THE°IS

95 WI”lUlHNlllHlleIHIHIHUIHJill”‘IlWHlHlllli

19‘)? 01812 5462

This is to certify that the

dissertation entitled

Unintended Effects of Fear Appeals:
The Role of Stage of Change, Threat, and Efficacy

presented by
Hyunyi Cho

has been accepted towards fulfillment
of the requirements for

PhoDo degree in MaSS Media

 

(/k/

Magmprofessor

 

Date 7/20/99

MSU is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

 

LIBRARY

Michigan State

University -

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

 

 

 

 

 

DATE DUE DATE DUE DATE DUE
NOV 9 94326020 13
MR1 7 7
—fi-E g ‘L 7“ APB 35w;

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1M gummy.“

UNINTENDED EFFECTS OF FEAR APPEALS:
THE ROLE OF STAGE OF CHANGE, THREAT, AND EFFICACY

By

Hyunyi Cho

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Mass Media PhD. Program
College of Communication Arts & Sciences

1999

ABSTRACT

UNTNTENDED EFFECTS OF FEAR APPEALS:
THE ROLE OF STAGE OF CHANGE, THREAT, AND EFFICACY

By

Hyunyi Cho

Doing social good, the conscious intention of health communication campaigns,
may produce unanticipated, and sometimes adverse, consequences. Campaign messages
using fear appeals, in particular, have long been recognized for their potential to induce
unintended effects, including the initiation of risky behaviors. While attempting to "do
good," health communication campaigners need to examine possible unintended effects
as well as intended effectiveness to keep their ethical mandate to "do no harm."

The present study represents an attempt to identify and investigate the conditions
under which unintended effects of fear appeals are likely to occur. Specifically, the roles
of stages of change, threat, and efficacy in producing unintended effects of fear appeals
were examined. These unintended effects include defensive avoidance, message
derogation, perceived manipulation, fatalism, hopelessness, reactance, and wishful
thinking. The extended parallel process model and the transtheoretical model were used
as theoretical frameworks. A lab experiment (n=274) was conducted to examine college
students' reactions to fear appeals used in skin cancer prevention messages. Threat was
manipulated in two (high, low) levels. Behavior was measured in a four-week follow-up.

The results indicated that in general, unintended effects of fear appeals are more

likely to occur in individuals in earlier stages of change than later stages of change.

While perceived threat energized both intended and unintended effects of fear appeals,
lack of efficacy among individuals in earlier stages of change largely rendered unintended
effects more conspicuous than among others.

After being exposed to fear appeal messages, college students who had not
intended to engage in skin cancer prevention reported greater likelihood of avoiding the
thoughts of skin cancer and its prevention, being fatalistic about the threat of skin cancer,
feeling hopeless, and perceiving that the messages were constraining their behavioral
freedom, than those who had intended to engage in or who had engaged in behavior for
skin cancer prevention. Concurrently, those who had not contemplated behavioral
changes for skin cancer prevention indicated less favorable attitude toward message
recommendations, fewer intentions to engage in preventive behavior, and fewer changes
in behavior, than those who had contemplated or had engaged in skin cancer prevention.

If those who do not engage or who do not intend to engage in preventive behavior
are the major intended audience of health communication campaigns and need
communication campaigns most, the results of this study hold significant implications
and challenges for message designers. The results suggest that those in greatest need of
communication campaigns may be the most likely to exhibit unintended effects and least
likely to respond with intended changes when being exposed to health communication
messages using fear appeals. Moreover, fear appeal messages are likely to be more
effective in reinforcing those who already plan or engage in preventive behavior than

persuading those who have yet to initiate changes toward prevention.

Copyright by
HYUNYI CHO
1999

ACKNOWLEDGMENTS

It is a pleasure to look back and express my deepest gratitude to many individuals
who helped me on this study. Professor Kim Witte spent many hours for me to answer
questions and to look at messages and measures. Her guidance and encouragement were
indispensable for this study. I thank Professors Charles Atkin, Robert Larose, and Bella
Mody for their guidance and support. Their invaluable comments helped me put and see
this study in better perspective.

I am fortunate to have many good MMPhD friends: Kathy Bradshaw, Alice Chan,
Xigen Li, Hugh Martin, Teresa Mastin, Charles St. Cyr, and Patricia Whalen. I thank
them for being such a great source of motivation and encouragement.

Yeonshin Kim, Sean Limon, and Professors Franklin Boster, Janice Bukobac,
Steve Edwards, and Bella Mody helped me out in accessing subjects. I also thank Ilene
Young, Amy Bauer, Sejung Choi, Yungkyun Choi, Joohyun Lee, Sunyoung Lee, and
Professor Bonnie Reece for their help in running experiments.

I thank Dongsung Hong and Sook Lee for their friendship.

My advisor Professor Charles Salmon has guided me every step of the way up to
this study and its completion. I've learned so much from him: about research, about
teaching, and about being a good person. His faith in me, which he did not flinch even
when I stumbled, sustained and will sustain me.

My parents made everything possible, including this wonderful venture to

America.

TABLE OF CONTENTS

 

 

 

LIST OF TABLES ....... -- -- - ..... - -- ..... ----IX
CHAPTER 1: INTRODUCTION -- - -1
Unintended Effects of Health Communication Campaigns ......................................... 7
Unintended Effects of Fear Appeals .......................................................................... 11
CHAPTER 2: LITERATURE REVIEW ......... -- -- _ - -19
Defining Fear Appeals ............................................................................................... 19
Theoretical Approaches to Unintended Effects of Fear Appeals ............................... 20
Drive Models ......................................................................................................... 21
Expectancy-Value Models ..................................................................................... 22
Protection Motivation Theory ............................................................................ 22
Health Belief Model .......................................................................................... 25
Psychological Reactance ........................................................................................ 26
Parallel Process Models ......................................................................................... 27
Parallel Process Model ....................................................................................... 27
Extended Parallel Process Model ....................................................................... 28
Fear Appeals and Audience Characteristics ................................................................ 31
The Transtheoretical Model ........................................................................................ 32
Linking the Extended Parallel Process Model and the Transtheoretical Model ......... 37
Hypotheses and Research Questions ........................................................................... 44

vi

CHAPTER 3: METHOD -- ..... 46

 

 

 

Overview .................................................................................................................. 46
Subjects .................................................................................................................... 46
Procedure ................................................................................................................. 47
Stimulus Materials ................................................................................................... 48
Measures .................................................................................................................. 49
Analysis ................................................................................................................... 57
CHAPTER 4: RESULTS ............. - - -- - ....... -- - -- 59
Message Validation ................................................................................................... 59
Manipulation Check .................................................................................................. 59
Hypothesis 1 .............................................................................................................. 64
Hypothesis 2 .............................................................................................................. 7O
Hypothesis 3 .............................................................................................................. 84
Hypothesis 4 .............................................................................................................. 87
Research Question 1 .................................................................................................. 89
Research Question 2 ................................................................................................ 103
CHAPTER 5: DISCUSSION. ..... - - -- 105
The Role of Stage of Change ................................................................................... 105
The Role of Efficacy ................................................................................................ 108
The Role of Threat ................................................................................................... 110
Relationship Between Unintended and Intended Effects ......................................... 111
Implications .............................................................................................................. 1 12

vii

Limitations ............................................................................................................... 1 15

 

APPENDICES -- -- - -- 117
Appendix A1: Informed Consent Form ...................................................................... 118
Appendix A2: High Threat Message .......................................................................... 119
Appendix A3: Low Threat Message ........................................................................... 121
Appendix B: Measures ............................................................................................. 123

REFERENCES -- - - - 135

 

viii

LIST OF TABLES

Table 1a. Typology of Unintended Effects of Health Communication Campaigns:
Predominantly Individual-Level, Direct, Short-Term, and Content/Target
Specific ............................................................................................................... 5

Table 1b. Typology of Unintended Effects of Health Communication Campaigns:
Predominantly Institutional-Level, Indirect, Long-Term, and

Content/Target Diffuse ....................................................................................... 6
Table 2. Unintended Effects of Fear Appeals ................................................................. 16
Table 3. Distribution of Stages of Change ...................................................................... 50
Table 4. Reliabilities of Measures .................................................................................. 56
Table 5. Zero-order Correlations among Variables ........................................................ 62

Table 6a. Mean Differences of Unintended Effects across Stages:
Results of Planned Comparison ........................................................................ 65

Table 6b. Mean Differences of Unintended Effects across Stages:

Results of Tukey-B Tests ................................................................................... 67
Table 7. Summary of Analysis of Covariance of Threat x Efficacy and

Adjusted Means ................................................................................................ 74
Table 8a. Summary of Analysis of Covariance of Threat x Response Efficacy and

Adjusted Means for Fatalism ............................................................................. 85
Table 8b. Summary of Analysis of Covariance of Threat x Self Efficacy and

Adjusted Means for Hopelessness ..................................................................... 86
Table 9. Summary of Analysis of Covariance of Threat x Response Efficacy and

Adjusted Means for Attitude toward Risk Behavior ......................................... 88
Table 10. Summary of Analysis of Covariance of Stage x Threat x Efficacy and

Adjusted Means ................................................................................................. 93
Table 11. Results of Partial Correlations between Unintended and Intended Effects ..... 104

CHAPTER 1

INTRODUCTION

Health communication campaigns are social practices (Salmon, 1989, 1992).
Influencing a large number of individuals’ beliefs, attitudes, and behavior (Rogers &
Storey, 1987), health communication campaigns constitute an agent of social change
(Kotler, 1978; Paisley, 1981). Concomitantly, they represent an outcome of social
process in which conflicting values and powers are negotiated (Rakow, 1985; Guttman,
1996:1997)

That social practices entail multiple functions was first noted by Merton (1957).
Pointing out that conscious intentions of social practices are independent of objective
outcomes, Merton suggested that manifest and latent functions of social practices be
distinguished. The former refers to “those objective consequences for a specified unit
which contribute to its adjustment or adaptation and were so intended, while the latter
refers to “unintended and unrecognized consequences of the same order” (p. 63). In this
light, Salmon (1989) argued that effectiveness, those outcomes intended by the
campaigner, be distinguished from effects, outcomes including unexpected ones, in the
evaluation of health communication campaigns.

Unanticipated consequences have been studied, although not explicitly, in a
broad range areas, such as medicine (e.g., Barsky, 1988a; 1988b; Illich, 1976; McKeown,
1979; Nesse & Williams, 1994; Sagan, 1987; Tenner, 1996), environmental control (e.g.,

Burton, Kates, & White, 1993; McPhee, 1989; Wijkman & Timberlake, 1984),

computerization (e.g., Kling, 1995; Weizenbaum, 1976; Zuboff, 1984), and interpersonal
(e.g., Mortensen, 1997) and mass communication (see examples below).

In the domain of medicine, for example, researchers have noted the paradoxical
effects of the improvements of the nation’s health status. The advances in life
expectancies, mortality rates, treatments, and preventive measures were not always
accompanied by the enhancement of the public’s perceptions of their own health (Barsky,
1988a; 1988b; Illich, 1976; Tenner, 1996). This discrepancy, aptly described as “the
failure of our success” (Barsky, 1988b, p. 415) and “doing better but feeling worse”
(Wildavsky, 1977, p. 105), is attributed to factors such as the following.

Progress in medical care has decreased the mortality of acute diseases. This
accomplishment, however, came at the cost of relative increases in chronic diseases.
Concurrently, the promotion of health and health-consciousness has undermined the
public's perceived health status by rendering them overly sensitive to symptoms and
feelings of illness (Barsky, 1988a; 1988b; Mechanic, 1983; Miller, Murphy, & Buss,
1981; Pennebaker, 1982).

Undesirable consequences have also been investigated in media effects research,
focusing mainly on the areas of commercial advertising and entertainment content
(Brown & Walsh-Childers, 1994; McGuire, 1988). The examination of the unfavorable
effects of commercial advertising has focused on issues involving special audience such
as children (e.g., Adler, Lesser, Meringoff, Robertson, Rossiter, & Ward, 1980;

Comstock & Paik, 1991; Kunkel & Gantz, 1992; Wartella, Heintz, Aidman, &

Mazzarella, 1990) and long term changes in values and culture (e. g., Holbrook, 1987;
Pollay, 1984; 1986a; 1986b; 1987).

A wealth of research has been done on the effects of entertainment program.
This line of research may technically not be considered as one on "unintended" effects
because intentions of entertainment programs may not be as clearly defined as planned
social actions such as health communication campaigns, or the intentions may be widely
divergent (e.g., to make money vs. to frighten). Nonetheless, the inauspicious nature of
their effects has received much attention from both researchers and the general public.
Studies have focused on violence on viewer aggression (e. g., Bandura, Ross, & Ross,
1963; Berkowiz, 1984; Gerbner, Gross, & Signorielli, 1980; Phillips, 1983), portrayal of
minorities on viewer stereotyping (e.g., Armstrong, Neuendorf, & Brentar, 1992; Atkin,
Greenberg, & McDermott, 1983; Elliot & Byrd, 1983), and erotic content on sexual
attitude and behavior (e.g., Donnerstein, Linz, & Penrod, 1987; Greenberg, Brown, &
Buerkel-Rothfuss, 1993; Zillmann & Bryant, 1982; 1984) (McGuire, 1988).

Unintended effects can be either positive or negative. For example, in response to
opponents' criticism that children's advertising promotes materialism, proponents
contended that advertising can help children be familiarized with the values of the real
world (Belch & Belch, 1995). However, research has tended to focus on negative
consequences, perhaps in recognition of the public's attention and the need of social
intervention.

Yet, health communication campaigns, due to their manifest intent to do good

rather than harm, have been underexamined for their possible adverse effects. Salmon

(1996) offered initially a compilation and typology of unintended effects of health
communication campaigns. The next section presents a review of his typology, which is

summarized in Tables la and 1b.

Table 1a

Typology of Unintended Effects of Health Communication Campaigns:
Predominantly Individual-Level, DirectLShort-Term, and Content/Target

 

 

Specific
Unintended Definitions Examples
Effects

Obfuscation Confusion and Lerman, Rimer, Trock,

misunderstanding of messages. Balshem, & Engstrom, 1990;
Roberts, Imrey, Turner,
Hosokawa, & Alster, 1986

Dissonance Psychological discomfort and Guttman & Zimmerman, 1996;
distress provoked by the Witte, 1992a; Witte &
incongruence between Morrison, 1995a
recommended health state and
audience’ actual state.

Iatrogenic Deleterious effects of Becker, 1993; Burgo & Perez,
therapeutic treatment causing 1986; CANDI, 1990; Pavich,
new pathology independent of 1986; Worth & Rodriguez,
the old. 1987; Wyatt, 1988

Boomerang The phenomenon in which Blood & Snyder, 1993;
audience respond to the Feingold & Knapp, 1977;
opposite way recommended by Kraus, Riggins, & Franti, 1975;
a message Reinarman & Levine, 1995;

Snyder & Blood, 1992; Stuart,
1974
Stigrnatization The phenomenon in which Kolata, 1992; Levine, 1991;

every action intended by a
campaign brings about equal
and opposite reaction.

Salmon, 1996; Wang, 1992

 

Table 1b

Typology of Unintended Effects of Health Communication Campaigns
Predominantly Institutional-Level,lndirect, Long-Term, and

 

 

 

Content/Target Diffuse
Unintended Definitions Examples
Effects
Epidemic of The phenomenon in which the Barsky, 1988a; 1988b; Becker,
Apprehension pervasiveness of fear appeals 1993; Illich, 1976; McKeown,
over the long term produce 1979; Mechanic, 1983; Miller,
unnecessary consciousness and Murphy, & Buss, 1981;
concern over health. Pennebaker, 1982
Desensitization Narcotizing dysfunction of the Becker, 1993; Downs, 1972;

Culpability

Social
Reproduction

Enabling

Opportunity
Cost

mass media which induces
desensitization toward health
issues.

The phenomenon of locating
the causes of problems within
the individual, rather than in
social and structural conditions.

The phenomenon in which
campaigns reinforce existing
social inequities.

The phenomenon in which
defining a condition as a matter
of public health produce and
promote industries.

The phenomenon in which
emphasis on individual
responsibilities impede the
progress in public health
through structural change.

Lazarsfeld & Merton, 1951;
Middlestadt, 1991

Guttman, 1997; Middlestadt,
1991; Milio, 1981; Minkler,
1990; Paletz, Pearson, &
William, 1986; Salmon, 1989;
Wallack, 1989

Ball-Rokeach, Grube, &
Rokeach, 1981; Fairly, Haddad,
& Brown, 1996; O’Keefe,
1971; Salmon, Wooten, Gentry,
Cole, & Kroger, 1995; Vidmar
& Rokeach, 1974

Barsky, 1988a; 1988b; DeJong
& Wallack, 1992; Fischer,
Richards, Berrnan, Krugman,
1989; Glanz, 1996

Alonzo, 1985; Antonovsky,
1995; Becker, 1993; Cassel,
1976; F aden & Faden, 1978

 

Unintended Effects of Health Communication Campaigns

Before reviewing the types of unintended effects, a discussion of sources of such
effects may be necessary. Unlike efficacy, "the level of desired effect of a program when
delivered and received under optimum conditions” (Flay, 1986, p. 468), effects of health
communication campaigns are determined by a confluence of factors operating in the real
world. Flay (1986, p. 468) stated that effectiveness or intended effects refer to “the level
of good over harm that a program achieves when received under typical real-world
conditions of availability and acceptance." So do unintended effects.

Therefore, sometimes inadequacies or lack of the factors that produce intended
outcomes may create unintended effects. For example, error in audience segmentation or
insufficient formative research can result in unanticipated consequences. Next, unlike
commercial messages, the placement of public health messages tend to not be under
control of the campaign planners. In this case, good segmentation and formative research
lose their effectiveness by being delivered to an unintended audience (Salmon, 1996).

According to Salmon (1996), however, there are other factors that make unique
contributions toward unintended effects. Dogmatism, by driving actions strictly out of
principle without regard for consequences for others (Merton, 195 7), tends to produce
distress among marginalized, such as the smokers. These unintended effects can be
largely categorized into either individual/short-term level or societal/long-term level

(Salmon, 1996).

Individual/Short-Term Level

Obfugation. One of the unintended effects of health communication campaigns
observed at individual level includes obfuscation, misunderstanding and subsequent
misguided behavior. For example, messages designed to reduce unnecessary physician
visits resulted in reductions in necessary visits among Afiican-Americans and Medicaid
patients (Roberts, Imrey, Turner, Hosokawa, & Alster, 1986).

Dissonance. Dissonance is another type unintended effects occurring at
individual level when recommended health state and audience’ actual state are
incongruent. Guttman and Zimmerman (1996) found that in the face of health education
messages promoting breast feeding and the society's disapproval of breast feeding in
public, infants-raising women felt guilty, deprived, and were frustrated.

Boomerang. Boomerang effects, in which audience react to the opposite way
recommended by a message, have long been recognized in various persuasion contexts,
including fear appeals in health communication. After being exposed to fear appeals,
audience reported reduced level of worry about risk (Stuart, 1974), less negative attitude
toward risk behavior (Feigngold & Knapp, 1977; Snyder & Blood, 1992), stronger
intentions to engage in risk behavior, and increased risk behavior (Stuart, 1974; Witte,
1992a).

Iatrogenic. The term "Latrp" is the Greek word referring "physician," and
"M" is "origin." Illich ( 1976) used this word to refer to adverse clinical, social, and
cultural impacts created by the medical establishment. Salmon (1996) used the term

"iatrogenic" to refer to unintended effects of health communication campaigns

engendering new pathology. For example, Kaplan (1985) reported that while dietary
changes were associated with reduced mortality of heart disease, this decrease was related
to increases in the mortality of cancer.

Stigmatization. Health communication messages that attempt to scare audience
with possible consequence of disabilities may foster stigmatization (Wang, 1992). Wang
argued that by suggesting that disabilities are undesirable, messages such as the following
can render the perceptions toward the disabled as negative: “If you think seat belts are
confining, think about a wheelchair;” “Last year. 1057 teenagers got so drunk they
couldn’t stand up. Ever" (p. 1099).

Societal/Long-Term Level

Over time, unintended effects of health communication campaigns may become
cumulative and render changes in societal level (Salmon, 1996).

EJidemic of apprelflon. The ubiquity of messages about health risks are
considered to have created a climate of fear (Becker, 1993; Tenner, 1996). This
phenomenon, which Thomas (1983) termed as “an epidemic of apprehension,” may
undermine public's sense of and well-being by raising consciousness and concern over
health unnecessarily high. Sense of insecurity produced by such messages may make the
public perceive that health risks “1qu in every aspect of daily life: the air we breathe, the
water we drink, the food we eat, the homes we live in, the substance we touch, and the
work we do” (F einstein & Esdaile, 1987, p. 113). Further, F einstein and Esdaile (1987)

suspected that such unnecessary fear, often aroused by messages lacking credible and

rigorous scientific standards, has generated proposals for changes in policies for health,
nutrition, and industrial safety.

Desensitization. Another type of unintended effects of health communication
campaign that occur over long-term at societal level is desensitization. Larzasfeld and
Merton (1951) predicted that over time, increasing exposure to campaign messages about
health issues may inadvertently render the public apathetic toward such issues.
Empirically, Kinnick and others’ (1996) study of compassion fatigue found that long-
terrn, diffusive media coverage was linked to desensitization and emotional burnout
toward issues including AIDS, homelessness, child abuse, and violent crime.

Culpabilig. By stressing individual responsibility, health communication
campaigns have fostered the tendency to locate the causes of illnesses and diseases within
individual. As a result, campaigns located culpability in individuals, instead of social,
economic, political, and cultural systems that may determine an individuals' health status
to a greater degree (Guttman, 1997; Milio, 1981; Paletz, Pearson, & William, 1986;
Rakow, 1989; 1997; Salmon, 1985; 1989; Wallack, 1989).

Social reproduction. Outcomes of health communication campaigns tend to
reinforce existing social inequities. For example, a 1995 study on the public's AIDS
knowledge found that low education group had lesser knowledge on both true- and false-
transmission of AIDS than the high education group (Salmon, Wooten, Gentry, Cole, & ’
Kroger, 1995).

Enabling. Health issues that are communicated in campaigns tend to enable

industries including food, condom, charity, medicine, and social marketing (Salmon,

1996). Designated driver campaign was embraced by and enabled the alcohol industry by
encouraging non-drivers' drinking and blurring attention on other issues such as underage
drinking (DeJong & Wallack, 1992). Similarly, stop underage smoking campaign
enabled the tobacco industry, by implying that a cigarette is an "illicit pleasure,"
"declaration of independence," and "self-identity" (Glantz, 1996, p. 156).

Opportunity cost. To the extent that health communication campaigns stress
individual responsibilities, opportunities to change individuals' socioeconomic
environments which have direct relationship with their health status may be lost (Alonzo,
1985; Becker, 1993; Cassel, 1976; Faden & Faden, 1978; Ferudenberg, 1981; Mechanic,
1988; Slater & Carlton, 1985).

Of Salmon's (1996) typology. three categories particularly pertain to unintended
effects of fear appeals: "boomerang," "dissonance," and "epidemic of apprehension."
And the fear appeal is one of the message strategies in health communication campaigns
that has most often been recognized for its unintended consequences. In this context, this
study focuses specifically on the identification and investigation of the conditions under
which unintended effects of fear appeals are likely to occur. The next section discusses
each of the unintended effects of fear appeals in detail. Table 2 offers a summary.

Unintended Effects of the Fear Appeal

One of the earliest studies on unintended effects of fear appeals is Johns Hopkins
University researchers' work on the impact of motion pictures used for venereal disease
control (Lashley & Watson, 1929). Specifically, Lashley and Watson were concerned

that fear appeals in such motion pictures may elicit a phobia encumbering intended

outcomes. The results of the study did not find support for this hypothesis. Lashley and
Watson surmised that fear arousal was not intense enough to trigger any unintended
effects. Over the years, however, studies of fear appeal have reported a range of
unintended effects that can be summarized into either boomerang or dissonance in
Salmon's (1996) typology.
Boomerang

Boomerang effects of fear appeals were observed in audience' attitude, intentions,
and behavior. Hovland et a1. (1953) and Witte (1992a) suggested and documented other
types of negative reactions to fear appeals that may bring about boomerang effects:
defensive avoidance, message derogation, and perceived manipulation. The following
sections discuss each of these effects.

