Journal Logo

APPLIED SCIENCES: Psychobiology and Behavioral Strategies

Can the Feeling Scale Be Used to Regulate Exercise Intensity?


Author Information
Medicine & Science in Sports & Exercise: October 2008 - Volume 40 - Issue 10 - p 1852-1860
doi: 10.1249/MSS.0b013e31817a8aea
  • Free


Speculation exists that a positive affective response experienced during exercise will lead to greater enjoyment of the exercise session, promote a positive memory of the exercise experience, and may ultimately play a significant role in predicting exercise adherence (7,9,22,23,32,33,40,42). The intensity at which exercise is performed is a determinant of the affective response (35). To establish the intensity that will produce a positive affective response across individuals, the extant literature has used a research protocol that manipulates exercise intensity and explores the resultant affective response as the dependent variable. Using this methodology, it is proving difficult, and possibly unrealistic, to pinpoint an intensity to which all individuals will have a positive affective response. Consequently, in an exercise prescription setting, there is still a risk of individuals being prescribed an exercise intensity that will result in an unpleasant exercise experience that may potentially reduce motivation for future exercise behavior. Therefore, perhaps a new approach to exercise intensity regulation is required to facilitate the experience of a positive affective response. One approach is to allow the individual to self-regulate their exercise intensity to result in a specific positive affective response.

Affect is regarded as a generic term that characterizes the subjective experience of any valenced (pleasant or unpleasant) state (12,21,34,38) and subsumes the concepts of emotions and moods (7). It has been argued that the quality of the subjective experience, whether it results in the individual feeling good or bad (basic affect), should be the key outcome of interest when trying to determine the association with continued exercise behavior (11). The measurement of affect has been a contentious issue. However, within the exercise context, Ekkekakis and Petruzzello (12) recommend that the Feeling Scale (FS) developed by Hardy and Rejeski (23) be used to measure the valence (pleasure-displeasure, good-bad) component of affect. This is an 11-point bipolar scale anchored from very good (+5) through neutral (0) to very bad (−5). Individuals are asked to report how they feel at specific time points. Recent studies investigating the exercise-affective response relationship have used the FS as the outcome measure of valence (7,8,22,32,36,41) and demonstrated its validity in this context (40). These studies have also recognized the importance of defining exercise intensity according to a fixed metabolic profile, namely, the ventilatory threshold (VT) or lactate threshold (LT), rather than being defined using a percentage of maximal capacity. This ensures that the exercise intensity is physiologically equivalent between individuals and eliminates one factor that may confound the exercise intensity-affect relationship. Results have shown that FS responses vary between different exercise intensities; specifically, as intensity increases, affect during exercise generally becomes less positive or more negative, particularly as the intensity increases above VT (7,9,10,32,36,41). Above VT, this decline in affective response is homogenous between individuals. Another key finding from this research is that there are substantial differences in affective responses between individuals to the same metabolically equivalent exercise intensity, particularly when the intensity is below or around the VT (7,10,32,36,41). At these intensities, individuals experience an increase in FS responses during exercise, some experience a decrease and some remain stable throughout. This pattern of responses has also been shown when exercise was prescribed at 65% V˙O2max (40), described by ACSM (1) as a moderate intensity and one at which individuals are encouraged to exercise. Potential explanations for this individual variability in affective response are found in the dual-mode model (7). The model proposes the interplay of cognitive processes and interoceptive cues in the generation of an affective response. At intensities around VT, the affective response is proposed to be influenced primarily by cognitive processes and the appraisal of the exercise experience (7). It is suggested that individuals appraise the exercise differently, and this cognitive individuality could explain the variability in affective response to a specific exercise intensity (36). If this is the case, when an exercise intensity is prescribed, there is a risk that some individuals may not experience a pleasant affective response because they appraise the intensity more negatively.

To try and reduce these interindividual differences in affective response, researchers have used a protocol where individuals are asked to self-regulate their exercise intensity (28,32,36) and choose an intensity which they prefer. The rationale being that individuals will use the cognitive appraisal process and the resultant affective response to self-regulate the exercise intensity to one that produces (or maintains) a positive affective response. Research using this protocol has shown that, on average, individuals choose to exercise at intensities around the ventilatory or LT (28,32,36), and the intensity is increased across the duration of the bout (28,32,36). In conjunction, although intensity increases across time, FS responses remain stable and positive, with an average of 3.75 (SD = 0.76) (32) and 2.60 (SD = 0.34) (36) and ranging from 2.17 to 2.48 (28), i.e., in the "good" to "fairly good" range. Taking a different approach, Hills et al. (24) asked sedentary obese and nonobese participants to walk 2 km at a pace they perceived was "walking for pleasure." The obese group chose to walk at approximately 70% HRmax; the nonobese group chose to walk at approximately 59% HRmax. Importantly, this study showed that individuals chose the same intensity to correspond to walking for pleasure over two consecutive trials and provides support for the idea that individuals can regulate their exercise intensity based on a perception of how walking for pleasure feels. However, it did not characterize what "walking for pleasure" meant to the participants, and this can be provided by using a validated measure of pleasure such as the FS. Additionally, by using the FS, individuals can anchor their feelings across the range of affective valence.

