Urban / Suburban Park Use : Links to Personal Identity ?

Public greenspaces provide an opportunity for community members to engage with the outdoors. In many locations, however, parks are under used. In an effort to gauge the potential for outdoor interaction and ecosystem education, we conducted a survey of residents from a central New Jersey, USA, county. Our correlation analysis indicated that park use could be related to socioeconomics and in particular education, environmental literacy, pet ownership, outdoor enjoyment and preferred environment. Variables relating to mood and other personal characteristics were more strongly associated with individual identity characteristics. Through multivariate analyses, we offer an organizing framework that can help tailor outdoor greenspace improvement/restoration and programming to identity categories. These categories are a combination of where an individual lives, enjoyment of the outdoors, education and socio-economics, sense of community, institutional trust, and pet ownership. key words: Urban park, Suburban park, Ecosystem education, Ecosystem restoration.

key words: Urban park, Suburban park, Ecosystem education, Ecosystem restoration.

IntRodUctIon
Green spaces provide human social and health benefits 1 .They can also address broader ecological needs for local and regional benefit 2 .With only one-fifth of the US population engaging in sports and recreation activity 3 , green spaces are, on average, under used.In this note, we provide evidence showing that activity designers seeking to engage the public in green space could benefit from understanding park use and environmental literacy at the local levels.
There has been much study into the benefits of maintaining public park lands (hereafter green space).While green space use and human health and mood, in particular, has garnered much research attention (see review in [1]), connections between these spaces and increases in community capital have also been drawn (e.g., [4]).Such investigations have brought to light the disparities between access to desired and undesired green spaces 5 .These authors provide evidence that parks can serve to further isolate socio-ecologically disadvantaged communities, underscoring the need to not only provide green life in these spaces but also sufficient amenities (e.g., safe location, motivating scenery, positive use, and ample facilities) 6 .
Given this link between access, advantage, and benefit of green space, urban environments are attracting researcher attention regarding public use (e.g., [7]).Such studies support the notion that urban green space use varies by location, age, gender, and race 8,9 .This variation aside, it is clear that these spaces are underused 3 compared to capacity and in some cases are seen as detrimental 5 .
There is, however, the potential to increase park use with targeted intervention and promotion of community activity (e.g., movements such as Ciclovia 10 ; or direct park amenification 11 ).Indeed, use of parks has been shown to increase when members of the surrounding community were engaged, versus not being involved, in decisionmaking about park features and amenities 12 .At the very least, simple consultation with residents about resource management was sufficient to increase positive perception of the benefits of a protected greenspace 13 .Beyond consultation and awareness, there is evidence to suggest that targeted activities/ events in greenspaces, not only increases park use, but also contributes to a more positive view of that park 14 .These authors caution, however, that direct park amenification is not sufficient to increase park use.Rather park use is correlated to a number of park user-related factors.These studies suggest that questions regarding the lifestyle practices and use of local and regional green spaces need to be addressed, especially across urban and exurban landscapes where a myriad of park perceptions and associated differences in park quality exist In response to this need, we designed a questionnaire to learn more about local and regional green space use.We report here the results from this preliminary survey.Based on these data, we make recommendations for the types of outdoor educational intervention that will foster environmental literacy.

