Investigating Women’s Preferences for Urban Green Spaces by Using the Fuzzy-Set Quantitative Comparative Analysis Method: The Case of Chengdu, China

Abstract

Urban green spaces (UGSs) for women align with the United Nations Sustainable Development Goals and contribute positively to women’s well-being. However, research on women’s preferences for green spaces is lacking. This study uses fuzzy-set qualitative comparative analysis to explore the causal complexity of multiple built environment characteristics that drive women’s preferences for UGS at different times. Results show that no single characteristic of the built environment is necessary and sufficient to determine women’s preferred UGS. Six configurational paths are identified as driving female preferences, among which time, security, and public service facilities have the strongest impact. This study aims to offer novel research perspectives and methodological support for the development of inclusive and sustainable UGS.

1. Introduction

1.1. Background

Urban green spaces (UGSs) are crucial elements of urban landscapes. They are defined as natural open spaces containing trees within built-up areas including parks, street trees, urban forests, residential lawns, and rooftop gardens [1]. In densely populated cities, UGSs facilitate interactions between people and the natural environment, serving as vital amenities that provide substantial physical and mental health benefits to urban residents navigating crowded and stressful environments [2,3,4]. Accordingly, improving the quality of UGS planning and design is crucial to enhance human well-being and achieve sustainable urban development, especially as human contact with nature diminishes [5,6].

The United Nation’s Sustainable Development Goal (SDG) 11.7 emphasizes the importance of providing safe, inclusive, accessible, and green urban spaces universally by 2030. Focusing particularly on women, children, the elderly, and persons with disabilities, this SDG calls for the creation of safe, resilient, and inclusive cities. It also affirms that women should have equal rights to engage in activities in their chosen UGS to enhance their well-being [7].

For more than a century, researchers from diverse disciplines have collaborated to articulate the importance of considering women’s concerns in urban space design to promote the equitable and inclusive development of cities and societies [8]. From the birth of the concept of “care-oriented urbanism” in the 1860s to the 1920s, women developed a broader definition of public infrastructure, advocating for the integration of “care-oriented urbanism” into sustainable development. In her 1980 article, “What Would a Non-sexist City Be Like?” Dolores Hayden described such a city as one that would transcend the traditional definitions of home, neighborhood, and workplace. Gradually, the notion took hold that a city supportive of women would be one where the separation between households is more relaxed, work is less rigid, domestic labor is recognized as a productive activity, and the care of children, the elderly, and the disabled is integrated into economic life [9]. The goal of helping women gain equal access to UGS is to provide all inhabitants—regardless of social distinction—equal opportunities and the urban qualities necessary for full human development [10].

Despite the socioeconomic changes in recent decades, the achievements of women in the labor force, the significant advances in gender equality, and the fact that the majority of female users no longer feel deeply disadvantaged in public spaces, the inequalities shaped by family and social identities persist and cannot be ignored. Women, relative to their proportion in society, are less likely to use UGS [11,12]. However, compared with other groups, housewives benefit especially from UGS in residential areas [13]. Owing to their triple role of producing, reproducing, and managing community, women usually have a more complex relationship with the built environment and are often a vulnerable group, even if their familial and social identities are responsibilities they willingly take on [14]. Given this complexity, it makes sense to perform an invisible assessment of women’s preferences. We attach great importance to ecofeminism and extract scientific ideas from its values [15,16]. However, our focus on gender concepts aims to understand women’s preferences and needs for design improvement rather than framing male and female users in opposition or critiquing societal oppression. The female participants in our study expressed hopes that green spaces would also appeal to their fathers, husbands, and sons. Therefore, we argue that UGS designed with women’s preferences in mind should not only meet the demands of women’s living spaces but also promote a balance between their lives and work. Additionally, domestic activities carried out in green spaces should be recognized as productive work. Through optimized spatial design, the burden of domestic work on women can be alleviated, and men can be encouraged to participate in these activities within UGS.

Feminist geographers have demonstrated how urban space influences women’s health and have revealed the ways that women’s limited access to urban space may restrict their access to knowledge, power, and resources [17]. However, the existing vision of urban space is essentially homogeneous, ignoring gender differences in green space demands [18]. This approach to non-discriminatory gender equality does not fully recognize and respect the differences in gender preferences in the use of space [19]. Research indicates that UGS can be designed to foster more harmonious and inclusive cities by incorporating women’s preferences. For instance, Vienna has successfully addressed the reduced use of green spaces by girls by dividing the existing UGS into smaller spaces, mitigating the competitive dynamics between girls and boys.

The majority of previous research on women’s preferences has focused on gender differences in individual perceptual factors from the perspective of female users. However, these studies often overlooked the interconnected nature of various factors within the built environment and how they jointly influence user choices at different times. Consequently, a mismatch emerges between user perceptions and the actual complexity of the built environment. We use fuzzy-set qualitative comparative analysis (fsQCA) to address this research gap.

1.2. Application of the Fsqca Method

Qualitative comparative analysis (QCA) is an asymmetric data analysis technique that leverages both qualitative and quantitative methods, analyzing data “by case” rather than “by variable” [20]. It combines qualitative inductive reasoning with quantitative empirical testing to help identify necessary and sufficient conditions for specific outcomes. The fsQCA method integrates fuzzy sets and fuzzy logic with QCA principles, allowing for variables to be assessed across a continuous spectrum from 0 to 1 [21,22]. It considers the combined effects of all variables and reveals a variety of potential causal relationships for exploring how six built environment characteristics—area, surrounding land use, accessibility, usability, security, and public service facilities (PSFs)—and time influence women’s preferences for UGS.