Attitude. For example, after exposure to anti-drug abuse public service

 

announcements, attitudes toward amphetamines changed significantly less negative than
prior to exposure (Feingold & Knapp, 1977). College students' attitude toward drinking
became less negative after being exposed to alcohol warning labels (Blood & Snyder,
1993; Snyder & Blood, 1992). Similarly, results of a drug education found that
participants significantly reduced their worry about drugs in general than non-participants
(Stuart, 1974).

Intentions. English smokers indicated stronger intentions to smoke after being
exposed to cigarette warning labels (Hyland & Birrell, 1979). Rogers et a1. (Kleinot &
Rogers, 1982; Rogers & Mewbom, 1976) found that high threat messages, when coupled

with low response efficacy, increased intentions to smoke or drink.

12

Behavior. Stuart (1974) found that participants of a drug education significantly
increased the purchase of marijuana and LSD than non-participants (Stuart, 1974). Witte
(1992a) found that when perceived efficacy was low, threatening messages prompted
individuals to engage in unprotected sex more often.

Defen_sive avoidance. After being exposed to a fear-arousing message,
individuals attempt to avoid the thoughts and communications of risk and prevention
presented in the message. While some researchers cast doubt on defensive avoidance
hypothesis, others' contention that lack of evidence stems from ethical consideration
delimiting the upper bound of fear manipulation is shared by many (see Mongeau, 1995).
Janis and Feshbach ( 1953) first suggested defensive avoidance as an unintended effect of
threat messages, and McCrae (1984) identified avoidance as a coping mechanism that
particularly deals with threat. Defensive avoidance can be most notable type of
unintended effects of fear appeals because it was inversely related to attitude, intentions,
and behavior (Rippetoe & Rogers, 1987; Witte, 1992a).

Message derogation. When perceived threat is high, individuals express dislike of
the message and its presenter. Witte (1992a) found that message derogation was
negatively related to intended attitudes.

Perceived manipulation. When perceived threat is high, individuals tend to
perceive that the message is deliberately trying to scare them. This type of reaction has
been reported in many fear appeal studies including Janis and F eshbach's (1953),

Leventhal, Singer, and Johns' (1965), and Witte's (1992a).

Dissonance

Rippetoe and Rogers (1987) adopted McCrae's (1984) coping mechanism,
fatalism, hopelessness, and wishful thinking, to better understand maladaptive responses
to threatening messages. These reactions can be best categorized as a form of dissonance
in Salmon's (1996) typology. Rippetoe and Rogers found that they occur when the levels
of perceived threat and efficacy are incongruent (i.e., when threat was high but efficacy
was low). Whether these reactions are linked to boomerang effects has yet to be
determined.

FLtaflspg. Researchers have suggested (Casey, 1995) and found (Rippetoe &
Rogers, 1987) that fatalism may obstruct intended effects of fear appeal messages. In
fatalistic reactions, instead of thinking about and acting upon ways to avert a possible
threat, individuals accept it as given and engage in non-action. Rippetoe and Rogers
reasoned that fatalism may be most unsalutary because by accepting the threat, the
individuals do not engage in any type of coping responses to threat at all.

Hopelessness. If fatalism is a result of lack of belief in message
recommendations, hopelessness is more likely to be an outcome of lack of confidence in
one's ability to perform recommended behavior (Rippetoe & Rogers, 1987). Rogers
(1983) described this reaction as "posture of resignation and helplessness" (p. 477).

Wishful thinking. Hovland et al. (1953) suggested that when message
recommendations are perceived to be not efficacious, a highly likely threat may motivate
individuals to form "magical" or "wishful" thoughts. McCrae (1984) found wishful

thinking as a coping reaction to threat. Specifically, subjects wished that a miracle would

happen, they could change the past, and the threat would go way. Rippetoe and Rogers
(1987) found highly threatening message on breast cancer elicited wishful thought among

those who perceived low efficacy of breast self examination.

15

Table 2 Unintended Effects of Fear Appeals

 

 

Unintended Definitions Examples
Effects

Defensive Avoidance of exposures to thoughts Janis & F eshbach, 1953;

avoidance and communications about threat after F ruin et al., 1991; Schafer
being exposed to fear appeal messages et al., 1993; Witte, 1992a;
(Hovland et al., 1953). Witte & Morrison, 1995a

Message Dislike of communication and Witte, 1992a; Witte et al.,

derogation complaints about the content (Hovland 1998
etaL,l953)

Perceived Belief that the communicator is Witte, 1992a, Witte et al.,

manipulation deliberately attempting to scare them 1998
(Hovland et al., 1953).

Fatalism The acceptance of a stressful situation Fruin et al., 1991; McCrae,
as unchangeable and complacency in 1984; Rippetoe & Rogers,
the face of danger because nothing can 1987; Schafer et al., 1993
be done (Rippetoe & Rogers, 1987, p.
598)

Hopelessness Learned helplessness resulting from Fruin et al., 1991; McCrae,
absence of belief in possible solutions 1984; Rippetoe & Rogers,
to a threat (Rippetoe & Rogers, 1987, 1987
p.598)

Wishful A response that prompts the use of Hovland et al., 1953; Fruin

thinking panaceas or unrealistic solutions to a et al., 1991; McCrae, 1984;

problem (Rippetoe & Rogers, 1987, p.
598)

Rippetoe & Rogers, 1987

 

Table 2 (cont'd).

 

 

Unintended Definitions Examples
Effects
More Overall evaluation of the behavior that F eingold & Knapp, 1977;
favorable involves health threat in a positive Blood & Snyder, 1993;
attitude manner. Snyder & Blood, 1992
toward risk
behavior

Less favorable Overall evaluation of the behavior that

attitude reduces health threat in a negative
toward manner.

recommended

behavior

Increased Expression of a greater degree of

intention to intention to engage in behavior that
engage in risk involves health threat.

behavior

Counter- Engagement in behavior that involves
productive health threat more frequently or to a
behavior greater degree.

Witte, 1992a

Hyland & Birrell, 1979;
Kleinot & Rogers, 1982;
Rogers & Mewbom, 1976;
Witte, 1992a

Witte, 1992a

 

The understanding of health communication campaign effects will not be

complete when efforts are confined to the evaluation of intended outcomes. Investigation

of unintended effects is particularly important for health communication campaigners

because they have an ethical mandate to “do no harm” even while attempting to “do

good” (Wang, 1992).

Because of the prominence of fear appeals illustrated above, the present study

attempts to identify and investigate the conditions under which unintended effects of fear

appeals may occur. The extended parallel process model (EPPM) and the transtheoretical

model (TM) were used as theoretical frameworks. A lab experiment (n=274) was
conducted to examine college students' reactions to fear appeals used in skin cancer
prevention messages. Threat was manipulated in two (high, low) levels. Behavior was
measured in a four-week follow-up.

This study is organized into four chapters. Chapter one introduced unintended
effects of health communication campaigns and fear appeals. Chapter two reviews the
literature on fear appeals, focusing on their unintended effects. Chapter three describes
the method of this study. Chapter four presents the results of this study. Chapter five
discusses the findings and their implications. It also presents the limitations of this study.
The appendices include informed consent form. high and low threat messages, and

measures.

18

CHAPTER 2

LITERATURE REVIEW

This chapter begins by discussing the definition of the fear appeal and theoretical
approaches to the unintended effects of the fear appeal. Drive, expectancy-value,
psychological reactance, and parallel-process models including the EPPM are discussed.
Next, this chapter describes the TM, a stages of change model, and how it relates to the
EPPM. Finally, it concludes by presenting hypotheses and research questions.

Defining_the Fear Appeal

Fear has been part of human condition, and the fear appeal shares equally long
history. For example, the Bible abounds with fear appeals: "...but you must not eat from
the tree of the knowledge of good and evil, for when you eat of it you will surely die"
(Genesis 2: 17). Death in this passage represents a key construct of the fear appeal which
researchers have termed "unfavorable consequences" (Hovland, Janis, & Kelley, 1953),
"magnitude of noxiousness" (Rogers, 1975), or "severity" (Witte, 1992b). However, fear
as persuasion tool has not been systematically studied until the early 19505 by Yale
school of psychologists led by Janis. Interestingly, Janis' (Janis & Feshbach, 1953)
seminal work on fear appeals focuses on their unintended and adverse effects of fear
appeals.

What is fear, and what is the fear appeal? What makes the fear appeal so
notorious for generating unintended effects, but still so prevalent in many communication

contexts including public health? Fear is a form of strong, negative affect (Fiske &

l9

Taylor, 1991). Appeals refer to those that motivate individuals to comply to message
recommendations (Hovland et al., 1953). An obvious corollary is that fear appeals are
those that motivate individuals to comply to message recommendations through the
arousal of fear.

However, a look at theories of the fear appeal makes one wonder if the definition
above correctly describes the fear appeal. Fear has faded away from the discussion of
fear appeals (see next section on Rogers, 1975; 1983) after Janis' work. Researchers
argued that motivation comes from the thoughts on threat, rather than the emotion of fear.
Fear has been reinstated recently, but still does not play a central role in inducing
intended effects of health communication campaigns (see next section on Witte, 1992b;
1994;1998)

To put the fear appeal in perspective for this study, a review of fear appeal
theories and their explanations for unintended effects is in order.

Theoretical Approaches to the Unintended Effects of the Fear App_eal

In general, four major theoretical approaches have been used to explain
unintended effects of fear appeals: drive, expectancy-value, psychological reactance, and
parallel process. Researchers (Boster & Mongeau, 1984; Dillard, 1995; Hale & Dillard,
1995; Witte, 1992b; 1998) have usually classified fear appeal theories into three major
categories: drive, expectancy-value, and parallel-process models. In addition to these
major theoretical frameworks for fear appeals, psychological reactance (Brehm, 1966) has

also been used to explain unanticipated effects of fear appeals.

20

Drive Models

The seminal work on unintended effects of fear appeals was done by Janis and his
colleagues (Janis, 1967; Janis & Feshbach, 1953; Hovland, Janis, Kelly, 1953). In their
studies, unintended as well as intended effects of fear appeals are explained in terms of
drive. Drive, the key construct of learning theory, refers to “emotional tension” (p. 78).

Hovland et al. (1953) used the term "threat appeal" to refer to a motivational
message portraying "negative consequences" of noncompliance to message
recommendations. They expected that such messages will arouse varying degrees of fear.
The experience of fear then elicits drive which motivates individuals to engage in
behaviors to reduce the drive. Therefore, the greater the fear, the greater compliance to
message recommendations.

However, Janis and F eshbach's (1953) experiment found a negative and linear
relationship between the level of fear and compliance to message recommendations on
dental hygiene. Subsequently, Janis and F eshbach hypothesized three conditions under
which intended effects of fear appeals may be hampered and defensive reaction and
therefore unintended effects may occur instead: when the intense experience of fear
renders individuals (1) fail to attend to messages and (2) be aggressive toward message
presenters, and (3) avoid thoughts and communications about risks presented in messages
afier exposure.

The hypothesis about defensive avoidance was supported, but the ones about
inattentiveness and aggression toward communicator were not. As a result, Janis and

Feshbach (1953) argued that avoidance of message cues after exposure will occur when

21

emotional tension aroused by high level of fear is not sufficiently reduced. A factor that
may reduce the tension, according to Janis and F eshbach, is reassurance which may come
either from the self or message recommendations. To be more specific, Hovland et al.
(1953) predicted that unintended effects will occur when message recommendations are
perceived to be ineffective in deterring the threat or to be "impossible to carry out" (p.
77). A seed of droughts about the role of self and response efficacy are found here.

Later, Janis (1967; Hovland et al., 1953) proposed a curvilinear hypothesis.
According to this hypothesis, as the level of fear increases, compliance to message
recommendation increases up to an optimal point. After reaching this point, compliance
decreases and defensive reactions becomes prominent, formulating an inverted-U shape
curve of compliance as a function of fear. The hypothesis was criticized by Leventhal
(1970), Rogers (1975; 1983), and Beck and Frankel (1981) for many reasons including its
post-hoe, unfalsifiable nature. Most of all, this curvilinear hypothesis is at odds with the
results of Janis' (Janis & Feshbach, 1953) earlier experiment which found negative and
linear association between fear and compliance (Hale & Dillard, 1994).
WCV-Value Models

Expectancy-value theories posit that behavior is determined by an individuals'
strength of expectations about consequences of behavior and the value that one assigns to
the consequences (e.g., Rotter, 1954). This theoretical approach has been adopted to
explain and predict health-related behavior.

Protection motivation theory. Roger’s (1975; 1983) original and revised

protection motivation theory (PMT) represents an expectancy-value approach to fear

22

appeals. These models depart from Janis’ drive approach because they propose that
compliance to message recommendations is an outcome of cognitive processes, rather
than emotional tension. Rogers (1975) contended that protection motivation, which
induces changes in attitude and behavioral intention, is the result of cognitive appraisal.
Thus, although Rogers called his PMT as a "theory of fear appeals," fear was absent in
his prediction of intended attitude change.

From this perspective, Rogers ( 1975) first delineated three key components of fear
appeals, magnitude of noxiousness, probability of occurrence, and efiicacy of
recommended response. These variables of the PMT work in a multiplicative manner to
elicit protection motivation and intention to comply to message recommendations.
Explanations for possible adverse effects of fear appeals were not included in the original
PMT.

Studies testing the original PMT found unintended effects resulting from
interactions between noxiousness or probability of threat and response efficacy (Kleinot
& Rogers, 1981; Rogers & Mewbom, 1976). When response efficacy was low, increases
in probability and/or noxiousness of threat induced greater intention to engage in risk
behavior, such as smoking and drinking.

Perhaps because of these findings, Rogers (1983) repostulated the original PMT.
New variables, including self efficacy, intrinsic (i.e., physical and psychological) and
extrinsic (i.e., social) rewards of risk behavior, and response costs were integrated into the
revised PMT. Most important to this discussion of unintended effects of fear appeals, the

revised PMT delineates the processes leading to maladaptive responses (i.e., the

23

responses message designers did not intend) including continuation or initiation of risk
behavior.

Overall, the revised PMT suggests three processes that may bring about
unintended effects. First, Rogers suggested that the intrinsic and extrinsic rewards of risk
behavior increase the likelihood of maladaptive responses to threat, while noxiousness
and probability of threat decrease such likelihood. Therefore, when the rewards of risk
behavior outweigh its noxiousness and probability, unintended effects are likely to occur.

Second, the revised PMT suggests that maladaptive responses may occur when
the costs of complying to message recommendations are perceived to be greater than self
efficacy and response efficacy of engaging in recommended behavior.

Finally, Rogers (1983) predicted that perceptions of low self efficacy or response
efficacy in the face of strong noxiousness and/or probability of threat will result in null or
boomerang effects of fear appeals, in which individuals reduce intention to engage in
protective behavior.

This prediction of the interactions between efficacy and threat perception
variables is consistent with the results of Rogers’ earlier studies (Kleinot & Rogers, 1982;
Rogers & Mewbom, 1976). Rippetoe and Rogers (1987) examined maladaptive
responses to messages on the threat of breast cancer, using the revised PMT. The results
indicated that high threat perceptions prompted both adaptive and maladaptive coping
responses and that what determines the type of coping responses is the level of self and

response efficacy, evidencing interaction between threat and efficacy.

24

The revised PMT’s utility in explaining unintended effects of fear appeals is
limited, however. The model’s predictions on the combinatory rule of its variables are
inconsistent. At one point, Rogers (1983) stated that severity, vulnerability, self efficacy,
and response efficacy work in an additive manner to elicit protection motivation.
Elsewhere in the same work, Rogers predicted a multiplicative combination of the same
four variables (Witte, 1998).

Health belief model. Another expectancy-value model that posits threat as a
motivational factor is the health belief model (HBM) (Rosenstock, 1974; Janz & Becker,
1984). Bearing much resemblance to fear appeal theories, the HBM predicts that
perceived susceptibility and severity motivate individuals to engage in preventive actions,
and perceived benefits and barriers determine the type of preventive action to be engaged.
The HBM has been used and supported in extensive domains of health communication
campaigns.

However, some pointed out that many of HBM studies were retrospective, which
limits the inference of causality (Kegeles, 1981). The HBM has also been criticized for
its lack of specificity in the description of combinatory rules of variables. Therefore, the
HBM may provide a set of variables that produce intended effects, but no theoretical
framework explaining neither intended nor unintended effects (Wallston & Wallston,
1984)

Possibly because of these factors, reports of unintended effects are rarely found in
reviews of the HBM studies (e.g., Montgomery et al., 1989; Rosenstock, 1974). One

exception is Joseph et al.'s ( 1987) longitudinal study of a cohort of gay men. Increased

25

susceptibility was associated with deleterious effects such as obsessive/compulsive
behavior, distress in social role playing, and distractive concerns and worries about
AIDS.

Psychological Reactance

Yet another explanation advanced for unintended effects of fear appeals was
psychological reactance (Brehm, 1966). According to reactance theory, individuals
experience reactance when they perceive their freedom to engage in different kinds of
behavior is threatened. As a result, individuals will be motivated to restore the behavioral
freedom by changing their attitude toward communicator negatively and/or engaging in
behavior opposite to the ones recommended.

Reactance theory has been supported in a range of areas including alcohol
prevention messages (Bensley & Wu, 1991) and patient compliance (Graybar,
Antonuccio, Boutilier, & Varble, 1989). For example, Allen, Sprenkel, and Vitale (1994)
predicted that after the change of US. legal drinking to 21 years of age, underage college
students would drink more than their legal age counterparts because of perceived
reduction in behavioral choices. The results lent support to reactance theory.

That reactance triggers reactions opposite to message recommendations appears to
be relatively clear. For example, derogation of fear-arousing messages was negatively
associated with attitude (Witte, 1992a). Similarly, Heilman and Garner (1975) report that
toward messages and communicators restricting their behavioral choices, subjects
expressed negative attitudes. However, research to date does not provide a clear picture

whether the feeling of fear creates reactance. Meta-analysis of fear appeal studies (Boster

26

& Mongeau, 1984; Mongeau, 1995) found that the relationship between fear and
compliance to message recommendations is positive and linear.

It may be necessary to distinguish threats to freedom from threats to health.
Brehm’s (1966) prediction would pertain to the former. It has yet to be determined
whether the latter will induce the experience of reactance as well. For example, Bensley
and Wu (1991) manipulated high and low threat to determine alcohol prevention
messages’ effects on college students. Although the results clearly supported the
predictions of reactance, what was actually manipulated in their study was the level of
dogmatism in message recommendations, not the level of susceptibility or severity of
alcohol use. Therefore, the construct of fear in and of itself may not necessarily elicit
reactance.

A fear-arousing health communication message has a potential to produce
reactance, however, because it by default includes behavioral recommendations (Rogers,
1975; 1983; Witte, 1992b). Hence, it may be perceived as more inhibitive than other
types of messages.

_P_ar_allel Process Models

P_arallel process model. Leventhal (1970; 1971; Leventhal, Safer, & Panagis,
1983) theorized that fear appeals can elicit two kinds of processes: danger control and
fear control. Danger control process refers to the one in which individuals think about
ways to avert the threat presented in fear appeal messages. Therefore, this is a cognitive

process leading to intended outcomes of fear appeals.

27

On the other hand, fear control process produces the effects health communication
campaigners would not intend in designing and delivering fear appeal messages. In this
process, the experience of fear motivates individuals to reduce the emotional tension by
denying the threat presented in messages or avoiding any cues that may create fear.
Therefore, fear control process accounts for unintended effects of fear appeals. Overall,
what motivates individuals to comply to message recommendations is not emotional
arousal, but cognitive appraisal, according to this model.

While some researchers lauded the model for its recognition of differential
processing of fear appeal messages, others pointed out its lack of specificity and
unfalsifiable nature (Boster & Mongeau, 1984; Dillard, 1994; Rogers, 1975; 1983; Witte,
1994). The extended parallel process model (Witte, 1992b; 1994; 1998) explicates fear
control as well as danger control processes.

The extended flallel process model. The EPPM (Witte, 1992b; 1994; 1998) is
built upon Leventhal’s (1970; 1971; Leventhal, Safer, & Panagis, 1983) parallel process
model. In explaining intended effects of fear appeals, the EPPM adopts Rogers’ (1975;
1983) protection motivation theory. In explaining unintended effects of fear appeals, it
partly borrows from Janis’ (1967) drive model.

The key components of the EPPM are described below (Witte, Cameron,
McKeon, & Berkowitz, 1996, p. 320). Fear refers to a negative emotion aroused by a
threat. Threat refers to a risk, harm, or danger. Perceived threat refers to the beliefs
about the risk. It comprises perceived susceptibility and severity. Perceived

susceptibility refers to personal beliefs about the probability of experiencing the threat.

28

Perceived severity refers to personal beliefs about the magnitude of the threat. Perceived
severity and susceptibility work in an additive manner.

Efficacy comprises response and self efficacy. Response efficacy refers to
personal beliefs about the effectiveness of message recommendations in controlling the
threat. Self efficacy refers to personal beliefs about one's ability to carry out the message
recommendations to control the threat. Self and response efficacy work in an additive
manner.

According to the EPPM, the processing of a fear appeal message can result in
three types of responses (Witte et al., 1996). When a threat does not elicit sufficiently
high level of susceptibility or severity, individuals are not motivated to process the
message, and the fear appeal does not result in any responses.

Next, when a threat is sufficiently high to reach a certain threshold level,
individuals are motivated to engage in either danger control process or fear control
process. First, individuals evaluate the threat. The greater the threat perceived, the
greater the motivation to appraise the level of efficacy. It is the level of perceived
efficacy that determines whether an individual will go through danger control or fear
control process.

D_anger control process. When individuals' perceive efficacy is stronger than
perceived threat, they engage in danger control process. This is primarily a cognitive
process which elicits protection motivation. In this process, individuals believe that the
they are susceptible to a threat which is severe, and that they are able to effectively avert

the threat. Therefore, protection motivation prompts them to ways of averting the threat

29

and to adopt message recommendations. EPPM research has measured danger control
process in terms of changes in attitude, behavioral intentions, and behavior. This danger
control process would be intended effects of health communication campaigns using fear
appeals, because health communication campaigns intend to induce attitude and behavior
changes (Rogers & Storey, 1987).

Fear control process. When the perceptions of threat is stronger than the
perceptions of efficacy, individuals engage in fear control process. This is primarily an
emotional process, prompted by defensive motivation. Fear control process occurs when
individuals believe that they are susceptible to a threat which is severe, but do not believe
themselves to be able to perform message recommendations and/or that the message
recommendations to be effective in averting the threat. High level of fear caused by this
process produces defensive motivation, resulting in coping responses that reduce fear and
prevent danger control responses from occurring. EPPM research has measured fear
control process in terms of defensive avoidance, message derogation, and perceived
manipulation. This type of responses would pertain to unintended effects of health
communication campaigns using fear appeals, because individuals avoid or react to
persuasive messages, and do not attempt to deter threats to their health condition.

The EPPM departs from the PMT and the original Parallel Process Model by
reinstating fear to fear appeals. Rogers (1975) posited that fear does not influence
intended attitude change neither directly not indirectly. Leventhal (1970) theorized that
fear control leading to unintended effects and danger control leading to intended effects

are mainly independent.

30

However, metaanalyses of fear appeals (Boster & Mongeau, 1984; Mongeau,
1995) have found positive and linear relationship between perceived fear and intended
effects. The EPPM offers a reconciliation by suggesting an indirect relationship between
fear and intended effects. Specifically, when perceived efficacy is high, the association
between fear and intended changes in attitude, intentions, and behavior is mediated by
perceived threat. As a result, fear has returned in the process of motivating individuals to
comply to message recommendations, although indirectly. Fear produces intended
effects only when efficacy is high; if not, fear primarily produces unintended effects
(Witte, 1992b).

