Adjustment of Measures of Strength and Power in Youth Male Athletes Differing in Body Mass and Maturation

Abstract

Adjustment for body mass and maturation of strength, power and velocity measures of young athletes is important for talent development. Seventy-four youth male athletes performed a ballistic leg press test at five loads relative to body mass. The data were analyzed in maturity groups based on years from peak height velocity: -2.5 to -0.9 y (n = 29); -1.0 to 0.4 y (n = 28); and 0.5 to 2.0 y (n = 16). Allometric scaling factors representing percent difference in performance per percent difference in body mass were derived by linear regression of log-transformed variables, which also permitted adjustment of performance for body mass. Standardized differences between groups were assessed via magnitude-based inference. Strength and power measures showed a greater dependency on body mass than velocity-related variables (scaling factors of 0.56 to 0.85 vs 0.42 to 0.14 %/%), but even after adjustment for body mass most differences in strength and power were substantial (7% to 44%). In conclusion, increases in strength and power with maturation are due only partly to increases in body mass. Such increases, along with appropriate adjustment for body mass, need to be taken into account when comparing performance of maturing athletes.

Full text

41
Official Journal of NASPEM and the
European Group of PWP
www.PES-Journal.com
ORIGINAL RESEARCH
Pediatric Exercise Science, 2014, 26, 41-48
http://dx.doi.org/10.1123/pes.2013-0029
© 2014 Human Kinetics, Inc.
Meylan was with the Sport Performance Research Institute,
Auckland University of Technology, Auckland, New Zealand,
at the time of this research and is currently with the Canadian
Sport Institute Pacic, Vancouver, Canada. Cronin and Hop-
kins are with the School of Sport and Recreation, Auckland
University of Technology, Auckland, New Zealand. Oliver is
with the Cardiff School of Sport, University of Wales Institute
Cardiff, Cardiff, UK.
Adjustment of Measures of Strength and Power in Youth
Male Athletes Differing in Body Mass and Maturation
Cesar Marius Meylan, John Cronin, and Will G. Hopkins
Auckland University of Technology
Jonathan Oliver
University of Wales Institute Cardiff
Adjustment for body mass and maturation of strength, power, and velocity measures of young athletes is
important for talent development. Seventy-four youth male athletes performed a ballistic leg press test at
ve loads relative to body mass. The data were analyzed in maturity groups based on years from peak height
velocity: –2.5 to –0.9 y (n = 29); –1.0 to 0.4 y (n = 28); and 0.5 to 2.0 y (n = 16). Allometric scaling factors
representing percent difference in performance per percent difference in body mass were derived by linear
regression of log-transformed variables, which also permitted adjustment of performance for body mass.
Standardized differences between groups were assessed via magnitude-based inference. Strength and power
measures showed a greater dependency on body mass than velocity-related variables (scaling factors of
0.56–0.85 vs. 0.42–0.14%/%), but even after adjustment for body mass most differences in strength and power
were substantial (7–44%). In conclusion, increases in strength and power with maturation are due only partly
to increases in body mass. Such increases, along with appropriate adjustment for body mass, need to be taken
into account when comparing performance of maturing athletes.
Keywords: adolescent, pediatrics, resistance training, strength
In many team sports, muscle power is regarded as a
dening physical attribute of elite players that needs to be
trained progressively and monitored from an early stage
of a player’s development. Power output is the product
of force and velocity and is dened and limited by the
force-velocity relationship (16). On this basis, maximal
power output may improve by an increased ability to
develop force at a given velocity and/or velocity at a
given force (4,5). Several cross-sectional (9,10,27,33) and
longitudinal (28) studies using cycling ergometers have
investigated the role of growth and maturation and asso-
ciated quantitative changes (e.g., in body mass, lean leg
volume) and qualitative changes (e.g., in intermuscular
coordination, motor unit recruitment) in muscle proper-
ties on the force-velocity-power relationship. However,
the applicability of the ndings to activities incorporat-
ing running and jumping is problematic, since cycling
requires limited use of the posterior chain hip extensors
and is not a weight-bearing exercise (36).
The vertical jump and its derivatives are some of the
most widely used movements to assess the power of the
leg musculature because of their simplicity. The jumps
can also be considered as some of the most “explosive”
tests, owing to both the short duration and the high
intensity of the movement (4). Researchers (2,31,32,34)
have investigated the force-velocity-power relationship
using loaded jump squat protocols to quantify the effect
of strength, competition level and training programs
on this relationship. The force-velocity-power prole
during ballistic jump movements has been shown to
differentiate stronger from weaker athletes (2), level of
play (34) and individual specic force-velocity relation-
ships within a group of athletes (32). Such an approach
can also provide insight into the mechanistic changes
responsible for power increase during growth and matu-
ration (35). An isoinertial loading protocol has also been
used concurrently for maximal strength prediction and
force-velocity-power proling (21). The benets of such
loading protocols are many, but no studies, to the authors’
knowledge, have investigated the role of maturity status
on the isoinertial force-velocity-power relationship,

42 Meylan et al.
which may be more relevant to the on-eld requirements
of the youth athlete.
