Journal Logo


Comparisons of RPE before, during, and after Self-Regulated Aerobic Exercise


Author Information
Medicine & Science in Sports & Exercise: March 2009 - Volume 41 - Issue 3 - p 682-687
doi: 10.1249/MSS.0b013e31818a0f09
  • Free


Ratings of perceived exertion (RPE) are widely used to assess effort sense associated with exercise. Laboratory use of RPE is predominantly used to capture perceptions of effort during incremental exercise tests. Clinicians and fitness professionals may also use RPE as an adjunct to more objective physiological measures in the prescription of exercise. These primary uses of RPE are consistent with the original concepts of RPE as a measure of momentary effort sense described by Borg (3). However, it is possible that measurement of effort sense before and/or after an acute exercise bout may provide use to clinicians and fitness professionals aiming to improve physical activity adoption and maintenance.

Measurement of predicted effort before exercise is an area that has not received attention in the scientific literature. The absence of this line of inquiry is significant given the possibility that perceptions of pending effort required for exercise may impact behavioral choices. One area of study that has considered the impact of anticipated perceptions on behavior is pain and fear research using a match-mismatch paradigm (15). This literature suggests that a tendency exists to poorly predict pain intensity, both through underprediction and overprediction when the painful stimuli is novel and that congruence between predicted and actual perceptions occurs only after numerous exposures to the stimuli (16). Moreover, the literature suggests the tendency for most inaccurate predictions to be overpredictions because such predictions allow for protection against danger or damage (2). Overpredictions of pain that are severe or rigid result in avoidance behaviors that are often maladaptive (2,14). Given the presence of such a phenomenon in pain experience, it is possible that similar experiences may be present with respect to predicted perceptions of exertion associated with exercise. The presence of this type of perceptual match-mismatch in exertion may well represent a barrier to regular physical activity participation, especially in individuals unaccustomed to exercise.

Although measurement of effort sense before exercise has not been studied, the assessment of exertion after the completion of exercise has been investigated. This assessment is commonly called a session RPE and was developed to quantify the global exertional experience during an entire session (6,7). Such an application of RPE departs from the more common use designed to assess exertional perceptions during an acute bout of exercise. Research has demonstrated that the session RPE is a valid method of assessing effort sense associated with a completed bout of aerobic and resistance exercise in both general and athletic populations (4,5). Furthermore, research suggests that the session RPE can be useful in subsequently quantifying exercise intensity during intermittent bouts of aerobic exercise (10).

Although investigations thus far suggest that session RPE can be an effective research and clinical tool, research to date has not assessed perceptions before, during, and after aerobic exercise at varying intensities associated with public health recommendations. Current guidelines for physical activity encourage individuals to participate in aerobic exercise for 20-30 min·d−1 across intensities ranging from light to moderate to vigorous (12). Furthermore, existing research has not compared exertional responses at time points before, during, and after bouts of continuous aerobic exercise. The present experiment was designed to fill this void. Therefore, this study compared RPE measured before, during, and after three different bouts of exercise performed at different self-selected intensities on a treadmill. By comparing the pattern of exertional responses, the present design will indicate if a perceptual match-mismatch is observed during aerobic exercise bouts that vary on intensity.


Participants and research design

Participants were 26 unpaid faculty, staff, and students (10 men, 16 women, mean age ± SD = 25.8 ± 6.2 yr) at a university in the southeastern United States. Participants were diverse in terms of fitness status and were recruited through email solicitation. The participants completed four trials separated by at least 48 h. The first trial was a maximal protocol to measure maximal oxygen uptake (V˙O2max). The remaining trials were submaximal and 30 min in duration at self-selected intensities according to instructions to exercise at "light," "moderate," and "vigorous" loads. All participants completed the sessions in order from least to most intense. Exposure to the trials of exercise in ascending order affords some measure of ecological validity because similar methods are used in practice by fitness enthusiasts and professionals. Additionally, pilot work suggested that the current methodology provided for appropriate, stable responses without requiring numerous familiarization trials. The exact intensity (treadmill speed) of these trials was selected by the participant through the following of instructions to adjust speed to elicit a work level that produced the prescribed level of exertion, the primary outcome variable.


