Original Article doi:10.1111/j.1365-2214.2007.00767.

Pictorial and verbal category-ratio scales for effort

estimation in children
B. Marinov, S. Mandadjieva and S. Kostianev
Department of Pathophysiology, Medical University of Plovdiv, Plovdiv, Bulgaria

Accepted for publication 3 April 2007

Keywords
Borg scale, children, effort
perception,
Pictorial-CERT,
repeatability
Correspondence:
Dr Blagoi Marinov,
Pathophysiology
Department, Medical
University of Plovdiv, 15 A
Vassil Aprilov Blvd., 4002
Plovdiv, Bulgaria
E-mail:
b_marinov@mail.orbitel.bg
Abstract
Background Research on the diverse aspects of exercise performance in childhood in the past
20 years has included an increase in the study of perceived exertion.
Objective The aim of this study was to compare children’s ratings of effort perception by means of
the Borg Category-Ratio Perceived Exertion (CR-10) Scale and a pictorial version of the Children’s
Effort Rating Table (Pictorial-CERT) scale, and to assess the long-term repeatability of the two scales.
Methods Fifty healthy children (25 girls and 25 boys; initially aged 10.4 ⫾ 0.5 years) participated in
three incremental treadmill tests until volitional exhaustion or a maximal gradient of 22% at
5.4 km/h was attained. The first two tests (T1 and T2) were at an interval of 1 month. The third test
(T3) took place 3 years later and utilized exactly the same protocol.
Results Perceived exertion correlated significantly with measures of exercise intensity – minute
ventilation, heart rate and oxygen uptake for the whole group. The range of correlations for all tests
was significantly higher for the Pictorial-CERT (r = 0.62–0.88 and r = 0.59–0.71 for the Pictorial-CERT
and CR-10 respectively). Intraclass correlation coefficients between T1 and T2 were significantly
higher for the Pictorial-CERT in comparison with the CR-10 (0.77 vs. 0.54, respectively; z = -2.07;
P = 0.038).
Conclusion The Pictorial-CERT is more appropriate for use with children of this age range and
appears to be more reproducible than the Borg CR-10 Scale. Concurrent and construct validity
evidence promotes the use of the Pictorial-CERT by junior children.

Research on the diverse aspects of exercise performance in chil-
dren in the past 20 years has included a marked interest in the
concept of perceived exertion. This has been accompanied by
the development of various scales which have been designed to
accommodate the cognitive abilities of children and improve
the accuracy of the measurement of perceived exertion. Prepu-
bescent children understand categorization, but find it easier to
comprehend and interpret pictures and symbols than words
and numbers (Eston & Lamb 2000; Robertson et al. 2000).
Following suggestions that a greater conceptual understanding
of the effort continuum in children would be facilitated by a
meaningful pictorial scales of perceived exertion (Noble &
Robertson 1996), several studies have proposed appropriate
child-specific rating scales. Some of them have incorporated
stylized figures to grade effort perception (Nystad et al. 1989;
Eston et al. 2000; Robertson et al. 2000). These include the Cart
and Load Effort Rating Scale (CALER; Eston et al. 2000), the
OMNI scale (Robertson et al. 2000), the rating of perceived
exertion in children (RPE-C) (Groslambert et al. 2001), the Bug
and Bag Exertion (BABE) Scale (Parfitt et al. 2007) and the
Curvilinear Scale (Eston & Parfitt 2006).
The Borg Category-Ratio Perceived Exertion (CR-10) Scale
(1982) is a measurement tool known to be highly sensitive
to detect general fatigue. It is an internationally recognized

© 2007 The Authors

Journal compilation © 2007 Blackwell Publishing Ltd 35
36 B. Marinov et al.

method of assessing perceived exertion, and it is used in con- Table 1. Anthropometric characteristics (mean ⫾ SD) of the studied
children during the first and second sessions
junction with standard metabolic measures in a wide range of
settings. Wilson and Jones (1991) have shown that Borg scale Parameter T1 or T2 T3

readings are reproducible in both the short and long term, but Age (years)
there is a critical age (7–11 years) for understanding of the Boys 10.5 ⫾ 0.5 13.5 ⫾ 0.5†
Girls 10.4 ⫾ 0.5 13.4 ⫾ 0.5†
concept of this scale. It is generally recognized that most pre- Height (cm)
pubescent children do not yet have the cognitive ability or Boys 144.5 ⫾ 6.2 163.5 ⫾ 7.5†
verbal capacity to understand the descriptors of exercise inten- Girls 145.0 ⫾ 7.2 160.1 ⫾ 5.6†
Weight (kg)
sity associated with adult-perceived exertion scales (Williams Boys 41.2 ⫾ 10.7 59.2 ⫾ 15.4†
et al. 1994; Eston & Lamb 2000; Robertson et al. 2000; Eston & Girls 35.2 ⫾ 6.2* 49.4 ⫾ 7.4*†
Parfitt 2006). BMI (kg/m2)
Boys 19.5 ⫾ 3.7 22.0 ⫾ 4.9†
The formation of the Children’s Effort Rating Table (CERT)
Girls 16.7 ⫾ 2.3* 19.3 ⫾ 2.6*†
scale (Eston et al. 1994; Williams et al. 1994) was an original Per cent body fat
attempt to tackle the problem of children’s comprehension of Boys 18.8 ⫾ 6.4 21.8 ⫾ 6.7†
exercise effort. This idea was extended further by Eston and Girls 20.9 ⫾ 8.2* 24.8 ⫾ 7.9*†