The force-velocity-power profile is affected by
certain properties of skeletal muscles that change during
growth and maturation. For example, increased muscle
cross-sectional area with age is likely to inuence the
force component of power, while greater sarcomere
length may enhance velocity capabilities (35). Other
factors such as motor-unit recruitment may affect both
aspects of the force-velocity relationship responsible
for power output (11). When comparing young athletes,
controlling for body mass may provide an insight into the
mechanisms responsible for the changes in force-veloc-
ity-power relationship during growth and maturation.
Strength and power variables are commonly expressed
using ratio scaling (e.g., power per unit of body mass)
but such scaling often fails to produce a size-free index.
Allometric scaling (e.g., power per unit of mass raised to
some exponent) is commonly accepted as a better method
to scale for body mass (37). Specic theoretical scaling
models for force, power, or speed have been suggested
(20), but to accurately account for body mass, exponents
specic to the performance test and the athlete group
should be applied (37). Therefore, the purpose of this
study was to use allometric scaling to investigate strength
and power relationships in a ballistic loading test with a
group of maturing male athletes.
Methods
Participants
Seventy-four males between 11 and 15 years of age vol-
unteered for this study. All participants were nominated
by their physical education teacher to be part of a school
sports academy. Participant characteristics are presented
in Table 1. The Human Research Ethics Committee of
Auckland University of Technology approved the study
and both the participants and their parents/guardians gave
their written consent/assent before the start of the study.
Testing Procedures
Participants attended one designated testing session
preceded by a familiarization session of all testing
procedures. Anthropometric measurements were taken
before performance testing. The standing height (cm),
sitting height (cm) and weight (kg) were measured and
the body mass index (BMI) calculated. The maturity
status of the athletes determined using years from peak
height velocity (PHV offset; 29) as well as the percent-
age of predicted adult stature (23). After determination
of maturity status, athletes were split into three maturity
groups for analysis (see Table 1).
Participants then undertook a 15-min standardized
warm-up using the different loads employed in the testing.
Performance testing consisted of three trials of ballistic
concentric squats on a supine squat machine (Fitness
Works, Auckland, New Zealand) at ve different relative
loads to body mass (%) in a randomized order: 80%,
100%, 120%, 140% and 160%. Before each load, par-
ticipants were asked to fully extend their leg to determine
the zero position, which was used to determine the end
of the pushing phase. A recovery of 30 s between trials
within load and 120 s between loads was given. The foot
position and knee angle (70°) were controlled for each
trial (6). The supine squat machine was designed to allow
novice participants to perform maximal squats or explo-
sive squat jumps, with the back rigidly supported, thus
minimizing the risk associated with such exercises in an
upright position (e.g., excessive landing forces, lumbar
spine exion and extension; Figure 1).
A linear position transducer (Celesco, Model
PT9510–0150–112–1310, USA) attached to the weight
stack measured vertical displacement relative to the
ground with an accuracy of 0.1 cm. These data were
sampled at 1000 Hz by a computer based data acquisition
and analysis program. The displacement-time data were
ltered using a low-pass fourth-order Butterworth lter
with a cut-off frequency of 50 Hz, to obtain position.
The ltered position data were then differentiated using
Table 1 Participant Characteristics (Mean ± SD) of the Maturity
Groups Based on Peak Height Velocity (PHV)
Variables Pre PHV
(
n
= 29)
Mid PHV
(
n
= 28)
Post PHV
(
n
= 16)
Age (y) 12.1 ± 0.7 13.4 ± 0.6 14.4 ± 0.4
PHV offset (y) –1.7 ± 0.5 –0.2 ± 0.5 1.0 ± 0.4
Height (cm) 152 ± 6 166 ± 8 173 ± 4
Relative height (%)
a
85.4 ± 2.4 91.7 ± 2.2 96.2 ± 1.7
Mass (kg) 40.7 ± 4.7 54.6 ± 8.7 63.1 ± 9.6
Leg length (cm) 74.2 ± 3.8 80.1 ± 5.2 82.2 ± 2.6
Body mass index (kg·m
–2
) 17.4 ± 1.6 19.8 ± 2.6 21.1 ± 2.9
Note. All differences between groups were clear.
a
Height as a percent of predicted adult height.

