Understanding values and attitudes toward recycling


Material Information

Understanding values and attitudes toward recycling predictions and implications for communication campaigns
Physical Description:
xi, 146 leaves : ; 29 cm.
Werder, Olaf
Publication Date:


Subjects / Keywords:
Mass Communication thesis, Ph.D   ( lcsh )
Dissertations, Academic -- Mass Communication -- UF   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 2002.
Includes bibliographical references (leaves 136-145).
Statement of Responsibility:
by Olaf Werder.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 029238662
oclc - 51020693
System ID:

This item is only available as the following downloads:

Full Text






Copyright 2002



This dissertation is dedicated to my mentor, Dr. Kim B. Rotzoll, who predicted the route I
was taking years before I entered it. May I strive to justify the faith and intuition.


While a dissertation is often considered the labor of life and pain of one single

individual to earn his or her rights of passage into the academic community, in actuality

there are many individuals without whose help this work would never be completed in its

present form. Consequently these people deserve my thanks for their contribution to this

dissertation. Without doubt, I was very fortunate to have found such an accomplished

scholar, advisor, teacher, and friend in Dr. Marilyn Roberts, who has entered my doctoral

life early. Without her, my experience would not have been as enjoyable and satisfying.

While her constant support and guidance have made me a better scholar, it is actually

hard to express in words the gratitude I feel for her encouragement. It has truly been a

privilege to work with her.

Several others have also been essential to my experience in the doctoral program

and in the completion of this dissertation; in particular the members of my committee. I

offer Dr. Debbie Treise my sincere thanks for helping me organize and simplify my ideas

on theory and application; and for sharing a kind word when I needed one. I thank Dr.

John Sutherland for his patience and belief in me while I struggled to learn the details of

statistical analysis; and for allowing me to learn from his expertise. I thank Dr. Cynthia

Morton for helping me find the connection between my personal ideals and academic

research, in introducing me to the field of social marketing; and for always being there

when I had questions. I thank Dr. Samuel Barkin for giving me a deeper understanding of

decision-making under risk, for helping me connect my interest in mass communication

with the broader topic of environmental risk analysis, and for simply being a friend.

I am indebted to my students, many of whom were possibly not quite aware of the

"double-life" I was leading as a teacher and researcher during that time. Their humor and

excitement in the classroom made my teaching role easier and gave me new enthusiasm

for continuing with my project every time I came out of the classroom. Their ability to

learn almost by themselves helped when I was engrossed with the research, and allowed

me to advance without "letting them down."

I extend special thanks to fellow doctoral students Stephynie Chapman, Andrew

Clark, Guy Golan, Jaemin Jung, Michael Palenchar, and (now faculty member) Kelly

Page for their wisdom and encouragement. I feel blessed to have these people as friends

in my life.

I feel indebted to the management and staff of the Florida Survey Research

Center. I have a deep appreciation for the generosity and wisdom of Dr. Michael

Scicchitano and Dr. Tracy Johns. Beyond assisting with the collection of the data, their

guidance was truly vital to this research.

I especially thank Dr. Lynda Kaid and her assistant Cynthia DeForest for giving

their time and resources in tracking down sources, for being willing to sponsor this

research, and for always lending a sympathetic ear when the research hit a dead end.

I thank Ken; Barbara and Johnny; Dr. Kim; Jason; Mary and her staff; Patty, Pam,

and Jodie; Ryan; Paisley; Charles; Florann and Jochen for supporting me, even though

they sometimes were not aware that they did. Their acceptance and support, even at weak

moments, gave me confidence to finish this project. Knowing that they exist in this world

is enough to make me happy and proud.

Finally, and above all, I want to thank my Mom and Dad, my sister Astrid and my

brother-in-law Andi, and my Grandpa and late Grandma. They are my light, model and

motivation. Words fail to describe how their love and support have made me into the

person I am today. While they pressed me when I needed it, they always saw the potential

in me, and their belief in me gave me faith to believe in myself. I am the person I am

because of these people.



ACKN OW LEDGM ENTS ..................................................... ........... ................ iv

A B STR A C T ........................................... ............ ... .......... .. ............. .. x


1 IN TRODU CTION ............................................. 1

Purpose of the Study .............................................. .........................................3
T heoretical R ationale.............................. .......... ........ ............................. ... 5
Scope and Lim stations .................................................... ................... .......... 8

2 REVIEW OF THE LITERATURE................................................................ 10

Social M marketing ..................................... ...... ............ ...... ..... .......... 10
Environmental Research in the Social Sciences............................. ................. 12
Environment as a Social Construct........................... ......................... 12
Environm ent in the M edia.............................................. ..... .... ......... 15
Environm ent and the Consum er ........................................ .......... ................17
Environmental Research ........................................... ..................... 19
Definition of the Recycling Term and Current Situation...................................20
Recycling in Florida............................ .. ... ................ 23
R recycling R esearch................................ ..................... ...... ............ 24
Theory of R seasoned A action ...............................................................26
Definition ...................................................... 26
Application in Recycling Research .................................... .................... 29
Move toward the Theory of Planned Behavior............................................ ....33
D definition ......................................... ............. ......... ..... ...... .......... ... 33
A application in R recycling Research .............................. .....................................37
Personal V alues and Recycling ....................................................... ................42
Introduction ........................... ....... ... ... ... ...... 42
V alues Research .................... ..... ....... .... ..... .... ... .. ........... 42
Schwartz Values M odel...................................... ................. ..... ..........44
Application in Environmental/Recycling Research.................................. .....46
Inclusion of Values in the Theory of Planned Behavior.................................. 52
Proposed M odel ........................................................ .. ..... ..........54
Summary, Research Questions, and Hypotheses ........................... ................ 58


Research Questions.......................... ....... ................................ 61
H ypotheses .................................................... ............... ......... ...... .......62

3 M ETH OD OLOGY ............................................................ .............................. 64
Operationalization of the M odel............... .............. ............................ 64
Sample............... ...................... .. .. ..............65
Survey D esign ............................ .... ..... .......... .... ....... ... ... ........ ......... 66
Explanatory Variables................ ....................... 67
R response V ariable ................................................... ............ .................. 69
R eliab ility ............... .. ... ..... .. ......... .............. .. ..... ........ ...... .. ... ........... 70
V alid ity ................ ................ .. .... ..... .. ...... ... ...... ... ...................... .. ........ .. ..... 7 1
Content V validity ........................................ ............ ....... ..................71
Convergent and Discriminant Validity..................................... ................. 71
Predictive V alidity ..................................................................................... 72
Construct Validity................................................................ 73
E external V alidity........................ ....... ....... ........................ ............... 73
Procedures .......................... ... .............. ..... .... ... ............. 74
M easurem ent ................... .......... ... ... ... ...... ....... .... .......... ...74
D ata Examination and Cleaning............................................. ..... .......... 75
Statistical Analyses.......... ................................................... 75
Data Aggregation........................................... ................ 76
Correlation Analysis ................................. ........................................76
Regression Analyses ................................ .............. ..........................76
Assumptions of M multiple Regression Tests...................... ............... .. ......77

4 RESULTS ....................... ................. ..... ............ 80
Prelim inary A nalyses.................................................. .......... ................. 80
Study Participants ....................................... ........................................... 80
D ata Exam nation Results........................................... ........................... 81
Regression Model Assumptions........................ ............................. 82
Analysis of the Theory of Planned Behavior ........................................ ..................83
V ariable Preparation ........................................ ................................. 84
R egression Study ................................................. .................................. 85
Analysis of the Proposed Values-Enhanced Model ......................................... 87
C orrelation Study................................... ............ ................................. 88
R egression Study ................................................... ............................... 90
Sum m ary of the H ypothese Tests.................... .................................................. 95
Exploratory Post-Hoc Analyses ........................................ .......................... 98
Sum m ary of the R research Q uestions...................................................................... 100

5 D ISC U SSIO N ........................ ........ ... .............................................. .... 103
Applicability of the Theory of Planned Behavior to Predict Recycling Interest ....... 103
Overview of the Hypotheses ............................................... 103
Comparison of the Effects of the Determinants of Intention........................... 104
Connection between Intention and Behavior .......................................... 109
Impact of Personal Values on Recycling Intentions............... ................. 110


Review of Different Recycling Belief Components............................... 110
O verview of H ypotheses................................................................... ......... 112
Improvement of Predictability Power of the Model........................................ 112
Relevance for Public Entities Creating Recycling PSA Campaigns..................... 115
Lim stations of the Study....................................... ......................... ................. 120
Suggestions for Future Research ........................................ ....................... 125
Conclusion ................................................ ... .... ... ... ............ 127

APPENDIX QUESTIONNAIRE ............................. ..................... ....... 130

LIST OF REFEREN CES............................................................................... .. 136

BIOGRAPHICAL SKETCH................................................................... 146


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Olaf Werder

August 2002

Chairman: Marilyn S. Roberts, Ph.D.
Major Department: Mass Communication

Since the 1980s, waste management issues have emerged as a key concern for

Florida, a state with rapid population growth. Because the influence of the legislature is

limited to supporting the counties in their recycling efforts, a vast discrepancy exists

among Florida's 67 counties.

While recycling figures and participation percentages might be difficult to com-

prehend, their role as an environmental problem is not. Because research into intrinsic

motivation considers fundamental factors in actual decision making, an important con-

cern of public entities in charge of community recycling has always been to determine

why people do or do not participate in these programs.

Environmental values have been found to be a key determinant for pro-

environmental behavior and, therefore, regulate the manner in which behavior occurs.

Social marketing efforts often compromise a person's values, in order to promote values

deemed socially more compelling by the sponsoring organization. Since individuals hold-

ing opposing values may be reluctant to comply with the marketing goals, it seems criti-

cal to incorporate personal values into a public campaign.

The purpose of the present study is first to test the Theory of Planned Behavior as

an explanation of recycling intentions. The theory maintains that attitudes, norms, and

perceived control elements determine behavioral intentions. Second, the study will ex-

plore more fully what role values play in explaining recycling intentions.

A telephone survey was conducted during the last two weeks of May 2002 in

Gainesville, Florida. It was hypothesized that residents' attitudes, subjective norms, and

perceived behavioral control would equally predict recycling intentions. It was further

hypothesized that the inclusion of values would increase the likelihood to explain recy-

cling intentions better.

The findings suggest that attitudes toward recycling are the most significant pre-

dictor of intentions. While certain values correlated well with the TPB model variables,

values did not significantly improve the parameters' predictability over intentions. How-

ever, the close correlations values have with beliefs and evaluations of recycling imply

that values can expand the applicability of behavioral models for recycling. On a practical

level, the study suggests that an understanding of values can improve communication

campaigns that aim to change or reinforce habits.


Since the 1980s, waste management issues have emerged as a key concern for

state and local governments in the United States, particularly in states with rapid

population growth. In Florida over 24.8 million tons of solid waste were collected in

1998 (Department of Environmental Protection 2001), an increase of over 500,000 tons

from 1995 (Department of Environmental Protection 1996). This translates into 9.1

pounds per person per day. While this figure is slightly lower than the estimated Florida

average of 10.2 pounds/person/day (DEP 2001), the corresponding recycling estimate is

dramatically below reality. Florida's prediction report for 1998 estimated that 10.1

million tons of waste (about 36% of the total) would be recycled. The actual recycling

rate for 1998 was 6.9 million tons, or 28% (DEP 2001). This means that 270 pounds per

Floridian per year will actually end up on landfills instead of being recycled.

Although Florida's average recycling rate of 38 to 40% ranked highest in the

nation (Environmental Protection Agency 1999), the state placed 56% of its total waste in

landfills and 16% in combustion. In other words, the recycling quota is in a negative

trend (-12% from 1995). With a population growth projection between 17.5 and 23

million people by 2020 (DEP 2001, Roe Littlejohn 1997), there is ample concern about

Florida's waste management.

Florida's Department of Environmental Protection and individual counties

administer the funds for education and information (media). As a result, the incorporated

municipalities have a great deal of discretion in structuring the actual recycling programs

within their boundaries (Martinez & Scicchitano 1998). The influence of the legislature is

limited to mandating certain waste reduction figures (currently 30% should be recycled

for counties with populations over 50,000) and introducing recycling goals for individual

counties. A vast discrepancy in recycling rates exists among Florida's 67 counties,

ranging from a recycling rate of 38% in Lee County to 5% in Hendry County (DEP

2001). Recycling success or failure cannot simply be explained by factors such as

urbanization, population size or citizens' educational and income levels.

Alachua County, the home of Florida's largest public university, is a rather rural

county in the northeast part of the state. With a population of about 212,000 it ranks only

nineteenth in population size among all counties. However, it shares the fourth rank with

Palm Beach County in recycling rates (DEP 2001). It not only surpassed counties with

metropolitan areas, such as Orlando, Miami, and Tampa, but also counties with

educational centers, such as Leon County, where Florida's second largest university is

located. If not demographics, what explains Alachua's success? The county and its

largest community, Gainesville, have managed to establish curbside recycling pick-up

service to 96% of single-family dwellings and 5% of multi-family dwellings (apartment

complexes). Among those with service, 80% of single-family homes and 4% of multi-

family homes participated, for a total participation rate in the county of 58% (DEP 2001).

Although a rate of 58% is compared to other counties an adequate level, it also means

that 42% of the population are not recycling. Simply judging from the service-to-

participation relationship, one notices that 17% of homes that could recycle are not

currently recycling. It seems that creation of service is not enough to overcome reluctance

to recycle in certain individuals.

While the amount of waste and participation percentages may be difficult to

comprehend, their roles in environmental problems are not (Bagozzi & Dabholkar 1994).

Given that environmental problems and social ills are interconnected, waste management

is a social problem with far reaching consequences in the areas of human health,

ecological balance, and the local economies (Starke 1991). Consequently, determining

consumers' reluctance to recycle created considerable research interest in disciplines such

as psychology and environmental policy (Arbuthnot 1977; Hines, Hungerford & Tomera


Thogersen (1996) describes recycling research as falling into two main theoretical

approaches. The first, applied behavioral analysis (Stem & Oskamp 1987) provides

information about reactions to extrinsic stimuli. The second, attitude research (Hopper &

Nielsen, 1991; Kok & Siero 1985) analyzes the cognitive (attitudinal) antecedents

believed to guide the behavior. Because research into intrinsic motivation considers

fundamental factors in actual decision making, it bodes well for future effort of

determining motivating factors of recycling (Bagozzi & Dabholkar 1994).

Purpose of the Study

The purpose of the present study is twofold. First the Theory of Planned Behavior

(Ajzen 1985) was tested as an explanation of recycling intention. This provided an

effective framework for studying the determinants of recycling behavior. The Theory of

Planned Behavior hypothesizes that intentions directly determine behavior and are

themselves influenced by attitudes toward the consequences, projected subjective norms

about others' opinions, and feelings about personal control over one's behavior and its

outcomes. Despite its successful application to recycling behavior (Oskamp et al. 1991;

Vining & Ebreo 1990), most studies have not operationalized the variables specifically,

but have deviated from the recommended practice of applying all determinants as

specified by the theory. Thus findings have been mixed. By applying the Theory of

Planned Behavior in the current study, the full model was tested. The findings were then

compared to an augmented model in which recycling-related personal values were

introduced as a predictor of recycling intentions (Schwartz 1992; Stern & Dietz 1994).

The second purpose of this study is to more fully examine the determinants of

attitudes, subjective norms and perceived behavioral control as they relate to recycling.

The study models and builds upon the research of Bagozzi and Dabholkar (1994), who

stated their purpose as follows:

"Previous research using attitude theory has investigated antecedents based upon
beliefs and evaluations and organized these into the summation of the products of
beliefs and evaluations. This approach works best for physical products or when
the consequences of acting are concrete and tangible. But because recycling
involves abstract goals and values and is highly subjective, the traditional
approach is less useful. Recycling-related beliefs are concrete judgments about
the consequences (positive or negative) of recycling and tend to focus more on
means or outcomes (e.g., inconvenience, saving money). Recycling goals, in
contrast, are abstract motives for recycling (by definition positive) and refer to
ends (e.g., provide for future generations). Recycling beliefs enter decision
making as reasons for or against acting; recycling goals are more conative and
may even be deontological moral values that motivate or compel one to act. Many
recycling goals, particularly higher-order ones, do not arise from decision making
but are a priori virtues. Another problem with the traditional approach is that it
does not address the hierarchical organization of the antecedents to attitudes,
subjective norms, and past behavior. Our second objective is thus to discover (a)
the key antecedents, (b) how they are structured, and (c) how they influence the
proximal causes (i.e., attitudes and subjective norms) of decisions to recycle" (p.

In other words, it is important to separate between beliefs and goals, or values that

lie beneath beliefs and are unencumbered by cognitive decision making processes.

