A study on children's multi-sensorial experiences of nature: design approaches and preferences for primary school architecture case studies in Glasgow, Scotland, UK

Acknowledging the importance of better connecting children with nature, nature-based proposals for educational buildings and landscape designs have been proposed as a way to nurture children's affinities for the natural world (Kellert, 2015; Louv, 2006). In this study children from 7 to 12 years old were highlighted in “the concrete operation stage” for their greater environmental experience range, compared to the under 6-year-old and over 12-year-old stages (Heerwagen and Orians, 2002; Piaget, 1964; Steiner, 1996, p. 11). This group is considered to exhibit a significant intellectual development engaging with surrounding creatures and natural settings (Kellert, 2002; Miller and Kuhaneck, 2008), and a higher connection to nature than teenagers (Hughes et al., 2019). Consequently, children at this stage are encouraged to have daily and meaningful interactions with nature.

However, in industrialized and urban settings children are spending less time outdoors. It is becoming increasingly apparent, in educational settings worldwide, that this reduction in their interaction with nature is being reflected in the experiences of children (Larson et al., 2019). In order to overcome this matter the impacts, and potential approaches, of a greening process that integrates nature and nature-based pedagogies into school settings, are being increasingly investigated to provide a better educational environment (Kuo et al., 2019). Previous research indicates that a greening process in school settings offers potential for positive contributions to children's cognitive development and performance (Dadvand et al., 2015; Li and Sullivan, 2016), and to their health and wellbeing (Bagot et al., 2015; van Dijk-Wesselius et al., 2018). A number of studies focusing on primary school architecture, have examined the impacts of nearby urban configurations, spatial features of studying and playing environments on the child–nature connected intensity. Recent research on ‘Child–Nature–Distance” by To and Grierson (2019) found that a well-planned and designed school could enhance children's proximity to nature, supporting findings that closer juxtaposition between primary school sites and urban green spaces provides more positive benefits to pupils through daily multi-sensorial experiences, and for various educational and recreational activities (Fischer et al., 2019; Liu and Chen, 2021). Within school playground sites, it has been shown that information-rich natural features, in contrast to barren and artificial surfaces and settings, expand the diversity of playing opportunities within safer and friendlier environments (van Dijk-Wesselius et al., 2018; Ndhlovu and Varea, 2018), enhance pupils' academic performance (Dadvand et al., 2015) and increase environmental stewardship (Jansson et al., 2014). In particular, providing a wide range of natural experiences contributes to a closer connection between children and their local natural environment (Danks, 2019). In the UK, existing policies and frameworks provide the context for the study of wilderness school environments, namely: Area guidelines for mainstream schools: Building Bulletin 103 (DfE, 2014), Building Better Schools: Investing in Scotland's Future (COSLA and The Scottish Government, 2009), and Creating Excellent Primary Schools (CABE, 2010), although these are suggestive rather than mandatory in the way designers currently practice. According to these sources, important features for sustainable school environments in the UK are naturalized outdoor spaces that not only enhance various children's activities but also provide good indoor environmental conditions through natural lighting and ventilation. Providing contact with nature through openings within classrooms is important for natural lighting, views and air quality that can positively influence students' health, academic performance, and behaviours (Küller and Lindsten, 1992; Li and Sullivan, 2016). Furthermore, previous studies also highlight benefits of natural settings in classrooms, such as indoor plants or vertical green walls (van den Berg et al., 2017; Han, 2009). Given the growing evidence for positive benefits of natural-based settings, and considering the significant amount of time children spend at primary schools, exploring feasible architectural interventions that can provide children with multi-sensorial experiences of nature at school has become an imperative. The model of a “forest school”, where woodland is utilized as an entire school setting, presents a prominent example of advancing pupils' hands-on experiences of nature (Harris, 2021). Nevertheless, in the majority of situations, children from urban areas do not have opportunities to attend such schools. A possible response to these issues is emerging through the application of biophilic design principles developed from the theory of biophilia. According to Dr Stephen Kellert (2018, pp. 147-152), notable features of biophilic design in educational architectures (e.g. natural lighting, wetlands, swales, courtyards, interior connections to the outside environment shapes and forms inspired by nature and natural-rich materials) could both upgrade the environmental qualities for all children's activities and enrich natural connections while still ensuring safety and security for them. These recent perspectives suggest that, in the design of a new school, or the retrofitting of an existing school, children's multi-sensorial experiences of the natural environment can be considered a priority design requirement to foster integral and beneficial goals of sustainable development (Bolten and Barbiero, 2020).

