Research Methodology: Endocrinologic Measurements
in Exercise Science and Sports Medicine
Anthony C. Hackney, PhD, CPH, FACSM*; Atko Viru, PhD, DSc
`
*The University of North Carolina at Chapel Hill, Chapel Hill, NC; 3Deceased; 4Tartu University, Tartu, Estonia
Objective: To provide background information on methodo-
logic factors that influence and add variance to endocrine
outcome measurements. Our intent is to aid and improve the
quality of exercise science and sports medicine research
endeavors of investigators inexperienced in endocrinology.
Background: Numerous methodologic factors influence hu-
man endocrine (hormonal) measurements and, consequently,
can dramatically compromise the accuracy and validity of
exercise and sports medicine research. These factors can be
categorized into those that are biologic and those that are
procedural-analytic in nature.
Recommendations: Researchers should design their stud-
ies to monitor, control, and adjust for the biologic and
procedural-analytic factors discussed within this paper. By
doing so, they will find less variance in their hormonal outcomes
and thereby will increase the validity of their physiologic data.
These actions can assist the researcher in the interpretation and
understanding of endocrine data and, in turn, make their
research more scientifically sound.
Key Words: hormones, biomedical sciences, sport research
design, study design
During the last 25 years, an increasing number ofexercise science and sports medicine researchershave begun to incorporate endocrinologic mea-
surements (eg, hormones, cytokines) into their designs.1,2
This approach has allowed for a heightened level of
research examining the physiologic mechanisms associated
with certain clinical and performance-related conditions
found in athletes.
Unfortunately, some exercise science investigators have
not always controlled certain factors (eg, time of day for
blood sampling) that can influence many of the hormones
within the human endocrine system. This lack of experi-
mental control has often resulted in the emerging research
being inconsistent, contradictory, and difficult to interpret.
The inadequate accounting for experimental controls may
be due to limited knowledge by exercise science and sports
medicine researchers in the area of endocrine methodology.
It is regrettable that this educational shortcoming occurs,
but in reality, few laboratories throughout the world
actively pursue exercise endocrinology as one of their
primary areas of study or research foci.
The factors that influence hormonal measurements and
contribute to variance in outcomes can be categorized as
coming from 2 potential sources: factors affecting biologic
variation (ie, affiliated with the physiologic function or
status of the participant) and factors affecting procedural-
analytic variation (ie, determined by the investigators
conducting the research).1 Regardless of whether the
source of variance is participant or investigator derived,
if it is not controlled or accounted for appropriately, then
hormonal measures can certainly be compromised. Such
compromises can call into question the validity and
scientific quality of a research study.
We developed this paper in order to serve as a primer for
exercise science and sports medicine researchers on the
methodologic factors involved in endocrine measurements,
so that they can improve the quality of their research
endeavors. This paper is by no means a complete and
exhaustive treatise on this topic, and readers desiring more
encompassing discussions are directed to more involved
texts.1,3
The field of endocrinology uses abbreviations for many
hormones. These abbreviations can sometimes be confus-
ing to novices in the field. To aid unfamiliar researchers,
the Table lists the most common hormones associated with
the area of exercise science and sports medicine and their
typical abbreviations.4
BIOLOGIC FACTORS
As noted in the opening text, factors that can influence
hormonal measurements can be categorized into 2 broad
areas: biologic and procedural-analytic. The biologic
factors are those that are determined to be connected in
some way to the physiologic function or status of the
participant at the time of specimen collection. These factors
can be viewed as endogenous in nature.
Sex. Until the onset of puberty, males and females
exhibit little difference in their resting hormonal profiles.
Once puberty is reached, however, males demonstrate
increased androgen steroid hormone production and
females show the characteristic menstrual cycle pulsatile
release of gonadotrophin and sex steroid hormones.5–7
Additionally, at puberty, resting levels of leptin (an
adipocyte cytokine, a low molecular-weight protein that
has endocrinelike actions on select physiologic processes
[eg, immune system]8) tend to become elevated in females
as compared with those in males.9 The differences that
manifest at puberty tend to persist through adulthood until
women become postmenopausal.8,9
Some sex-specific differences in the hormonal responses
to exercise also exist. These include an earlier and greater
rise in testosterone in males and a greater pre-exercise
growth hormone response in females. Additionally, the
Journal of Athletic Training 2008;43(6):631–639
g by the National Athletic Trainers’ Association, Inc
www.nata.org/jat
literature review
Journal of Athletic Training 631

magnitude of the sex steroid hormonal response to exercise
in females is influenced by the status and phase of the
menstrual cycle.10,11 Interestingly, the menstrual cycle
hormones can influence other hormones and their response
to exercise (eg, increased estradiol-b-17 leads to increased
growth hormone levels).10–12 The menstrual cycle is
discussed further in a later section of this paper. It is also
important to recognize that some hormones show little or
no differences between the sexes in response to exercise (eg,
water-balance hormones such as aldosterone and vaso-
pressin).5,10,11
For the above reasons, researchers should be cautious
when testing mixed-sex participant populations in their
studies, depending upon desired outcomes. To avoid
confounding results, the investigator needs to be certain
that the hormonal outcomes being measured are not
influenced by sex.
Age. In hormonal research, participants not matched for
age and maturation level may demonstrate increased
outcome variance. Prepubertal and postpubertal children
of the same sex do not typically display the exact same
hormonal responses or relationships.13,14 This is illustrated
by the well-documented increase in insulin resistance
observed as an adolescent goes through puberty.15
Such concerns also should be extended to the other end
of the age spectrum. A postmenopausal woman or
andropausal man could have drastically different hormon-
al responses than her or his prepausal counterparts. For
example, growth hormone and testosterone typically
decrease with age, whereas cortisol and insulin resistance
increase.16–18
These types of age-related differences can exist at rest, in
response to exercise, and even after an exercise training
program. For this reason, in designing studies (and unless
the researcher is studying age-related changes among
populations3), it may be important to match participants
by chronologic age or maturation level (or both) in order
to increase the homogeneity of the responses and decrease
interindividual variability.
