Planning and Design Implementation
Walking Choice in Medium-sized and Small Cities:
2024 Volume 12 Issue 3 Pages 139-160
Details
2024 Volume 12 Issue 3 Pages 139-160
The relationship between urban form and transportation mode choice is complex. Moreover, studies have reported inconsistent results, including on the correlation between urban form and walking choice. This research explores this correlation in the medium-sized and small cities of Bandung and Yogyakarta, Indonesia, respectively. In this study, activity-based daily trip behavioural data were employed, and the respondents were selected using stratified random sampling based on the walkability of the environment and the distance to the city centre. Further, logistic regression analysis and the chi-square test were used to identify the correlation between urban form and walking choice. The results reveal that urban form has a significant effect on the decision to walk, particularly in relation to transit trips, the distance to the city centre, and pedestrian-friendly environments. However, the decision to walk is also affected by socio-economic factors. In addition, specific factors, such as gender and vehicle ownership, have different effects compared to those reported in previous studies.
Recent studies have shown that contemporary urban forms generate various critiques since people consider them to have failed to produce habitable cities and to increase social isolation, environmental degradation, and economic injustice (Alberti, Booth, et al., 2007; Beatley and Manning, 1997; Frey, 1999; Haughton, 1999; Huang, Moudon, et al., 2019; Jabareen, 2006; Kang, 2018; Kramer, 2013; Newman and Kenworthy, 1989; Tiwari, Nigro, et al., 2023; Zhou and Homma, 2022). Recently, the urban form has been considered to contribute to climate change (Sun, Zhang, et al., 2022; Tian, An, et al., 2020). The tendency in urban areas seems to involve lengthening the travel distance and increasing motorised vehicle use and mode change (trip chaining). Motorised vehicle ownership among low-income communities is apparently increasing (Boarnet and Crane, 2001; Handy, 1993a; Pickrell and Schimek, 1999), while bicycle use/walking (non-motorised forms of travelling) is declining (Lee, 2020; Newman and Kenworthy, 1989). The trend of increasing motor vehicle ownership among low-income communities has no significant changes to date (Chiu and Guerra, 2023; Ghimire and Bardaka, 2023; Herwangi, 2018; Herwangi, Pradono, et al., 2015; Klein, Basu, et al., 2023; Mulyana, Saputri, et al., 2021).
The as-yet-undetermined cause of trip behaviour change has led to the emergence of sustainable city concepts, such as new urbanism, compact cities, neo-traditional development, and sustainable transportation, that encourage the development of a hypothesis that convincingly possesses the following argument: the change in trip mode choice is related to the change in urban form (Kang, 2018; McMillan, 2005). Debates have arisen on whether urban form impacts urban sustainability, particularly in relation to how it can be employed to solve the problem of motorised vehicle use and implement sustainable transportation (Breheny, 1992; Kang, 2017; Koh and Wong, 2013; Manaugh and Kreider, 2013; Roo and Miller, 2000; Talen and Koschinsky, 2013; Williams, 2016; Williams, Burton, et al., 2000). Ongoing debates are related to how urban form reduces the trip length and time, dependency on motorised vehicles, emissions, pollution, and accidents; increases the use of public transportation and bicycles; and promotes walking habits (Berrill, Nachtigall, et al., 2024; Sun, Zhang, et al., 2022; Tian, An, et al., 2020; Tiwari, Nigro, et al., 2023). These debates strengthen the idea that a compact urban form with a mixed land-use configuration, high intensity of spatial utilisation and density, and conventional road network form (general grid form) will encourage walking trips. However, the above-mentioned urban form aspects do not immediately present a convincing solution due to the uncertainty in how urban form affects trip mode choice (Williams, 2016). While studying the influence of urban form on travel behaviour, it is important to consider a variety of elements, including personal characteristics, the purpose of the journey, demographic effects, and individual preferences (Fillone and Mateo-Babiano, 2018; Lefebvre-Ropars and Morency, 2018). This uncertainty mainly concerns whether urban form manipulation can contribute to the sustainability of trip behaviour mobility considering different economic and cultural backgrounds and whether a compact urban form (as opposed to a sprawl) is the most effective form to implement sustainable transportation and trip patterns (Breheny, 1995; Chen and Felkner, 2020; Feitelson and Verhoef, 2001; Rickaby, 1987), especially for regional and interregional trips (Heidicar, 2000; Mohapatra, Mohanachandran, et al., 2023; Monteiro, Sousa, et al., 2024; Newman and Kenworthy, 2000). Breheny, for example, considers that this relationship oversimplifies the complex nature of trip behaviour, particularly in the context of live-work spaces (Breheny, 2004). The relationship between urban form and transportation can be proven empirically; however, the precise manner in which the relationship unfolds (including walking choice) remains unknown (S. L. Handy, 1996; Ribeiro and Hoffimann, 2018; Williams, 2016). Similarly, in studies on pedestrians, the correlation between urban form and walking choice has not been precisely established since it is, in general, related to the development of walking mode demand theory and trip route choice (Boarnet and Crane, 2001; Boarnet and Greenwald, 1998; Crane, 1996; Flegal, 1999; S. Handy, 1996).
