Did the built environment attenuate reductions in leisure walking during COVID-19? A quasi-panel study

Did the built environment attenuate reductions in leisure walking during COVID-19? A quasi-panel study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Did the built environment attenuate reductions in leisure walking during COVID-19? A quasi-panel study Mahsa Naseri, Alexa Delbosc, Liton Kamruzzaman This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3977307/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract The COVID-19 pandemic caused decreased physical activity levels due to isolation, travel restrictions, and facility closure. This meant that walking remained the main option for individuals to sustain their physical well-being and mental health. This study examines changes in walking behaviour during the early years of the pandemic, and how such changes were affected by the built environment characteristics of Melbourne neighbourhoods over the period of lockdowns in 2020. By evaluating the impact of built environment characteristics on leisure walking patterns during the 2020 lockdowns, we provide insights into the interplay between the built environment and physical activity. We found that self-reported leisure walking decreased notably during the COVID-19 restrictions. The influence of the built environment on leisure walking remained consistent throughout the pandemic stages. Factors such as green space, residential density, and land-use diversity demonstrated associations with leisure walking. The presence of more green spaces was linked to higher rates of walking, while moderate residential density was associated with the highest walking rates. Surprisingly, more diverse locations showed lower levels of walking, potentially due to pandemic-related closures of non-essential businesses or limited access to green spaces in these areas. These findings emphasize the importance of considering built environment characteristics in promoting and maintaining physical activity levels, even during times of restricted movement. Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction COVID-19-related restrictions led to a decline in physical activity levels for many people around the world (Runacres et al., 2021 ). In the first two years of the pandemic, most people experienced periods when fitness facilities and recreational centres were temporarily closed. Consequently, engaging in walking activities emerged as a readily accessible and effective means for many individuals to sustain their physical well-being and support their mental health. Walking can also facilitate social interaction while maintaining social distances, especially during the periods of pandemic when people were forced to stay in their neighbourhood and the risk of social isolation and loneliness increased (Lättman et al., 2023 ). Taken together, walking likely played an important role in supporting mental and physical health during the early pandemic years. Yet much of the transport research coming out of the pandemic years focussed on public transport ridership, commuting behaviour and working from home, with fewer studies looking at non-commute and active travel (Hook et al., 2023 , Cusack, 2021 , McElroy et al., 2022 ). Decades of research before the pandemic highlights the importance of the built environment on walking (Saelens and Handy, 2008 , Ewing and Cervero, 2010 ). One might hypothesise that the built environment would have a particularly salient impact on leisure walking during the pandemic, when many people were restricted to their neighbourhoods, forcing them to rely on their local built environment to facilitate (or discourage) walking. Yet to date, very few studies directly examined the relationship between leisure walking and the built environment during the pandemic (Lotfata et al., 2022 ). This paper aims to estimate the association between built environment characteristics and leisure walking behaviour during the first year of the pandemic. In particular, we hypothesised that one’s neighbourhood built environment can attenuate reductions in leisure walking during COVID-19 lockdowns. The COVID-19 lockdowns implemented in Melbourne provided a unique opportunity to assess this relationship, as Melbourne residents experienced some of the longest and most restrictive lock-down regulations in the world. Literature review COVID-19 and walking Urban mobility since 2020 was considerably impacted by COVID-19 when different actions were taken globally in order to reduce the spread of the virus, such as social distancing, movement restrictions, working from home and travel bans. Urban dynamics and mobility changed and this change was significant in the modal distribution of trips (Barbieri et al., 2020 ). Trends in walking were impacted by changes in work and exercise opportunities. As work from home orders were introduced, and people did not have access to out-of-home entertainment and exercise facilities, the distribution of active travel was affected. In response, people maintained their physical and mental health through recreational active travel such as walking (De Vos, 2020 ). The change in recreational walking during COVID-19 was not clear, with some studies finding increases and others showing decreases. Hunter et al. ( 2021 ) utilized data from mobile devices and area-level data from 1.62 million mobile phone users to examine how much people walked in 10 metropolitan areas across the United States. The time frame of the analysis spanned from mid-February 2020, which was before the implementation of lockdown measures, to late June 2020, when some of these restrictions were relaxed. Their results indicated that there was a 70% reduction in the number of walks and a 50% decrease in the average distance walked in all metropolitan areas during the pandemic. Although there was a steady increase in walking from mid-April 2020, after some commercial and business activities resumed, walking remained approximately 18% below pre-pandemic levels. A study by McElroy et al. ( 2022 ) found a similar trend. Using a self-reported survey from California as well as a national sample, they found that walking as a commute mode decreased during the pandemic. Both studies also found that unlike walking for commuting, recreational walking increased in comparison with pre-COVID period (McElroy et al., 2022 , Hunter et al., 2021 ). Similarly, Power et al. ( 2023 ) found recreational walking and utilization of trails, particularly those in proximity to urban areas, in Ireland increased. Their research demonstrated that trail usage (based on counter data) in 2020 has shown a general rise of 6% as compared to the previous year. Additionally, during the most severe COVID-19 lockdown periods in 2020, trails located within a 2 km radius of urban areas recorded an increase in average footfall counts of up to 102% compared to those situated outside of the 2 km range. Built environment and walking A long history of research explores how built environment characteristics affect travel behaviour (Ewing and Cervero, 2010 ). Walking is associated with having access to destinations, connected street network, and higher residential density. Mixed land-use is found to increase walking rates as they provide access to a greater variety of destinations within walking distance (Saelens and Handy, 2008 ). A higher residential density creates critical mass for more people to walk, to see others walking and thereby to promote a safer environment for walking (Kamruzzaman et al., 2016 ). Indirectly, higher density supports diverse land use patterns and ultimate creates a favourable condition for walking (Saelens and Handy, 2008 ); it could be. Street network pattern usually reflects connectivity, which has a positive effect on walking (Sarkar et al., 2015 , Marshall and Garrick, 2010 ) by providing convenient access to destinations, direct routes, increased safety options, and a more enjoyable walking environment. Green spaces have a positive effect on physical activity and walking (Sugiyama et al., 2013 ) because they provide individuals with more access to nature, which serves as an inviting and stimulating environment that encourages people to engage in outdoor activities. People who have more access to green space, or even street trees, tend to walk and be more physically active than those with poor access (Dadvand et al., 2016 , Sarkar et al., 2015 ). Therefore, the built environment characteristics can either facilitate or discourage walking in a neighbourhood. However, it should be noted that the effects of the built environment on recreational walking are less clear, compared to walking for transport (Saelens and Handy, 2008 ). In addition, recent studies have found that the built environment was a stronger predictor of walking for transport, whereas socio-demographics have a greater impact on walking for leisure (Yin et al., 2023 , Liu et al., 2021 ). The role of socio-demographics will be explored in greater depth in the next section. As the COVID-19 pandemic imposed restrictions on movement and encouraged individuals to stay close to their homes and neighbourhoods. With limited opportunities for travel and access to external recreational spaces, people relied heavily on their local built environment for physical activity. One might hypothesise that because many people were restricted to activity in their local neighbourhood during the pandemic, the built environment might have had an even greater impact on walking during the early pandemic years. Yet despite a long-standing history of research on the relationship between the built environment and active travel (Ewing and Cervero, 2010 ), at the time of writing only three studies explored this topic in the context of the COVID-19 pandemic. McCormack et al. ( 2022 ) conducted a qualitative study that explored how the built environment in Canada supported outdoor physical activity during the early stages of the COVID-19 pandemic. The study found that parks and pathways were important for maintaining outdoor activity, but crowding and lack of physical distancing were concerns. One potential limitation of this study is the small sample size, as only 12 adults were interviewed. This may not be representative of the larger population or capture the experiences of diverse groups. Additionally, the study did not consider an objective measure of built environment for the analysis. Hook et al. ( 2021 ) examined the impact of the COVID-19 lockdown on travel behaviour in Flanders, Belgium. Their findings revealed that the built environment did not have a significant impact on travel behaviour changes. This study used an "urbanization factor" to measure the built environment, but did not analyse specific characteristics of the built environment within that measure. Finally, Shaer et al. ( 2021 ) examined the impact of COVID-19 and related policies on active travel of Central Business District (CBD) residents compared to non-CBD residents in Shiraz, Iran. The study investigated the relationship between perceived built environment factors and active travel before and after the outbreak. The study found that the quality of built environment factors in the CBD led to longer average walking times among CBD residents compared to non-CBD residents. They used subjective measures to assess the built environment, which means that individuals may have different perceptions or interpretations of the environment. None of these studies used objective measures for specific built environment characteristics to isolate its effect. There is a lack of knowledge regarding how people in various neighbourhoods with different spatial characteristics respond to the COVID-19 pandemic. Data and methods Study area The study area of this research was Melbourne, the capital and second most-populous city of the Australia. Melbourne experienced over 260 days in lockdown (i.e. stay-at-home orders) during 2020–2021. In this study we focused on Melbourne’s second and longest lockdown, which lasted for 111 days started from early July 2020 until late October 2020. During this period, work from home orders were introduced, major events were cancelled and many schools and non-essential services were closed. In addition, unlike the previous lockdown, in the second lockdown a 5-kilometre travel restriction was introduced which forced people to stay in their local neighbourhood and use their local (outdoor) facilities for recreation purposes and outdoor activities should be limited to 2 hours per day. The unique characteristics of Melbourne’s lockdown, particularly the 5-kilometre travel restriction imposed in lockdown 2, made the city a salient case study for travel behaviour researchers. Data sources Two types of data were used in this study. First, we drew upon a travel behaviour survey of Melbourne residents to quantify their self-reported walking for leisure. Second, we used open spatial data sets to quantify the built environment characteristics of each survey respondent’s residential neighbourhood. Survey data We were provided access to the ‘C-19 Long Term Transport Impact Study’ in Melbourne (Currie et al., 2021 ) which is a quasi-longitudinal survey including 2,163 people who lived in Melbourne in mid-2020. The survey involved an online questionnaire which used a market research panel company to recruit a representative sample of Melbourne residents. The sampling frame employed quotas based on income, age, and location. The sampling frame was designed to recruit from three different locations: inner, middle, and outer Melbourne. The researchers identified the share of the population in each of the age and income cohorts for each region of Melbourne using the census. The sampling frame was created to ensure that the survey sample matched the census in terms of these characteristics. More information on the survey and sampling methodology can be found in (Currie et al., 2021 ). This data set was collected in two different legs; the first one occurred during Melbourne’s first lockdown and the second one captured people’s travel behaviour adaptation across both the first and second lockdown. As we were interested on behaviour changes during different stages of lockdown, we used the second wave of the data set (sample size 1,341 responses), which ran from 16 July to 8 August 2020, within the first month of Melbourne’s second lockdown period. The study as a whole focused on wide range of information with a focus on public transport use and working from home. For the purpose of this paper, we only studied self-reported walking for leisure (not commuting). The sociodemographic characteristics of this survey are provided in Table 1 . We included these demographics as control variables in our models, given the significant association between demographics and leisure walking (Yin et al., 2023 , Liu et al., 2021 ). Table 1 Sociodemographic characteristics of survey sample Socio-demographic characteristic Survey Sample Percentage Melbourne average (ABS 2021) Gender Male 37.9% 49.2% Female 62.0% 50.8% Income Low (0- $ 530 per week) 34.0% 32.1% Medium ( $ 530- $ 1870 per week) 48.7% 43.4% High (> $ 1870 per week) 17.3% 24.5% Age Youth (15–24) 11.8% 12.3% Adult (25–64) 72.0% 54.5% Older adult (> 65) 16.1% 15.1% Employment Employed 55.4% 88.5% Non-worker 36.2% 6.2% Not in labour force 8.4% 5.3% Education Non-university degree 40.5% 67.2% Higher education 59.5% 32.8% House hold type Single 20.0% 24.8 Couple with children (Family) 23.5% 37.93% Other 56.5% 36.37% Housing type Flat, unit or apartment 39.6% 15.6% Free-standing/ detached 48.4% 67.8% Town house/ semi detached 12.0% 16.2% Participants report the frequency of their leisure walking per week in three stages: pre-COVID, during the first lockdown and during second lockdown (the present, at the time of the survey). Responses were recorded on a seven-point scale: 0. Didn’t do this 1. 