Race. Many humoral constituents are known to vary
among people of different races.1,3 Only a few hormonal
differences, however, have been identified to exist. For
example, resting parathyroid hormone levels tend to be
higher in blacks than whites.19 White females tend to have
higher levels of estrogens than Asian females.1,20 Repro-
ductive hormone levels during gestational periods also may
vary across races (whites, blacks, Hispanics, Asians, and
Indians).20–23 Greater resting insulin levels and insulin
resistance have been noted in certain Native American
tribes (eg, Prima Indians); however, these differences may
be related more to obesity issues than to race.24
Hormonal responses to exercise and exercise training
related to race have not been well studied, and the limited
data available do not suggest drastically different response
outcomes. However, further research in this area is
necessary.1,23,24
Body Composition. Varying levels of adiposity can
greatly influence the cytokines released by adipose tis-
sue.3,8,9 These cytokine substances in turn have autocrine-
like, parcrinelike, and endocrinelike actions in the body
and influence metabolic, reproductive, and inflammatory
status.2,3,8,9 In addition, several cytokines have been linked
directly to increased hormonal levels (eg, increased
interleukin 6 is associated with increased cortisol).8 This
situation is compounded as adiposity reaches the level of
obesity and subsequently affects many hormones, poten-
tially to a far greater extent (eg, insulin and leptin levels
tend to be elevated at rest in many obese people).25–29
If a person’s level of adiposity increases (toward obesity),
the hormonal response to exercise and exercise training can
change drastically from that of a normal-weight person.
For example, in obese people, the catecholamine and
growth hormone response to exercise is reduced.29 Cortisol
responses to exercise have been elevated in some over-
weight-obese individuals, although reductions also have
been reported.28,29 Exercise training often results in a loss
of body mass, which helps to bring the responses of these
hormones more in line with that observed in normal-weight
people.29–33
To ensure that participants’ various levels of body
composition will not confound some hormonal outcomes,
investigators need to match their volunteers for adiposity
as closely as possible rather than simply matching body
weights. Exactly how close a match is needed is not known,
but grouping normal-weight, overweight (body mass index
$ 25.0 # 30.0 kg?m22), and obese (body mass index
.30.0 kg?m22) individuals in the same participant pool
can certainly complicate and add variance to some
hormonal outcomes.1,29
Mental Health. Certain mental health conditions are
associated with high anxiety levels (eg, posttraumatic stress
disorder), which can lead to enhanced sympathetic nervous
system and hypothalamic-pituitary-adrenal axis activity.34–36
Subsequently, resting levels of circulating catecholamines,
Table. Hormone Abbreviations Commonly Used in Exercise Sci-
ence and Sport Medicine Endocrinologic Research4
Name Abbreviation
Adrenocorticotropic hormone ACTH
Aldosterone ALD
Antidiuretic hormone ADH
Atrial natriuretic peptide ANP
Arginine vasopressin AVP
b-endorphin b-END
Catecholamines Cats
Corticotropin-releasing hormone CRH
Cortisol CORT
Epinephrine EPI
Estradiol-b-17 E2
Follicle-stimulating hormone FSH
Glucagon GLU
Gonadotrophin-releasing hormone GnRH
Growth hormone GH
Growth hormone-releasing hormone GHRH
Insulin IN
Insulinlike growth factor 1 IGF1
Leptin LP
Luteinizing hormone LH
Norepinephrine NOR
Progesterone P
Prolactin PRL
Reverse triiodothyronine rT3
Testosterone TEST
Thyrotropin-releasing hormone TRH
Thyroid-stimulating hormone TSH
Thyroxine T4
Triiodothyronine T3
632 Volume 43 N Number 6 N December 2008

adrenocorticotropic hormone, b-endorphin, and cortisol also
may be elevated in these conditions. In contrast, persons who
are depressed may have low arousal levels and suppressed
levels of the aforementioned hormones. Furthermore,
depression is sometimes accompanied by low activity levels
in the hypothalamic-pituitary-thyroid axis (ie, low levels of
thyrotropin-releasing hormone, thyroid-stimulating hor-
mone, thyroxine, and triiodothyronine).34–36
These alterations in resting hormonal levels can, in turn,
result in altered hormonal responses to exercise and
exercise training in individuals who have high levels of
anxiety or frustration.37–39 Responses may be heightened in
some cases and diminished in others.37–39
Asking a participant to complete a mental health
screening questionnaire can serve as an excellent tool to
determine if a potential emotional or psychological
problem exists that could confound hormonal measures.
A variety of such screening tools are available, and the
reader is directed to several references for discussions of the
topic.40,41 However, any such screening should be per-
formed by a trained, qualified individual.
Menstrual Cycle. The menstrual status (eumenorrheic
versus amenorrheic) and cycle phase (follicular, ovulation, or
luteal) in females can produce basal changes in key
reproductive hormones such as estradiol-b-17, progesterone,
luteinizing hormone, and follicle-stimulating hormone.
These changes can be large and dramatic within select
individuals. For example, the ovulatory and luteal phases
result in increases in all of the aforementioned hormones
above levels seen in the follicular phase (eg, 2-fold to 10-fold
greater in eumenorrheic females).42 Furthermore, as noted
earlier, certain reproductive hormones can, in turn, influence
other nonreproductive hormones at rest.11,43,44
Menstrual status and cycle-phase hormonal influences
can affect exercise and exercise training responses, too.
Therefore, researchers may need to conduct exercise testing
with females of similar menstrual status or in similar
phases of their cycle (or both). This precaution also is
applicable to females who are using oral contraceptives,
which can mimic some hormonal fluctuations similar to
cycle-phase changes.44,45
Circadian Rhythms. Many hormonal levels fluctuate and
display circadian variations. In some cases, these variances
are due to pulse generator aspects (ie, the spontaneous
release of select hypothalamic hormone-releasing factors
[hormones]46) within the endocrine regulatory axis. In
other cases, variances are related to humoral stimuli
changes brought on by participant behavior or environ-
mental factors.47,48 Circadian hormones can display
dramatic changes in levels due to their rhythm patterns,
cortisol being a prime example. Morning cortisol levels are
typically twice those found later in the day.49–51 The
magnitude of this effect is illustrated in the Figure, which
has been redrawn from the results of Hackney and Viru.49
When conducting exercise research, these fluctuations
and circadian variations need to be addressed. The
magnitude of exercise responses may not be similar at
different times of the day, even if the exercise intensity and
duration are held constant.1,49 Investigators should plan
accordingly to carefully control and replicate the time of
day in which research testing is conducted and the
hormonal specimen collected.52,53
PROCEDURAL-ANALYTIC FACTORS
The second category of factors influencing hormonal
measurements is made up of those factors that are
procedural or analytic in nature. That is, these factors are
determined, selected, or in some way controlled (poten-
tially) by the investigators conducting or the participant
Figure. The mean 24-hour cortisol responses of endurance-trained men (n = 17) on 3 separate days: a control baseline day with no
exercise, a high-intensity exercise day (1 hour in the morning and 45 minutes in the afternoon of interval training at 100% to 110% maximal
oxygen uptake [repeated bouts of 2 minutes of exercise and 2 minutes of recovery in each session]), and a moderate-intensity exercise day
(1 hour in the morning and 45 minutes in the afternoon of continuous aerobic training at 60% to 65% of maximal oxygen uptake). The figure
has been redrawn with permission.49
Journal of Athletic Training 633

involved in the research.1 These factors can be viewed as
exogenous in nature.