The EPPM has been tested in various health communication areas such as AIDS
(Witte 1992a), genital warts (Witte, Berkowitz, Cameron, & McKeon, 1998), tractor
safety practices (Witte, Peterson, Vallabhan, Stephenson, Plugge, Givens, Todd,
Becktold, Hyde, & Jarrett, 1993), and electromagnetic fields (McMahan, Witte, & Meyer,
1998a). In summary, the EPPM explains and predicts the conditions under which
intended and unintended effects of fear appeals are likely to occur. This study uses the
EPPM to examine unintended effects of fear appeals.
Fear Apmals and Audience Characteristics

The EPPM posits that individual differences may influence the effects of fear
appeals. Indeed, audience characteristics have long been noted to impact the effects and
effectiveness of fear appeals. For example, fear appeals were more effective for older
audience than younger audience (Boster & Mongeau, 1984; Carmel, 1995). Individuals

with high trait anxiety showed greater magnitude of defensive avoidance than those with

31

low trait anxiety (Witte & Morrison, 1995a). A study found that high fear was more
effective for high sensation seekers, while low fear was more effective for low sensation
seekers (Donohew, Sypher, & Buloski, 1991). On the other hand, Witte and Morrison
(1995b) found that high sensation seekers were not persuaded by fear appeal messages
regardless of fear arousal level. Horowitz (1969; 1972) found that audience who
volunteered to be exposed to fear appeals were more likely to be persuaded than non-
volunteers.

One of the audience characteristics that has yet to be applied to fear appeal
research to determine its influence on unintended as well as intended effects is stages of
behavior change. The next section describes the transtheoretical model, a theory of stages
of change. Using the EPPM and the TM, this study attempts to examine the effects of
fear appeals on individuals in different stages of change.

The Transtheoretical Model

A model that elucidates the dynamic process of health behavior change is the TM
(Prochaska & DiClemente, 1988; Prochaska, DiClemente, & Norcross, 1992). The TM is
a more integrative approach to behavior change as its key components, stages of change,
may subsume many different factors explained and predicted by other theories.

Based on the assumption that behavior change is an ongoing process involving a
series of stages, the TM has been tested and supported in diverse areas including smoking
cessation (Prochaska & Diclemente, 1983; De Vries & Backbier, 1994), alcohol
abstinence (Diclemente & Hughes, 1990), sun screen use (Rossi, 1989), dietary change

(Glanz, Patterson, Kristal, DiClemente, Heimendinger, Linnan, & McLerran. 1994), and

32

contraceptive use (Grimley, Riley, Bellis, & Prochaska, 1993). A similar pattern of
stages of change emerged with regard to these different kinds of behaviors (Prochaska et
al., 1994; Rimer, 1996).

Stages of Change

Within the model are three major components, stages and process of change and
decisional balance which are explained below (Prochaska et al., 1992).

Precontemplation. The stage at which individuals have no intention to stop a
risky behavior or start a healthy behavior in the foreseeable future. The TM measures the
intentions to change in six-month time frame. Thus, precontemplators are those who do
not intend to engage in changes within six months. Individuals at this stage do not intend
to engage in change for many possible reasons. For example, they may lack knowledge
of the risk involving their current behavior. Or, they may be unwilling to acknowledge
that their behavior puts them at risk.

Contemplation. The stage at which individuals consider stopping their risky
behavior or starting a healthy behavior in the foreseeable future. Individuals at this stage
have not engaged in behavioral change, but they intend to initiate changes within six
months. The TM also predicts that individuals may stay at this stage for a long period of
time due to the difficulty in evaluating the advantages and disadvantages of behavior
change.

Preparation. The stage at which individuals intend to change their risky behavior
or start a healthy behavior. Individuals at this stage have not engaged in behavior change,

but they intend to do so within a month.

33

ALtIQLI. The stage at which individuals engage in behavior change. Individuals in
the action stage have tried to stop their risky behavior or start healthy behavior, but for no
more than a month.

Maintenance. The stage at which individuals routinize the healthy behavior and
try to prevent regression into the risky behavior. Individuals at this stage have engaged in
behavioral change for more than six months.

Regression is possible at any point of these stages of change, and individuals may
or may not resume the progress.

Processespf Chapga

The stages of change explain when changes in behavioral intentions occur. The
processes of change provide what may help progress though the stages of change.
Processes of change comprise an array of coping strategies that individuals use to change
their behavior. Studies (DiClemente & Prochaska, 1982; Prochaska & Diclemente, 1984)
have found that the relative contribution of each process is different from stage to stage.

Precontemplation to contemplation. For example, the move from the
precontemplation to the contemplation stage was facilitated by consciousness raising,
dramatic relief, and environmental reevaluation processes.

Contemplation to preparation. The transition from contemplation to preparation
was expedited by self-reevaluation.

P_rapmtion to action. As persons move from preparation to action, self-liberation

process was effective.

34

Action to maintenance. Behavioral processes, such as reinforcement
management, helping relationships, counterconditioning, and stimulus control were
effective in moving individuals from the action to the maintenance stage (Prochaska et
al., 1992).

DecisionABalaflq

Individuals at different stages of a behavior change are different from each other
in their decisional balance (Janis & Mann, 1977) of perceived advantages and
disadvantages of the new behavior. Decisional balance is a mental balance sheet for
accounting potential gains and costs resulting from any decision. It is assumed that
individuals will not decide to change or continue a behavior unless they perceive
advantages to outweigh the disadvantages.

Prochaska and colleagues’ (1994) study of stages of change and decisional
balance for twelve risky behaviors found that across behaviors the advantages of change
were higher in the contemplation stage than in precontemplation and that the
disadvantages of change were lower in the action stage than in contemplation. This
finding may suggest that progress from the precontemplation to the contemplation stage
involves an increase in the awareness of the advantages of change, and that the transition
from the contemplation to the action stage involves a decrease in the prominence of the
disadvantages of change.

Relationship with Other Models
The TM serves as an important basis for dynamic audience segmentation and

message design. Baranowski (1989-1990) used the TM to recommend both individual-

35

and environmental-level strategies to move target audiences through each of the stages of
change. Andreasen (1995) explained how to apply social marketing strategies to the TM
categorization to help people move through the stages of change. Maibach and Cotton
(1995) described how the social cognitive theory (Bandura, 1986) can facilitate the
progress through stages.

Overall, researchers tend to suggest that in moving persons from
precontemplation to contemplation stage, awareness and knowledge of risk is important.
For contemplators to move to preparation stage, messages addressing advantages of
healthy behavior and disadvantages of risky behavior are recommended. To move
persons from preparation to action and maintenance, providing skills and social support is
emphasized.

Champion (1994) tested the relationship of the variables of the HBM and the TM
with regard to mammography adoption. Women in the contemplation and
action/maintenance stages had significantly higher benefit perceptions than those in the
precontemplation stage. Barrier perceptions were significantly higher for women in the
precontemplation stage than those in the contemplation and action stages. Women in
action/maintenance stages had significantly higher susceptibility and severity perceptions
than those in the precontemplation or contemplation stages.

As such, individuals' behavioral change can be facilitated if stage-appropriate
messages are delivered. Conversely, use of stage-inappropriate messages

may hinder progress or trigger a regression to an earlier stage (Maibach & Cotton, 1995).

36

Therefore, the TM and the EPPM can provide a useful framework to examine the
conditions under which unintended effects of fear appeals may occur. Research is needed
to determine the influence of dynamic audience characteristics of stages of change on the
effects of fear appeals. The following section discusses relationships among TM and
EPPM variables.

mking the Transtheoretical Model and the Extended Parallel Process Model
High Threat

According to the EPPM, whether a fear appeal message results in intended or
unintended effects depends on the level of perceived efficacy, comprising self and
response efficacy. Specifically, the EPPM predicts that high threat messages will bring
about intended effects on attitude, intentions, and behavior when perceived efficacy is
high. On the other hand, when perceived efficacy is low, high threat produces unintended
effects such as defensive avoidance, message derogation, and perceived manipulation.

Studies of the TM have found that the level of self efficacy increases as persons
move from precontemplation to maintenance stages (Gorely & Gordon, 1995; Grimley,
Riley, Bellis, & Prochaska, 1993; Marcus. Eaton. Rossi, & Harlow, 1994; Prochaska,
Harlow, Redding, & Snow, 1990).

Few studies have examined the changes in the level of response efficacy across
stages. However, it would be reasonable to assume that response efficacy increases as
well, as persons move from precontemplation to maintenance stages. Decision balance is
different from response efficacy, but would be somewhat similar as the consideration of

the former may implicitly include the latter. Prochaska et al. (1994) found that when

37

individuals moved from the precontemplation to contemplation stage, the advantages of
recommended behavior came to outweigh disadvantages across 12 risky behaviors, and
the advantages remain outweighing disadvantages from the contemplation through
maintenance stage in 6 out of the 12 behaviors.

Studies on behavior and coping strategies suggested differential effects of threat
on individuals who adopt healthy behavior and who do not. In Fruin, Pratt, and Owen’s
(1991) study, which examined the relationship between exercise status and coping
strategies, active persons responded with adaptive coping strategies, such as rational
problem-solving and intention to engage in recommended behavior, more strongly than
did inactive ones following messages on exercise and cardiovascular disease. Similarly, a
study on food safety (Schafer, Schafer, Bultena, & Hoiberg, 1993) found that individuals
who have engaged in healthy, preventive behaviors and had quality diets were more
likely to respond to a health threat with adaptive coping strategies. From the EPPM
standpoint, it can be conjectured that the response of adaptive coping strategies by those
who have performed recommended behavior (i.e., persons in action and maintenance
stages) may stem from high self and response efficacy developed from their experience.

Therefore, because the level of efficacy increases as persons move from the
precontemplation to maintenance stage, unintended effects of fear appeals are less likely
to occur in later stages than earlier stages.

In this study, the term “unintended effects of fear appeals” refers to the following
(refer to Table 2). The first group of such effects are the outcomes of the fear control

process in the EPPM: defensive avoidance, message derogation, and perceived

38

manipulation. Health communication campaigns using fear appeals intend changes in
attitudes, intentions, and behavior in the direction of message recommendations. Fear
appeal message designers would certainly not intend audience to avoid message cues and
distort or dislike the message.

Next, unintended effects of fear appeals in this study also include the maladaptive
coping strategies to threat examined in Rippetoe and Rogers' (1987) study: fatalism,
hopelessness, and wishful thinking. These coping strategies were first identified by
McCrae as those occurring in reactions to threat. Strategies such as defensive avoidance
and fatalism can be termed as adaptive in the sense that they may reduce distress invoked
by threat (Rippetoe & Rogers, 1987). However, Rippetoe and Rogers used the term
“maladaptive” because these strategies do not “directly manage the threat to physical
well-being by dealing with the reality of the situation” (p. 598). In the EPPM context,
fatalism, hopelessness, and wishful thinking would be outcomes of the fear control
process. Individuals attempt to control fear by becoming fatalistic or thinking wishfully
rather than to control danger by engaging in rational thinking or preventive behavior. In
addition, psychological reactance (Brehm, 1966) will also be examined as a form of
unintended effects of fear appeals as it may trigger boomerang effects in attitude,
intentions, and behavior.

High threat may produce different type of unintended effects in association with
different types of low efficacy. Earlier research on fear appeals (Hovland et al., 1953;

Rogers, 1975) focused on response efficacy and its relationship with threat. Bandura's

39

(1977) self-efficacy was integrated into later theories of the fear appeal including the
revised PMT and HBM and the EPPM.

When perceived response efficacy is low, individuals may respond to high threat
messages with fatalism. Because persons perceive that message recommendations are
not effective in controlling a serious danger to which they are susceptible, they may try to
control fear by accepting the situation and by not following message recommendations
(Rippetoe & Rogers, 1987).

Rippetoe and Rogers (1987) also suggested when perceived self efficacy is low,
persons may respond to high threat messages with hopelessness. It is because while
individuals do not accept the situation as unchangeable, they do not perceive that they are
able to carry out the message recommendations either. As a result, they may feel helpless
and possibly intensify their risky behavior.

Low Threat

According to the EPPM, low threat has no effects because it leads to minimal
message processing and changes in attitude, intentions, and behavior (Witte, 1994).
However, some research suggests that low threat, under certain conditions, may produce
unintended effects as well by decreasing perceived risk and increasing perceived benefits
of risk behavior. This line of reasoning is drawn from studies on the effects of anti-drug
abuse public service announcements (Feingold & Knapp, 1977) and government
warnings on alcoholic products (Blood & Snyder, 1993; Snyder & Blood, 1992).

Although these studies do not discuss the results in explicit terms of low threat, an

examination of the fear appeal messages used in the experiments suggests that it may be

40

the case. When compared to many of the experimental stimuli of high threat which
presented vivid and graphic pictures and emphasized the severity of and subjects’
susceptibility to disease (e. g., Janis & F eshbach, 1953; Janis & Terwilliger, 1962; Kleinot
& Rogers, 1982; Rippetoe & Rogers, 1992; Rogers & Mewbom, 1976; Witte, 1992a;
Witte & Morrison, 1995a), the warnings can be categorized as low threat messages.

The alcohol warning label used in Snyder and Blood’s (1992; Blood & Snyder,
1993) studies reads: “According to the Surgeon General, (1) women should not drink
alcoholic beverages during pregnancy because of the risk of birth defects; (2)
consumption of alcoholic beverages impairs your ability to drive a car or operate
machinery, and may cause health problems.” Apparently, the message does not likely to
elicit much perception of susceptibility among persons who are not pregnant, drivers, or
machine operators. Next, the severity of the consequences of drinking is described just as
“health problems.” Furthermore, the gruesome photos of patients with disease which
have ofien been used in high threat manipulation are absent in warnings. Indeed, Blood
and Snyder (1993) reported that the most important factor accounting for the boomerang
effects of their study was the perception that the likelihood of an accident due to drinking
was low.

Low threat, however, may just result in null effects, instead of unintended effects,
unless some other factors play a role of moderator or mediator. This study attempts to
empirically test if low threat results in unintended effects, and if so under what

conditions.

41

When the probability that threat will be averted by message recommendations
was perceived as low, alcohol warning increased the benefits of risky behavior. If
perceived response efficacy is low, current behavior involving risk may look more
attractive. Under this circumstance, individuals may decide to continue to take the risk,
instead of engaging in the recommended behavior. This unintended effect is most likely
to occur in precontemplation stage. Disadvantages of changing behavior have been found
to outweigh advantages, which may be analogous to response efficacy as discussed
previously, in this stage.

R_el_agionship between Unintended and Intended Effects

In general, the unintended effects of fear appeals are likely to be negatively
associated with intended effects (Leventhal, 1971; Witte, 1992a). For example, defensive
avoidance was found to be negatively associated with subsequent recall of fear appeal
message (Janis & Feshbach, 1953), behavioral intention (Rippetoe & Rogers, 1987), and
behavior (Witte, 1992a).

However, not all unintended effects are likely to interfere with intended effects.
For example, research on stress coping indicated that anxiety is positively associated with
social support seeking (Schacter, 1959; Vitaliano, Russo, Carr, Maiuro, & Becker, 1985).
Janis and F eshbach’s (1953) study found that while subjects exposed to high fear appeals
expressed more of perceived manipulation and message derogation, they were more
likely than others to favorably evaluate the quality of the message.

Hence, it was suggested that such responses to high threat may not trigger the

rejection of message (Hovland, Janis, Kelly, 1953; Janis & F eshbach, 1953). Somewhat

42

similarly, Leventhal, Singer, and Jones’ (1965) study found that while subjects exposed
to high threat messages were more angry than those exposed to low threat messages, they
indicated greater behavioral intention. Moreover, Witte (1992a) reported that perceived
manipulation is positively associated with behavior change.1

Witte’s (1992a) findings also included that message derogation is not related to
attitude change, behavioral intention, and behavior. Likewise, Blood and Snyder (1993)
found that reactance2 showed positive association with fear, but did not interfere with
behavioral intentions to result in boomerang effects. Rippetoe and Rogers (1987)
suggested that fatalism may not be related to intended effects and other unintended
effects because by accepting their fate, persons are not likely to engage in other modes of
coping, be it danger control or fear control. On the other hand, they suggested that
hopelessness may lead to counterproductive behavior as a reaction to their belief that they
cannot perform the recommended behavior successfully. However, if fatalism is an
outcome of low response efficacy in the face of high threat as discussed previously, it
may also trigger counterproductive behavior. Intention to increase risk behavior was
found in studies in which only response efficacy was manipulated (e.g., Kleinot &

Rogers, 1981; Rogers & Mewbom, 1976). Furthermore, fatalism was positively

 

lHovland, Janis, and Kelly (1953) pointed out that this may be a result of lack of cues raising subjects'
suspicion of the source’s intention. For example, Janis and Feshbach’s (1953) experimental stimuli were
presented as an official part of the high school’s hygiene program. In Witte’s (1992a) experiment, fear
appeal messages were presented as as part of AIDS education materials. Therefore, reactance (i.e.,
perceived manipulation and message derogation) to fear appeals used in health communication campaigns
which have manifest intent to do social good is not likely to result in message rejection.

2 While reactance is defined as message derogation and perceived manipulation in the EPPM (Witte,
1992a), Blood and Snyder’s (1993) measures of reactance are close to Brehm’s (1966), and include: “It’s
my life. I should be able to drink and drive if I want to;” Nobody should be telling me when or where 1 can
drink;” “I resent being told I can’t drink and drive.”

43

associated with behaviors that increase the risk of contracting AIDS (Kalichman, Kelly,
Morgan, & Rompa, 1997; McKirnan, Ostrow, & Hope, 1996). Therefore, fatalism, along
with hopelessness, may result in counterproductive behavior. On the basis of previous
discussion, the following hypotheses and research questions are proposed.
Hypotheses and Research Questions
H1: In general, the magnitude of unintended effects of fear appeals is likely to be
greater in earlier stages than later stages.
Hla: The magnitude of unintended effects is greater in the precontemplation stage
than in the contemplation stage.
Hlb: The magnitude of unintended effects is greater in the contemplation stage
than in the preparation stage.
ch: The magnitude of unintended effects is greater in the preparation stage
than in the action stage.
Hld: The magnitude of unintended effects is greater in the action stage
than in the maintenance stage.
H2: When perceived efficacy is low, perceived high threat is likely to bring about
greater degree of unintended effects than when perceived efficacy is high.
H3a: When perceived response efficacy is low, perceived high threat is likely to bring
about greater degree of fatalism than when perceived response efficacy is high.
H3b: When perceived self efficacy is low, perceived high threat is likely to bring about
greater degree of hopelessness than when perceived self efficacy is high.

H4: When perceived threat is low, attitude toward risk behavior is likely to be more

44

favorable when response efficacy is low than when response efficacy is high.
RQl: How do threat, efficacy, and stages of change interact in bringing about unintended
effects?

RQ2: How are unintended effects associated with intended effects?

45

CHAPTER 3

METHOD

Design

mm

A lab experiment involving messages on the topic of "skin cancer" was designed
to test the proposed hypotheses and to answer the research questions.

Threat (susceptibility, severity) was manipulated in two (high, low threat
message) levels. Efficacy was not manipulated but measured to examine its possible
main as well as interaction effect with threat and stages of change. Subjects’ stages of
change were measured immediately before the experiment.

The dependent variables included defensive avoidance, message derogation,
perceived manipulation, fatalism, hopelessness, reactance, wishful thinking, attitude
toward risk and recommended behavior, intentions to engage in preventive behavior, and
risk and preventive behavior. Behavior was measured four weeks after the experiment in
a follow-up questionnaire.

Subjects

Subjects of this study (n=274) were Michigan Stage University undergraduate
students recruited from introductory advertising, communication, and telecommunication
courses for extra credit and a drawing for gift certificates. To include subjects in all
stages of change, no prescreening of subjects in terms of skin cancer knowledge was

done.

46

Overall, 60.6 percent of the respondents were female, and 39.4 percent of them
were male. Respondents’ age ranged from 18 to 37 (SD =2.1), and the typical age was 20
(mean, median, mode = 20). Caucasians comprised 83.6 percent of the respondents.
Asians, African-Americans, and Hispanics represented 6.9, 6.2, and 1.5 percent of the
respondents, respectively.

In the four-week follow-up, 239 subjects participated, producing an attrition rate
of 12.7%.
Procedure

Subjects participated in the experiment in groups of varying size ranging upto
thirty. After a one-page questionnaire assessing their stages of change was administered,
subjects were randomly assigned to one of the two experimental conditions (high, low
threat).

As has been the case in many previous fear appeal studies (e.g., Janis & Feshbach,
1953; Witte, 1992b), the experiment was presented as a part of public health campaign
project. Subjects were told that the messages were in the early stages of development to
provide college students with facts about skin cancer and that their responses were
important for further improvement of the messages. To ensure full attention, subjects
were asked to read the messages carefully and to underline important words, phrases, or
sentences in each of the paragraphs.

Following the instructions, high or low threat messages including five photos
were distributed to the subjects. Immediately after reviewing the messages, subjects

completed posttest questionnaires. Four weeks later, subjects were visited in their

47

classrooms for the completion of a follow-up questionnaire. After this, subjects were
debriefed about the purpose and design of the study to clear up any possible
misunderstanding and were given skin cancer prevention brochures from the American
Cancer Society.

Stimulus Materials

Messages were developed with Witte’s (1993) recommendations and
information and verbatim quotes obtained from existing skin cancer prevention messages
from such organizations as the National Cancer Institute, the American Cancer Society,
and the American Academy of Dermatology. Each message manipulation consisted of
factual messages, a case of a fictitious skin cancer patient, and five photos (see Appendix
A2 and A3).

The high threat message described skin cancer in detail, presenting vivid
descriptions of its symptoms and consequences. College-aged persons’ susceptibility to
skin cancer was emphasized. For example, the case for the high threat message read:
“Jennifer, a 21-year old student at Michigan State University, discovered a freckle near
her eye... A year later, Jennifer became one of 9,500 patients to die from skin cancer.”
The high threat message also included graphic color photos of skin cancer patients’ body
parts.

The low threat message described skin cancer as a less severe disease with fewer
significant consequences. College—age persons’ susceptibility to skin cancer was
minimized by noting the fact that skin cancer most frequently affects persons of age over

50. For example, the case for the low threat message read: “A 52-year old farmer in New

48

Zealand went to see a doctor because he has developed an unusual mole on his face...
Because skin cancer rarely spreads, a biopsy often is the only test needed to determine the
stage of skin cancer.” The low threat message included color photos conveying neutral
images, such as clinical lab results (e. g., Witte, 1992a). Both high and low threat
messages were based on facts, and they were equated for length, order, complexity, and
quality. Recommended behavior included sunscreen use, wearing protective hat and
clothing, minimizing sun exposure at midday, and performing periodic skin self-
examination. The messages were subjected to pilot tests.
Measures

Seven-point Likert scales, ranging from “strongly agree” to “strongly disagree,”
were used, unless otherwise noted (see Appendix B for a full description of the
measures). To generate an index for each of the constructs, items in the same construct
were averaged. Cronbach’s alphas were used to evaluate the reliability of measures,
which are presented in Table 4.
Stages of Change

Following previous TM studies (DiClemente et al., 1991; Grimley et al., 1993;
Prochaska, 1994), a mutually exclusive stage classification algorithm was used so that all
subjects could be classified into only one stage. Each subject was asked to select one
item that best represents their current sunscreen-use behavior. Table 3 presents a

summary.

49

Table 3 Distribution of Stages of Change

 

 

Stage Classification %

Precontemplation Subjects did not use sunscreen consistently when 52.2
going out in the sun, and do not intend to start using it
consistently within the next 6 months.

Contemplation Subjects did not use sunscreen consistently when 19.7
going out in the sun, but intend to start using it
consistently within the next 6 months.

Preparation Subjects did not use sunscreen consistently when 4
going out in the sun, but intend to start using it
consistently within the next 30 days.

Action Subjects used sunscreen consistently when going out 9.9
in the sun, but for less than 6 months.

Maintenance Subjects used sunscreen consistently when going out 21.2

in the sun for more than 6 months.

 

Intention to use sunscreen in the next 6 months was used to differentiate
contemplators from precontemplators. Intention to use sunscreen in the next 30 days was

used to distinguish prepared subjects from contemplators. Subjects who have attempted

sunscreen use from one day to six months were categorized as actors. Those who have

used sunscreen for more than six months were classified as maintainers. Overall, 52.2

percent of subjects were in precontemplation stage and 19.7 percent in contemplation

stage. Four percent of subjects were in preparation stage, 9.9 percent in action stage, and

21.2 percent in maintenance stage.

