J. Phys. Ther. Sci.

30: 1221–1226, 2018

The Journal of Physical Therapy Science

Original Article

Clinical method to assess shoulder strength

related to front crawl swimming power in male
collegiate swimmers

Takenori Awatani, MS, RPT1, 2)*, Ikuhiro Morikita, PhD, MD2, 3), Seigo Mori, MS1),
Junji Shinohara, PhD1), Yasutaka Tatsumi, MS1)
1) Faculty of Sports Science, Kyushu Kyoritsu University: 1-8 Jiyuugaoka, Yahatanishi-ku, Kitakyushu,
Fukuoka 807-8585, Japan
2) Graduate School of Sport and Exercise Sciences, Osaka University of Health and Sport Sciences, Japan
3) Faculty of Physical Education, Osaka University of Health and Sport Sciences, Japan

Abstract. [Purpose] Although a correlation has been reported between shoulder strength and maximum swim-
ming power during arm-only swimming, the correlation between shoulder strength and maximum swimming pow-
er during front crawl swimming remains unclear. This study aimed to confirm the validity of a clinical assessment
method for shoulder strength related to front crawl swimming power. [Participants and Methods] Study participants
included 9 healthy male collegiate swimmers. Shoulder strength, including extension and internal rotation torque
and swimming power, were measured. [Results] Maximum swimming power was significantly correlated with
extension torque in the position of maximum shoulder abduction on the dominant side (r=0.844). No significant
correlations were observed between the swimming velocity-to-swimming power ratio and the rate of bilateral dif-
ferences in extension torque in the position of maximum shoulder abduction. [Conclusion] The extensor strength
in the position of maximum shoulder abduction was significantly correlated with the maximum swimming power,
suggesting that this assessment method is useful for front crawl swimmers. Notably, measurements on the dominant
side may provide useful data that are essential in training to improve front crawl swimming propulsion.
Key words: Correlation, Validity, Extensor strength
(This article was submitted May 10, 2018, and was accepted Jul. 3, 2018)

INTRODUCTION
The front crawl stroke consists of an arm pull and leg kick. In particular, arm-only swimming is used for swim training
and rehabilitation to improve propulsion1–3). The shoulder joint is one of the most important joints used during arm pull. We
investigated the relationship between arm-only swimming and shoulder strength and found that maximum swimming power
(MSP) was significantly associated with internal rotator strength in the position of abducted and external rotation (r=0.85;
p<0.001)4). Additionally, we reported that the swimming velocity-to-swimming power ratio (SVPR) was significantly cor-
related to the rate of bilateral difference in the shoulder extensor strength in the position of maximum shoulder abduction (r=
−0.728; p=0.006)4). Although the link between shoulder strength and MSP in arm-only swimming has been investigated, the
correlation between shoulder strength and MSP during front crawl swimming is still unclear.
Swimming power is important to improve swimming velocity; therefore, studies have shown that swimming power and
swimming velocity are significantly positively correlated5–7). Similarly, shoulder strength is important for improving MSP
because muscular strength and technique are necessary for propulsion. However, because differences in the trajectory of
the arm during front crawl swimming and arm-only swimming have been reported8), there is a possibility that the relation-

*Corresponding author. Takenori Awatani (E-mail: [email protected])

©2018 The Society of Physical Therapy Science. Published by IPEC Inc.
This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial No Deriva-
tives (by-nc-nd) License. (CC-BY-NC-ND 4.0: https://creativecommons.org/licenses/by-nc-nd/4.0/)

1221
ship between shoulder strength and MSP may be different between front crawl swimming and arm-only swimming. Thus,
to determine appropriate muscular strength measurement methods for swimming athletes’ rehabilitation and training, it is
necessary to conduct research regarding front crawl swimming and arm-only swimming.
It would be beneficial to determine the relationship between muscle strength and MSP and SVPR during front crawl swim-
ming is important for understanding how to improve training and rehabilitation. The purpose of this study was to confirm the
validity of a clinical method used to assess shoulder strength related to front crawl swimming power.