Environment. When conducting investigations, it is
important to remember that excessive exposure to hot or
cold ambient temperatures can stimulate various endocrine
gland hormones: for example, those involved in water
balance (aldosterone) or energy substrate mobilization
(cortisol).37,54,55 Even elevated ambient relative humidity
(water vapor) can induce this effect, primarily due to
compromised heat dissipation through evaporation adding
to the body core temperature.55 These effects can be further
augmented if hypoxemia is induced, as with exposure to
high altitudes.56–58
Many of the exercise and exercise training hormonal
responses are affected tremendously by environmental
factors. In particular, catecholamines, growth hormone,
aldosterone, antidiuretic hormone (vasopressin), adreno-
corticotropic hormone, and cortisol are susceptible to
environmental conditions and show exacerbated responses
in various conditions.1,37,54,55 To minimize these influences,
it is critical to conduct exercise testing in controlled,
standardized conditions, such as in a laboratory. However,
if conducting field research (where standardization can be
impossible), then it is important to measure and record
environmental factors and convey them in any report of the
data.
Nutrition. The prior nutritional status and practices of a
research participant, including diet composition, caloric
intake, and timing of meals, can greatly influence the
hormones associated with energy substrate mobilization
and use (eg, insulin, glucagon, epinephrine, growth
hormone, insulinlike growth factor, cortisol).1,59,60 The
exact nature of the effect (augmentation or attenuation)
depends on the interaction of these nutritional factors and
how severe the alterations are from the participant’s
normal nutritional regimes.1,29,34
The hormones listed above are critical during exercise to
ensure that energy metabolism meets the demands of
exercise. Thus, a participant’s altered dietary practices and
nutritional status can change energy substrate (glycogen)
storage and availability.60–62 This, in turn, can cause the
hormonal response to exercise to vary to some degree. For
example, Galbo et al59 demonstrated that the glucagon,
epinephrine, growth hormone, and cortisol responses to
exercise were greater after 4 days of a low-carbohydrate,
high-fat diet.
Typically, in clinical settings, participants should fast
prior to (eg, for 8 hours before) blood hormonal evalua-
tions. It is not always practical, however, for athletes to
comply with such requests due to their high demand for
caloric intake to maintain energy balance, anabolism, and
muscle glycogen reserves. Therefore, a modified approach
may be necessary for this special population. All the same,
for a repeated-measures research design, exercise and
sports medicine investigators should try to control and
standardize the dietary practices of their participants as
much as possible to mitigate the effects of different diets
among and within volunteers.34,59
The eating disorder anorexia nervosa is a special concern
relative to nutrition status due to its profound effect on the
endocrine system.1,45,63 Anorexics tend to have lower levels
of resting luteinizing hormone, follicle-stimulating hor-
mone, and estradiol-b-17.63 Anorexia also affects the
pituitary-thyroid gland axis. Specifically, the condition is
associated with suppression of triiodothyronine, somewhat
decreased thyroxine, elevation of reserve triiodothyronine,
and occasionally decreases in thyroid-stimulating hor-
mone.63 Such a thyroidal state is referred to as the
euthyroid sick syndrome hormonal profile and can accom-
pany severe body-weight loss.3,45,63 The adrenocortical axis
also is affected, with higher levels of cortisol due to
increased liberation of corticotrophin-releasing hormone.63
Growth hormone is increased, although insulinlike growth
factor 1 (IGF-1) levels (which facilitate the physiologic
actions of growth hormone) are suppressed in anorexia.63
Due to the psychological aspects of anorexia nervosa,64,65
this condition could be discussed organizationally with
mental health issues. However, this condition is also
biologic in nature and, consequently, has powerful effects
on a multitude of endocrine measurements.
Stress and Sleep. Emotional stress and sleep deprivation
are both known to affect certain hormones within the
endocrine system. For example, emotionally distraught
individuals typically have elevated basal catecholamine,
growth hormone, cortisol, and prolactin levels.1,66–68 Those
hormones with highly circadian patterns (eg, luteinizing
hormone, follicle-stimulating hormone, adrenocorticotro-
pic hormone, cortisol) can demonstrate shifts in their
characteristic patterns when sleep cycles are disrupt-
ed.36,39,66–70
Factors such as stress and sleep deprivation also can
influence the hormonal response to exercise and exercise
training. Investigators must attempt to control these
factors whenever possible. In fact, it is advisable to have
a pre-exercise questionnaire completed by a participant to
monitor and evaluate the level of these factors, and if a
predetermined status is not attained, then hormonal
measures and exercise testing should be rescheduled.
As a footnote to this issue, many investigators in the
exercise and sports medicine area rely on college students
as research volunteers. Such students can have high levels
of emotional stress due to the demands of their education
(eg, examination periods, projects being due, oral reports).
Care should be taken to not study student participants in
times of high emotional stress, because a multitude of
hormones can display atypical values and responses.36
Physical Activity. Time between exercise sessions can
affect the hormonal profiles of individuals.71,72 If inade-
quate amounts of time have elapsed (limited recovery),
some hormonal responses at rest or in the subsequent
exercise testing can be attenuated and others augmented.
Furthermore, the magnitude of this effect can be influenced
by the exertion required of the prior exercise (eg, high-
intensity intervals require longer recovery).
Ideally, the researcher may require a 24-hour recovery
period before a participant reports to the laboratory for
testing. However, athletes may find it difficult to reduce
their training or miss a workout session for experimental
purposes. A modified approach may be necessary, such as
limiting recovery to 8 or 12 hours, somewhat preventing
stress and anxiety (which, as noted earlier, can themselves
affect the endocrine system) as a result of missing less
training time.1,72–74
A powerful influence on resting and exercise hormo-
nal response of a participant is the exercise training
status: trained versus sedentary. Better-trained participants
634 Volume 43 N Number 6 N December 2008

typically display greater effects on the neuroendocrine
system response. Many hormones usually show attenuated
resting and submaximal exercise responses in trained
individuals, although some can actually be augmented
(eg, testosterone in resistance-trained individuals) in
response to submaximal and maximal exercise.2,75–79 An
extensive dialogue on the influence of exercise training on
hormonal profiles at rest and in response to exercise is
beyond the scope of this paper, but the reader is directed to
additional references for more in-depth discussions.2,3
Participant Posture. As a person changes position,
changes occur in the plasma volume component of the
blood. Standing upright results in reduced plasma volume
compared with the recumbent position.80 These shifts in
plasma fluid are in response to gravitational effects as well
as alterations in capillary filtration and osmotic pres-
sures.80 Large molecular-size hormones and those bound to
high-weight carrier proteins could be trapped in the
vascular spaces; thus, a loss of plasma fluid increases the
concentration of these hormones (hemoconcentration).