However, currently lacking in school design studies is an exploration of the relationship between spatial settings and children's connections with nature. Firstly, the established link between physical characteristics of school settings and the intensity of children's multiple sensorial connections with natural features (other than greenery), including what they hear, touch, taste and smell, suggests that further study is needed (Franco et al., 2017). A holistic understanding of multi-sensorial experiences of natural elements and stimuli can successfully deliver natural benefits for children because not all exposure to nature is pleasant (Ratcliffe, 2021) to children and their specific activities. This is one of the central issues of environmental psychology - how different architectural and natural attributes within a spatial environment provoke different affective states and interactions in various types of individuals, such as reactions of liking or disliking, engaging or averse, exciting or boring. Similarly, it is important for us to understand children's perspectives of ideal environments at primary schools to identify design features that can enhance their attraction to nature (Beheiry and Gabr, 2022; Ghaziani, 2021; Ghaziani et al., 2021).

Additionally, a research gap remains in a specific aspect of methodology. Although studies have shown the considerable benefits of potential sustainable or biophilic design strategies, findings were observed through students' academic performances, perceptions and satisfaction (Ghaziani et al., 2021; Peters and D'Penna, 2020; Tucker and Izadpanahi, 2017). These studies exclusively measured children's intellectual or emotional concepts, which largely depend on educational methods, while there was limited demonstration and exploration of natural-spatial experiences. Therefore, further research on children-nature connection can benefit from mixed approaches (or combined qualitative and quantitative methods) in order to advance our understanding of how children perceive and evaluate nature. These approaches can help to determine related factors and consequences of experiences associated with child–nature connectedness levels. For urban and architectural research aspects, Klaniecki et al. (2018) and Ives et al. (2017, 2018) shared similar suggestions for further studies: to utilize empirical explorations, investigate in different spatial scales and consider subjective meanings to understand the interactions and linkages between these factors.

To address these gaps in knowledge, we have undertaken an investigation into the association between the natural and physical characteristics of school settings located in different contexts, and have adopted child-centred analytical methods required for a comprehensive understanding of how school design can help to foster children's multi-sensorial experiences with nature (Deppeler and Aikens, 2020). The crucial question about primary school design emerges: “According to various spatial-natural features in primary school environments, what are differences in children's discovery of nature and feelings toward nature through visual and non-visual senses and their environmental preferences?” This investigation addresses limitations of previous interventions to increase nature connection in children in educational contexts, and follows a new approach to research schools' environments focusing on spatial configurations influencing the child–nature relationship (Barrable and Booth, 2020). Our aim is to support future decision-making for the design of primary school environments that can help to foster a reconnection between children and nature, which we consider an imperative. In summary, it is essential to follow these directions to provide practical implications for future sustainability and architectural research.

Responding to this imperative, we offer an exploration of how pupils (from 7 to 12 years old) experience nature, including the way they directly perceive nature through multiple senses, and how they express their environmental preferences at primary schools. Contextual differences allow us to identify the significance of particular site settings and architectural features. Here, we have adopted multiple case studies in various spatial scales and contexts of primary schools in Glasgow, Scotland with two main research objectives:

1. to explore children's multi-sensorial and affective dimensions toward perceived nature viewed as the basic core constructs of their natural consciousness and behaviour (Linzmayer et al., 2014). Distinct features of natural classifications were clarified through children's own perspectives, and particular features of natural environmental stimuli that compose sensory variety and variation explored (Moore, 1989; Sebba, 1991).

2. to understand children's demands and preferences with regard to nature and places at schools. The relationship between children's perspectives and spatial-natural features helps to encourage children to positively engage with nature. This requires exploration of where – children's favourite places, why and what for – their needs in the desired natural and built environments. Thus, an application of the child-centred method provides opportunities for children to express their environmental and spatial perspectives and interests increasing their subjective well-being (Mustapa et al., 2015).

Together these objectives define key factors influencing the child's multi-sensorial natural experiences, and provide insights into ways of reconnecting children and nature within an educational context.

Children were recruited from three private primary-schools of The Glasgow Academy (TGA) in Scotland, including Kelvinbridge (TGA_K), TGA_Milngavie (TGA_M) and TGA_Newlands (TGA_N). These schools were selected since children groups were taught with the same educational curriculum within considerably different learning environments (e.g. distinct locations in urban settings, spatial configurations and architectural features of classrooms and playground settings) (see Figure 1). With these differences, this study anticipated exploring the influences of school physical and landscape settings on child–nature multi-sensorial experiences through children's nature exploration and perception on nature. While acknowledging that the selected schools belong to a particular socio-cultural and economic context that will influence the child–nature connection, the researchers believe that the focus of this current study on applying a method for identifying sensory determinants of natural and non-natural spaces is transferable to educational architecture more widely.