Studies on the effects of urban form on walking choice have highlighted the difficulty in predicting the correlation between urban form and trip behaviour and the former’s effect on the latter (Boarnet and Greenwald, 1998; Kang, 2015; Krizek, 2003; Ribeiro and Hoffimann, 2018; Saelens, Sallis, et al., 2003; Vale, Saraiva, et al., 2016; Ye, Fei, et al., 2017a). Therefore, further and deeper knowledge on the interaction between urban form and the decision to choose to walk is still needed (Chen and Felkner, 2020; Fillone and Mateo-Babiano, 2018; S. L. Handy, 1996; Lefebvre-Ropars and Morency, 2018; Williams, 2016). Given this, the current paper explores the effect of urban form (and non-urban form) on the decision to walk.
Efforts to attain sustainable transportation are generally realised by reducing associated environmental impacts and applying economic instruments to decrease the use of private vehicles (Campisi, Basbas, et al., 2020; Ewing and Cervero, 2017; Williams, 2016). These efforts are carried out by developing an ideal urban form; thus, the use of non-motorised transportation, such as walking, is increasing. Various studies related to the use of transportation modes and urban forms have been conducted in different scientific fields. Approaches in the field of transportation have focused on the relationship between the physical environment and travel behaviour (Kaplan, Nielsen, et al., 2016), while studies in environmental psychology, community development, and public health have addressed the relationship between the physical environment and social behaviour as well as that between social behaviour and public sentiment and health-related physical activity (Adkins, Makarewicz, et al., 2017; Chudyk, McKay, et al., 2017; Laatikainen, Haybatollahi, et al., 2019; Manz, Mensink, et al., 2018; Sanders, Roe, et al., 2019; Solbraa, Anderssen, et al., 2018; Toker, 2015). In the context of urban development, walking is considered a solution for short-distance travel where the design of the built environment should provide convenience for pedestrians.
In an attempt to describe and explain people’s tendencies and decisions to walk, the choice of pedestrian routes, and their qualitative perceptions, experts have examined factors related to individual characteristics (e.g., age, gender, income, etc.), mobility opportunities (e.g., availability of public transport), types of trips (destination, frequency, time availability, etc.), and features of the walking environment (Barros, Martínez, et al., 2017; Kaplan, Nielsen, et al., 2016; Mateo-Babiano, 2016; Zhao, Nielsen, et al., 2018). Other studies have shown that the level of convenience, comfortability, and safety factors are useful to consider. Factors such as the pavement width; objects of architectural, urban, and environmental attraction; the density of shops, bars, services, and economic activities; vehicle–pedestrian separation; bicycle tracks (cyclability); seating facilities; areas providing shelter and shade; the width of vehicle roadways; and street lighting have a significant effect on the level of walkability (Blečić, Canu, et al., 2016). Furthermore, research specifically related to walking modes has revealed the influence of visual aspects of the landscape, roadside commercial activities, safety, ease of travel, and the availability of urban spaces on the choice of pedestrian routes (Foltête and Piombini, 2010; Frosch, Martinelli, et al., 2019; McIlroy, Plant, et al., 2019; Saelens and Handy, 2008; Vallejo-Borda, Cantillo, et al., 2020).
Urban form refers to the spatial configuration formed by the variables of a city; it not only represents land use, characteristics of the urban transportation system, and urban design features but also the socio-demographic and economic conditions of the city. According to various studies, the concept of the relationship between urban form and walking choice cannot measure how urban form directly influences the choice to walk. Urban form variables need to be translated into a representative proxy variable to make them measurable. The proxy variable that represents the urban form is diverse both at the city and neighbourhood scales.
