 

Illlllllllﬂllllﬂll

LIBRARY

Michigan State
1 University

 

 

 

 

This is to certify that the

dissertation entitled

Persuading Young Women to Obtain
Pap Tests: It's Better to Be Positive
Than Negative

presented by
Steve L. Robbins

has been accepted towards fulﬁllment
of the requirements for

 

 

 

Ph. D. degree in Communication
Major professor
1-23-96

Date

 

MS U i: an Afﬁrmative Action/Equal Opportuniry Institution 0-12771

 

PLACE N RETURN BOX to remove this chockom from your record.
TO AVOID FINES return on or before date duo.

    

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Afﬁrmative Action/Equal Opportunity Intuition
W1

 

PERSUADING YOUNG WOMEN TO OBTAIN PAP TESTS:
IT’S BETTER TO BE POSITIVE THAN NEGATIVE

BY

Steve L. Robbins

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Communication

1 996

ABSTRACT

PERSUADING YOUNG WOMEN TO OBTAIN PAP TESTS:
IT’S BETTER TO BE POSITIVE THAN NEGATIVE

BY

Steve L. Robbins

This study tested the effectiveness of persuasive messages encouraging young
women to get Pap tests based on Ajzen and F ishbein’s (1980) theory of reasoned action.
Speciﬁcally, it compared the effectiveness of positive and negative outcome belief
messages with a general information message. In addition, the study examined the validity
of the theory of planned behavior (TPB) in the context of Pap tests and offered an
additional component to the TPB in predicting intention and behavior. Results of the
study indicate that positive and negative outcome belief messages are no more effective at
inﬂuencing behavioral intention than a general informational message. However, a
positive outcome belief message was signiﬁcantly more effective at inﬂuencing behavior
than a general information message. Moreover, the positive belief message was more
effective at inﬂuencing both intention and behavior than a negative belief message. Results
of the study also lend support to the theory of planned behavior with the addition of a

perceived behavioral importance factor. A revised theory of planned behavior is oﬂ‘ered.

ACKNOWLEDGEMENTS

The author would like to thank a network of individuals for their immeasurable assistance
and support during the preparation for and during this project. First and foremost,
thanks and gratitude go to my Lord and Savior whose love, guidance, and providence
lights the path that on which I continue to follow. Second, to the two most important
women in my life, my mother and my wife. Thanks mom for raising me with an
unconditional love, and giving me the many opportunities which have shaped my life.
Thanks Donna for all the encouragement, understanding, and support during graduate
school. You made it easier and more fun than it would have been otherwise. Third, to my
“Number One” Nicholas. You have helped me again to see what is truly important in my
life — family and friends. Fourth, much gratitude to Chuck Atkin for being a great mentor
and friend. Thanks for giving me the skills and love for a vocation in which I am having
the time of my life. Also, special thanks to Bill Donohue for your valuable advice on and
insight into how to “quack.” To Kelly Morrision and Steve McComack for their
friendship and use of the Morrison & McComack Dwelling Place. Last of all, thanks to

my in-laws, Ed and Connie Lenneman, for your love, support, and beautiful daughter.

TABLE OF CONTENTS

Introduction ....................................................... 1

Review of Literature ......................................... I ....... 5

Method .......................................................... 32
Analyses and Results ............................................... 41
Discussion ............ ‘ ............................................ 58
Conclusion ........................................................ 65
Appendix A ....................................................... 67
Appendix B ....................................................... 68
Appendix C ....................................................... 69
References ........................................................ 70

iv

Introduction

Current statistics from the National Cancer Institute indicate that cervical cancer is the
third mostcommonly occurring cancer among women living in the United States (National
Cancer Institute, 1995). More than 50,000 cases of uterine cervix cancer cases occur
every year, and approximately 6,000 women die from invasive uterine cancer annually
(American Cancer Society, 1990). Moreover, approximately 600,000 women are
diagnosed annually with premalignant changes known as squamous intraepithelial lesion
(SIL) or cervical intraepithelial neoplasia (CIN) (National Cancer Institute, 1995). If these
premalignant conditions are detected early enough, there is an excellent chance of
preventing the onset of invasive uterine cancer. However, if left undetected SIL and CIN
can and often lead to invasive cancer with a progression time of as little as several months
(Bain & Crocker, 1983). Health professionals suggest that prevention and early detection
behaviors can signiﬁcantly reduce the number of deaths caused by cervical cancer
(Healthy People 2000, 1991). Early detection allows doctors to treat premalignant
conditions with a success rate approaching 100% (Healthy people 2000, 1991).

Use of the Pap test remains one of the best tools for detecting cervical cancer and it’s
premalignant forms in their early stages. The Pap test is credited by many for the
recorded decline in cervical cancer mortality in the 19703 and 19805 (Healthy People
2000, 1991). Unfortunately, a signiﬁcant portion of the female US. population is not

aware of or continue to disregard the recommendation to get regular Pap tests (Calle et a1.,

1993; Healthy People 2000, 1991).

While older women are at greatest risk for cervical cancer mortality (National Cancer
Institute, 1995), the group at most risk for never having a Pap test is never-married
women 18 to 39 years old and never—married white women (Calle et a1., 1993). Further
inspection of the data of young women indicates that those at highest risk for never
having a Pap test are those in the population of 18 to 24 year olds. Nationally bout one-
third of the young women in this age category has never had a Pap test, compared to only

10% of women 25 to 39 years of age. ‘

One of the reasons given for the low rate of Pap test behaviors among 18 to 24 year
old women is the belief that Pap tests are not needed until the onset of sexual activity.
While it is true that women should get regular Pap tests once they become sexually active,
the American Cancer Society recommends that even young women who are sexually
inactive should have an annual Pap test beginning at the age of 18 (American Cancer
Society, 1987). In addition, it is recommended that young women have Pap tests
conducted annually unless their doctor advises otherwise (American Cancer Society,
1987, Healthy People 2000, 1991).

The Pap test is a simple procedure that takes a relatively short amount of time to
complete. During the procedure, a doctor or nurse obtains a sample of the layer of tissue
that covers the cervix. Similar to skin, the tissue is constantly rejuvenating itself, and as
new cells grow and multiply, they force older, nearly dead cells toward the surface.

These are the cells sampled by the Pap smear and used to test for any abnormalities.

While this procedure is relatively painless, a number of women report a level of

3

discomfort. Discomfort and other inconveniences associated with the test are relatively
minor compared to the beneﬁts of obtaining annual Pap smears. Regular testing can
identify precancerous conditions which, if caught early, are nearly 100-percent curable
(American Cancer Society, 1991). If a woman does have these pre-cancerous conditions
the longer she goes without a Pap test, which can detect the onset of cancer, the more
serious and less treatable the cancer becomes. It is clear then that regular, annual Pap
tests are an important part of the preventive health behaviors of adult women.

In light of the great and potentially life-saving beneﬁts offered by Pap tests, and the
evidence which suggests that a large portion of women, especially the white, young, and
never-married, are not adopting Pap test behaviors, it is appropriate to develop
interventions and communication campaigns targeting this population for adopting Pap
test behaviors. Because there is some evidence that past behaviors are positively
correlated with future behaviors (Bentler & Speckart, 1981; McGuire, 1976), it seems
prudent to get young women to engage in Pap test behaviors so that when they are older
and at higher risk for cervical cancer, the probability will be greater that they will indeed
take preventive action. The question, then, is not if there should be an intervention to

increase Pap test behavior, but of what should the intervention consist?

The Present Study

Most health researchers and professionals are likely to agree that any health
communication intervention to change behaviors, beliefs, attitudes, etc., should be driven

by theory. Of the many theories which exist for explaining and predicting health

4

behaviors one which has received a lot of attention and empirical support is Ajzen and
Fishbein's (1980) theory of reasoned action (TRA) and an extension of that theory,
Ajzen's (1985) theory of planned behavior (TPB). The present study tests the latter
theoretical framework for its predictive validity in the behavioral domain of Pap test
intentions. It also investigates the effectiveness of persuasive messages based on the
theory of planned behavior in inﬂuencing beliefs, attitudes, intentions, and behaviors with
respect to Pap smear tests. First, this paper discusses the relationship between attitude
and behavior. It then describes each of the two relevant theories (TRA and TPB) and
provides a literature review of studies applying them. Third, after presenting the two
previous studies utilizing the theory of reasoned action to explain Pap test behaviors, this
paper offers another component (perceived behavioral importance) as an extension to the
theory of planned behavior. This new factor is hypothesized to enhance the predictive
utility of the theory in explaining young women's intentions to obtain Pap tests.

Multiple regression and path analytic techniques are used to test the hypothesized model
(the extension of the TPB). In addition, several hypotheses are offered based on previous
ﬁndings relevant to Pap test behaviors. Various other analytic techniques, discussed later,
are used to test the hypotheses and assess the effectiveness of TPB-based persuasive
messages in inﬂuencing the underlying beliefs associated with each of the exogenous belief
components of the theory of planned behavior. Methods and procedures of the study are
then described, and ﬁnally, ﬁndings are presented along with a discussion of the ﬁndings
and their implications for future research. We ﬁrst take up the controversial relationship

between attitude and behavior.

5
The Attitude-Behavior Relationship

The relationship between attitude and behavior has been and continues to be a matter of
great discussion for persuasion and social inﬂuence researchers (Eagley & Chaiken, 1993).
Early studies investigating the effect attitudes have on behavior suggested that attitudinal
dispositions could be used to help explain human behavior (Ajzen & Fishbein, 1980).
That is, one should be able to predict another’s behavior by knowing their attitude toward
the behavior. Indeed, this assumed relationship can be found in or inferred from a number
of different deﬁnitions of attitude. For example, attitudes have been deﬁned as
predispositions to respond (Osgood, Tusci, & Tannenbaum, 1957), dispositions to react
(Sarnoff, 1960), and as having an energizing effect on behavior (Allport, 1935). However,
the results of early attitudinal studies questioned the validity of an attitude-behavior
relationship.

One of these early studies was LaPiere's (1934) classic examination of racial prejudice.
In his investigation, LaPierre accompanied a Chinese couple to a number of business
establishments in the United States to ﬁnd out howmany would refuse service to the
couple. He found that of the 251 establishments they visited, only one refused service.
About six months later, each of the 251 establishments was sent letters asking if it would
accept members of the Chinese race as business patrons. Amazingly, 90% of the 128
establishments responding to the question said they would not. This ﬁnding suggested
that attitudes are not necessarily associated with behavior, and that the relationship

between the two variables is tenuous at best.

Another study by Cory (1937) examining the relationship between students’ attitudes

6

toward cheating and their actual cheating behavior also found that attitudes were not
strongly related to behavior. Cory’s investigation resulted in a r=.02 correlation between
attitude and behavior. Later empirical studies examining the attitude-behavior relationship
continued to ﬁnd a weak relationship between the two variables (Ehlrich, 1969; McGuire,
1969; Wicker, 1969). In 1955, prominent sociologist Herbert Blumer publicly criticized
and challenged the assumption that attitudes inﬂuence behavior. A few years later another
sociologist, Irwin Deutscher (1973), published a harsh critique of the attitude-behavior
relationship by providing past studies which were empirically weak in demonstrating a

causal relationship between attitudes and behavior.

The attack against an attitude-behavior relationship hit full steam in the 19603. It was
then that psychologist Alan Wicker (1969) published an article in which he wrote that
there is ”little evidence to support the postulated existence of stable underlying attitudes
within the individual which inﬂuence ...his actions" (p. 75). He based the his critique on
an extensive literature review of empirical studies testing a hypothesized relationship
between attitudes and behavior. In his review, he found the average correlation between
attitude and behavior to be .15 with a high correlation of .30. Based on this evidence,
Wicker concluded that, “taken as a whole, these studies suggest that it is considerably
more likely that attitude will be unrelated or only slightly related to overt behaviors than
attitudes will be closely related to actions (Wicker, 1969, p. 65). Abelson (1972) similarly
concluded that there was not enough evidence to support the hypothesis that attitudes

impact human actions.

Though the evidence was building against the once-assumed attitude-behavior

7

relationship, a number of theorists was unwilling to accept that such a relationship did
not exist. They ﬁrmly believed that an actor's attitudes had to have some underlying, if
not obvious, relationship with his/her behavioral actions. Some of these scholars,
examining other work in the area, concluded that the “no- relationship” position was too
pessimistic, and unfounded. Dillehay (1973) argued that some of these earlier studies that
found null or weak attitude-behavior results suffered from methodological shortcomings.
He noted in the LaPierre (1934) study that it was likely that different sets of people
responded to the actual face-to-face request for service and the mail survey which
followed six months later. That is, there is reason to suspect that desk clerks and host
staff handled the face-to-face service requests while hotel and restaurant managers
responded to the mail survey. Dillehay argued that the conclusion of Lapierre (i.e, no
relationship between attitude and behavior) is misguided since subjects who were part of

the ﬁrst wave of data collection were not the same as those in the second wave.

Other scholars (Ajzen & Fishbein, 1977) have noted at least two other shortcomings
in the early attitude-behavior studies. First, Ajzen and F ishbein (1977, 1980) have
indicated that a weak relationship between attitude and behavior should not be surprising
when the measures of attitude and behavior are at different levels of speciﬁcity. For
example, discovering another’s general attitude toward higher education is not a good
predictor of that person’s behavior in taking or not taking a college biology class. To
increase the probability of predicting that particular behavior, one must assess that
person’s attitude toward taking the biology class. Second, it is many times the case that
behavioral measures and their corresponding attitudinal measures are generally single acts

chosen “on an intuitive and arbitrary basis” (Fishbein & Ajzen, 1974, p. 65). That is,

8

behavior relevant to some attitude object may be affected by environmental factors rather
than by attributes associated with the attitude object. However, Fishbein and Ajzen
suggest that if one observes a person’s behavior toward an attitude object across myriad
settings, one would see that the person behaves similarly toward the attitude object
because the object itself would be inﬂuencing the measure of behavior, and not the
peculiarities of the settings (Cacioppo, Hawkins, & Petty, 1981).

A review of the attitude-behavior literature by Ajzen and Fishbein (1977) provided a
number of studies in which the previously discussed methodological shortcomings were
present. In those studies the relationship between attitude and behavior were non-
signiﬁcant. On the other hand, in studies which employed appropriate measures,
statistically signiﬁcant relationships between attitude and behavior were found each time
(26 out of 26 studies). Fishbein and Ajzen’s literature review was instrumental in turning

the tide toward a relationship between attitude and behavior (Cacioppo et. al., 1981).

While the work of Fishbein and Ajzen put more faith in an attitude-behavior
relationship, it did not necessarily explain away all the weak and null results.
Subsequently, a number of scholars believed and subsequently concluded that the
relationship between attitude and behavior must be mediated or moderated by some
unaccounted for variables (Ajzen & Fishbein, 1980). Such effects of other variables
would provide evidence that the concept of attitude is associated with behavior and is
important to the study of human behavior. Moreover, it would help to explain the
ﬁndings which suggested a weak or no relationship between attitude and behavior. These

researchers set out to ﬁnd those variables and situations that would provide evidence for

9

their belief that attitudes are associated with, and could predict behaviors across a wide

variety of domains.

