Pediatric Exercise Science, 1989,1 , 360368
The Relationship Between Perceived Exertion
and Heart Rate in Children and Adults
Mary C. Gillach, James F. Sallis, Michael J. Buono,
Patricia Patterson, and Philip R. Nader
This study examined the relationship between heart rate (HR) as a measure
of physiological strain and ratings of perceived exertion (RPE) in 193 children
(mean age= 11 yrs) and 188 adults (mean age =36 yrs) during submaximal
cycle ergometry. Two methods of correlating HR and RPE were compared.
Computing correlations (r) for each individual's data and then taking the group
mean produced very high rs, ranging from 0.92 to 0.95. Correlating HR
and RPE for the entire group at all powers simultaneously produced much
lower rs, ranging from 0.63 to 0.65. Correlations were essentially the same
for children and adults, and there was no evidence of a practice effect. The
results indicated that (a) children in this age group were as capable of expressing RPE as adults, and @) absolute levels of perceived exertion were
not predictive of physiological strain (as indicated by heart rate).
Exercise intensity is reflected in the response of many physiological processes including oxygen consumption, blood pressure, blood lactate levels, and
heart rate (HR). Because HR is linearly related to power and oxygen consumption (VOz) in adults (1, 18, 27), it is accepted as an accurate index of physiological strain.
An alternative method of monitoring exercise intensity is ratings of perceived exertion (WE). These ratings, obtained from subjects during testing, are
subjective estimates of strain that simultaneously reflect both physiological and
psychological variables. The most commonly used RPE scale is the 15-point scale
by Borg (6). This scale, which ranges from 6 to 20, was constructed to correspond
with the normal HR range for most healthy middle-aged subjects. Thus, a given
RPE value multiplied by 10 should approximate exercising HR (7, 8). Several
studies by Borg and others have confirmed this 1:10 ratio of RPE to exercise
HR in adults (6, 9, 13, 14, 17, 27) while others have not (12, 28). One study
found this ratio to be accurate in adults but not in children (2). Whether HR can
be reliably estimated by RPE remains unclear.
Mary C. Gillach, Michael J. Buono, and Patricia Patterson are with the Departments of Biology and Physical Education at San Diego State University, San Diego, CA
92182. James F. Sallis and Philip R. Nader, M.D., are with the Department of Pediatrics
at the University of California at San Diego, La Jolla, CA 92093.
Supported in part by NM grant HL 30872.
Perceived Exertion and Heart Rate
- 361
The positive linear relationship between HR and RPE over a range of powers
is well documented in adults (6, 11, 24, 25, 28, 29). Numerous studies report
correlations (r) ranging from 0.77 to 0.90 (4, 6, 10, 27). Based on these high
correlations, subjective ratings are presumed to accurately reflect physiological
strain in adults. However, the statistical methodology used to calculate r s in these
studies is often inadequately described. Few authors specified whether rs for individual subjects were used to calculate group mean rs or if rs were calculated
from the simultaneous analysis of all subjects' HR and RPE data. Most studies
that reported their methodology used the former method, which is likely to inflate
the correlation. It appears that the latter correlation method, which considers all
variables at each power, is more appropriate.
Few studies have examined how children perceive their physiological strain.
It is difficult to compare existing studies of adults and children because of differences in methodology. Studies of adults usually correlate absolute HR with
RPE while few studies in children did so (2, 3). Most studies of children correlate percent HR max with RPE, with resulting rs ranging from 0.69 to 0.94
in children over 9 years of age (22,30). Nevertheless, several investigators suggest that the reliability and validity of W E may be just as high in children as
in adults (2, 22, 30). Test-retest reliabilities of the use of RPE by children have
ranged from 0.37 to 0.89 (19, 22, 30) while reliabilities in adults have ranged
from 0.78 to 0.98 (20, 21, 28). Although a few studies have indicated that children are able to use W E , no studies have compared the validity of RPE scales
in children and adults from the same population using the same methods. The
purpose of the present study was to assess age-related differences in the relationship of RPE to physiological strain (HR).
Methods
Subjects
The subjects for this study were drawn from 95 Anglo and 111 Mexican American families (N=623) who were participating in the San Diego Family Health
Project (23). Families with fifth- or sixth-grade children were recruited from 12
elementary schools. For this analysis, 283 children ages 10 to 14 years (mean
age = 11 yrs) and 295 adults ages 22 to 55 (mean age =36 yrs), who completed
two submaximal cycle ergometer testing sessions 1 year apart, were studied. Subject characteristics for adults and children, by gender, are displayed in Table 1.