50

To minimize variance differences resulting from unequal cell sizes, stages of
change were reduced from five to three. Contemplation stage (i.e., the one in which
individuals intend to engage in preventive behavior in six months), was merged with the
preparation stage (i.e., the one in which individuals intend to engage in preventive
behavior in a month), because both stages have to do with behavioral planning, although
in different time spans. Action stage, (i.e., the one in which individuals have engaged in
preventive behavior for less than six months), was merged with the maintenance stage
(i.e., the one in which individuals have engaged in preventive behavior for more than six
months), because both stages have to do with actual practices of behavior, although in
different time periods.

Egg

Fear arousal (or = .95; mean = 3.02; SD = 1.72) was appraised by asking subjects
to rate six mood adjectives on a scale ranging from “not at all” to “very much.” The six
moods were: “frightened,” “tense,” “nervous,” “anxious,” “uncomfortable,” and
“nauseous,” as has done in many fear appeal studies (e. g., Leventhal, Singer, & Jones,
1965; Maddux & Rogers, 1983; Rippetoe & Rogers, 1987; Witte, 19923).

M

Threat (or = .89; mean = 4.57, SD = 1.39) was operationalized as susceptibility to
threat and severity of threat. Witte and colleagues’ (1996) scale was used to measure this
variable and served as a manipulation check. The susceptibility scale (or = .85; mean =
4.30; SD = 1.46) included three items: “I am at risk for skin cancer,” “It is likely that I

will develop skin cancer,” and “It is possible that I will develop skin cancer.”

51

The severity scale (or = .92; mean = 4.83; SD = 1.66) included three items: “I
believe that skin cancer is a severe health problem,” “I believe that skin cancer is serious
threat,” and “I believe that skin cancer is a significant disease.” Because susceptibility
and severity are dimensions of the threat construct (Witte et al., 1996), the items were
averaged to create an overall index.

m

Measurement of efficacy (or = .90; mean = 5.18; SD = .89), operationalized as
response efficacy and self efficacy, was based on Witte et al.’s scale (1996). Self-
efficacy (or = .90; mean = 4.56; SD = 1.12) was assessed with fourteen items including, “1
can easily perform periodic skin self-examination to prevent skin cancer” and “Wearing
clothing that covers my body to prevent skin cancer is no problem for me.”

Response efficacy (or = .92; mean = 5.81; SD = 1.06) was measured with seven
items including, “Minimizing my exposures to the sun at midday is effective in
preventing skin cancer.” Because self efficacy and response efficacy are dimensions of
the efficacy construct (Witte et al., 1996), the items were averaged to create an overall
index.

Demndent Measures

The dependent variables included defensive avoidance, message derogation,
perceived manipulation, fatalism, hopelessness, reactance, wishful thinking, attitude
toward risk and recommended behavior, intentions to engage in preventive behavior, and

preventive behavior.

52

In measuring defensive avoidance, message derogation, and perceived
manipulation, Witte’s (1992a) scales were used. Measures of fatalism, hopelessness, and
wishful thinking were based on Rippetoe and Rogers’ (1987) scales. Reactance measures
were based in Blood & Snyder’s (1993) items. Witte's (1992a) scale was used to measure
attitudes toward risk and recommended behavior. Reliability of all measures were
adequate except for reactance and wishful thinking variables.

Defensive avoidance. Subjects indicated their perceptions with regard to two
statements, “When I sunbathe. I tend to avoid thoughts of skin cancer,” “When I get a
- sunburn, I tend to avoid thinking about skin cancer,” and “When I suntan, I try to not
think about skin cancer” (or = .87; mean = 4.26; SD = 1.69).

Message derogation. To measure message derogation, subjects were asked if they
perceived that the message was “exaggerated,” “overblown,” and “overstated” (or = .86;
mean = 3.29; SD = 1.32).

Perceived manipulation. Perceived manipulation was assessed by asking subjects

’9 6‘

if they thought that the message was “manipulative, misleading,” “distorted,” and
“exploitative” (or = .86; mean = 2.92; SD = 1.13).

F atalism. To measure fatalism, subjects were asked to indicate the degree to
which they agree or disagree with the following two statements: “If you are destined to
die from skin cancer, you will; there is little you can do about it” and “Only time will tell

if I develop skin cancer; nothing can be done anyway but wait” (or = .69; mean = 2.28;

SD = 1.33).

53

Hopelessness. Items for hopelessness scale included the following three items:
“In this day and age, it sometimes seems a hopeless task to say healthy,” “Given what I
know about skin cancer, I sometimes feel it’s almost useless to try to stay health,” and
“When I think about the prospect of cancer, I sometimes feel like saying, "What’s the use
in staying health?"” (or = .81; mean = 2.55; SD = 1.33).

Reactance. This construct of psychological reactance was assessed with three
items including, “Nobody should be telling me when or how I can get sun,” “It’s my life.
I should be able to get sun if I want to,” and “I resent being told that I can’t get sun.”
Reliability of this scale was low (or = .50; mean = 3.56; SD = 1.09).

Wishful thinking. Items for wishful thinking measure included, “When faced
with the prospect of developing skin cancer, it helps me to dream of a world where there
are no diseases like cancer,” “I believe a miracle cure for cancer in the near future is the
answer to my fears about skin cancer,” and “Given what I know about skin cancer, I wish
I belonged to another time and another place.” Reliability of this scale was very low (or =
.37; mean = 2.79; SD = 1.16), suggesting that the scale may comprise more than one
dimensions.

m. Attitude toward risk (e.g., “In general, sunbathing is”) and
recommended behavior (e. g., “Using sunscreen when out in the sun to prevent skin
cancer is”) was measured with a semantic-differential scale including two sets of word
pairs: “positive -- negative” and “desirable -- undesirable.” Cronbach’s alphas for
attitude toward risk and recommended behavior were .78 (mean = 3.63; SD = 1.53) and

.89 (mean = 4.52; SD = 1.26), respectively.

54

Intentions. Intentions to engage in recommended behavior (or = .96; mean = 3.66;
SD = 1.37) were measured with items including, “I intend to use sunscreen consistently
when exposed to sun within the next 6 months” and “I intend to perform periodic skin
self-examination to prevent skin cancer.”

Behavior. Items measuring preventive behavior included, “Did you do anything
to prevent skin cancer since you first participated in this study,” “I have tried to minimize
my exposure to the sun at midday since I first participated in this study, “ I currently use
sunscreen when exposed to the sun to prevent skin cancer” (or = .90; mean = 2.85; SD =
1.35).

Demographic variables. Age, gender, race, skin type, risk behavior (e.g.,
frequency of outdoor activities, number of times sunburned and peeled during the last two
years, number of times visiting tanning parlors) and preventive behavior (e.g., use of
sunscreen and protective clothing, minimization of sun exposure at midday), and social
influence (e.g., friends’ and family members’ use of sunscreen, knowledge of other
individuals who have cancer) were asked. In analysis, these variables were used as

control variables.

55

Table 4 Reliability of the Measures

 

 

Index Cronbach’s Mean Standard
or Deviation
Fear .95 3.02 1.72
Threat .89 4.57 1.39
Susceptibility .85 4.30 l .46
Severity .92 4.83 1.66
Efficacy .90 5.18 .89
Response efficacy .92 5.81 1.06
Self efficacy .90 4.56 1.12
Defensive avoidance .87 4.26 1.69
Message derogation .86 3.29 1.32
Perceived manipulation .86 2.92 1.13
Fatalism .69 2.28 1.33
Hopelessness .81 2.55 1 .33
Reactance .50 3.56 1.09
Wishful thinking .37 2.79 1.16
Attitude
toward risk behavior .78 3.63 1.53
toward recommended .89 4.52 1.26
behavior
Intentions .96 3 .66 1 .37
Behavior .90 2.85 1 .35

 

Note. Seven-point Likert scales, ranged from "strongly agree to strongly disagree,"
were used.

56

Analysis

To test H1, planned comparison of means was done initially. Specifically,
contrast coefficients (-1 1 O, 0 -1 1) were used to compare adjusted means of dependent
variables in each of the three stages after controlling for confounds. Covariates were not
controlled in the analysis of H1 because stages of change may subsume such covariates.

In addition, to detect any other possible significant differences among means, a
Tukey-B test was done. The above described contrast tests limit number of comparisons
to k-l. A one-way analysis of variance was done first. When the overall E ratio was
significant, a Tukey-B test was done to determine the significance of mean differences.
Tukey-B test was chosen over Scheffe’ test which is too conservative (Kerlinger, 1986).
Like Scheffe’, the Tukey-B test makes all comparisons and is adequate for unequal n
sizes (Kirk, 1982).

To test H2, H3a, H3b, H4, and RQ2, the analysis of covariance (ANCOVA)
method was used to control for covariates and confounds exerting significant influences.
A regression approach was employed to adjust for unequal number of subjects on some of
the variables (Tabachnick & Fidell, 1996).

In testing H2, a median (4.67 on a 7-point scale) split was done for the efficacy
measure to perform 2 (high, low threat) x 2 (high, low efficacy) factorial analysis. In
testing H3a, a median (6.00 on a 7-point scale) split was done for the response efficacy
measure to perform 2 (high, low threat) x 2 (high, low response efficacy) factorial

analysis. In testing H3b, a median (4.57 on a 7-point scale) split was done for the

57

efficacy measure to perform 2 (high, low threat) x 2 (high, low self efficacy) factorial
analysis.

To answer RQl, a 2 (threat) x 2 (efficacy) x 3 (stage) ANCOVA, controlling
covariates and confounds, was done to detect significant main effects of and interaction
effects among the three variables. Partial correlations, controlling for confounds, were

run to answer RQ2.

58

CHAPTER 4

RESULTS

Message Validation

Multiple pilot tests were done to equate the high and low threat messages in terms
of accuracy, objectivity, ease of reading, understandibility, and amount of learning and to
appraise any differences in these dimensions. The results of final pilot study indicated no
significant differences in these confound check measures (p > .05).

The pilot tests also validated threat message manipulations. Subjects were
randomly given either high or low threat messages. Their perceived susceptibility,
severity, and fear were measured on a 7-point scale (see the next section for the
description of measures). Independent sample t-test were used to determine the
difference between means. Results indicated significant (p < .05) differences between
high and low threat messages in terms of susceptibility, severity, and fear.

Manipulation Checks

Susceptibilig. Subjects who were exposed to the high-threat message reported a
greater level of susceptibility of skin cancer (M = 5.00) than those who were exposed to
the low threat message (M = 3.59, t = -9.16, pf: 272, p < .001).

icy/pm. Subjects who were exposed to the high-threat message perceived that
skin cancer is more severe disease (M = 5.86) than those who were exposed to the low

threat message (M = 3.78, t = ~13.41, g: 272, p < .001).

59

Egg. Subjects who were exposed to the high threat message perceived a greater
level of fear (M = 4.23) than those who were exposed to the low threat message (M =
1.74, _t_ = -17.22, M: 263, p < .001).

Therefore, the results of susceptibility, severity, and fear manipulations presented
above indicate that the high and low threat messages produced significantly different
perceptions of threat.

Confound checka. Subjects who were exposed to the high and low threat
messages did not show any difference in their self efficacy or response efficacy. No
differences were found between high and low threat condition subjects in terms of their
perceived objectivity of the messages. However, subjects in the high threat condition
believed that the message was more accurate (M = 5.54) and clear (M = 5.74) and of
better quality (M = 5.52) than those who were in the low threat condition (M = 4.90, t = -
3.65, M: 272, p < .001; M = 5.11,t= -3.82, g": 271, p < .001; M = 4.50, t = —5.65, at:
271 , p < .001 , respectively). Subjects in the high threat condition also believed that they
understood (M = 6.07) and learned (M = 5.51) from the message more than those who
were in the low threat condition (M = 5.58, t = -3.34, M: 272, p = .001; M = 4.63, 1:
44.62, _c_1_f= 272, p < .001, respectively).

These differences occurred in spite of the fact that the high and low threat
messages were validated through pilot tests. As mentioned earlier, in the final pilot test,
such differences were not detected (p > .05 for all items) perhaps because of the
idiosyncrasy of the sample or small sample size (n= 14); pilot test subjects may have

primarily been precontemplators, who often lack knowledge of risk.

60

One possible reason for these differences may be that the prescreening of subjects
was not done for this study, because the purpose of this study is to determine individuals
in different stages of changes’ reactions to fear appeals. Furthermore, skin cancer is a
relatively well-known disease (Clarke, 1997; Hanley, 1996). Therefore, some subjects
may have perceived that the low threat message stating that college students are at low
risk and that treatment is easy and painless may not entirely accurate.

The results of thought-listing also shed some light. Subjects who were exposed to
the low threat message complained about the lack of pictures showing actual skin cancer
symptoms. Some subjects in the low threat condition questioned why they have to
engage in preventive behavior if college students are at low risk of skin cancer. These
factors may have affected low threat condition subjects’ perceptions of clarity and quality
and the degree to which they learned and understood the message.

To control for their influence, the confounds were controlled in analyses, as
reported in the following sections. Results of zero-order correlation among variables are

presented in Table 5.

61

Table 5 Zero-Order Correlations among Variables

 

Fear Threat Efficacy Defensive Message Perceived
avoidance derogation manipulation

 

Fear

Threat .67*

Efficacy .17* .25*

Defensive -.O4 -.00 -.06

avoidance

Message -.16* -.18* 25* .11

derogation

Perceived -.10 -.03 -.21* .05 .60*
manipulation

Fatalism .13* .10 -.15* .10 .11 .17*
Hopelessness .21* .19* -.12* .18* .09 .30*
Reactance -.18* -.17* -.11 .22* .27* .22*
Wishful .36* .28* .08 .07 -.08 .10
thinking

Attitude: .10 .08 .15* -.30* -.14* -.09
risk behavior

Attitude: .04 .09 .33* -. 15* -.11 -.04
recommended

behavior

Intentions .43* .40* .40* -.19* -.19* -.04
Behavior .43 .36* .22* -.17* -. 19* -.04

 

Note. *p < .05. In the scale (1-7) for attitude toward risk behavior, 1 denoted good,
desirable, and favorable.

62

Table 5 (cont'd)

 

Hopeless Reactance Wishful Attitude Attitude Intentions
Fatalism -ness thinking toward risk toward
behavior recommended
behavior

 

Fear

Threat

Efficacy

Defensive

avoidance

Message

derogation

Perceived

manipulation

Fatalism

Hopelessness .54*

Reactance .33* 22*

Wishful .39* .42* .04

thinking

Attitude: -.05 -.06 -.21* -.05

risk behavior

Attitude: -.11 -.00 -.11 -.01 .51 *
recommended

behavior

Intentions -.11 .00 -.26* .20* . 10 .27*
Behavior -.03 .07 -.22* .18 .16* .20* .70*

 

63

Hypothesis 1

H1 predicted that in general, the magnitude of unintended effects of fear appeals
are likely to be greater in earlier stages than later stages. As noted earlier, stages were
reduced from five to three. Subsequently, analysis tested the following reformulated
hypotheses. Hla: The magnitude of unintended effects is greater in the precontemplation
stage than in the contemplation/preparation stage; Hlb: The magnitude of unintended
effects is greater in the contemplation/preparation stage than in the action/maintenance
stage. The results, summarized in Table 6a and 6b, partially support H1.

Results of planned contrast tests are reported first. A significant main effect of
stage was found for defensive avoidance (E (2, 267) = 5.16, p < .01), perceived
manipulation (E (2, 268) = 3.10, p < .05), reactance (E (2, 270) = 6.21, p < .005), attitude
(E (2, 270) = 7.69, p = .001), intentions (E (2. 265) = 82.34, p < .001), and behavior (E (2,

225) = 28.09, p < .001).

64

Table 6a Mean Differences of Unintended Effects across Stages:
Results of Planned Comparison

 

 

Mean Square E
Defensive avoidance 14.14 5.16 .01
Message derogation 2.41 1.68 .19
Perceived manipulation 3.02 3.10 .05
Fatalism 3.79 2.18 .12
Hopelessness 3.97 2.26 .1 l
Reactance 6.92 6.21 .00
Wishful thinking 1.37 1.04 .35
Attitude 1 1.61 7.69 .00
Intentions 95.56 82.34 .00
Behavior 41 .71 28.09 00

 

65

Significant difference between those in the contemplation/preparation and
action/maintenance stage was found for defensive avoidance (t = -1.65, p < .1), perceived
manipulation (t = -2.39. p < .05), intentions (t = 3.92, p < .001), and behavior (1 = 3.18, p
< .01). Significant difference between those in the precontemplation and
contemplation/preparation stage was found for reactance (t = -2.66, p < .01), attitude, (1 =
3.04, p < .005), intentions (t = 7.39, p < .001), and behavior (1 = 3.78, p < .001).

Next, results of Tukey-B tests are reported. Table 6b provides a summary.

66

Table 6b Mean Differences of Unintended Effects across Stages: Results of
Tukey-B Tests

 

 

 

Stage
Precontemplation Contemplation/ Action/

Preparation Maintenance
Defensive Avoidance 4.52"(1.57) 4.22 (1.63) 3.763(183)
Message Derogation 3.36 (1.28) 3.30 (1.35) 3.13 (1.36)
Perceived Manipulation 2.92 (1.10) 3.05 (1.04) 2.75 (1.25)
Fatalism 2.47a(1.41) 2.003(1.14) 2.11 (1.28)
Hopelessness 2.71a(1.38) 2.52 (1.23) 2.22a(1.31)
Reactance 3.80“"(1.10) 3.308(101) 3.34"(1.04)
Wishful Thinking 2.73 (1.19) 2.96 (1.09) 2.74 (1.16)
Attitude 4.25“”(127) 4.763(113) 4.82"(1.29)
Intentions 2.88a(1.15) 4.113"(1.06) 4.87” (.93)
Behavior 2.333(123) 3.03ab(1.27) 3.81"(1.13)

 

Note. 3’ ° indicate significant differences (p_< .05) in means of each pair.

In parentheses are standard deviations.

67

Defensive avoidance. A significant mean difference (E (2, 269) = 4.72, p < .01) in
defensive avoidance was found between precontemplation stage (M = 4.52) and
action/maintenance (M = 3.76, p < .05) stage. Subjects who did not intend to use
sunscreen consistently within the next six months indicated greater defensive avoidance
than those who had been using it consistently for more than one month.

Fata_l_is_m_. A significant difference (E (2, 271) = 3.53, p < .05) was found in
fatalism between precontemplation stage (M = 2.47) and contemplation/preparation stage
(M = 2.00, p < .05). Subjects who did not intend to use sunscreen consistently within the
next six months indicated greater likelihood of fatalism than those who intended to start
using it consistently within the next one to six months.

Hopelessness. A significant mean difference (_E (2, 271) = 3.07, p < .05) in
hopelessness was found between precontemplation stage (M = 2.71) and
action/maintenance stage (M = 2.22, p < .05). Subjects who did not intend to use
sunscreen consistently within the next six months showed greater likelihood of feeling
hopeless than those who had been using it consistently for more than one month.

Reactance. Significant mean differences (I: (2, 272) = 6.75, p < .005) in reactance
were found between precontemplation stage (M = 3.80) and contemplation/preparation
stage (M = 3.30, p < .05) and between precontemplation (M = 3.80) and
action/maintenance (M = 3.34, p < . 05) stages. Subjects who did not intend to use
sunscreen consistently within the next six months indicated a stronger likelihood of
perceiving that the message was a threat to their freedom than those who intended to use

sunscreen within one to six months or those who had used it more than one month.

68

map. Significant mean differences (E (2, 272) = 6.68, p < .005) in attitude
were found between precontemplation stage (M = 4.25) and contemplation/preparation
stage (M = 4.76, p < .05) and between precontemplation (M = 4.25) and
action/maintenance (M = 4.82, p < . 05) stages. Subjects who did not intend to use
sunscreen consistently within the next six months indicated less favorable attitude toward
recommended behavior than those who intended to use sunscreen within one to six
months or those who had used it more than one month.

Intentions. As expected, significant mean differences (E (2, 267) = 82.75, p <
.0001) in intentions were found between precontemplation (M = 2.88) and
contemplation/preparation (M = 4.11, p < .05) stages, between contemplation/preparation
(M = 4.11) and action/maintenance (M = 4.87, p < .05) stages, and between
precontemplation (M = 2.88) and action/maintenance (M = 4.87, p < . 05) stages.

Behavior. In the four-week follow—up, significant mean differences (E (2, 227) =
27.82, p < .0001) in behavior were found between precontemplation (M = 2.33) and
contemplation/preparation (M = 3.03, p < .05) stages, between contemplation/preparation
(M = 3.03) and action/maintenance (M = 3.81, p < .05) stages, and between
precontemplation (M = 2.33) and action/maintenance (M = 3.81, p < . 05) stages.
Subjects who did not intend to use sunscreen consistently within the next six months
reported fewer behavioral changes for skin cancer prevention than those who intended to

use sunscreen within one to six months or those who had used it more than one month.

69

No significant differences were discovered among individuals in different stages
of change with regard to message derogation (E (2, 271) = .66, p > .05), perceived
manipulation (E (2, 270) = 1.21, p > .05), and wishful thinking (E (2, 264) = .92, p > .05).
Hypothesis 2

H2 stated that “When perceived efficacy is low, perceived high threat is likely to
bring about greater degree of unintended effects than when perceived efficacy is high.”
The results partially support H2, and are shown in Table 7.

Defensive avoidance. A significant main effect of efficacy was found for
defensive avoidance: E (1, 269) = 4.94, p < .05. A significant covariate was race (E (1,
269) = 9.34, p < .005). No other significant main or interaction effects were found.
Subjects who perceived low efficacy indicated greater degree of defensive avoidance
(adjusted M = 4.50) than those who perceived high efficacy (adjusted M = 4.04).

Message derogtion. A significant interaction between threat and efficacy was
found, E = (1, 272) = 5.92, p < .05, for message derogation. Significant covariates were
perceived accuracy of message (E (1, 272) = 40.08, p < .001) and minimization of
exposure to the sun at midday (E (1, 272) = 5.94, p < .05). When threat was high,
subjects who perceived low efficacy indicated greater degree of message derogation
(adjusted M = 3.53) than those who perceived high efficacy (adjusted M = 2.86).

Perceived manipulation. A significant main effect of efficacy was found, (E (1,
266) = 4.36, p < .05, for perceived manipulation. Significant covariates for this variable
were perceived accuracy of message (E (1 , 266) = 60.09, p < .001), friends’ use of

sunscreen (E (1, 266) = 7.89, p = .005), and number of times peeled for the past two years

70

(E (1, 266) = 4.60, p < .05). No other significant main or interaction effects were found.
Subjects with low perceived efficacy indicated greater degree of perceived manipulation
(adjusted M = 3.06) than those with high efficacy (adjusted M = 2.80).

Fa_tali_sr_n_. A significant main effect of threat (E ( 1, 272) = 6.00, p < .05) was
found for fatalism. Significant covariates were subjects’ current tan level (E (1 , 272) =
5.93, p < .05) and family members’ (E (1, 272) = 7.43, p < .01) sunscreen use. No other
significant main or interaction effects were found. Subjects who were in the high threat
condition expressed greater degree of fatalism (adjusted M = 2.47) than those who were
in the low threat condition (adjusted M = 2.09).

Hopelessness. A significant main effect of threat (E (1, 272) = 8.92, p < .005) was
found for hopelessness. Significant covariates were frequency of workout in the sun (E
(1, 272) = 5.80, p < .05) and friends’ (E (1, 272) = 9.31, p < .005) and family members’
(E (1, 272) = 8.44, p < .005) sunscreen use. No other significant main or interaction
effects were found. Subjects who were in the high threat condition expressed greater
degree of hopelessness (adjusted M = 2.78) than those who were in the low threat
condition (adjusted M = 2.31).

Reactance. A significant main effect of efficacy was found (E (1, 269) = 4.15, p <
.05) for reactance. Significant covariates included frequency of leisure activities in the
sun (E (1, 269) = 10.51, p < .01) and perceived amount of learning from message (E (1,
269) = 6.01 , p < .05). No other significant main or interaction effects were found.
Subjects who perceived low efficacy indicated greater reactance (adjusted M = 3.67) than

those who perceived high efficacy (adjusted M = 3.41).

71

Wishful think_ipg. A significant main effect of threat was found (E (l, 265) =
18.99, p < .001) for wishful thinking. No other significant main or interaction effects
were detected. A significant covariate was perceived accuracy of the message (E (1, 265)
= 3.93, p = .05). Subjects who were in the high threat condition showed greater tendency
of wishful thinking (adjusted M = 3.06) than those who were in the low threat condition
(adjusted M = 2.51).