TGA_K – is located in a high-density area of Glasgow; however, its site is surrounded by a river landscape and diverse green spaces. Three pupil groups, including TGA_K_7 (grade 7 – aged 11–12), TGA_K_4 (grade 4 – aged 8–9) and TGA_K_3 (grade 3 – aged 7–8), completed the natural exploration tasks within outdoor spaces that stretch from the entrance gate to the trim-trail playground area (K_a). The central hard-surfaced ground areas featured a flower-bed, flower-plant pots and wooden benches (K_c and d) while remaining areas had artificial grass-surfaced ground and higher degrees of greenness (K_b). Regarding three studied classrooms, the pupils of grades 3 (TGA_K_3_I) have outstanding naturalness values for visual and non-visual connections due to its large window system opening directly to outdoor greenspace while the classrooms of grades 4 (TGA_K_4_I) and 7 (TGA_K_7_I) have natural connections via looking and hearing only because of their closed windows during survey times.

TGA_M – is within a residential area in the suburb north of Glasgow; it offers substantially higher multi-sensorial naturalness than TGA_K and TGA_N due to a surrounding river landscape and a wilderness area which are parts of the school's facility. Children completed outdoor tasks in the area at the entrance gate (M_a) which is decorated with many flower-plants and surrounded by many high trees and bushes, and in the wild area (M_b) that is significantly featured with the earthy ground, high trees, bushes and diverse plants. Here, there was only one investigated classroom (TGA_M_I) where pupils TGA_M (grades 3 and 4, aged 7–9) participated and had only natural connections via looking and hearing. It is featured with a large area of glass-window at the side which is connected via an open-viewed corridor adjacent to outdoor space (e.g. bushes and wooden plants as hedgerows), a glass area above on another side and indooSr flower-plants for decoration.

TGA_N – places in the residential area of Mansionhouse Garden estate side and is covered by crossroads and private houses. There was only one group TGA_N (grade 3 pupils, aged 7–8) who did the task within their classroom and a central playground. Within this central hard-surfaced playground, each particular functional area is divided by different colours and play facility settings (N_b and N_c). The external landscape is designated with a hedgerow comprised of shrubs and high trees, some flower-plant pots and a flower bed for decoration (N_a). Consequently, fewer naturalness values and natural diversity for multi-sensorial experiences occur in this school playground. Regarding the studied classroom (TGA_N_I), it has two sides connecting directly to the outside through large glass-window systems at an approximate height (in closed position) for looking and hearing; besides, this room was featured with caterpillar glass boxes for indoor studying activities.

The participants were 75 pupils without special needs (53% males, 47% females) with all children being aged between 7 and 12 years (Mage = 8.92 years, SD = 1.50) as shown in Table 1.

This study received ethical clearance from the University of Strathclyde Ethics Committee. School approval, parental consent and pupil assent forms were obtained before surveys were undertaken.

The topic of (re)connecting children with nature has accumulated more interest within multiple disciplines. In line with current literature, research on this topic can benefit from combined qualitative and quantitative perspectives to help advance children's explanations of their perceptions and emotions toward nature (Beery et al., 2020; Chawla, 2020). Following these suggestions, this study integrates quantitative and qualitative analysis of data obtained by children's open-questions and diagrams to identify the matrix of natural elements and stimuli that children perceive within their classrooms and playgrounds, the descriptor terms for sensory notations based on their feelings and natural and non-natural features of their favourite to spaces. These methods were used to create fun environments that encourage children to work on tasks and express their views without pressure, and to offer sufficient information and in-depth details for analysis (Grant, 2017; Punch, 2002).

The data collection period was between April and May 2018, springtime in Scotland, UK Children had two working papers with a similar format: the classroom and playground tasks. The child-friendly format questionnaire was developed with support from an educational expert, working as a manager and teacher in the three schools, to ensure appropriate, clear and consistent communication with children. Children were encouraged to provide written statements and/or produce drawings that would illustrate their perception of natural elements. In keeping with previous research findings, self-selected participation methods varied (Literat, 2013). Generally, children with a less developed vocabulary, preferred to draw whilst other children, who were perhaps less confident in their drawing abilities, preferred to produce written responses. This method is appropriate for children at both 7–9 and 10–12 years old to match their personal preferences and in their familiar ways (Horstman et al., 2008). Thus, allowing the participants to decide whether in writing or drawing on the natural elements or stimuli, they could provide rich information about experiences of nature within a place. The children's paperwork was structured as follows:

1. Firstly, children responded to questions:

   • “What Nature do you see/hear/touch/taste/smell in your classroom/playground?”,

   • “What does it look like (colour/shape/action)?” and

   • “How do you feel?”