PARTICIPANTS AND METHODS

This study was approved by the Research Ethics Committee of Kyushu Kyoritsu University (approval no. 2017-11).
Nine healthy male collegiate front crawl swimmers participated in this study (age, 19.2 ± 1.2 years; body weight, 64.7 ±
9.3 kg; height, 169.4 ± 6.8 cm; all values expressed as mean ± standard deviation [SD]). Participants had no shoulder pain or
shoulder surgery during the past 6 months. They received an oral and written explanation of the study and provided informed
consent. Experienced examiners performed muscle strength and swimming power measurements.
Swimming power was measured after muscle strength measurements. The maximum isometric force was measured using
a hand-held dynamometer (HHD) (Mobie MM100C; Minato Medical Science Co., Ltd., Japan), and extremity length, using
a digital calliper (D-500; Niigata seiki Co., Ltd., Japan). Maximum isometric force and extremity length were used to cal-
culate torque4). Extension force in the position of maximum shoulder abduction and internal rotation force in the position of
abducted and external rotation were also measured. The measurement position of the extension force was maximum shoulder
abduction with the elbow extended and the forearm in neutral9) (Fig. 1). For the extension force measurements, participants
stayed in a prone position with their toes, abdomen, chest, and mentum in contact with the floor. The measurement positions
used to determine the internal rotation force were 90 degrees of abduction and 90 degrees of external rotation, with the
elbows in 90 degrees flexion and the forearms in the neutral position9) (Fig. 1). For the internal rotation force measurements,
participants stayed in a prone position with their toes, abdomen, chest, elbow of on the measurement side, and mentum
in contact with the floor. The opposite upper extremity made contact with the body. The HHD was placed on a floor and
touched the heads of the metacarpal bones on the palmar side during both measurements. Participants performed a 3-second
maximum isometric contraction 3 times per session with more than 5 minutes of rest between sessions. The examiner verified
the measurements, which were repeated if the position changed or an error occurred9). The average of 3 repetitions was
calculated. Upper extremity length was defined as the distance from the acromion to the distal head of the third metacarpal
bone on the dorsal side4). The forearm length was defined as the distance from the lateral joint line of the elbow to the distal
head of the third metacarpal bone on the dorsal side4). The extension torque (ET) was calculated from the upper extremity
length and the extension force. The internal rotation torque (IT) was calculated from the forearm length and the internal
rotation force. The rate of bilateral difference, defined as the bilateral difference between the dominant and non-dominant
sides, was calculated using the equation [absolute (difference of both sides) / mean of both sides ×100]4). The rate of bilateral
differences in ET (RBET) and the rate of bilateral differences in IT (RBIT) were calculated.
Swimming power was measured four times during swimming in a 25-m indoor pool. Front crawl swimming was per-
formed 25 m once and 15 m three times using a simple swimming power device (drag boat)7) (Fig. 2). The 15-m swim was
performed from the 5- to 20-m point. Swimming velocity was defined as the average velocity of 10 m and calculated using
transit time of the head in 10 m (from the 10- to 20-m point). The maximum swimming velocity (MSV) was defined as the
swimming velocity during the 25-m swim. Swimming performances were captured using a moving image recorded by a
digital video camera (HDR-CX270V; Sony Marketing Inc., Japan) (at 60 fps) poolside. Swimming velocity and the towing
force of the 15-m swim were measured using the drag boat at three load levels4).
The regression equation (towing force = regression coefficient × swimming velocity + regression constant) was used to

Fig. 1. Measurements of shoulder muscle strength using a hand-

held dynamometer. Fig. 2. The swimming power measurement using the drag boat.
a: The extension force in the maximum shoulder abducted posi- The swimming velocity was defined as the average velocity of
tion. 10 m and was calculated using transit time of the head in 10 m
b: The internal rotation force in the abducted and external rotated (from the 10-m point to the 20-m point). The swimming velocity
position. and the towing force of the 15-m swim were measured using the
drag boat for three load levels. The swimming power was calcu-
lated using the swimming velocity and the towing force.

