Perceived Accessibility for Women: First-Mile Trips to Transit-Oriented Developments

Abstract

First- and last-mile trips significantly influence public transport ridership. This study contributes to the understanding of perceived accessibility for women in first-mile trips by considering intersectionality of land use diversity and first-mile mode choices. Trips to 12 transit-oriented development (ToD) stations across Auckland city, New Zealand, were analyzed. An equity perspective is considered by selecting ToD sites with varying land-use diversity index scores and distances traveled. Around 500 public transport riders responded to an online survey and the sample closely represents Auckland’s population. Data were analyzed using exploratory factor analysis to determine the perceived accessibility and perceived personal security scores of women in comparison with men. Findings provide statistically significant evidence of gendered differences in access to ToDs. Women are more likely to walk to ToD stations than men and they are more likely to walk within a distance of 1 km (~0.62 mi). High-intensity land use positively influences women’s perception of ease of accessing stations. This suggests women’s restricted access to cars, and highlights that walkable ToDs are more beneficial for women. Overall, findings highlight the need for disaggregated data to incorporate a gendered perspective in the design of inclusive public transport systems. As more cities develop strategies to increase ridership, it is crucial to examine travel patterns through a gender equity lens to ensure the travel needs of women are considered, particularly those who are often marginalized.

Keywords

public transportation women accessibility equity rail income

There is an inherent connection between the development of transport infrastructure and the users who will derive the greatest benefit. Ensuring accessibility to public transport is essential for creating a connected and sustainable urban environment. Transit-oriented developments (ToD) aim to improve quality of life by promoting walkability and mixed land-use around stations ( 1 ). The basic principle of ToD suggests inclusive access to all socio-economic groups, by efficient, walkable, and connective mobility modes, at lower financial and environmental expenses ( 2 ). Implementation of large public transit-oriented infrastructure can sometimes lead to inequitable and unjust outcomes to any commuters who are marginalized, such as women ( 3 ).

The ease of making first- and last-mile trips to public transport stations is critical to ensure access by different population groups. Most women prefer to walk to public transit stations to reduce journey cost, and if the distance is short with a pedestrian-friendly environment ( 4 ). Gendered travel patterns for first-mile trips can vary greatly based on a person’s perception of safety, built environment, travel time, and distance to stations ( 5 , 6 ). The built environment around public transit stations has a strong positive influence on women’s walking trips compared with men ( 4 ). Alongside this, for women, fear of victimisation has a strong influence on their walking route, return journey time, and mode choice. In a study by Chowdhury et al., young ethnic-minority women expressed feeling anxious and constantly staying vigilant as they walked home from a public transport station ( 7 ). Souza et al. found that, in Brazil, women tend to change their walking routes either because of the presence of unknown men, or empty or poorly lit walkways ( 8 ). An aspect of accessibility to ToDs that has received relatively less attention is gendered differences in the first- and last-mile trips. Understanding the accessibility patterns from home (origin) is essential for ensuring equitable access to ToDs.

Gentrification caused by ToDs has been well documented, where the primary issue is marginalization of the lower-income population resulting in re-segregation ( 9 ). Displacement of marginalized population groups and rising residential property values near stations are common implications of public-transit-oriented development projects. Additionally, the creation of job opportunities and the impact of this on regional competitiveness further contribute to these discussions. It negatively affects the marginalized population groups, as gentrification causes less accessibility to new opportunities created by ToDs. Together, these factors exacerbate societal inequalities based on commuters’ income, age and gender ( 9 ).

The present study contributes by examining trip characteristics of first-mile trips to ToDs through an equity lens. Perceive accessibility to ToDs is explored for women’s first-mile trips, in comparison with men. This is undertaken in two ways. First, factors including gender, first-mile mode choice, and land use diversity index (LUDI) values are considered in perceived accessibility to ToD stations by intersectional groups of women with age. Second, perceived personal security is included as a factor of perceived accessibility to ToD stations with high and low LUDI values. An online survey was administered in Auckland, New Zealand, and collected responses from 514 participants. Results provide evidence on the gendered differences in first-mile trips to ToDs—particularly for walking trips—and commuters’ age group. Findings from the study are expected to be relevant for planning strategies by public transport authorities to enable gender-inclusive access to stations.

Literature Review

Gendered Differences in First- and Last-Mile Trips

Public transport users display a variety of behaviors, influenced by multiple factors, in their first- and last-mile trips. Travel time to access and egress from a station and the distance to a station were amongst the predominant factors influencing first- and last-mile mode choices ( 10 ). Commuters tend to choose public transport as the first-mile mode, when access or egress time is large ( 11 ). Bergman et al. found that walking is dominant for the first-mile mode choice if the distance is less than 0.5 mi, while public transport modes are dominant when the distance is longer than a mile ( 10 ).

Many studies have been conducted giving substantial attention to the influence of socio-demographic characteristics on first- and last-mile trips ( 11 ). Meng et al. found that commuters’ gender, age, and income predominantly influence the last-mile mode choice ( 12 ). The study suggested that men tend to walk more than women, and that low-income commuters are likely to choose cycling over buses. The study further noted that the tendency to cycle in the last mile increases with age. Contrary to this, in Portland, U.S., it was found that low-income commuters tend to use feeder buses more than high-income commuters, and that park-and-ride facilities increased tendency of driving to stations ( 10 ). In the Netherlands, car availability does not affect the first-mile mode choice, and bicycle infrastructure which is well-connected with the public transport system influences mode choice ( 13 ).

Several studies have established that women’s daily mobility patterns are strongly influenced by their perceived personal security ( 14 – 16 ). Only a few studies examined the influence of gender in first- and last-mile trips ( 16 ). Alam et al. discussed that first-mile trips are not gender-neutral ( 17 ). Perceived safety in and around stations can significantly affect gendered differences in mode choices. It can dictate which stations people choose to board or egress, and the first- and last-mile mode choice ( 15 , 16 ). The study also found that, in urban environments, women prefer walking more than men to access stations ( 18 ). In Brazil, around 54% of women tend to change their walking routes because of the presence of unknown men, 39% change their route because of empty walkways, and 37% change because of dark walkways with no illumination ( 10 ). Solving issues in accessibility to stops and the design of public transport infrastructure addressing women’s urban equity can effectively reduce women’s perception of safety ( 14 ).