Using rating scales to regulate exercise intensity is common in exercise prescription. Borg's (3,4) RPE is an example of a psychophysiological means of regulating exercise intensity and can be used in estimation and production protocols (17). Over 10 yr ago, Dishman (6) highlighted the utility of the RPE scale (3) as a tool to help individuals self-regulate exercise intensity and support motivation. Further, research by Eston et al. (13-16,18) has demonstrated that after a minimum training period of one or two practice sessions, individuals are able to use the RPE scale to reliably self-regulate exercise intensity. However, using the RPE scale to self-regulate does not ensure a positive affective response. The RPE scale is regarded as a valid and reliable measure of "what" the individual is feeling during exercise in terms of the heaviness and strain of physical work (3). This is conceptually distinct from the FS which measures "how" people feel during exercise, the affective response of pleasure/displeasure (23). Hardy and Rejeski (23) have shown that across increasing exercise intensities, RPE and FS have between 11% and 30% of shared variance. This suggests that the affective response will influence the RPE, and conversely, individuals will attach different FS responses to this physical work (as measured by the RPE scale) depending on how it is interpreted (23). These resulting FS responses have consequences for future exercise behavior (42).

We propose that the FS could be used to regulate intensity, such that individuals are asked to regulate their intensity to produce a particular positive affective response. In this situation, the affective response will always remain positive, and the exercise intensity chosen will become the dependent variable. From previous research, we can assume that the chosen intensity will sit below or around the VT and importantly be in the range that will provide health and fitness benefits according to the American College of Sports Medicine (ACSM) (1) guidelines. From an exercise prescription perspective, for the exerciser, this intensity regulation protocol may be more practical than, for example, having to continuously monitor HR to ensure the exercise is at the "correct" intensity. The FS has been shown to be a useful tool to mark the point at which individuals begin to exercise above their VT due to the surge of displeasure that occurs at this point (9). Ekkekakis et al. (9) concluded that by using the FS, "exercisers could also monitor when they begin to feel substantially worse than before and regulate their pace accordingly" (p. 157). Thus, they have encouraged the use of the valence dimension of affect as a means of regulating exercise intensity; however, the FS has not yet been validated for this purpose. There are many self-regulation processes important for the initiation and maintenance of exercise behavior, for example, goal setting (26), planning skills (30), and self-monitoring (29). Research has shown that regular exercisers have enhanced self-regulatory abilities compared with irregular exercisers in relation to exercise goals (26) and that self-regulatory skills improve after regular exercise (31). Thus, although previous research has shown that sedentary individuals are able to self-regulate their intensity (20,24,28,32,33,36), it is likely that practice may be required before individuals will be able to achieve effective self-regulation and that ability to self-regulate will vary between individuals.

The purpose of this study was to evaluate whether or not the FS could be used as a means of regulating exercise intensity. Specifically, it aimed to investigate the chosen exercise intensity that corresponded to FS values of 3 (good) and 1 (fairly good) and to examine the variability in the chosen intensity over consecutive exercise bouts. It was hypothesized that 1) individuals would exercise at a higher intensity in the FS 1 condition compared with the FS 3 condition. Previous research has shown less positive affective responses to higher exercise intensities and therefore asking individuals to feel less positive, which we assume will correspond to them choosing a greater exercise intensity. 2)Within each exercise session, intensity would increase over time as has been shown in previous intensity self-regulation studies (28,32,33,36). We would also anticipate that there would be little variability between exercise sessions conducted at the same FS value. However, because no research has examined whether this proposal and affective responses are influenced by other factors (e.g., preexercise affective state), this analysis is more exploratory in nature.



Seventeen sedentary women volunteered to participate (mean age = 44.8 yr, SD = 8.9; mean weight = 72.9 kg, SD = 13.3; mean body mass index [BMI] = 27.2, SD = 3.9; mean V˙O2max= 31.3 mL·kg−1·min−1, SD = 3.8). To be eligible, participants were required to have not exercised more than once a week for the previous 6 months. Participant characteristics are shown in Table 1. Participants completed a medical history questionnaire, which ensured that they did not have cardiorespiratory or muscular contraindications to participate in physical activity and that they were not currently taking any medications for health problems. All participants gave written informed consent before commencing the study. Ethical approval was granted by the University Ethics Committee.

Participant characteristics.


The Feeling Scale (FS).

Affective valence (pleasure/displeasure) was measured using the FS (23). Participants rate their current feelings on an 11-point bipolar scale ranging from +5 to−5, with verbal anchors of very good (+5), good (+3), fairly good (+1), neutral (0), fairly bad (−1), bad (−3), and very bad (−5). Previous research (40) found the FS to correlate between 0.51 and 0.88 with the valence scale of the Self-Assessment Manikin (27) and between 0.41 and 0.59 with the valence scale of the Affect Grid (37).

Ratings of Perceived Exertion (RPE).

The Borg 6-20 RPE Scale (4) assessed whole-bodyRPE. Participants state the number that reflects how hard the exercise feels on a 6-20 scale, ranging from no exertion at all [6] through somewhat hard [13] to maximal exertion [20]. In line with recommendations (4), participants were given standardized instruction on how to use the scale (p. 47) and had time to practice during the familiarization session.


