Restorativeness in Natural and Urban Environments: A Meta-Analysis. Issue 2 (April 2021)

• Record Type: Journal Article
• Authors: Menardo, Elisa; Brondino, Margherita; Hall, Rob; Pasini, Margherita
• Is Part Of: Psychological reports. Volume 124:Issue 2(2021)
• Page Start: 417
• Page End: 437
• Publication Date: 2021-04
• Subjects: Meta-analysis -- nature environment -- urban environment -- restorativeness -- environmental psychology; Psychology -- Periodicals; Psychiatry; Psychology; Psychologie -- Périodiques; Psychology; 150.5
• DOI: 10.1177/0033294119884063
• Languages: English
• ISSNs: 0033-2941

Abstract

In recent decades, there has been an increasing volume of research aimed out quantifying the extent to which the natural environment can assist in restoring mental, emotional, and physical well-being. It is commonly assumed that natural environments are more restorative than urban environments. However, studies addressing this issue use a variety of methods for data collection making it difficult to compare the findings of different studies. The research reported here uses a meta-analysis aimed at estimating how much natural environments are perceived as being more restorative than urban environments. We investigated the role of moderator variables such as research design, kind of natural environment, participants, measurement instruments used or the context in which the data were collected. PsycINFO, PsycARTICLES, Scopus, SpringerLINK, Web of Science online databases were used to identity all peer-review articles on restorativeness published to date ( k = 167). Reference sections of obtained papers were examined for additional studies. Only 22 studies met inclusion criteria (direct exposure to environment, comparison between one outdoor environment with natural element and one without natural element, and restorativeness measured by self-report scale) and were included in meta-analysis. Results show that natural environments are perceived to be more restorative than urban environments (Cohen's d (confidence interval) = 1.99 (1.38–2.61)). Significant heterogeneity between the study was found ( Q (19) = 503.16, p < .001) and variability within studies was very high ( I 2 = 97%). However, subsequent univariate moderator analyses were not significant. Other methodological differences (e.g., lighting conditions) could explain this variability. We concluded that the variability in studies is more likely to be due to individual differences (e.g., age, connections to nature, and environmental attitude) than the methodological differences.