Abstract
Common spaces in universities are crucial for the psychological well-being of university students, as they serve as spaces for their daily activities. This study aimed to discuss the factors and pathways that influence the characteristics of common spaces in universities with respect to the psychological restoration of university students. We selected 11 universities in Guangzhou, China for the survey: SCUT Wushan Campus, SCNU Shipai Campus, JNU Shipai Campus, SYSU South Campus, SCNU University Town Campus, GDUFS South Campus, SCUT University Town Campus, GDUT University Town Campus, GZHU University Town Campus, GZIC(SCUT) and HKUST(GZ), all of which are located in the Tianhe, Haizhu, Panyu and Nansha districts of Guangzhou. A questionnaire survey was conducted among 662 university students from the above 11 universities and 630 valid questionnaires were ultimately collected. Structural equation modelling (SEM) was employed to verify the hypotheses and assess the net effects of the variables, and fuzzy set qualitative comparative analysis (fsQCA) was used to explore the configurational patterns of high-level psychologically restorative effects. The results of the empirical research indicated the following: (1) the characteristics of the architectural scale and aesthetics, landscape richness, availability of rest facilities and compatibility of activity facilities within the university common spaces had a significant and direct effect on the psychological restoration of university students; (2) the three behavioural patterns of contact with nature, physical activity, and social interaction had a significant mediating role between the characteristics of university common spaces and psychological restoration of university students; (3) the fsQCA results revealed that four distinct patterns—integrated patterns of environment and facilities, static rest patterns, dynamic exercise patterns and interaction-driven patterns—each facilitated high levels of psychological restoration. These findings elucidate the mechanisms by which common space characteristics affect university students’ psychological restoration and offer valuable insights into campus design.
Supplementary Information
The online version contains supplementary material available at 10.1038/s41598-025-95771-8.
Keywords: Healthy environmental design, University common spaces, Psychological restoration, Behavioural activities, Green campus, SEM and FsQCA
Subject terms: Psychology, Environmental sciences, Environmental social sciences, Health care
Introduction
According to statistics from China’s Ministry of Education, the number of universities in the country increased from 1,002 in 1998 to 2,263 in 2008. As leaders in the reform and opening up, many existing universities in the Guangzhou region were constructed during this period. However, owing to constraints imposed during the construction cycle, the spatial layouts of university campuses in Guangzhou is monotonous. The common spaces on these campuses have gradually become fragmented, leading to the emergence of numerous negative spaces.
Relevant research has found that university students are prone to psychological distress due to academic pressure and anxiety. According to the statistical results of the 2022 Survey Report on University Students’ Mental Health in China, nearly 90% of university students experience psychological distress due to academic pressure1. Consequently, university students expect more psychological support from campus environments to alleviate stress and release negative emotions. However, with the development of university campuses, the design issues from earlier phases have not been rectified in accordance with the psychological needs of university students.
Common spaces serve as effective resources to facilitate the physical and psychological restoration of university students. They enhance the quality of life, sense of attachment, mental and physical health and social skills of students2. In the planning and design process, university common spaces are formed by the surrounding campus buildings, creating spatial nodes of varying scales and configurations. The quality of these spaces is enhanced through landscape design, which includes facilities for rest and other activities3. Examples include vegetation, waterfronts, squares, courtyards and sports fields. Therefore, the characteristics of university common spaces consist of elements relating to architectural and landscape environments and supportive facilities. These elements are correlated with attention restoration, stress alleviation and enhanced psychological health outcomes4,5.
However, research on the factors influencing psychological restoration associated with university common spaces is still in its preliminary stages. Moreover, such research has often overlooked how university students’ behavioural patterns within these environments affect their psychologically restorative outcomes. An in-depth analysis of the relationships and pathways among spatial environmental characteristics, students’ behavioural activities and psychological restoration is also currently lacking. Therefore, effective strategies for optimising the design of these common spaces have been neglected.
In previous studies, regression-based methods such as multiple regression analysis and structural equation modelling (SEM) have typically been employed to measure the psychologically restorative outcomes of university common spaces6,7. These methods focus solely on the net effects of individual variables, making it difficult to explain multiple concurrent causal relationships and equivalence issues resulting from the interdependence of various variables. In contrast, fuzzy-set qualitative comparative analysis (fsQCA) addresses these limitations by offering a framework for managing high levels of causal complexity, thereby complementing traditional correlational analyses8. FsQCA can assist in identifying sufficient causal conditions that contribute to outcomes and reveal the configurational effects of various causal pathways, which is an important dimension that traditional statistical analyses have not adequately addressed9. In addition, the restorative effects of common spaces in universities have a complex causal relationship, with the variables in the model affecting each other. As such, the fsQCA offers a more comprehensive solution for understanding these multifaceted variable interactions. Therefore, in this study, traditional variable-based SEM was integrated with the novel fsQCA to analyse the data. The effective combination of these two methodologies enhances the descriptive, predictive and explanatory power of scientific theories10.
This study systematically developed a model that identified factors influencing psychologically restorative outcomes. Using common spaces from 11 university campuses in Guangzhou as case studies, we employed SEM to validate the model and analyse the influencing pathways of the variables. Furthermore, fsQCA was conducted to examine the configuration of conditions that facilitated high levels of psychologically restorative outcomes. Our research introduces a quantitative approach to measuring environmental characteristics and psychological assessments in university campus designs. This approach transforms qualitative environmental features into quantitative indicators, enabling the study of nonquantifiable factors in a measurable manner and establishing effective pathways for promoting psychological recovery through environmental design. Our study has significant practical implications for improving the quality of university campus environments and promoting student mental health.
This study aimed to address the following questions: (1) What are the characteristics of university common spaces that facilitate psychological restoration in university students? (2) What are the influencing relationships and pathways among university common space characteristics, student behavioural activities and psychologically restorative outcomes? and (3) Which configurational model can be developed to achieve optimal psychologically restorative outcomes? By investigating these questions, we elucidated the pathways through which psychologically restorative outcomes operate. Our study offers targeted optimisation strategies to improve the quality of campus environments, thereby providing design advice for creating healthier university campuses.
Literature review
Restorative environment theory and practice
There are two classic theories relating to the effects of restorative environments on individuals: Attention Restoration Theory (ART) and Stress Reduction Theory (SRT). ART, proposed by Kaplan (1995), emphasises that restorative environments typically possess four characteristics: being away, extent, fascination and compatibility11. When individuals experience an environment embodying these four characteristics, their depleted attention can be partially restored12. Ulrich’s SRT posits that distinct stressful experiences can result in a decline in an individual’s cognitive ability13. However, when an environment encompasses specific positive factors, it can facilitate effective relaxation of individuals within the environment, reduce stress and foster a transformation from negative to positive emotions. These two theories complement each other and provide important measures for the psychologically restorative outcomes of spatial environments.
The concept of restorative environments has increasingly influenced the design of western university campuses, resulting in notable outcomes. For instance, American university campuses prioritise dynamic interactions and multicultural integration in their restorative environmental designs, emphasising cognitive restoration and creativity stimulation, while encouraging university student engagement with nature and physical activities14,15. In contrast, Chinese university campus designs are shaped by traditional educational philosophies and a highly competitive atmosphere that tends to favour functional and orderly designs that predominantly enhance restorative experiences through visual aesthetics16. Unfortunately, the significance of environmental restoration is not adequately interpreted or understood in China. Research indicates that highly stressed populations prefer environments characterised by strong natural elements and a sense of refuge17. However, the environmental design of China’s university campuses frequently neglects these preferences and instead relies on functional indoor spaces, such as dormitories and libraries.
University common space characteristics
There is a lack of research about the restorative effects of architectural environments within university campuses. To some extent, the influence of built environment characteristics on psychological restoration may be mediated by the concept of ‘being away’18,19. For instance, the ratio of the width (D) to height (H) of building enclosures can enhance an individual’s positive experience within a space, specifically when D/H ranges from one to two20. Additionally, the openness and arrangement of building enclosures can directly influence the perceived potential for restoration21. The historical elements of the architectural environment can draw an individual’s attention away from their daily needs22,23. Furthermore, variations in building facades can elicit a range of emotional experiences that positively or negatively affect an individual’s emotions24,25. For example, the diversity of building facade decorations in terms of both quantity and style enhances the richness of spatial visuals, evoking a stronger sense of fascination and being away, which is positively correlated with the potential for psychological restoration26. Overly dense building facades can intensify negative emotions, including frustration, tension and disappointment27. Previous studies have indicated that specific architectural characteristics can enhance psychological restoration28, and such research insights can be used to analyse the impact of built environment characteristics on the psychological restoration of university students in common spaces on campuses. As such, it is important to carefully consider design elements (such as the scale and configuration of building enclosures, historical components and building facades) in university common spaces, particularly within courtyards and plazas.
Most studies have indicated that a rich landscape positively affects the psychological restoration of university students29. The quantity and coverage of vegetation on university campuses are significantly and positively correlated with university students’ psychological well-being and academic performance30,31. In addition, a diverse array of plant species can effectively capture student attention and alleviate psychological stress32,33. Green and yellow plants alleviate attentional fatigue and improve work efficiency34–36. Expansive lawns are considered to be positive restorative visual components37,38. The area, shoreline length and morphology of water features on university campuses significantly enhance spatial richness, demonstrating clear effects on alleviating stress and reducing mental fatigue4,39.
Recent evidence suggests that both the quantity and quality of rest facilities positively influence psychological restoration among university students40,41. For instance, an ample number of comfortable seats can provide students with greater opportunities to rest, thereby enhancing their willingness to relax and relieve stress42. Facilities oriented towards natural landscapes can extend the duration and frequency of students’ contact with nature while increasing social opportunities, thus promoting emotional regulation and psychological restoration43. Furthermore, seating that is free from environmental disturbances and maintains a high level of cleanliness can also facilitate resting behaviours among university students to a certain extent3.
A limited number of studies have indicated that recreational venues can promote dynamic exercise behaviour among university students, while high-quality sports facilities contribute to psychological enjoyment and relaxation for university students44,45. The maintenance of recreational venues and fitness facilities after their establishment can encourage university students to engage in physical activity, thereby enhancing positive emotions43. Convenient access to recreational venues provides opportunities for students to exercise, and a diverse range of activity spaces with well-equipped fitness facilities can enhance the choice and duration of physical activity among university students, significantly promoting psychological restoration3.
University students’ behavioural patterns
The likelihood of engaging in leisure and recreation, socialising with friends, or participating in outdoor physical activities depends on the characteristics of the built environment46. The enclosure characteristics of the built environment in common spaces, along with human scale and openness, can enhance an individual’s physical activity levels and promote social interactions such as walking and conversing47,48. High-quality building facades, historical architecture, and inviting scales are critical for improving the quality of the walking experiences33,49–51. Simultaneously, the quality of the landscape environment can foster health-promoting behaviours, including interactions with nature, physical activity and social interactions52–55. For example, campus water landscapes featuring natural shorelines facilitate activities that connect university students to nature, such as reading, meditation and enjoying scenery, whereas landscapes with increased amounts of hard paving are more likely to attract university students to participate in physical activities, such as ball sports and running56. Furthermore, aesthetic water features and higher vegetation coverage can mitigate the urban heat island effect while promoting social activities57,58. Moreover, the availability of rest facilities can increase the number of visits to these space59. The rest facilities, which provide convenient views and high comfort, can promote the duration and frequency of university students’ interactions with nature and increase their social opportunities14. The compatibility of activity venues supports and encourages physical activity behaviours among university students60.
Researchers have recently examined the potential relationship between physical activity and psychological restoration. Engaging in sports creates a psychological distance or sense of ‘being away’, which differs from the daily environment, thereby facilitating psychological restoration61. In this respect, university students participating in outdoor sports can feel energised and unaffected by stress62,63. A recent study highlighted contact with nature as an important means of alleviating loneliness and enhancing mental health64. Contact with nature can further reduce psychological stress among university students, particularly in relation to exam-related pressure and the anxieties of campus life65. University students’ engagement in activities such as reading, meditating and observing plants and animals in campus green spaces can enhance their scholastic performance and well-being66. Additionally, social activities such as walking, conversing, gathering and recreational leisure with friends are considered restorative behavioural activities67–69. These activities are defined as ‘the restoration of relational resources’, and they contribute to emotional management and the maintenance of mental health61.