One line of research which has examined the attitude-behavior discrepancy has framed
the problem as one of measurement (Fishbein & Ajzen, 1974; Weigel and Newman,
1976). As previously discussed, Ajzen and Fishbein (1980; Ajzen, 1975) argue that in
order to accurately predict behavior from attitude, speciﬁc measures of attitude need to be
taken with respect to the behavioral criterion measure. Put differently, the accuracy in
predicting speciﬁc behaviors from attitudes will be greatly enhanced when the measure of
attitude is speciﬁed at the same level as the behavior in question. For example, in order to
use attitudes to predict whether a woman will get a Pap test within a speciﬁed period of
time, one must measure the woman’s speciﬁc attitudes toward that particular behavior,
not her general attitude toward Pap tests. Aj zen and Fishbein argue it is not adequate
enough to measure general attitudes toward Pap tests. A global or general measure of
attitude toward Pap tests will not do a good job of predicting the speciﬁc behavior of
getting a Pap test within a particular time frame. However, it has been shown that even

‘when speciﬁc measures of attitude are utilized, the relationship between attitudes and
behavior may still be weak (Jaccard et a1., 1977). Such ﬁndings have led investigators to
develop theories which include other variables that are hypothesized to predict behavior.

In their theory of reasoned action, Ajzen and Fishbein (1980) argue that actors'
intentions mediate the effect of attitude on behavior. They deﬁne intention as a
subjective probability to act (Ajzen & Fishbein, 1980). They claim that one’s "intention

to perform (or not perform) a behavior as the immediate determinant of the action" (p. 5).

10

In fact, they state that it is relatively easy to predict another’s behavior if one has
knowledge of that person’s intention to perform the behavior. However, they caution that
there is not always a one-to-one correspondence between intention and behavior, but
"barring unseen events, a person will usually act in accordance with his or her intention"
(p. 5). While it is at some level helpful to know that intention has a causal relationship
with behavior, Ajzen and F ishbein (1 980) admit that the notion that intention is a good
predictor of behavior does not give much insight to the reasons why people act. It is of
more value to be able to explain behavior. Accordingly, they posit a cognitive theory for
both predicting and explaining human behavior. That theory, the theory of reasoned

action, is taken up in the next section.

Theory of Reasoned Action

A number of different theories has been used to try to explain human behaviors. When it
comes to the adoption of health behaviors, one of the most applied theoretical
frameworks is the theory of reasoned action (Ajzen, 1975; Ajzen & Fishbein, 1980). In
general, the theory has been empirically supported across a wide variety of behavioral
domains (Ajzen, 1991; Sheppard, 1988). The theory posits that behavior is directly
inﬂuenced by an individual's intention to perform or not perform the behavior in question.
As the immediate and only direct determinant of behavior in the theory, Ajzen and
Fishbein suggest that, "behaviors are not really difﬁcult to predict" (p. 5) given knowledge
of an actor's intention. However, they also state that while being able to predict behavior

is useful, it is also valuable to understand why people behave. Therefore, their model

11

includes other variables which help to determine behavioral intention.

According to Ajzen and Fishbein, behavioral intention is a function of an individual's
beliefs about the behavior in question. These beliefs have been divided into two
conceptually distinct groups: behavioral and normative (Fishbein & Ajzen, 1975).
Behavioral beliefs (i.e. outcome beliefs) about outcomes are said to inﬂuence attitudes
toward the behavior while normative beliefs (i.e., beliefs about what referent others
believe about an object) are hypothesized to impact general subjective norms.
Speciﬁcally, attitude about the behavior is operationalized to be equivalent to the sum of
the products of salient beliefs about outcomes that would result if the behavior were to be
enacted (outcome beliefs or expectancies) and the value the individual places on each of

the outcomes (outcome evaluation or value). It is commonly expressed as:

11
Abeh = .11 biei
1:

where A be h is attitude toward the behavior, b is the belief that performing the behavior be h

leads to a consequence or outcome i; e is the individual's evaluation of the outcome i; and

n is the number of salient beliefs the individual holds about performing the behavior be h.

In the same manner, subjective norm is a function of salient beliefs about what

important individuals or reference groups think about what the individual should or

should not do (normative beliefs) multiplied by the actor's motivation to comply with

12

those important others. Subjective norm is algebraically expressed in this manner:

SN: I NBOMCi
. 1 1

l:

where SN stands for subjective norm; NB represents normative belief; MC is the
individual's motivation to comply with referent i; n is the number of relevant referents.
The relative importance of the attitudinal and normative components in inﬂuencing
behavioral intention is expected to vary across behavioral domains (Eagley & Chaiken,
1993; Kurland, 1995; Morrison, Gillmore, & Baker,1995). That is, for one particular
behavior the attitudinal component may account for more of the variance in behavioral
intention than the subjective norm component, and in another behavioral domain,
subjective norm may inﬂuence behavioral intention more so that attitude. For example, in
a study of ethical behaviors of insurance agents Kurland (1995) found that attitude
accounted for more variance in intention than subjective norm. On the other hand,
Morrison et a1. (1995) found that subjective norm was more important than attitude in
determining college students intentions to use condoms with casual partners. In terms of
producing persuasive messages, it is theoretically fruitful to know which component is
more heavily weighted as a determinant of intention so that any persuasive message

developed can target the more salient and important beliefs.

Because of the way the attitudinal and normative components of the theory of

reasoned action are operationalized, the theory of reasoned action has also been placed

13

under the class of theories known as expectancy-value theories. According to the theory
of reasoned action, any other variables external to the model only inﬂuence intention
through the mediating effect of attitude and subjective norm. However, Ajzen and

F ishbein (1980) state that they are open to the inclusion of other predictor variables if

they can be shown to signiﬁcantly enhance the predictive utility of the theory.

Ajzen and Fishbein (1980) suggest that one can target an individual's outcome and
normative beliefs with persuasive messages to increase the individual's behavioral
intention and, ultimately, the probability that actual behavior will occur. Speciﬁcally,
they suggest that persuasive messages should challenge, de-emphasize, and/or
counterargue any beliefs that might inhibit the performance of the recommended behavior
and support and maximize any beliefs which are positively associated with the behavior.
Several applications of the theory of reasoned action provide empirical support for
theory of reasoned action-based persuasive messages (i.e., messages with emphasize
salient positive outcome beliefs and de-emphasize salient negative outcome beliefs)
(Brubaker & Fowler, 1990; Fishbein et a1., 1980; Hoogstraten, 1985). However, in two of
the studies, one investigating testicular self-examination (Brubaker & Fowler, 1990) and
the other examining dental care behaviors (Hoogstraten, 1985), researchers found that
while theory of reasoned action-based messages (i.e., messages which target salient
negative and positive outcome beliefs) are useful in affecting beliefs and ultimately
behavior, they also found that non-theory based informational messages were just as
effective. These ﬁndings challenge Ajzen and Fishbein's (1980) claim that persuasive
messages that target salient beliefs are superior to messages which do not. In sum, the

few studies which have tested the claim that TRA-based persuasive messages are better

14

than messages which do not target salient outcome beliefs have resulted in conﬂicting
results. One goal of the present research is to test the effectiveness of a TRA-based

persuasive message in the domain of Pap test behavior.

Boundary Conditions of the TRA

According to F ishbein and Ajzen (1975; Ajzen & Fishbein, 1980), the magnitude of the
relationship between behavioral intention and actual behavior can be affected by three
conditions. First, correlations between intention and behavior can be impacted by the
extent to which the behavior in question is under the actor's volitional control. The theory
of reasoned action does not consider efﬁcacy issues directly. There is a number of
situations in which one may have strong intentions to perform an act, but is unable to do
so because of an obstacle or barrier. For example, young women may have a high level of
intent to get a Pap test, but they may ﬁnd the cost of getting one prohibitive. Second, the
relationship between intention and behavior can be affected by the extent to which the
measures of behavioral intention and actual behavior are at the same level of speciﬁcity.
For example, if the behavior one is exploring is that of getting a Pap test within a one
week time span, then measures of intent should also specify a one week time span. Third
the change in intentions between the measure of behavioral intention and the measure of
actual behavior. Violations of these boundary conditions negatively affect the predictive
utility of the theory. They caution researchers utilizing the theory. to be mindful of these

boundary conditions in their research investigations.

The many applications of the theory of reasoned action have yielded generally

15

supportive results when the theory's boundary conditions are met (Sheppard et a1., 1988)
. It has been used to predict and explain a variety of health behaviors including breast
self-examination (Hill et a1., 1985), testicular self-examination (Brubaker & Fowler, 1990;
Brubaker & Wickerman; 1990; Steffan, 1990), Pap test behavior (Hill et al. 1985; Hennig
& Knowles, 1990), alcohol use (Fishbein et a1., 1980; London, 1982; McCarty et a1.,
1983), seat-belt use (Budd et a1., 1984), and dental care and ﬂossing (Hoogstraten et a1.,

1985; Toneatto & Binik, 1987).

However, it is obvious that not all health behaviors are under the volitional control of
the actor. That is, there is a variety of health behaviors which have barriers that an actor
may have difﬁculty in overcoming, or may not be able to overcome without the use of
resources, skills, abilities, etc., not immediately available to him/her. For example, the
performance of getting a Pap smear test is contingent upon such things as knowledge
about when Pap tests are needed, money to pay for the test, access to a doctor or health
site, and time constraints. In these situations where actors have less than full control of
behavioral performance, it can be seen how positive attitudes and subjective norm
potentially have little effect on intention and behavior if the actor believes that he/she
cannot perform or has difficulty in performing the behavior. Critics of the theory of
reasoned action claim that the theory is not sufﬁcient in these situations where the actor
does not have full control over whether he/she performs the behavior (Sheppard et a1.,

1988).

16

Theory of Planned Behavior

In response to these critics Ajzen (1985) extended the theory of reasoned action to
include a factor that takes into account behaviors which are not completely under the
control of the actor. His theory of planned behavior includes perceived behavioral control

as another factor that potentially inﬂuences behavioral intention.

Perceived behavioral control is deﬁned as an individual's perceptions about his/her
ability in performing the behavior in question. Ajzen (1985) suggests that perceived
behavioral control is conceptually similar to Bandura's concept of self-efﬁcacy (Bandura,
1977; Bandura, 1982). According to Bandura (1982, p. 122) the concept of self-efﬁcacy
"is concerned with judgments of how well one can execute courses of action required to
deal with prospective situations." Previous studies provide support for the inclusion of a
self-efﬁcacy component in models designed to predict health-adoption behaviors (Beck &
Ajzen , 1991; Devellis, Blalock, & Sandler, 1990; Hill, Gardner & Rassaby, 1985). The
data from those studies suggest that an individual's ability to perform a behavior is
strongly inﬂuenced by his/her assessment of the requisite resources and abilities needed
(Bandura, Adams, Hardy, and Howells 1980; Beck & Ajzen, 1991; Brubaker &
Wickersharn; 1990).

The theory of planned behavior, like it's predecessor, has received a good deal of
empirical support (Ajzen, 1991; Madden, 1992). Ajzen and Madden (1986) were one of
the ﬁrst to offer a complete test of the theory. They used the theory to examine college
students' class attendance (study 1) and predictions of getting an "A" in a class (study 2).

The results of the study suggest that perceived behavioral control accounted for a

17

signiﬁcant portion of the variance in intention beyond that accounted for by attitude and
subjective norm. However, perceived behavioral control did not add signiﬁcantly to the
prediction of behavior when behavioral intention was controlled. This non-signiﬁcant
ﬁnding, they suggest, can be accounted for by the relatively high degree of control
students have over attending class. They argue that under conditions where actors have a

high degree of volitional control, the theory of reasoned action is sufﬁcient in predicting

behavior.

In their second study the researchers investigated the behavior of getting an "A" in a
college course. Data measuring students' attitudes, subjective norms, perceived behavioral
control and behavioral intention were collected at two different points of time; one at the
beginning of the course and one at the end. The students' grades in the course were used
as the measure of behavior. Analyses of the data from the ﬁrst wave yielded similar
results to their ﬁrst study. That is, perceived behavioral control added signiﬁcantly to the
prediction of intention, but added little to the prediction of actual behavior. Analyses of
the data collected from the second wave, however, were consistent with the predictions
offered by the theory of planned behavior. That is, perceived behavioral control enhanced
the prediction of behavior even after controlling for intention. They explained this result
by arguing that students' perceptions of control concerning getting an "A" in the course at
the beginning of the semester were inﬂated and unrealistic. More realistic control
perceptions were given at the end of the semester when students' had more information to
use in assessing their ability to get an "A." They conclude that perceived behavioral
control only has predictive utility with respect to behavior when perceptions of control

are accurate.

18

In another test of the theory of planned behavior Schiﬁer and Ajzen (1985)
investigated the behavior of weight loss. They found that college women's assessment of
control over losing weight had a signiﬁcant contribution on intentions to lose weight and a
moderate effect on actual weight loss. They suggest that the moderate effect of perceived
control on weight loss may be due to the unaccounted for effects of other factors. In sum,
applications of the theory of planned behavior have, in general, been supportive of the
model's validity. With respect to the domain of health, the theory has yielded favorable
results across a variety of health behaviors including weight loss (Schiﬁer & Ajzen, 1985),
condom use (Morrison et a1., 1995), exercise (Godin, 1990; Godin et a1., 1989), and cancer

screening (Devellis et. Al., 1990).

Applications to Pap Smear Tests

One behavior which has not been investigated with the theory of planned behavior is that
of getting a Pap Smear test. As noted earlier, Pap tests are strongly recommended for
early detection of cervical cancer. A literature review by this researcher shows that Pap

‘ test behaviors have been examined with the theory of reasoned action no less than two
times. Hill, Gardner, and Rassaby (1985) found support for the theory of reasoned
action in predicting intentions to obtain Pap tests and perform breast self-examinations
(the remaining discussion will only refer to Pap tests). In their study they tested three
models commonly used to explain the adoption of health behaviors (i.e., theory of
reasoned action, health belief model, and subjective probability model). Data measuring

components of each of the three theoretical frameworks were collected from N=123

19

Australian women ranging in age from 18 to 70 (mean years =34). A behavioral measure
was not taken. Their results indicated that each model had predictive utility with respect
to intentions to perform Pap tests. However, none of the theories accounted for large
amounts of variance in Pap test intentions. The two theories that accounted for the most
variance, the theory of reasoned action and the health belief model, accounted for 26%
and 30% of the variance, respectively. In choosing the "best" model the authors opted for
the theory of reasoned action because it is the most parsimonious of the two models.

