How Outdoor Environments in Kindergarten Support Children’s Autonomous Play Behavior: A Case Study of Beijing, China

Abstract

In high-density urban environments, outdoor kindergarten spaces are vital for children’s cognitive and social development, yet their design within constrained urban greenery poses a significant challenge. This study investigated how these environments support development through autonomous play. Conducted as a case study in three Beijing kindergartens, it employed a framework analyzing ten environmental elements across four dimensions: terrain space (e.g., open space, slopes), game facilities (fixed and movable), loose materials, and natural elements (water, plants). Behavioral observations were used to examine associations between these elements and children’s play behaviors. The findings suggest that diverse, naturalized, and adaptable combinations of elements may best foster autonomous play. While functional play was predominant, our analysis identified that a core combination of rigid fixtures, shielded places, and loose materials appears to optimally support this play type, which is primarily linked to solitary play. By strategically supplementing this core with elements like moving fixtures and loose objects, the environment can further encourage constructive, dramatic, and exploratory play—forms that show stronger associations with cooperative group play. This reveals a potential pathway through which sequenced environmental provisioning might scaffold the progression from individual to social play, thereby fostering socio-cognitive growth. Consequently, the study proposes three exploratory design principles: the differentiated allocation of elements to target specific play behaviors, deliberate naturalization of the setting, and incorporating dynamic adjustability for flexibility. These hypothesis-generating strategies aim to inform the design of kindergarten outdoor spaces, offering practical guidance for creating more sustainable and child-inclusive urban communities, though their generalizability requires further cross-context validation.

1. Introduction

Today, the global urban population is increasing rapidly. As a result, many countries will face numerous challenges in meeting the needs of their growing urban populations, including those related to environmental, social and economic issues [1]. In high-density urban environments in China, the scale and quality of children’s outdoor spaces in urban green space are often insufficient [2,3,4,5]. On the other hand, early childhood, as the beginning of an individual’s life course, is critical for establishing a solid foundation for a child’s social, emotional and overall well-being [6]. The early years of a child’s life build the basis for lifelong growth [7]. As a form of activity that conforms to children’s nature, autonomous play is conducive to promote children’s cognitive and social development [8,9,10,11]. The outdoor environment, with its characteristics of openness, diversity and naturalness, provides rich and suitable conditions for children to carry out autonomous play [12,13,14,15,16,17]. The outdoor space of kindergartens, as an easily overlooked sub-component of urban green spaces, is precisely the most important living and learning place for children. The quality of its space is particularly important for children’s comprehensive development [18,19,20,21,22].

In recent decades, researchers have explored the relationship between outdoor environment and children’s behavior. Studies have indicated that the quality of the physical environment is associated with early childhood development by influencing behavioral, cognitive, and emotional aspects [18,19,20]. Scholars have demonstrated that there is a positive correlation between the utilization of facilities and children’s outdoor play and sports activities [21]. Relative studies typically focus on specific types of spaces, such as parks, streets and public open spaces. The most important behavioral variables in the activities are collaboration and social interaction [22,23,24,25,26,27,28], participation and decision-making [24,25,29,30], types of play involved [31,32,33,34,35], motor play [36], and interaction with nature [37].

Study showed environments that support multiple types of play have the greatest value, and children’s play value continues to increase, from traditional playground designs that feature fixed equipment to complex, nature-inspired designs that focus on manipulation and dynamic elements [38]. The diversity and complexity of space and play materials are related to more exploratory behavior, participation, and social interaction [39]. Studies have demonstrated that outdoor environments in kindergartens should have a child-scale design, diversity and suitability of materials, spatial definition, comfort, cleanliness, and safety. Relative freedom, uncrowded spaces, and accessibility of equipment promote children’s exploration [26,28,40,41]. Many studies have focused on proving that the natural environment has a positive impact on children’s more diverse and creative forms of light physical activity and psychological development. Adding natural materials to outdoor environments can increase play with natural materials, autonomous play, and prosocial behavior, while reducing antisocial behavior and engagement in moderate-to-vigorous physical activity [23]. When natural and artificial areas are mixed, children seem more willing to participate in different types of cognitive, functional, and exploratory play compared to only natural or artificial areas [35]. Fixed functional devices [42,43,44,45] and loose parts [23,46,47,48] fail to establish clear conclusions about their association with bodily activity. This illustrates the complexity of the relationship between children and their environment.

In the study of children’s environmental classification, the issue with analyzing spatial features as factors affecting children’s play behavior might lie in the difficulty of quantifying and scoring different spatial factors. Children’s play behavior is not so much affected by the type of playground as by the play elements available to them [49]. Heft [50] was the first to propose a functional support classification for children’s activity environments, constructing preliminary functional characteristics provided by various environmental features in outdoor environments based on affordance theory, and categorizing the affordances that might occur in the environment into ten environmental elements. Lerstrup [51], based on Heft’s [50] research and combined with his own research, proposed optimization and modification suggestions for the classification of environmental affordance.

The method selection of related research mostly involves observation techniques. The Play Observation Scale (POS), revised by Rubi et al. based on the Patton/Piaget Scale, combines Patton’s social play classification and Piaget’s cognitive play classification to create a cross-dimensional play behavior classification, and adds five non play behaviors. In 2001, Rubi et al. revised the scale again and added the category of “exploration” to the classification of cognitive play behaviors. Previous studies have shown that POS has good reliability and validity, and can be used to demonstrate the impact of play environment on children’s play behavior, distinguish individual differences in children’s play levels during play, and demonstrate the interaction between cognitive and social levels of play [52].

Such studies have covered multiple regions in non-Western countries, including Singapore, China, Iran, Malaysia, Romania, Indonesia, etc. The research content is extensive, covering the design and evaluation of children’s outdoor play Spaces, the relationship between play behavior and environmental characteristics, and the impact of play on children’s social skills and scientific creativity, etc. In terms of research methods, scholars adopted a hybrid approach, integrating various means such as questionnaire surveys, interviews, GIS spatial analysis, behavioral observation and recording, as well as quantitative analysis [53,54,55,56]. In terms of conclusion, research indicates that children have diverse demands for play Spaces, including safety, fun, challenge, and the integration of natural elements [57,58,59]. In terms of design, it is necessary to consider promoting social interaction and physical activity [5,17]. Meanwhile, natural materials and green Spaces have a positive impact on children’s emotional health and social anxiety [59,60]. In addition, the design of game facilities should encourage cooperative and creative play [61,62].

Although research in the field of environmental behavior has made significant progress, there are still obvious shortcomings in the subdivision of children’s activity space and behavior. Currently, there is a lack of rich, mature and comprehensive quantitative evidence to support research on the relationship between outdoor spaces in kindergartens and children’s play behavior. The research on this field is mostly at the level of hypothesis, qualitative or semi-quantitative research or focus on the relationship between simple single element and the environment.

