This article was downloaded by: [Johann Christian Senckenberg]

On: 27 August 2014, At: 10:14

Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Research Quarterly for Exercise and Sport

Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/urqe20

Single versus Multiple Sets for Strength: A MetaAnalysis to Address the Controversy
a

Matthew R. Rhea , Brent A. Alvar & Lee N. Burkett

Department of Exercise and Wellness , Arizona State University , USA

Published online: 26 Feb 2013.

To cite this article: Matthew R. Rhea , Brent A. Alvar & Lee N. Burkett (2002) Single versus Multiple Sets for
Strength: A Meta-Analysis to Address the Controversy, Research Quarterly for Exercise and Sport, 73:4, 485-488, DOI:
10.1080/02701367.2002.10609050
To link to this article: http://dx.doi.org/10.1080/02701367.2002.10609050

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions

Research Note-Physiolol}Y

Research Querterly for Exercise end Sport

2002 by the American Alliance for Health,
Physical Education, Recreation and Dance

Vol. 73, No.4, pp. 4B5-4BB

Single Versus MUltiple Sets for Strength: A Meta-Analysis

to Address the Controversy

Downloaded by [Johann Christian Senckenberg] at 10:14 27 August 2014

Matthew R. Rhea, Brent A. Alvar, and Lee N. Burkett

Key words: resistance training, weight training

eight training can be used as a stressor to overload

the neuromuscular system and develop strength.
A number of weight training variables can be manipulated to overload this system. The main variables include
volume, intensity, and the amount of rest allowed between sets or workouts. Of these variables, volume has
received much interest among researchers and professionals, and considerable debate has arisen (Byrd, 1999)
regarding an increased strength benefit of multiple-set
programs.
The debate over the amount of volume needed to
elicit maximal strength gains has continued in recent
years with several narrative reviews (Carpinelli & Otto,
1999; Feigenbaum & Pollock, 1999) of research literature
comparing single and multiple sets of training. These
reviews have determined that single-set training programs elicit similar strength increases or health benefits
(especially in untrained individuals) compared to multiple-set programs due to the inability of most of these
studies to identify a statistical difference at the .05IeveI.
The reliance on probability values (p) places considerable importance on statistical power. If statistical
power is low, the possibility of committing a Type II error (failing to reject the Null hypothesis despite a true
difference existing) is increased. With this in mind, a

Submitted: October 29, 2001

Accepted: February 19,2002
MatthewR. Rhea, BrentA. Alvar, andLee N. Burkettare with
the Department of Exercise and WeI/ness at Arizona State
University.

ROES: December 2002

power analysis (Cohen, 1988) was performed on a random

sample of 10 studies from the literature comparing single
and multiple sets. The mean power was calculated to be
0.56. Cohen (1988) reported that a power below 0.80
would incur too great a risk of a Type II error. Statistical
significance is also heavily affected by large variance and
small sample sizes. This can result in large differences
between groups being deemed nonsignificant or in
small differences reaching statistical significance solely
based on sample size (and statistical power).
The body of research comparing single and multiple sets for maximal strength gains contains many studies performed with small sample sizes and low power;
therefore, reliance solely on p values may be misleading. Completing a meta-analysis could be a valuable asset in this situation. This procedure, popularized by
Glass ( 1982), combines the results ofindependent studies, sums sample sizes across studies, and increases statistical power. It involves calculating an effect size (ES),
which represents the magnitude, in standard deviation
units, of a treatment. The ES is an objective measure of
meaningfulness, which when combined with statistical
significance, provides a more accurate picture of treatment effects.
A meta-analysis has two major advantages over the
narrative literature reviews previously performed in this
body of research (Thomas & French, 1986). Procedures
for decision making are provided and reported in. a
meta-analytical procedure, whereas narrative reviewsrely
solely on the reviewer's jurisdiction. It also provides a
quantitative method for analyzing research findings
through calculating and comparing the ES. The purpose of this research was to systematically examine studies comparing single- and triple-set training programs
for strength by calculating and examining ESs.