However, their analyses also found that the barriers component of the health belief
model accounted for a signiﬁcant portion of the variance in addition to the variance
accounted for by attitude and subjective norm. Since the notion of barriers is highly
associated with the concept of self-efﬁcacy and perceived behavioral control, this ﬁnding
suggests that the component of perceived behavioral control might enhance the predictive
power of the theory of reasoned action in explaining intention and behavior. In fact, Hill
et al. (1985) conclude that a composite model consisting of the theory of reasoned action
and the barriers component (i.e., theory of planned behavior) might be best suited to
predict women's intentions to get a Pap test. In addition, their ﬁndings suggest that '
persuasive messages targeting older women to get Pap tests should maximize the positive
behavioral outcome beliefs of l) ﬁnds cancer in the early stages, and 2) give a sense of
relief to ﬁnd nothing amiss, and minimize the negative beliefs of 1) involves

embarrassment, and 2) is physically unpleasant.

In a study similar to the Hill et a1. study, Hennig and Knowles (1990) examined

intentions to get Pap tests among a sample of N=144 Australian women over 40 years old

20

(mean years = 54). They too found that attitude and subjective norm signiﬁcantly
inﬂuenced Pap test intentions, but the two components explained only 12% of the
variance in intention. One possible explanation for the small effect on the variance in
behavioral intention is that these researchers incorrectly obtained the measure of attitude.
These researchers operationalized attitude by summing up belief outcomes and belief
evaluations before multiplying the two to come up with attitude. This method of creating
the attitude measure has been cautioned against (Ajzen & F ishbein, 1980). The correct
method is multiplying each of the outcome beliefs with its corresponding belief evaluation
and then summing up the products. Hennig and Knowles methodological mistake
conceivably introduced non-random error into their study and produced the low variance
ﬁnding in behavioral intention.

While the Hill et al. (1985) and Hennig and Knowles (1990) studies lend general
support to the theory of reasoned action, their results with respect to the amount of
variance in intention accounted for by the theory's components was relatively low. Such
ﬁndings suggest one or a combination of reasons which may account for the results. First,
one or more of the boundary conditions speciﬁed for the theory of reasoned action may
have been violated. In the case of these two studies, it is possible that the subjects had
limited control over getting a Pap test. To the extent that this is true, Pap test behaviors
are not completely volitional and thus outside the scope of the theory of reasoned action.
Second, it is likely that other factors have a direct effect on intention which were not
accounted for in either of the studies. For example, the relationship between intention
and subjects’ attitudes and subjective norm with respect to getting a Pap test within two

years (target behavior in both studies) may be mediated or moderated by the subject’s

21

assessment of the importance of the behavior. If one deems a behavior to be unimportant,
while still holding positive attitudes toward the behavior and believing others think he/she
should do it, there is a strong possibility that he/she will not enact the behavior. This idea
of behavioral importance or salience is similar to the concept of attitude importance.

These two concepts are taken up in the next section.

Attitude Importance

In the social science literature the concept of attitude importance has been thought of as a
motivational concept that inﬂuences perception and behavior (Boninger, Krosmck &
Berent, 1995, Eagley & Chaiken, 1993). The literature on attitude strength does not
contain a strong formal deﬁnition of the concept, nor does it provide any well-established
operationalizations (Boninger et. a1., 1995). Subsequently, conceptualizations of attitude
importance include ideas of extremity (Tannenbaum, 1956), investment (Krosnick, 1990),
caring (Krosnick, 1989), emotional commitment (Abelson, 1988), signiﬁcance (Boninger,
et. a1., 1995), and accessibility (F azio, 1986). Deﬁnitions of attitude importance include,
“the degree to which a person is passionately concerned about and personally invested in
an attitude” (Krosnick 1990, p. 60) and “an individual’s subjective sense of the concern,
caring, and signiﬁcance he or she attaches to an attitude” (Boninger et. a1., 1995, p. 62).
These conceptualizations and deﬁnitions suggest that when a person attaches a high level
of personal importance to an attitude, he/she deeply and passionately cares about it and is
committed to it. Boninger et. a1. (1995, p. 62) argue that attitude importance is “a belief

that links an attitude to an attribute.”

22

Many of the ideas used in discussing attitude importance can be carried over to the
idea of behavioral importance. In this light, behavioral importance can be thought of as
the degree to which one cares about or is committed to a behavior. If an individual
attaches great personal importance to a behavior then that individual views the behavior
as personally salient, is passionate towards it, and is more likely to engage in the behavior
than not. For example, those who believe that it is very important to wear a seat-belt
every time they are in an automobile are likely to wear seat-belts and encourage others to
wear seat-belts. Their attachment of importance to that particular behavior (seat-belt use)
is expected to guide their seat-belt relevant behaviors. Like attitude importance,
behavioral importance is conceptualized to be a belief and is deﬁned as an individual’s
perception of the value, signiﬁcance, and relevance of a behavior. Behavioral importance,

like intention, is hypothesized to inﬂuence one’s intention to act.

Within the realm of expectancy-value theories, and more speciﬁcally the theory of
reasoned action, the question may arise, “How is the concept of behavioral importance
different than the concept of attitude?”While behavioral importance may be similar to the
concept of attitude, insofar as they both are hypothesized to inﬂuence intention,
conceptually they are argued to be quite different. Behavioral importance, as a belief, is
generally cognitive in nature while attitude has generally been thought of as more affective
than cognitive (Eagley & Chaiken, 1993). In order to obtain a clearer distinction between

the concepts it may be useful to further examine and deﬁne the concept of attitude.

23

The Concept of Attitude

The English word attitude comes from the Latin aptus which means “ﬁtness” or
“adaptedness.” Original uses of the term primarily referred to posture or bodily position
(e.g., the way one sits is sometimes perceived as indicative of one’s “attitude” at the
time). In his Expressions of Emotions in Man and Animals, Charles Darwin utilized
attitude as a motor concept, or the physical expression of an emotion (see Fleming, 1967).
Darwin viewed an attitude as a biological state of readiness to respond. Experimental
psychologist L. Lange referred to a “task-attitude” as a musculature preparation to
respond. He offered this behavioral conception of attitude following an 1888 reaction-
time experiment in which he found that subjects mentally prepared to press a telegraph
key in response to a signal did so more quickly than subjects who focused on the signal
rather than the response to the signal. In the same vein, English neurophysiologist Charles
Sherrington in 1906 referred to attitude as a person’s normal pose or posture, not as an
occasional indicant of a strong emotion or a response to a certain task set (see Fleming,
1967). While Darwin, Lange, and Sherrington originally conceptualized attitude as motor
states it was not long before the concept of attitude was viewed as a mental of cognitive

construct.

It was in 1918 that sociologist William 1. Thomas and poet-philosopher F lorain
Znaniecki published The Polish Peasant in Europe and America. This landmark volume in
social research examined the problems Polish immigrants faced in coming to the United

States. Much of what Thomas and Znaniecki presented centered around the concept of

attitude. In their work, the authors referred to attitude as a wholly affective and evaluative

24

concept. For them, an attitude was nothing short of a feeling(s) toward some object.
Attitudes were such things as “love of children,” “hatred of criminals,” and “respect for
science.” Thomas and Znaniecki’s statement of attitudes was historically important
because, for the ﬁrst time, attitudes had been separated from its physiological content
(see Fleming, 1967).

This more cognitive-based view of attitudes received greater support in the 1930’s
due to the inﬂuential work of such neo-behaviorists as Hull, Tolman, and Skinner whose
works divorced attitude from physiology (Petty, Ostrom, & Brock, 1981) . As more and
more researchers began examining the similarities between attitude and psychophysical
judgments it became commonplace to refer to attitudes as mostly, if not purely, cognitive.
By World War II this cognitive understanding of attitude was ﬁrmly entrenched in the
vocabularies of both academics and lay persons. However, while it was understood that

attitude was a cognitive concept, there was less agreement on how to deﬁne attitude.

A number of different deﬁnitions of attitude has been given in the past. Attitude has
been referred to as a psychological tendency with some degree of favor or disfavor
(Eagley & Chaiken, 1992), a mental state of readiness exerting inﬂuence upon an
individual's response to an object (Allport, 1935), an enduring organization of
motivational, emotional, perceptual, and cognitive processes with respect to some aspect
of the individual's world (Krech & Crutchﬁeld, 1948), an object evaluation that is stored
in memory (Judd, Drake, Downing & Krosnick), the affect for or against a psychological
object (Thurstone, 1931), a learned disposition to respond favorably or unfavorably with

respect to an object (Fishbein & Ajzen, 1975), and a person's evaluation of any

25

psychological object (Ajzen & Fishbein, 1980). The many varied deﬁnitions of attitude
make it difﬁcult to ﬁnd commonalities that weave through all. However, most
Contemporary attitude researchers would agree that an attitude is a learned, enduring, and
aﬂective evaluation of an object which has some degree of causal impact on behavior
(Eagley & Chaiken, 1993; Perloff, 1993). We next take up a discussion of these

components of attitude.

Attitudes are widely accepted to be more a product of learning rather than the result
of genetic forces (Eagley & Chaiken, 1993; Perloff, 1993; Petty, Ostrom, & Brock, 1981).
It is through the course of socialization that we learn the attitudes which we hold. A
number of different factors (e.g., family, friends, media) play an integral role in the
formation of our attitudes, and it would be very difﬁcult to single out any one factor as
the determinate of any of the attitudes we hold. While some investigators (McGuire,
1985; Schacter 1982) hold that genetic forces inﬂuence us to develop some attitudes or
direct us to behave in a particular manner toward attitude objects, there is little evidence
to support a genetic view of attitude formation (Eagley & Chaiken, 199; Perloff, 1993).

It is generally understood that attitudes are enduring mental dispositions rather than
ﬂeeting, moment-to-moment occurrences (Perloff, 1993). Unlike moods, which may
covary with the environment in which we ﬁnd ourselves, attitudes are said to be stable
formations that affect our thoughts and behaviors. Attitudes are unlikely to change

because we’ve had a bad day or because someone yelled at us. Rather, attitudes tend to be
enduring and, in many cases, very resistant to change.

Lastly, attitudes are said to have a strong affective dimension. Thurstone (1931), an

26

inﬂuential scholar in the area of attitude measurement, conceptualized attitude as the
amount of affect for or against an attitude object. Subsequent conceptualizations of the
attitude concept offered other dimensions beyond the affective one (see Scott, 1968;
Smith, Bruner & White, 1956). The properties of these other dimensions suggest three
classes of responses: 1) aﬂective -- one’s feelings about or emotions toward some object,
2) cognitive -- one’s thoughts, ideas, associations, and images relevant to some object, and
3) conative (i.e., behavioral) -- one’s behavioral response to some object. It has been
suggested (Ostrom, 1969) that the three response types can be used to assess one’s
attitude toward a designated stimulus object. Accordingly, aﬂect has to do with the
positive or negative feelings toward an object; cognition is concerned with the positive or
negative attributes one associates with a stimulus object; and conation deals with the
positive or negative behaviors which one displays as a response to a stimulus object.
While Ostrom (1969) suggested that all three responses could be used as indicants of an

attitude, they are said to be conceptually distinct and independent of one another.

The tri-partite View of attitude has been used by a number of researchers. Some have
used two or all three types of responses in creating their deﬁnitions of attitude (Allport,
1935; Katz & Stotland. 1959; Rosenberg & Hovland, 1960). However, there has been a
move to return to Thurstone’s conceptualization of attitude as mainly affective in nature
(Cacioppo, Hawkins, & Petty, 1981, Perloff, 1993). As examples, Bem (1970) deﬁnes
attitudes as likes and dislikes; Collins (1970) suggests an attitude is a feeling of good or
bad, fair or unfair toward an attitude object; Insko and Schopler (1972) present attitudes
as dispositions to favorably or unfavorably evaluate attitude objects; and Fishbein &

Ajzen (1975, p. l 1) say that, “the major characteristic that distinguishes attitude from

27

other concepts is its evaluative or affective nature.” By restricting the deﬁnition of
attitude to mainly being an affective evaluation, one is able to distinguish feelings (i.e., “I
hate Pap smears”) which constitute attitude, from cognition (i.e., “Pap smears are
important”) and conations (i.e., “ I rarely get Pap smears”). While we may deﬁne
attitude with mainly an affective component, it is not to say that the cognitive and
behavioral components should be dismissed. Rather, as stated previously, cognitions and
conations are given distinct and conceptually independent recognition.

To the extent that one utilizes Fishbein and Ajzen (1975; Ajzen and Fishbein's,1980)
deﬁnition of attitude as an affective evaluation of an object and the way in which they
operationalize attitude (e. g., expectancy-value), then one can make a claim that an attitude
toward some particular object, in the present case, a behavior, is an affective disposition
toward that behavior. This conceptualization does not fully represent the degree of value,
importance, or signiﬁcance one may place on an attitude object. That is, the level of
importance one places on an attitude object is different than one's affective disposition
toward it. This can be seen in a case in which an individual has very positive attitudes
toward a behavior (e.g., watching a particular television program), but places little
importance on watching that program. Accurately predicting intentions and behavior
based only on the actor's positive attitudes is tenuous at best. Conversely, an individual
may believe exercising three times a week is very important, but have negative attitudes
about exercising three times a week. It can be seen that holding a favorable or unfavorable
evaluation of a behavioral action is not the same as believing that performing the behavior
is important. Of course, one might expect the evaluation (affective) of a behavioral action

to be correlated with the importance (cognitive) placed on that action. However, there is

28

at least a conceptual distinction between the two dimensions which should be considered
in examining behaviors within the class of expectancy-value theories (e.g., theory of
reasoned action, theory of planned behavior). In fact, a study by Budd (1986) suggests
that a measure of belief salience can improve the predictive power of the theory of
reasoned action. While belief salience as Budd operationalizes it (i.e., measuring only the
ﬁve most important or accessible beliefs) is not the same as the factor of perceived
behavioral importance, it suggests that a measure of how important a behavior is to a
person is potentially useful in predicting if the person intends to actually perform the

behavior.

In the case of Pap test behavior, and health behaviors in general, a measure of
behavioral importance has the potential to exert an independent inﬂuence on behavioral
intention and/or moderate the effects of other variables. For example, take the case in
which a woman has positive attitudes about getting a Pap test, but does not feel it
important to get one. Some young women may know the beneﬁts of getting Pap tests
(i.e. hold positive attitudes), but may be misinformed and believe that they do not need to
get Pap tests until they are older and more susceptible to cervical cancer (i.e., low
behavioral importance). In this case, one might predict that these women are less likely to
intend to get a Pap test. In the case of older women, like those in the in the Hennig and
Knowles study, they might believe that getting regular Pap tests is not necessary if they
are no longer sexually active (i.e., not important), but still have positive attitudes toward
Pap tests. It is argued that knowing the value (i.e., perceived behavioral importance) an

actor places on a behavior enhances the prediction of intentions.