Urban form can be measured at the city, neighbourhood, and site scales (see Figure 1); these scales define the urban form and are commonly used to identify its effect on other variables (Barros, Martínez, et al., 2017; Cervero and Gorham, 1995; Ewing and Rong, 2008; Stojanovski, 2019; Toker, 2015). Some researchers have defined urban form on a city and neighbourhood scale using variables that include urban compactness, land use, public transportation availability, population density, and the level of settlement sprawl (Burgess, 2000; Cervero and Gorham, 1995; Crane, 1996; Ewing and Cervero, 2017; Ewing and Rong, 2008; Frank and Pivo, 1994; Jabareen, 2006; Newman and Kenworthy, 2000; Seong, Lee, et al., 2021; Stojanovski, 2019). Further, several others have defined urban form on a micro/meso scale (Burchfield, Overman, et al., 2006; Fulton, Pendall, et al., 2001; Heidicar, 2000; Hess, Moudon, et al., 1999; Kang, 2015, 2017, 2018; Lopez and Hynes, 2006; Neves, da Silva et al., 2021; Osorio, McCullen, et al., 2017; Ribeiro and Hoffimann, 2018; Seong, Lee, et al., 2021; Stojanovski, 2019; Ye, Fei, et al., 2017b).
This study employed city categories based on the guidelines of the United Nations (Department of Economic and Social Affairs, 2019): medium-sized cities (1 to 5 million inhabitants) and small cities/urban settlements (fewer than 500,000 inhabitants). The selected case studies involve Bandung and Yogyakarta (Figure 2), categorised as a medium-sized city and a small city/urban settlement, respectively, based on their dimensions/diameter involving pedestrian-friendly areas (Newman and Hogan, 1987). Bandung is 167.67 square kilometres (64.74 square miles) in area and is located 768 metres (2,520 feet) above sea level, with 2,507,888 inhabitants in 2019 (BPS-Statistics of Bandung Municipality, 2020). Yogyakarta is 32.5 square kilometres (12.5 square miles) in area and 113 metres (371 feet) above sea level, with 414,055 inhabitants in 2019 (BPS-Statistics of Yogyakarta Municipality, 2020).
(Sources: Wikipedia under Creative Commons License)
This research used activity-based analysis to examine trip behaviour data and trip mode usage and identify the characteristics of walking trips. Two trip models incorporating the walking mode were used in this study. The first is the binary daily trip model, which describes whether a person conducts daily trips by walking; a similar model has been employed in other studies (Cervero and Gorham, 1995; Handy, 1993b). This model is more suitable compared to one considering walking frequency as the independent variable. The case of Indonesian cities presents a different context. A study by the Institute for Transportation and Development Policy (2006) concluded that in Indonesian cities, very few daily utilitarian trips are made by walking. Not all utilitarian travellers work in the formal sector and require daily commutes. The daily trip model is effective in determining the number of utilitarian travellers but is ineffective in explaining the types of trips that occur and is only applicable to utilitarian trips. The second is a trip type model. In this study, the types of pedestrian trips were classified into three categories: the main mode of conducting trips (utilitarian), a transit mode to public transportation (transit), and a mode for recreational/sports activities (non-utilitarian).
Data collectionData was collected through a primary data survey (using direct field observations and questionnaire distribution – Table 1) and a secondary data survey. The respondents were selected based on the characteristics of their type of settlement/housing and the location of their residence. The first step was to identify the characteristics of settlements in both cities, which were grouped into two types: according to location – city centre or urban fringe – and according to the type of structured settlement – built by developers, with adequate vehicle access, or unorganised settlements/housing (with organically grown crops and narrow road access). Locations were then randomly selected for distributing questionnaires based on these characteristics (Figure 3). The questionnaires were distributed via the random selection of house numbers on weekdays from the afternoon until the evening (though not at midday because travellers still at work would not have had the opportunity to participate). The number of questionnaires was based on the population sampling of selected settlement/housing. 275 questionnaires were distributed in Bandung and 240 questionnaires in Yogyakarta. Not all respondents were willing to answer. After assessing the validity of the questionnaires, 223 questionnaires for Yogyakarta City and 172 for Bandung City qualified for further analysis.
| Bandung City | Yogyakarta City | ||
|---|---|---|---|
| Housing Environment | |||
| Unorganised | 77 | Unorganised | 143 |
| Organised | 95 | Organised | 80 |
| Location (proximity to city centre) | |||
| City Centre | 54 | City Centre | 79 |
| City fringe area | 118 | City fringe area | 144 |
The research was conducted in three stages (Figure 4). First, the urban form variables (at various scale levels) were constructed and their effect on the choice of walking mode was considered to explore the urban form. The selected meso-scale urban form variables included environmental quality (type of housing and level of environmental friendliness for pedestrians), the proximity of access to public transportation, and infrastructure (pedestrian support facilities). On the macro scale, the variables included proximity to the city centre (proximity/centrality), land use, built-up area, density, and city size (see Table 3). Second, the characteristics of trips made on foot were explored using a questionnaire addressing socio-demographic and economic factors and the travel characteristics and modes used based on daily activities and respondents’ perceptions regarding the urban form variable. Third, using descriptive statistical analysis, the factors that influence the decision to walk and the effect of the city shape on walking choice were measured.
