Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Dec 1.
Published in final edited form as: Health Psychol. 2017 Oct 9;36(12):1186–1194. doi: 10.1037/hea0000532

Physical activity and negative affective reactivity in daily life

Eli Puterman 1, Jordan Weiss 2, Mark R Beauchamp 3, Jacqueline Mogle 4, David M Almeida 5
PMCID: PMC5709165  NIHMSID: NIHMS916872  PMID: 29016149

Abstract

The results from experimental studies indicate that physically active individuals remain calmer and report less anxiety after the induction of a standardized stressor. The current study extends this research to real life, and examines whether daily physical activity attenuates negative affect that occurs in response to naturally occurring daily stressors. The current study used data from the second wave of the National Study of Daily Experiences, a sub-study of the second wave of the Midlife in the United States Study (MIDUS-II) of 2,022 individuals aged 33–84 questioned nightly for eight consecutive days about their general affect and affective responses to stressful events and their engagement in physical activity. Results indicated that while negative affect is significantly elevated on days with stressful events compared to days free of events in all individuals, these effects are attenuated in those who remain physically active when compared to those who were underactive. This was also true for any day participants were physically active. Importantly, negative affect in response to any specific stressor was reduced the closer in time that the stressor occurred to the bout of exercise in underactive participants, while, in active participants, negative affect in response to any stressor remained low throughout the entire day that participants reported that they were active. Given the significant mental and physical health implications of elevated affective reactivity observed in previous studies, the current study sheds further light on the importance of remaining physically active in times of stress.

Keywords: physical activity, affective reactivity, daily stressors

Regular physical activity is associated with improvements in a range of outcomes, including delayed cognitive decline (Sofi et al., 2011), reduced depressive symptomatology equal to pharmaceuticals (Hoffman et al., 2011), and reduced risk of early mortality (Arem et al., 2015). A growing body of research has sought to identify the extent to which physical activity boosts positive affect (PA) and/or reduces negative affect (NA). While it is clear that physical activity has direct effects on reports of daily experiences of affect, especially PA (Hogan, Mata, & Carstensen, 2013; Reed & Buck, 2009; Wichers et al., 2012), little is known about the extent to which physical activity might buffer against affective reactivity (i.e., changes in affect in response to experiences) in daily life. The current study examined whether physical activity mitigates NA in response to daily stressful events.

Literature Review

Physical Activity and Affect

Affect is the conscious experience of neurophysiological states, marked by its valence and arousal (Russell, 1980; Russell & Barrett, 1999). The effects of physical activity on PA have been consistently found within laboratory (Cox, Thomas, & Davis, 2001; Ekkekakis, Hall, & Petruzzello, 2005; Tate & Petruzzello, 1995), daily process field-based (Hogan et al., 2013; Mata et al., 2012; Poole, Steptoe, & Wawrzyniak, 2011; Wichers et al., 2012) and intervention (Reed & Buck, 2009) studies.

There is little evidence to support the effects of physical activity on the experiences of NA (Hogan et al., 2013; Mata et al., 2012; Poole et al., 2011; Wichers et al., 2012). For example, using ecological momentary assessments with repeated assessments 10 times per day for five consecutive days, Wichers and colleagues (2012) examined the lagged relationships between physical activity and PA and NA in over 500 female twins. Their results indicated that PA increased immediately after a bout of activity and lasted as long as three hours for women without histories of depression, and for an hour and a half for women with histories of depression. On the other hand, a bout of activity on any particular day had no immediate or lagged effects on NA. Mata and colleagues’ (2012) reported similar findings using similar methodologies (8 daily assessments for 7 days) in men and women.

Physical Activity and Affective Reactivity

Initial work by DeLongis, Coyne, Dakof, Folkman, and Lazarus (1982) revealed that the occurrence of daily stressors, ranging from work overload to interpersonal conflicts, had a significant impact on physical health (e.g., somatic complaints like chest pain, headaches) stronger than the effects of major life stressors. Since then, there has been an accrual of evidence that daily stressors predict daily physical and psychological well-being (Almeida & Kessler, 1998; DeLongis, Folkman, & Lazarus, 1988) and biological mediators of disease outcomes, such as daily cortisol output (Stawski, Cichy, Piazza, & Almeida, 2013) and inflammation (Gouin et al., 2011).

While daily events matter, how one responds affectively to these experiences appears to be especially important to health. Two recent studies, using data from the National Study of Daily Experiences, demonstrated that greater increases in NA on days with stressful events compared to days without stressful events was related to a greater incidence of affective disorders (Charles, Piazza, Mogle, Sliwinski, & Almeida, 2013) and chronic health conditions (Piazza, Charles, Sliwinski, Mogle, & Almeida, 2013) a decade later. In these studies, NA reactivity, defined as the changes in NA on stressor days compared to stressor-free days, reflects the expected byproduct (i.e., affect) of the stress and coping process.

How physically active and inactive individuals affectively respond to naturally occurring or laboratory induced stressors has been far less studied compared to the broader work on general affective states. Laboratory studies have shown that physically active individuals are less likely to experience increases in general NA and discrete negative emotions after the induction of a laboratory stressor compared to those who are physically inactive. Rimmele and colleagues (2007, 2009) demonstrated that both male athletes and sedentary untrained adult males increased in reported anxiety and decreased in calmness in response to a stressor induced in the laboratory. However, these effects were significantly less pronounced in the athletes. Mata, Hogan, Joorman, Waugh, and Gotlib’s (2013) laboratory study indicated that when depressed individuals were provided an opportunity to exercise prior to a sad mood induction task they subsequently displayed similar mitigated affective responses to non-depressed individuals, whereas those who were depressed but who did not exercise continued to display elevated NA with repeated inductions. To date, the extent to which physical activity mitigates NA reactivity to daily stressors in real life settings is unknown.

