original research article

Validity and reliability of the session RPE method for

monitoring exercise training intensity
L Herman (MS)1
C Foster (PhD, FACSM)1
M A Maher (PhD)2
R P Mikat (PhD, FACSM)1
J P Porcari (PhD, FACSM)1
1
Department of Exercise and Sport Science, University of Wisconsin-La Crosse, USA
2
Department of Biology, University of Wisconsin-La Crosse, USA

Abstract 4.3) and hard (81 v. 81%, 94 v. 94%, 91 v. 91% and 7.3 v.
7.4) exercise bouts. Correlations between repeated bouts
Objective. The Session Rating of Perceived Extertion for %VO2peak (r = 0.98), %HRpeak (r = 0.98), %HRreserve (r
(RPE) is a method of measuring exercise intensity that = 0.98) and Session RPE (r = 0.88) were significant and
may be useful for the quantitative assessment of exercise strong.
training programmes. However, there are inadequate data
regarding the validity and reliability of the Session RPE Conclusions. The results support the validity and reli-
method. This study was designed to evaluate both the ability of the Session RPE method of monitoring exercise
validity and reliability of the Session RPE method in com- intensity, although as might be predicted for a subjective
parison to objective measures (%HRpeak, %HRreserve and method the Session RPE was less precise than the objec-
%VO2peak) of exercise intensity. tive measures of exercise training intensity.

Methods. Fourteen healthy volunteers (7 male, 7 female)

performed 6 randomly ordered 30-minute constant-load
Introduction
exercise bouts at 3 different intensities, with each inten- The Session Rating of Perceived Exertion (RPE) method
sity being repeated. Oxygen consumption (VO2) and heart of monitoring exercise intensity was developed in 1995
10

rate (HR) were measured throughout each exercise bout as a modification of the Category Ratio (CR) RPE meth-
and normalised to maximal values obtained during a pre- od.
2,3,19,20,22
The modification involved asking the subject to
liminary maximal exercise test. Thirty minutes following give a global rating of the perceived exertion for the entire
the conclusion of each exercise bout, the subject rated exercise session ~30 minutes after the conclusion of an ex-
the global intensity of the bout using a modification of the ercise bout rather than rating the momentary level of exertion
Category Ratio (CR) (0 - 10) RPE scale. This rating was as is the usual practice with RPE.
2,3,19,20,22
At that time, lim-
compared to the mean value of objectively measured ex- ited objective data were presented in support of the validity
ercise intensity across the duration of the bout. of the technique, on the basis of both heart rate (HR) and
Results. There were significant non-linear relationships blood lactate responses. In the subsequent decade, several
2 papers have used the Session RPE method in a variety of
between Session RPE and %VO2peak (R = 0.76), %HRpeak
2 2 settings and have demonstrated its value relative to quan-
(R = 0.74) and %HRreserve (R = 0.71). There were no
4,7,8,9,16,17,23
significant differences between test and retest values of tifying exercise training intensities and as the
%VO2peak, %HRpeak, %HRreserve and Session RPE during intensity component of larger schemes of evaluating train-
6,7,10,11,13,21
the easy (47 v. 47%, 65 v. 66%, 47 v. 48% and 2.0 v. 1.9), ing programmes. Most notably, the Session RPE
moderate (69 v. 70%, 83 v. 84%, 74 v. 75%, and 4.2 v. method has been used to demonstrate inconsistencies be-
tween the training programmes designed by coaches and
11
Correspondence: executed by athletes, thus providing a plausible explana-
tion for the incidence of overtraining syndrome in high-level
C Foster 6
athletes. However, the validity of the Session RPE method
Department of Exercise and Sport Science has not been systematically tested against accepted objec-
University of Wisconsin-La Crosse tive methods of measuring exercise training intensity, such
La Crosse, WI 54601 as %VO2peak, %HRpeak, and %HRreserve; nor has the reliability
Tel: 54601608-785 8687 of the method been demonstrated. Accordingly, the purpose
Fax: 54601608-785 8172 of this study was to evaluate the validity and reliability of the
E-mail: [email protected] Session RPE method.