Assuming the role of a priori virtues, values are more difficult to influence by everyday

stimuli while at the same time determining to a large extent the motivation to act.

Theoretical Rationale

Most marketing and mass communication research has focused on studying

demographic variables, knowledge, or environmental concerns (Vining & Ebreo 1990;

Van Liere & Dunlap 1980). Previous researchers (Oskamp et al. 1991) sought to identify

demographic and psychographic profiles of environmentally concerned people in order to

use this information for product development or target segmentation. Environmentally

conscious or "green" consumer segments have become increasingly important for the

proliferation of products and services as well as corporate images (Elkington 1994;

Kassarijan 1973).

This mode of thinking seems to reflect the prevalence of the established

'consumption constellation construct,' (Lowrey et al. 2001), defined as "a cluster of

products and consumption activities associated with a social role" (Kamins & Assael

1987). It suggests that attitudes toward products symbolize information about the

consumer's self-identity. In the realm of social and environmental issues, this correlates

to a scenario in which a positive attitude toward a cause leads to a positive behavior with

respect to that cause.

A shortcoming of most research of environmental concerns is that the

conceptualization and measurement are too broad. Many previous studies cover both a

wide range of psychological reactions and a wide range of behaviors within the same

construct (Bagozzi & Dabholkar 1994). By mixing the many psychological reactions with

the many behaviors, the construct makes it difficult to predict specific behaviors beyond

tautological interpretations (Van Liere & Dunlap 1981). Ajzen and Fishbein (1980)

suggested that there is reason to expect that "attitudes toward a target may be unrelated to

a person's beliefs about the consequences of performing a specific action toward that

target" (p. 88). For example, a person may hold pro-environmental opinions, but in no

way be inclined to participate in household recycling activities. The particular individual

may not perceive his or her behavior as negatively impacting the environment.

A promising line of inquiry examines the behavioral and attitudinal causes of

recycling. Those using expectancy-value models (such as the Theory of Reasoned Action

or Theory of Planned Behavior) largely subscribe to a rational actor idea. In these

models, the individual weighs the costs and benefits of the outcomes of his or her

behavior and acts accordingly. Although it was found that economic sensibility is one of

two basic attitudes leading to sorting of recyclables (Israel 1991), it was not so much the

economic motives but the idealistic motives (protecting the environment) that stimulated

people to participate. De Young (1986) reported a close association between derived

satisfaction and intrinsic motivation. These motives allow for the inclusion of ethical and

altruistic positions, in which the individual might not be the beneficiary of the action.

Ultimately, both positions are results of the abstract goals and values origins of the


While personal values define individual goals, goal-directed behavior is

considered the substance for both rationality and altruism (Pierce 1979). Essentially both

the economic and the socio-psychological models of man emphasize two basic cognitive

components: values and beliefs (Shapiro 1969). Values are responsible for the selection

and maintenance of the ends or goals toward which human beings strive. Beliefs are

proscriptive convictions upon which humans act by preference (Allport 1961) creating an

intervening variable for behavior. Values in turn regulate the methods and manner in

which this striving takes place (Vinson et al. 1977).

There is debate about the degree to which value systems at the societal level

directly drive society-environment interactions. A growing body of literature on the

distribution of risk within and between nations treats decisions about technological risk as

revealing societal value preferences (Beck 1992; Salmon 1989). Some nations (Germany,

Great Britain, and the U.S.) are willing to accept the risk connected to advancing

technology, such as nuclear energy generation, while others (Denmark, Sweden) are more

concerned about the risk element and approach progress more carefully.

Public information campaigns take the form of social interventions that have been

prompted by the determination that some situation represents a social problem meriting

social action. These campaigns are logically seen as a value-laden activity, where people

bring their own moral judgments to an activity (e.g., anti-yard burning campaigns). Not

all persons will agree upon the ends pursued (stop yardburning) and the means used to

achieve these ends. At the center of this conflict is the fundamental tension between

social control and individual freedom (Salmon 1989). As such, social marketing or

communications efforts necessarily compromise certain values and interests, in order to

promote values and interests deemed more socially, economically or morally compelling

by the sponsoring organization of the change effort. As a result, individuals holding

opposing values will be reluctant to comply with the message content.

Previous research on the values origins of general environmental concerns and

specific environmental behavior, such as recycling, identified three "ethics"

corresponding to three classes of valued objects. These are: the homocentric (or socio-

altruistic) which focuses on other people; the ecocentric (or biospheric) which focuses on

nonhuman objects; and the egocentric (or egoistic) which focuses on the self (Stern et al.

1993; Merchant 1992). The first two have often been combined in subsequent research

(Stern & Dietz 1994) to form a general altruistic ethic toward human and nonhuman

entities. Schwartz (1992) eventually added a traditional orientation as a fourth "ethic,"

implying a behavior using internalized societal norms (such as local customs).

A communication study that examines the origins of recycling beliefs and

intentions will provide implications for public policy, marketing communications, and

future public service or advocacy campaigns that aim to affect behavioral change. By

providing additional information through extensions of the studies of Ajzen and Fishbein

(1980) and Bagozzi and Dabholkar (1994), we hope to further clarify the role of values in

citizens' attitudes and behaviors toward recycling. Values, as enduring beliefs in the

preference of a specific mode of conduct (Rokeach 1973), form the foundation for human

attitudes and perceptions of the world. Communicators need to understand the nature and

range of values (and how different value orientations affect both attitudes and consequent

behaviors), if they want to predict outcomes of policy decisions. Social marketers and

advocacy communicators must learn how to craft more successful advertising messages

for issues and causes of local, national and global concerns.

Scope and Limitations

Any sampling choice has the potential not to be representative of an entire area

population, let alone the population of the entire country. The current study should be

seen as offering a theoretical relationship between values differences and their potential

impact on behavior in a narrow geographic area. Although any limited survey will never

elucidate the tendencies of an entire population, it can point to other relevant avenues for

future research.

While the current study postulates an extension to an existing model of human

behavior regarding recycling, it also naturally excludes additional variables that may

moderate or explain recycling behavior, such as personal effort or the influence of

persuasive cues. While researchers do not always agree on the exact impact of a given

variable on recycling behavior, values do indeed have an effect, either by reinforcing

existing beliefs and attitudes or by obstructing or counteracting situational attitudes and


Finally, issues of measurement reliabilities in the model construction need to be

addressed. While it is hoped that proper pre-testing and variable definition will largely

eliminate confusion and misunderstanding, the potential of misunderstanding in defining

and measuring a rather amorphous and seemingly intuitive construct such as a value does

exist. Since people might not apply much cognitive effort to summon a personal value

(unlike the attitude toward an object or behavior), it is the researcher's responsibility to

minimize confounding influences.


This review consists of four main sections:

* A brief overview of social issues and environmental issues communication theory
with an emphasis of waste recycling literature.

* A discussion of the Theory of Reasoned Action and review of selected empirical
studies in which this theory was used as a framework for recycling.

* A discussion of the Theory of Planned Behavior and review of selected empirical
studies in which this theory was used as a framework for recycling.

* A review of values research, its use as a framework in recycling studies, and the de-
velopment of the final model to be used in the current study.

Social Marketing

Social issues are ideas that are of interest to many individuals within a society

(Fine 1981). They differ from commercial ideas (for tangibles or services) as they are not

only motivated by self-serving goals, but by a desire to help others. Most social issues

benefit other people more than the individual acting on a social message, while he or she

carries most of the burden or cost. For instance, avoiding littering to preserve nature's

pristine beauty when one has to carry the wrapping paper of a hamburger for hours be-

cause there are no receptacles around is generally considered a huge cost by most people,

while the benefit to the self does not seem evident.

The adoption of new social ideas is closely related to the formation of values, atti-

tudes, beliefs, interests, and opinions about issues. A belief (as well as the related con-

cepts) can be regarded as a mental acceptance of the validity of an idea. The totality of all

beliefs the belief system determines the position one takes on an issue, which often

prompts participation in social action (Fine 1981).

At first glance the adoption process for social ideas does not seem to be too dif-

ferent from that for commercial goods. At the heart of all marketing lies a philosophy of

consumer orientation. Goods and services are described as the solution to a problem. The

aim of marketing efforts is to position this idea in a way that potential consumers of the

good accept this description and act accordingly. Social marketing, or the marketing and

promotion of social issues, distinguishes itself from commercial marketing by two key

components. First, the "product" of social marketing is often amorphous, a mere idea of

what ought to be, such as physical health, pollution control, social justice, gender and

race equality. This makes it fairly difficult to attach a price to the "product," not only for

the change agent, but also the consumer. This, in turn, has effects on promotion, which is

often in an interactive relationship with price. It also affects measurement of the social

price of an idea for cost/benefit purposes. Ideas of "breaking even" and "getting what you

paid for" do not seem to apply well.

Second, related to this is the idea of consumer response within the larger field of

consumer research. One difference lies in the nature of the forces motivating purchase (or

adoption) behavior. The perceived (and actual) consequences could be more far reaching,

more involving, than the effects of an ordinary purchase. Being a Ford Truck owner is a

less profound statement than being a Greenpeace activist. It would seem that adoption

behavior is based on more subtle and indirect motivation than acquisition behavior. In

committing oneself to follow a particular movement, one probably undergoes a good deal

of forethought. Impulse buying is not very prevalent (Fine 1981). In adopting an idea,

reinforcement through gratification either occurs as a delayed reward (weight reduction,

quitting smoking) or it accrues not to the individual, but to society as a whole (pollution,

equal employment).

It makes sense that social consumer behavior is well suited to be studied within

the broader discipline of consumer psychology. The argument is that conceptual frame-

works (such as perception, cultural values, attitudes, group influences, personality, learn-

ing, and information processing) form an important role in understanding a consumer's

opinion and reaction toward a social issue. Wasik (1996) argued that "the ultimate exten-

sion of marketing is the selling of values" (p. 59).

One area that fits squarely under the above-mentioned agenda is the relationship

of people toward their natural environment. A heightened research interest in this rela-

tionship is documented by a subfield of social issues communication (the area of envi-

ronmental, green, or eco-communication). The next segment introduces this discipline

and discusses the particular circumstances that surround environmental issues.

Environmental Research in the Social Sciences

Environment as a Social Construct

Social theorists, studying environmental risk perception, concentrate almost ex-

clusively on the social and economic spheres and have tended to neglect the cultural

arena (Beck 1995). The environment is more often associated with the natural than the

social sciences, explaining in part the silence of sociology on this issue. When it was ex-

amined, the relationships between society and nature were seen as distinct spheres gov-

erned by different temporal mechanisms.

Throughout the history of Western thought there have been competing models of

the relationship between humans and nature. Some have depicted nature as a state of

chaos. Thomas Hobbes (1651/1958) viewed the natural human condition before the emer-

gence of civilized society as brutish and short. By contrast, his contemporary John Locke

(1690/1960) thought that nature was a state of humanitarian bliss; "natural laws" must

form the basis of a just society. "Nature," or our relationship with the physical environ-

ment, is socially constructed. "Nature" is culturally and historically constructed since our

perceptions are inextricably bound up with particular models of society that are dominant

at any one period in time.

Recently anthropological studies have concentrated on how certain social prob-

lems come to be defined as risks. A study by Douglas & Wildavsky (1982) suggested that

our selection of risks is influenced by social values and the way in which different cul-

tures operate. Competing public perceptions of risk are equally biased because they re-

flect different cultural meaning systems. Alex Wilson (1992) in his seminal book entitled

'The Culture ofNature' explored some of the ways in which "nature" is culturally con-

structed in modern society. He argued that nature cannot be separated from culture since

it is mediated through major social institutions and the culture industry.

For the German social theorist Ulrich Beck (1992), risk, appearing in the natural

environment (e.g., nuclear reactor explosions, greenhouse effect, groundwater contamina-

tion by leaching landfills) has become a central anchor for conflict in modern industrial-

ized society. He argued that riches are tangible goods that are understandable. Production

is the result of methodological thinking and execution. In contrast, the perception of eco-

logical devastation and the consequences of industrial growth are difficult to grasp. This

perception can depend to a much lesser extent on methodological knowledge, measure-

ment procedures, rules of accountability and acknowledgement in science and law, and

on information policies of suspect operations and cooperating authorities.

Perception of devastation must break through the wall of denial that stems from

the fact that most ecological disasters elude pinpointed scientific measure, making it hard

for scientists to generate laws. There is no trend in sight that experts are getting organ-

ized. Neither can those who report isolated cases escape playing the role of a "deviant

expert." In the ecological conflict, individuals or small groups can act with considerable

effect. The politics of the ecological question involve universal themes. The conflict even

passes through people. While one's heart may beat "Green," one's mind and routine con-

tinue in old habits.

Unlike social issues, ecological ones often face human inactivity. While our own

experience supports action on the social question, the ecological issue is not merely ab-

stract. It virtually requires that we ignore our own senses. Although poverty can be made

to disappear statistically, it remains painfully present for those who endure it. On the

other hand, air pollution from cars or land pollution from littering resembles examples of

'prisoner's dilemma' theories (Van Vugt et al. 1995). Often the menace can only be per-

ceived in defiance of the semblance of normality (Beck 1995).

Only by using complicated measuring instruments can the nature and degree of

the threat be determined. Thus, threats replace individual organs of perception with gov-

ernmental, bureaucratic, and scientific "organs." The blindness of everyday life with re-

spect to the omnipresent, abstract, scientific threats is a relative and revisable process. It

depends on the socially available knowledge and how much society considers it worth-

while to pay attention to these events that "at first glance" appear to be nonexistent. Ways

of acting need to be rewarded that simply raise into view what was previously invisible.

Democracy can be protected from perishing in the thicket of risk expertocracy. Those

who would open people's eyes to the ecological issue and keep them open must redirect

and inspire society's knowledge and perception through education and training. Beck

(1995) argued:

"Only if nature is brought into people's everyday images, into the stories they tell,
can its beauty and its suffering be seen and focused on. Seeing is cultural seeing,
attention is narrated attention. Our culture, and therefore, we ourselves, see and
hear in symbols, in which what is invisible or forgotten stands out and lives figu-
ratively. This does not just happen; rather it is done, often against resistance.
Knowledge of cultural sensitivities is just as significant for this work as are cour-
age or objective knowledge" (p. 56).

The social and economic importance of knowledge grows similarly, and with it

the power over the media to structure knowledge (science and research) and disseminate

it (mass media).

Environment in the Media

Many risk communication studies are based on the underlying assumption that it

is possible to judge the quality of reporting through the use of objective measures. Al-

though we cannot totally dispense with the objective ideal, objectivity is not necessarily

the same thing as accuracy. For instance, with culture variously being defined as incorpo-

rating values and norms, ideology, subjective states, rituals and discourse, events dis-

cussed by the mass media often tune into deeply held cultural beliefs. Particular issues

that attract attention tend to be mediagenic and often possess a powerful symbolic reso-


The importance of culture in the local context drives the framing of public under-

standing of environmental issues. Lay audiences often draw on local knowledge in mak-

ing sense of those issues. Though the news media play a significant role in shaping atti-

tudes, audience research suggests that we may take on different subjectivities in interpret-

ing media texts. In other words, our reading of media texts is framed by our pre-existing

attitudes and social and cultural backgrounds (Allan et al. 2000).

Media reporting on environmental issues is often risk-led. Coverage is often based

on anxieties, concerning threats to health posed by major incidents, accidents, or disas-

ters. People's attitudes about risk are largely focused on specific risks rather than an out-

look on environmental issues in general. Where some risks are concerned, there are im-

portant divergences of perception between policy-makers, scientists and the public. The

environment, like other substantive areas of media reporting, is largely mediated through

the "expert" as the voice of authority, using quantitative measures as a basis for risk per-

ception. They often marginalize lay views (Bell 1994), which are more likely to be influ-

enced by qualitative assessments. Risk "experts" are often critical of the mass media, ar-

guing that risks tend to be distorted and the media are too reliant on pseudo-experts. Me-

dia practitioners tend to treat issues in a rather emotive way, exploiting the human inter-

est factor.

Stories lacking emotive quality (such as household waste recycling) are usually

not communicated, or communicated in a way that would shape attitudes in favor of the

subject. With lack of interest from the editorial side, it seems to be the commercial com-

munication field that aims to reach and form ecological awareness.

Environment and the Consumer

Although concerns about environmental issues have objective roots, they are

shaped by the promotional activities of issue sponsors and culture representation (includ-

ing advertising, photography and art). A change in the perception of nature has led to a

growing recognition of the need to "manage" public opinion concerning the environment.

Since the mid 1980s significant energies have been channeled into the substantial risk-

management industry and corporate green advertising (Anderson 1997).