The Current Study

The current study conceptualized NA reactivity similar to naturalistic studies that differentiated between days with and days free of stressors (Charles et al., 2013) and to laboratory studies that examined affective responses related to specific stressors (Puterman et al., 2014). Building upon previous laboratory studies (Mata et al., 2013; Rimmele et al., 2007), the current study examined three main research questions. First, drawing from this extant body of research, it was hypothesized that NA reactivity would be mitigated in those who are regular exercisers when compared to those who are not. Second, it was hypothesized that NA reactivity would be alleviated to a greater extent on days that individuals report being physically active when compared to days without such activity. Finally, given the demonstrated relationship in time between physical activity and affect (Wichers et al., 2012), this study further tested whether mitigating effects of physical activity are a function of when the bout of exercise occurs in relation to the stressor. With this in mind, it was hypothesized that attenuation effects would be more pronounced when exercise occurs proximally to the experience of the stressor, when compared to when the exercise bout occurs more distally in time. In sum, by completing this series of detailed analyses the current study represented an opportunity to paint a naturalistic picture of physical activity as a means of activating stress resiliency, by comparing exercisers to non-exercisers and days with and without bouts of physical activity as they relate to general NA and stressor-specific NA reactivity.

Methods

The second wave of the Midlife in the United States Study (MIDUS II; http://midus.wisc.edu/), a longitudinal study of health and aging in the United States, contains a national sample of 4,963 individuals (ages: 32–84; 53% female). The original wave, MIDUS-I (n= 7,108), was surveyed between 1995–1996 and was reassessed between 2004–2006 (average follow-up interval: 9 years; range: 7.8–10.4). A random subsample of 3,600 MIDUS II participants was recruited to participate in the second wave of the National Study of Daily Experiences (NSDE II), a sub-study that examined their daily stressful experiences, affective states, and health behaviors. A total of 2,022 individuals (78% response rate) completed NSDE II (aged 33–84, mean 56 years of age (SD = 12.20); 57% female). MIDUS and NSDE study procedures were reviewed and approved by the Behavioral Sciences and the Health Sciences IRBs at the University of Wisconsin-Madison and The Pennsylvania State University.

Procedures

Participants in the NSDE II completed telephone interviews about their daily experiences over the course of eight consecutive evenings. The first and last day’s interviews lasted approximately 15–20 minutes and the other six interviews lasted approximately 10–15 minutes. They were asked questions about their affective states, stressors they experienced, health behaviors, and physical symptoms over the previous 24-hour period. After three consecutive non-response days, the research team terminated participation in NSDE II. A total of 14,912 interview days were completed in NSDE II’s 2,022 participants (average response: 7.4/8 days) (Almeida, McGonagle, & King, 2009).

Measures

Daily stressors

Daily stressors were assessed through the Daily Inventory of Stressful Events (Almeida, Wethington, & Kessler, 2002). The inventory asked respondents whether they experienced a stressor across seven domains: arguments with others, avoided an argument, stressor at work/school, stressor at home, discrimination, network stressor (event happened to a family member or close friend), or any other stressor. If any stressor occurred, respondents were asked what time it occurred. Responses were coded as ‘1’ if a stressor occurred in a specific domain or ‘0’ if not. Responses were also summed for number of stressors reported per day.

Active Day

Participants were asked how much time they spent “since this time yesterday” engaging in activities “that would cause you to break a sweat.” If affirmative, participants were then asked to report the time at which they began their physical activity and the amount of time they engaged in the activity. This brief daily physical activity measure was adapted from the 12-item NDSE II general measure of physical activity (Cotter and Lachman, 2010), which has been found to display good internal consistency and criterion validity. The U.S. Office of Disease Prevention and Health Promotion’s 2008 “Physical Activity Guidelines for Americans” (Physical Activity Guidelines Advisory Committee, 2008) recommends engaging in moderate levels of physical activity in bouts of no less than 10 minutes on most days that add up to 150 minutes per week. Moderate intensity is defined, in part, as a level of intensity that causes one to break a sweat. Thus, in the current study, a given day was considered an active day (1=Yes, 0 = No) if the bout of activity reported on that day made them break a sweat and lasted a minimum of 10 minutes. Thirty-three of the days with reported activity included bouts less than 10 minutes and were coded 0.

Activity Status

Bouts of activity that were reported as occurring for a minimum of 10 minutes were then summed across the 8 days of study participation to determine their physical activity status. The recommended 150 minutes per week (Physical Activity Guidelines Advisory Committee, 2008) equate to approximately 21.4 minutes per day, which would correspond to 171.4 minutes across the 8 days of study. Thus, participants who completed more than 172 minutes during their participation in the study were categorized as “active” and those who did not were categorized as “underactive.”

General NA

General NA was assessed by having participants rate on a scale from 0 (not at all) to 4 (all the time), how often they experienced each of the following 14 emotions during the previous 24 hours: restless or fidgety, nervous, worthless, so sad nothing could cheer me up, that everything was an effort, hopeless, lonely, afraid, jittery, irritable, ashamed, upset, angry, frustrated. These items come from the Nonspecific Psychological Distress Scale (Kessler et al., 2002) and a modified version of the Positive and Negative Affect Schedule (Watson, Clark, & Tellegan, 1988). General NA was calculated by taking the sum across all items, with higher scores reflecting higher NA (possible range scores 0 – 64; actual range from data 0 – 35.50). Measures derived from this instrument were found to be internally consistent (α = .91; see Raudenbush, Rowan, & Kang, 1991, for calculation method for nested data).

Stressor-specific NA

Stressor-specific NA was assessed by having participants rate on a scale from 0 (not at all) to 3 (all the time), the extent to which they felt each of the following four emotions in response to each of the stressors they experienced: angry, nervous or anxious, sad, shameful (Almeida, Stawski, & Cichy, 2011). Stressor-specific NA was calculated by taking the sum across four affective items and seven stressors, with higher scores reflecting higher negative affect (possible range scores 0 – 84; actual range from data 0 – 35). Measures derived from this instrument were found to be internally consistent (α = .85).

Covariates

Covariates included in all analyses were sex (Male = −1; Female = 1), age (centered around the sample mean = 56.24), education (less than high school = 0; only high school = 1; college or vocational degree = 2; postgraduate education in progress or completed = 3), racial identification (White = 0; Black = 1; Other = 2), chronic health conditions over previous year (sum score of 30 possible items ranging from respiratory, cardiovascular, arthritis, and other health conditions, centered around the sample mean = 2.54), daily smoking status (smoking day = 1; non-smoking day = −1), and daily alcohol use on each day of assessment [None = −1; Moderate =0 (1 to 2 drinks for males, 1 for females); Heavy = 1 (3 or more for males, 2 or more for females) based on daily alcohol consumption categories determined by U.S. Department of Health and Human Servies and U.S. Department of Agriculture (2015)].