Each participant completed an informed consent document, a demographic questionnaire, and a health status questionnaire. Participants classified as low risk for exercise participation according to the ACSM guidelines (e.g., lipids, smoking status, family history, fasting glucose) were invited to complete the study (1). All participants were cleared for participation through a personal physician or physicians affiliated with the research team. The sample was delimited in this manner because physician supervision was not readily available for maximal exercise testing. Participants were given instructions consistent with ACSM guidelines regarding the ingestion of food, fluids, alcohol, caffeine, and tobacco as well as sleep and clothing considerations (e.g., avoid alcohol, caffeine, and tobacco for three h before testing) (1). Participants provided informed consent in accordance with institutional guidelines before the first exercise session. All documents and procedures were approved by the university institutional review board.

Metabolic testing

A progressive, multistage protocol was performed on a Trackmaster treadmill (Full Vision, Newton, KS) to determine V˙O2max. The exercise protocol began at 4.0 km·h−1 mph and increased by 0.8 km·h−1 each minute (with no increases in grade) until speed reached 12.9 km·h−1, after which intensity was increased by increasing treadmill grade by 2% each minute with no further increases in speed. HR, RPE, blood pressure (BP), and expired gases were measured in accordance with standard exercise testing guidelines (1). HR was assessed continuously using a Polar™ heart rate monitor (Polar, USA). RPE was estimated each minute using the walk/run format of OMNI perceived exertion scale (17). BP was determined by auscultation every 3 min. Expired O2 and CO2 concentration were collected through an air cushion mask and analyzed continuously using a Vacumetrics MiniVista™ metabolic cart (Vacumetrics, Ventura, CA). V˙O2 values were averaged every 20 s, and V˙O2max was identified as the largest volume of oxygen consumed per minute during the test. Criteria for verifying maximal exertion were as follows: a peak HR within 10 beats of age-predicted maximal HR, a peak RPE of at least 9 (on a 0-10 scale), and a peak respiratory exchange ratio (RER) of at least 1.10. Upon conclusion of the test, participants completed a 3- to 5-min active recovery.

Experimental exercise trials

The primary instructions to the research participants during the submaximal trials were to exercise on the treadmill at self-selected intensities that correspond to "light," "moderate," and "vigorous." Participants were informed of the length of the trial before exercise and were encouraged to select intensities that corresponded with the verbal descriptor and could also be maintained for the entire 30-min trial. The initial intensity loading was variable in length and served as the graded warm-up period for each submaximal trial. Additionally, each trial was concluded with a brief graded cool down. The intensity of each trial resulted in warm-up periods that were variable in length such that the warm-up for the vigorous trial was longer than the warm-up trial for the light and the moderate trials (P < 0.05). Cool-down length, however, did not differ between trials (P > 0.05). Participants were given the standard instructions every 5 min during the trial to make any adjustment in speed necessary to maintain the requested self-selected intensity. Expired gases were collected during the fourth minute, whereas HR and RPE were collected during the fifth minute. All workload changes were made by the participant at the conclusion of each 5-min segment of the trial.


The primary variable of interest for this study was RPE, which was assessed before, during, and after each trial. Assessments taken before and after exercise used a written question, and assessments taken during exercise used a chart and an oral response. In each case, RPE was assessed using the response format associated with the OMNI walk/run scale (17). The OMNI scale includes written and illustrated descriptions of effort. This scale uses a single-item indicator of exertion scored on an 11-point scale (from 0 = "extremely easy," 2 = "easy," 4 = "somewhat easy," 6 = "somewhat hard," 8 = "hard," to 10 = "extremely hard"). RPE was assessed at three time points outside of the exercise trial: (i) immediately before exercise (Pre), (ii) immediately after exercise cool down (Post-0), and (iii) 15 min after exercise (Post-15). Each of these measurements was taken while the participant was seated comfortably in a reclining chair in an area adjacent to the exercise equipment. The preexercise assessment was taken immediately after the participant was informed that the intensity of the trial would be self-selected and anchored to a perceived intensity that was either "light," "moderate," or "vigorous." This assessment included the stem "how much exertion do you anticipate experiencing during this trial of exercise?" and served as the prediction RPE. Similarly, the postexercise assessment included the stem "how much exertion did you actually experience during this trial of exercise?" and served as the session RPE. Each of these assessments was taken while the participant was seated comfortably in a reclining chair in an area adjacent to the exercise equipment. RPE during the exercise trials was assessed at six time points at 5-min intervals associated with the 30-min exercise trial.