The impact of the built environment and transport infrastructure development is shown to have varying degrees of impact on women’s active transport mode usage for first- and last-mile journeys ( 17 , 19 ). For instance, a ToD development in New Orleans, U.S. with new bike lanes resulted in a higher uptake by women pre- and post-development ( 20 ). A study in Medellin, Columbia, showed that the shift from private automobile or taxi use to active transport or public transport use for women was 12% times more than men’s shift, after connecting walking paths and other urban mobility infrastructure developments to the metro stations ( 21 ). The study also elaborated that improved ease of access to the stations had a greater impact on women’s quality of life, as they are higher users of public transport. However, in the case of Seattle, U.S., it was found that a higher proportion of walking to stations was reported by male students who have less access to cars, compared with female students, after the opening of the Seattle Light Rapid Transit line with several pedestrian facility modifications ( 22 ).

Impact of Public-Transport-Oriented Developments

Public-transport-oriented developments include public transport infrastructure developments such as ToDs, LRTs, and bus rapid transit (BRT). Transit-induced gentrification and displacement resulting from less affordability of housing near stations and enhancement of regional competitiveness of cities by creating jobs and expansion of catchment areas are some main socio-economic implications of ToDs, in most of the areas where studies have been conducted ( 5 , 9 ). The concept of transport justice explains how inequitable and unjust outcomes to the community can be generated by implementing large infrastructure projects, without having housing regulations that protect communities which are vulnerable and do not benefit from the change ( 3 ). More recent work has found that such developments can cause gentrification, causing an inequitable share of benefits amongst different communities ( 9 ).

The concept of equitable ToDs (EToD) discusses the same concept of transport justice, where it enables communities to enjoy the benefits of mixed, dense, walkable urban development near transit stations, regardless of their age, gender, income, race, ethnicity, disability, or immigration status ( 23 ). While income is a significant factor in promoting social inclusiveness, EToD highlights the importance of considering various other factors from an equity perspective. Only a limited number of studies have explored the influence of public-transport-oriented developments on other socio-demographic characteristics. Although many studies suggest that such implementations contribute to the sustainable development of society, their impact on fostering social inclusion, as far as access to public transport is concerned, remains unclear ( 23 ).

Perceived Accessibility Score (PAC)

More recently, there has been a shift in focus to perceived accessibility. “Perceived accessibility” is the ease of living a high-quality life using the transport system, and it is suggested as a measure of subjective satisfaction with accessibility ( 24 ). Accessibility to public transport is considered an important measure in evaluation and planning of public transport to reduce social exclusion ( 25 ). Morris et al. divided measure of accessibility into two measures based on the number of opportunities and the level of satisfaction ( 26 ). Another study developed a measure which comprises of trip coverage (destination), temporal coverage (based on time of travel), and spatial coverage (closeness to the stop) ( 27 ). The most common study referenced is by Lattman et al., which provides four statements to determine the perceived accessibility score (PAC) ( 24 ).

Previous studies emphasize that measuring perceived accessibility comes with certain key components and implications ( 28 ). Overall, it was noted that most existing studies focus on the perceived accessibility of using public transport, rather than perceived accessibility to and from public transport, which are first- and last-mile journeys ( 29 ). Since accessibility is usually studied as spatial accessibility, perceived accessibility is often poorly reflected and has potential mismatches with the evaluation of accessibility based on prevailing measures ( 29 ). Perceived accessibility can be considered as a mediator between an individual’s travel behavior and the built enviornment, which again shows mismatches between objective accessibility and perceived accessibility ( 28 ). A key limitation of PAC is that PAC does not focus on the perceived safety, service quality, and an individual’s income, which are factors that are found to be significant for accessibility ( 30 ).

Mode choice strongly influences the PAC, where individuals are less likely to use non-motorized travel modes in instances where the perceived accessiblity by car is high ( 31 ). In Malmo, Sweden, cyclists reported a significantly higher PAC compared with car or public transport users for daily travel activities ( 32 ). However, a study from the UK found a weakly positive correlation between PAC and car availability, contrary to the usual assumption of higher PAC for car availability ( 28 ). Moreover, greater travel distances reports a negative influence on the perceived accessibility of public transport ( 33 ).

An evaluation of PAC based on the age of commuters revealed that commuters in their thirties and elderly riders testified to a lower level of PAC for daily activities using public transport than younger riders ( 24 ). A study conducted in Melbourne, Australia, captured a similar result where the PAC based on age confirmed that perceived accessibility tends to decrease when one ages ( 30 ).Women are also more likely to be dissatisfied with their perceived accessiblity for public transport than men ( 30 ).

The present study contributes by utilizing PAC measures to explore perceived accessibility of first-mile trips from an equity perspective for ToDs. In addition, gender differences in perceived accessibility are measured by including perceived safety as one of the measures of PAC. While most studies highlight the quality of travel, safety, and frequency of travel to be important determinants of perceived accessibility, gendered differences using PAC measures for ToDs has not been explored. To meet the demands of public transport, the system needs to be accessible to all user groups. From the literature review, it is more apparent that perceived accessibility can act as a barrier to public transport ridership. As such, the present study contributes to this knowledge to enable planners and operators to provide a system that is accessible, in particular for vulnerable population groups.

Methodological Approaches

Selected Sites

Overview of Selected Sites

Sites selected include major interchanges and BRT stations in Auckland, New Zealand. Auckland was selected as the case study as it has the largest public transport network in New Zealand. Twelve ToD stations were selected comprising of both inter-modal and intra-modal transfers: three multi-modal interchanges, five train stations, and four BRT stations. Auckland’s public transport network has four commuter rail lines: Eastern Line, Southern Line, Western Line, and Onehunga Line, and one BRT, known as the Northern Busway. All train lines and bus services terminate at Britomart Transport Interchange in the city centre. Of the 12 sites, majority are bus-train interchanges, except for those along the BRT line. The New Lynn Interchange (MULTI_WEST_2) is the first ToD designed in Auckland. It was opened in 2011 and features a weather protected interchange for bus services and train lines, a park-and-ride, and traffic calming features in adjacent roadways. Figure 1 shows their geographical distribution ( 34 ). There are no train lines in the northern part of Auckland. The public transport service is served by the Northern Busway, which was developed in 2008. The main BRT line is served with local bus feeder services at five stations through intra-modal transfers, and the surrounding land use was developed to meet the stations’ needs. In general, as shown in Figure 2, the multi-modal interchanges showcase a diverse land allocation in the ToD catchment area, compared with other selected intra-modal interchanges.