1.3. Aims and Objectives

The main research question of this study is as follows: what types of UGSs do women prefer? To explore this main question, we use two sub-questions: (1) Is there a single determinant among the built environment features of women-preferred UGS that significantly influences women’s preferences for UGS? (2) When combined, which built environment features of UGS are most favored?

Following this introduction, Section 2 details the construction of the indicator system used in this study. Section 3 describes an empirical study that applies fsQCA to analyze data on women’s preferences for UGS, including a comprehensive documentation of data collection and calculation methods. Section 4 presents the results, identifying the necessary and sufficient conditions for women’s preferences for UGS and addressing sub-questions 1 and 2. Lastly, Section 5 and Section 6 discuss the results and summarize this study, respectively.

2. Construction of the Indicator System

This research does not aim to identify new built environment characteristics of UGS affecting women’s preferences but instead uses empirical evidence to support prior findings. Including UGS characteristics unrelated to women’s preferences increases the case count required for fsQCA and complicates the results, hindering their operationalization. The impacts of time and various built environment characteristics on the gender-differentiated use of UGS have been demonstrated in previous research (

Table 1. Built environment characteristics affecting women’s preference for UGS.

Component	Sources	Description
Time	(Chidambaram & Scheiner, 2021 [23]; Esping-Andersen, 2009 [24]; Fan et al., 2011 [25]; Scheiner, 2016 [26]; Shen et al., 2018 [27]; Silm et al., 2013 [28])	Owing to large-scale urbanization and transportation development, urban residents now spend most of their time outside their residential areas [23]. Advancements in market goods and domestic services have reduced the time required for family caregiving. Younger urban middle- and upper-class women are moving away from traditional expectations of managing household duties. However, they still undertake more domestic chores and have less recreation time than men. They often face competing demands from home, social relationships, and work. Consequently, family activities tend to be spatially and temporally fixed. These activities act as “anchors” for scheduling other activities [24,25,26]. Additionally, childcare duties often include taking children to green space. This may increase women’s demand for green spaces more than men [29].
Area	(Garcia-Ramon et al., 2004 [30]; Liang et al., 2024 [31])	The group activities that women like to perform usually have a higher demand for urban green spaces (UGSs) than those preferred by men, so the size of the green space affects the choice of female users [30,31].
Surrounding land use	(Bolte et al., 2019 [32]; Mushkani & Ono, 2021 [33]; Pitarch-Garrido, 2018 [34])	The surrounding land use of UGS impacts people’s lifestyles and well-being and influences public space configurations [33]. Studies show that UGSs near residential areas are typically more vibrant [32,34]. Specifically, housewives benefit more from UGS in residential areas than other groups. This finding indicates that land use around UGS significantly impacts the presence of women.
Accessibility	(Burgess, 2008 [35]; Chen et al., 2024 [36]; Sadeghi & Jangjoo, 2022 [17])	Women are more likely to engage in caregiving and domestic responsibilities, and thus exhibit more complex daily mobility patterns. Generally, UGS designs do not accommodate these factors but are based instead on simplistic “home–work–home” daily trajectories. Consequently, current studies on UGS accessibility do not adequately capture women’s typical travel patterns. Examples of these patterns include “home–school–grocery store–home–work–school–home” [35,37]. This travel mode influences women’s preferences for UGS through public transportation and various street access routes.
Usability	(Hornblower et al., 2012 [38]; Wendel et al., 2012 [39])	Usability of green spaces is defined as the ease with which physical visits and activities can be initiated within a space [38]. Some studies use the area of a UGS to assess space availability. However, practical usage issues arise when parts of the green space are too steep or densely planted. These conditions make it difficult for women to navigate and fully utilize these areas [39].
Security	(Navarrete-Hernandez et al., 2021 [40]; Sadeghi & Jangjoo, 2022 [17]; Yavuz & Welch, 2010 [41])	Studies on the security of green spaces reveal that women generally feel less safe than men in urban spaces [41]. According to the vulnerability hypothesis, perceptions of security are shaped by an individual’s sense of personal vulnerability. Compared to men, women’s greater physical and social vulnerability contributes to their heightened perception of risk in the built environment. They feel more unsafe than men do in dark, deteriorating, or isolated public spaces. This feeling is particularly pronounced in areas showing signs of uncivil behavior, such as graffiti [17,40]. Women tend to avoid these areas.
Public Service facilities	(Bocarro et al., 2015 [42]; Sadeghi & Jangjoo, 2022 [17])	Public service facilities (PSFs) in UGS are predominantly linked with physical activities among males. By contrast, women’s physical activities often correlate with the presence of children and caregiving roles [42]. This reality influences their use of these facilities more for active transportation than for direct usage. Additionally, when caring for children, women often require additional PSFs, including improved resting areas, mother and baby rooms, restrooms, and shaded spaces. All these facilities influence women’s preferences for UGS.

).

3. Methodology

3.1. Research Methods

We choose fsQCA because of its three main strengths. First, fsQCA does not treat each antecedent variable as acting independently. Instead, it considers the combined effects of all variables. This provides a holistic view of how various conditions collectively influence the outcome [20]. Second, fsQCA can identify multiple pathways from antecedent variables to the same outcome, revealing various potential causal routes. Finally, unlike traditional methods that handle only symmetric relationships, fsQCA adeptly manages both the presence and absence of conditions. This feature accommodates the asymmetry inherent in antecedent variables.

