A COMMUNITY LANDSCAPE MODEL OF PRO-ENVIRONMENTAL BEHAVIOR …

The Pennsylvania State University

The Graduate School

College of Arts & Architecture

A COMMUNITY LANDSCAPE MODEL OF PRO-ENVIRONMENTAL BEHAVIOR:

THE EFFECTS OF LANDSCAPE AND COMMUNITY INTERACTION

ON RESIDENTIAL ENERGY BEHAVIORS IN TWO PENNSYLVANIA TOWNS.

A Dissertation in

Architecture &Human Dimensions of Natural Resources and the Environment

by

Stephen P. Mainzer

© 2017 Stephen P. Mainzer

Submitted in Partial Fulfillmentof the Requirements

for the Degree of

Doctor of Philosophy

May 2017

ii

The dissertation of Stephen P. Mainzer was reviewed and approved* by the following:

Charles Andrew ColeAssociate Professor of Landscape Architecture and EcologyChair of CommitteeDissertation Co-adviser

A.E. LuloffProfessor Emeritus of Rural SociologyDissertation Co-adviser

Mallike BoseAssociate Professor of Landscape Architecture

James FinleyProfessor of Forest ResourcesIbberson Chair of Forest Resource Management

Ute PoerschkeAssociate Professor of ArchitectureDirector of Graduate Studies

*Signatures are on file in the Graduate School

iii

AbstractWe are using more energy every year. Between 2001 and 2011, Pennsylvania residential electricity

sales increased by two and a half times the number of new customers, accounting for almost one

third of the state’s total electricity consumption. Our ability to meet demand by acquiring new energy

sources faces several challenges. Confusion surrounds the physical and economic accessibility of

remaining fossil fuel sources. Immense land use requirements and subsequent environmental impacts

challenge a total shift to renewable energy sources. The laws of thermodynamics limit the potential

for new technology to efficiently convert raw energy to consumable sources. As a result, any rational

strategy to meet future energy demands must involve conservation.

Conservation is a pro-environmental behavior, an act intended to benefit the environment

surrounding a person. I posit that a transdisciplinary model, the community landscape model of the

pro-environmental behavior, unifies the conceptually analogous - yet disparate - fields of landscape,

community, and behavior towards explaining residential energy conservation actions. Specifically,

the study attempted to describe links between the physical environment, social environment, and

conservation behaviors through a mixed-method framework. Two Pennsylvania townships – Spring

and East Buffalo townships – were selected from an analysis of housing, electricity consumption, and

land cover trends. Key informants from both townships informed the design of a survey instrument

that captured the utility consumption, residential conservation actions, energy and environmental

values, types and levels of community engagement, perceived barriers, and socio-demographic

information from 107 randomly selected households.

A mixed-method analysis produced evidence that place-based values and intention to participate in

the community were significantly linked to lower utility consumption in households. People who cared

deeply about their town were both more likely to attend community events and use less energy in their

home. Other less significant examples of influences from the physical and social environments are

presented in chapters 4 and 5.

iv

Table of Contents

List of Figures ...............................................................................................................................viList of Tables .................................................................................................................................viiPreface ...........................................................................................................................................viiiAcknowledgements ....................................................................................................................... ix

Chapter 1: Problem Statement ......................................................................................................11.1 Problem Description .........................................................................................................11.2 Background .......................................................................................................................11.3 Purpose and Research Questions .....................................................................................31.4 Significance .......................................................................................................................4

Chapter 2: Literature Review ........................................................................................................52.1 Introduction .......................................................................................................................52.2 Informing environmental problems through field analysis: toward a

community landscape theory of pro-environmental behavior .........................................5

Chapter 3: Methodology ...............................................................................................................223.1 Research Design ................................................................................................................223.2 Exploratory Site Analysis .................................................................................................233.3 Key Informant Interviews .................................................................................................303.4 Key Informant Data Analysis ............................................................................................323.5 Survey Design & Implementation ....................................................................................323.6 Spatial Analysis .................................................................................................................373.7 Survey Data Refinement & Analysis ................................................................................38

Chapter 4: Results .........................................................................................................................414.1 Summary of key informant interviews .............................................................................414.2 Summary of spatial data ....................................................................................................494.3 Survey and spatial data results ..........................................................................................49

Chapter 5: Discussion ...................................................................................................................735.1 Review of findings ............................................................................................................735.2 Validity of the model framework ......................................................................................735.3 Role of controls and sociodemographic factors ................................................................785.3 Role of conflict ..................................................................................................................795.5 Validity and reliability .......................................................................................................80

v

Chapter 6: Summary & Future Questions .....................................................................................846.1 Summary of study framework and results ........................................................................846.2 Implications of findings ....................................................................................................856.3 Implications for the field ...................................................................................................866.4 Future questions ................................................................................................................87

Bibliography .................................................................................................................................89

ApendiciesAppendix A: Key Informant Guide .........................................................................................95Appendix B: Survey Postcard .................................................................................................97Appendix C: Drop-off, Pick-up Survey Material....................................................................98Appendix D: Exploratory Data Visualization .........................................................................119Appendix E: Results of Correlation Analysis .........................................................................125

vi

List of Figures

Figure 1: Number of Acres Required to Produce Equivalent Energy ...........................................3Figure 2: A community landscape theory of pro-environmental behavior ...................................16Figure 3: Study framework ...........................................................................................................23Figure 4: Location of qualifying MCDs by quadrant ...................................................................28Figure 5: Location of Spring Township and East Buffalo Township ............................................29Figure 6: Example of individual home’s one-half mile land cover radius ...................................37Figure 7: Average residential electricity prices in Pennsylvania ..................................................77

vii

List of Tables

Table 3.1: Example site selection quadrants .................................................................................27Table 3.2: Comparison of study sites ............................................................................................29Table 3.3: Summary of key informants .........................................................................................30Table 4.1: Summary of key informant interviews ........................................................................41Table 4.2: Summary of key informant data for Spring Township.................................................42Table 4.3: Summary of key informant data for East Buffalo Township .......................................45Table 4.4: Summary of land cover data ........................................................................................49Table 4.5: Survey response rates by modality ..............................................................................50Table 4.6: Names and definitions of explanatory variables ..........................................................51Table 4.7: Factoring of energy values into a single component ....................................................55Table 4.8: Factoring of perceived barriers in two components .....................................................55Table 4.9: Descriptive statistics – energy behaviors (dependent variables) .................................55Table 4.10: Descriptive statistics – physical environment variable ..............................................57Table 4.11: Descriptive statistics – control variables ....................................................................58Table 4.12: Results of Shapiro-Wilk Tests for Normality .............................................................59Table 4.13: Description of skewness and kurtosis of energy behaviors .......................................59Table 4.14: Test for difference among variables ...........................................................................60Table 4.15: Summary of correlation between energy behaviors and variables ............................63Table 4.16: Priority rankings .........................................................................................................64Table 4.17: Average levels of community engagement ................................................................64Table 4.18: Comparison of variables that significantly correlate to dependent variable EBA .....66Table 4.19: Comparison of variables that significantly correlate to dependent variable EBJ ......66Table 4.20: Comparison of multivariate regression models ‘Past conservation actions’ ..............67Table 4.21: Comparison of multivariate regression models ‘Utility consumption’ ......................68Table 4.22: Summary of differences between townships among key variables ...........................70Table 4.23: Socio-demographic summary of sample ....................................................................71

Table A.1: Correlation between energy behaviors and environmental values ..............................125Table A.2: Correlation between energy behaviors and land cover ...............................................125Table A.3: Correlation between energy behaviors and types of social engagement .....................126Table A.4: Correlation between energy behaviors and types of community engagement ............126Table A.5: Correlation between energy behaviors and perceived barriers ...................................126Table A.6: Correlation between energy behaviors and housing typology factors ........................127Table A.7: Correlation between energy behaviors and socio-demographic factors ......................127

viii

Preface

This dissertation is an original intellectual product of the author, S. Mainzer. The human subjects

research reported in Chapter 4 was determined by the Pennsylvania State University Office of

Research Protections (STUDY00003702) to be exempt from formal Institutional Review Board (IRB)

review.

Chapter 2. A version of this material is currently under review for publication as Mainzer, Stephen

and A.E. Luloff (2017). Mainzer is the primary author and was responsible for all intellectual content

and research unless otherwise cited.

ix

Acknowledgements

My dissertation simply would not have been possible without the love and unconditional support

of my family. Kendall’s infinite belief in me superseded any questions about returning to graduate

school and galvanized my efforts to complete something that we could be proud of on schedule. This

work is for my children: Sebastian, whose inquisitive intellect will eventually reveal to him that his

future world is the product of our efforts today, and Maverick, who I believe will be fearless. Though

they never asked, I was home for dinner every night.

This transdisciplinary study greatly benefited from a collection of intellects that bridged parallel

fields. In addition to their immense knowledge, support, and trust, I would like to sincerely extend my

gratitude to the following for their unique contributions:

Andy Cole (ecologist), who created a space where my work and goals were respected, valued, and

supported, while serving as model for navigating academia as an introvert;

A.E. Luloff (rural sociologist), who imparted the value of having a backbone, pursuing work that

you care about, collaborating with others, and his immense understanding of people;

Mallika Bose (architecture and urban planning) and James Finley (forestry) whose critical thoughts

drove me to explore threads I would have otherwise neglected;

Stephanie Johnson-Zawadzki (social psychologist), who lit the inspirational catalyst for this study

during raging dialogues about the confounded state of environmental psychology and irrational

behavior, and for her willingness (and patience) in deciphering statistical queries using small

words; and

Sarah Eissler (rural sociologist), who stepped up and volunteered her valuable time towards data

collection at a critical junction in the study during a rare week where she was not off exploring

corners of the globe.

This dissertation was supported by grants from the Department of Landscape Architecture, the

Ibberson Chair for Forest Management, and the Stuckeman Center for Design Computing.

1

Chapter 1: Problem Statement

1.1 Problem Description

Exponential population growth in a world of finite natural resources necessitates conservation.

Energy is the prime resource essential for all basic life requirements including subsistence, warmth,

and mobility. Energy cannot be created; inefficient processes currently may only transfer it. Recent

data points to an unsustainable supply of fossil fuels and serious challenges for replacement by

renewable sources at current levels of consumption. In the absence of an emerging energy transition

strategy, conservation is a rational strategy. Behavioral modeling is a deep and established field of

theory that is inherently applicable to conservation. Pro-environmental behaviors minimize impacts

on the natural environment. Pro-environmental frameworks vary across the literature, but generally

consist of an interactive field consisting of an individual’s attitudes and beliefs, social norms, and an

external context. The interactional theory of community is conceptually similar to PE frameworks.

Community is imagined as an interactive field of locality, local society, and locally-oriented actions.

Landscape is a social product of community-level interaction that emerges when societies apply

their cultural values upon a locality. Logically, an exploration of the interaction between behavioral

modeling, community, and landscape suggest creative conservation strategies.

1.2 Background

Energy is the base resource necessary for human life, though its efficiency and available supply

is surprisingly limited. Natural resources provide energy essential for basic life requirements, such as

heating and mobility (Wackernagel & Rees 1996, Wackernagel et al. 1996) and production of goods

and services (farms, roads, consumable fuel, etc.) (Gever et al. 1986). Limitations imposed by the

Second Law of Thermodynamics constrain our ability to produce and efficiently consume energy. The

law states that 1) energy cannot be created or annihilated, and 2) because the universe is constantly

moving towards disorder, or entropy, energy is required to perform all work (Gever et al. 1986). It is

not possible to create more energy. The latter can only be converted from a naturally existing form to

a form useful for work. Most energy consumed (approximately two-thirds) during this process is lost

as heat. The remaining one-third of energy is useable for work, though it too is eventually converted to

2

heat (Pimental et al. 2010, Wackernagel & Rees 1996).

To put the challenge of population growth and energy use into perspective, Pimental et al. (2010)

suggest that a sustainable energy carry capacity for the United States would require approximately

29 quads (1 quad = 1015 BTU = 1,055 x 1018 joules) of energy. Gever et al. (1986) define carrying

capacity as the number of individuals that a given area can support indefinitely. At this hypothetical

carrying capacity, the United States would need to massively decrease consumption by almost 75% of

current levels and dedicate more than 20% of its total land to renewable energy production.

Energy consumption at the household level makes up a significant portion of total energy use

in Pennsylvania. Residential customers are purchasing more energy every year. In 2001, private

residential sales (in megawatt hours (MwH) accounted for 33% of total sales (28% in 2011). During

this ten-year period, the annual average increase in residential energy sales was 2.45%, while the

number of residential consumers grew at a rate of 1.01% (Energy Information Administration 2015).

Given the data, energy conservation at the household level is a critical and important step towards

sustainable demand.

Renewable energy has been largely hailed as the future solution to energy consumption. However,

there are several challenges associated with renewables. First, sustainable renewable resources must

be used at a rate lower or equal to their rate of regeneration (Daly 1990). Meadow et al.’s (2004)

World3 model illustrates that all renewable resources are currently being used up faster than they can

regenerate. More disconcerting is the amount of land required to produce renewable energy. Compared

to fossil fuels, renewable sources require significantly more land. The Nature Conservancy estimates

that the U.S. will require 67 million acres in 25 years to meet electricity and fuel demands assuming a

low percentage of renewable sources. A total shift to renewable sources could expand that estimate to

150 million acres1 (Outka 2011, McDonald et al. 2009). The potential mix of renewable sources and

their respective land requirements necessary to meet such demand is not uniform. Favored renewable

biofuels require massive amounts of land and while wind, hydro, and solar require significantly less

– though each necessitates more land than natural gas, coal, and nuclear (McDonald et al. 2009 and

Figure 1). This figure represents a small portion of total US land acres (approximately 6.5%) but

would consume a significant portion (approximately 23%) of US federally owned lands. The land

1 Presumably,thisfigureassumesasignificantportionoftherenewableresourcemixconsistsofbiofuels.SeeMcDonaldetal.(2009)foradescriptionoftheland-useintensityimpactsofrenewableandfossilfuels.

3

requirements increase to 36% if Alaska is discounted (Gorte et al. 2012). Much of the federally-owned

lands contain valued natural resources such as forests, national parks, and wildlife.

Community and landscape theory presents an opportunity to understand the constructs that

influence pro-environmental behaviors. Behavior is a function of the internal values and norms

of an individual and the influences of their environment (Ajzen 1991, Lewin 1946). Community

incorporates three critical elements: locality, local society, and locally-oriented actions in an interactive

field (Luloff 1998, Wilkinson 1991). Community is not a place, but it is created through collective

place-based interests and actions (Bender 1978, Luloff 1998). Community emerges when people care

about each other, a place, and act (Luloff 1998). Community development encourages agency in the

form of collective actions towards a common interest (Luloff 1998, Flint et al.. 2008). Landscape is

created through culturally-driven actions performed by local societies (Jackson 1997, 1994, Jacobs

1992, Lewis 1979). The physical qualities of a place are clues to the values and meaning that people

have attached to the land (Greider & Garkovich 1994).

1.3 Purpose and Research Questions

The purpose of this study is to demonstrate the validity of a unified model of community,

landscape, and pro-environmental behavior in describing the motivations of energy conservation

Figure 1: Number of Acres Required to Produce Equivalent Energy (Outka 2011 and McDonald et al. 2009)

140,000

120,000

100,000

80,000

60,000

40,000

20,000

0

Biom

ass

Win

d

Hydr

o

Petro

lem

Sola

r

Natu

ral G

as

Sola

r The

rmal

Coal

Geot

herm

al

Nucl

ear

EnergySourceType

Numbe

rofA

cres

4

actions. Towards that goal, the study proposes two research questions:

R1: Does the landscape influence residential energy conservation behavior?

R2: Does community engagement influence residential energy conservation behavior?

Further, the study attempts to decipher what types and levels of each possible motivator are relevant to

energy conservation behavior.

1.4 Significance

The basic information provided here should impart a sense of alarm. Clearly, action is needed. The

literature suggests several ways forward. First is a fairly unrealistic reduction in population growth that

would serve only to delay the inevitable surging demand for resources (Malthus 1798). Second is faith

in the technological acumen of the resource optimists to manage the laws of thermodynamics in such

a way as to maximize our use of solar resources and minimize future ecological land use impacts from

energy sprawl. Third, we could embark on an immense conservation commitment towards a balanced

relationship between energy consumption and regeneration or discovery levels. Of these options, this

study focuses on a practical and critically important approach to household energy conservation. This

study will provide new knowledge to the fields of architecture, landscape architecture, community

studies, rural sociology, and the general field of natural resources by identifying new, critical factors

to consider in the design of the physical landscape. Further the study will demonstrate the validity of

unified behavior, community, and landscape models in natural resources, conservation, and design

research.

5

Chapter 2: Literature Review

2.1 Introduction

The following text is an excerpt from a journal publication (Mainzer and Luloff 2017) that

reviews the relevant literature from behavioral psychology, landscape, and community theory.

The paper presents each discipline as an interactive field bonded by common elements that shape

actions and places. I suggest the unification of these fields provides a comprehensive framework for

describing social-environmental phenomena, what I refer to as a community landscape theory of pro-

environmental behavior. This paper seeks to accomplish three things: (1) to construct a rationale for

unifying commonalities among field-based concepts of behavior, landscape, and community; (2) to

describe a transdisciplinary field-theory based model of pro-environmental behavior; and (3) to present

a hypothetical application.

2.2 Informing environmental problems through field analysis: toward a community landscape

theory of pro-environmental behavior 2

Community has been defined as is a network of associations which emerges to solve problems

(Kaufman 1959). The interactive physical and social landscape of a community inform pro-

environmental behavioral intention, a key component to identifying solutions to local environmental

problems. This paper seeks to accomplish three things: (1) to construct a rationale for unifying

commonalities among field-based concepts of behavior, landscape, and community; (2) to describe

a transdisciplinary field-theory based model of pro-environmental behavior; and (3) to present a

hypothetical application. Outlining the convergence of these constructs formulates a framework that

may assist local problem solving through a holistic understanding of social, cultural, and biophysical

interactions.

How we treat the landscape, and the broader environment in which it is embedded, is a critical

element of interactional field theory. To date, relatively little work has been done applying interactional

field theory in non-traditional areas of study. That is, most attention has been placed on its role in

community development – particularly in rural and small areas – and in substantive domains such as

2 Mainzer,StephenandA.E.Luloff.2017.Informingenvironmentalproblemsthroughfieldanalysis:Towardacommunitylandscapetheoryofpro-environmentaltheory(currentlyinrevision).(Sectionnumberingaddedforthisdocument).

6

natural resource and environmental work (Larson et al. 2015). Its application in other well-suited areas

has largely been fugitive. We build upon recent work by Larson et al.’s (2015) to broad traditional

application of interactional community theory. Here, we anchor its relevance and utility to landscape

theory. We begin by defining community and landscape as multi-level social constructs.

Definitions of community in the literature are often vague or confounded (Flint et al. 2008, Luloff

et. al 2004). Indeed, the complex nature of community crossing levels and units of analysis have

introduced doubts to its theoretical application (Manfredo et al. 2004). However, it is precisely this

multi-level nature that infuses community development with the potential for real and sustainable

environmental action (Larson et al. 2015, Bridger & Luloff 1999). The interactional theory of

community (see Kaufman 1959, Wilkinson 1991, Theodori 2000, 2008, Flint et al. 2008, Larson et al.

2015) potentially improves the application of pro-environmental behavior models by describing the

interactive relationship among people, place, and actions.

Long-term environmental health is vulnerable to human actions (Bridger and Luloff 1999).

However, human action may also benefit the environment. The multi-level nature of community

involving individuals, societies, and ecologies potentially contributes to well-being at each level

(Larson et al. 2015). McHarg’s seminal work, Design with Nature (1969) is an enduring primer for

how humans may choose to act towards changing the landscape in such a way as to be suitable for a

given biological, physical, and social context. McHarg acknowledged that the multitude of landscape

dimensions present, consisting of physical and social characteristics, in any place are defined by the

social values of those who inhabit, work, or transition through that landscape (McHarg & Steiner

1998, McHarg 1969). Employing this methodology necessarily incorporates social values into

landscape design and implementation. Such actions require immense community support to realize

suggesting that changes in the landscape reflect an alignment of social values at the individual and

society level sufficient to affect individual, social, and ecological well-being.

We suggest that this process, which has taken place throughout human history, may be

deconstructed into three conceptual constructs: landscape, the community field, and the behavioral

field. An understanding of each, and more importantly, their interrelationships, vitally outlines a path

towards informing environmental problems.

7

2.2.1 Rationale for a placed-based model

Our efforts to better understand environmental actions at the community level differ from other

more well-known attempts to do the same. The Community Capital Framework (CCF) and Social-

ecological systems (SES) approaches have been exhaustively described elsewhere and it is beyond the

scope of this paper to fully review either framework. Rather, our aim is to identify the circumstances

under which a community level environmental problem may benefit from a field-based approach.

CCF attempts to quantify a comprehensive collection of interrelated resources (natural, cultural,

human, social, political, financial, and built capital) towards the goal of informing community

development strategies (Fey, Bregendahl, & Flora 2006). The underlying premise is that each capital

stock is related to the others and careful management of each yields community development. Emery

& Flora (2006) posit that community improvement begins with improved social networking which

improves other capital stocks. While this approach does well to acknowledges the critical role the

individual actors play in community development, it also distills community development to a linear

process of accumulating economic wealth. We argue that interactions at the community level are part

of a continual and emergent process (Kaufman 1959, Wilkinson 1991, Theodori 2000, 2008, Flint

et al. 2008, Larson et al. 2015) whereby changes in the environment, including each of the capital

stocks accounted for by CCF, influence subsequent actions intended to benefit the place within which

a community is rooted. Community development emerges through relationships between people and

place that are continually adapting to the physical and social environment.

SES frameworks attempt to embrace socio-ecological complexity by describing the multiple

temporal and spatial scales of a socio-ecological environment (Ostrom 2009). Such systems are in

constant transition seeking one of a multitude of potential stable states. The goal of SES approaches

is to encourage resilience by describing the function, structure, and feedback loops of social and

ecological systems (Stedman 2016). However, Stedman recognizes a potential shortcoming of recent

SES efforts to desegregate humans from natural systems. By minimizing the inherent subjectivity of

perception, interpretation, and actions in human actors, SES systems fail to accommodate inherent

variation in communities and places. Others also support the importance of agency (Wahl 2014) and

subjective indicators such as social norms, beliefs, and culture in SES frameworks (Jones & Tanner

2015).

8

CCF and SES lack indicators of human variation as it relates to place. Stedmen (2016) offers sense

of place methodology as a means of incorporating human variation into research. Places emerge where

human activities, social, and psychological processes are rooted in a physical setting and are strongly

related to the formation of beliefs, attitudes, and identities (Stedman 2002). Sense of place refers to

the meanings and attachments of a place formed by individuals and groups (Tuan 1977, Trentelman

2011, & Kudryavtsev, Stedman, & Krasny 2012). Meanings describe the symbolic importance

assigned to places, such as cultural values (Kudryavtsev, Stedman, & Krasny 2012). Attachments

refer to the emotional bond between people and their environment (Stedman 2003, Trentelman

2011, & Kudryavtsev, Stedman, & Krasny 2012). Drawing generalities about a place is dangerous

as meaning and attachment dynamics vary from place to place and may be multidimensional within

a place (Trentelman 2011). Communities and environments are highly dynamic and differentiated

(Leach, Mearns, & Scoones 1999). Inherent subjectivity enables environmental actions to be

perceived by community members as good for one group and harmful by another (Stedman 2016).

Stedman (2002) further argues that sense of place is derived from the strength of social relationships,

which are influenced by length of stay and social mobility. The influential power of the physical

setting compared to the socio-symbolic setting is debatable (see Tresselman 2009). Yet, the physical

environment influences place formation in indirect ways (Stedman 2003, Jackson 1994). Place

attachment is grounded in natural amenities (Kemmis 1990, Wilkinson 1991), environmental features

(Eisenhauer, Krannich, & Blahna 2000), and physical qualities that drive cultural development and

place meanings (Stedman 2003).

An interactional theory of community (see Kaufman 1959, Wilkinson 1991, Theodori 2000,

2008, Flint et al. 2008, Larson et al. 2015) incorporates the dynamic process of social relationships

through its emergent nature while grounding nebulous interactions in a place. Community is not a

scale of place. It is a setting for locally based social interactions (Trentleman 2009). When applied as

a mediating structure between landscape and behavior, the resulting framework encompasses much

of what CCF and SES models attempt while maintaining flexibility for human qualities. We do not

suggest that CCF and SES models are ineffective – each has a critical role in an appropriate scenario.

Stedman (2016) calls for researchers to further investigate how people emphasize different elements of

a system and offers sense of place as a possible way forward. We are attempting to answer that call by

9

proposing a simple model that accommodates the dynamic social and physical qualities of place at the

community level that influence pro-environmental behavior. It is unclear how place dynamics interact

with other factors that influence pro-environmental behavior (Kudryavtsev, Stedman, & Krasny

2012). We posit that a model based on the socio-physical landscape at the community level has the

potential to illuminate relationships between place and behavior. The following offers a primer on key

landscape, community, and behavior concepts with an emphasis on their isomorphic commonalities.

2.2.2 Landscape as a socially constructed place

The word environment encompasses the physical and social context as it relates to an individual.

Landscapes emerge from places when cultural meanings, values, and beliefs are applied to the physical

setting. The human landscape is an autobiography of our day-to-day qualities (Lewis 1979). Broadly,

landscapes demonstrate the cultural evolution of local people. The physical landscape represents the

cultural values of its inhabitants through their vernacular actions (Lewis 1979). Jackson (1994, 1997)

and Jacobs (1992) famously constructed literary narratives of how landscapes change over time in

response to internal, or externally imposed, changes in cultural values. Natural environments, and

by extension, natural resources, have no inherent meaning. Landscapes are a “symbolic environment

created by a human act of conferring meaning on nature an environment” (Greider & Garkovich

1994, 1). Landscapes reflect cultural values and beliefs that emerge from social interaction between

people in a place. When multiple overlapping cultures are present, multiple landscapes will reflect the

complexity of those local cultures (Greider & Garkovich 1994).