29
Hypotheses

One goal of the present research is test the theory of planned behavior as a model for
predicting women’s behaviors with respect to getting a Pap test. Past studies in the
behavioral domain of Pap tests have not formally tested the theory of planned behavior
(Hennig & Knowles, 1990; Hill et. a1., 1985). However, the results of the Hill et. a1.
(1985) study suggest that a barrier/efﬁcacy component might add predictive utility to the
theory of reasoned action. Adding such a component to the TRA essentially turns the
model into that expressed by the theory of planned behavior (Ajzen, 1985). The theory
of planned behavior is hypothesized to have better predictive utility, with respect to
behavioral intention, than the theory of reasoned action if Pap tests behaviors are not
completely volitional. The TPB's factor of perceived behavioral control, which extends
the theory of reasoned action, has been shown to contribute signiﬁcantly in accounting for
explained variance in behavioral intention in cases where the behavior is somewhat
beyond the complete control of the actor. In light of past research and the desire to
understand if the behavior of getting a Pap test is perceived by an actor to be under her

volitional control, the ﬁrst hypothesis is offered.
H1 :The factor of perceived behavioral control will account for a signiﬁcant amount of

the variance in Pap test intention over and beyond that accounted for by attitude and

subjective norm.

Support for this hypothesis would suggest that Pap test behaviors are perceived by

actors to be, to some degree, outside their complete control.

The second hypothesis tests the claim that perceived behavioral importance has an

30

inﬂuential impact on Pap test intention. The relatively small level of explained variance in
Pap test intentions in the Hill et a1. (1985) and Hennig and Knowles (1990) studies
suggest that other variables need to be added to the theory of reasoned action, at least in
explaining intentions regarding Pap tests. It was previously argued that perceived
behavioral importance (i.e., the value placed on performing the behavior) is a conceptually
distinct from attitude and could play a predictive role in explaining Pap test intentions.
Subsequently, the second hypothesis is offered.

H2: The factor of behavioral importance will account for a signiﬁcant amount of the

variance in Pap test intention and, over and beyond that accounted for by attitude

subjective norm, and perceived behavioral control.

The third and fourth sets of hypotheses tests Ajzen and Fishbein’s (1980) claim that
persuasive messages which target speciﬁc outcome beliefs are more effective than general
informational messages. Speciﬁcally, Ajzen and Fishbein claim that messages which
utilize salient outcome beliefs should have a stronger impact on intention and behavior

than a message containing general information about the behavior in question.
H3A2Persuasive messages targeting positive outcome beliefs will be more effective at

inﬂuencing Pap test intention than a general informational Pap test message.
H3B:Persuasive messages targeting positive outcome beliefs will be more effective at

inﬂuencing Pap test behavior than a general informational Pap test message.
H4A:Persuasive messages targeting negative outcome beliefs will be more effective at

inﬂuencing Pap test intention than a general informational Pap test message.

31
H4B:Persuasive messages targeting negative outcome beliefs will be more effective at
inﬂuencing Pap test behavior than a general informational Pap test message.
It was noted that past research empirically testing Ajzen and Fishbein's claim has
resulted in inconsistent ﬁndings. The present study attempts to provide more

information in order to substantiate or refute the claim.

32

Method

Subjects

The sample selected for this study is college-aged undergraduate females. Subjects of this
age make up an appropriate sample because recent studies suggest that young, never-
married women 18 to 39, and never married White women compose the largest groups of
females never having a Pap smear (Calle, et a1., 1993). Moreover, the same study
suggests that the youngest women aged 18 to 24 have the highest risk of never having a
Pap smear. Compared to the rest of the population, college age women ﬁt the
demographic proﬁle of women who are at high-risk of never having a Pap smear. In light
of past research on Pap test behavior and the theoretical framework applied in the present
research, it also is appropriate to use a young sample because young women's Pap smear

intentions and behaviors have never been examined with the theory of planned behavior.

Participants in the present investigation were N=183 female undergraduates from a
small, midwestem college. The mean age of the participants was M=2l .6 years with no
woman over age 23 and no one less than 18 years of age. 14% were ﬁrst year students,
16% were sophomores, 20% were juniors, 20% were seniors, and 30% indicated they
were 5th year seniors. Nearly one-ﬁfth of the participants (18%) had never heard of the
Pap smear test before coming to participate in the study, and 67% had never had a Pap
smear test conducted. 13% of the women indicated they were “not knowledgeable at all”
concerning Pap tests, 75% felt they were somewhat knowledgeable, and the remaining

12% said they were very knowledgeable. These Pap test behavior and knowledge

33
statistics further underscore the need for interventions encouraging young women to

obtain Pap smear tests.

Pilot Study

As suggested by Ajzen (1988) a prior elicitation questionnaire was administered to a
sample of young, undergraduate females (N=31). Those involved in the pilot study came
from the same population from which main study participants were recruited. The
elicitation questionnaire was designed to obtain modal salient behavioral, normative, and
control beliefs relevant to getting a Pap smear test. F ishbein and Ajzen (1980) suggest
that the predictive validity of the theory of reasoned action is enhanced when modal
salient beliefs are targeted in persuasive messages. Subsequent analyses of the
questionnaires provided the positive and negative outcome beliefs which were used to
develop the stimulus materials. In deciding which beliefs were the most salient, Fishbein
and Ajzen’s (1980) recommendation to take the top ﬁve to nine most mentioned beliefs
was used. However, an examination of questionnaires indicated that subjects only
mentioned six positive outcome beliefs and eight negative outcome beliefs, and some of
those were only mentioned once. In light of this data, this researcher decided that only
beliefs mentioned by at least ten percent of the subjects would be utilized. Table 1 below

shows those positive and negative outcome beliefs which met the ten percent criterion.

34

 

Table l

Salient Positive and Negative Outcome Beliefs

ﬂ Positive Outcome Beliefs

     
     
      

Negative Outcome Beliefs

 

 

 

 

 

  

 

 

l. Detects cancer early 1. Fear of negative result
2. Gives me peace of mind 2. Costs a lot of money
3. Learn more about my own health 3. Causes discomfort

4. Can save my life 4. Anxiety producing

 

 

Messages

Based on the ﬁndings of the elicitation questionnaire, persuasive messages emphasizing
positive outcome beliefs (message 1) and challenging negative outcome beliefs (message 2)
were developed. A general information message (message 3) was also developed from an
informational pamphlet distributed by the health center at the subjects' college. Each of
the three messages was one page in length and indicated the college health center as the
source. All three messages began with providing the same general information about Pap
tests. The positive belief message (“Pap Tests: Learn the Beneﬁts!”) then emphasized
the four outcome beliefs obtained from the prior elicitation questionnaire in the second
half of the message while the negative belief message (“Pap Tests: They’re Not as Bad as
You Think!”) went on to de-emphasize the four negative outcome beliefs. The general
information message.(“Pap Tests: What You Should Know!”) continued on with
providing more general information. The three messages used in the study are presented

in Apeendices A, B, and C.

35

Questionnaire

Based on the ﬁndings from the prior elicitation questionnaire and the previously cited Pap
test studies a survey instrument was designed to measure the components of the theory
of planned behavior. Items on the test instrument conformed to procedures suggested by
Ajzen and Fishbein (1980). Unless otherwise noted all items were measured on a 7-point
linear scale and asked about obtaining a Pap test by the end of the summer (a time span of
approximately three months).

Behavioral Beliefs: On a 7-point scale anchored by extremely likely (I) and extremely
unlikely (7) subjects were asked to give probability ratings that seven outcomes Would
result from having a Pap test. The seven beliefs were presented in a manner similar to
what is subsequently described: "My having a Pap test by the end of the summer
(belief)...." give me a sense of relief that I don't have cancer, reassure me about not having
cancer, cause me embarrassment with my doctor, be physically unpleasant, mean any
cancer found would be curable, cause me to worry until I got the test results, detect cancer
in the early stages.

Outcome Evaluation: To obtain measures of the value subjects place on each outcome
belief subjects were asked to rate how good or bad each outcome would be. These items
were presented like this, "For you, how good or bad would (outcome) be?" Subjects
responded on a 7-point scale with anchors, extremely bad (1) and extremely good (7).

Attitude: A molecular measure of attitude was obtained by smnming up the products
of each outcome belief and its corresponding outcome evaluation, in accordance with the

procedure prescribed by Aj zen and F ishbein (1 980). A global measure of attitude toward

36

having a Pap test by the end of the summer was obtained by utilizing respondents' scores
on 6 semantic differential items: bad-good, foolish-wise, unfavorable-favorable,

harmful-beneﬁcial, worthless-valuable, useless- useful.

Normative Beliefs: To assess subjects' normative beliefs, respondents were asked the
degree to which seven important individuals/ groups would be in favor or oppose the
subject having a Pap test by the end of the summer. Subjects were asked in this manner,
"How much would (individual/ group) be in favor or oppose you having a Pap test by the
end of the summer?" The important others were: my mother/stepmother, my
father/stepfather, my boyfriend/signiﬁcant other, my doctor, my close friends, my
roommates, my grandmother. Responses were recorded on a 7-point scale with anchors,
extremely oppose (1) and extremely in favor (7).

Motivation to Comply: Measures to assess the degree to which the subject wanted to
do what each of the seven referents wanted them to do with respect to having a Pap test
by the end of the summer was obtained on a 7-point scale. The scale was anchored by
don't want to at all (1) and want to very much (7), and was asked in this manner, "When
it comes to having a Pap test by the end of the summer, I want to do what my (referent)
wants me to do." ‘

Subjective Norm: Similar to the molecular measure of attitude, the molecular measure
of subjective norm was obtained by multiplying each normative belief score with its
corresponding motivation to comply score and summing up the seven products. To
obtain a global measure of subjective norm subjects were asked the extent to which they

agree with the following four items, "Most people who are important to me (1) think I

37

should get, (2) want me to get, (3) are in favor of me getting, (4) would like me to get) a
Pap test by the end of the summer." Responses were recorded on a 7-point scale
anchored by totally disagree (1) and totally agree (7). A general motivation to comply
question was also be asked, "When it comes to getting a Pap test by the end of the
summer, I generally want to do what most people who are important to me 1) think I
should do , (2) want me to do , (3) are in favor of me doing, and (4) would like me to do.
Again, each general belief was multiplied by its corresponding general motivation to
comply. Each of the products were then be summed to give a global measure of

subjective norm.

Barriers: Respondents were asked to indicate the degree to which seven items would
be barriers to them getting a Pap test by the end of the summer. The 7-point scale was
anchored by no barrier at all (1) and very big banier (7). The seven barrier items were:
lack of time, forgetting to do so, embarrassment associated with the test, fear of result,
physical discomfort associated with the test, indignity of examination, access to
doctor/health center, and cost of getting the test.

Barrier Control: Respondents' ability to overcome the seven barriers also was
assessed by asking subjects how easy or difﬁcult it would be for them to overcome each
of the barriers. Anchors of very difﬁcult (1) and very easy (7) were used. The seven
questions were framed as follows, "when it comes to getting a Pap test by the end of the

summer, how easy/difﬁcult would it be for you to over come (barrier)."

Perceived Behavioral Control: A molecular measure of perceived behavioral control

was obtained by summing each of the eight products resulting from the multiplication of

38

each barrier score with its respective barrier control score. A global measure of perceived
behavioral control was assessed using the following four items, 1) I have control over
whether I get a Pap test by the end of the summer, 2) I am able to get a Pap test by the
end of the summer, 3) It would be easy for me to get a Pap test by the end of the

summer, 4) Nothing is stopping me from getting a Pap test by the end of the summer.

Perceived Behavioral Importance: A global measure of importance was obtained by
asking subjects to respond to the following six items, 1) How important is it for you to
get a Pap test by the end of the summer? (not important at all-extremely important), 2)
How big a goal is it for you to get a Pap test by the end of the summer? (not a goal at all-
very big goal), 3) How motivated are you to get a Pap test by the end of the summers?
(not motivated at all-extremely motivated), 4) How much value do you place on getting a
Pap test within the next six months? (very little value-a great deal of value), (5) how
committed are yo to getting a Pap test by the end of the summer (not committed at all-
extremely committed), and (6) how signiﬁcant is it for you to get a Pap test by the end of
the summer (not signiﬁcant at all-extremely signiﬁcant). Each item was scored on a seven-

point scale.

Intention: Intention was measured with four items, 1) I intend to get a Pap test by
the end of the summer, 2) It is likely that I will get a Pap test by the end of the summer,
3) I expect to get a Pap test by the end of the summer, and 4) I plan to get a Pap test by
the end of the summer. Respondents were asked to indicate the degree to which they
agreed with each of the four statements on a 7-point scale, where 1 means totally disagree

and 7 means totally agree. The four items were factor analyzed obtain the measure of

39
intention.

Behavior: Because of the difﬁculty in obtaining an actual behavior measure
(measuring if respondents actually got a Pap smear test), a proxy for behavior was used in
order to assess the entire framework set forth in the theory of planned behavior. The
proxy in this case was a pre-addressed, stamped postcard respondents could ﬁll out to
receive more information about Pap smear tests and cervical cancer. The postcards were
attached to each of the questionnaires and subjects were told at the beginning of the data
collection session that if they wanted to receive more information they could ﬁll out the
post card and mail it back to the researcher. Although this measure of behavior is not

optimal, it does offer a behavioral proxy that can be used to test the study’s hypotheses.
In addition to the measures noted above, items to test the effect of the three messages

were also included in the questionnaire. Speciﬁcally, items were included to measure

persuasiveness, memorability, favorability, fear producing, motivating, credibility,

informativeness, and attitude toward Pap smears.

Design and Procedures

A quasi-experimental post-test only, no control group design with three experimental
groups was utilized. This particular design allows the comparison of scores among the
three message conditions. Participants were recruited with ﬂyers from various dorms
(underclassmen) and on-campus apartment complexes (upperclassmen) to help ensure
variance in age (within the age boundaries of traditional college-aged students). The ﬂyers

asked for participants for a women’s health study (nothing on the ﬂyer indicated the

40

healthissue under investigation). As an incentive, the ﬂyer stated that those who
participated in the study would be included in a drawing for cash prizes of $100, $50, and

$25.

Data collection was conducted by female undergraduate research assistants on two
consecutive nights at various spots on the college campus. When participants arrived,
they were asked to sign-in on a sheet of paper which would be used to select the winners
of the cash prizes. A research assistant then provided a short, general introduction to the
research project and then distributed a packet of materials which included the survey
instrument, one of the three messages (general information, N=55; negative belief
messages, N=68; positive belief messages, N=60), and a consent form. Participants in
any particular data collection session all received the same message. Each data collection
session lasted approximately 25 to 30 minutes. Participants were told during the
introduction that if they would like further information about Pap tests and cervical
cancer they could do so by completing and returning the pre-addressed stamped postcard
attached to the end of their survey instrument. Return of the card was used as the

behavioral measure.