3.2. Study Area

Our study area is a continuous, walkable 5 km section of UGS along the Jinjiang River in Chengdu. This section’s predecessor, the Jinjiang River Renovation Project, received the United Nations Habitat for Humanity Award in 1998. Chengdu, a mega-city in southwest China, has a population exceeding 20 million and has been one of the fastest-growing cities in the country over the last two decades. The Jinjiang River, downstream of the Dujiangyan Irrigation Project, is a tributary of the Yangtze River. In October 2017, Chengdu began building high-quality UGS along Jinjiang River. Residents engage in a diverse array of social activities within these green spaces. Chengdu is known for its eclectic, harmonious, and open cultural atmosphere. The literature suggests that in more egalitarian cultures, gender differences in preferences may become more pronounced as women gain greater freedom of choice [43,44]. In Chengdu, female users can choose UGS based on their preferences, contributing to the authenticity of the data.

We selected a continuous, walkable UGS along the Jinjiang River as our study area, spanning 5 km in the urban center of Chengdu. To ensure data processability, we segmented the UGS into 15 sections. These segments are delineated by urban arterial roads, with lengths ranging from 400 m to 850 m and areas between 5000 and 25,000 m². They have a service radius of 500 m, based on an assumed pedestrian speed of 4 km/h [31]. The continuous, walkable UGSs we chose minimize variance in the backgrounds and habits of female users. Offering a diverse selection of UGSs within their daily activities helps us better identify their preferences.

Originally, we attempted to identify women’s preferences directly through a questionnaire, but the results we obtained did not reflect their real thoughts well because many respondents had difficulties choosing and describing their real preferences through the questions. Therefore, we ended up only counting the demographic characteristics of the female users who answered the questionnaire, which facilitated the analysis of the invisible assessment at a later stage. A total of 202 questionnaires were distributed to female users of UGS in the study area to obtain information on their age, occupational background, and usage habits to better analyze the reasons for the emergence of women’s preferences for green spaces (Figure 1). Combined with the analysis of the identity background characteristics of the surveyed women, the results of this study basically cover female groups of all ages and social identity backgrounds, and in terms of the proportion of the number of people in these groups who use green spaces, the results of the study are more inclined to reflect the overall preference for green spaces of enterprise workers and freelancers aged 25–59 and with a lower level of education. Using the questionnaire data, a study of the correlation between frequency of green space use and peer companions reveals that women who use green space more frequently tend to travel alone or with children, that these women use green space twice a week or more, and that about 35% of female caregivers spend more than an hour in green space, which is much more than the average length of time. This suggests that caregiving activities influence women’s demand for green space use in our study area.

3.3. Data Collection and Analysis

In this study, the terms “accessibility”, “usability”, and “security” are neutral. We did not attach the absolute positive connotations often found in qualitative research. Overprovision of UGS features can lead to conflicts from various urban development perspectives, and excessive accessibility may complicate user demographics. Too much usability might reduce the actual green space available, while overemphasizing security could diminish user privacy. For instance, an overly secure area, resembling a lounge space filled with surveillance cameras, might deter visitors.

The data collection was conducted in two phases using diverse methods (

Table 2. Data collection details and procedures.

Phases	Methods and Activities	Date and Time	Objectives
First phase	Unobtrusive observation: watching, counting, photographing	Six days with good weather conditions in mid-October 2023: three weekdays (11, 19, and 26 October) during the day (09:00–18:00) and at night (19:00–22:00), and on three weekends (3, 14, and 22 October) during both day and night.	Three observers were assigned to count the number of women carrying out activities or resting in the 15 UGSs at the same time.
Second phase	Measurement fsQCA	From October 2023 to December 2023. Data processing for fsQCA began in December 2023.	Identifying the combination of UGS characteristics that affect the presence of women.

). The first phase involved unobtrusive behavioral observation in the designated area. In the second phase, field observations and measurements were complemented by data extracted from city maps.

(1)

Women’s preferences for UGS

In our fsQCA model, the outcome is women’s preferences for UGS. We hypothesized that the support of UGS for women’s activities, evidenced by the proportion of women visitors, determines its alignment with their preferences. A higher proportion of women visitors in UGSs suggests that women have a more positive experience in those spaces. From the 12 observation sessions, we calculated the average percentage of women visitors and drew the maps (Figure 2).

(2)

Time

According to heatmaps from online platforms and empirical knowledge of the study area, pedestrian traffic in the green space peaks during weekend nights. This is followed by weekday nights and then weekend days. The lowest traffic occurs on weekdays during the daytime. Therefore, this study quantified the temporal variable by assigning scores to each time period: weekend nights = 4, weekday nights = 3, weekend daytime = 2, and weekday daytime = 1.

(3)

Area

Research has shown that men participate significantly more in sports and activity spaces, while women tend to use plazas, walking paths, and children’s play areas more frequently [45]. Their activity preferences lead to varying demands for UGS [45]. Using roads as boundaries, the area of the 15 UGSs was calculated based on data extracted from publicly available online maps.

(4)

Surrounding land use

In the study area, most of the UGSs near residential areas are surrounded by old six-story residential buildings, and only a few green spaces are adjacent to newly built high-rise buildings. Unlike studies in the United States and European suburbs, the overall population density in our study area is high, and the people who use the green spaces are older. The buildings around the UGS in the non-residential area are mostly government or corporate office buildings and some high-rise hotels, and a few green spaces are adjacent to schools, hospitals, and other buildings. We quantified the residential area (${SR}_i)$ within a 100 m buffer zone surrounding each UGS and then assessed the current land use status (${LU}_i$). A high ${LU}_i$ value indicates a large proportion of residential area adjacent to the UGS. Conversely, a low ${LU}_i$ value signifies a small proportion of residential area around the UGS.

$${LU}_i={\displaystyle \frac{{SR}_i}{{TS}_i}}$$

• $L{U}_i=$ surrounding land use of $i_{\mathrm{th}}$ UGS;
• $S{R}_i=$ the area of surrounding residences of $i_{\mathrm{th}}$ UGS;
• $T{S}_i=$ total construction area of surrounding buildings of $i_{\mathrm{th}}$ UGS.

