Physical Activity and Sedentary Behavior During COVID-19: Trajectory and Moderation by Personality

Method

Results

Descriptive statistics for the two samples are in Table 1. All data satisfied the criteria for parametric analysis (e.g., skewness below 2).

OPEN IN VIEWER

Table 1.

Descriptive Statistics for the Two Samples.

Variable	UAS	PBSR
	Mean (SD) or %	Mean (SD) or %
Age	48.62 (16.08)	43.75 (18.64)
Sex (% female)	58	54
Education	11.30 (2.26)	3.90 (1.60)
Race (% White)	78	62
Ethnicity (% Hispanic)	16	16
Wave 1 physical activity	2.06 (2.18)	—
Wave 2 physical activity a	3.73 (2.35)	—
Wave 3 physical activity b	3.71 (2.30)	—
Wave 4 physical activity c	3.77 (2.30)
Wave 1 sedentary behaviors	—	6.44 (3.45)
Wave 2 sedentary behaviors d	—	7.10 (3.84)
Wave 3 sedentary behaviors e	—	7.23 (3.75)
Neuroticism	2.73 (0.82)	2.76 (0.78)
Extraversion	3.24 (0.79)	3.11 (0.63)
Openness	3.61 (0.62)	3.46 (0.63)
Agreeableness	3.98 (0.61)	3.61 (0.64)
Conscientiousness	4.00 (0.62)	3.71 (0.72)

Note. N _UAS = 6,702; N _PBSR = 3,992. UAS = Understanding America Study; PBSR = Psychological, Behavioral, and Social Response.

^a N = 5,156. ^b N = 5,830. ^c N = 5,874. ^d N = 1,770. ^e N = 1,759.

In the UAS, the average number of days of physical activity in the last week increased from about 2 days a week in early March to almost 4 days a week by the end of May (Table 1). The MLM analysis confirmed that this increase in physical activity during the pandemic was significant (Table 2). A quadratic term was added in an additional analysis, which revealed that physical activity followed a nonlinear trajectory and tended to plateau over time (b = −0.39, SE = .01, p < .001). Personality was associated with both the intercept and slope of physical activity. Consistent with the literature on personality and physical activity, higher neuroticism was related to a lower mean level of physical activity, whereas higher extraversion, openness, agreeableness, and conscientiousness were associated with a higher mean level of physical activity level (i.e., the association with the intercept). Consistent with the hypothesis, neuroticism was associated negatively with the slope of physical activity, whereas conscientiousness had a positive association with the slope; agreeableness was also positively associated with the slope (not hypothesized). These results suggested that higher neuroticism attenuated the increase in physical activity, whereas higher conscientiousness and agreeableness were related to a greater increase in physical activity over time (Table 2). Neither extraversion nor openness was related to the slope of physical activity.

OPEN IN VIEWER

Table 2.

Multilevel Modeling Results for Changes in Physical Activity (Understanding America Study).

Variable	b (SE)	p Value	b (SE)	p Value	b (SE)	p Value	b (SE)	p Value	b (SE)	p Value
Intercept	2.97 (.036)	<.01	2.98 (.036)	<.01	2.98 (.036)	<.01	2.99 (.036)	<.01	2.99 (.036)	<.01
Time	0.53 (.009)	<.01	0.53 (.009)	<.01	0.53 (.009)	<.01	0.53 (.009)	<.01	0.53 (.009)	<.01
Age	0.01 (.001)	<.01	0.01 (.001)	<.01	0.01 (.001)	<.01	0.01 (.001)	<.01	0.01 (.001)	<.01
Sex (male)	0.18 (.05)	<.01	0.25 (.04)	<.01	0.22 (.05)	<.01	0.28 (.05)	<.01	0.27 (.04)	<.01
Education	0.11 (.01)	<.01	0.12 (.01)	<.01	0.11 (.01)	<.01	0.12 (.01)	<.01	0.11 (.01)	<.01
Race (White)	0.37 (.05)	<.01	0.34 (.05)	<.01	0.35 (.05)	<.01	0.34 (.05)	<.01	0.35 (.05)	<.01
Ethnicity (Hispanic)	−0.50 (.06)	<.01	−0.50 (.06)	<.01	−0.48 (.06)	<.01	−0.47 (.06)	<.01	−0.48 (.06)	<.01
Neuroticism	−0.20 (.02)	<.01
Neuroticism × Time	−0.02 (.009)	.016
Extraversion			0.24 (.02)	<.01
Extraversion × Time			0.01 (.009)	.19
Openness					0.16 (.02)	<.01
Openness × Time					0.001 (.009)	.87
Agreeableness							0.14 (.02)	<.01
Agreeableness × Time							0.019 (.009)	.04
Conscientiousness									0.28 (.02)	<.01
Conscientiousness × Time									0.05 (.009)	<.01

Note. Coefficients are unstandardized coefficients (SE in parentheses).