The reason for this activity was that many businesspeople believed that the 1990s

started an "environmental decade" (Fisher 1990). In fact, consumer interest in environ-

mentalism, a phenomenon labeled "green consumerism" by Ottman (1992), grew in the

marketplace, primarily fueled by an increasing awareness of issues due to increased me-

dia coverage of disasters, such as the Exxon Valdez spill and the Bhopal killings. A gen-

eration who grew up with environmental education had meanwhile reached working and

voting age. It seems that the public had begun to realize that their consumption activities

contributed to environmental problems. "As a consequence there appears to be a growing

desire to protect the environment as evidenced by a seeming willingness of consumers to

avoid products that they believe contribute to environmental degradation" (Carlson et al.


The end result of "green advertising" can be understood as an attempt to engineer

change in society. A media or information campaign takes the form of social intervention

prompted by the determination that some situation represents a social problem meriting

social action. This campaigning effort is logically seen as a value-laden activity, as not all

persons will agree on the ends pursued and the means used to achieve these ends. At the

center of this conflict is the fundamental tension between social control and individual

freedom (Salmon 1989). As such, social marketing or communications efforts necessarily

compromise certain values and interests, in order to promote values and interests deemed

more socially, economically or morally compelling by the organization sponsoring the

change effort.

Depending on the context, this social situation can take the form of some individ-

ual or group, a change agent or agency, making such determinations. As a case in point,

some individuals might engage in behaviors which bring them pleasure or facilitate their

lives, but which also have a level of risk associated with them that a change agent consid-

ers too high. In general, any phenomenon that happens in and to the environment has

usually major consequences for humans (greenhouse effect, ozone layer, erosions,

groundwater contamination, and air pollution to name a few).

Both J. Walter Thompson (1990)'and The Roper Organization (1990) have con-

ducted large-scale surveys in the United States. The findings empirically supported the

success of campaigns in identifying relatively large and emerging consumer segments

with a definite lifestyle and propensity to "buy green" (Fuller and Allen 1995). The rea-

son for why "green consumerism" became one of the most accepted areas among all en-

vironmental issues seems to be found in the fact that it resembles most the marketing sce-

narios of commercial marketing. It is important to point out, though, that the bulk of

environmental marketing and social marketing in general has to deal with challenges

that are qualitatively different and unique to its field. Those comprise characteristics,

such as lacking demand (enticing positive behavior toward a service for which the target

audience sees no need), obscure benefits (encouragement of a behavior that leads to the

absence of a negative outcome), or third party benefits (payoff of a behavior goes to a

third party or society in general) (Andresen 1995; Kotler and Andresen 1991).

A case in point is the effort of local communities to deal with the reduction of

household trash, officially called municipal solid waste (MSW). As more landfills are

filled to their capacity, various municipalities in the U.S. have engaged in serious efforts

to promote waste avoidance in the form of reducing, reusing, or recycling waste. The ac-

ceptance rate to cooperate is different for cities and states. Household waste recycling

constitutes an example of the third-party benefit characteristic.

Environmental Research

As early as the 1970s, marketing efforts have attempted to identify the ecological

oriented consumer. A flurry of research was conducted to profile population segments

that showed environmental concern (Anderson and Cunningham 1972; Balderjahn 1988;

Kassarjian 1971). Throughout the 1980s other academic areas began concentrating on the

ecologically-conscious public as well, such as sociology (Van Liere and Dunlap 1981),

education (Hines, Hungerford and Tomera 1987), and psychology (Arbuthnot 1977).

Similar to marketing studies, these research projects concentrated for the most part on

descriptive information, such as demographics, with some focusing on personality and

psychological factors, such as alienation, attitude toward pollution and knowledge of en-

vironmental issues (Polonsky et al. 1995).

Overall, the relationships of demographic and socioeconomic variables with eco-

logical concern did sometimes result in inconsistent or contradictory findings with re-

spect to the direction of the assumed relationship. On the other hand, constructs such as

personality measures, dogmatism, and attitude studies showed some promise (Kinnear,

Taylor and Ahmed 1974). The idea behind these constructs was that each citizen has a

duty to the community and future generations in an environmentally responsible manner.

Schudson (1991) in his framework analyzed the consumption culture. His findings not

only echoed the psycho/social indicators, but also added the construct of social norms or

pressures as a guide for environmental behavior.

Unfortunately, much of the past behavioral science research has studied general

environmental concern rather than more restricted topics (Oskamp et al. 1991). After re-

viewing 23 articles that investigated factors relating to environmental concern, Van Liere

and Dunlap (1980) recommended that environmental concern should be studied in terms

of more specific environmental issues. Research should investigate people's beliefs and

attitudes of those issues concerning trade-offs to other valued goals. Dunlap and Van

Liere (1984) found, for instance, that traditional American values (e.g., support for eco-

nomic growth) were detrimental to maintaining a strong proenvironmental stance.

Recycling is used in this project as it is a good example for an action that typically

offers little direct benefit to the individual, but that often involves substantial personal

cost with respect to time and effort (Smith et al. 1994). Next, the attention is turned to the

specific environmental issue of recycling.

Definition of the Recycling Term and Current Situation

Municipal waste recycling, or post-consumer recycling, is a term applying to the

recycling of waste materials generated by personal consumption activity as opposed to

those generated directly by industrial processes (Fuller and Allen 1995). Recycling is de-

fined as "the extraction and reuse of useful substances found in waste" (American Heri-

tage Dictionary 1985). This definition implies a circular flow of product disposition, as

opposed to the traditional linear one, to reintegrate materials in the market (Fuller and

Allen 1995).

Of the twenty most industrially advanced democracies in the world, the U.S.

ranks fifteenth in paper recycling and nineteenth in glass recycling. According to the

Congressional Research Service, "Other countries use less packaging than the U.S., recy-

cle more of it, and are considering recycling policy measure stronger than measures gen-

erally being considered in America." Despite the fact that on a per capital basis, as well as

in absolute amounts, the U.S. is the largest generator of waste of any nation on earth, the

U.S. is least engaged in any of the above-mentioned activities (Hershkowitz 1998).

Using recycled materials helps avoid the air and water pollution typically caused

by manufacturing plants that solely rely on unprocessed virgin raw materials. Recycling

materials reduces the need to process and refine the raw materials for paper, plastics,

glass, and metals. Recycling lessens the toxic air emissions, effluents, and solid wastes

that these manufacturing processes create. Moreover, timber harvests, for instance, would

have to increase 80% over current levels without recycled fibers (Hershkowitz 1998), an

example of its influence on virgin resources and the entire 'ecoscape'. Recycling also im-

pacts energy production by saving more of it relative to the incineration of wastes for en-

ergy recovery. Aside from these indirect effects, recycling has direct positive effects re-

lated to health and ecological risks associated with human household and industrial


Landfills generate hazardous and uncontrolled air emissions and threaten surface

and groundwater supplies. They have contaminated aquifer drinking water supplies, wet-

lands, and streams throughout the U.S. The list of toxic and hazardous chemicals emitted

as gas or leaking as liquid from thousands of landfills defines a waste management option

with wide-ranging pollution impacts.

As Americans learned of these serious environmental problems posed by the dis-

posal of certain materials batteries, yard wastes, tires, etc. into landfills, recycling be-

gan to proliferate. The management of garbage became more complicated. As entrepre-

neurs and environmentalists demanded that valuable, useful, or dangerous materials in

the waste stream be separated for reprocessing or marketing, the logic of municipal waste

collection shifted in many communities. Operating budgets and administrative procedures

relating to sanitation programs were modified.

The media began to turn more proactive as well. What has previously been la-

beled a "non-event" had evolved into something that could be captured with pictures and

personal stories, showing the risk affiliated with anti-recycling behavior. As a result recy-

cling stories and stories of risks due to unchecked waste dumping became more numer-

ous. Since recycling is part of a larger web of interwoven economical, political, legal, and

cultural issues, the rate of recycling differs from state to state and community to commu-

nity within a state.

The execution of the recycling task has normally been placed in the jurisdiction of

a local municipality. In a response to escalating waste problems many states and munici-

palities have issued legislation that typically includes mandated curbside collection pro-

grams of recyclable goods. However, these mandates effectively removed the voluntary

cooperation aspect of recycling for those communities.

Recycling in Florida

In Florida, curbside recycling activities are voluntary. The state of Florida is actu-

ally a leading state on recycling. It is currently tied for third place with Tennessee and

Wisconsin, after Minnesota and New Jersey respectively, in the percentage of municipal

waste that gets recycled (EPA Data 1999). Florida's 67 counties show differing rates of

accomplishing the average statewide recycling goal of between 35 and 40% of all waste.

While more rural areas with low populations and infrastructure (Dixie County, Indian

River County) tend to be the low recycling candidates, and urban centers (Miami-Dade

County, Duval County) the leaders, there are surprising differences from that norm. The

top recycler in the state is Lee County (Ft. Myers), while the Tampa-St. Petersburg metro

area (Hillsborough and Pinellas Counties) as well as the capital city of Tallahassee (Leon

County) rank in the lower midfield, despite equal ordinances by the respective state de-

partment (Department of Environmental Protection 2001).

It is the local municipality's responsibility to provide special bins at no cost to the

household, which would then put those bins out alongside the regular garbage receptacles

on collection day. There is some evidence that this approach has met with consumer ac-

ceptance. Sixty to eighty percent of the eligible households (usually single unit homes) in

communities such as Jacksonville, Fort Myers, Daytona, and Gainesville participate in

the program. Computed over the entire community's population, this roughly translates

into the targeted acceptance rate of 30 to 40% (Department of Environmental Protection,

2001). These statistical averages cannot accurately reflect and demonstrate a unified re-

cycling participation rate within even the high-scoring communities. It constitutes, among

others, a reason for why communities and companies are still searching for better ap-

proaches to change recycling behavior.

Waste haulers, the companies hired by a local municipality to pick up household

waste, have been in a unique position to become functionaries in these channels, as a re-

sult of this acceptance rate. For example, in North and Central Florida, Waste Manage-

ment Inc., a major solid waste hauler, has formed a subsidiary, Recycle America, to im-

plement curbside collection contracts with local municipalities. "The process involves the

use of specialized, compartmentalized collection vehicles and also the operation of a cen-

tralized municipal reclamation facility (MRF), which is sponsored by a consortium of

local governments. The MRF is the central receiving facility, at which sorting, packaging

(baling, densification, etc.), and marketing activities take place. Since the geographic

coverage of waste-hauler contracts is often extensive, these systems can generate signifi-

cant steady volumes of materials over time" (Fuller and Allen 1995).

Recycling Research

Waste management issues have become a key concern of the government, the pri-

vate sector, and the general public (Taylor and Todd 1995). People appear sensitive to

environmental issues, and many seem to hold positive attitudes toward environmental

programs. Despite these positive attitudes, participation in different voluntary waste man-

agement programs varies widely (McCarthy and Shrum 1994). Notwithstanding a grow-

ing literature on the behavioral research on recycling (Ebreo 1999; Shrum et al. 1995;

Stern and Oskamp 1987), little is known about the factors that influence individual waste

management behavior, or how beliefs and attitudes relate to behavior. According to

Shrum, Lowrey and McCarthy (1994), most studies examine only a small number of

variables and create models that lack integrative power. In an attempt to build more theo-

retically integrated models to understand the relationship between beliefs, attitudes, and

behavior, more personality and values variables have been used recently (Guttierez 1996;

Park et al. 1998; Thogersen 1986;).

"Unfortunately, personality variables (e.g., altruism) are seldom actionable from a

public perspective. More direct measures of recycling concern and recycling knowledge

seem to be more salient means of segmentation and may result in improved marketing

strategies. In addition, much of the existing research has operationalized ecological con-

cern in terms of attitudinal responses about environmentally sound activity, e.g., the use

of recycling centers. However, since progress toward solving environmental problems is

likely to be dependent on pro-environmental behaviors more so than ecological con-

sciousness (Van Liere and Dunlap 1981), researchers should focus on consumers' actions

with respect to the environment [here: recycling] rather than simply their attitudes [here:

clean, safe environments]. Recent investigations have used multiple measures of ecologi-

cal concern that include some behavioral component. Unfortunately, these behavioral

measures often tap consumers' purchase activities to the exclusion of other forms of

ecologically sound behavior (e.g., conservation activities)" (Polonsky et al. 1995). The

recycling discussion in the industry has focused on the biodegradability and bio-safety of

the organic product or packaging in stores rather than the recycling activity itself

Most recycling models have analyzed the cognitive (attitudinal) antecedents or

dispositions believed to guide the behavior (Hopper and Nielsen 1991; Kok and Siero

2985; Vining and Ebreo 1992). The most popular model in attitude research on recycling

behavior has been the Theory of Reasoned Action (Thogersen 1996).

Theory of Reasoned Action


Throughout the history of social psychology the concept of attitude has played a

major role in explaining human action, viewing attitudes as behavioral disposition (Ajzen

and Fishbein 1980). Since the early 1900s a number of theories have been developed to

provide a framework for the attitude-behavior relationship that would provide explana-

tory and predictive information.

Despite concerns by some, e.g., Allport (1935), early studies seemed to confirm

the validity of unidimensional effects of attitudes on behavior. Findings, such as the one

by LaPiere (1934), raised doubts about this assumption. With the accumulation of nega-

tive results, alternative influences on behavior and explanations for the failure of attitude

as a predictor were needed. "By the late 1950s, a multicomponent view was adopted and

attitudes were viewed as a complex system comprising the person's beliefs about an ob-

ject, his feelings toward the object, and his action tendencies with respect to the object"

(Ajzen and Fishbein 1980).

One such theory, the theory of reasoned action (Fishbein 1967; Fishbein and

Ajzen 1975), suggested that a person's behavior is determined by his intentions to behave

in a specific way. His intention is, in turn, influenced by the person's attitude toward the

behavior and the perception of social pressures imposed to perform the behavior (Ajzen

and Fishbein 1980) (Figure 2-1)

Behavioral intention represents an individual's motivation to attempt to engage in

a certain behavior. The stronger a person's intention to perform the behavior, the greater

is the likelihood that it will happen. For instance, if the resident of a given city states that

he/she is extremely likely that he/she will recycle glass bottles, then it is conceivable that

he/she will ask for recycling bins and dispose of all empty glass bottles separate from the

other household waste.

The attitude toward performing the behavior is on average measured with a sim-

ple method of the semantic differential. Attitudes toward a concept in the model are re-

garded as the person's feelings of favorableness or infavorableness for that concept. The

perception of social pressures, also known as subjective norms, deals with the influence

of the social environment on intentions and behavior. It refers to and asks for a person's

perception that important others (known as referents) think that the person should or

should not perform the behavior in question (Fishbein and Stasson 1990).

Attitudes toward the behavior are determined themselves by behavioral beliefs

and evaluations of consequences, emanating from those beliefs (Figure 2-1). It is impor-

tant to note that within this model the object of the belief is the behavior of interest and

the associated attribute is a consequence of the behavior. The interest is not in a person's

beliefs about, say, the "church", but rather in the person's beliefs about "attending church

this Sunday". According to Fishbein and Ajzen (1975) these are the only attitudes that are

directly relevant for predicting and understanding human behavior. As defined by Ajzen

and Fishbein (1980),

"Attitudes are based on the total set of a person's salient beliefs. People usually
believe that performing a given behavior will lead to both positive and negative
consequences; their attitudes toward the behavior correspond to the favorability or
unfavorability of the total set of consequences, each weighted by the strength of
the person's beliefs that performing the behavior will lead to each of the conse-
quences" (p. 67).

Subjective norms are also a function of a person's beliefs, but in this case they are

not behavioral, but normative beliefs. They are measured by multiplying a person's belief

that specific referents think he/she should (or should not) perform the behavior with the

person's general motivation to comply with each referent (Figure 2-1).


Motivation to Normde


Figure 2-v1. Path model for the Theoy of Reasoned Action

Ajzen and Fishbein (1980) do not deny that other variables, such as age, educa-
tion, or personality traits, may be related to behavior. Unlike other behavavioral Behavior
they argue that "external variables will be related to behavior only if they are related to

one or more of the variables specified by our theory" (p. 82). The fields of social issues
and altruism, for instance, are areas, where personal traits are frequently used. Azen and

Fishbein argue that discussed personality traits are too generic to relate them to a specific
Motivation to Nor

behavFigure 2-1. Personalty traits (or valuthe Theory of Reasons) are usually viewed as a predisposition
Ajzen and Fishbein (1980) do not deny that other variables, such as age, educa-

tion, or personality traits, may be related to behaviors (eg., aggressiveness, caringlike other behaviors), but not any specific
they argue that "external variables will be related to behavior only if they are related to

one or more of the variables specified by our theory" (p. 82). The fields of social issues

and altruism, for instance, are areas, where personal traits are frequently used. Ajzen and

Fishbein argue that discussed personality traits are too generic to relate them to a specific

behavior. Personality traits (or value orientations) are usually viewed as a predisposition

toward a class of behaviors (e.g., aggressiveness, caring behaviors), but not any specific

action. While someone might, for example, be generally pro-environmentally predis-

posed, that same individual might still not recycle. However, Ajzen and Fishbein do grant

the question about the origins of behavioral beliefs (1980, see p. 90).