J. Phys. Ther. Sci. Vol. 30, No. 10, 2018 1222

calculate the regression coefficient and the regression constant. Swimming velocity and towing force values were obtained
during front crawl swimming when pulling the drag boat with three different load levels and during normal front crawl
swimming. Swimming power was calculated using swimming velocity and towing force (swimming power = towing force ×
swimming velocity)7). MSP was defined as the maximum value of the curve representing the relationship between swimming
power and swimming velocity7). SVPR was calculated using power-to-weight ratio (SVPR = MSV / MSP / kg)4).
Averages, SD, and 95% confidence intervals (CI) were calculated using basic statistics. Torque was used to analyse muscle
strength. The correlation between MSV and MSP, MSP and shoulder strength, shoulder extensor strength and shoulder
internal rotator strength, and SVPR and the rate of bilateral difference were examined. These relationships were evaluated
using Pearson’s product-moment correlation coefficient. Excel for Windows 2013 (Microsoft Japan Co., Ltd.) and R2.8.1
were used for statistical analysis; a p-value<0.05 was considered statistically significant.
Holm-Bonferroni correction was performed for statistically significant p values10, 11). Coefficient values of significant
correlations were defined based on the criteria of Hinkle et al.12): negligible, 0.00 to 0.30; low, 0.30 to 0.50; moderate, 0.50
to 0.70; high, 0.70 to 0.90; and very high, 0.90 to 1.00.

RESULTS
MSP was significantly correlated with ETD and ETN. No significant correlations were observed between MSP and ITD,
MSP and ITN, SVPR and RBET, or SVPR and RBIT.
Upper extremity length was 64.6 ± 3.1 cm (mean ± SD) on the dominant side and 64.6 ± 2.8 cm on the non-dominant side.
Forearm extremity length was 32.6 ± 1.5 cm on the dominant side and 32.6 ± 1.3 cm on the non-dominant side. Results of
measurements are shown in Table 1.
MSV was significantly associated with MSP (r=0.927; p<0.001) (Table 2). MSP was significantly correlated with ET

Table 1. Results of the muscle strength measurement

Mean (standard deviation)

EFD (N) 134.2 (21.9)
EFN (N) 130.5 (25.2)
IFD (N) 129.6 (20.3)
IFN (N) 123.6 (12.8)
ETD (Nm) 87.0 (16.7)
ETN (Nm) 84.6 (17.9)
ITD (Nm) 42.4 (8.1)
ITN (Nm) 40.4 (5.1)
RBET (%) 3.1 (5.5)
RBIT (%) 4.0 (10.5)
EFD: The extension force of the dominant side; EFN: The exten-
sion force of the non-dominant side; IFD: The internal rotation
force of the dominant side; IFN: The internal rotation force of
the non-dominant side; ETD: The extension torque of the domi-
nant side; ETN: The extension torque of the non-dominant side;
ITD: The internal rotation torque of the dominant side; ITN:
The internal rotation torque of the non-dominant side; RBET:
The rate of bilateral difference in extension torque; RBIT: The
rate of bilateral difference in internal rotation torque.

Table 2. Results of the front crawl swimming measurement

Pearson’s correlation coefficient between MSV

Mean (standard deviation)
Mean (95% Confidence Interval) assessment
MSV (m/s) 1.82 (0.09)
MSP (W) 99.62 (20.01) r = 0.927** (0.684 to 0.985) very high
MSP/kg (W) 1.55 (0.29) r = 0.807** (0.308 to 0.958) high
SVPR 1.21 (0.23)
Holm–Bonferroni correction were performed on statistically significant p values. *p<0.05, **p<0.01.
MSV: The maximum swimming velocity; MSP: The maximum swimming power; SVPR: The swimming velocity-
to-swimming power ratio.

