Human-nature connection (HNC) is a concept derived from investigating the formulation and extent of an individual’s identification with the natural world. This relationship is often characterized as an emotional bond to nature that develops from the contextualized, physical interactions of an individual, beginning in childhood. This outcome presents complexity in evaluating the development of HNC but suggests optimism in the pathways for enhancing lifelong HNC.

As urban populations increase, there is a growing recognition worldwide of the potential for urban green space to cultivate HNC and thus shape the environmental identity of urban residents.

The results of an online survey of 560 visitors to three community parks (managed primarily to provide a variety of physical, social and cultural opportunities) and three conservation parks (managed primarily to protect native plants and wildlife) in Madison, Wisconsin, USA, were used to investigate HNC.

Linear mixed effects models evaluated visitors’ HNC as a function of their (1) literacy and sentiment about wildlife species, (2) park experience, (3) number and frequency of nine childhood and adult recreation experiences, and (4) demographics.

Across the park response groups, the number and frequency of childhood and adult recreation experiences was significantly associated with HNC, and this positive association persisted in multiple recreation activities. Furthermore, species literacy and sentiment, visiting a park for 'Nature', and frequent and extended visitation also was significantly associated with HNC by park type.

Our research demonstrates the importance of lifelong recreation experiences in the development and enhancement of HNC and provides evidence for differences in the expression of HNC associated with particular attributes of urban park visitors and their views of wildlife.

Methodology

Study Area

Madison has a population of approximately 270,000 residents, covers approximately 260 km2, and is located in south central Wisconsin, USA (US Census Bureau, 2022).  Madison is currently the fastest growing city in Wisconsin and is home to the state capital and the University of Wisconsin-Madison (US Census Bureau, 2022). The study area is within the Yahara Watershed, now largely dominated by agricultural and urban land cover, and experiences four distinct seasons (Carpenter et al., 2007, Wisconsin State Climatology Office, 2010). 

The six selected parks were based on their classification as a community or conservation park; an estimated visitation rate; a central, western, or eastern location in Madison; and approval from the Madison Parks Division of the City of Madison (Figure 1).  The size of the community parks ranged from 19.07 ha to 101.50 ha, and the size of the conservation parks ranged from 24.39 ha to 39.17 ha.  The parks can be broadly described as mixed forest ecosystems with open grass areas and low levels of pavement and structural development.  Conservation parks contain native grasslands whereas community parks may contain native grasslands and/or mowed turf.  By definition, conservation parks are managed to protect native plant and wildlife species, resulting in the inclusion of vegetation and management practices supporting that objective (City of Madison Parks Division, 2022).  As a result of their conservation status, recreation therein is limited to physical activities such as hiking and snowshoeing and nature-based activities such as watching birds / wildlife and photography.  Dogs are not allowed in conservation parks.  Community parks are designed to provide a variety of physical, social, and cultural opportunities, including athletic fields and courts, playgrounds, and picnic shelters.  Community parks allow dogs that are leashed and licensed (City of Madison Parks Division, 2022).    

Study Population and Survey

We conducted an online survey to park visitors in three conservation parks and three community parks in Madison. Our research design was approved by the University of Wisconsin Education and Social/Behavioral Science Institutional Review Board as exempted research. We developed the survey in Google Forms and administered it in the parks using a park-specific quick response (QR) code printed either (1) on posters that were statically accessible to park visitors throughout the study period or (2) on postcards dynamically handed to park visitors at selected times during the study period.  The posters were visible outdoors in all six parks from 2021-09-04 through 2021-10-24 (high-use fall period) and from 2022-06-09 through 2022-08-24 (high-use summer period).  Postcards were distributed in the six parks on four Saturdays in both September and July from 10.00 to 12.00.  These dates and times were selected to coincide with the days and times with the highest number of park visitors, the availability of surveyors, and the approval of the Madison City Parks Division.  Each postcard had a unique three-digit number required to access the online survey.    Adults (18 years or older) were approached by the surveyor (lead author and/or student assistants trained in research ethics and project specifics) and invited to participate.  After verbally agreeing to participate (standard approach for exempted research), each potential respondent was asked three questions to check for nonresponse bias: (1) zip code, (2) year of birth, and (3) main reason for visitation. For poster and postcard respondents who continued on to take the online survey, the first question was a screening for informed consent, with only those who actively acknowledged consent continuing into the study’s content questions.  