colleagues (2000) with the reduction of the verbal descriptors *Gender difference (P < 0.05 for all).
†Differences between first and second session.
and inclusion of stylized figures. In an endeavor to fine-tune the
BMI, body mass index.
CERT, they created the so-called CALER (Cart and Load Effort
Rating) scale, presenting a figure of a child pulling a cart loaded
progressively with bricks, and the BABE Scale, which presented n = 25 Borg
PCERT Borg
a cartoon character stepping onto a box with a backpack con- n = 50 +
taining coloured bricks commensurate with the number on the n = 25 Borg PCERT PCERT

scale (Parfitt et al. 2007). More recently, Yelling and colleagues

(2002) introduced the Pictorial Children’s Effort Rating Table
(Pictorial-CERT), which uses all of the original numerical and
verbal descriptors of the CERT as proposed by Williams and
colleagues (1994). Although the reliability and validity of the
other alternative scales have yet to be unequivocally established,
studies in young children (aged 8–11) using CERT have shown
it to have greater validity when compared with the traditionally
adopted rating of perceived exertion (RPE) scale (Lamb 1996;
Leung et al. 2002). The validity of the Pictorial-CERT was
assessed recently by Roemmich and colleagues (2006), but to
the best of our knowledge, there are no data on its short- or
long-term repeatability.
The aim of this study was to compare the effort estimation by
means of the Pictorial-CERT and CR-10 scales in children and
to assess the long-term repeatability of the two scales.
T1 T2 T3
1 month 3 years
Figure 1. Design of the study. PCERT, Pictorial vision of the Children’s
Effort Rating Table.
presented in Table 1. The children were generally physically
active, but not engaged in sports training of any kind. Normal
subjects were chosen for the study because it was considered that
they would provide information that could be later used as a
reference to a patient population. Prior to the test procedures,
written informed consent was obtained from a parent or guard-
ian, and the associated risks and benefits were explained. The
procedures used in this study were approved by the Institutional
Ethics Committee at the Medical University of Plovdiv.

Methods Study design

Subjects
Fifty healthy children (25 boys and 25 girls; initially aged
10.4 ⫾ 0.5 years, and after 3 years) took part in the study. The
children did not suffer chronic diseases and took no medications
that might affect exercise performance and judgement. Descrip-
tive characteristics of the boys and girls recruited in the study are
The children performed a total of three incremental exercise
tests in two sessions. This is shown in Fig. 1. The first session
included two tests (T1 and T2) at an interval of 1 month.
The same treadmill protocol was used in all testing occasions.
Effort perception was assessed by only one scale at a time, and
the children were randomly assigned to a group utilizing one of
the measurement scales, alternating with the other on the

© 2007 The Authors

Journal compilation © 2007 Blackwell Publishing Ltd, Child: care, health and development, 34, 1, 35–43