Psychologically restorative outcomes
The primary objective of psychological restoration is to improve mental health, and this can be accomplished by enhancing the restorative effects of the campus environment to alleviate mental health issues among university students70. The different characteristics of common university spaces can have either positive or negative effects on students’ psychological restoration within these spaces28. Consequently, ‘the facilitating (or hindering) effects of the environmental characteristics of university common spaces on the psychological restoration of university students’ can be defined as ‘psychologically restorative outcomes’3.
Previous studies on psychological restoration have primarily focused on measurement indicators such as ‘attention restoration, psychological stress relief, regulation of negative emotions, recovery of vitality and a reduction in feelings of loneliness’ to evaluate psychological restoration outcomes among university students36,71. Attention restoration is a critical criterion used to assess the effectiveness of restorative environments11. In learning spaces, sustained focus or the prolonged use of directed attention can hinder the ability of university students to suppress distractions and maintain concentration72. Restorative environments can help university students regain their depleted attention and elicit positive psychological responses73. Stress relief is an important indicator of psychologically restorative outcomes13. Spatial design elements can significantly alleviate psychological stress by providing opportunities for rest and revitalisation74. The adjustment for negative emotions typically involves the restoration of positive emotions. When environmental factors align with expectations or needs, university students are likely to exhibit positive emotional responses45.
The recovery of vitality and a reduction in loneliness are recognised as key indicators of psychological restoration and are employed to assess psychological well-being7,75. A strong positive correlation exists between revitalisation and a reduction in loneliness. Restoring an individual’s mental vitality can enhance their emotional state, thereby increasing their willingness to engage with others and diminishing feelings of loneliness76. Conversely, reducing loneliness can alleviate anxiety and depression, which promotes vitality. Environmental designs aimed at facilitating the revitalisation of university students and reducing their feelings of loneliness should aim to improve areas situated near sources of stress, such as study areas, to create opportunities for psychological restoration77.
Model construction
According to the above literature review, positive environmental characteristics are a fundamental requirement for achieving a ‘restorative environment’, which not only provides opportunities for recovery, but also promotes recovery activities and experiences61. Specifically, the characteristics of common spaces on university campuses include the following: building enclosure scale and openness, architectural historical atmosphere, building facade design, the quantity and diversity of plants and their colours, lawn coverage, ornamental waterscape, the number and orientation of rest facilities and the comfort, privacy and hygiene they provide, the quantity and types of activity fields and their accessibility, the number of fitness facilities and the maintenance of these facilities. These 20 characteristics can significantly influence the psychological restoration of university students. High-quality restorative environments also provide adaptive opportunities for behavioural engagement66,78. Overall, the effect of spatial environments on psychologically restorative outcomes is mediated by greater contact with nature, more frequent outdoor physical activities, and increased social interactions79,80. The characteristics of common spaces on university campuses (the architectural scale and aesthetics, landscape richness, availability of rest facilities and the compatibility of activity facilities) can directly affect the psychological restoration of university students. These characteristics can also indirectly influence psychological restoration through the mediating effects of university students’ behavioural activities (such as contact with nature, physical activities and social interaction).
We proposed four sets of research hypotheses and constructed a hypothetical model based on these hypotheses (Fig. 1). The hypotheses are as follows:
Fig. 1.
Hypothetical model.
H1: The characteristics of university common spaces (architectural scale and aesthetics, landscape richness, availability of rest facilities and compatibility of activity facilities) have a direct and positive effect on the psychological restorative outcomes of university students.
H2: The characteristics of university common spaces have a direct and positive effect on the behavioural patterns (contact with nature, physical activity, and social interactions) of university students.
H3: University students’ behavioural patterns (contact with nature, physical activity, and social interactions) have a direct and positive effect on psychologically restorative outcomes.
H4: The characteristics of university common spaces have a positive effect on psychologically restorative outcomes through the mediating role of university students’ behavioural patterns.
Based on the proposed hypothetical model, the four research hypotheses were refined into four distinct groups (Fig. 2). Prior research on restorative environments offers substantial evidence supporting our hypotheses. Nevertheless, research specifically addressing the restorative aspects of common spaces on university campuses remains limited. Previous studies have primarily focused on the influence of landscape environmental characteristics on psychological restoration while often overlooking the restorative potential of architectural features within the campus29–31. Few studies have examined the effects of rest and activity facilities on the psychological restoration of university students3. Accordingly, we provide evidence to verify and supplement the psychological restorative effects of university common spaces from four aspects: architectural scale and aesthetics, landscape richness, availability of rest facilities and compatibility of activity facilities.
Fig. 2.
Hypothetical paths.
Research methodology
Research cases
This study was guided by the overarching objective of enhancing campus environmental quality as outlined in the ‘14th Five-Year Plan’ for the Development of Higher Education in Guangzhou (2022)81. As shown in Fig. 3, we selected the common spaces of 11 universities in Guangzhou, Guangdong Province, China as case studies. These included SCUT Wushan Campus, SCNU Shipai Campus, JNU Shipai Campus, SYSU South Campus, SCNU University Town Campus, GDUFS South Campus, SCUT University Town Campus, GDUT University Town Campus, GZHU University Town Campus, GZIC(SCUT) and HKUST(GZ). Approximately 100,000 university students live on these campuses. The selected cases included both traditional and newly constructed university campuses featuring a comprehensive range of common space types in five typical categories: vegetation, waterfronts, squares, courtyards and sports fields. Furthermore, these cases reflected the construction standards of universities and demonstrated significant representativeness.
Fig. 3.
Research cases from 11 universities. (Figure source: figure created by the first author; The map was created using Qgis 3.30 (open source); The software download link is https://www.qgis.org/download/).
Questionnaire design
The questionnaire consisted of four sections. The first section was the scale of university common space characteristics, which encompassed four dimensions comprising 20 items: architectural scale and aesthetics, landscape richness, availability of rest facilities and compatibility of activity facilities. The second section comprised the student behaviour pattern scale, which was predominantly informed by prior research3. This included the adjustment and redevelopment of certain measurement items according to the behavioural characteristics of university students. This section encompassed three dimensions with nine items: contact with nature, physical activity, and social interaction. section was the scale used to assess psychologically restorative outcomes. Based on the existing literature, five measurement items relating to psychological restoration were identified: restoration of attention, relief of psychological stress, regulation of negative emotions, recovery of vitality and reduced feelings of loneliness. Descriptive statistics for gender, age and grade level of the surveyed university students were calculated.
All measurement scales employed a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). A 7-point Likert scale is typically preferred over a 5-point scale because it is easier to manage and offers a more precise reflection of respondents’ evaluations82. In June 2023, we distributed questionnaires in the common spaces of 11 university campuses as a preliminary study and successfully collected 120 valid responses. We conducted reliability and validity tests on the preliminary survey data. The Cronbach’s α coefficient was 0.911, indicating that the preliminary survey questionnaire had high reliability. Exploratory factor analysis was performed on all measurement indicators. The Kaiser-Meyer-Olkin (KMO) statistic was 0.826 and Bartlett’s test of sphericity was considered significant at the 0.001 level. All eigenvalues were greater than 1, and the cumulative variance explained was 70.229%. The factor loadings of all the observed variables on the corresponding latent variables exceeded the standard of 0.5, indicating that the preliminary survey questionnaire had good construct validity. Therefore, we proceeded with the distribution of the formal survey questionnaires.
This study was approved by the Ethics Committee of the School of Architecture at South China University of Technology. Before completing the questionnaire, all university students were informed of the study’s objectives, processes, benefits and risks. All the participating university students provided informed consent. We confirmed that all methods were conducted in accordance with the relevant guidelines and regulations.
Data collection and analysis
In October 2023, a formal questionnaire survey targeting university students in common spaces across 11 campuses was conducted using randomly distributed paper questionnaires. A total of 662 questionnaires were distributed, of which 630 were valid, resulting in a recovery rate of 95.2%. The completed questionnaires indicated a reasonable distribution of respondents in terms of gender, age and grade, suggesting that the overall sample structure was sound and representative (Table 1).
Table 1.
Descriptive statistics of basic information of university students responding to questionnaires.
| Items | Category | Sample size | Proportion (%) |
|---|---|---|---|
| Gender | Male | 371 | 58.9 |
| Female | 259 | 41.1 | |
| Age | 18–20 | 345 | 54.8 |
| 21–24 | 236 | 37.5 | |
| 25–27 | 38 | 6.0 | |
| ≥ 28 | 11 | 1.7 | |
| Grade | First to second year undergraduate students | 337 | 53.5 |
| Third year and above undergraduate students | 235 | 37.3 | |
| Graduate students | 43 | 6.8 | |
| Doctoral students | 15 | 2.4 |
This study used SEM and fsQCA for data processing and validation of the conceptual model, and these two methodologies emphasise distinct aspects of the analysis. Structural equation modelling (SEM) is a variable-oriented technique that focuses on the net effects of independent variables on dependent variables. By contrast, fsQCA integrates qualitative and quantitative assessments to determine the extent to which cases belong to a specific set, thereby identifying the combination of conditions required for the occurrence of outcomes83–85. We combined these two methods to investigate the psychologically restorative outcomes of university common spaces, structured in the following steps: first, we used SEM to quantitatively assess the influencing factors and pathways, while also examining the mediating role of student behaviour patterns. Subsequently, we applied fsQCA to analyse the relationships among variable combinations, thereby identifying the antecedent conditions that contribute to high levels of psychologically restorative outcomes in university common spaces.
Data analysis and results
Results of SEM analysis
Normality tests
A normal data distribution is an essential prerequisite for parameter estimation using the maximum likelihood method (MLM). Previous research indicates that for sample sizes exceeding 300, the normality test relies on the dependent values of skewness and kurtosis. Specifically, an absolute skewness coefficient below two and an absolute kurtosis coefficient below seven can serve as reference values for assessing a normal distribution86,87. We performed skewness and kurtosis analyses of the questionnaire items using SPSS version 26.0. The absolute values of the skewness coefficients ranged from 0.008 to 0.663, whereas those of the kurtosis coefficients ranged from 0.048 to 1.581 (Supplementary Table S1). Both sets of values were below their respective standard thresholds. Furthermore, we employed Q-Q plots for validation, where all probability points were uniformly distributed around the 45-degree line (Supplementary Figure S1), suggesting that the sample observations aligned closely with the assumption of a normal distribution. This permitted the use of the MLM for parameter estimation.
Reliability and validity analysis
Reliability and validity tests on the collected data primarily assessed the reliability and accuracy of the scale, which was a crucial prerequisite for subsequent analyses. The survey data were analysed using SPSS software (version 26.0) for reliability and validity testing (Table 2, 3, 4). In questionnaire surveys, reliability testing typically relies on Cronbach’s α coefficient; a coefficient between 0.8 and 0.9 signifies excellent scale reliability. As indicated in Table 3, the Cronbach’s α coefficients for each latent variable exceeded 0.8, and the Corrected Item-Total Correlation (CITC) coefficients for each item with the overall sum were greater than 0.4. This suggested that the design of the observed variables for each latent variable was well structured, thus confirming the reliability of the survey questionnaire results.
Table 2.
