Operationalizing the theory of planned behavior

Abstract

The Theory of Planned Behavior (TPB) is an expectancy-value theory that provided a framework for the study of behavioral and normative beliefs affecting health behaviors. The purpose of this article is to describe operationalization of the TPB. The process of instrument development is explicated, with examples from two studies of BSE behavior in older women. Potential threats to reliability are discussed.

Full text

Researchin Nursing& Health.
1991, 14, 137-144
Operationalizing the Theory
of
Planned Behavior
Heather
M.
Young, Letha Lierman, Gail Powell-Cope, Danuta Kasprzyk,
and Jeanne
Q
.
Benoliel
The Theory of Planned Behavior (TPB) is an expectancy-value theory that provided a
framework for the study
of
behavioral and normative beliefs affecting health behaviors.
The purpose of this article is
to
describe operationalization of the TPB. The process of
instrument development is explicated, with examples from two studies of BSE behavior in
older women. Potential threats to reliability are discussed.
The Theory of Planned Behavior (TPB) (Ajzen,
1985) is a theory of human behavior that takes
into consideration the influence of personal eval-
uations, perceived social pressure, and perceived
control
in
predicting the intention to perform a
given behavior. The TPB is an extension of the
Theory of Reasoned Action (TRA) (Fishbein
&
Ajzen, 1975) that includes personal evaluations
(attitudes) and perceived social pressure (social
norms) as the most powerful antecedents to the
intention to perform a behavior. The TRA received
considerable attention as a model for understanding
and predicting health-related behaviors such as
family planning, weight
loss,
exercise, and im-
munization (Mullen, Hersey,
&
Iverson, 19871,
but was limited by the assumption that behavior
is under complete volitional control.
The TRA was challenged by work on self-ef-
ficacy (Bandura, 1977) and the Triandis Intention
Model (Davidson, Jaccard, Triandis, Morales,
&
Diaz-Guerrero, 1976). Because self-confidence
in
the ability to perform a behavior was identified
as an important construct in predicting behavior,
a number of authors proposed the inclusion of
similar constructs such
as
facilitating factors
(Triandis, 1977), perceived behavioral control
(Ajzen, 1985), and action control
(Kuhl,
1984)
in explanatory models
of
human behavior. The
Theory of Planned Behavior (TPB) is an expansion
of the TRA with the addition of the construct,
perceived behavioral control (Ajzen, 1985). The
revised model has been shown to yield greater
explanatory power than
the
original TRA for goal-
directed behaviors (Ajzen
&
Madden, 1986; Ajzen,
1987).
A
modified version of the TRA was used
in
a
longitudinal intervention study to understand,
predict, and modify breast self-examination (BSE)
behavior in middle-aged and older women (Lier-
man, Kasprzyk,
&
Benoliel,
in
press). Over the
course of this longitudinal study, we have benefitted
from the evolution
of
the TPB. The purpose of
this article is to describe methodological consid-
erations in operationalizing this model. The process
of instrument development is described, using ex-
amples from our research
in
BSE performance of
older women. This is followed by a discussion
of reliability and validity issues
in
operationalizing
the modified version of the TRA.
The Theory
of
Planned Behavior
The TPB is an expectancy value model (Ajzen
&
Fishbein, 1980; Montano, 1986). According to
the TPB, personal behaviors are under volitional
Heather
M.
Young, Ph(C), RN, is a doctoral candidate; Letha Lierman, PhD, RN, is a
research associate; Gail Powell-Cope, Ph(C), RN, is a doctoral candidate; Danuta Kasprzyk,
PhD, is a research associate; and Jeanne
Q.
Benoliel, DNSc, RN,
is
Elizabeth Soule
Professor of Nursing Emeritus in the School
of
Nursing University
of
Washington.
This research was supported by a National Center for Nursing Research, NIH Grant No.
This article was received on November
3,
1989,
was revised, and accepted for publication
Requests for reprints can be addressed to Heather
M.
Young, School
of
Nursing, SC-
1 R01 01618-01.
September
5, 1990.
72,
University of Washington, Seattle, WA
98195.
0
1991
John
Wiley
&
Sons.
Inc.
0160-6891/91/020137-08
$04.00

