Content uploaded by Daniel M. Landers
Author content
All content in this area was uploaded by Daniel M. Landers on Dec 29, 2014
Content may be subject to copyright.
Content uploaded by Chad D Rethorst
Author content
All content in this area was uploaded by Chad D Rethorst
Content may be subject to copyright.
392
The authors are with the Department of Kinesiology, Arizona State University, Tempe, AZ.
Journal of Sport & Exercise Psychology, 2008, 30, 392-410
© 2008 Human Kinetics, Inc.
The Anxiolytic Effects of Exercise:
A Meta-Analysis of Randomized Trials
and Dose–Response Analysis
Bradley M. Wipfli, Chad D. Rethorst, and Daniel M. Landers
Arizona State University
A meta-analysis was conducted to examine the effects of exercise on anxiety.
Because previous meta-analyses in the area included studies of varying quality,
only randomized, controlled trials were included in the present analysis. Results
from 49 studies show an overall effect size of −0.48, indicating larger reductions in
anxiety among exercise groups than no-treatment control groups. Exercise groups
also showed greater reductions in anxiety compared with groups that received
other forms of anxiety-reducing treatment (effect size = −0.19). Because only
randomized, controlled trials were examined, these results provide Level 1, Grade
A evidence for using exercise in the treatment of anxiety. In addition, exercise
dose data were calculated to examine the relationship between dose of exercise
and the corresponding magnitude of effect size.
Keywords: physical activity, mental health, anxiety
Anxiety is a mental disorder that has been estimated to affect approximately
8.3% of the U.S. adult population (Myers et al., 1984). Briefly, anxiety has been
defined as an unpleasant emotional state, which is often accompanied by fatiguing
or exhausting physiological symptoms (Greist & Jefferson, 2000). Anxiety disorders
vary in degree of severity from periods of mild discomfort to panic disorder, and
common symptoms include rapid heart rate, sweating, nausea, chills, trembling,
hyperventilation, a fear of doing something uncontrolled, and a fear of death.
The onset of an anxiety attack can arise from cues (phobias) or “out of the blue”
(spontaneous) (Greist & Jefferson).
Of the roughly 18 million adults affected by an anxiety disorder, it has been
estimated that only 23% are receiving treatment. Common treatments for anxiety
disorders are similar to those for clinical depression; it has been shown that 75% of
people who are affected by clinical depression are also affected by anxiety disorder
(Myers et al., 1984). The two most common treatments for these disorders are
various drug treatments (antidepressants) and psychotherapy. The use of drugs and
psychotherapy to treat anxiety disorders, however, has drawn criticism. Common
side effects that accompany antidepressants include weight gain, hyperglycemia,
Anxiolytic Effects of Exercise 393
hyperlipidemia, sexual dysfunction, and elevated blood pressure (Gardner, Bald-
essarini, & Waraich, 2005). Recently, antidepressant drugs classified as selective
serotonin reuptake inhibitors have been linked to increased risk for suicide or
suicidal thoughts in both children and adults (Ludwig & Marcotte, 2005).
In addition to these unwanted side effects, Andrews, Sanderson, Corry, and
Lapsley (2000) have described the need for more treatments for depression due to
the high percentage (approx. 50%) of people who may be clinically depressed but
are not receiving treatment. Because a large percentage of people who have some
type of anxiety disorder are not receiving treatment, the need for more forms of
treatment for depression that was described by Andrews can be extended to include
anxiety disorders as well. Given these criticisms, there is a general interest in finding
new methods of treatment for anxiety disorders that are effective and widely acces-
sible, and lack unwanted side effects. One potential method of treatment that has
generated a large amount of research is exercise (Moore & Blumenthal, 1998).
The effects of exercise on anxiety have been examined by previous meta-
analyses (Calfas & Taylor, 1994; Kugler, Seelbach, & Kruskemper, 1994; Landers
& Petruzzello, 1994; Long & Van Stavel, 1995; McDonald & Hodgdon, 1991;
Petruzzello, Landers, Hatfield, Kubitz, & Salazar, 1991; Schlicht, 1994). All these
meta-analyses reported that exercise was related to decreases in anxiety. More
specifically, in the Petruzzello et al. study, dependent measures of anxiety were
divided into three categories: self-reported state anxiety, self-reported trait anxi-
ety, and psycho-physiological measures of anxiety. While the effect sizes for all
three categories indicated that exercise was associated with a reduction in anxiety,
effect sizes ranged from 0.24 for state anxiety and 0.34 for trait anxiety, to 0.56
for psycho-physiological measures of anxiety. Results from moderator variables
in this study showed that studies using random assignment to groups had higher
effect sizes than other methods of subject assignment for measures of trait anxi-
ety. This same finding, however, was not seen for measures of state anxiety and
psycho-physiological measures of anxiety. Additional results from the analysis of
moderator variables showed that exercise was significantly better than other types
of treatment (i.e., relaxation or meditation) for reducing trait anxiety and psycho-
physiological measures of anxiety. These effect sizes were small in magnitude
(0.31 and 0.15, respectively).
Results from the Long and Van Stavel (1995) meta-analysis were similar to
those of Petruzzello et al. The overall effect sizes in this analysis were 0.45 for
within-group studies and 0.36 for studies that used a contrast group. Both of these
effect sizes indicate a moderate reduction in anxiety, and fall within the range that
was seen by Petruzzello et al. Moderator variable results from this quantitative
review also showed that higher effect sizes (0.64 for within-group studies and 0.44
for contrast group studies) resulted from studies that used random assignment to
groups.
To evaluate the efficacy of exercise in the treatment of anxiety disorders, it is
useful to examine the criteria for grading treatment recommendations that have
been refined by Guyatt, Cook, Sackett, Eckman, and Pauker (1998). Based upon
these criteria, a treatment that achieves Level 1, Grade A status receives the highest
level of recommendation. In order for evidence to reach Level 1, Grade A status,
the following criteria must be met: The evidence must come from randomized,
controlled trials; there must be a high benefit-to-risk ratio; the results must be clear
394 Wipi, Rethorst, and Landers
cut; and the evidence must come from a large sample. Large-scale randomized trials
are often extremely costly and time consuming to conduct and are therefore rare.
Level 1, Grade A evidence, however, can also be achieved through meta-analysis
by combining results of Level 2 studies, which are smaller randomized, controlled
trials (Guyatt et al.).
Because only 13 of the 104 studies in the Petruzzello et al. meta-analysis, and
15 of the 40 studies in the Long and Van Stavel meta-analysis were randomized,
controlled trials, they fail to establish Level 1, Grade A evidence. Therefore, the
purposes of this study were to update and strengthen earlier meta-analytic findings,
and to establish Level 1, Grade A evidence for using exercise as a treatment for
anxiety by including only randomized, controlled trials in the analysis.