1 time a week 2. 2 times a week 3. 3 times a week 4. 4 times a week 5. 5 times a week 6. More than 5 times a week Note that because these walking behaviours were measured retrospectively, changes between ‘pre-COVID’ and the second wave of COVID rely on respondent recall. Quantifying the built environment Because the C-19 survey was not originally intended for spatial analysis, respondents only provided their suburb rather than their street address. Therefore, we had to map respondent’s suburb on the most relevant spatial resolution used in the Australian census, Statistical Area 2 (SA2). These SA2 units contain on average 10,000 residents, their median size is approximately 9 square kilometres in urban areas and they are meant to represent a community that interacts spatially or socially. Based on the literature, we considered built environment indicators including residential density (Forsyth et al., 2007 ), green space area (Watts et al., 2013 ), and land-use mix (Cervero and Kockelman, 1997 ), connectivity (Bentley et al., 2018 ), and cycling infrastructure coverage (Buehler and Dill, 2016 ) which have been found to influence active travel behaviour. It is worth mentioning that cycling infrastructure density is not normally associated with leisure walking, we decided to kept this in the model as some bike infrastructure is shared with walkers. Moreover, by including cycling infrastructure, the model recognizes the interconnections between different modes of active transport and the importance of holistic planning and design. Considering cycling infrastructure alongside walking infrastructure helps capture the broader context of active travel behaviour and provides insights into how the built environment impacts various modes of active transport. In order to calculate residential density, we extracted dwelling numbers in each neighbourhood from the 2016 census data ( https://www.abs.gov.au/ ). Residentional area percentage was calculated based on the area of the residential blocks in each neighbourhood. Similarly, green space percentage was determined based on the parklnads area in each neighbourhood. Land-use diversity measure was calculated by simpson’s diversity index (formula 1) in which a is the total area of specific land-use in neighbourhood and A is the total area of all land-use categories). This index value ranges from 0 to 1 and the higher value, the more diverse land-use pattern (Kamruzzaman and Hine, 2013 ). Land-use diversity = 1- \(\sum \left(\frac{a}{A}\right)\) 2 (1) For cycling infrastructure measure, Vicroads and AURIN open data were downloaded and GIS methods were used to measure the length of bike paths in each neighbourhood. To remove the effect of neighbourhood area and have a normilized comparision in the next steps, we divided total bike paths length (km) by neighbourhood area (km 2 ). It is worth mentioning that cycling infrastructure density is not normally associated with leisure walking, as only some bike infrastructure is shared with walkers. For the connectivity index, we calculated the number of intersections with 3 legs or more and, applying the same logic as cycling infrastructure, we divided this value by neighbourhood area (km 2 ). Because the impact of the built environment on travel behaviour is likely to be non-linear (see for example Yin et al., 2023 , Tao et al., 2020 ), we divided these built environment characteristics into quartiles. The distribution of built environment variables is provided in Table 2 and the spatial distribution of them is provided in Fig. 1 . Table 2 Built environment measurement method and their quartile measurement Measure Measurement method Lowest quartile Second quartile Third quartile Highest quartile Residential Area percentage Area of residential blocks divided by neighbourhood area (Km 2 ) 0-34.47 34.47–58.89 58.89–75.68 75.68–97.66 Residential density Number of dwellings divided by neighbourhood area (Km 2 ) 0-581.79 581.79-918.825 918.825-2075.586 2075.586-6874.92 Green space percentage Area of parklands divided by neighbourhood area (Km 2 ) 0-6.09 6.09–11.38 11.38–18.38 18.38–66.31 Land use diversity Simpson diversity index 0-0.38 0.38–0.52 0.52–0.68 0.68–2.29 Cycling infrastructure Bicycle lane length (km) divided by neighbourhood area (Km 2 ) 0-0.71 0.71–1.66 1.66–3.24 3.24–7.09 Connectivity Number of intersections divided by neighbourhood area (Km 2 ) 2.36–54.07 54.07–65.85 65.85-105.42 105.42-225.03 Analytical methods First, we used descriptive statistic techniques to show the variation of leisure walking across different built environment settings and sociodemographic groups. Moreover, to visually represent the distribution of these walking behaviour categories, the participants were divided into four distinct groups based on their reported walking behaviour. Second, to investigate the relationship between the built environment, demographics, and time period on leisure walking rates, multivariate models were employed. These models allowed for the isolation of the individual effects of each variable and their interactions, enabling a more nuanced understanding of their influence on leisure walking and cycling rates. As the survey included self-reported leisure cycling and walking at three different time periods (before COVID, first lockdown, and second lockdown), it is necessary to use repeated measure modelling to investigate the combined influence of built environment, demographics, and time period on cycling and walking rates. Repeated measures data is non-independent as it is taken from the same individual over time, and traditional statistical methods that assume independence can lead to biased estimates of the effects. Therefore, we decided to analyse the data by considering multilevel modelling techniques. Multilevel modelling is a statistical technique that can account for the hierarchical or nested structure of the data, and is therefore suitable for analysing repeated measures data. As the dataset in this research has a negative binomial distribution, we utilized generalized linear mixed models, which are a type of multilevel modelling that use a link function to handle non-normally distributed outcomes. This allowed us to examine the relationship between leisure walking dynamics over the study period, built environment, and sociodemographic characteristics. We treated the explanatory variables (built environment and socio-demographics) as fixed effects in all models since they remained constant during the study period. Moreover, all models included COVID stage as a key variable of interest to find out whether the effect of built environment and sociodemographic characteristics changed during different stages of COVID-19 restrictions. This approach also allowed us to investigate potential interactions between different variables. In order to interpret the coefficients produced by these models, we graphed the estimated marginal means from the model results and utilized a contrast test to determine whether there were statistically significant differences between groups. Estimated marginal means provide the unstandardized estimate values of a dependent variable, when all independent variables are controlled for. By employing contrast tests, we were able to examine and quantify the significance of variations in walking and cycling rates based on different factors such as the built environment, socio-demographic characteristics, and COVID-19 restrictions. This comprehensive analysis offers significant findings regarding the relationship between the built environment, socio-demographics, and the influence of COVID-19 restrictions on walking and cycling rates. Descriptive results In this section, we first analysed changes in leisure walking during the stages of the study. Then we grouped participants based on their built environment and sociodemographic characteristics and compared the mean walking rates across different groups. Summary of leisure walking responses Table 3 indicates overall leisure walking across in the stages of the study. There is a decrease in self-reported weekly recreational walking during the first and second lockdown stages. Based on a generalized linear mixed model in which we consider only the effect of the stage of study, theses changes in means are statistically significant (F (2,4022) = 9.421, P < 0.001). Similarly, the percentage of people who did not walk for leisure increased from 19.2% before COVID to 22.9% in the first lockdown and 30.4% in the second lockdown. Based on the Cochran’s Q test results for walked / did not walk, the changes in walking during different stages of lockdown is statistically significant (χ2(2) = 104.478, p < 0.01). Table 3 Leisure walking distribution during the pandemic Before COVID-19 First lockdown Second lockdown Didn’t do this 19.2% 22.9% 30.4% 1 per week 13.3% 13.9% 13.0% 2 per week 17.4% 13.5% 14.4% 3 per week 16.6% 14.8% 10.1% 4 per week 10.3% 10.3% 9.3% 5 per week 10.3% 11.3% 10.7% 5 + per week 12.9% 13.2% 12.1% Total walked 80.8% 77.1% 69.6% Mean 2.68 2.63 2.36 Std. Deviation 2.00 2.09 2.14 Note: Change in mean and percentage who walked were both statistically significant Our analysis of the walking dynamic during COVID-19 lockdown in Melbourne indicates the more pronounced declines during the second lockdown compared to the initial lockdown and the period before the pandemic. It is worth to mention that 5km restriction implemented during the second lockdown and this particular measure played a crucial role in intensifying the declines experienced during the second lockdown. In the next step, participants were categorized into four groups based on their walking behaviour. Note that these categories are used just to visualise the extent to which walking decreased over time, and are not used in the multilevel modelling. Individuals who reported no walking activity were classified under the "no walking" group; those who engaged in walking one or two times per week were placed in the "Low walking" group. Participants who walked 3 or 4 times per week were assigned to the "Medium walking" group, while individuals who walked 5 or more times per week were categorized as "High walking." The distribution of these different walking behaviours in the study are visually presented in Fig. 2 . Most of these categories were very stable cross lockdown stages. However, the percentage of individuals who did not engage in any walking activity during a given week exhibited an upward trend over the course of the study, whereas the proportion of ‘medium’ walkers experienced the most change. Leisure walking dynamics and the built environment In this section, we take a descriptive look at the effect of the built environment on patterns of recreational walking. Figure 3 indicates the changes in recreational walking by considering different built environment characteristics. People who lived in neighbourhoods with higher green space percentage were more likely to walk before COVID and there is less change during different stages of lockdown, which means perhaps people were protected against the impact of COVID-19 on physical activity (Fig. 3 -a). There is a general belief among urban planners that land-use diversity promote walking (Seong et al., 2021 ), but in this study we have opposite observation (Fig. 3 -b). It is probably because land-use diversity did have as many benefits during lockdowns when non-essential services were closed. Moreover, Wood et al. ( 2010 ) found that ‘sense of community’ encouraged leisure walking, which was associated with lower level of land-use mix. High residential density (Fig. 3 -c) and residential area (Fig. 3 -d) lead to more walking. Perhaps seeing other people walking in the neighbourhood provided a sense of safety for people who wished to go for a walk. Connectivity is associated with walking for transport (Sugiyama et al., 2012 ) because it makes more destinations accessible for people. On the other hand, when it comes to recreational walking the quality and proximity of recreational destinations are more important. In this study we cannot find a clear pattern for connectivity (Fig. 3 -e). Higher quartiles of cycling infrastructure were associated with higher walking and less change between stages of the study (Fig. 3 -f). Some bike infrastructure is made up of paths shared between cyclists and people who are walking, so it is possible that neighbourhoods which have higher cycling infrastructure have more walkable paths that could motivate people to explore their surroundings. Multivariate modelling results In this section, our aim was to analyse how the built environment influenced leisure walking behaviour throughout different stages of the pandemic. To investigate this, we utilized generalized linear mixed models to examine the relationship between leisure walking patterns, the built environment, and the control variables (sociodemographic characteristics). Our specific focus was on exploring whether the influence of the built environment varied across the three study periods: pre-COVID, first lockdown, and second lockdown. We hypothesised that the impact of the built environment would be particularly significant during the lockdown stages, especially in the second lockdown when participants were subjected to a 5km travel restriction. Fixed effects models In this section, we report the initial fixed effect f-tests of different models to examine the effect of built environment and socio-demographic characteristics on leisure walking across the three COVID stages (before COVID, first lockdown and second lockdown). All three models included COVID stage as a key variable of interest (Table 4 ), and in all models, COVID stage was statistically significant even when controlling for all other variables. We consider first two models (Model 1 and 2) as the ‘base models’ without interaction effects. In Model 1 we evaluated the effects of COVID stage and built environment factors; we found that in addition to COVID stage, residential density, green space percentage, and land-use diversity have significant effects on leisure walking. In Model 2 we investigated whether COVID stage and built environment variables were still significant when you include demographic control variables. Indeed, in Model 2 income, age, employment, and education had a significant effect on leisure walking, yet these variables did not reduce the impact of COVID stage or built environment variables. As one of the aims of this study was to examine whether the effect of the built environment changed depending on the stage of COVID-19 lockdowns. The descriptive results shown in Fig. 3 suggest that some interactions between built environment and COVID stage may exist. For this reason in Model 3 we included the interaction between built environment factors and stage of COVID. However, contrary to the descriptive results in Fig. 3 , none of the interactions between COVID stage and built environment were statistically significant. This means the effect of the built environment does not change because of COVID-19 restrictions. In other words, locations with the most favourable built environment for walking saw the same degree of reduction in leisure walking as locations with poor built environment. Table 4 Fixed effects F-tests between built environment characteristics, socio-demographics, and leisure walking (* indicates the variable