Leisure and psychosocial functioning during the COVID-19 pandemic: Exploring data from adolescents and parents of adolescents

Abstract

During the COVID-19 pandemic, many people experienced life disruptions affecting leisure activities, such as discontinuation of after-school activities for adolescents and changes in the amount of leisure time available for parents. Because leisure activities have previously been associated with psychosocial functioning, the present study examined leisure behaviors and psychosocial functioning among 158 parents/guardians of adolescents and 116 of their children during the first summer of the COVID-19 pandemic in the United States. Results indicated that time spent engaging with TV, video games, and social media during the pandemic was associated with higher levels of depressive symptoms and anxiety for both parents and adolescents. Retrospective reports indicating decreased leisure time socialization were also associated with depressive symptoms during the pandemic in both groups. These results highlight the importance of leisure for both adolescents and parents of adolescents and its capacity to support or undermine psychosocial functioning in challenging times.

Keywords:

• COVID-19
• recreation; adolescents, parents, psychosocial functioning

The COVID-19 pandemic heavily impacted many aspects of everyday life, including leisure and psychosocial functioning.¹ However, the effects of the COVID-19 pandemic were not uniformly experienced (Tai et al., 2021). Recognizing the differential impact of living through difficult times at different developmental periods across the life span (Elder, 1975), the present study examined the extent to which leisure behaviors, and perceived changes in these behaviors, were associated with levels of depressive symptoms, anxiety, and emotional support among adolescents and parents of adolescents in 2020, during the first summer of the COVID-19 pandemic in the United States.

In this article, we first review research on different types of leisure activities and the evidence for their associations with psychosocial functioning as well as the impacts of the COVID-19 pandemic on leisure. We then describe the ways the pandemic may have been experienced differently by adolescents and parents of adolescents. Using a sample of U.S. parents/guardians and a sample of their adolescent children, the present study sought to identify whether each group’s leisure behaviors were associated with depressive symptoms, anxiety, and emotional support during the first summer of the pandemic.

Leisure and psychosocial functioning

Although there are many definitions of leisure, Weybright et al. (2019) found that U.S. participants understood the term to refer to recreational activities undertaken in time free from work. These activities often include, but are not limited to, physical activity, outdoor pursuits, reading, and watching TV or playing video games. Leisure has the potential to contribute to well-being (Coleman & Iso-Ahola, 1993; Kelly, 2012) and can aid in handling stressful situations (Caldwell, 2005; Kelly, 2012). However, an individual’s leisure choices are also shaped by constraints such as cultural norms, available resources, and confidence levels (Godbey et al., 2010).

Some patterns have been observed among specific types of leisure activities and psychosocial functioning. For example, leisure time physical activity is consistently associated with lower levels of mental illness and better health for both adults and adolescents (McPhie & Rawana, 2015; Penedo & Dahn, 2005). Social forms of leisure may also be beneficial as leisure supports (Chang et al., 2014; Coleman & Iso-Ahola, 1993), and can facilitate the development of friendships and social support systems, which are crucial for the well-being of both adolescents (Steinberg & Morris, 2001) and adults (Lin et al., 2014). Similarly, spending leisure time outdoors has been shown to support psychosocial functioning in youth and adults (Thomsen et al., 2018), and the limited research on reading suggests that it is also likely to be beneficial for youth (Mak & Fancourt, 2020). However, the evidence is less clear for the relationships between psychosocial functioning and other types of leisure activities. For example, screen-based leisure activities (such as playing video games and using social media) may provide opportunities to receive social and emotional support for adolescents (Odgers et al., 2020), but some types of social media use are also associated with depressive symptoms for adolescents (Vidal et al., 2020) and adults (Escobar-Viera et al., 2018). In sum, socializing, physical activity, and outdoor time have established positive relationships with psychosocial functioning. It is less clear how activities like reading, social media, and playing video games or watching TV may be associated with psychosocial functioning for adolescents and parents of adolescents. Furthermore, because the COVID-19 pandemic drastically affected opportunities to engage in different types of leisure activities (Moore et al., 2020) while simultaneously increasing stress and uncertainty (Achterberg et al., 2021), it is important to examine to what extent different types of leisure activities are associated with psychosocial functioning in the context of such widespread disruption, because the results may vary from those observed prepandemic.

Pandemic impacts on leisure

The COVID-19 pandemic drastically impacted physical health (Ioannidis, 2021) and psychosocial health (Dubey et al., 2020) and altered people’s ability to engage in their usual activities, including leisure activities. Leisure during the pandemic was affected by community-level factors such as government-mandated prevention methods and lockdown measures (Kaufman et al., 2021), which limited available opportunities. Personal characteristics like age, race, gender, and employment status also contributed to variation in people’s ability to engage in leisure activities (Chen et al., 2021; Mutz & Reimers, 2021). The COVID-19 pandemic and associated public health initiatives led to widespread disruptions to work and school, as well as disruptions of leisure activities via the closure of gyms, playgrounds, and other recreational facilities (Moore et al., 2020; Park et al., 2020). In addition to the challenges caused by the pandemic itself, the disruption of established leisure routines has been previously associated with worse psychosocial functioning in populations of adults and college students (Allen-Collinson & Hockey, 2007; Doerksen et al., 2014). Whether disruptions are due to an unexpected injury or simple fluctuation in weekly leisure behavior, rates of depression, anxiety, and stress increase when changes alter people’s typical patterns of leisure (Doerksen et al., 2014). However, these types of disruptions to leisure have not been studied among adolescents, and societal-level disruptions, such as those caused by the COVID-19 pandemic, have also not been widely studied from this perspective.