Dose–Response Analysis
In the quest to satisfy all of Hill’s (1965) criteria for causation, there have been
several attempts to find evidence of a dose–response relationship between exercise
and the associated mental health benefits. Much of this evidence has been summa-
rized by Dunn, Trivedi, and O’Neal (2001). This article, however, argues that even
though there is evidence to suggest a dose–response relationship between exercise
and mental health benefits, much of this evidence comes from Grade B and Grade C
studies, which consist of mainly correlational evidence. There have also been more
recent studies that provide support for the existence of a dose–response relationship
in this area (Dunn, Trivedi, Kampert, Clark, & Chambliss, 2002; Galper, Madhukar,
Barlow, Dunn, & Kampert, 2006), which also provide correlational evidence. Two
randomized, controlled trials (Dunn, Trivedi, Kampert, Clark, & Chambliss, 2005;
Singh et al., 2005) have shown evidence of a dose relationship with regards to mental
health. Dunn et al. examined four exercise groups, each of which were assigned an
energy expenditure (either 17.5 kcal·kg−1·week−1 or 7.0 kcal·kg−1·week−1) and an
exercise frequency (either 3 or 5 days per week). Results showed that both of the
higher-dose groups showed significantly lower depression scores after the 12-week
intervention than the lower-dose groups, regardless of exercise frequency. Results
from Singh et al. showed that older adults who completed a higher-intensity weight
training program showed significantly less depression postintervention than older
adults who completed a lower-intensity weight training program.
Although the aforementioned evidence provides some support for a dose–
response relationship between exercise and improved mental health, further evi-
dence is required to strengthen this link. A secondary purpose of this study, therefore,
was to provide additional evidence in support of a dose–response relationship
between exercise and improved mental health by examining this relationship in the
randomized, controlled studies that were collected for the meta-analysis.
Hypotheses
Hypothesis 1. Exercise groups show significantly greater reductions in anxiety
than no-treatment and placebo control groups, as indicated by an overall effect
size that is significantly different from zero. Based on results from previous meta-
analyses, it is predicted that the overall effect size is moderate in magnitude.
Anxiolytic Effects of Exercise 395
Hypothesis 2. Based upon findings from Petruzzello et al. meta-analysis, it is
predicted that exercise groups show greater reductions in anxiety than groups who
underwent a different form of anxiety-reducing treatment. In addition, it is predicted
that this effect size will be small in magnitude.
Hypothesis 3. Dose of exercise (kcal·kg−1·week−1) predicts a significant amount
of variance in effect size.
Method
Literature Search
An electronic literature search was conducted using the following databases:
PubMed, SportDiscus, PsycInfo, and Dissertation Abstracts International. This
search was conducted using various combinations of the following terms: exercise,
physical activity, physical exercise, running, jogging, walking, weight lifting, weight
training, anxiety, anxiety disorder, and mental health. In addition, references from
previous meta-analyses, narrative reviews, and all studies that were located were
referenced by hand for possible studies that met the inclusion criteria.
Inclusion Criteria
Only randomized, controlled studies that included a self-report measure of anxiety as
a dependent variable were included in the analysis. Psycho-physiological measures
were not included in the meta-analysis because very few randomized, controlled
studies employed these measures. Studies also had to include an exercise condition
that did not include another form of treatment, i.e., exercise plus psychotherapy or
drug therapy, as an independent variable. The analysis was also limited to studies
in the English language that were available as of January 2006.
Moderator Variable Coding
Each study that met the inclusion criteria was coded for both participant and design
characteristics. The first author coded the studies, and the second author recoded
the moderator variables from a random sample of 10 studies to examine coding
reliability. Studies were coded as part of a clinical population or a nonclinical
population, as determined by whether the participants were classified as clinically
anxious, with or without clinical depression, according to DSM-IV (American
Psychiatric Association, 1994) criteria. Effect sizes were also coded for gender, to
examine whether exercise has a differential effect for men and women, or when
interventions included both men and women. The final participant characteristic
that was coded was mean participant age.
Design characteristics that were coded included the duration of the exercise
intervention, type of exercise (aerobic, anaerobic, or a combination of the two), the
frequency of exercise, the intensity of the exercise, and the duration of the exercise
bout. For studies in which a range of intensities or bout durations were given (i.e.,
60–70% VO2max or 20–30 min of exercise), the lowest value of the range was used
396 Wipi, Rethorst, and Landers
because this is the only level for which it can be assumed that every participant
in the study reached. In addition, because there has been a great deal of interest
in whether improvements in aerobic fitness are at least partially responsible for
the mental health benefits of exercise (Etnier, Nowell, Landers, & Sibley, 2006),
studies were coded for the average percentage improvement in VO2max that was
realized by the participants.
Effect Size Calculation
A single average effect size was calculated for studies that used multiple different
measures of anxiety and studies that used multiple different exercise groups (i.e.,
one weight training group and one aerobic group). This treatment of effect size
measures leads to independence of effect size values and gives each study in the
analysis equal weight. Hedges’s g was used as the measure of effect size in this
meta-analysis, in which Hedges’s g = MExperimental − MControl / SDPooled, where
SDPooled =
Posttest means and standard deviations were used in effect size calculation. When
means and standard deviations were not available, an effect size was calculated
using F, t, and r values when available (Rosenthal, 1994). Because studies with
large sample sizes yield more precise estimates of true population parameters than
small studies, each standardized mean difference was multiplied by
(Hedges & Olkin, 1985). Finally, based upon the recommendations of Hedges and
Olkin (1985), along with Hunter and Schmidt (2004), a random effects model was
used, and effect sizes were weighted by the inverse of the variance to calculate the
overall effect size (Shadish & Haddock, 1994). In addition, gains effect sizes were
calculated for exercise and control groups, in which case the calculation for effect
size becomes MPost − MPre / SDPooled for each group, and
SDPooled =
Each effect size was coded such that a negative value indicates a reduction in
anxiety. After calculation, effect sizes were divided into two categories: those
that compared exercise to a no-treatment control group, and those that compared
exercise to some other form of treatment for anxiety. In each case, a one-sample t
test was used to compare the value of the overall effect to zero.
An overall Q value, which represents the total amount of variance among the
set of effect sizes, was calculated to test for homogeneity of variance among the
Anxiolytic Effects of Exercise 397
effect sizes. This value was tested against a χ2 distribution, in which df = k − 1,
where k = number of effect sizes. A significant Q value would indicate that the
data are heterogeneous, and warrant the examination of moderator variables by
partitioning the variance into Qw and Qb. Qb values are also tested against a χ2
distribution, in which df = number of categories of the moderator variable − 1.