is statistically significant) Model 1 Base model: built environment Model 2 Base model: BE + demographics Model 3 BE and time interaction COVID stage 9.884* 10.293* 10.317* Green space percentage 5.930* 4.881* 4.903* Land-use diversity 2.902* 3.326* 3.234* Residential density 4.600* 3.742* 3.755* Residential area percentage 1.610 1.967 1.947 Connectivity 0.468 0.292 0.283 Cycling infrastructure density 0.505 0.521 0.503 Age - 13.888* 13.703* Gender - 2.340 2.396 Employment - 0.180 0.192 Income - 6.827* 6.733* Education - 9.121* 8.973* Household type - 2.506 2.442 Employment changes - 2.861* 2.811* Residential area percentage * COVID stage - - 0.295 Residential density * COVID stage - - 0.215 Green space percentage * COVID stage - - 0.045 Land-use diversity * COVID stage - - 0.124 Cycling infrastructure density * COVID stage - - 0.149 Connectivity * COVID stage - - 0.121 Model fit statistics AIC 16614.645 16564.341 16629.070 BIC 16752.987 16777.937 17067.539 Log likelihood 16570.392 16495.745 16486.555 Note: * indicates the variable is statistically significant Across the four models, Model 2 is the best fit based on AIC (16564.314) but slightly poorer than Model 1 using BIC (16777.937 vs 16752.982); this is not surprising as BIC penalises additional parameters more strongly than AIC. For this reason, we use Model 2 for the next section of the results. Table 4 only presents the f-statistic, which is an indication of statistical significance but does not indicate the effect size of each variable. In all models, the effect of lockdown stage was significant and green space percentage, land use diversity, and residential density are leading built environment factors that have impact on leisure walking during COVID-19 restrictions. In the next stage of analysis, we interpret the estimated means for Model 2, which illustrate the average effect of a given independent variable when all other variables are controlled for. We also show the results of planned contrast tests to determine which levels of the independent variables are significantly different to each other. Estimated marginal means Figure 4 represents the estimated marginal means and planned contrast tests for the significant effects from Model 2: COVID stage, residential density, green space percentage, land use diversity, income, age, employment change and education. In all models, the effect of lockdown stage was significant. Planned contrasts showed that this decline was only significant for lockdown 2 relative to the other two time periods (Fig. 4 -a). This is likely because the 5km travel restrictions were only imposed in this lockdown period. The estimated marginal means for built environment characteristics confirm that the relationship between walking and built environment is not necessarily linear. Residential density had a significant effect on leisure walking, but Fig. 4 -b suggests that density only increases walking to a point, before it decreases at the highest density. This could be because in very highly populated areas, even walking outdoors could be perceived to impose a risk of contagion. Green space is another significant built environment factor that contrast analysis indicate significant difference between the lowest and highest quartile of green space percentage (Fig. 4 -c). This effect is also non-linear, with a clear plateau between the third and fourth quartile. This result confirms the history of research that shows a link between green space and leisure walking (Dadvand et al., 2016 , Sarkar et al., 2015 ). Land-use diversity is another significant factor in the model and based on the contrast analysis, difference between the lowest and highest quartile of land-use diversity is statistically significant (Fig. 4 -d). However, the direction of this effect was counter-intuitive, with the lowest quartile of diversity showing the highest leisure walking rates. As most business were closed during lockdown, perhaps land-use diversity did not have a practical meaning for people. The remaining graphs show the relationship between the control demographics and leisure walking. Income plays a significant role in predicting leisure walking behaviour, the higher income level, the higher walking rate (Fig. 4 -e). Even though walking is a cost-free form of exercise, and one might think it would be more appealing to low-income people, they had the lowest rates of walking. Age is one of the most significant demographic characteristics studied, where the level of leisure walking is increased as people gets older in this study (Fig. 4 -f). Moreover, based on the contrast analysis, differences between the adults’ and older adults’ walking level and young people are statistically significant. Employment change is another significant factor on recreational walking. People who lost a job during the pandemic had the highest walking rate compared to people with no change (Fig. 4 -g). The variance among people who found a job was very high, reflecting the small number of respondents in this category (N = 30). And finally, people with a higher education level are more likely to walk for recreational purposes (Fig. 4 -h). It is worth noting that although people with a higher degree are likely to also have a higher income, since both variables are included in the model, education has an independent effect. Discussion In the early years of the COVID-19 pandemic, many people faced significant restrictions on where and why they could travel. In Melbourne, for example, non-essential businesses closed, people were told to work from home, and they had to remain within 5km of their home. One might hypothesise that under these extreme circumstances, the design of one’s neighbourhood would have a particularly significant impact on how much people walked for leisure, as they were unable to go to a gym or visit distant parks. In this study, self-reported leisure walking reduced significantly between the first and second lockdown stages (when the 5km travel restriction was implemented). In particular, people who were already medium- or low-frequency walkers were somewhat likely to drop into low- or no-walking between the first and second lockdown stages (see Fig. 2 ). Descriptive results suggested that some built environment factors may have attenuated some of these reductions; for example, locations with more green space saw a lower reduction in leisure walking. Yet multivariate modelling found that our hypothesis was not supported: the effect of built environment on leisure walking was the same regardless of the COVID stage. Among the built environment factors that we studied in this model, green space, residential density and land-use diversity all influenced leisure walking regardless of pandemic stage. Places with more green space had higher rates of walking, a finding echoed by much past research (Dadvand et al., 2016 , Sarkar et al., 2015 ). The third-highest quartile of residential density had the highest walking, declining in the highest density quartile, perhaps due to infection concerns in the most populated areas. Land-use diversity had a counter-intuitive relationship, with lower walking in the most diverse locations. This could be due to the closure of non-essential businesses during the pandemic, or because more diverse locations had less greenspace. Demographic effects were largely consistent with past research, and the effects did not vary between pandemic waves. This study has a number of limitations. The first is that leisure walking was reported based on respondents’ recollection of how much they walked before COVID, so we are relying on their memory. Moreover, we did not ask them their attitude toward leisure walking and physical activity. The absence of this information may limit the comprehensive understanding of factors influencing leisure walking behaviour during pandemic. Furthermore, because we did not know the exact location of participants’ homes we could not use the smallest spatial resolution (SA1) to calculate the land use variables. Instead, we had to rely on the larger SA2 statistical area, which may not accurately reflect the land use immediately around a person’s home. Finally, we did not have a direct measure of ‘walkability’ based on the local design of footpaths and availability of pedestrian crossings. Yet taken as a whole, this study suggests that local built environment characteristics are an important determinant of walking for leisure whether or not people are restricted in their travel. Walking is a free and easily-accessed means of physical activity which can contribute to improved community health and wellbeing. The link between physical activity and both physical and mental health long been known (Warburton and Bredin, 2016 , Biddle et al., 2019 ). Supportive infrastructure plays an important role for people with inadequate access to private recreational facilities or limited mobility such as low-income people and youth. These vulnerable groups are already at greater risk of physical inactivity; therefore, it is important to consider them in neighbourhoods planning. This need was recognized by the United Nations who have codified the aim of providing “universal access to safe, inclusive and accessible, green and public spaces, particularly for women and children, older persons and persons with disabilities” in the UN Sustainable Development Goals (target 11.7). Declarations Author Contribution MN, AD and LK worked together on research conceptualisation and methodology. MN conducted data analysis, visualisation and wrote the initial paper draft. All authors reviewed and edited the manuscript. References Barbieri, D. M., Lou, B., Passavanti, M., Hui, C., Lessa, D. A., Maharaj, B., Banerjee, A., Wang, F., Chang, K. & Naik, B. 2020. A survey dataset to evaluate the changes in mobility and transportation due to COVID-19 travel restrictions in Australia, Brazil, China, Ghana, India, Iran, Italy, Norway, South Africa, United States. Data in brief, 33 , 106459. Bentley, R., Blakely, T., Kavanagh, A., Aitken, Z., King, T., Mcelwee, P., Giles-Corti, B. & Turrell, G. 2018. A longitudinal study examining changes in street connectivity, land use, and density of dwellings and walking for transport in Brisbane, Australia. Environmental health perspectives, 126 , 057003. Biddle, S. J., Ciaccioni, S., Thomas, G. & Vergeer, I. 2019. Physical activity and mental health in children and adolescents: An updated review of reviews and an analysis of causality. Psychology of Sport and Exercise, 42 , 146-155. Buehler, R. & Dill, J. 2016. Bikeway networks: A review of effects on cycling. Transport Reviews, 36 , 9-27. Cervero, R. & Kockelman, K. 1997. Travel demand and the 3Ds: Density, diversity, and design. Transportation research part D: Transport and environment, 2 , 199-219. Currie, G., Jain, T., Aston, L. & Reynolds, J. 2021. Long Term Impacts of COVID-19 on Melbourne Research. Cusack, M. 2021. Individual, social, and environmental factors associated with active transportation commuting during the COVID-19 pandemic. Journal of transport & health, 22 , 101089. Dadvand, P., Bartoll, X., Basagaña, X., Dalmau-Bueno, A., Martinez, D., Ambros, A., Cirach, M., Triguero-Mas, M., Gascon, M. & Borrell, C. 2016. Green spaces and general health: roles of mental health status, social support, and physical activity. Environment international, 91 , 161-167. De Vos, J. 2020. The effect of COVID-19 and subsequent social distancing on travel behavior. Transportation research interdisciplinary perspectives., 5 , 100121. Ewing, R. & Cervero, R. 2010. Travel and the built environment: A meta-analysis. Journal of the American planning association, 76 , 265-294. Forsyth, A., Oakes, J. M., Schmitz, K. H. & Hearst, M. 2007. Does residential density increase walking and other physical activity? Urban Studies, 44 , 679-697. Hook, H., De Vos, J., Van Acker, V. & Witlox, F. 2021. Does undirected travel compensate for reduced directed travel during lockdown? Transportation Letters, 13 , 414-420. Hook, H., De Vos, J., Van Acker, V. & Witlox, F. 2023. Evolutions in undirected travel (satisfaction) during the COVID-19 pandemic. Transportation research part F: traffic psychology and behaviour, 94 , 99-113. Hunter, R. F., Garcia, L., De Sa, T. H., Zapata-Diomedi, B., Millett, C., Woodcock, J., Pentland, A. S. & Moro, E. 2021. Effect of COVID-19 response policies on walking behavior in US cities. Nature communications, 12 , 3652. Kamruzzaman, M. & Hine, J. 2013. Self-proxy agreement and weekly school travel behaviour in a sectarian divided society. Journal of Transport Geography, 29 , 74-85. Kamruzzaman, M., Washington, S., Baker, D., Brown, W., Giles-Corti, B. & Turrell, G. 2016. Built environment impacts on walking for transport in Brisbane, Australia. Transportation, 43 , 53-77. Kerr, J., Rosenberg, D. & Frank, L. 2012. The role of the built environment in healthy aging: Community design, physical activity, and health among older adults. Journal of planning literature, 27 , 43-60. Kruger, J., Ham, S. A., Berrigan, D. & Ballard-Barbash, R. 2008. Prevalence of transportation and leisure walking among U.S. adults. Preventive Medicine, 47 , 329-334. Lättman, K., Olsson, L. E., Waygood, E. O. D. & Friman, M. 2023. Nowhere to go–Effects on elderly's travel during Covid-19. Travel Behaviour and Society, 32 , 100574. Liu, J., Xiao, L. & Zhou, J. 2021. Built environment correlates of walking for transportation. Journal of Transport and Land Use, 14 , 1129-1148. Lotfata, A., Gemci, A. G. & Ferah, B. 2022. The changing context of walking behavior: coping with the COVID-19 Pandemic in urban neighborhoods. Archnet-IJAR: International Journal of Architectural Research, 16 , 495-516. Marshall, W. E. & Garrick, N. W. 2010. Effect of street network design on walking and biking. Transportation Research Record, 2198 , 103-115. Mccormack, G. R., Petersen, J., Naish, C., Ghoneim, D. & Doyle-Baker, P. K. 2022. Neighbourhood environment facilitators and barriers to outdoor activity during the first wave of the COVID-19 pandemic in Canada: A qualitative study. Cities & Health , 1-13. Mcelroy, S., Fitch, D. T. & Circella, G. 2022. Changes in Active Travel During the COVID-19 Pandemic. Pandemic in the Metropolis: Transportation Impacts and Recovery. Springer. Paul, P., Carlson, S. A., Carroll, D. D., Berrigan, D. & Fulton, J. E. 2015. Walking for transportation and leisure among US adults—National Health Interview Survey 2010. Journal of Physical Activity and Health, 12 , S62-S69. Pollard, T. M. & Wagnild, J. M. 2017. Gender differences in walking (for leisure, transport and in total) across adult life: a systematic review. BMC Public Health, 17 , 341. Power, D., Lambe, B. & Murphy, N. 2023. Trends in recreational walking trail usage in Ireland during the COVID-19 pandemic: Implications for practice. Journal of Outdoor Recreation and Tourism, 41 , 100477. Runacres, A., Mackintosh, K. A., Knight, R. L., Sheeran, L., Thatcher, R., Shelley, J. & Mcnarry, M. A. 2021. Impact of the COVID-19 pandemic on sedentary time and behaviour in children and adults: a systematic review and meta-analysis. International journal of environmental research and public health, 18 , 11286. Saelens, B. E. & Handy, S. L. 2008. Built environment correlates of walking: a review. Medicine and science in sports and exercise, 40 , S550. Sarkar, C., Webster, C., Pryor, M., Tang, D., Melbourne, S., Zhang, X. & Jianzheng, L. 2015. Exploring associations between urban green, street design and walking: Results from the Greater London boroughs. Landscape and Urban Planning, 143 , 112-125. Seong, E. Y., Lee, N. H. & Choi, C. G. 2021. Relationship between land use mix and walking choice in high-density cities: A review of walking in Seoul, South Korea. Sustainability, 13 , 810. Shaer, A., Rezaei, M., Rahimi, B. M. & Shaer, F. 2021. Examining the associations between perceived built environment and active travel, before and after the COVID-19 outbreak in Shiraz city, Iran. Cities, 115 , 103255. Sugiyama, T., Giles-Corti, B., Summers, J., Du Toit, L., Leslie, E. & Owen, N. 2013. Initiating and maintaining recreational walking: a longitudinal study on the influence of neighborhood green space. Preventive medicine, 57 , 178-182. Sugiyama, T., Neuhaus, M., Cole, R., Giles-Corti, B. & Owen, N. 2012. Destination and route attributes associated with adults' walking: a review. Medicine and science in sports and exercise, 44 , 1275-1286. Tao, T., Wu, X., Cao, J., Fan, Y., Das, K. & Ramaswami, A. 2020. Exploring the nonlinear relationship between the built environment and active travel in the twin cities. Journal of Planning Education and Research , 0739456X20915765. Warburton, D. E. & Bredin, S. S. 2016. Reflections on physical activity and health: what should we recommend? Canadian Journal of Cardiology, 32 , 495-504. Watts, P., Phillips, G., Petticrew, M., Hayes, R., Bottomley, C., Yu, G., Schmidt, E., Tobi, P., Moore, D. & Frostick, C. 2013. Physical activity in deprived communities in London: examining individual and neighbourhood-level factors. PloS one, 8 , e69472. Wood, L., Frank, L. D. & Giles-Corti, B. 2010. Sense of community and its relationship with walking and neighborhood design. Social science & medicine, 70 , 1381-1390. Yin, C., Cao, J., Sun, B. & Liu, J. 2023. Exploring built environment correlates of walking for different purposes: Evidence for substitution. Journal of Transport Geography, 106 , 103505. Additional Declarations No competing interests reported. Supplementary Files Appendix.