Procedure
Assessments were conducted by pairs of exercise physiology graduate students
who were thoroughly trained and certified in the protocols. Testing was conducted
on calibrated Monark cycle ergometers using a modified Astrand-Ryhming
protocol of gradually increasing power (26). Subjects pedaled at 50 rpm at an
initial power of 25 watts (W), and the power was increased 25 W every 2 minutes.
Most subjects completed three to five workloads. HR was assessed at the end
ofaach 2-min stage with a cardiotachometer and photoplethethysmographattached
to the earlobe (CIC Instruments, Hempstead, NY, Model 60200). Radial palpation was used in rare instances when the cardiotachometer malfunctioned. W E
was measured at rest and at the end of each 2-min stage with Borg's original
15-point scale (6) (see Figure 1). Brief standard instructions in the use of the
362
- Gillach, Sallis, Buono, Patterson, and Nader
Table 1
Characteristics of Subjects in the San Diego Family Health Project
(means and SD)
Boys
N=144
Weight (kg)
Height (cm)
Body mass index (kglmz)
Resting heart rate
VO, max (mllkglmin)
Girls
N = 139
Men
N = 100
Women
N=195
41 .O
(1 1.9)
145.3
(9.1)
19.2
(3.7)
74.7
(10.8)
50.8
(1 1.6)
W E scale were provided before the test. Subjects were instructed to say or point
to the number on the scale that described how they were feeling at the time. Subjects' understanding of the scale was tested by asking for a rating at rest and during their most vigorous exercise. Instructions and the Borg scale were available
in Spanish when needed. Tests were preceded by a standardized screening form
on which subjects reported history of cardiovascular and other diseases, relevant
symptoms, and medication use. Exercise tests in adults were terminated at 70%
age-predicted HR rnax or at an RPE of 15. The endpoint criterion for children
was 85% age-predicted HR max.
Two procedures were used to compute r s between HR and RPE in both
children and adults. The first method involved calculating the correlation across
powers for each subject and then deriving a group mean. This was the method
most often reported in the literature. The second method simultaneously considered
each HR-RPE pair at all powers for all subjects in the group. For both methods,
a Fisher z test was employed to test for differences between groups and between
trials.
Results
Table 2 presents correlations between HR and RPE based on the mean of individual correlations across powers. The extremely high correlations in all groups
indicated a strong association between HR and RPE. At Test 1 only, there was
a statistically significant difference between correlations for children and adults.
However, given the magnitude of the difference, it is of little practical significance.
There were no differences between correlations at Test 1 and Test 2. Table 3
presents correlations between HR and W E based on the entire group. There are
nonsignificant differences between adults and children and between Test 1 and
Test 2. These correlations are consistently much lower than those in Table 1.
Perceived Exertion and Heart Rate
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
- 363
Very, very light
Very light
Fairly light
Somewhat hard
Hard
Very hard
Very, very hard
Figure 1 - Borg's RPE scale (6).
Table 2
Correlation (0 Between HR and RPE in Children
and Adults Based on the Mean of Individual Correlations
Subject
Test 1
Test 2
Children
Adults
*Significant difference between children and adults, p<.05;
Note: All correlations were significant at p<.OOl.
Table 3
Correlation (r) Between HR and RPE in Children
and Adults Based on Simultaneous Analysis of the Entire Group
Subject
Test 1
Children
Adults
Note: All correlations were significant at p<.001.
Test 2
364
- Gillach, Sallis, Buono, Patterson, and Nader
Correlations were conducted separately for boys, girls, men, and women. Within
age groups, correlations were essentially identical for both gender groups.
Figures 2 and 3 further clarify the correlations in Table 2. These scatterplots from Test 1 indicate that, in general, there is a positive linear association
between HR and RPE, and the relationships are similar in children and adults.
However, at any given RPE value there is a wide range of HRs such that the
prediction of HR from RPE is very imprecise.
Discussion
This study clearly indicated that the type of statistical method used greatly influences one's conclusions regarding the relationship between HR and RPE. Previous investigators have tended to rely on the method of calculating correlations
for each individual subject and then generating a group mean. This method produces high correlations because as power and HR increase, RPE also increases.
In all individual subjects there was a linear association. This suggests that children
ADULT GROUP DATA
Figure 2 - The relationshi
f perceived exertion in adu
Perceived Exertion and Heart Rate - 365
CHILDREN GROUP DATA
Figure 3 - The relationship between submaximal exercise heart rates and ratings
of perceived exertion in children.
and adults are able to discriminate changes in physiological strain. Thus the RPE
scale can be used to properly monitor changes in exercise intensity since each
subject reported increases in RPE as exercise intensity increased over several
powers.