Test results of exploratory factor analysis a.
| Latent variables | Observed variables | Component | |||||||
|---|---|---|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | ||
| F1 Architectural scale and aesthetics | A1 Appropriate scale of building enclosure | 0.787 | |||||||
| A2 High openness of building enclosure | 0.804 | ||||||||
| A3 Diverse forms of building enclosure | 0.782 | ||||||||
| A4 Strong architectural historical atmosphere | 0.772 | ||||||||
| A5 Varied building facades | 0.735 | ||||||||
| F2 Landscape richness | A6 Large number of plants | 0.757 | |||||||
| A7 Abundant plant types | 0.727 | ||||||||
| A8 Rich plant colours | 0.822 | ||||||||
| A9 Extensive lawn coverage | 0.785 | ||||||||
| A10 Highly ornamental waterscape | 0.789 | ||||||||
| F3 Availability of rest facilities | A11 Plentiful rest facilities | 0.840 | |||||||
| A12 Rest facilities with a view | 0.836 | ||||||||
| A13 Comfortable rest facilities | 0.859 | ||||||||
| A14 Good privacy of rest facilities | 0.832 | ||||||||
| A15 Hygienic rest facilities | 0.793 | ||||||||
| F4 Compatibility of activity facilities | A16 Plentiful activity fields | 0.831 | |||||||
| A17 Abundant types of activity fields | 0.798 | ||||||||
| A18 Accessible activity fields | 0.791 | ||||||||
| A19 Sufficient number of fitness facilities | 0.725 | ||||||||
| A20 Good level of maintenance of activity | 0.823 | ||||||||
| F5 Contact with nature | B1 Appreciation of animals and plants | 0.806 | |||||||
| B2 Sitting and reflection | 0.809 | ||||||||
| B3 Reading and learning | 0.788 | ||||||||
| F6 Physical activities | B4 Field activities | 0.725 | |||||||
| B5 Facility activities | 0.815 | ||||||||
| B6 Free activities | 0.780 | ||||||||
| F7 Social interaction | B7 Walking and conversing | 0.797 | |||||||
| B8 Chat gathering | 0.831 | ||||||||
| B9 Recreational interaction | 0.765 | ||||||||
| F8 Psychologically restorative outcomes | C1 Restoration of attention | 0.664 | |||||||
| C2 Relief of psychological stress | 0.657 | ||||||||
| C3 Regulation of negative emotions | 0.736 | ||||||||
| C4 Recovery of vitality | 0.777 | ||||||||
| C5 Reduce feelings of loneliness | 0.811 | ||||||||
Extraction method: principal component analysis.
Rotation method: Kaiser normalised maximum variance method.
a. The rotation converged after 6 iterations.
Table 3.
Test results of reliability, validity and confirmatory factor analysis of the hypothetical model.
| Latent variables | Observed variables | Standardised factor loadings | CITC | C.R. | AVE | Cronbach’s α |
|---|---|---|---|---|---|---|
| F1 Architectural scale and aesthetics | A1 Appropriate scale of building enclosure | 0.769 | 0.703 | 0.869 | 0.572 | 0.865 |
| A2 High openness of building enclosure | 0.822 | 0.736 | ||||
| A3 Diverse forms of building enclosure | 0.801 | 0.706 | ||||
| A4 Strong architectural historical atmosphere | 0.711 | 0.678 | ||||
| A5 Varied building facades | 0.668 | 0.636 | ||||
| F2 Landscape richness | A6 Large number of plants | 0.836 | 0.764 | 0.894 | 0.628 | 0.888 |
| A7 Abundant plant types | 0.814 | 0.731 | ||||
| A8 Rich plant colours | 0.760 | 0.740 | ||||
| A9 Extensive lawn coverage | 0.714 | 0.681 | ||||
| A10 Highly ornamental waterscape | 0.832 | 0.776 | ||||
| F3 Availability of rest facilities | A11 Plentiful rest facilities | 0.827 | 0.776 | 0.908 | 0.663 | 0.905 |
| A12 Rest facilities with a view | 0.830 | 0.775 | ||||
| A13 Comfortable rest facilities | 0.839 | 0.801 | ||||
| A14 Good privacy of rest facilities | 0.810 | 0.760 | ||||
| A15 Hygienic rest facilities | 0.762 | 0.714 | ||||
| F4 Compatibility of activity facilities | A16 Plentiful activity fields | 0.865 | 0.796 | 0.892 | 0.626 | 0.883 |
| A17 Abundant types of activity fields | 0.816 | 0.756 | ||||
| A18 Accessible activity fields | 0.861 | 0.774 | ||||
| A19 Sufficient number of fitness facilities | 0.632 | 0.603 | ||||
|
A20 Good level of maintenance of activity facilities |
0.757 | 0.728 | ||||
| F5 Contact with nature | B1 Appreciation of animals and plants | 0.825 | 0.747 | 0.864 | 0.680 | 0.860 |
| B2 Sitting and reflection | 0.880 | 0.774 | ||||
| B3 Reading and learning | 0.765 | 0.704 | ||||
| F6 Physical activities | B4 Field activities | 0.894 | 0.801 | 0.892 | 0.735 | 0.889 |
| B5 Facility activities | 0.787 | 0.745 | ||||
| B6 Free activities | 0.886 | 0.822 | ||||
| F7 Social interaction | B7 Walking and conversing | 0.809 | 0.693 | 0.823 | 0.610 | 0.818 |
| B8 Chat gathering | 0.840 | 0.720 | ||||
| B9 Recreational interaction | 0.685 | 0.610 | ||||
| F8 Psychologically restorative outcomes | C1 Restoration of attention | 0.857 | 0.763 | 0.874 | 0.584 | 0.882 |
| C2 Relief of psychological stress | 0.800 | 0.719 | ||||
| C3 Regulation of negative emotions | 0.655 | 0.640 | ||||
| C4 Recovery of vitality | 0.765 | 0.753 | ||||
| C5 Reduce feelings of loneliness | 0.707 | 0.732 |
N = 630.
Table 4.
Discriminant validity test results.
| Latent variables | F4 | F3 | F1 | F2 | F7 | F6 | F5 | F8 |
|---|---|---|---|---|---|---|---|---|
| F4 | 0.791 | |||||||
| F3 | 0.166 | 0.814 | ||||||
| F1 | 0.305 | 0.209 | 0.756 | |||||
| F2 | 0.365 | 0.267 | 0.351 | 0.792 | ||||
| F7 | 0.218 | 0.341 | 0.434 | 0.411 | 0.781 | |||
| F6 | 0.666 | 0.181 | 0.411 | 0.466 | 0.278 | 0.857 | ||
| F5 | 0.240 | 0.385 | 0.247 | 0.611 | 0.308 | 0.300 | 0.825 | |
| F8 | 0.470 | 0.365 | 0.454 | 0.564 | 0.501 | 0.582 | 0.530 | 0.764 |
The numbers on the diagonal represent the square roots of the average variance extracted from the latent variables, whereas the numbers below the diagonal indicate the correlation coefficients among the latent variables.
Validity testing primarily included exploratory factor analysis (EFA) and confirmatory factor analyses (CFA). First, an exploratory factor analysis (EFA) was conducted (Table 2); the KMO statistic was 0.919, and Bartlett’s test was significant at the 0.001 level, indicating that the potential variables were suitable for validity analysis. Factors were extracted using principal component analysis, and the cumulative explained variance was 72.417% with all eigenvalues exceeding 1. This finding indicated that the data sufficiently represented the original dataset. Common factors were extracted through orthogonal rotation, resulting in the identification of eight dimensions. The composition of these factors aligned with the hypotheses proposed in the model, suggesting that the scale employed in this study demonstrated strong structural validity.
Second, we conducted confirmatory factor analysis (CFA), which included tests for convergent and discriminant validity. Convergent validity was assessed based on three criteria: standardised factor loadings (greater than 0.5), composite reliability coefficients (CR greater than 0.7) and average variance extracted (AVE greater than 0.5, where a higher AVE indicated greater explanatory power of the observed variables for the latent variables). As shown in Table 3, the standardised factor loadings for all the observed variables exceeded 0.5, suggesting that the relationships between the latent and observed variables were statistically significant. The CR coefficients for all latent variables were above 0.7, indicating strong internal consistency among the measurement items. Moreover, the AVE values for all latent variables exceeded 0.5, indicating that our model effectively measured the variables and demonstrated high convergent validity.
The discriminant validity was evaluated by comparing the square roots of the AVE for each latent variable. The square root of the AVE for all latent variables surpassed the correlations with their corresponding variables, suggesting that the model exhibited good discriminant validity (Table 4).
Model fit test
The maximum likelihood estimation method was used to test the validity of the relevant hypotheses. The reference values for the model fit indices are listed in Table 5. The chi-square (χ²) value degrees of freedom (df) ratio (χ²/df) was 1.841 and below the benchmark of 2, indicating a good model fit. Additionally, the goodness of fit index (GFI), adjusted goodness of fit index (AGFI), root mean square error of approximation (RMSEA), comparative fit index (CFI), Normed Fit Index (NFI), Bollen’s incremental fit index (IFI) and Tucker-Lewis index (TLI) were adopted to estimate model fit, suggesting that the hypothesised model exhibited a good level of fit. The parameter estimation results of the model and standardised path coefficients are shown in Fig. 4.
Table 5.
Reference values of model fit indices.
| Fit indices | Reference value | Model value | Overall model fit |
|---|---|---|---|
| χ2/df | <2.00 | 1.841 | Yes |
| GFI | >0.90 | 0.916 | Yes |
| AGFI | >0.90 | 0.901 | Yes |
| RMR | <0.05 | 0.014 | Yes |
| RMSEA | <0.05 | 0.037 | Yes |
| CFI | >0.90 | 0.967 | Yes |
| NFI | >0.90 | 0.931 | Yes |
| IFI | >0.90 | 0.967 | Yes |
| TLI | >0.90 | 0.964 | Yes |
Fig. 4.
Parameter estimation results of the model and standardised path coefficients.
Common method bias test
Given that the self-assessment method of the questionnaire potentially leads to common method bias (CMB), we employed Harman’s single-factor test to examine whether CMB existed in the model88. The results indicated that the unrotated exploratory factor analysis extracted eight factors with eigenvalues greater than one, with the maximum variance explained by a single factor of 30.604% (less than 40%). Furthermore, recognising that the single-factor test could yield insignificant results under certain conditions, we incorporated a common method factor into the SEM and performed a single-factor confirmatory factor analysis involving all the indicators used in the hypothesis testing. First, we constructed a confirmatory factor analysis model M1, followed by model M2, which included the method factor. The comparison of the main fit indices between models M1 and M2 yielded the following results (Table 6): △GFI = 0.009, △CFI = 0.006, △NFI = 0.008, and △RMSEA = -0.003. The changes in all fit indices were less than 0.01, indicating that the inclusion of the common method factor did not significantly improve the model, and that there was no notable common method bias in the measurement.
Table 6.
Common method bias test results.
| Model | χ2/df | GFI | CFI | NFI | RMSEA | △GFI | △CFI | △NFI | △RMSEA |
|---|---|---|---|---|---|---|---|---|---|
| M1 | 1.841 | 0.916 | 0.967 | 0.931 | 0.037 | 0.009 | 0.006 | 0.008 | −0.003 |
| M2 | 1.748 | 0.925 | 0.973 | 0.939 | 0.034 |
Hypothesis testing
Analysis of path coefficients
The analysis of the path coefficient results (Table 7) revealed that F1 (architectural scale and aesthetics; β = 0.105, p < 0.05), F2 (landscape richness; β = 0.112, p < 0.05), F3 (availability of rest facilities; β = 0.087, p < 0.05), and F4 (compatibility of activity facilities; β = 0.095, p < 0.05) had significant direct effects on F8 (psychologically restorative outcomes), thus supporting hypotheses H1a, H1b, H1c, and H1d. In contrast, F1 (architectural scale and aesthetics; β = 0.005, p > 0.05) did not significantly affect F5 (contact with nature), suggesting that hypothesis H2a1 was not supported. Additionally, F2 (landscape richness; β = 0.546, p < 0.001) and F3 (availability of rest facilities; β = 0.238, p < 0.001) positively and significantly affected F5 (contact with nature), thereby supporting hypotheses H2a2 and H2a3. F1 (architectural scale and aesthetics; β = 0.174, p < 0.001), F2 (landscape richness; β = 0.209, p < 0.001 ), and F4 (compatibility of activity facilities; β = 0.537, p < 0.001) positively and significantly affected F6 (physical activity), thereby supporting hypotheses H2b1, H2b2, and H2b3. Moreover, F1 (architectural scale and aesthetics; β = 0.303, p < 0.001), F2 (landscape richness; β = 0.248, p < 0.001), and F3 (availability of rest facilities; β = 0.212, p < 0.001 ) positively and significantly affected F7 (social interaction), thus supporting hypotheses H2c1, H2c2, and H2c3.
Table 7.