41
Analyses and Results

Results indicate that 50% of subjects in the positive outcome belief message condition
returned their postcards, compared to 31% of those in the negative outcome belief
message condition, and 29% of those who received general information messages. A one-
way analysis of variance conﬁrms a signiﬁcant difference among the conditions
F(2,167)=5.50, p<.01 when it comes to behavior. While there was a signiﬁcant difference
in conditions in behavior, a one-way AN OVA did not indicate any difference in
conditions in terms of behavioral intention, F(2,180)=.23, p>.05. At ﬁrst glance these
results are puzzling since one would expect a difference among the conditions on the
measure of intention if there is a difference on the measure of actual behavior. However,
it must be reminded that the measure of behavior in this study was whether or not
subjects returned a postcard requesting more information about Pap tests. Though this
behavior requires some effort, it is not as effortful as getting an actual Pap test.
Moreover, there are presumably fewer negative outcomes associated with returning a
postcard than with actually going through the Pap test procedure . In other words, the
behavior being measured in the present study is relatively easy to do compared to the
behavior presented in the measure of intention. In this context, the obtained results are
not surprising.

Before subjects responded to items relevant to the components of the TRA and TPB,
they were asked to rate the message they read along seven (7) evaluative dimensions.

The ratings are presented in Table 2.

42

 

Table 2

Message Ratings

ltemGeneral InfoNegative BeliefPositive Belief
(Overall mean & SD) Mean SD Mean SD Mean SD

 

 

Persuasive . 4.76 1.25 5.25 1.08 5.32 1.17
(m=5.13, SD=1.18)

Memorable 4.67 1.06 4.99 1.1 l 4.98 1.26
(m=4.89, SD=1.15)

Favorable 4.70 1.27 5.12 1.21 5.13 1.26
(m=5.02, SD=1.25)

Fear Producing 3.56 1.50 3.34 1.67 4.32 1.50
(m=3.73, SD=1.62)

Motivating 4.56 l.l5 4.94 1.30 4.88 1.24
(m=4.8l, SD=1.24)

Credible ‘ 5.67 1.09 5.69 1.00 5.58 1.33
(m=5.65, SD=1.14)

Informative 6.05 .73 6.23 .83 6.12 .98
(m=6.12. SD=.85)
Note: Ratings are on a l to 7 scale with 7 being high.

 

In general, respondents found the messages to be persuasive (m=5.13), memorable
(m=4.89), favorable (m=5.02), motivating (m=4.81), credible (m=5.65), and informative
(m=6. 12). Subjects did not ﬁnd the messages frightening (m=3.73). The results of the
ratings do not indicate a large number of differences in means among the three message
categories. Negative and positive belief messages were rated signiﬁcantly more persuasive
and favorable than general informational messages, p<.05 in both cases. General
information messages were rate slightly less memorable than both negative and positive

belief messages, though differences were not signiﬁcant at the .05 level. Surprisingly,

43

positive belief message were rated signiﬁcantly more fear producing than both general
information and negative belief messages, p<.05 in both cases. One plausible reason for
this result is that in the positive outcome belief message, two of the targeted beliefs
mentioned cervical cancer while emphasizing the life saving potential of Pap tests. It is
likely that these two beliefs in their mentioning of cancer prompted cognitions associated
with death, and thus produced a moderately fearful reaction to the message. All three
messages were rated relatively high on the credible and informative dimensions. This
suggests that respondents probably did not discount or entirely disregard the information
they read. Other comparisons did not produce any other signiﬁcant differences among
the messages.

To test hypotheses one and two, multiple regression analyses were conducted.
Before any regression model was tested, valid factors to be used in the regression
equations were developed. In this study, conﬁrmatory factor analysis (CFA) was used to
provide valid factors for the testing of the regression model. in CFA, the researcher a
priori hypothesizes the existence of certain factors based on theory. The researcher then
develops items to measure those factors. After gathering data relevant to those items, the
researcher subjects the items to at least three separate tests. In the ﬁrst test, the face
validity of each items is assessed. That is, the researcher asks if the items in question
appear to be measuring the hypothesized factor and no other factors. Items which are
considered face valid are then tested for internal consistency. This statistical procedure
tests for the similarity of the item-to-total correlations of the within factor items. The
third test is a test of parallelism (external consistency). In this test, each of the items

within a factor are examined to see if they correlate similarly to outside factors within the

44

hypothesized model. Items which pass the tests for face validity, internal consistency,
and parallelism are then considered to be relatively valid indicators of their respective
hypothesized factors. Table 3 presents a list of the items which passed the content and
statistical tests along with their conﬁrmed factors. All factors were derived from the
global measures of attitude, subjective norm, perceived behavioral control, perceived
behavioral importance, and intention (latent variables). Bentler (1980) argues that
measures of latent variables should be used in causal modeling analyses because they are
more reliable than measures of the more manifest, distal variables (i.e, outcome beliefs,
outcome evaluations). Measures of these more distal variables include more error and are

thus less reliable. These latent factors were used in subsequent analyses.

45

 

 

Table 3
CFA Results: Conﬁrmed Latent Factors and Their Items
Factor
Loading:
Attitude (alpha = .92)
.77 For me, getting a Pap test by the end of the summer is extremely foolish/wise.
.74 For me, getting a Pap test by the end of the summer is extremely harmful/beneﬁcial.
.94 For me, getting a Pap test by the end of the summer is extremely worthless/worthwhile.
.89 For me, getting a Pap test by the end of the summer is extremely useless/useful.
.83 For me, getting a Pap test by the end of the summer is extremely unimportant/important

Subjective Norm (alpha = .90)

.88 When it comes to getting a Pap test by the end of the summer, I generally want to do what most people
who are important to me THINK I should do.

.81 When it comes to getting a Pap test by the end of the summer, I generally want to do what most people
who are important to me WANT me to do.

.79 When it comes to getting a Pap test by the end of the summer, I generally want to do what most
people who are important to me are IN FAVOR of me doing.

.85 When it comes to getting a Pap test by the end of the summer, 1 generally want to do what most

people who are important to me ENCOURAGE me to do.

Perceived Behavioral Control (alpha = .74)

.73 It would be easy for me to get a Pap test by the end of the summer.
.68 I am unable to get a Pap test by the end of the summer. (R)
.70 I have complete control whether I get a Pap test by the end of the summer.

Perceived Behavioral Importance (alpha = .95)

.86 How much value do you place on getting a Pap test by the end of the summer?
.91 How motivated are you to get a Pap test by the end of f the summer?

.95 How big a goal is it for you to get a Pap test by the end of the summer?

.90 How signiﬁcant is it for you to get a Pap test by the end of the summer?

.81 How committed are you toward getting a Pap test by the end of the summer?
.93 How important is it for you to get a Pap test by the end of the summer?

Intention (alpha = .92)
.86 I intend to get a Pap test by the end of the summer.
.90 I do not plan to get a Pap test by the end of the summer. (R)
.91 It is unlikely that I will get a Pap test by the end of the summer. (R).

 

Note: The letter “R” indicates the item was reversed scored.

Hypothesis one predicted that the addition of a perceived behavioral control (PBC)
factor would account for a greater amount of the variance in Pap test intention over and

beyond that accounted for by attitude and subjective norm. By adding this

46
control/efﬁcacy factor, the theory of reasoned action is turned into the theory of planned
behavior. To test this hypothesis two separate multiple regression analyses were
conducted. The ﬁrst analysis regressed intention onto attitude and subjective norm. The
second regression analysis regressed intention onto attitude, subjective norm, and
perceived behavioral control. Comparing the two st from these analyses allows one to
see if one model accounted for more variance in intention than the other. If the regression
equation containing perceived behavioral control accounts for more variance, one can
conclude that the additional factor (PBC) was the cause for the added variance accounted
for. Results show support for hypothesis one. The regression of intention onto attitude
and subjective norm was signiﬁcant (Multiple R=.57, R2=.33, F(2, 180) = 44.16, p <
.001). Attitude and subjective norm accounted for 33% of the variance in intention. The
regression of intention onto attitude, subjective norm, and perceived behavioral control
was also signiﬁcant (Multiple R=.62, R2=.38, F(3, 178) = 37.10, p < .001). With all
three factors which comprise the theory of planned behavior in the model, 38% of the
variance is accounted for in intention. The addition of perceived behavioral control factor
increased the variance accounted for in intention by 5%. These results suggest that PBC
is a useful factor in attempting to predict Pap test intentions. Moreover, they suggest

that getting a Pap test is not perceived by young women to be fully under their control.
While it is useful to know that attitude, subjective norm, and perceived behavioral
control have a combined effect on intention, it is also important to understand the relative

contributions each makes, independent of the other predictor variables. One way to

47
examine these independent effects is to examine the partial correlations of each
independent variable. A partial correlation analysis was conducted. Table 4 presents the
zero-order and partial correlations of the TRA-model exogenous predictors with

intention.

 

Table 4
Correlations between TRA predictor variables and intention

[Predictor Variables Zero-order Correlations (r) Partial Correlations (pr)

 

 

 

 

 

 

 

Subjective Norm .51 .36
Attitude .48 .3 1 ‘

 

 

The data indicate that subjective norm and attitude have approximately the same
impact on intention, with respect to one another, whether one examines the zero-order
correlations or the partial correlations. By partialling out the effect of one independent
variable while examining the correlation between the other independent variable and the
dependent variable, we are able to see the independent effect of each of the predictor
variables. We are also able to see their relative contributions to changes in the dependent
variable In the present case, we see that the association between subjective norm and
intention shrinks from r=.51 to pr=.36 when attitude is partialled out. The partial

correlation between subjective norm and intention is signiﬁcant, t=6.98 (180), p < .01.

48
Likewise, the relationship between attitude and intention drOps from r=.48 to pr=.31
' when subjective norm is partialled out. The partial correlation between attitude and
intention is also signiﬁcant, t=6.01 (180), p < .01. One can readily see that the strength
of the relationship between attitude and intention was enhanced by the relationship
between attitude and subjective norm. In the same way, the relationship between
subjective norm and intention was enhanced by the relationship between attitude and
subjective norm. Though subjective norm is slightly more correlated with intention than
is attitude with intention, the differences are within sampling error.

Squaring the partial correlations gives an estimate of the amount of variance attitude
and subjective account for, independently, in intention. One can see that subjective norm
accounts for 13% (prz = .362 =. l 3) of the variance in intention and attitude accounts for
10% (prz = .312 =. 10) of the variance. The summed independent effects of subjective
norm and attitude (23%) does not equal the R2 from the multiple regression analysis
(33%) because attitude and subjective are correlated and also have a combined effect,
apart from the individual effects. Multiple regression takes into account both
independent and combined effects (Cohen & Cohen, 1983).

Table 5 shows the zero-order and partial correlations of the theory of planned

behavior exogenous predictor variables (perceived behavioral control is added) with

intention.

49

 

Table 5

Correlations between TPC predictor variables and intention

 

 

 

IW Zero-order Correlations Partial Correlations
PBC (perc. beh. control) .36 .23
Subjective Norm .52 .35
Attitude .48 .29

 

 

 

 

Again, the relationship of each of the predictor variables with intention shrinks when

the effects of the other independent variables are partialled out. Each of the partial

correlations is statistically signiﬁcant at p = .01. For the partial correlation between PBC

and intention, t(l78)=4.82 , p < .01 , between subjective norm and intention, t(l 78)=7.34,

p < .01, and between attitude and intention, t(l78)=6.08, p < .01. Moreover, the data

show that perceived behavioral control accounts for 5% of the variance in intention,

subjective norm accounts for 12% of the variance, and attitude accounts for 8% of the

variance. Again, the summed independent effects (25%) does not equal R2 from the

multiple regression analysis (3 8%), suggesting a combined effect of the independent

variables. In sum, the data suggest that perceived behavioral intention is a valid and useful

factor in predicting young women’s intention to get a Pap test. It accounts for additional

variance over and beyond attitude and subjective norm.

Hypothesis two predicted that the factor of perceived behavioral importance (PBI)

50
would account for a signiﬁcant amount of the variance in Pap test over and beyond that
accounted for by the three factors in the theory of planned behavior (i.e., attitude,
subjective norm, perceived behavioral control). Multiple regression analyses were used to
test the hypothesis. Results of the analyses indicate strong support for this prediction.
The regression of intention onto attitude, subjective norm, perceived behavioral control
and perceived behavioral importance was found to be signiﬁcant (Multiple R = .81,
R2=.66, F(4, 177) =85.26, p < .001). The inclusion of perceived behavioral importance

substantially increased the amount of explained variance in intention from 38% to 66%.

To discover the relative, unique effects of each the four independent variables, partial
correlations were once again calculated. Table 6 provides the zero-order and partial
correlations of intention with the theory of planned behavior predictor variables along

with perceived behavioral importance.

 

Table 6

Correlations between TPB variables + Perceived Behavioral Importance with
Intention

Zero-order Correlations Partial Correlations

 

 

'edimr Variblaes
4 FBI (per. beh. importance) .79 .67
PBC (perc. beh. control) .36 .22

 

‘ Subjective Norm .53 .20

 

 

 

 

51

With the inclusion of perceived behavioral importance in the regression equation, we
see substantial drops in the strength of the association between the other three
independent variables and intention (subjective norm: r=.53 —> pr=.20; PBC: r=.36 —>
pF.22; attitude: r=.48 —> pr=.01). Indeed, once PBI is added to the equation, the
relationship between attitude and intention disappears. Signiﬁcance tests indicate a
statistically signiﬁcant partial correlation between PBI and intention, t=26.22, (177 df), p
< .01, PBC and intention, t=8.61, (177 df), p < .01, and subjective norm and intention,
t=7.83, (177 (If), p < .01. Each of these predictor variables makes statistically signiﬁcant
and unique contributions to intention. However, the non-signiﬁcant partial correlation
between attitude, t=.39, (177 df), p > .05 suggests that the relationship between attitude
and intention is spurious, or that attitude has only an indirect effect on intention (see
Cohen & Cohen, 1983). Since the association between attitude and intention weakened
substantially with the addition of perceived behavioral importance into the regression
equation, it is plausible that the attitude component is not necessary in a model
representing Pap test intentions. However, since past studies have indicated a correlation
between attitude and intention, it is likely that attitude has an indirect effect on intention,
possibly via a path that runs through perceived behavioral importance (i.e., perceived
behavioral importance is an intervening variable). The weakened relationships between
subjective norm and intention and perceived behavioral control and intention when

perceived behavioral importance was added also suggests perceived behavioral importance

52
has an intervening status in these two relationships . To check whether perceived
behavioral importance has an intervening effect, a causal model was tested. This model,

along with its corresponding path coefficients is depicted in Figure l.

 

Figure l

Causal Model with PBI as an Intervening Variable

 

 

| Attitude]

 

 

 

 

 

30:29:“ .29 ’ .80 .18 i Behavior]

 

 

'08

‘Path model with perceived behavioral importance as an intervening variable. The model provided a good
ﬁt to the data, x2: 3. 73, df= 6, p=.8l.