Application in Recycling Research

The Theory of Reasoned Action has been used successfully in the past 20 years in

a variety of behavioral outcome or intention research, both in naturalistic and experimen-

tal settings. A number of studies have examined the predictive power of the model to ex-

plain particular behaviors, as well as tested correlations between the variables in the

model. The theory was, for instance, used to explain reenlisting in the military (Shtiler-

man 1982), voting behavior (Fishbein, Jaccard, Davidson, Ajzen and Cohen 1980), hav-

ing an abortion (Smetana and Adler 1980) and breast-feeding vs. bottle-feeding of babies

(Manstead et al. 1983).

Within the realm of recycling behavior the theory of reasoned action has been

used to explain the influence of education on recycling intentions of students (Goldehar

1991), the differences in recycling behaviors between ethnic subgroups (Gamba 2000),

and the relationship of self-perceptions on recycling behavior (Park, Levine and Sharkey

1998). Two of these studies have applied the Theory of Reasoned Action with success to

explain recycling behavior, the third (Gamba 2000) found less support for the theory.

Goldenhar (1991) conducted two studies testing the influence factors on recycling

of college students, specifically those that live in on-campus dormitories. In the first

study, she tested in a decision-making model, how well the theory explains recycling be-

havior. In the second study she incorporated two types of interventions (educational and

feedback), which were developed to modify recycling attitudes, beliefs, and behavioral

intentions, in order to enhance recycling behavior. A questionnaire comprising the con-

cepts of the Theory of Reasoned Action was administered to 4,682 first-year students at

The University of Michigan. Baseline data were gathered from 3,706 out of 4,682 stu-

dents (80% response rate). Of those 3,706 students, 1,604 students also completed the

follow-up questionnaire (34% response rate overall). For her second study she used a

quasi-experimental design over eight residence halls to match them on size and randomly

assign them to one of four intervention conditions (two halls per group): (1) recycling

education, (2) feedback about recycling behavior, (3) education plus feedback, or (4) con-

trol. The intervention period lasted 5 months. Path analysis, used in the first study, indi-

cated that the Theory of Reasoned Action was useful in explaining self-reported recycling

behavior. The respondents' rated importance of recycling compared to other social issues

mediated the relationship between attitudes, beliefs, and behavioral intentions. Utilizing

multiple comparisons in the analysis of the second study, her results showed that there

were no significant group differences in terms of the students' attitudes, beliefs, rated im-

portance, recycling knowledge, or behavioral intentions. Students receiving monthly

feedback pertaining to the amount of material recycled in their residence, however, re-

ported participating in recycling to a greater degree than those receiving only the educa-

tional intervention or nothing at all.

Park, Levine, and Sharkey (1998) examined behavioral intentions to recycle

among students in Hawaii, using the Theory of Reasoned Action as a framework. Based

on prior findings that attitude toward the behavior is a better determinant of intentions to

recycle than subjective norms, they speculated that an individual's self-image (called

self-construals in the study) will have an influence on the weight of attitudinal and nor-

mative influence on intentions to recycle. Accepting the original theory, the interest of

Park, Levine, and Sharkey lies not so much in the relation between attitude toward the

behavior and subjective norm, but in the relative weight of each component in the theory.

Based on results from previous studies that found gender differences for condom

use (Greene, Hale, and Rubin 1997) and cultural differences between countries for prod-

uct purchase influences (Lee and Green 1991), Park et al. hypothesized that one's self-

perception in relation to others influences the attitude and norm variables of the Theory

of Reasoned Action. Data were gathered from 201 undergraduate students enrolled in up-

per division classes at the University of Hawaii with a diverse ethnic makeup (25% Japa-

nese, 18% Chinese, 14% Caucasian, 13% Filipino, 7% Hawaiian, 2% Hispanic, 2% Afri-

can-American, 19% other) to determine if the different culturally imposed self-images

have indeed an influence on recycling intentions.

While the researchers found that their test of the Theory of Reasoned Action sup-

ported the original predictions of the theory, the data were not consistent with the hy-

potheses raised involving self-construals. Instead, self-construals had direct effects on the

attitudes toward behavior and subjective norm measures. In other words, even though

self-construals affected attitude toward behavior and subjective norm, they did not influ-

ence either the relation between the two components or the relative weight of the two

components in predicting behavioral intention. Systematic effects on subjective norm,

however, revealed an effect of self-construals. The more interdependent one's self-

construal was that is, the more one aligns one's self-image with expectations and values

of others (a concept closely related to the "locus of control" idea) the higher the scores

were on subjective norm. Park et al. (1998) found that "these higher scores were a func-

tion of higher scores on motivation to comply" (p. 203). Though the data were not consis-

tent with their original assumptions of an influence of self-construals on the relative

weight of the attitudinal and normative factors, the researchers drew an interesting con-

clusion. Since self-construals have an obvious direct influence on the motivation to com-

ply factor, it appears that individual with high interdependent selves are more susceptible

to messages targeting both attitudinal and normative components, as those individuals see

the positive social consequences of recycling as more likely. Individuals high in inde-

pendence would most likely be better targeted with messages, aiming at behavioral out-

comes alone (p.206).

In a study that analyzed how different ethnic groups in a city engage in household

recycling, Gamba (2000) used the Theory of Reasoned Action as a predictive model. He

specifically examined the similarities and differences between Latino, European-

American, Asian, and Filipino residents of San Francisco in their recycling attitudes,

norms, intentions, and observed behaviors. A mail survey was conducted and observa-

tions of curbside recycling were made (walk-along on collection day, ride-along with col-

lection trucks) in selected areas of San Francisco for eight weeks. Data were gathered for

1092 respondent households. Gamba found that recycling participation was relatively

high and no discernable differences were found among the cultural groups. Unlike the

previous studies, Gamba found less support for the Theory of Reasoned Action in his

study. His regression analysis revealed less explanatory power of attitudes and subjective

norms on intention to recycle (5% for subjective norm and 14% for attitude). He also

found little variance in observed recycling behavior explained by a respondent's inten-

tion, although the latter did predict self-reported recycling well. He asserted that the fact

that recycling participation among his sample was already high and a widespread practice

in this urban area, the assumed model showed less empirical support. He also observed

that the mailing of the questionnaire alone and the follow-up reminders produced a sub-

stantial observed increase in average weekly participation for all cultural groups over the

duration of the study. This demonstrates the effect of making beliefs more salient in peo-

ple's mind. In conclusion, he suggests that a program should emphasize an individual's

intentions to recycle and basic knowledge of the program, stress the ease of participating

and use simple reminders as possible intervention strategies to increase curbside recy-

cling (p. 158).

Overall, findings from the three studies suggest that there is a relationship be-

tween an individual's recycling attitudes, beliefs, and behavior. In addition, feedback and

educational intervention strategies, as well as self-images and values constructs appear

useful in explaining and enhancing recycling behavior.

Based on the empirical research by Goldenhar (1991), Park, Levine, and Starkey

(1998), and Gamba (2000), the following premises are suggested as a foundation for the

current study's hypotheses and research questions:

1) Intentions to recycle are on average a sufficient predictor for actual recycling behav-
ior, in case intention is measured on an aggregate level.

2) Attitudes and subjective norms about recycling are influenced by personal and cul-
tural constructs, such as self-perceptions and values.

3) Attitudes and subjective norms alone are necessary but not sufficient determinants of
recycling intentions.

Move toward the Theory of Planned Behavior


The Theory of Reasoned Action was developed explicitly to deal with purely voli-

tional behaviors (Ajzen 1988). Problems arise, when the theory is applied to behaviors

that are not entirely under a person's volitional control. A well-known case in point

would be the failed attempt of people to quit smoking, although they seriously intended

to do so. Failure to enact the behavior may occur either because of a change in intentions

or because performance of the behavior failed.

A number of researchers have focused on the question of volitional control (Ban-

dura 1977; Kuhl 1981). Perhaps the best known example is the concept of internal and

external locus of control (Rotter 1966). It refers to the belief that one's outcomes are ei-

ther under the control of one's own behavior (internal) or under the control of such fac-

tors as powerful others or chance (external) (Ajzen 1988). Bandura (1977) introduced the

concept of perceived self-efficacy. The concept refers to the subjective probability that

one is capable of executing a certain course of action. In a somewhat related analysis of

action control, Kuhl (1981) introduced the concept of state versus action orientation, a

concept close to willpower. Action-oriented people are assumed to focus their attention

on action alternatives and to make use of their abilities to control their performance. In

contrast, state-oriented individuals focus their attention on their thoughts (their present,

past, or future) rather than taking action consistent with their intentions (Ajzen 1985).

Closely related to self-efficacy beliefs is Ajzen's (1985) concept of perceived be-

havioral control, a variable that is defined as one's perception of how easy or difficult it

is to perform the behavior (Eagly and Chaiken 1993). Ajzen (1985) states,

"the success of an attempt to execute the behavioral plan depends not only on the
effort invested (the strength of the attempt), but also on the person's control over
other factors, such as requisite information, skills, and abilities, including posses-
sion of a workable plan, willpower, presence of mind, time, opportunity, and so
forth" (p. 36).

Ajzen proposed an extension of the Theory of Reasoned Action, the Theory of

Planned Behavior (Ajzen 1985; Ajzen and Madden 1986). The addition of the third ante-

cedent of intention is the degree of perceived behavioral control. As a general rule, the

greater the perceived behavioral control, the stronger is the intention to perform the be-

havior under consideration. For example, if a person wants to recycle and thinks he or she

has control over this behavior (recycling bins are readily available, there is no extra cost,

recyclables can be put out together with the regular waste), the person is more likely to

actually recycle (Figure 2-2).

In summary, the four directly measured variables are: (1) behavioral intention

(BI), (2) attitude toward the behavior (A), (3) subjective norm (SN), and (4) perceived

behavioral control (PBC). These variables form the following equation:

BI (A + SN + PBC) = wlA + w2SN + w3PBC

with wl, w2 and w3 representing the relative contributions (weights) of attitude, subjec-

tive norm, and perceived behavioral control, respectively, to the prediction of behavioral

intention (Ringer Lepre, 2000).

Beliefs Attitude

Evalons Behavioral Behavior

Motivation to Norm

Beliefs Perceived

Figure 2-2. Path model for the Theory of Planned Behavior

The addition of the perceived behavioral control variable raises two interesting

questions regarding its meaning and value for the model. First, perceived behavioral con-

trol is added as an exogenous variable that has both a direct effect on behavior and an in-

direct effect on behavior through intention. According to the construction of the original

Theory of Reasoned Action, Ajzen and Fishbein (1980) argued that "the effects of exter-

nal variables are mediated by beliefs, and therefore, taking external variables into account

is not expected to improve prediction of intention (...) or behavior" (p. 91). Since the ad-

dition of the control variable does improve the prediction of intentions, it seems to open

the door for the addition of other "external" variables that could strengthen the model.

Second, Ajzen's claim of perceived behavioral control's synonymity with self-

efficacy as defined by Bandura (1977) has met with criticism (Fishbein and Stasson

1990). These studies assert that self-efficacy is a more internally based notion within an

individual. This contrasts with perceived behavioral control which includes an influence

by others or events. It should be plausible to allow for other internally-located factors

next to self-efficacy, such as willpower, interest, or ascription of responsibility, to flow

into the measure of perceived behavioral control. Finally, there seems to be some merit to

divide the perceived behavioral control variable into an external component (as defined

by Fishbein and Stasson (1990)), and an internal component (as defined by Bandura

(1977)). This would not influence the original theory substantially. Both of the other two

crucial variables- attitudes and norms are likewise comprised of two antecedent com-

ponents that form this variable.

More research seems needed at this point to clearly determine the appropriate

measure for use in the theory of planned behavior. There seems to be some value in

including measures of both external and internal control elements.

Based on the theoretical underpinnings of the Theory of Planned Behavior (Ajzen

1985), the following premise can be introduced as a foundation for this study's research


4) With the addition of the perceived behavioral control element in the Theory of
Planned Behavior the predictive power of the original Theory of Reasoned Action
model is increased, allowing for cases in which the behavior (recycling) is not under
complete volitional control.

Application in Recycling Research

Similar to the Theory of Reasoned Action, the Theory of Planned Behavior has

been used in multiple studies to explain behavior. Ajzen and his colleagues (Ajzen and

Madden 1996; Schifter and Ajzen 1985) were among the first to empirically test the the-

ory. These studies dealt with weight loss and class attendance topics. In either case the

new variable of perceived behavioral control showed strong predictive power in explain-

ing intention to behave in a certain way. The overall predictive ability of the Theory of

Reasoned Action was substantially improved. As a result, the Theory of Planned Behav-

ior appeared to be an improvement over the Theory of Reasoned Action to explain inten-

tion and actual behavior.

In the area of recycling and green consumerism, two studies applied the Theory of

Planned Behavior to explain the antecedents of recycling and composting intentions

within an integrated waste management behavior model (Taylor and Todd 1995), and the

influence of self-identity on attitudes and intentions to engage in shopping for organic

products (Sparks and Shepherd 1992).

Sparks and Shepherd (1992) based their study on reports that a relationship exists

between self-identity concepts and behavioral intentions that is independent of the role of

attitudes toward the behavior or social norms. Since all these studies seemed to have been

discussed in reference to the Theory of Reasoned Action, Sparks and Shepherd hypothe-

sized that the concept of self-identity could be covered well by the added variable of per-

ceived behavioral control. In other words, an adequate operationalization of the compo-

nents of the Theory of Planned Behavior would result in no independent relationship be-

tween a measure of self-identity and a measure of behavioral intentions.

To test the hypothesis, 236 randomly sampled members of the general public in a

medium-sized town in England returned a mailed questionnaire. The questionnaire in-

cluded the standard variables of the Theory of Planned Behavior (salient beliefs, outcome

evaluations, attitudes, subjective norm, perceived control, and intentions) as well as

measures of identification with green consumerism and health-consciousness and a

"green concern" index. Contrary to their expectations, the analysis revealed a substantial

independent effect for self-identity, an effect that persisted when a measure of past con-

sumption was included in the equation. They tentatively concluded that psychological

identification (pro-environmental self-concept) reflects more than an inference from past

behavior and acts as more than an index of values concerning external consequences of

action (p. 394).

This, in turn, supports proposals that both the Theory of Reasoned Action and the

Theory of Planned Behavior need to take account of the role of self-identity in influenc-

ing behavioral intention. While both the Theory of Planned Behavior and the Theory of

Reasoned Action were successfully replicated, the above issue raises questions for mod-

els of attitudes based on "expected utility" because choices seem to be influenced by a

multitude of considerations. These would include the socially or culturally fashioned

symbolic meanings of those choices, such as the social identities that the choices might

help to confer (p. 397). Identity-related symbolic outcomes that are supported by different

choices would likely be of great importance to people and their particular social milieu

(Giddens 1991).

Within the concept of the Theory of Reasoned Action and Theory of Planned Be-

havior, attitudes are formulated on the basis of utilitarian outcomes. These outcomes are

the result of a cost/benefit analysis of the individual. A closer examination of how atti-

tudes relate to subjective expected utility underpinnings of the Theory of Reasoned Ac-

tion and Theory of Planned Behavior is needed. There may be evidence to either question

the usefulness of the measured attitudes, or add an antecedent dimension to the model

that can sufficiently explain the formation of those attitudes.

Taylor and Todd (1995) were concerned primarily with the critique that the recy-

cling behavior literature lacks an integrated theoretically based model to understand the

relationships between environmental beliefs, attitudes, and behavior (Hopper and Niel-

sen, 1991). Taylor and Todd created an integrative model, based on the Theory of

Planned Behavior, which also included perceived innovation characteristics (Rogers

1983), facilitating conditions (Triandis 1979), and self-efficacy (Bandura 1977) as key

determinants of recycling intentions and behavior. The latter two variables were posi-

tioned as direct antecedents to Ajzen's perceived behavioral control variable within the

model. The first variable is based on beliefs about the perceived characteristics of an in-

novation (Rogers 1983).