Analysis of path coefficients.
| Hypothesis | Regression path | Unstandardised estimate | Standardised estimate | S.E. | C.R. (t-value) | P | Conclusion |
|---|---|---|---|---|---|---|---|
| H1a | F1→F8 | 0.134 | 0.105 | 0.052 | 2.592 | 0.010 | Supported |
| H1b | F2→F8 | 0.119 | 0.112 | 0.053 | 2.270 | 0.023 | Supported |
| H1c | F3→F8 | 0.104 | 0.087 | 0.045 | 2.334 | 0.020 | Supported |
| H1d | F4→F8 | 0.099 | 0.095 | 0.048 | 2.054 | 0.040 | Supported |
| H2a1 | F1→F5 | 0.006 | 0.005 | 0.047 | 0.130 | 0.896 | Not supported |
| H2a2 | F2→F5 | 0.526 | 0.546 | 0.044 | 12.029 | *** | Supported |
| H2a3 | F3→F5 | 0.257 | 0.238 | 0.043 | 6.033 | *** | Supported |
| H2b1 | F1→F6 | 0.260 | 0.174 | 0.057 | 4.577 | *** | Supported |
| H2b2 | F2→F6 | 0.263 | 0.209 | 0.048 | 5.459 | *** | Supported |
| H2b3 | F4→F6 | 0.656 | 0.537 | 0.049 | 13.383 | *** | Supported |
| H2c1 | F1→F7 | 0.288 | 0.303 | 0.044 | 6.530 | *** | Supported |
| H2c2 | F2→F7 | 0.197 | 0.248 | 0.037 | 5.405 | *** | Supported |
| H2c3 | F3→F7 | 0.190 | 0.212 | 0.038 | 4.936 | *** | Supported |
| H3a | F5→F8 | 0.257 | 0.231 | 0.052 | 4.903 | *** | Supported |
| H3b | F6→F8 | 0.238 | 0.280 | 0.044 | 5.433 | *** | Supported |
| H3c | F7→F8 | 0.283 | 0.211 | 0.058 | 4.908 | *** | Supported |
***p < 0.001; N = 630.
F5 (contact with nature; β = 0.231, p < 0.001), F6 (physical activity; β = 0.280, p < 0.001), and F7 (social interaction; β = 0.211, p < 0.001) exerted significant positive effects on F8 (psychologically restorative outcomes), thereby supporting hypotheses H3a, H3b, and H3c.
Tests of mediation effects
The bootstrap method for 5,000 resamples was used to analyse the mediating effects of the three types of university students’ behavioural patterns between university common spaces and psychological restoration. The mediating effect was deemed significant if both the bias-corrected and percentile estimates of the effect value did not include 0 within the 95% confidence interval and z ≥ 1.96 with two-tailed significance at p < 0.05.
As demonstrated in Table 8, the results of the total effects indicated that the standardised lower and upper limits of the total effect of F1 (architectural scale and aesthetics), F2 (landscape richness), F3 (availability of rest facilities) and F4 (compatibility of activity facilities) on F8 (psychologically restorative outcomes) did not contain zero, with a z-value of more than 1.96 and p < 0.05. These findings suggested that the total effects of the paths F1 → F8 (β = 0.218, p < 0.001), F2 → F8 (β = 0.348, p < 0.001), F3 → F8 (β = 0.186, p < 0.001) and F4 → F8 (β = 0.245, p < 0.001) were significant.
Table 8.
Direct, mediated and total effects of the standardised hypothesis model.
| Hypothesis | Indirect path | Point estimate | Product of coefficients | Bootstrapping | P (Two-tailed significance) | Conclusion | ||||
|---|---|---|---|---|---|---|---|---|---|---|
| Bias-corrected 95%CI | Percentile 95%CI | |||||||||
| SE | Z | Lower | Upper | Lower | Upper | |||||
| Standardised direct effects | ||||||||||
| H2a1 | F1→F5 | 0.005 | 0.048 | 0.104 | -0.087 | 0.100 | -0.088 | 0.098 | 0.917 | Not supported |
| H2b1 | F1→F6 | 0.174 | 0.040 | 4.350 | 0.095 | 0.251 | 0.097 | 0.252 | 0.000(***) | Supported |
| H2c1 | F1→F7 | 0.303 | 0.039 | 7.769 | 0.224 | 0.378 | 0.227 | 0.379 | 0.000(***) | Supported |
| H1a | F1→F8 | 0.105 | 0.042 | 2.500 | 0.023 | 0.187 | 0.022 | 0.186 | 0.012(*) | Supported |
| H2a2 | F2→F5 | 0.546 | 0.038 | 14.368 | 0.469 | 0.618 | 0.473 | 0.621 | 0.000(***) | Supported |
| H2b2 | F2→F6 | 0.209 | 0.042 | 4.976 | 0.125 | 0.292 | 0.126 | 0.294 | 0.000(***) | Supported |
| H2c2 | F2→F7 | 0.248 | 0.042 | 5.905 | 0.160 | 0.326 | 0.163 | 0.329 | 0.000(***) | Supported |
| H1b | F2→F8 | 0.112 | 0.052 | 2.154 | 0.008 | 0.216 | 0.009 | 0.217 | 0.031(*) | Supported |
| H2a3 | F3→F5 | 0.238 | 0.038 | 6.263 | 0.161 | 0.310 | 0.162 | 0.311 | 0.000(***) | Supported |
| H2c3 | F3→F7 | 0.212 | 0.040 | 5.300 | 0.132 | 0.289 | 0.131 | 0.288 | 0.000(***) | Supported |
| H1c | F3→F8 | 0.087 | 0.037 | 2.351 | 0.014 | 0.158 | 0.015 | 0.159 | 0.019(*) | Supported |
| H2b3 | F4→F6 | 0.537 | 0.033 | 16.273 | 0.469 | 0.598 | 0.472 | 0.600 | 0.000(***) | Supported |
| H1d | F4→F8 | 0.095 | 0.045 | 2.111 | 0.005 | 0.182 | 0.007 | 0.183 | 0.035(*) | Supported |
| H3a | F5→F8 | 0.231 | 0.050 | 4.620 | 0.133 | 0.330 | 0.130 | 0.327 | 0.000(***) | Supported |
| H3b | F6→F8 | 0.280 | 0.052 | 5.385 | 0.179 | 0.383 | 0.180 | 0.384 | 0.000(***) | Supported |
| H3c | F7→F8 | 0.211 | 0.045 | 4.689 | 0.119 | 0.296 | 0.120 | 0.297 | 0.000(***) | Supported |
| Standardised mediation effects | ||||||||||
| H4a1 | F1→F5→F8 | 0.001 | 0.011 | 0.091 | −0.022 | 0.023 | −0.022 | 0.022 | 0.928 | Not supported |
| H4a2 | F1→F6→F8 | 0.049 | 0.015 | 3.267 | 0.024 | 0.083 | 0.023 | 0.080 | 0.001(**) | Supported |
| H4a3 | F1→F7→F8 | 0.064 | 0.017 | 3.765 | 0.034 | 0.099 | 0.034 | 0.099 | 0.000(***) | Supported |
| H4b1 | F2→F5→F8 | 0.126 | 0.029 | 4.345 | 0.072 | 0.187 | 0.069 | 0.184 | 0.000(***) | Supported |
| H4b2 | F2→F6→F8 | 0.059 | 0.016 | 3.688 | 0.033 | 0.097 | 0.030 | 0.094 | 0.000(***) | Supported |
| H4b3 | F2→F7→F8 | 0.052 | 0.015 | 3.467 | 0.027 | 0.087 | 0.026 | 0.085 | 0.001(**) | Supported |
| H4c1 | F3→F5→F8 | 0.055 | 0.015 | 3.667 | 0.030 | 0.091 | 0.028 | 0.087 | 0.000(***) | Supported |
| H4c2 | F3→F7→F8 | 0.045 | 0.012 | 3.750 | 0.024 | 0.074 | 0.022 | 0.071 | 0.000(***) | Supported |
| H4d | F4→F6→F8 | 0.150 | 0.030 | 5.000 | 0.096 | 0.213 | 0.095 | 0.212 | 0.000(***) | Supported |
| Standardised total effects | ||||||||||
| F1→F8 | 0.218 | 0.039 | 5.590 | 0.142 | 0.296 | 0.139 | 0.295 | 0.000(***) | Supported | |
| F2→F8 | 0.348 | 0.040 | 8.700 | 0.268 | 0.426 | 0.270 | 0.427 | 0.000(***) | Supported | |
| F3→F8 | 0.186 | 0.035 | 5.314 | 0.116 | 0.255 | 0.115 | 0.254 | 0.000(***) | Supported | |
| F4→F8 | 0.245 | 0.034 | 7.206 | 0.177 | 0.310 | 0.177 | 0.310 | 0.000(***) | Supported | |
Standardised estimation of 5000 bootstrap samples; ***p < 0.001, **p < 0.01, *p < 0.05, N = 630.
In the direct effect test shown in Table 8, the standardised lower and upper limits of the direct effect of F1 (architectural scale and aesthetics) on F5 (contact with nature) included zero, with a z-value less than 1.96 and p > 0.05. This indicated that the direct effects of the path F1→F5 (β = 0.005, p > 0.05) were not significant; therefore, hypothesis H2a1 was not supported. The standardised lower and upper limits of the direct effect of F1 (architectural scale and aesthetics) on F6 (physical activity), F7 (social interaction), and F8 (psychologically restorative outcomes) did not contain zero, with a z-value of more than 1.96 and p < 0.05. This indicated that the direct effects of the path F1→F6 (β = 0.174, p < 0.001), F1→F7 (β = 0.303, p < 0.001) and F1→F8 (β = 0.105, p < 0.05) were significant; thus, hypotheses H2b1, H2c1 and H1a were supported.
The standardised lower and upper limits of the direct effect of F2 (landscape richness) on F5 (contact with nature), F6 (physical activity), F7 (social interaction) and F8 (psychologically restorative outcomes) did not contain zero, with a z-value of more than 1.96 and p < 0.05. This indicated that the direct effects of the path F2→F5 (β = 0.546, p<0.001), F2→F6 (β = 0.209, p<0.001), F2→F7 (β = 0.248, p<0.001) and F2→F8 (β = 0.112, p<0.05) were significant; therefore, hypotheses H2a2, H2b2, H2c2 and H1b were supported.
The standardised lower and upper limits of the direct effect of F3 (availability of rest facilities) on F5 (contact with nature), F7 (social interaction) and F8 (psychologically restorative outcomes) did not contain zero, with a z-value of more than 1.96 and p < 0.05. This indicated that the direct effects of the path F3→F5 (β = 0.238, p<0.001), F3→F7 (β = 0.212, p<0.001), F3→F8 (β = 0.087, p<0.05) were significant; therefore, hypotheses H2a3, H2c3 and H1c were supported.
The standardised lower and upper limits of the direct effect of F4 (compatibility of activity facilities) on F6 (physical activity) and F8 (psychologically restorative outcomes) did not contain zero, with a z-value of more than 1.96 and p < 0.05. This indicated that the direct effects of the path F4→F6 (β = 0.537, p<0.001), F4→F8 (β = 0.095, p<0.05 ) were significant; therefore, hypotheses H2b3 and H1d were supported.
The standardised lower and upper limits of the direct effect of F5 (contact with nature), F6 (physical activity) and F7 (social interaction) on F8 (psychologically restorative outcomes) did not contain zero, with a z-value of more than 1.96 and p < 0.05. This indicated that the direct effect of the path F5→F8 (β = 0.231, p<0.001 ), F6→F8 (β = 0.280, p<0.001) and F7→F8 (β = 0.211, p<0.001) was significant; therefore, hypotheses H3a, H3b and H3c were supported.
The mediation effect test is shown in Table 8, where the standardised lower and upper limits of the mediating effect of F1 (architectural scale and aesthetics) through F5 (contact with nature) on F8 (psychologically restorative outcomes) included zero, with a z-value of less than 1.96 and p > 0.05. This indicated that the mediating effect of the path F1→F5→F8 (β = 0.001, p > 0.05) was not significant; therefore, hypothesis H4a1 was not supported. The effects of F1 (architectural scale and aesthetics) on F8 (psychologically restorative outcomes) through F6 (physical activity) and F7 (social interaction) were significant. The mediating effects of the paths F1→F6→F8 (β = 0.049, p < 0.01) and F1→F7→F8 (β = 0.064, p < 0.001) were significant, supporting hypotheses H4a2 and H4a3.