In line with recommendations for the measurement of self-efficacy (19), a three-item measure of self-efficacy was developed for the purposes of the study. The items asked participants to rate how confident they were in their ability to exercise for 30 min when the exercise caused them to feel: 1) very good, 2) good, and 3) fairly good. Participants reported their confidence levels on a scale of 0% (not at all confident) through 50% (moderately confident) to 100% (completely confident), with 10-point increments given on the scale. Scores on the three items were averaged to provide an indicator of the strength of self-efficacy. Internal consistencies (α) at all eight times of administration were between 0.75 and 0.92.


Maximal graded treadmill test.

In the first visit to the laboratory, participants completed a maximal graded treadmill test to volitional exhaustion to establish maximal aerobic capacity (V˙O2max) and VT. This visit also served as a familiarization session for using the FS and RPE scale. Initially, participants completed an informed consent form that explained the procedures of the study, a self-report activity history questionnaire, and the medical history questionnaire. An HR monitor (Polar PE3000; Polar Electro Oy, Kempele, Finland) was fitted to the participant's chest. Participants were asked to sit quietly for 5 min, and HR at rest was obtained. A description of the RPE and FS was given, participants were instructed on how to use them, and a rating for the FS was given. The procedures for the test were then explained. Before the commencement of the test, age, height, and weight were recorded. The participant then stood on the treadmill and was fitted with a mouthpiece and noseclip.

The exercise test was based on the Balke-Ware treadmill graded exercise test protocol (1). Participants were required to walk on a Quinton Series 90 Q65 treadmill at an initial speed of 4 km·h−1 on a gradient of 0% for 3 min. The speed was then increased to 5.4 km·h−1 and the gradient increased to 2% (for four participants, the speed was kept at 4 km·h−1). Every minute thereafter, the gradient was increased by 1-2% while maintaining a speed of 5.4 km·h−1. The test continued until the participant reached the point of volitional exhaustion. Continuous breath analysis was performed using a Sensormedics 2900 Metabolic Cart (Sensormedics Corporation, Yorba Linda, CA) metabolic analysis system, and subsequently the data were averaged over 20-s intervals. The analyzer was calibrated before each test. The test was classified as a maximal effort if two of the standard criteria for reaching maximal oxygen uptake were reached: 1) a peak or plateau in oxygen consumption (changes of less than 2 mL·kg−1·min−1), 2) an RER of ≤1.1, and 3) reaching or exceeding a final HR of age-related maximum (220 bpm − age). HR, RPE, and FS were recorded every minute. The RPE and the FS scales were placed in front of the participants, and at every minute, they were asked to point to the value that characterized them at that time.

V˙O2max was determined as the highest V˙O2 value attained after reaching the standard criteria. The V-slope method was used to determine VT. V˙CO2 was plotted against V˙O2, and from visual inspection of the graph, VT was determined to be the point at which the first disproportionate increase in V˙CO2 occurred. Two investigators independently assessed the graphs, and if there was disagreement between these two investigators, a third investigator analyzed the graphs, and discussions between the three investigators were held until a consensus was reached on where VT lay. There was no need for the third investigator to be consulted in any of the cases.

Exercise sessions.

All 17 participants performed eight 30-min sessions of treadmill exercise on the Quinton Series 90 Q65 treadmill in a laboratory. Four consecutive exercise sessions were performed at an intensity perceived to correspond to an FS value of 3, "good." The other four sessions were performed at an intensity perceived to correspond to an FS value of 1, "fairly good." The order of condition was counterbalanced. Two exercise sessions were completed each week (at least 1 day apart) at the same time of the day.

The procedure for each exercise session followed the same pattern. The participant was fitted with an HR monitor (Polar PE3000), and resting HR was taken. The measure of self-efficacy was taken. The participant was then shown the FS, and its instructions and the procedure for the exercise session were explained. The participant was told, "You will be exercising on the treadmill for 30 min; we would like you to select a speed and gradient on the treadmill that will result in you feeling "good" (or "fairly good" depending on condition) so reporting an FS score of "3" (or "1") through the 30 min. You will have the opportunity to change the speed and the gradient every 5 min if you wish." (Participants were not allowed to change the intensity between each 5-min assessment.) The participant provided an FS score that corresponded to how she felt at that time. She was then given a few minutes to find the speed and gradient that she perceived made her feel "good" [or "fairly good"], and once she was comfortable with the speed and gradient she chose, the 30 min began. Participants chose to walk on the treadmill at varying gradients; one participant increased the intensity to a jog in the latter stages of each exercise session. The actual speed and gradient of the treadmill as well as HR were kept blind from the participant at all times. In the last 30 s of each 5-min period, treadmill speed and gradient were recorded as well as measures of HR and RPE to enable the calculation of exercise intensity. At this point, the participant was asked to confirm that she still felt "good" [or "fairly good"] and was given the opportunity to change the speed and gradient if she wished. Once the 30 min was over, participants were given the opportunity to warm down for 2 min and then gave a reading on the FS.

Statistical analysis.