(5)

Accessibility

In an empirical study of green spaces in Beijing, China, results showed significant gender differences in access to green spaces in daily travel, with men overall being more mobile in the city and having greater advantage in accessing green spaces. Although women had greater access to green spaces in terms of overall travel time, they were more restricted to selected routes, with age, occupation, household type, and mode of travel (e.g., car, bus, or walking) all being important factors intertwined with gender [46]. Another study on the commuting patterns of suburban residents in Beijing and Shanghai observed that women were less likely to commute by car, relying instead on public transportation more frequently than men [46]. Therefore, the present study focused on assessing the walking accessibility and public transportation accessibility of UGS [47].

The green spaces selected for this study are continuous, so using conventional community-level accessibility study methods (e.g., AG2SFCA) to assess each UGS would cause many study areas to overlap within the 500 m walking service radius. This overlap made it impossible to realistically assess the current state of accessibility for the UGS sections. On the basis of the field study and public city maps, we focused solely on the number of paths directly connecting to the UGS sections and the number of public transportation stations (bus and subway) within 500 m. The two indicators reflect the friendliness to different modes of transportation and together affect the accessibility of green spaces. Accordingly, we regarded the weight of both as the same in the calculation. The higher the value of ${AB}_i$, the more accessible the UGS by walking or public transportation. Conversely, the lower the value of ${AB}_i$, the less accessible the UGS. The field study indicated that users of UGS with higher accessibility values are generally more mobile, whereas users of UGS with lower accessibility values tend to be more stationary.

$${AB}_i=0.5 {\displaystyle \frac{{NR}_i}{L_i}}+0.5{\displaystyle \frac{{NP}_i}{L_i}}$$

• ${AB}_i=$ accessibility of $i_{\mathrm{th}}$ UGS;
• ${NR}_i=$ number of roads directly connected to $i_{\mathrm{th}}$ UGS;
• ${NP}_i=$ number of public transportation stations within 500 m of $i_{\mathrm{th}}$ UGS;
• $L_i=$ length of the $i_{\mathrm{th}}$ UGS boundary.

(6)

Usability

We adopted a framework developed by other scholars for assessing space usability. This framework divides the assessment into three steps: space enclosure calculation, space availability calculation, and weighted usability calculation [48]. Space enclosure is evaluated based on FOV values, which are defined by the UGS boundary and the height of the visitor’s visual range. This evaluation helps determine whether the UGS obstructs the visual communication of users. During data collection, we used the width of the walking path as the distance from the obstacle. A buffer zone 2 m wide was established around each UGS to extract obstacle data. Space availability determined if the UGS has sufficient usable area for activities. Usable area, which requires a gentle slope suitable for human activities, was calculated as the area not occupied by natural greenery and man-made facilities [48]. Through field research, we measured the proportion of space within the UGS that was shaded by natural greenery and the area occupied by both natural and man-made facilities. The higher the value of ${UB}_{\mathrm{i}}$, the more usable area the UGS has. Conversely, the lower the value of ${UB}_{\mathrm{i}}$, the more active space in this UGS is occupied by natural greenery, man-made facilities, and flood slope.

(7)

Security

Users’ perception of security is typically influenced by a combination of factors, including the security of the built environment itself, as well as the gender, age, health status, culture, and socioeconomic factors of the users [49]. Therefore, the security of the UGS built environment should not be equated with the perceived safety of users. To more accurately assess the security of each UGS, we refined the Weighted Aggregated Aum Product Assessment method using a previously proposed approach [50]. The weight distribution of the 10 selected indicators was determined (

Table 3. Mapping between UGS security factor groups and UGS security assessment criteria.

UGS Security Assessment Criteria	Unit	Weight	Optimum	Data Source
Monitor	(Number)	0.10	+	Observations on site
Impenetrable barriers along the pathways/length of all pathways	(%)	0.10	−	Observations on site
Illuminated pathways	(%)	0.14	+	Observations on site
Entrances	(Number)	0.07	+	Publicly available city maps
Attractive structures	(Number/Ha)	0.14	+	Observations on site
Maintenance	(10 points scales)	0.14	+	Observations on site
Road pavement	(%)	0.07	+	Observations on site
Informative and clear signage	(10-point scales)	0.03	+	Observations on site
Real estate price in the surroundings	(RMB per square meter)	0.07	+	Open data mapped with GIS data
Uncivilized behavior	(Number/1 h)	0.14	−	Observations on site

) based on the weighted results of urban park safety assessment factors provided by an expert group [50]. Next, a decision matrix was constructed to normalize the data for subsequent calculations. The additive relative importance and multiplicative relative importance were determined to calculate the comprehensive evaluation index of key factors.

The security assessment results indicate that a higher ${ST}_i$ value corresponds to a more open UGS, characterized by greater human construction, supervision, and maintenance, which collectively enhance security. Conversely, a lower ${ST}_i$ value indicates a more closed UGS, associated with increased privacy, reduced human intervention, a more natural and primitive state, and decreased security.

(8)

Public service facilities

The capability of UGS to supply the activity needs of more users can be assessed by combining the types of PSFs with their current utilization. The construction status of PSFs in the selected UGS is evaluated using a grading system. A high PSF grade indicates that the UGS can better meet the diverse activity needs of users. Public services for mothers and carers are highlighted because they are more representative of this group of women and because of the large gender difference in public services for mothers and carers compared to men.

If the PSF rating of a UGS is 1, it means the public facilities of the UGS only include resting facilities, such as chairs for users.