Additionally, F2 (landscape richness) had a significant effect on F8 (psychologically restorative outcomes) through F5 (contact with nature), F6 (physical activity) and F7 (social interaction). This indicated that the mediating effects of the paths F2→F5→F8 (β = 0.126, p < 0.001), F2→F6→F8 (β = 0.059, p < 0.001) and F2→F7→F8 (β = 0.052, p < 0.01) were significant, supporting hypotheses H4b1, H4b2 and H4b3. We also found that the provision of F3 (availability of rest facilities) had a significant effect on F8 (psychologically restorative outcomes) through F5 (contact with nature) and F7 (social interaction). This indicated that the mediating effects of the paths F3→F5→F8 (β = 0.055, p < 0.001) and F3→F7→F8 (β = 0.045, p < 0.001) were significant, supporting hypotheses H4c1 and H4c2. F4 (compatibility with activity facilities) had a significant effect on F8 (psychologically restorative outcomes) through F6 (physical activity). This indicated that the mediating effect of the path F4→F6→F8 (β = 0.150, p < 0.001) was significant, supporting hypothesis H4d.
The results indicated that the characteristics of university common spaces (F1 (architectural scale and aesthetics), F2 (landscape richness), F3 (availability of rest facilities) and F4 (compatibility of activity facilities)) had significant total and direct effects on F8 (psychologically restorative outcomes). Additionally, the mediating effects of F5 (contact with nature), F6 (physical activity) and F7 (social interaction) were partially significant, suggesting that the mediating effects identified in this study were incomplete and that the behavioural patterns of university students played a partial mediating role. According to Table 8, the overall effect of university common space characteristics on psychological restoration among university students was 0.997, with a direct effect of 0.398 and a mediating effect of 0.599 attributed to university students’ behavioural patterns. This finding indicated that for every unit, an increase in common space characteristics in universities increased the psychological restoration effect by 0.997 units. Specifically, 0.599 represented the effect of common space characteristics on psychological restoration mediated by university students’ behavioural patterns, whereas 0.398 reflected the direct effect of these characteristics on psychological restoration. Notably, this mediating effect accounted for approximately 60% of the total effect.
Results of FsQCA
Selection and calibration of variables
SEM analysis confirmed that university common space characteristics and university students’ behavioural patterns were significant antecedent variables influencing psychologically restorative outcomes. fsQCA can serve as a supplementary method to enhance the analysis of results derived from the interdependent interactions of multiple conditions based on the net effect investigation of the SEM and explains various plausible configurations from a holistic perspective in the context of complex causal relationships89. Therefore, we combined SEM with fsQCA to explore the combined effects of architectural scale and aesthetics, landscape richness, availability of rest facilities, compatibility of activity facilities, contact with nature, physical activity and social interaction (the antecedent conditions) on the psychological restoration of university students.
The influence of the first seven variables on psychological restoration was validated using a theoretical model and empirical results. To conduct the fsQCA, it was necessary to calibrate the antecedent condition variables (Table 9). Continuous variables were averaged and the data were calibrated according to the standards proposed by Ragin (2009) as follows: 5% (fully out), 95% (fully in) and 50% (crossover point)90.
Table 9.
Calibration of continuous variables.
| Variable name | Fully in (95%) | Crossover point (50%) | Fully out (5%) | |
|---|---|---|---|---|
| Conditional variable | F1 Architectural scale and aesthetics | 4.600 | 3.600 | 2.800 |
| F2 Landscape richness | 5.000 | 3.800 | 2.890 | |
| F3 Availability of rest facilities | 4.400 | 3.600 | 2.800 | |
| F4 Compatibility of activity facilities | 4.600 | 3.600 | 2.800 | |
| F5 Contact with nature | 4.667 | 3.667 | 2.667 | |
| F6 Physical activity | 5.517 | 4.333 | 3.333 | |
| F7 Social interaction | 5.000 | 4.167 | 3.333 | |
| Outcome variable | F8 Psychologically restorative outcomes | 4.400 | 3.800 | 3.000 |
Necessity analysis of conditions
Here, fsQCA software (version 3.0) was used to analyse the necessity of the condition variables. High levels of psychologically restorative outcomes (approaching a calibration value of 1) were used as the outcome variables, and the following were defined as condition variables: F1 (architectural scale and aesthetics), F2 (landscape richness), F3 (availability of rest facilities), F4 (compatibility of activity facilities), F5 (contact with nature), F6 (physical activity) and F7 (social interaction) and their subsets. The results indicated that the consistency of all the individual condition variables was less than 0.9, demonstrating that none of the individual condition variables met the necessity standard (Table 10). This suggested that high levels of psychologically restorative outcomes did not necessarily require the presence of any specific conditional variable; rather, they depended on the interactions between conditional variables. Therefore, multiple conditional variables needed to be combined to further analyse the sufficient conditions for achieving psychologically restorative outcomes.
Table 10.
Necessity analysis results for the conditions.
| Conditional variable | Consistency | Coverage |
|---|---|---|
| F1 Architectural scale and aesthetics | 0.702 | 0.768 |
| ~ F1 Architectural scale and aesthetics | 0.569 | 0.588 |
| F2 Landscape richness | 0.746 | 0.781 |
| ~ F2 Landscape richness | 0.538 | 0.581 |
| F3 Availability of rest facilities | 0.674 | 0.751 |
| ~ F3 Availability of rest facilities | 0.575 | 0.584 |
| F4 Compatibility of activity facilities | 0.699 | 0.773 |
| ~ F4 Compatibility of activity facilities | 0.563 | 0.575 |
| F5 Contact with nature | 0.752 | 0.754 |
| ~ F5 Contact with nature | 0.517 | 0.584 |
| F6 Physical activity | 0.779 | 0.759 |
| ~ F6 Physical activity | 0.492 | 0.575 |
| F7 Social interaction | 0.690 | 0.763 |
| ~ F7 Social interaction | 0.569 | 0.582 |
~ indicates the absence of the condition.
Sufficiency analysis of conditional configurations
Based on the aforementioned analysis, the antecedent configuration of university students’ psychological restoration was further examined. Three behavioural patterns —F5 (contact with nature), F6 (physical activity) and F7 (social interaction)—were used to investigate the relationship between the combined pathways of common space characteristics in universities and the adequacy of psychological restoration. Here, fsQCA 3.0 software was used to construct a truth table, setting the minimum acceptable threshold for raw consistency at 0.8 and the proportional reduction in inconsistency (PRI) consistency at ≥ 0.75. The configurations of the antecedent variables for psychological restoration are presented in Table 11. Considering the validity of the results and the moderation of complexity, we selected an intermediate solution for interpretation, identifying seven simplified configurations of conditions that could lead to a high level of psychologically restorative effect. The overall consistency of the seven configurations exceeded 0.8, indicating a good level of agreement, and the presence of all the conditional variables served as a sufficient condition for the results. The overall coverage of the seven configurations exceeded 0.5, demonstrating their ability to cover more than 50% of samples. In terms of coverage among the different configurations, the sample sizes of the seven configurations did not differ significantly, all having a certain degree of persuasiveness, and the overall model demonstrated strong explanatory power.
Table 11.
Configuration analysis of causal conditions for psychologically restorative outcomes.
| Solution | Causal condition | Consistency | Raw coverage | Unique coverage | Solution consistency | Solution coverage | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| F1 | F2 | F3 | F4 | F5 | F6 | F7 | ||||||
| 1 | ● | ● | ● | 0.884 | 0.516 | 0.516 | 0.884 | 0.516 | ||||
| 2 | ● | ● | ● | 0.902 | 0.504 | 0.024 | 0.856 | 0.638 | ||||
| 3 | ● | ● | ● | 0.888 | 0.540 | 0.060 | ||||||
| 4 | ● | ● | ● | 0.879 | 0.553 | 0.073 | ||||||
| 5 | ● | ● | ● | 0.907 | 0.474 | 0.062 | 0.858 | 0.612 | ||||
| 6 | ● | ● | ● | 0.890 | 0.495 | 0.083 | ||||||
| 7 | ● | ● | ● | 0.887 | 0.467 | 0.055 | ||||||
● indicate the presence of a condition, whereas blank cells represent “don’t care” conditions.
According to the fsQCA results (Table 11), there were seven conditional configurations of the pathways through which common spaces in universities influenced the psychological restoration of university students. Based on the previous theoretical framework, these seven conditional configurations were divided into two categories: category one primarily consisted of the direct effect pathways of university common space characteristics on psychologically restorative outcomes (Configurations 2 and 5), whereas category two comprised the mediating pathways related to the behavioural activity characteristics of university students in common spaces (Configurations 1, 3, 4, 6 and 7). FsQCA seeks to maximise coverage, which can lead to the magnification of subtle differences and the formation of similar yet distinct configurations91. Although the antecedent condition configurations vary, they can be classified under the same pattern if they arise from the same optimised solution. We identified four influencing patterns that promote high levels of psychological restorative outcomes. These were referred to as integrated patterns of environment and facilities, static rest patterns, dynamic exercise patterns and interaction-driven patterns (Fig. 5).
Fig. 5.
Configuration model.
Integrated patterns of environment and facilities
This model included two subpatterns composed of Conditional Configurations 2 and 5, which were primarily dominated by the spatial environment and facility characteristics. The core conditions of sub-pattern 1 consisted of F1 (architectural scale and aesthetics) * F2 (landscape richness) * F3 (availability of rest facilities), which mainly involved 15 key design elements, including appropriate scale of building enclosure, high openness of building enclosure, diverse forms of building enclosure, strong architectural history, varied building facades, a large number of plants, abundant plant types, rich plant colours, extensive lawn coverage, highly ornamental waterscape, plentiful rest facilities, rest facilities with a view, comfortable rest facilities, good privacy of rest facilities and hygienic rest facilities. The core conditions of sub-pattern 2 consisted of F1 (architectural scale and aesthetics) * F2 (landscape richness) * F4 (compatibility of activity facilities), which mainly involved 15 key design elements, including appropriate scale of building enclosure, high openness of building enclosure, diverse forms of building enclosure, strong architectural history, varied building facades, large number of plants, abundant plant types, rich plant colours, extensive lawn coverage, highly ornamental waterscape, plentiful activity fields, abundant types of activity fields, accessible activity fields, a sufficient number of fitness facilities and a good level of maintenance of activity facilities.
Static rest patterns
The core conditions of this pattern were primarily F2 (landscape richness) * F3 (availability of rest facilities) * F5 (contact with nature), which mainly involved 10 key design elements and three contacts with natural behaviours, including a large number of plants, abundant plant types, rich plant colours, extensive lawn coverage, highly ornamental waterscape, plentiful rest facilities, rest facilities with a view, comfortable rest facilities, good privacy of rest facilities, hygienic rest facilities, appreciation of animals and plants, sitting and reflection and reading and learning. The design of common spaces on university campuses should not only consider the spatial environment and facility characteristics, but also focus on the behavioural activity characteristics of university students, ensuring that these common spaces accommodate their static relaxation behaviour patterns of university students.
Dynamic exercise patterns
This model comprises two sub-patterns, Conditional Configurations 3 and 4, which are primarily dominated by architectural and landscape characteristics, activity facilities, and physical activity behaviours. The core conditions of sub-pattern 1 consisted of F1 (architectural scale and aesthetics) * F2 (landscape richness) * F6 (physical activity), which mainly involved 10 key design elements and three physical activity behaviours, including appropriate scale of building enclosure, high openness of building enclosure, diverse forms of building enclosure, strong architectural history, varied building facades, a large number of plants, abundant plant types, rich plant colours, extensive lawn coverage, highly ornamental waterscape, field activities, facility activities and free activities. The core conditions of sub-pattern 2 were F2 (landscape richness) * F4 (compatibility of activity facilities) * F6 (physical activity), which mainly involved 10 key design elements and three physical activity behaviours, including a large number of plants, abundant plant types, rich plant colours, extensive lawn coverage, highly ornamental waterscape, plentiful activity fields, abundant types of activity fields, accessible activity fields, sufficient number of fitness facilities, good level of maintenance of activity facilities, field activities, facility activities and free activities.