Exercise intensity was determined by calculating V˙O2 at the speed and gradient chosen by the participants using the formulae provided by the ACSM (1). In line with the exercise intensity nomenclature of the dual-mode model (7), this V˙O2 value was then represented as a percentage of the participants V˙O2 above or below V˙O2 at VT. Using SPSS 13.0 for Windows (SPSS Inc., Chicago, IL), a series of three-factor, condition (FS 1 and FS 3) × trial (exercise sessions 1, 2, 3, and 4) × time (5, 10, 15, 20, 25, 30 min), within-subjects repeated-measures ANOVA was conducted on the V˙O2, HR, RPE, and self-efficacy data. All statistically significant findings were followed up with nonorthogonal (repeated) planned comparisons to test the a priori hypotheses specified at the outset. Where the assumption of sphericity was violated, degrees of freedom were corrected with the Greenhouse-Geisser epsilon. Effect sizes are reported as η2 with values of 0.01, 0.06, and 0.14 regarded as small, medium, and large effects (39). Intraclass correlations (ICC) were calculated for V˙O2, HR, and RPE to examine the consistency in the mean intensity chosen in the four trials of each condition. To examine the variability in the intensities selected across the 30 min in each trial, the ICC and the coefficient of variation (CV) were calculated for V˙O2, HR, and RPE from individual data at the 5-, 15-, and 30-min time points. A repeated-measures ANOVA (condition × time) with Tukey post hoc tests were conducted to examine any significant changes in the CV.


The results for V˙O2 as a percentage of V˙O2 at VT showed significant main effects for trial [F(3, 48) = 7.27, P< 0.01, η2 = 0.31] and time [F(2.00, 31.96) = 29.15, P < 0.01, η2 = 0.65]. Repeated planned comparisons found that V˙O2 was significantly lower in trial 1 compared with trials 2, 3, and 4. There were no differences between trials 2, 3, and 4. Across time, V˙O2 increased significantly every 5 min until 20 min. From 20 to 30 min, V˙O2 was stable. There was no condition main effect. The average V˙O2 (SEM) for FS 1 was 8% (6%) greater than V˙O2 at VT and for FS 3 was 7% (6%) greater than V˙O2 at VT. For FS 1, the 95% confidence intervals ranged from 11.5% below−16% above VT in the first trial to 2.5% below−22.5% above by the fourth trial. For FS 3,95% confidence intervals ranged from 7% below−15% above in trial 1 to 5% below−18% above by trial 4. Intensity equated to 69% (3%) V˙O2max at FS 1 and 66% (2%) V˙O2max at FS 3. The mean and SEM for V˙O2 are found in Table 2.

Exercise intensity selected across time in each condition.

The results for HR expressed as a percentage of maximum HR (%HRmax) showed significant main effects for condition [F(1, 16) = 5.75, P < 0.05, η2 = 0.26] and time [F(1.77, 28.26) = 30.26, P < 0.01, η2 = 0.65]. %HRmax was greater in FS 1 than FS 3. Across time, %HRmax increased every 5 min until 20 min and from 20 min to 30 min (Table 2). The average %HRmax (SEM) for FS 1 was 68% (3%) and for FS 3 was 64% (2%).

RPE results demonstrated significant main effects for condition [F(1, 16) = 5.45, P < 0.05, η2 = 0.25] and time [F(2.07, 33.17) = 59.59, P < 0.01, η2 = 0.79]. Repeated planned comparisons showed that RPE was significantly higher for FS 1 than FS 3. Across time, RPE increased significantly every 5 min until 25 min into the exercise session. Average RPE (SEM) for FS 1 was 12.0 (0.3) and for FS 3 was 11.4 (0.3) (see Table 2).


The ICC for V˙O2, %HRmax, and RPE within trials were high for both the FS 1 and the FS 3 conditions. The ICC ranged between 0.98 and 0.99 for V˙O2, 0.92 and 0.98 for %HRmax, and 0.89 and 0.97 for RPE. This supports that the pattern of response over time within each trial was consistent for trials 1 to 4. However, the consistency across trials was not quite as high. ICC were calculated for each condition at 5 min, 15 min, and 30 min and ranged between 0.93 and 0.97 for V˙O2, 0.84 and 0.96 for %HRmax, and 0.85 and 0.95 for RPE (Table 3).

ICC and mean (SEM) CV for physiological variables at 5, 15, and 30 min across trials 1 to 4.


The average CV across the four trials for data captured at 5, 15, and 30 min ranged between 5% and 11% for V˙O2, %HRmax, and RPE (see Table 3). The ANOVA revealed a significant interaction [F(1.44, 23.04) = 4.94, P< 0.05, η2 = 0.24] and time main effect [F(1.46, 23.45) = 3.66, P < 0.05, η2 = 0.19] for V˙O2. Tukey post hoc follow-up tests indicated that the only significant difference was that CV was higher at 5 min than at 30 min in the FS 1 condition. CV data were stable across time in the FS 3 condition. There were no significant differences in CV for %HRmax or RPE.


The analysis of self-efficacy showed a significant main effect for trial [F(3, 48) = 5.19, P < 0.01, η2 = 0.25]. Repeated planned comparisons showed that self-efficacy was significantly higher in trial 4 compared with trials 1 (η2 = 0.37) and 2 (η2 = 0.24).