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-3977307","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":275829386,"identity":"ac47fd48-3094-4170-bf21-ff924f12c4c9","order_by":0,"name":"Mahsa Naseri","email":"","orcid":"","institution":"Monash University","correspondingAuthor":false,"prefix":"","firstName":"Mahsa","middleName":"","lastName":"Naseri","suffix":""},{"id":275829387,"identity":"c0b6a59b-02c5-482e-8eda-77e13165af4c","order_by":1,"name":"Alexa Delbosc","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABBUlEQVRIiWNgGAWjYDACdjDJzMAnAaIrwDw2sAhOwAwl2cBazpCshbGNCC38zczHJH4wWMuxSTcf/vBz3jY5c/bjzx4wVFgnNuDQInGYLdmwhyHdmE3mWJpk77bbxpY9OeYGDGfScWphOMxj+ICH4XBim0SOGTPjttuJGw7ksEkwth3GqUX+MP+Hg38YDtcDtRh/ZpwD1HL++TMJxn+4tRgc5mF8DLQlgU0ix0CasQGo5UaCmQRjA24thofZjI1lDNIN20B+6Tl229jgxhsziYRj6ca4tMgdb34m+abCWp4fFGI/am7LGZxPfybxocZaFqf3Ic5DF0jAq3wUjIJRMApGASEAACeoVNSrwrn0AAAAAElFTkSuQmCC","orcid":"","institution":"Monash University","correspondingAuthor":true,"prefix":"","firstName":"Alexa","middleName":"","lastName":"Delbosc","suffix":""},{"id":275829388,"identity":"bd41a544-43ec-4b07-b070-bcd5047081f5","order_by":2,"name":"Liton Kamruzzaman","email":"","orcid":"","institution":"Monash University","correspondingAuthor":false,"prefix":"","firstName":"Liton","middleName":"","lastName":"Kamruzzaman","suffix":""}],"badges":[],"createdAt":"2024-02-22 02:22:09","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3977307/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3977307/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":52037343,"identity":"7bceec17-99f4-4d75-9165-46cc712d6ba5","added_by":"auto","created_at":"2024-03-05 17:21:23","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":348739,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial distribution of built environment characteristic\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-3977307/v1/34cf859397111ecc97fb72d7.png"},{"id":52037345,"identity":"ddf92912-aa33-4b36-aafc-18dbd1abb0d0","added_by":"auto","created_at":"2024-03-05 17:21:23","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":199120,"visible":true,"origin":"","legend":"\u003cp\u003eDynamics of leisure walking\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-3977307/v1/5b74fba87d2c965aa0c31fde.png"},{"id":52037341,"identity":"bb97366b-d8fe-43d5-a213-af34dd77b512","added_by":"auto","created_at":"2024-03-05 17:21:23","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":73510,"visible":true,"origin":"","legend":"\u003cp\u003edynamics of leisure walking by considering built environment characteristic\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-3977307/v1/708c035b90c799ca003cac19.png"},{"id":52038096,"identity":"a72e5f07-7233-4358-9a39-ce8c1f855348","added_by":"auto","created_at":"2024-03-05 17:29:23","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":76112,"visible":true,"origin":"","legend":"\u003cp\u003eEstimated marginal means for significant fixed effects in Model 2\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-3977307/v1/2edd58e90eedf10517d329e5.png"},{"id":61368097,"identity":"7835c0c2-7b47-45d1-8d11-6791b3bc960a","added_by":"auto","created_at":"2024-07-30 02:14:17","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1340358,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3977307/v1/9e9f0c12-3dba-41cc-9705-088fbdf208e0.pdf"},{"id":52037344,"identity":"a80e8b36-ba1a-436b-b615-bfdb5bbb87a9","added_by":"auto","created_at":"2024-03-05 17:21:23","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":49581,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-3977307/v1/d3245e1e266133f3441db024.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Did the built environment attenuate reductions in leisure walking during COVID-19? A quasi-panel study","fulltext":[{"header":"Introduction","content":"\u003cp\u003eCOVID-19-related restrictions led to a decline in physical activity levels for many people around the world (Runacres et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). In the first two years of the pandemic, most people experienced periods when fitness facilities and recreational centres were temporarily closed. Consequently, engaging in walking activities emerged as a readily accessible and effective means for many individuals to sustain their physical well-being and support their mental health. Walking can also facilitate social interaction while maintaining social distances, especially during the periods of pandemic when people were forced to stay in their neighbourhood and the risk of social isolation and loneliness increased (L\u0026auml;ttman et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Taken together, walking likely played an important role in supporting mental and physical health during the early pandemic years. Yet much of the transport research coming out of the pandemic years focussed on public transport ridership, commuting behaviour and working from home, with fewer studies looking at non-commute and active travel (Hook et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2023\u003c/span\u003e, Cusack, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2021\u003c/span\u003e, McElroy et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eDecades of research before the pandemic highlights the importance of the built environment on walking (Saelens and Handy, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2008\u003c/span\u003e, Ewing and Cervero, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). One might hypothesise that the built environment would have a particularly salient impact on leisure walking during the pandemic, when many people were restricted to their neighbourhoods, forcing them to rely on their local built environment to facilitate (or discourage) walking. Yet to date, very few studies directly examined the relationship between leisure walking and the built environment during the pandemic (Lotfata et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThis paper aims to estimate the association between built environment characteristics and leisure walking behaviour during the first year of the pandemic. In particular, we hypothesised that one\u0026rsquo;s neighbourhood built environment can attenuate reductions in leisure walking during COVID-19 lockdowns. The COVID-19 lockdowns implemented in Melbourne provided a unique opportunity to assess this relationship, as Melbourne residents experienced some of the longest and most restrictive lock-down regulations in the world.\u003c/p\u003e"},{"header":"Literature review","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eCOVID-19 and walking\u003c/h2\u003e \u003cp\u003eUrban mobility since 2020 was considerably impacted by COVID-19 when different actions were taken globally in order to reduce the spread of the virus, such as social distancing, movement restrictions, working from home and travel bans. Urban dynamics and mobility changed and this change was significant in the modal distribution of trips (Barbieri et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Trends in walking were impacted by changes in work and exercise opportunities. As work from home orders were introduced, and people did not have access to out-of-home entertainment and exercise facilities, the distribution of active travel was affected. In response, people maintained their physical and mental health through recreational active travel such as walking (De Vos, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe change in recreational walking during COVID-19 was not clear, with some studies finding increases and others showing decreases. Hunter et al. (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) utilized data from mobile devices and area-level data from 1.62\u0026nbsp;million mobile phone users to examine how much people walked in 10 metropolitan areas across the United States. The time frame of the analysis spanned from mid-February 2020, which was before the implementation of lockdown measures, to late June 2020, when some of these restrictions were relaxed. Their results indicated that there was a 70% reduction in the number of walks and a 50% decrease in the average distance walked in all metropolitan areas during the pandemic. Although there was a steady increase in walking from mid-April 2020, after some commercial and business activities resumed, walking remained approximately 18% below pre-pandemic levels. A study by McElroy et al. (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) found a similar trend. Using a self-reported survey from California as well as a national sample, they found that walking as a commute mode decreased during the pandemic. Both studies also found that unlike walking for commuting, recreational walking \u003cem\u003eincreased\u003c/em\u003e in comparison with pre-COVID period (McElroy et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2022\u003c/span\u003e, Hunter et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eSimilarly, Power et al. (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) found recreational walking and utilization of trails, particularly those in proximity to urban areas, in Ireland increased. Their research demonstrated that trail usage (based on counter data) in 2020 has shown a general rise of 6% as compared to the previous year. Additionally, during the most severe COVID-19 lockdown periods in 2020, trails located within a 2 km radius of urban areas recorded an increase in average footfall counts of up to 102% compared to those situated outside of the 2 km range.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eBuilt environment and walking\u003c/h2\u003e \u003cp\u003eA long history of research explores how built environment characteristics affect travel behaviour (Ewing and Cervero, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). Walking is associated with having access to destinations, connected street network, and higher residential density. Mixed land-use is found to increase walking rates as they provide access to a greater variety of destinations within walking distance (Saelens and Handy, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). A higher residential density creates critical mass for more people to walk, to see others walking and thereby to promote a safer environment for walking (Kamruzzaman et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Indirectly, higher density supports diverse land use patterns and ultimate creates a favourable condition for walking (Saelens and Handy, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2008\u003c/span\u003e); it could be. Street network pattern usually reflects connectivity, which has a positive effect on walking (Sarkar et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2015\u003c/span\u003e, Marshall and Garrick, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2010\u003c/span\u003e) by providing convenient access to destinations, direct routes, increased safety options, and a more enjoyable walking environment. Green spaces have a positive effect on physical activity and walking (Sugiyama et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2013\u003c/span\u003e) because they provide individuals with more access to nature, which serves as an inviting and stimulating environment that encourages people to engage in outdoor activities. People who have more access to green space, or even street trees, tend to walk and be more physically active than those with poor access (Dadvand et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, Sarkar et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Therefore, the built environment characteristics can either facilitate or discourage walking in a neighbourhood.\u003c/p\u003e \u003cp\u003eHowever, it should be noted that the effects of the built environment on recreational walking are less clear, compared to walking for transport (Saelens and Handy, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). In addition, recent studies have found that the built environment was a stronger predictor of walking for transport, whereas socio-demographics have a greater impact on walking for leisure (Yin et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2023\u003c/span\u003e, Liu et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). The role of socio-demographics will be explored in greater depth in the next section.\u003c/p\u003e \u003cp\u003eAs the COVID-19 pandemic imposed restrictions on movement and encouraged individuals to stay close to their homes and neighbourhoods. With limited opportunities for travel and access to external recreational spaces, people relied heavily on their local built environment for physical activity. One might hypothesise that because many people were restricted to activity in their local neighbourhood during the pandemic, the built environment might have had an even greater impact on walking during the early pandemic years. Yet despite a long-standing history of research on the relationship between the built environment and active travel (Ewing and Cervero, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2010\u003c/span\u003e), at the time of writing only three studies explored this topic in the context of the COVID-19 pandemic.\u003c/p\u003e \u003cp\u003eMcCormack et al. (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) conducted a qualitative study that explored how the built environment in Canada supported outdoor physical activity during the early stages of the COVID-19 pandemic. The study found that parks and pathways were important for maintaining outdoor activity, but crowding and lack of physical distancing were concerns. One potential limitation of this study is the small sample size, as only 12 adults were interviewed. This may not be representative of the larger population or capture the experiences of diverse groups. Additionally, the study did not consider an objective measure of built environment for the analysis.