Statistical Approach

All analyses were completed with SPSS version 23. Descriptive statistics were computed for all variables followed by bivariate analyses to determine whether potential covariates (sociodemographic, smoking and alcohol status, and chronic health conditions) were related to the a priori predictors (active day, activity status) and criterion variables (general and stressor-specific negative affect).

To test the hypotheses that NA reactivity is a function of activity status or active day, several multilevel analyses, including all covariates, were completed with random intercepts and fixed slopes using MIXED syntax with full maximum likelihood estimation in SPSS. Multilevel models with full maximum likelihood estimation are robust to missing data (Raudenbush & Bryk, 2002; Snijders & Bosker, 2012). General NA was regressed on (1) stressor-day (yes/no), (2) activity status (active/underactive), (3) activity day (yes/no) and the interactions between stressor day (yes/no) with activity status (active/underactive) and activity day (yes/no). Stressor-specific NA was regressed on either activity status or activity day. MIXED syntax estimates the outcome (Y) as a function of one predictor variable (X), a moderator variable (Z) and their interaction (X*Z), producing four relevant coefficients: (1) β0, (2) βx, (3) βz, and (4) βx*z. Interpretation of the coefficients is similar to interpreting regression coefficients with two independent dichotomous variables coded 0 and 1 and including their interaction (see Cohen, Cohen, West and Aiken (2003) Chapter 9 for a detailed explanation). Briefly, using general NA as the outcome and stressor-day and activity status as the two predictors, β0 is the estimated mean NA when stressor-day (variable 1) and activity status (variable 2) are both 0 (i.e., days without stressors in underactive participants). βx is the estimated change in NA from its estimated mean on days without stressors, coded 0, to days with stressors, coded 1, but only in underactive participants (coded 0). Adding the βx to β0 thus provides the estimated mean NA on days with stressors in underactive participants. A significant interaction βx*z suggests that the change in estimated mean NA on stressor days compared to days free or stressors is different for underactive compared to active participants. To compute the estimated mean NA on days with and without stressors in active participants, activity status is reverse coded and a MIXED model rerun. In this new model, β0 and βx are the estimated mean NA on stressor-free days and the change to stressor-days in active participants. Again, βx can be added to β0 to compute an estimated mean NA on days with stressors in the active participants.

A 3-level model was used to test the hypothesis that the mitigating effects of physical activity on NA reactivity is a function of when the bout of exercise occurs in relation to the stressor. Since multiple stressors could occur on each day for each person, the first level of the model (reflecting within day variation) included each individual stressor-specific NA and the time it occurred subtracted from the time the bout of exercise started, allowing to account for the non-independence in the data within persons and within days. The second level of the model included the variables that varied across days within individuals: that particular day’s length of bout of exercise, number of stressors, and daily smoking and drinking status. Finally, the third level included the variables that only varied between participants: age, sex, race, education, and chronic conditions. To test the third hypothesis, a multilevel growth curve model determined the trajectory of stressor-specific NA as a function of the time difference between the bout of exercise and the occurrence of each individual stressor. Growth curve modeling with time treated as linear (time) and curvilinear (time2) variables estimates an intercept, β0, an initial rate of change, βtime, and curvature, βtime2 and any significant time effects suggest that stressor-specific NA is a linear or curvilinear function of the time difference between the bout of exercise and occurrence of the stressor.

When appropriate, a Pseudo-R2 (Hedeker & Gibbons, 2006) was calculated to determine the proportional reduction in the unexplained variance between a more basic model that does not include any variables and ones that includes covariates and predictors. A Pseudo-R2 is determined by subtracting the estimate of covariate parameter for the residual or intercept (depending on level of analysis) for the more advanced model including the new variables from the more basic model and then dividing this value from the more basic model’s estimate.

Results

Few participants were missing sociodemographic data (0.2% were missing race or education data). Eighty-seven percent of participants completed 7 or all of the 8 days of the study and a very small minority (2.9%) of participants missed 5 or more phone calls. A greater number of missing days were related to younger age (r = −.12, p < .001) and lower education [ANOVA F (3, 2014) = 5.72, p = .001]. Post-hoc analyses demonstrated that those without a high school diploma completed fewer assessments than those at the three other education levels (M’s range = 0.46 to 0.51). Race was also related to greater number of days not completed [ANOVA F (2, 2013) = 33.17, p < .001], and post-hoc analyses demonstrated that black participants completed fewer assessments than other participants (M’s range = 0.51 to 0.72).

Descriptive statistics for all variables are presented in Table 1. Participants were, on average, aged 56 and were predominantly white (84.5%). Nearly half (48.2%) of the participants had high school or equivalent education and another 45.5% had completed college. 89.2% reported zero or moderate amounts of alcohol (no more than 8 for women or 16 for men) during the 8 days of the study. A majority of participants were non-smokers (84.8%). Participants also reported, on average, 2.54 chronic health conditions over the previous year.

Table 1.

Descriptive statistics for sociodemographic, health behaviors, health conditions and general and stressor-specific reactivity for all participants included in the National Study of Daily Experiences II.

Characteristic
Age, X (SD) 56.24 (12.2)
Sex
 Female, N (%) 1157 (57.2)
 Male, N (%) 865 (42.8)
Race
 White, N (%) 1703 (84.5)
 Black, N (%) 228 (11.3)
 Other, N (%) 85 (4.2)
Education
 <High School, N (%) 123 (6.3)
 High School or equivalent, N (%) 973 (48.2)
 College completed, N (%) 538 (26.7)
 Postgraduate, N (%) 379 (18.8)
Chronic Health Conditions, X (SD) 2.54 (2.48)
Drinking Status
 None, N (%) 1055 (52.2)
 Moderate, N (%) 784 (38.8)
 Heavy, N (%) 183 (9.1)
Smoking Status N (%)
 No, N (%) 1711 (84.8)
 Yes, N (%) 307 (15.2)
Active Days, Number, X (SD) 3.35 (2.51)
Activity Status
 Active, N (%) 1077 (53.3)
 Underactive, N (%) 945 (46.7)
General Negative Affect, X (SD)* 2.92 (3.89)
Stressor-Specific Negative Affect, X (SD)* 4.03 (2.88)
Open in a new tab
*

Note. All data for general and stressor-specific negative affect were first averaged per participant, and then a general average for the full sample taken.