Statistical analysis

Analyses of the data proceeded in three phases. The first phase included descriptive analysis of sample and graded exercise test characteristics. The second phase used a series of one-way (time: 5, 10, 15, 20, 25, and 30 min) ANOVA on HR, V˙O2, and treadmill speed. A significant effect was predicted for both HR and V˙O2, with planned contrasts used to compare mean differences. The third phase used separate 3 (trial: light, moderate, and vigorous) × 9 (time: Pre, 5, 10, 15, 20, 25, 30 min, Post-0, and Post-15) and 3 (trial: light, moderate, and vigorous) × 4 (time: Pre, 30-min average, Post-0, and Post-15) repeated-measures ANOVA on RPE. This manner of analysis allowed for the determination of how exertion that is averaged across time to produce a grand mean compares to momentary exertion collected throughout the exercise trials. Significant time × trial interactions were predicted and followed by planned contrasts. Because these comparisons increase the risk for type I error, the P value for post hoc analyses of means was adjusted to a more conservative 0.01. Mean differences were used to determine effect size (ES) differences (13).


Graded exercise testing

The participants were normal to marginally overweight (mean body mass index ± SD = 27.2 ± 4.0 kg·m−2 for males and 23.4 ± 2.6 kg·m−2 for females) and moderately fit for their age (V˙O2max = 49.2 ± 5.6 mL·kg−1·min−1 for males and 43.8 ± 8.5 mL·kg−1·min−1for females). Data collected during the maximal test indicated that maximal effort was achieved for the participants. Specifically, criterion for maximal effort was reached in 21 of 26 participants for HR, 25 of 26 participants for RPE, and 25 of 26 for RER. Furthermore, maximal criterion was achieved by all participants on at least two of the three indicators.

Treadmill and physiological responses

Analysis of treadmill speed, HR, and V˙O2 for each trial of exercise assessed whether external work and physiological responses changed from the beginning to the end of exercise. For the treadmill data, there were no changes in speed from the beginning to the end of the light trial, F(5, 21) = 1.89, P > 0.05, ES = 0.24, change = 0.34 mph, or vigorous trial, F(5, 21) = 2.23, P > 0.05, ES = 0.18, change = 0.26 mph, but speed did increase significantly during the moderate trial, F(5, 21) = 2.17, P < 0.05, ES = 0.28, change = 0.37 mph. Collectively, these data indicate that self-regulation of exercise intensity resulted in external load changes that were modest or insignificant. For the V˙O2 data, there were no significant changes in metabolic rate from the beginning to the end of the light trial, F(5, 21) = 1.98, P > 0.05, ES = 0.24, change = 1.89 mL·kg−1·min−1, but V˙O2 did increase significantly during the moderate trial, F(5, 21) = 11.40, P < 0.01, ES = 0.53, change = 4.69 mL·kg−1·min−1, and vigorous trial, F(5, 21) = 5.94, P < 0.01, ES = 0.57, change = 5.11 mL·kg−1·min−1. Collectively, these data indicate that self-regulation of intensity resulted in small metabolic rate elevations from the beginning to the end of exercise. For the HR data, there were significant increases in cardiovascular response from the beginning to the end of each trial: light, F(5, 21) = 5.79, P < 0.01, ES = 0.40, change = 12 beats·min−1; moderate, F(5, 21) = 15.37, P < 0.01, ES = 1.08, change = 23 beats·min−1; and vigorous, F(5, 21) = 12.04, P < 0.01, ES = 1.23, change = 18 beats·min−1. Collectively, these data indicate that self-regulation of intensity resulted in significantly elevated HR during all trials.