Assessing Strength and Power in Youth 43
the nite-difference technique to determine velocity (v)
and acceleration (a) data, which were each successively,
ltered using a low-pass fourth-order Butterworth Filter
with a cut-off frequency of 6 Hz (13). The force (F)
produced during the thrust was determined by adding
the weight of the weight stack to the force required to
accelerate the system mass, which consisted of the mass
of the weight stack (mWS), the mass of the participant
(mP), and the mass of the sled (mS), so F = g(mWS) +
a(mWS + mP + mS), where g is the acceleration due to
gravity and a is the acceleration generated by the move-
ment of the participant. Following these calculations,
power was determined by multiplying the force by veloc-
ity at each time point. Mean force, velocity, and power
were determined from the means of the instantaneous
values over the entire push-off phase (until full leg exten-
sion, i.e., position 0). The external validity of the derived
measurements from a linear position transducer have been
assessed using the force plate as a “gold standard” device
(r = .81–0.96; 7,14,19), with the only major limitation of
underestimating force and power output (19).
Data Analysis
Concentric leg-press squat one-repetition maximum
(1RM) was estimated via the load-velocity relationship
(21). The 1RM velocity was not calculated in the current
study and this value (0.23 m·s
–1
) was extracted from
previous studies in adults (13,21) to be plotted on the
load-velocity curve to extract 1RM. A pilot study on 10
children involved in the current study found a Pearson
correlation of 0.94 (90% condence limits 0.80–0.98)
between the actual 1RM (118.5 ± 27.3 kg) and predicted
1RM (112.1 ± 23.0 kg) using a 1RM velocity of 0.23
m.s
-1
. Force-velocity (F-v) relationships were determined
by least-squares linear regressions using mean force and
velocity at each load. Individual force-velocity slopes
were extrapolated to obtain Fmax and Vmax, which cor-
responded to the intercepts of the F-v slope with the force
and velocity axes respectively (31,32). Since the power-
load relationship is derived from the product of force and
velocity, it was described by second-degree polynomial
functions and maximal power output (Pmax) and the opti-
mal load at which Pmax occurred was determined using
the power-load regression curve (13). The goodness-of-t
of the individuals’ quadratics was expressed as a correla-
tion coefcient calculated by taking the square root of
the fraction of the variance explained by the model, after
adjusting for degrees of freedom; the values were then
averaged. This method has been validated against vertical
jump height in a previous study (r = .67; 38) as well as
against vertical peak power produced in a countermove-
ment jump (0.89; 90% condence limits 0.83–0.94) and
10-m sprint time (-0.79; -0.61 to -0.81) in the current
population sample (unpublished observations).
Statistical Analysis
Data in the text and gures are presented as means ±
SD (SD). Initially, pairwise comparisons of performance
variables between groups were conducted without
taking body mass into account using with a custom-
ized published spreadsheet (17). Differences in means
between groups were expressed in percent units derived
via log transformation. Magnitudes of differences were
assessed by standardization of the log-transformed
performance measure: dividing the difference in means
by an SD. The appropriate SD was the square root of
the mean of the variances of performance in the two
groups of interest. The effect of body mass on each
performance variable was then investigated using an
allometric scaling model y = a.×
b,
where y is the per-
formance variable, x is body mass, a is a constant and
b is the allometric scaling factor. The model was linear-
ized by taking natural logarithms of both sides: ln(y) =
ln(a) + b.ln(x), allowing estimation of b as a slope in a
Figure 1 — Experimental set up on the supine squat machine. The linear position transducer was attached the weight stack to
provide displacement data output as the participant moved horizontally.

44 Meylan et al.
linear regression. Initially, separate linear regressions
were performed for each maturity group. The difference
in slope between groups was evaluated by expressing
each slope as the effect of two SD of ln (body mass),
then comparing these effects between pairs of maturity
groups by standardization. The appropriate SD was now
the standard errors of the estimate (SEE) from each
regression (because the SEE represents typical differ-
ences between subjects after adjustment of performance
to the same body mass). As the differences in scaling
factors between groups for a given performance measure
were mostly trivial but unclear, a multiple regression
model was devised to provide a single scaling factor and
a single SEE for each measure using the Linest function
in Excel. The model consisted of ln(body mass) as a
simple numeric predictor, an intercept representing the
allometric constant ln(a) for a reference PHV group, and
two dummy variables each coded as 0 or 1 to represent
data coming from each of the other two PHV groups.