14 SAJSM vol 18 No. 1 2006

validity and reliability.indd 14 3/13/06 2:59:33 PM

 Methods TABLE II. Modification of the Rating of Perceived Ex-
ertion Scale. Subjects rated the entire exercise ses-
The subjects were 14 physically active volunteers (7 males, sion 30 minutes after exercise in response to the ver-
7 females). All subjects completed a health screening ques- bal prompt ‘How was your workout?’
1
tionnaire which revealed no evidence that exercise testing Rating Verbal anchor
or training should be limited. The subjects provided informed 0 Rest
consent and the protocol for this study was approved by the 1 Very easy
Institutional Review Board for the Protection of Human Sub- 2 Easy
jects at the University of Wisconsin-La Crosse. Characteris- 3 Moderate
tics of the subjects are presented in Table I. Since there was 4 Sort of hard
no expectation of a gender-related effect on the relationship 5 Hard
between the Session RPE and objective markers of exercise 6
intensity, the results of all subjects were pooled into a single 7 Very hard
group for analysis. 8 Very, very hard
9 Near maximal
All subjects completed 7 different exercise sessions with
10 Maximal
a minimum of 2 days of rest between sessions. The first
session was an incremental test to fatigue to determine
maximal exercise responses for HR and peak oxygen uptake the RPE scale, the only verbal prompting was ‘How was your
(VO2peak) (highest 30-second VO2 observed during the test). workout?’
Each subject completed all of their exercise sessions on Statistical comparisons were made using repeated
either a cycle ergometer or treadmill, depending on their measures analysis of variance (ANOVA) for an intensity x
normal exercise habits. During the cycle ergometer protocol, trials design. Statistical significance was accepted when p
the starting power output was 25 W and was increased by < 0.05. Post hoc comparisons, when justified by ANOVA,
25 W every 2 minutes. During the treadmill protocol, the belt were performed using the Tukey test. Curve fitting and
-1
velocity was set at either 5.6 or 8.1 kph (1.56 or 2.23 m.s ), computation of correlation coefficients and/or coefficients of
depending on whether the subject regularly walked or ran determination were made using a least squares technique.
for exercise. The grade was 0% initially, and was increased
by 2% every 2 minutes. The next 6, randomly ordered, Results
sessions, were performed using the same mode of exercise. The tests were completed by all subjects without complica-
Each was 30 minutes in duration, was of constant intensity tions, and all subjects completed all tests. The mean serial
within the trial, and varied in intensity between trials. There responses of VO2, HR and RPE are presented in Fig. 1. The
were 3 different intensity levels, designed to correspond to low-intensity exercise bout satisfied criteria for a steady state
easy effort (~40 - 50% VO2peak), moderate effort (~60 - 70% of VO2, whereas the moderate and hard exercise bouts dis-
VO2peak), and hard effort (~80 - 90%VO2peak). Each subject played evidence of a slow component of VO2. All outcome
completed each intensity session twice. During every measures differed (p < 0.05) between the 3 intensity levels.
exercise session, respiratory gas exchange was measured There was no significant difference between exercise inten-
using open-circuit spirometry (Quinton Q-MC, Seattle, WA) sity, measured by %VO2peak, %HRpeak, %HRreserve and Ses-
and HR was measured using radiotelemetry (Polar Electo- sion RPE during repeat trials at the same exercise intensity
Oy, Finland). The RPE was assessed every 5 minutes bout (Fig. 2).
throughout each exercise bout using the CR scale (Table II).
The Session RPE estimate of exercise intensity was
Thirty minutes after completing the exercise session (after
also shown to be reliable using regression analysis. The
having time to shower and change clothes), each subject 2
relationships between day 1 and day 2 for Session RPE (R
rated the perceived intensity of the entire exercise session
= 0.78, standard error of estimate (SEE) = 1.2), %VO2peak
according to the Session RPE scale (Table II), which is a 2 2
(R = 0.96, SEE = 3.2), %HRpeak (R = 0.93, SEE = 3.7), and
modification of the conventional CR RPE scale, with verbal 2
%HRreserve (R = 0.93, SEE = 5.7) were statistically significant,
anchors changed to represent terms that are more relevant
and had a small SEE (Fig. 3).
to exercisers fluent in American idiomatic English. In place
of the usual specific verbal instructions normally used with The Session RPE method was compared with %VO2peak,
%HRpeak and %HRreserve to gain an appreciation of the degree
TABLE I. Mean (± standard deviation) characteristics to which the various methods of exercise intensity were
of the subjects 2
measuring the same thing. The R between the Session RPE
Men Women 2 2
and %VO2peak (R = 0.76), %HRpeak (R = 0.74), and %HRreserve
Age (years) 33 ± 16 23 ± 1 2 2
(R = 0.71) was strong, but not comparable to the R amongst
Height (cm) 180 ± 9 165 ± 7 the various objective measures of exercise intensity (Fig. 4).
Body mass (kg) 93 ± 17 58 ± 5 There was apparently a much weaker relationship between
-1
VO2peak (ml.kg ) 41.5 ± 9.9 44.9 ± 5.9 the Session RPE and the objective measures of exercise
-1
HRpeak (beats.min ) 170 ± 19 186 ± 22
intensity at the higher exercise intensities (Fig. 4).