Most of the publications involved in the current research on children’s activity space and behavior originate from the United States, Norway, Canada and other countries. The geographical evidences related to the global child friendly city movement in Western countries, Australia, and New Zealand is relatively concentrated [63]. In China, due to the high importance attached by kindergartens to the protection of children’s privacy, there are a relatively small number of related studies on the correlation between outdoor environment and children’s behavior in kindergartens compared to foreign countries. In view of this, this study will focus on this area by investigating the correlation between spatial characteristics and children’s autonomous play within the specific context of high-density urban kindergartens in China. It seeks to advance the methodological and conceptual approaches to observational research on environment-play relationships in the following ways. Methodologically, it employs a dynamic post-occupancy evaluation (POE) intermittent sampling approach during peak activity periods, utilizing tools like the Play Observation Scale (POS) to enhance the ecological validity and representativeness of behavioral data. Conceptually, it applies a child-centered, evidence-based POE framework specifically to play behavior observation. Analytically, it establishes a localized classification framework and integrates quantitative analysis of empirical data with qualitative insights to explore complex interactions among multiple environmental elements. The findings, underpinned by a large-scale sample, aim to strengthen the reliability of evidence in this field. The overarching goal is to provide correlational insights for design that may support the occurrence of such play behaviors, thereby contributing to the broader understanding of children’s cognitive and social development. This work will supplement the relatively lacking evidence in the Chinese context and offer references for the construction of child-friendly cities.

2. Materials and Methods

2.1. Research Design

This study aims to explore how different outdoor environmental elements correlate to children’s cognitive and social play during autonomous activities. It consisted of three phases: site investigation and classification of environmental elements, observation and recording of play behaviors, and correlation analysis (Figure 1). The study was approved by the participating kindergartens and the Human Research Ethics Committee of China Agricultural University (CAUHR-2021002). All data were conducted from 2022 to 2024 and used exclusively for analysis, and video materials were kept strictly confidential.

2.2. Study Sites

The three kindergartens selected for this study are typical examples that emphasize nature and autonomous play: namely Fengtai District Hangtiansanyuan Kindergarten (South Campus), Fengtai District Hangtiansanyuan Kindergarten (North Campus), and Haidian District Hongyuan Kindergarten. Their total areas are 7690, 6400, and 9800 square meters respectively, with autonomous play areas of 2853, 2670, and 3277 square meters. In this study, the “autonomous play area” refers to the outdoor space primarily located within the planned kindergarten land boundary yet outside the building footprint, which is designated and available for children’s autonomous play activities. One is private and the other two are public, but are inclusive kindergartens receiving government subsidies. The selection was based on two key criteria. First, regarding the emphasis on nature, all three kindergartens’ outdoor spaces were co-designed with the research team based on the “natural practice” concept, which prioritizes sustainability, low-impact intervention, and the use of natural and onsite materials, fully considering children’s developmental needs and safety. Second, regarding the emphasis on autonomous play, each kindergarten highly valued and had successfully implemented an autonomous play curriculum, where children could freely choose activities without teacher intervention, providing a suitable environment for observing such behaviors. The study team participated in the design of outdoor activity spaces for three kindergartens and has been continuously collaborating with them on scientific research and spatial ring updates.

The study sites were divided and measured. Based on usage functions, element configurations, natural boundaries, design intentions, etc., the outdoor activity spaces of the three kindergartens were divided into 54 independent sites as samples. These criteria were clarified through a three-step process: preliminary zoning based on physical boundaries and original design documents, review and confirmation by the original designers, and iterative refinement to align objective site features with design intent. Respectively, the three kindergartens were divided into 23, 13, and 18 sites (Figure 2).

2.3. Environmental Elements

Based on the environmental element classification by Heft and Lerstrup [50,51], 10 types of elements were identified: open space, sloping terrain, shielded places, rigid fixtures, moving fixtures, loose objects, loose material, water, plants, and animals. These are grouped into four main categories: terrain space, game facilities, open materials and natural elements (

Table 1. Definition of environmental elements.

Number	Element		Definition	Scene
1	terrain space	open space	The relatively open outdoor space supports activities like running, jumping, cycling, chasing, playing games, and competitions.
2		sloping terrain	The elevation changes include varied terrains such as slopes, steps, and small slopes, excluding the second-floor platforms inside large play facilities that are not connected to the ground.
3		shielded places	Including small hidden spaces and spaces that provide shelter.
4	game facilities	rigid fixtures	Fixed-position play facilities including large-scale comprehensive fixed facilities (such as finished slides, etc.) and other small-scale facilities (such as plum blossom posts, trestles, etc.).
5		moving fixtures	Facilities that can be moved. Facilities that cannot be moved due to children’s ability limitations do not fall into this category.
6	open materials	loose objects	Loose materials are low- or unstructured items—small, scattered objects that children can grasp.
7		loose material	Materials such as sand, mud, and crushed stone have small particles and are available in large quantities.
8	natural elements	water	All accessible water play facilities, including washbasins.
9		plants	Trees, shrubs, grasses, and horticultural plants.
10		animals	Raised animals.

). All three kindergartens were found to include these ten environmental types in their outdoor areas (Appendix A, Figure A1). The frequency of each element across all sites was counted to show their distribution.

A statistical analysis was conducted on the environmental elements across 54 sites. Each site was assigned a binary value (1 if present, 0 if absent) for each element. Because some elements changed during the three autonomous play sessions, data were collected three times, resulting in 162 total activity sites values (Appendix D 

Table A4. Summary table of environmental elements assignment results in active area.