1223
on the dominant (ETD) (r=0.844; p=0.017) and non-dominant (ETN) sides (r=0.779; p=0.04) (Table 3). No significant
correlations were observed between MSP and IT on the dominant (ITD) side (r=0.72; p=0.057), MSP and the non-dominant
(ITN) side (r=0.624; p=0.073), SVPR and RBET (r=−0.426; 95% CI, −0.85 to 0.332; p=0.253), or SVPR and RBIT (r=0.002;
p=0.995) (Table 3).
In addition, ETD was significantly associated with ETN (r=0.965; p<0.001), and ITD was significantly associated with
ITN (r=0.84; p=0.046) (Table 3). No significant correlations were seen between ETD and ITD (r=0.777; p=0.11), ETN
and ITN (r=0.667; p=0.299), ETD and RBET (r=0.042; p=0.914), ETN and RBET (r= −0.169; p=0.663), ITD and RBIT
(r=0.478; p=0.193), ITN and RBIT (r=0.212; p=0.585), or RBET and RBIT (r= −0.479; p=0.192) (Table 3).

DISCUSSION
This study found significant correlations between MSV and MSP (r=0.927) and MSP and ETD (r=0.844). The sample size
was calculated to be 6 when using an effect size of 0.844, significance level of 5%, and detection power of 0.8; therefore, the
number of participants in this study was valid.
MSP was significantly correlated with ETD (r=0.844) and ETN (r=0.779). In addition, ETD was highly correlated with
ETN (r=0.965). We reported that ITD was most strongly correlated with MSP in arm-only swimming4). These results differ
from those of this study. The stroke phases are classified as the entry and catch phase, the pull phase, the push phase, and the
recovery phase13). In addition, other report has classified the glide phase, the early pull-through phase, the mid-pull-through,
the late pull-through phases, the end of the pull-through phase, and the recovery phase14). The entry and catch phase13) and
the glide phase14) are the non-propulsive phase or the phases without backward movements. However, Maglischo15) noted
that the catch phase is the first propulsive phase. Riewald and Rodeo16) noted that the catch and insweep is the pull phase.
Therefore, the definition of the catch phase is different for each researcher. Colwin17) noted applying pressure on the water
during the same phase as the early pull-through phase or the catch phase. Therefore, the shoulder extension in the position of
maximum shoulder abduction is similar to the early pull-through phase; the hand reaches maximum forward extension and
begins a downward motion14). Furthermore, the shoulder internal rotation in the position of abducted and external rotation is
similar from the early pull-through phase to the mid-pull-through phase.

Table 3. The results of the Pearson’s product-moment correlation coefficient

ETD ETN ITD ITN RBET RBIT

r 0.844** 0.779** 0.72 0.624 - -
MSP 95%CI 0.409 to 0.966 0.238 to 0.951 0.107 to 0.936 −0.069 to 0.911 - -
assessment high high - - - -
r - - - - −0.426 0.002
SVPR 95%CI - - - - −0.85 to 0.332 −0.663 to 0.665
assessment - - - - - -
r 0.965** 0.777 0.788 0.042 -
ETD 95%CI 0.838 to 0.993 0.233 to 0.951 0.26 to 0.953 −0.64 to 0.687 -
assessment very high - - - -
r 0.737 0.667 −0.169 -
ETN 95%CI 0.143 to 0.941 0.005 to 0.922 −0.749 to 0.558 -
assessment - - - -
r 0.84* - 0.478
ITD 95%CI 0.397 to 0.965 - −0.273 to 0.867
assessment high - -
r - 0.212
ITN 95%CI - −0.526 to 0.768
assessment - -
r −0.479
RBET 95%CI −0.867 to 0.271
assessment -
Holm-Bonferroni correction were performed on statistically significant p values. *p<0.05, **p<0.01.
95%CI: 95% Confidence Interval; MSP: The maximum swimming power; SVPR: The swimming velocity-to-swimming power ratio;
ETD: The extension torque of the dominant side; ETN: The extension torque of the non-dominant side; ITD: The internal rotation
torque of the dominant side; ITN: The internal rotation torque of the non-dominant side; RBET: The rate of bilateral difference in exten-
sion torque; RBIT: The rate of bilateral difference in internal rotation torque.