Pupils were asked to observe and identify names of natural objects corresponding with each sense, and then briefly describe their emotional feelings.

1. Following, questions related to pupils' favourite natural elements and places at their schools are:

   • “What parts of nature do you like most in the school playground and garden?”,

   • “What is your favourite place at your school?” and

   • “Tell me why you like this place?”

With the support of teachers, short meetings with students were held before conducting surveys within the studied contexts. Work with each participant group occurred within a period of 50 and 60 min, including 10 min for the introduction and guidelines, 15 min for children working in the classrooms, then 5–10 min moving from the classrooms to playgrounds and lastly, between 20 and 25 min for the outdoor tasks and collections. In classrooms, children positioned themselves as their regular daily activities. During working time, the researcher interacted and discussed with children whilst teachers only viewed from a distance. When pupils struggled and asked for help, the researcher had short conversations with them following the structure of the survey to avoid them having biased opinions under the influence of the researcher. These conversations started with the researcher's repeating the question and asking the child what nature he/she wanted to report. Then, the researcher asked questions about how a particular natural element looked like, and how he/she felt about it. When the child responded with words describing a characteristic of a natural element or stating their feelings, the researcher repeated this and guided him/her to write it down.

Firstly, a matrix of children's multi-sensorial experiences with nature and the descriptor terms for emotional notations is explored through the children's reports (see Figure 2 for examples of paper-works collected after surveys from children). This study argues that children have physical contact with nature through multi-sensorial experience. “Nature” is defined as natural settings, ranging from those that are independent of the human-built environments (a wilderness) to those that are within human-designed and constructed landscapes according to Kellert's descriptions of children's physical natural experiences (Kellert, 2002). The natural attributes, are synthesized from facets of the environmental education (National Association for Environmental Education U.K., 2016; Palmer, 1998), nature and landscape design (Dansereau, 1975; Williams, 1985) and biophilic design (Browning et al., 2014; Kellert, 2008; Kellert and Calabrese, 2015), include: Air, Astro-sky, Minerals, Water, Weather, Wind, Light, Fauna species and Flora species.

The natural elements that children reported in words or images were arranged into these natural attributes while responses that were incompatible with these designated natural attributes (for example, football (net), fence, climbing frame, rope, people, plane, litter, ear, bench, teddy, tea, desk and paper) were excluded from the data analysis. The perceived extent of each natural type is demonstrated through frequencies of the coding technique according to places and sub-sensorial experiences are illustrated in Figures 3 and 4.

Next, children's responses regarding their emotional feelings toward recognized natural elements, from the raw data, the relevant themes, short descriptions and representative keywords were recorded and categorized for coding according to the concept of the circumplex model of affective appraisals (Russell, 1980). These themes were categorized into pleasant-unpleasant responses ranging from high to low stimulus (as shown in Table 2), including surprise, excitement and joy, satisfaction, appreciation, relaxation and calmness, boredom, sadness, craving, disgust, anxiety and fear.

The final questions related to pupils' environmental preferences at their school were examined based on distinctive design features in relation to their own choice of words. A favoured environment was evaluated through children's concerns about physical, natural, emotional and social meanings of learning and recreational activities at schools. Indoor environments were treated as a composite group, whereas outdoor environments included sub-categories according to different features of playground areas within three primary schools, as shown in Figure 1.

Data management and the quantitative-qualitative analysis of data collected from children were performed using Excel and NViVo 11 programs.

The word cloud of perceived natural elements perceived by children at each school are generated by NVivo and presented in Figure 3. Then, the distribution and classification of nature according to particular sensorial experiences, obtained from children's reports, are presented in Figure 4 (the colours represent the frequencies of natural attributes children perceived through five senses within classrooms and playgrounds). From these figures, it can be seen that pupils reported higher numbers of natural elements and types of each and total sensorial experiences at playgrounds than in classrooms. Children in classrooms, with the exception of looking, reported more perception of nature via hearing and touching than tasting and smelling. At playgrounds, the greater natural values were experienced by touch whereas tasting was least noted. The results also revealed that flora and fauna species were the dominant attributes that children from three schools responded to, with 53.0 and 24.9% in total responses respectively. The results are circumstantially described as follows.