Landscape perception is a product of interactions between human experiences, knowledge,

sociocultural context and its individual components, and holistic landscape (Zube et al. 1982). In

short, perception is formulated from the relationships between people and their socially oriented

environment. The study of landscape spans many fields (Groth 1997). It is not surprising then that

the basic theoretical structure of landscape perception is found in areas, such as other behavior and

community fields described in this paper (see Lewin 1935, Wilkinson 1970, Kaufman 1959, & Luloff

1998). Each locates sociocultural influences in a dynamic relationship with the environment as it is

observed through the actions of local cultures. Lewis refers to the cultural landscape as “everything

we can see when we go outdoors” (1979, 1). He speaks to the observation of human values in the

landscape through seven axioms for reading the landscape. While each is valuable in linking people

10

to the landscape, a few key points synthesized from these axioms and their corollaries are particularly

relevant to the physical-social landscape field. First and most importantly, all items in the landscape

reflect culture and each is a clue as to the historic, present, and future conditions of those cultures.

Landscapes that appear to be similar despite being in geographically separate areas are likely to reflect

similar cultures. The same inference can be made for different landscapes in separate areas (Lewis

1979). It is possible to read collective landscape-related actions, or inactions, of a local society by

observing what has been constructed or preserved. Second, society is resistant to change. Landscapes

that have changed necessarily reflect a corresponding change in culture (Lewis 1979). Change

provides an opportunity to observe the results of social action, or to identify a shift in community

interests, individual values, or in the perceived barriers associated with either. Third, geographic and

ecological context is important in studying landscapes (Lewis 1979).

McHarg (1969) argued that both ecological context and the social perception of such qualities

was critical to solving natural resource problems. He and his colleagues sought to outline methods

of landscape design that incorporate the underlying interactive biophysical culture of landscapes

with local social values to implement suitable changes in the landscape (McHarg & Steiner 1998).

These endeavors represent a practical application of field theory in real world scenarios. McHarg’s

methodology has endured and evolved in modern day applications through the coupled use of

geographic information systems and local human knowledge to produce sustainable landscape designs

(Steinitz 2013, Johnson & Hill 2002).

2.2.3 Interactive theory of community

Current landscape design theory has made great strides by incorporating multiple dimensions of a

dynamic environment. However, it lacks a central component that provides an understanding of how

value transitions across scales from the individual to society and into action. An interactional theory

of community (Kaufman 1959, Wilkinson 1991, Theodori 2000, 2008, Flint et al. 2008, Larson et

al. 2015) synergizes pro-environmental behavior and landscape perception theories by describing

the emergent process of placed-based community action. Community grounds behavioral theory in

a locality whose physical and social qualities are revealed through and implemented by landscape

perceptions.

11

Defining community is a key step in unraveling its role in a behavior-landscape field. Flint, Luloff,

and Finley (2008) suggest that an overuse of poorly defined community concepts has resulted in

ungrounded science. A most common, yet often inaccurate, historic conception of community is based

on the stereotype of a small New England town (Bender 1978). The small-town ideal of community

is formulated on the premise of community emerging from a bounded, socially homogeneous area

of shared norms. It was assumed that reliance on natural resources would optimally generate small

spatial units with similar demographic patterns and that people of similar demographics would have

analogous customs. Yet such definitions (small spatial unit, homogeneous social structure, shared

norms) fail to recognize internal diversity (Agrawal & Gibson 1999). Such simplistic assumptions

often overlooked the actual diversity of values within a place – communities are typically not

homogeneous groups of like-minded people (Flint et al. 2008). Further, Bender indicates community is

not defined solely by geography – experience is also critical (1978, 6).

Community structure has evolved from Lewin and other disciplines’ social field theory (Wilkinson

1970). An interactional theory of community, developed and championed by such authors as Kaufman,

Wilkinson, and Luloff, presents a locally oriented definition that connects people to place through

social interaction. Community, in this sense, is composed of three elements: a locality, a local society,

and locally oriented social actions forming a dynamic and emergent field (Kaufman 1959, Luloff 1998,

Wilkinson 1991 Flint et al. 2008). It is no accident that this definition utilizes the word local in each of

its components. Community is inherently place-oriented, though it is not found in place alone (Luloff

1998). Each of the three components of the interactional theory of community is described below.

2.2.3.1 Local Society

Community emerges from social interaction (Flint et al. 2008). Bender posits that the modern

definition of community is “an aggregate of people who share a common interest in a particular

locality” (1978, 5). This idea is born from a scale of social units ranging from the family to a small

city, and includes intermediate groups, neighborhoods, towns, and classes (Bender 1978). Such units

reflect both a functional and natural need for interaction, i.e., an elemental bonding between people

(Bender 1978). Tönnies’ Gemeinschaft and Gesellschaft typological constructs are fundamental

in describing the structure of two different types of elemental bonds. Neither concept translates

exactly into English, though each invokes powerful imagery of different types of social bonding.

12

Gemeinschaft encompasses traditional, family- or kin-oriented personal relationships traditionally

linked to rural society. Gesellschaft evolved from urbanization and industrialization to embody

a capitalist framework of modern, impersonal, competitive relationships (Tönnies 1957, Bender

1978). While Gemeinschaft is linked to community ideals and personal needs for interaction,

Gesellschaft arises to satisfy utilitarian societal needs through impersonal connections, though no less

interdependent. In this sense, personal community networks are nested within a greater society of

utilitarian functionality. Notably, the comparison of Gemeinschaft to Gesellschaft typology should not

be thought of as a strict dichotomy (Bender 1978). The Gemeinschaft and Gesellschaft likely exists in

most places as a coexisting sliding scale of overlapping structures.

2.2.3.2 Locality

The locality is intrinsically tied to the emergence of community, though a place is not a community

(Luloff 1998). People who live together interact in place-based ways and community emerges when

interactions give way to place-based actions (Flint et al. 2008, Wilkinson 1999). The locality is a

territory where people live and meet their daily needs (Hawley 1950, Wilkinson 1999), including

residential location, employment, and jurisdictional areas (Wilkinson 1999). Such processes are

critical from a perspective of ecological well-being. However, there are several challenges in over-

emphasizing the role of territory in community.

To be sure, community is grounded in territory, both by definition of requiring place-based

actions (Wilkinson 1999) and by the ecological processes inherent in gathering resources (Hawley

1950). Hawley (1950) defines community in terms of ecological, structural, spatial, and biotic terms.

The ecological community is a network of relationships that provide for daily needs. The structure

for achieving these needs is a symbiotic fabric held together by common interests, such as meeting

individual needs. Spatially, this fabric may extend as far as necessary to encompass the area in which

these needs are met. Biotic habitats are thusly formed in pursuit of a common interest (Hawley 1950).

These definitions seem to closely align with the locality, local society, and local action constructs from

the community literature (Kaufman 1959, Wilkinson 1991, Theodori 2000, 2008, Flint et al. 2008,

Larson et al. 2015). While Hawley’s ecological definition ostensibly contains the essential elements

of community, its ethos is fixated on purely utilitarian motivations. The purpose of these ecological

community interactions is to meet resource needs, reflecting an impersonal impetus (Bridger 1996).

13

Hawley’s communities lack a personal elemental bond between its participants.

Further, a reliance on geographic concepts risks an oversimplified view of the community.

First, as stated earlier, relying on spatial units falsely suggests homogeneity. Second, concern over

urbanization’s impact on community also speaks to the fallacy of a boundary-focused definition

of community. Advancements in transportation and technology contributed to increasingly wide

territory boundaries (Kaufman 1959). By this definition, whole cities could constitute communities.

This process would not necessarily erode communities, but rather extend a generalization of interests

throughout the population. Wilkinson suggests this could be a beneficial societal goal (1999).

However, arguments against such broad application of community indicate that at such a scale what

appear to be common interests could actually reflect hegemonic determinations made by the privileged

class (Wilkinson 1999). For these reasons a purely territorial definition of community is insufficient.

Rather, the locality must be defined in terms of the community’s interactions and actions.

2.2.3.3 Locally Oriented Actions

Since community actions are measured not by their success but by purposive positive effort, the

intention to act has immense value in community development (Luloff 1998). Individuals are members

of economic, social, cultural, and political groups and their actions are representative of the greater

society (Moran 2006). The interrelationship of local society, locality, and locally oriented actions

contribute to the construction of an interactive community field (Wilkinson 1999, Wilkinson 1970).

The community field does not, however, encompass all types of interactions (Kaufman 1959).

For example, both Gemeinschaft and Gesellschaft presume interactions, though only Gemeinschaft

infers the personal relationships necessary for community emergence and development. However,

involvement in a community field may be expressed along a range of actions (Kaufman 1959) that

speak to the Gemeinschaft-Gesellschaft typology. For example, on one end of the scale is residence

and sustenance (Gesellschaft) and on the other is contributing to major policy changes (Gemeinschaft).

Linking actions that identify or reinforce purposive and positive actions towards common interests

of the local society and place further develops the community. Community actions occur when

collective action is taken towards locality-based interests. Community agency forms when the capacity

for community action exists (Luloff 1998, Flint et al. 2008). It is not necessary that actions focus on

14

goal achievement, or that action be inspiring. In accord with TPB, efforts or intentions made toward

common interests are sufficient to strengthen the community field (Luloff 1998, Wilkinson 1999).

2.2.3 Behavioral field theory

Pro-environmental behavior is defined as a conscious effort to minimize the impacts of an

individual’s actions on the natural environment (Kollmuss & Agyeman 2002, Park & Ha 2011) or on

the “availability of materials or energy” in the environment (Stern 2000, 408). Lewin (1935) stated that

behavior was a function of person and their environment. A person is composed of her or his personal

values, beliefs, and attitudes with respect to a specific behavior, formulized as:

B = ʃ (P, E)

where B = behavior, P = the person, and E = the environment

(Lewin 1935).

A person’s environment is assembled from influential physical and social factors. This formula

may also be modified to argue that a person is the function of their behaviors in the environment, or

that the environment is the result of persons and their behaviors (Stokols 1977). These factors exist

in a dynamic and interactive field that allows researchers to isolate one factor and draw inferences

from the other two. For example, the ability to “read” social cues in constructed landscapes allows

a researcher to derive assumptions about personal values (Lewis 1979). Studying the dynamic

relationships within the interactive field of landscape, people, and behavior is critical to understanding

how each element influences the others. Further, behavior is analyzed in terms of the action taken.

People and environment are inseparable as a unit of analysis in research (Lazarus & Folkman 1987).

Ajzen’s Theory of Planned Behavior (TPB) advances Lewin’s field theory by further defining

the person and environment constructs. TPB is a predictive behavioral model where the individual’s

attitudes and/or beliefs interact with the perceived external context to form their intentions to act.

Attitudes encompass a wide range of internal influences, including ecocentric and/or anthropocentric

values, environmental concern, perceived consequences, and a sense of obligation (Ajzen 1991, Stern

2000, Park & Ha 2011).

TPB presents two key concepts; first, personal attitudes share a formative relationship with social

norms, thereby implying that individuals define some part of themselves in terms of their social

15

interactions. For example, community plays a role in the social definition of self (Wilkinson 1991).

Through the process of interacting with the environmental field, individuals construct or amend, in

part, their own values, which inform their behavioral intentions. Second, TPB proposes intention as a

precursor to behavior. However, intention may preclude behavior when barriers or other controlling

limits are in place, including landscape qualities. TPB provides a useful structure to organize a study of

personal and external values and a rationale for the link between values and behavior. It has previously

been applied in analyses of personal values and household energy use (Abrahamse & Steg 2011).

Behaviors are resistant to change. Ritual actions reaffirm compliance with social values and

confirm trust in the social fabric. Breaking these habits is immensely difficult and risks personal

feelings of instability. Further, in complex social systems values are difficult to aggregate, leading to

miscommunication and poorly informed habits (Moran 2006). Guagnano, Stern, and Dietz proposed

that behavioral change may be modeled by observing the influential strength of individual attitudes

and external context, referred to as ABC Theory (Guagnano, Stern, and Dietz 1995, Turaga 2010).

This theory presents a model of interactions between context and individuals. According to ABC

Theory, behavior is a function of internal (people) and external (society or place) factors and there

are predictable relationships within that interactive field (Guagnano et al. 1995, Stern 2000). When

attitudes (A) are favorable and action is easy to accomplish within a given context (C), then the

behavior (B) is likely to occur (Guagnano et al. 1995, Kollmuss & Agyeman 2002). Behavioral change

is a response to attitudinal or environmental forces existing within an interactive field that may be

sufficient to overcome ingrained behavioral habits.

2.2.4 A community landscape model of pro-environmental behavior

Lewin’s field theory (1935) and its evolved forms of TPB and ABC theory illustrate the effects of

the external environment and internal values on an individual’s behavior in an interactive and dynamic

structure. Behavior emerges through an active interplay between personal attitudes and environmental

forces. Landscapes are inherently social environmental constructs that are modified over time by

behaviors to reflect the cultural values of local communities. This process may occur slowly and

naturally over time or may be the result of explicit design methodology. It requires immense time,

energy, and political will (or a purposeful lack of all three) to make changes to the landscape,

especially in urban areas. If such a change represents a community level action, then it follows that

16

the local social and physical environment was conducive to such a change. An interactive theory of

community explicitly states that community emerges from a dynamic field consisting of a locality, a

local society, and locally-oriented action. Actions are the product of social interactions in a place and

actions strengthen the bonds of interaction between people in a place. Changes in the landscape serve

to strengthen community fields.

A review of the extant literature suggests clear commonalities between behavior, landscape,

and community fields. If community acts as a mediator, then the locality is analogous to the local

landscape and the local society to the social landscape where both landscape constructs combine

to form a general environment that affects behavior. Locally-oriented actions are simply behaviors,

influenced and affected by the physical and social environment. Behavioral intentions are informed by

an interactive multi-level community field and are moderated by perceived barriers to action. When

physical or social environmental barriers carry stronger influence than individual attitudes, behavior

is not likely. Figure 2 represents the conceptual overlapping of behavior, landscape, and community

fields as a basic model of relationships.

2.2.5 A hypothetical application

Significant social-ecological conflicts emerge when difference in values arise between an

individual and groups (Janssen, Lindahl, & Murphy 2015). How people perceive landscapes is derived

localenvironment

interactivefield

personalattitudes

socialenvironment

intentionsPE

behaviorperceivedbarriers

Figure 2: A community landscape theory of pro-environmental behavior

17

from a combination of phenomena and social interaction that define deeply valued meanings. As a

result, scientifically oriented descriptions of places may violently clash with human values (Greider

& Garkovich 1994). Here we present one example of a community-level conflict and a hypothetical

application of the community landscape model of pro-environmental behavior.

In “Design with Nature” McHarg (1969) describes the imperative role that dune ecology plays

in keeping coastal towns safe from storms through a comparison of intrinsic dune knowledge in the

Netherlands that is dangerously lacking in the United States. The coast line of New Jersey, U.S. is

highly urbanized with commercial and residential properties driven by aesthetically desirable access

to the ocean. In 1962, a storm along the east coast of the New Jersey caused $80 million of damage,

destroyed 2,400 homes, damaged 8,300 homes, and caused multiple fatalities and injuries (McHarg

1969). McHarg depicts a clear lesson: maintain a dune landscape, limit urban development to suitable

areas, or face the wrath of nature. Nearly 50 years later, on October 29, 2012 Hurricane Sandy

reminded the country of the same lesson, killing over 159 people, damaging over 650,000 homes, and

forcing the closure of thousands of businesses. Multiple federal organizations have since responded

or been created to aid recovery of the affected areas, including the Federal Emergency Management

Agency (FEMA), the Small Business Administration (SBA), National Disaster Recovery Framework

(NDRF), and the Hurricane Sandy Rebuilding Task Force. The SBA has committed $3.8 billion in

loans towards household and small business recovery. Congress approved $50 billion in support

through the Disaster Relief Appropriations Act (Hurricane Sandy Rebuilding Task Force 2013).

While the lesson briefly described here should be clear – there remains immense conflict at the

community level regarding how to best prepare local landscapes for the next 50-year storm event.

The federal government’s plan to construct dunes along a 50-mile stretch of the coastline, a $1

billion effort, was derailed for six years by 1,000s of oceanfront home owners refusing to grant the

government easements on their properties. Despite recent evidence from Hurricane Sandy that places

not sufficiently protected by dunes will suffer serious damage, holdouts remained firmly motivated

by a fear that townships would take advantage of the easements to construct undesirables on their

property – such as boardwalks or bathrooms. Bitter divides emerged throughout the shore. Neighborly

relationships were broken, shops began denying services to holdouts, and holdouts were publically

recognized and shamed in local newspapers. The severity of the storm motivated local courts to deny

18

financial compensation to residents in exchange for granting easements. The highest court in the state

granted permission for the government to take land as needed by eminent domain (Zernike 2013).

Nearly four years after Hurricane Sandy one of the towns that suffered the most damage, Ortley Beach,

celebrated the signing of the last required easement. Easements were negotiated by a combination

of exhaustive “backroom” work and eminent domain. The US Army Corps expects to begin work

in Spring 2017 and expect the project to be completed in 2018 – six years after the Hurricane Sandy

disaster. Ortley Beach spent $2 million attempting to reinforce its dunes while waiting for the last

holdouts to grant permissions (Wall 2016).

The New Jersey dune conflict illustrates a need to assess natural resource problems from the

perspectives of community, landscape, and behavior. The organization of the coastal towns’ values,

beliefs, social relationships, and how they inform changes in the landscape are central to the conflicts.

The debate has forced community members to express their values and beliefs with regards to public

and private property rights. Strong divides in the town lead to a fracturing of the township into

multiple communities. While each community shared a common place, divergent values separated

communities into conflicting groups of locals who choose whether or not to act to benefit their

common place. Electing to sign an easement was a pro-environmental action that was perceived by

residents to benefit the town in ways that were more than commensurate to the potential personal

losses of their individual properties. At a minimum, two conflicting communities existed within each

town – the first a group who supported signing easements and a second group who refused to sign.

Likely, multiple sub-communities emerged from a complex pool of values, attitudes, and beliefs

existed within each primary community.

The community landscape model of pro-environmental behavior could be applied to the New

Jersey scenario as a method of quickly deconstructing the problem into key characteristics that may

inform problem-solving strategies. Below we hypothesize how each element of the model could yield

critical information.

• Pro-environmental behavior – signing easement allowing federal government access to land for

recreating a dune landscape. The model is oriented around this action.

• Local environment - Where should dunes be constructed? Where are the most and least

important areas? Which areas have the least resistant path to construction? Such questions could

19

be simply addressed through non-reactive spatial analysis methods to address where conflict

may arise and how to proceed along a path of least to most resistance. In this hypothetical

scenario, some areas may be less resistant and in more immediate need of dune reconstruction

efforts than others. Holdouts in resistant areas may be interested to know exactly where (and

why) it is most critical to address dune landscapes.

• Social environment – What is the relationship of the holdout resident to the town? Which social

groups are they a part of that were affected by the storm? What are their most valuable social

relationships? Closely linked to the collective attitudes of social group, understanding where a

holdout lies in relation to others who are important to him or her informs their possible networks

of information and encouragement. Members of strongly bonded family, work, service, or other

types of groupings may find it difficult to diverge from the group. It is difficult to act upon

personal values when they are in opposition of strong social environment forces (Guagnano,

Stern, & Dietz, 1995).

• Personal attitudes – What are the individual’s personal values, attitudes, and beliefs regarding

the landscape? How much of a priority is their local environment compared to their personal

amenities? How confident are they with science predicting the next 50-year storm event

happening in their lifetime or within the lifetime of someone they share a strong emotional

bond? Understanding personal attitudes informs potential existing and future conflicts about

environmental issues within their communities.

• Intentions & Perceived Barriers: Do they want to sign an easement, but feel prohibited by

physical or social barriers? Individuals will act on their intentions unless prevented by perceived

barriers (Ajzen 1991). Identifying whether or not individuals have pro-environmental intentions

and if barriers exist may assist in developing strategies for removing barriers.

From this simple, yet flexible, deconstruction of the relevant mediating the action of signing

easements, we could develop specific strategies for engaging conflicts with environmental values,

social networks, and personal values at the community and individual level. We suggest that such an

approach would avoid the need for lengthy legal processes, political pressure, and public shaming.

However, we do not expect that conflict and resistance would be avoided altogether. Rather, coupling

social information with landscape data enables interventions to target highly resistant areas and

develop place-based strategies.

20

2.2.6 Conclusions

This brief review of the relevant field, behavioral, landscape, and community literature suggests

there are common isomorphies within each discipline. First, individual behavior is determined by

environmental and personal interactions. Second, landscape is a social construct of a local culture’s

values observed through the process of change and modification. Third, an interactional theory of

community describes the multi-level bonds and interactions that drive collective local actions. Such

synergies among disciplines, localities, and people suggest a unifying transdisciplinary model.

Pro-environmental changes to the landscapes, especially those entrenched in unsustainable

habitual behaviors, requires an understanding of the community-level forces driving current

behaviors sufficient to identify points of intervention – both physical and social. In a community,

social interaction serves to generalize individual values towards common interests (Wilkinson 1991).

Strengthening the community field implies that individual values are synthesized to a degree where

adequate cohesion is transformed into place-based behaviors through such actions as landscape

modification. That which strengthens or weakens the community similarly affects the likelihood of

behavior. If we are to make changes in the landscape to combat place-based environmental problems

such as climate change, energy consumption, or personal health attrition – then a community

landscape model of pro-environmental behaviors suggests we must incorporate a multi-level and

multi-disciplinary field analysis of the problem. This requires a transdisciplinary approach.

Transdisciplinary collaboration has been increasingly recognized as a necessary method for

addressing environmental problems, including; urban planning, public policy, environmental

degradation, oceanic and coastal management, sustainable development, and human-nature

interactions (Mitrany & Stokols 2005). Complex environmental problems suffer from a lack of

scientific clarity and disagreement of values by decision makers. Solving wicked problems requires

dialogue that transcends disciplinary boundaries to include the perspectives of multiple fields and

input from local stakeholders (Balint et al. 2011). Collaboration across traditional expert and layperson

confines generates new knowledge, processes, and tools that suggest locally suitable solutions to

problems.

Transdisciplinary approaches differ from interdisciplinary approaches, which favor narrowing and

specialization of technical knowledge, and multidisciplinary approaches, which synthesize different

21

perspectives, by attempting to unify knowledge through application (François 2006). Transdisciplinary

research looks for similarities among models of different system structures (François 2006) that

may extend across organizational, analytical, and geographic scales (Stokols 2006). This process

requires close collaboration that transcends traditional academic boundaries to produce shared

conceptual models (Mitrany & Stokols 2005). The multi-field model of community, landscape, and

pro-environmental behavior outlined in this paper seeks to accomplish this goal by linking common

points of related field structures across social disciplines. A focus on behavior both as a product and an

element that reinforces field bonds inherently grounds the model in application while engendering new

knowledge about the multi-level influential factors of such behaviors. Our discussion of a hypothetical

application of the proposed community landscape model of pro-environmental behavior in studying

New Jersey coastal town conflicts strongly suggests a need for transdisciplinary responses to sincere

human-natural resources challenges.

Lewin (1946) called for researchers to work collaboratively with communities to analyze and

solve social problems (Stokols 2006). Stedman (2016) has recently called for social and ecological

system researchers to expand their scholarship to include sense of place initiatives. While this process

has barriers, such as academic jargon, discipline pride, and fear of professional failure from favoring

breath over depth (Mitrany & Stokols) – environmental problems simply cannot wait for disciplines

to sort out their individual differences. We believe our model takes a simple yet comprehensive

step towards those lofty goals by synergizing theory across the scholarship of social psychologists,

landscape designers, sociologists, and many others.

We present a conceptual model bridging three fields from three distinct disciplines to diagram a

transdisciplinary way of informing placed-based pro-environmental behaviors. The next step is to

operationalize the model in collaboration with communities in helping develop creative strategies for

identifying and solving environmental problems.

22

Chapter 3: Methodology

3.1 Research Design

The research design for this dissertation adopts a mixed methodology. Mono-method designs

may suffer validity issues when it is difficult to separate the types of measurements used from the

method selected. Triangulating multiple methods compliments the strengths and weaknesses of

each method (Tashakori & Teddle 1998, Creswell 2003, and Hunter & Brewer 2006) and attempts

to maximize internal and external validity through convergence among results (Hunter & Brewer

2006). The methodology for this study borrows from the framework established in Luloff’s The

Doing of Rural Community Development Research (1999). Here, the methodology design transitions

from an exploratory macroanalysis of statewide growth and energy consumption trends to a rigorous

microanalysis of local energy and environmental values and energy consumption. This approach is

validated both by Luloff’s analysis of rural communities (1999) and by Ryan (2004), who argues that

locally-generated energy consumption data is more accurate than national datasets.

This study implements a mixed-method sequential design through four phases: exploratory site

analysis (non-reactive data), key informant interviews (qualitative data), and survey design and

implementation (quantitative data) and spatial analysis (non-reactive data). Through this mixed

model framework, the study seeks to balance the strengths and weaknesses associated with researcher

bias and internal and external validity. Non-reactive spatial and socioeconomic data were explored

to generate a minimally biased selection of study sites, though the ultimate scope of the analysis was

adjusted for convenience due to limited time and funding. Qualitative data gathered through key

informant interviews was intended to improve internal validity by grounding the study in the values,

perspectives, and barriers of the local townships. Interview data informed the construction of a survey

instrument which was distributed to a broad sample of homeowners in both townships to improve

the external validity of the study. Spatial analysis attempted to reduce survey bias through the use

of non-reactive land cover data. Figure 3 illustrates how the study framework attempted to achieve

convergence of findings through triangulation across qualitative and quantitative methods.

23

3.2 Exploratory Site Analysis

The purpose of the exploratory site analysis is to concatenate a series of data used to identify and

select two study sites from the population of 2,568 minor civil divisions (MCDs) in Pennsylvania.

