Outdoor activity and walking for leisure in the COVID-19 era: their correlates and mutual relationship

ABSTRACT

The pandemic has changed daily urban life, including leisure activities and travel patterns. Although several studies have tracked changes in leisure and travel behavior due to the pandemic, few have analyzed the effects of COVID-19 preventive behavior by considering the interaction between activities and walking for leisure. This study aims to investigate the mutual relationship between outdoor activities and walking for leisure during the pandemic among 537 residents in the Seoul metropolitan area using a three-level least squares (3SLS) approach. The results demonstrate a bidirectional association between outdoor activities and walking during the pandemic. Individuals who avoided long-distance leisure travel increased their frequency of short-distance walking, and postponing or canceling social interactions led to more outdoor leisure activities. Notably, those who refrained from going out for leisure exhibited reduced outdoor activities but an increase in walking for leisure. These findings highlight the significance of considering the synergy between leisure activities and walking amid the pandemic.

GRAPHICAL ABSTRACT

KEYWORDS:

• Urban
• three-stage least squares
• outdoor activity
• walking for leisure
• preventive behavior
• COVID-19 pandemic

1. Introduction

The prolonged lockdowns and social distancing measures caused by COVID-19 have drastically changed daily lives. These measures and restrictions were particularly restrictive in metropolitan areas, significantly affecting people’s daily activities (Abdullah et al. 2020). Notably, there were significant changes in citizens’ leisure activities and travel behavior during the pandemic (Parady, Taniguchi, and Takami 2020).

Existing evidence indicates that the use of public transportation declined due to travel restrictions and border closures, leading to an increase in the use of active modes such as walking and bicycling (e.g., M. H. Kim, Lee, and Gim 2021). Similarly, restrictions on public gatherings and the closure of public facilities have led to a decrease in indoor activities, while outdoor activities have increased (Parady, Taniguchi, and Takami 2020). Previous studies have found that changes in people’s leisure activities and mode choices are affected by the perceived risk of COVID-19 and the COVID-19 preventive behavior. However, the findings of these studies were inconsistent, indicating that more empirical evidence of the effects of these factors, particularly in the context of outdoor leisure activities, is needed (Abdullah et al. 2020; Shakibaei et al. 2021).

Moreover, previous studies have emphasized the interdependence between leisure activities and travel mode choice in the decision-making process of leisure behavior. Specifically, certain discretionary activities serve as determinants of travel mode choice, while certain activities are determined based on the chosen travel mode (Baumgartner et al. 2023; Krygsman, Arentze, and Timmermans 2007). Notably, there has been a significant increase in the frequency of short-distance outdoor activities, and the use of walking mode in residential areas during the pandemic period (Di Marino et al. 2023; Padmakumar and Patil 2022).

However, recent research related to COVID-19 has mainly dealt with travel mode choice and leisure activity in separate models, overlooking their potential correlation. Given their interdependence, a high correlation exists between short-distance outdoor activities and the use of non-motorized travel modes, such as walking and cycling (Espiner et al. 2023). Therefore, understanding the simultaneous occurrence of outdoor leisure activities and walking for leisure is crucial to gaining a better understanding of the changes in citizens’ leisure behavior due to the pandemic.

Meanwhile, the COVID-19 pandemic has underscored the importance and utility of urban open green spaces and pedestrian pathways. As a result, urban planners and designers are focusing more on how to improve the urban neighborhood environment in response to the potential future infectious diseases. Some studies have presented strategies for post-pandemic adaptation, including redesigning public spaces and implementing management plans for parks and green areas (e.g., Honey-Rosés et al. 2021; Ugolini et al. 2020). However, there is insufficient empirical evidence to support the development of measures for improving the urban neighborhood environment.

Thus, this study aims to investigate the mutual relationship between outdoor activities and walking for leisure during the COVID-19 pandemic and examine the effects of COVID-19 preventive behaviors on this association. Specifically, the study seeks to (1) explore the interrelationship between outdoor leisure activities and walking for leisure in an urban context during the COVID-19 pandemic, and (2) analyze the effects of preventive behavior adopted to avoid COVID-19 infections on them. To this aim, we employ a three-stage least squares (3SLS) simultaneous equation model using data collected from a sample of 537 residents in the Seoul metropolitan area. The study results can contribute to understanding how the leisure behavior patterns of urban residents have changed in the COVID-19 era. Furthermore, our findings can serve as empirical evidence for developing urban plans and policies to improve the urban neighborhood environment to address future infectious disease challenges.

