MEDICINE AND SCIENCE IN SPORTS AND EXERCISE

SYMPOSIUM
Vol. 14, No. 5, pp. 377-381, 1982

Psychophysical bases of perceived exertion

GUNNAR A.V. BORG
Department of Psychology
University of Stockholm
Box 5602
S-114 86 Stockholm, Sweden

ABSTRACT In my opinion perceived exertion is the single best in-

BORG, GUNNAR A.V. Psychophysical bases of perceived exertion.
Med. Sci. Sports Exercise, Vol. 14, No. 5, pp. 377-381, 1982. dicator of the degree of physical strain. The overall per-
There is a great demand for perceptual effort ratings in order to ceived exertion rating integrates various information, in-
better understand man at work. Such ratings are important com- cluding the many signals elicited from the peripheral
plements to behavioral and physiological measurements of physical
performance and work capacity. This is true for both theoretical working muscles and joints, from the central cardiovascular
analysis and application in medicine, human factors, and sports. and respiratory functions, and from the central nervous
Perceptual estimates, obtained by psychophysical ratio-scaling system. All these signals, perceptions, and experiences are
methods, are valid when describing general perceptual variation,
but category methods are more useful in several applied situations integrated into a configuration or “Gestalt’’ of perceived
when differences between individuals are described. A presentation exertion.
is made of ratio-scaling methods, category methods, especially
the Borg Scale for ratings of perceived exertion, and a new method Psychophysical Ratio-Scaling Methods
that combines the category method with ratio properties. Some
of the advantages and disadvantages of the different methods are
discussed in both theoretical-psychophysical and psychophysiol- The need to devise better methods to measure perceptual
ogical frames of reference. intensities led to the development of several so-called
“ratio-scaling methods”’ (13,14). It was the scientists’ goal
PERCEIVED EXERTION, RATIO SCALES, CATEGORY SCALES to develop methods that had the same metric qualities as
methods used in physics and physiology, i.e., methods with
Need of Methods to Quantify Symptoms
an absolute zero and with the same distance between all
scale values. One such method is “ratio production,” in
During recent decades we have become more interested
which subjects are asked to increase or decrease a certain
in how people feel, what aches and pains they have, and
variable stimulus until it is perceived to be a certain fraction
how difficult they perceive their work to be. Most scientists
or multiple of a standard stimulus. For example, in an
and practitioners in the health sciences agree that it is
experiment concerning brightness, taste, or loudness, each
important to understand subjective symptoms and how
subject is asked to set the variable stimulus to be half or
they relate to objective findings. Therefore, we must de-
twice as intense as a certain standard stimulus. When this
velop methods to quantify these subjective symptoms.
is performed throughout the whole intensity continuum,
These methods should be equally applicable to most people
regardless of gender, age, circumstances, and national
possibilities can be obtained to construct a psychophysical
origin.
scale, i.e., a scale describing how the perceived intensity
varies with the actual physical intensity.
Perceived Exertion as an Indicator of Physical Strain A simple example taken from everyday experiences
demonstrating the ratio-production method involves one’s
The individual’s perception of exertion during physical perception of speed while driving a car. The subject drives
work is interesting when studying man at work or leisure- 50 miles/h and then is instructed to decrease the car’s
time activities, in diagnostic situations or exercise pre- speed until he/she perceives it to be half as fast. The actual
scriptions, and in epidemiological evaluation of daily ex- speed is really about 85 miles/h—very much above the
ercise intensities. People primarily seek medical care be- physical half. This is in accordance with a power function
cause they feel ill, not for treatment of a specific disease. with an exponent of about 2, that is, the subjective speed
Medical assistance is most frequently sought by patients grows to the approximate square of the physical speed
who have noted a severe decrease of their physical working (2).
capacity and a subsequent subjective strain. One of the most popular ratio-scaling methods is “mag-
nitude estimation” (14), With this method the subjects are
Submitted for publication August, 1981.
presented stimuli of different intensities and are asked to
Accepted for publication June, 1982. assign numbers to them, depending upon how intensely