Partition Code	Partition Name	Open Space	Sloping Terrain	Shielded Places	Rigid Fixtures	Moving Fixtures	Loose Objects	Loose Material	Water	Plants	Animals
0	Bare soil and loose objects area	1	1	0	1	1	1	0	0	1	0
1	Variable facility area	0	0	0	0	0	0	0	0	1	0
2	Horticultural planting area	0	0	1	1	1	1	1	0	1	0
3	Terrain subsidence area	1	1	1	1	1	1	1	0	1	0
4	Sandpit and cloud ladder area	1	0	0	1	0	1	1	0	0	0
5	The bark area of the pool	1	0	0	0	0	0	1	0	1	0
6	Animal area	0	0	0	0	0	0	0	0	1	1
7	Separate the green areas	0	0	0	0	0	0	0	0	1	0
8	Music Classroom Area	0	0	1	1	1	1	0	0	1	0
9	Large floor mat area	1	0	1	0	1	1	0	0	0	0
10	Pergola area	0	0	1	0	1	1	1	0	1	0
11	Recreational green area	0	0	0	0	0	1	1	0	1	0
12	Small slide area	0	0	1	1	0	0	0	0	0	0
13	Variable facility area	0	0	0	0	1	1	1	0	1	0
14	Wooden platform area	1	0	0	0	0	1	1	0	1	0
15	Large Equipment Area	1	0	1	1	0	0	1	0	1	0
16	Bare soil and loose objects area	1	0	1	1	1	1	1	0	0	0
17	Maze area	0	0	1	1	0	0	0	0	0	0
18	Comprehensive water play area	0	0	0	1	0	0	0	1	1	0
19	Sandpit slide area	0	0	1	1	1	1	1	0	0	0
20	Recreational green area	0	0	0	0	0	0	0	0	1	0
21	Large empty area	1	0	0	0	0	0	0	0	0	0
22	Pergola area	0	0	1	1	1	1	1	0	1	0
0	Bare soil and loose objects area	1	1	0	1	1	1	0	0	1	0
1	Variable facility area	0	0	0	0	0	0	0	0	1	0
2	Horticultural planting area	0	0	1	1	1	1	1	0	1	0
3	Terrain subsidence area	1	1	1	1	1	1	1	0	1	0
4	Sandpit and cloud ladder area	1	0	0	1	0	1	1	0	0	0
5	The bark area of the pool	1	0	0	0	0	0	1	0	1	0
6	Animal area	0	0	0	0	0	0	0	0	1	1
7	Separate the green areas	0	0	0	0	0	0	0	0	1	0
8	Music Classroom Area	0	0	1	1	1	1	0	0	1	0
9	Large floor mat area	1	0	1	0	1	1	0	0	0	0
10	Pergola area	0	0	1	0	1	1	1	0	1	0
11	Recreational green area	0	0	0	0	0	1	1	0	1	0
12	Small slide area	0	0	1	1	0	0	0	0	0	0
13	Variable facility area	0	0	0	0	1	1	1	0	1	0
14	Wooden platform area	1	0	0	0	1	1	1	0	1	0
15	Large Equipment Area	1	0	1	1	0	0	1	0	1	0
16	Bare soil and loose objects area	1	0	1	1	1	1	1	0	0	0
17	Maze area	0	0	1	1	0	0	0	0	0	0
18	Comprehensive water play area	0	0	0	1	0	0	0	0	1	0
19	Sandpit slide area	0	0	1	1	1	1	1	0	0	0
20	Recreational green area	0	0	0	0	0	0	0	0	1	0
21	Large empty area	1	0	0	0	0	0	0	0	0	0
22	Pergola area	0	0	1	1	1	1	1	0	1	0
0	Bare soil and loose objects area	1	1	0	1	1	1	0	0	1	0
1	Variable facility area	0	0	0	0	1	1	0	0	1	0
2	Horticultural planting area	0	0	1	1	1	0	0	0	1	0
3	Terrain subsidence area	1	1	1	1	1	1	1	0	1	0
4	Sandpit and cloud ladder area	1	0	0	1	0	1	1	0	0	0
5	The bark area of the pool	1	0	0	0	0	0	1	0	1	0
6	Animal area	0	0	0	0	0	0	0	0	1	1
7	Separate the green areas	0	0	0	0	0	0	0	0	1	0
8	Music Classroom Area	0	0	1	1	1	1	0	0	1	0
9	Large floor mat area	1	0	1	0	1	1	0	0	0	0
10	Pergola area	0	0	1	0	1	1	1	0	1	0
11	Recreational green area	0	0	0	0	0	1	1	0	1	0
12	Small slide area	0	0	1	1	0	0	0	0	0	0
13	Variable facility area	0	0	0	0	1	0	0	0	1	0
14	Wooden platform area	1	0	0	0	0	1	1	0	1	0
15	Large Equipment Area	1	0	1	1	0	0	1	0	1	0
16	Bare soil and loose objects area	1	0	1	1	1	1	1	0	0	0
17	Maze area	0	0	1	1	0	0	0	0	0	0
18	Comprehensive water play area	0	0	0	1	0	0	0	1	1	0
19	Sandpit slide area	0	0	1	1	1	1	1	0	0	0
20	Recreational green area	0	0	0	0	0	0	0	0	1	0
21	Large empty area	1	0	0	0	0	0	0	0	0	0
22	Pergola area	0	0	1	1	1	1	1	0	1	0
0	Large floor mat area	1	0	0	0	0	1	0	0	0	0
1	Large empty area	1	0	0	0	0	1	0	0	0	0
2	Pier Facilities area	0	0	1	1	1	1	0	0	1	0
3	Separate the green areas	0	0	0	1	0	1	0	0	1	1
4	Large Equipment Area	0	0	1	1	0	0	0	0	0	0
5	Small cushion area	1	0	0	0	0	1	0	0	0	0
6	Comprehensive Equipment Area	0	0	1	1	0	0	0	0	0	0
7	Trampoline jumping area	0	0	0	1	0	0	0	0	1	0
8	Sensory integration balance region	0	0	0	1	0	1	0	0	1	0
9	Terrain slide area	0	1	1	1	0	1	1	0	1	0
10	Sandpit learning area	0	0	0	1	1	1	1	1	0	0
11	Small empty area	1	0	1	0	1	1	0	0	0	0
12	Separate the green areas	0	0	0	0	0	0	0	0	1	0
0	Large floor mat area	1	0	0	0	0	1	0	0	0	0
1	Large empty area	1	0	0	0	0	1	0	0	0	0
2	Pier Facilities area	0	0	1	1	1	1	0	0	1	0
3	Separate the green areas	0	0	0	1	0	1	0	0	1	1
4	Large Equipment Area	0	0	1	1	1	0	0	0	0	0
5	Small cushion area	1	0	0	0	0	1	0	0	0	0
6	Comprehensive Equipment Area	0	0	1	1	0	0	0	0	0	0
7	Trampoline jumping area	0	0	0	1	0	0	0	0	1	0
8	Sensory integration balance region	0	0	0	1	1	1	0	0	1	0
9	Terrain slide area	0	1	1	1	0	1	1	0	1	0
10	Sandpit learning area	0	0	0	1	1	1	1	0	0	0
11	Small empty area	1	0	1	0	1	1	0	0	0	0
12	Separate the green areas	0	0	0	0	0	0	0	0	1	0
0	Large floor mat area	1	0	0	0	0	1	0	0	0	0
1	Large empty area	1	0	0	0	0	1	0	0	0	0
2	Pier Facilities area	0	0	1	1	1	1	0	0	1	0
3	Separate the green areas	0	0	0	1	0	1	0	0	1	1
4	Large Equipment Area	0	0	1	1	0	0	0	0	0	0
5	Small cushion area	1	0	0	0	0	1	0	0	0	0
6	Comprehensive Equipment Area	0	0	1	1	0	0	0	0	0	0
7	Trampoline jumping area	0	0	0	1	0	0	0	0	1	0
8	Sensory integration balance region	0	0	0	1	1	1	0	0	1	0
9	Terrain slide area	0	1	1	1	0	1	1	0	1	0
10	Sandpit learning area	0	0	0	1	1	1	1	0	0	0
11	Small empty area	1	0	1	0	1	1	0	0	0	0
12	Separate the green areas	0	0	0	0	0	0	0	0	1	0
0	Sandpit water play area	1	0	0	1	1	1	1	1	1	0
1	Playground track area	1	0	0	0	1	1	0	0	0	0
2	Terrain slide area	1	1	1	0	0	1	0	0	1	0
3	Circular grass area	1	0	0	0	0	0	0	0	1	0
4	Outdoor classroom area	1	0	0	1	0	1	0	0	0	0
5	Equipment area to be built	1	0	0	1	0	1	0	0	1	0
6	Garden Exploration Area	0	0	0	0	0	1	0	1	1	0
7	Horticultural planting area	0	0	0	1	0	0	0	1	1	0
8	Export zone	0	0	0	0	0	0	0	0	1	0
9	Stage grandstand area	1	1	0	1	0	0	0	0	0	0
10	Large-scale field area	1	0	0	0	1	0	0	0	0	0
11 + 17	Perceive the path area	0	0	0	0	0	1	0	0	0	0
12	Terrain climbing area	0	1	1	1	0	0	0	0	1	0
13 + 15 + 18 + 20	Green area of the playground	0	0	0	0	0	0	0	0	1	0
14	Terrain Observation Area	0	1	0	0	0	0	0	0	1	0
16	Separate the green areas	0	0	0	0	0	0	0	0	1	0
19	Entrance area	0	0	0	0	0	0	0	0	1	0
21	Green area around the island	0	0	0	0	0	1	0	0	1	0
0	Sandpit water play area	1	0	0	1	1	1	1	1	1	0
1	Playground track area	1	0	0	0	1	1	0	0	0	0
2	Terrain slide area	1	1	1	1	0	1	0	0	1	0
3	Circular grass area	1	0	0	0	0	0	0	0	1	0
4	Outdoor classroom area	1	0	0	1	0	1	0	0	0	0
5	Equipment area to be built	1	0	0	1	0	1	0	0	1	0
6	Garden Exploration Area	0	0	0	0	0	1	0	1	1	0
7	Horticultural planting area	0	0	0	1	0	0	0	1	1	0
8	Export zone	0	0	0	0	0	0	0	0	1	0
9	Stage grandstand area	1	1	0	1	0	0	0	0	0	0
10	Large-scale field area	1	0	0	0	1	0	0	0	0	0
11 + 17	Perceive the path area	0	0	0	0	0	1	0	0	0	0
12	Terrain climbing area	0	1	1	1	0	0	0	0	1	0
13 + 15 + 18 + 20	Green area of the playground	0	0	0	0	0	0	0	0	1	0
14	Terrain Observation Area	0	1	0	0	0	0	0	0	1	0
16	Separate the green areas	0	0	0	0	0	0	0	0	1	0
19	Entrance area	0	0	0	0	0	0	0	0	1	0
21	Green area around the island	0	0	0	0	0	1	0	0	1	0
0	Sandpit water play area	1	0	0	1	0	1	1	1	1	0
1	Playground track area	1	0	0	0	1	1	0	0	0	0
2	Terrain slide area	1	1	1	1	0	1	0	0	1	0
3	Circular grass area	1	0	0	0	0	0	0	0	1	0
4	Outdoor classroom area	1	0	0	1	0	1	0	0	0	0
5	Equipment area to be built	1	0	0	1	0	1	0	0	1	0
6	Garden Exploration Area	0	0	0	0	0	1	0	1	1	0
7	Horticultural planting area	0	0	0	1	0	0	0	1	1	0
8	Export zone	0	0	0	0	0	0	0	0	1	0
9	Stage grandstand area	1	1	0	1	0	0	0	0	0	0
10	Large-scale field area	1	0	0	0	1	0	0	0	0	0
11 + 17	Perceive the path area	0	0	0	0	0	1	0	0	0	0
12	Terrain climbing area	0	1	1	1	0	0	0	0	1	0
13 + 15 + 18 + 20	Green area of the playground	0	0	0	0	0	0	0	0	1	0
14	Terrain Observation Area	0	1	0	0	0	0	0	0	1	0
16	Separate the green areas	0	0	0	0	0	0	0	0	1	0
19	Entrance area	0	0	0	0	0	0	0	0	1	0
21	Green area around the island	0	0	0	0	0	1	0	0	1	0

