Forested Areas and a Sense of Place: Comparing the Perceptions of Compact and Metropolitan Urban Populations


Places within our perceived environment, such as urban areas and city parks and forested tracts, are socially constructed. In order to advance environmental policies in the public sphere (e.g., principles of ecosystem management and forest preservation), resource managers must learn to craft discourse that persuades and mobilizes their intended audience (cf., Kraft & Wuertz, 1996). Such target constituencies, often represented by interest groups and politicians, define the perception of problems in the political arena. And these perceptions are generally driven by local conditions and the activities taking place within or adjacent to such environments (De Haven-Smith, 1988). Consequently, it becomes necessary for resource managers to use citizens' perceptions in developing communication strategies that actually address the core values of a population.

One viable approach to isolating the core values related to forested environments is to examine the discourse people use to describe the places in which they live, work, and play. A significant line of research and theory regarding the role of perception in identifying values for forested spaces has been associated with the idea of a "sense of place." Those interested in the construct seem to agree that a sense of place is the perception of what is most salient in a specific location, which may be reflected in value preferences or how that specific place figures in discourse. Although previous studies of environmental perception have shown that persons' sense of place are potential mediators of environmental policy preferences, the relationship between a person's sense of place, their view of themselves as related to the natural environment, and the role played by forests in forming the landscape of the mind remains something of a mystery to social scientists and resource managers alike.

In 1998, staff at the USDA Forest Service's North Central Research Station (NCRS) in Evanston (IL) provided a grant to Dr. James Cantrill of Northern Michigan University to interview of a larger representative portion of respondents to determine the extent to which forested environments are salient in persons' senses of place, how those perceptions may be associated with various activities that help define a sense of self in the environment, and if there are differences in the perception of self and place for those who reside in compact (i.e., at a distance from a large city) versus metropolitan urban settings. Upon consultation with the NCRS, the city of Palatine (IL) was chosen as one of the two research sites for this study. The city is located in the northwest suburbs of Chicago and is served by a major commuter train-line; easy access to Chicago, reasonable housing prices, and a favorable quality of life results in Palatine being one of the most desirable living communities in the metropolitan area (e.g., it was once recognized as a designated "Tree City" by a national organization). In contrast to the metropolitan character of Palatine, the second research site selected for this study could be described as a "compact" urban center. Marquette (MI) is the major city in the Upper Peninsula of Michigan and is a hub for regional activities and services. As a population center, the city provides a venue for a regional medical facility, Northern Michigan University, and a variety of light industry in addition to mining and forestry companies working the surrounding rural countryside.

Samples of citizens living in Palatine and Marquette were interviewed to gather data for this study. The procedures employed in choosing interview respondents generally replicated those used by Cantrill (1998). First, we sampled a youthful population (age 13 to 18) by contacting key informants in youth councils and educational institutions. Second, we sampled working class citizens and professionals by contacting key informants who were well acquainted with various strata in the community. Third, we sampled retired individuals, often relying on information gained from senior citizen centers and community outreach programs. Care was taken to insure that a representative sample of men and women in a variety of walks of life participated in the research. Approximately 274 individuals were contacted by trained interviewers, and 200 eventually consented to be interviewed.

Methods, Tools, and Data

The interview schedule used by Cantrill (1998) was modified for use in the communities of Marquette and Palatine. In addition to questions including the sex, occupation, length of residence in the research sites, descriptions of the boundaries of the area they called their "home," and age of the respondent, individuals were asked to describe (a) what it was like to live in the area now and in the past; (b) whether they planned on staying in the region; (c) what features contributed to their sense that the place was "special" in any way and the qualities that made them feel "connected" to the region in one way or the other; (d) what they believed would happen to the area in the future, as well as what they would prefer to occur; and (e) the types of outdoor recreational pursuits they enjoyed. Since we were interested in what features were salient regarding place depictions, respondents were not prompted to consider forested environments per se; the protocol probed for such perceptions only after the interviewees had the opportunity to generate free-association responses by asking respondents to describe (f) what they thought of area forests or parks in general and the role of trees in specific.

"Open-coding" procedures were adopted in the qualitative analysis of interview transcripts (Strauss & Corbin, 1990). This is a process of making comparisons between individuals' responses and within respondent's narratives or answers to questions. The method is quite flexible in that it permits line-by-line analysis of phrases or single words, examination sentences or paragraphs, or inspection of an entire interview to isolate and compare dominant themes and perceptions. In the current research, several stages of analytic coding were conducted. First, the primary author read through each transcript and set of field notes to inductively identify the dominant theme(s) embedded in each respondent's answers to each of the questions on the schedule. Another trained coder then used the open codings for each response to independently assess the inter-coder reliability of the initial themes found in 20% of the coded transcripts. Second, responses were divided into three categories reflecting respondents' (a) current sense of place, (b) their projected sense of place in the future and, (c) the specific role of forests in their sense of place (i.e., questions dealing with descriptions of local parks and forested areas and the role trees played in daily life). In turn, and based upon their general impressions growing out of the initial open coding process, two coders placed all themes found in each of the three sense of place categories into thematic clusters emphasizing a focus upon either social, economic, or environmental considerations. Third, "axial codings" (in which themes were compared across responses within question categories to develop a reduced coding scheme that could be applied across the entire data set) were calculated. Finally, in order to facilitate the comparison of qualitative and quantitative data, categories and characteristics of theme-embedded responses to questions within sense of place categories were combined to determine for each respondent (a) a general foci representing a person's predominant focus upon either social, economic, or environmental considerations (or combinations thereof), (b) a dominant themerepresenting the thematic category that best reflected trends in each of the sense of place divisions, and (c) the general valence of a respondent's overall impressions. At this stage, modal codings for foci, themes, and valences were used to assess general tendencies. All resulting codings were input into a spreadsheet for statistical analysis.