138
RESEARCH IN NURSING &HEALTH
Beliefs about
outcome of
performance
I
BSE
*
Attitude
r(-\
Evaluation of
Behavioral
1
Behavior
1
Beliefs about
expectations
4
social norm
Motivation to
comply with
others‘
*-
expectations
FIGURE
1.
Theory
of
Planned Behavior model. From “Prediction
of
goal-directed behavior: Attitudes, intentions,
and perceived behavioral control” by
I.
Ajzen and
T.J.
Madden, 1986, Journal
of
Experimental Social Psychology,
22,
p.
458.
Copyright 1986 by Academic Press. Adapted by permission.
control and
are
determined by behavioral intentions.
Behavioral intentions are a function of attitudes,
social norms, and perceived behavioral control
(see Fig. 1). Antecedents to attitudes and social
norms are behavioral and normative beliefs. An-
tecedents to perceived behavioral control
are
beliefs
about the ease with which one can execute the
behavior and beliefs about resources and obstacles
related to the behavior (Ajzen
&
Madden, 1986).
These beliefs are weighted by personal values.
Perceived behavioral control (an addition to the
TRA, Fig. 2) independently contributes to the
formation
of
behavioral intention. In addition,
perceived behavioral control affects behavior di-
rectly. These propositions have received empirical
support (Ajzen
&
Madden, 1986; Ajzen, 1987).
Mathematically, attitude is the sum of the prod-
ucts of beliefs about the behavioral outcome
(expectancy) and the evaluation of these outcomes
(value). Social norm is the sum of the products
of normative beliefs about the expectations of
others (expectancy) and the motivation to comply
with those expectations (value). Perceived be-
havioral control is the sum of the products of
factors affecting control over the behavior and the
relative strength of these factors. Ajzen and Fish-
bein
(
1980) also recommend that attitude be mea-
sured directly using semantic differential scales
in
relation to the target behavior and social norm
be measured directly using a global social norm
item. Likewise, Ajzen and Madden (1986) rec-
ommend a direct global measure of perceived be-
havioral control.
Attitudes, social norms, and perceived behav-
ioral control are considered the most powerful
predictors of intention; therefore, the model does
not include any background variables, such as
age, marital status, or education, which describe
the population (Ajzen
&
Fishbein, 1980). It
is
assumed that individual qualities, such as edu-