Exertional responses

Perceived exertion data analysis assessed perceived exertion before, during, and after trials of treadmill exercise of varied intensity. When exertion was analyzed by way of momentary assessments, there was a significant effect for time F(8, 200) = 30.80, P < 0.01, as well as a significant effect for trial, F(2, 50) = 112.71, P < 0.01. Additionally, there was a significant interaction between these factors, F(16,400) = 7.87, P < 0.01. Similarly, when exertion was analyzed using the average of values obtained during exercise, there was a significant effect for time, F(3,75) = 17.44, P < 0.01, as well as a significant effect for trial, F(2,50) = 137.78, P < 0.01. Additionally, there was a significant interaction between these factors, F(6,150) = 4.49, P < 0.01. Collectively, these data suggest that exertion values obtained before, during, and after self-regulated aerobic exercise are well matched when comparing predicted to session values and when viewing in-task as the final minute of exercise. These data are reported in Figures 1 and 2. Significant main effects for time were decomposed and analyzed through planned contrasts that are detailed in Table 1.

Exertion responses before, during, and after exercise using momentary exertions for in-task values. Data represent mean ± SE for RPE.
Exertion responses before, during, and after exercise using average exertion for in-task values. Data represent mean ± SE for RPE.
Within-trial comparisons of exertion.


This experiment was designed to investigate the exertional responses before, during, and after treadmill trials conducted at intensities reflective of current physical activity recommendations for health and fitness (12). The research design required participants to perform 30 min of exercise at self-selected intensities perceived as "light," "moderate," and "vigorous." Treadmill load was manipulated by the participants every 5 min as necessary to maintain the requested intensity. The exercise protocols used in this study resulted in sessions of aerobic exercise that simulate public health recommendations as evidenced by the cardiovascular and metabolic responses reported above. These data indicate that the manipulation was successful with respect to the intensities of the observed exercise trials.

The contrasts of predicted and session perceived exertion against in-task exertion reveal the importance of the in-task measurement model. One model is based on a series of momentary exertion responses, and the other uses an exertion score that represents the mean exertion over time or the grand mean derived from each momentary exertion. The present research design included both models along with predicted and session RPE to highlight the influence of measurement time on exertional response during exercise. Such considerations are important because they have implications on how perceptual findings are interpreted and integrated into the literature.

Patterns of responses related to predicted exertion suggest that perceptual expectations were generally not well matched to exertion measured during exercise. Predicted exertion for all trials was higher than that which was measured during the first half of the exercise session and the average of the entire 30-min session. However, momentary exertions taken at or near the end of each session were not different than the predicted exertion for all trials. Collectively, these results indicate that anticipated exertion generally exceed observed exertions. Such findings are largely in agreement with pain literature that has compared anticipated and actual pain with respect to various pain stimuli (2,16). Using a pain analogy, it is possible that overpredictions of exertion represent an effort to protect oneself against a surprising and potentially unpleasant level of exertion. However, it is also possible that the predictions observed in the present study reflect a tendency for the prediction to be congruent with the level of exertion that is associated with the latter stages of the exercise bout. Although this position seems tenable, the current project included only one length of exercise, and it is unclear if similar responses would be observed in shorter or longer bouts of exercise. It is possible that congruence in a session of different length may occur at some other portion of the exercise session.