377
878 MEDICINE AND SCIENCE IN SPORTS AND EXERCISE

they are perceived. The subjects are most often asked to subjective intensity is more “intense” than another; how-
match numbers to the perceived physical stimulus in such ever, neither the degree of intensity nor the position of
a way that the numbers vary directly in proportion to zero intensity can be accurately determined. An advantage
subjective intensities. of a category scale is that it is possible for direct inter-
individual comparisons to arise because the subjects react
Power Functions to the stimuli in a more “absolute” way. Consequently,
we can be rather certain that if the subject says a weight
Most psychophysical relations may be described by is “‘light,” it will seem relatively lighter to that person
power function, with exponents ranging from 0.3 to 3, than to someone who says it is “heavy.”
depending mainly upon the modality tested but also de-
pending on experimental conditions (9,15). The first studies Perceived Exertion—Heart Rates and Lactate
of perceived exertion in heavy physical work were per- Accumulation
formed in the end of the 1950’s by Borg and Dahlstrém
(6,7) and in the 1960's by Borg (2). Even in this “modality,” High correlation coefficients between ratings of per-
power functions described the perceptual variation with ceived exertion and heart rates indicate the differential
the physical intensity. The exponent was about 1.6, very value of the scale but not the general validity of the growth
close to that found for subjective force of handgrip, viz., function. Because the perceived exertion determined by
1.7 (12). ratio-scaling methods grew with an exponent of about 1.6,
It has been possible in most psychophysical studies to it was concluded by Borg (2) that an integration of central
prove that equal stimulus ratios produce equal response factors, such as heart rates, and peripheral factors (such
ratios (15). Thus, the Fechnerian log-function provides a as blood lactates, with an exponent of about 2) would
fundamentally inaccurate description of the general stim- better “explain” the psychophysical variation than any
ulus-response function, which instead should be repre- single physiological variable.
sented in the form of a power function. The most general
Borg’s RPE Scale
form of this equation that fits both psychophysical and
physiological functions was originally proposed by Borg
A new category scale for ratings of perceived exertion
(1):
was constructed by Borg (3) to increase linearly with the
R = a + c(S—b)",
exercise intensity for work on cycle ergometer. Because
where R is the intensity of the response, S is the stimulus
oxygen consumption and heart rate increase linearly with
intensity, a and b are constants showing the starting point
work load, this would be a convenient means of con-
of the function, c is the proportionality constant, and n
structing a scale, even if it did violate the true growth of
is the exponent.
the perceived intensities. The RPE scale (Table 1) has
become very popular and has been translated into many
Difficulties in Interindividual Comparisons
different languages, including French, German, Japanese,
One major drawback with ratio-scaling methods is that Hebrew, and Russian. In many studies, correlations of
they do not provide any direct “levels” for interindividual ratings and heart rates ranging from 0.80-0.90 have been
comparisons. Good general functions for a group of subjects found, but high correlations with other physiological var-
can be obtained, but it is difficult to compare the subjects iables (8,10) have also been found.
with each other because subjects are asked only to make The scale values range from 6 to 20 and can be used
relative comparisons. One subject may rate a 1-pound to denote heart rates ranging from 60-200 beats-min™'.
weight a “10” and a 2-pound weight “25,” while another
may assign “4” and “10” to the same weights. However,
the subject assigning the “25” rating to the 2-pound weight Table 1. The 15-grade scale for ratings of perceived exertion, the RPE Scale. (3)
does not mean that he perceives it to be heavier than the 6
subject who has rated it “10.” 7 Very, very fight
To overcome the difficulties associated with the ratio- 8
9 Very light
scaling methods, a scale for ratings of perceived exertion 10
was developed by Borg (2). The first was a 21-grade scale ll Fairly light
with verbal anchors similar to those used in a subsequent 12
13 Somewhat hard
RPE scale. It represents one kind of category scale, often 14
used in many applied situations when a simple but direct 15 Hard
estimation of the subjective intensity is needed, but when 16
17 Very hard
the metric properties of the scale are of less importance. 18
Some category scales may be interval scales, but most of 19 Very, very hard
them are rank-order scales. It may only be stated that one 20
 PSYCHOPHYSICAL BASES 379