The disruptions caused by the pandemic may have led to increases in some types of leisure and decreases in others, with differential associated impacts on psychosocial functioning. For example, outdoor spaces saw increased rates of visitation during COVID-19 lockdowns (Geng et al., 2021; Slater et al., 2020), which may be associated with improved physical and mental health for youth and adults (Thomsen et al., 2018). Physical activity is also known to improve physical and mental health (McPhie & Rawana, 2015; Penedo & Dahn, 2005) but, with the pandemic-related closure of gyms and other fitness facilities, people were no longer able to rely on their usual sources of physical activity (Castañeda-Babarro et al., 2020). The shift from exercising in gyms to exercising online (Kaur et al., 2020) reflects the trend of increasing social media use during the pandemic (Paschke et al., 2021). With the increase in social isolation during the COVID-19 pandemic, social media has been identified as a potential coping mechanism (Cauberghe et al., 2021). However, other research showed that virtual connections to friends had a negative effect on adolescents’ depressive symptoms during the pandemic, whereas in-person family interactions had a supportive effect (Ellis et al., 2020).

In sum, the COVID-19 pandemic may have precipitated increases in some types of leisure and decreases in others. Leisure is an important part of a healthy routine (Bailey et al., 2016), and disruptions to this leisure routine can have negative consequences for psychosocial functioning (Allen-Collinson & Hockey, 2007; Doerksen et al., 2014). As such, though it is possible that increases in certain leisure activities may be beneficial, it is also possible that even positive changes in one’s leisure routine may be associated with some distress.

Differential impacts of disruptive events

Although the COVID-19 pandemic disrupted the lives of people across the globe, historical events can have different impacts on individuals of different ages/generations, despite living through these events at the same time (Elder, 1975). For example, Elder (1975) noted that (bearing in mind heterogeneity within each age cohort) individuals who were approximately 10 years old at the onset of the Great Depression were less vulnerable to “family misfortune” (p. 184) than younger children, while also being young enough that they did not face the challenge of navigating scarce employment opportunities. The COVID-19 pandemic, another widely disruptive historical event, may also have had differential impacts on individuals in different age cohorts and developmental periods. Despite living in the same household, the COVID-19 pandemic may have affected adolescents and parents/guardians of adolescents differently, due to their different social roles and positions in their life course. The present study therefore focuses on potential differences in pandemic experiences between adolescents and parents/guardians of adolescents, while acknowledging that the extent to which individuals’ lives were disrupted varied considerably with factors such as age, race, employment status, and household income (Tai et al., 2021). Factors such as employment change during the pandemic and the severity of the local impact of COVID-19 may also have influenced both leisure experiences and psychosocial functioning for both adolescents and parents/guardians of adolescents and are therefore important to consider as well.

The pandemic context for adolescents

Psychosocial development is particularly important in adolescence (Sawyer et al., 2012; Steinberg & Morris, 2001), but COVID-19 restrictions prevented adolescents from attending in-person schooling and extracurricular activities for much of 2020, significantly altering the social lives of adolescents (Magson et al., 2021). Adolescents with lower levels of social support were more likely to experience adverse psychosocial impacts such as depression and anxiety during the pandemic (Campione‐Barr et al., 2021). Prior to the pandemic, adolescents were also more likely to experience depressive symptoms than adults (Deighton et al., 2019), and research has shown that adolescents were at an even greater risk of depression during the pandemic (Loades et al., 2020). In addition, recreation and leisure activities provide opportunities for psychosocial development, friendship formation, and skill building (Caldwell & Witt, 2011). The loss of structured, adult-led programs such as youth sports and school clubs may have reduced adolescents’ opportunities for positive development if they were not able to find similar opportunities within their families, outdoors, or online (Ettekal & Agans, 2020; Reed et al., 2022).

The pandemic context for parents of adolescents

Although there is considerable diversity in the ages of individuals parenting adolescents, this phase of the life course often involves challenges as families navigate the changes of the adolescent period (Soenens et al., 2019). Adults raising children experienced lower psychosocial functioning than other adults during the pandemic (Czeisler et al., 2021). Having their adolescent children continually home with them during pandemic lockdowns may have been particularly difficult for some parents, whereas others may have used this opportunity to build closer relationships (Pediconi, 2022). In addition, although many adults retained levels of family support, social support from friends, community, organizations, and society declined (Li et al., 2021). Adults also faced stress relating to finances and job adjustment, and parents confronted challenges associated with losing childcare/education (Chen et al., 2021). Overall, the stressors of the pandemic negatively impacted parents’ psychosocial functioning, increasing their levels of stress and symptoms of mental illness such as depression and anxiety (Achterberg et al., 2021).

The present study

We examined the relationships between leisure behaviors and psychosocial functioning among U.S. adolescents and parents/guardians of adolescents during the summer of 2020. Specifically, we were interested in whether engagement in leisure behavior (and changes in leisure behaviors) predicted psychosocial functioning during the COVID-19 pandemic and whether the relationships between leisure and psychosocial functioning differed between adolescents and parents/guardians of adolescents.

Our first set of hypotheses was that, consistent with the prepandemic literature, engagement in more social, outdoor, and/or physical activities during the pandemic would be associated with better psychosocial functioning (i.e., fewer depressive symptoms, lower levels of anxiety, and higher levels of perceived emotional support) for both adolescents and parents/guardians of adolescents. The literature regarding the effects of screen time provided less compelling evidence that leisure time spent watching TV, video gaming, or using social media would be associated with better psychosocial functioning. Therefore, we predicted that more leisure time spent in these pursuits would be associated with poorer psychosocial functioning for both adolescents and parents/guardians of adolescents. We additionally explored the association between reading and psychosocial functioning, despite insufficient literature to support a directional hypothesis. This first set of hypotheses sought to assess whether, even during the unique event of a pandemic, participation in leisure activities was consistently associated with psychosocial functioning for both adolescents and parents/guardians of adolescents.