485

Rhea, AlvaT, and Burkett

Method
Literature Search
Searches were performed for all English language
studies from 1962-2000 comparing single- and tripleset training programs. Computer searches of Science
Citation Index, National Library of Medicine, Sport Discus, ERIC, and Medline were performed. Hand searches
of relevant journals and reference lists obtained from
articles were conducted.

variance (Hedges & Olkin, 1985), with single set training

groups set as the control group. Testsfor homogeneitywere
also conducted (Hedges & Olkin, 1985). Analysis ofvariance with confidence intervals identified differences
among variables. Rosenthal's ( 1979) formula was used to
calculate the tolerance of the overall ES to negative experimental results. This formula calculates the estimated number ofstudies demonstrating null results that
would need to exist (that were not included in the analysis) to bring the overall ES to nonsignificance.

Downloaded by [Johann Christian Senckenberg] at 10:14 27 August 2014

Inclusion Criteria
Studies were included with both single- and tripleset training groups and compared identical strength
measures among the groups. In the event that necessary statistical or descriptive information was deficient,
solicitations were made from the primary investigators.
A total of 16 studies were found (Berger, 1962; Capen,
1956; de Hoyos et al., 1998; de Hoyos, Herring,
Garzarella, Werber, & Brechue, 1997; Haas, Garzarella,
de Hoyos, & Pollock, 1998, 2000; Jacobson, 1986;
Kraemer, 1997; Kraemer et al., 1997; Leighton, Holmes,
Benson, Wooten, & Schmerer, 1967; Messier & Dill,
1985; Reid, Yeater, & Ulrich, 1987; Rhea, Alvar, Ball, &
Burkett, 2002; Sanborn et al., 2000; Silvester, Stiggings,
McGown, & Bryce, 1982; Starkey et al., 1996).
Coding of Studies
The primary investigator read and coded each study
to identify descriptive information and program design.
Once the studies were coded, five ES were required for
each level or variable to be examined statistically. Program methodology was examined. To be classified as a
controlled study, both groups must have trained at equal
intensity with the same variation in programs (i.e., if the
multiple-set programs were periodized, the single-set
program must be periodized in the same manner). The
participants' status was divided into trained and untrained classifications. Participants must have been part
of a weight-training program for at least 1 year prior to
the study to be considered as trained. The length of the
training programs were divided into categories of 1-5,
6-10,11-15 weeks, and so on. Coder drift was assessed
(Orwin, 1994) by randomly selecting 10 studies for
recoding. Per case agreement was determined by dividing the variables coded the same by the total number of
variables. A mean agreement of 0.90 was required for
acceptance.
Statistical Data
Effect sizes were calculated using means and pooled
standard deviations and weighted by the inverse of the

486

Results
A total of 16 relevant studies were coded, yielding
93 ESs (see Table 1). Coder drift was assessed to be 0.96.
Thus, the coding process used in this analysis was found
to be consistent. The overall ES of all studies included
in the analysis was calculated to be 0.28 (SD= 0.56). The
mean ES of those studies demonstrating equated intensity and controlled variation was calculated to be 0.70
(SD = 0.92). In those studies failing to maintain this
control, the ESwas calculated to be 0.25 (SD= 0.70). This
difference approached significance, F(1, 35) = 2.55, p=
.12. The test for homogeneity was significant, QT (92) =
222.9, P < .001, warranting examination of potential
moderating variables or variables that resulted in ESs of
differing magnitude.
Training status was found to be a significant variable. Participants who were classified as trained exhibitedan ES (0.55; SD=0.59) that was significantly greater
than the ES of untrained participants, (0.23; SD= 0.54)
F(1, 55) = 4.03, p< .05. Only three lengths of training
categories produced enough ESs to be analyzed (see
Table 1). Training programs of 11-15 weeks and 21-25
weeks resulted in higher ESs (0.39 and 0.33, respectively) than 6-10-week programs (0.17), but this difference was statistically nonsignificant (p = .26), possibly
due to the lack of available ESs for longer training programs. More research is needed using training programs longer than 15 weeks. The tolerance of these
results to null experimental results was calculated to
be 571. Thus, 571 ESs demonstrating the null hypothesis would have to exist to bring the current ES to
nonsignificance.