second visit. We chose a period close to the upper limit of short
term to ensure that, on the second visit, the children would not
remember their ratings on the previously utilized scale. The
second session took part 3 years apart and consisted of only one
treadmill test (T3). The same protocol was applied to the same
children for verification of the long-term repeatability of these
RPE instruments. Both scales (Borg CR-10 and Pictorial-CERT)
were displayed one after another at each stage of the test, and the
order of their presentation was also randomized. Special atten-
tion was paid to ensure absolute standardization of procedures
and conditions, which is a prerequisite for unbiased comparison
between the tests.
Treadmill exercise test
The treadmill tests were performed in the morning with the
children habituated to both the general environment and the
actual procedures. The exercise was carried out on a motor-
driven, electronically controlled treadmill (TrackMaster™, JAS
Fitness Systems, Pensacola, FL, USA). The protocol imple-
mented in all of the tests was a validated modification of Balke
protocol similar to that used by Riopel and colleagues (1979),
except that in our laboratory we start with a rapid initial incre-
ment of 6%. The slope is then increased by 2% every minute
until volitional exhaustion or elevation of 22%. Thus, the pro-
tocol consisted of nine 1-min increments with constant velocity
of 5.4 km/h. The recovery period had standard 3-min duration
(2.7 km/h and zero elevation). Heart rate (HR) was monitored
electrocardiographically (Hellige, Freiburg, Germany), and the
oxygen saturation was traced with a pulseoxymeter (Pulseox
DP-8, Minolta, Osaka, Japan).
Gas exchange measurements
Throughout the test, gas exchange variables were determined
with an online computerized system (CardiO2™, Medical
Graphics, St Paul, MN, USA) using standard open-circuit
techniques. Subjects breathed through a mouthpiece (dead
space < 60 mL), and a pneumotachometer was used for record-
ing of tidal volume (VT; mL/min, BTPS) and minute ventilation
(VE; L/min, BTPS). Data were averaged every 30 s and used to
calculate oxygen uptake (VO2; mL/min, STPD), carbon dioxide
production (VCO2; mL/min, STPD) and respiratory exchange
ratio (RER). The system was calibrated before each test with
gases of known concentrations.
Effort perception
Perceived exertion was assessed by means of two different scales.
The CR-10 Scale (Borg 1982), presented in Table 2, has a more
Journal compilation © 2007 Blackwell Publishing Ltd, Child: care, health and development, 34, 1, 35–43
Pictorial and verbal scales for effort estimation 37
Table 2. Category Ratio Borg (CR-10) Scale.
0 Nothing at all
0.5 Very, very light
1 Very light
2 Light
3 Moderate
4 Somewhat heavy
5 Heavy
6
7 Very heavy
8
9
10 Very, very heavy (almost maximum)
• Maximum
Figure 2. Pictorial variant of the Children’s Effort Rating Table (Williams
et al. 1994; Yelling et al. 2002).
limited number range, and nearly all numbers are associated
with verbal expressions. The rationale for application of this
variant (CR-10) of the Borg scale was that we believe it is better
understood by the children in our study for cultural and devel-
opmental reasons.
The Pictorial-CERT is a scale combining child-specific verbal
descriptors of perceived exertion using the whole of the original
scale of Williams and colleagues (1994) and adjacent pictorial
content (Yelling et al. 2002; Fig. 2). The scale was designed to be
readily assimilated by the children on the basis of their own
experience and has been validated on a slightly older group of
children aged 12 and 15 years (Yelling et al. 2002).
A standard definition of perceived exertion adapted for chil-
dren and separate instructional sets for the Borg CR-10 Scale
and Pictorial-CERT were read to the subjects prior to the test.
On both scales, the participants were asked to rate the perceived
effort, and these values were processed in the subsequent analy-
sis. The children were asked to provide a rating of perceived
exertion in the final 30 s of each treadmill stage. The children
© 2007 The Authors
38 B. Marinov et al.

were not reminded of their previous ratings in order to avoid Table 3. Peak cardiopulmonary exercise variables (mean ⫾ SD)
bias. Exercise tests
Parameter T1 T2 T3 ANOVA

Other measurements VO2 (mL/min) 1291 ⫾ 282 1270 ⫾ 275 1912 ⫾ 399 *
VO2/kg (mL/min/kg) 34.1 ⫾ 3.0 33.6 ⫾ 2.9 35.5 ⫾ 3.5 †
Prior to the exercise, all the participants underwent a thorough VE (L/min) 40.5 ⫾ 11.6 40.4 ⫾ 12.0 56.6 ⫾ 13.8 *
BF (per min) 44.8 ⫾ 8.2 42.3 ⫾ 8.3 43.3 ⫾ 8.6 NS
clinical examination and were subjected to complete anthropo-
HR (beat/min) 178.2 ⫾ 7.5 175.7 ⫾ 7.6 170.5 ⫾ 10.8 *
metric measurements, including stature, body mass, and skin- VE/VO2 31.1 ⫾ 4.2 31.7 ⫾ 4.3 29.7 ⫾ 4.2 †
fold thickness over the triceps and subscapular regions. The RER 1.06 ⫾ 0.10 1.04 ⫾ 0.10 1.03 ⫾ 0.09 NS
measurements were performed on the right side of the body Statistical significance P < 0.05.
with Harpenden calipers (British Indicators, St Albans, UK). *Third test differs from the first and the second one.
†Third test differs from the second one.
The values of the skinfolds were added together, and the sum
NS, non significant; VO2, peak oxygen uptake; VO2/kg, relative oxygen uptake;
was used to calculate per cent body fat (Slaughter et al. 1988). VE, total exercise ventilation; BF, breathing frequency; HR, heart rate; VE/VO2,
For stratifying children by their body mass index (BMI) as ventilatory equivalent for oxygen; RER, respiratory exchange ratio.
‘normal’, ‘overweight’ and ‘obese’, we used the cut-off points
published by Cole and colleagues (2000), based on BMI centiles
for subjects aged 2–18 years. The adopted cut-off points for the children had considerably increased their body size and
overweight and obesity in the pooled reference values corre- related parameters – BMI and % body fat.
spond to BMI = 25 kg/m2 and BMI = 30 kg/m2 in adults. There were significant gender differences with regard to
According to these criteria, the overweight group consisted of 15 bodyweight, BMI and % body fat (P < 0.05 for all) at the time of
children in the first two tests (11 with overweight and 4 with the initial tests and the subsequent test. Boys were significantly
obesity) and 14 children in the third test (10 with overweight heavier than girls, but at the same time they had significantly
and 4 with obesity). lower % body fat.