The purpose of this study was to evaluate whether the FS could be used to guide the self-regulation of exercise intensity so that individuals experience an affective state of good and fairly good during exercise. We were interested to ascertain the chosen exercise intensity that corresponded to an FS value of 3 (good) and 1 (fairly good) and to examine whether the selected intensity was similar over consecutive exercise bouts. The results show that to achieve an affective state of good (FS 3), individuals exercise at a lower intensity than to achieve an affective state of fairly good (FS 1). The FS 3 condition equated to an RPE of 11.4 (SEM = 0.3), 64% HRmax (SEM = 2%), and V˙O2 of 20.7 mL·kg−1·min−1 (SEM = 0.8), which equates to approximately 6 METs. The FS 1 condition equated to an RPE of 12.0 (SEM = 0.3), 68% HRmax (SEM = 3%), and V˙O2 of 21.3 mL·kg−1·min−1 (SEM = 1.0), which also equates to approximately 6 METs. In both conditions, on average, the intensities were 7-8% above the individual's VT. Across the four bouts of exercise at each condition, individuals consistently selected the same intensity to elicit a feeling state of good and fairly good. Interestingly, across the 30 min in all of the exercise sessions, individuals increased the intensity to maintain the required affective state.

It has been proposed that within the exercise intensity-affect relationship, exercise intensity should be defined in relation to the VT (or LT) to ensure physiological equivalence between individuals (7). When compared with the V˙O2 value recorded at VT during the graded exercise test, V˙O2 values calculated from the exercise session data showed that participants exercised, on average, 2-4% above VT in the first exercise session and between 6% and 10% above VT in the three subsequent sessions. There was no difference in the chosen intensity between conditions using this physiological marker. The reported confidence intervals reflect that a greater proportion of the population worked above their VT, particularly by the fourth trial. Previous research, which has asked low active individuals to self-select their preferred exercise intensity, has also shown that individuals select intensities close to their VT and report average FS values between 3 and 4 (men) (32) and between 2 and 3 (women) (28, 36). From the results of this study and previous studies, it would seem that during self-regulated exercise protocols of 20-30 min, FS values of between 1 and 3 may signify that the individual is exercising close to their VT. Interestingly, however, if we prescribe an exercise intensity around VT, individuals report a lower average FS value of between 1 and 1.3 (36). This suggests that allowing individuals to self-regulate their own intensity results in a more pleasant exercise experience, without compromising exercise intensity. The V˙O2 data also showed that the intensity chosen in trial 1 of both conditions was significantly lower than the intensity chosen in trials 2 to 4. This would suggest that it may take one exercise session for individuals to recognize the exercise intensity that corresponds to an affective response of good or fairly good or indeed to appreciate what feeling good or fairly good actually means in the context of exercise. However, once this is completed, individuals seem to be able to replicate that intensity in subsequent sessions.

In support of our hypothesis, the %HRmax analysis showed that individuals chose a significantly greater exercise intensity to represent feeling fairly good than feeling good. On average, individuals chose to exercise at 68% HRmax (SEM = 11) in the FS 1 condition and 64% (SEM = 8) in the FS 3 condition. Importantly, the effect size data (η2 = 0.26) showed this to be a large effect (39). This difference in intensity between conditions was also shown in the RPE data. This significant difference between the conditions shows that the distinction between feeling good and fairly good during exercise can be a matter of only a 4% increase in intensity. Thus, even a very subtle increase in intensity can be enough to make the individual feel less positive about their exercise experience. Given the proposed link between affect and adherence (7,9,22,23,32,33,40,42), the influence that this "extra" intensity can have on affective responses may be enough to discourage future exercise participation because the experience was less pleasant than they are willing to tolerate.

Results also showed that during each exercise session, irrespective of condition, individuals chose to increase their exercise intensity across the 30 min (this pattern emerged for all physiological variables). This is in line with previous self-selected intensity research that has shown that individuals choose to increase their intensity across time while maintaining a positive affective state (28,32,36). These differences, between conditions and across time, may be explained through the cognitive appraisal process that underpins the affective response to exercise at intensities around VT (7). Rose and Parfitt (36) showed that the factors involved in this cognitive appraisal include perceptions of ability to cope with the intensity and complete the exercise session, perceiving the intensity as comfortable yet challenging and feeling in control of the intensity, the perception that the exercise is providing benefits and important outcomes, and the ability to dissociate from the symptoms of the exercise such that attention is focused on factors external to the exercise. The appraisal process that results in affective responses of good and fairly good may be different and would consequently influence the exercise intensity that was selected. Similarly, there maybe a difference in the cognitive appraisal at different time points that leads to the increase in intensity to maintain the same affective response. As Ekkekakis et al. (9) point out, "despite the importance of exercise intensity, relatively little is known about the processes of self-monitoring and self-regulation of exercise intensity particularly among formerly sedentary individuals" (p. 150). The self-efficacy results highlight the change in one aspect of this cognitive appraisal across trials. Self-efficacy increased from trials 1 and 2 to trial 4, showing that after three exercise sessions, individuals felt higher levels of perceived ability to exercise for 30 min at an intensity that lay in the fairly good to good range on the FS. This supports the basic premise that performance accomplishments (2) enhance self-efficacy and perhaps provides an explanation for why intensity was lower in trial 1 compared with trials 2 to 4. An alternative explanation may be that participants' self-efficacy responses in trial 4 were more accurate than the responses provided in trials 1 and 2 because individuals would have gained valuable experience of the exercise stimulus on which to base their efficacy beliefs. Thus, caution should be taken when interpreting the self-efficacy results. The role of self-efficacy and other potential cognitive explanations for the increase in intensity across time and trial and the specific factors that cause an affective response of good and fairly good should be the subject of future investigations.

