Table 2.
Summary of study characteristics and results.
| References, country | Study design | Sample size (age) | Green space exposure concept | Green space data source | Prosocial behaviour measure | Confounders adjusted in the model | Methods | Results in adjusted model | Quality |
|---|---|---|---|---|---|---|---|---|---|
| Amoly et al. (2014), Spain | Cross-sectional study | 2,111 (7–10 years) | a. Time spent playing in green spaces (a total number of hours during the last school period and summer holidays); b. Residential surrounding greenness in buffers of 100, 250, and 500 m; c. School greenness in a buffer of 100 m; d. Home-school greenness (average residential and school surrounding greenness in a buffer of 100 m, weighted by daily time spent at home and school); e. Residential proximity to a major green space (a binary variable indicating whether the child's home within 300 m of a major green space). |
Questionnaires; NDVI | Parent-reported prosocial scale from SDQ (a continuous variable). | Child's sex, school level, ethnicity, preterm birth, breastfeeding, exposure to environmental tobacco smoke, maternal smoking during pregnancy, responding person, parental educational achievement, parental employment status, and neighbourhood socioeconomic. | Quasi-Poisson mixed-effects models | No statistically significant association was found between all green space indicators and prosocial behaviour (non-significant in expected direction). | Fair |
| Andrusaityte et al. (2019), Lithuania | Cross-sectional study. | 1,489 (4–6 years) | a. Time spent in a city park (hours per week); b. Residential surrounding greenness in buffers of 100 m. |
Questionnaires; NDVI | Parent-reported prosocial scale from SDQ (a binary outcome: borderline/abnormal vs. normal). | Child's sex, birth weight, wheeze, asthma, allergy, BMI, breastfeeding, siblings, paracetamol and antibiotic usage during the first year of life, maternal education, tobacco smoke, age at childbirth. | Logistic regression | Increased time spent in city parks per 1 h per week was associated with decreased odds of borderline/abnormal prosocial behaviour: aOR = 0.98 (0.96, 0.99) (significant in expected direction). Non-significant association was found for residential surrounding greenness (non-significant in expected direction). |
Fair |
| Balseviciene et al. (2014), Lithuania | Cross-sectional study. | 1,468 (4–6 years) | a. Residential surrounding greenness in a buffer of 300 m; b. Proximity to the nearest city parks (transformed using the square root function in meters). |
NDVI | Parent-reported prosocial scale from SDQ (a continuous variable). | Child's age, sex, and parenting stress. | Linear regression | Analysis was stratified by mother's educational level. Increased distance to city parks was negatively associated with prosocial behaviour among lower education group: β = −0.029 (p < 0.05) (significant in expected direction). Residential greenness was negatively associated with prosocial behaviour among higher education group: β = −1.104 (p < 0.05) (significant in unexpected direction). |
Fair |
| Bates et al. (2018), USA | Experimental study (one- group post-test-only design) | 3,345 and 3,710 observations at the first (T1) and second (T2) time, respectively (age ranges from pre-kindergarten to 8th grade) | Schoolyard renovation by increasing the presence of natural components (e.g., grass, trees) and also the quality (e.g., aesthetics; facilities). | In-person observation | Positive social interaction, measured by behavioural mapping using System for Observing Children's Activity and Relationship during Play (SOCARP). It was measured two times (T1, T2) after schoolyard renovation. | No confounders adjusted in the analysis | Chi-square test | The percentage of observed positive social interaction or prosocial behaviour increased from T1 (27.10%) to T2 (35.20%) (p < 0.001) (significant in expected direction). | Poor (no pretest, no randomisation) |
| Carrus et al. (2015), Italy | Experimental study (two- group post-test-only design) | 39 (1.5–3 years) | Children's spending time in school green space vs. in internal space of school | In-person observation | Positive social interaction, measured by a behavioural checklist to record frequency of positive relational behaviours | No confounders adjusted in the analysis | ANOVA | After children were exposed to green space, more frequent positive relational behaviours were observed on days when children spent time in school green space compared to days when they did not (p = 0.038) (significant in expected direction). | Poor (no pretest, no randomisation) |