This was intended to make the scale easier to use because interindividual meaning. Category expressions denoting
a certain value on the scale, e.g., 13, would match ap- subjective intensities of heaviness in everyday experiences
proximately a heart rate of 180 beats. min™=! for 30-50- often refer to a special class of objects, such as suitcases,
year-old subjects. However, this close relationship was not rackets, or glassware. If someone believes that plastic glasses
intended to be taken too literally because the meaning of are ‘very light,” it may be interpreted that the perceived
a certain heart rate value as an indicator of strain depends heaviness underlying the expression is a certain distance
upon age, type of exercise, environment, anxiety, and other (which might be identified in relative z-scores) below the
factors. The advantage of not having to refer to a table mean in his sample of experiences, which stands in a
to interpret the meaning of a rating value has been great certain relation to another person’s experiences and to a
and has overshadowed the disadvantages of the scale. given population. If we exchange glasses with skis, the
With popular category scales the term “moderate” (or “absolute” subjective intensity will be much higher, but
a similar expression) is placed in the middle of the scale, the relative position in each sample may be the same.
and the terms “‘strong” and “weak” with the addition of
“rather” or “very” are placed symmetrically on each side The Range Model
of ‘““moderate.”” This does not create an interval scale, for
when comparing the responses, R,, Ro, Rg. . .R,...R,, it In a laboratory setting subjects are tested with more
cannot be said that the distance from R, to Rg is equal to “meaningless” stimuli, that is, specific objects or classes
that from R, to Ry. Because the RPE scale is, on the other of objects are not referred to as in the example above, the
hand, constructed to increase linearly with the exercise cognitive evaluation of the subjective intensity will depend
intensity, it may be stated that R,-R, = R3-R,, or when to a much higher degree upon the “‘absolute’”’ perceptual
stated in RPE units, 17-15 = 13-11 as with equal right intensity and its position in the total spectrum, which
heart rate (in beats-min—!) 170-150 = 130-110. ranges from an extremely weak intensity (e.g., R, is equal
to the threshold [Rj] or somewhat above that) to an ex-
“Experiential Value” of Category Expressions tremely strong intensity (e.g., R, = R,, i.e. a terminal [t]
value equal to a maximal intensity or somewhat below
It is true from a simple mathematical point of view that).
that a ratio scale is more accurate than an interval scale The verbal expressions used to report perceived exertion
(with equal distances but without a true zero), which in in treadmill running or ergometry cycling thus depend
turn is better than a rank-order scale. With a ratio scale more on sensory signals than the cognitive frame of ref-
all types of mathematical calculations can be performed erence. The physical stress causes a certain strain for each
and direct comparisons with physical and physiological individual depending upon his or her capacity relative to
measurements, obtained with true ratio scales, can be the position of the intensity in each individual’s “absolute”
made. However, the meaning of a certain value is not range. This simple “range-model” indicates a fundamental
only dependent upon where it belongs ona ratio scale, it principle for interprocess comparisons (1,2,4). It must be
is also given by the context in question. As an example, modified when comparing interindividual responses and
when I visited the United States for the first time in the when comparing responses from different situations. As a
1960's, someone talked with me about a certain athlete result, there may be some differences between an exercise
who was 6 ft tall. This description did not mean much to intensity based on that perceived by the subjects and based
me. I understood, of course, that he was taller than someone on heart rate data depending upon the kind of exercise.
5 ft tall and was shorter than someone 7 ft tall, but I did The relation between RPE and heart rate may change
not know if he was considered to be “tall” or “‘short.” somewhat from walking to running.
Similarly, an American unaccustomed to the metric system
will encounter the same problem when he is told that Use of Perceived Exertion in Exercise Prescriptions
someone is 1.90 m tall. It will be understood that his height
is 1.9 times greater than the true meter unit in Europe, Perceived exertion is often taken into account for ex-
but not whether the person is “‘tall’’ or “short.”” However, ercise prescriptions. However, those skeptical about the
if it is also noted that the person is two standard deviations valid application of perceived exertion often refer to the
above the mean in the Swedish male population, new discrepancies found when comparing RPE values and heart
information is obtained permitting relative comparisons rate data to illustrate difficulties encountered when it is
of individuals. Similarly, an “additional meaning” may used in clinical rehabilitation. For example, a patient
also be obtained when using some category scales. training with a “target heart rate” of 130 beats-min™',
Despite the fact that these scale values do not have the which can be achieved by walking strenuously, may rate
good ratio properties, we may have the advantage of scale the exertion as “hard” with an RPE of 15; however, his
anchors in certain “population norms” or “experiential heart rate may reach 150 beats: min—! when running, but
values.” Category expressions like “comfortable intensity he may perceive the exertion to be the same. To those
level” and “preferred radio volume” thus provide a good who object to the use of the RPE scale, it may be explained
380 MEDICINE AND SCIENCE IN SPORTS AND EXERCISE

that the discrepancy is understood and may be taken into Table 2. The new rating scale constructed as a category scale with ratio properties.
(5)
consideration when instructing the patient.
In defense of the use of perceived exertion it may be Nothing at all

OO
Very, very weak (just noticeable)

wn
said that there is little evidence that a certain heart rate Very weak
is a better indicator of “dangerous strain” than a certain Weak (light)

OO
ONS WH
Moderate
perceived exertion. On any given day one may run and Somewhat strong
achieve a heart rate of 150 and feel “fine” with an RPE Strong (heavy)
of 13, while on another day the same exertion may cause
the runner to feel “bad” with an RPE of 17 as a result Very strong

of physical and emotional negative factors.