Urban form, as considered in this research, involves the city, neighbourhood, and site scales (see Figure 1). These three scales, used in defining urban form, are commonly employed in research studies aiming to investigate the effect of urban form with other variables (Ewing and Rong, 2008). Burgess (2000) defined urban form at the city scale using variables including city compactness, land use, public transportation availability, population density, and the level of housing sprawl, while other studies have defined urban form at the site scale (Ewing and Rong, 2008; Fulton, Pendall et al., 2001; Kang, 2015; Lopez and Hynes, 2003). Measurements using the city (macro) and neighbourhood (meso) scales were conducted to distinguish between the characteristics of the two cities of the case study. Meanwhile, measurements on the individual scale were conducted to assess the effect of urban form on trips involving the walking mode. Table 2 presents the variables used to identify the characteristics of the cities considered herein and those used in the model.
| Urban form factor | Measuring tool | Variable used in the model |
|---|---|---|
| Level of environmental compactness | Spatial Utilisation Intensity | Type of Housing Environment (organised/unorganised) |
| City centre | Moran’s I | Distance to city centre |
| Public Transportation | Availability of and accessibility to public transportation | Distance from residence to the nearest public transportation stop |
Remarks:
Moran’s I was calculated using the formula:
| Category | Sub Variable | Description | Code |
|---|---|---|---|
|
Urban form and Environmental Factor |
Housing location | Distance from the respondent’s residence to the city centre: 1 if respondent resides in the city centre, and 0 otherwise. | Location |
| Level of environmental compactness | Type of respondent’s housing (organised or unorganised): 1 if respondent resides in organised housing, and 0 otherwise. | Housing | |
| Availability of public transportation | Distance from residence to public transportation stop | Pub_tranprt | |
| Pedestrian friendliness | The perception of whether the pedestrian path around the residence is comfortable for walking: 1 if respondent considers the surrounding environment comfortable for walking, and 0 otherwise. | Ped_friend | |
| Pedestrian trip | Pedestrian path design | The result of principal component analysis (PCA) from pedestrian path design variables: width of pedestrian path, material of pedestrian path, lighting of pedestrian path, supporting facility of pedestrian path, connectivity of pedestrian path, and security and safety of pedestrian path. After these were transformed into one variable, the score factor for each case was used as the variable for analysis. | Fac_score |
| Decision to walk | Respondent’s decision to walk daily: 1 if respondent walks daily, and 0 otherwise. | Walking | |
| Maximum walking distance | Maximum distance that can be used by respondent in walking | Max_walk | |
| Private vehicle ownership | Number of private vehicles owned | Vehicle | |
| Purpose of daily walk | The purpose of daily walking is divided into three categories: walking as the main mode, walking as a transit mode, and walking as a recreational/sports activity. | Purp_walk1, Purp_walk2, Purp_walk3 | |
| Socio-economic Condition | Per capita income | Per capita income is calculated by dividing the household income by a conversion factor: (a + 0.7c) 0.5, where a is the number of adults and c is the number of children in the household. This conversion follows MacKerron and Mourato’s approach (MacKerron & Mourato, 2009). Household income is calculated from the median range of the income category. | Income |
| Gender | Gender: 1 if the respondent is female, and 0 if male. | Sex | |
| Age | Respondent’s age | Age | |
| Level of Education | Level of education: 1 if respondent is a high-school graduate, and 0 otherwise. | Education | |
| Occupation | Does the respondent have a job? 1 if respondent has a job, and 0 otherwise. | Occupation |
Simple association analysis between two variables was performed using the chi-square test (χ2), which is only used to determine whether a variable is significantly associated with another particular variable. The level of association cannot be ascertained via this test but rather via multivariate analysis using a logistic regression model.
Factors affecting the decision to carry out daily walking trips are determined using logistic regression analysis since the dependent variable is bivariate (0: not walking and 1: walking). The regression formula is F (walking) = f (urban form, socio-economics), where the dependent variable used in the model is the decision to walk, and the independent variable includes income, age, vehicle, fac_score, and max_walk (continuous variables) and housing, location, sex, occupation, education, walking, pub-tranprt, and ped_friend (discrete variables).