| Dopko et al. (2019), Canada | Experimental study (two- group post-test-only design) | 80 (mean age = 10.49 years) | Children' spending time outdoors at the nature school vs. indoors at the museum | In-person observation | Using two tasks: a. A windfall task by asking children to imagine that they received money and what they decided on four available options (buy things they want, give to charity, spend on gifts for other people, and save for the future). Children who decided for charity and spending on gifts for other people represent higher prosociality. b. A tangram task by asking children to imagine that they assigned 11 tangrams from three categories: easy, medium, and hard to someone else in their class. Children who assigned more tangrams in easy and medium categories, and few in hard category represent higher prosociality. |
No confounders adjusted in the analysis | Paired sample t-test | Windfall task: Mean score for spending money on charity was statistically higher among children visiting nature school than museum: β = 3.66 (0.06, 7.26) (significant in expected direction). Mean score for spending money on gift was lower among children visiting nature school than museum: β = −4.15 (−8.32, 0.03) (non-significant in unexpected direction). Tangram task: Mean score for assigning easy tangram was statistically higher among children visiting nature school than museum: β = 0.74 (0.01, 1.46) (significant in expected direction). Mean score for assigning hard tangram was statistically lower among children visiting nature school than museum: β = −1.29 (−2.15, −0.42) (significant in expected direction). |
Poor (no pretest, no randomisation) |
| Mayfield et al. (2017), USA | Experimental study (two- group pretest-post-test design) | Two elementary schools for each intervention and control groups. This study included 3,588 SOCARP scans representing 1,196 child recess days with 3 rotation conducted. | The intervention was carried out by improving the quality of playground through adding playground marking with colourful interactive games. In addition, intervention schools received equipment to use with the game and training sessions for teachers. | In-person observation | Positive social interaction, measured by behavioural mapping using System for Observing Children's Activity and Relationship during Play (SOCARP). | Scans nested within days nested with schools | Mixed- effects regression analysis | There was a non-significant decrease in prosocial behaviour in the verbal or physical manner before and after the intervention (non-significant in unexpected direction). | Fair |
| McEachan et al. (2018), UK | Longitudinal study | 2,594 (aged 0 at baseline, 4 years at follow up) | a. Satisfaction with green space (asked among a sub-sample of 832 (32%) only) b. Time spent playing outside (minutes per week calculated for winter and summer months - asked among a sub-sample of 832 (32%) only) c. Residential surrounding greenness in buffers of 100 m, 300 m, and 500 m |
Questionnaires; NDVI | Parent-reported prosocial scale from SDQ (a continuous variable) | Child's age, sex, maternal age, cohabitation status, maternal education, subjective poverty, household size, neighbourhood deprivation index, mother's smoking behaviour, and mother's treatment record of mental disorder | Linear regression | Analysis was stratified by ethnicity (white British vs. south Asian). Satisfaction with green space was significantly associated prosocial behaviour among south Asian children only: β = 0.20 (0.02, 0.38) (significant in expected direction). Time spent playing outside was not associated with prosocial behaviour among both ethnicities (non-significant in expected direction for south Asian children and non-significant in non-reported direction for white British children). Residential greenness in all buffer distances were not associated with prosocial behaviour among both ethnicities (non-significant in expected direction). |
Good |
| Odgers et al. (2012), UK | Cross-sectional study | 2,024 (12 years) | Percentage of green space in a buffer of 0.5 mile (measured only among a sub-sample of 200 neighbourhoods) | A systematic social observation using Google Street view | A combined parent and teacher's reports of Revised Rutter Parent Scale for School-Age Children (a continuous variable) | No confounders adjusted in the analysis | Linear regression | No association was observed between percentage of green space and prosocial behaviour (non-significant in unexpected direction). | Poor (no control for confounders) |