).

2.4. Play Behavior Classification and Observation

(1)

Reparatory stage

The on-site recording tools used in this study to collect video information on children’s activities include cameras, mobile phones, surveillance cameras and an Unmanned Aerial Vehicle (UAV is only used in one authorized kindergarten, and its function is the same as other video tools. To ensure behavioral naturalness, the UAV was deployed at a flight altitude exceeding 80 m before the children entered the play area, making it visually and audibly indistinguishable from a bird at ground level, thus minimizing the likelihood of children noticing it during activities) (Figure 3). 24 college student volunteers were recruited, and a backup volunteer database was created. They received training in behavior observation while showing interest in children’s landscapes.

(2)

Behavioral observation stage

The study recorded 40 min videos of autonomous outdoor play across 21 sessions. Before each session, we set up equipment and assigned volunteers to different areas to ensure full coverage of children’s play behavior. After selection, videos from 9 of these autonomous play sessions (3 sessions per kindergarten) were included for analysis. A total of 1100 child-sessions from children aged 3 to 6 across the three kindergartens participated in these 9 outdoor autonomous play sessions, yielding 22,000 data points. For each kindergarten, the three observed sessions involved largely the same class of children, with attendance varying slightly each time due to individual absences: Hangtiansanyuan Kindergarten (South Campus) had 179, 159, and 200 participants respectively; Hangtiansanyuan Kindergarten (North Campus) had 121, 109, and 117 participants respectively; and Hongyuan Kindergarten had 69, 73, and 70 participants respectively. After accounting for repeated participation across sessions, over 400 unique children were involved, with a gender ratio close to 1:1.

The Play Observation Scale (POS) was selected as the foundation for this study after reviewing multiple behavioral scales. It was then adapted through learning, refinement, and expansion to suit outdoor observations in Beijing kindergartens, aiming to explore the relationship between children’s behavior types and their environment. Cognitive play was classified from low to high complexity as functional play, constructive play, dramatic play, games-with-rules, and exploratory play. Social play was categorized from less to more interaction as solitary play, parallel play, and group play. Combining both dimensions created cross-coded categories (e.g., A1 = Solitary functional play) (

Table 2. Principles for observing children’s play behaviors.

Behavior Type		Coding	Definition	Example	Scene
cognitive play	functional play	A	Activities carried out to enjoy the physical sensations created mainly consist of simple exercise activities. Activities aimed at enjoying physical sensations mainly involve simple exercises.	Running, crawling, walking on a balance board, riding a tricycle, and skipping rope.
	constructive play	B	Children play to manipulate objects for creating and constructing things, distinguishing them from their functional use.	Building with blocks, molding mud, cutting leaves, gardening, and playing musical instruments.
	dramatic play	C	Any element of pretend play, such as role-playing, imaginative activities, or attributing life to inanimate objects.	Nurse play, house play, wooden boat paddling, gun battle play, and simulated cooking.
	games-with-rules	D	Accept and adapt to pre-established rules. There should be competitive elements among the children.	Running races, playing “Wooden Man”, bouncing balls, and rock-paper-scissors.
	exploratory play	E	Focus on observing an object to gather visual details about its physical properties or auditory information.	Observe, listen, touch, and examine objects.
social play	solitary play	1	Children play separately from others, often at a distance of more than 1 m and with different toys. However, distance is not always the only criterion.
	parallel play	2	Children engage in independent play while still paying attention to others. Their toys are similar to those of nearby children.
	group play	3	Children play with two or more peers, and the activities share a common goal centered on the group.

).

(3)

Video interpretation coding

Based on weather conditions, effective activity duration, the number of children, video recording quality and completeness, etc., the best three observation datasets were selected for each kindergarten, totaling 9 for statistical analysis.

The behavior mapping method was used to digitally record and overlay children’s activities in ArcGIS 10.7 software. Videos were segmented every 2 min using Pot Player 1.7.21916. For each dataset, 20 evenly spaced time points were selected per session. In ArcGIS 10.7, the location information and behavior of children were recorded and analyzed for each observation data by manually marking and encoding, with annotations added to behaviors occurring in the target observation area. We provided unified training for all researchers involved in video interpretation work and conducted cross-validation tests on the same video sample to ensure that everyone has a uniform standard for interpreting the information in the video. Finally, the data was imported into Microsoft Excel (version 2601) and SPSS 26 for statistical analysis.