J. Phys. Ther. Sci. Vol. 30, No. 10, 2018 1224

 According to another report8), front crawl swimming has a larger forward amplitude than arm swimming, but the down-
ward and backward amplitudes are small. Additionally, during front crawl swimming, there is a correlation between swim-
ming velocity and downward amplitude8); the velocity is fast if the downward amplitude is large. In other words, front
crawl swimming involves a great forward reach and generates swimming velocity by downward movement. Because front
crawl swimming is faster than arm swimming, a fast stroke rate is required18). In addition, extensor strength in the position
of maximum shoulder abduction, which is similar to the downward movement of the early pull-through phase, was cor-
related with MSP in this study. Generally, the early pull-through phase may be more important for front crawl swimming
than for arm-only swimming. Therefore, measuring extensor strength in the position of maximum shoulder abduction is an
important assessment method for front crawl swimmers. In particular, the dominant side may be one of the most important
measurements. However, Deschodt et al.8) reported that the downward amplitude of arm swimming had no correlation with
swimming velocity. In addition, leg kicking raises the lower extremity and keeps the body horizontal, and the downward
movement of the upper extremity was reported to have an effect on lower extremity sinking19). Internal rotator strength in the
position of abducted and external rotation, which is similar to the position during backward movement phases, was correlated
with swimming velocity. Therefore, because arm swimming cannot compensate for the effect of lower limb sinking due to
downward movements, internal rotator strength in the position of abducted and external rotation may be correlated.
We reported that RBET was highly correlated with SVPR in arm-only swimming (r= −0.728)4). Furthermore, Toussaint
and Truijens20) noted that each stroke cycle has varying propulsion and drag. These fluctuations are considered to be associ-
ated with swimming efficiency, which is determined on the basis of the following indicators: the stroke index, the efficiency
of propulsion generation of the arm stroke, and the coefficient of variation of the hip intra-cyclic velocity variation21). There-
fore, swimming efficiency conserves propulsion generated by each arm. RBET may be associated with swimming efficiency
because its magnitude is negatively associated with SVPR. However, this study did not observe that SVPR was significantly
correlated with RBET; this differed from our past results4). Because front crawl swimming includes the arms and various
factors of the limbs, only shoulder strength was not related to SVPR. Therefore, future studies including propulsion efficiency
and the coefficient of variation for velocity are needed.
These data show a significant correlation between ETD and MSP and validate this tool in front crawl swimmers. Athletes
with a shoulder injury should seek to improve MSP and shoulder muscle strength during training and rehabilitation for front
crawl swimming. In these cases, shoulder muscle strength should be evaluated. This study suggests that muscle strength of
front crawl swimmers can be suitably evaluated using extensor strength measurements in the position of maximum shoulder
abduction during training and rehabilitation. In addition, this measurement method may have high specificity for front crawl
swimmers. Notably, measurements on the dominant side may provide useful data that are essential in training to improve
front crawl swimming propulsion. However, the relationship between the strength of other muscles and MSP remains unclear.
Altogether, these data may help create strength training programs and improve swimming performance in front crawl swim-
mers. A limitation of this study was that the assessments were performed on dry land and not during swimming under water.
Future studies that include motion research or electromyographic analysis are needed, as they were not used in this study.

Conflict of interest
None declared.

ACKNOWLEDGEMENTS

The authors thank the athletes who participated in this study. The authors also thank Editage (www.editage.jp) for Eng-
lish language editing.