SAJSM vol 18 No. 1 2006 15

validity and reliability.indd 15 3/13/06 2:59:33 PM

 Fig. 3. Scatter plots of individual responses of Session
RPE, %VO2peak, %HRpeak and %HRreserve during Trial 1 and
Trial 2, with all three intensities combined.

Discussion
The Session RPE has already been shown to be of value in
terms of evaluating exercise intensity and monitoring training
6,7,10,11,13,21
in a number of situations. The primary outcome of
this study is the demonstration that the Session RPE method
is not only reliable during repeat challenge by the same exer-
Fig. 1. Serial responses of %VO2max, %HRmax and RPE
cise stimulus, but is well related to widely accepted objective
during the course of the easy, medium and hard exercise
measures of exercise training intensity. As such, we interpret
bouts. The two trials at each intensity were combined to
the present data as supporting the validity of the Session
produce the figures.
RPE method as an alternative method of monitoring exer-
cise training. The present data are comparable with previ-
2,3,19,20,22
ous studies which have shown RPE to be a valid
and reliable measure of momentary exercise intensity. We
chose not to measure blood lactate responses during either
incremental or constant intensity exercise, which is the other
widely accepted objective method of measuring exercise in-
tensity. This decision was based on practical concerns rather
than any particular limitation of using lactate as a monitoring
tool. Other studies demonstrating session RPE and lactate
responses during similar exercise bouts would be of interest,
and would provide a test of our early results with the Session
10
RPE method compared with blood lactate responses.
In this study, we used constant-load exercise bouts as
the method for providing the exercise challenge. This was
done because it was much more convenient to represent
the exercise intensity using objective methods during
steady-state exercise. Previous studies from our laboratory
have compared the Session RPE method with HR-based
measures of exercise intensity during non-steady state
8,9
Fig. 2. Mean responses of %VO2peak, %HRmax, %HRreserve exercise, with substantially similar findings to the present
and Session RPE during Trial 1 and Trial 2 of the three data. Thus, we believe that the collective experience suggests
exercise intensity bouts. that the Session RPE method is a valid marker of exercise
intensity under a variety of situations. Other studies from our
laboratory have demonstrated that the ‘drift’ in Session RPE

16 SAJSM vol 18 No. 1 2006

validity and reliability.indd 16 3/13/06 2:59:34 PM

 exercise intensity. This raises the more fundamental question
implicit in the TRIMP concept related to the signal for
adaptations to exercise. Clearly, exercise intensity changes
the milieu internal in ways that are different from exercise
duration. Until this signalling mechanism is elucidated, it may
not be critical whether one or the other of the comparatively
crude subjective methods of estimating exercise intensity is
used.

References
1. Balady GJ, Chaitman B, Driscoll D, et al. Recommendations for cardiovascular
screening, staffing and emergency policies at health/fitness facilities: A Joint
Position Statement by the American College of Sports Medicine and the Ameri-
can Heart Association. Med Sci Sports Exerc 1998; 30: 1009-18.
2. Borg G, Hassman P, Langerstrom M. Perceived exertion in relation to heart rate
and blood lactate during arm and leg exercise. Eur J Appl Physiol 1987; 65:
679-85.
3. Borg, G, Ljunggren G, Ceci R. The increase of perceived exertion, aches
and pains in the legs, heart rate and blood lactate during exercise on a bicycle
Fig. 4. Scatter plots of individual responses of Session ergometer. Eur J Appl Physiol 1985; 54: 343-9.