methodS
We selected residents from a stratified random sample of NJ postal addresses in Middlesex County.This survey encompassed 6 townships and residents were all within 0.5miles of a green space/ park.This area of New Jersey was chosen because of the increasing levels of urbanization across a broad range of socio-economic backgrounds.
We designed and validated a questionnaire with multiple sections: local and regional park use, environmental literacy and self-report about environmental literacy, scientific practices, trust in information sources, characteristics about oneself and demographic information.We asked about preferred park features but report those results elsewhere.The full questionnaire can be found in the Appendix.All work was conducted with Rutgers University Institutional Review Board approval.
The process of design began with a list of constructs (Table 1).Given our interest in offering outdoor educational interventions, we choose to focus on environmental and science literacy, personality, trust, and demographics as these have been shown to be important drivers environmental behavior change 15 .Items were originally selected form Jordan's previous work; with the exception of the environmental literacy questions that were taken from a standardized Environmental Literacy assessment 16 .A few additional questions were added to fill in gaps.
To validate the items, we invited 10 people who were representative of the age 20-70 living in central New Jersey target area to complete the questionnaire.These individuals were recruited through opportunity sampling and no compensation was provided.The lead author interviewed individuals following questionnaire completion.Through these discussions, we sought to understand where there was a gap between participant and interviewer ideas regarding the questionnaire responses.Modifications to questions were made with the participant present.
Following this, a subset of questionnaires were sent to our target audience generated from a random search of white pages.comaddresses in the target area.One hundred questionnaires were sent at random with 21 returns.Based on these, further question modifications were made in an effort to reduce ambiguity as noted by both lead authors.We then asked the same 10 volunteers to read for understanding the questionnaire.After this edit, the questionnaire was mailed.Three hundred and eighty questionnaires were mailed (and not returned as undeliverable) with 74 completed returns; making a 19.5% response rate.
Survey responses were first inspected in a Multiple Correlation Analysis to determine associations and possible trends.The authors designed this method in an effort to holistically view correlations that exists between questionnaire approaches.To help make meaning of our associations, we next analyzed our data using Multiple Correspondence Analysis, which is an ordination technique that is able to reduce data redundancy in large sets of categorical data.Individuals are plotted along several composite dimensions made up of the variables (or in our case, the questions) in our study.From these dimensions, we ran a Cluster Analysis to determine which variables likely provide the greatest explanation for differences between individuals in our study population.To help avoid spurious clustering, we chose to retain the clusterings that provided 7 or fewer clusters.From there we used discriminant Function Analysis (dFA) to provide a sense of the robustness of our clusters were.dFA requires us to provide the clusters and associated variables, while the analysis indicates how well group membership could be predicted.We found that two clusterings resulted in a high percentage of accurate placements (More detail available by request).The results of the correlations and two clusterings are described below.

ReSULtS
We found that variable categories, such as the trust constructs, tended to be highly correlated with another.We also found a number of the education and environmental literacy variables to correlate with self-report personality characteristics and experiences.In summary:

demographics
Females tended to score higher in environmental knowledge and older individuals were less likely to report better health and to travel to regional parks.

by town
People who felt a part of a community within a township were more likely to report themselves as smart but were less likely to report optimism.These individuals also reported greater issue awareness within that town.Community feeling within a township varied by town as did park quality and rating of oneself as smart tended to vary with town.

education, environmental Literacy, and environmentalists
People who report more education were more likely to report more affluence, issue awareness within a town, pro-environmental views, better health, greater optimism and mood, and a better schooling experience.They were also more likely to say that the community should pay for environmental care People who rated themselves as having pro-environmental views tended to have greater environmental knowledge, a higher regard for learning, some understanding of science, and a greater awareness of parks in their area.Individuals with pro-environmental views also tended to rate themselves as being less anxious or depressed, preferring less built environments, having a better sense of how science is done, enjoying the outdoors, enjoying learning, and were less likely to report that environmental problems can always be fixed later.These individuals also tended to report that they would pay for environmental care.These individuals reported less trust in local businesses, corporations, or social groups.It is important to note, however, that a high number of correlation analyses conducted in the same analysis could result in some spurious significant correlations given the likely increase in type I error.
Because we are interested in understanding park use and targeting environmental literacy type interventions, we felt that the clustering procedure would allow us to determine similarities and differences among individuals.As previously mentioned, we retained two clusterings: Figure 1 (with 6 groups) and Figure 2 (with 4 groups).The former yielded clusters that were most characterized by differences in environmental literacy, education, enjoyment of outdoors, preferred environment types, pets, and trust.The latter figure yielded 4 major clusters that were most characterized by town, issue awareness, and community feel.

dIScUSSIon
In summary, our correlation analysis indicated that park use could be related to socioeconomics and in particular, education, environmental literacy, pet ownership, outdoor enjoyment and preferred environment.These variables can be related to perception of self with respect to the environment.Variables such as mood and other personal characteristics, which did not correlate as strongly with park use, are more strongly associated with individual characteristics (Figure 1).Most individuals felt, however, that responsibility tended to rest more with the government in terms of environmental care as opposed to self.

