12.5: The Measurement of Affective Responses to Exercise

Up to this point, we have briefly described how hedonic theory relates to exercise behavior and we have distinguished the concepts of core affect, emotion, and mood (for a more comprehensive review, see Ekkekakis, 2013). In the following sections, we focus on several methodological considerations for practitioners seeking to measure affective responses to exercise. Specifically, we emphasize measure selection, the timing and frequency of measurement, expected individual variability in affective responses to exercise, reducing measurement bias, and considerations for decreasing measurement error.

12.5.1. Measure Selection

How should an exercise professional choose which measures or instruments to use? There are two general approaches to selecting measures to assess affective responses to exercise. The first is by far the easiest and involves choosing a measure that "has been used by others before." However, as is often the case, the easiest approach is not necessarily the best approach. Alternatively, Ekkekakis (2012, 2013) outlined a three-step approach to measure selection. The first step involves choosing whether one wants to measure affect, mood, or emotion. As described earlier, these are three related, but distinct, psychological constructs. In the second step, the exercise professional should choose a conceptual model that is appropriate for the construct identified in the first step. The final step involves choosing a measure based on the adopted conceptual model and psychometric considerations, such as evidence of validity and reliability among people with similar characteristics (e.g., age, sex, health status) to the client of the exercise professional. We endorse this approach, and strongly recommend further examination of readings on these topics (Ekkekakis, 2012, 2013; Ekkekakis & Zenko, 2016).

For example, let’s imagine an exercise professional who is interested in measuring core affective responses to exercise. Which measure or measures should be used? There are many possibilities, and some are more appropriate than others. In this scenario, the first step is completed. The exercise professional identified core affect as the psychological construct to be measured. Sound understanding of the differences between core affect, emotion, and mood is necessary to complete this step. Further, if a practitioner chooses to measure a specific emotion, then a complete understanding of that specific emotion is required. If, for example, a practitioner is interested in assessing guilt, then they should understand how guilt is different than shame (Miceli & Castelfranchi, 2018).

The second step requires the exercise professional to adopt a theoretical conceptualization of the construct identified in the first step (in this example: core affect). Earlier, we completed this step by adopting the conceptualization proposed by Russell and colleagues (Russell, 1980; Russell & Barrett, 1999). Accordingly, core affect is a combination two independent (i.e., orthogonal) dimensions, namely pleasure-displeasure (sometimes called affective valence) and activation (sometimes called arousal). Further, the dimension of pleasure-displeasure is considered to be bipolar (Russell & Carroll, 1999). The justification and support for this conceptualization, known as the circumplex model of affect (Russell, 1980) has been discussed elsewhere in greater detail (e.g., Ekkekakis, 2008, 2012, 2013). We encourage students and readers to consult with this literature for a more complete understanding of the theoretical underpinnings of the circumplex model of affect.

In the third step, a practitioner should select a measure that aligns with the conceptual model that was adopted (in this example, the circumplex model of affect; Russell, 1980). This conceptual model suggests that the exercise professional should choose one measure that corresponds with the dimension of activation, and a separate measure corresponding with pleasure-displeasure. Here, there are several candidate measures such as the Felt Arousal Scale (to measure activation; Svebak & Murgatroyd, 1985) and the Feeling Scale (ranging from "I feel... very bad to very good"; Hardy & Rejeski, 1989) or Empirical Valence Scale (ranging from most unpleasant imaginable to most pleasant imaginable; Lishner et al., 2008) to measure pleasure-displeasure. Notice how pleasure-displeasure is considered to be bipolar (Russell & Carroll, 1999) and the Feeling Scale and Empirical Valence Scale are both bipolar, indicating conceptual alignment.