_Att_itu_d§. A significant main effect of efficacy (E (1, 272) = 11.25, p = .001) was
found for attitude toward recommended behavior. Significant covariates included use of
protective clothing (E (1, 272) = 8.59, p < .005) and minimization of sun exposure at
midday (E (l , 272) = 6.95, p < .01 ). No other significant main or interaction effects were
found. Subjects who perceived low efficacy indicated less favorable attitude (adjusted M
= 4.28) toward recommended behavior than those who perceived high efficacy (adjusted
M = 4.75).

Intentions. A significant interaction between threat and efficacy was found for
intentions: E (1 , 265) = 4.69, p < .05. Significant covariates were minimization of
exposure to the sun at midday (E (1, 265) = 53.98, p < .001) and preexisting sunscreen
use including frequency of use (E (1, 265) = 12.81, p < .001) and reapplication every two
hours (E (1, 272) = l 1.42, p < .001). Subjects who were exposed to the high threat
message indicated fewer intentions to engage in preventive behavior when perceived
efficacy was low (adjusted M = 3.48) than when perceived efficacy was high (adjusted M

= 4.25).

72

Behavior. In the four-week follow-up, a significant main effect of threat was
found (E (1 , 227) = 19.45, p < .001) for behavior. No other significant main or
interaction effects were found. Significant covariates were preexisting protective
behavior such as use of protective clothing when out in the sun (E (1, 227) = 13.76, p <
.001) and minimization of sun exposure at midday (E ( 1, 227) = 6.01, p < .05). Subjects
who had been in the high threat condition reported greater behavioral change for skin
cancer prevention (adjusted M = 3.18) than those who were in the low threat condition

(adjusted M = 2.50).

73

Table 7 Summary of Analysis of Covariance of Threat x Efficacy and Adiusted Mea_p§

a. Defensive Avoidance

 

 

 

 

 

Source of Variation d_f Mean Square F p
Covariate l 25.68 9.34 .00
Threat 1 .39 .14 .71
Efficacy 1 13 .59 4.94 .03
Threat x efficacy 1 5.38 1.96 .16
Error 265 2.75
M. Significant covariate was race.
Threat

Low High

My Low 4.32 4.68

High 4.14 3.93

 

74

b. Message Derogation

 

 

Source of Variation at: Mean Square F p
Covariate 1 1 55.62 40.08 .00
Covariate 2 1 8.24 5.94 .02

Threat 1 2.86 2.06 .15

Efficacy 1 6.42 4.63 .03

Threat x efficacy 1 8.21 5.92 .02
Error 267 1.39

 

Note. Significant covariates were perceived accuracy of the message and minimization
of sun exposure at midday, respectively.

 

 

Threat
Low High
E_ffi_capy Low 3.39 3.53
High 3.42 2.86

 

75

c. Perceived Manipulation

 

 

Source of Variation gl_f Mean Square E p
Covariate 1 1 58.41 60.69 .00
Covariate 2 1 7.59 7.89 .01
Covariate 3 l 4.42 4.59 .03

Threat 1 .67 .70 .40

Efficacy 1 4.20 4.36 .04

Threat x efficacy 1 1.50 1.56 .21
Error 260 .96

 

Note. Significant covariates were perceived accuracy of the message, friends' use of
sunscreen, and number of times peeled during the past two years, respectively.

 

 

Threat
Low High
Effigy Low 3.03 3.08
High 2.93 2.67

 

76

d. Fatalism

 

 

Source of Variation at: Mean Square E p
Covariate 1 1 12.45 7.43 .01
Covariate 2 1 9.92 5.93 .02

Threat 1 10.05 6.00 .02

Efficacy 1 3.00 1.78 .18

Threat x efficacy 1 4.58 2.73 .10
Error 267 1.68

 

Note. Significant covariates were level of family members' use of sunscreen and level of
tan, respectively.

 

 

Threat
Low High
Efficacy Low 2.07 2.72
High 2. 12 2.24

 

77

e. Hopelessness

 

 

Source of Variation fl Mean Square E p
Covariate l 1 15.15 9.31 .00
Covariate 2 l 13.72 8.44 .00
Covariate 3 1 9.43 5.80 .02

Threat 1 14.52 8.92 .00

Efficacy l 4.20 2.58 .11

Threat x efficacy 1 .16 .10 .76
Error 266 1.63

 

Note. Significant covariates were friends use of sunscreen, family members' use of
sunscreen, and frequency of workout in the sun, respectively.

 

 

Threat
Low High
My Low 2.47 2.89
High 2.16 2.68

 

78

f. Reactance

 

 

Source of Variation gl_f Mean Square E p
Covariate 1 1 11.60 10.51 .00
Covariate 2 l 6.05 6.02 .02

Threat 1 1.80 1.63 .20

Efficacy 1 4.58 4.15 .04

Threat x efficacy 1 .05 .05 .83
Error 264 1.10

 

Note. Significant covariates were amount of time spent for leisure activities in the sun
and perceived amount of learning from the message, respectively.

 

 

Threat
Low High
Efficacy Low 3 .77 3.57
High 3 .48 3.34

 

79

g. Wishful Thinking

 

 

Source of Variation fl Mean Square E p
Covariate 1 4.89 3.93 .05
Threat 1 18.72 15.02 .00
Efficacy 1 .001 .001 .98
Threat x efficacy 1 .10 .08 .78
Error 261 1.25

 

Note. Significant covariate was perceived accuracy of message.

 

 

Threat
Low High
Em Low 2.50 3.08
High 2.53 3.07

 

80

h. Attitude

 

 

Source of Variation d_f Mean Square F p
Covariate 1 1 11.12 8.59 .00
Covariate 2 1 9.00 6.95 .01

Threat 1 1.22 .94 .33

Efficacy 1 14.56 11.25 .00

Threat x efficacy 1 1.78 1.37 .24
Error 267 1.29

 

Note. Significant covariates were use of protective clothing when out in the sun and
minimization of sun exposure at midday, respectively.

 

 

Threat
Low High
My Low 4.43 4.13
High 4.74 4.77

 

81

i. Intentions

 

 

Source of Variation d_f Mean Square E p
Covariate 1 1 25.49 26.77 .00
Covariate 2 l 12.19 12.81 .00
Covariate 3 1 10.87 11.42 .00

Threat 1 11.18 11.74 .00

Efficacy 1 14.75 15.49 .00

Threat x efficacy 1 4.47 4.69 .03
Error 259 .95

 

Note. Significant covariates were frequency of sunscreen use, minimization of sun
exposure at midday, and reapplication of sunscreen every two hours, respectively.

 

 

Threat
Low High
Effigy Low 3.33 3.48
High 3.57 4.25

 

82

j. Behavior

 

 

Source of Variation d_f Mean Square F p
Covariate 1 1 18.82 13.76 .00
Covariate 2 1 8.22 6.01 .02

Threat 1 26.60 19.45 .00

Efficacy 1 2.94 2.15 .14

Threat x efficacy 1 .51 .37 .54
Error 222 1.37

 

Note. Significant covariates were use of protective clothing when out in the sun and
minimization of sun exposure at midday, respectively.

 

 

Threat
Low High
E_ffic_a_c_y Low 2.43 3.02
High 2.57 3.35

 

83

Hypothesis 3

H3a stated that “When perceived response efficacy is low, perceived high threat is
likely to bring about greater degree of fatalism than when perceived response efficacy is
high.” The results are described in Table 8a and 8b. Significant main effects of threat (E
(l, 272) = 4.80, p < .05) and response efficacy (E (l, 272) = 6.76, p = .01) were found for
fatalism. Significant covariates were race (E (1 , 272) = 4.34, p < .05) and the level of tan
(E (1, 272) = 6.28, p < .05). No significant interaction effect was found. Subjects who
were in the high threat condition expressed greater degree of fatalism (adjusted M = 2.45)
than those who were in the low threat condition (adjusted M = 2.11). At the same time,
subjects who perceived low response efficacy indicated greater degree of fatalism
(adjusted M = 2.49) than those who perceived high response efficacy (adjusted M =
2.07).

H3b stated that “When perceived self efficacy is low, perceived high threat is
likely to bring about greater degree of hopelessness than when perceived self efficacy is
high.” Significant main effects of threat (E (1, 272) = 18.18, p = .005) and self efficacy
(E (1, 272) = 7.80, p < .01) were found for hopelessness. A significant covariate was
frequency of workout in the sun (E (1, 272) = 8.19, p = .005). No significant interaction
effect was found. Subjects who were in the high threat condition indicated greater degree
of hopelessness (adjusted M = 2.63) than those who were in the low threat condition
(adjusted M = 2.30), while subjects who perceived low self efficacy greater degree of
hopelessness (adjusted M = 2.76) than those who perceived high self efficacy (adjusted M

= 2.32).

84

Table 8a Summary of Angysis of Covariance of
Threat x Response Efficacy and Adiusted Means for Fatalism

 

 

Source of Variation d_f Mean Square E p
Covariate 1 1 11.47 6.95 .01
Covariate 2 1 8.58 5.20 .02

Threat 1 10.25 6.21 .01
Response efficacy 1 10.16 6.16 .01
Threat x efficacy 1 3.81 2.31 .13

Error 267 1.65

 

Note. Significant covariates were family members' use of sunscreen and level of tan,
respectively.

 

 

Threat
Low High
Lfficapy Low 2.17 2.80
High 2.02 2.17

 

85

Table 8b Summg of Analysis of Covariance of
Threat x Self Efficacy and Adiusted Means for Hopelessness

 

 

Source of Variation d_f Mean Square E p
Covariate 1 1 19.69 12.41 .00
Covariate 2 1 14.63 9.22 .00
Covariate 3 l 12.50 7.88 .01
Covariate 4 1 7.24 4.57 .03

Threat 1 15.95 10.05 .00

Efficacy l 7.20 4.54 .03

Threat x efficacy 1 1.19 .75 .39
Error 265 1.59

 

Note. Significant covariates were family members' use of sunscreen, friends' use of
sunscreen, frequency of workout, and level of tan, respectively.

 

 

Threat
Low High
My Low 2.53 2.88
High 2.07 2.69

 

86

Hymthesis 4

H4 stated that “When perceived threat is low, attitude toward risk behavior is
likely to be more favorable when response efficacy is low than when response efficacy is
high.” ANCOVA method was used to test this hypothesis. This hypothesis was not
supported. No significant main effect of threat (E (1, 272) = 1.80, p = ns) or efficacy (E
(1, 272) = .00, p = ns) or interaction (F (1, 272) = , p = ns) was discovered. Significant
covariates were frequency of going to a tanning parlor (E (1, 272) = 63.49, p = .001),
level of tan (E (1 , 272) = 11.62, p = .001), minimization of exposure to the sun at midday
(E (1, 272) = 5.23, p < .05), and use of protective clothing (E ( 1, 272) = 4.66, p < .05).

Results are described in Table 9.

87

 

Table 9 Summg of Analysis of Covariance of Threat x Response Efficacy and
Adiusted Means for Attitude toward Risk Behavior

 

 

Source of Variation d_f Mean Square E p
Covariate l 1 98.78 64.00 .00
Covariate 2 l 18.07 11.71 .00
Covariate 3 l 8.03 5.20 .02
Covariate 4 1 7.24 4.69 .03

Threat 1 1.19 .76 .38

Efficacy 1 .01 .00 .95

Threat x efficacy 1 .07 .05 .83
Error 265 1.54

 

Note. Significant covariates were frequency of going to a tanning parlor, level of tan,
minimization of sun exposure at midday, and use of protective clothing when out in the
sun, respectively.

 

 

Threat
Low High
My Low 3.57 3.67
High 3.55 3.72

 

88

Research Question 1

RQl asked, “How do threat, efficacy, and stages of change interact in bringing
about unintended effects?” The results are presented in Table 10.

Defensive avoidance. A significant main effect of stage was found for defensive
avoidance: E (2, 269) = 3.60, p < .05. No other significant main or interaction effects
were found. A significant covariate was frequency of going to a tanning parlor: E (1 ,
272) = 25.71, p < .001. Subjects in the precontemplation stage indicated greater degree
of defensive avoidance (adjusted M = 4.57) than those in the contemplation/preparation
stage (adjusted M = 4.30), who in turn indicated greater degree of defensive avoidance
than those in the action/maintenance stage (adjusted M = 3.89), regardless of threat or
efficacy levels.

Message derogation. A significant interaction between threat and efficacy was
found, E (1 , 271) = 10.10, p < .005, for message derogation. Significant covariates were
perceived accuracy of the message (E (1, 272) = 30.51, p < .001) and minimization of sun
exposure at midday message (E (1 , 272) = 4.76, p < .05). When threat was high, subjects
indicated greater degree of message derogation when perceived efficacy was low
(adjusted M = 4.15) than when perceived efficacy was high (adjusted M = 2.90).

Perceived manipulation. A significant interaction between threat and efficacy was
found for perceived manipulation: E (1, 264) = 4.92, p < .05. Significant main effects of
stage (E (2, 264) = 3.62, p < .05) was also found. Significant covariates were perceived
accuracy (E (1, 264) = 22.76, p < .001) and quality (E (1, 264) = 5.56, p < .05) ofthe

message, friends’ use of sunscreen (E (1, 264) = 7.75, p < .01), and number of times

89

peeled during the past two years (E (1, 264) = 4.70, p < .05). When threat was high,
subjects who perceived low efficacy indicated greater degree of perceived manipulation
(adjusted M = 3.27) than those who perceived high efficacy (adjusted M = 2.65).
Subjects in precontemplation (adjusted M = 2.98) and contemplation/preparation
(adjusted M = 3.13) stages indicated greater degree of perceived manipulation than those
who were in action/maintenance stage (adjusted M = 2.66).

Ea_ta_1is_m_. A significant main effect of threat was found for fatalism: E (1 , 271) =
5.78, p < .05. No other significant main or interaction effects were found. Significant
covariates were level of tan (E (1, 271) = 7.49, p < .01) and family members’ use of
sunscreen (E (1, 264) = 4.20, p < .05). Subjects who were in the high threat condition
(adjusted M = 2.42) indicated greater level of fatalism than those who were in the low
threat condition (adjusted M = 1.99) across different stages of change and levels of
efficacy.

Hopplessness. A significant main effect of threat was found for hopelessness: E
(1, 271) = 13.15, p < .001. No other significant main or interaction effects were found.
Significant covariates were friends’ (E (1 , 271) = 9.09, p < .005) and family members’ (E
(l, 271) = 4.76, p < .05) sunscreen use and frequency of workout in the sun E (1, 271) =
5.64, p < .05. Subjects who were in the high threat condition indicated greater level of
hopelessness (adjusted M = 2.82) than those who were in low threat condition (adjusted
M = 2.19) across different stages of change and levels of efficacy.

Reactance. A significant main effect of stage was found, E (1 , 268) = 4.55, p <

.05, for reactance. No other significant main or interaction effects were found.

90

Significant covariates were amount of time spent per week for leisure activities in the sun
during summer (E (l , 268) = 8.84, p < .01) and perceived amount of learning from the
message (E (1, 268) = 5.26, p < .05). Subjects in precontemplation stage indicated
greater degree of reactance (adjusted M = 3.73) than those in contemplation/preparation
stage (adjusted M = 3.36) who in turn indicated greater reactance than those in
action/maintenance stage (adjusted M = 3.29), regardless of the levels of threat or
efficacy.

Wishful thinking. A significant main effect of threat was obtained, E (1, 264) =
12.23, p = .001, for wishful thinking. No other significant main or interaction effects
were found. A significant covariate was perceived accuracy of the message (E (1, 264) =
4.57, p < .05). Subjects in the high threat condition indicated greater degree of wishful
thinking (adjusted M = 3.05) than who were in the low threat condition (adjusted M =
2.48), across stages of change and levels of efficacy.

mag. A significant main effect of efficacy was found, E (l, 271) = 6.63, p <
.05, for attitude. No other significant main or interaction effects were found. Significant
covariates were frequency of going to a tanning parlor (E (1, 271) = 7.10, p < .01), use of
protective clothing (E (1 , 271) = 5.13, p < .05), and minimization of sun exposure at
midday (E (1, 271) = 5.41, p < .05). Subjects who perceived low efficacy indicated less
favorable attitude toward recommended behavior (adjusted M = 4.37) than those who
perceived high efficacy (adjusted M = 4.77), across different stages of change and levels

of threat.

91

Intentions. A significant interaction between threat and efficacy was found, E (1,
263) = 5.89, p < .05, for intentions to engage in preventive behavior. Significant main
effects of stage (E (1 . 263) = 16.15, p < .001) was also found. Significant covariates were
minimization of sun exposure at midday (E (1 , 263) = 10.92, p = .001), frequency of
sunscreen use (E (1. 263) = 8.39. p < .005), and reapplication of sunscreen every two
hours (E (1, 263) = 5.44, p < .05). Subjects indicated fewer intentions to engage in
preventive behavior when threat was high and perceived efficacy was low (adjusted M =
3.72) than perceived efficacy was high (adjusted M = 4.46). Subjects in
precontemplation stage indicated fewer intentions (adjusted M = 3.34) than those in
contemplation/preparation stage (adjusted M = 3.89) who in turn indicated fewer
intentions than those in action/maintenance stage (adjusted M = 4.38).

Behavior. Significant main effects of stage (E (2, 226) = 24.81, p < .001) and
threat (E (1, 226) = 18.33, p < .001) were found for behavior. No other significant main
or interaction effects were found. A significant covariate was use of protective clothing
when exposed to the sun (E (1, 226) = 37.06, p < .001). Subjects in precontemplation
stage reported fewer behavioral changes (adjusted M = 2.41) than those in
contemplation/preparation stage (adjusted M = 2.86) who in turn reported fewer
behavioral changes than those in action/maintenance stage (adjusted M = 3.77). Fewer
behavioral changes were made when threat was low (adjusted M = 2.66) than high

(adjusted M = 3.37).

92

Table 10 Summag of Analysis of Covariance of Stage x Threat x Efficacy and

Adiusted Means

a. Defensive Avoidance

 

 

 

 

 

Source of Variation d_f Mean Square E p
Covariate l 64.56 25.71 .00
Stage 2 9.03 3.60 .03
Threat l 1.43 .57 .45
Efficacy 1 3 .74 1 .49 .22
Stage x threat 2 .30 .12 .89
State x efficacy 2 5.17 2.06 .13
Threat x efficacy 1 7.87 3.13 .08
Stage x threat x efficacy 2 .24 .09 .91
Error 257

Mtg. Significant covariate was frequency of going to tanning parlor.
S_taga
Precontemplation Contemplation/ Action/
Preparation maintenance
M
Low High Low High Low High
_Efi'flay Low 4.24 4.63 4.22 4.73 3.86 4.64
High 4.86 4.53 4.15 4.09 3.68 3.39

 

93

b. Message Derogation

 

 

Source of Variation d_f Mean Square F p
Covariate 1 l 52.81 38.51 .00
Covariate 2 l 6.53 4.76 .03
Stage 2 1.49 1.09 .34
Threat 1 .44 .32 .57
Efficacy l 7.91 5.77 .02
Stage x threat 2 .61 .44 .64
State x efficacy 2 2.69 1.96 .143
Threat x efficacy 1 13.85 10.101 .00
Stage x threat x efficacy 2 2.75 2.01 .14
Error 25 1.3

 

Note. Significant covariates were perceived accuracy of the message and minimization of
sun exposure at midday.

 

 

Stage
Precontemplation Contemplation/ Action]
Preparation maintenance
Threat
Low High Low High Low High
Efficacy Low 3.45 3.27 3.44 4.15 3.16 3.89
High 3.55 3.16 3.60 2.90 3.24 2.50

 

94

c. Perceived Manipulation

 

 

Source of Variation df Mean Square E p
Covariate 1 1 20.93 22.76 .00
Covariate 2 1 7.13 7.75 .01
Covariate 3 1 5.11 5.56 .02
Covariate 4 1 4.32 4.70 .03
Stage 2 3.33 3.62 .03
Threat l .23 .25 .62
Efficacy 1 4.94 5.37 .02
Stage x threat 2 .97 1.06 .35
State x efficacy 2 .79 .86 .43
Threat x efficacy 1 4.52 4.92 .03
Stage x threat x efficacy 2 2.55 2.77 .06
Error 249 .92

 

Note. Significant covariates were perceived accuracy of the message, fiiends' use of
sunscreen, perceived quality of the message, and number of times peeled during the
previous two years.

 

 

Stage
Precontemplation Contemplation] Action/
Preparation maintenance
Threat
Low High Low High Low High
Efficacy Low 3.11 3.01 3.25 3.59 2.34 3.21
High 3.01 2.78 2.75 2.94 2.87 2.22

 

95

d. Fatalism

 

 

Source of Variation d_f Mean Square E p
Covariate l l 12.55 7.49 .01
Covariate 2 1 7.03 4.20 .04
Stage 2 3.84 2.29 .10
Threat 1 9.68 5.78 .02
Efficacy 1 2.86 1.71 .19
Stage x threat 2 .75 .45 .64
State x efficacy 2 .49 .29 .75
Threat x efficacy 1 3.67 .2.19 .14
Stage x threat x efficacy 2 .01 .01 .99
Error 258 1.68

 

Note. Significant covariates were family members' use of sunscreen and level of tan.

 

 

Stage
Precontemplation Contemplation/ Action/
Preparation maintenance
Threat
Low High Low High Low High
Efficacy Low 2.17 2.78 1.67 2.65 2.08 2.60
High 2.35 2.47 1.62 2.02 2.05 2.02

 

96

e. Hopelessness

 

 

Source of Variation d_f Mean Square E p
Covariate 1 1 14.77 9.09 .00
Covariate 2 1 9.15 5.64 .02
Covariate 3 1 7.73 4.76 .03
Stage 2 .73 .45 .64
Threat 1 21.35 13.15 .00
Efficacy 1 4.94 3.04 .08
Stage x threat 2 3.88 2.39 .09
State x efficacy 2 1.57 .97 .38
Threat x efficacy 1 .88 .54 .46
Stage x threat x efficacy 2 1.24 .76 .47
Error 257 1.62

 

Note. Significant covariates were friends' use of sunscreen, frequency of workout in the
sun, and family members' use of sunscreen.

 

 

flags
Precontemplation Contemplation/ Action/
Preparation maintenance
Threat
Low High Low High Low High
Efficacy Low 2.60 2.63 2.03 3.34 2.18 3.17
High 2.45 2.76 1.96 2.62 1.86 2.43

 

97

f. Reactance

 

 

Source of Variation d_f Mean Square E p
Covariate 1 1 9.46 8.84 .00
Covariate 2 1 5.63 5.26 .02
Stage 2 4.86 4.55 .01
Threat 1 .02 .02 .89
Efficacy 1 1.47 1.38 .24
Stage x threat 2 .38 .35 .70
State x efficacy 2 .42 .40 .67
Threat x efficacy 1 .10 .09 .77
Stage x threat x efficacy 2 1.89 1.76 .17
Error 255 1.07

 

Note. Significant covariates were frequency of leisure activities in the sun and perceived

amount of learning.

 

 

Stage
Precontemplation Contemplation/ Action/
Preparation maintenance
Threat
Low High Low High Low High
Efficacy Low 3.97 3.58 3.48 3.60 3.15 3.49
High 3.64 3.72 3.10 3.25 3.49 3.05

 

98

g. Wishful Thinking

 

 

 

 

 

Source of Variation d_f Mean Square E p
Covariate 1 5.82 4.57 .03
Stage 2 .52 .41 .67
Threat 1 15.56 12.23 .00
Efficacy l .00 .00 1.00
Stage x threat 2 .00 .00 1.00
State x efficacy 2 .44 .35 .71
Threat x efficacy 1 .54 .43 .51
Stage x threat x efficacy 2 1.21 .95 .39
Error 252 1.27

1%, Significant covariate was perceived accuracy of the message.
Sgga
Precontemplation Contemplation/ Action/
Preparation maintenance
thcat
Low High Low High Low High
EM Low 2.53 3.04 2.62 3.16 2.12 3.10
High 2.33 2.98 2.54 3.16 2.72 2.85

 

99

h. Attitude

 

 

Source of Variation Qf Mean Square E p
Covariate 1 1 8.99 7.10 .01
Covariate 2 l 6.86 5.41 .02
Covariate 3 1 6.50 5.13 .02
Stage 2 2.79 2.20 .11
Threat 1 .79 .62 .43
Efficacy l 8.39 6.63 .01
Stage x threat 2 .71 .52 .57
State x efficacy 2 .22 .17 .84
Threat x efficacy 1 .52 .41 .52
Stage x threat x efficacy 2 .68 .56 .59
Error 257 1.27

 

Note. Significant covariates were frequency of going to a tanning parlor, minimization of
sun exposure at midday, and use of protective clothing.