Figure 1.

Selected transit-oriented development sites.

Figure 2.

Land use allocation of selected transit-oriented development sites.

For this study, each station is denoted with its location and an arbitrary station number. Each station was named based on the area of Auckland where the site is located. Accordingly, altogether there are eight railway transit stations—two sites in West Auckland, four sites in East Auckland, and two sites in South Auckland—as well as four BRT stations in North Auckland. Figure 2 illustrates the type of land use surrounding each station, LUDI values, and average household income level.

Nationally, there is no breakdown of income levels and the only references to income breakdowns are: the income level divisions of the national census questionnaire–2018, annual incomes for tax rates by the Inland Revenue Department (IRD) of New Zealand–2021, and income limits for the community services card (CSC) ( 35 ). Annual income breakdowns for tax brackets from IRD provide a threshold value for high-income levels ( 36 ). CSC is a system which supports low-income individuals and families with health care costs and public transport costs, and income limits in CSC provide a threshold value for low-income levels ( 37 ). Accordingly, based on these threshold values, income levels for this study were categorised as: income less than $30,000 (low-income), $30,001–$60,000 (low-middle income), $60,001–$90,000 (middle-income), $90,001–$120,000 (middle-high income), and income more than $120,000 (high-income).

Land Use Diversity Index (LUDI)

LUDI is a measure that is used for spatial patterns and indicates the diversity or mixedness of land use ( 38 ). In general, land use near main corridors or interchanges typically has higher LUDI values, whereas lower LUDI values are common for residential developments. The higher the LUDI value, the more diverse the surrounding land use of the station would be, which is an indicator of a beneficial ToD.

Considering the definition of ToDs, a walkable distance of 1.0 km radius around each site was selected ( 1 ). Land use data were obtained using Auckland Council’s GeoMaps mapping service ( 34 ). LUDI values for each site are calculated using Simpson’s diversity index, which ranges between 0 and 1 ( 30 ). Equation 1 is commonly adopted to determine the degree of land-use mix through the proportion of different land uses in the surrounding area of a station. The proportion of land use area is weighted based on the total area of the site, to calculate the LUDI value. Therefore, n_i is the land use area of each land use category, and N_i is the total land area of the respective site. For this study, seven land use categories were considere: residential (A1), commercial and retail (A2), open spaces (A3), public facilities (A4), terrain (A5), station area (A6), and other uses (A7) ( 39 ).

An example of LUDI calculation is shown in Equation 2 for MULTI_EAST_4 site, which has a high proportion of commercial, retail, and public facilities. It was calculated as 0.729, meaning it has a high land use mix. For the study, a LUDI value greater than 0.701 was categorised as high and less than 0.701 was categorised as low, based on the median value of the calculated LUDIs (0.701).

D = 1 - \frac{\sum n_{i}^{2}}{N_{i}^{2}}

(1)

where

D = Simpson’s diversity index,

n _i = land area of ith land use category, and

N _i = total land area.

D_{EAST_4} = 1 - \frac{[A 1_{EAST_4}^{2} + A 2_{EAST_4}^{2} + A 3_{EAST_4}^{2} + A 4_{EAST_4}^{2} + A 5_{EAST_4}^{2} + A 6_{EAST_4}^{2} + A 7_{EAST_4}^{2}]}{[A 1 + A 2 + A 3 + A 4 + A 5 + A 6 + A 7]_{EAST_4}^{2}}

(2)

D_{EAST_4} = 1 - \frac{[296, 962 . 3_{EAST_4}^{2} + 1, 041, 684 . 7_{EAST_4}^{2} + 287, 548 . 1_{EAST_4}^{2} + 166, 214 . 1_{EAST_4}^{2} + 56, 475 . 7_{EAST_4}^{2} + 22, 728 . 1_{EAST_4}^{2} + 549, 126 . 2_{EAST_4}^{2}]}{[2, 420, 739.2]_{EAST_4}^{2}}

$D_{EAST_4} = 0.729$ (high LUDI)

MULTI_EAST_1 site, a middle-income suburb, has the highest LUDI value and TRAIN_SOUTH_1 site is a very low-income area with the lowest LUDI value. Compared with all other sites, those in South Auckland have the lowest LUDI valuess. In sites such as MULTI_EAST_1 and MULTI_EAST_4, although the LUDI value is high, a high proportion of the surrounding is unused lands or parking lots. In several sites, such as MULTI_EAST_1, TRAIN_EAST_2, BRT_NORTH_2, and BRT_NORTH_4, a significant land area is allocated for water bodies.

Survey Design and Description of Variables

An online survey was distributed by an independent company. The survey was restricted to those who are current public transport users. The questionnaire comprises of three sections. The first section includes questions on basic socio-economic characteristics (gender, age) and typical trip details that were commonly used in previous studies, including trip-frequency, distance traveled, mode choice, and trip-purpose. Response categories for these socio-demographic questions were adopted from the national census questionnaire ( 35 ). The second section includes questions on participants’ first-mile trips (mode choice, travel time, distance, reason for mode choice), and the origin of journey station (typical wait time, CCTV availability, etc.). It is assumed that the last station of a public transport journey in Auckland would be Britomart station, as most people come to the city centre for either employment or educational purposes. The built environment characteristics were obtained based on the questions asked about the first station of the public transport journey, and land use data from Auckland Council’s GeoMaps mapping service were used to further study the station area characteristics ( 29 ). The third section comprises of six Likert scale statements to evaluate the PAC of commuters in an equity perspective.

The PAC that was developed by Lattman et al. is based on key aspects of perceived accessibility, as follows ( 24 ).

PAC 1: “I have adequate access to my preferred activities by transport mode X.”

PAC 2: “By transport mode X, it is possible to do all my preferred activities.”

PAC 3: “I’d be able to continue living the same way, if X was my only travel mode.”

PAC 4: “It is easy to do my activities by X.”

The first statement is a measure of outcome indicators of accessibility. It captures the actual use and satisfaction of accessibility, rather than the hypothetical possibility ( 24 ). The second statement discusses process indicators, which capture the possibilities of travel to opportunities in a certain mode of travel. The third statement captures the potential of interaction opportunities for an individual, and was based on the definition of accessibility. The final statement captures the “ease” of reaching activities, in relation to the definition of accessibility. Lattman et al. discussed that this statement should be dependent not only on time and cost components, but also on the perceived safety, perceived quality, and frequency of travel as well ( 24 , 25 ).