Bivariate analyses (t-tests, Pearson correlations, ANOVAs) revealed that the number of days participants reported being physically active was related to being older (r = 0.07, p =.001), male [t (2016) = 2.14, p = .03], having fewer chronic health conditions (r = −0.08, p < .001), and not smoking [t (2016) = 2.81, p = .005]. Race (p = .24), education (p = .08), and alcohol status (p = .15) were unrelated to number of days being physically active. Similar significant associations between covariates and activity status (active vs. underactive) were apparent, whereby active participants were older [t (2020) = −2.05, p = .04], male [χ2 (2) = 26.62, p < .001] and had fewer health conditions [t (1953) = 2.82, p = .005]. Two-hundred and eight (10.3%) participants reported no days with any stressors and were excluded from the analyses since the occurrence of any stressors is integral to the concept of affective reactivity. Of those who reported any days with stressors (N = 1,814), fifty-six percent (N = 1,062) of participants were classified as physically active. Physical activity was not associated with exposure to daily stressors. Of those who reported any days with stressors, active and underactive participants had similar number of days with the occurrence of stressors, 3.21 days compared to 3.16 days respectively [t(1812) = 0.58, p = .56]. 39.5% of active days had a stressful event, with similar numbers on non-active days (38.3%) [χ2(1) = 2.24, p = .13]. Number of stressful events reported also did not differ on active days versus inactive days across all participants [t(14889) = −0.37, p = .71].

Preliminary analyses: Independent associations between covariates, main study variables and estimated general and stressor-specific NA

General and stressor-specific NA were regressed on the covariates in two separate multilevel models. In each model, whites were the referent group, as was postgraduate education. Younger adults (β = −0.62, S.E. = 0.01, 95%C.I. = −0.08, −0.05, p < .001) and those with more chronic health conditions (β = 0.47, S.E. = 0.03, 95%C.I. = 0.41, 0.53, p < .001) reported greater estimated general NA. Additionally, black participants (β = 0.63, S.E. = 0.25, 95%C.I. = 0.15, 1.12, p = 0.011) had greater estimated general NA than white participants, while participants in other racial groups (p = .65) did not differ from white participants. Participants with a college degree but did not pursue postgraduate education (β = −0.57, S.E. = 0.23, 95%C.I. = −1.02, −0.13, p = .012) and participants who completed only high school or GEDs (β = −0.44, S.E. = 0.21, 95%C.I. = −0.85, −0.03, p = .034) had lower estimated general NA than those who pursued postgraduate education. Those with less than a high school diploma had similar levels to those who pursued postgraduate education (p = .92). Estimated general NA on days that participants drank heavily (p = 0.07) or moderately (p = .59) was similar to days with no alcohol. Days participants reported smoking was related to greater estimated general NA than days when smoking was not reported (β = 0.40, S.E. = 0.10, 95%C.I. = 0.22, 0.58, p < .001). Significant estimates were similar with regards to stressor-specific NA, though women also had significantly greater estimates of stressor-specific NA compared to men (β = 0.35, S.E. = 0.07, 95%C.I. = 0.21, 0.49, p < .001), and on moderate drinking days, participants’ estimates for stressor-specific NA was higher than on days with no drinks (β = 0.63, S.E. = 0.16, 95%C.I. = 0.32, 0.94, p < .001). Including all the covariates accounted for 18% and 20% (Pseudo-R2 = 0.18 and 0.20) of the previously unexplained variation in general and stressor-specific NA, respectively.

In independent models including all covariates, general NA was regressed on (1) stressor day (NO = 0/Yes = 1; βstressorday), (2) activity status (NO = 0/Yes = 1; βactivitystatus), and (3) physically active day (NO = 0/Yes = 1; βactiveday). Average general NA on stressor days was significantly higher (X̄ = β0 + βstressorday = 4.68) than general NA on stressor-free days (X̄= 2.26) (β0 = 2.26, S.E. = 0.19, .95%C.I. = 1.89, 2.62, p < .001; βstressorday = 2.43, S.E. = 0.06, .95%C.I. = 2.31, 2.55, p < .001). Active participants reported lower general NA across all days (X̄= 3.24) than underactive participants (X̄= 3.76) (β0 = 3.76, S.E. = 0.22, .95%C.I. = 3.34, 4.19, p < .001; βactivitystatus = −0.52, S.E. = 0.15, .95%C.I. = −0.82, −0.22, p = .001). Days with a bout of exercise had significantly lower general NA (X̄= 3.37) than days without a bout of exercise (X̄= 3.58) (β0 = 3.58, S.E. = 0.20, .95%C.I. = 3.18, 3.97, p < .001; βactiveday = −0.21, S.E. =0.07, .95%C.I. = −0.34, −0.08, p = .002). Pseudo-R2’s demonstrated an additional 10%, 1%, and 1% reduction in unexplained variance in general NA models including stressor-day, activity status, and activity day, respectively.

Hypothesis 1. Individuals who reach levels of physical activity comparable to those recommended by the Centers for Disease Control and Prevention have attenuated affective reactivity compared to those who do not

General NA was regressed on two dichotomous variables, stressor day and activity status, their interaction, and all covariates. Results indicated a significant interaction, βinteraction = −0.38, S.E. =0.12, .95%C.I. = −0.62, −0.15, p = .001. As seen in Figure 1, general NA was significantly lower on stressor-free days (β0 = 2.10, S.E. = 0.20, .95%C.I. = 1.70, 2.49, p < .001) and affective reactivity was significantly attenuated (βstressorday = 2.26, S.E. = 0.08, .95%C.I. = 2.10, 2.41, p < .001) in physically active participants compared to those who were underactive (β0 = 2.44, S.E. = 0.21, .95%C.I. = 2.03, 2.85, p < .001; βstressorday = 2.64, S.E. = 0.09, .95%C.I. = 2.47, 2.81, p < .001). The difference in estimated general NA on days with stressors compared to days without stressors corresponded to a 14.4% reduction [100-((active βstressorday /underactive βstressorday)*100)] in affective reactivity between active and underactive participants.

Figure 1.

Figure 1

Open in a new tab

Estimated general negative affect on days with and without stressors in active and underactive participants (+/− SE).

