Anthropogenic climate change and biodiversity loss are major threats not only to non- human living beings but also to our own survival1,3,8,9. Despite there being apparent international consensuses to better preserve nature (e.g., the Convention on Biological Diversity4 and Aichi Biodiversity Targets), to limit global warming to 2°C (e.g., the Paris Agreements) and to create a sustainable, equitable world (e.g., the Sustainable Development Goals - SDGs), actions remain limited and sustainable targets have failed to be achieved5,10. Why do citizens and governments still find it hard to consider the health and well-being of humans as dependent upon the health of the natural world1,3,8?
One possible explanation is cultural. Monotheist religions11,12 and influential modern philosophers13,14 advocated a discontinuity between the inner-worlds of humans and those of other living beings, which has led people to consider themselves disconnected from, and dominant over, the rest of nature (the modern “Western” worldview). Another explanation is the increased urbanization of the world’s human population. Today, more than 55% of people live in urban areas, and this proportion is expected to increase to 68% by 205015. A major consequence is the increased disconnection of people from experience of natural habitats16. The combination of both psychological and physical disconnects from the natural world can result in devaluation of nature, thereby legitimizing and facilitating destructive practices towards nature by individuals and societies (Fig. 1).
What can be done to modify these destructive trends? International scientific assessments, such as the MA, IPCC and IPBES, indicate that healthy natural systems are crucial to reach sustainability1–3. The IPBES further claims that sustainable goals will not be achieved without a “transformative change” including an increase in “awareness of connectivity in the environmental crisis and new norms regarding interactions between humans and nature” (Fig. 1). A key societal trait that is relevant to achieving sustainability is Human-Nature Connectedness12–14 (HNC), which has been described and defined in different, often complex, ways (Supplementary information (SI) Appendix 1). We define HNC as the extent to which humans include the natural world in the self or see themselves as part of nature. A number of recent studies found that estimates of HNC were positively correlated both with pro-environmental behaviour17,18 and with human welfare19,20.
Due to paucity of data, prior synthetic analyses of the roles of HNC were forced to lump studies with disparate methodologies, whether they be experimental or correlational, and whether or not they used metrics of HNC that had been validated by methods that we describe in SI Appendices 2 and 3. Here, we take advantage of the rapidly-increasing literature to restrict our database to studies using at least one of the validated quantitative metrics of HNC13, 14,19,20-23 and to separate experimental from correlational studies. Thus, we provide, for the first time, a robust, coherent and global synthesis of the HNC literature with separate meta-analyses of experimental and correlational studies (Extended data Fig. 1 and Extended data Table 1). This allows us to identify practices that have the potential for increasing HNC, and investigate whether HNC is a key element of sustainability, in particular nature conservation and human welfare.
Experimental studies
Existing experimental studies are strongly biased towards adults and towards high-income, industrialized countries with a "Western" worldview (Extended data Fig. 2). Among them, we identified 6 types of experimental designs: (1) exposure to real nature: direct contact with nature, either outdoors or indoors; (2) exposure to virtual nature: videos or pictures of nature; (3) mindfulness: focusing one’s attention on one's inner self and one's environment in the present moment; (4) environmental education: exposure to naturalist, scientific and ecological knowledge of the natural world; (5) combination of exposure to real nature + environmental education and, (6) combination of exposure to real nature + mindfulness.
We combined two types of comparison: (1) sequential comparison of trait measurements made on the same participants before and after experimental treatment and (2) simultaneous comparison of control and treatment groups (see Methods and SI Appendix 4).
All experimental designs, except environmental education, affected HNC significantly and positively (Fig. 2 and Extended data Table 2). The clearest positive effects on HNC were those of mindfulness, with or without experience of real nature, while estimated positive effects of environmental education were low. These analyses of experimental studies show that exposing individuals to nature, either real or virtual, does improve HNC. They also suggest exciting lines of future research to improve the efficiency of experimental programs designed to raise HNC by including mindfulness practices and/or by trying less anthropocentric transmission of scientific information.
Correlational studies
Like experimental studies, correlational studies are strongly biased toward adults from industrialized countries. Meta-analysis of their results shows that HNC is negatively correlated with materialism/consumerism and political conservatism and positively correlated with the following suite of behaviours, opinions and personality traits: naturalist knowledge; time spent in natural outdoor spaces; engagement in mindfulness practices; pro-environmental values; humanistic values; happiness and good health.