However, the RPE scale was also designed to reflect the level of physiological strain such that RPE could be used to estimate exercise HR. For example,
Bar-Or (2) found that RPE x 10 estimated HR adequately for middle-aged and
older adults but not for younger subjects. This use of W E was clearly not supported in either children or adults in the present study. Figures 2 and 3 indicate
that the range of HRs at every level of RPE was very large for children and adults.
Since each RPE value was associated with HRs that ranged over 40 to 100 bpm,
there is no useful predictive power of W E . Since these correlations were based
on submaximal tests, the distribution of HRs was truncated. This restriction of
range could have suppressed the correlations. In addition, RPE-HR correlations
may be lower than correlations between RPE and other physiological indicators
in children (3). Thus, present methods could have underestimated the associations.
366
- Gillach, Sallis, Buono, Patterson, and Nader
The confidence in these results is increased since almost identical results
were found for children and adults and for males and females. There were no
indications that practice improved the HR-RPE association. It is reasonable to
hypothesize that the ability of a 35-year-old subject to accurately report his or
her level of exertion would be better than that of a 10-year-old, given differences
in cognitive abilities. However, no such differences were found. Given that very
few subjects had previous experience with exercise testing or rating their exertion level, improvement with practice could be expected. This was tested by evaluating subjects' first exercise test and later their third test. A second test was
conducted at the 3-month point but those data were not analyzed for this paper.
Although the practice hypothesis was not supported by the present study, the
measurement schedule was not optimal for testing the hypothesis. For maximal
effect, a practice session should probably be held a few days prior to the testing
session.
Ratings of perceived exertion are commonly used as indices of exercise
intensity when prescribing exercise regimens (5, 15, 16). This practice is based
on the assumption that RPE reflects physiological strain, but it appears that such
an assumption is unwarranted because many previous studies may have overestimated associations between RPE and physiological strain. The results of the
present study suggest that, while RPE-HR associations may be reliable over testing
sessions in adolescents (19). RPE does not accuratelv reflect HR in either children or adults. These fihdings are essentially in agriement with the conclusion
of Bar-Or (2), who found high WE-HR associations only in middle-aged and
older adults. Adults in the present study averaged 36 years of age. Using RPE
to prescribe and monitor exercise intensity is therefore a questionable practice,
since at any given RPE, HR may vary by 40 to 100 bpm. In healthy individuals
this degree of error would render RPE useless. In clinical -populations,
basing
exercise intensity on RPE could be dangerous.
The primary conclusion from this study is that RPE is useful for tracking
changes in exercise intensity within a session, but that absolute levels of perceived
exertion are not related to indices of physiological strain. Therefore, using RPE
X 10 as an estimate of heart rate may be dangerously misleading. ~ a t i n of
~s
perceived exertion should not be used as substitutes for heart rate or other physiological measures. The associations between RPE and HR were similar in children and adults.
References
1. Astrand, I. Aerobic work in men and women with special reference to age. Acta
Physiol. Scand. 49 (Suppl. 169): 1-92, 1960.
2. Bar-Or, 0 . Age-related changes in exercise perception. In: Physical Work and Effort, G . Borg (Ed.). Oxford: Pergamon Press, 1977, pp. 255-266.
3. Bar-Or, O., and S.L. Reed. Rating of perceived exertion in adolescents with neuromuscular disease. In: The Perception of Exertion in Physical Work, G . Borg and
D. Ottoson (Eds.). Southampton, U.K. : MacMillan, 1986, pp. 137-148.
4. Bar-Or, O., J.S. Skinner, E.R. Buskirk, and G. Borg. Physiological and perceptual
indicators of physical stress in 40 to 60-year-old men who vary in conditioning level
and in body fat. Med. Sci. Sports 4:96-100, 1972.
Perceived Exertion and Heart Rate
- 367
5. Birk, T.J., and C.A. Birk. Use of ratings of perceived exertion for exercise prescription. Sports Med. 4:l-8, 1987.
6. Borg, G. Physical Performame and Perceived Exertion. Lund, Sweden: Gleerup, 1962.
7. Borg, G. Perceived exertion as an indicator of somatic stress. Scand. J. Rehab. Med.
2-3:92-98, 1970.
8. Borg, G. Perceived exertion: A note on "history" and methods. Med. Sci. Sports,
5:90-93, 1973.
9. Borg, G., J-G. Karlsson, and L. Ekelund. A comparison between two work tests controlled subjectively and by heart rate. In: Physical Work and Effort, G. Borg (Ed.).
Oxford: Pergamon Press, 1977, pp. 239-254.