As shown in Table 1, four Likert scale statements were developed in accordance with the perceived accessibility measure developed by Lattman et al. ( 24 ). In addition, two statements related to safety in walking the first mile were included to determine perceived safety for access to stations.

Table 1.

Statements for Perceived Accessibility Scale (PAC)

PAC	Key aspect of accessibility	Statements developed by Lattman et al. ( 24 )	Statement developed for the present study
1	Actual adequacy of choice of mode as an outcome indicator of access	“I have adequate access to my preferred activities by X.”	“I live close to the first station so that I can walk to and from it.”
2	The potential of opportunities to travel to preferred activities of interest	“By X, it is possible to do all my preferred activities.”	“There are a lot of activities in the surrounding area both during the day and at nighttime.”
3	The perceived possibility and potential opportunities of travel	“I’d be able to continue living the same way, if X was my only travel mode.”	“The area around my first station has improved pedestrian connections with refuge island and zebra crossings signage and other pedestrian-related and streetscape improvements.”
4	Ease of accessing activities	“It is easy to do my activities by X.”	“It is easy to access my first station.”
5	Perceived safety	na*	“I feel safe walking to and from my first station during the day.”
6			“I feel safe walking to and from my first station after dark.”

Note: na = not applicable, X = mode choice.

Statements 5 and 6 were developed by the author based on previous studies on perceived safety in access to transit-oriented developments.

Statistical Methods

PACs were calculated using factor analysis ( 24 ). As shown in the conceptual model in Figure 3, PAC scores were calculated for factors including gender, LUDI value and first mode. Exploratory factor analysis (EFA) and confirmatory factor analysis have been commonly adopted to generate factor loadings in evaluating PAC ( 24 ). Following a similar approach, EFA based on principle axis factoring was preferred. This approach is mostly used in PAC measures, when there are no predictions on the existing factors and their relationship to variables ( 40 ). Assuming the four measures of accessibility are unidimensional, principal axis factoring was used. One of the main limitations of EFA is that it assumes linearity between the factors and observed variables, and that the data are normally distributed ( 40 ). The Kaiser–Meyer–Olkin (KMO) values and Bartlett’s test of sphericity are normally used to specifiy that the correlation matrix was appropriate for the factor analysis ( 40 ). The standard minimum KMO value for factor analysis is 0.05 ( 40 ). In this study, a Bartlett’s test of sphericity value below 0.05 was used to specify that the correlation matrix was not random ( 40 ). In addition, Cronbach’s alpha analysis value of above 0.70 provided a satisfactory overall correlation and an internal consistency reliability. The data analysis was conducted using IBM SPSS software (version 29).

Figure 3.

Conceptual model of present study.

Results

A total of 514 responses were received from the online survey of current public transport riders. Out of the 514 responses, 445 responses were deemed usable for analyzing first-mile journey questions after removing incomplete data. A summary of participants’ demographics and first-mile journeys is provided in Table 2. Complete responses from 514 participants were obtained. The sample sizes were close to the Auckland census data. The participant demographics recorded an almost 50% split between men and women, similar to Auckland census data. With regard to age ranges, the sample size is close to Auckland census data for participant ages 18–22 years and 65 or over; however, there is a slight difference for age ranges 23–44 years and 45–64 years. Table 3 provides a list of variables used for the mode choice model with definitions.

Table 2.

Summary of Participant Demographics and First-Mile Journeys

Variable	Participants (n = 514)	Percentage (%)
Gender
Male	268	52.1
Female	243	47.3
Non-binary/third gender	3	0.6
Age (years)
18–22	70	13.6
23–44	243	47.3
45–64	153	29.8
65 or over	48	9.3
Ethnicity
Māori	30	5.8
Pacific Islander	36	7.0
New Zealand European	191	37.2
Asian	129	25.1
Southeast Asian	22	4.3
European	23	4.5
African	5	1.0
Other	78	15.1
Trip purpose
Work	234	45.5
Education	56	10.9
Errands	69	13.4
Social	112	21.8
Other	43	8.4
Trip frequency
Less than 3 times a week	380	73.9
More than 3 times a week	134	26.1
First-mile journeys
Mode choice
Walk	206	44.7
Car	204	44.3
Cycle	14	3
Dropped off	37	8
Distance traveled in first-mile mode
Less than 600 m	67	13
600 m–1 km	140	27.2
1–2 km	97	18.9
2 km or more	210	40.9
Travel time
Less than 5 min	121	26.2
5–10 min	135	29.3
10–15 min	103	22.3
15 min or more	102	22.1

Table 3.

Descriptive Statistics for Perceived Accessibility (PAC) of Full Dataset (n = 514)

PAC item (statement)	Correlations					Mean	SD
	1	2	3	4	5
PAC 1	na	na	na	na	na	2.71	1.472
PAC 2	0.386*	na	na	na	na	2.28	1.376
PAC 3	0.319*	0.255*	na	na	na	2.75	1.399
PAC 4	0.288*	0.322*	0.213*	na	na	2.41	1.203
PAC 5	0.239*	0.296*	0.111*	0.307*	na	2.35	1.256
PAC 6	0.364*	0.403*	0.155*	0.303*	0.366*	2.15	1.344

Note: SD = standard deviation; na = not applicable.

p < 0.001.

Perceived Accessibility for All Participants

Table 3 represents the descriptive statistics for all six PAC items, including the statements for perceived safety (PAC 5 and PAC 6). The Cronbach’s alpha analysis provides a satisfactory result for the overall six item correlation (α = 0.708).

Initially, an EFA based on principle axis factoring was conducted for all participants based on their first-mile mode choice, land use mix, and gender, to have an overall picture on differences in PAC, as represented in Table 4. Mode choices include walking to station and use of car to station, where the use of car includes both its use as a mode as well as getting dropped off.

Table 4.