A significant Qb value indicates that the moderator variable contributes to the
variance among effect sizes. Weighted effect sizes and standard errors were then
calculated for each category within the moderator variables, along with 95%
confidence intervals (Hedges, 1994). For each moderator variable that revealed a
significant Qb test, pairwise comparisons between all effect sizes were examined,
which indicate whether the effect sizes are significantly different from one another
(Hedges & Olkin, 1985).
Dose–Response Analysis
The volume of exercise used in each study was estimated in units of energy
(kcal·kg−1·week−1) expended during exercise, assuming 5 kcal per milliliter of O2
consumed. Total O2 consumed per kilogram per week was calculated as
Total O2 consumed = (VO2max) × (intensity) × (bout duration) × (frequency)
where total O2 units are mL·kg−1·week−1, VO2max units are mL·kg−1·min−1, aver-
age exercise bout intensity is in %VO2max, average bout duration is in minutes,
and exercise frequency is number of exercise bouts per week. Only those studies
reporting VO2max were used, and pre- and postintervention VO2max measurements
were averaged in the calculation (Powers & Howley, 2003).
After calculating exercise dose, the data were examined for outliers. Any
exercise dose that was greater than three standard deviations away from the mean
were considered outliers, and were removed from the analysis. Pearson’s product–
moment correlation was then calculated between exercise dose and the correspond-
ing effect sizes for the exercise groups. In addition, regression analysis was used to
further examine the relationship between exercise dose and effect size.
Results
Exercise Versus No-Treatment Controls
A total of 140 studies were located for possible inclusion in the analysis. Of these
140 studies, 63 met inclusion criteria, and 43 had sufficient information for calcu-
lation of effect sizes. Corresponding authors from the remaining 20 studies were
contacted and asked to provide additional data that would allow for effect size
calculation. Correspondence was received from a number of authors, six of whom
were able to provide sufficient data for effect size calculation, resulting in 49 studies
in the final analysis. In total, 3,566 individuals participated in the 49 studies that
were included in the analysis. The overall weighted effect size was −0.48 (95%
confidence intervals −0.63, −0.33), which is significantly different from zero. This
indicates that people in exercise groups experienced 1/2 of a standard deviation
larger reduction in anxiety than people in control groups. The average standard
398 Wipi, Rethorst, and Landers
deviation on the State Trait Anxiety Inventory (Spielberger & Krasner, 1988), which
was the most commonly used measure in these studies, was 7.13 points. An effect
size of −0.48 indicates that, on average, anxiety scores for people in exercise groups
will drop 3.42 more points than people in nonexercising control groups.
Moderator Variables
The mean intercoder reliability rate (Cronbach’s α) was found to be 0.93, and the
test for homogeneity of variance was found to be significant, Q = 136.67, p < .001,
thus warranting the examination of moderator variables. Only three Qb values were
found to be significant, and pairwise comparisons revealed four significant differ-
ences between moderator variable categories: Unpublished studies had significantly
higher effect sizes than published studies, participants who were 31–45 years old
had significantly higher effect sizes than those who were 45 years and above, and
participants who exercised 3–4 times per week had significantly higher effect sizes
than those who exercised 1–2 times per week or 5+ times per week.
Further analysis of moderator variables revealed that four group effect sizes
were found to be nonsignificantly different from zero: The effect sizes for partici-
pants who were under 18, for participants who exercised either 1–2 times per week
or 5+ times per week, and for those who had fitness improvements of 6–10%. It
should be noted, however, that all four of these categories, along with many others,
contain few effect sizes. More randomized trials in such areas may reveal significant
moderators of the relationship between exercise and anxiety in the future. Further
results from the analysis of moderator variables are presented in Table 1.
Gains Effect Sizes
Forty-two studies included enough information for calculation of gains effect
sizes. The overall weighted gains effect size for control groups was −0.08, (95%
CI = −0.13, −0.04). The overall weighted gains effect size for exercise groups was
−0.43, (95% CI = −0.46, −0.36).
File Drawer Test
A file drawer test (Hunter & Schmidt, 2004) was conducted to examine the potential
impact of unpublished studies that were not included in the meta-analysis. This test
revealed that it would take 120 missing randomized, controlled trials, all showing
null results, to reduce the magnitude of the overall effect size to −0.15, which, under
the current confidence intervals, would be a nonsignificant effect. Furthermore, the
fact that unpublished studies had higher effect sizes in this meta-analysis shows
that the potential for publication bias in this area is small.
Exercise Versus Other Treatments
A total of 27 randomized, controlled trials compared an exercise group to some
other form of treatment for anxiety. The types of treatment that were compared with
exercise are listed in Table 2. These studies resulted in 28 effect sizes, a popula-
tion of 1,924 participants, and an overall effect size of −0.19. This effect size was
found to be significantly different from zero, t(27) = −2.21, p < .05, indicating that
399
Table 1 Moderator Variables of the Relationship Between Exercise
and Anxiety
df Qb
Category
k
Effect size
SE
95% CI
Gender
2 1.86 Mixed 37 −0.41 0.04 −0.50, –0.33*
Male 6 −0.53 0.14 −0.81, –0.25*
Female 6 −0.31 0.09 −0.49, –0.12
Age
3 8.28* <18 3 −0.18 0.16 −0.49, 0.12
18–30 10 −0.50 0.08 −0.66, –0.34*
31–45 17 −0.51a0.07 −0.65, –0.37*
>45 17 −0.31b0.06 −0.42, –0.19*
Population
1 0.44 Clinical 3 −0.52 0.18 −0.87, –0.17*
Nonclinical 46 −0.40 0.04 −0.47, –0.32*
Type of exercise
2 4.03 Aerobic 44 −0.40 0.04 −0.48, –0.32*
Anaerobic 1 −1.15 0.40 −0.37, –1.94*
Combined 2 −0.52 0.16 −0.83, –0.21*
Intervention duration
3 2.74 Acute bout 8 −0.39 0.10 −0.59, –0.18*
4–9 weeks 8 −0.59 0.11 −0.81, –0.36*
10–14 weeks 14 −0.40 0.07 −0.53, –0.26*
15 + weeks 18 −0.38 0.06 −0.49, –0.27*
Frequency
2 16.37* 1–2 times/week 5 −0.16a0.10 −0.34, 0.03
3–4 times/week 32 −0.51b0.05 −0.42, –0.61*
5+ times/week 3 −0.13a0.13 −0.38, 0.12
Exercise bout duration
2 2.67 1–30 min 23 −0.41 0.05 −0.51, –0.30*
31–60 min 16 −0.31 0.06 −0.44, –0.19*
61–90 min 3 −0.61 0.20 −0.99, –0.22*
Exercise intensity
1 0.89 Moderate† 13 −0.31 0.07 −0.44, –0.17*
Hard†† 15 −0.40 0.06 −0.47, –0.32*
VO
2max
improvements
2 0.59 1–5% increase 3 −0.26 0.12 −0.48, –0.02*
6–10% increase 4 −0.22 0.15 −0.52, 0.06
11% + increase 11 −0.33 0.07 −0.48, –0.19*
Sources
1 4.48* Journal 41 −0.37 a0.04 −0.45, –0.29*
Unpublished 8 −0.61 b0.11 −0.82, –0.40*
Note. Differing superscripts indicate that pairwise comparisons revealed significant differences between
the effect sizes.