Spearman rank or Pearson product-moment bivariate correlation analysis was conducted to test the multicollinearity in the logistic regression model. Logistic regression does not require normality and other classical assumptions, although one can still screen the data for outliers. Logistic regression interpretation is carried out using the odds ratio, and in such analyses, the interpretation of the feasibility of the model is conducted using the chi-square value, which is expected to be smaller than that in the statistics table.
To compare the characteristics of travelling on foot in the two cities considered herein, a marginal effect was calculated by comparing the magnitude of the effects of the variables. These two effects were compared and ranked according to their order. The use of the magnitude of this effect has been standardised such that it can be compared between two different samples.
The city of Bandung is characterised by a low level of grouping that tends to spread randomly. A high concentration of spatial utilisation is present at the centre of activities, unlike Yogyakarta, where the concentration of spatial utilisation is relatively equally spread throughout the city (Table 4).
| Measuring Tool | Bandung | Yogyakarta |
|---|---|---|
| Spatial Utilisation Intensity | The concentration of spatial utilisation is high and mixed around the city centre and less equally distributed. Meanwhile, high spatial utilisation is also found in some primary and secondary centres in the eastern part of the city. | The concentration of spatial utilisation is high and mixed and tends to spread. Housing conditions, particularly in the city centre, are characterised by high utilisation intensity served by a narrow road network. |
| Moran’s I (using public facility distribution) |
0.1722 The level of grouping tends to be low and is spread randomly, with a strong influence from the city centre. The high intensity of spatial utilisation around the city centre and secondary centres is one of the factors attracting walking mode trips. |
0.4623 This tends to be a city with a high level of grouping; however, dependency on the city centre tends to be low due to the spatial utilisation intensity being spread equally over the entire city and its fringes. |
| Availability of and Accessibility to Public Transportation | Regional range service and various types of modes (public transportation, ojek, city bus, and taxi). | Not attaining evenly distributed regional range service, long waiting period, and limited operational time (only until 6 p.m.). Types of public transportation include the city bus (TransJogja), ojek (motorcycle), and taxi. |
The characteristics of walking trips in Bandung and Yogyakarta are presented in Table 5, and the walking frequency and longest walking distance are given in Table 6.
| City | Frequency | Percentage | Cumulative % |
|---|---|---|---|
| Bandung | |||
| Daily walking trip (1) | |||
|
|
109 | 63.4 | 63.4 |
|
|
63 | 36.6 | 100 |
| Type of Walking Trip (2) | |||
|
17 | 9.9 | 9.9 |
|
83 | 48.3 | 58.1 |
|
72 | 41.9 | 100 |
χ2 (1) with motorised vehicle ownership = 3.521E1, p<0.001
χ2 (1) with type of daily trip = 1.345E1, p<0.05
χ2 (1) with the frequency of walking = 9.356 (n.s.)
| Yogyakarta | |||
|---|---|---|---|
| Daily trip with walking mode (1) | |||
|
|
94 | 42 | 42.2 |
|
|
129 | 57.8 | 100 |
| Type of Walking Trip (2) | |||
|
44 | 19.7 | 19.7 |
|
26 | 11.7 | 31.4 |
|
153 | 68.6 | 100 |
χ2 (1) with motorised vehicle ownership = 3.976E1, p<0.001
χ2 (1) with type of daily trip = 3.239 (n.s.)
χ2 (1) with the frequency of walking = 6.676 (n.s.)
Remarks: n.s. = not significant
* The case where respondents answered that they did not travel daily on foot does not imply that they did not walk; however, walking was not carried out for daily trips.