 

 

Latent variable causal modeling was used to check the ﬁt of the measurement model
with the data. Bentler (1980) argues that this model of testing causal models is superior
to manifest variable causal modeling because the former method eliminates measurement
error that is introduced by the measurement of the more distal, manifest variables (e. g.,
outcome beliefs, outcome evaluation, normative beliefs, motivation to comply, etc.). Any
measure has with it an associated error component, so the more measures one makes, the
more errors are introduced to the model. For example, in this study, a molecular measure

of attitude (i.e., the sum of the belief X evaluation products) is a less reliable measure of

53
attitude than a global measure (i.e., latent factor) because more items need to be measured
in the molecular measure. Reducing error in the model enhances reliability.

The model was tested using Hunter and Hamilton’s (1992) PATH program. Results
of the analysis suggest a good ﬁt to the data, x2 = 3.73, df=6, p =8]. 95% conﬁdence
intervals were used to check the signiﬁcance of each of the path coefficients (betas). For
the path between attitude and PBI the lower and upper bounds of the conﬁdence interval
are .33, .50, thus b=.43 is signiﬁcant. For subjective norm and PBI, the lower and upper
bounds are .21, .37 respectively, thus b=.29 is signiﬁcant. The 95% conﬁdence interval
for the path between PBC and PBI is —. 10, < b >. 1 8; since the conﬁdence interval
includes zero (0), this path is non-signiﬁcant. The path between PBI and intention has a
conﬁdence interval of .77 < b> .84, thus b = .80, is signiﬁcant. The path between
intention and behavior has a conﬁdence interval between .11, and .26, thus b= .18 is

signiﬁcant.

The non-signiﬁcant path between PBC and PBI along with the data which indicate
that PBC has a signiﬁcant impact on intention suggests that the path between PBI and
intention needs to be direct. That is, PBC and PBI become the immediate determinants of
intention while attitude and subjective norm remain in their positions as the determinants
of PBI. Moreover, intention remains the only predictor of behavior. The adjusted model

along with corresponding path coefﬁcients are presented in Figure 2.

54

 

 

Figure 2
Adjusted Causal Model
Att i t u d e .44

 

 

 

 

 

 

\ .76 .18

f / PB' ——> Intention ___> Behavior

Subjective
Norm _ 31 /
.15

*Adjusted causal model. The model ﬁts the data well, x2=2.44, df=6, p=.93.

 

 

 

 

 

 

 

 

 

 

 

The adjusted model was also tested using PATH. Results indicate the data ﬁt the
model very well, x2=2.44, df=6, p=.93. The chi-square statistic was reduced ﬁ'om
x2=3.73 (previous model) to x2=2.44 (adjusted model). Moreover, all path coefﬁcients
are signiﬁcant at the .05 level of signiﬁcance. To check the amount of variance in
intention accounted for by PBI and PBC a multiple regression analysis was conducted.
The analysis regressing intention onto PBI and PBC was found to be signiﬁcant (Multiple
R=.80, R2=.64, F(2,179)=161, p<.001. Furthermore, the amount of variance in intention
accounted for by PBC and PBI is virtually equivalent to the amount of variance accounted
for by attitude, subjective norm, PBI and PBC (R2=.66). This suggests that a more
parsimonious model can be created without sacriﬁcing any predictive power by deleting

attitude and subjective norm. A check of the correlations among all variables in the model

55
also support such a conclusion. Table 7 shows attitude and subjective norm are highly
correlated with perceived behavioral importance suggesting multicollinearity among the
independent variables. According to Cohen and Cohen (1983) when multicollinearty
exists, it is appropriate to combine the correlated variables into a single index or delete the
more peripheral ones. Since a factor analysis did not support the single index option, it

was decided that attitude and subjective norm should be dropped from the model.

 

Table 7

Correlations Among Model Variables

 

 

Int PBC PBI SN Att Beh
Intention 1.00 .36“ .80* .52"' .48* .18"
PBC — 1.00 .27“ .28“ .24“ .00
PBI — — 1.00 .53“ .59* .21“
SN — — — 1.00 .48* .16“
Art — — — — 1.00 .14*
Beh — -— - - _ — 1.00

 

* indicates signiﬁcance at .05 level

 

Hypothesis 3A predicted that persuasive messages using positive outcome beliefs
will have a greater impact on Pap test intention than will general information messages.
To test this hypothesis, a t-test comparing independent group means was conducted. In

the positive belief condition, the mean score for intention is, m = 12.07, sd = 5.78. The

56
mean in the general information condition is, m = 11.40, sd = 5.70. While the results are
in the right direction, the hypothesis is not supported, t(113) = .62, p > .05. Table 8
presents the means and standard deviations of intention and behavior according to

message condition.

 

Table 8

Means and Standard Deviations of
Intention and Behavior According to Message Type

 

 

Intention Behavior
Message
Type Mean SD Mean SD
General Info 11.40 5.70 1.29 .46
Negative Belief 10.00 5.57 1.31 .47

Positive Belief 12.07 5.78 1.50 .51

 

Note: The measure of intention is the sum of three items. Each item was measured on a seven point scale
with 1 being low and 7 being high. Therefore, scores on the measure of intention can range from 3 to 21.

Scores on the measure of behavior can range from 1 (did not return postcard) to 2 (returned po_stcard).

 

Hypothesis 3B predicted that persuasive messages targeting salient positive beliefs
would have a greater impact on Pap test behavior (whether or not they return a postcard
requesting for more information) than messages containing general Pap test information.
Since the dependent variable (behavior) has a binomial distribution a Z-test for the
equality of two proportions (see Kanji, 1993) was used to test the hypothesis . Results

of the analysis suggest the null hypothesis that the proportions are equal can be rejected,

57
Z = 2.33, p < .05. The means and standard deviations are m = 1.50 (proportion = 50%),
sd = .51, and m = 1.29 (proportion = 29%), sd = .46 for the positive belief and general

information message conditions respectively.

Hypothesis 4A predicted that persuasive messages using negative outcome beliefs
will have a greater inﬂuence on Pap test intention than messages containing general
information about Pap tests. A t-test comparing independent group means did not
support this hypothesis, t(121) = -1.37, p > .05. The mean in the negative belief
condition is m = 10.00, sd = 5.57 while the mean in the general information condition is m

= 11.40, sd = 5.70.

Hypothesis 48 predicted that persuasive messages targeting salient negative outcome
beliefs would have a greater impact on Pap test behavior than a general information
message. Though an analysis of the proportions in each condition (negative belief: P =
31%; general information: P = 29%9) show the results are in the right direction, the

difference in the proportions is not signiﬁcant, Z = .25, p > .05.

Additional t and Z-tests were conducted to assess whether or not positive outcome
belief messages impacted Pap test intention and behavior differently than negative
outcome belief messages. These analyses indicate that positive belief messages (m =
12.07) inﬂuenced intentions to get a Pap test more so than negative belief messages (m =
10.00), t(126) = -2.06, p < 05. Likewise, positive belief messages (P = 50%) did a better
job of inﬂuencing respondents to send in a postcard (behavior) than negative belief

messages (P = 31%), 2= 2.16,p < .05.

58

Discussion

This paper examined the impact of various message types on Pap test intention and
behavior. Speciﬁcally, it investigated Ajzen and Fishbein’s (1980) claim that messages
employing pre-determined salient outcome beliefs, based on the theory of reasoned
action, are more persuasive than messages containing general information. Previous
studies (Brubaker & Fowler, 1990; Hoogstraten et. A1, 1985) offered mixed results when
testing Ajzen and Fishbein’s claim. This paper sought to offer some evidence to either

reﬁrte or substantiate the claim within the behavioral domain of Pap test behavior.

Additionally, the present study tested the theory of planned behavior (TPB) with
respect to Pap test behavior. Past studies (Hennig & Knowles, 1990; Hill et. A1, 1985)
utilizing the theory of reasoned action to examine Pap test behavior suggested that a
control component (i.e., self-efﬁcacy/perceived behavioral control) might add predictive
utility to the model. The additional component would extend the theory of reasoned
action into the theory of planned behavior. Support for the added component would
suggest that Pap test behavior is to some degree not completely under the control of an
individual. That is, there is any number of barriers to getting a Pap test which many
women perceive as unbreakable without the help of someone or something. Finally, this
study offered an additional component which was hypothesized to add predictive utility
to the TPB. Perceived behavioral importance was predicted to account for a greater

amount of variance in intention, over and beyond that accounted for by attitude,

59
subjective norm, and perceived behavioral control.

The results of the present study offers some support for the theory of planned
behavior within the domain of Pap test behavior. Hypothesis 1, which predicted that
perceived behavioral control would enhance the predictive utility of the theory of
reasoned action, was supported. A multiple regression analysis indicated perceived
behavioral control accounted for an additional 5% of the variance in intention to get a Pap
test. Moreover, the partial correlation between PBC and intention was found to be
signiﬁcant, t(178) = 4.82 , p < .01. These results suggest that Pap test behaviors are not
completely under the volitional control of young women, or at least not perceived to be.
There are at least two reasons why the young women in this study may have perceived
getting a Pap test was beyond their control. First, some may have believed that getting a
Pap test is expensive and felt, as college students, they could not afford it. Second, some
may have seen access to a doctor/ gynecologist or health center as a barrier they could not
overcome by themselves. The health center at the college the subjects attended does not
provide Pap tests. Though local health agencies provide free Pap testing, it is likely the
case that many of the subjects did not have this information. Barriers such as these can
be overcome with some effort and information. These results suggest information
campaigns targeting young women for Pap tests should address these and other barrier

issues.

Hypothesis 2 was also supported. This hypothesis predicted that the addition of

perceived behavioral importance would account for an increase in variance in intention

60
over and beyond what was accounted for by the TPB predictors. Indeed, with perceived
behavioral importance in the regression equation, the amount of variance accounted for in
intention increased from 38% (TPB ) to 66% (TPB + perceived behavioral importance).
When the relative contributions of each of the four predictor variables were examined with
partial correlations, the associations between intention and the three theory of planned
behavior predictors (attitude, subjective norm, PBC) were substantially reduced. In fact,
the partial correlation between attitude and intention was only .01, compared to a zero-
order correlation of .48. These results suggested that PBI was intervening in the
relationship between intention and all three TPB predictors. This model was tested using
path analysis.

The path analysis with a model depicting (1) intention as the only determinant of
behavior, (2) PBI as the only determinant of intention, and (3) attitude, subjective norm
and PBC as the immediate determinants of PBI provided strong support for the model, x2
= 3.73, df = 6, p = .81. All path coefﬁcients were statistically signiﬁcant in the model
except for the path between PBC and PBI. Since earlier analyses showed a signiﬁcant
relationship between PBC and intention, an adjusted causal model was created to include
this particular path, and delete the non-signiﬁcant path between PBC and PBI. The
revised model provided an even better ﬁt than the previous model, x2 = 2.44, df = 6, p =
.93. Moreover, after examining the amount of variance in intention accounted for by PBC
and PBI, it was seen that those two factors accounted for virtually the same amount of

variance explained by attitude, subjective norm, PBC and PBI combined. This being the

61
case, this researcher suggests that attitude and subjective norm are not necessary
predictors of intention, as laid forth by Ajzen (1986) in his theory of planned behavior.
A more parsimonious model predicting intention and behavior can be used. This model is

depicted in Figure 3.

 

Figure 3

A Revised Theory of Planned Behavior

 

Perceived
Behavioral

 

 

 

 

Intention ——> Behavior

 

 

 

 

 

 

Behavioral
Control

\
LPerceived /

 

 

 

Tests of hypotheses 3A and 3B provided mixed results. There was no support for
the hypothesis that messages based on positive outcome beliefs were more effective than
general informational messages, as Ajzen and Fishbein (1980) claim. This null ﬁnding
could be the result of several reasons. First, respondents may not have completely read
the message they were given. Since the differences in the messages would not be noticed
by a reader until they were half-way done reading, those not carefully reading the second-

half of the message would not be greatly affected no matter which message they read.

62
However, since the mean ratings of the infonnativeness of the messages was relatively
high (low 65 on a 7-point scale), it is likely that most of the subjects read the entire
message. Another reason may be that any messages about Pap tests, whether they target
positive outcome beliefs, negative outcome beliefs, or general knowledge may elicit the
same thought processes and cognitions about such tests; especially in a sample in which
two-third of the subjects had never had a Pap test. Finally, getting a Pap test can take a
lot of effort and planning, and intentions to get a Pap test may not develop in the short

amount of time subjects had to read the message and respond to the questionnaire.

There was support for hypothesis 3B. Subjects who read the message targeting
positive outcome beliefs were much more likely to return a postcard requesting more
information than those who read the informational message. In fact 50% of those in the
positive belied message condition returned a postcard compared to 29% in the general
information condition. Again, these results must be interpreted with caution since
returning a postcard was used as a proxy for the actual behavior of getting a Pap test
conducted. The act of returning a postcard requires less effort and planning, and probably
does not have as many negative cognitions associated with it than actually getting a Pap

test.

Hypothesis 4A and 48 predicted that messages targeting negative outcome beliefs
would be more effective in inﬂuencing intention and behavior than general information
messages. Neither of the hypotheses was supported. The reasons listed previously to

account for the null ﬁnding on hypothesis 3A might also account for the null ﬁnding in

63

hypothesis 4A.

The null ﬁndings for hypothesis 3A, 4A, and 4B challenge the assumption set forth
in the theory of reasoned action that persuasive messages must target underlying outcome
beliefs in order to impact behavioral intention and actual behavior. The data indicate that
general information messages work just as well as messages challenging negative outcome
beliefs and those emphasizing positive outcomes. These ﬁndings are consistent with
those of Brubaker and Fowler (1990) and Hoogstraten et. a1 (1985) and suggest that
Ajzen and Fishbein’s (1980) model of the belief structure underlying intention and
behavior is incomplete or misspeciﬂed. Indeed, the model developed from this study
(revised theory of planned behavior) suggests it is not necessary to target outcome beliefs.
It is only necessary to develop messages which primarily increase the perception of the

importance of a behavior. Those message may or may not target outcome beliefs.

However, when it comes to deciding whether to one should target negative or
positive outcome beliefs in a persuasive message, the data from the present study suggest
that messages targeting positive outcome beliefs are more effective at inﬂuencing intention
t(126)=-2.06, p<.05, and behavior t(l l6)=-2.17, p<.05. One reason that might account for
this ﬁnding is that young women with any knowledge of Pap tests may only focus on
negative beliefs and not positive outcomes. That is, the consequences of getting a Pap
test are far more salient for young women than the beneﬁts, and what they need is not
more information bringing up the negative beliefs they hold (even if these beliefs are

challenged in the message), but information detailing the many beneﬁts Pap tests have for

64
young females. In this study, more than 70% of subjects indicated they were somewhat
knowledgeable about Pap tests, but only 34% had ever had a Pap test performed on them.
These statistics suggest that the subjects who felt somewhat knowledgeable may have had
the “wrong” knowledge. In other words, many of the respondents may only have had
negative information and or information which did not emphasize the beneﬁts of or need
for Pap tests. These data suggest that any information campaign encouraging young

women to get Pap tests should focus on the positive outcomes and beneﬁts of Pap tests.