*p < .05.
†Moderate = 40–59% VO2max/HRR or 55–69% HR max.
††Hard = 60–84% VO2max/HRR or 70–89% HR max.
400 Wipi, Rethorst, and Landers
exercise is slightly, yet significantly, better at reducing anxiety than other common
forms of treatment. Exercise was found to be equal to, or better than, all of the other
forms of treatment except pharmacotherapy; the effect size for pharmacotherapy
was 0.11, which is considered to be very small. Further results from this analysis
are summarized in Table 2.
Studies With Insufficient Data
There were 14 randomized, controlled trials that examined the effects of exercise
on anxiety for which no effect size was able to be calculated. Of these 14 studies,
12 showed results indicating that participation in exercise resulted in lower anxiety
scores when compared with controls. Nine of these 12 studies showed results that
were statistically significant (p < .05), whereas the other three showed nonsignificant
anxiety reductions in exercise groups compared with control groups.
Dose–Response Analysis
Twelve of the randomized, controlled trials provided adequate information to
examine a dose–response relationship. Because some of the studies included
multiple exercise groups, there were a total of 21 “doses” of exercise that could
be examined in this analysis. One outlier was excluded from the analysis; the
exercise dose for this study was more than 4 standard deviations above the mean.
Pearson’s product–moment correlation between exercise dose and effect size was
found to be nonsignificant, r = .00. Multiple regression analysis revealed that a
quadratic trend explained the largest amount of variance, R2 Multiple = 0.185, but
this model was not statistically significant, F(2, 18) = 1.92, p > .05. The trend in
the data showed that effect size magnitude increased as exercise approached a dose
of 12.5 kcal·kg−1·week−1, and then began to decrease as exercise dose increased.
These data are displayed in Figure 1.
Table 2 Effect Sizes of Types of Treatment Compared
With Exercise
Treatment
k
Effect size
Cognitive/behavioral therapy 2 0.00
Group therapy 3 −0.09
Light exercise (stretching, yoga) 6 −0.15
Relaxation/meditation 9 −0.23
Stress management education 5 −0.45
Pharmacotherapy 2 0.11
Music therapy 1 −0.05
Total: 28 −0.19*
*One-sample t test, p < .05
Anxiolytic Effects of Exercise 401
Discussion
The effect size of −0.48 for exercise groups compared with control groups rep-
resents Level 1, Grade A evidence for using exercise as a treatment for anxiety
disorders. These results indicate that exercise alone can be effective at reducing
anxiety because none of these studies included groups that combined exercise with
some other form of treatment. In addition, the comparison of exercise to other
forms of treatment revealed that exercise is as effective, and nearly as effective
as the two most common treatments for anxiety disorders—psychotherapy and
pharmacotherapy, respectively. This finding supports the use of exercise as one of
the frontline treatments for anxiety. In addition, research in the area of depression
has shown that the combination of pharmacotherapy and exercise is also effective
for reducing depression, and may therefore be effective for reducing anxiety as
well (Blumenthal et al., 1999).
As mentioned earlier, self-report questionnaires were the only measure of
anxiety included in this meta-analysis. Self-report measures are often criticized
for potential bias in the form of behavioral artifacts. It may be worthwhile to note,
Figure 1 — The relationship between exercise dose and effect size.
402 Wipi, Rethorst, and Landers
therefore, that the results of this meta-analysis are supported by animal literature.
Animal research is able to provide an unbiased look at some of the mental variables
that have shown an association with exercise. For example, mice that are depressed
and anxious tend to show less free-roaming behavior, less sexual activity, and gain
weight, which are also common symptoms of humans who are depressed and/or
anxious. After an exercise training program, these animals will show behavioral
changes, such as more free roaming, more sexual activity, and weight loss, which
are considered to be unbiased measures of changes in anxiety and/or depression.
(Landers & Arent, 2007).
The overall effect size that was found in this meta-analysis is similar, yet
slightly larger, than the effect sizes that were found in previous meta-analyses.
This is consistent with findings from earlier meta-analyses that showed larger effect
sizes for studies that used random assignment to groups, which is one marker of
study quality. Attrition rates may provide another indication of study quality. The
average attrition rate for exercise groups in this meta-analysis was 7.1%, whereas
the average attrition rate for control groups was 4.4%. The attrition rates indicate
that the randomized, controlled trials in the analysis were high-quality studies, and
also show that exercise is an effective, viable option for alleviating symptoms of
anxiety. The majority of the population in this study, however, was not receiving
clinical treatment, and future studies should focus on exercise among those with
clinically diagnosed anxiety disorders.
The analysis of moderator variables indicates the strength and consistency of
the results. It should also be noted that reductions in anxiety occurred with minimal
improvements in VO2max; also notable, however, is the increase in effect size that
occurred as fitness improvements increased to 11% or higher improvement. These
results show that exercise can be useful for reducing symptoms of anxiety in practi-
cally any population, and in any dose. The different effect sizes that were seen at
varying exercise frequencies are likely a result of the small number of studies in
the 1–2 times per week and 5+ times per week categories. The data for participants
31 to 45 years old compared with participants 45 years old and above, however,
comes from a larger number of studies. Although the pairwise comparisons revealed
this difference to be statistically significant, the significance of this difference in
an applied setting is minimal. Exercise was effective in both of these groups, and
the effect sizes are similar in magnitude. Even though these results could be inter-
preted as simply showing that something is better than nothing, the comparison of
exercise to other treatments, such as stress management and education, provides a
refutation to this interpretation.
Some trends emerged in the analysis of moderator variables that may have
reached statistical significance with more effect sizes in certain categories. For
example, groups that exercised either 61–90 min per session experienced larger
reductions in anxiety (effect size = −0.61) than groups who exercised for 1–30
or 31–60 min per session (effect sizes = −0.41, and −0.31, respectively). Despite
the disparity between these effect sizes, the differences did not reach statistical
significance. This trend seems to indicate that more exercise leads to greater reduc-
tions in anxiety, which led to the question of whether a dose–response relationship
could be established from the randomized, controlled studies that were used in the
analysis.
Anxiolytic Effects of Exercise 403
Although the R2 in the dose–response analysis was relatively robust, explaining
18.5% of the variance in effect size, the quadratic trend did not show significance.