The online survey consisted of 30 questions, grouped into four categories: (1) literacy and sentiment about wildlife species, (2) recreation and park experience, (3) HNC, and (4) demographics.  For species literacy and sentiment, respondents were asked questions evaluating (1) the correct photographic identification of six mammal species, each considered a generalist and likely present in the study parks, and (2) visitor sentiment about each species (Figure 2).  For recreation activity, respondents were asked questions about (1) the number and frequency of childhood and adult experiences with bird / wildlife watching, camping, canoeing / kayaking, fishing, gardening, hiking, hunting, nature photography, and picnicking; (2) the main reason for visitation; (3) prior visitation; (4) length of visit; and (5) distance of residence to the park.  For HNC, the abbreviated six-item short form of the Nature Relatedness Scale (NR-6) was used, with four statements from NR-Self (1-4) and two statements from NR-Experience (5 and 6):

• My connection to nature and the environment is a part of my spirituality.
• My relationship to nature is an important part of who I am.
• I feel very connected to all living things and the earth.
• I always think about how my actions affect the environment.
• My ideal vacation spot would be a remote, wilderness area.
• I take notice of wildlife wherever I am.

Demographic questions included age group, educational level, and gender.  The survey responses were in the form of a short answer (only identification of species), exclusionary checkboxes, or a 5-point Likert scale response (“Never” to “Very Often” or “Disagree Strongly” to “Agree Strongly”).  Wildlife literacy and sentiment questions were accompanied by a corresponding species-specific color photo (Figure 2).  Species sentiment was measured by species-specific exclusionary responses: 'I am happy they live at the park’, ‘I think they are important for the park ecosystem', 'I am concerned about their impact on human safety', 'I am concerned that they bring disease', 'I think they are a nuisance', or 'I am unsure how I feel or do not care’.  We piloted the survey with a focus group before administering it in the six parks to identify possible issues such as unclear language or challenges in viewing on mobile devices and adjusted our final survey accordingly.  All survey responses were anonymous.       

Analysis

Initial exploratory analysis included a random effect for park type (community and conservation) and a random effect and interaction term for survey type (postcard and poster).  The type of park was a significant factor, and the models afterwards were separated into two model sets, one for community park visitors and one for conservation park visitors.  A random effect was included for the parks sampled (3 community parks or 3 conservation parks) within the corresponding model set.  The type of survey was not a significant random effect, and the data of each type of survey were combined based on the type of park.  No differences were found between the potential and actual respondents by postcard with respect to zip code, year of birth, and main reason for visitation. This suggests that nonresponse bias was unlikely. 

Mixed-effects linear models were applied using the ‘lme’ function in the 'nlme' package (v3. 1-152; Pinheiro et al., 2021) of the R software, version 4.2.1 (R Core Team, 2019).  As our work forwards investigation on the specific factors associated with HNC (using the mean NR-6 score of a respondent) rather than the conventional application of NR-6 as a predictor of pro-environmental behavior or self-assessed well-being, we evaluated factors independently rather than collectively. Separate models were developed for community and conservation park survey data to evaluate HNC as a function of factors within four categories: (1) species literacy and positive species sentiment; (2) number, frequency, and type of outdoor recreation activities of childhood and adulthood; (3) main reason for visitation, prior visitation, length of visit, and distance of residence to the park; and (4) demographic factors (age category, educational level, and gender).  Species literacy was calculated as the average of responses recorded in six species: '1' for a correct answer and '0' for an incorrect answer.  Positive species sentiment was calculated as '1' for 'I am happy they live at the park' or 'I think they are important for the park ecosystem' or '0' for all other responses, recorded in six species.  The frequency of participation in childhood and adult recreation activities was calculated for each respondent as the average of responses from 1 (never) to 5 (very often) for the three activities respondents listed as the most frequent.  This method was chosen to reduce penalties for individuals who participated in fewer activities compared to those who participated in a greater number of activities.  The main reason for visiting had one of three options: 'Passive' included playground, picnic, and socialization, 'Active' included walking, exercise, and sports, and a third category 'Nature' included responses specific to nature-based activities (enjoying nature, watching birds / wildlife and nature photography).  The category 'Nature' is a subset of activities derived from activities otherwise categorized as 'Passive' but specific to nature-related recreation (Çetinkaya et al., 2017; Metin et al., 2017; Sampath et al., 2020).  The assumptions for the models, including linearity and normality, were graphically checked using residual versus fitted value plots and QQ plots of residuals.  To correct for the large number of tests performed, we used the false discovery rate (FDR) method (Benjamini & Hochberg, 1995, Benjamini et al., 2005). We set our false discovery rate at 5%. F-test p-values and confidence intervals reported in our paper are FDR-adjusted and can be compared to 0.05 to determine significance. For categorical predictor variables, if the overall F-test was significant by the FDR method, the means were separated using Tukey’s HSD method.