Study sites were selected by observing three metrics: household growth, change in residential energy

use, and change in land cover between 2001 and 2011. Household growth represents a rough measure

of urbanization over time (Theodori & Luloff 2000). Land cover illustrates how urbanization has

impacted land cover patterns. Residential energy use identifies which MCDs are consuming greater

amounts of energy. These years were selected to: (1) match all years to a common base in order to

minimize error; and (2) match the availability of land cover data.

3.2.1 Data Gathering and Refinement

Typically, the unit of scale for statewide exploratory analysis is the county. However, Krannich and

Luloff (1993) note that counties often contain a high degree of internal variation. For this reason, this

phase adopts the minor civil division (MCD) as the unit of analysis. The U.S. Census Bureau defines

an MCD as a primary county subdivision more commonly known as towns, townships, boroughs,

cities, or districts with governmental and administrative functions (US Census 2015). Data for each

MCD within Pennsylvania was gathered using the methods and sources outlined below.

3.2.1.1 Land Cover Data

Changes in geospatial land cover data are observed over time. The National Land Cover Database

(NLCD) provides data identifying land cover changes between the years 2001 and 2011. For this

reason, the availability of land cover data established the timescale for this study. However, MCD

Phase 1:Exploratory Site Analysis

Quantitative, non-reactiveLand cover dataU.S. Census dataElectricity sales

Quantitative, non-reactiveLand cover data(Spatial analysis)

Quantitative, reactive survey questions

(Statistical analysis)

Qualitativeinterview notes

(Latent content analysis)

Phase 2:Key Informant Interviews

Phase 3:Survey

Phase 4:Spatial Analysis

Figure 3: Study framework

24

boundaries and/or annotation have changed slightly between the observed years. The resolution of

these conflicts is important, because the unique geographic identification number for each MCD links

all land cover and sociodemographic data across space and time. Conflicts between the 2001 and 2011

datasets were identified and resolved as follows:

• Two (2) conflicts contained minor differences in the text of the MCDs name. Text was edited to

match.

• Five (5) conflicts observed a change in their unique geospatial identification number (GEOID).

GEOIDs were edited to match the 2011 version.

• Five (5) conflicts did not appear to exist in the 2011 dataset. These were removed from both

datasets and were not included in future analysis. Boundaries for all MCDs were visually

checked to confirm that there had been no significant merging or other changes over time.

The NLCD classification includes 18 unique land cover types (see National Land Cover Database

Classification 2011, http://www.mrlc.gov/nlcd11_leg.php) consisting of major and sub-classifications.

For simplicity, land cover types were condensed into the following seven (7) groups:

• Open Water;

• Developed: (Developed Low intensity, Developed Medium Intensity, Developed High Intensity,

Developed Open Space);

• Barren Land;

• Forest: (Deciduous Forest, Evergreen Forest, Mixed Forest);

• Shrub: (Dwarf Shrub, Shrub, Grassland, Sedge, Lichens, Moss);

• Agriculture: (Pasture / Hay, Cultivated Crops); and

• Wetlands: (Woody Wetlands, and Emergent Wetlands).

GIS software was used to generate individual tables for each MCD summarizing the total acreage

of land that has changed for each land cover type. ESRI ArcMap’s Zonal Histogram tool created a land

cover table of each 30-meter cell for each individual MCD. The number of cells for each land cover

type was converted to square meters and number of acres. In this manner, the number of acres of each

land cover type for each MCD in 2001 and 2011 was calculated. The percent change of each land

cover type was calculated to compare changing land cover patterns.

Initially, the land cover data analysis produced a significant amount of erroneous data. All 2000 or

25

2011 land cover data with a “0”, indicating no acres of that land cover type were present, produced an

error when calculating percent change from 2001 to 2011. The error was resolved by observing what

percent of data each land cover type contained errors. The majority of errors were contained in Water

(21%), Barren (48%), Shrub (43%), and Wetlands (35%) land cover types. By comparison, Developed

(<1%), Forest (1%), and Agriculture (9%) contained relatively few errors. Water, Barren, Shrub, and

Wetlands land cover types were removed from further analysis. The removal of these land cover types

significantly reduced the amount of error in the data. Removing these land cover types did not affect

the analysis, as few, if any, homes are located in these areas. Coincidentally, the Developed, Forest,

and Agriculture land cover types contained more than 95% of the total land cover in PA. The resulting

data set contained fewer than 5% errors. All remaining errors were coded as “No Data”, effectively

removing them from further analysis.

3.2.1.2 Household Data

The number of households for each MCD was obtained from the 2011 U.S. American

Communities Survey and the 2000 U.S. Dicentennial Census. Different datasets were required, as the

US Census does not conduct the same level of survey each year. Further, no census was conducted at

the MCD level in 2001. The number of existing households for each MCD was converted to describe

the percent change between 2001 and 2011.

3.2.1.3 Energy Consumption Data

Energy consumption at the MCD level is not readily available. Instead, a model was constructed

to approximate energy consumption for each MCD. Energy consumption for residential, commercial,

industrial, and transportation customers is available in one-year increments from the Energy

Information Administration (EIA). In 2011, there were over 120 energy utilities in Pennsylvania. The

Pennsylvania Public Utility Commission identifies eleven primary electrical distribution companies

(EDCs) operating in Pennsylvania (2002, 2011). Based on EIA data, in 2001 these EDCs accounted

for more than (85%) of total energy consumption and (86%) of residential consumption. In 2011,

the EDCs share fell but maintained more than (61%) of total consumption and (67%) of residential

consumption. This trend may be a result of the emergence of renewable energy policy in Pennsylvania.

In 2004, the Alternative Energy Portfolio Standard (AEPS) required all EDCs to provide 18% of

26

their energy from advanced or renewable sources by 2020. Wind farms have been the major producer

of renewable energy during this period, increasing from 0.2% to 24.5% of all renewable electricity

generated between 2000 and 2011 (Commonwealth Economics 2013, http://www.pennaeps.com/

aboutaeps/). However, the eleven primary EDCs remain the dominant market share of electricity

distribution and were used as the basis for generating energy consumption estimates for each MCD.

This process was accomplished by digitizing the distribution area for each EDC, available from the

Electric Power Outlook for Pennsylvania (Pennsylvania Public Utility Commission 2011, 2002) into

GIS data. MCDs were then assigned to each distribution area by geographic location. Discrepancies

between overlapping boundaries were adjusted by assigning the MCD to whichever distribution area

the majority of its land area is located within.

Presumably, MCDs with a greater number of homes and a higher median age of home experience

a higher share of electricity consumption. Residential sales were distributed based these two factors

within each MCD. US Census data provided the number and median age of homes in each MCD. The

following equation determined relative energy consumption for each MCD:

EIy = (TEDy) x (MHHy / THHy) x PMA

Where:

EIy = Index of relative residential electricity consumption for a given year

TEDy = Total electricity distributed by a utility

MHHy = Number of households in an MCD in given year

THHy = Total number of households in a utility’s distribution area in a given year

PMAy = Percentile rank (youngest to oldest) of median household age in an MCD in a given year

Seven (7) MCDs did not have household age data available and were removed from the dataset

resulting in a final sample size of 2,569. This formula was applied to the 2001 and 2011 datasets and

the percent change in EI was calculated.

3.2.2 Site Selection Model

In order to identify potential study sites, MCDs were classified into quadrants of a 2x2 grid

consisting of high and low household and energy change (see Table 3.1). Typically, this method was

27

used to select four (4) study sites, one from each quadrant (see Luloff 1999). However, to limit the

scope of this study, two of the four quadrants were emphasized: high household changes with high

energy change (Q1) and high household change with low energy change (Q2). The median value

for household and energy change (-6.7% and 1.1%, respectively) determined the boundary between

quadrants. Additionally, MCDs with less than 500 total households were removed to disqualify very

rural areas that may prove difficult for further study. Land cover data was compared between the

remaining study candidates to determine two polarizing sites for further study.

The analysis produced 812 MCD candidates within Q1 and 165 candidates within Q2 (see Figure

4). However, few couplings between places of similar size (number of households) in Q1 and Q2 were

apparent. Further, many of the qualifying MCD candidates were ultimately considered to be unfeasible

due to long driving distances, the need for multiple visits to conduct fieldwork, and limited time and

funding available for the study. Spring Township (Q1) and East Buffalo Township (Q2) emerged as a

pairing of relatively similar size, positive household growth, land cover changes, contrasting energy

trends, and relative proximity minimizing travel related costs and time expenditures (see Figure 5).

Table 3.1: Example site selection quadrants

HighΔenergyconsumption

LowΔenergyconsumption

HighΔhouseholdgrowth Q1 : Study Site 1 Q2: Study Site 2

Median % change

LowΔhouseholdgrowth NA NA

Median % change

28

3.2.3 Description of Study Sites

Total residential energy consumption in a given area is expected to increase as the number

of households increase. Spring Township and East Buffalo Township demonstrate opposing

household energy trends between 2001 and 2011 (see Table 3.3). While Spring Township

contained approximately 1,100 more households than East Buffalo Township in 2011, the areas

share approximately the same household density, suggesting similar urban form. Spring Township

experienced a rapid increase in the number of households (22%) compared to the state’s median

household rate of growth (-6.7%). Spring Township’s estimated residential energy consumption

significantly increased by 70% during this same period. In comparison, East Buffalo Township

experienced a slower rate of household growth (7%), yet was well above the state median. However,

estimated residential energy consumption in East Buffalo Township decreased (-1%), which is a

unique trend in the state. Both townships lost forest and agricultural land to development of low-

density residential land. Spring Township experienced significant household growth compared to

East Buffalo Township’s relatively static growth. Both townships appear to be suburbanizing, a trend

which is expected to increase energy consumption. Despite this, East Buffalo experienced virtually no

increase in energy consumption.

The two townships exhibit several other notable differences. First, the median age of households

Figure 4: Location of qualifying MCDs by quadrant

-50%

0%

50%

100%

-50% 0% 50% 100% 150% 200%

ΔHou

seho

lds

ΔEnergy

Q1SpringTownship

EastBuffaloTownshipQ2

29

Esri, HERE, DeLorme, MapmyIndia, © OpenStreetMap contributors, andthe GIS user community

Figure 5: Location of Spring Township and East Buffalo Township

SpringTownship EastBuffaloTownship

Table 3.2: Comparison of study sites

Q1:SpringTownship Q2:EastBuffaloTownship Difference

Total # of households (2011) 3,129 2,023 1,106

Acres 17,368 9,702 7,666

Density 0.18 0.21 -0.03

Changes (2001-2011)

Household growth 22% 7% 15%

Electricity sales 70% -1% 71%

Developed acres 12% 2% 11%

Forest acres -1% -1% -1%

Agriculture acres -3% -1% -2%

Median year structure built 1982 1970 12

House heating fuel

Utility gas 20.9 11.9 9.0

Bottled, tank, or LP gas 1.2 1.3 -0.1

Electricity 40.3 50.6 -10.3

Fuel oil, kerosene, etc. 29.1 31.9 -2.8

Coal or coke 3.5 1.0 2.5

Wood 3.5 2.0 1.5

Solar energy 0.0 0.0 0.0

Other fuel 1.2 0.3 0.9

Median income $54,397.00 $67,352.00 $ 12,955.00

Median home value $162,700.00 $211,400.00 $48,700.00

30

in East Buffalo Township are approximately 12 years older than homes in Spring Township. As stated

earlier, presumably older homes are less efficient and consume a greater amount of energy. Older

homes in East Buffalo Township should suggest higher energy consumption – yet this is not the

case. Second, while the sources of energy are relatively similar between the two townships, Spring

Township relies slightly more on utility gas than East Buffalo. Third, East Buffalo Township’s median

income and home value are significantly higher than Spring Township’s (by $12,000 and $49,000,

respectively) suggesting a greater concentration of wealth in East Buffalo Township. Fourth, East

Buffalo Township experienced an influx of new residents from outside of the county at nearly twice

the rate of Spring Township.

3.3 Key Informant Interviews

The purpose of this phase of the research design is two-fold: 1) to identify the perceptions, values,

and perceived barriers with regards to local landscapes and energy consumption, and 2) to identify

recent local issues related to land development or energy of which the general public is aware. This

information was used in the design of a survey instrument with a locally relevant frame of reference

for the respondent. Framing is an important process that identifies communities with different values

within a place (Jansujwicz et al. 2013). A survey designed specifically for a place should receive

higher response rates and improve the internal reliability of the study design.

A total of 18 key informants were interviewed over a three-month period. Key informants (KIs)

Table 3.3: Summary of key informants

SpringTownship EastBuffaloTownship

Position Gender Date Position Gender Date

Town Manager M 2/19/16 Town Manager F 4/28/16

Zoning Official M 2/19/16 Planning Official M 4/28/16

Forest Manager M 4/19/16 Recreation Manager F 4/28/16

Newspaper Editor M 2/12/16 Radio Editor F 5/17/16

Education / Energy M 4/19/16 Education / Energy F 4/28/16

Education / Energy M 4/19/16 Local Energy Expert F 5/31/16

Small Business Owner M 4/20/16 Small Business Owner M 5/11/16

Fire / Service M 4/25/16 - - -

Community Leader F 5/2/16 Community Leader F 5/9/16

Bank M 5/5/16 - - -

31

included elected officials and township managers, community leaders, educators, and members of

local media outlets who lived or worked in the township (Luloff 1999, Elmendorf & Luloff 2001,).

KIs were also identified through a snowball method, where KIs were asked to suggest other residents

or colleagues who may be important informants (Luloff 1999, Elmendorf & Luloff 2001, Keefer et

al. 2003, Eser & Luloff 2003). Three additional KIs were identified in this manner. KIs were initially

contacted by publically available phone numbers or email addresses and were invited to participate

in a short interview. The majority of informants in Spring Township were male, while the majority of

informants in East Buffalo Township were female. The gender skew was a function of who held key

roles (manager, directors, community leader, etc.) in each township.

Most interviews (13) were held at the KIs office or place of business. The remaining interviews

were conducted through phone (4) or at a café (1). I took notes regarding key ideas, quotes, and

non-verbal cues on a laptop during the interviewer, though KIs were offered the opportunity to

remove the laptop from the interview setting (and be replaced with hand written notes) if they were

uncomfortable.

A short interview guide was prepared to facilitate dialogue with KIs; the interview typically lasted

about 45 minutes (see Appendix A). The interview guide followed a pyramidal funneling strategy

where questions transitioned from broad impersonal issues to specific personal topics (Hay 2000).

The structure was intended to allow the KI to become comfortable with the interviewer before talking

about their personal values or beliefs. The guide consisted of open-ended questions with prompts to

encourage a well-rounded conversation. KIs were asked about their perceptions of the town in terms

of its physical and social qualities, recent changes and residents’ reactions to them, specific landscapes

residents cared about, and how residents defined and prioritized environmental and energy values. At

the conclusion of each interview, I reviewed my notes, and added additional commentary about the

interview and methodology.

A short summary of the findings from each township was sent to each of the KIs for their review to

ensure that the findings were accurate and comprehensive (see Chapter 4: Results). Three KIs replied

and offered additional context for their responses which served to clarify related survey questions.

32

3.4 Key Informant Data Analysis

Data gathered from the KI interviews was subjected to a latent content analysis. This type of

analysis searches for common themes in qualitative data, which may not have been directly expressed

by KIs (Tashakori & Teddle 1998, Creswell 2003). Typical themes included:

• Setting and context;

• Perspectives;

• Ways of thinking about people and objects;

• Process;

• Activities;

• Strategies;

• Relationships and social structures;

• Pre-assigned codes.

The goal of the analysis was to identify five to seven underlying themes in the data (Creswell 2003).

Given the research topic’s focus on social and environmental values and relationships, the analysis

focused on locating themes related to setting and context, perspectives, activities, relationships and

social structures.

My notes were collated into a single transcript for each township and imported into the QSR Nvivo

software. Each transcript was read in its entirety twice and coded for possible themes. Each theme

was then organized as major node with several related minor nodes that represent key distinctions

within a theme (such as a series of specific land covers that compose a broad rural setting). Nodes

were summarized in two ways that illustrate how particular themes tended to dominate the interview.

Frequency describes the number of times a theme was mentioned. Coverage, the percentage of

transcribed notes within a theme, was used to assess which themes were most relevant to each

township and were strong candidates for inclusion in the survey.

3.5 Survey Design & Implementation

The purpose of the survey was to improve the generalizability of KI data by gathering primarily

quantitative data from a broad sample in each township. The survey was designed to gather data about

each component of the proposed Community Landscape Theory of Pro-environmental Behavior (see

33

Chapter 2, Figure 2). While the theory determined the survey structure, the KI data informed the

language of the survey. KI data enabled the survey to be designed specifically for each township by

including explicit landscapes, community activities, and barriers prevalent in that area.

3.5.1 Survey Design

The survey was organized into six (6) major sections: environmental and energy values,

community engagement, perceived barriers, socioeconomic data, energy related behaviors, and

environmental controls (see Appendix C). Environmental values were composed of three scales, place

attachment, landscape values, and general place values in a Likert scale battery. Place attachment

was addressed through three questions asking to what degree respondents cared for their town, felt it

was unique, and their willingness to move away from the town. Landscape values asked respondents

to indicate how strongly they valued five different types of land covers, farmland, forest, mountains,

green space, and water, which were highlighted in the KI interviews. General place values were other

qualities that were often mentioned in interviews, such as sense of community in the town, having

small town feelings, and access to arts and culture.

Energy values were gathered through a series of Likert scale questions that asked respondents how

strongly they felt about energy consumption, conservation, sources, and the environmental impacts of

energy. Community engagement was addressed through a battery of questions regarding with whom

and how often respondents socialized, participated in local issues, attended local community events,

and volunteered their time. Perceived barriers were gathered by a Likert scale battery of questions

about conditions that may have prevented the respondent from conserving energy. Survey respondents

were also asked to provide information about their home and sociodemographic status (for use as

control variables) and their home address for spatial analysis.

To improve the likelihood of respondents providing data, energy behaviors were presented in three

different forms. Respondents were asked to provide electricity and gas information from a specific

month’s utility bill. Presumably, respondents might not have had convenient access to this information.

Thus, a second question was included asking respondents to estimate the dollar amount of a specified

month’s electricity and gas utility bill. Kroon et al. (2014) stated that sophistication of fuel sources

increased as household income increased. This transitory process migrates households from the use

34

of biofuels (firewood) for primarily subsistence purposes to transition fuels (charcoal, coal); advanced

fuels (electricity, gas, biofuels) required more advanced technologies (2011). This process does not

occur in exact phases. It is more likely that fuel stacking occurs where energy sources overlap. For

this reason, single measures, such as electricity utility costs, do not sufficiently capture household

energy use. Respondents were given the opportunity to report energy consumed from fuel oil, bottled

tank gas, coal, wood, solar, hydro, wind, or other alternative sources. They were also asked if they had

performed a series of energy conservation behaviors, and/or how likely it was that they would perform

those behaviors in the future. To control for environmental factors that may affect household energy

consumption, respondents were asked to report on a series of conditions about their current home,

including age, square footage, adjacent tree cover, energy efficient appliances, etc. (United State Green

Building Council 2010).

3.5.2 Survey Distribution Methods

According to the U.S. Census classifications, Spring Township and East Buffalo Township are

considered urban clusters (U.S. Census Bureau 2010). However, both are best described as agricultural

centers with significant patches of natural forest and relatively low-density development. Functionally,

they are rural places. Surveying rural places presents challenges due to distances between homes and a

traditional lack of trust of people from outside the community. Further, surveys continue to experience

a significant decline in response rates (Steele et. al 2001, Allred & Davis 2010). Given the challenge

of distributing a survey to two rural places with limited time and funding, four (4) separate distribution

methods were employed representing a significant effort over a two-month period.

3.5.2.1 Hybrid Postcard Method

Hybrid methods of distributing surveys offers an enticing balance between response rates and

costs. Kaplowitz et al. (2004) assessed the response rates of various combined email and postcard

distributions and reminders compared to a traditional mail survey. Each postcard an email contained

a web link to an online survey. While the mail survey received the highest response rate (31.5%),

email/postcard combinations received response rates between 20% and 29.7% at nearly a tenth of the

cost of mail surveys. Their results demonstrate the viability and financial efficiency of online surveys

coupled with print media. To maximize the limited funding available for this study, a hybrid postcard

35

distribution method was employed according to the following method.

Households in both Spring and East Buffalo townships have access to online utilities sufficient

to complete the survey. The National Broadband Map (NBM) was created through an initiated

by the National Telecommunications and Information Administration (NTIA) and the Federal

Communications Commission (FCC) to visualize broadband access across every neighborhood in the

United States. The NBM shows that nearly all residents of Centre (95%) and Union (87%) counties

(Spring township and East Buffalo township, respectively) have access to minimum online utilities.

The NBM describes minimum utilities as those with wired access to downloads speeds greater than

768k per second. All residents have access to wireless utilities.

A list of all residential addresses in each township was acquired from the Centre County GIS

Department and the Union County Tax Assessment Office. To ensure only residential addresses were

included, address lists were checked to confirm that: (1) the property type listed was residential; (2)

property owners listed the same address as the property; and (3) properties were physically located

within the study area. To confirm physical location within the study area, the address lists were

imported into ArcOnline mapping software and visually confirmed. After removing non-qualifying

properties, the address lists contained 2,420 properties in Spring Township and 1,780 properties in East

Buffalo Township. Approximately 20-25% of properties were removed from each address list.

A targeted sample size of approximately 300 households per township was chosen to achieve a

95% confidence interval and a 5% margin of error. Six-hundred (600) households were randomly

selected from each township by assigning each household a randomly generated number, ordering

from high-to-lowest value, and selecting the highest 600 households. A total of 1,200 postcards

were sent to the identified households. The postcard contained a brief introduction to the research

study, a web link to an online survey, and a quick response (QR) code for use with mobile devices.

Postcards were reviewed for content and approved for use of the Penn State University and Stuckeman

School logos by the school’s marketing and communications department (see Appendix XX). As an

incentive, all postcards also advertised a random drawing for a $25 gift card for eligible respondents

who completed all survey questions. Postcards were printed, checked for accurate mailing addresses

according to US Postal records, and mailed by Vistaprint.

The survey methodology planned for two additional waves of postcard reminders to recipients

36

who had not yet responded. However, the first wave received a very low response rate that was not

commensurate with the cost of mailing postcards (see Chapter 4: Results). Instead, three alternative

distribution methods were implemented. The following sections outline each method.

3.5.2.2 Key Informant Convenience Method

Following the postcard distribution, KIs were contacted for their assistance in gathering responses.

Each KI was contacted via email and requested to send the web link and a PDF of the postcard to 3-5

of their friends or colleagues who live within the township. KIs were informed that because they had

already participated in the study, they were not eligible to participate in the survey. Both Spring and

East Buffalo township managers offered to host the web link with a short description of the survey

on the township’s website (however East Buffalo Township was unable to do so due to technical

difficulties with the website). Notably, this method adds to the number of respondents at the expense

of external validity.

3.5.2.3 Drop-off, Pick-up Method

A Drop-off, Pick-up (DOPU) method often provides data within a shorter period of time then

mail surveys but at a trade-off of significant time and labor investments (Steele et al. 2001). Alfred

& Davis (2010) recommended that researchers visit each household three times; first to describe the

survey’s purpose and schedule for pick-up during a face-to-face conversation with the potential survey

respondent, second to pick-up or leave a reminder, and third to pick-up or leave a self-addressed

envelope. (Due to the limited funding, the self-addressed envelope was omitted). To ensure that

surveys were not offered to households who had already declined to participate via the postcard

survey, a new set of households were randomly selected. Participants for the postcard survey were

selected from the top 600 randomly ordered households, whereas participants for the DOPU survey

were selected from the bottom 150 of the ordered households.

Three colleagues (one in Spring Township and two in East Buffalo) attempted drop-offs on a

Sunday morning through late afternoon. The day was selected based on the availability of participants

and to avoid conflict with a Penn State home football game. Surveys were left in a specialized clear

plastic bag on the front door if there was no answer at a household. If someone answered the door,

participants requested to speak with the head of household who needed to be over 18 years of age.

37

The head of household was determined to be the person who was most familiar with the home’s utility

bills. Participants returned two days later for first pick-up, left a reminder flyer, and again two days

later for second pick-up.

3.5.2.4 Facebook Advertisement

An advertisement was placed that targeted all Facebook users older than 18 years of age within 3

miles of Spring Township and East Buffalo Township. The advertisement increases the visibility of a

Facebook page (similar to a small website) that directs users to the online survey. The advertisement

was active for one week. Facebook advertises based on a proprietary internal algorithm that fluctuates

ad exposure daily to achieve optimal results. Through this method, Facebook estimated that between

560 and 1,800 users will have viewed the advertisement as part of their normal activities on the

website.

3.6 Spatial Analysis

As a compliment to the survey questions about environmental values, land cover data surrounding

each respondent’s household was gathered. The resulting variables indicated the amount of each land

cover type that was within one-half mile of a household. Home addresses (street, city, state, and zip

!

Esri, HERE, DeLorme, MapmyIndia, © OpenStreetMap contributors, andthe GIS user community

Figure 6: Example of individual home’s one-half mile land cover radius

38

code) were provided as part of the survey. ArcGIS Online mapping software was used to geolocate

each address. All addresses were visually verified to be within the townships’ boundary. ArcMap

software was used to create a one-half mile buffer around each address as a representation of the

households’ immediate physical environments. One-quarter mile is a commonly used measure to

indicate a comfortable walking distance for most people. One-half mile was chosen to represent both

the walking and visual experience of the immediate environment (see Figure 6). Land cover data from

the most recent (2011) National Land Cover Database, the same as used for selection of the townships,

was extracted within the buffer for each address. This data was summarized into major land cover

categories – developed (open, low, medium, and high), forest, agriculture, and all others – that match

language used in the survey. Each land cover type was converted to a percent of the total land area to

describe the land cover mix for each households’ environment. This data was included in the survey

dataset for further statistical analysis.