An additional feature of the present study involved the use of quantifiable survey data to compare individuals' senses of place with their sense of self in that place. In order to view the relationship between one's preference for certain activities and how one perceives the environmental backdrop for those activities, an "Activities Survey" was constructed which reflected the range of things people could do in the Marquette and Palatine regions. This survey largely consisted of a series of five-point Likert-type scales reflecting respondents' opinions of and current participation in a variety of activities, as well as their belief that they would engage in those activities if they lived somewhere else. Four topical dimensions cast into two divisions were employed in constructing the 24 distinct activities listed on the survey. The first division dealt with activitiesthat could be considered environmentally benign and the remainder were arguably harmful to the environment. The second division dealt withsettings subdivided between (a) built and (b) natural, largely forested environments; half of those settings were ones in which individuals participate in groups and half reflected more solitary contexts for activities. Overall, the survey was intended to represent a range of ways in which respondents might be connected to acting in various human and natural contexts, as well as how salient or valued activities and settings appeared to be at present and in the future.

Previous research (i.e., Cantrill & Chimovitz, 1993; Cantrill & Masluk, 1996) has established that inductively generated themes characterizing a person's depictions of the environment can be compared with survey data through the use of both parametric and nonparametric statistics to more clearly identify differences between populations and subgroups. In the context of the current study, the ability to match quantitative with qualitative data allowed us the opportunity to discern which attributes of a sense of place and self, possibly influenced by various demographic factors, mediated reactions to forested environments.

Discussion of Results

At a very general level, and throughout the process of open and axial coding, it became apparent that unless specifically prompted to do so those interviewed rarely made reference to forests or trees per se. Indeed, most of our respondents focused on largely social aspects in describing the area they lived in. When trees were mentioned, it was typically in the context of social activities with family and friends or, occasionally, the loss of elm trees due to Dutch Elm disease. Even a cursory examination of the features mentioned by Palatine and Marquette respondents suggests that most features were associated with settings that highlighted human activity in the environment rather than something that existed independent of human use. Certainly, the Marquette sample referenced more "natural" features than did the Palatine group. Overall, however, in both locations features foregrounding trees or forests (or settings such as parks where trees might be a prominent aspect of the landscape) accounted for little more than 25% of all the features that were mentioned. In particular, one's affinity for environments or activities that include the presence of trees or forests does not seem to be an integral component in these individuals environmental selves. Finally, given the results of a regression analysis, we would be hard pressed to assume that location (as in compact vs. metropolitan cities) matters when it comes to whether or not forested areas seemed important to our respondents.

Among the major rationales for conducting the research reported here was a fundamental assumption that, in order to effectively advocate the wise use of resources related to forested environments, agencies must discover and take into account the self interest-driven perceptions of local populations regarding trees and forests. Although the basic reasoning remains intuitively appealing, the findings of this study suggest that the process of developing an effective discourse focusing on the relationship between local citizens and trees or forests may not be as simple as might be expected. The empirical results noted earlier point to a number of unanticipated conclusions. In either location, our respondents voluntarily referenced trees or forests only to a quite modest degree when describing their sense of place. Their current and projected senses of place seem bounded by social issues and into a socio-economic context that relegated forested environments to the role of a stage or background upon which they interact with only a human component of the local environment. In short, people do not seem to define themselves in relation to trees in either compact or metropolitan urban settings and it is not apparent that most of those activities depended on forests per se. Furthermore, insofar as we observed a marked preference for describing "home" in terms of social or economic dimensions in both sampling locations, we cannot conclude that forested environments such as parks (or trees in general) exert much of an influence on one's sense of place.

Despite the apparent lack of salience of forests and trees for our respondents, this research may nonetheless assist in the management of forest environments for urban populations. The research findings outlined in this report suggest, in particular, resource managers need to go beyond merely identifying the sense of place embraced by a local population. Consequently, while this study may not make the task of decision making in applied resource management contexts any easier, it does provide a warrant for further research and caution in advising the construction of advocacy campaigns. The findings also suggest that to better understand social and psychological aspects of landscape ecology and ecosystem management we may have to go back and rethink some bedrock assumptions about concepts such as sense of place and the environmental self. And, in conclusion, the drive for conceptual retrenchment suggested here may better serve the agency in building public policy consensus regarding land use and community relationships with existing forested areas (cf. May, 1994).


  • Those interested in reviewing the technical report for this USDA-funded study (including all statistical measures and comparisons) may contact the Principal Investigator: Jim Cantrill, Northern Michigan University. Information regarding the North Central Research Station may be obtained by accessing: Additional background material for this study is contained in the case study referencesincluded below.

Books and Publications

  • Cantrill, J. G. (1998). The environmental self and a sense of place: Communication foundations for regional ecosystem management. Journal of Applied Communication Research, 26, 301-318.

  • Cantrill, J. G., & Chimovitz (1993). Culture, communication, and schema for environmental issues: An initial exploration. Communication Research Reports, 10, 47-58.

  • Cantrill, J. G., & Masluk, M. D. (1996). Place and privilege as predictors of how the environment is described in discourse. Communication Reports, 9, 79-84.

  • De Haven Smith, L. (1988). Environmental belief systems: Public opinion on land use regulation in Florida. Environment and Behavior, 17, 176-199.

  • Kraft, M. E., & Wuertz, D. (1996). Environmental advocacy in the corridors of government. In J. G. Cantrill and C. Oravec, (Eds.), The symbolic Earth: discourse and our creation of the environment (pp. 95-122). Lexington: The University Press of Kentucky.

  • May, S. (1994). Superior parks: Communications plan for protected areas in the Lake Superior basin. Unpublished manuscript, Parks Canada.