According to the innovation literature, three perceived characteristics of an inno-

vation have been found to influence adoption behavior: relative advantage, complexity,

and compatibility. Since relative advantage and complexity have been found to be impor-

tant predictors of attitude, Taylor and Todd expected those characteristics to influence

attitude formation in the context of the Theory of Planned Behavior. Compatibility, a

component of facilitating conditions (Triandis 1979), was estimated to influence per-

ceived behavioral control. In the altered design of the Theory of Planned Behavior model,

relative advantage refers to the degree an innovation provides benefits that supersede

those of its precursor (a concept consistent with the notion of perceived costs and bene-

fits). Complexity represents the degree to which an innovation is perceived to be difficult

to understand and use (p. 611). Among the new three control variables, forming the per-

ceived behavioral control structure, facilitating conditions relate to access to resources

necessary to perform the behavior. Self-efficacy correlates to the perceived ability to

carry out the behavior. Perceived compatibility is defined as the degree to which the in-

novation fits with the potential adopter's existing values, lifestyle, previous experiences,

and current needs (p. 612).

After pilot-testing the constructs, data were gathered through a survey over 761

respondents in a mid-sized city for both recycling and composting intentions. Both fit sta-

tistics and path analyses suggested that the integrative model explained the assumed func-

tioning of recycling intentions well. Taylor and Todd 1995) pointed out that intentions to

recycle were positively influenced by attitude and perceived behavioral control, but were

negatively influenced by subjective norm. The somewhat surprising result in regards to

the subjective norm was assumed to be due to the relative maturity of the recycling pro-


This argument was also raised by Gamba (2000) as a potential interpretation of

his inconclusive findings regarding the Theory of Reasoned Action effects on household

recycling. Normative influences were important determinants of subjective norm, ex-

plaining 75% of its variance (p. 620). Finally, efficacy and resource-facilitating condi-

tions were positively related to perceived behavioral control, though compatibility was

not. Taylor and Todd argue that although recycling does not seem to be perceived by

people as being compatible with their daily routines or lifestyles, it did not weaken the

control they felt over their behavior. It suggests that, given adequate knowledge, people

may be willing to overcome personal inconvenience to realize the more global benefits of

recycling (p. 620). Taylor and Todd maintain that while the original Theory of Planned

Behavior is a useful starting point, the integrated waste management model can provide a

better understanding of the complex relationships that influence waste management inten-

tions and subsequent behavior.

Based on these empirical research studies, applying the Theory of Planned Behav-

ior, the following premises are added as foundations for this study's hypotheses and re-

search questions:

5) The addition of perceived behavioral control in studies predicting recycling intentions
and behavior has shown to improve predictability of the Theory of Reasoned Action.

6) Attitude, subjective norms, and perceived behavioral control all seem to provide
equally significant explanatory power for behavioral intentions and behavior.

7) A stricter separation of the perceived behavioral control variable into control beliefs
and perceived external facilitation conditions will strengthen this variable.

8) The inclusion of antecedents to the attitudinal, normative, and control beliefs in the
form of self-concepts or personal values has been found to improve predictive ability
of the entire model.

Personal Values and Recycling


As the two empirical studies using the Theory of Planned Behavior have shown, a

growing number of researchers have begun to study the development of environmental

attitudes as well as underlying cultural or personal values systems, influencing recycling

behavior directly. Research streams emerged from the literature on norm activation the-

ory, perceived risk, self-concept effects, psycho-social variables, and the new environ-

mental paradigm (NEP). Each stream can be considered a relatively large autonomous

field of its own. Together they may be viewed as parts of a larger, more generalized

framework, which incorporated ideas about nature of values proposed by Schwartz

(1977, 1992) and Rokeach (1973, 1979). Schwartz's theories have been applied by Stem

(1987, 1992) within the context of environmental concern research (Young 1997). The

following discussion assesses values research and explains and defends its use in a modi-

fied model of the Theory of Planned Behavior within the current study. This will provide

a conceptual framework.

Values Research

Values and norms can predispose individuals to hold certain attitudes and react in

predictable ways toward environmental problems (Dunlap and Van Liere 1978; Stem,

Dietz and Kalof 1993). It seems that there are compelling theoretical reasons for assum-

ing that the study of a person's values is likely to be useful. The study will use the defini-

tion of Rokeach (1973), who stated:

A value is an enduring belief that a specific mode of conduct or end-state of exis-
tence is personally or socially preferable to an opposite or converse mode of con-
duct or end-state of existence. A value system is an enduring organization of be-
liefs concerning preferable modes of conduct or end-states of existence along a
continuum of relative importance (p. 5).

The key elements of this definition are comprised of the words "enduring", "be-

lief', and "end-state of existence". The endurance quality of a value stems from the fact

that a value is learned initially in isolation from other values in an absolute manner. One

can, to use the example of "honesty" as a value, not be just a little honest. At the same

time, one can also not be sometimes honest, and sometimes not. Honesty, as an example

for an end-state of existence, is always desirable over others. Although people learn

through experience and maturation to integrate values into a hierarchical system, the be-

havioral outcome in a specific scenario will be determined by the relative importance of a

specific value compared to others. For example, honesty as a value might be subordinated

to another value (say: freedom), if this value is seen as more important in the situation,

but it will never be compromised in its own right.

A value is also considered a prescriptive, or proscriptive, belief, or "a belief upon

which a man acts by preference" (Allport 1961). Values have cognitive, affective, and

behavioral components (Rokeach 1973). "It has a behavioral component in the sense that

it is an intervening variable that leads to action when activated" (Rokeach 1973, p.7).

The distinction of beliefs toward a mode of conduct versus end-states of existence

separates values into two kinds: instrumental values and terminal values. Since values

form a functionally interconnected system, once they are internalized, subsequent values

research (Kahle 1984; Schwartz 1992) has not maintained this strict separation.

Values researchers have concluded that values are multi-faceted standards. Values

help individuals to rationalize beliefs, attitudes, and behaviors that would otherwise be

personally or socially unacceptable so that one will end up with personal feelings of mo-

rality and competence, and in the end enhanced self-esteem.

An attitude refers to an organization of several beliefs around a specific object or

situation (Rokeach 1968). In contrast, a value refers to a single belief of a very specific

kind. Attitudes are focused on some specified object or situation, while values transcend

them. Since values are also considered standards, applying to all kinds of situations, they

are believed to occupy a more central position than attitudes within one's personality

makeup and cognitive system. Values are determinants of attitudes as well as behaviors.

Values also differ from social norms. A value refers to an end-state of existence

and transcends specific situations. In contrast, a social norm refers to only one mode of

behavior connected in a prescriptive fashion to a specific situation. For example, Navaho

Indians should refrain from having ceremonials at the time of an eclipse of the moon

(Kluckhohn 1951). This behavior is subject to sanctions from the Navaho society, and

only apply to the eclipse scenario. Second, a value is more personal and internal, whereas

a norm is consensual and external to the person (Rokeach 1973).

It can be assumed that values may have a rightful place as antecedents to beliefs,

attitudes, norms, and actions in a model, such as the Theory of Planned Behavior.

Schwartz Values Model

Schwartz (1970) proposed that people are aware and concerned with others' well-

being and consequently act out of a sense of moral obligation to help others. In other

words, they act altruistically. "These moral norms may be internalized wholly or par-

tially, or they may be perceived as expectations held by significant others" (Schwartz

1970, p. 130). Schwartz's term for a non-internalized norm is "social norm", which is de-

fined the same as in the Theory of Reasoned Action. His term for an internalized norm is

"personal norm". He referred to it as "internalized values" (Schwartz and Howard 1980).

People, who, e.g., hold a great concern for the environment, generally have a

great concern for others' welfare as well. If they are aware of adverse consequences to

others, as a result of their behavior, they behave in a pro-environmental fashion. They

will do so as well, if they ascribe a personal responsibility to themselves to act altruisti-

cally and reduce the negative consequences. Thus values influence behavior when they

are activated by situational concern (Karp 1996).

Schwartz (1992, 1994) extended his research on values, administering a global

survey with Likert-type questions that inquired on the importance of 56 value items as

"guiding principles" in respondents' lives. His new theory of values was predicated on

the Rokeach scale. He identified ten motivational goals (e.g., conformity, security, he-

donism), which were further collapsed into four identifiable clusters, representing the ex-

tremes of two basic conditions (Young 1997). This is illustrated in Figure 2-3.

Self-Transcendence 4 -Self-Enhancement
To Change
Self-direction Achievement

Conformity Power


Figure 2-3. Schwartz values dimensions (with 4 exemplary motivational types)

The openness to change versus conservation dimension indicates the degree to

which individuals are motivated to independent action and willing to challenge them-

selves for both intellectual and emotional realization (Karp 1996).

[The dimension] arrays values in terms of the extent to which they motivate peo-
ple to follow their own intellectual and emotional interests in unpredictable and
uncertain directions versus to preserve the status quo and the certainty it provides
in relationships with close others, institutions, and traditions (Schwartz 1992, p.

The second dimension contrasts values oriented toward the pursuit of self-interest

(self-enhancement) with values related to a concern for the welfare of others (self-


It arrays values in terms of the extent to which they motivate people to enhance
their own personal interests (even at the expense of others) versus the extent to
which they motivate people to transcend selfish concerns and promote the welfare
of others, close and distant, and of nature (Schwartz 1992, p. 43).

As traditionally understood, the concern for others in the environmental literature

usually refers to people. However, if the subject holds "ecological values", there is no

reason why the "other" could not be nonhuman (Thogersen 1996). Consequently, some

scholars include "biospheric altruism" behaviors judged with reference to ecological

values within the domain of morality (Ster, Dietz and Kalof 1993). Overall,

Schwartz's work on individual values will be used as a basis for the current study in re-

gards to environmental values, related to recycling.

Applications in Environmental/Recycling Research

The Theory of Reasoned Action and the Theory of Planned Behavior have be-

come the most popular analytical framework for recycling and more general proenviron-

mental research. Both are a variety of Subjective Expected Utility (SEU) models that as-

sume that action is motivated by a desire to maximize private utility (Thogersen 1996).

The general assumption for recycling research was that action is indeed either triggered

by selfish motives (gains, cost avoidance) or adherence to accepted social norms in the

society. Scholars, such as Thogersen (1996), have argued that recycling should be treated

as an instance of prosocial behavior, because of its benefits to society and the environ-


"In affluent industrial societies, environmental behaviors like recycling are typi-
cally classified within the domain of morality in people's minds. Attitudes regard-
ing this type of behavior are not based on thorough calculation, conscious or un-
conscious, of the balance of costs and benefits. Rather, they are a function of the
person's moral beliefs, that is, beliefs in what is the right or wrong thing to do"
(Thogersen 1996, p.537).

This alternative, theoretical approach that seeks to explain behavior with values

and morality has been applied in studies, testing Schwartz's altruism model (Guagnano,

Ster and Dietz 1995; Hopper and Nielsen 1991; Vining and Ebreo 1992), and more ad-

hoc based models (Derksen and Garttrell 1993; De Young 1986; Stem, Dietz and Kalof

1993). Two studies applied the concept ofproenvironmental values and personality traits

directly to explain proenvironmental behavioral intentions. Karp (1996) analyzed the in-

fluence of value orientations on the intention to protect the environment. Allen and Fer-

rand (1999) applied the influence of Geller's concept of "actively caring for the environ-

ment" (1995) to determine proenvironmental behavior.

Karp (1996) based his study on Schwartz's theory and measure of values (1992),

accepting the assumption that values play a role in specific situations when activated by a

set of altruistic concerns. The idea of an impact for specific situations was criticized by

attitude theory (Ajzen and Fishbein 1980). Subsequent studies (Ster and Dietz 1994)

have demonstrated that values are a good predictor of specific behaviors. The goal of

Karp's study was to clarify the role of values in predicting environmental behavior by

using a complete Schwartz Scale of Values to test the effect of values. Values are arrayed

along two dimensions, one spanning between self-enhancement and self-transcendence,

another spanning between openness to change and conservation. Within this 2x2 values

dimension matrix exist ten value categories, called motivational types by Schwartz.

Those include values, such as benevolence, security, power, and hedonism, among oth-

ers. Karp hypothesized that individuals, who hold self-transcendence and openness to

change values are the most likely to engage in proenvironmental behavior. Those that

combine self-transcendence with conservation values engage in proenvironmental behav-

ior, if based on a normative standard (rules).

Finally, those that hold self-enhancement values are less likely to act in a proenvi-

ronmental fashion. This behavior is only modified by the openness to change dimension.

Openness to change might lead self-enhancement individuals to engage in proenviron-

mental behaviors, if there is a link to self-interest, e.g., buying organic food for health

reasons (p. 116). In the study Karp conducted, he measured values along the Schwartz

(1992) Scale of Values (a 9-point scale, ranging from "opposed to my values" to "of su-

preme importance to me") and proenvironmental behavior on a self-reported scale of ac-

tivities (a 5-point scale, ranging from "never" to "always"). Through factor analysis the

behaviors were reduced to three categories, called "Good Citizen" (frequent proenviron-

mental behaviors), "Activist" (infrequent proenvironmental behavior), and "Healthy Con-

sumer" (targeted proenvironmental behavior).

The main finding of the study was that the effect of self-transcendence and open-

ness to change, as well as a biospheric motivation type, has a strong positive effect on all

three categories of environmental behavior. The other combinations of values dimen-

sions, however, were not fully supported by the findings, except for the extreme opposite

values combination to the above (self-enhancement and conservation), which showed a

negative effect on the composite behavioral category as well as the "Activist" category.

Overall, Karp found that an understanding of the relationship between values and

behavior increases the likelihood of determining what triggers people to engage voluntar-

ily in proenvironmental behaviors. This would help to engage in programs that promote

noncoercive solutions to problems rather than forcing a behavior to avoid a free-rider ef-

fect. Since the findings were no entirely conclusive, Karp suggests that in addition to us-

ing values alone, one should consider addressing rational considerations (perceived indi-

vidual efficacy, estimation of consequences) as well (p. 131). This suggestion does point

to the usefulness of a merging of two theory streams a theory based on rational choice,

such as the Theory of Planned Behavior, with a values-based theory, such as the

Schwartz model.

Related to the idea of perceived efficacy is the concept of actively caring, devel-

oped by Geller (1995). The actively caring perspective is analogous to the concept of

self-transcendence (Maslow 1971; Schwartz 1992). Actively caring explains behaviors,

executed to benefit others, make others feel better, or influence other's behavior in a de-

sired direction (Geller 1995). Geller developed a model of actively caring, a form of al-

truistic motivation that primarily looked at the psychological internal determinants of an

"actively caring" attitude. Geller's actively caring variable mediates the relationship be-

tween environmentally responsible behaviors and personality factors related to self-

affirmation (self-esteem, belonging, self-efficacy, optimism, and personal control).

Allen and Ferrand (1999) tested this hypothesis empirically in a study on 121 stu-

dents in New York. Although Geller's model is related to Schwartz's norm activation

model of altruism (1977), it contains specific precursors lacking in the Schwartz model.

In addition, unlike the Schwartz model, which includes an adherence to a social norm,

Geller's notion of caring is directly arrived from an internal feeling of sympathy and per-

sonal psychological makeup. Hence, data testing Schwartz's model would be inconsistent

with Geller's hypothesis (Allen and Ferrand 1999, p.341). The purpose was to directly

test Geller's model.

Allen and Ferrand (1999) asked 121 undergraduate students at a liberal-arts col-

lege in New York to fill out a lengthy questionnaire that assessed self-esteem, feelings of

belonging, sense of personal control regarding environmental problems, sympathy for

others, and the extent to which they engaged in a variety of environmentally friendly be-

haviors. To measure the predictors, existing scales were used, whenever possible, from

previous research. Personal control was measured with the Environmental Action Internal

Control Index (Smith-Sebastano 1992). Self-esteem was measured using the Texas Social

Behavior Inventory (Helmreich, Stapp and Ervin 1974). Belonging was measured using

the Social Connectedness Scale from Lee and Robbins (1995). Since Geller had not iden-

tified a specific measure of actively caring, the researchers chose Davis' (1983) measure

of sympathy.

Path analyses were conducted to test the mediational aspect of Geller's model

(p.344). Allen and Ferrand's findings suggest that sympathy is an important predictor of

environmentally responsible behavior. Sympathy is facilitated by feelings of personal

control, which supports Geller's notion that actively caring mediates the relation between

self-affirmation factors (e.g., personal control) and environmentally friendly behavior.

Unlike personal control though, self-esteem and belonging did not predict environmen-

tally friendly behavior well. The researchers speculated that their measure of personal

control was the one, most specifically related to environmental problem solving and con-

cern. People felt specifically empowered or self-affirmed in relation to environmental is-

sues. If it is the case that self-empowerment feelings unlike self-affirmation goals need to

precede dispositional factors to predict behavior, Geller's theory (1995) does not explain

actively caring properly and needs to be changed.

Similarly to Karp (1996), Allen and Ferrand caution as well to interpret their find-

ings in a way that suggests that environmentally friendly behavior is entirely a function of

sympathy and altruism. As Ster, Dietz and Kalof(1993) demonstrated in finding egois-

tic motives in addition to altruistic motives as antecedents of environment-protecting be-

havior, environmentally friendly behavior appears to be multiply determined.