In the PBSR, the number of hours spent being sedentary increased from about 6.5 hr/day before the pandemic to about 7.25 hr/day during the pandemic (Table 1). The MLM analysis confirmed that this increase in sedentary behavior was significant (Table 3). Further analysis revealed that sedentary behaviors followed a nonlinear trajectory, illustrated by an increase and a plateau (b = −0.19, SE = .08, p < .05). Consistent with the literature on personality and sedentary behavior, higher neuroticism was related to a higher mean level of sedentary behavior, whereas higher extraversion, agreeableness, and conscientiousness were related to a lower mean level of sedentary behavior. Openness was unrelated to the intercept of sedentary behavior. In contrast to the hypothesis, there was no association between any of the five traits and the slope of sedentary behavior (Table 3).

OPEN IN VIEWER

Table 3.

Multilevel Modeling Results for Changes in Sedentary Behaviors (Psychological, Behavioral, and Social Response Study).

Variable	b (SE)	p Value	b (SE)	p Value	b (SE)	p Value	b (SE)	p Value	b (SE)	p Value
Intercept	6.66 (.07)	<.01	6.69 (.07)	<.01	6.67 (.07)	<.01	6.67 (.07)	<.01	6.67 (.07)	<.01
Time	0.35 (.04)	<.01	0.34 (.04)	<.01	0.34 (.04)	<.01	0.35 (.04)	<.01	0.36 (.04)	<.01
Age	0.01 (.003)	<.01	0.006 (.003)	.04	0.004 (.003)	.24	0.006 (.003)	.04	0.01 (.003)	<.01
Sex (male)	0.21 (.10)	.04	0.14 (.10)	.16	0.10 (.10)	.36	0.04 (.11)	.71	0.01 (.10)	.93
Education	0.03 (.03)	.37	0.02 (.03)	.40	−0.002 (.03)	.96	−0.006 (.03)	.85	0.006 (.03)	.86
Race (White)	0.42 (.11)	<.01	0.47 (.11)	<.01	0.48 (.11)	<.01	0.48 (.11)	<.01	0.46 (.11)	<.01
Ethnicity (Hispanic)	−0.03 (.15)	.83	0.04 (.15)	.79	0.02 (.15)	.92	0.01 (.15)	.92	−0.00 (.15)	.99
Neuroticism	0.50 (.05)	<.01
Neuroticism × Time	0.002 (.04)	.95
Extraversion			−0.61 (.05)	<.01
Extraversion × Time			0.01 (.04)	.75
Openness					−0.00 (.05)	.98
Openness × Time					0.006 (.04)	.88
Agreeableness							−0.15 (.05)	<.01
Agreeableness × Time							−0.02 (.04)	.69
Conscientiousness									−0.42 (.06)	<.01
Conscientiousness × Time									−0.03 (.04)	.42

Note. Coefficients are unstandardized coefficients (SE in parentheses).

Discussion

The present study examined the change in physical activity and sedentary behavior during the acute phase of the COVID-19 pandemic and tested whether personality traits were related to change in these behaviors. Both physical activity and sedentary behavior increased overall. Lower neuroticism and higher conscientiousness and agreeableness were associated with a greater increase in physical activity over time. Higher neuroticism and lower extraversion, agreeableness, and conscientiousness were also associated with average levels of sedentary behavior but not to change over time. These results extend recent findings on both changes in patterns of physical activity and sedentary behavior found in Europe to the United States, and the association between personality and precautionary behaviors (Aschwanden et al., in press; Blagov, 2020; Bogg & Milad, 2020) by showing that personality traits are also related to the adoption of health-related behaviors during the pandemic.

This study builds on preliminary evidence from Europe on changes in patterns of physical activity and sedentary behavior during the lockdown caused by the pandemic. Specifically, Cheval and colleagues (2020) found that engagement in vigorous physical activity declined and sedentary behavior increased during the pandemic, whereas moderate physical activity increased. Although UAS did not differentiate between moderate and vigorous physical activity, we found an increase in the average number of days that individuals exercised during the pandemic. The number of hours spent sedentary likewise increased. This pattern supports anecdotal evidence reported in the news that although gyms and other fitness centers were closed across the country during the lockdown, there has been an increase in the time spent outside, particularly walking around one’s neighborhood compared to before the pandemic (San Francisco Chronicle, 2020).

The relation between personality and physical activity in the present study was consistent with the larger literature in this area: Individuals higher in neuroticism tend to engage in less physical activity and more sedentary behavior, whereas individuals higher in conscientiousness, extraversion, and openness tend to be more active (Allen, Magee, et al., 2017; Allen, Walter, & McDermott, 2017; Rhodes & Smith, 2006; Sutin et al., 2016; Wilson & Dishman, 2015). In addition, the results support the link between conscientiousness and less sedentary behaviors (Allen, Walter, & McDermott, 2017). The present study adds that these basic associations replicate during the pandemic and that personality is associated with changes in physical activity across the pandemic. Again, consistent with the broader literature on personality and health, we found that lower neuroticism and higher conscientiousness were associated with a greater increase in physical activity.