| Park et al. (2016), South Korea | Experimental study (one- group pretest-post-test design) | 336 (5–7 years) | Participation in 24-session horticultural activity program that included indoor and outdoor activities, such as transplanting, planting seeds, making and applying eco-friendly fertilizer, observing vegetable plants, harvesting, etc. | In-person observation | Teacher-reported of prosocial behaviour using the revised questionnaire with four subscales (helping, sharing, cooperation, kindness) (a continuous variable) | No confounders adjusted in the analysis | Paired sample t-test | All prosocial behaviour scales (helping, sharing, cooperation, kindness) increased from pretest to post-test (significant in expected direction). | Fair |
| Richardson et al. (2017), UK | Longitudinal study | 2,909 (aged 4 years at baseline, 6 years at follow-up) | a. Percentage of park space in a buffer of 500 m b. Percentage of total natural space in a buffer of 500 m c. Garden access (indicating whether the child had access to a private garden) |
Land cover map; Questionnaire | Parent-reported prosocial scale from SDQ (a continuous variable) | Child's age, sex, screen time, household income, educational attainment, carer's mental health, and neighbourhood socio-economic status | Linear regression | Analysis was stratified by the child's sex and household educational level. Percentage of total natural space was significantly associated with prosocial behaviour among girls: β = 0.14 (p < 0.01) and among high education households: β = 0.12 (p < 0.05) (significant in expected direction). Percentage of parks was not significantly associated with prosocial behaviour among all sub-sample groups (non-significant in expected direction). Access to private garden was not significantly associated with prosocial behaviour among all sub-sample groups (non-significant in unexpected direction). |
Good |
| Sobko et al. (2018), Hong Kong | Cross-sectional study | 299 (2–5 years) | Connectedness to nature (enjoyment of, empathy for, responsibility toward, and awareness of nature) | Questionnaire | Parent-reported prosocial scale from SDQ (a continuous variable) | No confounders adjusted in the analysis | Structural equation modelling | Greater responsibility toward nature was significantly associated with improved prosocial behaviour: β = 0.77 (significant in expected direction). | Poor (no control for confounders) |
| Van Aart et al. (2018), Belgium | Longitudinal study | 172 (6–12 years at baseline, 9–15 years at follow-up) | a. Percentage of semi-natural and forested area in a buffer of 2,000 m b. Percentage of agricultural area in a buffer of 300 m |
Land cover map | Parent-reported prosocial scale from SDQ (a continuous variable) | Child's age, sex, and parental socio-economic status | Linear regression | Percentage semi-natural and forested area was not associated with prosocial behaviour (non-significant in unexpected direction). Percentage of agricultural area was not associated with prosocial behaviour (non-significant in expected direction). |
Fair |
| van Dijk-Wesselius et al. (2018), Netherlands | Experimental study (two- group pretest-post-test design) | About 700 (7–11 years) | The intervention was carried out by increasing the presence of natural components (e.g., grass, trees) and also the quality of schoolyards (e.g., aesthetics; facilities). | In-person observation | a. Prosocial orientation assessed by self-administrated Social Orientation Choice Card (SOCC) (a binary variable) b. Self-reported prosocial scale from SDQ (a continuous variable) |
Child's sex, grade level | Multi-level analysis | Analysis was stratified by grade levels (4, 5, and 6). Proportion of prosocial orientation in grades 4 and 5 in intervention compared to control group increased from baseline to the follow-up, but there was a significant decrease in grade 6 (significant in expected and unexpected directions). There was no significant increase of self-reported prosocial behaviour (non-significant in non-reported direction). |
Fair |
| Whitten et al. (2018), Australia | Cross-sectional study | 26,848 (mean age = 11.92 years) | Connectedness to nature | Questionnaire (self-report) | Self-reported prosocial scale from SDQ (a continuous variable) | Child's sex, social supports, empathy, attention, and neighbourhood socio-economic status | Linear regression | Increased connection to the nature was associated with higher prosocial behaviour: β = 0.12 (p < 0.001) (significant in expected direction). | Fair |