Our second set of hypotheses centered around retrospectively reported changes in the frequency of participation in leisure activities from before the COVID-19 pandemic to the summer of 2020, predicting that perceived changes in leisure activity frequency (self-reported increases or decreases) would be associated with poorer psychosocial functioning for both adolescents and parents/guardians of adolescents. However, we expected that this relationship would be impacted by the type of activity, such that perceived increases in the frequency of outdoor time and/or physical activity or decreases in social media and/or TV watching and video gaming would be associated with better psychosocial functioning during the pandemic. In addition, we expected that participants’ leisure experiences and psychosocial functioning during the pandemic would likely vary by demographic characteristics (e.g., generation, race, gender) and life circumstances (e.g., employment status, socioeconomic status, urban versus rural residential area, local prevalence of COVID-19), so we accounted for these factors in all analyses, although our focus was on the examination of potential differences between adolescents and parents/guardians of adolescents.

Methods

The present study used a cross-sectional survey of U.S. adolescents and parents/guardians of adolescents to examine how the leisure activities of individuals in these two different phases of the life course were related to their psychosocial functioning during the first summer of the COVID-19 pandemic.

Procedures

Data for this study were collected using an online survey distributed between May and August 2020. Survey distribution followed a modified snowball sampling technique, whereby approximately 100 youth development practitioners were invited by the research team via email to share the survey announcement with the families they serve using their existing communication systems (e.g., newsletters, emails). These practitioners were selected based on the social and professional networks of the research team, with efforts made to identify practitioners across a wide variety of U.S. states. Parents/guardians were asked to consent for both themselves and their oldest adolescent child (aged 12–17) to participate in the study and to provide their child’s email address. Using this information, we sent a separate survey and assent form to the adolescent. Because the survey link was shared on social media, we received over 500 ineligible, invalid, and otherwise unusable surveys. We therefore screened participants based on their responses to attention check items as well as study eligibility criteria (e.g., being the parent of an adolescent aged 12 to 17 years; see Agans et al., 2023 for details). All participants (parents/guardians and adolescents) who met the eligibility screening criteria received a $10 gift card after completing the survey, which took an average of 21 min for parents and 16 min for adolescents, with longer outlier times among 15 parents and 13 adolescents.

Participants

Participants were 158 parents/guardians (22% male, 77% female, 1% non-binary; average age = 46, SD = 6.18) and 116 of their adolescent children (48% male, 50% female, 1% non-binary, 1% “something else fits better”; average age = 15, SD = 1.58). Most of these families were White (82% of parents; 79% of adolescents), with smaller percentages of Black/African American (6% of parents; 7% of adolescents), Asian/Asian American (6% of parents; 5% of adolescents), Hispanic/Latino/Latina (1% of parents; 4% of adolescents), multiracial (2% of parents; 4% of adolescents), and other races/ethnicities (3% of parents; 1% of adolescents). Of the parents/guardians, 76% were currently employed and 60% reported changes in their employment during the pandemic. Of the adolescents, 24% were currently employed and 25% reported changes to their employment. The majority of participants lived in households where parents reported an annual household income over $100,000 (55%), with 18% earning $75,000 to $99,000 per year, 12% earning $50,000 to $74,000 per year, and 15% earning less than $50,000 per year. Participants represented 31 U.S. states (21% from New York, 18% from Pennsylvania, 11% from California, and <10% from each other state) and a wide variety of residential areas (27% towns, 22% suburbs, 19% rural areas, 18% smaller cities, 14% large cities). Participants reported the local COVID-19 infection rate in their community at the time of the survey as moderate (parent mean = 2.45, SD = 1.07 on a scale from 1 to 5; adolescents mean = 2.50, SD = 1.01).

Measures

Demographics

All participants reported their own age, race/ethnicity, gender, current employment, and employment changes during COVID-19. Options for gender were male, female, transgender, non-binary, questioning/not sure of gender identity, and something else fits better. Due to the small number of individuals reporting genders other than male and female, we binarized this measure into male and non-male categories. Similarly, due to the predominantly White sample, race was binarized for analysis. Parents/guardians also provided information about their household income in 2019 and reported their current employment status with options including homemaker; employed and working; employed but on leave (e.g., furlough); unemployed; looking; not in the labor force due to retirement, disability, or other; and other. Parents also responded to the statement “I am experiencing changes in my employment due to COVID-19” on a 5-point Likert scale, where 1 = not at all and 5 = a great deal. Adolescents were asked to report for both “currently” and “this time last year” how much time they spend in paid employment on a 5-point Likert scale, where 1 = not at all and 5 = more than 4 h.

Local COVID-19 impact

All participants were asked to respond to the statement “People in my community are being diagnosed with COVID-19” on a 5-point Likert scale, where 1 = not at all and 5 = a great deal.