If the rating is 2, in addition to the resting facilities, the UGS also provide users with basic amenities such as restrooms and mother and baby rooms.

If the rating is 3, in addition to the above PSF, the UGS provides users with exercise equipment, children’s play equipment, and other activity equipment.

If the rating is 4, in addition to the above PSF, the UGS has high-quality landscape architectural features.

3.4. Data Calibration

The physical characteristics of the built environment of the 15 UGS segments were quantified through a systematic and scientifically based assessment methodology (

Table 4. Assessment results for the 15 segments of UGS.

Alternatives	$\mathit{A}{\mathit{R}}_{\mathit{i}}$	$\mathit{L}{\mathit{U}}_{\mathit{i}}$	$\mathit{A}{\mathit{B}}_{\mathit{i}}$	$\mathit{U}{\mathit{B}}_{\mathit{i}}$	$\mathit{S}{\mathit{T}}_{\mathit{i}}$	$\mathit{P}\mathit{S}{\mathit{F}}_{\mathit{i}}$
S1	25,200	0.4	13.0952	0.9017	0.6321	4
S2	6760	0.45	9.4675	0.2721	0.5883	2
S3	8700	0.75	13.7931	0.5519	0.5622	1
S4	13,225	0.7	10.4348	0.3826	0.5864	4
S5	16,560	0.3	9.7826	0.7092	0.5715	2
S6	5600	0.8	10.0000	0.2323	0.5662	3
S7	5000	0.35	10.0000	0.3083	0.4852	3
S8	23,275	0.75	12.0301	0.722	0.6145	3
S9	13,300	0.45	7.5188	0.497	0.4486	3
S10	6960	0.9	12.069	0.4243	0.5997	2
S11	8700	0.7	10.3448	0.5398	0.5864	1
S12	15,000	0.95	10.0000	0.6585	0.5464	3
S13	11,000	0.85	10.0000	0.6003	0.6046	2
S14	11,000	0.95	9.0909	0.5327	0.6349	4
S15	6600	0.9	9.0909	0.4151	0.5829	4

). Consistent with existing studies, this research used the direct calibration method [20] to convert the data into fuzzy-set affiliation scores. Using the direct method, we set calibration thresholds at 95%, 50%, and 5% to define the “fully in”, “cross-over”, and “fully out” states, respectively (

Table 5. Calibration anchors of each fuzzy set.

Indicator	Variable	Description	95% Fully In	50% Cross-Over	5% Fully Out
WP	Women’s presence	The proportion of women in UGS	0.78	0.5	0.32
TI	Time	Time at observation	4	2.5	1
AR	Area	The area of UGS	25,200	11,000	5000
LU	Surrounding land use	The proportion of surrounding residences	0.95	0.75	0.3
AB	Accessibility	Calculated from connecting roads and public transport stops	1	0.3955	0
UB	Useability	Calculated from the space enclosure and the space availability	0.9017	0.5327	0.2323
ST	Security	Calculated from 10 security estimation criteria	0.6349	0.5864	0.4486
PSF	Public service facilities	Classification of public service facilities	4	3	1

). Since cases with a 0.5 affiliation score are typically excluded from analysis, we adjusted this value to 0.501 to render more realistic conclusions [51].

4. Results

4.1. Necessity Analysis of Conditions

Referring to mainstream QCA research, this research first tested whether a single condition (including its non-sets) constitutes a necessary condition influencing women’s preferences. In fsQCA, a condition is considered necessary for an outcome if it consistently exists whenever the outcome occurs [20]. Consistency is an important measure in determining a necessary condition. A condition is deemed necessary for an outcome if the level of consistency exceeds 0.9 [52]. The results of the necessary condition test, which assesses women’s preferences using fsQCA 3.0 software, were analyzed. Clearly, the consistency level of all the conditions does not exceed 0.9. Therefore, among the seven conditions, there are no necessary conditions that influence women’s preferences, addressing question 1 (

Table 6. Analysis of necessary conditions.

Outcome Variable: WP	Built Environment Description	Consistency	Coverage
TI	Higher vitality (night)	0.730185	0.735467
~TI	Lower vitality (daytime)	0.519675	0.523434
AR	Large area	0.615614	0.715241
~AR	Small area	0.682894	0.607055
SLU	Surrounded by many residential areas	0.638283	0.644447
~SLU	Surrounded by few residential areas	0.660456	0.663640
AB	More accessible, with highly mobile users	0.716153	0.674227
~AB	Less accessible, with more regular users	0.628686	0.680798
UB	More usable areas but less natural greenery	0.633187	0.703316
~UB	Fewer usable areas but more natural greenery	0.725883	0.668801
ST	More open, with more human construction, supervision, and maintenance	0.713108	0.687907
~ST	More secluded, with less human intervention and higher natural state	0.668465	0.704387
PSF	Many types and quantities of PSFs	0.630738	0.668737
~PSF	Few types and quantities of PSFs	0.654201	0.627555

Notes: “~” is a negation (“NOT”).

).

4.2. Sufficiency Analysis of Conditional Grouping

From a set-theoretic perspective, the analysis of groupings explores whether the set represented by multiple conditions is a subset of the outcomes. Previous research suggests that the consistency level for determining adequacy should be at least 0.75 [52]. However, it is also important to consider the distribution of cases in the truth table and the researcher’s familiarity with those cases. The truth table rows (groupings) with outcomes of 0 and 1 should both be represented and balanced [22]. The frequency threshold should also include at least 75% of observed cases. Having met these best practice criteria, this study established a frequency threshold of 1. It set the consistency threshold for sufficient conditions at 0.8 and required a PRI consistency exceeding 0.75.