Findings related to in-task exertion provide insight into how individuals self-regulate exercise over a period that is reflective of public health recommendations for physical activity participation. Specifically, perceived exertion increased significantly from the beginning to the end of exercise despite standard instructions to maintain an intensity that corresponded to light, moderate, and vigorous labels used in physical activity recommendations. During the light trial, for example, the 5-min exertion value was positioned along the OMNI scale just above "easy" and increased to a point midway between "easy" and "somewhat easy" at 10 min. RPE continued to climb throughout the session and was just below "somewhat easy" near the conclusion of the exercise session. The 1.5-U increase in RPE occurred without significant treadmill speed increases and in the presence of instructions to maintain a self-selected light intensity throughout the 30-min session. A similar pattern of results was observed for the moderate and the vigorous trials, but the magnitude of RPE increase was greater over the duration of the trial (moderate: 2.5-U increase; vigorous: 3.5-U increase). The presence and the magnitude of this perceptual drift are consistent with research paradigms that have used constant-load exercise sessions (9,11). Similar RPE changes during sessions that allowed for self-selected load manipulation to maintain a prescribed intensity seem unexpected at first glance but are in line with previous research because the exercise loads were generally stable in a manner similar to exercise constant-load exercise. It is possible that the upward drift of RPE is related to the established slow component of V˙O2 (8), but it is also possible that fatigue contributed to the upward adjustments in RPE as a function of exercise time. This possibility for fatigue, however, is only relevant when using a more liberal definition of fatigue that includes sensations of tiredness in addition to the more conservative definition that requires decrements in muscular performance (20). More specifically, tiredness of some kind may have occurred, but the maintenance of treadmill speed throughout the exercise session indicates that muscular performance was not impaired. Although any fatigue present in the current investigation would have been limited in scope, compensatory muscular contractions exhibited as exercise duration progressed could have contributed to increases in RPE. Additionally, the planned manipulation was to select prescription labels (light, moderate, and vigorous) that were not included on the OMNI RPE scale, which expresses intensity as variations of "easy" and "hard." It is possible that the intensity descriptions that were used were subject to a general interpretation that varied somewhat between subjects based on previous physical activity experience. These descriptions did not constrain RPE but did facilitate a disconnect between self-selected intensity labels and reports of exertion.

Development of the session RPE measurement has provided additional flexibility in exertion assessment with respect to both aerobic and resistance exercise (4,5). The current research extends prior work and raises methodological issues that need to be addressed as research involving session RPE moves forward. Only one study to date has compared session RPE responses measured during the recovery period after aerobic exercise (18). This study used cycle ergometer exercise prescribed at constant-load exercise intensities based on ventilatory threshold. Session RPE responses were taken 30 min after the completion of exercise and indicated that the exertion values were within ranges expected based on the metabolic requirements of the exercise task. The methodology used in the present study differed both in the regulation of the exercise intensity (self-regulated vs constant load) and timing of session RPE assessment (immediate and 15-min after vs 30-min after). Results of the current study indicate that the session RPE is generally similar to predicted RPE and the in-task RPE taken near the end of exercise. However, the session RPE was different than RPE values taken throughout most of the exercise and the average RPE for the entire trial. Such results indicate that reflections on the exertional experience of exercise are focused on the closing minutes of an exercise session rather than a reflection on the entire exercise bout. This tendency to recall exertional perceptions experienced during the session primarily as it ended suggests a perceptual memory fade that may have implications on physical activity recall and general perceptions of the exercise experience.

A primary limitation of this study relates to the sample. The current sample is rather narrow with respect to age and aerobic fitness. Such limitations in the sample prevent broad generalizations to the overall population, particularly to older and less fit individuals. However, the purpose of the research was to investigate the exertion patterns associated with current physical activity recommendations, and the younger, more fit participants in the current study do represent the population segment most likely to be physically active (19). As such, the current study design provides an appropriate, ecologically valid population to study these relatively novel considerations related to perceived exertion. Additionally, there is the possibility that the request for a predicted exertion for each session may have created a tendency on the part of the participants to generate a level of exertion during exercise that confirmed their preexercise prediction. Such a possibility could be tested in future studies by comparing exertional patterns in a group that is asked to predict exertion against a group that is not asked to make preexercise predictions of exertion. The presentation of the trials in ascending order for intensity may have also created an ordering effect that could have created biased exertional responses. A final limitation noted here is that exercise modality was limited to treadmill exercise in a laboratory environment. Future investigations might provide additional insight through similar research designs that make use of other exercise modalities and a comparison of the environments associated with both exercise and recovery period. Additionally, some attention should be directed to the procedures and descriptions associated with session RPE measurement because there is considerable variability in how this variable is both defined and measured. Lastly, it might also be valuable for future designs that include assessments of RPE before, during, and after exercise to vary the length of exercise, possibly using deception, to determine more fully the impact of exercise duration on predicted and session exertion.