Regarding classroom results, the group of K_7_I pupils head the amounts of natural elements and diversity even though this classroom involves fewer advantages of direct experiences of nature than rooms of K_4_I and K_3_I. They explored nature primarily through looking and hearing and offered the highest ratios of plants and animals. With the advantages of opened windows confronted to greenspace and water landscape, weather conditions (with and without wind) and available indoor natural setting for study (“tadpoles”), the amounts of perceived natural elements and types that K_3_I pupils reported were much higher than children of K_4_I room. The notable outcome was a dominance of animal species and flora species in regards to senses of look, hear and smell.

Moving to outdoors, flora species was the most significant natural feature that was figured out by five senses, especially looking, touching and smell by both three groups. The oldest group K_7_O continued heading the capacity of exploring nature within their playground. They also highlighted in looking and hearing sound of dogs and birds. With surrounding water landscape, they not only saw and heard sounds of water features but also figured out mineral elements while younger groups paid more attention on the flower bed area, especially the K_3_O pupils who head the proportion of flora-related answers. Besides, the windy and gentle rainy situation during the survey time of K_3_O group influenced the variety of nature children could experience with additions of “rain drops” and “wind” that were perceived through both visual and non-visual senses compared to results of K_4_O group.

Pupils at Milngavie explored the least numbers of natural elements and types within classroom. The proportion of flora-related responses significantly lead the experienced natural types and primarily distributed into vision. Pupils here remarkably noted indoor-decorative flower-plants.

Within the entrance gate area (M_a), the results showed the dominance of the flora group that was mainly distributed into the exploration of looking, touching and smelling. Although moving to wilderness area (M_b), the total figure of perceived natural elements was slightly lower than within the entrance gate area. What stands out here was a greater numbers of insect species, more plentiful botany and mineral elements, that were experienced through looking, touching and smelling, than within the other sites. Additionally, “lake” and “river” were water features that children explored through all senses except smell.

Within the classroom of Newlands, two primary perceived natural groups were fauna (e.g. “birds”, “squirrel” and “caterpillar” boxes) and flora species (e.g. “leaf/leaves”, “wood” and “tree(s)”) that were mostly collated into looking, hearing and touching experiences.

Related to outdoor exploration, the discrepancy between visual and non-visual experiences was not significant like other TGA school cases despite the highest ratio of vision. Flora and fauna head the proportions of perceived natural types. Children had the greatest concerns of five sensorial experiences with trees, plants and flowers while responses of animals were mainly sorted into looking and hearing.

Overall, results showed pupils exhibit both pleasant and unpleasant feelings toward nature through direct multi-sensorial exposure within school environments and also indicated the dominance of positive feelings, including senses of surprise, excitement, satisfaction, appreciation, relaxation and calmness from children (Table 3).

Vegetations and animal species accounted for the highest percentages of positive responses according to their large frequencies of explorations. In particular, trees, plants with leaves and colourful flowers mostly generated senses of happiness, satisfaction and aesthetical appreciation through looking, touching and smelling. Pet-orientated species (e.g. dog, cat, birds and squirrels), and insects (e.g. butterflies, minibeasts, caterpillars, tadpoles, bug, ants, bees and worm) predominantly incur pleasant feelings of looking and hearing. With regard to water, most responses from Kelvinbridge and Milngavie pupils indicated a greater distinct attribute of the water element with a lower arousing degree than stimulus from other natural types. Children reported their relaxation and calmness when looking and hearing divergent moving states of water. In contrast, unpleasant feelings were described when they directly tasted, smelt or touched other natural elements, for instance, “tired” – smell a flower, “grumpy” - smell a leaf, “sick” – look at skunk cabbage, “sad” – look at the green bushes and brown leaves, “a bit scared” – see a woodlouse, “dirty” and “yucky” – touch, taste and smell soil. Through the children's descriptions here it might be deduced that sensorial emotions vary with greater natural diversity. And pleasant and unpleasant feelings, ranging from low to high arousal levels, are evoked when their explorations are unlimited.

Following the predominant positive feelings, flora and fauna groups were mainly favoured with 63.9 and 29.3% respectively. Among these, the beauty of colourful flowers were strongly identified. With regard to more unusual site conditions, river and brook were identified by pupils of Kelvinbridge and Milngavie. Warm weather was identified by some pupils expressing their feelings of comfort during the spring season; weather conditions and thermal comforts ultimately influences children's perceptions on natural favour. This is an important feature of experiences within natural environments in a cold climate and should be considered in designing schools since a windy space can diminish children's comfort and therefore their willingness to use outdoor spaces for hands-on experiences.