Interaction-driven patterns
This model included two sub-patterns composed of Conditional Configurations 6 and 7, which were primarily characterised by architectural and landscape characteristics, rest facilities, and social interaction behaviours. The core conditions of sub-pattern 1 were F1 (architectural scale and aesthetics) * F2 (landscape richness) * F7 (social interactions), which mainly involved 10 key design elements and three social interaction behaviours, including appropriate scale of building enclosure, high openness of building enclosure, diverse forms of building enclosure, strong architectural history, varied building facades, a large number of plants, abundant plant types, rich plant colours, extensive lawn coverage, highly ornamental waterscape, walking and conversing, chat gathering and recreational interaction. The core conditions of sub-pattern 2 were F2 (landscape richness) * F3 (availability of rest facilities) * F7 (social interaction), which mainly involved ten key design elements and three social interaction behaviours, including a large number of plants, abundant plant types, rich plant colours, extensive lawn coverage, highly ornamental waterscape, plentiful rest facilities, rest facilities with a view, comfortable rest facilities, good privacy of rest facilities, hygienic rest facilities, walking and conversing and chat gathering and recreational interaction.
Discussion
In this study, four research hypotheses were proposed and the empirical findings provided strong support for the hypothesised model. The results of the SEM analysis confirmed the existence of two pathways through which university common spaces influenced the psychological restoration experienced by university students. These pathways were as follows: (1) direct pathway, characteristics of university common spaces → psychologically restorative outcomes; and (2) mediated pathway, characteristics of university common spaces → university students’ behavioural patterns → psychologically restorative outcomes. Building on this foundation, we utilised the fsQCA method to validate the effectiveness of the two previously discussed pathways and offered a more comprehensive framework for enhancing high levels of psychologically restorative outcomes. Furthermore, we investigated various plausible configurations aimed at achieving improved psychologically restorative outcomes, ultimately identifying four distinct configuration patterns.
Therefore, we here discuss the effects of the pathways and configuration patterns of common spaces characteristics on psychologically restorative outcomes with the aim of designing university campus common spaces that better support psychologically restorative outcomes, considering university students’ behavioural patterns.
Direct pathways and configurational patterns of university common space characteristics affecting psychologically restorative outcomes
According to the SEM results, F2 (landscape richness; β = 0.112) had the stronger direct effect on psychologically restorative outcomes. This study further revealed that a diverse variety of plant species and quantities as well as varied water landscape areas and forms in common spaces in universities can enhance the visual appeal of the environment, alleviate psychological stress and facilitate psychological restoration. This finding is consistent with the results of Ha and Kim (2021) and Halecki et al. (2023), who indicated that landscapes characteristics with higher biodiversity and rich elements significantly improve spatial attractiveness, while also influencing the quality of psychological restoration. The effect of F1 (architectural scale and aesthetics; β = 0.105) on psychologically restorative outcomes ranked second. Previous studies have not adequately addressed the influence of the architectural environmental characteristics on these effects. This study demonstrated that an appropriate degree of enclosure openness in buildings can evoke psychological content distinct from typical learning environments, eliciting pleasurable psychological experiences for university students, thereby enhancing their levels of psychological restoration. Additionally, the direct effects of F3 (availability of rest facilities; β = 0.087) and F4 (compatibility of activity facilities; β = 0.095) on psychologically restorative outcomes were relatively small. This finding is inconsistent with the results of Han et al. (2022) and Luo et al. (2022), who suggested that a sufficient number of rest and activity facilities are more beneficial for psychological restoration. We believe that the primary reason for this discrepancy in research findings is the relatively abundant supply of rest and activity facilities in urban parks and squares, which differ significantly from the spatial environment of university campuses in terms of their environmental characteristics.
However, existing research on this topic has typically relied on SEM methods that view architectural environments, landscape environments, rest facilities and activity facilities as isolated antecedents7,92,93. Our research provides a combinatorial pattern (integrated patterns of environment and facilities) for the mechanism directly influencing high-level psychologically restorative outcomes, based on the results of the fsQCA. This pattern comprises two sub-patterns, and an analysis of the configurational paths of sub-patterns 1 (F1 * F2 * F3) and 2 (F1 * F2 * F4) revealed that an appropriate architectural scale, favourable architectural aesthetic characteristics, rich landscape characteristics, and a combination of highly supportive rest facilities and highly compatible activities effectively enhanced psychologically restorative outcomes. Consequently, in university common spaces, students are more likely to experience high levels of psychological restoration when spatial elements including suitable architectural environments, diverse landscape environments and high-quality facility support are present simultaneously.
Mediating pathways and configurational patterns of university students’ behavioural activities affecting psychologically restorative outcomes
Through empirical research using SEM, F2 (landscape richness; β = 0.126) and F3 (availability of rest facilities; β = 0.055) each had an indirect effect on F8 (psychologically restorative outcomes) through F5 (contact with nature). This suggests that rich landscape characteristics in university common spaces can provide university students with more opportunities to contact nature, whereas sufficient rest facilities can extend the duration of their contact with the natural environment, thereby enhancing their psychological restoration. These results agree with the research of Hami and Abdi (2021), and Nie et al. (2024), which confirmed that the characteristics of landscape environments and rest facilities can promote psychological restoration. However, these studies examined only linear causal relationships through single spatial elements and neglected the combinatorial effects of spatial environmental characteristics. Using the fsQCA method to study the combinatorial effects, we found that the configurational pathway for contact with nature, F2 * F3 * F5 (static rest patterns), had an overall coverage of 51.6%. The analysis of the net effects via SEM revealed that the explanatory variance (R²) through the mediating pathways of contact with nature, F2→F5→F8 and F3→F5→F8, was 42.7%. This indicated that the combinatorial effects had greater explanatory power. In university common spaces, a combination of rich landscape characteristics and highly supportive rest facilities can enhance university students’ behavioural experiences of contact with nature, thereby promoting high levels of psychologically restorative outcomes.
Additionally, the SEM analysis revealed that F1 (architectural scale and aesthetics; β = 0.049) can indirectly affect F8 (psychologically restorative outcomes) through F6 (physical activity). Previous studies have overlooked the fact that suitable architectural enclosure scales, forms, and aesthetic characteristics in university common spaces can improve psychological restoration by enhancing university students’ willingness to engage in physical activities94. F2 (landscape richness; β = 0.059) and F4 (compatibility of activity facilities; β = 0.150) both had an indirect effect on F8 (psychologically restorative outcomes) through F6 (physical activity). This result is consistent with the findings of Markevych et al. (2017), Sun et al. (2023), and Wang et al. (2021). Landscape environmental quality was also positively correlated with university students’ physical activity behaviours. The number of activity fields and compatibility of facilities can provide environmental support and promote physical activity among university students, thereby benefiting emotional regulation and psychological restoration. Prior research focused primarily on how the mediating role of physical activity promotes the net effects of psychological restoration46,95–97. However, by comparing the results of the fsQCA configurational effects and the SEM net effects, we found that the overall coverage of the configurational paths of physical activity, F1 * F2 * F6 and F2 * F4 * F6 (dynamic exercise patterns), was 63.8%. The net effect explained variance (R²) of the mediating paths F1 → F6 → F8, F2 → F6 → F8, and F4 → F6 → F8 was 52.6%. Moreover, we confirmed that the variation in the intensity of the combined effect of physical activity was more explanatory than that of a single mediating path. This indicates that a combination of suitable architectural scales, better architectural aesthetic characteristics, and rich landscape characteristics, or a combination of rich landscape characteristics and highly compatible activity facilities, can enhance opportunities for university students to engage in physical activities, thereby triggering a high level of psychologically restorative outcomes.
F1 (architectural scale and aesthetics; β = 0.064), F2 (landscape richness; β = 0.052), and F3 (availability of rest facilities; β = 0.045) all had indirect effects on F8 (psychologically restorative outcomes) through F7 (social interaction). Previous studies have rarely investigated how social interactions mediate the relationship between the spatial environment and mental health. Furthermore, only a few studies have measured the mediating role of social interaction on psychological restoration from the single perspective of landscape environmental characteristics, overlooking the effect of rest facility characteristics on social interaction behaviour and psychologically restorative outcomes98,99. From the fsQCA results, we found that the overall coverage of the configurational paths of social interaction, F1 * F2 * F7 and F2 * F3 * F7 (interaction-driven patterns), was 61.2%. Meanwhile, the results of the SEM net effects indicate that the explained variance (R²) of the mediating paths of social interaction, F1 → F7 → F8, F2 → F7 → F8 and F3 → F7 → F8, is 30.6%. This finding suggests that the combined effect of social interactions is more explanatory than the singular mediating path. To investigate the potential effect of these predictors on psychologically restorative outcomes, we analysed the effective configurations of landscape characteristics, rest facility characteristics, and university students’ social interaction behaviours in university common spaces. Few studies have examined the combined effects of these factors100–102, and our findings complement those of these previous studies.
Study limitations
Overall, this study reveals the effects of university common space characteristics on the psychological restoration of university students. We conducted research using a two-phase approach, and the proposed model and methodology are more comprehensive and specific than those used in existing studies. Initially, we employed traditional SEM methods to analyse the causal relationships among common spaces in universities, student behavioural activities and psychologically restorative outcomes, thereby validating the effectiveness of the proposed hypothetical model. Subsequently, we developed four causal configuration patterns to achieve enhanced psychologically restorative outcomes using the novel fsQCA method. Nevertheless, this study has several limitations that require further investigation. First, the hypothetical model developed in this study requires further refinement. We primarily based the model on the characteristics of university students’ behavioural activities as mediating factors, whereas moderating factors (such as the time and frequency of university students’ visits to common spaces) were not included. Future studies should further investigate these factors. Second, our study employed only quantitative samples for the analysis; qualitative data from in-depth interviews and behavioural observations also have theoretical and methodological value and should thus be explored further in future studies.
Conclusions and recommendations
In this study, field investigations of university common spaces were conducted from the perspective of the restorative environment. This involved spatial characteristic analysis and behavioural activity extraction, leading to the development of a restorative model for university common spaces. Moreover, SEM was employed to validate the hypothetical model and analyse the influencing pathways related to psychologically restorative outcomes. Furthermore, the fsQCA method was used to identify the configurational combinations of antecedent conditions that triggered high levels of psychologically restorative outcomes. The following conclusions were drawn:
(1) The study results confirm that four characteristics of university common spaces (architectural scale and aesthetics, landscape richness, availability of rest facilities and compatibility with activity facilities) directly influence university students’ psychological restoration. Additionally, these characteristics exert psychologically restorative outcomes through the mediating role of university students’ behavioural activities (such as contact with nature, physical activity and social interactions). (2) Different environmental characteristics have varying effects on psychologically restorative outcomes. Seven configurational paths triggering high levels of psychologically restorative outcomes were identified using the fsQCA method to examine the effects of factor combinations. (3) Based on the characteristics of the hypothetical model and configurational paths, we identified and constructed four triggering patterns (integrated patterns of the environment and facilities, static rest patterns, dynamic exercise patterns and interaction-driven patterns) that could achieve high levels of psychologically restorative outcomes. These patterns can be used to optimise the design of university common spaces and provide practical guidance for planning healthy campuses.
Our research will help designers, planners and decision-makers acquire a deeper understanding of the influence of university common space characteristics on the psychological restoration of university students. Furthermore, this study offers a replicable research methodology for investigating the restorative environments on university campuses. Based on the four triggering patterns that promote high levels of psychological restoration, we provide optimised design recommendations for university common spaces, emphasising four key aspects: an integrated experience of the environment and facilities, static relaxation, dynamic exercise and interaction-driven engagement.
In terms of the integrated experience of the environment and facilities, diverse spatial forms must be constructed by leveraging the enclosure characteristics of architecture while incorporating historical and cultural elements to cultivate a recognisable spatial atmosphere that fosters positive restorative effects from a visual perception perspective. Enhancing the multisensory experience of the landscape environment is crucial for maximising its restorative potential, thereby increasing the opportunities for university students to interact with restorative spaces. The number and positioning of rest facilities must align with the environmental characteristics of common spaces to optimise campus environments for psychological restoration. Furthermore, activity venues and facilities should be integrated seamlessly into the surrounding environment to enhance the appeal of these spaces and offer university students a restorative experience.