Conversely, increasing plasma fluid decreases the concen-
tration of these hormones (hemodilution).37,81 These
adjustments in fluid volume to move in or out of the
vascular space due to posture shifts typically require 10 to
30 minutes.80,81
When blood is drawn to assess hormone levels in exercise
research, the participant’s position during specimen collec-
tion should be controlled and should be reported in
publications. This type of information is most certainly
necessary if a postural change lasts 10 minutes or longer.51,81
Specimen Collection. Proper precautions must be taken
in the collection and storage of blood specimens to ensure
their viability for hormonal analysis. In clinical and
exercise-related blood work, venous blood is the specimen
typically used. If the specimen is being obtained by
venipuncture, the tourniquet should remain on the
volunteer’s arm for 1 minute or less. Greater lengths of
time can result in fluid movement from the vascular bed
due to hydrostatic pressures.81 Once collected, the blood
sample should be centrifuged at 46C to separate the plasma
(if the collection tube contains anticoagulant) or allowed to
clot (if the collection tube is sterile) and then centrifuged
for serum. If centrifugation cannot be done immediately,
then the blood sample should be placed on ice, but
centrifuging without delay is recommended. Once separat-
ed, the plasma or serum should be stored at a temperature
of 2206C to 2806C until later analysis. The plasma or
serum should be stored in airtight, cryofreeze tubes with
screw caps, which allow for a longer storage period. It is
also advisable to split specimens into several aliquots if
multiple hormonal analyses are to be conducted. Once a
sample is thawed, it has a relatively short shelf life in a
refrigerator, and repeated unthawing and refreezing cycles
can degrade certain hormonal constituents.82–84 Care
should be taken to ensure that the assay procedures
employed are specific for plasma or serum, because in some
cases these procedures are not interchangeable (eg,
adrenocorticotropic hormone is measured in plasma).
Furthermore, examining the literature may be necessary
to determine if one form of blood component is more
popular or prevalently used in research.
In blood specimens, either plasma or serum is used for
biochemical analysis, but some hormonal measures also
can be assessed in urine and salivary samples. In general,
plasma and serum provide very similar values in hormonal
analysis and seldom is one considered better than the other
in blood analysis.83 Be aware, however, that specific assay
procedures do, in some situations, have a preferred bodily
fluid for analysis. Thus, the researcher must know which
substance each hormonal assay requires as the analyte and
plan accordingly. This type of information is provided by
the manufacturer of the analytical supplies and compo-
nents used in the assay procedures.
Urine and saliva are attractive specimens to collect
because they are noninvasive in nature. They do, however,
have certain drawbacks. Urine analysis tends to be limited
primarily to steroid-based hormones, and 24-hour urine
samples must be collected, a tedious and demanding
process for the participant. Urine measurements may not
always reflect real-time hormonal status either, because
urine can sit in the bladder for hours before being voided.
Saliva allows for easier sampling and can reflect hormonal
status in a more real-time fashion. However, saliva
primarily allows only for steroid hormonal assessments
(ie, constituents that can cross from the blood into the
salivary gland).85 Furthermore, saliva is limited to free
hormonal concentrations, because the protein-bound
constituents typically cannot pass through the salivary
gland. Research suggests that the blood and saliva levels
can mirror each other in their changes but not perfectly, so
associations (r values) of 0.7 to 0.8 are typically
found.1,82,85 Investigators must determine if these limita-
tions preclude the use of these biologic fluids in their
studies.82,85,86
Analytical Assays. A variety of biochemical analytical
methods (assays) exist for measuring hormones in biologic
specimens. Chromatographic, receptor, and immunologic
assays are all available. Perhaps the most prevalent
contemporary technique in use is immunologic assays, which
have variations such as chemoluminescence immunoassay
(CLIA), radioimmunoassays (RIA), enzyme immunoassays
(EIA), enzyme-linked immunoassays (ELISA), and electro-
chemoluminescence immunoassays (ECLIA).87–89 Each
technique has its strengths and weaknesses, and the
discussion of each is beyond the scope of this paper, but
the reader is directed to additional references for more
background and explanation.90–92
Exercise and sports medicine researchers need to know
the particular aspects of the hormonal assay techniques
they plan to use in their studies. Specifically, they should be
aware of the precision of the assay (How accurate is it?),
sensitivity of the assay (How small a change can it detect?),
and the specificity of the assay (How much cross-reactivity
is there with similar-looking chemical structures in the
specimen?). Ideally, the researcher wants the most precise,
highly sensitive, and specific assay obtainable, but cost
considerations can affect decision making in these matters.
It is advisable for the researcher to report precision,
sensitivity, and cross-reactivity values in publications to
allow readers to determine the quality of the analytical
techniques and procedures of the assays used. Additionally,
it is desirable to report the coefficient of variation (CV)
within and between assays for each hormone measured.
These data allow the reader to determine how well the
analytical technical procedures were carried out.92,93 One
step to mitigate the potential between-assays CV is to
Journal of Athletic Training 635

collect and analyze biologic samples in batches of
specimens and not as isolated specimens on a day-by-day
basis. However, caution is necessary here, because batches
that are too large can influence the outcome by creating
‘‘end-of-run effects’’ within the assay. That is, running a
large number of samples in a single batch may compromise
the precision of the technician performing the assay (ie,
procedural fatigue) or the kinetics of the specific assay may
be influenced by the length of time it takes to pipette the
various components in assay (ie, in adding the chemical
reagents to the first sample tubes versus the last tubes, too
much time has transpired, resulting in different lengths of
time for chemical reactions to take place within the
specimen tubes).92,93
Data Transformations. Before conducting statistical
analysis of the collected hormonal data, it may be
necessary to transform the data. Two of the most common
transformations typically seen in the literature are (1)
expressing the data as a percentage change from some
precondition basal value and (2) conducting a logarithmic
conversion of the data. The first is typically done to
account for relative changes in hormonal concentrations
when absolute magnitude of change may be misleading.