65
Conclusion

The results of this study should be interpreted with care. Although this study provided
support for a revised theory of planned behavior, it may be that the results cannot be
generalized to other behaviors. It would be helpful for future research to test the revised
model in different behavioral domains and contexts. In addition, this study only used a
proxy for the actual behavior of getting a Pap test. Though this did not impact the results
of analyses which did not include behavior in the equation/model being analyzed (e. g.,
multiple regression analyses with intention as the dependent variable), it does call into
question the results of the full revised model. Future research should test the full model
with a behavior measure which corresponds to the behavior asked about in items
measuring intention, perceived behavioral control, and perceived behavioral importance.
Finally, the data from this study were obtained from young, white women and should not

be generalized to older, and/or non-white populations.

In summary, Ajzen and Fishbein’s (1980) theory of reasoned action and its
extension, the theory of planned behavior, provide an initial understanding of Pap test
behaviors and intentions. Intention has been shown to be the immediate determinant of
behavior, and acts as a mediator for the effects of other variables. However, this study
also suggests that, at least within the behavioral domain of Pap tests, the models need
some adjustment. Speciﬁcally, the data suggest that a perceived behavioral importance
factor can be used to replace attitude and subjective norm as one of the two immediate

determinants of intention, the other being perceived behavioral control. Such a model is

66

more parsimonious and explains as much variance in intention as the theories of reasoned
action and planned behavior

The guidelines presented by Ajzen and Fishbein (1980) for designing persuasive
messages to affect intention and behavior are not necessarily supported by the results of
the present study. Ajzen and Fishbein’s (1980) claim that persuasive messages which
target the underlying belief structure associated with a behavior are more effective than
general informational messages received mixed support. Though belief based messages
were not any more effective at inﬂuencing intention than general information messages,
positive outcome belief messages did have a greater impact on behavior. Again the
behavior measured in this study was only a proxy for actual Pap test behavior, but the
return of a postcard requesting more information is a behavior which is meaningful. If it is
the case that young women have incorrect and/or negative information conceming Pap
tests, it would be helpful to be able to encourage them to seek out more information. For
those who may be uninformed, or misinformed, requesting more information about Pap

tests is probably a good ﬁrst step toward actually getting one.

APPENDICES

67

Appendix A: Positive Outcome Belief Message

PAP TESTS: Learn the Benefits

Cervical Cancer

In 1995, the American Cancer Society estimates that more than 80,000 women will be diagnosed with
precancerous or cancerous conditions of the cervix (the lower end of the uterus). Health professionals suggest
that the use of annual Pap tests to screen for cervical cancer can greatly reduce the thousands of deaths that result
from cervical cancer.

What is a Pap Test?

The Pap test is a method of detecting precancerous and cancerous conditions of the cervix. It is considered to be
the best cancer-screening tool available. The procedure was named for Dr. George Papnicolaou, who developed
the test about 40 years ago.

How Is The Pap Test Done?

During a Pap test, a small sample of cell tissue is taken from the cervix with a swab. Like skin, the tissue
constantly regenerates itself. As new cells at the bottom of the layer grow and multiply, they force older, nearly
dead cells toward the surface — the area sampled in a Pap test. The cells are then examined under a microscope for
.any abnormalities.

Where Can I Get a Pap Test?
Pap tests can be obtained from your family doctor, an OB/GYN, a medical clinic. or local health department.

How Often Should I Get a Pap Test?

Doctors recommend that all women have the Pap test performed annually as part of a pelvic examination. Regular
Pap tests should be obtained by women who are sexually active or who have reached the age of 18. All women
who are 18 or over should have annual Pap tests whether they are sexually active or not.

Can Pap Tests Detect Cancer Early Enough?

Pap tests are highly effective at detecting cancer in its early stages when it is most curable. In fact, precancerous
conditions that are identiﬁed in their early stages are 100% curable!

Can I Have Cervical Cancer Without Knowing It?

The answer to that question is YES. Oﬁen there is no pain associated with cancer of the cervix. Obtaining
annual Pap tests can give you peace of mind and reassure you that you don't have cancer, as well as catch cancer
in its earliest stages when it is much easier to cure.

What Can I Learn From Pap Tests?

Many women say getting an annual test helps them learn more about their own health and become more health
conscious. Moreover, studies show that women who get regular Pap tests tend to take other health actions that
protect them from preventable illnesses.

Pap Tests Can Save Your Life!

In recent years the number of deaths due to cervical cancer has decreased 75%. Health experts credit the
widespread use of the Pap test as the primary reason for the signiﬁcant reductions in cervical cancer deaths.
Indeed, there are many beneﬁts to getting regular Pap tests.

68

Appendix B: Negative Outcome Belief Message

PAP TESTS: They’re Not As Bad As You Think

Cervical Cancer

In 1995, the American Cancer Society estimates that more than 80,000 women will be diagnosed with
precancerous or cancerous conditions of the cervix (the lower end of the uterus). Health professionals suggest
that the use of annual Pap tests to screen for cervical cancer can greatly reduce the thousands of deaths that result
from cervical cancer.

What is a Pap Test?

The Pap test is a method of detecting precancerous and cancerous conditions of the cervix. It is considered to be
the best cancer-screening tool available. The procedure was named for Dr. George Papnicolaou, who developed
the test about 40 years ago.

How Is The Pap Test Done?
During a Pap test, a small sample of cell tissue is taken from the cervix with a swab. Like skin, the tissue
constantly regenerates itself. As new cells at the bottom of the layer grow and multiply, they force older, nearly

dead cells toward the surface - the area sampled in a Pap test. The cells are then examined under a microscope for
any abnormalities.

Where Can I Get a Pap Test?
Pap tests can be obtained from your family doctor, an OB/GYN, a medical clinic, or local health department.

How Often Should I Get a Pap Test?

Doctors recommend that all women have the Pap test performed annually as part of a pelvic examination. Regular
Pap tests should be obtained by women who are sexually active or who have reached the age of 18. All women
who are 18 or over should have annual Pap tests whether they are sexually active or not.

What If I Have An Abnormal Pap Test?

Many women believe that an abnormal Pap smear can only mean one thing: cervical cancer. In fact, an abnormal
Pap smear often indicates a minor problem, such as inﬂammation or vaginal infection. If your smear is abnormal,
(as are some 10% of all Pap tests), the most likely cause is a minor condition. While it’s important to follow up
on an abnormal result, there’s no cause for panic.

Aren’t Pap Tests Expensive?

Contrary to popular belief , Pap tests are relatively inexpensive and sometimes free. Many public health centers
offer Pap tests free of charge or for a very minimal fee. Any cost for a Pap test is far outweighed by the potential of
saving your life.

Aren’t Pap Tests Embarrassing?

Some women ﬁnd that getting a Pap test can cause embarrassment with their doctor. However, most say that the
embarrassment quickly goes away as they get to know their doctor. And, if you have a preference for a male or
female doctor, you can often request to have your test done by a doctor that you feel most comfortable with.

Don’t Pap Tests Cause Physical Discomfort?

While some women experience a little discomfort during the test, it takes only a few seconds to perform and is
painless. Health experts emphasize that the lifesaving potential of the Pap test greatly outweighs the minimal
discomfort some women experience.

69

Appendix C: General Information Message

PAP TESTS: What You Should Know

Cervical Cancer

In 1995, the American Cancer Society estimates that more than 80,000 women will be diagnosed with
precancerous or cancerous conditions of the cervix (the lower end of the uterus). Health professionals suggest
that the use of annual Pap tests to screen for cervical cancer can greatly reduce the thousands of deaths that result
from cervical cancer.

What is a Pap Test?

The Pap test is a method of detecting precancerous and cancerous conditions of the cervix. It is considered to be
the best cancer-screening tool available. The procedure was named for Dr. George Papnicolaou, who developed
the test about 40 years ago.

How Is The Pap Test Done?

During a Pap test, a small sample of cell tissue is taken from the cervix with a swab. Like skin, the tissue
constantly regenerates itself. As new cells at the bottom of the layer grow and multiply, they force older, nearly
dead cells toward the surface — the area sampled in a Pap test. The cells are then examined under a microscope for
any abnormalities.

Where Can I Get a Pap Test?
Pap tests can be obtained from your family doctor. an OB/GYN. a medical clinic. or local health department.

How Often Should I Get a Pap Test?

Doctors recommend that all women have the Pap test performed annually as part of a pelvic examination. Regular
Pap tests should be obtained by women who are sexually active or who have reached the age of 18. All women
who are 18 or over should have annual Pap tests whether they are sexually active or not.

Who Evaluates a Pap Smear?

The sample of tissue obtained from a Pap smear along with some personal information are sent to a laboratory. A
specially trained technologist examines the tissue under a microscope and searches for abnormalities.

How Is the Pap Smear Evaluated?

The Bethesda Classiﬁcation System is a recently developed means of evaluating the Pap smear. It was devised at a
1988 workshop by the national Cancer Institute. This system provides three important categories of information
used in evaluation.

New Directions for the Pap Smear

Advances in technology are combining with the traditional Pap smear technique. the new Bethesda System, and
better sampling and interpretation guidelines to make this test as effective as possible.

PAPNET: This is a computerized method for analyzing Pap smears. The PAPNET computer duplicates the
process that the human eye and mind use to identify abnormal cells. Currently it is only being used to retest Pap
smears that have been interpreted by technologists, but eventually it is hoped that it will replace much of the
screening now done by hand.

ViraPap: This is a test that was approved by the FDA in 1989. It is used to screen for the presence of human
papillomavirus in Pap smear samples.

LIST OF REFERENCES

70

References

Abelson, R. P. (1972). Are attitudes necessary? In B. T. King & E. McGinnies (Eds).
Attitudes, conﬂict and social change. New York: Academic-Press.

Abelson, R. P. (1988). Conviction. American Psychologist, 43, 4, 267-275.

Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior.
Englewood Cliffs, NJ: Prentice Hall.

Ajzen, I., & Madden, T. J. (1986). Prediction of goal-directed behavior: Attitudes,
intentions, and perceived behavioral control. Journal of Experimental Social Psychology. ,
22, 453-474.

Allport G. W. (1935) Attitudes. In C. Murchinson (Ed.), Handbook of social psychology
W01. 2). Worchester, Mass: Clark University Press.

Babrow, A. S., Black, D. R., & Tiffany, S. T. (1990). Beliefs, attitudes, intentions and a
smoking-cessation program: A planned behaviour analysis of communication campaign
development. Health Communication, 2, 3, 145-163

Bagozzi, RR, Yi, Y., & Baumgartner, J. (1990). The level of effort required for behaviour as

a moderator of the attitude-behaviour relation. European Journal of Social Psychology,
20 , 45-59.

Bagozzi, RP. (1989). An investigation of the role of affective and moral evaluations in the
purposeful behaviour model of attitude. British Journal of Social Psychology, 28, 97-113.

Bagozzi, R.P, Baumgartner, H. & Yi, Y. (1992). Stated versus action orientation and the

theory of reasoned action: An application to coupon usage. Journal of Consumer
Research, 18, 505-518.

Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist, 3 7,
122-147.

Beck, L., & Ajzen, I. (1991). Predicting dishonest actions using the theory of planned

 

71

behavior. Journal of Research in Personality, 25, 285-301.
Bern, D. J. (1970). Beliefs, attitudes, and human aﬂairs. Belmont, CA: Brooks/Cole.

Bentler, RM. (1980). Multivariate analysis with latent variables: Causal modeling. Annual
Review of Psychology, 31,419-456.

Bentler, P.M., & Bonett, D.G. (1980). Signiﬁcance tests and goodness of ﬁt in the analysis
of covariance structures. Psychological Bulletin, 88,588-606.

Bentler, P.M., & Speckart, G. (1979). Models of attitude-behavior relations. Psychology
Review, 86, 452-464.

Bentler, P.M., & Speckart, G. (1981). Attitudes “cause” behaviors: A structural equation
analysis. Journal of Personality and Social Psychology, 40, 226-23 8.

Bentler, P.M., & Mooijaart, A. (1989). Choice of structural model via parsimony: A
rationale based on precision. Psychological Bulletin, 106, 315-317.

Boldero, J. (1995). The prediction of household recycling of newspapers: The role of
attitudes, intentions, and situational factors. Journal of Applied Social Psychology, 25.
440-462.

Boninger, D.S., Krosnick, J .A., & Berent, MK. (1995). Origins of attitude importance: Self-
interests, social identiﬁcation, and value relevance. Journal of Personality and Social
Psychology, 68, 61-80.

Boyd, B., & Wandersman, A. (1991). Predicting undergraduate condom use with the
F ishbein an d Ajzen and the Triandis attitude-behavior models: Implications for public
health interventions. Journal of Applied Social Psychology, 21, 1810-1830.

Brubaker, R. G., & Fowler, C. (1990). Encouraging college males to perform testicular self-
examination: Evaluation of a persuasive message based on the revised theory of reasoned
action. Journal of Applied Social Psychology, 17, 17, 1411-1422.

Budd, R.J., & Spencer, GP. (1985). Exploring the role of personal normative beliefs in the
theory of reasoned action: The problem of discriminating between alternative path
models. European Journal of Social Psychology, 15, 299-313.

72

Budd, RJ. (1986). Predicting cigarette use: The need to incorporate measure of salience in
the theory of reasoned action. Journal of Applied Social Psychology, 16, 663-685.

Burnkrant, R.E.,-& Page, T.J. (1988). The structure and antecedents of the normative and
attitudinal components of Fishbein’s theory of reasoned action. Journal of Experimental
Social Psychology, 24, 66-87.

Cacioppo, T., Hawkins, S. G., & Petty, R. E. (1981). The nature of attitudes and cognitive
responses and their relationships to behavior. In R. E. Petty, T. M. Ostrom, & T. C.
Brock (Eds), Cognitive responses in persuasion. Hillsdale, NJ: Lawrence Earlbaum
Associates, Inc.

Calle, E.E., Flanders, W.D., Thun, M.J., & Martin, L.M. (1993). Demographic predictors of
mammography and pap smear screening in US women. American Journal of Public
Health, 83, 53-60.

Chan, D.K., & Fishbein, M. (1993). Determinants of college women’s intentions to tell their
partners to use condoms. Journal of Applied Social Psychology, 23, 1455-1470.

Cohen, J ., & Cohen, P. (1983). Applied Multiple Regression/Correlation Analysis for the
Behavioral Sciences (2nd ed.). Hillsdale, NJ: Lawrence Earlbaum.

Collins, B. E. (1970). Social psychology. Reading, Mass. : Addison-Wesley.