The nadir of this curve occurs at approximately 12.5 kcal·kg−1·week−1, which is
slightly lower than the current public health dose that has been recommended by the
American College of Sports Medicine and Centers for Disease Control for physi-
cal health (Pate et al., 1995). Unfortunately, the sample size for the dose–response
analysis was relatively small; only 24% (12/49) of the randomized, controlled trials
that were included in the analysis provided adequate information to calculate an
exercise dose. In addition, nearly half of the effect sizes in the meta-analysis were
derived from studies that did not report the intensity of exercise used. Detailed
information in more studies would likely have led to stronger conclusions to be
drawn from the dose–response analysis in this study. Furthermore, if research in
this area is to have any clinical significance, it is crucial that future studies provide
both a clear description of the exercise paradigm that was used, and provide detailed
information about the quantity of exercise that was used.
It should also be noted that the studies that were included in the meta-analysis
used a fairly homogeneous set of interventions, as evidenced by the large discrepan-
cies in the number of effect sizes in certain categories. For example, the majority
of studies examined aerobic exercise in a nonclinical population, with a frequency
of 3–4 times per week. Many of the moderator variables may have shown signifi-
cant differences, and the dose response analysis may have been able to yield more
conclusive results, had there been more variation among the exercise paradigms.
Future research should attempt to fill some of the gaps that are present in the lit-
erature. Specifically, it may be beneficial to examine the effects of both low and
maximal intensity exercise on anxiety, or for more studies to examine the effects
of anaerobic exercise on anxiety in randomized, controlled trials. It is also crucial
that future research be conducted with clinical populations. Clinical patients are
the people who potentially have the most to gain from this type of research, and
the effect size for clinical populations from this meta-analysis (−0.52) supports the
recommendation to use exercise as a method of treatment for anxiety disorders.
This effect size, however, is based upon only three studies, and additional studies
with clinical populations would provide a much stronger recommendation.
Evidence showing that exercise can help alleviate anxiety and depression
is fairly strong, but the methods by which exercise affects anxiety and depres-
sion remain largely unknown, providing another area of concentration for future
research. There are numerous theories that have been proposed, both psychological
and physiological. Some of the explanations, such as the thermogenic hypothesis,
the time-out hypothesis, the Pitts–McClure hypothesis, and the behavioral artifact
hypothesis have not been supported in the literature, and are no longer considered
tenable (Landers & Arent, 2007). The literature investigating these hypotheses
has been summarized by Landers and Arent (2001). There are many psychologi-
cal theories that are still under investigation, including increases in self-esteem,
self-concept, and self-efficacy. Physiological hypotheses that have been developed
to explain the anxiolytic and antidepressive effects of exercise include the brain-
derived neurotrophic factor hypothesis, the endorphin hypothesis, the endocan-
nabinoid hypothesis, the hypothalamic-pituitary-adrenal axis hypothesis, and the
norepinephrine hypothesis. These hypotheses remain tenable and are supported
404 Wipi, Rethorst, and Landers
primarily by animal models, but a paucity of research has examined these hypoth-
eses in humans (Landers & Arent, 2007).
Conclusions
In conclusion, the evidence presented in this meta-analysis provides support for
using exercise for the alleviation of anxiety. Because only randomized, controlled
trials were included in the analysis, this study presents Level 1, Grade A evidence
for the use of exercise as a treatment for anxiety disorders. Although the dose–
response analysis portion of this study does not show a definitive relationship
between amount of exercise and reductions in anxiety, the quadratic shape of the
curve may give a good indication of the relationship, and it is consistent with cur-
rent physical activity recommendations. Future research in the area should focus
on varied exercise paradigms, attempt to use a clinical population whenever pos-
sible, and attempt to examine multiple doses of exercise in an experimental setting.
Research with humans into the mechanisms by which exercise can reduce anxiety
may also be useful in the future.
References
Note. References preceded by an asterisk are listed in the appendix table that
follows.
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental
disorders (4th ed.). Washington, DC: American Psychiatric Publishing.
Andrews, G., Sanderson, K., Corry, J., & Lapsley, H.M. (2000). Using epidemiological data
to model efficiency in reducing the burden of depression. The Journal of Mental Health
Policy and Economics, 3, 175–186.
*Antunes, H.K.M., Stella, S.G., Santos, R.F., Bueno, O.F.A., & de Mello, M.T. (2005).
Depression, anxiety, and quality of life scores in seniors after an endurance exercise
program. Revista Brasileira de Psiquiatria (Sao Paulo, Brazil : 1999), 27, 266–271.
*Bahrke, M., & Morgan, W. (1978). Anxiety reduction following exercise and meditation.
Cognitive Therapy and Research, 2, 323–333.
*Bahrke, M.S., & Smith, R.G. (1985). Alterations in anxiety of children after exercise and
rest. American Journal of Corrective Therapy, 39, 90–94.
*Bartholomew, J., Morrison, D., & Ciccolo, J. (2005). Effects of acute exercise on mood
and well-being in patients with major depressive disorder. Medicine and Science in
Sports and Exercise, 37, 2032–2037.
*Berger, B., Friedmann, E., & Eaton, M. (1988). Comparison of Jogging, the relaxation
response, and group interaction for stress reduction. Journal of Sport and Exercise
Psychology, 10, 431–447.
*Blumenthal (1991). Long-term effects of exercise on psychological functioning in older
men and women. Journal of Gerontology, 46, 352-361.
Blumenthal, J.A., Babyak, M.A., Moore, K.A., Craighead, W.E., Herman, S., Khatri, P., et
al. (1999). Effects of exercise training on older patients with major depression. Archives
of Internal Medicine, 159, 2349–2356.
*Blumenthal, J.A., Sherwood, A., Babyak, M.A., Watkins, L.L., Waugh, R., Georgiades,
A., et al. (2005). Effects of exercise and stress management training on markers of
cardiovascular risk in patients with ischemic heart disease. Journal of the American
Medical Association, 293, 1626–1634.
Anxiolytic Effects of Exercise 405
*Broocks, A., Bandelow, B., Pekrun, G., George, A., Meyer, T., Bartmann, U., et al. (1998).
Comparison of aerobic exercise, clomipramine, and placebo in the treatment of panic
disorder. The American Journal of Psychiatry, 155, 603–609.
*Brown, M.A., Goldstein-Shirley, J., Robinson, J., & Casey, S. (2001). The effects of a
multi-modal intervention trial of light exercise and vitamins on women’s mood. Women
& Health, 34, 93–112.
Calfas, K.J., & Taylor, W.C. (1994). Effects of physical activity on psychological variables
in adolescents. Pediatric Exercise Science, 6, 406–423.
*Castro, C.M., Wilcox, S., O’Sullivan, P., Baumann, K., & King, A. (2002). An exercise
program for women who are caring for relatives with dementia. Psychosomatic Medi-
cine, 64, 458–468.