The results indicated that outdoor areas for various recreational activities are the predominately favoured places with 72.5% of the respondents. Regarding TGA_K, the oldest group K_7 had a significant inclination toward indoor spaces (65%); their least favourite option was the sport playground. Oppositely, approximately a half of younger children selected outdoor spaces, especially the adventurous playground (K_a). At Milngavie, a half of participants reported the main artificial grass-surfaced playground (M_c) as the place they like most, while a minority group (20%) selected the wilderness area (M_b). In Newlands, two different areas for playing outdoor, including the general setting for all pupils (N_b) and a particular setting for playing sports (N_c), accounted for equal importance (38%), while 23% of participants preferred indoor places, including classroom and other indoor spaces, such as auditorium, gymnasiums, library, ICT room and dining hall. Beyond the place designations, children's reasons for favouring particular places revealed that they selected a place for at least one reason regarding their comfort, interests, habits and the need for social interactions; including:

Collectively, the important results of children's environmental preferences are as follows:

1. Indoor environments: children demanded individual activities, specific functions and artificial facilities. They highly noted features of spaces for their comfort satisfaction (for example, big, warm, quiet, cosy and spacious) while natural concerns were less given.

2. Outdoor structured playing areas: children mostly stressed physical activities and interactions with others; they felt fun here. The challenges of playing were also mentioned by children who favoured trim-trail or adventured playground for non-natural facilities.

3. Wilderness areas: they mentioned specific activities, such as climb trees, explore and play loose parts there. Their descriptions showed the differences between wilderness spaces and outdoor developed spaces relating to higher creative, challenging levels; the former addressed that nature was playing facilities for their activities while the artificial settings were related to the latter.

In regard to differences in perceiving nature, the results show the older children were, they increasingly explored more various natural classifications. Particularly, within the TGA_K school where three groups explored, the oldest group heads the amounts of natural elements and diversity within indoor environments, even though their classroom involves fewer advantages of direct experiences of nature than rooms of younger groups as well within the same playground (as shown in Figure 3). In relation to gender, girls explored plants with flowers more than boys ( f_female = 23, f_male = 11), and it could be explained by their greater intensively attractions toward various types of colourful flower-plants within classrooms and playgrounds for decoration purposes.

Regarding feelings toward perceived nature, there were no differences between aged groups. However, female pupils showed higher preferences to express perceptions and attitudes toward nature in the given tasks; their responses were double that of boys. They also reported more positive feelings (e.g. senses of excitement, fun, satisfaction and appreciation, relaxing and calmness) than male pupils who described high arousal levels and unpleasant feelings toward their recognised natural objects, for example, through expressions of surprise, boredom, disgust and fear.

What stands out in children's spatial preferences is both age and gender differences regarding favoured features of playground settings. In particular, with 58.5% responses, pupils aged 11–12 showed higher preferences to indoor spaces than children aged 7–9 (only 20.7%) who mostly select outdoor spaces as their favourite places at schools. About gender difference, 31.4% female pupils favour adventurous and energetic activities with the grassed surface in Kelvinbridge school, or even nature-based challenges of the wilderness area in the Milngavie school, while only 12.5% male pupils selected this option.

This study examined the issues influencing child–nature multi-sensorial connections within primary schools. With respect to the main research question, the obtained results reveal how differently children experience nature through visual and non-visual senses within various spatial-natural features in primary school environments, from indoor to outdoor and from low to high naturalness spaces. Our findings contribute impacts of architectural features and landscape settings on pupils' natural explorations and environmental preferences.

Firstly, this study reveals the significant influences of the nearby urban configurations on the quantitative and qualitative degrees of children's natural exploration through visual and non-visual sensorial modalities, for example the wilderness area of Milngavie and profound urban nature of Kelvinbridge that compose greenery and water landscapes. Responding to the restraints within urban planning would require a prioritising of decisions pertinent to the selection of sites for new schools based on proximity distance and regular daily access to natural areas, especially limited conditions of urban school sites with small sizes, hard surfaced playgrounds or poor performance landscape design. This has been seen in a study by Fischer et al. (2019) that the neighbouring vacant wild sites were planned, managed and occupied to enhance students' gardening, recreational activities and food experiences at the Caspar David Friedrich secondary school. Proper pathways to connect nearby natural areas outside a school could be improved in the re-accommodation of functional layout, and by increase in built envelopes' permeabilities to maximize naturalness values.