THEORY
OF
PLANNED BEHAVIOR
/
YOUNG ET
AL.
139
cational attainment or ethnic origin, indirectly affect
attitudes by shaping beliefs and evaluations.
Operationalizing
the
TPB
Ajzen and Fishbein (1980) elaborate specific
guidelines for construct measurement and ques-
tionnaire development. Data are derived from the
population of interest to ensure that salient beliefs
are included. Investigators using the TRA have
shown that the predictive power of the model is
greatest when attitudes and behavior are at the
same level of specificity (Mullen, Hersey,
&
Iver-
son, 1987). For example, prediction is enhanced
when considering attitude toward performing BSE
monthly, measures of intention to perform BSE
monthly, and reports of monthly BSE behavior.
If one were to measure attitude toward general
health promotion activities including exercise, diet,
and health monitoring in order to predict BSE
intention and behavior, less prediction would be
expected.
The model is operationalized
in
five important
steps as outlined by Ajzen
&
Fishbein
(I
980) and
expanded by Ajzen and Madden
(
1986). The first
three steps
in
operationalizing
the
TPB are the-
oretical and can be executed without the involve-
ment of the population of interest.
To
understand
the contribution of salient beliefs and referents in
the decision to perform a specific behavior con-
sultation with members of the target population
is
required. The
five
steps are summarized below.
1. Select the behavior of interest, and define
it
in
terms of its action, target, context, and
time elements.
Define the corresponding behavioral intention.
Define general attitude, social norm, and per-
ceived behavioral control.
2.
3.
4.
Elicit the salient behavioral, normative, and
perceived control beliefs about the target be-
havior from a representative sample.
Develop questionnaire items from
the
salient
behavioral, normative, and perceived control
beliefs.
5.
APPLICATION
OF
THE MODEL
TO
THE STUDY
Of
BSE
The overall pupose of our research program was
to develop and test a BSE teaching and support
intervention targeting middle aged and older women
using the TRA as a guide. As mentioned above,
operationalization of the model
in
this study was
expanded
as
the TPB evolved. Following iden-
tification of salient behavioral and normative be-
liefs, an instructional program was devised that
had as its intent the modification of these beliefs
through provision of educational materials and
social support (Lierman, Young, Powell-Cope,
Georgiadou, Benoliel, 1990). Details of this pro-
gram and model testing results (evidence of con-
struct validity) are reported elsewhere (Lierman,
Young, Kaspryzk,
&
Benoliel, 1990; Powell-Cope,
Lierman, Kasprzyk, Young,
&
Benoliel, in press;
Lierman, Kaspryzk,
&
Benoliel,
in
press). The
following section is focused on the development
of the Beliefs and Attitudes Questionnaire. In-
strument development was conducted according
to the guidelines of Ajzen and Fishbein
(1
980)
discussed in the previous section.
Samples
The Beliefs and Attitudes Questionnaire was de-
veloped and refined in two samples. Guided
in-
Attitude
Behavior
'I
Social
norm
Behavioral
in
tent ion
-
Perceived
Behavioral
Control
FIGURE
2.
Theory
of
Reasoned Action model. AjzeniFishbein, Understanding Attitudes and Predicting Social
Behavior,
0
1980,
p.
84.
Adapted
by
permission
of
Prentice-Hall, Inc., Englewood
Cliffs,
NJ.