*Cramer, S.R., Nieman, D.C., & Lee, J.W. (1991). The effects of moderate exercise train-
ing on psychological well-being and mood state in women. Journal of Psychosomatic
Research, 35, 437–449.
*Crocker, P.R., & Grozelle, C. (1991). Reducing induced state anxiety: effects of acute
aerobic exercise and autogenic relaxation. Journal of Sports, Medicine, and Physical
Fitness, 31, 277–282.
*De Geus, E.J., Van Doornen, L.J., De Visser, D.C., & Orlebeke, J.F. (1990). Existing
and training induced differences in aerobic fitness: their relationship to physiological
response patterns during different types of stress. Psychophysiology, 27, 457–478.
*Deivert, R.G. (1990). Efficacy of an aerobic exercise program as treatment for depression
and anxiety in alcohol and chemically dependent adults. Unpublished doctoral dis-
sertation, The Pennsylvania State University.
*DePalma, M.T. (1989). The Effects of exercise on anxiety, depression, and type A behavior.
Unpublished doctoral dissertation, Cornell University.
*DiLorenzo, T.M., Bargman, E.P., Stucky-Ropp, R., Brassington, G.S., Frensch, P.A., &
LaFontaine, T. (1999). Long-term effects of aerobic exercise on psychological outcomes.
Preventive Medicine, 28, 75–85.
*Doan, B.T., Plante, T.G., Digregorio, M.P., & Manuel, G.M. (1995). Influence of aerobic
exercise and relaxation training on coping with test-taking anxiety. Anxiety, Stress,
and Coping, 8, 101–111.
*Dugmore, L.D., Tipson, R.J., Phillips, M.H., Flint, E.J., Stentiford, N.H., Bone, M.F., et al.
(1999). Changes in cardiorespiratory fitness, psychological wellbeing, quality of life,
and vocational status following a 12 month cardiac exercise rehabilitation programme.
Heart (British Cardiac Society), 81, 359–366.
Dunn, A.L., Trivedi, M.H., Kampert, J.B., Clark, C.G., & Chambliss, H.O. (2002). The DOSE
study: A clinical trial to examine efficacy and dose response of exercise as treatment
for depression. Controlled Clinical Trials, 23, 584–603.
Dunn, A.L., Trivedi, M.H., Kampert, J.B., Clark, C.G., & Chambliss, H.O. (2005). Exercise
treatment for depression: efficacy and dose response. American Journal of Preventive
Medicine, 28, 1–8.
Dunn, A.L., Trivedi, M.H., & O’Neal, H.A. (2001). Physical activity dose-response effects
on outcomes of depression and anxiety. Medicine and Science in Sports and Exercise,
33(Supplement), S587–S597.
*Eby, J.M. (1984). An investigation into the effects of aerobic exercise on anxiety and
depression. Unpublished doctoral dissertation, University of Toronto.
*Emery, C.F., Schein, R.L., Hauck, E.R., & MacIntyre, N.R. (1998). Psychological and
cognitive outcomes of a randomized trial of exercise among patients with chronic
obstructive pulmonary disease. Health Psychology, 17, 232–240.
Etnier, J.L., Nowell, P.M., Landers, D.M., & Sibley, B.A. (2006). A meta-regression to
examine the relationship between aerobic fitness and cognitive performance. Brain
Research. Brain Research Reviews, 52, 119–130.
406 Wipi, Rethorst, and Landers
Galper, D.I., Madhukar, T.H., Barlow, C.E., Dunn, A.L., & Kampert, J.B. (2006). Inverse
association between physical activity and mental health in men and women. Medicine
and Science in Sports and Exercise, 38, 173–178.
Gardner, D.M., Baldessarini, R.J., & Waraich, P. (2005). Modern antipsychotic drugs: A
critical overview. Canadian Medical Association Journal, 172, 1703–1711.
*Gowans, S.E., De Hueck, A., & Abbey, S.E. (2002). Measuring exercise induced mood
changes in fibromyalgia: A comparison of several measures. Arthritis and Rheumatism,
47, 603–609.
*Gowans, S.E., De Hueck, A., Voss, S., Silaj, A., Abbey, S.E., & Reynolds, W.J. (2001).
Effects of a randomized, controlled trial on mood and physical function in individuals
with fibromyalgia. Arthritis Care and Research, 45, 519–529.
Greist, J.H., & Jefferson, J.W. (2000). Anxiety disorders. In H.H. Goldman’s (Ed.) Review
of General Psychiatry—5th Ed. New York, NY: McGraw-Hill.
Guyatt, G.H., Cook, D.J., Sackett, D.L., Eckman, M., & Pauker, S. (1998). Grades of recom-
mendation for antithrombotic agents. Chest, 114, 441–444.
Hedges, L.V. (1994). Fixed effects models. In H. Cooper & L.V. Hedges (Eds.), The handbook
of research synthesis (pp. 285–299). New York: Russell Sage Foundation.
Hedges, L.V., & Olkin, I. (1985). Statistical methods for meta-analysis. Orlando, Florida:
Academic Press.
Hill, A.B. (1965). The environment and disease: association or causation. Proceedings of
the Royal Society of Medicine, 58, 293–300.
*Hilyer, J.C., Wilson, D.G., Cillon, C., Caro, L., Jenkins, C., Pencer, W.A., et al. (1982).
Physical fitness training and counseling as treatment for youthful offenders. Journal
of Counseling Psychology, 29, 292–303.
Hunter, J.E., & Schmidt, F.E. (2004). Methods of meta-analsys: Correcting error and bias
in research findings. Thousand Oaks, CA: Sage.
*Jorgensen, C. (1986). Aerobic conditioning in the therapeutic treatment of chronic schizo-
phrenia. Unpublished doctoral dissertation, Northern Arizona University.
*King, A., Taylor, C., & Haskell, W. (1993). Effects of differing intensities and formats
of 12 months of exercise training on psychological outcomes in older adults. Health
Psychology, 12, 292–300.
*King, A.C., Barr Taylor, C., Haskell, W.L., & DeBusk, R.F. (1989). Influence of regular
aerobic exercise on psychological health: A randomized, controlled trial of healthy
middle-aged adults. Health Psychology, 8, 305–324.
*Klein, M.H., Greist, J.H., Gurman, A.S., Neimeyer, R.A., Lesser, D.P., Bushnell, N.J., et al.
(1985). A comparative outcome study of group psychotherapy vs. exercise treatments
for depression. International Journal of Mental Health, 13, 148–177.
*Koukouvou, G., Kouidi, E., Iacovides, A., Konstantinidou, E., Kaprinis, G., & Deligiannis,
A. (2004). Quality of life, psychological, and physiological changes following exercise
training in patients with chronic heart failure. Journal of Rehabilitation Medicine, 36,
36–41.