3.7 Survey Data Refinement & Analysis

Survey data was analyzed through the following steps: refinement of data, factor analysis,

descriptive statistics, modality and normality tests, correlation, and sample comparison. The following

sections briefly describe the rationale and methods associated with each. Results are discussed in detail

in Chapter 4: Results.

3.7.1 Data Refinement

Survey data from across survey modes of distribution and from both townships was combined

into a single dataset. Each survey response was coded with a unique identification number. (DOPU

survey’s that were analogue coded were also annotated with an identification number on the paper

survey). The following steps were taken to ensure integrity of the data:

1. Dummy variables for each of the four survey modes and two townships were created.

2. Eight survey questions were reverse coded to be consistent with directionality of Likert scale

responses.

3. Thirteen responses that contained no data were removed from the dataset. (This was likely due

to respondents clicking on the online survey link, but not answering any questions).

4. Responses in kilowatt-hours (electricity) or cubic feet (gas) were converted to a common

39

measure (joules) through multiplication of energy conversion rates. (Washington University

2005, Iowa State University 2008). The same process was applied to responses of alternative

fuel sources.

5. Responses in dollar amount (from utility payments) were converted to joules by dividing the

dollar amount by the local utility distributor’s August rates (West Penn Power, Columbia Gas,

Citizen’s Electric Company, and UGI Central Penn Gas).

The dataset was formatted for use in the IBM SPSS statistical software. This involved coding all

non-responses to individual survey questions as -99.

3.7.2 Data Analysis

The survey received relatively low response rates which resulted in a small sample size (see

Chapter 4, Table 4.5 for more information). The analytical techniques described here were intended to

maximize the statistical relevance of the data through reduction of variables. Factor analysis of survey

batteries served two functions. First to evaluate the validity of scales within a battery, and second

to condense the total number of variables. Batteries of environmental values, energy values, and

perceived barriers were factored into one to three components. A standard of 0.50 was used to assess

which component best aligned with an individual variable. Other survey items were summarized as a

sum, such as yes/no questions, or an average of responses for Likert scale or categorical questions.

To better understand the shape of the data, dependent variables (the four different types of

energy behaviors), control variables (energy values, perceived barriers, housing characteristics, and

sociodemographic status), and determinate variables (environmental values, community engagement,

and land cover) captured by the survey were described in terms of their mean, standard deviation, and

variance. Control variables with low variance (1.1 or less) were considered not likely to be influential

and were removed from further analysis to limit the number of accountable variables and maximize

the utility of further statistical analysis. The community landscape framework being tested here is

composed of determinate and dependent variables. All determinate variables were included regardless

of variance to allow further analysis to assess the role of each in influencing energy behaviors.

Energy behaviors were tested for differences among survey and township modality and for

normality, using the Kruskal Wallis or Mann-Whitney tests and Shapiro-Wilk test, respectively.

40

Significant differences observed in the modality tests resulted in the inclusion of both survey and

township dummy variables in all further analysis. To achieve normality in two of the four energy

behaviors, eight (8) extreme outliers were removed from the data set based on a visual evaluation of

Q-Q normality distribution plots of the data. Correlation tests between energy behaviors and variables

representing physical and social environments and community engagement were performed while

controlling for energy values, housing characteristics, perceived barriers, and sociodemographic

conditions at the 10%, 5% and 1% significance levels. Variables that correlated at the 10% or less

level were considered significant and were included in a subsequent series of block regression models.

The regression models identified which, if any, variables significantly influenced energy consumption

behaviors by removing variables that did not explain variance in behaviors. The remaining significant

variables were included in a reduced regression model (Zawadzki et al. 2016, Flint and Luloff 2007).

This method improved the analytical rigor of the study by deciphering if variables remained significant

in the context of other variables. The results of the block regression model were tested for collinearity

through a partial correlation analysis. Last, sample data was compared between townships and across

U.S. Census sociodemographic data to assess the generalizability of the study.

41

Table 4.1: Summary of key informant interviews

Lengthofinterview(in minutes)

Timelivingorworkingintown(number of years)

Spring Township East Buffalo Township Spring Township East Buffalo Township

Average 39 53 16 22

Median 38 55 13 14

Min 25 35 1.5 5

Max 55 75 40 60

Chapter 4: Results

4.1 Summary of key informant interviews

Eighteen key informant interviews were conducted over a period of four months generating over

fourteen hours of quantitative and qualitative data. Much of the interview data pertains to key values,

priorities, and concerns related to a town’s surrounding landscape and energy conservation attitudes.

Interviews lasted between 25 and 75 minutes. The length of time an informant had lived or worked in

the town varied from 1.5 to 60 years, but a typical informant had more than ten years of experience in

their town (Table 4.1).

4.1.1 Summary of Spring Township interviews

Much of the Spring Township interviews focused on place-based values associated with

rural landscapes and other valued qualities of a small town. These themes were reflected in brief

conversations about land preservation priorities. Energy was minimally discussed, as informants

uniformly agreed it was not a priority of most town residents. Many themes were interrelated and

supported a deep aesthetic connection to the rural landscape. The major themes are described below.

Table 4.2 lists a summary of themes by frequency (number of times mentioned) and coverage (length

of note text related to a theme).

Rural Setting

Informants most often mentioned the rural setting as the quality they valued the most about their

town. This theme was described in two ways: visual aesthetic quality of the landscape and loosely

defined relationships with neighbors. Farmland was mentioned most often to describe what informants

42

liked about the rural setting, though they also valued the diversity of forests, mountains, general

green space, and water bodies that offered scenic and recreational opportunities. Informants described

their relationship with these landscapes as personal and passive. Words like “peaceful”, “quiet”, and

“natural beauty” were used to explain the enjoyment of being alone in the landscape or the sense of

escaping in nature. These landscapes were also associated with valued activities such as hunting and/or

fishing.

Preservation topics amplified the value of the rural setting. Informants were concerned about

potential development of the natural setting from urbanization or sprawl. Frequently, informants

expressed interest in preserving forests and agricultural lands. One respondent said:

It’s a gift that we get to be here. We have to be really careful about how we treat it,

and what we leave for our kids and their kids.

Informants valued the convenient relationship between their town and the natural setting. Informants

generally preferred to live in a home that enabled access to the natural landscape and that was close,

Table 4.2: Summary of key informant data for Spring Township

Themes Frequency Coverage

Place values 160 28.1%

Rural Setting 35 3.0%

Farmland 18 2.9%

Forest 9 2.0%

Mountains 8 1.1%

Green 6 1.1%

Water 4 0.8%

Small Town 29 4.9%

Sense of Community 26 7.2%

Location 15 2.9%

Education & Culture 10 2.2%

Priorities 99 18.4%

Landscape 22 4.0%

Preservation 14 2.2%

Urbanization & Sprawl 12 2.7%

Water Quality 12 1.6%

Energy 39 9.1%

43

but not too close, to neighbors. One respondent best summarized this idea as:

All the benefits of the farm, and none of the responsibilities.

Location

This balance between developed and natural land appeared to be a reoccurring theme. While

appreciative of the natural benefits surrounding the town, residents also pragmatically enjoyed easy

access to more urbanized areas (State College) and local industrial development for the employment,

entertainment, and cultural opportunities provided by each. Though there were concerns of the town

becoming more of a suburb of State College, most appreciated what Spring Township offered …

The best of both worlds.

Small town

While these informants appreciated access to urban areas, they also valued the quiet atmosphere

and safety of their self-defined small town. They enjoyed that Spring Township did not feel

overcrowded and liked seeing the same people around town due to the limited number of subdivisions.

Again, words like “quiet” and “peaceful” were used to identify qualities they liked about the town.

Underlying their small town was a mild isolationism theme. Informants appreciated that Spring

Township could be “[their] spot,” and were concerned about emerging conflicts associated with

urbanization pressure from State College. One informant suggested suburbanization would cause an

unwelcome and upsetting sociodemographic shift.

Community

Spring Township informants placed a high value on their sense of community. There were ample

opportunities for the town’s residents to gather at fairs and parades – particularly the fire department’s

annual carnival – and other seasonal events. The carnival appears to require a massive community

effort each year and has relied on volunteers to implement. Informants described the event as:

A lot of work, but at the end of the week everyone feels connected.

A real demonstration of teamwork.

Informants believed the town had a “neighborly feel” and the more committed people were to being

44

involved, the more they would appreciate the town. Neighbors actively helped each other especially in

response to tragedies. There was also a perception that families tended to stay in town for generations

and that wayward individuals found their way back. Families with five to six generations of volunteers

supported the fire department, and there were several examples of two to three generation firefighters

active at a time.

Energy

There was a consensus that energy was not a concern to residents or something they thought about

until there was a strong economic impact, such as an increase in energy utility bills. Several quotes

summarized the lack of concern regarding energy consumption:

When it hits their wallets.

Throw more wood in the stove, its cheap.

Just want to flip the switch.

When they start putting a price tag on it.

It’s always there.

However, there appeared to be what one informant referred to as a steep divide regarding energy

issues. While most residents did not think about energy, one informant suggested it was because they

simply were not aware of alternative actions – further stating it was a problem of education.

All they know is throwing more wood on the stove.

There was some evidence of conservation efforts in Spring Township. Act 129 is a refund program

that mandated utility providers to offer incentives for energy improvements to homes or businesses.

Several town ordinances were recently updated to allow solar or other alternative energy sources.

The energy market has opened to allow consumers to find alternative energy sources. One informant

asserted that people “definitely cared” about energy. However, these examples appeared to be in the

minority in Spring Township. An informant summarized the town’s landscape values and energy

priorities as

We use it [energy] everyday, but it’s not in our faces as much as our landscape.

45

4.1.2 Summary of East Buffalo Township interviews

Much of the dialogue with East Buffalo Township informants centered on the relationship between

what informants referred to as the “town and country” dynamic. The concept of town or country

permeated virtually all themes and priorities. The relationship between the developed downtown area

and the surrounding landscapes was the source of much value and conflict regarding changes to either

environment. Notably, the town that informants referred to (Lewisburg) was technically outside the

East Buffalo Township study area, but residents did not typically distinguish between these two minor

civil divisions (MCDs).

Town & country

Informants described the town and country relationship in East Buffalo as both a benefit and source

of conflict due to differing values between town and country residents. Residents enjoyed access to

the arts and cultural opportunities associated with Lewisburg and nearby Bucknell University, while

valuing easy access to the scenic beauty of natural landscapes that were within a short drive or bike

ride outside of town. An informant claimed that the town has a …

… magic mix of accessibility from farmland to downtown.

Most descriptions of the landscape focused on its rural quality and general natural beauty. Informants

Table 4.3: Summary of key informant data for East Buffalo Township

Themes Frequency Coverage

Values 138 35.8%

Town & Country 36 12.2%

Differing Values 24 10.0%

Benefits 12 2.3%

Development 23 10.4%

Education & Culture 12 5.0%

Sense of Community 12 3.4%

Scenic Beauty 12 2.1%

Flooding Concerns 7 2.6%

Priorities 51 19.0%

Energy 25 10.4%

Landscape 26 8.7%

46

had a strong personal connection to the surrounding scenic landscapes, particularly the forests,

agriculture, and adjacent Susquehanna River. Few specific examples were offered other than passive

bike- or walk-friendly options. Regarding the river, one informant claimed:

Susquehanna River is at the heart of people’s identity.

Others claimed a “love / hate relationship” with the river due to recent flooding events and future

concerns. While the scenic lands appeared to garner the most positive attention, the state of the river

drew the most attention regarding priorities. The town had significant floodplain issues related to

the river that affected 30% to 40% of the local housing stock. Actions taken by these householders

had spiraling effects throughout the area such as changes in the tax base and districting for the local

schools.

There appeared to be a socio-geographic divide in East Buffalo that drove divergent values

and priorities. Town residents were generally regarded as highly educated, liberal affiliates of

Bucknell University, whereas residents outside the town were perceived as less educated and deeply

conservative. The difference in values between these social groups played out in housing, commercial

development, and recreation planning conflicts with strong social stigmatism overtones. One informant

recalled a conflict concerning where to locate a work force housing project that rapidly devolved into

an “ugly and racist dialogue” and community backlash towards “those kinds of people.” The housing

project was stalled while town supervisors were voted out of office. The social divide between town

and country was exemplified in such conflicts where town residents were accused of being intelligent

but out of touch with reality and country residents were stereotyped as reactionaries unwilling to

change or plan for the future.

Community

East Buffalo Township has a strong sense of community despite differing values between the

town and country. Residents were often described as willing to “pitch in and volunteer” and highly

likely to participate in community events. Informants listed several popular community events such

as the Arts Festival, Lewisburg Live, and organized volunteer cleanings of the river and rail trail. The

social network of East Buffalo was perceived to be small, personal, and yet easy to permeate for new

residents in the area. While there was a perception of safety in East Buffalo, there was also a concern

that an influx of new people to the area would threaten neighborliness, community, and safety.

47

Energy

Energy issues reflected similar sentiments about the assumed socioeconomic effects of change.

The recent natural gas boom has driven energy dialogue in East Buffalo. While there was no natural

gas mining within East Buffalo, activity in adjacent townships has both encouraged residents’ hope

for economic windfalls and raised concerns about a potential sociodemographic shift in the town that

did not align with the town’s sense of community. Levels of concern or awareness people expressed

about energy issues appeared to be related to the town and country divide. One informant described

the ideological energy divide as:

It depends on what circle you’re in.

Overall, current low energy costs minimized energy concerns. Both residents and local businesses

primarily conceptualized energy concerns as financial decisions. Local businesses were interested in

reducing their energy demand and appeared to be engaged in an open dialogue of successes, failures,

and recommendations. These actions were a result of efforts to lower operating costs rather than

focusing on environmental values. Residents viewed natural gas as an opportunity for cheap and

independent energy that would have a direct effect on their personal finances.

Oh! Natural gas will be cheaper. I’ll do that.

Everyone can agree on lowering their energy bill.

Residents did not appear to consider a bigger picture or long-term effects of mining, particularly

regarding its effects on the environment. Energy decisions were uniformly financial decisions, not

environmental considerations.

While we’re waiting for a clean economy, let’s continue to use fossil fuels.

Let’s get the money out while we can.

Similar to other prevalent themes in East Buffalo, energy was surrounded by discordant social attitudes

that reinforced the town and country divide. Natural gas activity promoted concerns about an invasion

of low-income workers, that one informant described as:

They’re not from here.

48

Residents also worried that the next cycle of natural gas mining activity would drive a housing

bust in the township. There was also apprehension that natural gas activity would introduce wealthy

executives who would influence key community dimensions, such as decisions about the local high

school.

4.1.3 Comparison of Spring and East Buffalo townships interviews

Spring and East Buffalo townships shared an appreciation for their small town and rural

environment and easy access to natural landscapes, particularly agricultural lands. Residents and local

businesses of both towns viewed energy as primarily an economic factor and acted to minimize costs,

but did not consider environmental impacts of energy decisions. Community engagement in Spring

and East Buffalo townships was professed to be very active, with each town offering many community

events and volunteering opportunities. Both towns also enjoyed opportunities for employment, arts,

and culture from nearby local universities.

Despite surface appearances, there was one key difference between the two towns’ community-

level interactions across landscape and energy dimensions. While the values and priorities of landscape

and energy issues in each township appeared to be relatively similar, there were significant community

conflicts surrounding each issue that were unique to East Buffalo Township. Whereas Spring Township

residents quietly enjoyed their connections between small town and natural landscapes, East Buffalo

Township’s town and country settings were a flashpoint for conflict. Such divisive attitudes were

attached to changes in the landscape, especially natural gas mining or new urban development,

which were perceived to be associated with sociodemographic changes to the town that threatened its

community fabric.

4.1.4 Results of member checking for Spring and East Buffalo townships

Three KIs replied to a brief summary of the interviews with additional feedback and comments.

One KI in Spring Township felt that in comparison to dialogue about the landscape, energy issues were

overly generalized and lacked specific connections to township. A KI from East Buffalo Township

noted that sprawl-based land use conflicts were not limited to the loss of farmland and rural character,

but also included negative economic impacts to the Lewisburg downtown area. A second KI from East

Buffalo suggested the conflict regarding where to relocate the town’s high school was largely a result

49

of a small, yet outspoken group. This KI also suggested the loss of agricultural land was a common

misperception among town residents. The KI further clarified that the township was divided into

ten zoning districts and development only occurred in approved zoning areas. Notably, this did not

contradict the notion that agricultural land was being lost, only that development was occurring where

it was planned to occur. The KI also drew a distinction between Lewisburg residents, who placed a

high value on the river and were impacted by its flooding, and East Buffalo residents who were largely

unaffected by the state of the river.

4.2 Summary of spatial data

The land cover mix within one-half mile of respondents’ households in Spring and East Buffalo

townships was relatively similar with several exceptions (Table 4.4). The predominant land cover

types were low-density development, forest, and agriculture. Approximately one-percent or less of

the average respondent’s environment contained other land cover types. Households in developed

areas contained a similar mix of intensity. However, in East Buffalo, the land cover mix contained

more developed land (~16%) at the expense of forest and agricultural land. Most of the difference in

developed land was developed open space that typically consisted of very low density residential and/

or parks or similar spaces.

4.3 Survey and spatial data results

The data suggested there were factors that influenced a person’s energy consumption behaviors

in a myriad of different ways across levels and types of interactions with the physical and social

environment. A noteworthy non-response bias in the sample limits the degree to which findings may

be generalized about each township.

Table 4.4: Summary of land cover data

Variable SpringTownship EastBuffaloTownship Difference Average

Developed 52.7% 69.0% -16.3% 63.2%

Open Space 17.5% 34.6% -17.1% 28.5%

Low Intensity 26.8% 24.5% 2.4% 25.3%

Med Intensity 6.9% 7.4% -0.5% 7.2%

High Intensity 1.5% 2.5% -1.0% 2.1%

Forest 19.1% 11.7% 7.5% 14.3%

Agriculture 27.1% 18.5% 8.5% 21.6%

Other 1.1% 0.8% 0.3% 0.9%

50

4.3.1 Summary of survey response rates

The survey generated 107 responses with an average response rate of 13.7% and an average cost

of $16.38 per response. Four separate methods of distributing the survey were attempted with varying

results (see Table 4.5). The initial hybrid postcard method resulted in few responses, especially given

the disproportionately large cost per response. Subsequent methods fared better at a drastically reduced

cost per response. Notably, the success of the survey method varied between townships. The Facebook

advertisement and online survey produced the most results at the least cost. While the KI Convenience

method consisted of email requests and required no direct costs (such as those associated with printing,

postage, or online services), indirect costs associated with time and travel to build a trusted network of

KIs were considerable. The Facebook advertisement proved to be a cost-effective method of gathering

survey responses, though the more expensive Drop-off, pick-up method was more productive on

average. In both townships, the KI Convenience method was an effective method that drew upon the

social network built during the KI interviews.

Table 4.5: Survey response rates by modality

HybridPostcard KIConvenience Drop-off/Pick-up Facebook Summary

5-Sep 19-Sep 2-Oct 19-Oct

Spring Township

No. of Responses 7 8 9 17 41

Response Rate 1.2% NA 18.4% 20.7% 13.4%

Cost per response $50.14 * $28.83 $5.88 $21.00

East Buffalo Township

No. of Responses 12 15 27 12 66

Response Rate 2.0% NA 27.6% 12.4% 14.0%

Cost per response $31.92 * $9.61 $8.33 $13.53

Total

No. of Responses 19 23 36 29 107

Response Rate 1.6% NA 23.0% 16.6% 13.7%

Cost per response $38.63 * $22.75 $6.90 $16.38

*See description of indirect costs in preceding paragraph

51

4.3.2 Description of variables

The initial survey provided over 150 individual items. Given the small sample size, it was not

possible to analyze all items. Moreover, the purpose of this study was to demonstrate the validity of

a community landscape framework of behavior by casting a wide net of physical and social items

that was grounded in the relevant literature. Factor analysis and a statistical description of items

illuminated which items were likely to influence behavior and reduced the number of tested variables.

Table 4.6 lists the dependent, determinant, and control variables:

Table 4.6: Names and definitions of explanatory variables

Variable Name Description Sources

Modality

Township Dummy variable controlling for township

Survey mode Categorical variable controlling for survey distribution type (postcard/online, KI convenience, drop-off / pick-up, and Facebook advertisement)

Physical environment

Care for township Eisenhauer et al. (2000)

Feeling that town is unique Eisenhauer et al. (2000)

Willingness to move from township Eisenhauer et al. (2000)

Farmland Respondent’s value of farmland Key informant interviews

Forest Respondent’s value of forest lands Key informant interviews

Water quality Respondent’s value of water quality Key informant interviews

Green space Respondent’s value of green space (i.e. parks, recreation, etc.) Key informant interviews

Mountains Respondent’s value of mountains Key informant interviews

New development Respondent’s value of new development Key informant interviews

Sense of community in township Respondent’s sense of community in their township Key informant interviews

Small town character Respondent’s value of small town character Key informant interviews

Arts and culture Respondent’s value of access to arts and culture Key informant interviews

Open space land cover % developed open space NLCD (2011)

Low-intensity land cover % developed low intensity NLCD (2011)

Medium-intensity land cover % developed medium intensity NLCD (2011)

High-intensity land cover % developed high intensity NLCD (2011)

Developed land cover % combined developed types NLCD (2011)

Forest land cover % combined forest types NLCD (2011)

Agriculture land cover % combined agriculture types NLCD (2011)

Other land cover % combined other types NLCD (2011)

52

Social environment

Family interactions Frequency of engagements with family Barret et al. (2015)

Close friends interactions Frequency of engagements with close friends Barret et al. (2015)

Acquintances interactions Frequency of engagements with acquintances Barret et al. (2015)

Neighbor interactions Frequency of engagements with neighbors Barret et al. (2015)

Community groups interactions Frequency of engagementswith community groups Barret et al. (2015)

Other social groups interactions Frequency of engagements with other social groups Barret et al. (2015)

Volunteering Number of hours volunteering Barret et al. (2015)

Past community actions SUM of CELA1, CELA2, CELA3, and CELA4

Proposed land development actions Summary of past actions regarding proposed land development Theodori & Luloff (2015)

Far or forest preservation actions

Summary of past actions regarding preserving farm or forest lands Theodori & Luloff (2015)

Natural gas or water quality actions

Summary of past actions regarding either natural gas mining (East Buffalo) or water quality (Spring) Theodori & Luloff (2015)

Cost of energy actions Summary of past actions regarding cost of energy Theodori & Luloff (2015)

Intended community actions MEAN of CELI1, CELI2, CELI3, and CELI4

Proposed land development intentions

Summary of intention to act regarding proposed land development Theodori & Luloff (2015)

Far or forest preservation intentions Summary of intention to act regarding preserving farm or forest lands Theodori & Luloff (2015)

Natural gas or water quality intentions

Summary of intention to act regarding either natural gas mining (East Buffalo) or water quality (Spring) Theodori & Luloff (2015)

Cost of energy intentions Summary of intention to act regarding cost of energy Theodori & Luloff (2015)

Energy behaviors

Past conservation actions SUM of past behaviors

Future conservation intentions MEAN of future intentions

Searched in a library for ways to conserve energy in my home.

LEED for Homes (2010)

Searched the Internet for ways to conserve energy in my home.

LEED for Homes (2010)

Met with an expert about ways you can conserve energy in your home.

LEED for Homes (2010)

Purchased an ENERGY STAR (or similar) certified appliance. LEED for Homes (2010)

Installed energy efficient windows. LEED for Homes (2010)

Purchased energy efficient light bulbs. LEED for Homes (2010)

Opened the doors and windows instead of turning on the air conditioner.

LEED for Homes (2010)

53

Planted trees or other vegetation to provide shade for your home.

LEED for Homes (2010)

Purchased a hybrid vehicle for personal use.Invited a professional to audit my home's energy consumption.

LEED for Homes (2010)

Installed or have a professional install new or additional insulation to my home

LEED for Homes (2010)

Participated in an energy rebate program. LEED for Homes (2010)

Participated in a green electricity program. LEED for Homes (2010)

Estimated utility payments Estimated month's electricity and gas utility payment

Utility consumption Utility reading for electricity and gas

Control variables

Value of energy conservation MEAN of energy values*

Perceived community barriers MEAN of perceived barriers related to community perceptions*

Perceived barriers MEAN of all other perceived barriers*

Shade coverage of home LEED for Homes (2010)

Number of ENERGY STAR appliances in home LEED for Homes (2010)

Home ownership Own or rent

Home typology Single family detached, mobile home, duplex, townshouse, or appartment

Home size Square footage of home LEED for Homes (2010)

Number of bedrooms

Home age LEED for Homes (2010)

Number of years residing in current community Austin et al. (2015)

Age Austin et al. (2015)

Gender Austin et al. (2015)

Marital status Austin et al. (2015)

Number of people in household Austin et al. (2015)

Under age 18 Austin et al. (2015)

Over age 65 Austin et al. (2015)

Race/ethnicity Austin et al. (2015)

Political affliation Austin et al. (2015)

Education Austin et al. (2015)

Employment status Austin et al. (2015)

Income Austin et al. (2015)*see Table 4.7 and Table 4.8 for factor analysis of items

54

Survey question batteries related to energy values and perceived barriers, which contained a

combined 18 individual items, were reduced to a smaller number of grouped components through a

factor analysis. This procedure helps reduce the aggregate number of variables included in further

analyses. A factor value of greater than 0.5 was used to suggest a strong alignment with a component.

Energy items collapsed with strong alignment into a single factor (see Table 4.7). Three items (battery

questions number 2, 4, and 10; see Appendix C) from the perceived barriers battery did not align with

any factor and were excluded from further analysis. Once removed, the remaining items aligned with

two factors: one related to perceptions of community barriers and one containing all other barriers3

(see Table 4.8). Community barriers described a person’s feeling of distance between their feelings

and actions about energy conservation and their perception of their community’s feelings and actions.

Other barriers referred to a person’s feelings about energy conservation in general.