| Variable | Mean | Std. Deviation | Min | Max |
|---|---|---|---|---|
| Bandung | ||||
| Weekly Walking Frequency | 4.51 | 2.726 | 0 | 7 |
| Longest Walking Distance (m) | 416.86 | 171.015 | 100 | 1000 |
|
χ2 walking distance frequency with daily utilitarian trip = 1.521E1, p<0.001 χ2 the longest distance of daily utilitarian trip = 0.984 (n.s.) |
||||
| Yogyakarta | ||||
| Weekly Walking Frequency | 4.34 | 2.848 | 7 | |
| Longest Walking Distance (m) | 399.10 | 168.99 | 900 | |
|
χ2 walking distance frequency with daily utilitarian trip = 1.521E1 (n.s.) χ2 the longest distance of daily utilitarian trip = 1.821E1 (n.s.) |
||||
Remarks: n.s. = not significant
In Bandung, the average frequency of trips made on foot by the respondents exceeds four times per week. However, more than 40% of the trips are of the non-utilitarian or recreational/sport types. The average distance of the trips made by the respondents is approximately 400 metres, and the walking frequency is not significantly associated with daily walking trips. Although some respondents do not travel daily on foot (during weekdays), they continue to walk as part of sports and/or recreational activities (109 respondents). The chi-square test (χ2) between the frequency of walking and the type of daily trip reveals an association between the two variables (p<0.001) and, hence, a possibility that different types of daily trips influence the frequency of walking. This indicates that a person who travels in a utilitarian manner walks more often than a non-utilitarian pedestrian. On the other hand, there is no association between the longest walking distance and the type of trip, implying that all types of trips have a similar maximum distance. Some literature sources have indicated that the average walking distance is approximately 400 metres to 0.8 kilometres (Hess, Moudon, et al., 1999; Horak, Kukuliac, et al., 2022).
In Yogyakarta, no relationship exists between daily walking trips and the types of daily trips, and further, no association is present between the frequency of walking types and the types of daily trips. Walking in Yogyakarta City is an activity unaffected by the type of trip. This implies that travel activities depend on individual initiative, with very little influence from external factors, such as economic reasons that make it “compulsory” to travel on foot (or, in other words, most types of travel that occur in Yogyakarta City comprise utilitarian trips). In addition, in this city, no association was found between the longest distance on foot and the type of daily trip. Table 7 presents the correlation between the characteristics of walking trips.
| Variable | Variable | City of Bandung | City of Yogyakarta |
|---|---|---|---|
| Daily walking trip | Type of daily trip |
χ2 = 1.345E1, p<0.05 An association exists between the two variables. It is possible that the specific type of daily trip will decide whether a person conducts daily walking trips or not. |
χ2 = 3.239 (n.s.) There is no correlation between daily walking trips and the type of daily trip. This implies that most respondents “are not forced to walk”; their actions depend more on their willingness to walk. |
| Frequency of walking |
χ2 = 9.356 (n.s.) Not statistically significant if it is associated with daily walking trips. Even though they do not carry out daily walking trips (on working days), some respondents continue doing walking trips as part of sports and/or recreational activities (109 respondents). |
χ2 = 6.676 (n.s.) No correlation between the frequency and type of trip, strengthening the findings that walking trips in Yogyakarta comprise an activity unaffected by the type of trip. |
|
| Motorised vehicle ownership |
χ2 = 3.521E1, p<0.001 The low number of pedestrians making daily walking trips is correlated with motorised vehicle ownership in Bandung. |
χ2 = 3.976E1, p<0.001 Trip activities depend on the initiative of the pedestrian, and the effect of external factors, such as being “obliged” to walk, is minimal (or, in other words, most types of trips in Yogyakarta are utilitarian). |
|
| Type of daily trip | Frequency of walking |
χ2 = 1.521E1, p<0.001 An association exists between the two variables, and the differences in the types of daily trips may influence the frequency of walking. This implies that pedestrians conducting utilitarian trips will be more inclined to walk than those making non-utilitarian trips. |
χ2 = 1.521E1 (n.s.) No association between frequency of walking and type of daily trip strengthens the finding on the absence of the utilitarian pedestrian group who are “forced” to walk in Yogyakarta. |
| Farthest distance |
χ2 = 0.984 (n.s.) No association between the longest walking distance and the type of trip, implying that the maximum distances of all types of trips are almost similar. Some literature sources have indicated that the average walking distance is approximately 400 metres to 0.8 kilometres. |
No association between the longest walking distance and the type of trip, implying that the maximum distances of all types of trips are almost similar. Certain literature studies have revealed that the average walking distance is approximately 400 metres to 0.8 kilometres. |
|
| Perception of the pedestrian facility condition | Daily trip |
χ2 = 6.313 (p<0.05) A correlation between the two variables is present, indicating that the pedestrian path condition possibly affects daily trips in Bandung. |
χ2 = 0.801 (n.s.) The general design of the pedestrian path and daily trips show that no correlation exists between the two. This is probably related to the finding above and could mean that daily walking trips only occur in certain places. |
| Type of trip |
χ2 = 6.635 (n.s.) No association between the type of trip and the condition of the pedestrian path. This shows that all pedestrians conducting walking trips have similar notions – that is, the pedestrian path is less than appropriate for walking trips. |
χ2 = 6.940 (n.s.) No association between the condition of the pedestrian path and the type of trip, showing that different types of trips only occur in certain locations. |
Remarks: n.s. = not significant
Factors affecting daily walking trips in the city of Bandung are based on gender (men tend to conduct more walking trips compared to women), environmental compactness, and residential location (the determining factor in making the decision to walk). These three factors have a negative effect on the decision to walk (see Table 8). In Yogyakarta (see Table 8), factors affecting daily walking trips are the per capita income, which shows a significant negative effect (the higher the income, the lower the willingness to walk), and the urban form variable of proximity to the city centre (walking also occurs in the fringe areas), which also exhibits a negative effect. The level of education, on the other hand, has a positive effect on the decision to walk.