For example, a cortisol change from 10.0 to 12.0 mg/dL is
much different from a 2.0 to 4.0 mg/dL (20% versus 100%)
change, even though the absolute magnitude is identical. A
100% increase in the hormonal concentration may have
many more profound physiologic effects than the smaller
percentage. In the second case, logarithmic transformation
is normally performed due to a large degree of variance in
the participant data, resulting in a nonnormal distribution.
This can be due to sample size issues, variance in the
analytical technique, or the physiologic nature of the
hormone being studied. Regardless of the transformation
used, it is vital that the researcher report in the publication
if and how the data were manipulated before conducting
the statistical analysis and the rationale for performing the
transformation.87,94
A third data transformation that is used less frequently is
the area under-the-curve (AUC) procedure. This is carried
out when serial specimen samples from a participant are
tested (repeated-measures design). These serial values are
plotted, and then the area under the plotted responses is
integrated, thus collapsing several data values into one
response and potentially eliminating some of the variability
associated with many hormonal measurements.95 Some
investigators favor this approach, feeling that the overall
response of the hormone and gland in question can be
better quantified. Yet the procedure can be influenced by
the number of serial samples collected and the circadian
rhythm of the hormone. That is, highly variable (pulsatile)
hormones require more frequent specimen sampling, and
misleading results can occur if the sampling is too
infrequent.96
Data Analysis. Of course, the statistical procedures
applied to any research study are dictated by the design
of that study. Most research in the exercise and sports
medicine area tends to employ parametric analysis (eg, t
test, analysis of variance, Pearson correlation). These
analytical procedures work well with endocrine data,
provided the underlying assumptions for their use are not
violated.97 Furthermore, many North American journals
prefer this form of analysis due to the robust nature of the
technique and the reduced likelihood of making a type I
error (indicating that findings are significant when they are,
in fact, not significant). Nevertheless, nonparametric
analysis (eg, Wilcoxon, Mann-Whitney U, Friedman tests)
can be equally applicable for endocrine use when study
designs are not excessively complex and sample sizes are
relatively small.97 As above, it is vital that the researcher
report for publication what the specific statistical analy-
ses being used were and what the rationale was for their
use.96–99
Once assays are performed and statistical results
obtained, the researcher needs to try and understand the
data in order to interpret the magnitude of treatment
outcomes and other effects. In this interpretative process,
many researchers focus intently on obtaining statistical
significance. Although such significance is important, a
key question that has to be addressed in the data is the
issue of statistical significance versus practical (clinical)
significance within the hormonal findings. To address
that question, the researcher must take into account the
smallest clinically important positive and negative re-
sponse values or levels of the effect being researched: that
is, the smallest change values or levels that matter.
Studies can be statistically significant and yet largely
insignificant clinically. It is important to note that large
sample sizes can produce a statistically significant result,
even though limited or no practical importance is
associated with the finding.100
Effect sizes (ESs) are an increasingly important index
used to quantify the degree of practical significance of
study results.101 Once computed, the ES statistic can be a
useful indicator of the practical or clinical importance of
research results, because it can be operationally defined; it
is possible to give the observed ES a rating such as
negligible-trivial, moderate, or important-very large.102
From such ratings, the researcher can discern the form
and quantity of significance obtained in the study.
Furthermore, the ES statistic has 2 advantages over
traditional statistical significance testing: (1) it is indepen-
dent of the size of the sample, and (2) it is a scale-free
index. Therefore, ES can be interpreted uniformly in
different studies regardless of the sample size and the
original scales of the variables being examined.101
SUMMARY AND CONCLUSIONS
Over the last 25 years, exercise science and sports
medicine researchers have steadily increased the number of
studies being published that examined hormones and the
endocrine system. Regrettably, not all investigators work-
ing in this area of research are entirely aware of the factors
that must be accounted for and controlled in order to
ensure that data are valid and accurate. In this paper, we
have reviewed some of the key biologic and procedural-
analytic factors that can confound endocrine data and add
variance to hormonal outcome measurements. If research-
ers inexperienced with endocrinology design their studies to
monitor, control, and adjust for the factors mentioned
here, they will find more consistency in their endocrine data
and, thus, enhance the legitimacy of their research. Such
actions can greatly aid investigators in interpreting and
understanding endocrine data and, in turn, make their
research more scientifically sound.
636 Volume 43 N Number 6 N December 2008

ACKNOWLEDGMENTS
The main aspects of this paper were presented as part of a
symposium presentation at the 2000 National Athletic Trainers’
Association meeting in Nashville, Tennessee. This paper is
dedicated to my coauthor, Dr Atko Viru, who was a great
mentor, colleague, and friend. Regrettably, Dr Viru died while
this paper was under revision.
REFERENCES
1. Trembly MS, Chu SY, Mureika R. Methodological and statistical
considerations for exercise-related hormone evaluations. Sports
Med. 1990;20(2):90–108.
2. Kraemer WJ, Ratamess NA. Hormonal responses and adaptations
to resistance exercise and training. Sports Med. 2005;35(4):339–361.
3. McMurray RG, Hackney AC. The endocrine system and exercise.
In: Garrett W, ed. Sports Medicine. New York, NY: Williams &
Wilkins; 2000:135–162.
4. International Union of Pure and Applied Chemistry (International
Union of Biochemistry and Molecular Biology). Recommendations
on organic & biochemical nomenclature, symbols & terminology.
http://www.chem.qmul.ac.uk/iupac/. Accessed August 19, 2008.
5. Warne GL, Kanumakala S. Molecular endocrinology of sex
differentiation. Sem Reprod Med. 2002;20(3):169–180.
6. Webb ML, Wallace JP, Hamill C, Hodgson JL, Mashaly MM.
Serum testosterone concentration during two hours of moderate
intensity treadmill running in trained men and women. Endocr Res.
1984;10(1):27–38.
7. Bunt JC, Bahr JM, Bemben DA. Comparison of estradiol and
testosterone levels during and immediately following prolonged
exercise in moderately active and trained males and females. Endocr
Res. 1987;13(2):157–172.
8. Pedersen BK, Hoffman-Goetz L. Exercise and the immune system:
regulation, integration, and adaptation. Physiol Rev. 2000;80(3):
1055–1081.
9. Foster DL, Nagatani S. Physiological perspectives on leptin as a
regulator of reproduction: role in timing puberty. Biol Reprod.