Kugler, J., Seelbach, H., & Kruskemper, G.M. (1994). Effects of rehabilitation exercise
programmes on anxiety and depression in coronary patients: A meta-analysis. The
British Journal of Clinical Psychology, 33, 401–410.
Landers, D.M., & Arent, S.M. (2001). Exercise and mental health. In R.N. Singer, H.A.
Hausenblas, & C.M. Janelle (Eds.), Handbook of sport psychology (2nd ed., pp.
740–765). New York: Wiley.
Landers, D.M., & Arent, S.M. (2007). Physical activity and mental health. In G. Tenenbaum
& R. Eklund (Eds.), Handbook of sport psychology (3rd ed., pp. 469–491). Hoboken,
NJ: John Wiley & Sons Inc.
Landers, D.M., & Petruzzello, S.J. (1994). Physical activity, fitness, and anxiety. In C.
Bouchard, R.J. Shephard, & T. Stevens (Eds.), Physical activity, fitness, and health.
Champaign, IL: Human Kinetics.
Anxiolytic Effects of Exercise 407
*Lion, L.S. (1978). Psychological effects of jogging: a preliminary study. Perceptual and
Motor Skills, 47, 1215–1218.
*Lobitz, W.C., Brammell, H.L., Stoll, S., & Niccoli, A. (1983). Physical exercise and anxiety
management training for cardiac stress management in a nonpatient population. Journal
of Cardiac Rehabilitation, 3, 683–688.
*Long, B.C. (1984). Aerobic conditioning and stress inoculation: A comparison of stress-
management interventions. Cognitive Therapy and Research, 8, 517–542.
Long, B.C., & Van Stavel, R. (1995). Effects of exercise training on anxiety: A meta-analysis.
Journal of Applied Sport Psychology, 7, 167–189.
Ludwig, J., & Marcotte, D.E. (2005). Anti-depressants, suicide, and drug regulation. Journal
of Policy Analysis and Management, 24, 249–272.
McDonald, D.G., & Hodgdon, J.A. (1991). The psychological effects of aerobic fitness
training: Research and theory. New York: Springer-Verlag.
*McEntee, D.J., & Halgin, R.P. (1999). Cognitive group therapy and aerobic exercise in the
treatment of anxiety. Journal of College Student Psychotherapy, 13, 37–55.
Moore, K.A., & Blumenthal, J.A. (1998). Exercise training as an alternative treatment
for depression among older adults. Alternative Therapies in Health and Medicine, 4,
48–56.
Myers, J.K., Weissman, M.M., Tischler, G.L., Holzer, C.E., Leaf, P.J., Orvaschel, H., et al.
(1984). Six-month prevalence of psychiatric disorder in three communities: 1980 to
1982. Archives of General Psychiatry, 41, 959–967.
*Newton, M., Mutrie, N., & McArthur, J.D. (1991). The effects of exercise in a coronary
rehabilitation program. Scottish Medical Journal, 38, 38–41.
*Norris, R., Carroll, D., & Cochrane, R. (1992). The effects of physical activity and exercise
training on psychological stress and well-being in an adolescent population. Journal
of Psychosomatic Research, 36, 55–65.
*Norvell, N., & Belles, D. (1993). Psychological and physical benefits of circuit weight
training in law enforcement personnel. Journal of Consulting and Clinical Psychol-
ogy, 61, 520–527.
Pate, R.R., Pratt, M., Blair, S.N., Haskell, W.L., Macera, C.A., & Bouchard, C. (1995).
Physical activity and public health: A recommendation from the Centers for Disease
Control and Prevention and the American College of Sports Medicine. Journal of the
American Medical Association, 273, 402–407.
*Petajan, J.H., Gappmaier, E., White, A.T., Spencer, M.K., Mino, L., & Hicks, R.W. (1996).
Impact of aerobic training on fitness and quality of life in multiple sclerosis. Annals
of Neurology, 39, 432–441.
Petruzzello, S.J., Landers, D.M., Hatfield, B.D., Kubitz, K.A., & Salazar, W. (1991). A
meta-analysis on the anxiety reducing effects of acute and chronic exercise. Sports
Medicine (Auckland, N.Z.), 11, 143–182.
*Pierce, T.W., Madden, D.J., Siegel, W.C., & Blumenthal, J.A. (1993). Effects of aerobic
exercise on cognitive and psychological functioning in patients with mild hypertension.
Health Psychology, 12, 286–291.
Powers, S.K., & Howley, E.T. (2003). Exercise Physiology: Theory and Applications to
Fitness and Performance. New York, NY: McGraw-Hill.
Rosenthal, R. (1994). Parametric measures of effect size. In H. Cooper & L.V. Hedges
(Eds.), The handbook of research synthesis (pp. 231–244). New York: Russell Sage
Foundation.
*Roth, D., & Holmes, D. (1987). Influence of aerobic exercise training and relaxation train-
ing on physical and psychologic health following stressful life events. Psychosomatic
Medicine, 49, 355–365.
*Roth, D.L. (1989). Acute emotional and psychophysiological effects of aerobic exercise.
Psychophysiology, 26, 593–602.
408 Wipi, Rethorst, and Landers
*Roth, D.L., Bachtler, S.D., & Fillingim, R.B. (1990). Acute emotional and cardiovascu-
lar effects of stressful mental work during aerobic exercise. Psychophysiology, 27,
694–699.
*Setaro, J.L. (1985). Aerobic exercise and group counseling in the treatment of anxiety and
depression. Unpublished doctoral dissertation, University of Maryland.
Schlicht, W. (1994). Does physical exercise reduce anxious emotions? A meta-analysis.
Anxiety, Stress & Coping. International Journal (Toronto, Ont.), 6, 275–288.
Shadish, W.R., & Haddock, C.K. (1994). Combining estimates of effect size. In H. Cooper
& L.V. Hedges (Eds.), The handbook of research synthesis (pp. 231–244). New York:
Russell Sage Foundation.
Singh, N.A., Stavrinos, T.M., Scarbek, Y., Galambos, G., Liber, C., & Fiatrone-Singh, M.A.
(2005). A randomized controlled trial of high versus low intensity weight training versus
general practitioner care for clinical depression in older adults. Journal of Gerontoloty:
Medical Sciences, 60A, 768–776.
*Sorenson, M., Anderssen, S., Hjerman, I., Holme, I., & Ursin, H. (1999). The effect of exer-
cise and diet on mental health and quality of life in middle-aged individuals with elevated
risk factors for cardiovascular disease. Journal of Sports Sciences, 17, 369–377.
Spielberger, C.D., & Krasner, S.S. (1988). The assessment of state and trait anxiety. In
R. Noyes, Jr., M. Roth, & G.D. Burrows (Eds.), Handbook of Anxiety Classification,
Etiological Factors and Associated Disturbances (Vol. 2, pp. 31–51). Amsterdam:
Elsevier.