| Variable | Bandung | Yogyakarta |
|---|---|---|
| Urban form variable | ||
|
|
-0.184** | -0.080 |
|
|
0.131 | -0.148* |
|
|
-0.007 | 0.019 |
|
|
0.074 | 0.084 |
|
|
-0.111** | 0.017 |
| Socio-economic variable | ||
|
|
-0.789 | -0.046** |
|
|
-0.260** | 0.038 |
|
|
0.001 | -0.004 |
|
|
-0.086 | -0.016 |
|
|
-0.117 | 0.306*** |
|
|
-0.039 | 0.039 |
Remarks:
A total of 36.6% of the respondents in Bandung and 57.8% of those in Yogyakarta conduct walking trips. The ‘Income’ variable is modified by multiplication by a factor of 10E8.
Influential: *p<0.01; ** p<0.05; *** p<0.001
As presented in Table 9, the urban form factors affecting the decision to walk are based on the type of trip. In Bandung, these factors are the proximity to public transportation and to the city centre (positive influence) and the level of environmental compactness (negative influence) for utilitarian trips. On the contrary, for transit trips, the proximity to public transportation and that to the city centre have a negative influence on the decision to walk, while the level of environmental compactness has a positive influence. Finally, for non-utilitarian trips, the level of environmental compactness has a negative effect, whereas pedestrian-friendly environments have a positive effect. Regarding Yogyakarta, utilitarian trips are positively affected by the condition of the pedestrian path. The proximity to public transportation and the condition of the pedestrian path has a negative effect on transit trips, and there is no urban form variable affecting non-utilitarian trips.
| Variable | Bandung | Yogyakarta | ||||||
|---|---|---|---|---|---|---|---|---|
| (1) | (2) | (3) | (1) | (2) | (3) | |||
| Urban form variable | ||||||||
|
|
0.043** | -0.120* | 0.0514 | 0.011 | -0.128** | 0.080 | ||
|
|
-0.029** | 0.277** | -0.256** | -0.044 | 0.025 | 0.031 | ||
|
|
0.035** | -0.174* | 0.059 | -0.002 | 0.072 | -0.047 | ||
|
|
-0.002 | 0.059 | -0.042 | 0.035** | -0.055** | -0.002 | ||
|
|
-0.007 | -0.117 | 0.131*** | -0.182 | n.a. | 0.052 | ||
| Socio-economic variable | ||||||||
|
|
-0.032** | -0.101** | -0.863 | -0.076** | -0.345** | -0.012 | ||
|
|
-0.020** | -0.012** | -0.002** | -0.012* | 0.037* | 0.009 | ||
|
|
0.001 | 0.076 | 0.087 | -0.002 | -0.004 | -0.009 | ||
|
|
-0.075 | -0.09 | 0.05 | -0.050 | -0.064 | 0.071 | ||
|
|
-0.041 | -0.024 | -0.011* | 0.312*** | 0.119** | 0.101* | ||
|
|
-0.481*** | -0.112** | 0.267*** | 0.017 | -0.063** | 0.062* | ||
|
Remarks:
|
|
|||||||
The results of the above analysis reveal the differences in the effect of urban form on walking trips in Bandung and Yogyakarta and are summarised as follows:
The results of this study indicate that walking choice is not completely influenced by the variables of city form. In some cases, an individual’s decision to walk is also affected by socio-economic factors. For individuals who have limited economic capacity, the comfort/quality of the environment does not encourage walking, especially for utilitarian trips. These results limit the relationship that can be established between environmental variables and walking decisions, as referred to in previous studies (Adkins, Makarewicz, et al., 2017; Blečić, Canu, et al., 2016; Foltête and Piombini, 2010; Frosch, Martinelli, et al., 2019; McIlroy, Plant, et al., 2019; Saelens and Handy, 2008; Seong, Lee, et al., 2021; Vallejo-Borda, Cantillo, et al., 2020).