The structure of this study is shown in Figure 1: the literature review explores previous studies on the interaction between leisure activities and walking, the influence of COVID-19 preventive behavior, and the effect of the perceived risk of COVID-19 infection and highlights how this study differs from previous works. The methodology section describes the conceptual model, data collection process, variables, and analysis method. The results section presents the findings from the 3SLS simultaneous equation model analysis. Finally, the discussion and conclusions sections provide a summary of the findings and their implications.

Figure 1. Flow Chart of Research Process.

2. Literature review

2.1. Interaction between outdoor activities and walking for leisure

Urban outdoor activities encompass all leisure and recreational activities taking place in natural green areas (e.g., mountains, seas, rivers surrounding the city) or artificial open spaces (e.g., outdoor sports facilities and parks) within urban settings (Elbakidze et al. 2023; Phipps 1991). These activities primarily rely on non-motorized transport (NMT) modes such as walking or cycling due to their proximity. Particularly, walking, as a NMT mode, is advantageous for sustainable urban transport, providing health benefits, affordability, accessibility, and social engagement (De Vos et al. 2019).

In the context of leisure behavior, walking can be described either as a general form or as a specific form. Walking as a mode of leisure travel is primarily aimed at moving to a specific destination for leisure or recreation (Buckley, Stangl, and Guinn 2017). Conversely, walking as part of outdoor activities is aimed at enjoying the surrounding scenery, resting, recreation, or exercising, with an emphasis on the act of walking itself (Cerin, Leslie, and Owen 2009; Elbakidze et al. 2023). As such, the definition of walking varies depending on its context and intent (Lanzendorf 2002). Walking and outdoor activities provide urban residents with opportunities for physical activity (De Vos 2020; Sarangi, Manoj, and Bhosley 2022), social interaction (Kaźmierczak 2013), and improving mental health (Elbakidze et al. 2023).

Previous research on the relationships between various outdoor activities (i.e., Beil and Hanes 2013; Mackett and Paskins 2008) indicates that activity influences walking. For example, Mackett and Paskins (2008) examined the relationship between children’s participation in physical activity and travel modes. They found that children who participated in numerous shopping and club activities experienced increased automobile travel, whereas children who participated in outdoor play activities experienced more walking trips. Similarly, Beil and Hanes (2013) found that people who engaged in various outdoor activities also walked a great deal. Overall, previous studies have shown that the more people engage in outdoor activities, the more walking is used as a mode of travel. It is thus reasonable to predict that as outdoor activities increase, walking trips will also increase.

By contrast, research on the reciprocal effect of walking on activity (Brustad 1993; Tudor-Locke, Ainsworth, and Popkin 2001) has primarily been conducted in the fields of physical activity and public health. For example, Brustad (1993) reported that the more children walked, the more they engaged in out-of-home rather than at-home activities; in turn, this increased the overall frequency of physical activity. Tudor-Locke et al. (2001) also confirmed that the more people walk, the greater the likelihood of increased outdoor activity. These studies show that it is reasonable to assume that the frequency of walking for leisure affects the amount of outdoor activity as a form of leisure behavior.

2.2. The effect of COVID-19 preventive behavior

Since leisure activities are not imperative, at least in terms of livelihood, they are more vulnerable to being reduced or eliminated by preventative measures, such as social distancing, that minimize face-to-face interaction (Parady, Taniguchi, and Takami 2020). Previous studies showed that people who take steps to prevent infectious diseases like COVID-19 are more likely to stay at home or avoid out-of-home activities (e.g., De Vos 2020). It is argued that non-compulsory activities, such as leisure activities, are drastically curtailed as people restrict their activities and movement. One result of such changes is increased physical and mental problems.

In general, previous research demonstrates that COVID-19 preventive behavior has a considerable impact on travel mode choices and outdoor activities. Long-distance tourism and travel have been most affected by COVID-19 (Neuburger and Egger 2021), in line with the way previous infectious epidemics, like SARS (Kuo et al. 2008) and H1N1 (Lee et al. 2012), negatively impacted tourism demand. This effect is linked to the fact that people who are more aware of health risks (and thus perceive a higher risk of contracting an infectious disease while traveling) are more likely to cancel or postpone their travel (Cahyanto et al. 2016; Neuburger and Egger, 2021).