Practitioners should consider the available psychometric evidence supporting its validity and reliability among individuals similar to their clients. Furthermore, practitioners should make practical considerations, such as the timing and frequency of measurement. For example, administering a 15-item questionnaire multiple times throughout an exercise session might be irritating and cause a large burden on participants (and could lead to them reporting feeling worse because of it). The candidate measures described in this section are all single-item scales that can be used repeatedly during an exercise session with little burden on the exerciser. This is advantageous because it provides the exercise practitioner latitude to choose to measure affective responses multiple times throughout an exercise session. We discuss this highly impactful consideration, next.

12.5.2. Timing and Measurement Frequency

The timing and frequency at which exercisers are asked to report their affective responses during exercise can profoundly influence the interpretations of those collecting the measurements. For example, analyses of affective responses only prior to and following exercise often suggest that people "feel better" after exercise, regardless of the intensity of exercise. However, when affective responses are also collected during exercise, it becomes clear that how one feels during exercise is far from universal (Backhouse et al., 2007; Bixby et al., 2001; Hall et al., 2002; Parfitt et al., 1994). Many exercisers report increased pleasure (or less displeasure) during low-to moderate-intensity exercise. In contrast, when exercise moves beyond moderate intensity, some exercisers report feeling better, while others report feeling worse. Finally, when exercise intensity is near maximal, most exercisers report feeling worse (Ekkekakis et al., 2005, 2011). However, regardless of how the exercisers reported feeling during exercise, within a very short time after ceasing near-maximal-intensity exercise (i.e., seconds to minutes), many people exhibit an "affective rebound" where they report feeling better (Box et al., 2020; Ekkekakis et al., 2005, 2011). As it turns out, one of the most common maxims of exercise science, that "exercise makes people feel better", may be partly attributable to suboptimal measurement timing and frequency. Importantly, practitioners who assume that affective responses to exercise are universal may create situations where their clients experience increasing displeasure during exercise but report feeling better afterwards simply because the unpleasant exercise experience has finished. If these and similar instances of displeasure during exercise are experienced repeatedly, it may reduce the probability that the exerciser continues to adhere.

Therefore, to achieve a clearer understanding of the affective responses to exercise of a client, we recommend that measurement of affective responses occurs frequently before, during, and after exercise (see Ekkekakis et al., 2020) When measuring pleasure or displeasure during exercise, a practitioner assesses how the exerciser feels right now, at the moment, to measure the moment-to-moment changes in the experienced pleasure or displeasure of the exerciser. For example, in addition to reading appropriate instructions to familiarize the client or patient, a practitioner using the Feeling Scale (Hardy & Rejeski, 1989) might ask the exerciser “How do you feel right now?” at regular intervals. A practitioner should aim to achieve a good balance of limiting burden (i.e., not asking too many questions) but also achieving an accurate representation of how the client feels during exercise. To balance burden with measurement accuracy, the time between measurement instances can and should be increased or decreased based on the mode, duration, and intensity of the exercise. For example, to most faithfully measure how one feels during a one-hour, steady-state bicycle ergometer test, it may make the most sense to measure the affective responses of a client at five-minute intervals (i.e., when affective responses are expected to be relatively stable). On the other hand, during a graded-exercise test to exhaustion, a practitioner should consider measuring affective responses just prior to the end of each stage to adequately capture the more rapid changes in affective responses that are to be expected with short stages of exercise. Depending on the graded-exercise test protocol, this might require measurements every one, two, or three minutes. Further illustration of the importance of measurement timing and frequency during a graded-exercise test is presented in Figure 12.1.

Figure 12.1: A Hypothetical Measurement Scenario During a Graded-Exercise Test

Note. A practitioner who measures affective valence only before and after exercise may conclude that the graded-exercise test to exhaustion made the exerciser feel more positive (i.e., a post-exercise value of “4”, compared to a pre-exercise value of “1”). This is illustrated by the dashed arrow. However, the reality is that the affective responses of the client were much more complex. The hypothetical client felt more positive at the beginning of the session, but affective responses rapidly became less pleasant and more unpleasant (indicated by negative values) after five minutes of exercise (around the gas-exchange ventilatory threshold), before rebounding and becoming more pleasant after the exercise had ended(possibly because the client was relieved to stop exercising). This figure is adapted from an illustration and discussion presented by Backhouse and colleagues (2007).