140
RESEARCH
IN
NURSING &HEALTH
Table
1.
Demographic Characterlstlcs
of
Samples
1
and
2
Characteristic
Age (years)
Range
Mean
(SD)
>
12
years
<
12
years
Married
Widowed/divorced/never married
Education
[n
(YO)]
Marital status
[n
(%)I
Sample
1
(N
=
93)
52-90
70.5 (8.2)
86 (92)
7
(8)
52 (56)
41 (46)
Sample
2
(N
=
175)
48-78
61.4 (7.2)
161 (92)
14
(8)
121 (69)
54
(30)
terviews and pilot testing were completed
in
a
sample of
93
volunteers recruited from women’s
organizations of local churches (Sample
1).
On-
going instrument refinement and testing were con-
ducted in a sample of 175 women enrolled in our
longitudinal
BSE
study (Sample
2).
The recruitment
protocol for this study consisted of random selection
of potential participants from members of a health
maintenance organization. The participants first
were approached through a letter from their phy-
sicians describing the study and inviting their par-
ticipation. Informed consent was obtained from
all women in both samples. Demographic char-
acteristics are summarized in Table 1.
Procedure
The
TPB
was operationalized in our study of
BSE
behavior according to the forenamed steps.
The behavioral target was specified
as performance of
BSE
monthly in a 6-month
period.
In
the longitudinal study, participants were
asked to report retrospectively how many times
they had performed
BSE
in the last 6 months.
Two additional behavioral measures were included:
a
BSE
performance diary kept by the participants
and observation by trained interviewers of actual
performance proficiency.
To maintain consistency with the be-
havioral target, the behavioral intention was spec-
ified as,
“I
intend to perform
BSE
monthly for
the next 6 months,” answered on a 7-point Likert-
type scale ranging from
extremely likely
to
ex-
tremely unlikely.
The general attitude and social norm
were identified; for the TPB, general perceived
behavioral control is identified during this step.
Regarding
BSE,
the general attitude was, “My
performing
BSE
monthly for the next
6
months
would be
.
.
.
”
answered on 7-point scales with
the following dimensions:
goodlbad; benejciall
Step
1.
Step
2.
Step
3.
harmful; necessarylunnecessary; usefullfutile;
importantlunimportant.
The general social norm
was. “Most people important to me think I should
perform
BSE
monthly,” answered on a 7-point
Likert scale ranging from
strongly agree
to
strongly
disagree.
At the time of preliminary instrument devel-
opment we were using the TRA, hence we did
not explicitly seek perceived behavioral control
items. A measure of this construct was developed
later as some of the behavioral beliefs elicited in
the next phase corresponded to the recommen-
dations of Ajzen and Madden
(1986)
for perceived
behavioral control items. In particular, one item
(“Performing
BSE
would be difficult,” answered
on a 7-point Likert scale ranging from
strongly
agree
to
strongly disagree)
is
an
appropriate direct
measure of perceived behavioral control. Four items
related to perceived skill in performing
BSE
formed
a perceived control subscale. Finally, two items
related to worry about
BSE
and the impact of
breast cancer tap perceived control at an affective
level. Ajzen and Madden (1986) also suggest the
use of an item, “How much control do you have
over whether you do or do not (perform
BSE)?”
answered on a 7-point scale ranging from
complete
control
to
very little control;
we did not include
such an item.
Step
4.
This step consisted of
29
in-depth
interviews with a subset of Sample
1.
Qualitative
data were elicited regarding beliefs about
BSE
performance, cancer and cancer treatment, social
influences on
BSE
behavior, and experience with
breast cancer. To elicit salient beliefs about the
behavior of interest, the women were asked to
list both advantages and disadvantages associated
with the performance of
BSE
and to discuss any
other factors that they considered
in
relation to
monthly performance of
BSE.
To obtain normative
beliefs, the women were asked to identify indi-
viduals or groups who would approve andlor dis-

THEORY
OF
PLANNED BEHAVIOR
/
YOUNG ET AL.
141
Table
2.
Example of Belief and Evaluation Items
My performing BSE would allow me to detect breast cancer in an early stage.
Disagree Agree
1
2
3
4
5
6
7
strongly somewhat slightly neither slightly somewhat strongly
Detecting breast cancer in an early stage would be
Bad Good
1
2
3
4
5
6
7
extremely quite slightly neither slightly quite extremely
approve of their monthly performance of
BSE,
or who might have any other influence on this
behavior.
Step
5.
This step comprised content analysis
and instrument construction. The most frequent
qualitative responses were incorporated into the
pilot instrument as belief items about
BSE
per-
formance and normative beliefs about the expec-
tations of others (expectancy component). This
procedure enhanced the content validity of the
instrument. For each belief, a corresponding eval-
uation item was included (value component).
An example of a behavioral belief and corre-
sponding evaluation item
is
shown in Table
2.
The attitude score is derived from the sum of the
products of these items. The scores of each pair,
scaled from
-3
to
+3,
are multiplied, then all
the products are summed to obtain an overall at-
titude score. Table
3
displays an example of a
normative belief (expectancy) and corresponding
motivation to comply item (value). Social norm
is calculated in a similar fashion to attitude. Item
pairs are multiplied and then summed to obtain
a social norm score.
All salient beliefs were included in the pilot
instrument which was administered consecutively
to three groups of women (subsets of Sample
1)
and modified after each round based on feedback
from the participants. Instrument testing using
Sample
1
was conducted to establish internal con-
sistency reliability (Lierman et al., 1990). Based
on descriptive statistical analysis, items with small
variance were deleted as were certain items with
high intercorrelations. Cronbach’s alpha coefficient
for the 16-item attitude scale was
.70
and for the
9-item social norm scale was .86.
In the longitudinal study (Sample
2),
salient
behavioral and normative beliefs were incorporated
in
an ongoing fashion in the form of additional
items, ensuring that the instruments retained content
validity. Participants were interviewed on three
occasions with the
BSE
intervention conducted
upon completion of the first interview. During the
second interview, open-ended questions were asked
Table
3.
Example of Normative Belief and Motivation to Comply Items
Do you think your sister feels you should perform BSE on a monthly basis?
1.
I
don’t have any sisters
2.
I
don’t know
3. Definitely no
4.
Probably no
5.
She is neutral
6.
Probably yes
7.
Definitely yes
1.
I
don’t have any sisters
2.
Not at all
3.
Slightly
4.
Moderate amount
5.
A great deal
6.
I
don’t know
How much do you try to do what you feel your sister thinks you should do?