Leisure time activities

Items measuring time use were adapted from the Behavioral Risk Factor Surveillance System Survey (Centers for Disease Control and Prevention, 2001). The question stems “This time last year, how often did you spend time participating in …” and “Currently, how often do you spend time participating in …” were completed with a list of activities including reading, socializing, watching TV/gaming, social media, and being outdoors. Aside from socializing, these items did not specify whether the activities took place with other people. Participants responded to these questions using a Likert scale ranging from 1 (none) to 6 (every day). Physical activity was measured using items adapted from the Youth Risk Behavior Surveillance System (Centers for Disease Control and Prevention, 2019), including “On how many days were you physically active for a total of at least 60 minutes per day?” both “In a normal week prior to the pandemic” and “In the past 7 days,” with response options ranging from 0 to 7.

Depressive symptoms

Depressive symptoms were measured using the Center for Epidemiologic Studies Depression Scale (Radloff, 1977), which has been validated in adult and adolescent samples (Radloff, 1991). The question stem “Below is a list of the ways you might have felt or behaved. Please tell me how often you have felt this way during the past week …” was followed by 20 items (e.g., “I felt everything I did was an effort”). Participants responded using a 4-point Likert scale, where 0 = rarely or none of the time (less than 1 day) and 3 = most or all of the time (5–7 days). Responses were summed to calculate a scale score. Cronbach’s alpha was .75 for parents and .78 for adolescents.

General anxiety disorder

Anxiety symptoms were measured using the Generalized Anxiety Disorder 7-item scale (Spitzer et al., 2006), which has also been validated in adolescent samples (Mossman et al., 2017). The question stem “Over the last 2 weeks, how often have you been bothered by the following problems?” was followed by seven items (e.g., “Feeling nervous, anxious, or on edge”). A 4-point Likert scale, where 0 = not at all sure and 3 = nearly every day, was used and a scale score was calculated by summing the responses to each item. Cronbach’s alpha was .92 for parents and .89 for adolescents.

Emotional support

Emotional support was measured using the psychometrically validated Patient-Reported Outcomes Measurement Information System Emotional Support short form v2.0 (Reeve et al., 2007), which has previously been used with adolescents (Belson et al., 2022). The Patient-Reported Outcomes Measurement Information System is a widely used multidomain measure of health, and the emotional support items are part of the social health domain. The question stem “Please tell us how often the following statements are true for you …” was followed by four items (e.g., “I have someone who makes me feel appreciated”). Participants responded using a 5-point Likert scale, where 1 = never and 5 = always. Scale scores were calculated by summing the item responses. Cronbach’s alpha was .92 for parents and .93 for adolescents.

Data analysis

We used multivariate regression analyses to examine the relative influence of leisure activities, demographic characteristics, and perceptions of local COVID-19 impact on variance in psychosocial functioning during the pandemic for both adolescents and parents/guardians of adolescents (Hypothesis 1).² We tested differences by generation (i.e., parents versus adolescents) by including in our model interaction terms between leisure activities and generation, when generation had a significant main effect. For Hypothesis 2, we used nonparametric Wilcoxon signed-rank tests to assess whether changes in leisure activities were associated with the three indicators of psychosocial functioning due to the nonnormal distribution of the data. Perceived change scores were computed by subtracting current participation levels from retrospectively reported prepandemic participation. Scores were reduced to three categories (decrease for scores less than zero, no change for scores of exactly zero, and increase for scores above zero), and one-way analysis of variance (ANOVA) tests were used to assess their relationship to psychosocial functioning.

Results

Hypothesis 1: Type of leisure activity will affect associations with psychosocial functioning

As noted above, we predicted that engagement in more social, outdoor, and/or physical activities during the pandemic would be associated with better psychosocial functioning (i.e., fewer depressive symptoms, lower levels of anxiety, and higher levels of perceived emotional support), whereas leisure time spent watching TV, video gaming, or using social media would be associated with poorer psychosocial functioning for both adolescents and parents/guardians of adolescents. We additionally explored the association between reading and psychosocial functioning without a directional hypothesis.

For depressive symptoms, the initial model (including race, gender, generation, income, residential area, employment status, employment change during COVID, and local COVID-19 impact as predictors) was not significant. After adding the six leisure activities, R² improved from .078 to .222 and the model reached significance, F(14, 195) = 3.70, p < .001. The only variables with significant main effects on depressive symptoms in this model were generation (β = 3.44, SE = 1.27), gender (β = 2.94, SE = 1.26), employment change (β = 2.65, SE = 1.19), and time spent watching TV/playing video games (β = 2.72, SE = 0.70). Being an adolescent, not being male, having more changes in employment, and spending more time watching TV or playing video games were associated with higher rates of depressive symptoms. Table 1 shows these regression models. Based on these results, we also tested a model with interactions between leisure activities and generation, but the interaction terms did not reach significance.

Table 1. Regression analysis for predictors of depressive symptoms among youth and parents.

Parameters	Model 1	Model 2	Model 3
	Unstandardized coefficient (std. error)	Unstandardized coefficient (std. error)	Unstandardized coefficient (std. error)
Intercept	11.24 (4.48)*	0.62 (5.13)	−3.09 (9.54)
Demographics
Gender	2.28 (1.19)	2.94 (1.25)*	2.69 (1.28)*
Race	1.78 (1.48)	0.58 (1.41)	0.67 (1.44)
Generation	2.97 (1.34)*	3.44 (1.27)**	5.95 (5.17)
Income	−0.87 (0.51)	−0.71 (0.50)	−0.57 (0.52)
Residential area	−0.00 (0.46)	0.24 (0.43)	0.17 (0.44)
Employment status	0.15 (1.37)	0.46 (1.31)	0.58 (1.34)
Change in employment	2.20 (1.25)	2.65 (1.19)*	2.33 (1.22)
Local COVID severity	0.36 (0.55)	0.42 (0.52)	0.44 (0.53)
Leisure activities
Physical activity		−0.09 (0.27)	0.85 (0.86)
Outdoor time		−0.34 (0.59)	−1.92 (2.05)
Reading		0.40 (0.56)	0.40 (2.11)
Socializing		0.82 (0.53)	3.01 (2.05)
TV/video games		2.71 (0.69)***	2.76 (2.10)
Social media		0.88 (0.51)	1.29 (1.79)
Interactions
Generation * Physical activity			−0.66 (0.56)
Generation * Outdoor time			1.14 (1.24)
Generation * Reading			0.00 (1.22)
Generation * Socializing			−1.27 (1.18)
Generation * TV/video games			−0.14 (1.33)
Generation * Social media			−0.20 (1.04)
Model statistics
df	8	14	20
p-Value	0.052	<0.001	<0.001
R^2	0.078	0.222	0.237
R^2 adjusted	0.038	0.162	0.150