2.5. Correlation Analysis Method

The Spearman correlation coefficient was used to examine the correlation between children’s play behaviors and environmental factors in SPSS 26. As a non-parametric measure, it assesses correlations by evaluating the monotonic correlation between variables. It reflects the direction of the correlation between an independent variable X and a dependent variable Y. The coefficient ranges from −1 to 1, with values representing negative and positive linear correlations respectively. The closer the absolute value approached 1, the stronger the linear correlation was observed.

This study analyzed correlations among environmental elements, between cognitive and social play behaviors, and between environmental elements and play behaviors.

3. Results

3.1. Environmental Elements in Kindergarten Outdoor Spaces

Based on the statistical results of environmental element frequency across all sites, plants appeared most frequently in kindergarten outdoor spaces (105 times), followed by loose objects (89), rigid fixtures (80), and open space (63). Animals occurred least often (6 times), followed by water (14) and sloping terrain (21) (Figure 4).

The correlation among 10 types of environmental elements in 54 activity sites is analyzed according to the following classification of correlation strength (based on the absolute value of the correlation coefficient): very strong (0.8–1.0), strong (0.5–0.8), moderate (0.3–0.5), weak (0.1–0.3), and very weak or no correlation (0.0–0.1). Based on this, moving fixtures show the strongest link with loose objects, with a moderate positive correlation (0.463), and a moderate positive correlation with loose material (0.335). Open space has a moderate negative correlation with plants (−0.367) and a weak positive correlation with loose objects (0.264). Shielding space shows a weak positive correlation with loose material (0.257) and a weak negative correlation with water (−0.2). Rigid fixtures have a weak positive correlation with loose material (0.217). Sloping terrain correlates weakly positively with plants (0.169), while open space and animals show a weak negative correlation (−0.156) (Figure 5).

3.2. Spatial Characteristics and Correlation of Children’s Play Behaviors

In terms of play behavior frequency, functional play accounts for the largest proportion (66%) among cognitive play types, followed by constructive play (18%), dramatic play (10%), and exploratory play (5%). Games-with-rules have the lowest proportion (1%). Regarding density, functional play has the highest rate (0.47 times/m²), followed by constructive play (0.13). Dramatic and exploratory play show lower densities (both below 0.10), while games-with-rules have the lowest (below 0.01). Functional play not only dominates in both frequency and density but also shows higher volatility and dispersion. Significant differences exist across activity sites, with constructive, dramatic, and exploratory play showing similar distribution patterns. Games-with-rules rarely occur.

Among social play types, parallel play is most common (48%), followed by group play (32%), and solitary play is least frequent (20%). In terms of density, parallel play has the highest value (0.35 times/m²), followed by group play (0.21) and solitary play (0.18). Overall, the three social play types show similar density distributions, with parallel play having the highest values but also greater variability. Group and solitary play densities are relatively close (

Table 3. Behavioral density characteristics of children’s cognitive and social play.

	Minimum Value	Maximum Value	Average Value	Standard Deviation	Variance
functional play density	0.00	6.84	0.47	0.91	0.83
constructive play density	0.00	2.24	0.13	0.36	0.13
dramatic play density	0.00	2.96	0.08	0.31	0.09
games-with-rules density	0.00	0.56	0.00	0.05	0.00
exploratory play density	0.00	0.94	0.04	0.11	0.01
solitary play density	0.00	1.75	0.18	0.29	0.09
parallel play density	0.00	5.14	0.35	0.68	0.47
group play density	0.00	3.24	0.21	0.45	0.21

).

From a spatial distribution perspective of play behaviors, ArcGIS was used to record the spatial locations of children’s play behaviors, and the spatial distribution of the location points was determined by the average nearest neighbor value. Results show that cognitive and social play in the three kindergartens exhibited significant clustering, indicating clear spatial preferences. Kernel density maps illustrate the distribution and clustering levels of these play types across venues (Appendix B Figure A2 and Figure A3). Except for games-with-rules in Hangtiansanyuan Kindergarten (South Campus), which had a z-value greater than 2.58 (z = 3.64), all other play behaviors showed p-values of 0 and z-values below −2.58.

Due to the extremely low overall number of children’s games-with-rules observed in the Hangtiansanyuan Kindergarten (South Campus), there is no condition to compare the differences in the distribution of hotspots of children’s games-with-rules behavior in different facility activity sites. Therefore, the z-value in the average nearest neighbor analysis is greater than 2.58 (z = 3.64), indicating a significantly discrete distribution of the data.

From the perspective of correlations between cognitive and social play, functional play shows very strong positive correlations with solitary (0.837) and parallel (0.850) play, and a strong positive correlation with group play (0.712). Constructive play is strongly positively correlated with solitary (0.524), parallel (0.526), and group play (0.550). Dramatic play correlates strongly with solitary (0.504) and group play (0.554), and moderately with parallel play (0.463). Games-with-rules show a weak positive correlation with parallel play (0.216). Exploratory play has strong positive correlations with solitary (0.589), parallel (0.538), and group play (0.623) (Figure 6, Appendix C 

Table A1. Spearman correlation analysis of children’s cognitive play behavior and social play behavior (two-tailed).

		Functional Play	Constructive Play	Dramatic Play	Games-with-Rules	Exploratory Play
solitary play	Correlation coefficient	0.837 **	0.524 **	0.504 **	0.083	0.589 **
	Sig. (two-tailed)	0.000	0.000	0.000	0.295	0.000
parallel play	Correlation coefficient	0.850 **	0.526 **	0.463 **	0.216 **	0.538 **
	Sig. (two-tailed)	0.000	0.000	0.000	0.006	0.000
group play	Correlation coefficient	0.712 **	0.550 **	0.554 **	0.121	0.623 **
	Sig. (two-tailed)	0.000	0.000	0.000	0.126	0.000

** At the 0.01 level (two-tailed), the correlation is significant.

).

3.3. Correlation Between Play Behavior and Environmental Elements

From the perspective of correlations between cognitive play and environmental factors, functional play density is positively linked to open space, shielded places, rigid fixtures, moving fixtures, loose objects, and loose material, and negatively correlated with plants and animals. Among these, rigid fixtures show a strong correlation (0.551), while shielded places (0.414), loose material (0.357), plants (−0.319), and moving fixtures (0.311) have moderate correlations. Loose objects (0.256), animals (−0.179), and open space (0.171) show weak correlations.

Constructive play density correlates positively with shielded places, moving fixtures, loose objects, and loose material. Moving fixtures (0.459), loose objects (0.447), and loose material (0.419) have moderate correlations, while shielded places (0.205) show weak correlation.

Dramatic play density is positively associated with shielded places, rigid fixtures, moving fixtures, loose objects, and loose material, and negatively with plants. Moving fixtures (0.604) show a strong correlation, while shielded places (0.423), loose objects (0.355), and loose material (0.346) have moderate correlations. Rigid fixtures (0.247) and plants (−0.177) show weak correlations.