Costa, P.M., Jessor, R., & Donovan, J .E. (1989). Value on health and adolescent
conventionality: A construct validation of a new measure in problem-behavior theory.
Journal of Applied Social Psychology, 19, 841-861.

Deutscher, I. (1973). What we say/what we do: Sentiments and acts. Glenview, IL: Scott,
F oresman.

DeVellis, B.M., Blalock, S.J., & Sandler, RS. (1990). Predicting participation in cancer
screening: The role of perceived behavior control. Journal of Applied Social Psychology,
20, 639-660.

DeVries, H., Dijkstra, M., & Kuhlrnan, P. (1988). Self-efﬁcacy: The third factor besides
attitude and subjective norm as a predictor of behavioural intentions. Health Education

73

Research, 3, 273-282.

DeVries, H., Backbier, E., Kok, G., &Dijkstra, M. (1995). The impact of social inﬂuences in
the context of attitude, self-efﬁcacy, intention, and previous behavior as predictors of
smoking onset. Journal of Applied Social Research, 25, 237-257.

Dinning, W.D., & Crampton, J. (1989). The Kantz health opinion survey: Correlations with
preventive health behaviors and intentions. Psychological Reports, 64, 59-64.

Duckro, P.N., & Magaletta, PR. (1994). The effect of prayer on physical health:
Experimental evidence. Journal of Religion and Health, 33, 211-219.

Eagly, A. H., & Chaiken, S. (1993). The Psychology of Attitudes. New York: Harcourt Brace
Janovich.

Eagly, A.H., Mladinic, A., & Otto, S. (1994). Cognitive and affective bases of attitudes

toward social groups and social policies. Journal of Experimental Social Psychologz, 30,
113-137.

Emmons, R.A., & Diener, E. (1986). A goal-affect analysis of everyday situational choices.
Journal of Research in Personality, 20, 309-326.

Fazio, R. H. (1986). How do attitudes guide behavior. In R. M. Sorrentino & E. T.
Higgins (Eds), Handbook of motivation and cognition: Foundations of social behavior
(pp. 204-243). New York: Guilford Press.

F ishbein, M. (1967). Attitudes and the prediction of behavior. In M. F ishbein (Ed).
Readings in attitude theory and measurement. New York: Wiley.

F ishbein, M., & Ajzen, I. (1972). Attitudes and opinions. Annual Review of Psychology, 23,
487-544.

Fishbein, M., & Ajzen, I. (1973). Attribution responsibility: A theoretical note. Journal of
Experimental Social Psychology, 9, 148-153.

Fishbein, M., & Ajzen, I. (1974). Attitudes toward objects as predictors of single and
multiple behavioral criteria. Psychological Review, 81, 59-74.

74

F ishbein, M., & Ajzen, I. (1975). Belief," attitude, intention, and behavior: An introduction to
theory and research. Reading, Mass: Addison-Wesely.

F ishbein, M., & Ajzen, I. (1981). On construct validity: A critique of Miniard and Cohen’s
paper. Journal of Experimental Social Psychology, 17, 340-350.

F ishbein, M., Traﬁmow, D., Middlestadt, S.E., Helquist, M., Francis, C., & Eustace, MA.
(1995). Using an AIDS KABP survey to identify determinants of condom use among

sexually active adults from St Vincent and The Grenadines. Journal of Applied Social
Psychology, 25,1-20.

Fishbein, M., & Stasson, M. (1990). The role of desires, self-predictions, and perceived
control in the prediction of training session attendance. Journal of Applied Social
Psychology, 20, 173-198.

Flay, B.R., & Petraitis, J. (1994). The theory of triadic inﬂuence: A new theory of health
behavior with implications for preventive interventions. Advances in Medical Sociology,
4, 19-44.

Fleming, D. (1967). Attitude: The history of a concept. Perspectives in american History, 1,
287-365.

Fredricks, A.J., & Dossett, D.L. (1983). Attitude-behavior relations: A comparison of the
F ishbein-Ajzen and the Bentler-Speckart models. Journal of Personality and Social
Psychology, 45, 501-512.

F rese, M., Stewart, J ., & Hannover, B. (1987). Goal orientation and planfulness: Action

styles as personality concepts. Journal of Personality and Social Psychology, 52, 1182-
1194.

Gray, L.N., Stafford, M.C., & Tallman, I. (1991). Rewards and punishments in complex
human choices. Social Psychologt Quarterly, 54, 318-329.

Hennig, P., & Knowles, A. (1990). Factors inﬂuencing women over 40 years to take
precautions against cervical cancer. Journal of Applied Social Psychology, 20, 19, 1612-
1621.

Hill, D., Gardner, G., & Rassaby, J. (1985). Factors disposing women to take precautions

75

against breast and cervix cancer. Journal of Applied Social Psychology, 15, 1, 59-79.
Hunter, J ., and Hamilton, A. (1992). PATH Software.

Insko, C. A., & Schopler, J. (1972). Experimental social psychology. New York: Academic
Press.

Jaccard, J. (1981). Attitudes and behavior: Implications of attitudes toward behavioral
alternatives. Journal of Experimental Social Psychology, 1 7, 286-307.

Jaccard, J., Helbig, D.W., Wan, C.K., Gutrnan, M.A., &Kritz-Silverstein, DC. (1990).
Individual differences in attitude-behavior consistency: The prediction of contraceptive
behavior. Journal of Applied Social Psycholog, 20, 575-617.

Johnson, B.T., & Eagly, AH. (1989). Effects of involvement on persuasion: A meta-
analysis. Psychological Bulletin, 106, 290-314.

Kanji, G. K. (1993). 100 Statistical Tests. Thousand Oaks, CA: Sage Publications.

Katz, D., & Stotland, E. (1959). A preliminary statement to a theory of attitude structure and
change. In S. Koch (Ed), Psychology: A study of science (V 01. 3). New York: McGraw-
Hill.

Kendzierski, D. (1990). Decision making versus decision implementation: An action control
approach to exercise adoption and adherence. Journal of Applied Social Psychologi, 20,
27-45. '

Krosnick, J. A. (1989). Attitude importance and attitude accessibility. Personality and
Social Psychological Bulletin, 15, 297-308.

Krosnick, J. A. (1990). Government policy and citizen passion: A study of public issues in
contemporary America. Political Behavior, 12, 59-92.

Kurland, NE. (1995). Ethical intentions and the theories of reasoned action and planned
behavior. Journal of Applied Social Psychology, 25, 297-313.

Liska, AB. (1984). A critical examination of the causal structure of the Fishbein/Ajzen

76

attitude-behavior model. Social Psychology Quarterly, 47, 61-74.

Madden, T.J., Ellen, P.S., & Ajzen, I. (1992). A comparison of the theory of planned
behavior and the theory of reasoned action. Personality and Social Psychology Bulletin,
18, 3-9.

McCaul, K.D., O’Neill, H.K., & Glasgow, RE. (1988). Predicting the performance of dental
hygiene behaviors: An examination of the Fishbein and Ajzen model and self-efficacy
expectations. Journal of Applied Social Psychology, 18, 114-128.

McGuire, W. J. (1976). The concept of attitude and their relations to behaviors. In H. W.-
Sinaiko & L. A. Broedlings, (Eds), Perspectives on attitude assessment: surveys and their
alternatives. Champaign, IL: Pendelton.

Miller, L.E., & Grush, J .E. (1986). Individual differences in attitudinal versus normative
determination of behavior. Journal of Experimental Social Psychology, 22, 190-202.

Miniard, P.W., & Cohen, J.B. (1981). An examination of the F ishbein-Ajzen behavioral-
intentions model’s concepts and measures. Journal of Experimental Social Psychology,
1 7, 308-339.

Morrison, D.M., Gillmore, M.R., & Baker, SA. (1995). Determinants of condom use among
high-risk heterosexual adults: A test of the theory of reasoned action. Journal of Applied
Social Psychology, 25, 651-676.

Mulaik, S.A., James, L.R., Alstine, J ., Bennett, N., Lind, S., & Stilwell, CD. (1989).
Evaluatin of goodness-of-ﬁt indices for structural equation models. Psychological
Bulletin, 3, 430-445.

Netemeyer, R.G., & Burton, S. (1990). Examining the relationships between voting behavior,
intention, perceived behavioral control, and expectation. Journal of Applied Social
Psychology, 20, 661-680.

Netemeyer, R.G., Burton, S., & Johnston, M. (1991). A comparison of two models for the
prediction of volitional and goal-directed behaviors: A conﬁrmatory analysis approach.
Social Psychology Quarterly, 54, 87-100.

Norman, S.A., Talbott, E.O., Kuller, L.H., Krampe, B.R., & Stolley, PD. (1991).

77

Demographic, psychosocial, and medical correlates of pap testing: A literature review.
American Journal of Preventive Medicine, 7, 219-226. ‘

Ostrom, T. C. (1969). The emergence of attitude theory: 1930-1950. In A. Greenwald, T.
Brock, & T. Ostrom (Eds), Psychological foundations of attitudes. New York: Academic
Press.

Page], M.D., & Davidson, AR. (1984). A comparison of three social-psychological models
of attitude and behavioral plan: Prediction of contraceptive behavior. Journal of
Personality and Social Psychology, 4 7, 517-533.

Petkova, K.G., Ajzen, I., & Driver, B.L. (1995). Salience of anti-abortion beliefs and
commitment to an attitudinal position: On the strength, structure, and predictive validity
of anti-abortion attitudes. Journal of Applied Social Psychology, 25, 463-483.

Petty, R. E., Ostrom, T. M., & Brock, TC. (1981). Cognitive Responses in Persuasion.
Hillsdale, NJ: Lawrence Earlbaum Associates, Inc.

Pfau, M. (1994). Impact of product involvement, message format, and receiver sex on the
efﬁcacy of comparative advertising messages. Communication Quarterly, 42, 244-258.

Raats, M.M., Shepherd, R., & Sparks, P. (1995). Including moral dimensions of choice
within the structure of the theory of planned behavior. Journal of Applied Social
Psychology, 25, 484-494.

Roberson, L., Korsgaard, M.A., &Diddams, M. (1990). Goal characteristics and satisfaction:
personal goals as mediators of situational effects on task satisfaction. Journal of Applied
Social Psychology, 20, 920-941.

Ronis, D.L., & Kaiser, MK. (1989). Correlates of breast self-examination in a sample of
college women: Analysis of linear structural relations. Journal of Applied Social
Psychology, 19, 1068-1084.

Rosenberg, M. J ., & Hovland, C. I. (1960). Cognitive, affective, and behavioral components
of attitudes. In M. J. Rosenberg, C. I. Hovland, W. J. McGuire, R. P. Abelson, & J. W.
Brehm (Eds), Attitude organization and change. Nw Haven, Conn: Yale University Press.

Rotter, J .B. (1966). Generalized expectancies for internal versus external control of

78

reinforcement. Psychological monographs: General and Applied, 80, 1-28.

Schiﬁer, D.E., & Ajzen, I. (1985). Intention, perceived control, and weight loss: An

application of the theory of planned behavior. Journal of Personality and Social
Psychology, 49, 843-851.

Schlegel, R.P., D’Avemas, J.R., Zanna, M.P., DeCourvill, N.H., & Manske, SR. (1992).
Problem drinking: A problem for the theory of reasoned action? Journal of Applied Social
Psychology, 22, 358-385.

Scott, W. A. (1968). Attitude measurement. In G. Lindzey & E. Aronson (Eds), The
Handbook of Social Psychology (2nd Ed. Vol. 2). Reading, Mass. : Addison-Wesley.

Shelley, J.M., Irwig, L.M., Simpon, J .M., & Macaskill, P. (1991). Evaluation of a mass-
media-led campaign to increase pap smear screening. Health Education Research, 6, 267-
277.

Shepherd, G.J., & O’Keefe, DJ. (1984). Separability of attitudinal and normative inﬂuences
on behavioral intentions in the F ishbein-Ajzen model. Journal of Social Psychology, 122,
287-288.

Sheppard, B.H., Hartwick, J ., & Warshaw, PR. (1988). The theory of reasoned action: A
meta-analysis of past research with recommendations for modiﬁcations and future
research. Journal of Consumer Research, 15, 325-343.

Smith, M. B., Bruner, J. S., & White, R. W. (1956). Opinions and Personality. New York:
Wiley.

Stasson, M., & Fishbein, M. (1990). The relation between perceived risk and preventive
action: A within-subject analysis of perceived driving risk and intentions to wear
seatbelts. Journal of Applied Social Psychology, 20, 1541-1557.

Steffen, V.J. (1990). Men’s motivation to perform the testicle self-exam: Effects of prior

knowledge and an educational brochure. Journal of Applied Social Psychology, 20, 681 -
702.

Suarez, L., Nichols, D.C., & Brady, CA. (1993). Use of peer role models to increase pap
smear and mammogram screening in Mexican-American and Black women. American

 

79

Journal of Preventive Medicine, 9, 290-296.

Sutton, 8., & Hallett, R. (1989). Understanding seat-belt intentions and behavior: A decision-
making approach. Journal of Applied Social Psychology, 19, 1310-1325.

Tannenbaum, P. H. (1956). Initial attitude toward source and concept as factors in attitude
change through communication. Public Opinion Quarterly, 20, 413-425.

Thurstone, L. L. (1931). The measurement of social attitudes. Journal of A bnormal and
Social Psychology, 26, 249-269.

Toneatto, T., & Binik, Y. (1987). The role of intentions, social norms, and attitudes in the
performance of dental ﬂossing: A test of the theory of reasoned action. Journal of
Applied Social Psychology, 1 7, 593-603.

Vallerand, R.J., Pelletier, L.G., Guerrier, J .P, Deshaies, P., & Mongeau, C. (1992). Ajzen and
Fishbein’s theory of reasoned action as applied to moral behavior: A conﬁrmatory
analysis. Journal of Personality and Social Psychology, 62, 98-109.

Ward, J .E., Boyle, K., Redman, S., & Sanson-Fisher, R.W. (1991). Increasing women’s
compliance with opportunistic cervical cancer screening: A randomized trial. American
Journal of Preventive Medicine, 7, 285-291.

Warshaw, P.R., & Davis, ED. (1985). Disentangling behavioral intention and behavior
expectation. Journal of Experimental Social Psychology, 21, 213-228.

Weinstein, ND. (1993). Testing four competing theories of health-protective behavior.
Health Psychology, 12, 324-333.

Wicker, A. W. (1969). Attitude versus actions: The relationship of overt and behavioral
responses to attitude objects. Journal of Social Issues, 25, 4, 41-78.

Wilson, D.K., Wallston, K.A., & King, J .E. (1990). Effects of contract framing, motivation
to quit, and self-efﬁcacy on smoking reduction. Journal of Applied Social Psychology, 20,
53 1-547.

 

MICHIGAN srnIE UNIV. LIBRARIES
l“ll”NilNillll“Willi”“1WWWIIHHUI
31293015648169