In summary, the design allowed for the production of three 30-min trials of self-regulated exercise that generally reflect current physical activity recommendations. Findings from this experiment indicate that the sessions of aerobic exercise produced by intensity labels used in contemporary recommendations are associated with workloads within an expected range. We are aware of no other study that has concurrently assessed the relationships among preexercise, in-task, and postexercise exercise assessments of perceived exertion. Additionally, the two in-task measurement models described within this research raise important methodological questions regarding the point of reference that is used by participants when asked to anticipate or reflect on the exertional experience of an exercise bout. It appears as though both predicted and session RPE values are well matched to the concluding minute of aerobic exercise but are mismatched with respect to the majority of the session and the average exertion for the session. Although the current project did not investigate implications of perceptions of exercise on maintenance of exercise behavior, it is possible that future decisions regarding exercise behavior may be impacted by exertional perceptions before and after aerobic exercise.

This research project was funded in part through a College of Education Mini-Grant program through the University of South Florida.

Results of the present study do not constitute endorsement by ACSM.


1. American College of Sports Medicine. ACSM's Guidelines for Exercise Testing and Prescription. Philadelphia: Lippincott, Williams & Wilkins; 2006. p. 22-78.
2. Arntz A. Why do people tend to overpredict pain? On the asymmetries between underpredictions and overpredictions of pain. Behav Res Ther. 1996;34:545-54.
3. Borg G. Perceived exertion as indicator of somatic stress. Scand J Rehabil Med. 1970;2:92-8.
4. Day ML, McGuigan MR, Brice G, Foster C. Monitoring exercise intensity during resistance training using the session RPE scale. J Strength Cond Res. 2004;18:353-8.
5. Foster C, Florhaug JA, Franklin J, et al. A new approach to monitoring exercise testing. J Strength Cond Res. 2001;15:109-15.
6. Foster C. Monitoring training in athletes with reference to overtraining syndrome. Med Sci Sports Exerc. 1998;30(7):1164-8.
7. Foster C, Daines E, Hector L, Snyder AC, Welsh R. Athletic performance in relation to training load. Wis Med J. 1996;95:370-4.
8. Gaesser GA, Poole DC. The slow component of oxygen uptake kinetics in humans. Exerc Sport Sci Rev. 1996;24:35-71.
9. Glass SC, Chvala AM. Preferred exertion across three common modes of exercise training. J Strength Cond Res. 2001;15:474-9.
10. Green JM, Yang Z, Laurent CM, et al. Session RPE following interval and constant-resistance cycling in hot and cool environments. Med Sci Sports Exerc. 2007;39(11):2051-7.
11. Green JM, Pritchett RC, Crews TR, Tucker DC, McLester JR, Wickwire PJ. RPE drift during cycling in 18 degrees C vs 30 degrees C wet bulb globe temperature. J Sports Med Phys Fitness. 2007;47:18-24.
12. Haskell WL, Lee I, Pate RR, et al. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Med Sci Sports Exerc. 2007;39(8):1423-34.
13. Hedges LD, Olkin I. Estimation of a single effect size. In: Hedges LD, Olkin I, editors. Statistical Methods for Meta-analysis. New York: Academic Press; 1985. p. 76-104.
14. McCracken LM, Gross RT, Sorg PJ, Edmands TA. Prediction of pain in patients with chronic low back pain: effects of inaccurate prediction and pain-related anxiety. Behav Res Ther. 1993;31:647-52.
15. Rachman S, Lopatka C. Accurate and inaccurate predictions of pain. Behav Res Ther. 1988;26:291-6.
16. Rachman S, Arntz A. The overprediction and underprediction of pain. Clin Psychol Rev. 1991;11:339-55.
17. Robertson RJ. Perceived Exertion for Practitioners: Rating Effort with the OMNI Picture System. Champaign: Human Kinetics; 2004. p. 184.
18. Sweet TW, Foster C, McGuigan MR, Brice G. Quantitation of resistance training using the session rating of perceived exertion method. J Strength Cond Res. 2004;18:796-802.
19. U.S. National Center for Health Statistics. Healthy People 2000 Review, 1998-1999. Hyattsville (MD): Public Health Service; 1999. p. 29-37.
20. Wilmore JH, Costill DL. Metabolism and basic energy systems. In: Wilmore JH, Costill DL, editors. Physiology of Sport and Exercise. Champaign: Human Kinetics; 1999. p. 114-54.


©2009The American College of Sports Medicine