Statistical methods
All values are expressed as mean ⫾ SD. The results from peak
exercise data and their relationships with age and anthro-
pometric variables were assessed using descriptive statistics,
independent- and paired-samples t-tests, and correlation analy-
sis in SPSS for Windows (SPSS Inc., Chicago, IL, USA). Data
were analysed by a repeated-measures anova. Post hoc compari-
sons were assessed by the Bonferroni method. Intraclass corre-
lation analysis (ICC) was used to assess the test–retest reliability.
Fisher z-transformation was employed to test differences in the
correlation coefficients between groups and tests (MedCalc 9,
MedCalc Software, Mariakerke, Belgium). A level of P < 0.05
was considered to be statistically significant. The agreement
between the two methods of measurement of perceived exertion
was examined by means of the Bland and Altman method
(1986).
Results
Descriptive characteristics
A summary of the anthropometric characteristics is presented
in Table 1. As expected, over the 3-year period between the tests,
Peak cardiopulmonary variables
All children in the three exercise tests attained mean RER values
in excess of 1.00, providing evidence that a maximal effort had
been achieved. However, HR in the first and the second tests
reached mean values slightly below 180 per min.
The values of oxygen uptake and parameters of ventilation
(total exercise ventilation – VE, breathing frequency – BF, ven-
tilatory equivalent for oxygen – VE/VO2) at the time of the first
(T1) and the second (T2) tests were very close, without sig-
nificant gender differences (Table 3). Therefore, an excellent
reproducibility between the tests existed, and the correlation
coefficients between any pair of parameters ranged from 0.945
to 0.998 (P < 0.001 for all).
For VO2, significant main effects were found for treadmill
gradient (F1,50 = 134.4; P < 0.01) and gender (F1,50 = 12.7;
P < 0.01). Most of the functional parameters show growth-
dependent changes. It is evident from Table 3 that there are
significant differences in cardiopulmonary parameters between
T1 and T2 (the first session) and T3 (the second session). In T3,
children had significantly higher VO2 and total exercise ventila-
tion (VE). The HR during T3 was also slightly lower compared
with the values attained in T1 and T2 (P < 0.05 for both).

© 2007 The Authors

Journal compilation © 2007 Blackwell Publishing Ltd, Child: care, health and development, 34, 1, 35–43
 Pictorial and verbal scales for effort estimation 39

Table 4. Perceived effort (mean ⫾ SD) on the Borg CR-10 Scale and Table 5. Correlations (mean ⫾ SD) for perceived effort assessed by the
PCERT for three different exercise intensities in the three tests two scales with some physiologic measures of exercise intensity
Exercise intensity T1 or T2 T3 Borg (CR-10) PCERT
1st stage (6% slope) Physiological measurement T1 or T2 T3 T1 or T2 T3
PCERT 2.3 ⫾ 0.9 2.4 ⫾ 0.7
All participants (n = 50)
Borg 1.2 ⫾ 1.1* 1.1 ⫾ 0.9*
VO2 0.710 0.701 0.823* 0.878*
5th stage (14% slope)
HR 0.634 0.673 0.751* 0.793*
PCERT 3.7 ⫾ 1.2† 3.9 ⫾ 0.9†
VE 0.512 0.593 0.618* 0.635*
Borg 2.2 ⫾ 1.3†* 2.1 ⫾ 1.2†*
Boys (n = 25)
9th stage (22% slope)
VO2 0.702 0.725 0.815* 0.854*
PCERT 6.2 ⫾ 1.0† 6.0 ⫾ 1.3†
HR 0.619 0.697 0.728* 0.815*
Borg 4.6 ⫾ 1.3†* 4.1 ⫾ 1.2†*
VE 0.486 0.603† 0.594 0.648
Significant differences (P < 0.05). Girls (n = 25)
*Between scales. VO2 0.715 0.678 0.832* 0.897*
†Between stages. HR 0.589 0.588 0.764* 0.773*
PCERT, Pictorial vision of the Children’s Effort Rating Table. VE 0.520 0.534 0.642* 0.615*

Significance level for all correlations P < 0.05.