Discussion
The results of this meta-analysis demonstrate that
three sets of training increase strength to a greater degree than single-set training (ES = 0.28). Studies that
maintained stringent methodological control of train-

ROES: DeCflmber 2002

Downloaded by [Johann Christian Senckenberg] at 10:14 27 August 2014

Rhea, Alvar, and Burkett

ing intensity and variation exhibited a larger difference

between training programs (ES = 0.70). The differences
in ESs between those studies that controlled training intensity and program variation demonstrate an important
aspect of research in strength improvements. To investigate the impact ofvolume on strength development it is
important to hold all other variables constant. Because of
the greater methodological control, the ES ofO.70 is most
likelya more accurate measure of the greater strength gains
that are achieved with multiple set training.
Another finding of this analysis is that the difference
in strength gains is more pronounced in trained individuals. While there is a significant difference between gains
for untrained individuals favoring multiset programs,
this difference increases as the individual becomes
more accustomed to strength training. Theoretically, as
the individual becomes more accustomed to a weight
training program, more volume is needed to maximally
overload the neuromuscular system (Fleck & Kraemer,
1997).
As previously mentioned, a power analysis identified low statistical power among the individual studies
included in this analysis. This lack of power appears to
have resulted in the inability to detect the differences
between single and multiset programs for strength despite its existence (Type II Error), as only a few studies
(Berger, 1962; Kraemer, 1997; Kramer, Stone, O'Bryant
et al., 1997; Kramer, Stone, O'Bryant, Conley et al., 1997)
reached the .05 levels of significance. By combining the
results of these studies, power was increased and the ES
identified a significant difference between single- and
multiple-set training protocols for strength gains. The
magnitude of the ES is influenced by participants' training status, length of training program, and methodological control of the study.

Table 1. Meaneffect sizes and 95% confidence intervals

Variable

Overall
Controlled
methodology
Uncontrolled
Training status
Trained participants
Untrained participants
Length oftraining (weeks)
6-10
11-15
21-25

Confidence
SO
Lower Upper

ES

93

.28

.56

.22

.32

53
40

.70
.25

.92
.70

.31
-.16

1.10
.65

38

.46
.19

.50
.51

.40
.11

.51
.28

37

.17

48

.39
.33

.69
.56
.79

-.06
.23
-.32

.40
.56
.99

19

Note. n = number of effect sizes; ES = effect size; SO = standard

deviation.

ROES: December 2002

References
References marked with an asterisk indicate studies included in the meta-analysis.
*Berger, R (1962). Effect ofvaried weight training programs
on strength. Research Quarterly, 33, 168-18l.
Byrd, R (1999). Strength training: Single versus multiple sets.
Sports Medicine, 27,409-416.
*Capen, E. (1956). Study offour programs of heavy resistance
exercise for development of muscular strength. Research
Quarterly, 27, 132-142.
Carpinelli, R N., & Otto, R M. (1999). Strength training:
Single versus multiple sets-The authors' reply. Sports
Medicine, 27,412-416.
Cohen, j. (1988). Statisticalpower analysisfor the behavioral sciences (2nd ed.). Hillsdale, NJ: L. Erlbaum Associates.
*de Hoyos, D., Abe, T., Garzarella, L., Hass, C.J., Nordman,
M., & Pollock, M. L. (1998). Effects of6 months of highor low-volume resistance training on muscular strength
and endurance [abstract]. Medicine & Science in Sports &
Exercise, 30(Suppl.), S165.
*de Hoyos, D., Herring, L., Garzarella, L., Werber, G., &
Brechue, W. F. (1997). Effect of strength training volume
on the development of strength and power in adolescent tennis players [abstract]. Medicine & Science in Sports
& Exercise, 29(Suppl.), S164.
Feigenbaum, M. S., & Pollock, M. L. (1999). Prescription of
resistance training for health and disease. Medicine &
Science in Sports & Exercise, 31, 38-45.
Fleck, S., & Kraemer, W.J. (1997). Designingresistance training
programs (2nd ed.). Champaign, IL: Human Kinetics.
Glass, G. V. (1982). Meta-analysis-An approach to the synthesis ofresearch results. JournalofResearch in Science Teaching, 19(2),93-112.
*Hass, GJ., Garzarella, L., de Hoyos, D., & Pollock, M. L. (1998).
Effects of training volume on strength and endurance in
experienced resistance trained adults [abstract]. Medicine
& Science in Sports & Exercise, 30(Suppl.), S115.
*Hass, C. J., Garzarella, L., de Hoyos, D., & Pollock, M. L.
(2000). Single versus multiple sets in long-term recreational weightlifters. Medicine & Science in Sports & Exercise, 32, 235-242.
Hedges, L. V, & Olkin, I. (1985). Statisticalmethods for metaanalysis. Orlando, FL: Academic Press.
*jacobson, B. (1986). A comparison of two progressive weight
training techniques on knee extensor strength. Athletic
Training, 21,315-318,390.
*Kraemer, W.J. (1997). A series of studies-The physiological basis for strength training in American football: Fact
over philosophy. Journal of Strength and Conditioning Research, 11,131-142.
*Kramer,J., Stone, M., O'Bryant, H. S., Conley, M. S.,johnson,
R L., Nieman, D. C., Honeycutt, D. R, & Hoke, T. P.
(1997). Effect of single vs. multiple sets of weight training: Impact ofvolume, intensity, and variation. Journal of
Strength and ConditioningResearch, 11, 143-147. .
*Leighton,]. R, Holmes, D., Benson.]., Wooten, B., & Schmerer,
R (1967). A study on the effectivenessof ten different methods ofprogressive resistance exercise on the development of