\u003c/p\u003e \u003cp\u003eHook et al. (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) examined the impact of the COVID-19 lockdown on travel behaviour in Flanders, Belgium. Their findings revealed that the built environment did not have a significant impact on travel behaviour changes. This study used an \"urbanization factor\" to measure the built environment, but did not analyse specific characteristics of the built environment within that measure.\u003c/p\u003e \u003cp\u003eFinally, Shaer et al. (\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) examined the impact of COVID-19 and related policies on active travel of Central Business District (CBD) residents compared to non-CBD residents in Shiraz, Iran. The study investigated the relationship between perceived built environment factors and active travel before and after the outbreak. The study found that the quality of built environment factors in the CBD led to longer average walking times among CBD residents compared to non-CBD residents. They used subjective measures to assess the built environment, which means that individuals may have different perceptions or interpretations of the environment.\u003c/p\u003e \u003cp\u003eNone of these studies used objective measures for specific built environment characteristics to isolate its effect. There is a lack of knowledge regarding how people in various neighbourhoods with different spatial characteristics respond to the COVID-19 pandemic.\u003c/p\u003e \u003c/div\u003e"},{"header":"Data and methods","content":"\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eStudy area\u003c/h2\u003e \u003cp\u003eThe study area of this research was Melbourne, the capital and second most-populous city of the Australia. Melbourne experienced over 260 days in lockdown (i.e. stay-at-home orders) during 2020\u0026ndash;2021. In this study we focused on Melbourne\u0026rsquo;s second and longest lockdown, which lasted for 111 days started from early July 2020 until late October 2020. During this period, work from home orders were introduced, major events were cancelled and many schools and non-essential services were closed. In addition, unlike the previous lockdown, in the second lockdown a 5-kilometre travel restriction was introduced which forced people to stay in their local neighbourhood and use their local (outdoor) facilities for recreation purposes and outdoor activities should be limited to 2 hours per day. The unique characteristics of Melbourne\u0026rsquo;s lockdown, particularly the 5-kilometre travel restriction imposed in lockdown 2, made the city a salient case study for travel behaviour researchers.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eData sources\u003c/h2\u003e \u003cp\u003eTwo types of data were used in this study. First, we drew upon a travel behaviour survey of Melbourne residents to quantify their self-reported walking for leisure. Second, we used open spatial data sets to quantify the built environment characteristics of each survey respondent\u0026rsquo;s residential neighbourhood.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eSurvey data\u003c/h2\u003e \u003cp\u003eWe were provided access to the \u0026lsquo;C-19 Long Term Transport Impact Study\u0026rsquo; in Melbourne (Currie et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) which is a quasi-longitudinal survey including 2,163 people who lived in Melbourne in mid-2020. The survey involved an online questionnaire which used a market research panel company to recruit a representative sample of Melbourne residents. The sampling frame employed quotas based on income, age, and location. The sampling frame was designed to recruit from three different locations: inner, middle, and outer Melbourne. The researchers identified the share of the population in each of the age and income cohorts for each region of Melbourne using the census. The sampling frame was created to ensure that the survey sample matched the census in terms of these characteristics. More information on the survey and sampling methodology can be found in (Currie et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThis data set was collected in two different legs; the first one occurred during Melbourne\u0026rsquo;s first lockdown and the second one captured people\u0026rsquo;s travel behaviour adaptation across both the first and second lockdown. As we were interested on behaviour changes during different stages of lockdown, we used the second wave of the data set (sample size 1,341 responses), which ran from 16 July to 8 August 2020, within the first month of Melbourne\u0026rsquo;s second lockdown period.\u003c/p\u003e \u003cp\u003eThe study as a whole focused on wide range of information with a focus on public transport use and working from home. For the purpose of this paper, we only studied self-reported walking for leisure (not commuting). The sociodemographic characteristics of this survey are provided in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. We included these demographics as control variables in our models, given the significant association between demographics and leisure walking (Yin et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2023\u003c/span\u003e, Liu et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSociodemographic characteristics of survey sample\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eSocio-demographic characteristic\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSurvey Sample Percentage\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMelbourne average (ABS 2021)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eGender\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e37.9%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e49.2%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e62.0%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e50.8%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eIncome\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLow (0-\u003cspan\u003e$\u003c/span\u003e530 per week)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e34.0%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e32.1%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMedium (\u003cspan\u003e$\u003c/span\u003e530-\u003cspan\u003e$\u003c/span\u003e1870 per week)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e48.7%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e43.4%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHigh (\u0026gt; \u003cspan\u003e$\u003c/span\u003e1870 per week)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e17.3%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e24.5%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYouth (15\u0026ndash;24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e11.8%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e12.3%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAdult (25\u0026ndash;64)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e72.0%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e54.5%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOlder adult (\u0026gt;\u0026thinsp;65)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e16.1%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e15.1%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eEmployment\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEmployed\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e55.4%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e88.5%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNon-worker\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e36.2%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e6.2%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNot in labour force\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e8.4%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e5.3%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eEducation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNon-university degree\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e40.5%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e67.2%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHigher education\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e59.5%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e32.8%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eHouse hold type\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSingle\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e20.0%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e24.8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCouple with children (Family)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e23.5%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e37.93%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOther\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e56.5%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e36.37%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eHousing type\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFlat, unit or apartment\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e39.6%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e15.6%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFree-standing/ detached\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e48.4%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e67.8%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTown house/ semi detached\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e12.0%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e16.2%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eParticipants report the frequency of their leisure walking per week in three stages: pre-COVID, during the first lockdown and during second lockdown (the present, at the time of the survey). Responses were recorded on a seven-point scale:\u003c/p\u003e \u003cp\u003e0. Didn\u0026rsquo;t do this\u003c/p\u003e \u003cp\u003e1. 1 time a week\u003c/p\u003e \u003cp\u003e2. 2 times a week\u003c/p\u003e \u003cp\u003e3. 3 times a week\u003c/p\u003e \u003cp\u003e4. 4 times a week\u003c/p\u003e \u003cp\u003e5. 5 times a week\u003c/p\u003e \u003cp\u003e6. More than 5 times a week\u003c/p\u003e \u003cp\u003eNote that because these walking behaviours were measured retrospectively, changes between \u0026lsquo;pre-COVID\u0026rsquo; and the second wave of COVID rely on respondent recall.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eQuantifying the built environment\u003c/h2\u003e \u003cp\u003eBecause the C-19 survey was not originally intended for spatial analysis, respondents only provided their suburb rather than their street address. Therefore, we had to map respondent\u0026rsquo;s suburb on the most relevant spatial resolution used in the Australian census, Statistical Area 2 (SA2). These SA2 units contain on average 10,000 residents, their median size is approximately 9 square kilometres in urban areas and they are meant to represent a community that interacts spatially or socially.\u003c/p\u003e \u003cp\u003eBased on the literature, we considered built environment indicators including residential density (Forsyth et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2007\u003c/span\u003e), green space area (Watts et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2013\u003c/span\u003e), and land-use mix (Cervero and Kockelman, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e1997\u003c/span\u003e), connectivity (Bentley et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), and cycling infrastructure coverage (Buehler and Dill, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) which have been found to influence active travel behaviour. It is worth mentioning that cycling infrastructure density is not normally associated with leisure walking, we decided to kept this in the model as some bike infrastructure is shared with walkers. Moreover, by including cycling infrastructure, the model recognizes the interconnections between different modes of active transport and the importance of holistic planning and design. Considering cycling infrastructure alongside walking infrastructure helps capture the broader context of active travel behaviour and provides insights into how the built environment impacts various modes of active transport.\u003c/p\u003e \u003cp\u003eIn order to calculate residential density, we extracted dwelling numbers in each neighbourhood from the 2016 census data (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.abs.gov.au/\u003c/span\u003e\u003cspan address=\"https://www.abs.gov.au/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). Residentional area percentage was calculated based on the area of the residential blocks in each neighbourhood. Similarly, green space percentage was determined based on the parklnads area in each neighbourhood.\u003c/p\u003e \u003cp\u003eLand-use diversity measure was calculated by simpson\u0026rsquo;s diversity index (formula 1) in which a is the total area of specific land-use in neighbourhood and A is the total area of all land-use categories). This index value ranges from 0 to 1 and the higher value, the more diverse land-use pattern (Kamruzzaman and Hine, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2013\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eLand-use diversity\u0026thinsp;=\u0026thinsp;1- \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\sum \\left(\\frac{a}{A}\\right)\\)\u003c/span\u003e\u003c/span\u003e\u003csup\u003e2\u003c/sup\u003e (1)\u003c/p\u003e \u003cp\u003eFor cycling infrastructure measure, Vicroads and AURIN open data were downloaded and GIS methods were used to measure the length of bike paths in each neighbourhood. To remove the effect of neighbourhood area and have a normilized comparision in the next steps, we divided total bike paths length (km) by neighbourhood area (km\u003csup\u003e2\u003c/sup\u003e). It is worth mentioning that cycling infrastructure density is not normally associated with leisure walking, as only some bike infrastructure is shared with walkers.\u003c/p\u003e \u003cp\u003eFor the connectivity index, we calculated the number of intersections with 3 legs or more and, applying the same logic as cycling infrastructure, we divided this value by neighbourhood area (km\u003csup\u003e2\u003c/sup\u003e).\u003c/p\u003e \u003cp\u003eBecause the impact of the built environment on travel behaviour is likely to be non-linear (see for example Yin et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2023\u003c/span\u003e, Tao et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), we divided these built environment characteristics into quartiles. The distribution of built environment variables is provided in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e and the spatial distribution of them is provided in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBuilt environment measurement method and their quartile measurement\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMeasure\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMeasurement method\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLowest quartile\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSecond quartile\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eThird\u003c/p\u003e \u003cp\u003equartile\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eHighest quartile\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eResidential Area percentage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eArea of residential blocks divided by neighbourhood area (Km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0-34.