 

 

Stage
Precontemplation Contemplation/ Action/
Preparation maintenance
Threat
Low High Low High Low High
Efficacy Low 4.45 3.95 4.31 4.56 4.68 4.25
High 4.57 4.51 5.00 4.96 4.79 4.82

 

100

i. Intentions

 

 

Source of Variation d_f Mean Square F p
Covariate 1 l 9.33 10.92 .00
Covariate 2 1 7.17 8.39 .00
Covariate 3 1 4.65 5.44 .02
Stage 2 13.81 16.15 .00
Threat 1 9.63 1 1.26 .00
Efficacy 1 8.32 9.73 .00
Stage x threat 2 1.20 1.40 .25
State x efficacy 2 .23 .27 .77
Threat x efficacy 1 5.04 5.89 .02
Stage x threat x efficacy 2 .02 .02 .98
Error 250 .86

 

Note. Significant covariates were minimization of sun exposure at midday, frequency of
using sunscreen, and reapplication of sunscreen every two hours when out in the sun.

 

 

Stags
Precontemplation Contemplation/ Action/
Preparation maintenance
Threat
Low High Low High Low High
Efficacy Low 3.09 3.20 3.53 3.94 4.18 4.03
High 3.19 3.86 3.49 4.61 4.42 4.91

 

101

j. Behavior

 

 

 

 

 

Source of Variation df Mean Square E p
Covariate 1 43.37 37.06 .00
Stage 2 29.03 24.81 .00
Threat 1 21.45 18.33 .00
Efficacy 1 .005 .005 .95
Stage x threat 2 .29 .24 .78
State x efficacy 2 .74 .64 .53
Threat x efficacy 1 .40 .34 .56
Stage x threat x efficacy 2 .10 .09 .92
Error 214 1.17

Mata. A significant covariate was use of protective clothing when out in the sun.
ma
Precontemplation Contemplation/ Action/
Preparation maintenance
Lh_re_at
Low High Low High Low High
Em Low 2.19 2.73 2.37 3.06 3.55 4.16
High 2.01 2.72 2.47 3.54 3.37 4.01

 

102

Research Question 2

RQ2 asked, “How are unintended effects associated with intended effects?”
Partial correlations. controlling for confounds. were run to answer this research question,
as summarized in Table 1 1.

Defensive avoidance was negatively associated with attitude (5 = -.20, p <.Ol ),
intentions (; = -.l9, p <01), and behavior ([ = -.17, p <.05). Reactance was also
negatively related to attitude ([ = -.14. p <.05), intentions (g = -.28, p <.001), and behavior
(; = -.24, p <.001). The more subjects engaged in defensive avoidance or perceived the
messages were threats to their freedom, the fewer changes in attitude, intentions, and
behavior were made. Similarly, message derogation was inversely related to intentions (g
= -.19, p <01) and behavior (; = -.23. p = .001). The more subjects derogated the
message the fewer changes in intentions or behavior were made.

In contrast, wishful thinking was positively associated with intentions (; = .23, p =
.001) and behavior (; = -.23. p = .001). The more individuals wished for panaceas after
exposure to the threat messages, the more changes in intentions and behavior toward
prevention were made. Perceived manipulation, fatalism, and hopelessness were not

related to intended effects.

103

Table 11 Results of Partial Correlations between Unintended and Intended Effects

 

 

 

 

Intended Effects

Unintended Effects Attitude Intention Behavior
Defensive -.20** -.19*** -.l7*
avoidance
Message derogation -.O6 —.19* * -.23 * *
Perceived -.OO -.07 -.05
manipulation
Fatalism -. 12 -.10 -.02
Hopelessness -.02 .01 .08
Reactance -.14* -.28*** -.24***
Wishful thinking .00 .23 * ** .23 ***

* p < .05

** p S .01

* ** p S .005

104

CHAPTER 5

DISCUSSION

The results of this study illustrate an interplay of threat, efficacy, and stages of
change that may elicit unanticipated consequences of fear appeals in health
communication campaigns.

The Role of Stage of Change

The results of H1 highlights the importance of considering dynamic audience
characteristics, stages of change, in effective message design. Findings indicate that in
general, unintended effects of fear appeals are more likely to occur in earlier stages of
behavior change than in mid or later stages of change.

Specifically, after being exposed to fear appeal messages, college students who
were in the precontemplation stage were more likely to: avoid the thoughts of skin cancer
and its prevention, be fatalistic about the threat of skin cancer, feel hopeless, and perceive
that the messages were constraining their behavioral freedom, than those who were in the
contemplation/preparation stage or the action/maintenance stage.

At the same time, after being exposed to fear appeal messages, those who were in
the precontemplation stage reported less favorable attitude toward message
recommendations, fewer intentions to engage in preventive behavior, and fewer changes
in preventive behavior.

The results of H1 hold significant implications as well as challenges for message

designers. Those who do not engage or who do not intend to engage in preventive

105

behavior may be the major intended audience of health communication campaigns, and
need communication campaigns most. The results indicated, however, that this group of
individuals are most likely to exhibit unintended effects of fear appeals.

Concurrently, findings of H1 are reminiscent of a long-held view on the effects of
persuasion in general, that creating responses is more difficult than reinforcing existing
responses (Klapper, 1960; Stiff, 1994). Of all the stages, individuals in the
precontemplation stage may need changes in attitude, intentions, and behavior most
because they do not at all intend to or engage in preventive behavior. However, results
indicate that this group of individuals are least likely to respond with intended changes
after having been exposed to fear appeals. At the same time, individuals in the
contemplation/preparation stage or action/maintenance stage, who have already intended
or engaged in preventive behavior indicate greater changes in intended attitude,
intentions, and behavior than those in precontemplation stage. Therefore, fear appeals
used in this study appears to have been more effective in reinforcing existing intentions
and behavior than eliciting new attitude, intentions, and behavior among those who need
such change.

The importance of these findings can be underscored by noting that most often,
the majority of any at-risk population are at precontemplation stage (Prochaska, 1994).
Specifically, Prochaska stated that those who do not intend to change behavior in the near
future typically comprises 50% of any given population. In this study, about 52% of

subjects were precontemplators (see Table 3).

106

Then, what can planners of health communication campaigns do to minimize
possibility of unintended effects among precontemplators? Why is it that those who do
not plan behavioral changes are more prone to responses unintended by prevention
messages using fear appeals? By the same token, why are those who plan and implement
behavioral changes more likely to report intended responses? Other findings of this study
provide some understanding.

Results of RQl indicate that for some type of unintended effects, stages of change
alone can determine their magnitude. Specifically, stage of change was an overriding
factor determining the magnitude of defensive avoidance and reactance. Therefore, the
refusal to change and insistence on risk-related behavior may be an outcome of avoidance
of thoughts on risk and its prevention or resistance against communication that appears to
delimit their behavioral freedom. As a result, those individuals may have less knowledge
than those in contemplation/preparation stage or action/maintenance stage.

The TM characterizes precontemplators as those who are do not recognize their
risk or who refuse to acknowledge risk or who do not want to undertake healthier
behavior (Prochaska et al., 1994). Results of RQl lend support to these characterizations,
suggesting that the obstinacy of such audience in health/risk communication contexts
may come from their avoidance and reactance tendency. Future fear appeal messages for
precontemplators should be designed in a way that the audience would not perceive as
being prohibitive. On the other hand, the tendency of avoidance among precontemplators
may be a manifestation of their lack of perceived efficacy. Because precontemplators are

those who have not engaged in preventive behavior, they would not be as confident as

107

others in their ability to perform message recommendations. For this reason, they may
like to avoid think about the risk that may be relevant to them.
The Role of Efficacy

The results of H2 suggest that even after being exposed to information and aware
of threat, individuals may still display unanticipated responses to fear appeals, however.
Although the TM does not provide systematic characterizations of individuals in different
stages of change, studies of the model (Gorely & Gordon, 1995; Grimley etal., 1993;
Marcus et al., 1994; Prochaska et al., 1990) have found a consistent pattern across a
dozen different types of behavior: As individuals move from the precontemplation to the
maintenance stage, self efficacy and response efficacy increase. And, the EPPM predicts
that intended changes in attitude, intentions, and behavior are not likely to occur when
perceived efficacy is lower than perceived threat.

The results of H2 and H3 illustrate that whether fear appeals in health
communication campaigns will produce intended or unintended effects is largely
determined by the level of perceived efficacy. Even though the threat was perceived as
severe and relevant, college students indicated the likelihood of derogating the fear-
arousing messages about skin cancer and expressing fewer intentions to engage in
behavioral changes, when perceived efficacy was low.

Further, when perceived efficacy was low, college students tried to avoid the
thoughts of skin cancer and its prevention, believed that the message was manipulative

and restricting their behavioral choices, and expressed less favorable attitude toward

108

message recommendations than when perceived efficacy was high, regardless of
perceived threat level.

In addition, this study found that lack of different type of efficacy result in
different kind of unintended effects. Perceived efficacy comprises two distinct
dimensions of self efficacy and response efficacy (Witte, 1992b). When college students
believed that message recommendations are not efficacious in averting the threat, they
became fatalistic, by just accepting the threat as a fact of life. On the other hand, when
they believed that they are not able to perform recommended behavior, college students
felt as if they were without hope (H3a, H3b).

Also, results of RQl lend further support to the above reasoning that lack of
efficacy makes unintended effects more conspicuous in earlier stages. Even after stages
of change was considered as a factor eliciting unintended effects (i.e., in stage x threat x
efficacy AN COVA), efficacy still remained as a key factor determining the effects and
effectiveness of fear appeals. Specifically, message derogation, perceived manipulation,
and intentions emerged as outcomes of interaction between threat and efficacy, and
efficacy was the determinant of attitude toward recommended behavior, across different
stages of change.

The obvious implication for planners of health communication campaigns is that
in fear appeals should convince that message recommendations are efficacious in
deterring the risk and that the audience is capable of performing the recommendations.

Individuals in action/maintenance stages can have reinforcement on their

109

perceived efficacy from their own trial and repeat exercise of recommended behavior, as
studies suggest (Fruin et al., 1991; Schafer et al., 1993).

For those who have yet to initiate behavior change, fear appeal messages will play
a primary role in persuading them about efficacy. That neither self efficacy nor response
efficacy are personality traits (Bandura, 1986) suggests campaigns can create intended
changes.

The Role of Threat

Overall, the results indicate that perceived threat can energize both intended and
unintended effects of fear appeals, and this is consistent with Rippetoe and Rogers'
(1987) and Witte's (1992a) findings. First, the hypothesized boomerang effect of low
threat changing attitude toward risk behavior more favorable was not supported (H4).
When threat was not perceived as significant or relevant, no changes in attitude toward
risk behavior occurred, even though message recommendations were not believed as
efficacious. As the EPPM predicts, low threat might not motivate either intended or
unintended responses.

Next, results indicate that high threat can energize either intended or unintended
effects. The likelihood of wishing for panacea was amplified when the level of perceived
threat was high, regardless of the level of efficacy. At the same time, more behavioral
changes were reported when perceived threat was high (RQl).

Therefore, simply lowering the level of threat may not bring about any changes at
all, either intended or unintended. Lowering the level of threat to reduce unintended

effects may mean losing the potential to produce intended effects, because low threat

110

results in null effects. To elicit intended changes for prevention, first of all, a health risk
should be portrayed to an audience as significant and relevant. While keeping the level of
threat sufficiently high, message designers should take a great deal of care to tailor fear
appeals to their audience's efficacy level, to elicit intended effects and to minimize the
potential of unintended effects.

Re_lations_hip between Unintended and Intended Effects

Determining the relationship between unintended effects and intended effects
would provide an understanding of the seriousness of the former. Unintended effects
would be most unwholesome when they interfere the occurrence of intended effects.

Results of RQ2 indicate that defensive avoidance, message derogation, and
reactance are the most adverse of all types of unintended effects, because they decrease
the likelihood of intended effects, including attitude, intentions, and behavior.
Importantly, the magnitude of defensive avoidance and reactance were mainly determined
by the stages of change individuals were in (RQl). Therefore, future research should
examine what makes precontemplators avoid the thoughts of risk and its prevention and
to react and to remain resistant to fear appeals messages.

Results also suggest that not all unintended effects of fear appeals are deleterious:
Wishful drinking was positively associated with intentions and behavior. Wishful
thinking is a mode of coping threat (McCrae, 1984). Because fear appeals are used in
health communication campaigns to change attitude, intentions, and behavior, wishful

thinking in and of itself can be an unanticipated, undesirable response.

111

However, it is not without salutary effects because it helps individuals to deal
with the emotional distress facing threat (Rippetoe & Rogers, 1987). Moreover, it seems
that wishful thinking does not hinder individuals from engaging in rational problem-
solving approach, as it strengthened the likelihood of behavior change.

How about fatalism, hopelessness, and perceived manipulation, which were not
associated with intended effects? Can they be safely ignored because they do not affect
intended responses negatively? Although this study found no relationship, another study
found fatalism energized risk behavior of AIDS among gay and bisexual men (Kalichman
et al., 1997). Future research should determine the type and magnitude of unintended
effects' impact on intended effects in diverse health issues.

Implications

The discussion of implications of this study should be preceded by the fact that
the magnitude of unintended effects was relatively smaller than that of intended effects
(see Table 5). Thus, although significant differences were found among stages as
predicted, they may represent "potential" of unintended effects, rather than "actual"
unintended effects.

However, the relatively small magnitude of unintended effects may stem from the
characteristics of the health issue examined in this study: skin cancer. Skin cancer may
be more preventable than diseases such as breast cancer, and the impact of sun exposure
on skin cancer is more cumulative (especially for the subjects of this study who were

college students) when compared to that of unprotected sex on AIDS. Therefore, the

112

magnitude difference between unintended and intended effects should not be taken as
representative of fear appeals advocating all kinds of disease prevention.

Then, what should designers of public health campaigns do if they are aware that
they are likely to cause some unintended effects? Should communicators avoid
communicating altogether because they might cause harm to some audiences? This may
not be the best solution because doing so may mean loss of other audiences' opportunity
to be exposed to and to benefit from health information. For example, in this study, fear
appeals were likely to reinforce individuals in action/maintenance stage's preventive
behavior, while showing the potential to induce unintended effects among
precontemplators.

In this light, a question that health communication campaigners should ponder
upon regarding the potential of unintended effects may be "whose good and whose
harm." Next, health communication campaigners should also determine how much good
would be good enough for them to communicate and how much harm would constitute
enough reason to not communicate to a given type of audience.

To do so, planners of health communication campaigns should clearly define
objectives. Like objectives of any campaigns, a health communication objective should
include measurable goals, target audience, and time frame. Clearly defining target
audience is important because a message that has the potential to produce negative
unintended effects for a type of audience may not do so for another type of audience. For
example, should fear appeal messages' potential of producing unintended effects among

precontemplators deter its delivery to maintainers who are more likely to benefit from it?

113

Clearly defining the time frame is also important because depending on when you
measure the effects, unintended consequences may appear more or less conspicuous. For
example, Witte (1992a) noted that unintended effects of fear appeals such as perceived
manipulation may go away over the long term. At the same time, some unintended
consequences of fear appeals such as desensitization may not be readily apparent in the
short term due to their cumulative nature.

Efforts to control unintended effects can occur at each of the phases of a
campaign: formative, process, and summative evaluations. In the formative evaluation
phase, planners of health communication campaigns may use in-depth interviews, focus
groups, and surveys to detect possible unintended effects of messages under
development. Planners may also want to pretest a message to diverse types of audience
in addition to target audience members because public health messages often have the
potential to reach unintended audience. Next, in the process evaluation phase, efforts can
be made to monitor whether messages are delivered at intended times to intended
audience. Such efforts may help campaigners diagnose and develop plans for summative
evaluation. If process evaluation indicated messages were delivered to unintended
audience, summative evaluation may ascertain not only intended effects, but also the
magnitude of unintended effects.

Further, as Guttman (1997) urged, planners of health communication campaigns
should be aware of values (i.e., intervention sponsor's) that may play a role in

determining the above suggested criteria.

114

Limitations

This study has some limitations. First, lab experiment may not have been the best
method to examine the key underlying premise of this study, that campaigns take place in
real world and therefore may produce unintended effects. Certainly, the results of
experimental research differ from those of field studies and hence the former lacks
external validity. Exposures to fear appeals were forced, rather than natural, in this study.
This experimental design does not allow the examination of the long-term effect of
repeated exposures to fear appeals either. However, a study in field setting would make it
difficult to determine the causality of unintended effects unambiguously.

Another limitation is the confounds. The high threat message was perceived to be
more accurate, clear, of better quality, understandable, and informative. Certainly,
greater care and efforts should have been taken for the equivalence of the messages. The
purpose of this study dictated that health issue to be examined should be one that is
relatively well-known and would involve comparable number of subjects per stage. With
a well-known health issue such as skin cancer, balancing the requirements of equivalence
and experimental induction was a complex task. however. So the efforts for experimental
induction turned out to have compromised their purity. Another look at the messages
suggest that the sentences in the high threat message to increase susceptibility may have
rendered the high threat message more informative and interesting as well (e.g., "the
[sun's] rays can reach down into three feet of water").

In addition, as mentioned earlier, the magnitude of unintended effects were less

than that of intended effects. In part, this difference might have come from the fact that

115

this study was done in Michigan, not in areas such as Arizona in which skin cancer may

have far more relevance to its residents.

116

APPENDICES

ll7

Appendix A]

Informed Consent Form
Dear students:

We are currently developing messages to increase college students’ knowledge about skin
cancer. Our objective is to provide college students with the facts about skin cancer.
Specifically, we would like to correct college students’ any misunderstanding of skin
cancer resulting from misinformation or lack of information.

While examining the message, please underline phrases or sentences that you think are
important in each paragraph. After you examined the message, you will be asked to
complete a questionnaire on your reactions to the message. Your thorough examination
of both the pictures and written components and reactions to them are very important in
our project. It will take about 15 minutes to complete the questionnaire. Four weeks
later, you will be asked to complete a follow-up questionnaire in classroom.

Your responses to both questionnaires will remain anonymous and will not be associated
with you in any manner. However, for us to be able to analyze data, we ask you to create
your own code which consists of the following: father’s first initial, mother’s first initial,
your first initial, the month you were born, and the last two digits of your telephone
number. Please do not put this code on this page, but on the questionnaire as indicated.
This way, we can guarantee your anonymity.

Your participation is voluntary. You may choose not to answer certain questions or
discontinue your participation at any time without penalty. We do hope, however, that
you will participate in this project because your thoughts and opinions will greatly help
us out.

You will receive an extra credit by participating in this project. Moreover. a drayvirygwill
be held to provide 10 students with $10 cormons for a locaMestaurant when you complete
the follow-up questionnaire. If you would like to obtain copies of the results or if you
have any questions regarding this study, please contact Hyunyi Cho at 432-1310.

Thank you. We greatly appreciate your assistance in this project.

Please indicate if you are willing to participate in this study by signing your name below:

(Printed Name)

(Signature)

118

Appendix A2
High Threat Message

Facts about Skin Cancer

Skin cancer is the most common and most rapidly increasing form of cancer in the
United States. One in six Americans will develop skin cancer during their life time.

Each year, an estimated one million people in this country learn that they have skin
cancer. According to current estimates, 40 to 50 percent of Americans who live to age 65
will have skin cancer at least once.

Who is at risk?

You are. You may think that the Michigan sun is not strong enough to cause skin
cancer. This is M true. According to the Michigan Department of Community Health,
the incidence and mortality rates for skin cancer in Michigan have increased markedly,
and have far surpassed the percentage increases for total cancer rates. College students
may not aware that they have more sun exposure (for example, walking between classes,
going to afternoon football games, and outdoor sporting activities) than others. The fact
is that getting burned just once can lead to skin cancer, and you can get burned on a
cloudy day. You should note that snow reflects the sun’s rays. The rays can reach down
into three feet of water. American Academy of Dermatology warns that dark brown or
black skin is MOI a guarantee against skin cancer. No one is immune to the damaging
effects of the sun. Other less important causes of skin cancer include repeated medical
and industrial X-ray exposure, scarring from disease or burns, and occupational exposure
to such compounds as coal and arsenic.

How does skin cancer affect you?

The sun’s damaging effects begin at an early age. While you are out in the sun,
changes other than skin color occur beneath the surface of the skin. Your skin becomes
dry and leathery. You will soon look much older than you are because of deep wrinkling
and sagging of the skin. Moreover, sun exposure damages the blood vessels in the
sublayer of you skin. It affects your immune system, therefore decreasing your body’s
natural ability to fight off infection, cancer, and other diseases. Those who die of skin
cancer tend to do so at an earlier age than those who die of other cancers. As a result,
skin cancer ranks higher than many more common cancers in terms of average years of
life lost per person dying. The cure for skin cancer is often worse than the disease.
Treatment includes surgery, freezing, radiation therapy, and chemotherapy. Many times a
combination of methods are used, causing a prolonged period of debilitating pain.

A case

Jennifer, a 21-year old student at Michigan State University, discovered a freckle
near her eye. Within a week, she developed a small bump underneath what is now a
mole. Thinking little of it, it basically went unnoticed. Very soon, the mole became
irregularly shaped and somewhat sore. At the advice of her roommate, Jenny finally went
to see her doctor. Soon she developed the most serious type of skin cancer, malignant
melanoma. Blisters formed, oozed, and crusted over. As they broke again, they let the
sappy liquid drain from the open wound. The infected area became larger. As her skin
cancer spread, skin rotted off her face, exposing the tender red underlying tissues. The

119

oozing sores typical of skin cancer gave her excruciating pain. A year later, Jennifer
became one of 9,500 patients to die from skin cancer.
Types of skin cancer

There are three kinds of skin cancer that can destroy your skin and permanently
damage your internal organs. These are basal cell carcinoma, squamous cell carcinoma,
and malignant melanoma.

Basal cell carcinoma is the most common type of skin cancer. It starts as a small,
shiny, pearly bump on face, neck, hands, and trunk. Many times it goes unnoticed. Cells
from this tumor can be dislodged into the bloodstream and gain a foothold in other
organs, such as the lung, liver, and stomach. Otherwise, this cancer will continue to grow
at a fairly rapid pace until it damages underlying tissues around the growth. It can
extend below the skin to the bone.

Squamous cell carcinoma will start out as red, scaly patches or nodules. This kind
of cancer usually shows up and deforms the face and lips. Like basal cell carcinoma,
squamous cells can spread to other parts of the body.

Malignant melanoma is the worst of all skin cancers is. It starts near a mole and
attacks the color of the skin, hence the name malignant melanoma. Many people die
from this type of cancer because it spreads so quickly throughout the body. Melanoma
suddenly appears without warning. Survival rates for melanoma go as low as 14%,
depending upon the degree that the cancer spreads. It is typically a slow and painful
death.

Any UV rays that get to your skin can be detrimental. College-aged persons are
not immune to skin cancer. On the contrary, college students are in a VERY HIGH-
RISK GROUP for developing skin cancer. Many think skin cancer is easy to treat. This
is untrue. If not detected and treated, skin cancer can be deadly.

T 0 reduce the risk of skin cancer:

0 Minimize your exposure to the sun at midday.

0 Wear hats or clothing that cover your body.

0 Perform periodic skin self-examination.

0 Apply sunscreen with at least a SPF—1 5 or higher to all areas of the body

exposed to the sun.

0 Reapply sunscreen with at least a SPF-15 or higher every two hours and after

swimming or perspiring.

120

Appendix A3
Low Threat Message

Facts about Skin Cancer

With physical fitness being on everyone's mind, outdoor activities are becoming
extremely popular. More and more people enjoy physical exercise outside. According to
the National Sporting Goods Association, 75.3 million people enjoy jogging, 50.7 million
people participate in bicycling, and 46.4 million people take part in swimming, just to
name a few. It is important for college students to be concerned about their health. Yet,
some of them may be overly worried about skin cancer because they are not well aware
of facts about this disease.