Factor Loadings for All Factors (n = 445)

Statement	First-mile mode choice		Land use mix		Gender
	Walking (n = 206)	Car (n = 239)	High LUDI value (n = 303)	Low LUDI value (N = 208)	Women (n = 211)	Men (n = 234)
PAC 1 (walkability)	0.543	0.475	0.597	0.422	0.593	0.606
PAC 2 (activities of interest)	0.393	0.530	0.517	0.534	0.436	0.581
PAC 3 (perceived opportunities)	0.471	0.607	0.536	0.557	0.442	0.527
PAC 4 (ease of access)	0.716	0.639	0.716	0.531	0.572	0.644
PAC 5 (safety during the day)	na		na		0.639	0.658
PAC 6 (safety after dark)	na		na		0.263	0.477
Eigenvalue	1.844	1.951	2.050	1.783	2.701	2.594
% variance	46.09	48.77	51.25	44.59	45.02	43.23

Note: PAC = perceived accessibility; LUDI = land use diversity index; na = not applicable.

As expected, perception of walkability to stations (PAC 1) and ease in accessing stations (PAC 4) are the most influential for walking commuters, in comparison with commuters who use cars. This is an expected result, as the sense of high walkability explains the basic definition of a ToD, where land use mix provides the possibility to easily access stations and other amenities within walking distance. However, the PAC for perceived possibility to access preferred activities of interest (PAC 2) and the potential opportunities to travel (PAC 3) are the most influential for car users compared with walking commuters. This result is reasonable, as Auckland is a car-centric city.

Commuters from higher LUDI-value station surroundings report higher perceived accessibility in PAC 1 and PAC 4. This result signifies as an indicator of a beneficial ToD, as perception of walkability and ease in accessing stations is higher than in low-LUDI-value sites. When land use diversity is high, the possibility of reaching opportunities including the first station of public transport journey is high, which can also be explained through this result. However, there is no significant difference for PAC 2 and PAC 3, based on the LUDI value. This could be because most selected sites had the option of park-and-ride. This enables commuters in both high- and low-LUDI-value areas to have similar opportunities in accessing preferred activities of interest.

PAC score differences based on the gender of commuters suggest there is a deficiency of equity in perceived accessibility and perceived personal security. Women report a lower PAC s core than men, in relation to all PAC items. These findings are reasonable, as mobility could be limited for an able-bodied women compared with an able-bodied man, based on differences in physical strength ( 19 ). Perceived personal security-related statements were tested only for gender, to identify gender-sensitive differences in perception of safety in accessibility. It was noted that, for both safety during the day (PAC 5) and after dark (PAC 6), women’s PAC score is less than men’s. Previous studies have found that women’s perceived personal security is strongly influenced by the built environment ( 16 – 18 ). The lower scores for PAC 2 and PAC 3 by women suggests their perception of potential opportunities and possibilities in accessing preferred activities of interest and perceived opportunities created by built environment is less than men. Therefore, these findings align with previous studies, and reasonably prove that equity in access to stations can effectively increase women’s mobility.

Gendered Differences in Perceived Accessibility and Personal Security

In this section, the data set was disaggregated to evaluate the gendered differences in perceived accessibility and personal security to understand the possible causes for these differences, based on first mile mode choice and LUDI value. Mode choices include walking to station and use of car to station, where the use of car includes both the use of private automobile as well as getting dropped off. As a result, three EFA were conducted: (a) EFA for PAC based on gender and first mode, (b) EFA for PAC based on gender and LUDI value, and (c) EFA for walking commuters’ PAC based on gender and LUDI value.

Gendered Differences for First-Mile Mode Choice

Findings represented in Table 5 suggest the first-mile mode choice is correlated with gendered differences in both perceived accessibility and personal security. The result for PAC 1 and PAC 4 for walking commuters shows that the perception of women who walk on walkability and ease of access to stations is slightly higher than that of men who walk. Moreover, of all participants, around 65% of the men are employed in comparison with 45% of women, and more women selected “errands” as their trip purpose. Therefore, the finding is reasonable, as the income level of most women is lower than men’s, and walking is more dominant in women in Auckland because of their multifaceted gendered roles ( 41 ).

Table 5.

Factor Loadings for Gender and First-Mile Mode (All Sites)

Statement	Women (n = 211)		Men (n = 234)
	Walking (n = 101)	Car (n = 110)	Walking (n = 105)	Car (n = 129)
PAC 1 (walkability)	0.676	0.537	0.606	0.527
PAC 2 (activities of interest)	0.519	0.334	0.581	0.631
PAC 3 (perceived opportunities)	0.543	0.488	0.527	0.673
PAC 4 (ease of access)	0.665	0.480	0.644	0.710
PAC 5 (safety during the day)	0.617	0.712	0.658	0.736
PAC 6 (safety after dark)	0.336	0.143	0.477	0.565
Eigenvalue	1.257	2.884	2.595	3.567
% variance	46.58	43.77	34.51	41.63

Note: PAC = perceived accessibility.

There are no significant gendered differences for PAC 2 and PAC 3, in walking commuters. Walking women’s perception of personal security after dark is less than that of walking men. Both men and women who walk have reported a lower score for PAC 6 than PAC 5. However, there is no significant gendered difference in walking commuters’ perceived personal security during the day. Therefore, these findings suggest that, overall, both genders’ perception of safety is influenced by the time of travel. Further to this, aligning with previous findings, women are more sensitive about their safety after dark.

In car users, higher scores for PAC 2, PAC 3, and PAC 4 suggests that men find more perceived opportunities and ease in accessing stations and preferred activities of interest than women who use cars. This finding again reasonably proves that women have a higher tendency to walk than men; this may be because their multifaceted gender roles make their destinations less accessible by car ( 41 ). There is no significant gendered difference in perception of using a car during the daytime; however, women who use a car report a notably lower PAC score for safety after dark than men. The high score for PAC 5 in car users, compared with walking commuters in both genders, suggests that perceived personal security influences mode choice. Scores for PAC 6 show that, contrary to expectations, women feel safer walking after dark, while men feel safer using the car after dark. This could be because of the lower tendency of women to drive, or because of women’s fear of getting into a taxi driven by a man after dark.