As the seven conditions are not deemed necessary for identifying women’s preferences, no explicit counterfactual analysis was performed. Instead, this study focused on whether the presence or absence of these conditions leads to women’s preferences and identified intermediate solutions. The fsQCA 3.0 method generates three types of solutions, namely, complex, simple, and intermediate, with each varying in complexity. Aligning with existing research, this study reported the intermediate solution [51], complemented by the parsimonious solution.

Table 7. Analysis of sufficient conditions.

Configuration	Solution
	1a	1b	2a	2b	2c	3
Time	●	●	●	●	●
Area	⊗	•	●	●	●	●
Surrounding land use	●	●	⊗	⊗	⊗	⊗
Accessibility		•	•			●
Usability	⊗	•		•	⊗	⊗
Security	⊗	⊗	•		⊗	●
Public service facility	●	●	•		●	●
Raw coverage	0.2575	0.1784	0.2644	0.2141	0.2155	0.2278
Unique coverage	0.0755	0.0160	0.0744	0.0457	0.0060	0.0384
Consistency	0.9629	0.9540	0.9638	0.9700	0.9789	0.9773
Overall solution coverage	0.4927
Overall solution consistency	0.9532

Note: “●” = presence of core condition. “⊗” = absence of core condition. “•” = presence of peripheral condition. “” = absence of peripheral condition.

displays the results analyzing the seven conditionally formed groupings (paths) regarding women’s presence. According to the literature, results are depicted using solid circles to indicate the presence of a condition, forked circles for its absence, and empty spaces to denote an ambiguous state where the condition may be present or absent [51].

The consistency level of the six groupings listed in the table exceeds 0.95 for both individual and overall solutions. The overall solution coverage is at 0.49. The six groupings are considered sufficient combinations of conditions to supply women’s preferences for UGS. Our analysis of individual conditions revealed that time, area, and surrounding land use are core factors playing a significant role in influencing women’s preferences. The relationship between the conditions indicates a notable symmetry between the area of group 1a and the surrounding residential land use, as well as groups 2a, 2b, and 2c. This symmetry emphasizes that not all conditions necessarily promote women’s preferences for various UGS constructions. On the basis of the core factors and their explanatory logic, this research distilled the information into three combinations. These combinations address question 2 (Figure 3).

(1)

Pathways leading to women’s preferences for UGS around residential areas

According to the results of fsQCA, ~AR and SLU in path 1a (TI*~AR*SLU*~UB*~ST*PSF) are core conditions. They demonstrate the main characteristics of sections under this path. These sections are located in small green areas, each less than 10,000 m², on the periphery of residential areas. Most users are residents from nearby areas. TI is a core condition, suggesting strong time-based influences on women’s preferences for this type of UGS. ~UB and ~ST are also core conditions for this path. ~UB indicates that the UGSs feature more greenery and fewer paved plazas. ~UB does not imply that the UGSs are in a dangerous state but rather suggests that these areas are secluded, minimally disturbed by humans, more natural, and pristine. The UGSs create private and intimate spaces by obscuring views with greenery and dimming lights to enhance privacy. The core condition is PSF, indicating that the UGSs provide various public service facilities to supply user needs. This path has the second highest coverage (0.2575) and the highest unique coverage (0.0755) among all the groupings (Case 6-3, Case 6-4, Case 15-3, and Case 15-4).

In path 1b (TI*AR*SLU*AB*UB*~ST*PSF), TI as the core condition suggests a significant time influence on women’s preferences for the UGS. SLU and PSF, also core conditions, indicate that the surrounding area, similar to path 1a, is predominantly residential and has adequate PSFs. The core condition ~ST suggests that the built environment of the UGS is less safe, similar to path 1a. Peripheral conditions AR, AB, and UB indicate that the UGSs in this path are larger than those in path 1a, offering better access for walking or public transportation and ample land for activities. The path exhibits the lowest consistency (0.9540) and coverage (0.1784) among all groupings, encompassing two cases (Case 12-3 and Case 12-4).

(2)

Pathways leading to women’s preferences for UGS around non-residential areas

TI, AR, and ~SLU are core conditions in paths 2a (TI*AR*~SLU*AB*ST*PSF), 2b (TI*AR*~SLU*~AB*UB*~ST*~PSF), and 2c (TI*AR*~SLU*~AB*~UB*~ST*PSF). These paths show that UGSs around non-residential neighborhoods are larger than those at the residential edges. These UGSs typically range from 10,000 m² to 25,000 m² and are significantly influenced by time.

In path 2a, AB, ST, and PSF are peripheral conditions. They demonstrate that the UGS in this path caters to women’s preferences through enhanced accessibility, security, and provision of PSFs. Path 2a exhibits the highest coverage (0.2644) across four cases (Case 1-3, Case 1-4, Case 4-3, and Case 4-4).

In path 2b, UB, ~AB, ~ST, and ~PSF serve as peripheral conditions. The UGS of this path is relatively inaccessible but hosts a stable population. The space experiences minimal human intervention, retains a high degree of natural primitiveness, and offers limited diversity in PSFs. However, it meets women’s preferences by providing more available area. Path 2b exhibits a consistency of 0.9700 and a coverage of 0.2141, including only two cases (Case 5-3 and Case 5-4).

In path 2c, ~UB, ~ST, and PSF are the core conditions, while ~AB serves as the peripheral condition. The UGSs of this path are less accessible than those in Path 2b, yet the spaces have a stable user base, extensive green areas, and minimal human construction, supervision, and maintenance. Despite these, the UGSs meet women’s preferences by offering sufficient and diverse basic service facilities. Path 2c has the highest consistency among all groupings at 0.9789 and a coverage of 0.2155, including only two cases (Case 9-3 and Case 9-4).