Who is at risk?

College-aged persons are rarely affected by skin cancer. According to the US.
Department of Health and Human Services, the leading causes of death among college-
aged persons are accidents and adverse effects, motor vehicle accidents, suicide, and
homicide. Skin cancer afflicts older people. There are two main reasons. First, most
skin cancer is caused by lifetime exposure to ultraviolet (UV) radiation. Because the
development of skin cancer is influenced by UV exposure accumulated over a prolonged
period of time, skin cancer most frequently affects persons of age over 50.

Next, the risk of developing skin cancer is affected by where a person lives. For
example, in Australia and South Africa, where they have high amount of UV radiation,
incidence of skin cancer is high. Similarly, skin cancer incidence in the United States is
positively correlated with annual dosages of surface solar UV radiation received at each
geographic location. People in areas such as Texas need to be cautious of skin cancer
because of the year-round strong sun. You at Michigan State where the sun is so rare
belong to a very low-risk group for developing skin cancer. Other less important causes,
including repeated exposures to industrial X-ray and such compounds as coal and arsenic,
are not relevant to college students either.

How is skin cancer treated?

Skin cancers are easy to detect and most can be cured. Physicians generally are
able to spot any skin cancer cells during regular medical exams. In cases of early
detection, the cancer is safely treated. Skin cancer accounts for much less than 1 percent
of all cancer deaths. It is encouraging to know that skin cancer is now almost 100 percent
curable if found early and treated promptly. Doctors can remove skin cancer much like
they would remove a wart.

A Case

A 52-year old farmer in New Zealand went to see a doctor because he has
developed an unusual mole on his face. It was found to be a form of basal cell
carcinoma. Because it was in an incipient stage, the mole was completely removed at the
time of the biopsy at the doctor’s office. First, a local anesthetic was given to numb the
area for a short period of time. Then, the doctor removed the growth and looked at the
tissue under a microscope. This is called a biopsy. Because skin cancer rarely spreads, a
biopsy often is the only test needed to determine the stage of skin cancer. To make sure
that the cancer has not spread to other parts of the body, a special x-ray test was done.

121

Types of skin cancer

There are three kinds of skin cancer. These are basal cell carcinoma, squamous
cell carcinoma, and malignant melanoma. Basal cell and squamous cell carcinomas are
sometimes called nonmelanoma skin cancer.

Basal cell carcinoma develops in the layers of cells that form the base of the
epidermis, usually on the face, neck, hands, and trunk. It appears as pearly, translucent
raised area. Basal cell carcinoma accounts for more than 90 percent of all skin cancers in
the United States. However, it is encouraging to know that the cure rate for basal cell
carcinoma is greater than 95%, and basal cell carcinoma accounts for less than 0.1% of
patient deaths due to cancer.

Squamous cell carcinoma is a red, raised, scaly patch or nodule with well defined
outline. It appears on the rim of the ear or the face. Like basal cell carcinoma, the cure
rate of squamous cell carcinoma is over 95%. The fatality rate from basal cell and
squamous cell carcinomas is less than 1 percent.

Malignant melanoma is a blemish with irregular borders that may turn shades of
red, blue, or white. Melanoma has its beginnings in melanocytes, the skin cells which
produce the dark protective pigment called melanin. Melanoma consists of melanocytes
which have been transformed into cancer cells. Melanoma is much less common than
basal cell or squamous cell carcinomas. Doctors have reported a decrease in the number
of problems associated with this cell. Melanoma detected early has an excellent chance
of being cured.

College students are a very low risk group for developing skin cancer. Skin
cancer is prevalent among older people of age over 50. Any doubt you have about any
place on your skin can be cleared up by a five minute visit with your doctor. Treatment is
quick, easy, and basically painless.

To reduce the risk of skin cancer:

0 Minimize your exposure to the sun at midday.

0 Wear hats or clothing that cover your body.

0 Perform periodic skin self-examination.

0 Apply sunscreen with at least a SPF-15 or higher to all areas of the body

exposed to the sun.

0 Reapply sunscreen with at least a SPF-15 or higher every two hours and afier

swimming or perspiring.

122

Appendix B

Measures

Father’s first initial
Mother’s first initial
Your first initial

Month you were born
Last two digits of your telephone number

Below is a question about your current lifestyle. Please place a check next to the ONLY ONE sentence
that best describes your current lifestyle.

Do you use sunscreen consistently (i.e., each time exposed to the sun) when you go out in the sun?

1. No, and 1 do not intend to start using it consistently within the next 6 months.

 

N

. No, but I intend to start using it consistently within the next 6 months.

 

b)

. No, but I intend to start using it consistently within the next 30 days.

 

A

. Yes, I have been using it consistently but for less than 6 months.

 

5. Yes, 1 have been using it consistently for more than 6 months.

 

123

Father’s first initial

Mother’s first initial

Your first initial

Month you were born
Last two digits of your telephone number

Please answer the following questions with the message you just examined in mind (please circle on a

number).

Not
at all

How much did this message make you feel frightened: 1

How much did this message make you feel tense:
How much did this message make you feel nervous:
How much did this message make you feel anxious:

How much did this message make you feel
uncomfortable:

I

l

1

1

How much did this message make you feel nauseous: l

2

2

2

3

3

4

4

5

5

6

6

Very
much

7

7

Please indicate how you feel at this time about each of the following statement by circling a number. 7
means “strongly agree,” 4 means “neutral,” and 1 means “strongly disagree.” Some of the questions may
sound similar. However, for us to be able to obtain all aspects of your perceptions, please answer all

questions.

. I am at risk for skin cancer.

9)

CT

. It is likely that I will develop skin cancer.

c. It is possible that I will develop skin cancer.

Strongly
Disagree

(1. I believe that skin cancer is a severe health problem. 1

e. I believe that skin cancer is a serious threat.
f. I believe that skin cancer is a significant disease.
c. Reapplying sunscreen with at least SPF-15

or higher every two hours to prevent skin cancer is
no problem for me.

1

124

Neutral

Strongly
Agree

7

7

Strongly Strongly

Disagree Neutral Agree
(1. I can easily reapply sunscreen with at least SPF-15
or higher every two hours to prevent skin cancer. 1 2 3 4 5 6 7
e. Reapplying sunscreen with at least SPF-15 or
higher after swimming or perspiring to prevent
skin cancer is no problem for me. 1 2 3 4 5 6 7
f. I can easily reapply sunscreen with at least
SPF-15 or higher after swimming or perspiring
to prevent skin cancer. I 2 3 4 5 6 7
g. Wearing clothing that covers my body to
prevent skin cancer is no problem for me. I 2 3 4 5 6 7
h. I can easily wear clothing that covers my
body to prevent skin cancer. 1 2 3 4 5 6 7
i. Wearing a hat that provide shade for my face
to prevent skin cancer is no problem for me. 1 2 3 4 5 6 7
j. I can easily wear a hat that provide shade for
my face to prevent skin cancer. 1 2 3 4 5 6 7
k. Minimizing my exposure to the sun at midday
to prevent skin cancer is no problem for me. 1 2 3 4 5 6 7
l. I can easily minimize my exposure to the sun
at midday to prevent skin cancer. 1 2 3 4 5 6 7
m. Performing periodic skin self-examination
to prevent skin cancer is no problem for me. 1 2 3 4 5 6 7
n. I can easily perform periodic skin
self-examination to prevent skin cancer. 1 2 3 4 5 6 7
a. Applying sunscreen with at least SPF-15
or higher to all areas of my body exposed
to the sun is effective in preventing skin cancer. 1 2 3 4 5 6 7
b. Reapplying sunscreen with at least SPF-15
or higher every two hours is effective in preventing
skin cancer. 1 2 3 4 5 6 7
c. Reapplying sunscreen with at least SPF-15
or higher after swimming or perspiring is
effective in preventing skin cancer. 1 2 3 4 5 6 7
d. Wearing clothing that covers the body is
effective in preventing skin cancer. 1 2 3 4 5 6 7
e. Wearing a hat that provides shade for my face is
effective in preventing skin cancer. 1 2 3 4 5 6 7

125

Strongly
Disagree Neutral
f. Minimizing my exposure to the sun at midday is
effective in preventing skin cancer. 1 2 3 4
g. Performing periodic skin self-examination is
effective in preventing skin cancer. 1 2 3 4
1. When I was reading the message and looking at
the pictures, my instinct was to: l 2 3 4
Want to
protect myself
from skin cancer
2. When I was reading the message and looking at
the pictures, my instinct was to: 1 2 3 4
Want to
think about
skin cancer
3. When I was reading the message and looking at
the pictures, my instinct was to: l 2 3 4
Want to

do something to
reduce my chance of
developing skin cancer

Strongly
Agree

5 6 7
Not want to

to protect

myself from

skin cancer

5 6 7
Not want to
think about
skin cancer

5 6 7

Not want to

do something to
reduce my chance of
developing skin cancer

Thank you. Below are sets of word pairs. Please place a check next to the word that best describes how

you personally feel about each behavior.

 

 

 

 

 

 

l. Minimizing my exposure to the sun at midday during summer is:

Good : : ' ‘ ' ' Bad

Undesirable Desirable
Favorable Unfavorable

2. In general, suntanning is:

Good Bad

Undesirable Desirable
Favorable Unfavorable
Strongly
Disagree Neutral

3. I intend to use sunscreen consistently
when exposed to sun within the next 6 months. 1 2 3 4
4. For the next month, each time when
I’m exposed to the sun, I intend to use sunscreen. l 2 3 4
5. When faced with the prospect of developing
skin cancer, it helps me to dream of a world where
there are no diseases like cancer. 1 2 3 4
6. I intend to buy a sunscreen to prevent skin cancer
during the next four weeks. 1 2 3 4
7. Nobody should be telling me when or
how I can get sun. 1 2 3 4

126

Strongly

Agree
5 6 7
5 6 7
5 6 7
5 6 7
5 6 7

 

 

 

 

 

Strongly Strongly
Disagree Neutral Agree

8. There are so many ways to get cancer today.

it’s just a matter of time; 1 might as well just try to

accept it. 1 2 3 4 5 6 7

9. This message was misleading. I 2 3 4 5 6 7

10. In this day and age. it sometimes seems

a hopeless task to stay healthy. 1 2 3 4 5 6 7

11. I intend to use sunscreen consistently

when exposed to sun within the next 30 days. 1 2 3 4 5 6 7

12. This message was: 1 2 3 4 5 6 7
Overstated Neutral Not at all

Overstated

13. This message was: 1 2 3 4 5 6 7

Exaggerated Neutral Not at all
Exaggerated
14. Applying sunscreen with at least SPF-15 or higher when out in the sun is:
Good : : : : ' ' Bad
Undesirable ' ‘ Desirable
Favorable Unfavorable
15. In general, aftemoons out in the sun are:
Good ' ' Bad
Undesirable Desirable
Favorable Unfavorable

 

16. Applying sunscreen with at least SPF -15 or higher to all areas of the body exposed to the sun is:

 

 

Good Bad
Undesirable Desirable
Favorable Unfavorable

 

l7. Reapplying sunscreen with at least SPF-15 or higher every two hours is:

 

 

 

Good Bad
Undesirable Desirable
Favorable Unfavorable
Strongly Strongly
Disagree Neutral Agree
18. It’s my life. I should be able to get sun if
1 want to. l 2 3 4 5 6 7
19. The next time I’m exposed to the sun,
I intend to use sunscreen. l 2 3 4 5 6 7
20. This message was distorted. l 2 3 4 5 6 7
21. I intend to use sunscreen consistently
when exposed to the sun. 1 2 3 4 5 6 7
22. I believe a miracle cure for cancer in the near future
is the answer to my fears about skin cancer. 1 2 3 4 5 6 7

23. Reapplying sunscreen with at least SPF-15 or higher after swimming or perspiring is:

 

 

Good : : : : : : Bad
Undesirable : : : : : : Desirable
Favorable : : : : : : Unfavorable

 

24. In general, sunbathing is:

 

 

Good : : : : : : Bad
Undesirable : : : : : : Desirable
Favorable : : : : : : Unfavorable

 

25. Wearing clothing to cover my body when out in the sun is:

 

 

 

 

 

 

 

 

 

 

 

Good : : . . Bad
Undesirable : : : : : : Desirable
Favorable : : : : : : Unfavorable
Strongly Strongly
Disagree Neutral Agree
26. Given what 1 know about skin cancer, I sometimes
feel it’s almost useless to try to stay healthy. 1 2 3 4 5 6 7
27. The next time I’m exposed to the sun, I intend
to reapply sunscreen with at least SPF 15 or higher
after swimming or perspiring to prevent skin cancer. 1 2 3 4 5 6 7
28. This message was: 1 2 3 4 5 6 7
Distorted Neutral Not at all
Distorted
29. This message was: 1 2 3 4 5 6 7
Overblown Neutral Not at all
Overblown
30. Wearing a hat to provide shade for my face is:
Good : : ' ' ' Bad
Undesirable : : : : : : Desirable
Favorable : : : : : : Unfavorable
31. In general, sporting activities out in the sun are:
Good : : : : : . Bad
Undesirable : : : : : : Desirable
Favorable : : : : : : Unfavorable
Strongly Strongly
Disagree Neutral Agree
32. Given what 1 know about skin cancer, I wish
I belonged to another time and another place. 1 2 3 4 5 6 7
33. I intend to minimize my exposure to the sun
at middays to prevent skin cancer. 1 2 3 4 5 6 7
34. Going to tanning parlors is:
Good : : : : : : Bad
Undesirable : : : : : : Desirable
Favorable : : : : : : Unfavorable

 

128

35. I intend to talk to my friends about using
sunscreen to prevent skin cancer within
the next four weeks.

36. This message was exploitative.

37. The next time I’m exposed to the sun,
I intend to apply sunscreen with at least SPF-15
or higher to prevent skin cancer.

38. This message tried to manipulate me.

39. The next time I’m exposed to the sun,

I intend to apply sunscreen with at least SPF-15
to all areas of my body exposed to the sun to
prevent skin cancer.

40. I resent being told that 1 can’t get sun.

41. The next time I’m exposed to the sun,
I intend to reapply sunscreen with at least SPF - l 5
or higher every two hours to prevent skin cancer.

42. If you are destined to die from skin cancer.
you will; there is little you can do about it.

43. The next time I’m exposed to the sun,
I intend to wear clothing that covers my body
to prevent skin cancer.

44. When I think about the prospect of cancer,
I sometimes feel like saying,
“What’s the use in staying healthy?”

45. The next time I’m exposed to the sun,
I intend to wear a hat that provide shades
for my face to prevent skin cancer.

46. Only time will tell if 1 develop skin cancer:
nothing can be done anyway but wait.

47. I intend to perform periodic skin self-examination

to prevent skin cancer.

48. When I think about the prospect of skin cancer.

I sometimes feel like eating and drinking too much.

49. I intend to use sunscreen each time when
exposed to the sun within the next 6 months.

50. I intend to use sunscreen consistently

when I’m exposed to the sun within the next month.

129

Strongly Strongly
Disagree Neutral Agree
1 2 3 4 5 6 7
l 2 3 4 5 6 7
l 2 3 4 5 6 7
l 2 3 4 5 6 7
l 2 3 4 5 6 7
1 2 3 4 5 6 7
l 2 3 4 5 6 7
l 2 3 4 5 6 7
l 2 3 4 5 6 7
1 2 3 4 5 6 7
l 2 3 4 5 6 7
l 2 3 4 5 6 7
1 2 3 4 5 6 7
l 2 3 4 5 6 7
l 2 3 4 5 6 7
1 2 3 4 5 6 7

Strongly
Disagree

51. I intend to buy a sunscreen to prevent skin cancer
during the next six months. 1
52. When 1 sunbathe, I tend to avoid thoughts of

skin cancer. 1
53. When I get a sunburn. I tend to avoid thinking

about skin cancer. 1

54. When 1 suntan, 1 try to not think about skin cancer. 1

55. Using sunscreen when out in the sun to prevent skin c

Good : :

 

Undesirable

 

Favorable

 

56. Performing periodic skin self-examination is:
Good ' ' ‘

 

Undesirable

 

Favorable

 

Please think about the message you read in answering the following questions.
Strongly
Disagree

1. This message was an accurate description

of skin cancer. 1
2. This message was an objective description

of skin cancer. 1
3. This message was clearly written.

4. I clearly understood this message.

5. I learned a lot about skin cancer from this message.

6. The quality of arguments in this message was good. 1

Thank you. Here are some questions about your lifestyle.

a. Please place a check next to phrases that best describe your skin type.

Always burns; never tans
Burns easily; tans minimally

Rarely burns; tans a lot to dark

99993.“:-

b. Do you know of anyone who has cancer? (Please check all that apply.)

No._
1 do:

2

2

ancer is:

Desirable
Unfavorable

Desirable
Unfavorable

Burns moderately; tans gradually to light brown
Bums minimally; always tans well to moderately

Never burns; deeply pigmented; insensitive

Neutral

4

Neutral

4

Strongly

Agree
6 7
6 7
6 7
6 7
Strongly
Agree
6 7
6 7
6 7
6 7
6 7
6 7

1. _Myself 2. ____Family 3.__Friend 4. _Acquaintance 5. _Neighbor

130

c. Do you know of anyone who has skin cancer? (Please check all that apply.)
No.__
1 do:
1. _Myself 2. _Family 3. _Friend 4. _Acquaintance 5. _Neighbor

d. In general, do you work out in the sun?
1. _Not at all 2. _Rarely 3. _Some ofthe time
4. _1/2 ofthe time 5. _Most of the time

e. How often do you currently use sunscreen?

1. Always when exposed to the sun 2. Often when exposed to the sun
3. Sometimes when exposed to the sun 4. Rarely when exposed to the sun
5. Never when exposed to the sun

f. Please rate how tan you are now. (Please check only one.)

1. Not at all 2. A little tan 3. Medium tan 4. A lot oftan
5. Very tan 6. N/A (Black skin) 7. I don’t try to tan
g. What SPF factor do you most ofien use? (Please circle on a number.)
None 2 4 8 15 25 30
Strongly Strongly
Disagree Neutral Agree

3. Most of my family members use sunscreen
when out in the sun. 1 2 3 4 5 6 7

b. Most of my friends use sunscreen when
out in the sun. 1 2 3 4 5 6 7

a. In general, I use sunscreen when
out in the sun more than 2 hours per day. 1 2 3 4 5 6 7

b. In general, I wear a hat to protect against
overexposure to the sun. 1 2 3 4 5 6 7

c. In general, I wear protective clothing against
overexposure to the sun. 1 2 3 4 5 6 7

d. In general, I minimize my exposure to

the sun at middays. l 2 3 4 5 6 7
e. How often do you go to a tanning parlor?

1. _Never 2. _Less than 1 week per year 3. _Less than 2 weeks per year

4. _Less than 3 weeks per year 5. _Less than 1 month per year

6. _Less than 6 weeks per year 7. _More than 6 weeks per year
f. How many hours do you spend out in the sun during the summer per week? hours per week
g. How many hours per week do you spend in the sun doing leisure activities? hours per week
h. I have gotten sunbumed time(s) during the last 2 years.
i. l have peeled time(s) during the last 2 years.

131

Please indicate the degree to which you agree or disagree with each of the following statements by circling a
number. Please, answer all questions.

Strongly Strongly
Disagree Neutral Agree
1. I intend to begin a new job without much
advance planning on how 1 will do it. 1 2 3 4 5 6 7
2. I usually think about what 1 am going to do before
doing it. 1 2 3 4 5 6 7
3. 1 often do things on impulse. l 2 3 4 5 6 7
4. I very seldom spend much time on the details of
planning ahead. 1 2 3 4 5 6 7
5 . 1 like to have new and exciting experiences and
sensation even if they are a little frightening. l 2 3 4 5 6 7
6. Before I begin a complicated job, I make
careful plans. 1 2 3 4 5 6 7
7. I would like to take off on a trip with no
preplanned or definite routes or timetable. l 2 3 4 5 6 7
8. I enjoy getting into new situations where you can’t
predict how things will turn out. 1 2 3 4 5 6 7
9. I like doing things just for the thrill of it. 1 2 3 4 5 6 7
10. I intend to change interests frequently. 1 2 3 4 5 6 7
ll. I sometimes like to do things that are a little
frightening. 1 2 3 4 5 6 7
12. I’ll try anything once. 1 2 3 4 5 6 7
13. I would like the kind of life where one is
on the move and traveling a lot, with lots of
changes and excitement. I 2 3 4 5 6 7
14. I sometimes do ‘crazy’ things just for fun. 1 2 3 4 5 6 7
15. I like to explore a strange city or section of town
by myself, even if it means getting lost. 1 2 3 4 5 6 7
16. I prefer friends who are excitingly unpredictable. l 2 3 4 5 6 7
17. I often get so carried away by new and exciting
things and ideas that I never think of
possible complications. 1 2 3 4 5 6 7
18. I am an impulsive person. 1 2 3 4 5 6 7

132

Strongly Strongly
Disagree Neutral Agree

19. I like ‘wild’ uninhibited parties. 1 2 3 4 5 6 7

Please indicate the degree to which you agree or disagree with each of the following statements by circling
a number.

Strongly Strongly
Disagree Neutral Agree
1. I become frustrated when I am unable to make free
and independent decisions. 1 2 3 4 5 6 7
2. I become angry when my freedom of choice is
restricted. 1 2 3 4 5 6 7
3. It irritates me when someone points out things
which are obvious to me. 1 2 3 4 5 6 7
4. Regulations trigger a sense of resistance in me. 1 2 3 4 5 6 7
5. I find contradicting others stimulating. l 2 3 4 5 6 7
6. When something is prohibited, 1 usually think
"that's exactly what I am going to do." 1 2 3 4 5 6 7
7. I resist the attempts of others to influence me. 1 2 3 4 5 6 7
8. It makes me angry when another person is
held up as a model for me to follow. 1 2 3 4 5 6 7
9. When someone forces me to do something,
I feel like doing the opposite. l 2 3 4 5 6 7
10. I consider advice from others to be an intrusion. 1 2 3 4 5 6 7
11. Advice and recommendations induce me to do
just the opposite. I 2 3 4 5 6 7

Thank you. Lastly, we’d like to ask a few questions about you to help us classify our data.
a. How old were you on your last birthday?

b. What is your gender? 1. _Male 2. _Female

c. What best describes your race or ethnic group?

1. _Caucasian 2. _African American 3. _Hispanic 4. _Asian or Pacific Islander
5. Other (please specify)

 

*Please do NOT talk with any of your classmates about the message until the end of the term.

Thank you very much for your time and cooperation. We greatly appreciate your help.

133

We are now interested in what you were thinking about while you were reading the message.

You might have had ideas all favorable to the recommendations in the message, all opposed, all irrelevant
to the recommendations in the message, or a mixture of the three.

Any case is fine; simply list what it was that you were thinking while reading the message. Simply write
down the first idea you had in the first box, the second idea in the second box, and so on.

Please put only one idea or thought in a box.

You should try to record only those ideas that you were thinking during the time you were reading the
message.

Please state your thoughts and ideas as concisely as possible...a phrase is sufficient. We have deliberately
provided more space than we think most people will need to insure that everyone would have plenty of
room to write the ideas they had during the message.

 

 

 

 

 

 

 

 

 

134

 

REFERENCES

Adler, R., Lesser, G., Meringoff, L.. Robertson, T., Rossiter, J ., & Ward, S.
(1980). The effects of television on children: Review and recommendations. Lexington,

MA: Lexington Books.

Atkin, C., Greenberg, B., & McDermott, S. (1983). Television and race role
socialization. Journalism Quarterlyg60, 407-414.

Andreasen, A. (1995). Marketing social change. San Francisco, CA: Jossey-
Bass.

Armstrong, G. B., Neuendorf, K. A., & Brentar, J. E. (1992). TV entertainment,
news, and racial perceptions of college students. Journal of Communication 42, 153-
176.

 

Bandura, A. (1986). SociaM’oundations of thought and action: A social cogm'tive
theog. Englewood Cliffs, NJ: Prentice Hall.

Bandura, A., Ross, D., & Ross, S. A. (1963). Imitation of film-mediated
aggressive models. Joumal of Abnormal and Social Psychology, 66, 3-11.