487

Downloaded by [Johann Christian Senckenberg] at 10:14 27 August 2014

Rhea. Alvar, and Burkett

strength, flexibility, girth and bodyweight Journal of theAssociation ofPhysical and MentalRehabilitation, 21(3), 78-81.
*Messier, S., & Dill, M. (1985). Alterations in strength and
maximal oxygen uptake consequent to Nautilus circuit
weight training. Research Quarterly for Exercise and Sport,
56,345-351.
Orwin, R (1994). Evaluating coding decisions. In Cooper
and Hedges (Eds.), The handbook of research synthesis (pp.
139-162). New York: Russell Sage Foundation.
*Reid, C. M.,Yeater, R A, & Ullrich, I. H. (1987). Weight training and strength, cardiorespiratory functioning and body
composition of men. BritishJournal of Sports Medicine,
21(1),4Q-44.
*Rhea, M. R, Alvar, B. A, Ball, S. B., & Burkett, L. N. (2002.)
Three sets of weight training superior to one set with
equal intensity for eliciting strength. Journal of Strength
and ConditioningResearch, 16, 490-494.
Rosenthal, R (1979). The file-drawerproblem and tolerance for
null results. Psychowgy Bulletin, 86,638--641.
*Sanbom, K, Boros, R, Hruby, R, Schilling,B., O'Bryant, H. S.,
Johnson, R L., Stone, M. E., & Stone, M. H. (2000). Shortterm performance effects ofweight training with multiple
sets not to failure vs. a single set to failure in women. Journal
ofStrength and Conditioning Research, 14,328-331.

488

*Silvester,L., Stiggings, C., McGown, C., & Bryce, G. R (1982).

The effect of variable resistance and free-weight training on strength and vertical jump. National Strength and
ConditioningJournal, 3(6), 30-33.
*Starkey, D. B., Pollock, M. L., Ishida, Y, Welsch, M. A.,
Brechue, W. E, Graves,]. E., & Feigenbaum, M. S. (1996).
Effect of resistance training volume on strength and
muscle thickness. Medicine & Science in Sports & Exercise,
28,1311-1320.
Thomas.j., & French, K (1986). The use of meta-analysis in
exercise and sport: A tutorial. Research Quarterly for Exercise and Sport, 57, 196-204.

Authors' Note
Please address all correspondence concerning this article to Matthew Rhea, Department of Exercise and
Wellness, Arizona State University, East Campus, 6113
S. Kent, CLRB, Mesa,AZ 85212.
E-mail: matthew.rhea@asu.edu

ROES: December 2002