(3)

Pathway leading to women’s preferences for UGS irrespective of time

AR, ~SLU, AB, ~UB, ST, and PSF under path 3 (AR*~SLU*AB*~UB*ST*PSF) are all core conditions. The independence from time conditions indicates that the UGSs under this path supply women’s preferences throughout the time period. The UGSs under this pathway are characterized by a lower percentage of surrounding residential areas. They typically feature high construction quality with good accessibility, security, and adequate access to all types of basic services. The consistency of this grouping (0.9773) is second only to grouping 5, with four cases covered (Case 4-1, Case 4-2, Case 4-3, and Case 4-4).

To test the robustness of the conclusions, we increased the consistency threshold to 0.85 while keeping all other conditions unchanged. The results show that the newly generated grouping aligns with the original grouping. By raising the PRI threshold from 0.75 to 0.8 and keeping all other conditions constant, we identified a significant subset relationship between the configurations of the new and original groupings, suggesting that the conclusions are relatively robust [51].

4.3. Validation Analysis Based on Case Interviews

Case interviews with female users carry inherent sampling bias limitations, as they primarily survey existing UGS users while potentially overlooking perspectives from non-users or infrequent users. Nonetheless, this methodology remains one of the most effective approaches to validate public preferences through direct user engagement.

In UGS adjacent to residential areas, female users prioritized proximity to residence and high naturalness to facilitate simple personal and family-oriented activities. These female respondents emphasized the synergistic effects of temporal availability, service facilities, and accessibility on their spatial experiences:

I prefer nearby green spaces where I can take leisurely walks and rest spontaneously. The lush vegetation creates a tranquil atmosphere…I frequently engage in family conversations here.

(Teacher, 46 years old, UGS #6)

I consistently visit in the evenings. The primary users are middle-aged and elderly residents…The expansive plaza facilitates group dancing. Although shaded by trees, I don’t perceive safety concerns.

(Office worker, 30 years old, UGS #12)

For green spaces in non-residential zones, female users emphasized temporal availability, safety considerations, and service facilities as critical determinants. Some transient female users appreciated the spatial openness and inclusivity that enable them to experience urban cultural ambiance. Several noted that dense vegetation and controlled accessibility create natural buffers against urban noise, facilitating uninterrupted recreation:

Retirees playing cards dominate daytime use, with litter accumulation peaking in the afternoon. Post-dinner, female users increase significantly—mothers with children and dance groups emerge. Men typically congregate near buildings or seating areas, while women prefer open plazas.

(Sanitation worker, 51 years old, UGS #1)

A homemaker remarked

“I visit weekly with my child due to the well-maintained playgrounds and clean restrooms. Our local green space lacks comparable facilities and peer interaction opportunities…I avoid shrub-dense areas due to insect proliferation.”

(Homemaker, 39 years old, Green Space #1)

A student noted

“The evening ambiance combines city night view with street performances. Although dimly lit, the space feels romantically secure against the metropolitan backdrop.”

(High school student,16 years old, Green Space #5)

Green space #4 maintained consistent female patronage despite its low security. Users emphasized that safety perceptions were activity- and time-dependent rather than spatially determined. The privacy provisions and comprehensive amenities of the space satisfied user needs:

“The dense vegetation provides ideal privacy for leisure…Evening use feels secure through group activities. The plaza’s facilities—including nursing stations and first-aid rooms—accommodate multigenerational care.”

(Respondent, 55 years old, Green Space #4)

In order to mitigate the Hawthorne effect, whereby people deliberately change their behavior or verbal expression for the better when they realize they are being watched or observed, the interviews were conducted in the local dialect, through respectful conversations with groups of different occupations and ages. Our findings confirmed that the six identified pathways reflect gendered spatial needs. Whether driven by professional obligations, personal interests, or caregiving responsibilities, users critically evaluated how built environment characteristics facilitate their activities. When encountering suboptimal conditions, users adopted avoidance and protective strategies rather than pursuing spatial modifications, demonstrating internalized negotiation processes with their environments.

5. Discussion

5.1. What Types of UGSs Do Women Prefer?

This study conducted implicit assessments of women’s preference levels and influencing factors across 15 UGSs through unobtrusive observations, unstructured conversational interviews, and on-site quantification of built environment characteristics by employing fsQCA. We deliberately avoided direct questionnaire surveys about preferences because users often struggle to precisely quantify their needs and perceptual differences, while survey data may lack the authenticity of our direct gender ratio observations. Our approach better captures the outcomes of women’s free choices, which are potentially rooted in subconscious lifestyle habits and instinctive preferences that resist verbal articulation. Therefore, they offer unique reference value for gender preference research. The fsQCA method was originally more commonly used in management and economic analyses. This study introduced it into urban design to expand upon the results of traditional correlational analysis. This method is capable of handling both the presence and absence of conditions, revealing various potential causal relationships. Furthermore, fsQCA provides a new perspective for research on female preferences. Building on the research achievements of other scholars, this study developed an indicator system suitable for fsQCA, which evaluates female preferences and the characteristics of the built environment within a spatiotemporal framework and offers a certain degree of reference value for research in this field.

This research’s findings support studies that demonstrated how surrounding land use, accessibility, security, and public service amenities influence women’s presence in urban spaces [17,33]. We likewise support findings that various complex factors limit women’s access to green spaces [18]. The research framework proposed in this study takes these complexities into account and yields new findings from empirical tests conducted in Chengdu, China.