*Taylor, M.W. (1991). Effects of initial stress level, social support, and participation in an
exercise or music condition on the posttreatment stress, depression, and anxiety of
nurses. Unpublished doctoral dissertation, St. John’s University.
*Thaxton, L. (1982). Physiological and psychological effects of short-term exercise addic-
tion on habitual runners. Journal of Sport Psychology, 4, 73–80.
*Thorsen, L., Skovlund, E., Stromme, S., Hornslien, K., Dahl, A., & Fossd, S. (2005).
Effectiveness of physical activity on cardiorespiratory fitness and health-related quality
of life in you and middle-aged cancer patients shortly after chemotherapy. Journal of
Clinical Oncology, 23, 2378–2388.
*Van den Berg-Emons, R., Balk, A., Bussmann, H., & Starn, H. (2004). Does aerobic train-
ing lead to a more active lifestyle and improved quality of life in patients with chronic
heart failure? European Journal of Heart Failure, 6, 95–100.
*Veale, D., LeFevre, K., Pantelis, C., de Souza, V., Mann, A., & Sargeant, A. (1992). Aero-
bic exercise in the adjunctive treatment of depression: A randomized controlled trial.
Journal of the Royal Society of Medicine, 85, 541–544.
*Williams, P., & Lord, S. (1997). Effects of group exercise on cognitive functioning and mood
in older women. Australian and New Zealand Journal of Public Health, 21, 45–52.
*Wilson, L.M. (1985). The effects of an exercise conditioning program on reducing the stress
response in nurses. Unpublished doctoral dissertation, Wayne State University.
*Zentner, R. (1981). Psychological effects of a running program. Unpublished doctoral
dissertation, University of Oregon.
Manuscript submitted: January 18, 2007
Revision accepted: March 29, 2008
409
Appendix Table Characteristics of Studies Used in the Meta-Analysis
Author (year)
a
Average corrected
effect size
Exercise
type
Age
(mean or range)
N
Gender
Intervention
duration (weeks)
Bout duration
(minutes)
Frequency
(times/week)
Antunes (2005) −0.89 Aerobic 66.97 46 Male 24 20 3
Bahrke (1978) 0.29 Aerobic 51.9 75 Male Acute 20 Acute
Bahrke (1985) −0.13 Aerobic 10.7 65 Mixed Acute 15 Acute
Bartholomew (2005) 0.04 Aerobic 38.1 40 Mixed Acute 30 Acute
Berger (1988) −0.35 Aerobic 19 153 Mixed 12 20 3
Blumenthal (1991) −0.31 Aerobic 67 101 Mixed 16 30 3
Blumenthal (2005) −0.33 Aerobic 63 142 Mixed 16 45 3
Broocks (1998) −1.40 Aerobic 46 37 Mixed 10 NA 3
Brown (2001) −0.54 Aerobic 42 104 Female 8 20 5
Castro (2002) −0.18 Aerobic 62 85 Female 52 30 3
Cramer (1991) −0.40 Aerobic 34 35 Mixed 15 45 5
Crocker (1991) −1.33 Aerobic 20 85 Mixed Acute 40 Acute
De Geus (1990) −0.18 Aerobic 23.7 26 Male 7 90 4
Deivert (1990)b−2.33 Aerobic 31 40 Mixed 8 20 3
DePalma (1989)b−0.93 Aerobic 20.27 77 Mixed 12 20 3
DiLorenzo (1998) −0.51 Aerobic 32 111 Mixed 12 24 4
Doan (1995) −0.08 Aerobic 19 52 Mixed Acute 15 Acute
Dugmore(1999) −0.57 Combined 55 124 Mixed 52 NA 3
Eby (1984)b0.08 NA 19–31 39 Mixed NA 60 3
Emery (1998) −0.28 Aerobic 66.6 74 Mixed 10 45–90 3–5
Gowans (2001) −1.14 Aerobic 47 31 Mixed 23 30 3
Gowans (2002) −0.92 Aerobic 47 31 Mixed 23 30 3
Hilyer (1982) −0.39 Combined 15–18 43 Male 20 90 3
Jorgensen (1986)b−0.27 Aerobic 36.4 11 Mixed 6 60 3
King (1989) 0.31 Aerobic 48 113 Mixed 24 50 5
King (1993) −0.28 Aerobic 50–65 300 Mixed 52 40 3
(continued)
410
Author (year)
a
Average corrected
effect size
Exercise
type
Age
(mean or range)
N
Gender
Intervention
duration (weeks)
Bout duration
(minutes)
Frequency
(times/week)
Koukouvou (2004) −1.74 Aerobic 52.5 26 Male 24 60 3
Lion (1978) −0.55 Aerobic NA 6 Mixed 8 NA 3
Lobitz (1983) −0.99 Aerobic 36.5 18 Mixed 7 NA 3
Long (1984) −1.02 Aerobic 39.9 73 Mixed 10 90 3
McEntee (1999) −0.53 Aerobic NA 70 Mixed 6 20 3
Newton (1991) −0.41 Aerobic <70 22 Mixed 10 60 3
Norris (1992) −0.08 Aerobic 16.5 80 Mixed 10 25 2
Norvell (1993) −1.15 Anaerobic 32.84 29 Male 16 20 3
Petajan (1996) 0.11 Aerobic 40 46 Mixed 15 50 3
Pierce (1993) −0.25 NA 45 99 Mixed 16 50 3
Roth (1987) −0.02 Aerobic 18.9 36 Mixed 11 30 3
Roth (1989) −0.83 Aerobic 20.8 80 Mixed Acute 20 Acute
Roth (1990) 0.07 Aerobic 20.5 57 Female Acute 10 Acute
Setaro (1985)b−1.38 Aerobic 18–35 75 Mixed 10 NA 2
Sorensen (1999) −0.50 Aerobic 44.9 219 Mixed 52 60 3
Taylor (1991)b−0.21 Aerobic 39.1 102 Female 6 20 3
Thaxton (1982) −0.84 Aerobic 36 33 Mixed Acute 30 Acute
Thorsen (2005) 0.10 Aerobic 39.1 111 Mixed 14 30 2
Van den Berg-Emons
(2004) −0.17 Aerobic 58.6 34 Mixed 12 60 2
Veale (1992) −0.58 Aerobic 35.5 83 Mixed 12 NA 3
Williams (1997) −0.15 Aerobic 71.7 187 Female 42 60 2
Wilson (1985)b−1.28 Aerobic 31 34 Female 16 40 3
Zentner (1981)b−0.12 Aerobic 41 40 Mixed 10 60 3
aThe citations are found in the References section.
bAll the articles were published in journals; the unpublished citations bear the superscript b.
Appendix Table continued