*p < 0.05; **p < 0.01; ***p < 0.001

The initial model for general anxiety including only race, gender, generation, income, residential area, employment status, employment change during COVID, and local infection rate was significant, F(8, 200) = 2.87, p = .005, with main effects for race (β = 1.88, SE = 0.90, p = .038) and gender (β = 2.20, SE = 0.72, p = .003). In this model, White individuals and male individuals reported lower levels of anxiety. When leisure activities were added, R² increased from 0.107 to 0.227, F(14, 200) = 3.90, p < .001. In this model, only gender (β = 1.92, SE = 0.78, p = .013), TV/video gaming (β = 0.93, SE = 0.42, p = .029), and social media use (β = 0.95, SE = 0.32, p = .003) had significant effects. Specifically, non-male gender, spending more time watching TV/playing video games, and more time on social media were associated with higher anxiety for both adolescents and parents/guardians of adolescents. Table 2 shows these models.

Table 2. Regression analysis for predictors of general anxiety among youth and parents.

Parameters	Model 1	Model 2
	Unstandardized coefficient (std. error)	Unstandardized coefficient (std. error)
Intercept	−0.30 (2.61)	−4.56 (3.06)
Demographics
Gender	2.19 (0.72)**	1.91 (0.76)*
Race	1.87 (0.89)*	1.20 (0.86)
Generation	0.64 (0.80)	0.69 (0.77)
Income	−0.45 (0.30)	−0.29 (0.30)
Residential area	0.06 (0.26)	0.22 (0.25)
Employment status	0.83 (0.82)	0.86 (0.80)
Change in employment	0.85 (0.75)	0.89 (0.72)
Local COVID severity	0.43 (0.33)	0.48 (0.31)
Leisure activities
Physical activity		−0.15 (0.16)
Outdoor time		−0.14 (0.36)
Reading		0.28 (0.33)
Socializing		0.47 (0.32)
TV/video games		0.93 (0.42)*
Social media		0.94 (0.31)**
Model statistics
df	8	14
p-Value	0.005	<0.001
R^2	0.107	0.227
R^2 adjusted	0.069	0.169

*p < 0.05; **p < 0.01; ***p < 0.001

Finally, the model for emotional support followed a similar pattern, with the initial model being significant, F(8, 208) = 3.01, p = .003, with a main effect for race (β = −2.43, SE = 0.63, p < .001). White individuals reported significantly higher levels of emotional support. R² improved from 0.111 to 0.203 with the addition of leisure activities, F(14, 201) = 3.41, p < .001, and main effects were observed in this model for race (β = −2.55, SE = 0.62, p < .001), physical activity (β = 0.27, SE = 0.12, p = .023), outdoor time (β = 0.62, SE = 0.26, p = .018), and social media use (β = −0.49, SE = 0.23, p = .032). In this model, White participants and those engaging in more physical activity, more outdoor time, and using less social media reported higher levels of emotional support. Table 3 shows these regression models.

Table 3. Regression analysis for predictors of emotional support among youth and parents.

Parameters	Model 1	Model 2
	Unstandardized coefficient (std. error)	Unstandardized coefficient (std. error)
Intercept	14.62 (1.84)***	11.69 (2.19)***
Demographics
Gender	0.79 (0.50)	1.36 (0.54)*
Race	−2.46 (0.62)***	−.254 (0.61)***
Generation	0.53 (0.56)	0.64 (0.55)
Income	−0.10 (0.21)	−0.12 (0.21)
Residential area	0.17 (0.19)	0.05 (0.18)
Employment status	−0.15 (0.58)	−0.06 (0.57)
Change in employment	0.44 (0.53)	0.64 (0.51)
Local COVID severity	0.12 (0.23)	0.17 (0.22)
Leisure activities
Physical activity		0.27 (0.11)*
Outdoor time		0.61 (0.25)*
Reading		−0.24 (0.24)
Socializing		−0.00 (0.23)
TV/video games		0.43 (0.30)
Social media		−0.48 (0.22)*
Model statistics
df	8	14
p-Value	0.003	<0.001
R^2	0.111	0.203
R^2 adjusted	0.074	0.144

*p < 0.05; **p < 0.01; ***p < 0.001.