At school playgrounds, the more wilderness children are exposed to, the greater numbers of natural objects and types they can explore. A significant difference between natural land typologies and hard-surfaced playgrounds relates to diversity of nature. For example, the earthy ground of Milngavie site offers diversity of underground flora, fauna and mineral elements that could not found at hard-surface grounds. Although fewer natural elements were reported in wilderness area than other places, the research saw Milngavie pupils spend more time to discover and observe natural objects (such as they overturned a log to find snails or woodlice), and discuss with friends during survey time instead of entirely focused on current tasks. Through discovery of new things, various pleasant and unpleasant arousal levels were reported toward living and non-living natural elements. Likewise, water landscapes provide rich sources for children's discovery of aquatic ecosystems, abiotic components and offer more diverse soundscapes. Among natural classifications, this study found that flora species are widely experienced and favoured features at school sites. Colourful flower-plants are the most appealing feature that attract children's attention and engagements with multiple senses, especially female pupils. Simultaneously, seasonal changes and weather conditions generate a variety of natural stimuli from incorporation of different elements. With variations in the colours and smells of vegetation, sounds of birds, movement of rivers and the rustling of leaves in the wind, the intensity of stimuli perceived varies, and children are able to experience more natural diversity through a comprehensive sensorial experience, even when in classrooms. Within the natural world, a child of primary-school age has an opportunity for unlimited to discovery, to know, to feel and then to place in memory the diversified and complex attributes of natural experience (Cudworth, 2021; Herrington and Lesmeister, 2006; Wilson, 1984). And green schools, if considered a potential part of the biophilic urban realm, can contribute to human health and wellbeing goals; the natural ground project “Urban Jungle” of Merrylee Primary School in central Glasgow offers an example.

Another important finding is the significance of nature-based playgrounds to the social dimension of children's favoured places. The playground is considered an important social setting (e.g. gender, class, race and identity) in educational contexts, and it's common that boys are dominant playground users while girls are often excluded from much of the area of a playground (Blatchford et al., 2003; Paechter and Clark, 2007). This might be explained by the design of typical playgrounds which are mostly concrete or asphalt-covered surfaces supporting games and activities, historically for boys and older groups. To provide an equitable and healthy playground for all pupils, greening is now considered as an appropriate approach (Dyment and Bell, 2008; Lucas and Dyment, 2010). The results of our study support this approach, in particular by identifying female pupils' inclination for outdoor activities within grassed surface, or earthy ground areas. We have found that younger girls favour adventurous energetic activities, suggesting there are advantages in designing playgrounds with a nature-based surfaced cover and playing facilities, in order to balance the distribution of gender and age ranges.

Through a detailed investigation of students' perceptions of the ideal environmental school according to their needs and desires, these results are in keeping with views of children and young people in previous studies (Aminpour, 2021; Burke and Grosvenor, 2003, pp. 45-53); the playground and school landscape must be for their playing and adventurous activities. However, current design regulations for school buildings provide constraints, due to safety concerns, that result in schools that are often over-designed, over-structured and over-artificially equipped in order to control and monitor children's activities (Malone and Tranter, 2003). The gap between children and nature has been increasing as a result, and opportunities for children's non-visual experiences with nature have been diminishing. Our findings support a nature-based playground settings in which natural diversity must be yielded as an approach to nurturing children's habits and interests in nature, in a broader sense, that can help narrow the widening gap between humans and nature (Grierson, 2009).

Firstly, the influences of architectural features and spatial layout within buildings and school sites were illustrated through children's reports. For instance, their exploration of nature is enhanced within classrooms with windows facing outdoor nature. This can be seen in the cases of N_I, K_3_I, and K_4_I classrooms where varieties of plants and animal species were highly noted. In contrast, a more adverse location of M_I revealed an insufficient connection between pupils and outdoor nature. Furthermore, this study explored additional conditions of openings' features (i.e. scale, orientational position, height and operational status) on the degrees of nature pupils perceived through five senses within spatial conditions with adequate openings to outside, vision is the strongest and the most efficient connection between children and nature from indoor places compared to non-visual sensorial experiences. Thus, the inclusion of permeable features in building envelopes and the spatial layout can efficiently establish pathways for children's multi-sensorial natural experiences within indoor environments. Responding to this, adaptable and flexible features in learning arrangements (Schabmann et al., 2016) and building elements (i.e. flexible canopy (Kwon et al., 2019) or moveable shading elements (Kükrer and Eskin, 2021)) can further be applied to alter the visual and non-visual permeabilities in specific situations, especially adverse weather conditions or unfavourable temporary factors in surrounding environments. Advanced technology innovations in architectural design allows for these accessible pathways to be implemented, with improved environmental performance.