Based on these empirical research studies, applying the concepts of values and

personality traits, the following premises are added as foundations for this study's hy-

potheses and research questions:

9) Value orientations seem to provide explanatory power for the origins of behavioral,
normative, and control beliefs.

10)A succinct value orientation scale, such as the one provided by Schwartz, delivers
adequate explanation for proenvironmental behaviors, such as recycling.

11) Perceived behavioral control is sufficiently explained by internal beliefs of empow-
erment (self-efficacy) and beliefs about the extent of control of external factors.

12) Values influence behavioral intention about recycling through their proximal relation-
ship with attitudes, norms, and control perceptions.

Inclusion of Values in the Theory of Planned Behavior

Ajzen's Theory of Planned Behavior (1985) specifies in a mathematical way the

relationship among beliefs, attitudes, and behaviors (Petty and Cacioppo 1981). The the-

ory is based on the assumption that "humans are rational animals that systematically use

or process the information available to them" and that "the information is used in a rea-

sonable way to arrive at a behavioral decision" (Ajzen and Fishbein 1980).

The assumption of rationality has a favored position in economics (Tversky and

Kahneman 2000) and related disciplines. A society that is influenced by scientific think-

ing usually holds up the "rational choice model" of decision making as an ideal to which

we should aspire (Miller 1999). In other words, when confronted with an environmental

problem (such as solid waste), it is argued that we should develop a comprehensive un-

derstanding of the problem, explore all possible alternatives, engage in logical decision-

making, and seek evaluative feedback on the consequences of our actions.

In practice, "rationality" can take many forms. Being rational simply means that

one takes orderly steps toward achieving a reasonable coherent goal, as irrational as it

might appear to a neutral observer. Rationality is simply a mental model composed of

two broad sets of ideas, what people believe and value (their ideology), and how they

seek to achieve their valued goals (their preferred mode of reasoning or conduct) (Miller

1999). While two different people may use different rationalities, they are similar in that

their behavior is embedded in a set of values. It follows that all problem-solving behavior

is subjective; it cannot be "objective" in the sense of being totally detached from personal

and cultural values (p. 12).

This study does not argue that Fishbein and Ajzen assume a positivistic rationality

in their theory, it appears questionable why personal values as part of the "other vari-

ables" should be entirely exogenous to the model and can only indirectly (through be-

liefs) affect behavior or behavioral intent (Ajzen and Fishbein 1980, p. 82). Deducing

from the above arguments, it is possible that values can have a direct influence on behav-

ior, as demonstrated in the results of the previously tested empirical studies. In general,

researchers should consider including values as an antecedent internal variable set to be-

liefs with stable theoretical relations to behavior.

Furthermore, the main argument against the inclusion of personality traits which

would include values into the model has been that a general measure of, e.g., altruism

will not correlate well with any specific single behavior (p. 89). Values and attitudes are

organized in a hierarchical construct that renders values the determinants of attitudes. For

example, a held value, such as "a world of beauty" will be expressed in a specific situa-

tion in the form of a belief, such as "preference for a highway without litter". This might

ultimately result in a behavior, such as participating in an "adopt-a-highway" program to

act on the attitude. Granted, in this example the flow is sequential, and no direct value-

behavior relationship is present. However, there is no reason to exclude values from the

"value-belief-attitude-intention-behavior" sequence of operation, just because it is the

antecedent to the following.

In their own work, Ajzen and Fishbein (1980) opened the door for questions re-

garding the origins of beliefs (p. 90). The current study argues that values are the origins

of the beliefs and should be included in the model as an internal variable.

Proposed Model

Previous studies have demonstrated that values research applied to environmental

issues, including recycling, has become more robust. The work by Stern, Dietz and Kalof

(1993) incorporated Schwartz's work (1992) on universal values into a specification of

environmental concern (Young 1997). The combined model which Young (1997) has

called the Stern-Schwartz model proposes that individuals make decisions and form

attitudes about environmental issues by processing these situations through a system of

heuristics, where values, beliefs, and attitudes influence an individual's propensity to act.

The key component for the current study is the interactive inclusion of values orientations

to the attitudinal, normative, and control determinants of behavioral intention in the

model of the Theory of Planned Behavior. The idea of values enhanced variables will be

borrowed for use in the present study.

Value orientations underlie all beliefs, attitudes, and behavioral intentions; thus

they are postulated as causally antecedents to all other variables within the modified The-

ory of Planned Behavior. People can hold multiple value orientations to certain degrees,

which can vary across individuals. Individual attitudes toward recycling emanate from

three value orientations.

First, an egoistic value orientation predisposes people perform a type of cost/

benefit analysis with regard to recycling. Persons take either pro- or anti-recycling

stances in accordance with their assessment of the personal costs associated with the

problem (Stern et al. 1993; Young 1997). This could include such actions as supporting a

citywide household recycling program only if utility fees or taxes are not adversely im-

pacted by it. Empirical work by Ster and Dietz (1994) has shown that the egoistic value

is conceptually and empirically equivalent to Schwartz's Self-Enhancement dimension.

Literature on environmental risk rests on this value orientation in particular (Douglas and

Wildavsky 1982; Wilson 1992). Kahneman and Tversky (1979) have shown that if deci-

sions include great risk, people tend to minimize losses rather than maximize gains. This

is consistent with an egoistic value orientation.

Second, an altruistic value orientation describes individuals' concern about the

impact of the waste problem and their non-recycling behavior on others. Individuals are

likely to act to reduce the negative effects. People with this value orientation recycle in

order to provide long-term availability of natural resources to future generations (Young

1997). A potential behavioral intention could include an active and ongoing recycling and

clean-up support of the neighborhood so that people within (e.g., children at play) are

safe from pollution and run-off. Studies by Ster and Dietz (1994) have shown that the

altruistic value orientation is included in Schwartz's Self-Transcendence dimension.

A related value, the biospheric value orientation, predisposes people to be con-

cerned about the consequences of not recycling on the earth itself This orientation is also

conceptually included within Schwartz's Self-Transcendence dimension. Biospheric val-

ues are essentially synonymous with the New Environmental Paradigm (NEP). This

worldview, developed by Catton and Dunlap (1978), understands humans as part of the

natural world and governed by its rules. Recycling decisions are made as a result of con-

cerns what non-recycling would do to the natural environment (poisoning, endangered

species, etc.). Slightly different from the altruistic value orientation, behavior resulting

from biospheric values could include recycling activities motivated by desires to keep the

environment itself (not fellow citizens) safe from toxins and non-degradable trash. Bio-

spheric values have played a prominent role in the thinking of environmentalists (Naess

1989). Empirical analyses (Ster & Dietz, 1994) have failed to reveal a clear distinction

in the general public between valuing nature in itself and valuing nature because of the

human benefit. In the current study, the altruistic and biospheric orientations will be re-

garded as one category.

Third, a traditional value orientation predisposes individuals to act according to

established, internalized norms and cultural paradigms. Decisions on recycling are made

based on an adherence to an agreed-upon status quo within the community that a person

belongs to ("this is how it's done around here"). This orientation is conceptually related

to Schwartz's Conservation dimension. This dimension could manifest itself in a recy-

cling behavior that is largely motivated by how someone grew up, and how the neighbor-

hood thinks about recycling.

Finally, Schwartz' fourth dimension, Openness-to-change, will not be used via a

related values orientation. The reason is that this dimension is implicitly part of both the

egoistic and altruistic values orientation in this model.

With all other variables equal to the Theory of Planned Behavior, the proposed

model is illustrated in Figure 2-4.

---- -J \I Behavioral
Stable theoretical rltions Perceived o Ctrol
-.- Facilitation
Possible explanations for rela-

Figure 2-4. Path model for the values-enhanced model

In summary, seven variables are measured directly. They are: (1) behavioral in-

tention (BI), (2) attitude toward the behavior (A), (3) subjective norm (SN), (4) perceived

behavioral control (PBC), and (5 through 7) the values (V). These variables form the fol-

lowing equation:

BI -V(A +SN+ PBC) =a fPVra,A + 2VrSN + 3jVgoPBC+e

with fi, P and a3 representing the relative contributions (weights) of attitude, subjective

norm, perceived behavioral control, enhanced by the dominant values respectively, to the

prediction of behavioral intention.

Three of the determinants of intention (A, SN, and PBC) are, in turn, determined

by underlying belief structures, while the values construct is determined by the most

dominant values orientation. Stated formally, A is the sum of attitudinal beliefs (abi) mul-

tiplied by an evaluation of their outcome (ei), that is,

A = Z abie,.

SN is the product of the individual's normative beliefs regarding the influence of a par-

ticular referent (nbj) and the motivation to comply with that referent (mc,), that is,

SN = E nbjmcj.

PBC is the result of the sum of beliefs about personal control, i.e. the perceived difficulty

(or ease) with which to execute the behavior, (cbk) multiplied by the perceived facilitation

of the control factor (pk), that is,

PBC = E cbkpfk.

Finally, the dominant values orientation that is linked to the respective determi-

nant, is the result of the different values, an individual possesses. The subscripts rt, o,

and t. refer to the rational, altruistic, and traditional value orientation (V), as follows:

Val = (E veg) (Y v.a, + vbio)
VIra = (X vI,)
Vego = (Y VgJ.)

Summary, Research Questions. and Hypotheses

Summary of the Literature

The review of the literature on pro-environmentalism and recycling provides the

background and structure for the current study. Some of the most important points are

summarized below:

1: Recycling is an activity that despite its social benefits and potential benefits to

the individual's future well being is not done universally.

2: People cite different reasons for why they do not recycle, such as lack of

opportunity, lack of knowledge, doubts about making a difference, degree of difficulty,

and lack of interest.

3: Information campaigns to entice people to engage in recycling have usually fo-

cused on risks of non-recycling (fear appeal) or ease of engaging in it.

4: Research about recycling discovered that both rational and moral/ethical

thoughts determine people's behavior.

5: The Theory of Reasoned Action has proven to be a model that is useful to de-

termine and predict the variables that influence behavioral intention on recycling and

proenvironmental behavior in general.

6: The Theory of Planned Behavior has proven to be a model that strengthens the

determination and prediction of variables, influencing behavioral intention on recycling

and proenvironmental behavior in general.

7: Values orientation has shown to determine belief and attitude orientation to-

ward a proenvironmental behavior, including recycling.

8: Beliefs about the control over one's behavior are determined by one's inner

self-empowerment thoughts as well as perceptions of control over external factors.

9: Both empirical studies, adhering to models of rationality and those adhering to

altruism, suggest a strengthening of their predictive power through borrowing ideas from

the opposite concept.


Throughout the literature review several premises have been proposed as a foun-

dation for the current study. They will be used to formulate the research questions and

hypotheses in the current study. Those premises are:

1) Intentions to recycle are on average a sufficient predictor for actual recycling

behavior, in case intention is measured on an aggregate level.

2) Attitudes and subjective norms about recycling are influenced by personal and

cultural constructs, such as self-perceptions and values.

3) Attitudes and subjective norms alone are necessary but not sufficient determi-

nants of recycling intentions.

4) With the addition of the perceived behavioral control element in the Theory of

Planned Behavior the predictive power of the original Theory of Reasoned

Action model is increased, allowing for cases in which the behavior (recy-

cling) is not under complete volitional control.

5) The addition of perceived behavioral control in studies predicting recycling

intentions and behavior has shown to improve predictability of the Theory of

Reasoned Action.

6) Attitude, subjective norms, and perceived behavioral control all seem to pro-

vide equally significant explanatory power for behavioral intentions and be-


7) A stricter separation of the perceived behavioral control variable into control

beliefs and perceived external facilitation conditions will strengthen this vari-


8) The inclusion of antecedents to the attitudinal, normative, and control beliefs

in the form of self-concepts or personal values has been found to improve

predictive ability of the entire model.

9) Value orientations seem to provide explanatory power for the origins of be-

havioral, normative, and control beliefs.

10) A succinct value orientation scale, such as the one provided by Schwartz, de-

livers adequate explanation for proenvironmental behaviors, such as recycling.

11) Perceived behavioral control is sufficiently explained by internal beliefs of

empowerment (self-efficacy) and beliefs about the extent of control of exter-

nal factors.

12) Values influence behavioral intention about recycling through their proximal

relationship with attitudes, norms, and control perceptions.

Research Questions

After a thorough review of the literature, it is conceivable that personal, social or

traditional values can directly and indirectly influence behavioral intentions. In so doing,

values related to a specific behavior could become the origin of beliefs regarding the in-


RO1: What roles do values orientations play in explaining recycling intention?

Furthermore, specific values orientations seem to be closely related to a particular

determinant of behavior. In other words, a rational motivation to act seems to stem from a

rational values base, while e.g., a behavior resulting from adherence to norms comes

from a values orientation based on traditions.

RQ2: Can attitudes, social norms, and perceived control dominance in reference to recy-

cling intentions be traced back to their specific underlying values orientations?

Finally, if the different values are connected to different determinants of intention,

we could explain their effect on behavior and the origins for the respective recycling be-


R3: Will the likelihood of recycling intentions be explained better if we include values

to the belief-behavior model?


After formulating the research questions, several hypotheses were developed.

They are as follows:

HI: Behavioral beliefs, outcome evaluations, normative beliefs, motivations to

comply, control beliefs, and perceived facilitation will predict attitudes to-

ward, subjective norms about, and perceived control about intentions to en-

gage in recycling.

H2: Attitudes, subjective norms, and perceived behavioral control will predict be-

havioral intention to engage in recycling.

H3: Attitudes and perceived control will be major predictors of intention to en-

gage in recycling.

H4a: Egoistic value orientations will be positively related to the control compo-

nent and negatively related to the attitudinal and normative components.

H4b: Altruistic/biospheric value orientations will be positively related to the atti-

tudinal component and negatively related to the control and normative com-


H4c: Traditional value orientations will be positively related to the normative

component and not significantly related to the attitudinal and control com-


H4d: Rational value orientations (the resultant value of the difference of egoistic

value orientations and altruistic/biospheric value orientations) will be posi-

tively related to the attitudinal component and negatively related to the nor-

mative and control components.

H5a: The inclusion of rational value orientations to attitudes will make a signifi-

cant contribution in the prediction capability of attitude for recycling inten-


H5b: The inclusion of traditional value orientations to subjective norms will make

a significant contribution in the prediction capability of subjective norms for

recycling intention.

H5c: The inclusion of egoistic value orientations to perceived behavioral control

will make a significant contribution in the prediction capability of perceived

behavioral control for recycling intention.

H6: The inclusion of values will improve the predictability of the attitudinal,

normative and control variables to explain recycling intentions in the Theory

of Planned Behavior.


This study uses survey methodology to assess differences in values, beliefs, atti-

tudes, and behavioral intention among the sample. It is broken out into two main parts.

First a correlation analysis will be conducted to confirm the hypothesized variables from

the question item set and explore the relationships between the variables set forth in the

theoretic part. Second, a regression analysis will be conducted to test the hypotheses and

research questions.

This chapter is broken out into five sections. The first section deals with the

model operationalization and sampling strategy. The second section discusses the survey

design as well as variable and scale development issues. The third section analyzes reli-

ability and validity issues regarding measurement. The fourth section explains procedures

and data cleaning techniques. And the fifth section details statistical analyses.

Operationalization of the Model

The purpose of the current research was to develop and test the influence of per-

sonal values on the recycling intentions of residents in the Gainesville, Florida area, using

the Theory of Planned Behavior as a guide. The model is operationalized according to the

outline by Ajzen and Fishbein (1980) and its expansion by Ajzen and Madden (1986).

The first three steps are more theoretical in nature, while the final two steps are empirical

and necessitate the involvement of the population of interest. The five steps are as fol-

lows (Young et al. 1991):

1. Select the behavior of interest and define it in terms of its action, target, context, and
time elements.
2. Define the corresponding behavioral intention.
3. Define general attitude, social norm, and perceived behavioral control. Define the val-
ues set.
4. Elicit the salient behavioral, normative, and perceived control beliefs about the target
behavior from a representative sample.
5. Develop or adjust questionnaire items from the salient behavioral, normative, and
perceived control beliefs.

A comparative analysis will be performed over the most recent and representative

studies on recycling that have used either the Theory of Reasoned Action or the Theory

of Planned Behavior (Bagozzi and Dabholkar 1994; Gamba 2000; Goldenhar 1991; Park

et al. 1998; Todd and Taylor 1995). The most frequently cited beliefs are subsequently

used as questionnaire items in the study. A similar analysis is performed over studies that

have used values orientations (Guttierez Karp 1996; Schwartz 1992; Stern & Dietz 1994).