Next, stressor-specific NA across the seven domains of daily stressors was regressed on activity status, all covariates, and additionally, number of stressors reported that day. Results indicated that active participants had similar levels of stressor-specific NA across the seven domains to the underactive participants (β0 = −0.29, S.E. = 0.19, .95%C.I. = −0.66, −0.08, p = .12; βactivitystatus = 0.02, S.E. = 0.12, .95%C.I. = −0.21, 0.25, p = .86).

Hypothesis 2. Affective reactivity is mitigated on active compared to inactive days

General NA was regressed on two variables, stressor-day and exercise-day, their interaction, covarying age, sex, race, education, chronic health conditions, daily smoking, daily alcohol, and activity status (recoded −1 and 1 so that estimated NA corresponds to the average person in the study). Results indicated a significant interaction (βinteraction = −0.45, S.E. = 0.11, .95%C.I. = −0.67, −0.23, p < .001). As seen in Figure 2, affective reactivity was significantly attenuated (i.e. 17.0% reduction) on active (β0 = 2.28, S.E. = 0.19, .95%C.I. = 1.90, 2.66, p < .001; βstressorday = 2.19, S.E. = 0.08, .95%C.I. = 2.02, 2.35, p < .001) compared to inactive (β0 = 2.30, S.E. = 0.19, .95%C.I. = 1.92, 2.68, p < .001; βstressorday = 2.64, S.E. = 0.08, .95%C.I. = 2.48, 2.79, p < .001) days.

Figure 2.

Figure 2

Open in a new tab

Estimated negative affect on days with and without stressors on days participants were active compared to days they were not (+/− SE).

Next, stressor-specific NA across the seven domains of stressors was regressed on whether the day was an active day or not, including all covariates, activity status, and the number of reported stressors. Results indicated that active and inactive days had similar levels of stressor-specific NA across the 6 domains s (β0 = −0.25, S.E. = 0.18, .95%C.I. = −0.61, 0.10, p = 0.17; βactiveday = −0.06, S.E. =0.09, .95%C.I. = −0.24, 0.12, p = .51).

Hypothesis 3. Stressor-specific negative affect is attenuated the more proximal the occurrence of the stressor is to the report of an activity bout

A 3-level multilevel growth curve model examined whether NA in response to any stressor was a function of the amount of time between the bout of exercise and the occurrence of the stressor, including all covariates, number of stressors reported during the day and activity status. Linear and curvilinear effects were not significant (p’s > .50). Results did reveal a significant interaction with activity status for the curvilinear effects only (βinteraction = − 0.01, S.E. = 0.003, .95%C.I. = −0.01, −0.001, p = .043) suggesting that active and underactive participants significantly differed in how the timing of a bout of activity, relative to the occurrence of a stressor, impacted NA in response to the stressor. Follow-up analyses within active participants suggested stable stressor-specific NA with no significant changes throughout the day (β0 = 2.66, S.E. = 0.24, .95%C.I. = 2.20, 3.12, p < .001; βtime = −0.004, S.E. = 0.02, .95%C.I. = −0.03, 0.19, p = .70; βtime2= −0.002, S.E. = 0.002, .95%C.I. = −0.005, 0.001, p = .27). For underactive participants, though, stressor-specific NA varied as a function of the timing between the stressor and the bout of activity, evidenced by a significant curvilinear time effect (β0 = 2.11, S.E. = 0.48, .95%C.I. = 1.17, 3.05, p < .001; βtime = −0.01, S.E. = 0.02, .95%C.I. = −0.05, 0.04, p = .77; βtime2= 0.01, S.E. = 0.003, .95%C.I. = 0.001, 0.013, p = .023). Figure 3 illustrates these effects.

Figure 3.

Figure 3

Open in a new tab

Stressor-related negative affect as a function of time between the occurrence of the stressor and when physical activity occurred among active and underactive participants.

Discussion

In the current study, physical activity was not related to exposures to stressors but was related to NA reactivity. While all participants significantly increased their experience of NA on stressor days compared to stressor-free days, individuals who were physically active did not increase to the same extent as less physically active individuals. Results indicated a 14% reduction in general NA on stressor days compared to days free of stressors between participants who were active compared to those who were underactive. General NA reactivity was also reduced (17%) on days in which participants reported a bout of exercise compared to days without exercise, independent of their general activity status. On days with stressors, reported specific NA in response to stressors did not differ by general activity status or activity day status. However, the current study did find evidence that on active days, specific NA experienced in response to each individual stressor was a function of the time a bout of activity occurred from the time a stressor occurred, but only for those who were underactive. This was not the case for those who reached a level of activity comparable to Centers for Disease Control and Prevention recommendations (Physical Activity Guidelines Advisory Committee, 2008) during the course of the study, as these individuals reported stable stressor-specific NA regardless of the time between the activity bout and the occurrence of the stressor.

Few studies have examined the effects of physical activity on psychological or physiological parameters in daily life. A recent 3-month exercise-based intervention study in inactive adults demonstrated that students randomized to an exercise group displayed elevated daily and nightly heart rate variability (HRV) during final exam week compared to those in the control group (von Haaren et al., 2016). HRV is a measure of the balance in autonomic control between the parasympathetic and sympathetic nervous systems and elevated values infer a calmer or less activated physiologic arousal state. The current study is the first to extend the advantages of physical activity, albeit observationally, to affective responses to real-life daily stressors. It is unclear, though, whether providing physically inactive adults with an opportunity to become physically active through intervention ameliorates daily NA reactivity.

Unclear, as well, is whether individuals who are physically active simply do not react to the same extent to stressors as those less active or whether they react similarly but recover faster. Laboratory research indicates that in response to an acute laboratory stressor, active individuals remain calmer compared to those less active. The current study indicates that NA reported in response to any specific stressor in physically active individuals is stable, independent of the time between the stressor and the activity bout whereas less active individuals experience dampened negative affect closer in time between the two events. It seems, then, that the buffering potential of a bout of exercise on the experience of NA on stressor days and in response to specific stressors becomes embedded in those who are active on a regular basis. This study does not reveal whether a bout of exercise suppresses or reconfigures the end-of-day recall of the intensity of the affective experience of a stressor that occurred hours earlier or mitigates the actual experience of a stressor in real-time. Laboratory stressor and exercise manipulations are required to disentangle these possibilities.