There may also be a physiological basis for the increase in exercise intensity over time. The dual mode model suggests that at intensities proximal to the VT (the intensity chosen to represent both FS 1 and FS 3), along with the cognitive appraisal process, the affective response will be influenced to some extent by interoceptive cues. Therefore, changes in thermogenic, respiratory, and muscular cues brought about by the exercise may have prompted the individual to increase the intensity to maintain the specified FS response. However, it is interesting to note that RPE (which were taken before the individual deciding to increase their intensity) drifted across time, suggesting that intensity was increased to maintain the specific FS response despite the individual reporting higher levels of perceived exertion. This may again allude to the role of cognitive appraisal as influencing the decision to increase intensity. However, future research should also consider the collection of blood samples, core body temperature, and respiratory gases during the exercise bout to provide physiological data to help identify potential physiological mechanisms influencing affective responses.

To be convinced that the FS is a reliable method of regulating exercise intensity, it is important to show that, within each condition, across trials, individuals choose the same exercise intensity to represent feeling good and fairly good. The high ICC for V˙O2, %HRmax, and RPE are indicative that the individuals worked at consistent intensities in both the FS 1 and the FS 3 conditions. Further, the CV data support that there was low variability in the chosen intensity across trials for both conditions. The CV ranged between 5% and 11% across all variables, which equates to a 5-bpm difference in HR and 1.1% variance in V˙O2 expressed as a percentage of V˙O2 at VT. These are accepted variances that are indicative of a steady-state exercise intensity (25). Although the variability is low,individuals did improve across time in the FS 1 condition, with the CV for V˙O2 decreasing from 5 min (11%) to 30 min (6%). These data may suggest that it is harder to replicate an intensity to represent feeling fairly good compared with that representing feeling good at the outset of an exercise session, but that this improves over the 30-min session. It is a little unexpected, but nevertheless encouraging, that the sedentary women in the study were able to use their feeling states to regulate the exercise intensities they selected with such a high level of consistency. Ekkekakis et al. (9) suggested that most sedentary women find it difficult to estimate and regulate their exercise intensity with any degree of accuracy, but that this ability can be improved with practice. Our results would support this suggestion, showing that the ability to regulate intensity improves with one bout of practice and can even improve across one 30-min exercise session.

A limitation of the study is that direct measurements of V˙O2 using online gas analysis were not taken. It was felt that measuring V˙O2 directly during the exercise would interfere with the affective response being generated from the exercise and would reduce the external validity of our findings. Instead, we chose to use the ACSM (1) equation, which estimates the V˙O2 associated with specific treadmill speeds and gradients, and we are willing to accept the loss of sensitivity in the measurement of V˙O2 to gain greater external validity. The use of this procedure may account forthe different pattern of results that occurred in the V˙O2 data compared with the HR and the RPE data. This problem may be circumvented in the future by measuring respiratory gases on cessation of the exercise bout to provide more accurate V˙O2 information of the final exercise intensity. A further limitation is that to allow for the calculation of V˙O2, we restricted any intensity changes to each 5-min point rather than allowing the individual to change the intensity at their leisure. This procedure may mean that between the 5-min periods, individuals may not have been exercising at the specific FS rating because they could not change the intensity to maintain this rating. However, it should be noted that participants confirmed at each 5-min data point that they were feeling good or fairly good as required.

The strength of this study is the implications it has for professionals involved in exercise prescription. Exercise prescribers can be confident that if they allow individuals to self-regulate their exercise intensity using the good to fairly good range of the FS, they will naturally select an intensity recommended by ACSM (1) to be beneficial for health and fitness. More importantly, from a psychological perspective, using this method of intensity regulation may make it more likely that these individuals will have a pleasant exercise experience and will not be at risk of experiencing a potentially discouraging unpleasant affective state or a reduction in affective state during exercise. This is not guaranteed when exercise is prescribed at a specific intensity based around HR values or RPE, or in relation to VT (32,36,40). Incorporating the self-regulation of exercise intensity using affect could supplement other self-regulation strategies such as goal setting (26), planning skills (30), and self-monitoring (29) to further improve rates of exercise initiation and adherence. Obviously, the results of this study are generalizable only to a sedentary population of healthy women exercising on a treadmill; we encourage further research on the utility of the FS as an intensity regulation tool in other populations and exercise modes. Although we would expect other populations to be able to use the scale to regulate their intensity, the resulting intensity may be different. For example, it would be anticipated that the intensity required to elicit a good or fairly good affective response will vary with exercise ability and experience. As individuals continue to exercise and increase their fitness, the intensity that represents an FS 1 or FS 3 (measured by %HRmax or % V˙O2max) may increase. However, it may also be expected that when measured as a percentage of VT, the intensity would not change because we would expect VT to occur at a higher physiological load with increased fitness (5). We might also expect within-subject variability in the intensity, which represents a given FS response on different exercise sessions. Depending on day-to-day circumstances, for example, general levels of fatigue, preexercise affective state, mode of exercise, and the exercise intensity that will represent feeling good and fairly good is likely to differ. In practical terms, this variability does not impact on the usefulness of the FS scale to regulate intensity, although it may mean that chosen intensities sit below, or more positively, above recommended health guidelines. This method may also provide the extra motivation required to exercise on those days when the individual does not feel like exercising because he or she knows that the exercise can be regulated to provide a pleasant outcome. Potential within-subject differences on intensity regulation and its causes and consequences would be an interesting avenue for future research. Because our study was conducted in a laboratory-based environment, future research should also investigate whether or not individuals can use the FS scale to regulate intensity and what intensity is chosen when in other physical activity environments, for example, in an unsupervised home-based exercise program or when walking outdoors.