Additionally, one of the most important findings is related to children's attention to indoor natural elements. Experiential, and regular, learning related to nature evokes their awareness through five senses, particularly through touching and smelling. The results of pupils' emotional reactions, especially female groups, showed greater biophilic feelings toward species which are often considered as “disgusting”, for instance, caterpillars, worms and tadpoles (Almeida et al., 2014; Matchett and Davey, 1991). This positive alteration of children's perception could be explained by an increased familiarity within the studying environments. Results suggest that there are advantages in reconsidering indoor natural settings to upgrade pupils' direct experiences and nurture their biophilic feelings toward a wide range of natural elements.

Although the aged range was not equally distributed among three schools, the results of children's favourite places have shown that the late middle childhood (9–12 years old) have an increased trend to favour indoor environments than the early middle childhood (7–9 years old). These results support the steady changes of preference for non-natural and social needs over nature as indicated in previous studies (Aminpour, 2021; Liefländer et al., 2013; Müderrisoğlu and Gültekin, 2013), and findings that gaps between the child and nature also increasingly expand with age progression (Lyons, 1983). This suggests that different strategies need to be adapted to particular groups for enhancing direct natural experiences. For example, further investigations should be focused on other indoor functional spaces beyond classrooms for older students, such as the library, dining hall, ICT room, auditorium and gymnastic rooms, in order to apply design strategies to entire school environments.

Taken together, the findings of this study support the benefits of nature-based settings in school design, including indoor and outdoor environments, to children's multi-sensorial experiences, knowledge and biophilic feelings toward nature. The work is presented in the belief that nature can become an intrinsic driver of children's educational activities. Correspondingly, natural elements and stimuli can be physical and aesthetic components in the social and environmental evolution. When biophilic design approaches are applied to entire school environments, children's negative feelings toward unfamiliar and uncommon natural species can alter, our younger generations' environmental attitudes and behaviours towards the natural world can change, and we might be better placed to bridge gender differences. In particular, following the global Covid-19 pandemic, providing enhanced experiences of nature offers an important means of increasing children's well-being (Friedman et al., 2022) and environmental awareness (Rios et al., 2021). This study therefore attempts to contribute to the growing body of research that informs these approaches.

This research has identified and described key issues influencing children's multi-sensorial natural explorations, and their environmental preferences at primary schools. Our findings suggest that applications of the biophilic design, natural diversity and the provision of accessible and flexible pathways, are potential approaches to promoting children's multi-sensorial experiences of nature within their studying and playing environments. We believe these findings, and suggestions for primary school design, can support future policies and guidelines. Children's direct involvement in this research, through the application of a mixed methodology, has allowed us to better to understand their particular needs and spatial experiences (Hofmann, 2014, pp. 40-41), and the extent to which the relationship between their own perceptions and the physical-natural features surrounding them brings nature closer to them and encourages their engagement with it.

However, a limitation is that the data was gathered in the spring season only; thus, there is a need for further investigation on children's perception and preferences based on seasonal variations. Future research should also explore sub-sets of different educational-socio-cultural and economic factors to examine how these influence the child–nature connection, and inform educational architecture. These directions would broaden the present study for the decision-making process and identifying design parameters for sustainable architecture (Grierson and Moultrie, 2011; Grierson and Salama, 2016). Today's architect is challenged by new concepts on educational methods and the impacts of new emerging technologies, by complex, disordered and uncertain situations (Salama, 2020). Requirements of spatial and functional settings for children's health and wellbeing at school environments following the COVID-19 pandemic are likely to result in sweeping changes in the relationship between children and nature (Rios et al., 2021). In this sense, urban and architectural designers in general, are challenged and compelled to offer adaptable approaches, as Forsyth (2020) argues:

For this reason, the need for further multidisciplinary research on how we can maintain and enhance our children’s connection with nature through informed, well designed, changes in the built environment, is both vital and urgent.

This research was assisted by The Glasgow Academy in Glasgow, Scotland. The authors would like to express their gratitude to all staff, teachers, pupils and their parents of three primary-schools Kelvinbridge, Milngavie and Newlands for their participation and support during the investigation. The authors would like to acknowledge the financial support of the Ministry of Education and Training of the Vietnamese Government for a Ph.D. funding at The University of Strathclyde (Glasgow, Scotland, United Kingdom).