Hypothesis 2: Differences between prepandemic and pandemic leisure activities will be associated with psychosocial functioning

We examined to what extent participants’ time in each activity varied from prepandemic (retrospective report) to the time of the survey. We used Wilcoxon signed-rank tests due to the nonnormal distribution of the data and found that parents reported statistically significant differences in their level of participation in all six leisure activities. Specifically, parents reported spending more time outdoors (M of 3.57 versus 3.94, p < .001), engaging in more physical activity (M of 3.56 versus 3.94, p = .007), doing more reading (M of 1.28 versus 1.41, p = .019), spending less time socializing (M of 1.53 versus 1.13, p < .001), watching more TV/playing more video games (M of 1.61 versus 1.88, p < .001), and using more social media (M of 1.61 versus 1.82, p = .008) during the pandemic. For adolescents, however, significant differences were only observed in the form of less time socializing (M of 2.03 versus 1.25, p < .001), more time watching TV / playing video games (mean of 1.64 versus 1.91, p < .001), and more social media use (M of 1.63 versus 1.89, p = .001).

We used these retrospective prepandemic reports and reports of leisure time use during the pandemic to compute change scores and identify participants in each generation who perceived that their time spent in each type of activity had decreased (i.e., retrospective reports were higher than current reports), increased (i.e., retrospective reports were lower than current reports), or exhibited no change (i.e., retrospective reports matched current reports). We used ANOVA to test for differences in these change scores between adolescents and parents/guardians of adolescents, but because no statistically significant differences were observed, Table 4 reports the number of overall participants in each category ANOVA tests indicated that individuals who perceived decreases in their socializing during the pandemic exhibited more depressive symptoms compared to those who did not report differences in their socializing behavior [F(2,245) = 4.23, p = .016; mean difference = 2.53, p = .029, eta squared = .033], but there were no other statistically significant associations with indicators of psychosocial functioning (see Table 5).

Table 4. Number and percentage of participants in each change category for each type of leisure activity.

	Decrease	No change	Increase
Outdoor time	n = 40 (15.2%)	n = 133 (50.4%)	n = 91 (34.5%)
Physical activity	n = 72 (27.0%)	n = 99 (37.1%)	n = 96 (36.0%)
Reading	n = 40 (15.0%)	n = 162 (60.7%)	n = 65 (24.3%)
Socializing	n = 125 (47.0%)	n = 116 (43.6%)	n = 25 (9.4%)
TV/video gaming	n = 27 (10.1%)	n = 132 (49.2%)	n = 109 (40.7%)
Social media	n = 24 (9.1%)	n = 178 (67.2%)	n = 63 (23.8%)

Table 5. ANOVA tests for differences in psychosocial health for parents and adolescents whose leisure involvement decreased, increased, or exhibited no change.

	Decrease		No change		Increase		df	F
	M	SD	M	SD	M	SD
Depressive symptoms
Change in physical activity	20.7	7.9	17.7	6.8	18.6	8.2	2, 202	1.33
Change in outdoor time	18.6	8.4	19.3	7.9	18.8	7.4	2, 199	0.14
Change in reading	20.7	8.5	18.3	7.7	19.2	7.4	2, 201	0.77
Change in socializing	20.0	7.1	17.3	7.7	21.1	9.8	2, 200	3.05
Change in TV/video gaming	18.5	6.5	18.3	8.4	19.7	7.3	2, 202	0.27
Change in social media	18.8	8.1	18.6	7.7	20.1	8.1	2, 199	0.37
General anxiety
Change in physical activity	5.0	4.0	5.0	5.3	5.7	4.9	2, 206	0.88
Change in outdoor time	5.0	5.7	5.4	4.9	5.1	4.2	2, 204	0.18
Change in reading	6.6	5.1	4.7	4.8	5.6	4.3	2, 206	2.24
Change in socializing	5.9	5.0	4.5	4.1	6.0	6.2	2, 205	3.01
Change in TV/video gaming	4.4	4.0	5.6	5.2	5.1	4.5	2, 207	0.79
Change in social media	5.8	6.3	4.9	4.7	6.0	4.6	2, 204	0.94
Emotional support
Change in physical activity	16.7	2.9	16.9	3.6	16.8	3.5	2, 206	0.19
Change in outdoor time	16.3	3.7	16.9	3.1	17.1	3.6	2, 204	1.08
Change in reading	16.5	3.9	16.9	3.4	16.8	3.1	2, 206	0.21
Change in socializing	16.7	3.5	17.1	3.1	16.7	3.7	2, 205	0.27
Change in TV/video gaming	17.1	3.4	16.7	3.3	16.9	3.4	2, 207	0.25
Change in social media	16.9	3.4	17.1	3.2	16.2	3.8	2, 205	0.98

*p < 0.05; **p < 0.01; ***p < 0.001.

Discussion

The present study focused on the leisure behavior and psychosocial functioning of U.S. adolescents and parents/guardians of adolescents during the summer of 2020, exploring the relationship between leisure and psychosocial functioning in the unique historical context of the COVID-19 pandemic. Sudden changes to leisure have been noted to negatively impact well-being (Allen-Collinson & Hockey, 2007; Doerksen et al., 2014), and as recreation facilities closed and stay-at-home orders were put in place, leisure began to look different across the United States for people of all generations (Geng et al., 2021; Stockwell et al., 2021). Additionally, COVID-19 restrictions limited socialization (Magson et al., 2021), resulting in compromised psychosocial functioning for both adolescents and parents/guardians of adolescents (Chen et al., 2021; Magson et al., 2021). However, the extent to which these changes in leisure behavior may be associated with psychosocial functioning has not been previously examined or assessed from a life course perspective (Elder, 1975). The present study therefore examined associations between participation in different leisure activities and psychosocial functioning in the context of the global pandemic among U.S. adolescents and parents/guardians of adolescents.