Measurement of perceived exertion
A comparison of the corresponding verbal characteristics of the
scales revealed a significant overlap. Children’s average ratings
in both scales define the standardized workload achieved
through the modified Balke protocol as slightly below ‘heavy’
(Borg) or ‘hard’ (Pictorial-CERT).
The mean level of effort perception assessed with the CR-10
Scale during T1 and T2 (first session) was 4.60 ⫾ 1.3 [mean
centred coefficient of variation (CoV%) = 28.5%; range = 2 to 8
(corresponding verbal expression ‘light’ to ‘very heavy’]. When
the mean level of effort perception was evaluated with the pic-
torial variant of the CERT, children gave higher mean values of
6.2 ⫾ 1.3 [CoV% = 20.8%; range = 4 to 10 (‘just feeling a strain’
to ‘so hard I’m going to stop’)]. Only a few children gave ratings
exceeding 5 (heavy) on the CR-10 Scale, raising concerns about
the proper understanding of these grades.
Perceived effort on the Borg CR-10 Scale and Pictorial-CERT
for three different exercise intensities in the three tests is pre-
sented in Table 4.
For perceived effort ratings, significant main effects were
found for scale (F1,50 = 202.5; P < 0.05) and treadmill gradient
(F1,50 = 119.8; P < 0.05). No significant interactions (P > 0.05)
were found for perceived effort, HR, VE and VO2 data. Correla-
tion coefficients for perceived effort rated by the two scales with
some physiologic measures of exercise intensity (oxygen uptake,
HR and exercise ventilation) are presented in Table 5.
Effort perceptions in the three tests (T1/T2 and T3) assessed
with both the CR-10 and the Pictorial-CERT were significantly
correlated (P < 0.001) with VO2, HR and exercise ventilation
for all children. The Pearson correlation coefficients for the
Pictorial-CERT were significantly higher (P < 0.05; Fisher
z-transformation) than for the CR-10 Scale. Neither of the
*Scales significantly different.
†Tests significantly different.
PCERT, Pictorial vision of the Children’s Effort Rating Table.
scales in any of the tests displayed gender differences in effort
perception in our sample. There were also no significant differ-
ences in the correlations between the physiological measure-
ments and rating scales in T1/T2 vs. T3, with the exception of
significantly lower correlation coefficient between the CR-10
Scale and VE in boys.
The CR-10 ratings decreased slightly in the 3-year follow-up
test: from 4.60 ⫾ 1.3 in T1 to 4.08 ⫾ 1.2 in T3 (P = 0.004),
while the Pictorial-CERT values changed imperceptibly from
6.16 ⫾ 1.3 to 6.02 ⫾ 1.3 (P = 0.267). The CoV% in the third test
was also lower for the Pictorial-CERT compared with the CR-10
Scale (21.6% vs. 29.4%). In addition, the correlations between
the two scales (CR-10 and Pictorial-CERT) increased signifi-
cantly at the time of T3 (r = 0.642 vs. 0.835 for T1/T2 and T3,
respectively; z = -2.21; P = 0.027).
ICCs of the ratings from the CR-10 and the Pictorial-CERT
in T1/T2 and T3, which can be interpreted as a measure of
the test–retest reliability (or long-term repeatability), were
also highly significant: r = 0.540 and r = 0.773 respectively
(P < 0.001 for both). However, the ICC for the Pictorial-CERT
was significantly higher (z = -2.07; P = 0.038). In line with these
results, the limits of agreement of the mean values of the
Pictorial-CERT score were narrower compared with the CR-10
Scale. A comparison between the two scales (T1/T2 and T3)
with the mean ratings at peak exercise, mean differences, limits
of agreement, and ICCs are presented in Table 6.
A Bland and Altman diagram plotting differences between
the two scales in T3 against the respective mean values is pre-
sented in Fig. 3. The solid line represents the bias estimated by

© 2007 The Authors

Journal compilation © 2007 Blackwell Publishing Ltd, Child: care, health and development, 34, 1, 35–43
40 B. Marinov et al.

Table 6. Comparison of mean (SD) scores of different scales in different tests, and test–retest reliability (ICC) (n = 50)

Mean (SD) Mean (SD) Mean (SD) Limits of

RPE scale First scale Second scale Difference* agreement† P-value‡ ICC (95% CI)
Borg (T1 or T2) & Borg (T3) 4.6 (1.3) 4.1 (1.2) 0.52 (1.22) -1.87 + 2.91 0.004 0.54 (0.31–0.71)
PCERT (T1 or T2) & PCERT (T3) 6.2 (1.0) 6.0 (1.3) 0.14 (0.88) -1.58 + 1.86 0.267 0.77 (0.63–0.86)
PCERT (T1 or T2) & Borg (T2/T1) 6.2 (1.0) 4.6 (1.3) 1.56 (1.09) -0.58 + 3.70 <0.001 0.64 (0.44–0.78)
PCERT (T3) & Borg (T3) 6.0 (1.3) 4.1 (1.2) 1.94 (0.74) 0.49 + 3.39 <0.001 0.84 (0.73–0.90)

*Mean difference between the first and the second test.

†Mean difference ⫾1.96 SD.
‡P-value on differences between RPE scales from paired t-test.
ICC, intraclass correlation coefficient; CI, confidence interval; T1 and T2, tests in the first session; T3, test in the second session.
PCERT, Pictorial vision of the Children’s Effort Rating Table.

5.0 Table 7. Comparison of the ratings obtained with the two scales in the
first and the second sessions in the groups of normal-weight and
overweight children
4.0 Borg CR-10 PCERT
Body mass T1 or T2 T3 T1 or T2 T3
Normal 4.3 ⫾ 1.1 3.8 ⫾ 1.1 5.9 ⫾ 1.0 5.8 ⫾ 1.3
3.0
(n = 35) (n = 36) (n = 35) (n = 36)
Overweight 5.3 ⫾ 1.5 4.8 ⫾ 1.4 6.7 ⫾ 1.7 6.7 ⫾ 1.4
(n = 15) (n = 14) (n = 15) (n = 14)
PCERT - Borg CR-10

2.0 P-value 0.041 0.026 0.035 0.033

PCERT, Pictorial vision of the Children’s Effort Rating Table.