Games-with-rules density has a weak negative correlation with plants (−0.242). Exploratory play density is positively correlated with shielded places, rigid fixtures, moving fixtures, loose objects, and loose material. Loose material (0.467), moving fixtures (0.371), shielded places (0.351), and loose objects (0.300) show moderate correlations, while rigid fixtures (0.290) have a weak correlation (Figure 7, Appendix C 

Table A2. Spearman correlation analysis between environmental factors and children’s cognitive play behavior (two-tailed).

		Functional Play	Constructive Play	Dramatic Play	Games-with-Rules	Exploratory Play
open space	Correlation coefficient	0.171 *	0.118	0.097	0.079	−0.029
	Sig. (two-tailed)	0.030	0.135	0.219	0.317	0.715
sloping terrain	Correlation coefficient	0.101	−0.117	0.015	−0.069	−0.053
	Sig. (two-tailed)	0.202	0.139	0.853	0.384	0.501
shielded places	Correlation coefficient	0.414 **	0.205 **	0.423 **	0.098	0.351 **
	Sig. (two-tailed)	0.000	0.009	0.000	0.214	0.000
rigid fixtures	Correlation coefficient	0.551 **	0.140	0.247 **	−0.035	0.290 **
	Sig. (two-tailed)	0.000	0.076	0.002	0.656	0.000
moving fixtures	Correlation coefficient	0.311 **	0.459 **	0.604 **	0.027	0.371 **
	Sig. (two-tailed)	0.000	0.000	0.000	0.730	0.000
loose objects	Correlation coefficient	0.256 **	0.447 **	0.355 **	0.020	0.300 **
	Sig. (two-tailed)	0.001	0.000	0.000	0.801	0.000
loose material	Correlation coefficient	0.357 **	0.419 **	0.346 **	−0.107	0.467 **
	Sig. (two-tailed)	0.000	0.000	0.000	0.174	0.000
water	Correlation coefficient	−0.033	−0.037	−0.020	−0.050	−0.097
	Sig. (two-tailed)	0.678	0.641	0.803	0.524	0.219
plants	Correlation coefficient	−0.319 **	−0.132	−0.177 *	−0.242 **	−0.037
	Sig. (two-tailed)	0.000	0.093	0.024	0.002	0.641
animals	Correlation coefficient	−0.179 *	−0.073	−0.137	−0.035	0.137
	Sig. (two-tailed)	0.023	0.356	0.083	0.658	0.083

** At the 0.01 level (two-tailed), the correlation is significant. * At the 0.05 level (two-tailed), the correlation was significant.

).

From the perspective of correlations between social play and environmental factors, solitary play density is positively associated with shielded places, rigid fixtures, moving fixtures, loose objects, and loose material, and negatively correlated with plants. Shielded places (0.478), rigid fixtures (0.495), moving fixtures (0.444), loose objects (0.358), and loose material (0.474) show moderate correlations, while plants (−0.195) show weak negative correlation.

Parallel play density is positively linked to shielded places, rigid fixtures, moving fixtures, loose objects, and loose material, and negatively correlated with plants. Moderate correlations are found for shielded places (0.431), rigid fixtures (0.418), moving fixtures (0.397), loose objects (0.302), and loose material (0.387), with a weak negative correlation for plants (−0.260).

Group play density correlates positively with open space, shielded places, rigid fixtures, moving fixtures, loose objects, and loose material, and negatively with plants. Moderate positive correlations are observed for open space (0.302), shielded places (0.359), rigid fixtures (0.317), moving fixtures (0.463), loose objects (0.399), and loose material (0.447), while plants show moderate negative correlation (−0.315) (Figure 8, Appendix C 

Table A3. Spearman correlation analysis between environmental factors and children’s social play behavior (two-tailed).

		Solitary Play	Parallel Play	Group Play
open space	Correlation coefficient	0.098	0.048	0.302 **
	Sig. (two-tailed)	0.213	0.545	0.000
sloping terrain	Correlation coefficient	0.065	−0.003	0.033
	Sig. (two-tailed)	0.414	0.974	0.673
shielded places	Correlation coefficient	0.478 **	0.431 **	0.359 **
	Sig. (two-tailed)	0.000	0.000	0.000
rigid fixtures	Correlation coefficient	0.495 **	0.418 **	0.317 **
	Sig. (two-tailed)	0.000	0.000	0.000
moving fixtures	Correlation coefficient	0.444 **	0.397 **	0.463 **
	Sig. (two-tailed)	0.000	0.000	0.000
loose objects	Correlation coefficient	0.358 **	0.302 **	0.399 **
	Sig. (two-tailed)	0.000	0.000	0.000
loose material	Correlation coefficient	0.474 **	0.387 **	0.447 **
	Sig. (two-tailed)	0.000	0.000	0.000
water	Correlation coefficient	−0.063	−0.060	−0.031
	Sig. (two-tailed)	0.427	0.448	0.696
plants	Correlation coefficient	−0.195 *	−0.260 **	−0.315 **
	Sig. (two-tailed)	0.013	0.001	0.000
animals	Correlation coefficient	−0.102	−0.082	−0.114
	Sig. (two-tailed)	0.195	0.301	0.149

** At the 0.01 level (two-tailed), the correlation is significant. * At the 0.05 level (two-tailed), the correlation was significant.

).

4. Discussion

4.1. Environmental Elements in Kindergarten Outdoor Spaces That Support Autonomous Play

(1)

Elements that support cognitive play

The design of outdoor spaces should comprehensively consider the parallel development of various cognitive play to meet the diverse needs of children aged 3 to 6 [64]. Well-designed and organized play spaces can provide opportunities to stimulate multi-faceted development capabilities [65], and can also support various types of play [66].

This study identified 10 environmental elements that support cognitive and social play through different combinations. Functional play is strongly supported by rigid fixtures, shielded places, and loose material, which frequently co-occur. Their combination promotes functional play, enables efficient use of vertical space, creates meaningful play cycles, increases activity duration, and reduces transition time between activities [49]. This is an effective strategy for optimizing outdoor spaces in high-density urban environments.

For example, the sandpit-slide site at Hangtiansanyuan Kindergarten (South Campus) features a second-floor pergola with rigid fixtures such as slides and climbing nets. The sandpit below provides loose material play opportunities and acts as a buffer zone for the slide (Figure 9). This site has become a hotspot for functional play, validating the above findings and demonstrating its effectiveness in space optimization. This combination also plays a key role in supporting solitary play.

This study found that moving fixtures, loose objects, and loose material strongly support constructive play. These elements often co-occur, and their combined use greatly benefits this type of play. Moving fixtures offer flexibility for changing play scenes, while loose objects expand children’s material choices. Loose material enhances constructive play interest and variability and is closely linked to group play (Figure 10).

Moving fixtures, shielded places, loose objects, and loose material are key to dramatic play. These four elements often appear in pairs, and their combined use creates a colorful, safe, and imaginative play environment. Moving fixtures allow quick layout changes for diverse drama scenes. Shielded places provide private and secure spaces, helping children engage more fully. Loose objects and loose material serve as rich props for dramatic play (Figure 11).

Exploratory play is strongly supported by loose material, moving fixtures, and shielded places, which often occur in pairs. Placing moving fixtures within shielded places creates a safe and varied environment for exploration, while loose material offers diverse materials for discovery. Both constructive and dramatic play support a range of social interactions, from solitary to group play.