Baranowski, T. (1989-1990). Reciprocal determinism at the stages of behavior
change: An integration of community, personal, and behavioral perspectives.
International Quarterly of Health Education, 10, 297-327.

Barsky, A. J. (1988a). The paradox of health. New England Journal of Medicine.
ME, 414-418.

Barsky, A. J. ( 1988b). Worried sick: Our troubled quest for wellneg. Boston:
Little Brown.

Becker, M. H. (1993). A medical sociologist looks at health promotion. Journal
of Health and Social Behavior, 34, 1-6.

Belch, G. E., & Belch, M. A. (1995). Introduction to advertising and promotion.
Irwin.

Berkowitz, L. (1984). Some effects of thoughts on anti- and prosocial influences
of media events: A cognitive neoassociation analysis. Psychological Bulletin, 95, 410-
417.

Blood, D., & Snyder, L. B. (1993). Why warnings boomerang: The failure of the

Surgeon General’s alcohol warnings to affect young adults. Paper presented at the annual
meeting of the Association for Educators in Journalism and Mass Communication.

135

Brehm, J. W. (1966). A theory of psychological reacta_npa New York: Academic
Press.

Bensley, L. S., & Wu, R. (1991). The role of psychological reactance in drinking
following alcohol prevention messages. Journal of Applied Social Psychology, 21, 1111-
1124.

Brown, J. D., & Walsh-Childers. K. (1994). Effects of media on personal and
public health. In J. Bryant & D. Zillmann (Eds), Media effects: Advances in theory and
research (pp. 417-436). Hillsdale: NJ: Lawrence Erlbaum.

Burton, 1., Kates, R. W., & White, G. F. (1993). The environment as hazard.
New York: Guilford Press.

Cacioppo, J. T., & Petty, R. E. (1981). Social psychological procedures for
cognitive response assessment: The thought -listing technique. In TV. Merluzzi, C.R.
Glass, & M. Genest (Eds), Cognitive assessment (pp. 309-342).

Carmel, S. (1994). The role of age and an expanded Health Belief Model on
predicting skin cancer protective behavior. Health Education Research. 9. 433-447.

Champion, V. L. (1994). Beliefs about breast cancer and mammography by
behavioral stage. OncologyNursinfiomm, 21. 1009-1014.

Clarke, V. A. (1997). Skin type and Optimistic bias in relation to the sun
protection and suntanning behaviors of young adults. Journal of Behavioral Medicine.
M, 207-222.

Comstock, G., & Paik, H. (1991). Television and the American child. San Diego,
CA: Academic Press, Inc.

De Vries, H., & Backbier, E. (1994). Self-efficacy as an important determinant of

quitting among pregnant women who smoke: The @ pattern. Preventive Medicine. 23.
167-174.

DiClemente, C. C., & Prochaska, J. O. (1982). Self-change and therapy change of
smoking behavior: A comparison of processes of change in cessation and maintenance.

Addictive Behfliors, 7, 133-142.

DiClemente, C. C., & Hughes, S. (1990). Stages of change profiles in outpatient
alcoholism treatment. Joumfial of Substance Abuse&. 217-235.

136

DiClemente, C. C., Prochaska, J. O., Fairhurst, S. K., Velicer, W. F., Velasquez,
M. M., & Rossi, J. S. (1991). The process of smoking cessation: An analysis of
precontemplation, contemplation, and preparation: Stages of change. Journal of
Consulting and Clinical Psychology, 59, 295-304.

Donnerstein, E., Linz, D., & Penrod, S. (1987). The question of pornography:
Research findings and policy implications. New York: Free Press.

Donohew, L., Sypher, H. E., & Bukoski, W. J. (1991). Persuasive communication
and drug abuse prevention: Hillsdale, NJ: Lawrence Erlbaum.

Elliot, T. R., & Byrd, E. K. (1983). Attitude change toward disability through
television portrayal. Jouraal of Applied Rehabilitation Counseling, 14, 35-37.

Feingold, P. C., & Knapp, M. L. (1977). Anti-drug abuse commercials. Journal
of Communication 27, 20-28.

 

Feinstein, A. R., & Esdaile, J. M. (1987). Incidence, prevalence, and evidence:
Scientific problems in epidemiological statistics for the occurrence of cancer. American
Journal of Medicine. 82. 113-123.

F iske, S. T., & Taylor, S. E. (1986). Social cognition. McGraw-Hill.

Flay, B. R. (1986). Efficacy and effectiveness trials (and other phases of research)
in the development of health promotion programs. Preventive Medicine. 15. 451-474.

F ruin, D., Pratt, C., & Owen, N. (1991). Protection motivation theory and
adolescents’ perceptions of exercise. Joumiof Applied Social Psychology, 22, 55-69.

Gerbner, G., Gross, L., Morgan, M., & Signorielli, N. (1980). The
“mainstreaming” of America: Violence profile number 11. Journal of Communication
M, 10-29.

 

Glantz, S. A. (1996). Editorial: Preventing tobacco use--The youth access trap.
American Journal of Public Health. 86. 156-158.

Glanz, K., Patterson, R. E., Kristal, A. R., DiClemente, C. C., Heimendinger, J .,
Linnan, L., & McLerran, D. F. (1994). Stage of change in adopting health diets: Fat,
fiber, and correlates of nutrient intake. Health Education Quarterly. 21. 499-519.

Gorely, T., & Gordon, S. (1995). An examination of the transtheoretical model

and exercise behavior in older adults. Journal of Sport & Exercise Psychology, 17, 312-
324.

137

Graybar, S. R., Antonuccio, D. O., Boutilier, L. R., & Varble, D. L. (1989).
Psychological reactance as a factor affecting patient compliance to physician advice.
Scandinavian Journal of Behgioral Theories, 18, 43-51.

Greenberg, B. S., Brown, J. D., & Buerkel-Rothfuss, N. L. (1993). Media sex,
and the adolescent. Cresskill, NJ: Hampton Press.

Grimley, D. M., Riley, G. E., Bellis, J. M., & Prochaska, J. O. (1993). Assessing
the stages of change and decision-making for contraceptive use for the prevention of

pregnancy, sexually transmitted diseases, and acquired immunodeficiency syndrome.
Health Education Quarterly, 20, 455-470.

Guttman, N. (1996). Values and justifications in health communication
interventions: An analytic framework. Journal of Health Communication 1, 365-396.

 

Guttman, N. (1997). Beyond strategic research: A value-centered approach to
health communication interventions. Communication Theory, 7. 95-124.

Hale, J. L., & Dillard, J. P. (1995). Fear appeals in health promotion campaigns:
Too much, too little, or just right? In E. Maibach & R. L. Parrott (Eds), Designing health
messages: Approaches from communication theory and public health practice (pp. 65-
80). Thousand Oaks, CA: Sage.

Hanley, J. M. (1996). Positive attitudes towards suntanning and reported
tendency to engage in lifestyle behaviors that increase risk of skin cancer. Psychological
Reports, 79, 417-418.

Holbrook, M. B. (1987). Mirror, mirror, on the wall, what’s unfair in the
reflection on advertising? Journal of Marketing. 51. 95-103.

Hovland, C. I. (1959). Reconciling conflicting results derived from experimental
and survey studies of attitude change. American Psychologist. 14. 8-17.

Hovland, C. I., Janis, I. L., & Kelly, H. H. (1953). Communication and
persaasion: Psychological studies of opinion change. Yale University Press: New Haven,
CT.

Hyland, M., & Birrell, J. (1979). Government health warnings and the
“boomerang” effect. P_sychological Reports. 44. 643-647.

Illich, I. (1976). Medical nemesis: The expropriation of heal_t_h. New York:
Pantheon.

138

Janis, I. L. (1967). Effects of fear arousal on attitude change: Recent
developments in theory and experimental research. In L. Berkowitz (Ed.), Advances in
experimental social psychology (pp. 166-225). New York: Academic Press.

Janis, I. L., & Feshbach, S. (1953). Effects of fear-arousing communications.
The Joml of Abnormal and Social Psychology. 48. 78-92.

Janis, I. L., & Mann, L. (1977). Decision making: A psychological analysis of
conflict choice and commitment. New York: The Free Press.

 

Janis, I. L., & Terwilliger, R. F. (1962). An experimental study of psychological
resistances to fear arousing communications. Journal of Abnormal and Socia_l
Psychology. 65. 403-41 0.

Janz, N.K. & Becker, M.H. (1984). The health belief model: A decade later.
_Iiealth Education Quarterly, _1_1, 1-47.

Joseph, J. G., Montgomery, S. B., Emmons, C. A., Kessler, R. C., Ostrow, D. G.,
Wortman, C. B., O'Brien, K., Eller, M., & Eshleman, S. (1987). Magnitude and
determinants of behavioral risk reduction: Longitudinal analysis of a cohort at risk for
AIDS. Psychology and Health. 1, 73-96.

Kalichman, S. C., Kelly, J. A., Morgan, M., & Rompa, D. (1997). F atalism,
current life satisfaction, and risk for HIV infection among gay and bisexual men. Journal
of Con_sulting and Clinical Psychology. 65. 542-546.

Casey, M. K. (1995). F atalism and the modification of the Extended Parallel
Process Model. Paper presented at the annual meeting of the Speech Communication
Association, San Antonio, TX.

Horowitz, I. A. (1969). Effects of volunteering, fear arousal, and number of
communications on attitude change. Journal of Personality and Social Psychology, 1, 34-
37.

Horowitz, I. A. (1972). Attitude change as a function of perceived arousal. The
Journal of Social Psychology, 87, 117-226.

Kegeles, S. S. (1981). Review of the Health Belief Model and personal health
behavior. Social Science; and Medicine. 14C, 227, 1981.

Kerlinger, F. N. (1986). Foundations of behafivioral research. Harcourt Brace
Jovanovich College Publishers.

Kirk, R. E. (1982). Egaerimental design: procedures for the behavgal sciences.
Belmont, CA: Brooks/Cole.

139

Klapper, J. T. (1960). The effects of mass communication. New York: Free
Press.

Kleinot, M. C., & Rogers, R. W. (1982). Identifying effective components of
alcohol misuse prevention programs. Journal of Studies on Alcohol. 43. 802-811.

Kling, R. (1995). Computerization and controversy. San Diego: Academic Press,
1995.

Kotler, P. (1978). The elements of social action. In G. Zaltrnan (Ed.), Processes
and phenomena of social change (pp. 169-189). Huntington, NY: Robert E. Krieger.

Kunkel, D., & Gantz, W. (1992). Children’s television advertising in the
multichannel environment. Journal of Communication 42, 134-152.

 

Leventhal, H. (1970). Findings and theory in the study of fear communications.
In L. Berkowitz (Ed.), Advances in experimental social psychology (pp. 119-186). New
York: Academic Press.

Leventhal, H. (1971). Fear appeals and persuasion: The differentiation of a
motivational construct. American Journal of Public Health. 61. 1208-1224.

Leventhal, H., Safer, MA., & Panagis, D.M. (1983). The impact of
communications on the self-regulation of health beliefs, decisions, and behavior. Health

Education Quarterly, 10, 3-29.

Leventhal, H., Singer, R. P., Jones, S. (1965). Effects of fear and specificity of
recommendation upon attitudes and behavior. Journal of Personality and Social

Psychology, 2, 20-29.

Maddux, J. E., Rogers, R. W. (1983). Protection motivation and self-efficacy: A

revised theory of fear appeals and attitude change. Jomal of Experimental Social
Psychology. 19. 469-479.

Maibach, E. W., & Cotton, D. (1995). Moving people to behavior change. In E.
Maibach & R. L. Parrott (Eds), Designing health messages: Approaches from
communication theory and public health practice (pp. 41-64). Thousand Oaks, CA: Sage.

Marcus, B. H., Eaton, C. A., Rossi, J. S., & Harlow, L. L. (1994). Self-efficacy,
decision-making, and stages of change: An integrative model of physical exercise.

Journal of Applied Social Psychology, 24, 489-508.

McCrae, R. R. (1984). Situational determinants of coping responses: Loss, threat,
and challenge. Journal of Personality and Social Psychology. 46, 919-928.

140

McGuire, W. J. (1988). The mith of massive media impact: Savagings and
salvagings. In G. Comstock (Ed.), Public Communication and Behavior (pp. 174-257).
Academic Press.

McKeown, T. (1979). The role of medicine: Dream. mirage. or nemesis?
Princeton, NJ: Princeton University Press.

McKiman, D. J ., Ostrow, D. G., & Hope, B. (1996). Sex, drugs, and escape: A
psychological model of HIV-risk sexual behaviors. AIDS Care. 8. 655-669.

McMahan, S., Witte, K., Meyer, J. (1998). The perception of risk message
regarding electromagnetic fields: Extending the extended parallel process model to an
unknown risk. Health Communication, 10, 247-259.

McPhee, J. (1989). The control of nature. New York: Farrar, Straus, & Girous.

Mechanic, E. (1983). Handbook of health, health care. an_d the health professions.
New York: Free Press.

Merton, R. K. ( 1957). Social theory and social structure. Glencoe, IL: The Free
Press.

Miller, L. C., Murphy, R., Buss, A. H. (1981). Consciousness of body: Private
and public: Joumal of Personality and Social Psychology. 41. 397-406.

Mongeau, P. A. (1995). Another look at fear-arousing persuasive appeals. In M.
Allen & R. W. Preiss (Eds), Persuasion: Advances through Meta Analysis (pp. 53-58).

Montgomery, S.B., Joseph, J.G., Becker, M.H., Ostrow, D.G., Kessler, R.C. &
Kirscht, J .P. (1989). The health belief model in understanding compliance with
preventive recommendations for AIDS: How useful? AIDS Educafion aad Prevention. 1.
303-323.

Nesse, R. M., & Williams, (1994). Why we get sicl_<: The new science of
Darwinian medicine. New York: Vintage Books.

Pennebaker, J. W. (1982). The psychology of physical syrmnoms. New York:
Springer-Verlag.

Phillips, D. P. (1983). The impact of mass media violence on US homicides.
American Sociological Review. 48. 560-568.

141

Pollay, R. W. (1984). The identification and distribution of values manifest in
print advertising, 1900-1980. In R. E. Pitts & A. G. Woodside (Eds), Personal values
and consumer behavior (pp. 111-135). Lexington, MA: Lexington Books.

Pollay, R. W. (1986a). The distorted mirror: Reflections on the unintended
consequences of advertising. Journal of Marketing, 50, 18-36.

Pollay, R. W. ( 1986b). The quality of life in “ the padded sell”: Common
criticisms of advertising’s cultural character and international public policies. Current
Issues and Research in Advertising, 9, 173-250.

Pollay, R. W. (1987). On the value of reflections on the values in “the distorted
mirror.” Journal of Marketing. 51. 104-109.

Pollay, R. W. (1989). Campaigns, change and culture: On the polluting potential
of persuasion. In C. T. Salmon (Ed), Informatiorfimpaigns: Balancing social values
and social change (pp. 19-53). Newbury Park, CA: Sage.

Prochaska, J. O. (1991). Assessing how people change. Cancer 67, 805-807.

 

Prochaska, J. O. (1994). Strong and weak principles for progressing from
precontemplation to action on the basis of twelve problem behaviors. Health Psychology,
1;, 47-51.

Prochaska, J. O., & DiClemente, C. C. (1984). Stages and processes of self-
change of smoking: Toward an integrative model of change. Journal of Consulting and
Clinicflsychology. 51, 390-395.

Prochaska, J. O., & DiClemente, C. C. (1984). The transtheoretical approach:
Crossing traditional boundaries of change. Homewood, IL: Dorsey Press.

Prochaska, J. O., DiClemente, C. C. (1988). Measuring process of change:
Applications to the cessation of smoking. Journal of Consulting and Clinical Psychology,
M, 520-528.

Prochaska, J. O., DiClemente, C. C., & Norcross, J. C. (1992). In search of how
people change: Applications to addictive behaviors. American Psychologist. 47, 1102-
1114.

Prochaska, J. O., Harlow, Redding, & Snow (1990).

Prochaska J. O., Redding, C. A., Harlow, L. L., Rossi, J. S., & Velicer, W.F.

(1994a). The transtheoretical model of change and HIV prevention: A review. Health
Education Quarterly, 21, 471-486.

142

Prochaska J. O., Velicer, W. F ., Rossi, J. S., Goldstein, M. G., Marcus, B. H.,
Rakowski, W. R., F iore, C., Harlow, L. L., Redding, C. A., Rosenbloom, D., Rossi, S. R.
(1994b). Stages of change and decisional balance for 12 problem behaviors. Health
Psychology, M, 39-46.

Rakow, L (1989). Information and power. In C. T. Salmon (Ed.), Information

campaigns: Balancing social values and social change (pp. 164-184). Newbury Park,
CA: Sage.

Rakowski, W., Jube, C. E., Marcus, B. H., Prochaska, J. O., Velicer, W. F ., &
Abrams, D. B. (1992). Assessing elements of women’s decisions about mammography.
Health Psychology. 11, 111-1 18.

Rimer, B. K. (1996). Perspectives on intrapersonal theories in health education
and health behavior. In K. Glanz, F. M. Lewis, & B. K. Rimer. (Eds), Health behavior
and health education (pp. 139-148). San Francisco: Jossey-Bass.

Rippetoe, P. A., & Rogers, R. W. (1987). Effects of components of protection-
motivation theory on adaptive and maladaptive coping with a health threat. Journal of
Personality and Social Psychology. 52, 596-604.

Rosenstock, I.M. (1974). Historical Origins of the Health Belief Model. Health
Education Monographs, 2, 328-335.

Rogers, R. W. (1975). A protection motivation theory of fear appeals and attitude
change. Journal of Pflchology. 91. 93-114.

Rogers, R. W. (1983). Cognitive and physiological processes in fear appeals and
attitude change: A revised theory of protection motivation. In J. Cacioppo & R. E. Petty
(Eds), Social psychophysiology (pp. 153-176). New York: Guilford.

Rogers, R. W., & Mewbom, C. R. (1976). Fear appeals and attitude change:
Effects of a threat’s noxiousness, probability of occurrence, and the efficacy of coping
responses. Joumal of Persongitag and Social Psychology. 34. 54-61.

Rogers, E. M., & Storey, J. D. (1987). Communication campaigns. In C. R.
Berger & S. H. Chaffee (Eds), Handbook of communication science (pp. 817-846).
Beverly Hills, CA: Sage.

Rossi, J. S. (1989). Exploring behavioral approaches to UV risk reduction. In A.
Moshell & L. U. Blankenbaker (Eds), Sunlight, ultraviolet radiation and the skin (pp. 91-
93). Bethesda, MD: National Institutes of Health.

Rotter, J. B. (1954). Social learning and clinical psychology. Englewood Cliffs,
NJ: Prentice Hall.

143

Sagan, L. A. (1987). The health of nations. New York: Basic Books.

Salmon, C. T. (1989). Campaigns for social “improvement”: An overview of
values, rationales, and impacts. In C. T. Salmon (Ed), Information campaigns:
glancing social values and social change (pp. 19-53). Newbury Park, CA: Sage.

Salmon, C. T. (1992). Bridging theory “of” and theory “for” communication
campaigns: An essay on ideology and public policy. In S. A. Deetz (Ed),
Communication Yearboow (pp. 346-358). Newbury Park, CA: Sage.

Salmon, C. T. (1996). The other side of the coin: A typology of unintended
effects of social marketing. Paper presented at the second annual social marketing
conference, Fairfax, VA.

Schafer, R. B., Schafer, E., Bultena, G., Hoiberg, E. (1993). Coping with a health
threat: A study of food safety. Journal of Applied Psychology. 23. 386-394.

Schacter, S. (1959). The psychology of affiliation. Stanford, CA: Stanford
University Press

Snyder, L. B., & Blood, D. J. (1992). Caution: The Surgeon General’s alcohol
warnings and alcohol advertising may have adverse effects on young adults. Journal of

Applied Communication Research, 2, 37-53.

Stiff, J. B. (1994). Persuasive communication. New York: The Guilford Press.

Tabachnick, G. B., & F idell, L. S. (1996). Using multivariate statistics. Harper
Collins College Publishers.

Tenner, E. (1996). Why things bite back: Technology and the revenge of
unintended consequences. New York: Knopf.

Thomas, L. (1983). An epidemic of apprehension. Discover 4, 78-80.

 

Velicer, W. F., DiClemente, C. C., Rossi, J. S., Prochaska J. O. (1990). Relapse
situations, and self-efficacy. Addictive Behavior. 15. 271-283.

Vitaliano, P. P., Russo, J., Carr, J. E., Maiuro, R. D., & Becker, J. (1985). The
ways of coping checklist: Revision and Psychometric properties. Multivariate Behavioral

Research. 20. 3-26.

Wallstone, B. S., & Wallstone, K. A. (1984). Social psychological models of
health behavior: An examination and integration. In A. Baum & J. Singer (Eds),

144

Handbook ostychology and health (Vol. 4, pp. 23-53). Mahwah, NJ: Lawrence
Erlbaum.

Wang, C. (1992). Culture, meaning and disability: Injury prevention campaigns
and the production of stigma. Social Science and Medicine. 35, 1093-1102.

Wartella, E., Heintz, K. E., Aidman, A. J ., & Mazzarella, S. R. (1990). Television
and beyond: Children’s video media in one community. Communication Research 17,
45-64.

 

Weizenbaum, J. (1976). Computer power and human reason. San Francisco: W.
H. Freeman.

Wijkman, A., & Timberlake, L. (1984). Natural disasters: Acts of god or accts of
man? London: Earthscan.

Wildavsky, A. (1977). Doing better and feeling worse: The political pathology of
health policy. Daedalus, 106, 105-25.

Witte, K. (1992a). The role of threat and efficacy in AIDS prevention.
International Quarterly of Communig Health Education, 12, 225-248.

Witte, K. (1992b). Putting the fear back into fear appeals: The extended parallel
process model. Communication Monographs. 59, 330-349.

Witte, K. (1993). Message and conceptual confounds in fear appeals: The role of
threat, fear, and efficacy. Southern Communication Journal 58, 147-155.

 

Witte, K. (1994). Fear control and danger control: A test of the extended parallel
process model (EPPM). Communication Monographs. 61. 113-134.

Witte, K. (1998). Fear as motivator, fear as inhibitor: Using the Extended Parallel
Process Model to explain fear appeal successes and failures. In P.A. Andersen & L.K.
Guerrero (Eds), The Handboolg of communication and emotion: Research. theory.
applications. and contexts mp. 423-450). San Diego, CA: Academic Press.

Witte, K., & Morrison, K. (1995a). Examining the influence of trait
anxiety/repression-sensitization on individuals’ reactions to fear appeals.

Witte, K., & Morrison, K. (1995b). The use of scare tactics in AIDS prevention:
The case of juvenile detention and high school youth. Joumal of Applied
Communication Research 12, 128-142.

 

145

Witte, K. Sampson, J ., Liu, W. Y., & Morrison, K. (19950). Addressing cultural
orientations in fear appeals: Promoting AIDS-protective behaviors among Hispanic
immigrant and African-American adolescents.

Witte, K., Cameron, K. A., McKeon, J. K., & Berkowitz, J. M. (1996). Predicting
risk behaviors: Development and validation of a diagnostic scale. Journal of Heal_th_
Communication, 1, 317-341.

Witte, K., Berkowitz, J. M. Cameron, K. A., & McKeon, J. K. (1998). Preventing
the spread of genital warts: Using fear appeals to promote self-protective behaviors.
Health Education and Behavior.

Witte, K., Peterson, T. R., Vallabhan, S., Stephenson, M. T., Plugge, C. D.,
Givens, V. K., Todd, J. D., Becktold, M. G., Hyde, M. K., & Jarrett, R. (1993).
Preventing tractor-related injuries and deaths in rural populations: Using a persuasive
health message (PHM) framework in formative evaluation research. International
Quarterly of Community Health Education, 13, 219-251.

Zillman, D., & Bryant, J. (1982). Pornography, sexual callousness, and the
trivialization of rape. Journal of Communication. 32, 10-21.

Zillman, D., & Bryant, J. (1984). Effects of massive exposure to pornography. In
N. M. Malamuth & E. Donnerstein (Eds), Pornography and sexual aggression (pp. 115-
141). Orlando: Academic Press.

Zuboff, S. (1984). In the age of the smart machine. New York: Basic Books.

146