Gendered Differences for ToDs with High and Low LUDI Values

Table 6 represents the findings of PAC based on gender and land use mix of selected sites. It was noted from the findings that, as expected, PAC 1 is higher in high-LUDI-value sites, regardless of gender. However, contrary to men, women in high-LUDI-value sites find it easier to access stations (PAC 4) compared with women in low-LUDI-value sites. Men do not report a significant difference in PAC 4. However, there is a significant gendered difference in PAC 4 for commuters from low-LUDI-value sites, as men find it notably easier to access stations than women. The study also implied that, when the distance from ToD increases to more than 1 km, using a car is the dominant first mode choice for women, while men prefer walking. Therefore, the difference in PAC score for low-LUDI-value sites is reasonable, as mobility could be limited for an able-bodied women compared with an able-bodied man, based on differences in physical strength ( 19 ). The result for both PAC 2 and PAC 3 implies men’s perception of the possibility of reaching preferred activities of interest and perceived opportunities is higher than women, regardless of the LUDI value. It provides significant evidence of inequity for women in the accessibility of ToD stations. The score for PAC 2 and PAC 3 is least for women from low-LUDI-value sites, which indicates the importance of high-intensity land use mix around stations which results in a more accessible environment for marginalized commuters.

Table 6.

Factor Loadings for Gender and Land Use Diversity Index (LUDI) (All First-Mile Modes)

Statement	Women (n = 211)		Men (n = 234)
	High LUDI value	Low LUDI value	High LUDI value	Low LUDI value
PAC 1 (walkability)	0.677	0.520	0.644	0.558
PAC 2 (activities of interest)	0.486	0.335	0.578	0.559
PAC 3 (perceived opportunities)	0.482	0.335	0.531	0.500
PAC 4 (ease of access)	0.670	0.344	0.649	0.658
PAC 5 (safety during the day)	0.665	0.615	0.581	0.778
PAC 6 (safety after dark)	0.226	0.408	0.433	0.550
Eigenvalue	2.491	3.747	2.632	5.312
% variance	41.52	34.78	43.87	47.65

Note: PAC = perceived accessibility.

Further to this, PAC 5 scores imply that women feel the safest in high-LUDI-value sites, while men feel safest in low-LUDI-value sites. Women’s perception of safety after dark (PAC 6) is less than during the daytime, regardless of the LUDI value. This finding is reasonable, as it supports previous works which discussed women’s safety perceptions when walking at nighttime ( 17 , 18 ). However, both men and women feel the least safe in high-LUDI-value sites after dark. To determine if there are differences for high-LUDI-value sites, crime data from the New Zealand Police was used ( 42 ). Seven sites had a high number of victimization incidents (>800 in the year 2023–2024). These sites were MULTI_EAST_4 (incidents = 4,656), TRAIN_WEST_1 (incidents = 2,063), MULTI_WEST_2 (incidents = 1,652), BRT_NORTH_1 (incidents = 1,474), TRAN_EAST_2 (incidents = 1,019), TRAIN_SOUTH_2 (incidents = 971), and TRAIN_SOUTH_1 (incidents = 842). Four of the seven sites are also high-LUDI-value sites, which suggests high-LUDI-value sites report a high crime rate. Therefore, the lowest score for PAC 6, regardless of gender, in high-LUDI-value sites is reasonable.

Overall, it was noted that high-intensity land use mix positively influences women’s perception of ease of accessing stations, while it negatively influences their perception of safety after dark.

Gendered Differences in Walking Trips to High- and Low-LUDI-Value Sites

Findings in Figure 4 represent walking trips, based on the LUDI value. Both men and women in high-LUDI-value sites are more likely to walk for the first-mile trip when the distance to stations is less than 1 km. Walking is the dominant mode choice for women when distance to stations is less than 1 km, while men prefer walking longer distances. When the distance from ToD increases to more than 1 km, use of a car is the dominant first mode choice for women, while men prefer walking and driving their own vehicle.

Figure 4.

Walking commuters’ perceived accessibility score based on gender and land use diversity index (LUDI).

Overall, the percentage of walking trips when residing in a distance less than 1 km from stations was higher in high-LUDI-value sites than in low-LUDI-value sites, as expected, indicating that high-intensity land use mix positively influences walkability, regardless of gender. Moreover, walking is the dominant mode choice for women when the distance to the station is less than 1 km, while men prefer walking longer distances. Therefore, women from high-LUDI-value sites who walk find it easiest to access stations, while women from low-LUDI-value sites who walk find it the least easy. The higher percentage of walking trips in high-LUDI-value sites also implies that walking commuters’ perception of the possibility of reaching preferred activities of interest and perceived opportunities is greater in high-LUDI-value sites than low-LUDI-value sites, regardless of gender. However, women residing less than 1 km from stations in both high- (56.8%) and low- (52.6%) LUDI-value sites have a higher tendency to walk than men in high- (55.1%) and low- (47.3%) LUDI-value sites.

This finding is reasonable, because of women’s multifaceted gender roles. They access activities of interest more compared with men, which enables them to have a higher perception of potential opportunities to access by walking. Furthermore, the high number of walking trips closer to high-LUDI-value sites indicates better walkable infrastructure development in high-LUDI-value sites than in low-LUDI-value sites. This can also result in high perceived personal security in high-LUDI-value sites for women.

Overall, it was noted that women who walk have a greater tendency to walk the first mile when residing less than 1 km from stations, compared with men who walk. Moreover, women’s walkability and ease of access is influenced positively when the LUDI value is high.

Discussion

The aim of this study was to examine perceived accessibility for women in first-mile mode choice to ToD stations, in comparison with men. It contributes by considering the intersectionality of gender and age for first-mile mode choices, along with varying land use and trip characteristics, to examine perceived accessibility.

Overall, results show evidence of gendered difference in perceived accessibility and perceived personal security for first-mile trips. Women have a higher likelihood of walking than men. Of the 445 participants, around 65% of the men are employed in comparison with 45% of women, and more women selected “errands” as their trip purpose. These findings provide evidence of inequity for accessibility of ToD stations. A study by Beyazit et al. showed similar findings—that women are more likely to walk because of their multifaceted roles ( 41 ). Results also suggest that living in or near a ToD site significantly influences women’s perception of walking, compared with men. Walking is the dominant mode choice for women when the distance to stations is less than 1 km, while men prefer walking longer distances. When the distance from ToD increases to more than 1 km, being dropped off is the dominant first mode choice for women, while men prefer walking and driving their own vehicle. These findings are reasonable, as mobility could be limited for an able-bodied women compared with an able-bodied man, based on differences in physical strength ( 19 ). Findings from previous studies indicate that women tend to make shorter, yet multiple, trips compared with men and, thus, are more reliant on private vehicles ( 19 ).