It is noteworthy that rule-based play constituted only 1% of observed play activities in this study. This low occurrence may be attributed to the nature of autonomous play, where teacher intervention is minimized, and children’s self-organized competitive or rule-governed games are less common at this developmental stage. Teacher interviews corroborated that rule-based play in outdoor settings often relies on guided facilitation. While functional play predominates in high-density environments due to its space-efficient and easily supported nature, designers should also consider how to create environments that gently scaffold the emergence of more complex play forms, including rule-based play, through flexible layouts, multi-purpose equipment, and adaptable loose parts that can be reconfigured into game structures.

Together, the analysis of functional, constructive, dramatic, exploratory, and rule-based play suggests that the combination of three categories of environmental elements—game facilities, terrain space, and open materials—effectively supports various types of cognitive play and naturally encourages social interaction through play. Functional play, dramatic play, and exploratory play all involve the combination of these three elements: game facilities + terrain space + open materials (F + S + M), with at least one element from each category. Constructive play uses a combination of two elements: game facilities + open materials (F + M). The presence of two types of open materials highlights the importance of material richness in supporting constructive play, indicating weaker reliance on terrain space. Similarly, although dramatic play also uses the F + S + M combination, it includes two types of open materials, reflecting its distinct material requirements (

Table 4. Core supporting element combinations for cognitive and social play.

Behavior Type		Core Supporting Element Combination	The Roles of Each Element in the Combination	The Main Associated Social Play
cognitive play	functional play	rigid fixtures (F) + shielded places (S) + loose material (M)	Create a fast play cycle, enable vertical and combined space use, and extend activity duration.	solitary play
	constructive play	moving fixtures (F) + loose objects (M) + loose material (M)	Moving fixtures arrange scenes flexibly, loose objects add diversity, and loose material enhances interest and variability.	group play
	dramatic play	moving fixtures (F) + shielded places (S) + loose objects (M) + loose material (M)	Moving fixtures create scenes, shielded places provide private stages, and loose objects and loose material serve as props that inspire imagination.	solitary play group play
	exploratory play	loose material (M) + moving fixtures (F) + shielded places (S)	Placing moving fixtures in shielded areas creates a safe and diverse exploration environment, with loose material as the core exploration objects.	solitary play group play
social play	solitary play	rigid fixtures (F) + shielded places (S) + loose material (M) + moving fixtures (F) + loose objects (M)	Use the five environmental elements supporting functional play to create semi-shaded spaces in corners, helping reduce children’s anxiety.
	parallel play	shielded places (S) + rigid fixtures (F)	Shielded places provide a safe environment, and rigid fixtures serve as the main game facilities.
	group play	moving fixtures (F) + loose material (M) + loose objects (M) + open space (S)	Moving fixtures, loose objects, and loose materials support collaborative handling, assembly, and shaping. Open space allows for group interaction.

Note: The four major categories of the 10 environmental elements in the table are S = terrain space, F = game facilities, M = open materials.

).

(2)

Elements that support social play

The study found that shielded places and rigid fixtures better support parallel play than group play. In contrast, moving fixtures, loose materials, and loose objects more effectively support group play by enabling children to communicate and collaborate through movement, combination, and shaping. Open space also supports group play better, likely due to the need for larger areas to accommodate interactions. When designing environments, priority should be given to combining game facilities, terrain space, and open materials (F + S + M) to encourage group play. At the same time, spaces for solitary play should also be considered. Rigid fixtures, shielded places, loose materials, moving fixtures, and loose objects all provide some support for solitary play—which are the same elements that best support functional play. Therefore, a combination of F + S + M is recommended. Within both game facilities and open materials, including two environmental elements in each category can better promote functional and solitary play. It is also suggested to create semi-shaded spaces in corners to reduce anxiety caused by enclosed interiors (Table 4).

4.2. Key Elements and Pathways of Support for Cognitive and Social Play

However, in high-density urban settings, space is limited, so environmental design must focus on key elements. The study has found that in outdoor spaces of three kindergartens, functional play occurred more frequently than other types of cognitive play and was the main activity for developing children’s sensory-motor skills [67], consistent with their age-related cognitive development. Functional play enhances executive function and language ability through physical activities [68], laying the foundation for social interaction and serving as the starting point for social play [69,70].

On the other hand, the top five environmental elements supporting functional play are rigid fixtures, shielded places, loose material, moving fixtures, and loose objects. These elements also strongly support constructive, dramatic, and exploratory play. Therefore, prioritizing functional play in environmental design helps support other types of cognitive play. Functional play serves as the foundation for other cognitive play, both in terms of children’s cognitive development patterns and its high occurrence rate and environmental priority.

The study found that functional play areas often naturally trigger social interaction—such as taking turns on slides or cooperating during sand play. Constructive play significantly promotes group play requiring cooperation and sharing [71]. Dramatic play lies at the core of expression and social exploration [72,73,74] and supports all levels of social play. Exploratory play is closely linked to all forms of social play, especially group play, and encourages both individual and collective curiosity [75], which aligns with this study’s findings. Overall, cognitive play rooted in functional play stimulates social interaction across multiple levels.

Therefore, in high-density urban areas, ensuring sufficient space and core elements through the combination of rigid fixtures + shielded places + loose material (F + S + M), with a focus on functional play, serves as the foundation for maximizing environmental benefits. Based on this, the strategic addition of moving fixtures and loose objects can further support cognitive play such as constructive, dramatic, and exploratory play.

Our findings indicate that a potential sequence in which environmental support for foundational functional play—centered on the combination of rigid fixtures, shielded spaces, and loose materials—serves as the starting point for triggering more complex play behaviors. This support is first linked to solitary play, promoting basic cognitive and sensorimotor development. Furthermore, through the strategic addition of elements such as moving fixtures and loose objects, constructive, dramatic, and exploratory play can be further encouraged. These more advanced forms of play show stronger associations with group play, fostering cooperation and higher-level cognitive development.

This observed potential pathway—from functional play → solitary play and from constructive/dramatic/exploratory play → group play—aligns with classical theoretical frameworks of children’s play development, which posit that social play often progresses from individual to parallel to group interactions [76]. Thus, the multi-level support chain illustrated in our study (Figure 12) can be interpreted as a plausible hypothesis in a well-supported cognitive play environment, where play behaviors and social interactions evolve progressively, thereby promoting the sequential development of children’s cognitive and social abilities. However, these interpretations remain correlational and should be viewed as design-informed hypotheses rather than causal conclusions.