The survey captured a wide range of data across spatial and conceptual constructs. The following

describes the data in terms of relative hierarchy, centrality, and variance. Energy behaviors represent

either specific actions or utilities consumed. Due the differences in calculation of each behavior, the

average values of each is generally not meaningful. However, the data does indicate that, on average,

respondents estimated they spent more on utilities than they actually consumed. Three of the four

energy behaviors – past conservation actions, estimated utility payments, and utility consumption -

displayed significant variance (Table 4.9). Future conservation intentions displayed measurably lower

variance.

Few physical environment variables displayed strong variance, especially among those related to

land cover (see Table 4.10). However, there appeared to be meaningful practical differences in the land

cover mix of households. A standard deviation of 0.10 to 0.30 indicated a 10% to 30% difference in

land cover. For example, a household environment composed of 60% developed land was essentially

a different environment than one composed of 30% developed land. In some cases, such as forest and

agriculture land cover, such differences might double or triple the acreage of forest or agriculture land

in an area. The average local environment of respondents heavily favored developed lands over forest

or agriculture. Practically this makes sense as most of the surveys were collected in neighborhoods

throughout the townships.3 Perceivedbarrieritemswerealsosubjectedtoalternativerotations(Varimax,Quartimax,Equamax,andobliquerotations)inthefactoranalysis.Theresultsproducedtwocomponentsconsistingof2-4itemseach.Thesecomponentsweretestedforcorrelationwithdependentvariables.Theresultsdidnotdifferfromthosepresentedinsection4.3.3

55

Table 4.8: Factoring of perceived barriers in two components

Variables Perceivedbarriers

Perceivedcommunity

barriers

Where I live prevents me from conserving energy in my home. 0.70 0.37

Recent changes in my town discourage me from conserving energy in my home. 0.55 0.40

I do not have the money to make changes to conserve energy in my home. 0.58 -0.40

Even if I had a good reason, I do not know how to conserve more energy in my home. 0.70 -0.08

I want to conserve more energy in my home, but other things are more important. 0.67 -0.24

I do not have the time to make changes to conserve energy in my home. 0.74 -0.15

I do not have the time to make changes to conserve energy in my home. 0.40 0.56

I make more of an effort to conserve energy in my home than most people in my community. -0.35 0.78

Energy conservation is not a priority for most people in my community. 0.20 0.64

Extraction Method: Principal Component Analysis.

Table 4.7: Factoring of energy values into a single component

Variables Valueofenergyconservation

I care deeply about how much energy I consume. 0.88

The source of my energy matters to me. 0.86

I do not think about how much energy my home consumes. 0.73

I care about the environmental impacts of energy consumption. 0.68

I do not support investments in solar or other alternative energy sources. 0.61

I consider myself to be an energy conservationist. 0.79

Extraction Method: Principal Component Analysis.

Overall respondents agreed with most environmental values. Most variables averaged 4.0 or

higher. The most notable exception being support for new development, which key informants

reported is an oft-contentious issue. Respondents appeared to differ on how they valued their

willingness to move from their town and how they valued green space, water quality, and new

development. Two of these variables, willingness to move and new development, shared relatively low

averages and high variance, suggesting division among respondents feelings towards these topics. The

Table 4.9: Descriptive statistics – energy behaviors (dependent variables)

Variable N Mean Std.Deviation Variance

Past conservation actions 78 6.4 2.5 6.1

Future conservation intentions 84 2.9 0.9 0.9

Estimated utility payments 65 6.E+09 3.E+09 1.E+19

Utility consumption 53 4.E+09 3.E+09 7.E+18

56

other significantly varying variables, water quality and green space were among the highest valued

physical environment factors. (see Appendix D for boxplots of key variables).

In contrast to physical environment factors, the social environment factors exhibited wider ranging

averages and higher variance across nearly all variables. Past community actions and volunteering

both indicated particularly strong differences among responses. Of the 16 possible actions respondents

may have taken towards community issues provided in the survey, respondents took an average of 4.7

actions, and as few as zero or as many as nine. More significant were the differences in volunteering.

Respondents averaged ten hours per month, though most ranged from zero to 22 hours. That translated

to a real difference of zero to five hours per week of time committed to volunteering. The ways in

which people interacted in social networks also varied, especially in how often respondents engaged

with family and neighbors. As with energy behaviors, intended community actions displayed low

variance.

Given the low sample size, variance was a key measure in determining which control items to

include in further analytical models. Items with low variance (less than 1.0) expressed no practical

difference among the collected responses and were not likely to explain differences in energy

behaviors among respondents (see Table 4.10 and 4.11). Such items were removed from the analysis

unless otherwise noted. The following housing and sociodemographic variables were used as

controls for other potentially moderating factors in the households’ environments. In general, there

was little variance across housing characteristics. However, as expected, home size and age varied

significantly and both were seen as critical factors in energy consumption. Several sociodemographic

variables showed meaningful variance, including number of years residing in the respondent’s current

community, age, number of people in household, political affiliation, education, and income. These

variables and the two housing characteristic variables were included in further analysis. While value of

energy conservation, and perceived barriers, (both community and general) did not meet the variance

threshold, each represented a critical component of the community landscape framework and were also

included in the analysis as control variables.

Based upon this assessment, an analysis of determinant variables of energy conservation behaviors

included the following physical and social environmental factors: environmental values, land cover

mix, social networking, volunteering, past community actions and future community intentions. The

57

Table 4.10: Descriptive statistics - physical environment variable

Variable N Mean Std.Deviation Variance

Physical environment

Care for township 86 4.1 0.8 0.6

Feeling that town is unique 87 3.7 0.9 0.7

Willingness to move from township 86 3.6 1.0 1.1

Farmland 87 4.3 0.7 0.5

Forest 87 4.3 0.9 0.7

Water quality 87 4.4 1.3 1.6

Green space 87 4.3 1.2 1.5

Mountains 87 4.3 0.9 0.8

New development 87 2.9 1.0 1.0

Sense of community in township 87 4.0 0.8 0.6

Small town character 87 3.9 0.8 0.7

Arts and culture 87 4.1 0.8 0.6

Open space land cover 59 0.3 0.2 0.0

Low-intensity land cover 59 0.3 0.1 0.0

Medium-intensity land cover 59 0.1 0.1 0.0

High-intensity land cover 59 0.0 0.0 0.0

Developed land cover 59 0.6 0.3 0.1

Forest land cover 59 0.1 0.2 0.0

Agriculture land cover 59 0.2 0.2 0.0

Other land cover 59 0.0 0.0 0.0

Social environment

Family interactions 86 3.1 1.7 2.7

Close friends interactions 86 3.5 1.4 1.9

Acquintances interactions 86 2.3 1.3 1.7

Neighbor interactions 86 2.0 1.6 2.5

Community groups interactions 86 1.4 1.4 1.9

Other social groups interactions 86 2.2 1.5 2.3

Past community actions 81 4.7 3.9 15.6

Intended community actions 84 1.8 0.5 0.2

Volunteering 82 10.2 11.8 138.6

58

Table 4.11: Descriptive statistics - control variables

Variable N Mean Std.Deviation Variance

Personal values

Value of energy conservation 86 4.1 0.6 0.4

Perceived barriers

Perceived community barriers 86 2.4 0.6 0.4

Perceived barriers 86 2.7 0.6 0.4

Housing typology

Shade coverage of home 75 1.2 0.7 0.5

Number of ENERGY STAR appliances in home 75 1.8 0.8 0.6

Home ownership 75 2.0 0.0 0.0

Home typology 74 1.0 0.0 0.0

Home size 69 2130.5 973.4 947467.8

Number of bedrooms 74 3.4 0.7 0.5

Home age 73 44.5 25.8 667.7

Socio-demographic

Number of years residing in current community 73 20.0 14.8 219.0

Age 72 55.4 12.7 160.8

Gender 74 1.5 0.5 0.3

Marital status 73 4.6 1.0 1.1

Number of people in household 74 2.7 1.2 1.3

Under age 18 73 0.6 1.1 1.1

Over age 65 73 0.5 0.9 0.8

Race/ethnicity 73 1.0 0.2 0.0

Political affiliation 72 3.3 1.3 1.6

Education 74 5.0 1.2 1.4

Employment status 74 5.3 1.0 1.0

Income 67 6.1 1.4 2.0

59

Table 4.12: Results of Shapiro-Wilk Tests for Normality

Variable Statistic Significance df

Past conservation actions 0.96 0.14 48

Future conservation intentions 0.93 0.01 48

Estimated utility payments 0.91 0.00 48

Utility consumption 0.96 0.12 48

Table 4.13: Description of skewness and kurtosis of energy behaviors

Variable N MeanSkewness Kurtosis

Statistic Std. Error Statistic Std. Error

Past conservation actions 78 6.44 -0.25 0.27 -0.47 0.54

Future conservation intentions 84 2.94 0.15 0.26 -1.03 0.52

Estimated utility payments 65 5.77E+09 1.07 0.30 1.19 0.59Utility consumption 53 4.32E+09 0.36 0.33 -0.30 0.64

analysis also included value of energy conservation, perceived barriers, sociodemographic factors,

and the control variables home size and age. Other non-determinate variables which did not display

significant statistical or practical significance were removed from the analysis.

Tests for correlation between determinate variables and energy behaviors assume normality among

behaviors. Two energy behaviors, past conservation actions and utility consumption were normally

distributed while future conservation intentions and estimated utility payments did not illustrate

normality due to skewness and or kurtosis (see Table 4.12, 4.13 and Appendix D). Consequently, they

were removed from the analysis. The energy behavior variables past conservation actions and utility

consumption subsequently became the key dependent variables of the study.

Key determinant and dependent variables were tested for independence to determine whether

there were significant differences among responses among the four survey distribution modalities

and/or between the two townships. The results show there were differences between town and survey

modalities across all variables (see Table 4.14). Dummy variables for both town and survey modes

were included in the analysis to control for these differences.

60

Table 4.14: Test for difference among variables

Variable Township** Survey*

Dependent variables

Past conservation actions 0.38** 0.15*

Future conservation intentions 0.80** 0.30*

Estimated utility payments 0.01** 0.47*

Utility consumption 0.11** 0.02*

Physical environmentCare for township 0.34* 0.86*Feeling that town is unique 0.04* 0.79*Willingness to move from township 0.99* 0.28*Farmland 0.88* 0.40*Forest 0.82* 0.21*Water 0.23* 0.19*Green space 0.03* 0.24*Mountains 0.67* 0.54*New development 0.75* 0.32*Sense of community in township 0.11* 0.48*Small town character 0.43* 0.03*Arts and culture 0.01* 0.63*

Open space land cover 0.00*** 0.01***Low-intensity land cover 0.12*** 0.21***Medium-intensity land cover 0.08*** 0.22***High-intensity land cover 0.05*** 0.18***Developed land cover 0.09*** 0.98***Forest land cover 0.02*** 0.69***Agriculture land cover 0.52*** 0.99***Other land cover 0.83*** 0.51***

Social environmentFamily interactions 0.12* 0.01*Close friends interactions 0.10* 0.25*Acquintances interactions 0.03* 0.40*Neighbor interactions 0.43* 0.05*Community groups interactions 0.45* 0.88*Other social groups interactions 0.49* 0.69*Volunteering 0.71* 0.93*Past community actions** 0.38** 0.15**Intended community actions** 0.80** 0.30***Kruskal Wallis test, **Mann-Whitney test, ***Independent Samples Median test

61

4.3.3 Results of correlation analysis

The correlation analysis results illustrated significant relationships between both physical and

social environmental factors and energy behaviors. However, the results differed between energy

behaviors with respect to the qualities of the physical environment or the types of social / community

engagement that appeared to influence energy conservation. The role of the control variables,

other than housing characteristics, appeared to have had minimal influence. Appendix E lists a full

description of correlations between energy behaviors and all determinate variables. All significant

variables are discussed below (see Table 4.15). Positive effects associated with past conservation

actions represented more conservation actions and negative effects associated with utility consumption

indicated less energy consumed – each was an indicator of energy conservation. The strength of most

correlations ranged from -0.55 to 0.56, though most were significant at the 5% level.

Higher percentages of agriculture land cover also positively influenced energy conservation.

However, higher percentages of developed land cover appeared to discourage conservation.

Households in developed areas, especially medium intensity areas, were less likely to take actions to

conserve energy. This effect was consistent across intensity and was more significant when levels of

development intensity were combined. Absent and not significant was the lowest level of development

intensity (open space). Such land cover typically contained very low-density residential housing and

80% of the land area was vegetated (Anderson et. al. 1976).

A range of different types of social engagements positively influenced energy behavior. Spending

more time with close friends and community groups, as well increased hours of volunteering

encouraged energy conservation. Community engagement, including past actions and future intentions,

also positively influenced energy conservation. These variables represented a diverse series of actions

between people and their communities across levels. The significant types of interactions ranged from

personal networking among close friends to investment alongside a local society in township issues.

There were meaningful differences in how physical and social variables correlated with each

energy behavior. Past conservation actions was positively influenced by actionable behaviors that

either had occurred or were likely to occur in the future, such as social engagement with community

groups, volunteering, past community actions, and future community intentions. However, this may

not apply to households in developed areas, as they appeared to be less likely to take action. By

62

comparison, utility consumption appeared to be positively influenced by intangible factors, such

as care for place (and forestlands), relationships with close friends, and intention to engage with

community issues. The exception was past community level actions with respect to energy cost issues,

though the correlation effect was very low. The exception here may be specific to households in

medium intensity developed areas. Agricultural land cover appeared to be a strong indicator of energy

conservation in both behaviors.

In summary, it appeared as though intangible values about place, personal relationships, and strong

intentions tended to encourage general utility conservation, whereas social actions tended to encourage

specific energy conservation actions. Land cover mix significantly mattered. Households in developed

areas were less likely to conserve, while household in agricultural areas were more likely to conserve

in multiple ways.

A mix of effects was observed from housing characteristics, perceived barriers, and

sociodemographic controlling variables. The number of ENERGY STAR appliances in a home

positively influenced conservation in both types of behavior. This made logical sense as purchasing

an appliance was a specific action similar to those described in the past conservation actions survey

battery and an action that was expected to reduce utility consumption. Also expected was a strong

correlation between home size and utility consumption. However, home age was not significant.

Marital status was also significantly correlated to both behaviors, though in opposing ways. Married,

or other types of attachment within households, appeared to encourage conservation actions while

negatively influencing utility conservation.

The survey captured the priority of natural lands, energy conservation, and community relative

to a series of other priorities identified in the KI interviews (see Table 4.16). The apparent priorities

of respondents were somewhat incongruent from previous results. For example, beyond the expected

top priorities of family, safety, and employment, preservation of natural lands was, on average, the top

priority of respondents. The importance of natural lands would support previous results describing

the value of place, forest, and agricultural lands, especially their influence on energy behaviors. Yet,

energy conservation and community were of similarly moderate concern. This suggested respondents

were not cognizant of a potential link between place, community, and energy. Oddly, the lowest ranked

priority was urban land development. Engagement with community issues regarding proposed new

63

Table 4.15: Summary of correlation between energy behaviors and determinate and controlling variables

Variables Pastconservationactions Futureconservationintentions

Physical environment factors

Care for township -0.23 -0.38**

Feeling that town is unique -0.03 -0.37**

Willingness to move from township -0.33* 0.10

Low-intensity land cover -0.54* 0.22

Medium-intensity land cover -0.53* 0.39***

High-intensity land cover -0.55* 0.30

Developed land cover -0.44** 0.23

Forest land cover 0.22 0.04

Agriculture land cover 0.40** -0.36***

Social environment factors

Close friends interactions 0.31 -0.35*

Community groups interactions 0.39* -0.33

Volunteering 0.35* -0.09

Past community actions 0.35* -0.12

Cost of energy actions 0.30 -0.14**

Intended community actions 0.40** -0.43**

Proposed land development intentions 0.56* -0.45**

Far or forest preservation intentions 0.39* -0.44**

Cost of energy intentions 0.40** -0.45**

Controlling variables

Number of ENERGY STAR appliances in home 0.43* -0.4**

Home size -0.10 0.44***

Home age 0.11 -0.20

Perceived community barriers 0.33** 0.23

Marital status 0.30** 0.33**

* Significant at the 10% level, ** Significant at the 5% level, *** Significant at the 1% level

Controlling for survey modality, township, energy values, perceived barriers, home square footage, home age, and socio-demographic factors

64

Table 4.16: Priority rankings

Variables MEAN

Family 2.24

Safety 3.63

Employment 4.03

Preservation of natural lands & water 4.40

Cost of living 4.70

Energy conservation 5.49

Community 5.55

Access to arts & culture 7.09

Urban land development 7.74

Table 4.17: Average levels of community engagement

Variables MEAN

Past community actions - proposed land development actions 1.68

Intended community actions - proposed land development actions 2.01

Past community actions - farm or forest preservation actions 1.07

Intended community actions - farm or forest preservation actions 1.84

Past community actions - natural gas or water quality actions 1.10

Intended community actions - natural gas or water quality actions 1.84

Past community actions - cost of energy actions 0.89

Intended community actions - cost of energy actions 1.66

land development had the highest average levels of previous and intended engagement (see Table

4.17). It was possible that this item of the priorities question was too generic to capture real concern

about new development.

4.3.4 Results of regression analysis

A multivariate regression model illustrated the explanatory power of significantly correlated

variables for each behavior. The purpose of the block regression model was to eliminate variables

whose effects were confounded by other influences. To simplify the model, variable were tested

for collinearity (see Table 4.18 and Table 4.19). In cases were there was collinearity among two or

more variables, all except the variable with the highest correlation coefficient with the behavior were

65

removed from the model.

There was significant covariation between land cover types in both models. There was also

virtually no difference in coefficients among land cover types. As a result, all were removed from

both regression models. Past community actions and future community actions also co-varied. Based

on previous results, the intentions variable was more likely to influence behavior and the past actions

variable was removed from the model. Several variables displayed covariation; volunteering and

willingness to move from the town, and two types of place attachment. Willingness to move and

feeling that the township is unique were removed from the model.

Variables were grouped into four blocks according to the proposed community landscape

framework consisting of control variables, sociodemographic variables, determinate variables

(landscape and community factors), and perceived barriers. Variables that were significant (less than or

equal to the 10% level) in any of the blocks were included in a subsequent reduced regression model.

The reduced models presented a more clear description of the discrete effects of variables while also

increasing the statistical freedom of the model (number of degrees of freedom).

Past conservation actions (see Table 4.20) were influenced by number of ENERGY STAR

appliances, volunteering, and perceived barriers (both general and community). The most influential

indicator (0.37) appeared to be respondents who had previously purchased energy efficient appliances.

As expected, as respondents reported stronger feelings of barriers to energy conservation the number

of energy conservation actions decreased. However, a stronger sense of barriers that widen the social

gap between respondents and their community appeared to increase the number of energy conservation

actions.

Utility consumption (see Table 4.21) was influenced by home size, number of ENERGY STAR

appliances, and factors from both the physical and social environment. Both care for township and

future community intentions were associated with decreased energy consumption. To explore these

effects in more detail, a second reduced model that replaced future intentions with specific local

issues (proposed land development, farm or forest preservation, natural gas or water quality, and cost

of energy). Reduced Model 2 improved the explanatory power of the variables in multiple ways; 1)

the standardized coefficient of care for township decreased from -0.30 to -0.44, 2) the standardized

coefficient of future community intentions decreased from -0.23 to -0.38, and 3) the amount of

66

variation explained by the models increased from 0.42 to 0.54. These changes reflect substantial

improvements in explanatory power. However, the significance of the dummy variable for township

suggests that these effects likely pertain more towards East Buffalo Township than Spring Township.

Chapter 5 discusses the implications of these effects.

Table 4.18: Comparison of variables that significantly correlate to dependent variable EBA

Willingnesstomovefrom

township

Low-intensitylandcover

Medium-intensitylandcover

High-intensitylandcover

Agriculturelandcover

Communitygroups

interactionsVolunteering

Past communityactions

Low-intensity land cover -0.30*

Medium-intensity land cover 0.01 0.63***

High-intensity land cover 0.13 0.53*** 0.97***

Agriculture land cover 0.17 -0.69** -0.56*** -0.52***

Community groups interactions 0.11 -0.31* -0.12 -0.03 0.04

Volunteering -0.30* -0.07 -0.17 -0.28 0.18 0.21

Past community actions -0.23 0.14 -0.02 -0.01 -0.23 0.11 0.04

Intended community actions -0.26 -0.14 -0.18 -0.18 -0.02 0.29* 0.16 0.68***

df = 32

* Significant at the 10% level, ** Significant at the 5% level, *** Significant at the 1% level

Controlling for survey modality, township, energy values, perceived barriers, home square footage, home age, and socio-demographic factors

Table 4.19: Comparison of variables that significantly correlate to dependent variable EBJ

Carefortownship Feelingthattownisunique

Medium-intensitylandcover

Agriculturelandcover

Closefriendsinteractions

Pastcommunityactions

Feeling that town is unique 0.48***

Medium-intensity land cover 0.02 0.04

Agriculture land cover 0.24 0.21 -0.55***

Close friends interactions -0.15 0.08 -0.09 -0.09

Past community actions -0.18 -0.03 -0.02 -0.23 0.48***

Intended community actions 0.05 0.06 -0.18 -0.01 0.29* 0.68***

df = 33

* Significant at the 10% level, ** Significant at the 5% level, *** Significant at the 1% level

Controlling for survey modality, township, energy values, perceived barriers, home square footage, home age, and socio-demographic factors

67

Table 4.20: Comparison of multivariate regression models, given as standardized coefficients of variable ‘Past conservation actions’

Variables 1 2 3 4 ReducedModel

Control variables

Survey mode -0.15 -0.23* -0.24* -0.14 -0.06

Township -0.18 -0.19 -0.18 -0.19

Value of energy conservation 0.21 0.06 0.03 -0.09

Number of ENERGY STAR appliances in home 0.46*** 0.46*** 0.46*** 0.40*** 0.37***

Home size 0.04 0.14 0.03 -0.10

Home age 0.12 0.06 0.03 -0.04

Socio-demographic variables

Number of years residing in current community -0.20 -0.22 -0.18

Age -0.04 0.00 0.05

Number of people in household -0.09 -0.04 0.06

Political affliation 0.29 0.37** 0.31* 0.17

Education -0.27 -0.28 -0.21

Income 0.11 0.13 0.16

Determinate variables

Community groups interactions 0.01 0.02

Volunteering 0.27** 0.20 0.25**

Intended community actions 0.16 0.14

Perceived barriers

Perceived barriers -0.27** -0.20*

Perceived community barriers 0.27** 0.29***

R2 0.34 0.44 0.54 0.62 0.49

F value 4.18 2.76 3.06 3.57 8.54

df = 54

* Significant at the 10% level, ** Significant at the 5% level, *** Significant at the 1% level

68

Table 4.21: Comparison of multivariate regression models, given as standardized coefficients of variable ‘Utility consumption’

Variables 1 2 3 4 ReducedModel1

ReducedModel2

Control variables

Survey mode 0.11 0.07 0.12 0.20

Township 0.33** 0.36** 0.38** 0.32** 0.19 0.29**

Value of energy conservation -0.15 -0.07 -0.08 -0.21

Number of ENERGY STAR appliances in home -0.24 -0.3* -0.24 -0.21 -0.20 -0.23*

Home size 0.46*** 0.57*** 0.73*** 0.66*** 0.53*** 0.47***

Home age 0.09 0.07 0.15 0.08

Socio-demographic variables

Number of years residing in current community -0.02 -0.03 0.06

Age -0.17 -0.12 -0.09

Number of people in household 0.00 -0.01 0.05

Political affliation 0.06 0.15 0.14

Education -0.26 -0.18 -0.15

Income 0.09 0.09 0.08

Determinate variables

Care for township -0.36** -0.36** -0.30** -0.44***

Intended community actions -0.22 -0.26* -0.23*

Proposed land development intentions 0.28

Far or forest preservation intentions -0.27

Natural gas or water quality intentions 0.11

Cost of energy intentions -0.38**

Perceived barriers

Perceived barriers -0.02

Perceived community barriers 0.33** 0.23* 0.20

R2 0.36 0.41 0.53 0.59 0.42 0.54

F value 3.53 1.84 2.38 2.53 4.98 4.94df = 44 (Models 1-6)

df = 48 (Reduced Model 1)

df = 47 (reduced Model 2)

* Significant at the 10% level, ** Significant at the 5% level, *** Significant at the 1% level

69

4.3.5 Comparison of the sample across township

An underlying question of this study was what factors might explain how two towns can grow

in similar ways, yet have drastically different energy consumption trends. A test for independence

between township samples revealed that while there were significant differences in several variables,

there were no statistical or practical differences in energy consumption among household respondents

(Table 4.20). As a result, any differences in physical or social environmental factors were unlikely to

explain township energy trends.

4.3.6 Generalizability of the survey sample

Compared to The American Community Survey (2015), this survey successfully captured data

from a representative sample across household size, race and ethnicity, education, employment, and

income, with a few exceptions (see Table 4.23). The sample included a disproportionate number of

respondents with graduate or professional degrees and those who were working part-time. The sample

was also slightly skewed towards households with higher incomes. The survey was less successful

at capturing a representative sample across age and marital status. Respondents on average were 10

(Spring Township) to 20 (East Buffalo Township) years older than the U.S. Census median age for

each township. Married couples comprised approximately half of all resident in both towns. However,

this sample heavily skewed in favor of households with married couples. Nearly all (80% to 90%) of

respondents reported being married.

As a result of the statistically negligible differences in energy consumption between the two

townships, a skew towards higher than median ages of respondents, and a disproportionately high

percentage of married respondents, the sample neither explained difference in township behaviors nor

were the findings generalizable descriptions of the townships. However, results may be generalizable

to the aggregate population of both townships, given the caveat that they were likely to be most

applicable to a slightly older, more educated, and wealthier strata of the population.