The current study has several strengths. The study was well powered, which allowed capturing the occurrence of a wide array of real life stressors among individuals who were both active and underactive. The use of phone calls to survey the participants also allowed the team of research assistants to clarify any participants’ questions that may otherwise have remained unanswered in self-reported online or paper surveys. The current study was further able to control for several important sociodemographic and behavioral factors associated with stressor exposure or physical activity, including age, sex, race, education, health conditions, and even smoking and alcohol use. Future research should attempt to capture dietary practices on a daily basis, based on findings that eating highly palatable foods improves both psychological and physiological stress responses (Macht & Mueller, 2007; Tomiyama, Dallman, & Epel, 2011).

The current study, however, does have limitations. End-of-day recall represents a significantly improved assessment technique over retrospective reports to understand the unfolding of stress on a daily basis and affect (Anita DeLongis, Hemphill, & Lehman, 1992), though ecological momentary assessments may be more appropriate than end of day recall to adequately assess the relationship between a bout of activity and NA in response to a stressful event, as done in the final analysis. Furthermore, the emotional and time burden on participants when assessing daily mood and stressors requires brevity in the questions, usually a one-item face valid question amenable to repeated assessment (Iida, Shrout, Laurenceau, & Bolger, 2012).

Another limitation is that physical activity was assessed with a self-report measure. Most self-report instruments, such as the International Physical Activity Questionnaire (Booth, 2000), provide a retrospective assessment of physical activity accrued over a prolonged period of time, typically one week. In light of concerns with memory decay associated with such weekly retrospective measures, daily assessments of life experiences (Almeida et al., 2002; Kahneman, Krueger, Schkade, Schwarz, & Stone, 2004) have received increased usage within the behavioral sciences. Additionally, while the brief measure of physical activity embedded within the NSDE II interview protocol aligns well with current physical activity guidelines (Physical Activity Guidelines Advisory Committee, 2008) and displays sound content validity (vis a vis content relevance and representativeness (Messick & Samuel, 1995)), it should be noted that this brief measure has not been formally tested for convergent validity (e.g., in relation to objective measures such as accelerometry). In the current study, participants were asked whether they engaged in activities that made them break a sweat, one important factor in determining intensity, but not the only one. Intensity (light, moderate, or vigorous) is also defined by shortness of breath, fatigue and capacity to carry on a conversation. Results indicate that 56% of the sample was active during the 8 days of data collection using the definition that only included sweating and with reference to the recommended guidelines of 150 minutes of moderate-to- vigorous activity per week with each bout lasting at least 10 minutes, similar to the U.S. national average reported in 2011 (Centers for Disease Control and Prevention, 2015). Studies using accelerometers for objective activity measurement indicate a significantly reduced percent of the population active at levels recommended by the CDC (Tucker, Welk, & Beyler, 2011). Future studies should capture activity during the day with multi-method approaches including accelerometry data for real-time activity and repeated daily self-reported assessments for activities not captured adequately by these devices (i.e. swimming and biking), and that define moderate intensity with the full range of defining components. Such an enterprise would enable researchers to better understand, with greater sensitivity, the relationship between physical activity, the occurrence of stressful events and negative affect as they unfold throughout the day.

Physical activity has a wide range of beneficial effects on mental and physical health. The current study demonstrated that negative affective reactivity, previously demonstrated to be significantly associated with long-term mental and physical health problems, is less pronounced if individuals are physically active. While physical activity may not reduce the likelihood of the occurrence of stressful events, it may bolster people’s capacity to psychologically manage the stressor in a more effective way. The current study’s findings are observational in nature however, and intervention studies investigating changes in affective reactivity in previously inactive individuals who are provided the opportunity to become active are strongly needed.

Acknowledgments

This research was undertaken, in part, thanks to funding to EP from the Canada Research Chairs program and, in part, from the National Heart, Lung and Blood Institute of the National Institutes of Health under award number R00 HL 109247. The Midlife in the United States (MIDUS) investigation was supported by NIA grants P01-AG020166 and R01-AG019239. The original study was supported by the John D. and Catherine T. MacArthur Foundation Research Network on Successful Midlife Development. The funding sources had no involvement in the study design; data collection, analysis, or interpretation; nor the writing and submission of this manuscript.

Contributor Information

Eli Puterman, University of British Columbia.

Jordan Weiss, University of Pennsylvania.

Mark R. Beauchamp, University of British Columbia

Jacqueline Mogle, Pennsylvania State University.