Several such analyses were performed to obtain the
pairwise comparisons of the ln(a), representing mean
difference between PHV groups after adjustment for
body mass. The standard error provided by Linest for
the coefcient of the dummy variable was used to cal-
culate 90% condence limits for the group differences.
Magnitude of differences in mean performance between
maturity groups was evaluated via standardization
within the allometric analysis using the SEE from the
multiple regression model.
Threshold values for assessing magnitudes of stan-
dardized effects were 0.20, 0.60, 1.2 and 2.0 for small,
moderate, large and very large respectively (18). Uncer-
tainty in each effect was expressed as 90% condence
limits and as probabilities that the true effect was substan-
tially positive and negative. These probabilities were used
to make a qualitative probabilistic mechanistic inference
about the true effect (18): if the probability of the effect
being either substantially positive or substantially nega-
tive was <5% (very unlikely), the effect was deemed clear
and reported as the magnitude of the observed value,
with the qualitative probability that the true value was
at least of this magnitude. The scale for interpreting the
probabilities was as follows: 25–74%, possible; 75–94%,
likely; 95–99.5%, very likely; >99.5%, most likely (18).
The effect was otherwise deemed unclear, because the
span of its 90% condence interval (CL) was consistent
with a true effect that could be substantially positive and
negative. Use of a 90% CL allows for decisive outcomes
with sample sizes that are one-third those for outcomes
based on null-hypothesis testing with 80% power for 5%
signicance (18).
Results
Subject characteristics for the three maturity groups are
presented in Table 1. The differences between groups for
age and maturity (PHV offset and predicted adult height)
were large to extremely large, while the differences in
height, mass and leg length were at least small.
Before adjustment for body mass, the differences
in 1RM, Pmax and Fmax between Pre PHV and other
maturity groups ranged from large to very large, while the
differences between Mid and Post PHV were moderate
to large (Table 2). The difference in Vmax was unclear
between Pre and Mid PHV but large between these two
groups and Post PHV, which in turn inuenced differ-
ences in the F-v prole: the Pre PHV group was more
velocity dominant and the Mid PHV group more force
dominant compared with the Post PHV group. Optimal
load for Pmax derived from the power-load quadratic
curve (goodness of t: R
2
= .79; SEE = 20.7 W) expressed
as percent of body mass for Pre, Mid and Post PHV
groups were respectively 93 ± 14, 93 ± 18 and 90 ± 14
(mean ± SD).
Figure 2 shows an example of the allometric rela-
tionship between a performance measure (Pmax) and
body mass with each maturity group having a separate
slope. Table 3 shows the results of allometric analyses
for each performance variable with a single slope tted
to the three maturity groups. After adjustment for body
mass, Mid-PHV athletes had mostly moderate differences
from Pre-PHV athletes. The differences in performance
between Mid PHV and Post PHV were moderate to large
in velocity-dependent variables (Pmax, Fmax/Vmax
slope, Vmax) but remained small to unclear in force-
dependent variables (1RM, Fmax). Finally, differences
between Pre- and Post-PHV groups was moderate to large
for all variables except for the Fmax/Vmax slope (Figure
3), where the difference was unclear.
Discussion
The large differences in maturity status and age between
the groups were reected in somatic growth differences.
PHV is usually preceded by an accelerated increase in
leg length and followed by an increase in trunk velocity
growth and peak weight velocity (0.5–1 y post PHV) (25).
Further, children passing through the age of peak weight
velocity experience a change in the height-to-mass ratio
represented by an increase in BMI (25). The current study
demonstrated a regular increase in BMI with maturity as
well as a reduced difference in leg length once PHV was
passed. These ndings provide condence that the groups
were representative of the normal growth and maturation
associated with human development.