70

Table 4.22: Summary of differences between townships among key variables

Variable Mann-Whittneytest*

MEANvalue

Spring Township East Buffalo Township

n = 16-30 n = 36-57

Past conservation actions 0.60 6.19 6.56

Utility consumption 0.12 3,490,735,235.88 4,708,803,705.62

Utility consumption / sqft** 0.77 2,291,188.55 2,336,549.29

Care for township 0.34 4.03 4.18

Feeling that town is unique 0.04 3.50 3.88

Willingness to move from township 0.99 3.59 3.54

Developed land cover 0.03 53% 69%

Agriculture land cover 0.04 27% 19%

Close friends interactions 0.10 3.13 3.64

Community groups interactions 0.45 1.53 1.39

Volunteering 0.71 8.74 10.89

Past community actions 0.38 4.24 4.94

Cost of energy actions 0.80 0.83 0.93

Intended community actions 0.51 1.78 1.81

Proposed land development intentions 0.34 1.92 2.06

Far or forest preservation intentions 0.86 1.83 1.84

Cost of energy intentions 0.14 1.53 1.73

Number of ENERGY STAR appliances in home 0.03 1.50 1.90

Home size 0.02 1,815.65 2,287.89

Home age 0.59 42.96 45.27

Perceived community barriers 0.24 2.62 2.78

Marital status 0.39 4.73 4.53 *Reject null hypothesis that samples are not different at 5% level**EBJ normalized by home square footage

71

Table 4.23: Socio-demographic summary of sample

VariableSpringTownship EastBuffaloTownship

Census Sample Census Sample

Total population /n 7,619 41 6,443 66

Number of years residing in current community NA 19.0 NA 20.9

Age 42.9 52.0 38.4 59.5

Gender Male 49.4 41.9 51.8 53.1

Female 50.5 58.1 48.1 46.9

Marital status Married 56.8 90.3 45.9 79.6

Living with partner NA 3.2 NA 8.2

Widowed 8.1 0.0 2.3 0.0

Divorced /Separated 13.4 6.5 7.1 4.1

Single 21.6 0.0 44.7 8.2

Number of people in household 2.4 2.6 2.3 2.7

Under age 18 25.1 16.1 25.6 18.8

Over age 65 13.1 16.1 6.2 18.8

Race/ethnicity White 96.0 96.8 94.7 97.9

Not white 4.0 3.2 5.3 2.1

Political affiliation Liberal NA 3.3 NA 33.3

Moderate Liberal NA 13.3 NA 25.0

Moderate NA 46.7 NA 16.7

Moderate Conservative NA 23.3 NA 18.8

Conservative NA 13.3 NA 6.3

Education Did not complete high school 13.8 0.0 6.7 0.0

High school or equivalent 95.3 12.9 95.4 0.0

Some college or post high school training NA 16.1 NA 8.2

Associate's or 2-year vocational degree NA 12.9 NA 10.2

Bachelor's degree 28.2 35.5 56.3 28.6

Graduate / professional degree 9.5 22.6 26.2 53.1

72

VariableSpringTownship EastBuffaloTownship

Census Sample Census Sample

Employment status Full-time 72.2 77.4 41.7 60.0

Part-time 15.4 3.2 35.0 6.0

Retired NA 19.4 NA 32.0

Non-employed 4.6 0.0 2.6 0.0

Student NA 0.0 NA 0.0

Homemaker NA 0.0 NA 2.0

Income Less than $15,000 3.9 0.0 5.5 0.0

$15,000 to $24,999 10.1 0.0 6.1 0.0

$25,000 to $34,999 12.2 3.7 8.2 2.3

$35,000 to $49,999 15.8 18.5 14.3 6.8

$50,000 to $74,999 23.4 25.9 15.1 15.9

$75,000 to $99,999 13.5 25.9 7.9 27.3

$100,000 to $149,999 13.2 14.8 21.1 22.7

$150,000 or more 7.0 11.1 21.7 25.0

All census data from The American Community Survey 2015 4

4 AmericanCommunitySurveydatawasusedinsteadofU.S.CensusdatabecausemorerecentdatawasavailableattheMCDscale.

73

Chapter 5: Discussion

5.1 Review of findings

The results of the key informant interviews, spatial data, and survey suggested a myriad of

landscape- and community-oriented factors affected energy behaviors. Significant landscape

factors included land cover types and placed-based descriptors, while community factors included

volunteering, spending time with friends and community organizations, past participation in local

issues, and intentions to participate in future local issues. Home square footage, previous purchases

of ENERGY STAR appliances, and perception of barriers were important controlling factors. There

was no evidence of influence from sociodemographic characteristics. In this chapter, these findings are

discussed in relation to the proposed model, hypotheses, and relevant literature.

5.2 Validity of the model framework

The community landscape model of pro-environmental behavior (See Chapter 2, Figure 1)

attempted to unify disparate areas of study by demonstrating key linkages across emergent fields.

I hypothesized that (1) landscape connections and (2) community engagement influenced energy

conservation behaviors in homes. The research questions posited by these claims were founded on the

theories that people were innately bonded to place (Jackson 1997, 1994, Grieder and Garkovich 1994,

Lewis 1979) and that community emerged from a caring among people that was rooted in a place and

produced intentions to act (Luloff 1998, Wilkinson 1970). Below I present the evidence relevant to

each hypothesis.

5.2.1 Evidence of landscape connections

Each of the three analytical methods – latent content analysis of interview data, correlation

analysis, and regression analysis – produced evidence describing a multifaceted relationship between

landscape and energy behaviors. A clear sense of place emerged from the key informants about each

township. The rural setting, town and country dynamic, and deep value and concern for agricultural

land were consistent themes in how informants conceptualized their township.

These themes were also observed in both the correlation and regression analyses. Land cover

meaningfully influenced energy behavior, but covariation among variables, mixed directionality, and

74

low significance (mostly at p>0.10) limited how effectively discrete findings could be interpreted.

Developed land cover types shared strong (significant at p>0.10) covariation in most cases. Also, the

directionality of developed land cover types and previous energy actions was the opposite of what was

expected: respondents in more intensely developed environments had previously undertaken fewer

energy conservation actions. One exception were respondents in environments with higher percentages

of medium intensity displayed lower energy utility consumption. It is possible that there were factors

inherent to urbanized environments that acted as motivators or barriers to conservation actions –

but none of the housing typology, sociodemographic, or spatial factors captured by the survey that

might illuminate such barriers, such as housing type, home ownership, or income, were significantly

correlated to either energy behavior. Agricultural land cover appeared to be positively related to

conservation behaviors, though covariation with developed land cover types potentially confounded

which land cover type actually influenced behaviors. While difficult to discern the exact influence of

land cover, the evidence presented here suggested that land cover, or possibly a combination of land

cover types, was linked to energy conservation behaviors. Whether a particular land cover increased or

decreased conservation was not clear.

Place-based descriptors provided a stronger link between landscape and energy behaviors. The

results suggested the degree to which people cared about a place influenced how they conserved

energy. Key informants profusely lauded the landscape qualities that gave their township meaning and

value. Survey responses illustrated the degree to which they valued (or cared for) their town increased

as utility consumption decreased in both the correlation and regression models. The perception of

the township as unique shared a similar relationship, but the effect dematerialized in the regression

analysis. However, care for the town and perception of uniqueness co-varied suggesting these concepts

might be tightly connected.

Landscape emerges over time as humans transpose their cultural actions upon the land (Lewis

1979, Jacobs 1992, Jackson 1994, 1997). It is problematic to draw generalities about how a group may

apply meaning or attach to a landscape (Trentelman 2011) – yet agreement across key informants’

description of the townships coupled with statistical evidence that people cared about the township

suggested convergence between data sources. Taken together, the data suggested this sample of

township residents cared about the same conceptualization of place and that it influenced their energy

75

conservation behaviors. Remarkably, the effect of caring for the township strengthened when the

community factors of intention to participate in energy-specific local issues replaced general local

issues in the regression model.

5.2.2 Evidence of community engagement

Community relies on people expressing caring through action. The emergence of community was

observed through participation in local issues and volunteering, each of which was linked to energy

conservation behaviors. In both townships, key informants articulated ways in which community

emerged through groups of people taking actions they believed benefited others and the town - though

there was a stark contrast in the tone of these actions. In Spring Township, there was agreement about

the value of agricultural land and concern regarding suburban type development. Further, in crisis

situations neighbors generally came together to lend each other assistance. In East Buffalo, land-based

issues appeared to have incited conflict between groups holding differing values. These conflicts

often played out in town meetings and were perceived by informants to be associated with strong

sociodemographic undertones. These examples (and others reported in Chapter 4) were important but

unsubstantiated observations that communities formed from deeply valued local issues. The survey

built upon these claims by capturing community-level actions as past participation in local issues and

intention to participate in future local issues.

Both past actions and future intentions correlated in the expected direction with energy behaviors.

People who had both engaged with their community in the past and intended to do so in the future

had also completed more energy conservation actions, while people with higher intentions reported

lower levels of energy utility consumption. The effects were consistent whether analyzed as a

summary of past actions and future intentions, or as individual issues (land development, forest

and farm preservation, water quality or natural gas mining, and energy costs). However, the effects

dematerialized in the regression analysis - except for the relationship between intentions and

utility consumption. As noted above, intention to participate in energy-specific issues increased the

explanatory power of caring for township by decreasing energy utility consumption. Notably, past

actions and future intentions co-varied for both types of behavior, which confounded which type of

engagement likely influenced conservation.

76

Key informants expressed a strong sense of community in both townships that manifested as

heavily-attended local events, such as the fire departments carnival and parade, Night Out at the

Park, farmer’s markets, and trail cleanups. These events were often organized and successful because

of large groups of volunteers. Volunteering was linked to energy conservation actions in both the

correlation and regression models. In both cases, respondents who volunteered a greater number of

hours each month also completed more energy conservation actions. Volunteering was a purposive

action with the pro-environmental goal of improving the quality of life for people or a place. Despite

representing a community-level action, volunteering did not co-vary with either past participation or

future intention to participate. Volunteering was, therefore, functionally unique from other types of

community participation.

Different types and levels of socializing were also considered as correlates for community

engagements. Socializing alone does not constitute community development, but it is a useful

metric in describing types and strengths of social bonds of a place. Levels of social interaction with

friends and with community organizations were significantly correlated to utility consumption and

conservation actions, respectively. Yet neither effect held in the regression analysis. This was not

surprising given the low level of significance (p>0.10) of both variables and the small sample size for

the correlation analysis.

5.2.3 Role of perceived barriers

Ajzen (1991) and the proposed community landscape model of pro-environmental behavior

describe that behavioral intentions activate actions unless mediated by barriers. The study accounted

for barriers in two ways; asking key informants about how people prioritized energy and survey

questions regarding barriers to energy conservation. With few exceptions, key informants stated

that energy was simply not a priority while the cost of energy remained perceivably low. Survey

respondents supported this claim by ranking energy conservation as a priority below the cost of living,

though the spread (5.49 and 4.70, respectively) was not large compared to other items (see Chapter 4,

Table 4.16). However, the perception of cheap home energy was a misconception. Between 2001 and

2016 the average price (cents per kilowatt-hour) of electricity rose more than 50% from $0.09 to $0.14.

Over the last year, prior to when key informant interviews were initiated (February 2016), the seasonal

flux in residential electricity prices had leveled out at a 10-year high point (See Figure 7). Despite

77

these facts, the perception that energy was cheap persisted. This implied that people did not distinguish

residential energy sources from all other types, such as natural gas, petroleum, etc. Likely, there were

other external mediating factors acting as conservation barriers that were driving perceptions of energy

costs that did not provide informative feedback to residents.

The role of perceived barriers followed the study’s framework by acting to mitigate energy

conservation behaviors. Surprisingly, some expected barriers acted as motivators. The survey data

analysis assembled various potential conservation barriers into two groups; the first summarized

feelings that conservation created a social gap between the respondent and their community, and

the second encompassed a combination of other potential barriers. The effects of barriers most

acutely pertained to energy conservation actions. As expected, increased perception of barriers was

related to decreased number of conservation actions in the regression model (not significant in the

correlation model). However, in both the correlation and regression model, perception of community

barriers increased as the number of conservation actions increased. The greater a respondent felt that

their conservation actions separated them from the community, the more actions they completed.

Respondents appeared to be motivated by their uniqueness in their community. This observation defied

the expectation that perception of social norms should motivate similar behaviors (Stern 2000, Ajzen

Figure 7: Average residential electricity prices in Pennsylvania from January 2001 to May 2016 (Energy Information Administration 2017)

78

1992, Lewin 1946). There was some inconsistency in this claim, as perceived community barriers

operated as expected towards energy utility consumption, though with low significance (p>0.10)

5.2.4 Evidence of an interactive field

The evidence presented here confirmed the study’s hypotheses. Its results demonstrated bonds

between landscape, community, and energy conservation behavior. Validation of the proposed model

required evidence that these elements were part of an interactive field. Below I discuss two findings

that consistently converged across all three (key informant, correlation, and regression) analytical

models: care for township and intention to participate in local issues, specifically energy cost issues as

proxies for landscape and community constructs.

The theoretical frameworks of Lewin (1946), Ajzen (1991), Wilkinson (1970), Lewis (1979),

and Greider and Garkovich (1994), among others, suggested interaction between constructs across

social and physical environment influenced behavioral intentions. Key informants claimed residents

of both townships cared deeply about the landscape qualities of their town, particularly those that

emphasized the small town or rural aesthetic of neighborliness and access to natural landscapes. Both

towns shared concern for the preservation of qualities. East Buffalo, in particular, had a history of

conflict regarding planning and development decisions that threatened residents’ conceptualization of

these qualities. People have taken action to attend and vocalize their values at town meetings or among

other social groups. It was under such circumstances, when people were acting upon their attachment

to a place, where community emerged. Consequently, the results of the survey observed that how

deeply a respondent cared about their township effected the ways in which they consumed energy in

their home. Respondents who cared more for their town consumed less energy. The same relationship

was shared by respondents who intended to participate in local issues – such as those described by

key informants. More importantly, the effect of each increased when the specificity of the local issues

changed. That change signified interaction between landscape, community, and an energy behavior.

5.3 Role of controls and sociodemographic factors

Three housing typology factors were included in the analysis: home size (in square feet), home

age, and amount of ENERGY STAR appliances in the home. Home age, despite the assumption

that age was closely linked to utilities consumption and high variance within the sample, was not

79

significant in either model. The other typology factors contributed to energy behaviors in expected

ways. Residents of larger homes consumed greater amounts of energy. Home size was the most

powerful effect in both the correlation and regression models except for land cover (which as

discussed previously, experienced substantial covariation). Homes with more ENERGY STAR

appliances had slightly lower amounts of consumption and their residents completed more energy

conservation actions. It was notable that the presence of energy efficient appliances was more

influential than all other motivators of energy conservation actions. Respondents who had previously

purchased energy efficient appliances were more likely to complete additional conservation actions.

This finding aligns with the maxim that the best predictor of future actions are past actions.

Sociodemographic factors demonstrated virtually no effect on energy behaviors. It was not

surprising to find little influence, even from typically pro-environmental factors such as age,

education, and political affiliation. Van Liere and Dunlap’s (1980) exploration of the social

dimensions of environmental concern produced little evidence that social factors consistently

motivated environmental values. Oddly, marital status correlated to increased energy conservation

actions and utility consumption. One explanation for this phenomena was that married couples might

be more likely to have children or others living within the same home. Larger households might

increase demand for utilities while encouraging additional conservation efforts to minimize their

cost. However, the number of people in the household was included in the correlation and regression

models as a control – but was not significant in either.

5.3 Role of conflict

An underlying question of this study was how two seemingly similar townships could exhibit

dramatically different energy consumption trends. However, the data collected illustrated few

meaningful differences (see Chapter 4, Table 4.22). Most importantly, energy behaviors did not differ

in statistical or practical ways between the two townships during the data collection period. Among

key variables that were included in the analysis, only feelings that the town was unique, the percent

of developed land, amount of ENERGY STAR appliances, and home size statistically differed. Yet,

across all community-oriented factors (close friends interactions, volunteering, past community

actions, and intended community actions) and most sense of place measures (care for township and

feeling that town is unique), East Buffalo Township reported higher values. East Buffalo Township

80

respondents cared more for their town and were more active in their community. This claim was also

supported by the significance of the township dummy variable in the regression model (see Chapter 4,

Table 4.21).

The ways in which key informants described town relationships may offer a possible explanation

for why East Buffalo Township respondents demonstrated these characteristics. Conflict between

social groups, particularly among those who identified with the town as opposed to the adjacent

country, was a common theme mentioned by East Buffalo Township’s key informants. Responses

about the types of landscapes that people valued were typically coupled to examples of recent debates

or social conflicts. Such examples included relocating the high school, commercial development,

and suburbanization of agricultural lands. In contrast, Spring Township informants shared many of

the same issues, though they were described as common concerns that did not produce community-

level actions. Informants suggested land use conflicts were aggravated by underlying concerns about

incoming residents changing the sociodemographic profile of the town. It was possible that the more

affluent residents of East Buffalo Township would be more concerned about the immigration of lower

social classes than the residents of Spring Township.

The concept of conflict implied something that tears at the community fabric. However, in East

Buffalo, conflict appeared to be the way community groups expressed and acted upon their divergent

values. While community development at the township scale may suffer because of these conflicts,

these actions may have galvanized smaller scale community groups. The survey did not capture

feelings or sense of conflict, and so there was no converging evidence to support such a claim. Yet

the information provided suggested conflict provided opportunities for community development

and an outlet for residents to express the care for their township - and measurable increases across

both dimensions of the community landscape model of pro-environmental behavior. Logically this

suggested that areas with community-level conflicts about the landscape may be especially primed to

motivate pro-environmental behaviors. At a minimum, the data described disparity among community

emergence at differing levels and scales.

5.5 Validity and reliability

The findings of the study appeared to support several claims regarding the motivators of pro-

environmental behavior. A relatively low sample size and response rates that were less than the

81

literature prescribed challenged the validity and reliability of these claims. Below I discuss the

ramifications of the study framework and the validity of results it provided.

The internal validity of the study was supported by convergence across the Nvivo analysis of

qualitative data, correlation model, and regression model. The study model framework was based on

achieving the benefits of a mixed model approach. Key informant interviews improved the internal

validity of a subsequent survey by adapting context sensitive questions for each study site. The survey

attempted to generalize findings of the interviews about a population while non-reactive spatial land

cover data attempted to remove bias or misconception inherent in survey responses. Claims made by

informants proved to be significant in two independent statistical models. Further, the block model

approach of the regression analysis objectively removed co-varying variables to isolate what each

variable was actually measuring in the subsequent reduced model.

An inconsistency within the study framework was that differences in township energy

consumption illustrated in the site selection model did not manifest in data collected by the survey.

Data from the Energy Information Administration (EIA 2011, 2001) showed a nearly 70% difference

in residential electricity consumption over a 10 year period between the two township. However, the

difference between household energy consumption was slight. The survey captured additional energy

sources (see Appendix C), but all were reported in negligibly small amounts. One possible explanation

was that electricity sales data were gathered between 2001 and 2011. While this study was initiated

in 2014, actual energy behavior data was not gathered until 2016. It was possible that during this gap

in data collection the differences in energy consumption between the two townships normalized in an

unobserved way. The survey captured alternative sources of energy consumption (fuel oil, bottled gas,

coal, and wood), but few respondents reported using these sources. The timing of the survey may have

impacted these results. The survey was distributed in September and October. It is unlikely that most

respondents would be relying on alternative fuel sources during relatively warm months. Results may

have differed if the survey was distributed during winter months.

The external validity of the study suffered from survey non-response bias. The ideal number

of surveys to achieve a 95% confidence level was approximately 300 surveys per township. The study

captured 107 surveys (41 in Spring Township and 66 in East Buffalo Township). Both the hybrid

postcard and the drop-off / pick-up distribution methods resulted in lower than expected response rates

82

(Alfred & Davis 2010, Kaplowitz et al. 2004, Steele et al. 2001). Millar and Dillman (2011) suggest

that while response rates to web surveys may be improved if accompanied by postal invitations,

offering incentives after completion of the survey (such as gift card drawings) minimally improves

responses by less than three-percent. Web surveys that offer cash tokens in advance of completing the

survey fare better – but this strategy was not feasible given the limited funding available.

Two additional environmental factors possibly explained the results. First, the data collection

period took place during a highly active period of the 2016 United States of America Presidential

Election. Pennsylvania was a battleground state in the election, demonstrated by frequent political

signage throughout both townships. Elections typically coincided with high volumes of polling. It

was possible residents were simply saturated with having been asked for their opinion. At least two

potential respondents skeptically asked whether this study was associated with the election or “the

government” before accepting the survey. Second, responses to the survey suggested the topic of

the study was not a priority for most respondents – let alone those who did not complete the survey.

The survey was advertised as “your community, your landscape, your energy.” However responses

indicated these ideas were not priorities (see Chapter 4, Table 4.17). The lack of interest in the topics

may be linked to the misconception of cheap energy discussed above.

The survey captured a fairly representative sociodemographic sample of each township, but

was biased towards those with higher incomes and education, and married households. A comparison

of the sociodemographic profile of each township to U.S. Census data (see Chapter 4, Table 4.23)

suggested that while the total number of surveys were lower than expected, the responses collected

match several key sociodemographic strata for each township. The survey sample captured similar

proportions across gender, education, income, race/ethnicity, and employment status of respondents

compared to the township population. However, the survey also captured a higher proportion sample

of households with married couples, with graduate degrees, and higher incomes.

Convergence across the study framework suggested the study was internally valid, but a significant

non-response bias and bias towards specific sociodemographic strata challenged how confidently

the results could be applied to each township. However, key informants stated in multiple ways how

residents of each township were actively building community through shared place-based values,

concerns, and in the case of East Buffalo, conflict. The results of the statistical models substantiated

83

the informants’ claims that the community landscape model of pro-environmental behavior effectively

described interactions between the motivators and barriers of energy behaviors in these two towns.

While the low number of responses (especially in Spring Township) discouraged the study’s ability to

deconstruct these effects separately in each town, there was evidence that the effects were occurring

in two different places. To further support the ecological validity of the findings, additional inquiry is

needed in additional locations.

84

Chapter 6: Summary & Future Questions

6.1 Summary of study framework and results

The Earth has a finite amount of natural resources. Currently there are significant challenges

with cost effective extraction of fossil fuels, land use requirements of renewable energy sources, and

technological limitations to converting fossil or renewable resources to consumable forms of energy

such as electricity. These challenges persist despite an increasing energy demand. For example,

residential electricity consumption in Pennsylvania increased by two and a half times the number of

new customers over a ten-year period (2001-2011). Conservation must be a part of any rationally-

proposed strategy to meet future energy demand.

Conservation is a pro-environmental behavior. An interactive field composed of a person’s

internal values and their external environment informs behavioral intention. This field shares similar

conceptual constructs with the fields of community and landscape. Community emerges when people

act to benefit each other and a place. Landscape is formed when people apply meaning to the land

through acts of creations informed by cultural values. Analogous isomorphs across the fields of

behavior, community, and landscape suggest a unified model describing behavior as a function of

emergent interactions between the local landscape and community. This study attempted to validate

such a model through two related inquiries: (1) does the landscape influence residential energy

conservation; and (2) does community engagement influence residential energy conservation – and if

so, what types and levels of each are relevant?

To address these questions, this study adopted a mixed-method framework consisting of spatial

analysis of potential sites, qualitative key informant interviews, quantitative survey questions, and

spatial analysis of local land cover mix. The framework triangulated data of multiple types from

different sources to gain an understanding of local place (internal validity) and extract representative

findings (external validity). Two Pennsylvania townships were selected from over 2,500 minor civil

divisions, commonly called towns, through an evaluation of changes in home growth, electricity

sales, and land cover between 2001 and 2011. Spring Township and East Buffalo Township were

selected because during this period they exhibited similar changes in land cover and home growth but

85

drastically different amounts of electricity consumption. Eighteen key informants from both townships

were interviewed to gain an understanding of the values, priorities, and perceptions of local landscapes

and energy consumption. A subsequent survey was distributed in four ways: a hybrid postcard and

online survey, through key informant contacts, a traditional drop-off, pick-up survey, and through

Facebook advertisements.

While each survey distribution method elicited responses, the independent and overall response

rate was lower than typical. These conditions informed an analytical strategy that focused on

identifying statistically important determinant variables and removing other indicators unlikely to

contribute to explaining variance in the model. The survey captured environmental values (or the ways

in which households valued their local landscape), various types and levels of community engagement,

perceived barriers associated with energy conservation, housing typologies, and sociodemographic

characteristics. Survey respondents reported two types of residential energy consumption (utility

readings and estimated utility costs) and two types of conservation (past conservation actions and

intention to complete future conservation actions).

The results of the analysis illustrated several ways in which landscape and community influenced

energy conservation. In the most clear example, a partial correlation and block model regression

analysis both demonstrated an interaction between higher levels of care for place, intention to

participate in community issues, and lower energy utility consumption. Key informants’ claims that

people cared deeply for their town and would participate in community events in ways they perceived

would benefit the town support this relationship. Further, the effects of the relationship between

place, community intentions, and consumption increased when the community issue specified energy

costs. These observations denoted interactions between elements of the landscape and community

that informed conservation energy behavior, thereby validating the proposed model. Other supporting

evidence included the significance of volunteering in increasing energy conservation actions and

confounded links (in the form of significant correlations with covariation among land cover types)

between land cover and energy behaviors.

6.2 Implications of findings

Low response rates and negligible differences in energy consumption reported by survey

respondents limited how the study’s findings might explain energy consumption trends in either

86

township. Despite this, these findings have both theoretical and practical implications. The proposed

community landscape model of pro-environmental behavior was essentially an evolution of Lewin

(1946) and Ajzen’s (1991) classic models. Demonstrable proof of the link between a person, their

environment, and their actions lends support to a deep volume of social, environmental, and behavioral

literature founded on their work. Likewise, the findings provided additional examples of sense of

place, place attachment, and community emergence through participation and conflict in rural settings.

Practically, the findings are useful in informing strategies for encouraging energy conservation.