RPE v. %VO2peak, %HRpeak, and %HRreserve and of %VO2peak 4. Day ML, McGuigan MR, Brice GA, Foster C. Monitoring work intensities during
resistance training using a session RPE scale. Journal of Strength and Condi-
v. %HRpeak and %HRreserve and of %HRpeak v. %HRreserve. tioning Research 2004; 18: 353-8.
These plots demonstrate that although Session RPE is 5. Fitz-Clarke JR, Morton RH, Banister EW. Optimizing athletic performance by
influence curves. J Appl Physiol 1991; 71: 1151-8.
significantly related to objective markers of exercise in-
6. Foster C. Monitoring training in athletes with reference to overtraining syndrome.
tensity, it is less well related to the objective markers of Med Sci Sports Exerc 1997; 30: 1164-8.
exercise intensity than the objective markers are to each 7. Foster C, Daines E, Hector L, Snyder A, Welsh R. Athletic performance in rela-
other. tion to training load. Wisconsin Medical Journal 1996; 95: 370-4.
8. Foster C, Florhaug J, Franklin J, et al. A new approach to monitoring exercise
training. Journal of Strength and Conditioning Research 2001; 15: 109-15.
with progressively longer performance at a fixed exercise
9. Foster C, Florhaug J, Hrovatin L. Monitoring of athletic training. Vlaams Tijd-
intensity is comparable to the drift in HR during prolonged schrift voor Sportgeneeskunde en Sportwetenshappen 1999; 80: 47-54.
8,9
exercise. 10. Foster C, Hector L, Welsh R, Schrager M, Green M, Snyder A. Effects of spe-
cific versus cross-training on running performance. Eur J Appl Physiol 1995; 70:
We have not tested the Session RPE under very prolonged 367-72.
and exhaustive exercise such as, for example, marathon 11. Foster C, Heimann K, Esten P, Brice G, Porcari J. Differences in perceptions
running. In a situation of competitive performance (or even of training by coaches and athletes. South African Journal of Sports Medicine
2001; 8: 3-7.
severe training bouts) of a very prolonged duration, the
12. Foster C, Hoyos J, Earnest C, Lucia A. Regulation of energy expenditure during
Session RPE would approach maximal values, even though prolonged athletic competition. Med Sci Sports Exerc 2005; 37: 670-5.
the objective exercise intensity would remain within a clearly 13. Impellizzeri FM, Rampinni E, Coutts AJ, Sassi A, Marcora SM. Use of RPE-
based training load in soccer. Med Sci Sports Exerc 2004; 36: 1042-7.
submaximal zone. This problem remains to be explored in
14. Lucia A, Hoyos J, Carvajal A, Chicharro JL. Heart rate response to professional
that it presents a fundamental problem in terms of monitoring road cycling: The Tour de France. Int J Sports Med 1999; 20: 167-72.
exercise training. However, as this is also a limitation with 15. Lucia A, Hoyos J, Santally A, Earnest C, Chicharro JL. Tour de France vs Vuelta
HR-based methods of monitoring exercise intensity, this a Espana: Which is harder? Med Sci Sports Exerc 2003; 35: 872-8.

drift does not represent a limitation unique to the Session 16. McGuigan MR, Egan AD, Foster C. Salivary cortisol responses and perceived
exertion during high intensity and low intensity bouts of resistance exercise.
RPE method. Our underlying assumptions in studies using Journal of Sport Science and Medicine 2004; 3: 8-15.
the Session RPE method as the intensity component of 17. McGuigan MR, Foster C. A new approach to monitoring resistance training.
Strength and Conditioning Journal 2004; 26: 42-7.
monitoring training has been to use the Session RPE to
18. Morton RH, Fitz-Clarke JR, Banister EW. Modeling human performance in run-
replace objective measures of exercise intensity within the ning. J Appl Physiol 1990; 69: 1171-7.
training impulse (TRIMP) approach originally developed 19. Noble B, Robertson R. The Borg scale: development, administration, and exper-
5 18
by Fitz-Clarke et al. and Morton et al. and widely used imental use. In: Washburn R, Mittelmeiner K, eds. Perceived Exertion. Cham-
6,7,12,14,15,21 paign, Ill: Human Kinetics Publishers, 1996: 59-92.
by others. The TRIMP approach is a very useful
20. Pandolf K. Advances in the study and application of perceived exertion. Exerc
way of combining exercise intensity and duration into a Sport Sci Rev 1983; 11: 118-58.
single number representation of the stimulus for adaptation 21. Seiler KS, Kjerland GO. Quantifying training intensity distribution in elite endur-
ance athletes: Is there evidence for an ‘optimal’ distribution. Scandinavian Jour-
provided by any exercise bout, which has been limited by the nal of Medicine and Science in Sports (in press).
absence of a simple method of expressing exercise intensity. 22. Skinner J, Hutsler R, Bergsteinova V, Buskirk E. The validity and reliability of a
During very prolonged exercise, the drift of Session RPE (or rating scale of perceived exertion. Med Sci Sports 1973; 5: 94-6.
of HR) acts as a multiplier of exercise duration and suggests 23. Sweet TW, Foster C, McGuigan MR, Brice G. Quantitation of resistance training
using the session RPE method. Journal of Strength and Conditioning Research
a much larger TRIMP than might have been calculated using 2004; 18: 796-802.
the RPE or HR early during the exercise bout as a marker of

SAJSM vol 18 No. 1 2006 17

validity and reliability.indd 17 3/13/06 2:59:34 PM