In static rest patterns, the design of viewing and rest facilities enhances opportunities for university students to contact nature, thereby supporting behaviours related to meditation and relaxation, which are beneficial for psychological restoration. In dynamic exercise patterns, the establishment of multifunctional sports venues configured flexibly and equipped with arranged activity facilities promotes physical activity among university students, facilitating their psychological restoration. Furthermore, in the interaction-driven model, the design of shared interactive communication spaces increases the opportunities for university students to engage with their peers, contributing to enhanced cognitive development and positive emotional states.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Acknowledgements
This research is supported by the National Natural Science Foundation of China (Grant No. 52108011); the Guangdong Basic and Applied Basic Research Foundation (Grant No. 2024A1515012129 and Grant No. 2023A1515011137); the Fundamental Research Funds for the Central Universities (Grant No. 2024ZYGXZR048); the State Key Laboratory of Subtropical Building and Urban Science, South China University of Technology (Grant No. 2024ZB06); Guangzhou Basic and Applied Basic Research Foundation (Grant No. 2024A04J9930); the Department of Housing and Urban-Rural Development of Guangdong Province (Grant No. 2021-K2-305243); the Department of Education of Guangdong Province (Grant No. 2021KTSCX004). It is also partly supported by the China Scholarship Council (CSC) scholarship under the CSC Grant No. 202406150137.
Author contributions
Hongyan Wen: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Resources, Software, Visualization, Writing – original draft, Writing – review & editing. Hanzheng Lin: Data curation, Writing – review & editing, Visualization. Xiao Liu: Conceptualization, Methodology, Software, Investigation, Project administration, Resources, Writing – original draft, Writing – review & editing, Funding acquisition. Weihong Guo: Supervision, Resources, Validation, Writing – review & editing. Bao-Jie He: Conceptualization, Methodology, Project administration, Validation, Writing – review & editing.
Data availability
Data are available from the corresponding author for reasonable requirements.
Declarations
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Hongyan Wen and Hanzheng Lin contributed equally to this work.
References
- 1.Fang, Y., Wang, L. & Chen, Z. 2022 Survey report on college students’ mental health. In (eds Fu, X. L., Zhang, K., Chen, X. F. et al.) ‘The Blue Book of Mental Health: Report on National Mental Health Development in China (2021–2022)’. (Social Science Literature Publishing House, 2023).
- 2.Hipp, J. A., Gulwadi, G. B., Alves, S. & Sequeira, S. The relationship between perceived greenness and perceived restorativeness of university campuses and student-reported quality of life. Environ. Behav.48, 1292–1308 (2016). [Google Scholar]
- 3.Guo, W., Wen, H. & Liu, X. Research on the psychologically restorative effects of campus common spaces from the perspective of health. Front. Public Health11, 1131180 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Halecki, W., Stachura, T., Fudała, W., Stec, A. & Kuboń, S. Assessment and planning of green spaces in urban parks: A review. Sustain. Cities Soc.88, 104280 (2023). [Google Scholar]
- 5.Hunter, M. R., Gillespie, B. W. & Chen, S. Y.-P. Urban nature experiences reduce stress in the context of daily life based on salivary biomarkers. Front. Psychol. 722 (2019). [DOI] [PMC free article] [PubMed]
- 6.Grigoletto, A. et al. Restoration in mental health after visiting urban green spaces, who is most affected? Comparison between good/poor mental health in four European cities. Environ. Res.223, 115397 (2023). [DOI] [PubMed] [Google Scholar]
- 7.Gulwadi, G. B., Mishchenko, E. D., Hallowell, G., Alves, S. & Kennedy, M. The restorative potential of a university campus: objective greenness and student perceptions in Turkey and the United States. Landsc. Urban Plan.187, 36–46 (2019). [Google Scholar]
- 8.Thomann, E. & Maggetti, M. Designing research with qualitative comparative analysis (QCA): approaches, challenges, and tools. Sociol. Methods Res.49, 356–386 (2020). [Google Scholar]
- 9.Sukhov, A., Friman, M. & Olsson, L. E. Unlocking potential: an integrated approach using PLS-SEM, NCA, and FsQCA for informed decision making. J. Retail Consum. Serv.74, 103424 (2023). [Google Scholar]
- 10.Rao, X., Qiu, H., Morrison, A. M., Wei, W. & Zhang, X. Predicting private and public Pro-Environmental behaviors in rural tourism contexts using SEM and FsQCA: the role of destination image and relationship quality. Land11, 448 (2022). [Google Scholar]
- 11.Kaplan, S. The restorative benefits of nature: toward an integrative framework. J. Environ. Psychol.15, 169–182 (1995). [Google Scholar]
- 12.Kaplan, R. & Kaplan, S. The Experience of Nature: A Psychological Perspective (Cambridge University Press, 1989).
- 13.Ulrich, R. S. et al. Stress recovery during exposure to natural and urban environments. J. Environ. Psychol.11, 201–230 (1991). [Google Scholar]
- 14.Alves, S., Betrabet Gulwadi, G. & Nilsson, P. An exploration of how biophilic attributes on campuses might support student connectedness to nature, others, and self. Front. Psychol.12, (2022). [DOI] [PMC free article] [PubMed]
- 15.Movassaghi, K. S. Requirements for the Degree (University of Louisiana at Lafayette, 2020).
- 16.Zhang, Z., Jiang, M. & Zhao, J. The restorative effects of unique green space design: comparing the restorative quality of classical Chinese gardens and modern urban parks. Forests15, 1611 (2024). [Google Scholar]
- 17.Tabrizian, P., Baran, P. K., Van Berkel, D., Mitasova, H. & Meentemeyer, R. Modeling restorative potential of urban environments by coupling viewscape analysis of lidar data with experiments in immersive virtual environments. Landsc. Urban Plan.195, 103704 (2020). [Google Scholar]
- 18.Bornioli, A. & Subiza-Pérez, M. Restorative urban environments for healthy cities: a theoretical model for the study of restorative experiences in urban built settings. Landsc. Res.48, 152–163 (2023). [Google Scholar]
- 19.Weber, A. M. & Trojan, J. The restorative value of the urban environment: A systematic review of the existing literature. Environ. Health Insights12, 117863021881280 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Ashihara, Y. Exterior Design in Architecture.27–30 (Architecture & Building, 1985).
- 21.Xiang, L., Cai, M., Ren, C. & Ng, E. Modeling pedestrian emotion in high-density cities using visual exposure and machine learning: tracking real-time physiology and psychology in Hong Kong. Build. Environ.205, 108273 (2021). [Google Scholar]
- 22.Liu, X., Moayedi, H., Ahmadi Dehrashid, A., Dai, W. & Thi, Q. T. Developments and evolution of housing architecture in the post-corona era with a health-oriented approach. Build. Environ.265, 111936 (2024). [Google Scholar]
- 23.Reece, R. et al. Exposure to green, blue and historic environments and mental well-being: A comparison between virtual reality head-mounted display and flat screen exposure. Int. J. Environ. Res. Public Health19, 9457 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Lee, J. H. & Ostwald, M. J. The ‘visual attractiveness’ of architectural facades: measuring visual complexity and attractive strength in architecture. Archit. Sci. Rev.66, 42–52 (2023). [Google Scholar]
- 25.Serra, J., Iñarra, S., Torres, A. & Llopis, J. Analysis of facade solutions as an alternative to demolition for architectures with visual impact in historical urban scenes. J. Cult. Herit.52, 84–92 (2021). [Google Scholar]
- 26.Yang, Z. et al. A study on visual impact assessment of the enclosing wall entourage of Chinese traditional private garden. Environ. Impact Assess. Rev.105, 107427 (2024). [Google Scholar]
- 27.Mao, Y., Qi, J. & He, B. J. Impact of the heritage building façade in small-scale public spaces on human activity: based on spatial analysis. Environ. Impact Assess. Rev.85, 106457 (2020). [Google Scholar]
- 28.Wen, H. et al. An assessment of the psychologically restorative effects of the environmental characteristics of university common spaces. Environ. Impact Assess. Rev.110, 107645 (2025). [Google Scholar]
- 29.Browning, M. H. & Rigolon, A. School green space and its impact on academic performance: A systematic literature review. Int. J. Environ. Res. Public Health16, 429 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Li, D., Chiang, Y. C., Sang, H. & Sullivan, W. C. Beyond the school grounds: links between density of tree cover in school surroundings and high school academic performance. Urban For. Urban Green.38, 42–53 (2019). [Google Scholar]
- 31.Mason, L., Ronconi, A., Scrimin, S. & Pazzaglia, F. Short-Term exposure to nature and benefits for students’ cognitive performance: a review. Educ. Psychol. Rev.34, 609–647 (2022). [Google Scholar]
- 32.Li, S. Y. et al. How do Spatial forms influence psychophysical drivers in a campus city community life circle?. Sustainability15, 10014 (2023). [Google Scholar]
- 33.Sun, D., Ji, X., Lyu, M., Fu, Y. & Gao, W. Evaluation and diagnosis for the pedestrian quality of service in urban riverfront streets. J. Clean. Prod.452, 142090 (2024). [Google Scholar]
- 34.Hong, X. C. et al. The physiological restorative role of soundscape in different forest structures. Forests13, 1920 (2022). [Google Scholar]
- 35.You, Y. et al. Effects of tree leaf color on human physical and mental recovery from a looking-Up perspective. (2024).
- 36.Zheng, S., Zhou, Y. & Qu, H. Physiological and psychological responses to tended plant communities with varying color characteristics. J. Res.35, 32 (2024). [Google Scholar]
- 37.Deng, L. et al. Empirical study of landscape types, landscape elements and landscape components of the urban park promoting physiological and psychological restoration. Urban For. Urban Green.48, 126488 (2020). [Google Scholar]
- 38.Lin, H., Hong, X. C., Wen, C. & Hu, F. The historical sensing of urban forest based on the indicators of CES and landscape categories: A case of Kushan scenic area in China. Ecol. Indic.166, 112440 (2024). [Google Scholar]
- 39.Rout, A. & Galpern, P. Benches, fountains and trees: using mixed-methods with questionnaire and smartphone data to design urban green spaces. Urban For. Urban Green.67, 127335 (2022). [Google Scholar]
- 40.Lis, A. et al. Evaluation of sense of safety and privacy in parks in relation to the topography, the presence of dense vegetation and other people in the area. Landsc. Urban Plan.242, 104948 (2024). [Google Scholar]
- 41.Wu, Y. et al. Integrating restorative perception into urban street planning: A framework using street view images, deep learning, and space syntax. Cities147, 104791 (2024). [Google Scholar]
- 42.Nie, X., Wang, Y., Zhang, C., Zhao, Y. & Kirkwood, N. The varied restorative values of campus landscapes to students’ well-being: evidence from a Chinese university. BMC Public Health24, 487 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Zhang, X. Improving english teaching strategies from the perspective of college students’. Mental Health J. Cases Inf. Technol.26, (2024).