The elevated RPE value may be used equally with heart

OW
Very, very strong (almost max)

—
rate in determining a “risk factor.” Neither a single RPE Maximal
value nor a heart rate measure may be used alone as an
accurate indicator of ‘dangerous strain.” They comple-
ment each other.
A “perfect” or “excellent” indicator of “dangerous technical terminology, we should use a simple number
strain” must involve an integration of all important risk range, e.g., 0-10. We have chosen to denote 10 as “very,
factors, such as arrhythmias, blood pressure elevations, ST- very strong” or “very, very heavy” for the heaviest exercise
depressions, body temperature changes, blood lactate lev- or physical work perceived by the subject, e.g., lifting
els, and hormonal excretions. A single heart rate must be weights or running. To be able to cover a reasonably good
used in relation to the other strain variables and understood range of intensities, the opposite end of the scale, identified
to be just one factor in a complicated pattern of interacting as a “very, very weak” intensity or a “very, very light”
factors, A patient’s perceived exertion is considered in work, had to be set to 0.5. The different verbal expressions
exercise prescription because it is related closely to the were then placed where they belonged according to their
heart rate but it also integrates some other important strain ratio properties. When using this scale, people are per-
variables. mitted to use decimals and also to go beyond 10 and 0.5.
The scale is presented in Table 2.
A New Category Scale with Ratio Properties Recent studies by Borg (5) of this new category scale
with ratio properties have given psychophysical functions
In recent years we have tried to develop a category of about the same appearance as those obtained with mag-
scale with ratio properties, that is, a simple category scale nitude estimation. Exponents of about 1.6 have thus been
for differential use that has the positive attributes of a obtained for perceived exertion in cycle ergometer exercise.
general-ratio scale. In developing such a scale we have In another psychophysiological study (11) a close corre-
referred to our previous studies of perceived exertion and lation between ratings according to this new scale and
the known relationship between the RPE scale and a ratio both blood lactate and muscle lactate levels were obtained.
scale. We have also referred to studies of quantitative
semantics, “range problems,” and psychophysical rela- When to Use a Certain Scale
tionships (5).
The main idea is that numbers should be anchored by There may not be one perfect scale for all kinds of
verbal expressions that are simple and understandable by subjective intensities in all kinds of situations. Perhaps we
most people. The expressions should in turn be placed in should use different scales depending upon the purpose
the correct position on a ratio scale, where the expressions of the study. At present I think, however, that the old
belong according to their quantitative meaning. If, for RPE scale is the best one for most simple applied studies
instance, we determine 4 to represent “light,” 2 should of perceived exertion, for exercise testing, and for pre-
represent half of that intensity. We then have to choose dictions and prescriptions of exercise intensities in sports
an expression with an inherent meaning of being about and medical rehabilitation. The new category scale with
half as intense as “light,” such as “very light.” ratio properties may be especially suitable for determining
To make the scale easy to use for the lay population other subjective symptoms, such as breathing difficulties,
and not restricted to those familiar with mathematical or aches, and pain.

REFERENCES
1. Bore, G. Interindividual scaling and perception of muscular force. 8. Bore, G. Perceived exertion as an indicator of somatic stress. Scand.
Kungl. Fysiol. Sallsk. forh. 31(12):117-125, 1961. J. Rehab. Med. 2:92-98, 1970.
2. Borc, G. Physical Performance and Perceived Exertion. Lund, 4. Bore, G. Subjective effort in relation to physical performance and
Sweden: Gleerup, 1962, pp. 1-63. working capacity. In: Psychology: from Research to Practice, H.L.
PSYCHOPHYSICAL BASES 381

Pick et al. (Eds.). New York: Plenum Publishing Corp., 1978, pp. 10. MIHEVIC, P.M. Sensory cues for perceived exertion: a review. Med.
383-361. Sci. Sports Exercise, 18:150-163, 1981.
. Bore, G. A category scale with ratio properties for intermodal and 11. NOBLE, B., G. Bore, and I. Jacoss. Validation of a category-ratio
interindividual comparisons. In: Proceedings of the 22nd Interna- perceived exertion scale: blood and muscle lactates and fiber types.
tional Congress of Psychology. Leipzig: VEB Deutscher Verlag, Paper presented at the PanAmerican Congress and International
1980. (In press) . Course on Sports Medicine and Exercise Science, Miami, 1981. Int.
. Borc, G. and H. DAHLstROM. Psykofysisk undersékning av arbete J. Sports Med. 2:279, 1981. (Abstract)
pi cykelergometer. Nordisk Medicin 62:1383-1386, 1959. 12. STEVENS, J.C. and J.D. MACH. Scales of apparent force. J. Exp.
. Borc, G. and H. DAHLsTROM. The perception of muscular work. Psychol. 58:405-418, 1959.
Umea veten. skapliga skriftserie 5:1-26, 1960. 13. STEVENS, S.S. On the psychophysical law. Psychol. Rev. 64:158-181,
. Bors, G. and B. NOBLE. Perceived exertion. In: Exercise and Sports 1957.
Sciences Review, J.H. Wilmore (Ed.). New York: Academic Press, 14. STEVENS, S.S. Matching functions between loudness and ten other
1974, pp. 181-158. continua. Perception and Psychophysics 1:5-8, 1966.
. MARKS, L. Sensory Processes: The New Psychophysics. New York: 15. STEVENS, S.S. Issues in psychophysical measurement. Psychol. Rev.
Academic Press, 1974, pp. 1-289. 78:426-450, 1971.