10. Borg, G., and H. Linderholm. Perceived exertion and pulse rate during graded exercise in various age groups. Acta Med. Scand. 172: 194-206, 1967.
11. Borg, G., and B. Noble. Perceived exertion. In: Exercise and Sport Science Reviews,
J. Wilmore (Ed.). New York: Academy Press, 1974, pp. 131-153.
12. Burke, E.J., and M.L. Collins. Using perceived exertion for the prescription of exercise
in healthy adults. In: Clinical Sports Medicine, R.C. Cantu (Ed.). Boston: Collarnore
Press, 1983, pp. 93-105.
13. Burke, E.J., and T.J. Keenan. Energy cost, heart rate and perceived exertion during
elementary backstroke. Phys. Sportsmed. 12:55-59, 1984.
14. Carton, R.L., and E.C. Rhodes. A critical review of the literature on rating scales
for perceived exertion. Sports Med. 2:198-222, 1985.
15. Dishman, R.K., R.W. Patton, J. Smith, R. Weinberg, and A. Jackson. Using perceived exertion to prescribe and monitor exercise training heart rate. Inter. J. Sports
Med. 8:208-213, 1987.
16. Eston, R.G., B.L. Davies, and J.G. Williams. Use of perceived effort ratings to control exercise intensity in young healthy adults. Euro. J. Appl. Physiol. 56:222-224,
1987.
17. Grirnby, G., J. Bjure, A. Mattias, B. Ekstrom-Jodal, G. Tibbin, and L. Wilhelmser.
Work capacity and physiological responses to work. Am. J. Cardiol. 30:37-42, 1972.
18. Hellerstein, H.K., and R. Ader. ~ e l a t i o n s hbetween
i~
per cent maximal oxygen uptake (% max V02)and per cent maximal heart rate (% MHR) in normals and cardiacs.
Circulation, 44 (Suppl. 2): 71-76, 1971.
19. Komi, P.U., and S-L. Karppi. Genetic and environmental variation in perceived
exertion and heart rate during bicycle ergometer work. In: Physical Effortand Work,
G . Borg (Ed.). Oxford: Pergamon, 1977, pp. 91-99.
20. Lollgen, H., H-V. Ulmer, R. Gross, G. Wilbert, and G.V. Nieding. Methodological
aspects of perceived exertion rating and its relation to pedalling rate and rotating mass.
Euro. J. Appl. Physiol. 34:205-215, 1975.
21. Lollgen, H., H-V. Ulmer, and G.V. Nieding. Heart rate and perceptual response to
exercise with different pedalling speeds in normal subjects and patients. Euro. J. Appl.
Physiol. 37:297-304, 1977.
22. Miyashita, M., K. Onodera, and I. Tabata. How Borg's RPE-scale has been applied
to the Japanese. In: R e Perception of Exertion in Physical Work, G. Borg and
D. Ottosen (Eds.). Southampton, UK: MacMillan, 1986, pp. 27-34.
23. Nader, P.R., J.F. Sallis, J. Rupp, C. Atkins, T. Patterson, and I. Abramson. San Diego
Family Health Project: Reaching families through the schools. J. School Health,
56:227-231, 1986.
368
- Gillach, Sallis, Buono, Patterson, and Nader
24. Pandolf, K.B. Advances in the study and application of perceived exertion. Exer. Sport
Sci. Rev. 11:118-158,1983.
25. Sargeant, A.J., and C.T. Davies. Perceived exertion during rhythmic exercise involving different muscle masses. J. Hum. Ergonom. 2:3-11,1973.
26. Siconolfi, S.F., E.M. Cullinane, R.A. Carleton, and P.D. Thompson. Assessing
VOzmaxin epidemiological studies: Modification of the Astrand Ryhming test. Med.
Sci. Sports Exer. 14:335-338,1982.
27. Skinner, J.S., G.Borg, and E.R. Buskirk. Physiological and perceptual reactions to
exertion of young men differing in activity and body size. In: Exercise and Fitness,
B.D. Franks (Ed.). Chicago: The Athletic Institute, 1969,pp. 53-66.
28. Skinner, J.S., R. Hustler, V. Bergsteinova, and E.R. Buskirk. The validity and reliability of a rating scale of perceived exertion. Med. Sci. Sports, 5:94-96,1973.
29. Skimer, J.S., R. Hustler, V. Bergsteinova, and E.R. Buskirk. Perception of effort
during different types of exercise and under different environmental conditions. Med.
Sci. Sports, 5:llO-115,1973.
30. Ward, D.S., and 0. Bar-Or. Usefulness of a RPE scale for exercise prescription with
obese youth. Unpublished manuscript, University of South Carolina, 1987.