In sum, individuals with high HNC had deeper knowledge of nature, spent more time in natural outdoor spaces, engaged more in mindfulness practices and were happier and healthier than those with low HNC. It is not surprising, then, that high HNC individuals displayed more pro-environmental values, expressed as concern and attitudes toward the ecological environment. They were also more humanistic, in the sense of more strongly expressing their moral responsibilities to other humans, and their sense of being part of communities within society (Fig. 2; see Methods summary, Methods and SI Appendix 4 for detailed categories and Extended data Table 3, Extended data Figs. 3 to 6).
Additionally, one-third of the correlational papers (26 out of 108) addressed causal hypotheses using either structural equation models or path analysis. Although these approaches have limited ability to infer causality, these studies nonetheless suggest impacts on HNC of both contact with nature and mindfulness, in agreement with results from experimental data (Fig. 2 and Extended data Table 2). In the opposite direction of causality, they also suggest impacts of HNC on pro-environmental values, human well-being and health and pro-environmental behaviours.
Taken together, correlational data support conclusions from experimental studies but also add to them, showing HNC to be positively linked to values and behaviours that enhance ecological sustainability and well-being, and negatively linked to non-environmental or anti-environmental values. Future experimental research is needed to test causal relationships, particularly the hypothesis suggested by our meta-analysis that raising HNC would enhance both human welfare and conservation of nature.
Implications for Policy
Our meta-analyses robustly show that the extent to which people feel part of the natural world can be enhanced by very simple interventions involving contact with nature and mindfulness practices, at least for industrialized cultures that form the bulk of studies. They also show that validated HNC indices are positively linked to human welfare and nature conservation. Thus, improving HNC through contact with nature and mindfulness can be a valuable way to help individuals to understand and experience how much human welfare and nature conservation are interconnected21. As an example, “greening” schoolyards could improve citizens’ HNC and well-being and thereby foster support for sustainable policies, adding to its known positive effects on urban biodiversity and climate adaptation (the virtuous circle in Fig. 1).
Surprisingly, we found little impact of environmental education on HNC, and we hypothesize that this is due to the traditional anthropocentric transmission of scientific knowledge. While this result remains to be further tested, studies in non-industrialized populations28 and children29 suggest that modifying the way scientific information is transmitted using non-anthropocentric knowledge about non-human species can have a positive impact on HNC. Examples of non-anthropocentric approaches include focusing on similarities between humans and other species22–26 and taking the perspectives of other species and empathizing with them27.
If we are to develop efficient research and conservation programs aimed at simultaneously protecting humans and nature worldwide21, we urgently need longitudinal studies in children and in non-industrial traditional cultures. Children from industrialized societies are known to show a strong affinity for non-humans29, which tends to fade with age30,31 and develop into low ecological concern at adulthood32,33 along with the acquisition of anthropocentric cultural norms34. We know much less about equivalent developmental changes in traditional societies, although pro-environmental behaviours may be more likely to persist into adulthood in societies with a high interdependency between humans and nature28. Accumulation of knowledge about cultural and developmental norms and values in diverse societies, both traditional and industrial, should assist in building international environmental and educational policies and empowering citizens and governments to take actions and achieve sustainable targets at a global scale.
Methods summary
For experimental data, we performed the meta-analysis on Standardized Means Difference (SMD; Hedges’ g). For correlational data, meta-analyses were done on Pearson correlations after Fisher transformations. All estimates were calculated for each factor with the rma.mv function from the metaphor package35 in R, which permits to fit meta-analytic multivariate fixed- and random/mixed-effects models with or without factors via linear (mixed-effects) models. Since there were sometimes several measures of HNC and factors in one individual paper and/or one study, we included random effects including lab (multiple data from the same individual paper) and study (multiple data points from the same subjects) to take non-independence of data into account. Since heterogeneity of the true effects within a study (within a lab) could be (or not be) identical, we added a higher-level random effect at the estimate level so that random effects = ~ 1|lab/study/estimate.id. Then, we transformed the estimates (SMD and Fisher’s Z) into R estimates.
Correlational studies include 9 factors that we organized into 4 broader categories: (1) nature conservation: pro-environmental values (personal interest in and appreciation of nature) and pro-environmental behaviours (self-reported and observed pro-environmental behaviours as well as pro-environmental intentions); (2) human welfare: psychological health, social well-being, physical health and humanistic values (value of human beings including human freedom and progress); (3) lifestyle characteristics: physical contact with real nature (e.g., time spent in nature), environmental education (scientific or naturalist knowledge) and mindfulness practices, and (4) non-environmental values: “materialistic” values (values focusing on wealth, possessions, image, and status including consumerism) and conservatism (political values emphasizing traditional social institutions).