A s o t h e r s t u d i e s h a v e s h o w n , socioeconomics plays an important role in park use
and more importantly in park perception (e.g., [5]).Living conditions and sense of community, as well as practices such as pet ownership and child-rearing also seem to play an important role in predicting an individual's park use.Most importantly we see that certain elements of environmental ethic do not play a role in local park use, yet to be environmentally literate does relate to sense of community within a town and education.
Based on our findings we suggest that programs encouraging greenspace use, especially as they pertain to environmental literacy, should be framed with identity.Identity frames are defined as a cognitive framework or scheme belonging to individuals and their characteristics, or groups of individuals as identified from social experiences 17,18 .Identity frames can affect decision and interpretations of information.For example, personal identity can affect political actions and beliefs (e.g., with respect to policy 19 ).As evident in our data, greenspace use is more complicated than labeling individuals as preferring the outdoors or in or as being "environmental" or not.
Local parks were viewed as being along a gradient of poor to sufficient quality.Certainly, efforts targeted toward community may aid in developing a local environmental appreciation and a greater desire to participate in park upkeep.Post-hoc analysis indicated that the less affluent townships were associated with lower park quality than the more affluent townships.Targeting pet owners and may be a good angle for community involvement in local park upkeep, while targeting parents as a resource group might work for more regional level parks.
Although covering a small geographic area, our study has implications for how park use and intervention could be framed to motivate different groups.In more urban and sub-suburban areas, local government could be engaged in activity to promote environmental quality/literacy because trust of government and local community organizations tends to be high in individuals who prefer this type of environment.When targeting urban spaces, encouraging "late modern" type of environmental and space use ideals might also be important.In more suburban places, efforts to work with local groups and media may encourage greenspace use.However, with more rural-preferring individuals, trust in general might be lower, and therefore working with social networks may prove more fruitful.While the above may be somewhat speculative at this point and does rely in part on individuals living in their preferred environments, it would not be difficult to test some of these ideas with a larger sample.

entities that are closer in space are more similar to each other. cluster descriptors are based on variables that are in common with that group. In italics are hypothesized identity titles for individuals in these groups. Further analysis is required to test these.
responsibility for environmental care tended to report a lower mood in terms of anxiety and depression.trustIfan individual trusted an entity, they were more likely to trust another.With more education comes less trust in schools.Higher trust in schools is associated with rating of lower park quality (as is PersonalityI n d i v i d u a l s w h o p l a c e d ex t e r n a l

Ideas about Science and the environment 9
Efforts to link environmental quality to health and Please fill-in, circle, or check the most appropriate response.note:Here we consider parks to be green spaces that range in type and use, and are outdoor spaces that typically facilitate activities (e.g., sports, boating, hiking, etc.), rest, or picnics..Circle the response that most closely aligns with your definition of science as a way of knowing.Please circle only one.Rank in order from 1 to 10 the issues you believe Americans should be making effort to fix. 1 being the issue that should have the most effort, down to 10 with the issue needing the least effort.(strongly agree) to 4 (strongly disagree), please rate the extent to which you agree or disagree with the following.Scientists have not determined the best solution for disposing of nuclear waste.In the U.S., what do we do with it now?do we… a. Use is as nuclear fuel b.Sell it to other countries c. dump it in landfills d.Store and monitor the waste e. don't know Using a scale of 1 (strongly agree) to 4 (strongly disagree), please rate the extent to which you agree or disagree with the following.