Due to the pandemic, people were reluctant to use public transit (e.g., planes, trains, buses), a significant travel mode to tourist attractions. These modes were perceived as vulnerable to infectious diseases and were thus likely to be avoided (Haq, Shahbaz, and Boz 2020). As a result, we might expect that many people settled on leisure activities closer to their homes where walking or bicycles could be used (Abdullah et al. 2020; Liu, Cao, and Pei 2022), as these modes are identified as relatively safe ways to avoid overcrowding or crowded places.

Concerns regarding COVID-19 infection intensified during the pandemic period, raising the perceived risk of multi-use venues or indoor locations (Shamshiripour et al. 2020). Indoor sports activities were believed to be the most dangerous, whereas outdoor sports activities were perceived to be less risky (M. H. Kim, Lee, and Gim 2021). Previous research shows that outdoor camping, hiking, and urban park use increased dramatically (Ugolini et al. 2020). Urban green spaces encourage social contact (Kaźmierczak 2013) while promoting physical activity and mental wellness (Grigoletto et al. 2023). Given this, we can expect that social interactions increased in urban parks and broad open spaces, which were relatively safe from infection during this period, as compliance with preventive behaviors like social distancing greatly restricted social behaviors in other venues.

2.3. The effect of the perceived risk of COVID-19 infection

Human mobility and activity patterns are generally restricted during a pandemic, and individual risk perception of disease infection is emphasized as a determinant explaining these behavioral changes (Park and Reisinger 2010; Sharifpour, Walters, and Ritchie 2014). In particular, previous studies revealed that the perceived risk of infectious diseases such as SARS, H1N1, or Ebola had a considerable impact on leisure and travel activities (Cahyanto et al. 2016; C. Kim et al. 2017). Similarly, a recent study found that the perceived risk of infection during the COVID-19 pandemic significantly affected how people chose travel modes (e.g., Haq, Shahbaz, and Boz 2020; M. H. Kim, Lee, and Gim 2021) and engaged in recreational activities (e.g., Abdullah et al. 2020; De Vos 2020; Shakibaei et al. 2021).

In a study on the perceived risk of COVID-19 infection and travel mode choices (e.g., Barbieri et al. 2021; M. H. Kim, Lee, and Gim 2021; Liu, Cao, and Pei 2022; Ma et al. 2022), the higher people’s perceived risk of infectious disease in enclosed travel modes (e.g., airplanes and buses), the lower the ridership of public transit; people instead shifted to private travel modes (e.g., automobiles and walking) to avoid face-to-face contact. Additionally, NMT modes, such as walking and biking, gained popularity as the sense of safety from COVID-19 infection increased (Liu, Cao, and Pei 2022; Shakibaei et al. 2021; Shamshiripour et al. 2020).

Preferences for non-work-related activities, such as leisure, are also highly volatile under the influence of infectious diseases (e.g., Bucksky, 2020). For example, Abdullah et al. (2020) demonstrated that when public knowledge bearing on the perceived risk of COVID-19 infection increases, outdoor activities grow rather than indoor ones. Bayrsaikhan et al. (2021) confirmed that a higher sense of a given city’s safety from COVID-19 infection is associated with increased outdoor leisure space visits. Parady et al. (2020) and Shakibaei et al. (2021) also found that outdoor-based activities increased rather than indoor-based activities (e.g., shopping and eating out), although overall leisure activities declined during the pandemic.

Previous research indicates that the perceived risk of COVID-19 infection affects both travel modes and leisure activities, although the underlying causes may be distinct. In this context, the perceived risk of COVID-19 infection via travel modes such as walking, public transit, or automobiles refers to the risk of infection that may occur while moving to a destination and thus should be distinguished from the perceived risk attached to the locations where leisure activities take place. In other words, views about the safety of transit modes (e.g., public transit, private automobile, and walking) may influence the frequency of walking regardless of destination.