First, no single built environment factor alone constitutes a necessary or sufficient condition for achieving women-friendly green spaces. Women-friendly UGSs emerge from the concurrent interplay of multiple built environment characteristics, demonstrating a principle of equifinality; that is, diverse combinations of spatial attributes can lead to similar gendered outcomes. Specifically, six distinct configuration pathways were identified, revealing that varied arrangements of accessibility, safety infrastructure, and service provision may equivalently enhance women’s spatial preferences.

Second, the difference between weekdays and weekends does not significantly affect women’s activities in UGS. However, there are significant differences between daytime and nighttime activities. Most feminist researchers have focused on daytime activities, arguing that these better reflect differences in choices related to family and social identity [53]. However, we reveal that women in our study area prefer using UGS at night. This preference aligns with their group activities, exercise routines, and family walks, which predominantly occur in the evening to meet their specific activity types and spatial requirements [54]. We find that the demand of female users for accessibility, usability, and security in UGS during nighttime is lower than anticipated. Instead, they prefer UGSs that offer high privacy, limited accessibility, a stable number of users, minimal human intervention, and more natural settings. This preference may be linked to the better security conditions in our study area.

Third, the security of UGS and PSFs frequently emerge as core conditions influencing women’s preferences across the six identified paths. In our case studies, UGS sections near residential areas often lack security features yet consistently offer PSFs; these conditions are inversely related in terms of their presence. Conversely, in non-residential areas, security features and PSFs in UGS sections usually exhibit synergy, enhancing their attractiveness to women. However, this is not the case in path 2c. This phenomenon can be explained by the tendency of women to visit UGS sections near residential areas in groups. These groups often include children, spouses, or friends. Consequently, their demand for diverse and numerous PSFs surpasses their need for spatial security. In non-residential areas, however, security concerns are predominant. This situation necessitates that PSFs, particularly those for resting and sports, be situated in secure environments. This alignment ensures the availability of both security and PSFs. The new group-based findings validate and extend previous research on women’s perceptions of security in urban spaces [40,55], pinpointing specific circumstances under which women require enhanced security in UGS sections.

5.2. Implications for Urban Planning and Management

Various other types of green spaces form in the city besides the continuous linear green space we chose, but almost all public green spaces are closely related to the characteristics of the built environment discussed in this text, and our findings based on women’ freedom to choose according to their own preferences are generalizable. The identification of six paths that align with women’s preferences for UGS provides constructive suggestions for renovating UGS sections that initially overlooked women’s needs. Existing UGS sections may not meet all criteria perfectly. However, the findings of this research can help managers select the most suitable paths to enhance the women-friendliness of UGS sections, taking into account their current environmental attributes. Managers can use the three major categories of paths to align existing UGS conditions with the identified paths. This alignment will help determine the most suitable renovation strategy to attract women. For instance, a small UGS at the edge of a residential area with ample greenery and minimal pavement could enhance the diversity and quantity of PSFs. This enhancement would align with path 1a to better meet women’s preferences. Similarly, if a small UGS at the edge of a residential area already provides a range of PSFs, it could expand green spaces and reduce visibility. This change would enhance privacy and align with path 1a. If the conditions of a path are complex and challenging to implement, extracting the core conditions can simplify the process and improve practicability. For example, a large UGS with limited access away from residential areas might follow path 2c by reducing pavement and enhancing both security and the diversity and quantity of PSFs. By detailing the presence or absence of specific conditions, these six paths significantly improve the practical applicability of theoretical guidance.

Under this new UGS design paradigm, designers must comprehensively consider the interaction mechanisms between various built environment characteristics according to the usage patterns of specific user groups rather than indiscriminately applying identical standards to male and female users via conventional assumptions. A green space that values women’s preferences is neither exclusionary to males nor prioritizes feminine preferences at the expense of universal design principles. Instead, it requires systematically analyzing gendered demand differentials, translating women’s subjective needs into objective spatial manifestations, thereby progressively optimizing green infrastructure to enhance the well-being of women across all age groups. Equitable design transcends mere compliance with female-prioritized features, such as accessibility and safety. It necessitates multidimensional considerations encompassing user demographics (e.g., age and group composition), activity typologies, spatial perception patterns, and behavioral rhythms. Particularly in neighborhood green space design—which holds profound significance for middle-aged female users—these small-scale, cost-effective renovation projects (typically below municipal park standards) serve fixed communities. Implementing evidence-based design informed by localized female preference research can yield disproportionate social benefits relative to investment.

5.3. Study’s Limitations and Further Research Opportunities

We acknowledge three main limitations of this study. First, although it proposes six paths to make UGS attractive to women, other configurations may be more suitable for local conditions. These alternatives require further exploration through additional case studies in various contexts. Second, this study mainly focuses on the physical characteristics of the built environment of UGS while paying less attention to the esthetic aspects. Taking into account the UGS esthetics could reveal more mechanisms underlying the gendered use of UGS. Third, in Chinese cities such as Shanghai, Guangzhou, and Nanjing, the differences in green space preferences can be attributed to a number of complex factors, as citizens in each city have different hobbies and rhythms of life, and their preferences for green spaces are influenced by seasonal temperatures and even by the length of the day and night. Comparative studies between different cities may be a meaningful direction for future research and promote urban sustainability in China.

6. Conclusions

This study aims to understand the mechanisms underlying women’s preferences for green spaces in order to improve their access to high-quality UGS and provide valuable insights for future research and decision-making. Results reveal that no single built environment characteristic is a necessary and sufficient condition for determining women’s favorite UGS. Instead, six configuration pathways are identified that can drive women’s preferences for green spaces within the scope of this study. Among these pathways, time, security, and PSFs have the most significant impact on women’s preferences. It is anticipated that this study will bring greater attention to caring urbanism and foster dialog in this field among urban planning and design researchers and policymakers, particularly in non-Western and developing countries.