The large increase in strength and power from the
onset of PHV found in the current study could be attrib-
uted partly to the increase in body mass and associated
change in muscle cross-sectional area during growth
and its direct relationship with force (22). To determine
the role of maturation on strength, power, and velocity
capabilities independently of body mass, the dependent
variables were allometrically scaled for body mass. A
trivial and unclear difference in scaling factors for a given
variable between the groups was found and a single scal-
ing factor was calculated to compare athletes of different
maturity status. The body mass scaling factor for 1RM,
Fmax and Pmax (i.e., the b exponent) varied between

Assessing Strength and Power in Youth 45
Table 2 Force, Power, and Velocity Characteristics (Mean ± SD) of the Maturity Groups Based on Peak
Height Velocity (PHV)
Difference Between Groups (%) with 90% Confidence Limits
Variables
Pre PHV
(
n
= 29)
Mid PHV
(
n
= 28)
Post PHV
(
n
= 16) Mid-Pre PHV Post-Mid PHV Post-Pre PHV
1RM (kg) 77 ± 12 107 ± 20 126 ± 18 39 (28, 50) very
likely large
18 (9, 29) likely
moderate
64 (52, 77) most
likely very large
Pmax (W) 275 ± 65 400 ± 101 567 ± 102 44 (30, 61) likely
large
44 (21, 61) likely
large
108 (88, 131) most
likely very large
Fmax (N) 770 ± 120 1090 ± 200 1220 ± 170 40 (29, 51) very
likely large
13 (3, 23) possibly
moderate
57 (45, 70) most
likely very large
Vmax
(m·s
–1
)
1.42 ± 0.29 1.43 ± 0.23 1.98 ± 0.43 1 (–7, 10) unclear 37 (22, 53) likely
large
38 (23, 55) likely
large
Fmax/Vmax –570 ± 130 –780 ± 180 –660 ± 230 –38 (–54, 24)
possibly large
18 (4, 29) possibly
moderate
–14 (–33, 3) likely
small
Note. 1RM = estimated one repetition maximum weight, Pmax = maximal power along the load spectrum, Fmax = estimated maximal force from force-velocity
relationship, Vmax = maximal velocity from force-velocity relationship, Fmax/Vmax = ratio between Fmax and Vmax.
Table 3 Differences in Mean Performance Between Maturity Groups After Adjustment for Body Mass in an
Allometric Analysis with a Single Scaling Factor in the Three Groups
Performance SEE (%)
Difference (%) in Performance Adjusted for Body Mass
Scaling factor
(%/%) Mid-Pre PHV Post-Mid PHV Post-Pre PHV
1RM 0.69 (0.49, 0.89) 14 (12, 16) 13 (4,24) likely moderate 7 (–1, 16) likely small 21 (8,31) possibly large
Pmax 0.85 (0.57, 1.14) 20 (18, 24) 13 (0, 28) likely small 27 (14, 43) possibly
large
44 (23, 69) likely large
Fmax 0.56 (0.35, 0.77) 15 (13, 17) 19 (8,30) likely moderate 4 (–5, 13) unclear 23 (9, 39) possibly
large
Vmax 0.42 (0.16, 0.68) 18 (16, 21) –11 (–20, 0) possibly
moderate
29 (16, 43) possibly
large
15 (–1, 34) possibly
moderate
Fmax/Vmax 0.14 (-0.23, 0.52) 26 (23, 31) –33 (–55, -13) likely
moderate
19 (30,7) possibly
moderate
–7 (–32, 14) unclear
Note. The standard error of the estimate (SEE) is the within-group between-athlete standard deviation in performance used to assess the magnitude of the
differences. All data in parentheses are 90% confidence limits. PHV = peak height velocity, 1RM = estimated one repetition maximal, Pmax = maximal power
along the load spectrum, Fmax = estimated maximal force from force-velocity relationship, Vmax = maximal velocity from force-velocity relationship, Fmax/
Vmax = ratio between Fmax and Vmax.
0.56 and 0.85. These mean results were dissimilar to the
standard ratio usually used (b = 1) and only the exponent
for 1RM (b = 0.69) fell near the theoretical value of 0.67
proposed by Jaric et al. (20). The value of 0.85 for Pmax
was similar to that in previous research investigating
allometric scaling of power output for body mass during
countermovement jump (b = 0.90; 26), cycling in adults
(b = 0.92; 15) and children (b = 0.84–0.97; 33) while the
theoretical value of 0.67 still fell within the 90% CL of
the analysis. The CL of the Vmax scaling factor across
the three groups did not reach the zero value suggested
elsewhere (20; Table 3). Further, previous studies (32)
have used the ratio standard (b = 1) to scale the F-v
prole while this value approached zero in the current
study. In summary, theoretical allometric scaling can be
used with reasonable condence in strength and power
measures while velocity and F-v prole scaling still need
further investigation.
Although there were clear differences in perfor-
mance between groups after adjustment for maturity, the
width of the condence intervals for the differences and
for the associated scaling factors represent considerable
uncertainty. More data should be acquired to reduce the
uncertainty before the differences and the scaling factor
for body mass are used in practical settings. Adjustment
of an athlete’s performance score to a given body mass

46 Meylan et al.
within a maturity group (e.g., to the mean) would then
be achieved by multiplying the score by (mean mass/
athlete’s mass)
b
. To compare this athlete with those in
another maturity group with, for example, 13% higher
Figure 3 — Force-velocity relationship across the load spec-
trum for Pre (dashed line), Mid (dotted line) and Post (solid line)
peak height velocity (PHV) groups. The intercepts on the y and
x axes represent maximal force and maximal velocity capabili-
ties, respectively, estimated from the force-velocity relationship
across the ve loads. The arrow indicates where maximal power
occurred on the load spectrum for each maturity group.