Our first hypothesis was that engaging in more social, outdoor, and physical activities during the pandemic would be associated with better psychosocial functioning (and vice versa for social media and TV/video gaming) for both adolescents and parents/guardians of adolescents. Our results do not support the first part of this hypothesis. Within our regression models accounting for all of the leisure activities together, outdoor time and physical activity were only associated with higher levels of emotional support for adolescents and parents of adolescents. This limited relationship with psychosocial outcomes for outdoor time and physical activity during the pandemic fails to support previous research that found physical activity to be associated with well-being both prepandemic (McPhie & Rawana, 2015; Penedo & Dahn, 2005) and during the pandemic (Slater et al., 2020). However, these prior studies did not consider the additive associations between multiple types of leisure and psychosocial functioning. Parents and adolescents could engage in multiple forms of leisure to add structure to their days, with some linked to better psychosocial functioning and others linked to worse. When considering multiple forms of leisure together, engaging with social media and watching TV or playing video games stood out as being significantly associated with poor psychosocial functioning, likely counteracting the potential positive effects of other types of leisure time activities.

Our findings regarding the negative effects of social media and TV/video gaming were consonant with prior literature. We found that TV/video gaming and social media use were positively associated with anxiety during the pandemic, and TV/video gaming was also positively associated with depressive symptoms. These findings were consistent for both adolescents and parents of adolescents and support research linking social media use during the pandemic with high levels of anxiety among adolescents (Ellis et al., 2020). Despite social media’s demonstrated ability to promote social connection (Burke & Kraut, 2014), the benefits of social media sites like Facebook are related to the sites’ ability to facilitate real-world interactions (Ellison et al., 2011), which were largely unavailable during this time. In addition, an uptick in the consumption of COVID-19-related news, commonly featured on social media and TV, has been associated with an increase in depressive symptoms during the early stages of the pandemic (Price et al., 2022). However, because the present study did not assess causal relationships, it is also possible that increased use of social media, TV watching, and video gaming could be a coping mechanism used in response to high levels of depression and anxiety (Landers-Potts et al., 2017; Odgers et al., 2020).

In addition, the present study supported widespread evidence that the burdens and responsibilities experienced during the pandemic were not uniformly distributed (Chen et al., 2021; Mutz & Reimers, 2021). We found that, consistent with prior research (Deighton et al., 2019; Loades et al., 2020), depressive symptoms during the pandemic were high for adolescents, higher than the rates we observed for parents/guardians of adolescents. This finding also supports developmental research suggesting that widespread societal disruptions have a differential effect depending on one’s position in the life span (Elder, 1975). In addition, although non-male participants reported higher rates of depressive symptoms and anxiety, White and non-male participants reported the highest levels of emotional support. The observation that emotional support was higher for White individuals in the present study may be reflective of the heightened levels of stress and grief that members of marginalized racial communities experienced during the COVID-19 pandemic (Sneed et al., 2020; Tai et al., 2021) and the racial reckoning spurred by the murder of George Floyd (Rogers et al., 2020), which began concurrently with data collection for this study. Our findings that non-male individuals reported more emotional support are consistent with research showing men to be less likely than women to give and receive emotional support (Burleson et al., 2005), despite the fact that women and mothers faced greater levels of stress (Dias et al., 2020) and anxiety (Zhang et al., 2021) during the pandemic. Thus, though the pandemic has had widespread global impacts, both our research and that of other scholars indicates that the impacts to individuals varied considerably.

Our second hypothesis was that differences between prepandemic retrospective reports and current reports of frequency of engagement in various types of leisure activities would be associated with parents’ and adolescents’ psychosocial functioning during the pandemic. This hypothesis drew on research conducted prior to the pandemic on disruptions to leisure activities (Allen-Collinson & Hockey, 2007; Doerksen et al., 2014). We found that parents reported significant differences in all measured forms of leisure activities; however, differences in adolescents’ reports of leisure activities were limited to social media, TV/video gaming, and socializing. During the pandemic, many adolescents became socially isolated due to the closure of schools and other social settings (Magson et al., 2021), and our observed increase in adolescents’ self-reports of social media use aligns with prepandemic studies showing that socially isolated adolescents often use electronic communication (i.e., social media and texting) to reduce feelings of loneliness (Landers-Potts et al., 2017).

However, despite the prevalence of perceived differences in leisure behavior from prepandemic to the time of the study, the hypothesized relationship between leisure disruption and psychosocial functioning was not supported. Although prior studies have shown increasing levels of physical activity and outdoor time to be associated with improved states of psychosocial functioning (Eigenschenk et al., 2019; Mammen & Faulkner, 2013), participants who reported engaging in more physical activity and outdoor recreation during the pandemic did not show better psychosocial functioning than other participants. Similarly, although the inverse relationship between social connection and depression is well understood (Gariépy et al., 2016), we did not observe significant associations between changes in socializing and psychosocial functioning. These null results run counter to prepandemic research on leisure disruptions (Allen-Collinson & Hockey, 2007; Doerksen et al., 2014) and suggest that perceived disruptions to participants’ leisure may have been overshadowed by other major life disruptions caused by the pandemic. Specifically, because the pandemic affected primary routines required for survival (e.g., hygiene, sleep, eating), the disruption of secondary routines (e.g., leisure, work) may have had less impact on overall stress (Hou et al., 2020). Another possible explanation of this null result is that participants, in attempting to preserve their health and well-being during the pandemic, may have maintained leisure routines by substituting activities they deemed to be less risky or healthier in a pandemic context or by finding alternative ways to participate in leisure activities (e.g., Kim et al., 2022).