Meanwhile, determinants affecting the frequency of outdoor activities include the risk perception of natural green spaces (e.g., mountains and rivers) and open spaces (e.g., outdoor sports facilities, playgrounds, and parks) in the urban area. After the outbreak of COVID-19, urban natural areas and open spaces were acknowledged as relatively safe (Ugolini et al. 2020), and outdoor climbing and park visiting did, in fact, increase (Bae and Chang, 2020).

In summary, previous studies of leisure travel during COVID-19 revealed that both outdoor leisure activities and pedestrian travel increased during this period, but without addressing the association between these two variables. Moreover, the impact of the perceived risk of COVID-19 infection has been extensively studied, whereas the effects of COVID-19 preventive behavior are insufficient. This study differs from previous research by considering the bidirectional relationship between walking and outdoor activities in the context of COVID-19, while also confirming the impact of COVID-19 preventive behavior.

3. Methods and data

3.1. Data

This study focused on the Seoul metropolitan area, South Korea. The Seoul metropolitan area, with more than half (51%) of the entire Korean population, is a central location with easy availability of various types of leisure activities and means of transportation. Figure 2 shows the transport system including information on the location of the metro and bus stop in the Seoul metropolitan area. Therefore, it can be seen as the most appropriate place to understand the changes in leisure activities and mobility patterns of urban residents brought about by COVID-19.

Figure 2. Transport system (subway and bus) in Seoul Metropolitan area.

The primary data¹ were collected via a web-based survey of 537 residents (20 and above) in the Seoul metropolitan area. This online survey² was conducted by a specialized research institute (http://www.entrustsurvey.com/.) from September 23 to 7 October 2020. The respondents voluntarily participated in the survey and were selected based on gender and age using a quota sampling technique based on previous research (Bawazir et al. 2018; Jang and Baek, 2019) on infectious diseases. In quota sampling, the entire sample is categorized, and subjects are selected from each category. Specifically, the age criterion was set to include only individuals aged 20 and above, excluding respondents under the age of 20. This exclusion was due to the high likelihood of systematic error arising from the substantial influence of adults on the outdoor activities and travel of those under the age of 20. The obtained sample comprises individuals in their 20s (20.3%), 30s (34.4%), 40s (34.6%), 50s (8.6%), and those 60 and above (2.2%). In terms of gender, the sample is distributed as 272 females (50.7%) and 265 males (49.3%).

This study utilized inferential statistics on a sample segmented by demographic characteristics (age and gender) through a non-probabilistic quota sampling method. Unlike descriptive statistics, which depend solely on sample representativeness, inferential statistics prioritize the adequacy of variation within the sample data. Thus, the variability of sample is sufficient for use in inferential statistics, as indicated in Table 1.

Table 1. Descriptive statistics (n = 537).

Variable	Code name	Description	Mean	St. Dev.	Min	Max
Leisure walking	Walking	Frequency of walking for leisure	2.541	1.757	0	5
Outdoor activities	Outdoor activities	Frequency of outdoor leisure activities	2.112	0.954	1	5
COVID-19 preventive behavior	SDP1	Postponing or canceling visits to suburban areas for travel/leisure activities	2.839	1.564	1	5
	SDP2	Postponing or canceling social exchange activities	3.424	1.363	1	5
	SDP3	Reducing outings for purposes of eating out or leisure	3.951	1.101	1	5
Perceived risk of COVID-19 infection (travel modes)	RP_public	Perception of safety from COVID-19 infection (public transit)	−2.698	2.301	−5	5
	RP_car	Perception of safety from COVID-19 infection (car)	4.140	0.894	1	5
	RP_walk	Perception of safety from COVID-19 infection (walking)	3.473	0.952	1	5
Perceived risk of COVID-19 infection (outdoor activity spaces)	RP_open	Perceived risk of COVID-19 infection (open spaces: parks, playgrounds, outdoor sports facilities, etc.)	−0.670	2.263	−5	5
	RP_nature	Perceived risk of COVID-19 infection (urban natural green spaces: mountains, beaches, rivers, etc.)	0.531	2.465	−5	5
Socio-demographics	Gender†	Gender (0 = male, 1 = female)	0.506	0.500	0	1
	Age†	Age (years)	38.192	9.865	20	76
	Edu	Education (0= less than bachelor’s degree, 1= bachelor’s degree or higher)	0.624	0.166	0	1
	Marriage	Marital status (0 = single, 1 = married)	0.527	0.499	0	1
	Income	Household income (million won)	505.337	337.331	0	4.500
	num_HH	Household size (persons)	2.974	1.209	0	8
	num_prechild	Number of children at preschool level (persons)	0.166	0.449	0	3

†The sample comprises 265 (49.3%) males and 272 (50.7%) females in their 20s, 30s, 40s, 50s, and 60s and over, with 109 (20.3%), 184 (34,3%), 186 (34.6%), 46 (8.6%), and 12 (2.2%) respondents for each group.