Examining the influence of LUDI for first-mile trips showed that results are similar for both genders. Perception of walking was found to improve with an increase in LUDI value. Findings indicate the importance of high-intensity land use mix around stations, which results in a more walkable environment regardless of gender. Moreover, high-intensity land use mix positively influences women’s perception of ease of accessing stations, while it negatively influences their perception of safety after dark, compared with men. The findings provide empirical evidence that ToDs lead to sustainable travel patterns in commuters. Similar to results of previous studies, this could be because of either the proximity to stations, or because of design features of the development which includes high-density, pedestrian-focused land use mix with commercial and retail amenities ( 43 ).

There are some limitations identified from the analysis. Future research would benefit from a larger sample of women’s mode choice for first- and last-mile trips. In addition, a more critical geographical information systems analysis of the built environment can provide more insights into the effects of LUDI values. Given that Auckland is a car-centric city, the percentage of cyclists is very low, even more so for those who cycle for the first-mile trips. A country with a higher percentage of cyclist can provide a better understanding of active mode choice for women.

Conclusion

First- and last-mile trips are critical components of public transport journeys, as barriers can result in a journey not being taken. While studies have examined ToD access and gentrification, there is limited work in examining perceived access by different socio-economic groups. The present study contributes to this knowledge gap by examining the perceived accessibility of first-mile trips to ToDs by women, compared with men. Factors are explored from an equity lens by examining the intersectionality of gender and age in first-mile trip mode choice.

Data was collected in Auckland, New Zealand. Responses from 445 current public transport users were analyzed using EFA to determine PAC scores, considering walking and use of car as first-mile mode choices. The findings contribute by providing insights into influential factors for perceived access to ToD stations by women from different age groups. The key findings are:

Women who walk find their first-mile trips to stations more walkable and easier to access to stations than men who walk.

Men who use cars find more perceived opportunities, more ease in accessing stations, and preferred activities of interest than women who use cars.

High-intensity land use (higher LUDI value) positively influences women’s perception of ease of accessing stations, while it negatively influences their perception of safety after dark.

During the daytime, women feel the safest in high-LUDI-value sites, while men feel safest in low-LUDI-value sites.

During the nighttime, when considering the mode choice, women feel safer walking the first mile, compared with using a car, while men feel safer in using the car compared with walking.

Results from the study can assist public transport authorities in the planning and design of ToDs to be more inclusive and beneficial for women, particularly vulnerable groups. This study underscores the significance of addressing gendered differences in accessibility in first-mile trips to ToDs. The research highlights that catering to women’s needs requires more than measures such as dedicated train compartments or CCTV cameras. It calls for a comprehensive approach for inclusive transport policy and planning strategies that includes collecting detailed travel data to better understand the distinct travel patterns of women. It is expected that this study would contribute to inform transport policymakers on gender-based differences to improve accessibility of ToD stations. The study advocates for gender-inclusive ToD policies that consider an equity lens by making stations more accessible for women.

Footnotes

Author Contributions

The authors confirm contribution to the paper as follows: study conception and design: Suraweera, A; Chowdhury, S.; data collection: Chowdhury, S.; analysis and interpretation of results: Suraweera, A.; draft manuscript preparation (original): Suraweera, A; draft manuscript review and editing: Chowdhury, S. and Welch, T. All authors reviewed the results and approved the final version of the manuscript.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Subeh Chowdhury

Timothy Welch

References

Calthrope

The New American Metropolis. Princeton Architectural Press, New York, NY, 1993.

Institute for Transportation and Development Policy. TOD Standard 3.0. ITDP, New York, NY, 2017.

Karner

Bills

Golub

Emerging Perspectives on Transportation Justice. Transportation Research Part D: Transport and Environment, Vol. 116, 2023, p. 103618.

Patnala

P. K.

Parida

Chalumuri

R. S.

Gender Differentials in Travel Behavior Among TOD Neighborhoods: Contributions of Built Environment and Residential Self-selection. Travel Behaviour and Society, Vol. 31, 2023, pp. 333–348.

Park

Farb

Chen

First-/Last-Mile Experience Matters: The Influence of the Built Environment on Satisfaction and Loyalty Among Public Transit Riders. Transport Policy, Vol. 112, 2021, pp. 32–42.

Liao

Shibata

Ishii

Koohsari

M. J.

Inoue

Oka

Can Neighborhood Design Support Walking? Cross-sectional and Prospective Findings from Japan. Journal of Transport and Health, Vol. 11, 2018, pp. 73–79.

Chowdhury

Patel

Giridharan

Ceccato

Formation of Fear and Adaptive Behavior in Young Ethnic Minority Women Riding Public Transport. Transportation Research Record: Journal of the Transportation Research Board, 2023. 2678: 687–697.

Souza

A. C. S.

Bittencourt

Taco

P. W. G.

Women’s Perspective in Pedestrian Mobility Planning: The Case of Brasília. Transportation Research Procedia, Vol. 33, 2018, pp. 131–138.

Bardaka

Transit-Induced Gentrification and Displacement: Future Directions in Research and Practice. Transport Reviews, Vol. 44, No. 3, 2023, pp. 567–571.

10.

Bergman

Gliebe

Strathman

Modeling Access Mode Choice for Inter-Suburban Commuter Rail. Journal of Public Transportation, Vol. 14, No. 4, 2011, pp. 23–42.

11.

Cervera

Walk-and-Ride: Factors Influencing Pedestrian Access to Transit. Journal of Public Transportation, Vol. 3, No. 4, 2001, pp. 1–23.

12.

Meng

Koh

P. P.

Wong

Y. D.

Influence of Socio-Demography and Operating Streetscape on Last-Mile Mode Choice. Journal of Public Transportation, Vol. 19, No. 2, 2016, pp. 38–54.

13.

Givoni

Rietveld

The Access Journey to the Railway Station and Its Role in Passengers’ Satisfaction with Rail Travel. Transport Policy, Vol. 14, No. 5, 2007, pp. 357–365.

14.

Huang

Moudon

A. V.

Zhou

Stewart

O. T.

Saelens

B. E.

Light Rail Leads to More Walking Around Station Areas. Journal of Transport and Health, Vol. 6, 2017, pp. 201–208.

15.

Pulugurtha

S. S.

Vasudevan

Nambisan

S. S.

Dangeti

M. R.

Evaluating Effectiveness of Infrastructure-Based Countermeasures for Pedestrian Safety. Transport Research Record: Journal of the Transportation Research Board, 2012. 2299: 100–109.