1.0

The values of BMI (kg/m2) for normal and overweight groups

0.0
-1.0
2 3 4 5 6 7
(PCERT + Borg CR-10)/2
Figure 3. Bland and Altman plot of the two scales in the third test (T3).
The petals represent the number of cases in a single locus; the solid
reference line: the mean difference; upper dashed line: mean + 2 SD; and
lower dashed line: mean - 2 SD. PCERT, Pictorial vision of the Children’s
Effort Rating Table.
the mean difference of two scales’ values, and the dashed lines
are drawn at a distance of mean ⫾ 2 SD. The results from appli-
cation of this method show wide limits of agreement of the
mean values (bias = 1.93, 0.46 and 3.42), providing evidence
that the difference between the numerical expressions of the
two scales is substantial.
Effect of overweight on effort perception
The effect of excess body mass (overweight and obesity) on the
mean ratings of effort perception is presented in Table 7.
were 16.2 ⫾ 1.4 vs. 22.5 ⫾ 2.2 at the time of T1/T2 respectively.
For T3, the corresponding values were 18.5 ⫾ 1.7 vs. 26.0 ⫾ 3.6
respectively. It is evident that the greater the BMI, the higher the
increase in effort perception. There were significant differences
8 9
between the groups across all trials, regardless of which per-
ceived effort rating scale was used. Effort perception ratings
correlate significantly with BMI and % body fat in the two
sessions (r = 0.3–0.4; P < 0.05 for all).
Discussion
The study observed that children in the 10th year of their life
find it easier to use a combination of styled images and charac-
teristic verbal phrases associated with the Pictorial-CERT. In
addition, the long-term repeatability of the Pictorial-CERT was
significantly better than that of the CR-10 Scale.
In this study, we used identical treadmill protocols to ascer-
tain unbiased comparisons of perceptual understanding. The
treadmill exercise test involved brisk walking until achievement
of volitional exhaustion or peak VO2. This is a recognized test
for assessing aerobic performance in children. It avoids the
problem of the maximal test being too strenuous and unlikely to

© 2007 The Authors

Journal compilation © 2007 Blackwell Publishing Ltd, Child: care, health and development, 34, 1, 35–43