3.2. Variables

This study used variables such as frequency of walking for leisure, frequency of outdoor leisure activities, perceived risk of COVID-19 infection, COVID-19 preventive behavior, and demographic characteristics. The frequency of walking for leisure was operationalized as the average weekly amount of leisure walking in the last month after the COVID-19 pandemic; because it expected leisure walking to have a higher variance than outdoor activities. The frequency of outdoor leisure activities was operationalized as the number of times such activities were performed in the last week after the COVID-19 pandemic. For the analysis, we used the values added up with the “frequency of recreation activities in parks and green spaces” and the “frequency of using outdoor sports facilities.”

We employed instrumental variables (IVs) in our conceptual model to control for confounding and measurement errors because they could adjust for both observed and unobserved confounding effects (Gallant and Jorgenson 1979). Thus, drawing on findings of previous studies, the risk perception of COVID-19 infection, which affects walking and outdoor activity, was included as an instrumental variable (e.g., Haq, Shahbaz, and Boz 2020; Ma et al. 2022; Shamshiripour et al. 2020).

Specifically, the instrumental variables of the walking model were set as the perceived risk of COVID-19 infection from using the travel mode. This variable indicates the perceived risk of COVID-19 infection when traveling to one’s destination (e.g., Haq, Shahbaz, and Boz 2020; M. H. Kim, Lee, and Gim 2021), while it has no association with the frequency of leisure activities. Thus, instrumental variables in the walking model were constructed by measuring how safe the three travel modes of walking, public transit, and car were perceived to be during the COVID-19 pandemic, based on a 5-point Likert scale (1 “Not safe at all” to 5 “Very safe”).

Meanwhile, the instrumental variables of the outdoor activities model included the risk perception of COVID-19 infection from outdoor activity place (e.g., Abdullah et al. 2020; Bayrsaikhan et al. 2021; Shakibaei et al. 2021), which was predicted to be unrelated to travel modes. The perceived risk of COVID-19 infection from each outdoor place was measured using an 11-point Likert scale. The use of an 11-point Likert scale allowed for the simultaneous examination of both the negative and positive impacts of risk perception on urban outdoor leisure places in the COVID-19 pandemic. Therefore, regarding psychometric properties, Leung (2011) indicated that having more scale points may reduce skewness, and the 11-point scale exhibits the smallest kurtosis and is closest to normal distribution. Thus, we adopted a symmetric 11-point Likert scale.

Exogenous variables were constructed to capture COVID-19 preventive behavior. Personal preventive behaviors related to infectious diseases such as COVID-19 are closely related to changes in people’s outdoor activities and travel behavior (Neuburger and Egger, 2020). Thus, this study assumed that personal behaviors for the prevention of COVID-19 infection would affect outdoor leisure activities and walking for leisure, focusing specifically on three exogenous variables: postponing or canceling visits to suburban areas for travel/leisure activities, postponing or canceling social exchange activities, and reducing outings for purposes of eating out or leisure. All variables were measured using a 5-point Likert scale. The adoption of this scale was based on its demonstrated efficacy in assessing indicators pertaining to behavior and perception, as evidenced by prior research on MERS infection (Yang and Cho, 2017).

Finally, individual characteristic variables (gender, age, educational background, marital status) and household characteristic variables (household income, number of households, and number of household preschool children) were employed as control variables. All measured variables were standardized before analysis; Table 1 shows the descriptive statistics of the variables. Overall, variations in all research variables appeared to be sufficient for statistical reasoning.

3.3. Analysis

The purpose of this study is to examine the mutual relationship between the frequency of walking for leisure and the frequency of outdoor leisure activities of urban residents during the COVID-19 pandemic and the effects of COVID-19 preventive behavior on this relationship. It is possible that these two endogenous variables can be modeled separately. However, using separate single-equation models (e.g., ordinary least squares) would fail to consider the potential correlation between the frequency of walking for leisure and the frequency of outdoor leisure activities. These two variables are interrelated, allowing the endogenous variable in one equation to potentially be the independent variable in the other.