REFERENCES

1) Ogita F, Taniguchi S: The comparison of peak oxygen uptake between swim-bench exercise and arm stroke. Eur J Appl Physiol Occup Physiol, 1995, 71:
295–300. [Medline] [CrossRef]
2) Ogita F, Tabata I: Effect of hand paddle aids on oxygen uptake during arm-stroke-only swimming. Eur J Appl Physiol Occup Physiol, 1993, 66: 489–493.
[Medline] [CrossRef]
3) Gergley TJ, McArdle WD, DeJesus P, et al.: Specificity of arm training on aerobic power during swimming and running. Med Sci Sports Exerc, 1984, 16:
349–354. [Medline] [CrossRef]
4) Awatani T, Morikita I, Mori S, et al.: Relationship between isometric shoulder strength and arms-only swimming power among male collegiate swimmers:
study of valid clinical assessment methods. J Phys Ther Sci, 2018, 30: 490–495. [Medline] [CrossRef]
5) Toussaint HM, Vervoorn K: Effects of specific high resistance training in the water on competitive swimmers. Int J Sports Med, 1990, 11: 228–233. [Medline]
[CrossRef]
6) Ria B, Falgairette G, Robert A: Assessment of the mechanical power in the young swimmer. J Swim Res, 1992, 6: 11–15.
7) Mori S, Shimono A, Taguchi M, et al.: Usefulness of swimming power measurement using simplified apparatus. Japanese Journal of Sciences in Swimming
and Water Exercise, 2015, 18: 10–19 (In Japanese). [CrossRef]

1225
8) Deschodt VJ, Arsac LM, Rouard AH: Relative contribution of arms and legs in humans to propulsion in 25-m sprint front-crawl swimming. Eur J Appl Physiol
Occup Physiol, 1999, 80: 192–199. [Medline] [CrossRef]
9) Awatani T, Morikita I, Shinohara J, et al.: Intra- and inter-rater reliability of isometric shoulder extensor and internal rotator strength measurements performed
using a hand-held dynamometer. J Phys Ther Sci, 2016, 28: 3054–3059. [Medline] [CrossRef]
10) Holm S: A simple sequentially rejective multiple test procedure. Scand J Stat, 1979, 6: 65–70.
11) Abdi H: Holm’s sequential Bonferroni procedure. In: Salkind N, Dougherty DM, Frey B (eds.). Encyclopedia of research design. Thousand Oaks: Sage, 2010,
pp 573–577.
12) Hinkle DE, Wiersma W, Jurs SG: Applied statistics for the behavioral sciences, 5th ed. Boston: Houghton Mifflin, 2003, p 109.
13) Chollet D, Chalies S, Chatard JC: A new index of coordination for the crawl: description and usefulness. Int J Sports Med, 2000, 21: 54–59. [Medline] [Cross-
Ref]
14) Heinlein SA, Cosgarea AJ: Biomechanical considerations in the competitive swimmer’s shoulder. Sports Health, 2010, 2: 519–525. [Medline] [CrossRef]
15) Maglischo E: Swimming fastest. Champaign: Human Kinetics, 2003, pp 103–116.
16) Riewald S, Rodeo S: Science of swimming faster. Champaign: Human Kinetics, 2015, pp 23–28.
17) Colwin CM: Breakthrough swimming. Champaign: Human Kinetics, 2002, p 51.
18) Gourgoulis V, Boli A, Aggeloussis N, et al.: The effect of leg kick on sprint front crawl swimming. J Sports Sci, 2014, 32: 278–289. [Medline] [CrossRef]
19) Yanai T: Rotational effect of buoyancy in frontcrawl: does it really cause the legs to sink? J Biomech, 2001, 34: 235–243. [Medline] [CrossRef]
20) Toussaint HM, Truijens MJ: Biomechanical aspects of peak performance in human swimming. An Biol, 2005, 55: 17–40. [CrossRef]
21) Seifert L, Toussaint HM, Alberty M, et al.: Arm coordination, power, and swim efficiency in national and regional front crawl swimmers. Hum Mov Sci, 2010,
29: 426–439. [Medline] [CrossRef]

J. Phys. Ther. Sci. Vol. 30, No. 10, 2018 1226