This study confirms or supports the new urbanism approach in urban planning and design (Black, Collins, et al., 2001; Crane, 1994; Handy and Clifton, 2001; Hess, Moudon, et al., 1999; Kang, 2017; Krizek, 2000; Shriver, 1997). The study found that to a certain extent, high-density street network configurations enable the occurrence of walking trips (Gordon and Richardson, 1995; Kang, 2017; Seong, Lee, et al., 2021). In the context of developing countries, the arguments presented by new urbanism are not entirely valid, considering the presence of other factors or variables, such as non-urban form or city size variables (socio-economic), that are more influential regarding the walking mode choice. This study reveals that walking choice is affected by personal preferences or considerations (lifestyle, for instance) and socio-economic factors related to conducting walking trips (Frank and Pivo, 1994; Newman and Kenworthy, 2000; Zhao, Nielsen, et al., 2018).
The main findings of this study strengthen the argument that walking trips are affected by urban form factors. The results of the analysis reveal that walking trip patterns are determined by individual pedestrian decisions. However, the decision to conduct walking trips is not only affected by socio-economic factors but also by the external environment, as proven by the study. One can conclude that environmental conditions, from the meso (vicinity of a residence) to macro (city) scales, also have an effect in determining the pattern character and walking trip behaviour. The study results also lead to the conclusion that urban form is imperative in analysing walking trip patterns and behaviours in a city. The walking choice is also affected by urban form variables on the macro scale. Therefore, the approach and model of the correlation between urban form and walking choice should consider urban form factors at a scale wider than the meso scale.
In Bandung and Yogyakarta, the decision to walk is influenced not only by urban form variables but also by socio-economic factors (Neves, da Silva, et al., 2021). In some cases, walking choice is determined by economic limitations, where people have no choice but to walk. The purpose of trips is diverse and not singular, where the location of residences is influenced more by land prices such that it is not suitable for the place of work; this shows that the decision to walk is influenced not only by factors on the micro (site) and meso (neighbourhood) scales but also by aspects of the wider environment (macro scale). Therefore, developing policies that encourage walking and that are linked to the integration of more efficient and affordable public transport modes is important. Meanwhile, efforts to increase walking for non-utilitarian purposes can focus on a more equal distribution of service centres.
Planning pedestrian facilities using a technical approach that does not consider the patterns and characteristics of travelling on foot will prove inefficient and ineffective. This study highlights the importance of developing pedestrian facilities in the city centre and around housing areas both on the outskirts of the city and in the city centre. In the context of transportation, this study supports the policy of using mixed modes with the main modes of public transportation to reduce dependence on private vehicles and promote the development of transit facilities, parking, and public transportation stops at urban activity centres. Increasing the access or proximity to public transportation is important not only in the city centre but also in suburban areas. This study contributes to the development of spatial structure and land use configuration policies by distributing urban activities to reduce dependency on the city centre. Although the study results indicated a weak relationship, it is important to consider the proximity, concentration, grouping, and continuity of activities to reduce travel distances (Ding, Wang, et al., 2017). The development of complete service centres with public and social facilities will reduce the use of private vehicles for travel purposes. In line with the development of rail-based transportation and the national policy regarding the development of transit-oriented means (Ministry of Agrarian Affairs and Spatial Planning/National Land Agency, 2017), with reference to the Guidelines for Transit Oriented Development (TOD), the TOD areas should be considered in encouraging the development of walking modes.
This study has limitations in measuring urban form and a relatively small sample size. Data limitations have an impact on adjusting the variables in measuring urban form. For example, this study did not use the distance of the respondent's residence to the city centre, but only identified whether the respondent's residence was in the city centre or not. This resulted in the data for the city centre variable is considered as binary data with relatively low variance. Further research is recommended using actual distances in the city centre variable using geographic information systems thar will enrich the model and identify other things that have not been identified in this study. Likewise, the use of a larger sample may provide a more complete picture of this issue. Regarding travel patterns, this study found three different types of daily trips, in addition to utilitarian and non-utilitarian trips, with the addition of transit trips. By classifying into three types of daily walking trips, the model of daily trip types and their influencing factors, including urban form factors, could be different if applied in other cities. Especially in cities with very poor public transport, as in the case of some small or medium-sized cities in Indonesia or developing countries in general. Further research should consider other types of trips that may occur, especially in cities with different characteristics from the case study.
The authors declare that they have no conflicts of interest regarding the publication of the paper.