The context of this study is a random digit dialing telephone survey. This method

was chosen as it seems appropriate for investigating recycling in a natural setting and

reaching a representative sample ofrecyclers and non-recyclers. Since recycling services

in the greater Gainesville area are primarily offered to single-family households, a study

targeting a completely random population that, e.g., includes over-proportionally a popu-

lation living in multi-family dwellings would skew its findings too much.

A total of 400 people will be surveyed during the last two weeks of May 2002 by

telephone in the greater Gainesville, Florida area. A random-digit dialing procedure over-

laid by the appropriate ZIP code classification will be used to draw from the population

of all potential recycling households with working telephones, regardless of if the number

was directory listed or not (Bagozzi & Dabholkar 1994). The person responsible for recy-

cling will be asked for to select a respondent in each household. Within the ordered ZIP

code area, a computer-assisted method will be used to generate the last four digits of the

phone number. Due to this technique the sample can be considered a random digit dialing

(RDD) sample (Klecka and Tuchfarber 1978; Miller 1991).

In order to minimize Type I and Type II errors, and to be able to detect moderate

levels of change, 400 respondents will be recruited, classified as the heads of household.

This number was calculated in accordance with an alpha level of .05, and a range of accu-

racy of the estimate of plus/minus 5% within the population percentage. In other words,

the 95% confidence interval should be the sample percentage plus or minus 5%. Accord-

ing to Kalton (1983), this specification requires that 1.96 SE(p) = 5%, where p is the

sample percentage and SE the standard error. With the use of a random sample, SE(p) is

the square root of PQ / n', where P is the population percentage, Q = 100 P, and n' is

the estimate of that sample size. Thus, 1.96 times the square root of PQ / n' = 5, or: n' =

1.962PQ / 5.

In order to determine n', a value is needed for P. Since PQ is largest at P = Q =

50%, a very conservative choice is to set P equal to 50%. With this choice, n' = 384,

which would constitute the maximum required sample size.

Survey Design

The survey instrument includes five scales. It combines elements from Stern,

Dietz, and Kalof's (1993), and Stern and Dietz's (1995) previously validated instruments

for the values dimensions (V), and elements from Ajzen and Fishbein's (1980) previously

validated instruments for attitudes (A), subjective norms (SN), and perceived behavioral

control (PCB). Subscales will be replicated directly and have shown medium to high reli-

ability, as will be reported later. The dependent measure is composed of the behavioral

intention (BI) scale, also replicated from Ajzen and Fishbein (1980). The variables and

scale items of the model are discussed below.

Explanatory Variables

values are operationalized in this study by a scale that consists of 16 values items.

Three of those dimensions that had been previously used in the Ster et al. (1994) study

hypothesized to load on three of the Schwartz (1992) dimensions. In addition, the authors

generated the 'biospheric' dimension. Borrowing from Schwartz's methodology, Ques-

tion 15 of the survey instrument (Appendix, p.129f) was created using seven-point

Likert-scale questionnaire items. They ask respondents if a particular value is "important"

to their overall life's value system, with the scale ranging from "extremely unimportant"

to "extremely important" (Schwartz, 1992). Subscale items of the four values dimensions

are hypothesized as:

Factor One: Egoistic Values Factor Three: Biosperic Values
Vego 1 Authority Vbio 1 Unity with nature
Veg, 2 Social Power Vbio 2 Protecting the environment
Vego 3 Wealth Vbio 3 Respecting the earth
Vego 4 Influence

Factor Two: Altruistic Values Factor Four: Traditional Values
Van 1 A world at peace Vi, 1 Honoring parents and elders
V.i 2 -Equality V" 2 Self-discipline
V.t 3 Social justice V" 3 Clean
V. 4- Helpful V. 4 Politeness
Vt, 5 Social order

Attitude is operationalized by defining it as the attitudinal beliefs about the conse-

quences of performing a particular behavior. Following instructions by Ajzen and

Fishbein (1980), several measures do combine to get the overall score for attitude. In or-

der to receive a measure of "attitude toward the behavior of recycling household waste",

a direct attitude measure, using seven-point Likert-scale questionnaire items (e.g., "Recy-

cling is a beneficial activity (unimportant criterion important criterion)") was assessed.

Along with the direct measure, a combined measure is calculated. The combined measure

is computed by adding the products of pairs of seven-point Likert-scale behavioral belief

questionnaire items (e.g., "Recycling reduces landfill use and waste (strongly disagree -

strongly agree)") and seven-point Likert-scale outcome evaluation questionnaire items

(e.g., "I like to decrease landfill use and messy trash (extremely unimportant- extremely

important)") (Questions 9a-c, lOa-g and 1 la-g in Appendix, p.127f).The behavioral belief

measures were drawn from previous research studies that inquired about the most salient

beliefs and outcome evaluations about engaging in recycling.

Subjective norm is operationalized as an individual's normative beliefs concern-

ing the influence of a particular referent (e.g., family, friends) over the participant per-

forming a particular behavior. Following instructions by Ajzen and Fishbein (1980), a

direct measure, using a seven-point Likert-scale subjective norm questionnaire item (e.g.,

"Most people who are important to me think I should recycle in the next week (strongly

disagree strongly agree)"), was assessed first. This was again contrasted to a combined

measure, tabulated by summing seven-point Likert-scale normative belief questionnaire

items (e.g., "How much do you agree with the statement that your neighbors think that

you should recycle (strongly disagree strongly agree)") multiplied by seven-point

Likert-scale motivation to comply questionnaire items (e.g., "How likely it is that you

would want to do what your neighbors thinks you should do? (extremely unlikely ex-

tremely likely)"). The normative belief measures were drawn from previous research

studies that had asked respondents to list people who might have an influence over their

decision to engage in curbside recycling (Questions 12 through 14 in Appendix, p. 129).

Perceived behavioral control is operationalized as the beliefs about the control an

individual feels he or she has over performing a particular behavior. Following Ajzen's

(1985) and Taylor and Todd's (1995) instructions, several measures combine to achieve

an overall perceived behavioral control score. Direct perceived behavioral control meas-

ures were assessed using seven-point Likert-scale questionnaire items (e.g., "Whether or

not I recycle is completely up to me (strongly disagree strongly agree)"). A combined

measure, computed by summing the products of pairs of seven-point Likert-scale control

belief questionnaire items (e.g., "Recycling takes too much effort (strongly disagree -

strongly agree)") and seven-point Likert-scale perceived control factor facilitation ques-

tionnaire items (e.g., "I don't like to participate in activities if they make my life more

difficult (extremely unimportant- extremely important)") was again contrasted to the di-

rect measure. Control beliefs and perceived facilitation beliefs were also drawn from pre-

vious research studies that asked for a list of items/ feelings that might facilitate or ob-

struct an engagement in recycling (Questions 9d, 10h-1 and 1 lh-1 in Appendix, p. 127f).

Response Variable

The response variable in this study is behavioral intention. This variable was cho-

sen as the variable of interest, because the Theory of Planned Behavior states that behav-

ioral intention directly predicts behavior (Ajzen 1985), unless intention precedes actual

behavior with a huge time-lag. Therefore, it was of interest to find how independent vari-

ables relate to the reported behavioral intentions. Behavioral intention was defined as

how likely or unlikely it is that a respondent would engage in a particular behavior, which

in this study means recycling. Similar to the other variables the instructions of Ajzen and

Fishbein (1980) were followed, using seven-point Likert-scale questionnaire items (e.g.,

"During the next 30 days, how likely is it that you will take part in a city-sponsored recy-

cling program (extremely unlikely extremely likely)") to obtain a behavioral intention

score (Questions 7 and 8 in Appendix, p. 127).


Reliability refers to the degree to which a measure is free of variable measure-

ment error. If we assume that the "true" score remains constant (e.g., that the person's

"true" attitude has not changed), a perfectly reliable instrument will yield the same results

on different occasions (Fishbein and Ajzen 1975). This will assure generalizability of the

study's results.

The most appropriate method to measure reliability of a group of items that are

hypothesized to measure separate aspects of the same concept is called internal consis-

tency. The term refers to the consistency or cooperation that should exist between a sub-

set of questions in measuring the same idea. The benefit of this technique is that it re-

quires only a single test administration, which provides subsequently a unique estimate of

reliability. The most popular of these estimates is given by Cronbach's alpha (Carmines

and Zeller 1979).

In this study, items that were assumed to measure a concept (e.g., attitude) will be

compared, using Cronbach's alpha. In total, tests will be conducted for the four values

dimensions, attitudes, and the perceived behavioral control.


Validity, in general, refers to the degree to which an instrument measures the

"true" score it was designed to measure (Fishbein and Ajzen 1975). For surveys, it refers

to the items or scales in a questionnaire. Assessing the validity of a measuring instrument

can take several forms. The most appropriate ones for the current study are discussed be-


Content Validity

This type of validity depends on the extent to which an empirical measurement re-

flects a specific domain of content (Carmines and Zeller 1979). As it is usually difficult

to objectively measure an abstract theoretical concept, such as "value" precisely, content

validity on average refers to the "mutual acceptance of the universe of content" (Cron-

bach and Meehl 1955) by a group of knowledgeable reviewers.

As far as the current study is concerned, a thorough review of the literature was

conducted by the researcher to show a holistic picture of the concept that allows compar-

ing and contrasting of the study's measures. Furthermore, academics in the College of

Journalism and Communication and the College of Political Science at the University of

Florida examined the literature and concepts, and agreed upon the fit of the measures

with the studied concepts.

Convergent and Discriminant Validity

Convergent validity is achieved when an instrument that forms a valid measure of

a construct correlates highly with another valid measure toward the same construct.

Campbell and Fiske (1959) furthermore argued that an instrument should also have dis-

criminant validity. If the same method or instrument (e.g., the Likert procedure) is used to

measure different variables (attitude toward different objects), different results should be

obtained (Fishbein and Ajzen 1975).

For application to the Theory of Reasoned Action, Fishbein and Ajzen (1975)

tested the convergent validity of the measures of its concepts (beliefs, attitudes, and in-

tentions). They found that single self-report scales of attitude toward e.g., religiosity cor-

related highly with four traditional attitude scales (Guttman, Likert, Thurstone, and se-

mantic differential scales). Schwartz (1992) found similar results in tests of the values

measures. Davidson (1973) established empirical support for convergent and discriminant

validity of intentional measures, using 'true-false' and 'likely-unlikely' scales to assess a

variety of family planning concepts (e.g., intentions). Since this study's measures follow

the specifications of both the Ajzen and Fishbein (1980) and the Schwartz (1992) models,

it is thought that this study's measures establish validity properly.

Predictive Validity

Predictive validity refers to the ability of an instrument to estimate an important

future behavior or event (Nunally 1978). Both the Theory of Reasoned Action and the

Schwartz Values Theory apply models that are primarily designed to make predictions.

Previous empirical research has established the studies' measures of e.g. behavioral in-

tention, attitudes, and beliefs. Following the guidelines outlined in these theories strictly,

the current study's measures are thought to predict the concepts equally well.

Construct Validity

Fundamentally, construct validity is concerned with the extent to which a particu-

lar measure relates to other measures consistent with theoretically derived hypotheses

concerning the concept (or construct) that are being measured (Carmines and Zeller

1979). Thus, it focuses on the extent to which a measure performs in accordance with

theoretical expectations of contributing to a single concept.

For the constructs in this study, taken from the Theory of Planned Behavior, the

measures followed the exact specifications of Ajzen and Fishbein (1980), who have con-

structed a valid questionnaire to test the Theory of Planned Behavior. In reference to

these concepts, validity of the study's measures is assumed.

The values construct measures follow research based on the Schwartz Values

Model (Dietz and Stern 1994; Schwartz 1992; Young 1997), and can equally be consid-

ered valid. To assure unidimensionality of the four values dimensions (egoistic, bio-

spheric, altruistic, and traditional) in this study, item loadings are established. If items do

not load on the specific constructs or load on multiple constructs, they have to be as-

sumed as weak predictors for the values dimension. They might make up a distinct, but

related construct, and will be treated accordingly (e.g., taken out of further consideration

for a particular values dimension or merged to form a unidimensional construct).

External Validity

External validity answers the question "to what populations, settings, treatment

variables, and measurement variables an effect can be generalized" (Campbell and

Stanley 1963). Because the sample in this study was drawn at random from a general

population in Gainesville, we can assume that there should not be a problem to generalize

findings to the larger population of the sampling area (Gainesville).



A 30-item questionnaire was created and designed for telephone survey technol-

ogy. A pilot-test was conducted to assess reliability and validity issues. Once these issues

were found acceptable for the questionnaire, it will be administered to the sample.

The data will be collected in May 2002. Responses will be gathered by profes-

sional telephone callers under the supervision of the Florida Survey Research Center. The

callers completed a one-to two-hour training session and have between two to four years

of experience making calls. The random samples of residential telephone exchanges will

be provided by Genesys Sampling Systems. A variety of efforts will be used to reduce

bias due to nonresponse, including making weekend calls as supplements to weekday

evening calls, performing multiple callbacks, and accommodating requests for interview

appointments (Martinez & Scicchitano 1998).

This study implements a household survey that asks to speak to the person most

familiar with the family's recycling. Once the respondents are willing to participate, the

interviewer briefly discusses the purpose of the questionnaire of investigating attitudes

toward the environment and recycling behavior. After the introduction, the interviewer

explains the answer categories of the Likert scale and begins the interview. Initially, the

interviewer will inform the respondents that this particular study has been approved by

the Instructional Review Board at the University of Florida. The interviewer then reads

the respondents the approved informed consent and explains the rights of participation in

the study. The interviewer also advises them again that any personal information that is

given will be kept completely confidential, and that names and responses will be kept

anonymous. The questionnaire is estimated to take about 15-18 minutes to complete.

Data Examination and Cleaning

Surveys will be visually inspected to look for obvious respondent errors. For ex-

ample, if a whole section is unanswered, that particular survey is discarded. Before pre-

liminary analysis, the data will be transformed to ensure proper analysis with SPSS. This

transformation step includes recoding of values questions to eliminate negative numbers,

and reverse-coding of negatively worded questions to assure consistency (Young 1997).

To further examine the data for errors frequency of all variables will be run. This

procedure will identify any items that may be outside an acceptable range for a specific

variable. Problems that surface through this procedure will be subsequently corrected. In

a next step, the nature of the respondent answers will be examined as well. If a problem

surfaces, the case will be identified and the problem corrected.

Statistical Analyses

The level of significance for the statistical tests for this study is .05. This equates

to an acceptance of risk by this study that out of 100 samples, a true null hypothesis

would be rejected five times (Polit and Hungler 1999). After aggregation the collected

data will be analyzed in several ways, depending on the hypotheses and research ques-

tion, they related to.

Data Aggregation

The items hypothesized to form the variables attitude, subjective norm, and per-

ceived behavioral control will be subjected to a reliability analysis, and indices created

for each respectively by averaging the means of the responses and combining the items

measuring the three beliefs and three evaluations/ motivations into one weighted variable

respectively. The weighted variables will be used in the correlation and regression analy-

sis (Ringer-Lepre 2000).

A reliability analysis will also be conducted for the three values orientation to

check, if the theoretical values differences would persist for this study. Then an index

will be created for each values orientation, constructed as an average of the orientations.

A fourth values orientation index, the rational (or self-driven) orientation, will be created

as well from the difference of egoistic and altruistic/biospheric values. These weighted

values variables will be used in the correlation and regression analyses.

Correlation Analysis

Hypotheses 4a through 4d will be tested using bivariate correlation analysis. The

correlation table will provide the significance of relationship between each of the four

values orientations and attitude, subjective norm and perceived behavioral control.

Regression Analyses

The remaining hypotheses will be tested using simple or multiple linear regres-

sion. Hypotheses 1 through 3 are using multiple regression to test the original Theory of

Planned Behavior. This means that the attitude, subjective norm, and perceived behav-

ioral control factors will be regressed against the two aggregated recycling intention fac-

tors (plan to recycle, do not plan to recycle). Hypotheses 5a through 5c uses simple linear

regression, testing the partial effect of one of the determinants of intention (attitude, sub-

jective norm, or perceived behavioral control) without the addition of their respective

values orientation and with their interaction of this value orientation, both multiplicative

(e.g., Vt x A) and additive (e.g., Vr, + A). The three resulting goodness-of-fit values

(R2) will be compared to observe, if the original R2 value (that without addition of a val-

ues term) has significantly improved by either method. The results of those three regres-

sions will then be used in Hypothesis 6, comparing the original Theory of Planned Be-

havior regression equation to the values-enhanced regression equation, applying the most

appropriate interactive term for each variable.