There may be a number of reasons why neuroticism and conscientiousness were associated with changes in physical activity. Individuals higher in neuroticism, for example, express more concerns about the virus, including concerns about contracting COVID-19 (Aschwanden et al., in press; Bogg & Milad, 2020), and have experienced more negative affect during the crisis (Kroencke et al., 2020). These concerns and negative affect may lead individuals higher in neuroticism to avoid physical activity and outside environments that may have culminated in declines in physical activity. Individuals higher in conscientiousness, in contrast, have engaged in more preventative behavior to avoid catching the virus (Aschwanden et al., in press; Blagov, 2020; Bogg & Milad, 2020) and may be particularly likely to follow the physical activity guidelines during the COVID-19 crisis because of their propensity to comply with social norms (Jackson et al., 2010). Further, as the crisis constrained daily life, their self-discipline and organization may facilitate the ability and dedication to maintain an active lifestyle.

Some of the findings were surprising. For example, we did not expect agreeableness to be associated with either the baseline level of physical activity or increases in physical activity over time. This result is, in fact, in contrast to cross-sectional studies that find no association between agreeableness and physical activity and longitudinal studies that find agreeableness to be associated with decreases in physical activity over time (Allen et al., 2017). It is, however, consistent with recent findings that indicate that agreeableness predicts engagement in more precautions to reduce the risk of virus (Aschwanden et al., in press; Blagov, 2020; Bogg & Milad, 2020). Individuals high in agreeableness also want to get along well with others and close others that may compel them to engage in physical activity together, such as going for a walk. Individuals high in agreeableness may thus be both more likely to adhere to recommendations for physical activity and to maintain relationships with others.

It was likewise surprising that neither extraversion nor openness was related to change in physical activity. Both of these traits have been linked previously with greater engagement in physical activity (Sutin et al., 2016; Wilson & Dischman, 2015), and, in the present study, both traits were associated with higher average physical activity. Individuals high in extraversion tend to be active and vigorous (Armon & Shirom, 2011) and may derive a lot of enjoyment from physical activity (Courneya & Hellsten, 1998). These characteristics were true prior to the pandemic, and individuals high in extraversion may already have found their optimal level of activity. Further, individuals higher in openness engage in a variety of activities (Stephan et al., 2014), which may include physical activity during the pandemic but may also include greater engagement in other areas of interest (e.g., reading books, listening to podcasts, cooking new recipes).

The present study provides new evidence that personality predicts behavioral responses to the COVID-19 crisis. In addition to precautionary behaviors that both reduce the risk of contracting the virus and spreading it to other people (Aschwanden et al., in press; Blagov, 2020; Bogg & Milad, 2020), personality traits are associated with engagement in crucial health-related behaviors during the acute phase of the coronavirus crisis. This study has theoretical implications for existing models on personality and health (Friedman & Kern, 2014; Turiano et al., 2015). These models postulate that personality traits predispose individuals to engage in behaviors that contribute to their health over time. This study extends these models by showing that the association between personality and physical activity and sedentary behavior observed in nonpandemic times (e.g., Allen et al., 2017; Sutin et al., 2016) generalizes to the COVID-19 crisis.

The present study had several strengths that included a longitudinal design with repeated measures of physical activity and sedentary behaviors during the acute phase of the coronavirus crisis and an assessment of personality before the acute phase of the crisis in two large samples of adults. Both studies had adequate power to test the associations. Some limitations should also be considered. First, there were selection effects into the longitudinal assessments for both samples that may limit the generalizability of the findings. For example, in the PBSR, individuals with follow-up sedentary behaviors data had more adaptive personality profiles, characterized by lower neuroticism and higher conscientiousness, that lead to less sedentary behaviors. Second, the UAS did not have an assessment of physical activity using the same item prior to the pandemic. Therefore, this study examined the change in physical activity during the pandemic, and it was not possible to test whether the pandemic had an impact on usual patterns of physical activity. Third, this study relied on self-reported physical activity and sedentary behavior. More research using objective measures like accelerometer is needed. Finally, the data were collected in the acute phase of the pandemic in the United States. The patterns observed may or may not generalize to other countries or if the pandemic lingers.

In sum, this study provides further evidence for the role of personality for the prediction of behavioral responses to COVID-19 crisis. Higher neuroticism was related less physical activity and more sedentary behavior, whereas higher conscientiousness, agreeableness, and to a lesser extent higher extraversion and openness were related to more favorable patterns of physical activity during the acute phase of the coronavirus pandemic.