relate to daily activities. In the present study, children attained
HRs that ranged from 168 to 192 beats per min. This range is
similar to that reported in previous studies that have used the
Borg RPE scale (Ward et al. 1991) and Pictorial-CERT (Roem-
mich et al. 2006). The testing protocol may have been a factor
which accounted for why some children did not appear to
achieve maximal HRs (Armstrong & Welsman 1994). This
could be due to the muscular effort required and the associated
local lactate production in the legs to walk up such steep gradi-
ents near the terminal point of the test.
Several studies have validated the original CERT proposed by
Williams and colleagues (1994). Studies that have compared the
CERT with the 6–20 RPE during stepping in children aged
5–9 years (Williams et al. 1994), and during cycling exercise in
children aged 8–11 years (Lamb 1996) and 10–11 years (Leung
et al. 2002), provided further support for its use.
In the present study, the relationship between physiological
measures and rating scales indicated that the Pictorial-CERT
consistently correlated more highly with measures of exercise
intensity (VO2, HR, VE) than the CR-10 Scale. A study by Leung
and colleagues (2002) on 69 Chinese children assessed the
concurrent validity and the reliability of a Chinese-translated
(Cantonese) version of the Borg 6–20 RPE and the CERT
during cycling. The correlations for CERT, power output, HR
and oxygen uptake were consistently higher than those for the
6–20 RPE Scale. They also reported higher reliabilities (ICCs)
for the CERT (0.96 and 0.89 for the CERT and the Borg 6–20
Scale respectively).
When testing concurrent validity, we found significant cor-
relation coefficients for VO2, HR, VE and the RPE scales, which
ranged from moderate to very high (0.512–0.878) in the whole
group. These correlations are comparable to those established
by Utter and colleagues (2002) on similarly aged children tested
with the Borg 6–20 Scale and a children’s variant of the OMNI
scale. A test of construct validity showed a moderate to high
positive linear relationship between the CR-10 Scale and the
Pictorial-CERT (0.642 and 0.839 in T1/T2 and T3 respectively).
The relationship between perceived effort and oxygen uptake
(VO2), using variations of the CERT, has been documented in
several studies (Leung et al. 2002; Roemmich et al. 2006). In our
sample the correlations ranged between 0.701 and 0.878,
depending on the scale and the test. Regarding the relation
between the CR-10 and VO2, our findings are similar to other
studies (Mahon & Marsh 1992), and the correlations range from
0.0.678 to 0.725.
A good relationship has been reported between RPE and
HR (Borg 1982), but a causal link has been rejected. Our
study observed significant correlations between HR and effort
Journal compilation © 2007 Blackwell Publishing Ltd, Child: care, health and development, 34, 1, 35–43
Pictorial and verbal scales for effort estimation 41
perception determined by the Pictorial-CERT. This finding is
in agreement with most of the other studies that have vali-
dated scales of perceived exertion (Eston & Williams 1986;
Groslambert et al. 2001), differing only by the magnitude of
the correlations.
Product–moment correlation analysis between the results of
two measurement methods cannot always be regarded as an
indicator of agreement. Highly significant correlation coeffi-
cient can conceal a considerable lack of agreement (Bland &
Altman 1986). A plot of the difference between the methods
against their mean may be more informative (Bland & Altman
1986). The lack of agreement between the Borg and Pictorial-
CERT was shown in Fig. 3, where the discrepancies between the
ratings reached ⫾2.1 scale points. These differences can be
designated to the specific rating concept of the scales and the
heterogeneous cognitive level of development of the children
(Ward & Bar-Or 1990).
Fatness and excess bodyweight do not necessarily imply a
reduced ability to consume oxygen, but obesity appears to have
an adverse effect on submaximal aerobic capacity (Goran et al.
2000). In line with Ward and Bar-Or (1990), a greater awareness
of fatigue was found in the obese children (Marinov et al. 2002).
In the present study, overweight and obese children demon-
strated significantly greater RPE, irrespective of the trials and
the rating scales.
We also confirmed the practical applicability of the Pictorial-
CERT in 10- to 11-year-old boys and girls. According to the ICC
analysis in our study, the CR-10 Scale gives way to the Pictorial-
CERT regarding its long-term repeatability (r = 0.540 and
r = 0.773, respectively; P = 0.038). It is worth noting that only
ICCs higher than 0.7 were taken to be reliable (Williams et al.
2001).
Somatic and lung growth results in significantly higher body
mass, lung volumes and ventilation in 13-year-old children.
These changes are consistent with the higher VO2 and VE at peak
exercise and are consistent with data in the literature (Cotes
1993). It is unlikely that there will be a significant change in
perceived exertion in this particular age interval, which is to
some extent confirmed by the reliability results after 3 years.
Support for this observation is provided by another study in
which effort perception was estimated by the CR-10 Scale in the
age span of 7–14 years in a mixed gender sample of 614 children
(Marinov 2003).
There were several limitations to our study. The applied
treadmill protocol was able to elicit HR and VO2 responses
slightly above 80% of the predicted maximum. We could not
extend the analysis beyond this level. In addition, we utilized a
continuous protocol, and the influence of fatigue on effort per-
© 2007 The Authors
42 B. Marinov et al.
ception cannot be totally eliminated (Eston & Parfitt 2006). For
efficacy reasons, we preferred a saturated testing scheme (T1, T2
and T3 with combined reading) and careful standardization of
the protocol.
Conclusion
Children have a better comprehension of perceived exertion
when a combination of pictorial and verbal descriptors is used.
The Pictorial-CERT appears to be more reproducible than the
CR-10 Scale, and would be more sensitive because it offers finer
adjustment. Concurrent and construct evidence is presented in
this study, which justifies the use of a pictorial variant of the
CERT to estimate RPE during treadmill exercise. These findings
confirm the potential of pictorial child-specific scales for use
with prepubescent children and suggest that further research
using such scales is desirable.
Acknowledgement
We would like to thank Professor Roger Eston for his helpful
comments and final revision of the manuscript.
References
Armstrong, N. & Welsman, J. R. (1994) Assessment and
interpretation of aerobic fitness in children and adolescents.
Exercise and Sport Sciences Reviews, 22, 435–476.
Bland, J. & Altman, D. (1986) Statistical methods for assessment of
agreement between two methods for clinical measurement. Lancet,
I, 307–310.
Borg, G., (1982) Psychophysical bases of perceived exertion. Medicine
and Science in Sports Exercise, 14, 377–381.
Cole, T., Bellizzi, C., Flegal, M. & Dietz, H. (2000) Establishing
standard definition for child overweight and obesity worldwide:
international survey. British Medical Journal, 320,
1240–1243.
Cotes, J. E. (1993) Lung function through life: determinants and
reference values. In: Lung Function. Assessment and Application in
Medicine, 445–513. Blackwell Science Publications, Oxford, UK.
Eston, R. & Lamb, K. (2000) Effort perception. In: Pediatric Exercise
Science and Medicine (eds N. Armstrong & W. Van Mechelen),
pp. 87–91. Oxford University Press, Oxford, UK.
Eston, R. & Williams, J. (1986) Exercise intensity and perceived
exertion in adolescent boys. British Journal of Sports Medicine, 20,
27–30.
Eston, R., Lamb, K., Bain, A., Williams, A. & Williams, J. (1994)
Validity of a perceived exertion scale for children: a pilot study.
Perceptual and Motor Skills, 78, 691–697.
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd, Child: care, health and development, 34, 1, 35–43
Eston, R. G. & Parfitt (2006) ‘Effort perception’. In: Paediatric
Exercise Physiology (ed. N. Armstrong), pp. 275–298. Elsevier,
London, UK.
Eston, R. G., Parfitt, G., Campbell, L. & Lamb, K. L. (2000) Reliability
of effort perception for regulating exercise intensity in children
using the Cart and Load Effort Rating (CALER) Scale. Pediatric
Exercise Science, 12, 388–397.
Goran, M., Fields, D. A., Hunter, G. R., Herd, S. L. & Weinsier, R. L.
(2000) Total body fat does not influence maximal aerobic capacity.
International Journal of Obesity and Related Metabolic Disorders, 24,
841–848.
Groslambert, A., Hintzy, F., Hoffman, M. D., Dugue, B. & Rouillon, J.
D. (2001) Validation of a rating scale of perceived exertion in young
children. International Journal of Sports Medicine, 22, 116–119.
Lamb, K. (1996) Exercise regulation during cycle ergometry using the
CERT and RPE scales. Pediatric Exercise Science, 8, 337–350.
Leung, M. L., Chung, P. K. & Leung, R. W. (2002) An assessment of
the validity and reliability of two perceived exertion rating scales
among Hong Kong children. Perceptual and Motor Skills, 95,
1047–1062.
Mahon, A. D. & Marsh, M. L. (1992) Reliability of the rating of
perceived exertion at ventilatory threshold in children.
International Journal of Sports Medicine, 13, 567–571.
Marinov, B. (2003) Physiologic and pathophysiologic aspects of
exercise gas exchange in school children. Doctoral Thesis. Medical
University, Plovdiv, Bulgaria (in Bulgarian).
Marinov, B., Kostianev, S. & Turnovska, T. (2002) Ventilatory
efficiency and rating of perceived exertion in obese and non-obese
children performing standardized exercise. Clinical Physiology and
Functional Imaging, 22, 254–260.
Noble, B. & Robertson, R. (1996) Historical development of
perceived exertion. In: Perceived Exertion, pp. 59–77. Human
Kinetics, Chamaign, IL, USA.
Nystad, W., Oseid, S. & Mellbye, E. B. (1989) Physical education for
asthmatic children: the relationship between changes in heart rate,
perceived exertion, and motivation for participation. In: Children
and Exercise XIII (eds S. Oseid & K. Carlson), pp. 369–377. Human
Kinetics, Champaign, IL, USA.
Parfitt, G., Shepherd, P. & Eston, R. G. (2007) Reliability of effort
production using the children’s CALER and BABE perceived
exertion scales. Journal of Exercise Science and Fitness (in press).
Riopel, D., Taylor, A. & Hohn, A. (1979) Blood pressure, heart rate,
pressure rate product and echocardiographic changes in healthy
children during treadmill exercise. American Journal of Cardiology,
44, 697–704.
Robertson, R., Goss, F., Boer, N., Peoples, J., Foreman, A., Dabayebeh,
I., Millich, N., Balasekaran, G., Riechman, S., Gallagher, J. &
Thompkins, T. (2000) Children’s OMNI scale of perceived
exertion: mixed gender and race validation. Medicine and Science
in Sports Exercise, 32, 452–458.
Roemmich, J. N., Barkley, J. E., Epstein, L. H., Lobarinas, C. L.,
White, T. M. & Foster, J. H. (2006) Validity of PCERT and OMNI
walk/run ratings of perceived exertion. Medicine and Science in
Sports Exercise, 38, 1014–1019.