1999;60(2):205–215.
10. Ruby BC, Robergs RA. Gender differences in substrate utilisation
during exercise. Sports Med. 1994;17(6):393–410.
11. Bunt JC. Metabolic actions of estradiol: significance for acute and
chronic exercise responses. Med Sci Sports Exerc. 1990;22(3):
286–290.
12. Hackney AC, McCracken-Compton MA, Ainsworth BA. Substrate
metabolism responses to submaximal exercise in the midfollicular
and midluteal phases of the menstrual cycle. Int J Sport Nutr.
1994;4(3):299–308.
13. Hackney AC, McMurray RG, Judelson DA, Harrell JS. Relation-
ship between caloric intake, body composition, and physical activity
to leptin, thyroid hormones, and cortisol in adolescents.
Jpn J Physiol. 2003;53(6):475–479.
14. Horswill CA, Zipf WB, Kien CL, Kahle EB. Insulin’s contribution
to growth in children and the potential for exercise to mediate
insulin’s action. Pediatr Exerc Sci. 1997;9(1):18–32.
15. Amile SA, Caprio S, Sherwin RS, Plewe G, Haymond MW,
Tamborlane WV. Insulin resistance of puberty: a defect restricted to
peripheral glucose metabolism. J Clin Endocr Metab. 1991;72(2):
277–282.
16. Isurugi K, Fukutani K, Takayasu H, Wakabayashi K, Tamaoki B.
Age-related changes in serum luteinizing hormone and follicle-
stimulating hormone level in normal men. J Clin Endocr Metab.
1974;39(5):955–957.
17. Purifoy EE, Koopmars LH, Tatum RW. Steroid hormones and
aging: free testosterone, testosterone and androstenedione in normal
females age 20–87 years. Hum Biol. 1980;52(2):181–191.
18. Orentreich N, Brind JL, Rizer RL, Vogelman JH. Age changes and sex
differences in serum dehydroepiandrosterone sulfate concentrations
throughout adulthood. J Clin Endocr Metab. 1984;59(3):551–555.
19. Aloia JF, Feuerman M, Yeh JK. Reference range for serum
parathyroid hormone. Endocr Pract. 2006;12(2):137–144.
20. Adlercreutz H, Gorbach SL, Goldin BR, Woods MN, Dwyer JT,
Hamalainen E. Estrogen metabolism and excretion in Oriental and
Caucasian women. J Natl Cancer Inst. 1994;86(14):1076–1082.
21. Benn PA, Clive JM, Collins R. Medians for second-trimester
maternal serum alpha-fetoprotein, human chorionic gonadotropin,
and unconjugated estriol: differences between race or ethnic groups.
Clin Chem. 1997;43(2):333–337.
22. Mittelmark RA. Hormonal responses to exercise in pregnancy. In:
Mittelmark RA, Wiswell RA, Drinkwater BL, eds. Exercise in
Pregnancy. Baltimore, MD: Williams & Wilkins; 1991:175–184.
23. Wang C, Christenson P, Swerdloff R. Editorial: clinical relevance of
racial and ethnic differences in sex steroids. J Clin Endocr Metab.
2007;92(7):2433–2435.
24. Abbott WG, Foley JE. Comparison of body composition, adipocyte
size, and glucose and insulin concentrations in Pima Indian and
Caucasian children. Metabolism. 1987;36(6):576–579.
25. Ivandic A, Prpic-Krizevac I, Sucic M, Iuric M. Hyperinsulinemia
and sex hormones in healthy premenopausal women: relative
contribution of obesity, obesity type, and duration of obesity.
Metabolism. 1998;47(1):13–19.
26. Hansen BC, Jen KL, Belbez Pek SB, Wolfe RA. Rapid oscillations
in plasma insulin, glucagon, and glucose in obese and normal weight
humans. J Clin Endocrinol Metab. 1982;54(4):785–792.
27. Florkowski CM, Collier GR, Zimmet PZ, Livesey JH, Espiner EA,
Donald RA. Low-dose growth hormone replacement lowers plasma
leptin and fat stores without affecting body mass index in adults with
growth hormone deficiency. Clin Endocrinol (Oxf). 1996;45(6):769–
773.
28. Pasquali R, Vicennati V. Activity of the hypothalamic-pituitary-
adrenal axis in different obesity phenotypes. Int J Obes Relat Metab
Disorders. 2000;24(suppl 2):S47–S49.
29. McMurray RG, Hackney AC. Interactions of metabolic hormones,
adipose tissue and exercise. Sports Med. 2005;35(5):393–412.
30. Hurley BF, Nemeth PM, Martin WH III, Hagberg JM, Dalsky GP,
Holloszy JO. Muscle triglyceride utilization during exercise: effect of
training. J Appl Physiol. 1986;60(2):562–567.
31. Rahkila P, Soimajarvi J, Karvinrn E, Vihko V. Lipid metabolism
during exercise, II: respiratory exchange ratio and muscle glycogen
content during 4 h bicycle ergometry in two groups of healthy men.
Eur J Appl Physiol Occup Physiol. 1980;44(3):246–254.
32. Pasman WJ, Westertrep-Plantenga MS, Saris WHM. The effect of
exercise training on leptin levels in obese males. Am J Physiol.
1998;274(2, pt 1):E280–E286.
33. Ryan AS, Partley RE, Elahi D, Goldberg AP. Changes in leptin and
insulin action with resistive training in postmenopausal women.
Int J Obes Relat Metab Disord. 2000;24(1):27–32.
34. Hackney AC. Stress and the neuroendocrine system: the role of
exercise as a stressor and modifier of stress. Expert Rev Endocrinol
Metab. 2006;1(6):783–792.
35. Dorn LD, Burgess ES, Dichek HL, Putman FW, Chrousos GP,
Gold PW. Thyroid hormone concentrations in depressed and
nondepressed adolescents: group differences and behavioral rela-
tions. J Am Acad Child Adolesc Psychiatry. 1996;35(3):299–306.
36. Vaernes R, Ursin H, Darragh A, Lambe R. Endocrine response
patterns and psychological correlates. J Psychosom Res. 1982;26(2):
123–131.
37. Hackney AC. Exercise as a stressor to the neuroendocrine system.
Medicina (Kaunas). 2006;42(10):788–797.
38. Hamner MB, Hitri A. Plasma beta-endorphin levels in post-
traumatic stress disorder: a preliminary report on response to
exercise-induced stress. J Neuropsychiatry Clin Neurosci. 1992;4(1):
59–63.