David M. Almeida, Pennsylvania State University

References

  1. Almeida DM, Kessler R. Everyday stressors and gender differences in daily distress. Journal of Personality & Social Psychology. 1998;75(3):670–680. doi: 10.1037//0022-3514.75.3.670. [DOI] [PubMed] [Google Scholar]
  2. Almeida DM, McGonagle K, King H. Assessing daily stress processes in social surveys by combining stressor exposure and salivary cortisol. Biodemography and Social Biology. 2009;55(2):219–37. doi: 10.1080/19485560903382338. http://doi.org/10.1080/19485560903382338. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Almeida DM, Stawski RS, Cichy K. Combining checklist and interview approaches for assessing daily stressors: The Daily Inventory of Stressful Events. In: Contrada R, Baum A, editors. The Handbook of Stress Science: Biology, Psychology, and Health. New York: Springer; 2011. pp. 583–595. [Google Scholar]
  4. Almeida DM, Wethington E, Kessler RC. The Daily Inventory of Stressful Events: An interview-based approach for measuring daily stressors. Assessment. 2002;9(1):41–55. doi: 10.1177/1073191102091006. http://doi.org/10.1177/1073191102091006. [DOI] [PubMed] [Google Scholar]
  5. Arem H, Moore SC, Patel A, Hartge P, Berrington de Gonzalez A, Visvanathan K, … Matthews CE. Leisure time physical activity and mortality: A detailed pooled analysis of the dose-response relationship. JAMA Internal Medicine. 2015 doi: 10.1001/jamainternmed.2015.0533. http://doi.org/10.1001/jamainternmed.2015.0533. [DOI] [PMC free article] [PubMed]
  6. Bogg T, Slatcher RB. Activity mediates conscientiousness’ relationship to diurnal cortisol slope in a national sample. Health Psychology: Official Journal of the Division of Health Psychology, American Psychological Association. 2015;34(12):1195–9. doi: 10.1037/hea0000243. http://doi.org/10.1037/hea0000243. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Booth M. Assessment of physical activity: an international perspective. Research Quarterly for Exercise and Sport. 2000;71(2 Suppl):S114–20. [PubMed] [Google Scholar]
  8. Centers for Disease Control and Prevention. Nutrition, physical activity and obesity data, trends and maps database. 2015 Retrieved from https://nccd.cdc.gov/NPAO_DTM.
  9. Charles ST, Piazza JR, Mogle J, Sliwinski MJ, Almeida DM. The wear and tear of daily stressors on mental health. Psychological Science. 2013;24(5):733–41. doi: 10.1177/0956797612462222. http://doi.org/10.1177/0956797612462222. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Cohen J, Cohen P, West SG, Aiken LS, Mahwah USNJ. Applied multiple regression/correlation analysis for the behavioral sciences. 3. Lawrence Erlbaum Associates Publishers; 2003. http://doi.org/10.4324/9780203774441. [Google Scholar]
  11. Cotter KA, Lachman ME. No strain, no gain: psychosocial predictors of physical activity across the adult lifespan. Journal of Physical Activity & Health. 2010;7(5):584–94. doi: 10.1123/jpah.7.5.584. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cox R, Thomas T, Davis J. Positive and negative affect associated with an acute bout of aerobic exercise. Journal of Exercise Physiology. 2001;4(4):13–20. [Google Scholar]
  13. DeLongis A, Coyne JC, Dakof G, Folkman S, Lazarus RS. Relationship of daily hassles, uplifts, and major life events to health status. Health Psychology. 1982;1(2):119–136. http://doi.org/10.1037/0278-6133.1.2.119. [Google Scholar]
  14. DeLongis A, Folkman S, Lazarus RS. The impact of daily stress on health and mood: psychological and social resources as mediators. Journal of Personality and Social Psychology. 1988;54(3):486–95. doi: 10.1037//0022-3514.54.3.486. [DOI] [PubMed] [Google Scholar]
  15. DeLongis A, Hemphill K, Lehman D. A Structured Diary Methodology for the Study of Daily Events. In: Bryant FB, Edwards J, Tindale RS, Posavac EJ, Heath L, Henderson E, Suarez-Balcazar Y, editors. Methodological issues in applied social psychology. Vol. 2. Boston, MA: Springer US; 1992. pp. 83–109. http://doi.org/10.1007/978-1-4899-2308-0. [Google Scholar]
  16. 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. Journal of Sports Sciences. 2005;23(5):477–500. doi: 10.1080/02640410400021492. http://doi.org/10.1080/02640410400021492. [DOI] [PubMed] [Google Scholar]
  17. Gouin JP, Glaser R, Malarkey WB, Beversdolf D, Kiecolt-Glaser JK, Gouin Glaser R, Malarkey WB, Beversdorf D, Kiecolt-Glaser JJ-P, Beversdorf D. Chronic stress, daily stressors, and circulating inflammatory markers. Health Psychology. 2011;31(2):264–268. doi: 10.1037/a0025536. http://doi.org/10.1037/a0025536. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Hedeker D, Gibbons RD. Longitudinal Data Analysis. Hoboken: Wiley; 2006. [Google Scholar]
  19. Hoffman BM, Babyak MA, Craighead WE, Sherwood A, Doraiswamy PM, Coons MJ, Blumenthal JA. Exercise and pharmacotherapy in patients with major depression: one-year follow-up of the SMILE study. Psychosomatic Medicine. 2011;73(2):127–33. doi: 10.1097/PSY.0b013e31820433a5. http://doi.org/10.1097/PSY.0b013e31820433a5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hogan CL, Mata J, Carstensen LL. Exercise holds immediate benefits for affect and cognition in younger and older adults. Psychology and Aging. 2013;28(2):587–94. doi: 10.1037/a0032634. http://doi.org/10.1037/a0032634. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Iida M, Shrout PE, Laurenceau J-P, Bolger N. APA handbook of research methods in psychology, Vol 1: Foundations, planning, measures, and psychometrics. Washington: American Psychological Association; 2012. Using diary methods in psychological research; pp. 277–305. http://doi.org/10.1037/13619-016. [Google Scholar]
  22. Kahneman D, Krueger AB, Schkade DA, Schwarz N, Stone AA. A Survey Method for Characterizing Daily Life Experience: The Day Reconstruction Method. Science. 2004;306(5702):1776–1780. doi: 10.1126/science.1103572. http://doi.org/10.1126/science.1103572. [DOI] [PubMed] [Google Scholar]
  23. Lachman ME, Agrigoroaei S. Promoting Functional Health in Midlife and Old Age: Long-Term Protective Effects of Control Beliefs, Social Support, and Physical Exercise. PLoS ONE. 2010;5(10):e13297. doi: 10.1371/journal.pone.0013297. http://doi.org/ARTN. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Macht M, Mueller J. Immediate effects of chocolate on experimentally induced mood states. Appetite. 2007;49(3):667–74. doi: 10.1016/j.appet.2007.05.004. http://doi.org/10.1016/j.appet.2007.05.004. [DOI] [PubMed] [Google Scholar]
  25. Mata J, Hogan CL, Joormann J, Waugh CE, Gotlib IH. Acute exercise attenuates negative affect following repeated sad mood inductions in persons who have recovered from depression. Journal of Abnormal Psychology. 2013;122(1):45–50. doi: 10.1037/a0029881. http://doi.org/10.1037/a0029881. [DOI] [PubMed] [Google Scholar]