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e34.47\u0026ndash;58.89\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e58.89\u0026ndash;75.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e75.68\u0026ndash;97.66\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eResidential density\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNumber of dwellings divided by neighbourhood area (Km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0-581.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e581.79-918.825\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e918.825-2075.586\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e2075.586-6874.92\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGreen space percentage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eArea of parklands divided by neighbourhood area (Km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0-6.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e6.09\u0026ndash;11.38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e11.38\u0026ndash;18.38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e18.38\u0026ndash;66.31\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLand use diversity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSimpson diversity index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0-0.38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.38\u0026ndash;0.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.52\u0026ndash;0.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.68\u0026ndash;2.29\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCycling infrastructure\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eBicycle lane length (km) divided by neighbourhood area (Km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0-0.71\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.71\u0026ndash;1.66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1.66\u0026ndash;3.24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e3.24\u0026ndash;7.09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eConnectivity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNumber of intersections divided by neighbourhood area (Km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.36\u0026ndash;54.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e54.07\u0026ndash;65.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e65.85-105.42\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e105.42-225.03\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eAnalytical methods\u003c/h2\u003e \u003cp\u003eFirst, we used descriptive statistic techniques to show the variation of leisure walking across different built environment settings and sociodemographic groups.\u003c/p\u003e \u003cp\u003eMoreover, to visually represent the distribution of these walking behaviour categories, the participants were divided into four distinct groups based on their reported walking behaviour.\u003c/p\u003e \u003cp\u003eSecond, to investigate the relationship between the built environment, demographics, and time period on leisure walking rates, multivariate models were employed. These models allowed for the isolation of the individual effects of each variable and their interactions, enabling a more nuanced understanding of their influence on leisure walking and cycling rates.\u003c/p\u003e \u003cp\u003eAs the survey included self-reported leisure cycling and walking at three different time periods (before COVID, first lockdown, and second lockdown), it is necessary to use repeated measure modelling to investigate the combined influence of built environment, demographics, and time period on cycling and walking rates. Repeated measures data is non-independent as it is taken from the same individual over time, and traditional statistical methods that assume independence can lead to biased estimates of the effects. Therefore, we decided to analyse the data by considering multilevel modelling techniques. Multilevel modelling is a statistical technique that can account for the hierarchical or nested structure of the data, and is therefore suitable for analysing repeated measures data. As the dataset in this research has a negative binomial distribution, we utilized generalized linear mixed models, which are a type of multilevel modelling that use a link function to handle non-normally distributed outcomes. This allowed us to examine the relationship between leisure walking dynamics over the study period, built environment, and sociodemographic characteristics. We treated the explanatory variables (built environment and socio-demographics) as fixed effects in all models since they remained constant during the study period. Moreover, all models included COVID stage as a key variable of interest to find out whether the effect of built environment and sociodemographic characteristics changed during different stages of COVID-19 restrictions. This approach also allowed us to investigate potential interactions between different variables.\u003c/p\u003e \u003cp\u003eIn order to interpret the coefficients produced by these models, we graphed the estimated marginal means from the model results and utilized a contrast test to determine whether there were statistically significant differences between groups. Estimated marginal means provide the unstandardized estimate values of a dependent variable, when all independent variables are controlled for. By employing contrast tests, we were able to examine and quantify the significance of variations in walking and cycling rates based on different factors such as the built environment, socio-demographic characteristics, and COVID-19 restrictions. This comprehensive analysis offers significant findings regarding the relationship between the built environment, socio-demographics, and the influence of COVID-19 restrictions on walking and cycling rates.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eDescriptive results\u003c/h2\u003e \u003cp\u003eIn this section, we first analysed changes in leisure walking during the stages of the study. Then we grouped participants based on their built environment and sociodemographic characteristics and compared the mean walking rates across different groups.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eSummary of leisure walking responses\u003c/h2\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e indicates overall leisure walking across in the stages of the study. There is a decrease in self-reported weekly recreational walking during the first and second lockdown stages. Based on a generalized linear mixed model in which we consider only the effect of the stage of study, theses changes in means are statistically significant (F (2,4022)\u0026thinsp;=\u0026thinsp;9.421, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Similarly, the percentage of people who did not walk for leisure increased from 19.2% before COVID to 22.9% in the first lockdown and 30.4% in the second lockdown. Based on the Cochran\u0026rsquo;s Q test results for walked / did not walk, the changes in walking during different stages of lockdown is statistically significant (χ2(2)\u0026thinsp;=\u0026thinsp;104.478, p\u0026thinsp;\u0026lt;\u0026thinsp;0.01).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eLeisure walking distribution during the pandemic\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eBefore COVID-19\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFirst lockdown\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSecond lockdown\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDidn\u0026rsquo;t do this\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e19.2%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e22.9%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e30.4%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1 per week\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e13.3%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e13.9%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e13.0%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2 per week\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e17.4%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e13.5%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e14.4%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3 per week\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e16.6%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e14.8%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e10.1%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4 per week\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e10.3%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e10.3%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e9.3%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5 per week\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e10.3%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e11.3%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e10.7%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5\u0026thinsp;+\u0026thinsp;per week\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e12.9%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e13.2%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e12.1%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal walked\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e80.8%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e77.1%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e69.6%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.36\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStd. Deviation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.14\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003cp\u003eNote: Change in mean and percentage who walked were both statistically significant\u003c/p\u003e \u003cp\u003eOur analysis of the walking dynamic during COVID-19 lockdown in Melbourne indicates the more pronounced declines during the second lockdown compared to the initial lockdown and the period before the pandemic. It is worth to mention that 5km restriction implemented during the second lockdown and this particular measure played a crucial role in intensifying the declines experienced during the second lockdown.\u003c/p\u003e \u003cp\u003eIn the next step, participants were categorized into four groups based on their walking behaviour. Note that these categories are used just to visualise the extent to which walking decreased over time, and are not used in the multilevel modelling. Individuals who reported no walking activity were classified under the \"no walking\" group; those who engaged in walking one or two times per week were placed in the \"Low walking\" group. Participants who walked 3 or 4 times per week were assigned to the \"Medium walking\" group, while individuals who walked 5 or more times per week were categorized as \"High walking.\" The distribution of these different walking behaviours in the study are visually presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. Most of these categories were very stable cross lockdown stages. However, the percentage of individuals who did not engage in any walking activity during a given week exhibited an upward trend over the course of the study, whereas the proportion of \u0026lsquo;medium\u0026rsquo; walkers experienced the most change.\u003c/p\u003e\u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eLeisure walking dynamics and the built environment\u003c/h2\u003e \u003cp\u003eIn this section, we take a descriptive look at the effect of the built environment on patterns of recreational walking. Figure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e indicates the changes in recreational walking by considering different built environment characteristics. People who lived in neighbourhoods with higher green space percentage were more likely to walk before COVID and there is less change during different stages of lockdown, which means perhaps people were protected against the impact of COVID-19 on physical activity (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e-a).\u003c/p\u003e \u003cp\u003eThere is a general belief among urban planners that land-use diversity promote walking (Seong et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), but in this study we have opposite observation (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e-b). It is probably because land-use diversity did have as many benefits during lockdowns when non-essential services were closed. Moreover, Wood et al. (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2010\u003c/span\u003e) found that \u0026lsquo;sense of community\u0026rsquo; encouraged leisure walking, which was associated with \u003cem\u003elower\u003c/em\u003e level of land-use mix.\u003c/p\u003e \u003cp\u003eHigh residential density (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e-c) and residential area (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e-d) lead to more walking. Perhaps seeing other people walking in the neighbourhood provided a sense of safety for people who wished to go for a walk.\u003c/p\u003e \u003cp\u003eConnectivity is associated with walking for transport (Sugiyama et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2012\u003c/span\u003e) because it makes more destinations accessible for people. On the other hand, when it comes to recreational walking the quality and proximity of recreational destinations are more important. In this study we cannot find a clear pattern for connectivity (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e-e).\u003c/p\u003e \u003cp\u003eHigher quartiles of cycling infrastructure were associated with higher walking and less change between stages of the study (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e-f). Some bike infrastructure is made up of paths shared between cyclists and people who are walking, so it is possible that neighbourhoods which have higher cycling infrastructure have more walkable paths that could motivate people to explore their surroundings.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eMultivariate modelling results\u003c/h2\u003e \u003cp\u003eIn this section, our aim was to analyse how the built environment influenced leisure walking behaviour throughout different stages of the pandemic. To investigate this, we utilized generalized linear mixed models to examine the relationship between leisure walking patterns, the built environment, and the control variables (sociodemographic characteristics). Our specific focus was on exploring whether the influence of the built environment varied across the three study periods: pre-COVID, first lockdown, and second lockdown. We hypothesised that the impact of the built environment would be particularly significant during the lockdown stages, especially in the second lockdown when participants were subjected to a 5km travel restriction.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eFixed effects models\u003c/h2\u003e \u003cp\u003eIn this section, we report the initial fixed effect f-tests of different models to examine the effect of built environment and socio-demographic characteristics on leisure walking across the three COVID stages (before COVID, first lockdown and second lockdown). All three models included COVID stage as a key variable of interest (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e), and in all models, COVID stage was statistically significant even when controlling for all other variables.