- 44.Peng, Z., Zhang, R., Dong, Y. & Liang, Z. A. Study on the relationship between campus environment and college students’ emotional perception: A case study of Yuelu mountain National university science and technology City. Buildings14, 2849 (2024). [Google Scholar]
- 45.Stepansky, K., Delbert, T. & Bucey, J. Active student engagement within a university’s therapeutic sensory garden green space: pilot study of utilization and student perceived quality of life. Urban For. Urban Green.67, 127452 (2022). [Google Scholar]
- 46.Dzhambov, A. M. et al. Protective effect of restorative possibilities on cognitive function and mental health in children and adolescents: A scoping review including the role of physical activity. Environ. Res.233, 116452 (2023). [DOI] [PubMed] [Google Scholar]
- 47.Bower, I., Tucker, R. & Enticott, P. G. Impact of built environment design on emotion measured via neurophysiological correlates and subjective indicators: A systematic review. J. Environ. Psychol.66, 101344 (2019). [Google Scholar]
- 48.Liu, X., He, J., Xiong, K., Liu, S. & He, B. J. Identification of factors affecting public willingness to pay for heat mitigation and adaptation: evidence from Guangzhou, China. Urban Clim.48, 101405 (2023). [Google Scholar]
- 49.Lin, H. et al. Historical sensing: the Spatial pattern of soundscape occurrences recorded in poems between the Tang and the Qing dynasties amid urbanization. Humanit. Soc. Sci. Commun.11, 730 (2024). [Google Scholar]
- 50.Santos, T., Ramalhete, F., Julião, R. P. & Soares, N. P. Sustainable living neighbourhoods: measuring public space quality and walking environment in Lisbon. Geogr. Sustain.3, 289–298 (2022). [Google Scholar]
- 51.Zhang, L., Zhang, R. & Yin, B. The impact of the built-up environment of streets on pedestrian activities in the historical area. Alex Eng. J.60, 285–300 (2021). [Google Scholar]
- 52.Hino, K., Yamazaki, T., Iida, A., Harada, K. & Yokohari, M. Productive urban landscapes contribute to physical activity promotion among Tokyo residents. Landsc. Urban Plan.230, 104634 (2023). [Google Scholar]
- 53.Lin, H. et al. How social media data mirror spatio-temporal behavioral patterns of tourists in urban forests: A case study of Kushan scenic area in Fuzhou, China. Forests15, 1016 (2024). [Google Scholar]
- 54.Sun, P., Lu, W. & Jin, L. How the natural environment in downtown neighborhood affects physical activity and sentiment: using social media data and machine learning. Health Place79, 102968 (2023). [DOI] [PubMed] [Google Scholar]
- 55.Wilson, B., Neale, C. & Roe, J. Urban green space access, social cohesion, and mental health outcomes before and during Covid-19. Cities152, 105173 (2024). [Google Scholar]
- 56.Wang, R., Jiang, W. & Lu, T. Landscape characteristics of university campus in relation to aesthetic quality and recreational preference. Urban For. Urban Green.66, 127389 (2021). [Google Scholar]
- 57.Hami, A. & Abdi, B. Students’ landscaping preferences for open spaces for their campus environment. Indoor Built Environ.30, 87–98 (2021). [Google Scholar]
- 58.Xie, Q., Lee, C., Lu, Z. & Yuan, X. Interactions with artificial water features: A scoping review of health-related outcomes. Landsc. Urban Plan.215, 104191 (2021). [Google Scholar]
- 59.Altaher, Y. & Runnerstrom, M. G. Psychological restoration practices among college students. J. Coll. Stud. Dev.59, 227–232 (2018). [Google Scholar]
- 60.Neil-Sztramko, S. E., Caldwell, H. & Dobbins, M. School-based physical activity programs for promoting physical activity and fitness in children and adolescents aged 6 to 18. Cochrane Database Syst. Rev. (2021). [DOI] [PMC free article] [PubMed]
- 61.Hartig, T. Restoration in nature: Beyond the conventional narrative. In Nature and Psychology (eds. Schutte, A. R., Torquati, J. C. & Stevens, J. R.) 67 89–151 (Springer International Publishing, 2021).
- 62.Marselle, M. R. et al. Pathways linking biodiversity to human health: A conceptual framework. Environ. Int.150, 106420 (2021). [DOI] [PubMed] [Google Scholar]
- 63.Twohig-Bennett, C. & Jones, A. The health benefits of the great outdoors: A systematic review and meta-analysis of greenspace exposure and health outcomes. Environ. Res.166, 628–637 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 64.Astell-Burt, T. et al. Green space and loneliness: A systematic review with theoretical and methodological guidance for future research. Sci. Total Environ.847, 157521 (2022). [DOI] [PubMed] [Google Scholar]
- 65.Ning, W., Yin, J., Chen, Q. & Sun, X. Effects of brief exposure to campus environment on students’ physiological and psychological health. Front. Public Health11, 1051864 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 66.Markevych, I. et al. Exploring pathways linking greenspace to health: theoretical and methodological guidance. Environ. Res.158, 301–317 (2017). [DOI] [PubMed] [Google Scholar]
- 67.Bornioli, A., Parkhurst, G. & Morgan, P. L. The psychological wellbeing benefits of place engagement during walking in urban environments: A qualitative photo-elicitation study. Health Place53, 228–236 (2018). [DOI] [PubMed] [Google Scholar]
- 68.Herranz-Pascual, K. et al. Going beyond quietness: determining the emotionally restorative effect of acoustic environments in urban open public spaces. Int. J. Environ. Res. Public Health16, 1284 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 69.Shrestha, T., Di Blasi, Z. & Cassarino, M. Natural or urban campus walks and vitality in university students: exploratory qualitative findings from a pilot randomised controlled study. Int. J. Environ. Res. Public Health18, 2003 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 70.Boyd, F. Between the library and lectures: how can nature be integrated into university infrastructure to improve students’ mental health. Front. Psychol.13, 865422 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 71.Liu, X. The analysis of the influence of the network Chinese culture communication on college student sport psychology. Rev. Psicol. Deporte31, 186–193 (2022). [Google Scholar]
- 72.Du, Y., Zou, Z., He, Y., Zhou, Y. & Luo, S. Beyond blue and green spaces: identifying and characterizing restorative environments on Sichuan technology and business university campus. Int. J. Environ. Res. Public Health19, (2022). [DOI] [PMC free article] [PubMed]
- 73.Sun, S. et al. The psychological restorative effects of campus environments on college students in the context of the COVID-19 pandemic: A case study at Northwest A&F university, Shaanxi, China. Int. J. Environ. Res. Public Health18, 8731 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 74.Lee, K., Bae, H. & Jang, S. Effect of exercise combined with natural stimulation on Korean college students’ concentration and positive psychological capital: A pilot study. Healthc. Basel Switz.10, (2022). [DOI] [PMC free article] [PubMed]
- 75.Bai, Y., Wang, R., Yang, L., Ling, Y. & Cao, M. The impacts of visible green spaces on the mental well-being of university students. Appl. Spat. Anal. Policy17, 1105–1127 (2024). [Google Scholar]
- 76.Zhou, S., Wang, S., Liu, H., Green space exposure’s influence on mental well-being during Covid-19 campus lockdowns: A satisfaction mediating pathway. J. Environ. Eng. Landsc. Manag.32, 128–142 (2024). [Google Scholar]
- 77.Foellmer, J., Kistemann, T. & Anthonj, C. Academic greenspace and well-being - Can campus landscape be therapeutic?? Evidence from a German university. Wellbeing Space Soc.2, 100003 (2021). [Google Scholar]
- 78.Guo, L. H. et al. Does social perception data express the spatio-temporal pattern of perceived urban noise? A case study based on 3,137 noise complaints in Fuzhou, China. Appl. Acoust.201, 109129 (2022). [Google Scholar]
- 79.Nordbø, E. C. A., Raanaas, R. K., Nordh, H. & Aamodt, G. Disentangling how the built environment relates to children’s well-being: participation in leisure activities as a mediating pathway among 8-year-olds based on the Norwegian mother and child cohort study. Health Place64, 102360 (2020). [DOI] [PubMed] [Google Scholar]
- 80.Putra, I. G. N. E., Astell-Burt, T., Cliff, D. P., Vella, S. A. & Feng, X. Do physical activity, social interaction, and mental health mediate the association between green space quality and child prosocial behaviour? Urban For. Urban Green.64, 127264 (2021). [Google Scholar]
- 81.The People’s Government of Guangdong Province. 14th Five-Year Plan for the Development of Higher Education in Guangzhou (2022). https://www.gd.gov.cn/zzzq/zcjd/content/post_3616899.html (2023).
- 82.Finstad, K. Response interpolation and scale sensitivity: evidence against 5-point scales. J. Usability Stud.5, 104–110 (2010). [Google Scholar]
- 83.Pappas, I. O. & Woodside, A. G. Fuzzy-set qualitative comparative analysis (fsQCA): guidelines for research practice in information systems and marketing. Int. J. Inf. Manag.58, 102310 (2021). [Google Scholar]
- 84.Rihoux, B. & Ragin, C. C. Configurational Comparative Methods: Qualitative Comparative Analysis (QCA) and Related Techniques51 (Sage, 2009).
- 85.Woodside, A. G. Moving beyond multiple regression analysis to algorithms: calling for adoption of a paradigm shift from symmetric to asymmetric thinking in data analysis and crafting theory. J. Bus. Res.66, 463–472 (2013). [Google Scholar]
- 86.Kim, H. Y. Statistical notes for clinical researchers: assessing normal distribution (2) using skewness and kurtosis. Restor. Dent. Endod.38, 52–54 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 87.West, S. G. Structural equation models with nonnormal variables: problems and remedies. Struct. Equ. Model. Concepts Issues Appl. (1995).
- 88.Baumgartner, H., Weijters, B. & Pieters, R. The biasing effect of common method variance: some clarifications. J. Acad. Mark. Sci.49, 221–235 (2021). [Google Scholar]
- 89.Chuah, S. H. W., Tseng, M. L., Wu, K. J. & Cheng, C. F. Factors influencing the adoption of sharing economy in B2B context in China: findings from PLS-SEM and FsQCA. Resour. Conserv. Recycl.175, 105892 (2021). [Google Scholar]
- 90.Ragin, C. C. Redesigning Social Inquiry: Fuzzy Sets and Beyond (University of Chicago Press, 2009).
- 91.Mustafa, S., Zhang, W., Shehzad, M. U., Anwar, A. & Rubakula, G. Does health consciousness matter to adopt new technology?? An integrated model of UTAUT2 with SEM-f sQCA approach. Front. Psychol.13, 836194 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 92.Malekinezhad, F., Courtney, P., H bin Lamit & Vigani, M. Investigating the mental health impacts of university campus green space through perceived sensory dimensions and the mediation effects of perceived restorativeness on restoration experience. Front. Public Health8, 578241 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 93.Yusli, N. A. N. M., Roslan, S., Zaremohzzabieh, Z., Ghiami, Z. & Ahmad, N. Role of restorativeness in improving the psychological well-being of university students. Front. Psychol.12, (2021). [DOI] [PMC free article] [PubMed]
- 94.Black, N., Johnston, D. W., Propper, C. & Shields, M. A. The effect of school sports facilities on physical activity, health and socioeconomic status in adulthood. Soc. Sci. Med.220, 120–128 (2019). [DOI] [PubMed] [Google Scholar]
- 95.Clément, J. F. et al. Use of physical activity as a coping strategy mediates the association between adolescent team sports participation and emerging adult mental health. Ment. Health Phys. Act.27, 100612 (2024). [Google Scholar]
- 96.Pasanen, T. P., White, M. P., Wheeler, B. W., Garrett, J. K. & Elliott, L. R. Neighbourhood blue space, health and wellbeing: the mediating role of different types of physical activity. Environ. Int.131, 105016 (2019). [DOI] [PubMed] [Google Scholar]
- 97.Poortinga, W. et al. Associations of reported access to public green space, physical activity and subjective wellbeing during and after the COVID-19 pandemic. J. Environ. Psychol.97, 102376 (2024). [Google Scholar]
- 98.Cardinali, M. et al. Examining green space characteristics for social cohesion and mental health outcomes: A sensitivity analysis in four European cities. Urban For. Urban Green.93, 128230 (2024). [Google Scholar]
- 99.Liu, Q. et al. More meaningful, more restorative? Linking local landscape characteristics and place attachment to restorative perceptions of urban park visitors. Landsc. Urban Plan.197, 103763 (2020). [Google Scholar]
- 100.Barron, S. & Rugel, E. J. Tolerant greenspaces: designing urban nature-based solutions that foster social ties and support mental health among young adults. Environ. Sci. Polic.139, 1–10 (2023). [Google Scholar]
- 101.Chen, S., Sleipness, O., Christensen, K., Yang, B. & Wang, H. Developing and testing a protocol to systematically assess social interaction with urban outdoor environment. J. Environ. Psychol.88, 102008 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 102.Clarke, M. et al. Factors that enhance or hinder social cohesion in urban greenspaces: A literature review. Urban For. Urban Green.84, 127936 (2023). [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data are available from the corresponding author for reasonable requirements.