Slaughter, M., Lohman, T., Boileau, R., Horswill, C., Stillman, R., Van
Loan, M. & Bemben, D. (1988) Skinfold equations for estimation
of body fatness in children and youth. Human Biology, 60,
709–723.
Utter, A. C., Robertson, R. J., Nieman, D. C. & Kang, J. (2002)
Children’s OMNI Scale of Perceived Exertion: walking/running
evaluation. Medicine and Science in Sports Exercise, 34, 139–144.
Ward, D. & Bar-Or, O. (1990) Use of the Borg scale in exercise
prescription for overweight youth. Canadian Journal of Sport
Sciences, 15, 120–125.
Ward, D. S., Jackman, J. D. & Galiano, F. J. (1991) Exercise intensity
reproduction: children versus adults. Pediatric Exercise Science, 3,
209–218.
Journal compilation © 2007 Blackwell Publishing Ltd, Child: care, health and development, 34, 1, 35–43
Pictorial and verbal scales for effort estimation 43
Williams, J., Eston, R. & Furlong, B. (1994) CERT: a perceived
exertion scale for young children. Perceptual and Motor Skills, 79,
1451–1458.
Williams, J., Singh, S., Sewell, L., Guyatt, G. & Morgan, M. (2001)
Development of a self-reported Chronic Respiratory Questionnaire
(CRQ-SR). Thorax, 56, 954–959.
Wilson, R. C. & Jones, P. W. (1991) Long-term reproducibility of
Borg scale estimates of breathlessness during exercise. Clinical
Science, 80, 309–312.
Yelling, M., Lamb, K. & Swaine, I. (2002) Validity of a pictorial
perceived exertion scale for effort estimation and effort production
during stepping exercise in adolescent children. European Physical
Education Review, 8, 157–175.
© 2007 The Authors