\u003c/p\u003e \u003cp\u003eWe consider first two models (Model 1 and 2) as the \u0026lsquo;base models\u0026rsquo; without interaction effects. In Model 1 we evaluated the effects of COVID stage and built environment factors; we found that in addition to COVID stage, residential density, green space percentage, and land-use diversity have significant effects on leisure walking. In Model 2 we investigated whether COVID stage and built environment variables were still significant when you include demographic control variables. Indeed, in Model 2 income, age, employment, and education had a significant effect on leisure walking, yet these variables did not reduce the impact of COVID stage or built environment variables.\u003c/p\u003e \u003cp\u003eAs one of the aims of this study was to examine whether the effect of the built environment changed depending on the stage of COVID-19 lockdowns. The descriptive results shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e suggest that some interactions between built environment and COVID stage may exist. For this reason in Model 3 we included the interaction between built environment factors and stage of COVID.\u003c/p\u003e \u003cp\u003eHowever, contrary to the descriptive results in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, none of the interactions between COVID stage and built environment were statistically significant. This means the effect of the built environment \u003cem\u003edoes not change\u003c/em\u003e because of COVID-19 restrictions. In other words, locations with the most favourable built environment for walking saw the same degree of reduction in leisure walking as locations with poor built environment.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eFixed effects F-tests between built environment characteristics, socio-demographics, and leisure walking (* indicates the variable is statistically significant)\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eModel 1\u003c/p\u003e \u003cp\u003eBase model: built environment\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eModel 2\u003c/p\u003e \u003cp\u003eBase model: BE\u0026thinsp;+\u0026thinsp;demographics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eModel 3\u003c/p\u003e \u003cp\u003eBE and time interaction\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCOVID stage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e9.884*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10.293*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e10.317*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGreen space percentage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5.930*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.881*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.903*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLand-use diversity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2.902*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3.326*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3.234*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eResidential density\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.600*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3.742*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3.755*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eResidential area percentage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.610\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.967\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.947\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eConnectivity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.468\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.292\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.283\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCycling infrastructure density\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.505\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.521\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.503\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e13.888*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e13.703*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.340\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.396\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEmployment\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.180\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.192\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIncome\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6.827*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e6.733*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEducation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9.121*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e8.973*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHousehold type\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.506\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.442\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEmployment changes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.861*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.811*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eResidential area percentage * COVID stage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.295\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eResidential density * COVID stage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.215\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGreen space percentage * COVID stage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.045\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLand-use diversity * COVID stage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.124\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCycling infrastructure density * COVID stage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.149\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eConnectivity * COVID stage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.121\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c4\" namest=\"c1\"\u003e \u003cp\u003eModel fit statistics\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAIC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e16614.645\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16564.341\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e16629.070\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBIC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e16752.987\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16777.937\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e17067.539\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLog likelihood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e16570.392\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16495.745\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e16486.555\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c4\" namest=\"c1\"\u003e \u003cp\u003eNote: * indicates the variable is statistically significant\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eAcross the four models, Model 2 is the best fit based on AIC (16564.314) but slightly poorer than Model 1 using BIC (16777.937 vs 16752.982); this is not surprising as BIC penalises additional parameters more strongly than AIC. For this reason, we use Model 2 for the next section of the results.\u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e only presents the f-statistic, which is an indication of statistical significance but does not indicate the effect size of each variable. In all models, the effect of lockdown stage was significant and green space percentage, land use diversity, and residential density are leading built environment factors that have impact on leisure walking during COVID-19 restrictions.\u003c/p\u003e \u003cp\u003eIn the next stage of analysis, we interpret the estimated means for Model 2, which illustrate the average effect of a given independent variable when all other variables are controlled for. We also show the results of planned contrast tests to determine which levels of the independent variables are significantly different to each other.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eEstimated marginal means\u003c/h2\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e represents the estimated marginal means and planned contrast tests for the significant effects from Model 2: COVID stage, residential density, green space percentage, land use diversity, income, age, employment change and education.\u003c/p\u003e \u003cp\u003eIn all models, the effect of lockdown stage was significant. Planned contrasts showed that this decline was only significant for lockdown 2 relative to the other two time periods (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e-a). This is likely because the 5km travel restrictions were only imposed in this lockdown period.\u003c/p\u003e \u003cp\u003eThe estimated marginal means for built environment characteristics confirm that the relationship between walking and built environment is not necessarily linear. Residential density had a significant effect on leisure walking, but Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e-b suggests that density only increases walking to a point, before it decreases at the highest density. This could be because in very highly populated areas, even walking outdoors could be perceived to impose a risk of contagion.\u003c/p\u003e \u003cp\u003eGreen space is another significant built environment factor that contrast analysis indicate significant difference between the lowest and highest quartile of green space percentage (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e-c). This effect is also non-linear, with a clear plateau between the third and fourth quartile. This result confirms the history of research that shows a link between green space and leisure walking (Dadvand et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, Sarkar et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2015\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eLand-use diversity is another significant factor in the model and based on the contrast analysis, difference between the lowest and highest quartile of land-use diversity is statistically significant (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e-d). However, the direction of this effect was counter-intuitive, with the \u003cem\u003elowest\u003c/em\u003e quartile of diversity showing the highest leisure walking rates. As most business were closed during lockdown, perhaps land-use diversity did not have a practical meaning for people.\u003c/p\u003e \u003cp\u003eThe remaining graphs show the relationship between the control demographics and leisure walking. Income plays a significant role in predicting leisure walking behaviour, the higher income level, the higher walking rate (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e-e). Even though walking is a cost-free form of exercise, and one might think it would be more appealing to low-income people, they had the lowest rates of walking. Age is one of the most significant demographic characteristics studied, where the level of leisure walking is increased as people gets older in this study (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e-f). Moreover, based on the contrast analysis, differences between the adults\u0026rsquo; and older adults\u0026rsquo; walking level and young people are statistically significant. Employment change is another significant factor on recreational walking. People who lost a job during the pandemic had the highest walking rate compared to people with no change (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e-g). The variance among people who found a job was very high, reflecting the small number of respondents in this category (N\u0026thinsp;=\u0026thinsp;30). And finally, people with a higher education level are more likely to walk for recreational purposes (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e-h). It is worth noting that although people with a higher degree are likely to also have a higher income, since both variables are included in the model, education has an independent effect.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn the early years of the COVID-19 pandemic, many people faced significant restrictions on where and why they could travel. In Melbourne, for example, non-essential businesses closed, people were told to work from home, and they had to remain within 5km of their home. One might hypothesise that under these extreme circumstances, the design of one\u0026rsquo;s neighbourhood would have a particularly significant impact on how much people walked for leisure, as they were unable to go to a gym or visit distant parks.\u003c/p\u003e \u003cp\u003eIn this study, self-reported leisure walking reduced significantly between the first and second lockdown stages (when the 5km travel restriction was implemented). In particular, people who were already medium- or low-frequency walkers were somewhat likely to drop into low- or no-walking between the first and second lockdown stages (see Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Descriptive results suggested that some built environment factors may have attenuated some of these reductions; for example, locations with more green space saw a lower reduction in leisure walking.\u003c/p\u003e \u003cp\u003eYet multivariate modelling found that our hypothesis was \u003cem\u003enot\u003c/em\u003e supported: the effect of built environment on leisure walking was the same regardless of the COVID stage. Among the built environment factors that we studied in this model, green space, residential density and land-use diversity all influenced leisure walking regardless of pandemic stage. Places with more green space had higher rates of walking, a finding echoed by much past research (Dadvand et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, Sarkar et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). The third-highest quartile of residential density had the highest walking, declining in the highest density quartile, perhaps due to infection concerns in the most populated areas. Land-use diversity had a counter-intuitive relationship, with lower walking in the most diverse locations. This could be due to the closure of non-essential businesses during the pandemic, or because more diverse locations had less greenspace. Demographic effects were largely consistent with past research, and the effects did not vary between pandemic waves.\u003c/p\u003e \u003cp\u003eThis study has a number of limitations. The first is that leisure walking was reported based on respondents\u0026rsquo; recollection of how much they walked before COVID, so we are relying on their memory. Moreover, we did not ask them their attitude toward leisure walking and physical activity. The absence of this information may limit the comprehensive understanding of factors influencing leisure walking behaviour during pandemic. Furthermore, because we did not know the exact location of participants\u0026rsquo; homes we could not use the smallest spatial resolution (SA1) to calculate the land use variables. Instead, we had to rely on the larger SA2 statistical area, which may not accurately reflect the land use immediately around a person\u0026rsquo;s home. Finally, we did not have a direct measure of \u0026lsquo;walkability\u0026rsquo; based on the local design of footpaths and availability of pedestrian crossings.