4.3. Environmental Design Strategies for Outdoor Space in Kindergartens

Compared with adults, children at play are more spontaneous and have less inclination to sustain an activity. In environmental psychology, the elements that trigger action within a place are termed affordability. According to the theory of availability, diverse environments offer the possibility for its creative use [77,78]. Research on children’s outdoor behaviors reveals that children will choose places that offer various activity opportunities [79,80,81]. Beckwith’s [82] research suggested that the higher the diversity and flexibility of the equipment, the higher its sociability. A well-designed layout of play facilities should create a safe yet challenging environment that encourages children’s interest in play and exploration. Based on empirical findings, this study proposes the following design strategies:

(1)

Differentiated allocation of environmental elements

This study found that shielded places strongly support dramatic play. This may be because they can form semi-private spaces of varying sizes and openness, acting as stages or quiet corners. These spaces accommodate different group sizes and provide props such as stage setups, musical instruments, and game tools, inspiring creativity in dramatic play. This aligns with previous findings that semi-private spaces encourage dramatic play [83]. Rigid fixtures effectively support functional play by offering climbing, balancing, and sliding activities. Moving fixtures better support dramatic, constructive, and group play due to their flexibility and adaptability—they can be reconfigured, moved, or adjusted based on children’s needs and play scenarios. However, it is important to ensure that the difficulty level matches children’s abilities to avoid anxiety. Loose objects significantly support constructive and dramatic play. Their variability in shape, size, texture, and color stimulates creativity. Loose materials like sand and soil better support exploratory play. They enhance site flexibility and create natural, dynamic environments. When combined with other elements, they also promote group interaction.

(2)

Naturalization strategy

Combining plants and sloping terrain: The study found that plants and sloped landforms frequently co-occurred in the three kindergartens’ outdoor areas. This combination creates a rich environment for play and exploration. It simulates a miniature natural landscape, stimulating curiosity, encouraging ecological awareness, and connecting children with nature. Within this natural setting, movable, novel elements—such as stainless-steel slides, tunnels, and climbing stakes—can be added to integrate natural and diverse play experiences (Figure 13).

Optimization of plant configuration: Avoid excessive planting in open spaces, as plants and open space are negatively correlated. Overplanting can reduce available open space, affect accessibility and activities, block light and visibility, and increase maintenance challenges. Instead, plants with high permeability and play value should be combined with sloping terrain, loose material, and moving fixtures to create semi-enclosed or sheltered spaces that support exploratory and constructive play. This approach is supported by findings showing a stronger negative correlation between plants and group play than with solitary or parallel play, indicating that plants may significantly limit group interaction. This suggests that group play typically requires more open space.

Use of natural materials: More and more initiatives around the world are advocating for the naturalization of school grounds through woodlands, gardens and nature play [84,85,86,87,88,89]. On-site observations show that naturally shed materials like leaves and fruits provide loose objects for play. In this study, deciduous broad-leaved trees (e.g., poplar and ginkgo) at Hangtiansanyuan Kindergarten (South Campus) supplied abundant natural loose materials in autumn, enriching constructive, dramatic, and exploratory play with a sense of natural wildness (Figure 14).

(3)

Dynamic adjustability principle

The above research shows that the diversity and flexibility of environmental elements are key to encouraging children’s autonomous play. However, the relationship between environment and play is not one-way but dynamic and interactive. For example, children often reconfigure or adapt environmental elements during play, creating new scenes and ways to play. Specific instances included stacking tires to form a climbable terrain feature over a depression, and using loose parts in an open space to construct the content and form of a slide. This interaction suggests that when designing kindergarten outdoor spaces in high-density urban areas, designers should consider not only the initial setup of environmental elements but also their adjustability and adaptability—to meet evolving play needs while balancing safety with challenge.

Based on exploratory empirical findings from three Beijing kindergartens, this study proposes the following design strategies. It is crucial to note that these should be considered as context-sensitive suggestions requiring further validation in diverse settings. These strategies are derived from observed patterns in a specific high-density urban context and may not be directly transferable to all kindergarten settings. Their effectiveness likely depends on cultural, climatic, and pedagogical factors, underscoring the need for adaptive and locally responsive design approaches.

4.4. Limitation

The findings and implications of this study are bounded by several methodological and contextual limitations that must be considered when interpreting the results.

First, the study’s context and sample limit generalizability. The research was conducted in three urban kindergartens in Beijing. Specifically, the sample focused on inclusive, government-subsidized kindergartens, which represent the predominant type in Beijing’s public preschool system but may not reflect conditions in elite private institutions or under-resourced rural/peripheral schools. The findings may therefore not be directly transferable to other cultural or socio-economic contexts where play behaviors, educational philosophies, and spatial resources may differ.

Second, key methodological simplifications shape the scope of inference. To manage the complexity of outdoor environments and address our core design-oriented question—how to configure and combine environmental elements within functional areas—we employed two pragmatic analytical strategies: (1) using functional zones as the unit of analysis, which efficiently reveals area-level correlations between spatial configuration and collective play density but aggregates data, masking individual children’s play trajectories, preferences, and social dynamics; and (2) binary coding of environmental elements (presence/absence), which enables systematic examination of elemental co-occurrence and combinations but does not capture gradients in quality, size, quantity, or condition that likely influence play affordances. Consequently, our findings are best interpreted as identifying probable spatial prerequisites and associative patterns at the zone level, offering a foundational, design-relevant evidence base rather than elucidating individualized causal pathways or nuanced effects of element properties.

Third, specific environmental and temporal factors may have influenced outcomes. The autonomous play sessions in this study were conducted between September and November 2023. This autumn period coincides with plant dormancy in Beijing, which may have reduced vegetation’s aesthetic appeal and interactive potential, possibly contributing to the observed weak or negative correlations between plants and play. While the study window was selected for its conducive climatic conditions (minimizing weather-related confounds), seasonal and microclimatic variations warrant further investigation for a comprehensive understanding of outdoor play dynamics.

Collectively, these limitations underscore that our conclusions are provisional and context-sensitive. They highlight the critical need for future research to validate and extend the proposed principles across diverse kindergarten types, cultural settings, and seasons, employing more granular measures of environmental attributes and individual child tracking to deepen the evidence base for child-friendly outdoor design.

5. Conclusions

Through empirical research conducted in three urban kindergartens in Beijing, this study explored associations between ten environmental elements and children’s autonomous play. The findings from this specific context suggest that different types of play tend to rely on specific environmental combinations. Accordingly, spatial design in similar high-density settings could emphasize tailored configurations and adaptable elements to meet diverse play needs. Functional play was the most frequently observed type of cognitive play in this study and appears to be strongly supported by a combination of rigid fixtures, shielded places, and loose materials. Especially under spatial constraints, prioritizing these key elements may help create efficient play cycles. Thus, optimizing environmental element combinations aimed at functional play could potentially support more complex play forms, which in turn might encourage a progression from individual to group play, thereby promoting children’s development.

These interpretations and the resulting design implications are bounded by several important limitations. The study was conducted in a specific urban and cultural context (Beijing), during a particular season, and within government-subsidized kindergartens, which may limit the generalizability of the findings. The observed correlations do not imply causation, and the proposed design strategies are exploratory in nature.

Notwithstanding these contextual boundaries, this study offers methodological and practical contributions. The analytical framework—dividing spaces into functional samples—provides a replicable approach to quantifying environment-play relationships in high-density contexts. The findings, while preliminary, suggest that strategic allocation of environmental elements could better support autonomous play. The proposed principles (e.g., naturalization, dynamic adjustability) are offered as context-sensitive hypotheses for translating sustainable development goals into children’s daily experiences, providing a preliminary blueprint for child-friendly urban communities under spatial constraints.

Looking forward, the transferability of these patterns requires validation. Future research should expand across diverse regions, cultures, and kindergarten types, examine long-term impacts, and apply new technologies. Such work is essential to advance evidence-based practices in child-friendly outdoor design, ultimately contributing to healthier generations and more sustainable cities.