They suggest that energy conservation campaigns might find success in programs that inspired

empathy with the township and encouraged participation in energy-centric activities. Further,

such programs may find greater success in targeted landscapes with prescribed land cover mixes;

however, more research is required to ascertain the role of land cover. Areas previously entangled in

social conflict because of the loss of farmland and forests to low-density residential or commercial

development may be particularly amendable to such programs. Additional research is also necessary

to demonstrate that the model is valid in studying alternative pro-environmental behaviors. For

example, Chapter 2 described a hypothetical application of the community landscape model for pro-

environmental behavior in coastal towns facing community-level conflicts regarding dune ecosystem

reconstruction. The findings outline potential starting points – place attachment, participation in local

issues, volunteering, and land cover - for identifying the motivators that lead to signing easements that

enabled dune reconstruction.

6.3 Implications for the field

Lewin (1946) called for researchers to adopt action-oriented agendas that directly engaged local

social issues. The community landscape model of pro-environmental behavior proposed here attempts

to answer Lewin’s call by linking the fields of physical design and social science. Such a bond

encourages designer and planners to engage both the human and physical dimensions of place toward a

holistic understanding of the problem. Such a transdisciplinary approach (1) deconstructs disciplinary

silos of knowledge and methodology by advancing a unified theory that bridges traditionally design-

oriented and social science oriented fields, and (2) advocates for the importance of capturing the non-

discipline biased values of local people in the design process.

Human-environment conflicts, particularly those involving natural resources, inherently result from

87

disagreement about how cultural values are applied in different places (Balint et al. 2011). Physical

clues to conflicts are readily visible in how landscapes form and change over time in response to

contemporary application of cultural values. How the landscape should change is often the considered

the domain of designers and planners, especially in areas where people live and work. Seminary work

by Jane Jacobs, J.B. Jackson, Pierce Lewis (and others) defined design and planning efforts as studies

of cultural activities in space. The continued feasibility of design and planning fields’ claim of a role in

addressing human-environment conflicts necessitates an evolved approach to studying the physical and

social dimensions of place.

The model presented by this study and the supportive findings illustrate the critical importance

of designers and planners in addressing human-environment conflicts. Landscape is not just an

isomorphic component across community and behavioral fields. Rather, the findings suggest that it

is the central catalyst from which community and pro-environmental actions emerge. Respondents’

care for place encouraged higher levels of community participation and subsequent conservation.

How people conceptualized and attached to their local landscape (though the exact dimensions were

unclear) was an essential component in activating positive physical and social actions. The findings

suggested designers and planners were well positioned to engage critical socio-environmental

issues. But to adequately do this they must adopt transdisciplinary methods to effectively gather and

decipher multi-dimensional information about a place and its people. Reluctance to do so risks an

over-emphasis on the role of managing physical capital at the expense of neglecting the role of human

values and community networks.

6.4 Future questions

The study addressed both proposed research questions and provided theoretical and practical

implications for the allied fields of study and practice. In addition, the study raised critical questions

for future research. First, the generalizability of the study suffered from low response rates, limiting

how confidently the findings might be applied. Applying the proposed model across differing socio-

physical geographies concerning different pro-environmental behaviors would support the ecological

validity of the model. Second, the study discovered broad relationships between place attachment,

sense of place, and behavior. Future studies should aim to decipher the specific ways in which peoples’

bonds with place influence their conservation actions. Similarly, future studies should seek to discern

88

the specific types and levels of community engagement that encourage conservation. This study

identifies one such example in volunteering, though likely there are many others to discover. Third,

though unsubstantiated by the survey, the key informant data indicated that conflict played a direct

role in encouraging community engagement and indirect role in inspiring conservation. Conflicted

areas may be primed for pro-environmental strategies and programs, but more research is required to

describe the productive conditions of conflict.

The purpose of the community landscape model of pro-environmental behavior is to inform how

designers, planners, and researchers may address severe environmental threats with critical human

dimensions. The evidence presented in this study describes the broad ways in which behavior is

influenced by bonds between people, landscape, and community. Further application of the proposed

model will explore its utility and test its validity in different geographic contexts as a tool that informs

strategies for resolving human-environment conflicts.

89

Bibliography

Abrahamse, W. & Steg, L. (2011) Factors related to household energy use and intention to reduce it: The role of psychological and socio-demographic variables. Human Ecology Review. 18(1), 30-40.

Anderson, James R., Hardy, Ernest E., Roach, John T. & Richard E. Witmer. (1976). A Land Use and Land Cover Classification System for use with Remote Sensor Data. United States Department of the Interior. Geological Survey Professional Paper. http://landcover.usgs.gov/pdf/anderson.pdf. Accessed December 21, 2016.

Agrawal, A., & Gibson, C. C. (1999). Enchantment and disenchantment: the role of community in natural resource conservation. World Development. 27(4), 629-649.

Ajzen, I. (1991). The theory of planned behavior.Organizational Behavior and Human Decision Process. 50, 179–211.

Allred, Shorba Broussard & Amy Ross-Davis. (2010). The drop-off and pick-up method: an approach to reduce non-response bias in natural resource surveys. Small-scale Forestry. 10. November 2010.

Balint, P., Stewart, R, Desai A., & Walters L.C. (2011). Wicked Environmental Problems: Managing Uncertainty and Conflict. Island Press. Washington, D.C.

Barrett, Austin, Lauren Abbott, Sarah Eissler, Carolyn Fish, Lacey Goldberg, Stephen Mainzer, & Max Olsen. (2015). Community perceptions of local food: a mixed methods approach in State College, PA. State College, PA. The Hamer Center for Community Design. The Pennsylvania State University.

Bender, Thomas. (1978). Community and Social Change in America. Rutgers Press University. New Brunswick, NJ.

Bridger, J. C., & Luloff, A. E. (1999). Toward an interactional approach to sustainable community development. Journal of Rural Studies, 15(4), 377-387.

Commonwealth Economics, LLC. (2013). Energy in Pennsylvania: Past, Present, and Future. Submitted to Commonwealth of Pennsylvania Department of Environmental Protection. February 2013.

Dietz, T., Stern, P. C., & Guagnano, G. A. (1998). Social structural and social psychological bases of environmental concern. Environment and Behavior. 30(4), 450-471.

Eisenhauer, B. W., Krannich R. S., & D. J. Blahna. (2000). Attachments to special places on public lands: An analysis of activities, reasons for attachments, and community connections. Society Natural Resources. 13, 421–441.

Electric Data Browser (http://www.eia.gov/electricity/data/browser/) Energy Information Administration, Washington, DC.

90

Elmendorf, William F, & A.E. Luloff. (2001). Using qualitative data collection methods when planning for community forests. Journal of Arboriculture. 27(3), 139-151.

Emery, Mary & Cornelia Flora. (2006). Spiraling-up: mapping community transformation with community capitals framework. Community Development. 37(1), 19-35.

Eser, Sezer Göncüoğlu & A.E. Luloff. (2003). Community controversy over a proposed limestone quarry. Society and Natural Resources. 16, 793-806.

Fey, Susan, Corry Bregendahl, & Cornelia Flora. (2006). The measurement of community capitals through research. Online Journal of Rural Research & Policy. 1(1), 1-28.

Flint, C.G., A.E. Luloff, & J.C. Finley. (2008). Where is “community” in community-based forestry? Society and Natural Resources. 21(6), 526-537.

Flint, C.G. & A.E. Luloff. (2007). Community activeness in response to forest disturbances in Alaska. Society and Natural Resources. 20. 431-450.

François, C. (2006). Transdisciplinary unified theory. Systems Research and Behavioral Science. 23(5), 617-624.

Gever, John, Kaufman, R., Skole, D., & Vörösmarty, C. (1986). Beyond Oil: The Threat to Food and Fuel in the Coming Decades. Ballinger Publishing Company. Cambridge, MA.

Greider, T. & L. Garkovich. (1994). Landscapes: the social construction of nature and the environment. Rural Sociology. 59(1), 1-24.

Gorte, Ross W., Vincent, Carol Hardy, Hanson, Laura A., & Marc R. Rosenblum. (2012). Federal Land Ownership: Overview and Data. Congressional Research Service. CRS Report for Congress. 7-5700. R42346. February 8, 2012.

Guagnano, G. A., Stern, P. C., & Dietz, T. (1995). Influences on attitude-behavior relationships a natural experiment with curbside recycling. Environment and behavior. 27(5), 699-718.

Hawley, Amos. (1971). Human Ecology: A Theory of Community. The Ronald Press Company. New York, NY.

Hay, Iain. (2000). Qualitative Research Methods in Human Geography. Oxford University Press.

Hunt, J.D. (1973). Natural resource use and the hierarchy of needs. The Journal of Environmental Education. 4(4), 20-21.

Hurricane Sandy Rebuilding Task Force. (2013). Hurricane Sandy Rebuilding Strategy.

Ibtissem, Mustapha H. (2010). Application of value beliefs norms theory to energy conservation behavior. Journal of Sustainable Development. 3(2). 129-139.

Iowa State University. (2008). Biomass Measurements and Conversions. www.extention.iastate.edu/agdm. File C6-88.

Jackson, J.B. (1994). A Sense of Place, a Sense of Time. Yale University. New Haven, CT.

91

Jackson, J.B. (1997). The future of the vernacular. Groth, Paul. (1997). Frameworks for Cultural Landscape Study. In Groth & Bressi (eds.) Understanding Ordinary Landscapes. Yale University. New Haven, CT. 145-156.

Jacobs, J. (1961). The Death And Life Of Great American Cities. Random House. New York, NY.

Janssen, Marco A., Therese Lindahl, & James L. Murphy. (2015) Advancing the understanding of behavior in social-ecological systems: results from lab and field experiments. Ecology and Society. 20(4).

Jones, L. & Tanner T. (2015). Measuring ‘subjective resilience’ using people’s perceptions to quantify household resilience. ODI Working Paper 423.https://www.odi.org/sites/odi.org.uk/files/odi-assets/publications-opinion-files/9753.pdf.

Kaplowitz, Michael D., Hadlock, Timothy D., & Ralph Levine. (2004). A comparison of web and mail survey response rates. American Association for Public Opinion Research. 68(1), 94-101.

Kaufman, H. F. (1959). Toward an interactional conception of community. Soc. F., 38, 8.

Keefer, Matthew J, Finley, James C., Luloff, A.E, & Marc E. McDill. (2003). Understanding loggers’ perceptions. Journal of Sustainable Forestry. 17(3), 81-102.

Kemmis, D. (1990). Community and Politics of Place. Norman: Cambridge University of Oklahoma Press.

Kollmuss, A. & Julian Agyeman. (2002). Mind the gap: why do people act environmentally and what are the barriers to pro-environmental behavior? Environmental Education Research, 8(3), 239-260.

Kroon, B. van der, Brouwer, R. & Beukering, P.J.H. (2014). The impact of the household decision environment on fuel choice behavior. Energy Economics. 44, 236-247.

Kudryavtsev, Alex, Richard C. Stedman, & Marianne E. Krasny. (2011). Sense of place in environmental education. Environmental Education Research. 18(2), 229-250.

Larson, E. C., Luloff, A. E., Bridger, J. C., & Brennan, M. A. (2015). Community as a mechanism for transcending wellbeing at the individual, social, and ecological levels. Community Development. 46(4), 407-419.

Larson, Lincoln R., Richard C. Stedman, Caren B. Cooper, & Daniel J. Decker. (2105). Understanding the multi-dimensional structure of pro-environmental behavior. Journal of Environmental Psychology. 4, 112-124.

Lazarus, R. & Folkman, S. (1987). Transactional theory and research on emotions and coping. European Journal of Personality. 1. 141-169.

Leach, Melissa, Robin Mearns, & Ian Scoones. (1999). Environmental entitlements: dynamics, institutions in community-based natural resource management. World Development. 27(2), 225-247.

Lewin, Kurt. (1935). A Dynamic Theory of Personality. New York, NY, US: McGraw-Hill. 286 pp.

92

Lewin, K. (1946). Action research and minority problems. Journal of Social Issues. 2, 34–36.

Lewis, P. F. (1979). Axioms for reading the landscape: some guides to the American scene. The interpretation of ordinary landscapes: Geographical essays, 11-32.

Luloff, A.E. (1998). What Makes a Place Community? St. John Quick Bendigo Lecture. La Trobe University, Bendigo.

Luloff, A.E. (1999). The doing of rural community development research. Rural Society. 9(1), 313-337

McDonald, Robert I., Farigone, Joseph, Joe, Keisecker, Millar, William M., & Jimmie Powell. (2009). Energy sprawl or energy efficiency: climate change policy impacts on natural habitat for the United States of America. PLoS One. 4(8). doi:10.1371/journal.pone.0006802.

Manfredo, M., J. Vaske, D.R. Field, & P. Brown (eds.) (2004). The concept of community in natural resource management. Society and Natural Resources: A Summary of Knolwedge. Jefferson, MO: Modern Litho; (Luloff, A.E., R.S. Krannich, G.L. Theodori, C.K. Trentelman, and T. Williams). 249-259.

McHarg, I. L., & Mumford, L. (1969). Design With Nature. New York: American Museum of Natural History.

McHarg, I. L. & Steiner, F (eds.) (1998). To Heal the Earth: Selected Writing of Ian L. McHarg. Island Press. Washington, D.C.

Millar, Morgan M. & Don A. Dillman (2011). Improving response to web and mixed-mode surveys. Public Opinion Quarterly. 72(2). 249-269.

Mitrany, M., & Stokols, D. (2005). Gauging the transdisciplinary qualities and outcomes of doctoral training programs. Journal of Planning Education and Research. 24(4), 437-449.

Moran, Emilio F. (2006). People and Nature: An Introduction to Human Ecological Relations. Blackwell Publishing. Malden, MA.

National Broadband Map. http://www.broadbandmap.gov/. Accessed June 30, 2014.

Ostrom, Elinor. (2009). A general framework for analyzing sustainability of social-ecological systems. Science. 325(24), 419-422.

Outka, Uma (2011). The renewable energy footprint. Stanford Environmental Journal. 30. 241-308.

Park, J., & Ha, S. (2012). Understanding pro-environmental behavior. International Journal of Retail & Distribution Management. 40(5), 388-403.

Pennsylvania Alternative Energy Credit Program. http://www.pennaeps.com/aboutaeps/. Accessed December 19, 2016.

Pimentel, D., Whitecraft, M., Scott, Z. R., Zhao, L., Satkiewicz, P., Scott, T. J., & Moe, T. L. (2010). Will limited land, water, and energy control human population numbers in the future? Human Ecology. 38(5), 599-611.

93

Stedman, Richard C. (1999). Sense of place as an indicator of community sustainability. The Forestry Chronicle. 75(5), 765-770).

Stedman, Richard C. (2002). Toward a social psychology of place: predicting behavior from place-based cognitions, attitudes, and identity. Environment and Behavior. 34(5), 561-581.

Stedman, Richard C. (2003). Is it really just a social construction?: The contribution of the physical environment to sense of place literature. Society and Natural Resources. 16, 671-685.

Stedman, Richard C. (2016). Subjectivity and social-ecological systems: a rigidity trap (and sense of place as a way out). Sustainability Science. August. 1-11.

Steele, Jennifer, Bourke, Lisa, Luloff, A.E., Liao, Pei-shan, Theodori, Gene L., & Richard S. Krannich. (2001). The drop-off/pick-up method for household survey research. Journal of Community Development Society. 32(2), 238-248.

Steinitz, Carl. (2012). A Framework for Geodesign: Changing Geography by Design. ESRI Press. Redlands, CA.

Stern, Paul C. (2000). Toward a coherent theory of environmentally significant behavior. Journal of Social Issues. 56(3). 407-424.

Stokols, D. (1977). Origins and directions of environment-behavioral research. In Perspectives On Environment And Behavior. Springer US. 5-36.

Stokols, D. (2006). Toward a science of transdisciplinary action research. American Journal Of Community Psychology. 38(1-2), 63-77.

Stokols, D. (2011). Transdisciplinary action research in landscape architecture and planning. Landscape Journal. 30, 1-11.

Theodori, G.L. & Luloff, A.E. (2015). Perceptions of oil and natural gas development in the Eage Ford Shale: A summary of findings from a 2015 survey. Houston Advanced Research Center. October 2015.

Theodori, G. L., & Luloff, A. E. (2000). Urbanization and community attachment in rural areas. Society & Natural Resources. 13(5), 399-420.

Theodori, G. L. (2008). Constraints to the development of community. Community Development. 39(2), 91-110.

Trentelman, Carla K. (2009). Place attachment and community attachment: a primer grounded in the lived experience of a community sociologist. Society and Natural Resources. 22, 191-210.

Trentelman, Carla K. (2011). Place dynamics in a mixed amenity place: Great Salt Lake, Utah. Human Ecology Review. 18(2), 126-138.

Tuan, Yi-Fu. (1977). Space and Place: The Perspective of Experience. University of Minnesota Press, Minnesota.

United State Green Building Council. (2010). LEED ® for Homes Rating System. January 1, 2010.

94

U.S. Census Bureau. (2010). 2010 Census Urban and Rural Classification and Urban Area Criteria. https://www.census.gov/geo/reference/ua/urban-rural-2010.html. Accessed December 19, 2016.

Van Liere, Kent D. & Riley E. Dunlap. (1980). The social bases of environmental concern: a review of hypotheses, explanations, and empirical evidence. Public Opinion Quarterly. 44(1). 181-197.

Wackernagel, Mathis & Rees, William. (1996). Our Ecological Footprint: Reducing Human Impact on the Earth. New Society Publishers. Gabriola Island, B.C. Canada.

Wackernagel, M., Onisto, L., Bello, P., Callejas Linares, A., Susana LópezFalfán, I., Méndez Garcı́a, J. & Guadalupe Suárez Guerrero, M. (1999). National natural capital accounting with the ecological footprint concept. Ecological Economics. 29(3). 375-390.

Wahl, D. (2014). Exploring Pathways to Transformations in Post-Disaster-Event Communities: A Case Study on the Mad River Valley, Vermont. MS Thesis, Stockholm University.

Wall, Karen. (2016). Long-Awaited Dune Project Should Begin Next Spring, Army Corps Says. Patch Media. June 24, 2016.

Wilkinson, Kenneth P. (1970). The community as a social field. Social Forces. 48(3). 311-322.

Wilkinson, Kenneth P. (1991). The Community in Rural America. First Social Ecology Press. Middleton, WI.

Washington University. (2005). Energy in Natural Processes and Human Consumption – Some Numbers. ENVIR215. Spring 2005. http://www.ocean.washington.edu/courses/envir215/energynumbers.pdf.

Zawadzki, Stephanie J., Schwartz, Forrest G., Blair, Jordan C.E., Larson, Eric C., & Jennifer N. Newton. (2016). Understanding recycling while tailgating: applying an information-motives-behavior skills approach. The Contributions of Social Sciences to Sustainable Development at Universities. World Sustainability Series. 289-305.

Zernike, Kate. (2013). Trying to Shame Dune Holdouts at Jersey Shore. New York Times. New York. September 4, 2013.

Zube, E. H., Sell, J. L., & Taylor, J. G. (1982). Landscape perception: research, application and theory. Landscape Planning. 9(1). 1-33.

95

Appendix A: Key Informant Guide

Key Informant Interview Guide

Name: _____________________________ Title: ______________________________

Interview Location: ___________________ Date of Interview: ____________________

Start time: __________________________ Stop time: __________________________

Thank you for agreeing to be a part of this research study.

The research study focuses on how people value their community, where they live and

work, and the people they interact with in their community. I have an interest in exploring how

places influence a person’s willingness to act pro-environmentally, or how willing they are to

take actions that benefit their environment.

This interview is an important part of my dissertation research in partial fulfillment of a

PhD in Architecture from the Pennsylvania State University. Your time and effort is greatly

appreciated.

Your participation in this interview is completely voluntary; we can stop at any time and

you may feel free to skip any questions you would prefer not to answer. Your participation in the

interview implies your consent to be part of this research study.

Your responses and your personal identity will remain confidential in the dissertation. All

information, data, and notes acquired from this interview will be securely stored, per Internal

Review Board for human subjects (STUDY00003702) requirements.

I would be happy to share my notes from this interview with you, if you would like to

clarify your responses or offer additional comments and/or guidance.

Any questions? Ready to begin?

96

1. What makes [TOWN NAME] a great place to live and work?

[If interviewee replies places, locations, landscapes, or natural resources then ask about

community activities, social groups, events, etc.]

[If interviewee replies with social features, then ask about places, landscapes, natural

resources, etc.]

2. What would people miss most about [TOWN NAME] if it were to change or disappear?

3. What has changed since you have lived/worked in [TOWN NAME]?

[If physical, then ask about social changes.]

[If social, then ask about places and landscape/natural resource changes.]

4. Could that have been avoided? Why or why not?

5. What do you think [TOWN NAME] values the most?

6. When it comes to landscape/natural resources of [TOWN NAME], what do people care about?

a. What evidence is there of that?

7. How much of a priority are environmental or natural resource issues in [TOWN NAME]?

a. What are some examples?

8. When it comes to energy conservation, what do people care about?

a. What evidence is there of that?

9. How much of a priority are energy conservation issues in [TOWN NAME]?

a. What are some examples?

10. Are there others you think I should speak with?

97

Appendix B: Survey Postcard

98

WHAT IS THIS?This questionnaire is an important part of a study that seeks to understand how Pennsylvanians value their landscape, communities, and energy resources. Your participation is greatly appreciated.

HOW LONG WILL IT TAKE?The questionnaire will require approximately 15 minutes to complete. Participation is strictly voluntary and you may refuse to participate at any time. If you choose to participate, please answer all questions as honestly and accurately as possible. BEFORE YOU BEGIN:Several questions will refer to your AUGUST utility bill(s) for electricity and gas. It may be helpful to have these bills accessible before you begin the survey.

Stuckeman School of Architecture & Landscape Architecture229 Stuckeman Family Building

University Park, PA 16802

Your community. Your landscape. Your energy.

Spring Township

To be eligible for a random drawing for a $25 gift card to Barnes & Noble, you must answer ALL questions.A representative of Penn State’s Stuckeman School will return on Tuesday, October 4th to collect the survey.

Appendix C: Drop-off, Pick-up Survey Material

99

2

1. Please indicate your level of agreement with each of the following statements about SPRING TOWNSHIP.Please check one for each statement.

2. Please indicate your level of agreement with each of the following statements about energy use. Please check one for each statement.

Strongly disagree Disagree Neither agree

nor disagree Agree Strongly agree

I care deeply about Spring Township. � � � � �Spring Township is a unique place to live. � � � � �If I could, I would move away from Spring Township. � � � � �Preservation of farmland is important to me. � � � � �Being able to recreate in the forest is important to me. � � � � �Water quality is not important to me. � � � � �Access to green space (i.g. parks) is not important to me. � � � � �Being able to see the mountains is important to me. � � � � �I do not like seeing new development in Spring Township. � � � � �Having a sense of community in Spring Township is important to me. � � � � �I like that Spring Township feels like a small town. � � � � �I enjoy access to arts and culture. � � � � �

Strongly disagree Disagree Neither agree

nor disagree Agree Strongly agree

I care deeply about how much energy I consume. � � � � �The source of my energy matters to me. � � � � �I do not think about how much energy my home consumes. � � � � �I care about the environmental impacts of energy consumption. � � � � �I do not support investments in solar or other alternative energy sources. � � � � �I consider myself to be an energy conservationist. � � � � �

100

3

6. Please rank the following issues from MOST important (#1) to LEAST important (#9)

5. Which of the following community events have you attended?Please check all that apply.

Each of the following questions explore how you interact with your community. Please read each question and indicate your answer in the space provided. 3. How often do you get together or meet with the following types of people? Please check one for each statement.

4. On average, how many hours a month do you spend on community or volunteer activities? _______ number of hours

Urban land development _____Access to arts & culture _____Preservation of natural lands and water _____Community _____Safety _____Energy conservation _____Employment _____Family _____Cost of living _____

� The annual Pleasant Gap Fire Company Carnival

� The annual Easter Egg Hunt

� A farmers' market (if checked, which one?) ______________________� Participated in a group bicycle ride, kayaking, or hike.

� Other community event (please include the name of the event) ______________________� None

Never A few times a year Once a month A few times a

monthMore than

once a week Once a week

Family � � � � � �

Close friends � � � � � �

Acquaintances � � � � � �

Neighbors � � � � � �Community councils or groups � � � � � �Other social groups or organizations � � � � � �

101

4

Below is a list of actions you may or may not have taken in response to ENVIRONMENTAL ISSUES in or near your community. 7. In PART I, please indicate whether or not you have engaged in such an action. In PART II, please indicate your

likelihood of doing so in the future.Please check the responses that best describe your answers.

Part IHave you?

Part IIHow likely are you to do this in the future?

No Yes Not likely Somewhat likely Very likely

Attended a public meeting to get information and learn more about PROPOSED LAND DEVELOPMENTS. � � � � �Contacted a local elected official or governmental agency about PROPOSED LAND DEVELOPMENTS. � � � � �Met with a community group or organization to talk about PROPOSED LAND DEVELOPMENTS. � � � � �Talked to friends or neighbors about PROPOSED LAND DEVELOPMENTS. � � � � �Attended a public meeting to get information and learn more about PRESERVING FARM OR FOREST LANDS. � � � � �Contacted a local elected official or governmental agency about PRESERVING FARM OR FOREST LANDS. � � � � �Met with a community group or organization to talk about PRESERVING FARM OR FOREST LANDS. � � � � �Talked to friends or neighbors about PRESERVING FARM OR FOREST LANDS. � � � � �Attended a public meeting to get information and learn more about WATER QUALITY. � � � � �Contacted a local elected official or governmental agency about WATER QUALITY. � � � � �Met with a community group or organization to talk about WATER QUALITY. � � � � �

Talked to friends or neighbors about WATER QUALITY. � � � � �Attended a public meeting to get information and learn more about COST OF ENERGY. � � � � �Contacted a local elected official or governmental agency about COST OF ENERGY. � � � � �Met with a community group or organization to talk about COST OF ENERGY. � � � � �

Talked to friends or neighbors about COST OF ENERGY. � � � � �

102

5

The next few questions ask about general issues regarding your community and energy conservation.8. Please state your level of agreement with each of the following statements regarding your community.