In conclusion, this study has shown that individuals are able to use the FS to self-regulate their exercise intensity to elicit an affective state of good and fairly good. Although there is some discrepancy in results depending on the physiological variables examined, the results suggest that exercising to feel fairly good requires greater exercise intensity than exercising to feel good. However, both these stimuli lie close to the individual's VT. The advantage of using the FS over other methods of intensity regulation such as monitoring HR or using the RPE scale is that it facilitates the experience of a positive affective response from the exercise, and this should provide extra motivation for the individual to consider repeating their exercise experience. However, the implications of regulating exercise intensity using the FS and specifically using FS 1 and FS 3 as potential optimum affective targets on exercise behavior and other psychological outcomes have yet to be established.

The authors thank Ms. Sarah Gee who assisted with data collection and the two anonymous reviewers for their insightful comments on the manuscript. This work was supported by a departmental grant from the School of Physical Education, University of Otago, to the first author. The 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. 7th ed. Philadelphia (PA): Lippincott Williams & Wilkins; 2006. p. 366.
2. Bandura A. Self-Efficacy: The Exercise of Control. New York: W.H. Freeman; 1997. p. 604.
3. Borg GAV. An Introduction to Borg's RPE-Scale. Ithaca (NY): Mouvement; 1985. p. 30.
4. Borg GAV. Borg's Perceived Exertion and Pain Scales. Champaign (IL): Human Kinetics; 1998. p. 104.
5. Carter H, Jones AM, Doust JH. Effect of six weeks of endurance training on lactate minimum speed. J Sports Sci. 1999;17:957-67.
6. Dishman RK. Advances in Exercise Adherence. Champaign (IL): Human Kinetics; 2006. p. 406.
7. Ekkekakis P. Pleasure and displeasure from the body: perspectives from exercise. Cogn Emot. 2003;17:213-39.
8. Ekkekakis P, Acevedo EO. Affective responses to acute exercise: toward a psychobiological dose-response model. In: Acevedo EO, Ekkekakis P, editors. Psychobiology of Physical Activity. Champaign (IL): Human Kinetics; 2006. pp. 91-110.
9. Ekkekakis P, Hall EE, Petruzzello SJ. Practical markers of the transition from aerobic to anaerobic metabolism during exercise: rationale and a case for affect-based exercise prescription. Prev Med. 2004;38:149-59.
10. Ekkekakis P, Hall EE, Petruzzello SJ. Variation and homogeneity in affective responses to physical activity of varying intensities: an alternative perspective on dose-response based on evolutionary considerations. J Sports Sci. 2005;23:477-500.
11. Ekkekakis P, Petruzzello SJ. Analysis of the affect measurement conundrum in exercise psychology: I. Fundamental issues. Psychol Sport Exerc. 2000;1:71-88.
12. Ekkekakis P, Petruzzello SJ. Analysis of the affect measurement conundrum in exercise psychology: IV. A conceptual case for the affect circumplex. Psychol Sport Exerc. 2002;3:35-63.
13. Eston RG, Lamb KL, Parfitt CG, King N. The validity of predicting maximal oxygen uptake from a perceptually regulated graded exercise test. Eur J Appl Physiol. 2005;9:221-7.
14. Eston RG, Parfitt G, Campbell L, Lamb KL. Reliability of effort perception for regulating exercise intensity in children using a Cart and Load Effort Rating (CALER) scale. Pediatr Exerc Sci. 2000;12:388-97.
15. Eston R, Lambrick D, Sheppard K, Parfitt G. Prediction of maximal oxygen uptake in sedentary males from a perceptually regulated, sub-maximal graded exercise test. J Sports Sci. 2008;26:131-9.
16. Eston RG, Williams JG. Reliability of ratings of perceived effort for regulation of exercise intensity. Brit J Sport Med. 1988;22:153-5.
17. Eston RG, Williams JG. Control of exercise intensity using heart rate, perceived exertion and other non-invasive procedures. In: Eston RG, Reilly T, editors. Kinanthropometry and Exercise Physiology Laboratory Manual: Tests, Procedures and Data. Volume 2: Exercise Physiology, 2nd ed. London: Routledge; 2001. p. 213-34.
18. Faulkner JA, Parfitt G, Eston RG. Prediction of maximal oxygen uptake from the ratings of perceived exertion and heart rate during a perceptually-regulated sub-maximal exercise test in active and sedentary participants. Eur J Appl Physiol. 2007;101:397-407.