39. Gerra G, Volpi R, Delsignore R, et al. ACTH and beta-endorphin
responses to physical exercise in adolescent women tested for anxiety
and frustration. Pyschiatry Res. 1992;41(2):179–186.
40. Cohen S, Kamarck T, Mermelstein R. A global measure of perceived
stress. J Health Soc Behav. 1983;24(4):385–396.
Journal of Athletic Training 637

41. Beck AT, Epstein N, Brown G, Steer RA. An inventory for
measuring clinical anxiety: psychometric properties. J Consult Clin
Psychol. 1988;56(6):893–897.
42. Landgren B, Aedo A, Diczfalusy E. Hormonal changes associated
with ovulation and luteal function. In: Flamigni C, Givens J, eds.
The Gonadotropins: Basic Science and Clinical Aspects in Females.
London, England: Academic Press; 1982:200–212.
43. Hackney AC, Cyren HC, Brammeier M, Sharp RL. Effects of the
menstrual cycle on insulin-glucose at rest and in response to exercise.
Biol Sport. 1993;10(2):73–81.
44. Vanheest JL, Mahoney CE, Rodgers CD. Oral contraceptive use
and physical performance. In: Kramer WJ, Rogol A, eds. The
Endocrine System in Sports and Exercise. Oxford, England: Black-
well Publishing; 2005:250–260.
45. Loucks AB. Physical activity, fitness and female reproductive
morbidity. In: Bouchard C, Shepard RJ, Stephens T, eds. Physical
Activity, Fitness and Health: International Proceedings and Con-
sensus Statement. Champaign, IL: Human Kinetics; 1994:943–
954.
46. Matsumoto AM, Bremner WJ. Modulation of pulsatile gonadotro-
pin secretion by testosterone in man. J Clin Endocrinol Metab.
1984;58(4):609–614.
47. Rose RM, Kreuz LE, Holaday JW, Sulak KJ, Johnson CE. Diurnal
variation of plasma testosterone and cortisol. J Endocrinol. 1972;
54(1):177–178.
48. Rose SR, Nisula BC. Circadian variation of thyrotropin in
childhood. J Clin Endocrinol Metab. 1989;68(6):1086–1090.
49. Hackney AC, Viru A. Twenty-four-hour cortisol response to
multiple daily exercise sessions of moderate and high intensity. Clin
Physiol. 1999;19(2):178–182.
50. Weitzman ED. Circadian rhythms and episodic hormone secretion.
Annu Rev Med. 1976;27:225–243.
51. Goodman HM. Endocrinology concepts for medical students. Adv
Physiol Educ. 2005;25(1–4):213–224.
52. Hackney AC, Zack E. Physiological day-to-day variability of select
hormones at rest in exercise-trained men. J Endocrinol Invest.
2006;29(6):RC9–RC12.
53. Schulz P, Knabe R. Biological uniqueness and the definition of
normality: part 2—the endocrine ‘‘fingerprint’’ of healthy adults.
Med Hypotheses. 1994;42(1):63–68.
54. Finberg JP, Berlyne GM. Renin and aldosterone secretion following
acute environmental heat exposure. Isr J Med Sci.
1976;12(6):844–847.
55. Galbo H, Houston ME, Christensen NJ, et al. The effect of water
temperature on the hormonal response to prolonged swimming.
Acta Physiol Scand. 1979;105(3):326–337.
56. Mordes JP, Blume FD, Boyer S, Zheng MR, Braverman LE. High
altitude pituitary-thyroid dysfunction on Mount Everest. New
Engl J Med. 1983;308(19):1135–1138.
57. Rastogi GK, Malhotra MS, Srivastava MC, et al. Study of the
pituitary-thyroid functions at high altitude in man. J Clin Endocrinol
Metab. 1977;44(3):447–452.
58. Hoyt RW, Honig A. Body fluid and energy metabolism at high
altitude. In: Fregley MJ, Blatteis CM, eds. Handbook of Physiology,
Section 4: Environmental Physiology. New York, NY: Oxford
University Press; 1996:1277–1289.
59. Galbo H, Holst JJ, Christensen NJ. The effect of different diets and
of insulin on the hormonal response to prolonged exercise. Acta
Physiol Scand. 1979;107(1):19–32.
60. Phinney SD, Horton ES, Sims EA, Hanson JS, Danforth E Jr,
LaGrange BM. Capacity for moderate exercise in obese subjects
after adaptation to a hypocaloric, ketogenic diet. J Clin Invest.
1980;66(5):1152–1161.
61. Jezova-Repcekova D, Vigas M, Klimes I. Decreased plasma cortisol
response to pharmacological stimuli after glucose load in man.
Endocrinol Exp. 1980;14(2):113–120.
62. Bonen A, Belcastro AN, MacIntyre K, Gardner J. Hormonal
responses during intense exercise preceded by glucose ingestion.
Can J Appl Sport Sci. 1980;5(2):85–90.
63. Støving RK, Hangaard J, Hansen-Nord M, Hagen C. A review of
endocrine changes in anorexia nervosa. J Psychiatr Res. 1999;33(2):
139–152.
64. Casper RC. Recognizing eating disorders in women. Psychophar-
macol Bull. 1998;34(3):267–269.
65. So¨dersten P, Bergh C, Zandian M. Psychoneuroendocrinology of
anorexia nervosa. Psychoneuroendocrinology. 2006;31(10):1149–1153.
66. VanHelder T, Radomski MW. Sleep deprivation and the effect on
exercise performance. Sports Med. 1989;7(4):235–247.
67. Aakvaag A, Bentdal O, Quigstad K, Walstad P, Ronningen H,
Fonnum F. Testosterone and testosterone binding globulin (TeBG)
in young men during prolonged stress. Int J Androl. 1978;1(6):22–31.
68. Aakvaag A, Sand T, Opstad PO, Fonnum F. Hormonal changes in
serum in young men during prolonged physical strain. Eur J Appl
Physiol Occup Physiol. 1978;39(7):283–291.
69. Diamond P, Brisson GR, Candas B, Peronnet F. Trait anxiety,
submaximal physical exercise and blood androgens. Eur J Appl
Physiol Occup Physiol. 1989;58(7):699–704.
70. Hackney AC, Feith S, Pozos R, Seale J. Effects of high altitude and
cold exposure on resting thyroid hormone concentrations. Aviat
Space Environ Med. 1995;66(4):325–329.
71. Viru AM, Hackney AC, Valja E, Karelson K, Janson T, Viru M.
Influence of prolonged continuous exercise on hormonal responses to
subsequent exercise in humans. Eur J Appl Physiol. 2001;85(6):578–585.