Figure 2 — Maximal power (Pmax) and body mass relation-
ship of the different maturity groups based on peak height
velocity (PHV). The lines shown are the least-squares regression
lines provided by separate allometric analyses for each group.
Axes are logarithmic to show the modeled relationships as lines
rather than curves.
scores (Mid-Pre for 1RM in Table 3, the value specic
to this athlete population), the mean mass in this formula
would be that of the older group, and the resulting score
would be multiplied by 1.13. This formula could be
incorporated into a spreadsheet to allow practical assess-
ment of athletes in the original units of the performance
test. This approach would enhance talent identication
processes and reduce the risk of selecting bigger and
more mature athletes on the basis of greater absolute
scores on a given test.
The remaining difference in performance between
the maturity groups after adjustment for body mass
was most likely associated with maturation-related
changes in qualitative neuromuscular factors. Increases
in strength and power could be due to greater percentage
of motor unit activation, development of Type II bers,
increased fascicle length and improved motor coordina-
tion (11,22,35). Selective hypertrophy of Type II bers
(12) may be inuenced by the increase in testosterone
(24), which begins approximately one year before PHV
(8). The surge in testosterone could partly explain the
moderate divergence of relative strength between boys
at Pre PHV and Mid PHV observed in the current study
(Table 3). In the current study, quantitative and qualitative
changes of the neuromuscular system during growth and
maturation appeared to explain the increase in muscle
strength, but qualitative changes seem to play a greater
role in explaining the differences from Pre to Mid PHV.
The relationship between strength and power dictates
that an individual cannot possess a high level of power
without rst being relatively strong. This assertion is
supported by the robust relationship that exists between
maximal strength and maximal power production (4). The
concurrent large changes in relative strength (1RM and
Fmax) and power (Pmax) from Pre to Post PHV support
this contention. However, the difference in Pmax between
Mid PHV and Post PHV was not associated with a change
in strength of similar magnitude (Table 3). The velocity
dominant F-v prole of the Post PHV group compared
with the Mid PHV group (Figure 3) could explain why
their similar relative strength did not lead to a similar
Pmax. As the optimal load for Pmax was early in the
load spectrum for all groups, it seems that the ability
to produce force at fast velocity was a more important
determinant of Pmax and advantageous to the velocity
dominant prole of Post PHV. It can be concluded that the
shift in the F-v relationship at Mid PHV was associated
with a reduced ability to produce high velocity contrac-
tions and optimally use maximal strength for Pmax.
During the growth spurt around PHV, a disturbance of
motor coordination explained by the differential timing
of growth in both leg and trunk length has been observed
and referred to as “adolescent awkwardness” (1,30).
A reduced ability of motor unit synchronization and
intermuscular coordination during fast movement would
compromise both Vmax and Pmax (3) and thereby could
explain the reduced gain in both variables and shift in the
F-v prole for the boys at Mid PHV.

Assessing Strength and Power in Youth 47
4. Cormie P, Mcguigan MR, Newton RU. Developing maxi-
mal neuromuscular power: part 2 - training considerations
for improving maximal power production. Sports Med.
2011; 41(2):125–146. PubMed doi:10.2165/11538500-
000000000-00000
5. Cronin J, Mcnair PJ, Marshall RN. Velocity specic-
ity, combination training and sport specic tasks. J Sci
Med Sport. 2001; 4(2):168–178. PubMed doi:10.1016/
S1440-2440(01)80027-X
6. Cronin JB, Henderson ME. Maximal strength and power
assessment in novice weight trainers. J Strength Cond Res.
2004; 18(1):48–52. PubMed
7. Cronin JB, Hing RD, Mcnair PJ. Reliability and validity
of a linear position transducer for measuring jump per-
formance. J Strength Cond Res. 2004; 18(3):590–593.
PubMed
8. Croix MD. Advances in paediatric strength assessment:
changing our perspective on strength development. J
Sports Sci Med. 2007; 6(3):292–304. PubMed
9. Dore E, Bedu M, Franca NM, Diallo O, Duche P, Van
Praagh E. Testing preak cycling performance: effects
of braking force during growth. Med Sci Sports Exerc.
2000; 32(2):493–498. PubMed doi:10.1097/00005768-
200002000-00035
10. Dore E, Diallo O, Franca NM, Bedu M, Van Praagh E.
Dimensional changes cannot account for all differences in
short-term cycling power during growth. Int J Sports Med.