Then, a variable cannot be independently viewed as an endogenous or exogenous variable in this case. In other words, if urban residents’ restraints on leisure activities increase, the frequency of outdoor leisure activities decreases, leading to a decrease in the quantity of leisure walking. Alternately, if walking for leisure decreases or increases, the frequency of leisure activities would consequentially decrease or increase. To consider a mutual relationship of this sort, it is suitable to use a simultaneous equation model (Zellner and Theil 1992) that can better capture the interrelationship between outdoor leisure activities and leisure walking than a single equation model. So, we jointly estimated two models – the frequency of walking for leisure and the frequency of outdoor activities – using three-stage least squares (3SLS). Figure 3 shows the research conceptual model based on existing studies.

Figure 3. Conceptual model.

4. Results

Table 2 presents the results of the 3SLS analysis. The interaction between outdoor activity and walking for leisure was found to be significant (coef. = 0.368, p < 0.1) in the walking frequency model. Moreover, walking for leisure had a positive effect on outdoor leisure activities (0.331, p < 0.01) in the outdoor leisure activity model.

Table 2. Three-Stage Least Squares (3SLS) regression results.

	Model 1: Walking			Model 2: Outdoor activities
	Coef.		p> z	Coef.		p> z
Walking	–		–	0.331***		0.005
Outdoor activities	0.368*		0.089	–		–
SDP1	0.079*		0.094	0.043		0.187
SDP2	−0.085		0.144	0.117***		0.002
SDP3	0.190***		0.003	−0.176***		0.000
RP_public	−0.121**		0.029	–		–
RP_car	−0.181***		0.004	–		–
RP_walk	0.353***		0.000	–		–
RP_open	–		–	0.158***		0.000
RP_nature	–		–	0.151***		0.001
Gender	−0.378		0.228	−0.036*		0.072
Age	−0.010		0.852	−0.021		0.576
Edu	0.176*		0.057	−0.086		0.195
Marriage	0.031		0.404	−0.005		0.841
Income	−0.032		0.461	0.056*		0.051
num_HH	−0.068		0.536	0.056		0.457
num_prechild	0.069		0.519	−0.019		0.795
_cons	0.165		0.141	0.034		0.629
R-squared	0.114			0.087
N	537			537

Note: * p < 0.1, ** p < 0.05, *** p < 0.01.

The instrumental variables used in the study, including “the perceived risk of infection with COVID-19 on public transit” (−0.121), “the perceived risk of infection with COVID-19 in an automobile” (−0.181), and “the perceived risk of infection with COVID-19 when walking” (0.353), had a significant effect on walking frequency. Specifically, the belief that public transit and automobile use are safe from COVID-19 had a negative effect on walking frequency. In other words, a higher belief that public transit and automobiles are safe from COVID-19 results in a decrease in walking frequency, while a greater perception of walking being safe from COVID-19 infection led to an increase in walking frequency.

Similarly, the instrumental variables for outdoor leisure activities, such as the perceived safety of urban “open spaces” (0.158) and “natural green spaces” (0.151) from COVID-19, had a positive effect on outdoor leisure activities. This result corroborates previous findings and proves the validity of the instrumental variable selection used in each model.

Exogenous variables such as postponement or cancellation of out-of-town leisure visits (SDP1) had a statistically insignificant effect on the frequency of outdoor leisure activities, but increased walking frequency (0.079). On the other hand, postponing or canceling social interaction (SDP2) had a statistically insignificant effect on walking for leisure but a positive effect on outdoor leisure activities (0.117). These results demonstrate that when out-of-town leisure visits are postponed or canceled, walking for leisure increases, and also, when social interactions are delayed or canceled, outdoor leisure activities increase.

Notably, for those who stated they avoided going out for leisure and instead spent time at home (SDP3), the frequency of outdoor activities decreased (−0.176) while the frequency of walking for leisure increased (0.190). Finally, among the control variables, education had a statistically significant effect on walking (0.176), while gender and income had statistically significant effect of −0.036 and 0.056 on outdoor leisure activities, respectively. The model used in this study ensured the validity of the findings by taking into account demographic and social characteristics, including individual and household characteristics.