Please check one for each statement.Strongly disagree Disagree Neither agree

nor disagree Agree Strongly agree

Where I live prevents me from conserving energy in my home. � � � � �

Living in my town enables me to conserve as much energy as possible. � � � � �

Recent changes in my town discourage me from conserving energy in my home. � � � � �

There are many ways to conserve more energy in my home. � � � � �

I do not have the money to make changes to conserve energy in my home. � � � � �Even if I had a good reason, I do not know how to conserve more energy in my home.

� � � � �I want to conserve more energy in my home, but other things are more important.

� � � � �

I do not have the time to make changes to conserve energy in my home. � � � � �

Attempting to conserve energy would make me an outsider in my community. � � � � �

I do not see evidence of energy conservation in my community. � � � � �I make more of an effort to conserve energy in my home than most people in my community.

� � � � �

Energy conservation is not a priority for most people in my community. � � � � �

9. How do you typically commute to work? Please check all that apply.

10. How long is your commute to work in a typical week?

� Car � Bicycle

� Carpool � Walk

� Public transportation � Other ____________________________

Number of MILES ____________________

Number of MINUTES ____________________

103

6

Utility gas $ ____________________

Electricity $ ____________________

Now, we would like to ask some questions about your household’s use of energy.11. For each of the following sources of energy, please use your AUGUST utility bill to report the amount of energy consumed.

Please enter the amount consumed for each fuel source OR “0” if none was consumed.

13. For each of the following sources of energy, please report the amount consumed within the month of AUGUST. Please enter the amount consumed for each fuel source OR “0” if none was consumed.

14. Do you currently use any of the following renewable systems to supply electricity to your home? Please check all that apply and please provide the total amount of electricity consumed.

12. For each of the following sources of energy, please report or estimate the amount in dollars from your AUGUST utility bill.

Example:

Utility gas in hundred cubic feet ____________________ CCF

Electricity in kilowatt-hours ____________________ KWH

Fuel oil (kerosene, etc.) in gallons ____________________ gal.

Bottled tank or gas in pounds ____________________ lbs.

Coal in pounds ____________________ lbs.

Wood in number of cords ____________________ cords

Other ____________________

Solar in kilowatt-hours ____________________ KWH

Wind in kilowatt-hours ____________________ KWH

Hydro in kilowatt-hours ____________________ KWH

Other in kilowatt-hours ____________________ KWH

104

7

Now, we would like to ask some questions about your home and household. Please check one for each question.16. How often is your home covered by shade from surrounding structures or vegetation?

17. How many of your major household appliances (such as your refrigerator, dishwasher, laundry, heating and cooling system, or television) are ENERGY STAR certified?

� Never

� Less than half a day

� More than half a day

� Always

� None

� Some

� Most

� All

15. Below is a list of actions you may or may not have taken in response to energy conservation in your home. In PART I, please indicate whether or not you have engaged in such an action. In PART II, please indicate your likeli-hood of doing so in the future. Please check the responses that best describe your answers.

Part IHave you?

Part IIHow likely are you to do this in the future?

No Yes Not likely Somewhat likely Very likely

Searched in a library for ways to conserve energy in your home. � � � � �Searched the Internet for ways to conserve energy in your home. � � � � �Met with an expert about ways you can conserve energy in your home. � � � � �Purchased an ENERGY STAR (or similar) certified appliance. � � � � �Installed energy efficient windows � � � � �Purchased energy efficient light bulbs. � � � � �Opened the doors and windows instead of turning on the air conditioner. � � � � �Planted trees or other vegetation to provide shade for your home. � � � � �Purchased a hybrid vehicle for personal use. � � � � �Invited a professional to audit your home’s energy consumption. � � � � �Installed or had a professional install new or additional insulation to your house. � � � � �Participated in an energy rebate program. � � � � �Participated in a green electricity program. � � � � �

105

8

This information, as with all information provided in this survey, will be used only for statistical analysis and will remain strictly confidential.

23. How long have you lived in your current community? ________________________ number of years

24. In what year where you born? (e.g. 1975) ________________________ year

25. What gender do you consider yourself to be?

26. What is your current marital status?

� Male

� Female

� Other

� Married � Divorced / Separated

� Living with partner (but not married) � Single

� Widowed � Other ________________________

27. For the following questions, please enter the number of people.

How many people live in your household, including yourself? ____________________ number of people

How many members of your household are under the age of 18? ____________________ number of people

How many members of your household are over the age of 65? ____________________ number of people

19. What type of household is your primary residence?

20. Approximately how large is your home in square feet (sqft)? ________________________ sqft

21. How many bedrooms are in your home? ________________________ number of bedrooms

22. In what year was your home constructed? (e.g. 1996) ________________________ year

� Single detached home

� Mobile home

� Duplex

� Townhome

� Apartment

18. Do you own or rent your home?

� Own

� Rent

� Other

106

9

30. What is the highest level of education that you have completed?

31. What is your current employment status?

28. What race/ethnicity do you consider yourself?

29. How do you describe yourself politically?

� Did not complete high school � Associate or 2-year vocational degree

� High school or equivalent � Bachelor’s degree

� Some college or post high school training � Graduate / professional degree

� Full-time � Non-employed

� Part-time � Student

� Retired � Homemaker

� American Indian � Hispanic

� Asian � White (non Hispanic)

� Black or African American � Other ________________________

Liberal Moderate Liberal Moderate Moderate Conservative Conservative

� � � � �

32. Which of the following categories best describes your total 2015 household income from all sources before taxes?

� Less than $15,000

� $15,000 to $24,999

� $25,000 to $34,999

� $35,000 to $49,999

� $50,000 to $74,999

� $75,000 to $99,999

� $100,000 to $149,999

� $150,000 or more

This is the final section of the survey. This information (and all other information provided) will only be used for spatial analysis and will remain strictly confidential.33. What is your address?

Number and street ________________________________________

City ________________________________________

State ________________________________________

Zipcode ________________________________________

107

10

You have now completed the survey. THANK YOU very much for your time and effort!

Please feel free to use this space to provide any additional comments.

108

WHAT IS THIS?This questionnaire is an important part of a study that seeks to understand how Pennsylvanians value their landscape, communities, and energy resources. Your participation is greatly appreciated.

HOW LONG WILL IT TAKE?The questionnaire will require approximately 15 minutes to complete. Participation is strictly voluntary and you may refuse to participate at any time. If you choose to participate, please answer all questions as honestly and accurately as possible. BEFORE YOU BEGIN:Several questions will refer to your AUGUST utility bill(s) for electricity and gas. It may be helpful to have these bills accessible before you begin the survey.

Stuckeman School of Architecture & Landscape Architecture229 Stuckeman Family Building

University Park, PA 16802

Your community. Your landscape. Your energy.

East Buffalo Township

To be eligible for a random drawing for a $25 gift card to Barnes & Noble, you must answer ALL questions.A representative of Penn State’s Stuckeman School will return on Tuesday, October 4th to collect the survey.

109

2

1. Please indicate your level of agreement with each of the following statements about EAST BUFFALO TOWNSHIP.Please check one for each statement.

2. Please indicate your level of agreement with each of the following statements about energy use. Please check one for each statement.

Strongly disagree Disagree Neither agree

nor disagree Agree Strongly agree

I care deeply about East Buffalo Township. � � � � �East Buffalo Township is a unique place to live. � � � � �If I could, I would move away from East Buffalo Township. � � � � �Preservation of farmland is important to me. � � � � �Being able to recreate in the forest is important to me. � � � � �Water quality is not important to me. � � � � �Access to green space (i.g. parks) is not important to me. � � � � �Being able to see the mountains is important to me. � � � � �I do not like seeing new development in East Buffalo Township. � � � � �Having a sense of community in East Buffalo Township is important to me. � � � � �I like that East Buffalo Township feels like a small town. � � � � �I enjoy access to arts and culture. � � � � �

Strongly disagree Disagree Neither agree

nor disagree Agree Strongly agree

I care deeply about how much energy I consume. � � � � �The source of my energy matters to me. � � � � �I do not think about how much energy my home consumes. � � � � �I care about the environmental impacts of energy consumption. � � � � �I do not support investments in solar or other alternative energy sources. � � � � �I consider myself to be an energy conservationist. � � � � �

110

3

6. Please rank the following issues from MOST important (#1) to LEAST important (#9)

5. Which of the following community events have you attended?Please check all that apply.

Each of the following questions explore how you interact with your community. Please read each question and indicate your answer in the space provided. 3. How often do you get together or meet with the following types of people? Please check one for each statement.

4. On average, how many hours a month do you spend on community or volunteer activities? _______ number of hours

Urban land development _____Access to arts & culture _____Preservation of natural lands and water _____Community _____Safety _____Energy conservation _____Employment _____Family _____Cost of living _____

� National Night Out in a park.

� Helped clean up a park or trail.

� A farmers' market (if checked, which one?) ______________________� Participated in a group bicycle ride, kayaking, or hike.

� Other community event (please include the name of the event) ______________________� None

Never A few times a year Once a month A few times a

monthMore than

once a week Once a week

Family � � � � � �

Close friends � � � � � �

Acquaintances � � � � � �

Neighbors � � � � � �Community councils or groups � � � � � �Other social groups or organizations � � � � � �

111

4

Below is a list of actions you may or may not have taken in response to ENVIRONMENTAL ISSUES in or near your community. 7. In PART I, please indicate whether or not you have engaged in such an action. In PART II, please indicate your

likelihood of doing so in the future.Please check the responses that best describe your answers.

Part IHave you?

Part IIHow likely are you to do this in the future?

No Yes Not likely Somewhat likely Very likely

Attended a public meeting to get information and learn more about PROPOSED LAND DEVELOPMENTS. � � � � �Contacted a local elected official or governmental agency about PROPOSED LAND DEVELOPMENTS. � � � � �Met with a community group or organization to talk about PROPOSED LAND DEVELOPMENTS. � � � � �Talked to friends or neighbors about PROPOSED LAND DEVELOPMENTS. � � � � �Attended a public meeting to get information and learn more about PRESERVING FARM OR FOREST LANDS. � � � � �Contacted a local elected official or governmental agency about PRESERVING FARM OR FOREST LANDS. � � � � �Met with a community group or organization to talk about PRESERVING FARM OR FOREST LANDS. � � � � �Talked to friends or neighbors about PRESERVING FARM OR FOREST LANDS. � � � � �Attended a public meeting to get information and learn more about NATURAL GAS MINING. � � � � �Contacted a local elected official or governmental agency about NATURAL GAS MINING. � � � � �Met with a community group or organization to talk about NATURAL GAS MINING. � � � � �Talked to friends or neighbors about NATURAL GAS MINING. � � � � �Attended a public meeting to get information and learn more about COST OF ENERGY. � � � � �Contacted a local elected official or governmental agency about COST OF ENERGY. � � � � �Met with a community group or organization to talk about COST OF ENERGY. � � � � �

Talked to friends or neighbors about COST OF ENERGY. � � � � �

112

5

The next few questions ask about general issues regarding your community and energy conservation.8. Please state your level of agreement with each of the following statements regarding your community.

Please check one for each statement.Strongly disagree Disagree Neither agree

nor disagree Agree Strongly agree

Where I live prevents me from conserving energy in my home. � � � � �

Living in my town enables me to conserve as much energy as possible. � � � � �

Recent changes in my town discourage me from conserving energy in my home. � � � � �

There are many ways to conserve more energy in my home. � � � � �

I do not have the money to make changes to conserve energy in my home. � � � � �Even if I had a good reason, I do not know how to conserve more energy in my home.

� � � � �I want to conserve more energy in my home, but other things are more important.

� � � � �

I do not have the time to make changes to conserve energy in my home. � � � � �

Attempting to conserve energy would make me an outsider in my community. � � � � �

I do not see evidence of energy conservation in my community. � � � � �I make more of an effort to conserve energy in my home than most people in my community.

� � � � �

Energy conservation is not a priority for most people in my community. � � � � �

9. How do you typically commute to work? Please check all that apply.

10. How long is your commute to work in a typical week?

� Car � Bicycle

� Carpool � Walk

� Public transportation � Other ____________________________

Number of MILES ____________________

Number of MINUTES ____________________

113

6

Utility gas $ ____________________

Electricity $ ____________________

Now, we would like to ask some questions about your household’s use of energy.11. For each of the following sources of energy, please use your AUGUST utility bill to report the amount of energy consumed.

Please enter the amount consumed for each fuel source OR “0” if none was consumed.

13. For each of the following sources of energy, please report the amount consumed within the month of AUGUST. Please enter the amount consumed for each fuel source OR “0” if none was consumed.

14. Do you currently use any of the following renewable systems to supply electricity to your home? Please check all that apply and please provide the total amount of electricity consumed.

12. For each of the following sources of energy, please report or estimate the amount in dollars from your AUGUST utility bill.

Example:

Utility gas in hundred cubic feet ____________________ CCF

Electricity in kilowatt-hours ____________________ KWH

Fuel oil (kerosene, etc.) in gallons ____________________ gal.

Bottled tank or gas in pounds ____________________ lbs.

Coal in pounds ____________________ lbs.

Wood in number of cords ____________________ cords

Other ____________________

Solar in kilowatt-hours ____________________ KWH

Wind in kilowatt-hours ____________________ KWH

Hydro in kilowatt-hours ____________________ KWH

Other in kilowatt-hours ____________________ KWH

114

7

Now, we would like to ask some questions about your home and household. Please check one for each question.16. How often is your home covered by shade from surrounding structures or vegetation?

17. How many of your major household appliances (such as your refrigerator, dishwasher, laundry, heating and cooling system, or television) are ENERGY STAR certified?

� Never

� Less than half a day

� More than half a day

� Always

� None

� Some

� Most

� All

15. Below is a list of actions you may or may not have taken in response to energy conservation in your home. In PART I, please indicate whether or not you have engaged in such an action. In PART II, please indicate your likeli-hood of doing so in the future. Please check the responses that best describe your answers.

Part IHave you?

Part IIHow likely are you to do this in the future?

No Yes Not likely Somewhat likely Very likely

Searched in a library for ways to conserve energy in your home. � � � � �Searched the Internet for ways to conserve energy in your home. � � � � �Met with an expert about ways you can conserve energy in your home. � � � � �Purchased an ENERGY STAR (or similar) certified appliance. � � � � �Installed energy efficient windows � � � � �Purchased energy efficient light bulbs. � � � � �Opened the doors and windows instead of turning on the air conditioner. � � � � �Planted trees or other vegetation to provide shade for your home. � � � � �Purchased a hybrid vehicle for personal use. � � � � �Invited a professional to audit your home’s energy consumption. � � � � �Installed or had a professional install new or additional insulation to your house. � � � � �Participated in an energy rebate program. � � � � �Participated in a green electricity program. � � � � �

115

8

This information, as with all information provided in this survey, will be used only for statistical analysis and will remain strictly confidential.

23. How long have you lived in your current community? ________________________ number of years

24. In what year where you born? (e.g. 1975) ________________________ year

25. What gender do you consider yourself to be?

26. What is your current marital status?

� Male

� Female

� Other

� Married � Divorced / Separated

� Living with partner (but not married) � Single

� Widowed � Other ________________________

27. For the following questions, please enter the number of people.

How many people live in your household, including yourself? ____________________ number of people

How many members of your household are under the age of 18? ____________________ number of people

How many members of your household are over the age of 65? ____________________ number of people

19. What type of household is your primary residence?

20. Approximately how large is your home in square feet (sqft)? ________________________ sqft

21. How many bedrooms are in your home? ________________________ number of bedrooms

22. In what year was your home constructed? (e.g. 1996) ________________________ year

� Single detached home

� Mobile home

� Duplex

� Townhome

� Apartment

18. Do you own or rent your home?

� Own

� Rent

� Other

116

9

30. What is the highest level of education that you have completed?

31. What is your current employment status?

28. What race/ethnicity do you consider yourself?

29. How do you describe yourself politically?

� Did not complete high school � Associate or 2-year vocational degree

� High school or equivalent � Bachelor’s degree

� Some college or post high school training � Graduate / professional degree

� Full-time � Non-employed

� Part-time � Student

� Retired � Homemaker

� American Indian � Hispanic

� Asian � White (non Hispanic)

� Black or African American � Other ________________________

Liberal Moderate Liberal Moderate Moderate Conservative Conservative

� � � � �

32. Which of the following categories best describes your total 2015 household income from all sources before taxes?

� Less than $15,000

� $15,000 to $24,999

� $25,000 to $34,999

� $35,000 to $49,999

� $50,000 to $74,999

� $75,000 to $99,999

� $100,000 to $149,999

� $150,000 or more

This is the final section of the survey. This information (and all other information provided) will only be used for spatial analysis and will remain strictly confidential.33. What is your address?

Number and street ________________________________________

City ________________________________________

State ________________________________________

Zipcode ________________________________________

117

10

You have now completed the survey. THANK YOU very much for your time and effort!

Please feel free to use this space to provide any additional comments.

118

WHAT IS THIS?A questionnaire was offered to you or left on your door on Sunday, October 2nd. The questionnaire is an important part of a study that seeks to understand how Pennsylvanians value their landscape, communities, and energy resources. Your participation is greatly appreciated.

Stuckeman School of Architecture & Landscape Architecture229 Stuckeman Family Building

University Park, PA 16802

Your community. Your landscape. Your energy.

FINAL REMINDER

To be eligible for a random drawing for a $25 gift card to Barnes & Noble, you must answer ALL questions.

A representative of Penn State’s Stuckeman School will return for a FINAL PICK-UP ON THURSDAY, OCTOBER 6TH to collect the survey.

Please place the completed questionnaire in the bag provided and hang on your door by so that we will not need to disturb you again.

119

Appendix D: Exploratory Data Visualization

120

121

122

123

124

125

Appendix E: Results of Correlation Analysis Table A.1: Correlation between energy behaviors and environmental values

Variables Pastconservationactions Utilityconsumption

Care for township -0.23 -0.38**

Feeling that town is unique -0.03 -0.37**

Willingness to move from township -0.33* 0.10

Farmland 0.13 0.15

Forest -0.07 0.00

Water 0.21 0.09

Green space 0.07 -0.03

Mountains 0.20 -0.05

New development -0.07 -0.16

Sense of community in township 0.17 -0.25

Small town character 0.22 -0.18

Arts and culture -0.01 0.03

df = 27

* Significant at the 10% level, ** Significant at the 5% level, *** Significant at the 1% level

Controlling for survey modality, township, energy values, perceived barriers, home square footage, home age, and socio-demographic factors

Table A.2: Correlation between energy behaviors and land cover

Variables Pastconservationactions Utilityconsumption

Open space land cover 0.08 -0.05

Low-intensity land cover -0.54* 0.22

Medium-intensity land cover -0.53* 0.39***

High-intensity land cover -0.55* 0.30

Developed land cover -0.44** 0.23

Forest land cover 0.22 0.04

Agriculture land cover 0.40** -0.36***

Other land cover 0.30 -0.25

df = 23

* Significant at the 10% level, ** Significant at the 5% level, *** Significant at the 1% level

Controlling for survey modality, township, energy values, perceived barriers, home square footage, home age, and socio-demographic factors

126

Table A.3: Correlation between energy behaviors and types of social engagement

Variables Pastconservationactions Utilityconsumption

Family interactions 0.06 -0.02

Close friends interactions 0.31 -0.35*

Acquintances interactions -0.04 0.21

Neighbor interactions -0.15 0.02

Community groups interactions 0.39* -0.33

Other social groups interactions 0.25 -0.01

Volunteering 0.35* -0.09

df = 22

* Significant at the 10% level, ** Significant at the 5% level, *** Significant at the 1% level

Controlling for survey modality, township, energy values, perceived barriers, home square footage, home age, and socio-demographic factors

Table A.4: Correlation between energy behaviors and types of community engagement

Variables Pastconservationactions Utilityconsumption

Past community actions 0.35* -0.12

Proposed land development actions 0.33 -0.33

Far or forest preservation actions 0.21 -0.20

Natural gas or water quality actions 0.14 0.03

Cost of energy actions 0.30 -0.14**

Intended community actions 0.40** -0.43**

Proposed land development intentions 0.56* -0.45**

Far or forest preservation intentions 0.39* -0.44**

Natural gas or water quality intentions 0.06 -0.22

Cost of energy intentions 0.4** -0.45**

df = 22

* Significant at the 10% level, ** Significant at the 5% level, *** Significant at the 1% level

Controlling for survey modality, township, energy values, perceived barriers, home square footage, home age, and socio-demographic factors

Table A.5: Correlation between energy behaviors and perceived barriers

Variables Pastconservationactions Utilityconsumption

Perceived barriers -0.15 0.08

Perceived community barriers 0.33** 0.23

df = 29

* Significant at the 10% level, ** Significant at the 5% level, *** Significant at the 1% level

Controlling for survey modality, township, energy values, home square footage, home age, and socio-demographic factors

127

Table A.6: Correlation between energy behaviors and housing typology factors

Variables Pastconservationactions Utilityconsumption

Shade coverage of home 0.06 -0.01

Number of ENERGY STAR appliances in home 0.43* -0.40**

Home ownership 0.00 0.00

Home typology 0.00 0.00

Home size -0.10 0.44***

Number of bedrooms 0.07 0.28

Home age 0.11 -0.20

df = 30

* Significant at the 10% level, ** Significant at the 5% level, *** Significant at the 1% level

Controlling for survey modality, township, energy values, perceived barriers, and socio-demographic factors

Table A.7: Correlation between energy behaviors and socio-demographic factors

Variables Pastconservationactions Utilityconsumption

Number of years residing in current community 0.02 -0.05

Age -0.06 -0.13

Gender -0.02 -0.14

Marital status 0.30** 0.33**

Number of people in household -0.03 0.08

Under age 18 -0.14 -0.10

Over age 65 -0.18 0.18

Race/ethnicity 0.07 0.13

Political affliation -0.04 0.09

Education -0.06 -0.12

Employment status 0.08 0.09

Income 0.22 -0.01

df = 27

* Significant at the 10% level, ** Significant at the 5% level, *** Significant at the 1% level

Controlling for survey modality, township, energy values, perceived barriers, home square footage, and home age

Curriculum Vitae

EducationPh.D. (dual title) Architecture and Human Dimensions of Natural Resources & the Environment. 2017. The Pennsylvania State University, State College, PA.MS Landscape Architecture. 2011. The Pennsylvania State University, State College, PA.BLA. 2005. The Pennsylvania State University, State College, PA.

Academic EmploymentInstructor of Landscape Architecture, Stuckeman School. The Pennsylvania State University. 2016 - Instructor & Course Author of Geodesign, Stuckeman School. The Pennsylvania State University. 2012 - 2016.Research Assistant, Stuckeman Center for Design Computing. Stuckeman School. The Pennsylvania State University. 2014 - 2015.Research Assistant, Immersive Environment and GIS Laboratory. Stuckeman School. The Pennsylvania State University. 2010 - 2011.Teaching Assistant, Stuckeman School. The Pennsylvania State University. 2009 - 2011.

Professional Practice2012 - 2014 Planner. Renaissance Planning Group. Orlando, FL.2008 - 2009 Landscape Designer. PBS&J, Inc. Alexandria, VA.2006 - 2008 Landscape Designer. LandDesign, Inc. Alexandria, VA.2005 - 2006 Designer. Hargreaves Associates. Boston, MA.Summer 2005 Intern. EDAW. San Francisco, CA & Denver, CO.

PublicationsZawadzki, Stephanie J., Stephen Mainzer, Rae Anne McLaughlin, & A.E. Luloff. Close, but not too close: landmarks and their influence on housing values. Journal of Land Use Policy. 62. 351-360.(in revision) Mainzer, Stephen & A.E. Luloff. Informing environmental problems through field analysis: toward a community landscape theory of pro-environmental behavior.Barrett, Austin, Lauren Abbott, Sarah Eissler, Carolyn Fish, Lacey Goldberg, Stephen Mainzer, & Max Olsen. 2015. Community perceptions of local food: a mixed methods approach in State College, PA. State College, PA. The Hamer Center for Community Design. The Pennsylvania State University. ISBN 978-1-941659-04-5.

Conference PresentationsMainzer, Stephen. Community and Landscape Driven Pro-Environmental Behavior: Preliminary Findings of Household Energy Consumption in Two Rural Towns. Psychology of Architecture Conference. December 2016. Austin, TX. Mainzer, Stephen. Towards a transdisciplinary action research design studio: adapting a human dimensions of natural resources education structure. Environmental Design Research Association. May 2016. Raleigh, NC. Barrett, Austin, Lauren Abbott, Sarah Eissler, Carolyn Fish, Lacey Goldberg, Stephen Mainzer, & Max Olsen. 2015. Community perceptions of local food: a mixed methods approach. International Symposium on Society and Research Management. June 2015. Charleston, NC.Zawadzki, Stephanie J., Stephen Mainzer, & A.E. Luloff. 2015. Close, but not too close: landmarks and their influence on housing values. International Symposium on Society and Research Management. June 2015. Charleston, NC.Mainzer, Stephen & Timothy Murtha. 2011. Creating resilient neighborhoods: the relationship between ecologic principles and private development in Greendale, WI. Council of Educator in Landscape Architecture Conference. March 2011. University of Southern California, Los Angeles, CA. Mainzer, Stephen. 2010. Creating accessible suburbs: a background on suburban housing and public housing projects in the U.S. No)Boundaries Geography Student Conference. March 2010. Pennsylvania State University. State College, PA.

Recognitions2017 Alumni Dissertation Award, Penn State University2016 Best Student Paper Award, Environmental Design Research Association Conference2011 American Society of Landscape Architects, Student Honor Award2011 Penn State University College of Arts and Architecture, Creative Achievement Award2011 Landscape Architecture Foundation, University Olmsted Scholar