2000; 21(5):360–365. PubMed doi:10.1055/s-2000-3783
11. Dotan R, Mitchell C, Cohen R, Klentrou P, Gabriel D, Falk
B. Child-adult differences in muscle activation–a review.
Pediatr Exerc Sci. 2012; 24(1):2–21. PubMed
12. Glenmark B, Hedberg G, Kaijser L, Jansson E. Muscle
strength from adolescence to adulthood–relationship to
muscle bre types. Eur J Appl Physiol Occup Physiol.
1994; 68(1):9–19. PubMed doi:10.1007/BF00599235
13. Harris NK, Cronin JB, Hopkins WG. Power outputs of a
machine squat-jump across a spectrum of loads. J Strength
Cond Res. 2007; 21(4):1260–1264. PubMed
14. Harris NK, Cronin J, Taylor K-L, Jidovtseff B. Under-
standing position transducer technology for strength
and conditioning practitioners. Strength Cond J. 2010;
32(4):66–79. doi:10.1519/SSC.0b013e3181eb341b
15. Hazir T, Kosar N. Assessment of gender differences in
maximal anaerobic power by ratio scaling and allometric
scaling. Iso Exerc Sci. 2007; 15(4):253–261.
16. Hill AV. The heat of shortening and the dynamic constant
of muscle. In: Procceeding of the Royal Society of London.
London, 1938.
17. Hopkins WG. A spreadsheet to compare means of two
groups. Sportscience: A peer-reviewed journal and site
for sport research [Internet]. 11, 2007. Available from:
http://sportsci.org/
18. Hopkins WG, Marshall SW, Batterham AM, Hanin J.
Progressive statistics for studies in sports medicine and
exercise science. Med Sci Sports Exerc. 2009; 41(1):3–13.
PubMed doi:10.1249/MSS.0b013e31818cb278
19. Hori N, Newton RU, Andrews WA, Kawamori N,
Mcguigan MR, Nosaka K. Comparison of four different
Conclusion
A number of original ndings in this study have impor-
tant implications to assessment and training practice for
youth populations. Previous specic theoretical scaling
models of body mass for force, power, or speed have
assumed geometric similarity across all youth popula-
tion. Such approach can be used approximately with
force, strength, and power measures but would seem
problematic in velocity measures, given the results of
this study. The scaling factors need to be determined
over a large sample size before they can be used con-
dently in practical settings or otherwise need to be
calculated for every new sample. Practitioners should
also not compare athletes of different maturity status
with the assumption that adjustment for body mass
accounts for all maturational effects on strength, power,
and velocity capabilities, because qualitative factors
were also responsible for the difference in performance
between groups.
The development of power was associated not only
with a strength increase during maturation but also with
a change in velocity capability, as expressed by the F-v
relationship. Around PHV, there was a reduced ability
to use the same relative percentage of maximal force at
high velocity compared with the other two groups, which
affected the development of power. From these ndings it
may be inferred that maturity-specic training programs
should be considered. For example, to increase power and
reduce the negative shift of the F-v relationship, training
at the onset of PHV should concentrate on fast velocity
movement and high rates of force development with
movements that require a considerable level of muscle
coordination. Future longitudinal studies should investi-
gate the maturity effect on the dose-response of training
methods focusing on force versus velocity.
Finally, it needs to be acknowledged that athletic
performance such as jumping, sprinting or change of
direction are ultimately dened by the net impulse gener-
ated into the ground, often in a very short amount of time.
When assessing the determinants of athletic performance
in youth, future research may benet from a force-time
analysis in addition to or instead of the force-velocity
approach adopted in this paper.
References
1. Beunen G, Malina RM. Growth and physical performance
relative to the timing of the adolescent spurt. Exerc Sport
Sci Rev. 1988; 16(1):503–540. PubMed
2. Cormie P, Mcguigan MR, Newton RU. Influence of
strength on magnitude and mechanisms of adaptation to
power training. Med Sci Sports Exerc. 2010; 42(8):1566–
1581. PubMed doi:10.1249/MSS.0b013e3181cf818d
3. Cormie P, Mcguigan MR, Newton RU. Developing maxi-
mal neuromuscular power: Part 1–biological basis of maxi-
mal power production. Sports Med. 2011; 41(1):17–38.
PubMed doi:10.2165/11537690-000000000-00000

48 Meylan et al.
methods to measure power output during the hang power
clean and the weighted jump squat. J Strength Cond Res.