Another increasingly popular form of exercise, namely high-intensity interval exercise, comes with its own unique challenges for exercise professionals interested in understanding the affective responses of their clients. Because interval exercise is typically characterized by short bursts of vigorous-to supramaximal-intensity exercise interspersed with periods of low-intensity recovery, even small changes in measurement timing and frequency can obscure how well the affective responses of a client are understood. Unpleasant affective responses typically rebound following cessation of vigorous-to maximal-intensity intervals (Box et al., 2020). Therefore, to best capture the range of affective responses experienced during the frequent ups-and-downs of interval exercise, exercise professionals should ensure that their measurement timing and frequency also faithfully represents the anticipated patterns of affective responses to interval exercise. Specifically, each measurement of affect during interval exercise should be completed as close to the end of each interval as possible, as opposed to the beginning, middle, or afterward (Decker & Ekkekakis, 2017). Measurements only at the beginning or middle of intervals may not capture the anticipated peaks and valleys of affective responses during each interval. Furthermore, if measurements of affective responses about the exercise interval are solicited during the subsequent rest interval, the rapid affective rebound effect of resting may bias the interpretation of the affective experience of the client. This would also render the measure as one of remembered pleasure, or how pleasant or unpleasant one felt previously, rather than experienced pleasure (i.e., pleasure experienced at that moment). Examples of appropriate and problematic timing and frequency of measurement of affective responses during interval exercise are illustrated in Figures 12.2 and 12.3, respectively (also see Box & Petruzzello, 2021).

Figure 12.2. Appropriate Measurement Timing and Frequency During an Interval Exercise Session

Note. It is likely impossible to collect data at the precise end of each interval. However, to capture accurate data, practitioners should aim to measure as close to the end of each interval as possible. For example, 10-seconds before the end of each interval will help to improve the accuracy of the representation of the exercise affective experience. Using this approach, the exerciser appears to have highly variable affective responses to exercise (i.e., many affective rebounds during rest periods, and affective declines during intervals) and appears to experience negative affective valence toward the end of the exercise session (measured during Interval 5 and Interval 6). In this example, the practitioner measured affective responses consistently during both rest and interval periods.

Figure 12.3. Problematic Measurement Timing During the Same Interval Exercise Session

Note. In this case, measurements were collected at the beginning of exercise and during each rest interval of the same exercise session illustrated in Figure 12.2, obfuscating the true variability of affective responses throughout the interval exercise by missing the peaks and valleys of affective responses. Using this approach, the exerciser appears to have much more stable affective responses to exercise and appears to never experience negative affective valence.

12.5.3. Individual Variability

As reviewed earlier, individual variability in affective responses is to be expected, especially around the transition from moderate-to vigorous-intensity (i.e., near the ventilatory threshold; Ekkekakis, 2009; Ekkekakis et al., 2005, 2011). This dose-dependent variability is also discussed further in Chapter 11 by Jones and Zenko (2021). For example, in a seminal study by Van Landuyt and colleagues (2000), participants exercised for 30 minutes at 60% of their estimated VO₂ max, or maximal oxygen consumption. According to the American College of Sports Medicine (Garber et al., 2011), this corresponds to a “moderate” exercise intensity. Core affect was assessed during exercise using the Feeling Scale (Hardy & Rejeski, 1989) and the Felt Arousal Scale (Svebak & Murgatroyd, 1985). On average, there was no change in pleasure or displeasure during exercise, while activation increased. However, when the researchers examined individual differences in affective responses, they noted that 44.4% of participants felt increasing levels of pleasure during exercise, 14.3% exhibited essentially no change, and 41.3% felt less pleasant. Practically, these results suggest that it is prudent to analyze changes in affective responses during exercise on both the individual and group level, as opposed to just at the group level. If, for example, Van Landuyt and colleagues (2000) only assessed affective responses at the group level, they would have incorrectly concluded that exercise elicited essentially no change in the pleasure or displeasure experienced by the participants. However, because the researchers analyzed affective responses within each individual, we now know that some people will feel more pleasure during moderate-intensity exercise, while others will report feeling less.