The main findings of this study are summarized in Figure 4. First, the study found a bidirectional association between outdoor activities and walking for leisure. Further, cancellation of out-of-town leisure activities (SDP1) was associated with increased walking for leisure. It also found that while delayed or canceled social interaction activities (SDP2) did not show statistically significant impacts on walking frequency, they did lead to an increase in outdoor activities. Given that the increase in outdoor activities significantly impacted the rise in walking frequency, it likely had a positive effect on overall walking frequency. Finally, when all other were the same conditions (ceteris paribus), individuals who avoided going out for leisure (SDP3) showed decreased outdoor activities while increased walking frequency. This indicates a decrease in the use of enclosed modes of transportation, such as buses and subways, and an increase in walking instead.

Figure 4. Coefficients of the 3SLS model.

5. Discussion

Our results indicate a mutual relationship between outdoor activities and walking for leisure. In other words, there is a positive and synergistic relationship between the frequency of outdoor leisure activities and walking for leisure during the COVID-19 pandemic. This finding is consistent with previous research by Tudor-Locke et al. (2001) and Song et al. (2020), which demonstrated that walking frequency rises when outdoor activity increases, and vice versa. We also found that reducing out-of-town leisure and social interaction activities leads to increased walking for leisure. These findings are related to social distancing measures used during the pandemic, such as bans on large gatherings and indoor facilities (Kim et al. 2021; Shamshiripour et al. 2020).

These changes have the potential to influence the mental health of individuals (De Vos 2020), resulting in a preference for walking to increase subjective well-being. Existing evidence (Kaźmierczak 2013; Sarangi, Manoj, and Bhosley 2022) shows that non-motorized mode use improves physical activity, social interaction, and mental health. Indeed, Abdullah et al. (2020) and Sarangi et al. (2022) present initial evidence supporting the claim that during the pandemic, individuals engage in leisure activities like walking or cycling near their residences.

Increases in walking can provide insight into the prioritization of policy initiatives aimed at improving the physical environment. To achieve this, there is a need to focus on creating broader pedestrian areas in residential areas to improve the pedestrian environment and ensure easy access to outdoor activity spaces. Therefore, improving the pedestrian environment can increase both the quantity and quality of outdoor activity and walking for citizens.

An alternative way is to expand broader pedestrian areas by adapting existing roads or railways into pedestrian-only thoroughfares, which would expand pedestrian areas and improve accessibility by linking surrounding facilities and green spaces. This may be more acceptable, especially in residential regions with limited budgets and challenges securing new places.

For instance, Seoul has implemented policy initiatives such as the “Seoul Road Diet” and the “Seoullo 7017” project. The “Seoul Road Diet” is a part of the Walking City, Seoul policy, which reduces road lanes to create pedestrian-only streets. Seoullo 7017 is another case of renovating an outdated elevated highway into pedestrian-only thoroughfares.

These pedestrian-only thoroughfares can enhance accessibility and connectivity with nearby utilities and public spaces through a grid-like design. Urban planners and policymakers should consider integrating these pedestrian roads with the existing transportation system as they can serve as vital connecting hubs for nearby areas. Our study also indicates a positive interaction between walking and outdoor leisure activities. Given this, the expansion of pedestrian areas could have a synergistic effect on increasing the level of service of the pedestrian environment in the residential areas if it leads to improved access to existing parks or green spaces.

Another suggestion is to apply the pedestrian mall model or the shared streets model as an alternative to expanding broader pedestrian areas in neighborhood zones.

We found that restrictions on social interactions due to COVID-19 have led to an increase in walkable outdoor activities. Thus, pedestrian spaces need to be utilized in social distancing situations not only for outdoor activities or mobility but also as alternative places for disconnected social interaction. In this regard, the qualitative level of urban pedestrian environment services can be improved by designing more diverse utilities and securing pedestrian-oriented spaces transforming streets in neighborhoods or commercial districts into pedestrian malls.

In addition, our results also underscore the importance of walking mode for leisure, indicating an increase in the use of walking mode more than enclosed modes such as public transport and cars for leisure travel during the pandemic. This suggests that pedestrian-friendly roadways in residential areas are required to meet the increasing demand for outdoor activities and walking in the post-pandemic era. The conversion of vehicle-oriented “mixed-use” roadways into pedestrian-oriented ’shared’ streets is a crucial consideration to consider when designing pedestrian-friendly environment.