  26. Mata J, Thompson RJ, Jaeggi SM, Buschkuehl M, Jonides J, Gotlib IH. Walk on the bright side: physical activity and affect in major depressive disorder. Journal of Abnormal Psychology. 2012;121(2):297–308. doi: 10.1037/a0023533. http://doi.org/10.1037/a0023533. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Messick S, Samuel Validity of psychological assessment: Validation of inferences from persons’ responses and performances as scientific inquiry into score meaning. American Psychologist. 1995;50(9):741–749. http://doi.org/10.1037/0003-066X.50.9.741. [Google Scholar]
  28. Physical Activity Guidelines Advisory Committee. Physical Activity Guidelines Advisory Committee Report, 2008. Washington DC: 2008. [Google Scholar]
  29. Piazza JR, Charles ST, Sliwinski MJ, Mogle J, Almeida DM. Affective reactivity to daily stressors and long-term risk of reporting a chronic physical health condition. Annals of Behavioral Medicine. 2013;45(1):110–20. doi: 10.1007/s12160-012-9423-0. http://doi.org/10.1007/s12160-012-9423-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Poole L, Steptoe A, Wawrzyniak AJ. Associations of objectively measured physical activity with daily mood ratings and psychophysiological stress responses in women. Psychophysiology. 2011;48(8):1165–1172. doi: 10.1111/j.1469-8986.2011.01184.x. [DOI] [PubMed] [Google Scholar]
  31. Puterman E, Epel E, O’Donovan A, Prather AA, Aschbacher K, Dhabhar FS. Anger is associated with increased IL-6 stress reactivity in women, but only among those low in social support. International Journal of Behavioral Medicine. 2014;21(6):936–945. doi: 10.1007/s12529-013-9368-0. http://doi.org/10.1007/s12529-013-9368-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Raudenbush SW, Bryk AS. Hierarchical Linear Models: Applications and Data Analysis Methods. 2. Sage Publications Inc; 2002. [Google Scholar]
  33. Raudenbush SW, Rowan B, Kang SJ. A multilevel, multivariate model for studying school climate with estimation via the EM algorithm and application to U.S. high-school data. Journal of Educational and Behavioral Statistics. 1991;16(4):295–330. http://doi.org/10.3102/10769986016004295. [Google Scholar]
  34. Reed J, Buck S. The effect of regular aerobic exercise on positive-activated affect: A meta-analysis. Psychology of Sport and Exercise. 2009;10(6):581–594. http://doi.org/10.1016/j.psychsport.2009.05.009. [Google Scholar]
  35. Rimmele U, Seiler R, Marti B, Wirtz PH, Ehlert U, Heinrichs M. The level of physical activity affects adrenal and cardiovascular reactivity to psychosocial stress. Psychoneuroendocrinology. 2009;34(2):190–198. doi: 10.1016/j.psyneuen.2008.08.023. [DOI] [PubMed] [Google Scholar]
  36. Rimmele U, Zellweger B, Marti B, Seiler R, Mohiyeddini C, Ehlert U, Heinrichs M. Trained men show lower cortisol, heart rate and psychological responses to psychosocial stress compared with untrained men. Psychoneuroendocrinology. 2007;32(6):627–35. doi: 10.1016/j.psyneuen.2007.04.005. http://doi.org/10.1016/j.psyneuen.2007.04.005. [DOI] [PubMed] [Google Scholar]
  37. Russell JA. A circumplex model of affect. Journal of Personality & Social Psychology. 1980;39(6):1161–1178. [Google Scholar]
  38. Russell JA, Barrett LF. Core affect, prototypical emotional episodes, and other things called emotion: Dissecting the elephant. Journal of Personality and Social Psychology. 1999;76(5):805–819. doi: 10.1037//0022-3514.76.5.805. [DOI] [PubMed] [Google Scholar]
  39. Snijders TAB, Bosker RJ, Roel J. Multilevel analysis: an introduction to basic and advanced multilevel modeling. Sage; 2012. [Google Scholar]
  40. Sofi F, Valecchi D, Bacci D, Abbate R, Gensini GF, Casini A, Macchi C. Physical activity and risk of cognitive decline: a meta-analysis of prospective studies. Journal of Internal Medicine. 2011;269(1):107–17. doi: 10.1111/j.1365-2796.2010.02281.x. http://doi.org/10.1111/j.1365-2796.2010.02281.x. [DOI] [PubMed] [Google Scholar]
  41. Stawski RS, Cichy KE, Piazza JR, Almeida DM. Associations among daily stressors and salivary cortisol: findings from the National Study of Daily Experiences. Psychoneuroendocrinology. 2013;38(11):2654–65. doi: 10.1016/j.psyneuen.2013.06.023. http://doi.org/10.1016/j.psyneuen.2013.06.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Stephan Y, Sutin AR, Terracciano A. Physical activity and personality development across adulthood and old age: Evidence from two longitudinal studies. Journal of Research in Personality. 2014;49:1–7. http://doi.org/10.1016/j.jrp.2013.12.003. [Google Scholar]
  43. Tate AK, Petruzzello SJ. Varying the intensity of acute exercise: implications for changes in affect. The Journal of Sports Medicine and Physical Fitness. 1995;35(4):295–302. [PubMed] [Google Scholar]
  44. Tomiyama AJ, Dallman MF, Epel ES. Comfort food is comforting to those most stressed: Evidence of the chronic stress response network in high stress women. Psychoneuroendocrinology. 2011;36(10):1513–9. doi: 10.1016/j.psyneuen.2011.04.005. http://doi.org/10.1016/j.psyneuen.2011.04.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Tucker JM, Welk GJ, Beyler NK. Physical activity in U.S.: adults compliance with the Physical Activity Guidelines for Americans. American Journal of Preventive Medicine. 2011;40(4):454–61. doi: 10.1016/j.amepre.2010.12.016. http://doi.org/10.1016/j.amepre.2010.12.016. [DOI] [PubMed] [Google Scholar]
  46. U.S. Department of Health and Human Servies and U.S. Department of Agriculture. 2015-202 Dietary Guidelines for Americans. 8. Washington DC: 2015. [Google Scholar]
  47. von Haaren B, Ottenbacher J, Muenz J, Neumann R, Boes K, Ebner-Priemer U. Does a 20-week aerobic exercise training programme increase our capabilities to buffer real-life stressors? A randomized, controlled trial using ambulatory assessment. European Journal of Applied Physiology. 2016;116(2):383–94. doi: 10.1007/s00421-015-3284-8. http://doi.org/10.1007/s00421-015-3284-8. [DOI] [PubMed] [Google Scholar]
  48. Watson D, Clark LA, Tellegan A. Development and validation of brief measures of positive and negative affect: the PANAS scales. Journal of Personality and Social Psychology. 1988;54(6):1063–1070. doi: 10.1037//0022-3514.54.6.1063. http://doi.org/doi:10.1037/0022-3514.54.6.1063. [DOI] [PubMed] [Google Scholar]
  49. Wichers M, Peeters F, Rutten BPF, Jacobs N, Derom C, Thiery E, … van Os J. A time-lagged momentary assessment study on daily life physical activity and affect. Health Psychology. 2012;31:135–144. doi: 10.1037/a0025688. http://doi.org/10.1037/a0025688. [DOI] [PubMed] [Google Scholar]

RESOURCES