12.5.4. Neutral Measurement by the Exercise Professional

The measurement fidelity of exercise affective responses can also be influenced by the person doing the measurement. When an exercise professional solicits affective responses from a client, their demeanor can subtly (or sometimes, clearly) bias the affective responses of the client. For example, simply smiling, nodding, thanking, or telling a client "good job" after receiving ratings on a measure may be interpreted by clients that the response is the one the exercise professional prefers them to make. In this situation, the exercisers may continue to rate their affective responses in this manner because they think it will make their trainer happy (i.e., provide the "socially desirable" response versus their true feelings). Exercise professionals who wish to collect accurate measures of affective responses to exercise among their clients should do their best to ask questions in a non-leading manner with neutral facial expressions and tone of voice. Whichever tone of voice and facial expression the professional adopts should be used consistently throughout exercise so that subtle changes in demeanor do not unduly influence the affective responses of the client.

Photo by Cliff Booth from Pexels

12.5.5. Environmental and Situational Considerations for Reducing Measurement Error

The effects that the exercise environment and individual situations can have on affective responses should be considered by exercise professionals. For example, the concept of affect can be, as with many psychological constructs, difficult to comprehend. If the person being asked about affect does not understand what is being asked, it is unlikely that they will provide data with any value. Therefore, using standardized instructions for measures, as well as asking whether the concept makes sense to the individual, can help to reduce instances of measurement error. One way that exercise practitioners can check for comprehension is to ask exercisers to explain what it is they are being asked to report. If there is a lack of understanding, the practitioner can try to clarify the concept further.

The presence of music and video can also influence how individuals perceive exercise (Lind et al., 2009; Jones et al., 2014). It has been demonstrated that when exercise is at the light- to moderate-intensity range, music and/or video can help the individual to dissociate, or distract, from the bodily sensations associated with exercise (e.g., increased breathing rate, lactate accumulation). Put another way, having a distraction, such as music and video, can help to attenuate unpleasant feelings during exercise. For some, if the same exercise intensity is performed without music and video, their exercise affective responses may not be the same as if they were in the presence of the distracting audiovisual stimuli. Therefore, exercise professionals should consider these effects and use distractors carefully and not as a means to mask displeasure during exercise.

Some people may find exercising in front of mirrors or others unpleasant and/or embarrassing, as they perceive (whether real or imagined) that their bodies are on display and are being evaluated by others (i.e., social-physique anxiety; Sabiston et al., 2014; also see Chapter 7 by Vani et al., 2021). Exercisers who are beginners or those with low motor skillfulness may also experience displeasure if they feel they are being evaluated critically by others. Exercise professionals should probe for and be cognizant of these potential effects among their clients, ideally prior to beginning exercise. For some individuals, exercising alone, away from mirrors and the prying eyes of others, may improve their affective responses. Similarly, leadership style (e.g.., focusing on health vs. focusing on appearance) of exercise leaders in group fitness settings may influence affective responses to exercise (Raedeke et al., 2007).

Last, but not least, exercise environments that are too hot or cold may also impact how the exerciser feels and should be considered. A hot room may raise body temperature to the point where exercise becomes increasingly uncomfortable. At the same time, heat typically leads to more sweating(Harrison, 1986) and could exacerbate the influence of social-physique anxiety, as exercisers may perceive that they appear to others to be struggling with exercise due to their excessive sweating. In short, there are many potential environmental and situational influences on affective responses to exercise; practitioners should be mindful of these influences and aim to assess affective responses in a standardized setting.