In sum, although disruptions to leisure activities may be difficult under normal circumstances (Allen-Collinson & Hockey, 2007; Doerksen et al., 2014), our results suggest that changes in leisure activities during times of general life disruption may be less problematic. However, following our findings from Hypothesis 1, adolescents, their parents, and programs that support youth and families should be attentive to the connection between TV/video gaming and social media use and rates of anxiety and depressive symptoms (also see Ellis et al., 2020; Price et al., 2022). Notably, the usual benefits of physical activity could not be detected when accounting for variance in anxiety and depressive symptoms by TV/video gaming and social media use. In the postpandemic era, these findings remain relevant for supporting positive psychosocial functioning. Specifically, our findings suggest that future research should examine and intervene on leisure activities holistically, considering the balance of various types of leisure activities that an individual may engage in throughout the day.

Limitations and future directions

The present study’s ability to assess the specific impacts of the COVID-19 pandemic on the leisure and psychosocial health of adolescents and parents/guardians of adolescents is limited by the use of a sample of parents and adolescents living in a wide variety of U.S. states with variation in COVID-19 restrictions. We also did not examine family-level effects in the present study. Although the study recruited parents and adolescents from the same families, our research questions did not lead us to examine the bidirectional influences parents and adolescents may have had on each other’s leisure and psychosocial functioning, nor did we collect data on other family members (e.g., other children living in the home). In addition, the use of retrospective questions to measure participants’ pre-COVID-19 lives introduces the possibility of recall bias (Hipp et al., 2020), and our lack of prepandemic data on psychosocial functioning limits our ability to measure change in these factors. Instead, our study focused on participants’ perceptions of difference between their current and prepandemic lives. Longitudinal research is needed to examine changes in leisure behaviors more objectively. Future research should also consider using more nuanced measures of social media use and socialization, because social media is often used as a means of connecting with family and friends (e.g., Landers-Potts et al., 2017), and our measure did not specify whether socializing took place in person or online. Measures of physical activity that differentiate between leisure time physical activity and obligatory exertion for work or transportation should also be used.

Furthermore, our relatively small sample size and use of snowball sampling via practitioners limits the study’s generalizability. The recruitment of participants linked to youth development programs may have biased the study, because adolescents engaged in programming aimed at promoting positive development may have had different profiles of leisure participation than their unaffiliated peers. The sample’s relative homogeneity also contributed to the study’s inability to compare experiences across different demographic groups. The study sample’s lack of diversity in race, gender, and socioeconomic status points to a need for future research to utilize more diverse samples to better understand the effects of the pandemic on groups for whom its influence may have been more harmful (Dias et al., 2020; Sneed et al., 2020; Tai et al., 2021). However, due to the increased stress and burden placed upon these individuals by the pandemic and by the racial justice movement in the summer of 2020 (Rogers et al., 2020) while recruitment for this study was ongoing, it was also less likely that these individuals would prioritize participation in this type of research project. Future research should include larger, more diverse samples and examine how both leisure activities and psychosocial functioning may respond to family dynamics (e.g., whether adolescents who spent more time engaged in leisure with their parents also report better psychosocial functioning), especially in the context of pandemic lockdowns. In sum, further research is needed to better understand under what circumstances, and for whom, different types of leisure activities are associated with psychosocial functioning.

Conclusion

Although the COVID-19 pandemic was experienced globally, the ways in which it affected individuals and their communities varied. This study therefore sought to examine leisure behavior and psychosocial functioning among U.S. adolescents and parents/guardians of adolescents in the summer of 2020, drawing on Elder’s (1975) observation of differential effects of societal disruptions on people at different points in the life span. Our results show that adolescents experienced more depressive symptoms than parents/guardians and that parents/guardians of adolescents perceived more differences than adolescents between their prepandemic and current leisure behavior. We also observed some associations between leisure activities and psychosocial functioning: Individuals who reported decreased socializing and higher rates of screen-based leisure (i.e., social media and TV/video games) during the summer of 2020 also reported poorer psychosocial functioning. However, we also did not observe some expected associations between leisure behavior and psychosocial functioning (e.g., physical activity and outdoor time were not associated with depressive symptoms or anxiety in the present study). Future research should continue to explore the links between leisure and psychosocial functioning, especially under challenging circumstances such as those present in the summer of 2020. In light of the established role of leisure in supporting psychosocial functioning (Caldwell, 2005; Kelly, 2012) and the continued influence of the COVID-19 pandemic on the ways in which people engage in society, including through leisure (e.g., Moore et al., 2020; Park et al., 2020; Paschke et al., 2021), it will be important to continue to monitor the relationship between leisure activities and well-being.

Ethics statement

The Pennsylvania State University Institutional Review Board approved this study (protocol number 00015003). All participants provided consent/assent; parental consent was obtained for adolescents.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by The Pennsylvania State University College of Health and Human Development. The study was also supported by NIH /NCATS Grant #UL1TR000127 and UL1TR002014 in its use of the REDCap survey platform.

Notes

1 There is some variation in the literature with regard to how mental health, mental illness, and distress are understood in relation to each other. Payton (2009) has shown that there is more to mental health than absence of distress and dysfunction and that these factors should be understood as discrete phenomena rather than as a continuum of related experiences. We therefore use the term psychosocial functioning to encompass indicators of well-being (e.g., emotional support) and indicators of distress/disorder (e.g., symptoms of depression and anxiety).

2 Because our sample consists of parents/guardians and their adolescent children, we tested whether multilevel models were necessary to account for the nested data structure. Likelihood ratio test results indicated that multilevel models did not better fit the data for models of depression and emotional support. Although multilevel models were a better fit for generalized anxiety disorder, the conclusions were substantively the same, so we report the single-level models here for parsimony.