Assumptions of Multiple Regression Tests

Since multiple and simple linear regression are the primary methods for testing

the hypotheses (Agresti 1997; Norusis 1994), attention needs to be paid to its assump-

tions. The assumptions underlying multiple regression concern both the dependent and

independent variables and the relationship between those. Unlike many other statistical

tests, the analysis of assumption violation must be performed after the estimation of the

regression model. According to Hair et al. (1987), "the basic issue is whether, in the

course of calculating the regression coefficients and predicting the dependent variable,

the assumptions of regression analysis have been met (p. 172)." The major assumptions

are (Hair et al. 1987):

1. Linearity of the phenomenon: There is an assumed linear relationship between the
group of independent variables as well as between each independent variable and the
dependent variable. An analysis of partial regression plots between each independent
variable and the dependent variable was suggested by Hair et al. (1987) to assess this

assumption. A curvilinear pattern of residuals would indicate a non-linear relation-

2. Constant variance of the error term: This assumption refers to the concept of homo-
scedasticity (equal variance). Hair et al. (1987) recommended plotting the studen-
tized residuals against the predicted dependent variable values and comparing them
to a null plot (a random plot of points).

3. Independence of error terms: Regression analysis assumes independence of the pre-
dicted value. Predictions are not sequenced by other variables .Plots of residuals
against possible sequencing variables are useful to identify non-independence.

4. Normality of the error term distribution: Normal probability plots, comparing stan-
dardized residuals to a normal distribution (straight line), are a useful method for
identifying this condition (Hair et al. 1987).

This study will examine studentized residuals, outliers, influential observations,

and multicollinearity to test for assumption violations as outlined by Hair et al. (1987).

Partial regression plots will be used to examine the linearity of relationships. Cases that

are identified as violating these assumptions will be deleted from further specification.

To identify outliers, visual inspection of partial regression plots as well as indi-

vidual leverage values will be used. The latter indicate the distance between a single case

and the center of all observations. According to Neter et al. (1990), values greater than

2p/n were scanned, whereby p = the number of regression parameters in the function in-

cluding the intercept term, and n = sample size. The typical regression function for this

study includes the intercept and the variables attitude, subjective norm, and perceived be-

havioral control for a total of four regression parameters with a sample of 400.

As a second method to detect outliers, studentized deleted residuals will be used.

Following Neter et al. (1990), absolute values of the studentized deleted residuals will be

compared to a t-distribution with n-p-1 degrees of freedom.

The Variance Inflation Factor (VIF) will be used to discover multicollinearity ef-

fects. Neter et al. (1990) suggest that multicollinearity between the independent variables

exists, if a VIF value in excess of 10 for any of the independent variables is present. Hair

et al. (1987) suggest a process that this study followed. First, all condition indices above a

threshold value of 15, a conservative value (Hair et al. 1987), will be identified. Among

condition indices exceeding 15, variables with variance proportion above 90% will be

identified. A .90 or higher between two or more coefficients will indicate multicollinear-


For each regression, the Enter variable function will be used. Consistent with the

hypotheses of this study, this approach enters all variables simultaneously. After the data

are adjusted for violations of assumptions a second analysis will be conducted. In all

cases, the first elimination of outliers produces results that will be judged to adequately

meet the assumptions of multiple regression.


This chapter consists of three parts. First, a discussion of the descriptive statistics

about the study sample and assumptions of the regression method. Next, the analysis of

the original Theory of Planned Behavior, and an examination of the value-added model.

Finally, the results of the respective hypothesis tests and answers to the research ques-

tions are presented. The key method used for the statistical tests was multiple linear re-


Preliminary Analyses

A discussion of the demographic statistics, the results of the data examinations,

and the results of the tests for violations of the regression model assumptions follows.

Study Participants

Four hundred residents in the Gainesville, Florida area were surveyed during the

last two weeks of May 2002 by telephone. The participants ranged in age from 18 to 89

with a mean age of 40.9 years (SD=18.6). There were slightly more women in the sample

with approximately 62 percent of the respondents being female (n = 247). Most respon-

dents (96%) had a high school diploma or more. There was a statistically significant dif-

ference (t=3.16, p=.002) in educational levels between recyclers and nonrecyclers. Both

segments lived an average of 13 years in the Gainesville area with recyclers slightly

longer (+2 years). The household income demographics showed a propensity to recycle

that was slightly more pronounced among more affluent people. While 62% of house-


holds with an income above $35,000 tended to recycle, only 49% of households below

$35,000 did so. There was no statistical difference between recyclers and nonrecyclers as

far as their political orientations were concerned. While recyclers tended to lean slightly

more liberal (30%), the nonrecycler segment was concentrated in the moderate category

(40%). Details on the demographic characteristics are summarized in Table 4-1.

Table 4-1. Demographic characteristics of recyclers and nonrecyclers

Recyclers Nonrecyclers t Significance Total
Characteristic (N= 365) (N = 35) (N= 400)
Age (years) 3.24 .001
Range 18-89 19-81 18-89
Mean (SD) 41.1 (18.5) 32.3(16.6) 40.9 (18.6)
Education [n (%)] 3.16 .002
College+ 311(85) 26 (74) 337 (84)
No college 50(14) 9 (26) 59(15)
Family income [n (%)] 0.05 .964
> $35,000 225 (62) 17(49) 242 (61)
< $35,000 108 (30) 16(46) 124 (31)
Political orientation 1.51 .146
[n (%)]
Conservative 88 (24) 6 (17) 94 (23)
Moderate 134 (37) 14 (40) 148 (37)
Liberal 109 (30) 8(23) 118(29)
Residency (years) 1.59 .114
Range 0-75 0-52 0-75
Mean (SD) 15.2(15.3) 12.4(12.8) 14.9(15.1)

Data Examination Results

Surveys were visually inspected to control for obvious respondent errors. Incom-

plete surveys and surveys that seemed to be answered the same way throughout the sur-

vey were discarded. Multiple samples were ordered by the Florida Survey Research In-

stitute to arrive at the contracted number of 400 respondents. As a result none of the 400

surveys had to be discarded due to survey errors. Surveys were then given case identifica-

tion numbers for further analysis. The dataset arrived in an ASCII format. The data were

first entered into Excel and then visually inspected for data entry error. If an error was

found on a specific row, the survey was retrieved by case number and the record was cor-

rected. The entire file was translated to SPSS for preliminary analysis. In addition, nega-

tively worded questions (Ql 1.8-11.12, Q12.8-12.12) were reverse-coded for consistency.

After the data entry, a preliminary inspection of the data was conducted to un-

cover potential confounding effects in the sample. Frequencies of all variables were run

to further examine the data for error. The data were inspected to determine if items fell

outside the acceptable range for a particular variable. In a next step, the nature of the re-

spondent answers was examined. Cases were checked for "Don't know" and "Refused"

answers by sorting the cases in an ascending fashion from the lowest to highest value.

Cases with "Don't know" and "Refused" answers were considered missing variables and


Similarly, all weighted variables, such as the attitude belief and outcome evalua-

tion total, were checked for missing values. Cases with data of that nature were evaluated

as being missing and listwise deletion was used to eliminate them. Out of a total of 400

surveys, 87, (22%) were discarded because they had missing values in one or more of the

model variables. A total of 313 cases were retained for further analysis.

Regression Model Assumptions

Regression analyses involve a series of assumptions about the relationships of the

variables being measured to each other. These assumptions again are as follows:

1. Linearity of the phenomenon
2. Constant variance of the error term
3. Independence of the error term
4. Normality of the error term distribution

Violation of these assumptions can lead to serious problems in interpreting the re-

sults of the study. To test for assumption one and two, standardized residuals were plotted

against the predicted individual independent variables. The scatterplots showed a random

distribution void of observable pattern. As for assumption three, although data were not

collected and recorded sequentially, it is plausible that 'time' may have influenced the

residuals. The plot of standardized residuals against the sequencing variable showed no

discernible pattern. Finally, to test for normality (assumption four), a histogram of re-

siduals was constructed and superimposed with a normal distribution curve. The distribu-

tion of residuals appeared to be approximately normal.

Multicollinearity did not surface as a problem in the analyses of this study. This

was in part due to the fact that empirically tested variables and variable relationships of

tried theories were used. Both the Theory of Planned Behavior and the Schwartz Values

Model have been tested extensively by subsequent research. The variables behaved in the

detected fashion in this study as well.

Analysis of the Theory of Planned Behavior

This section discusses the results of the first three hypotheses, testing the Theory

of Planned Behavior as an explanation of recycling intention. The first hypothesis pro-

posed that belief and outcome variables would predict attitude, subjective norms and per-

ceived behavioral control. The second hypothesis predicted that attitude, subjective

norms, and perceived behavioral control would explain recycling intentions. The third

hypothesis suggested that attitude and perceived behavioral control would be the most

significant influence factors on recycling intentions.

Variable Preparation

Before testing the hypotheses, it was necessary to recode the attitude variable and

create an index. Three items on the questionnaire were designed to measure respondents'

attitude toward recycling. Each question was worded "Participation in recycling is" fol-

lowed by a different response scale for each question.

The seven-point scales measured the respondents' evaluation of how wise/foolish,

important/unimportant, and beneficial/harmful they perceived recycling. A reliability

analysis yielded an alpha of .76 for the three items. The attitude index was created by av-

eraging the means of the responses of the three items.

As far as the variables subjective norm and perceived behavioral control were

concerned, there was only one item determining a general measure of each variable. The

measure for subjective norm was worded: "Please tell me much you agree with the state-

ment that most people who are important to you think that you should recycle" with an-

swers ranging from strongly disagree (1) to strongly agree (7). The measure for perceived

behavioral control was worded: "Whether or not I recycle is completely up to me" with

an identical answer distribution. Since those two variables were composed of single-item

measures, it was unnecessary to create an index for use in the regression model. All three

variables were subsequently recorded to change their one to seven scales into bipolar -3 to

+3 scales to be consistent with the following predictor variables.

New weighted variables for the three explanatory variables in the model were re-

quired. The items measuring behavioral beliefs and outcome evaluations, normative be-

liefs and motivation to comply and control beliefs and perceived facilitation were com-

bined into one weighted variable each. These three weighted variables were subsequently

used in the regression model.

To accomplish this, the items measuring behavioral beliefs, normative beliefs,

control beliefs, outcome evaluation, motivation to comply, and perceived facilitation

were recorded to change their one to seven scales into bipolar -3 to +3 scales. Then, each

of the seven items measuring behavioral beliefs was multiplied with its matched item

measuring outcome evaluation. For example, the two items "Recycling reduces landfill

use and waste" (behavioral beliefs) and "I like to decrease landfill use and messy trash"

(outcome evaluation) were a matched pair and their product formed a weighted variable.

Similarly each of the five items measuring normative belief was multiplied with

its matched item measuring motivation to comply, as were the five items measuring con-

trol belief with their matched item measuring perceived facilitation.

Regression Study

Hypotheses one through three used linear regression to analyze the Theory of

Planned Behavior. Hypothesis one stated that the explanatory variables behavioral beliefs

and outcome evaluations will predict attitudes toward recycling, normative beliefs and

motivations to comply will predict subjective norms about recycling, and control beliefs

and perceived facilitation will predict perceived control about intentions to engage in re-

cycling. The regression analysis showed a significant and positive correlation between

the weighted variables representing behavioral beliefs and outcome evaluations and the

attitude index (r=.381, p<.001). The regression analysis also showed a significant and

positive correlation between the weighted variables representing normative beliefs and

motivation to comply and the general subjective norm variable (r=.230, p<.001). Finally,

the regression showed a significant and positive correlation between the weighted vari-

ables representing control beliefs and perceived facilitation and the general variable

measuring perceived behavioral control (r=. 195, p<.05). Consequently, it was concluded

that these weighted variables were significant predictors of attitude, subjective norm, and

perceived behavioral control.

Hypotheses two and three maintained that attitudes, subjective norms, and per-

ceived behavioral control would predict behavioral intention to engage in recycling, and

that attitudes and perceived control would be the major predictors of intention to engage

in recycling. A regression analysis revealed a significant and positive correlation between

attitude and behavioral intention (r=.208, p<.001). The correlation between subjective

norm and behavioral intention was positive, but not significant (r=.078, p<. 10), as was

the correlation between perceived control and behavioral intention (r=.043, p=.25).

The three variables then were entered into the model using stepwise regression

analysis (Table 4-2). The results showed that only attitude was a significant predictor of

behavioral intention (Table 4-3). As residents' attitudes toward recycling increase, so do

residents' intentions to engage in recycling activities.

Table 4-2. Stepwise model

Measure Step r
Attitude 1 .208
Subjective Norm 2 .078
Perceived Behavioral Control 3 .043

Table 4-3. Intention to recycle household waste

Measure St. Beta Signifiance .......
Attitude .208 3.39 .001
Subjective Norm .078 1.26 .210
Perceived Behavioral Control .047 0.77 .444

From the results of these hypotheses tests, a path model showing the application

of the Theory of Planned Behavior concerning recycling was developed (Figure 4-1).

Beliefs Attitude
Sr.381l\ r=.20s
Outcome w=.200X
Evaluations Relative importance r=.213 Behavioral .186 Behavior

Normti scodssubjective norm n '
Normative -- ,4 -
Beliefs w2=.053 e ,, --'b
Subjective r-.078**
Motivationto .20 Norm r.043** .0

Beliefs \ Perceived ,,-
195 Behavioral
Perceived Cnr

Figure 4-1. Path model for the Theory of Planned Behavior.
** indicates non-significant path

Analysis of the Proposed Values-Enhanced Model

This section discusses the results of the final eight hypotheses tests. These hy-

potheses examined the roles values orientations play in explaining recycling intention,

and if the likelihood of recycling intentions can be explained better if one includes values

to the belief-behavior model of Ajzen and Fishbein.

Before testing the hypotheses, the four values variables were recorded and an in-

dex created for each of them. Four items on the questionnaire were designed to measure

respondents' egocentric values orientation. Four items were designed to measure respon-

dents' altruistic values orientation. Three items were designed to measure respondents'

biospheric values orientation. Five items were designed to measure respondents' tradi-

tional values orientation. Each question was worded "How important are the following

principles in your life" followed by a response scale from "extremely unimportant" to

"extremely important" for each question.

A reliability analysis was conducted for each values dimension to see how well a

question set measured each construct. The analyses yielded the following: an alpha of .67

for the four egoistic values items; an alpha of .87 for the four altruistic values items; an

alpha of.81 for the three biospheric values items; an alpha of.84 for the combined altru-

istic/biospheric values items; and an alpha of .66 for the five traditional values items. The

constructed rational values orientation computed as the difference between egoistic and

altruistic values was checked by examining the correlation between the questions per-

taining to those values. The analysis yielded an alpha of .63 on the eight items. Four val-

ues indices were created by averaging the means of the responses of the items.

Correlation Study

Hypotheses four (a) through four (d), aimed to corroborate the basic structure of

the effects of the four values dimensions (egoistic, altruistic, traditional, rational) on the

three determinants of behavioral intention respectively via correlational analysis.

Hypothesis four (a) stated that the egoistic value orientation would be positively

related to the control component and negatively related to the attitudinal and normative

components. The Pearson correlation (Table 4-4) showed that the egoistic values dimen-

sion correlated significantly with perceived behavioral control (r=.207, df=313, p<.001).

This value was not significantly correlated with any other component.

Hypothesis four (b) affirmed a positive relationship between the altruis-

tic/biospheric value orientation and the attitudinal component and a negative relationship

between this value and the control and normative components. The altruistic/biospheric

values dimension correlated significantly but negatively with attitude (r--.227, df=313,

p<.001) as well as with subjective norm (r=-.127, df=313, p<.01). It did not correlate sig-

nificantly with perceived behavioral control (r=.047, df=313, p=.45).

Hypothesis four (c) maintained that the traditional value orientation is positively

related to the normative component and not significantly related to the attitudinal and

control components. While the traditional values dimension did indeed correlate signifi-

cantly with subjective norm, it correlated negatively (r=-.175, df-313, p<.001). Further-

more, it did not significantly correlate with either attitude (r--.092, df=313, p=. 14), or

perceived behavioral control (r-.088, df=313, p=.16).

Finally, hypothesis four (d) stated that the rational value orientation (the resultant

value of the difference of egoistic value orientations and altruistic/biospheric value orien-

tations) is positively related to the attitudinal component and negatively related to the

normative and control components. The correlation analyses confirmed this hypothesis in

part (Table 4-4). The rational values dimension was positively correlated and statistically

significant with attitude (r=. 172, df=313, p<.001), but not significantly correlated to sub-

jective norm (r=.019, df-313, p=.76). On the other hand, the findings showed a weak

positive correlation with perceived behavioral control (r=.148, df=313, p<.01).

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E2L7JRAK7_M1INVA INGEST_TIME 2013-03-21T13:16:35Z PACKAGE AA00013636_00001