\u003c/p\u003e \u003cp\u003eYet taken as a whole, this study suggests that local built environment characteristics are an important determinant of walking for leisure whether or not people are restricted in their travel. Walking is a free and easily-accessed means of physical activity which can contribute to improved community health and wellbeing. The link between physical activity and both physical and mental health long been known (Warburton and Bredin, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, Biddle et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Supportive infrastructure plays an important role for people with inadequate access to private recreational facilities or limited mobility such as low-income people and youth. These vulnerable groups are already at greater risk of physical inactivity; therefore, it is important to consider them in neighbourhoods planning. This need was recognized by the United Nations who have codified the aim of providing \u0026ldquo;universal access to safe, inclusive and accessible, green and public spaces, particularly for women and children, older persons and persons with disabilities\u0026rdquo; in the UN Sustainable Development Goals (target 11.7).\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eMN, AD and LK worked together on research conceptualisation and methodology. MN conducted data analysis, visualisation and wrote the initial paper draft. All authors reviewed and edited the manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eBarbieri, D. M., Lou, B., Passavanti, M., Hui, C., Lessa, D. A., Maharaj, B., Banerjee, A., Wang, F., Chang, K. \u0026amp; Naik, B. 2020. A survey dataset to evaluate the changes in mobility and transportation due to COVID-19 travel restrictions in Australia, Brazil, China, Ghana, India, Iran, Italy, Norway, South Africa, United States. \u003cem\u003eData in brief,\u003c/em\u003e 33\u003cstrong\u003e,\u003c/strong\u003e 106459.\u003c/li\u003e\n\u003cli\u003eBentley, R., Blakely, T., Kavanagh, A., Aitken, Z., King, T., Mcelwee, P., Giles-Corti, B. \u0026amp; Turrell, G. 2018. A longitudinal study examining changes in street connectivity, land use, and density of dwellings and walking for transport in Brisbane, Australia. \u003cem\u003eEnvironmental health perspectives,\u003c/em\u003e 126\u003cstrong\u003e,\u003c/strong\u003e 057003.\u003c/li\u003e\n\u003cli\u003eBiddle, S. J., Ciaccioni, S., Thomas, G. \u0026amp; Vergeer, I. 2019. Physical activity and mental health in children and adolescents: An updated review of reviews and an analysis of causality. \u003cem\u003ePsychology of Sport and Exercise,\u003c/em\u003e 42\u003cstrong\u003e,\u003c/strong\u003e 146-155.\u003c/li\u003e\n\u003cli\u003eBuehler, R. \u0026amp; Dill, J. 2016. Bikeway networks: A review of effects on cycling. \u003cem\u003eTransport Reviews,\u003c/em\u003e 36\u003cstrong\u003e,\u003c/strong\u003e 9-27.\u003c/li\u003e\n\u003cli\u003eCervero, R. \u0026amp; Kockelman, K. 1997. Travel demand and the 3Ds: Density, diversity, and design. \u003cem\u003eTransportation research part D: Transport and environment,\u003c/em\u003e 2\u003cstrong\u003e,\u003c/strong\u003e 199-219.\u003c/li\u003e\n\u003cli\u003eCurrie, G., Jain, T., Aston, L. \u0026amp; Reynolds, J. 2021. Long Term Impacts of COVID-19 on Melbourne Research.\u003c/li\u003e\n\u003cli\u003eCusack, M. 2021. Individual, social, and environmental factors associated with active transportation commuting during the COVID-19 pandemic. \u003cem\u003eJournal of transport \u0026amp; health,\u003c/em\u003e 22\u003cstrong\u003e,\u003c/strong\u003e 101089.\u003c/li\u003e\n\u003cli\u003eDadvand, P., Bartoll, X., Basaga\u0026ntilde;a, X., Dalmau-Bueno, A., Martinez, D., Ambros, A., Cirach, M., Triguero-Mas, M., Gascon, M. \u0026amp; Borrell, C. 2016. Green spaces and general health: roles of mental health status, social support, and physical activity. \u003cem\u003eEnvironment international,\u003c/em\u003e 91\u003cstrong\u003e,\u003c/strong\u003e 161-167.\u003c/li\u003e\n\u003cli\u003eDe Vos, J. 2020. The effect of COVID-19 and subsequent social distancing on travel behavior. \u003cem\u003eTransportation research interdisciplinary perspectives.,\u003c/em\u003e 5\u003cstrong\u003e,\u003c/strong\u003e 100121.\u003c/li\u003e\n\u003cli\u003eEwing, R. \u0026amp; Cervero, R. 2010. Travel and the built environment: A meta-analysis. \u003cem\u003eJournal of the American planning association,\u003c/em\u003e 76\u003cstrong\u003e,\u003c/strong\u003e 265-294.\u003c/li\u003e\n\u003cli\u003eForsyth, A., Oakes, J. M., Schmitz, K. H. \u0026amp; Hearst, M. 2007. Does residential density increase walking and other physical activity? \u003cem\u003eUrban Studies,\u003c/em\u003e 44\u003cstrong\u003e,\u003c/strong\u003e 679-697.\u003c/li\u003e\n\u003cli\u003eHook, H., De Vos, J., Van Acker, V. \u0026amp; Witlox, F. 2021. Does undirected travel compensate for reduced directed travel during lockdown? \u003cem\u003eTransportation Letters,\u003c/em\u003e 13\u003cstrong\u003e,\u003c/strong\u003e 414-420.\u003c/li\u003e\n\u003cli\u003eHook, H., De Vos, J., Van Acker, V. \u0026amp; Witlox, F. 2023. Evolutions in undirected travel (satisfaction) during the COVID-19 pandemic. \u003cem\u003eTransportation research part F: traffic psychology and behaviour,\u003c/em\u003e 94\u003cstrong\u003e,\u003c/strong\u003e 99-113.\u003c/li\u003e\n\u003cli\u003eHunter, R. F., Garcia, L., De Sa, T. H., Zapata-Diomedi, B., Millett, C., Woodcock, J., Pentland, A. S. \u0026amp; Moro, E. 2021. Effect of COVID-19 response policies on walking behavior in US cities. \u003cem\u003eNature communications,\u003c/em\u003e 12\u003cstrong\u003e,\u003c/strong\u003e 3652.\u003c/li\u003e\n\u003cli\u003eKamruzzaman, M. \u0026amp; Hine, J. 2013. Self-proxy agreement and weekly school travel behaviour in a sectarian divided society. \u003cem\u003eJournal of Transport Geography,\u003c/em\u003e 29\u003cstrong\u003e,\u003c/strong\u003e 74-85.\u003c/li\u003e\n\u003cli\u003eKamruzzaman, M., Washington, S., Baker, D., Brown, W., Giles-Corti, B. \u0026amp; Turrell, G. 2016. Built environment impacts on walking for transport in Brisbane, Australia. \u003cem\u003eTransportation,\u003c/em\u003e 43\u003cstrong\u003e,\u003c/strong\u003e 53-77.\u003c/li\u003e\n\u003cli\u003eKerr, J., Rosenberg, D. \u0026amp; Frank, L. 2012. The role of the built environment in healthy aging: Community design, physical activity, and health among older adults. \u003cem\u003eJournal of planning literature,\u003c/em\u003e 27\u003cstrong\u003e,\u003c/strong\u003e 43-60.\u003c/li\u003e\n\u003cli\u003eKruger, J., Ham, S. A., Berrigan, D. \u0026amp; Ballard-Barbash, R. 2008. Prevalence of transportation and leisure walking among U.S. adults. \u003cem\u003ePreventive Medicine,\u003c/em\u003e 47\u003cstrong\u003e,\u003c/strong\u003e 329-334.\u003c/li\u003e\n\u003cli\u003eL\u0026auml;ttman, K., Olsson, L. E., Waygood, E. O. D. \u0026amp; Friman, M. 2023. Nowhere to go\u0026ndash;Effects on elderly\u0026apos;s travel during Covid-19. \u003cem\u003eTravel Behaviour and Society,\u003c/em\u003e 32\u003cstrong\u003e,\u003c/strong\u003e 100574.\u003c/li\u003e\n\u003cli\u003eLiu, J., Xiao, L. \u0026amp; Zhou, J. 2021. Built environment correlates of walking for transportation. \u003cem\u003eJournal of Transport and Land Use,\u003c/em\u003e 14\u003cstrong\u003e,\u003c/strong\u003e 1129-1148.\u003c/li\u003e\n\u003cli\u003eLotfata, A., Gemci, A. G. \u0026amp; Ferah, B. 2022. The changing context of walking behavior: coping with the COVID-19 Pandemic in urban neighborhoods. \u003cem\u003eArchnet-IJAR: International Journal of Architectural Research,\u003c/em\u003e 16\u003cstrong\u003e,\u003c/strong\u003e 495-516.\u003c/li\u003e\n\u003cli\u003eMarshall, W. E. \u0026amp; Garrick, N. W. 2010. Effect of street network design on walking and biking. \u003cem\u003eTransportation Research Record,\u003c/em\u003e 2198\u003cstrong\u003e,\u003c/strong\u003e 103-115.\u003c/li\u003e\n\u003cli\u003eMccormack, G. R., Petersen, J., Naish, C., Ghoneim, D. \u0026amp; Doyle-Baker, P. K. 2022. Neighbourhood environment facilitators and barriers to outdoor activity during the first wave of the COVID-19 pandemic in Canada: A qualitative study. \u003cem\u003eCities \u0026amp; Health\u003c/em\u003e\u003cstrong\u003e,\u003c/strong\u003e 1-13.\u003c/li\u003e\n\u003cli\u003eMcelroy, S., Fitch, D. T. \u0026amp; Circella, G. 2022. Changes in Active Travel During the COVID-19 Pandemic. \u003cem\u003ePandemic in the Metropolis: Transportation Impacts and Recovery.\u003c/em\u003e Springer.\u003c/li\u003e\n\u003cli\u003ePaul, P., Carlson, S. A., Carroll, D. D., Berrigan, D. \u0026amp; Fulton, J. E. 2015. Walking for transportation and leisure among US adults\u0026mdash;National Health Interview Survey 2010. \u003cem\u003eJournal of Physical Activity and Health,\u003c/em\u003e 12\u003cstrong\u003e,\u003c/strong\u003e S62-S69.\u003c/li\u003e\n\u003cli\u003ePollard, T. M. \u0026amp; Wagnild, J. M. 2017. Gender differences in walking (for leisure, transport and in total) across adult life: a systematic review. \u003cem\u003eBMC Public Health,\u003c/em\u003e 17\u003cstrong\u003e,\u003c/strong\u003e 341.\u003c/li\u003e\n\u003cli\u003ePower, D., Lambe, B. \u0026amp; Murphy, N. 2023. Trends in recreational walking trail usage in Ireland during the COVID-19 pandemic: Implications for practice. \u003cem\u003eJournal of Outdoor Recreation and Tourism,\u003c/em\u003e 41\u003cstrong\u003e,\u003c/strong\u003e 100477.\u003c/li\u003e\n\u003cli\u003eRunacres, A., Mackintosh, K. A., Knight, R. L., Sheeran, L., Thatcher, R., Shelley, J. \u0026amp; Mcnarry, M. A. 2021. Impact of the COVID-19 pandemic on sedentary time and behaviour in children and adults: a systematic review and meta-analysis. \u003cem\u003eInternational journal of environmental research and public health,\u003c/em\u003e 18\u003cstrong\u003e,\u003c/strong\u003e 11286.\u003c/li\u003e\n\u003cli\u003eSaelens, B. E. \u0026amp; Handy, S. L. 2008. Built environment correlates of walking: a review. \u003cem\u003eMedicine and science in sports and exercise,\u003c/em\u003e 40\u003cstrong\u003e,\u003c/strong\u003e S550.\u003c/li\u003e\n\u003cli\u003eSarkar, C., Webster, C., Pryor, M., Tang, D., Melbourne, S., Zhang, X. \u0026amp; Jianzheng, L. 2015. Exploring associations between urban green, street design and walking: Results from the Greater London boroughs. \u003cem\u003eLandscape and Urban Planning,\u003c/em\u003e 143\u003cstrong\u003e,\u003c/strong\u003e 112-125.\u003c/li\u003e\n\u003cli\u003eSeong, E. Y., Lee, N. H. \u0026amp; Choi, C. G. 2021. Relationship between land use mix and walking choice in high-density cities: A review of walking in Seoul, South Korea. \u003cem\u003eSustainability,\u003c/em\u003e 13\u003cstrong\u003e,\u003c/strong\u003e 810.\u003c/li\u003e\n\u003cli\u003eShaer, A., Rezaei, M., Rahimi, B. M. \u0026amp; Shaer, F. 2021. Examining the associations between perceived built environment and active travel, before and after the COVID-19 outbreak in Shiraz city, Iran. \u003cem\u003eCities,\u003c/em\u003e 115\u003cstrong\u003e,\u003c/strong\u003e 103255.\u003c/li\u003e\n\u003cli\u003eSugiyama, T., Giles-Corti, B., Summers, J., Du Toit, L., Leslie, E. \u0026amp; Owen, N. 2013. Initiating and maintaining recreational walking: a longitudinal study on the influence of neighborhood green space. \u003cem\u003ePreventive medicine,\u003c/em\u003e 57\u003cstrong\u003e,\u003c/strong\u003e 178-182.\u003c/li\u003e\n\u003cli\u003eSugiyama, T., Neuhaus, M., Cole, R., Giles-Corti, B. \u0026amp; Owen, N. 2012. Destination and route attributes associated with adults\u0026apos; walking: a review. \u003cem\u003eMedicine and science in sports and exercise,\u003c/em\u003e 44\u003cstrong\u003e,\u003c/strong\u003e 1275-1286.\u003c/li\u003e\n\u003cli\u003eTao, T., Wu, X., Cao, J., Fan, Y., Das, K. \u0026amp; Ramaswami, A. 2020. Exploring the nonlinear relationship between the built environment and active travel in the twin cities. \u003cem\u003eJournal of Planning Education and Research\u003c/em\u003e\u003cstrong\u003e,\u003c/strong\u003e 0739456X20915765.\u003c/li\u003e\n\u003cli\u003eWarburton, D. E. \u0026amp; Bredin, S. S. 2016. Reflections on physical activity and health: what should we recommend? \u003cem\u003eCanadian Journal of Cardiology,\u003c/em\u003e 32\u003cstrong\u003e,\u003c/strong\u003e 495-504.\u003c/li\u003e\n\u003cli\u003eWatts, P., Phillips, G., Petticrew, M., Hayes, R., Bottomley, C., Yu, G., Schmidt, E., Tobi, P., Moore, D. \u0026amp; Frostick, C. 2013. Physical activity in deprived communities in London: examining individual and neighbourhood-level factors. \u003cem\u003ePloS one,\u003c/em\u003e 8\u003cstrong\u003e,\u003c/strong\u003e e69472.\u003c/li\u003e\n\u003cli\u003eWood, L., Frank, L. D. \u0026amp; Giles-Corti, B. 2010. Sense of community and its relationship with walking and neighborhood design. \u003cem\u003eSocial science \u0026amp; medicine,\u003c/em\u003e 70\u003cstrong\u003e,\u003c/strong\u003e 1381-1390.\u003c/li\u003e\n\u003cli\u003eYin, C., Cao, J., Sun, B. \u0026amp; Liu, J. 2023. Exploring built environment correlates of walking for different purposes: Evidence for substitution. \u003cem\u003eJournal of Transport Geography,\u003c/em\u003e 106\u003cstrong\u003e,\u003c/strong\u003e 103505.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-3977307/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3977307/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe COVID-19 pandemic caused decreased physical activity levels due to isolation, travel restrictions, and facility closure. This meant that walking remained the main option for individuals to sustain their physical well-being and mental health. This study examines changes in walking behaviour during the early years of the pandemic, and how such changes were affected by the built environment characteristics of Melbourne neighbourhoods over the period of lockdowns in 2020. By evaluating the impact of built environment characteristics on leisure walking patterns during the 2020 lockdowns, we provide insights into the interplay between the built environment and physical activity. We found that self-reported leisure walking decreased notably during the COVID-19 restrictions. The influence of the built environment on leisure walking remained consistent throughout the pandemic stages. Factors such as green space, residential density, and land-use diversity demonstrated associations with leisure walking. The presence of more green spaces was linked to higher rates of walking, while moderate residential density was associated with the highest walking rates. Surprisingly, more diverse locations showed lower levels of walking, potentially due to pandemic-related closures of non-essential businesses or limited access to green spaces in these areas. These findings emphasize the importance of considering built environment characteristics in promoting and maintaining physical activity levels, even during times of restricted movement.\u003c/p\u003e","manuscriptTitle":"Did the built environment attenuate reductions in leisure walking during COVID-19? A quasi-panel study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-03-05 17:21:18","doi":"10.21203/rs.3.rs-3977307/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"68a4c7b4-aa47-440c-bfd5-d1ea7153af90","owner":[],"postedDate":"March 5th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-07-30T02:06:09+00:00","versionOfRecord":[],"versionCreatedAt":"2024-03-05 17:21:18","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-3977307","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3977307","identity":"rs-3977307","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}