The Woonerf in the Netherlands is a successful implementation of the shared streets model. By removing the boundary between car lanes and sidewalks, and applying pedestrian-centric design elements such as trees and benches, Woonerf creates an atmosphere where pedestrians feel prioritized. Furthermore, by sharing parking spaces as a community and play area, it was possible to design a walking environment that could be used by diverse entities consistently. This emphasizes the need to take into account not only traffic management functions and regulations but also the application of spatial design features to improve the convenience and safety of citizens walking while also creating a pedestrian-centered coexistence walking environment.

6. Conclusions

Walking is an essential component of sustainable urban transport strategies, while outdoor activities are crucial for urban green space planning. There is growing interest in the role of walking for leisure in urban settings, particularly during the COVID-19 pandemic, due to the physical and mental health benefits of outdoor recreation activities. Therefore, this study aimed to investigate the simultaneous relationship between outdoor leisure activities and walking mode for leisure and to examine how COVID-19 preventive behaviors affect this relationship among Seoul residents during the pandemic.

This study has two major contributions. Academically, this study contributes to the existing literature by examining the interaction between walking and outdoor activities during the pandemic and analyzing the effects of COVID-19 preventive behaviors. In addition, the study extends the methodological scope of the existing body of knowledge by utilizing a 3SLS simultaneous equation approach.

Practically, our findings provide insight into the simultaneous relationship between outdoor activities and walking for leisure. The study also highlights the importance of considering the impact of COVID-19 prevention behaviors on leisure and travel behavior. These findings can inform interventions that promote outdoor activities and encourage people to walk more during the pandemic. Given the expected changes in the leisure and travel behavior after COVID-19, it is crucial to identify the link between leisure activities and travel modes as well as to observe the impact of preventive behavior to develop suitable countermeasures.

Despite the significant contribution, this study is not without limitations. First, we did not examine the satiation changes in the patterns of leisure activities and walking before the COVID-19 outbreak. Future studies could investigate changes in leisure activities and walking patterns after the COVID-19 pandemic. More specific insights into improving urban leisure spaces and transportation planning strategies could be gained by connecting predicted post-pandemic outdoor leisure activities and walking mode alternatives.

Second, we did not investigate the single linear causal relationship between two main factors, walking and outdoor activities, nor did we examine interaction effects such as trade-offs between variables. Further research should examine multiple interaction effects between variables. Third, our study employed non-probability quota sampling to ensure sufficient variability for inferential statistics, which limits the generalizability of our findings to groups homogeneous with the composition of our sample. Future research should use probability sampling techniques to strengthen the generalizability of the findings.

Ethics statement

The questionnaires were authorized by the Institutional Review Board of the researcher’s institution (IRB No. 1908/002–022).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the National Research Foundation of Korea [NRF-2021S1A3A2A01087370].

Notes on contributors

Tsolmon Bayarsaikhan

Tsolmon Bayarsaikhan is a Ph.D. candidate in environmental management at Seoul National University, Seoul, South Korea. Her current research interests include the study of urban leisure environments, focusing on leisure places and behavior in urban areas.

Moon-Hyun Kim

Moon-Hyun Kim, Ph.D., is an associate research fellow at Korea Institute of Public Administration, South Korea. His research primarily focuses on urban planning and transportation, with an emphasis on sustainable urban mobility, transportation policy, and the integration of transportation systems within urban development.

Tae-Hyoung Tommy Gim

Tae-Hyoung Tommy Gim, Ph.D., is an associate professor in the Graduate School of Environmental Studies and is jointly affiliated with the Interdisciplinary Program in Landscape Architecture and the Environmental Planning Institute at Seoul National University, Seoul, South Korea. He is also the director of the SNU Integrated Planning Lab. His fields of expertise include land use-transportation-environment interactions and quantitative methods.

Notes

¹ Dataset can be downloaded from: https://drive.google.com/drive/folders/1qFbpAfo9QCcZnBEEEcm1cqYcUF4nojJj?usp=sharing.

² The questionnaires were authorized by the Institutional Review Board of the researcher’s institution (IRB No. 1908/002–022)