2007; 21(2):314–320. PubMed
20. Jaric S, Mirkov D, Markovic G. Normalizing physical
performance tests for body size: a proposal for standardiza-
tion. J Strength Cond Res. 2005; 19(2):467–474. PubMed
21. Jidovtseff B, Harris NK, Crielaard J, Cronin JB. Using the
load-velocity relationship for 1RM prediction. J Strength
Cond Res. 2011; 25(1):267–270. PubMed doi:10.1519/
JSC.0b013e3181b62c5f
22. Jones DA, Round JM. Muscle development during child-
hood and adolescence. In: The young athlete, H Hebestreit
and O Bar-Or (Eds.). Malden: Blackwell Publishing Ltd,
2008, pp. 18–26.
23. Khamis HJ, Roche AF. Predicting adult stature without
using skeletal age: The Khamis-Roche method. Pediatrics.
1994; 94(4 Pt 1):504–507. PubMed
24. Krotkiewski M, Kral JG, Karlsson J. Effects of castration
and testosterone substitution on body composition and
muscle metabolism in rats. Acta Physiol Scand. 1980;
109(3):233–237. PubMed doi:10.1111/j.1748-1716.1980.
tb06592.x
25. Malina R, Bouchard C, Bar-Or O. Growth, Maturation,
and Physical Activity. Champaign: Human Kinetics, 2004.
26. Markovic G, Jaric S. Scaling of muscle power to body
size: The effect of stretch-shortening cycle. Eur J Appl
Physiol. 2005; 95(1):11–19. PubMed doi:10.1007/s00421-
005-1385-5
27. Martin RJ, Dore E, Hautier CA, Van Praagh E, Bedu
M. Short-term peak power changes in adolescents of
similar anthropometric characteristics. Med Sci Sports
Exerc. 2003; 35(8):1436–1440. PubMed doi:10.1249/01.
MSS.0000079074.47756.AB
28. Martin RJ, Dore E, Twisk J, Van Praagh E, Hautier CA,
Bedu M. Longitudinal changes of maximal short-term peak
power in girls and boys during growth. Med Sci Sports
Exerc. 2004; 36(3):498–503. PubMed doi:10.1249/01.
MSS.0000117162.20314.6B
29. Mirwald RL, Baxter-Jones AD, Bailey DA, Beunen GP.
An assessment of maturity from anthropometric mea-
surements. Med Sci Sports Exerc. 2002; 34(4):689–694.
PubMed doi:10.1097/00005768-200204000-00020
30. Philippaerts RM, Vaeyens R, Janssens M, Van Renterghem
B, Matthys D. R. Craenet al. The relationship between
peak height velocity and physical performance in youth
soccer players. J Sports Sci. 2006; 24(3):221–230. PubMed
doi:10.1080/02640410500189371
31. Rahmani A, Viale F, Dalleau G, Lacour JR. Force/velocity
and power/velocity relationships in squat exercise. Eur J
Appl Physiol. 2001; 84(3):227–232. PubMed doi:10.1007/
PL00007956
32. Samozino P, Rejc E, Di Prampero PE, Belli A,
Morin JB. Optimal force-velocity profile in ballistic
movements–altius: citius or fortius? Med Sci Sports
Exerc. 2012; 44(2):313–322. PubMed doi:10.1249/
MSS.0b013e31822d757a
33. Santos A, Welsman JR, De Ste Croix M, Armstrong N.
Age- and sex-related differences in optimal peak power.
Pediatr Exerc Sci. 2002; 14(2):202–212.
34. Sheppard JM, Cormack S, Taylor KL, Mcguigan MR,
Newton RU. Assessing the force-velocity characteristics of
the leg extensors in well-trained athletes: the incremental
load power prole. J Strength Cond Res. 2008; 22(4):1320–
1326. PubMed doi:10.1519/JSC.0b013e31816d671b
35. Van Praagh E, Dore E. Short-term muscle power during
growth and maturation. Sports Med. 2002; 32(11):701–
728. PubMed doi:10.2165/00007256-200232110-00003
36. Van Praagh E. Testing anaerobic performance. In: The
young athlete, H Hebestreit and O Bar-Or (Eds.). Malden:
Blackwell Publishing Ltd, 2008, pp. 453–468.
37. Welsman JR, Armstrong N. Scaling for size: Relevance to
understanding effects of growth on performance. In: The
young athlete, H Hebestreit and O Bar-Or (Eds.). Malden:
Blackwell Publishing Ltd, 2008, pp. 50–62.
38. Yamauchi J, Ishii N. Relations between force-velocity
characteristics of the knee-hip movement and vertical jump
performance. J Strength Cond Res. 2007; 21(3):703–709.
PubMed