * 0.23 0.1 0.07
10-min into exercise * 0.32 0.29
20-min into exercise * 0.02
30-min into exercise *
10-min into recovery
20-min into recovery
30-min into recovery
SAI Somatic to intensity 10 rec 20 rec 30 rec
Baseline 0.33 0.42 0.51
Start of exercise 0.78 0.88 0.96
10-min into exercise 1 1.09 1.18
20-min into exercise 0.64 0.73 0.82
30-min into exercise 0.65 0.74 0.82
10-min into recovery * 0.09 0.2
20-min into recovery * 0.11
30-min into recovery *
SAI Somatic hi intensity Baseline Start 10 ex 20 ex 30 e
Baseline * 0.25 1.16 1.16 1.08
Start of exercise * 0.95 0.95 0.87
10-min into exercise * 0.01 0.08
20-min into exercise * 0.08
30-min into exercise *
10-min into recovery
20-min into recovery
30-min into recovery
SAI Somatic hi intensity 10 rec 20 rec 30 rec
Baseline 0.51 0.44 0.6
Start of exercise 0.87 0.77 0.95
10-min into exercise 2.08 1.91 2.14
20-min into exercise 2.07 1.9 2.12
30-min into exercise 1.95 1.79 2.01
10-min into recovery * 0.06 0.13
20-min into recovery * 0.18
30-min into recovery *
SAI Co anitive to intensity Baseline Start 10 ex 20 ex 30 ex
Baseline * 0.17 0.5 0.71 0.67
Start of exercise * 0.35 0.57 0.52
10-min into exercise * 0.19 0.14
20-min into exercise * 0.06
30-min into exercise *
10-min into recovery
20-min into recovery
30-min into recovery
SAI Co anitive to intensity 10 rec 20 rec 30 rec
Baseline 0.79 0.73 0.9
Start of exercise 0.66 0.6 0.78
10-min into exercise 0.27 0.24 0.39
20-min into exercise 0.09 0.06 0.22
30-min into exercise 0.16 0.03 0.12
10-min into recovery * 0.03 0.14
20-min into recovery * 0.14
30-min into recovery
SAI Cognitive hi intensity Baseline Start 10 ex 20 ex 30 ex
Baseline * 0.27 0.68 0.55 0.52
Start of exercise * 0.39 0.26 0.23
10-min into exercise * 0.15 0.2
20-min into exercise * 0.05
30-min into exercise *
10-min into recovery
20-min into recovery
30-min into recovery
SAI Cognitive hi intensity 10 rec 20 rec 30 rec
Baseline 0.77 0.8 0.81
Start of exercise 0.48 0.49 0.52
10-min into exercise 0.08 0.09 0.16
20-min into exercise 0.25 0.27 0.31
30-min into exercise 0.31 0.33 0.37
10-min into recovery * 0 0.09
20-min into recovery * 0.11
30-min into recovery *