16.

Dunckel-Graglia

Brook

S. S.

Women-Only Transportation: How “Pink” Public Transportation Changes Public Perception of Women’s Mobility. Journal of Public Transportation, Vol. 16, No. 2, 2013, pp. 85–105.

17.

Alam

M. M.

Kurshitavili

Gonzalez

K. D.

Carvajal

K. G.

Baruah

Is a Mile for One a Mile for All? A Knowledge Synthesis Report on Gender and Mobility. World Bank, Washington, D.C., 2022.

18.

Chowdhury

Role of Gender in the Ridership of Public Transport Routes Involving Transfers. Transportation Research Record: Journal of the Transportation Research Board, 2019. 2673: 855–863.

19.

Jubainville

H. D.

Vanier

Women’s Avoidance Behaviours in Public Transport in the Ile-de-France Region. Crime Prevention and Community Safety, Vol. 19, No. 3, 2017, pp. 183–198.

20.

Parker

K. M.

Rice

Gustat

Ruley

Spriggs

Johnson

Effect of Bike Lane Infrastructure Improvements on Ridership in One New Orleans Neighborhood. Annuals of Behavioral Medicine, Vol. 45, No. 1, 2013, pp. 101–107.

21.

Milan

B. F.

Creutzig

Lifting Peripheral Fortunes: Upgrading Transit Improves Spatial, Income and Gender Equity in Medellin. Cities: The International Journal of Urban Policy and Planning, Vol. 70, 2017, pp. 122–134.

22.

The City of Chicago. Chicago EToD Policy Plan. The City of Chicago, IL, 2021.

23.

Currie

Stanley

Investigating Links Between Social Capital and Public Transport. Transport Reviews, Vol. 28, No. 4, 2008, pp. 529–547.

24.

Lattman

Friman

Olsson

L. E.

Perceived Accessibility of Public Transport as a Potential Indicator of Social Inclusion. Social Inclusion, Vol. 4, No. 3, 2016, pp. 36–45.

25.

Farrington

, Rural Accessibility, Social Inclusion and Social Justice: Towards Conceptualisation. Journal of Transport Geography, Vol. 13, No. 1, 2005, pp. 1–12.

26.

Morris

J. M.

Dumble

P. L.

Wigan

M. R.

Accessibility Indicators for Transport Planning. Transportation Research Part A: General, Vol. 13, No. 2, 1979, pp. 91–109.

27.

Maman

S. A.

Nicholas

A Composite Index of Public Transit Accessibility. Journal of Public Transportation, Vol. 14, No. 2, 2011, pp. 69–87.

28.

van der Vlugt

A.-L.

Curl

Wittowsky

What About the People? Developing Measures of Perceived Accessibility from Case Studies in Germany and the UK. Applied Mobilities, Vol. 4, No. 2, 2019, pp. 142–162.

29.

Pot

F. J.

Wee

B. V.

Tillema

Perceived Accessibility: What It Is and Why It Differs from Calculated Accessibility Measures Based on Spatial Data. Journal of Transport Geography, Vol. 94, 2021, p. 103090.

30.

Chau

H.-W.

Gaisie

Jamei

Chan

Lattman

Perceived Accessibility: Impact of Social Factors and Travel Modes in Melbourne’s West. Applied Sciences, Vol. 14, No. 15, 2024, p. 6399.

31.

Scheepers

C. E.

Wendel-Vos

G. C.

van Kempen

E. E.

de Hollander

van Wijnen

H. J.

Maas

Den Hertog

F. R.

, et al. Perceived Accessibility Is an Important Factor in Transport Choice—Results from the AVENUE Project. Journal of Transport & Health, Vol. 3, No. 1, 2016, pp. 96–106.

32.

Lattman

Olsson

L. E.

Friman

A New Approach to Accessibility – Examining Perceived Accessibility in Contrast to Objectively Measured Accessibility in Daily Travel. Research in Transportation Economics, Vol. 69, 2018, pp. 501–511.

33.

Olsson

L. E.

Friman

Lattman

Accessibility Barriers and Perceived Accessibility: Implications for Public Transport. Urban Science, Vol. 5, No. 3, 2021, p. 63.

34.

Auckland Council. GeoMaps Mapping Service. Auckland Council. https://www.aucklandcouncil.govt.nz/geospatial/geomaps/Pages/default.aspx. Accessed January 4, 2024.

35.

Stats NZ- Tatauranga Aotearoa. 2018Census Place Summaries. Stats NZ- Tatauranga Aotearoa. https://www.stats.govt.nz/tools/2018-census-place-summaries/. Accessed January 5, 2024.

36.

Inland Revenue Department – New Zealand. Tax rates for Individuals. Inland Revenue Department – New Zealand. https://www.ird.govt.nz/income-tax/income-tax-for-individuals/tax-codes-and-tax-rates-for-individuals/tax-rates-for-individuals. Accessed February 2, 2024.

37.

Work and Income. Community Services Card. Work and Income. https://www.workandincome.govt.nz/products/a-z-benefits/community-services-card.html. Accessed February 2, 2024.

38.

Bobbitt

. Simpson’s Diversity Index: Definition & Examples. Statology. https://www.statology.org/simpsons-diversity-index/. Accessed February 5, 2024.

39.

Takayama

Watanabe

Chang

H. C.

Morimoto

Study on the Characteristics of Japanese Transit Oriented Development as Seen from Long-term Land-use Changes. Transportation Research Procedia, Vol. 48, 2020, pp. 2313–2328.

40.

Bryne

B. M.

Structural Equation Modeling with Mplus, 1st ed. Routledge, New York, NY, 2011.

41.

Beyazit

Waite

I. A.

Balik

Erturan

Keşan

Improving Women’s Accessibility to Public Transport Through Participatory Street Experiments: The Case of Maltepe, Istanbul. Journal of Urban Mobility, Vol. 4, 2023, p. 100062.

42.

New Zealand Police. Crime Snapshot-2024. https://www.police.govt.nz/crime-snapshot. Accessed June 2, 2024.

43.

Boschmann

E. E.

Brady

S. A.

Travel Behaviors, Sustainable Mobility, and Transit-Oriented Developments: A Travel Counts Analysis of Older Adults in the Denver, Colorado Metropolitan Area. Journal of Transport Geography, Vol. 33, 2013, pp. 1–11.