72. Hackney AC. The neuro-endocrine system, overload training, and
regeneration. In: Lehmann M, ed. Ulm International Conference
Proceeding: Performance, Overload Training and Regeneration.
London, England: Plenum Press; 1999:173–186.
73. Viru A, Karelson K, Smirnova T. Stability and variability in hormonal
responses to prolonged exercise. Int J Sports Med. 1992;13(3):230–235.
74. Hartley LH, Mason JW, Hogan RP, et al. Multiple hormonal
responses to graded exercise in relation to physical training. J Appl
Physiol. 1972;33(5):602–606.
75. Richter EA, Sutton JR. Hormonal adaptation to physical activity.
In: Bouchard C, Shephard RJ, Stephen T, eds. Physical Activity,
Fitness and Health: International Proceedings and Consensus
Statement. Champaign, IL: Human Kinetics; 1994:331–342.
76. Luger A, Deuster PA, Kyle SB, Gallucci WT, Montgomery LC,
Gold PW. Acute hypothalamic-pituitary-adrenal responses to the
stress of treadmill exercise: physiologic adaptations to physical
training. New Engl J Med. 1987;316(21):1309–1315.
77. Hackney AC, Sinning WE, Bruot BC. Reproductive hormonal
profiles in endurance-trained and untrained men. Med Sci Sports
Exerc. 1988;20(1):60–65.
78. Remes K, Kuoppasalmi K, Adlercreutz H. Effect of long-term
physical training on plasma testosterone, androstenedione, luteiniz-
ing hormone and sex-hormone–binding globulin capacity. Scand J
Clin Lab Invest. 1979;39(8):743–749.
79. Ha¨kkinen K, Pakarinen A. Acute hormonal responses to two
different fatiguing heavy-resistance protocols in male athletes. J Appl
Physiol. 1993;74(2):882–887.
80. Westendorp RG, Roos AN, Riley LC, Walma S, Frolich M,
Mienders AE. Chronic stimulation of atrial natriuretic peptide
attenuates the secretory responses to postural changes. Am J Med
Sci. 1993;306(6):371–375.
81. Fawcett JK, Wynn V. Effects of posture on plasma volume and
some blood constituents. J Clin Path. 1960;13:304–313.
82. Chen YM, Cintron NM, Whitson PA. Long-term storage of salivary
cortisol samples at room temperature. Clin Chem. 1992;38(20):304.
83. Calam RR. Reviewing the importance of specimen collection. J Am
Med Technol. 1977;39(6):297–300.
84. Sonntag O. Hemolysis as an interference factor in clinical chemistry.
J Clin Chem Biochem. 1986;24(2):127–139.
85. Obminski Z, Klusiewicz A, Stupnicki R. Changes in salivary and
serum cortisol concentrations in junior athletes following exercises
of different intensities. Biol Sport. 1994;11(1):49–57.
86. Caraway WT. Chemical and diagnostic specificity of laboratory tests:
effect of hemolysis, lipemia, anticoagulants, medications, contami-
nants, and other variables. Am J Clin Pathol. 1961;37(5):445–464.
638 Volume 43 N Number 6 N December 2008

87. Kopchick JJ, Sackmann-Sala L, Ding J. Primer: molecular tools
used for the understanding of endocrinology. Nat Clin Pract
Endocrinol Metab. 2007;3(4):355–368.
88. Bowers LD. Analytical advances in detection of performance-
enhancing compounds. Clin Chem. 1997;43(7):1299–1304.
89. Dudley RF. Chemiluminescence immunoassay: an alternative to
RIA. Lab Med. 1990;21(4):216–222.
90. Shah VP, Midha KK, Findlay JWA, et al. Bioanalytic method
validation: a revisit with a decade of progress. Pharm Res.
2000;17(12):1551–1557.
91. De Ronde W, van der Schouw YT, Pols HAP, et al. Calculation of
bioavailable and free testosterone in men: a comparison of 5
published algorithms. Clin Chem. 2006;52(9):1777–1784.
92. Rosner W, Auchus RJ, Azziz R, Sluss PM, Raff H. Position
statement: utility, limitations, and pitfalls in measuring testosterone:
an Endocrine Society position statement. J Clin Endocrinol Metab.
2007;92(2):405–413.
93. Rodbard D. Statistical quality control and routine data processing
for radioimmunoassays and immunoradiometric assays. Clin Chem.
1974;20(10):1255–1270.
94. Fraser CG, Harris EK. Generation and application of data on
biological variation in clinical chemistry. Crit Rev Clin Lab Sci.
1989;27(5):409–437.
95. Hackney AC, Premo MC, McMurray RG. Influence of aerobic
versus anaerobic exercise on the relationship between reproductive
hormones in men. J Sports Sci. 1995;13(4):305–311.
96. Veldhuis JD, Johnson ML. Deconvolution analysis of hormone
data. Methods Enzymol. 1992;210:539–575.
97. Kingle RD, Johnson GF. Statistical procedures. In: Tietz NW, ed.
Textbook of Clinical Chemistry. Philadelphia, PA: WB Saunders;
1986:287–355.
98. Pincus SM, Hartman ML, Roelfsema F, Thorner MO, Veldhuis JD.
Hormone pulsatility discrimination via coarse and short time
sampling. Am J Physiol Endocrinol Metab. 1999;277(5, pt 1):E948–
E957.
99. Matthews DR. Time series analysis in endocrinology. Acta Paediatr
Scand Suppl. 1988;347:55–62.
100. Hopkins WG. Measures of reliability in sports medicine and science.
Sports Med. 2000;30(1):1–15.
101. Hojat M, Xu G. A visitor’s guide to effect sizes: statistical
significance versus practical (clinical) importance of research
findings. Adv Health Sci Educ Theory Pract. 2004;9(3):241–
249.
102. Cohen J. Statistical Power Analysis for the Behavioral Sciences.
2nd ed. Englewood, NJ: Lawrence Erlbaum Associates; 1988:
116–173.
Anthony C. Hackney, PhD, CPH, FACSM, and Atko Viru, PhD, DSc, contributed to conception and design; acquisition and analysis and
interpretation of the data; and drafting, critical revision, and final approval of the article.
Address correspondence to Anthony C. Hackney, PhD, CPH, FACSM, The University of North Carolina at Chapel Hill, CB # 8700,
Fetzer Building, Chapel Hill, NC 27599. Address e-mail to ach@email.unc.edu.
Journal of Athletic Training 639