19. Feltz DL, Chase MA. The measurement of self-efficacy and confidence in sport. In: Duda JL, editor. Advances in Sport and Exercise Psychology Measurement. Morgantown (VA): Fitness Information Technology; 1998. p. 65-80.
20. Focht BC, Hausenblas HA. State anxiety responses to acute exercise in women with high social physique anxiety. J Sport Exerc Psychol. 2003;25:123-44.
21. Frijda NH. Moods, emotion episodes, and emotions. In: Lewis M, Haviland JM, editors. Handbook of Emotions. New York: Guilford; 1993. p. 381-403.
22. Hall EE, Ekkekakis P, Petruzzello SJ. The affective beneficence of vigorous exercise revisited. Br J Health Psychol. 2002;7:47-66.
23. Hardy CJ, Rejeski WJ. Not what, but how one feels: the measurement of affect during exercise. J Sport Exerc Psychol. 1989;11:304-17.
24. Hills AP, Byrne NM, Wearing S, Armstrong T. Validation of the intensity of walking for pleasure in obese adults. Prev Med. 2006;42:47-50.
25. James DVB, Sandals LE, Wood DM, Jones AM. Pulmonary gas exchange. In: Winter EM, Jones AM, Davidson RCR, Bromley PD, Mercer TH, editors. Sport and Exercise Physiology Testing Guidelines. London: Routledge; 2007. p. 101-11.
26. Karoly P, Ruehlman LS, Okun MA, Lutz RS, Newton C, Fairholme C. Perceived self-regulation of exercise goals and interfering goals among regular and irregular exercisers: a life space analysis. Psychol Sport Exerc. 2005;6:427-42.
27. Lang PJ. Behavioural treatment and bio-behavioural assessment: computer applications. In: Sodowski JB, Johnson JH, Williams TA, editors. Technology in Mental Health Care Delivery Systems. Norwood (NJ): Ablex; 1980. p. 119-37.
28. Lind E, Joens-Matre RR, Ekkekakis P. What intensity of physical activity do previously sedentary middle-aged women select? Evidence of a coherent pattern from physiological, perceptual, and affective markers. Prev Med. 2005;40:407-19.
29. Lutz RS, Karoly P, Okun MA. The why and the how of goal pursuit: self-determination, goal process cognition, and participation in physical exercise. Psychol Sport Exerc. Doi:10.1016/j.psychsport.2007.07.006.
30. Macdonald A, Palfai T. Predictors of exercise behavior among university student women: utility of a goal-systems/self-regulation theory framework. Pers Individ Dif. 2008;44:921-31.
31. Oaten M, Cheng K. Longitudinal gains in self-regulation from regular physical exercise. Br J Health Psychol. 2006;11:717-33.
32. Parfitt G, Rose EA, Burgess WM. The psychological and physiological responses of sedentary individuals to prescribed and preferred intensity exercise. Br J Health Psychol. 2006;11:39-53.
33. Parfitt G, Rose EA, Markland D. The effect of prescribed and preferred intensity exercise on psychological affect and the influence of baseline measures of affect. J Health Psychol. 2000;5:231-40.
34. Ortony A, Clore GL, Foss MA. The referential structure of the affective lexicon. Cogn Sci. 1987;11:341-64.
35. Reed J, Ones DS. The effect of acute aerobic exercise on positive activated affect: a meta-analysis. Psychol Sport Exerc. 2006;7:477-14.
36. Rose EA, Parfitt G. A quantitative analysis and qualitative explanation of the individual differences in affective responses to prescribed and self-selected exercise intensities. J Sport Exerc Psychol. 2007;29:281-309.
37. Russell JA, Weiss A, Mendelsohn GA. Affect Grid: a single item scale of pleasure and arousal. J Pers Soc Psychol. 1989;57:493-502.
38. Scherer KR. On the nature of and function of emotion: a component process approach. In: Scherer KR, Ekman P, editors. Approaches to Emotion. Hillsdale (NJ): Lawrence Erlbaum; 1984. pp. 293-317.
39. Stevens J. Applied Multivariate Statistics for the Social Sciences. Mahwah (NJ): Erlbaum; 2002. p. 699.
40. Van Landuyt LM, Ekkekakis P, Hall EE, Petruzzello SJ. Throwing the mountains into the lakes: on the perils of nomothetic conceptions of the exercise-affect relationship. J Sports Exerc Psychol. 2000;22:208-34.
41. Welch AS, Hulley A, Ferguson C, Beauchamp MR. Affective responses of inactive women to a maximal incremental exercise test: a test of the dual-mode model. Psychol Sport Exerc. 2007;8:401-23.
42. Williams DM, Dunsiger S, Ciccoli JT, Lewis BA, Albrecht AE, Marcus BH. Acute affective response to a moderate-intensity exercise stimulus predicts physical activity participation 6 and 12 months later. Psychol Sport Exerc. 2008;9:231-45.


©2008The American College of Sports Medicine