142
RESEARCH
IN
NURSING
a
HEALTH
to identify new factors that might have influenced
performance; these were incorporated as items in
the third round of interviews. This practice ensured
that important factors which became salient over
the course of the longitudinal study were included.
The new factors identified by the women as
influencing their performance corresponded con-
ceptually to perceived behavioral control as defined
by the TPB. For example, during the second in-
terview, many women mentioned that disruptions
in their normal schedule related to the holidays
or to family visits had influenced their performance
of BSE. This comment was developed into be-
havioral beliefs and evaluation items that were
included in the third interviews to measure per-
ceived control (barriers to performance of BSE).
DISCUSSION
A major strength of the measurement approach
taken by Ajzen and Fishbein
(1980)
is the grounding
of the instrument development process
in
the
qualitative findings of in-depth interviews with
members of the target population. This practice
enhances content validity and ensures that the
items are salient for both the study population and
the target behavior. Interestingly, when items
thought to be important by the researchers were
included, their predictive ability was much lower
than the items that were grounded
in
the population
of interest. This finding further supports the power
of the prescribed method of instrument devel-
opment.
Despite the iterative process of instrument de-
velopment a number of threats to reliability remain
as a theory is applied to the empirical world. Our
most important concerns are related to (a) the
problems encountered when a participant does not
harbor a specific expectancy; (b) the notion of
evaluating beliefs acontextually; (c) the conceptual
difficulty participants have in answering complex
items; and (d) the use of an expectancy value
model in health research, an arena in which therapy
and research can blur. These concerns warrant
elaboration.
First, we question the reliability of attitude
scores when a specific belief is not endorsed by
a participant. The calculation of an attitude score
is
dependent
on
responses to belief (expectancy)
and evaluation (value) items. This is simple when
a participant agrees with the belief item; for
example, most people agree with the statement,
“My performing BSE would allow me to detect
breast cancer in an early stage.” The corre-
sponding evaluation item, “Detecting breast
cancer
in
an early stage would be good
. . .
bad,” rarely is difficult for a participant to answer.
When a specific belief is not held by the par-
ticipant, however, evaluation of that belief can
be problematic. For example, “My performing
BSE would be difficult” is a belief that was
identified with sufficient frequency to warrant
its inclusion in the instrument, but the majority
of the women disagreed with this statement.
When asked to respond to the evaluation item,
“Performing an examination that is difficult for
me would be good
. . .
bad,” those who disagreed
that it was difficult were unsure how to answer.
Should they imagine that it was difficult for
them or imagine what it would be like for some-
one who found it difficult?
A
second reliability issue rklates to the evaluation
component. In asking a respondent to evaluate an
item (on a Likert scale ranging from
extremely
good
to
extremely
bad) one assumes that the re-
spondent holds an absolute opinion about that
belief and that it is possible to make this evaluation
acontextually.
In
evaluating the factors that support
or interfere with their performance of BSE, many
women answered that it depends on the circum-
stances surrounding the factor. For example,
“Putting
off
the performance of monthly BSE’
could be considered extremely bad if,
in
the view
of the respondent, it was due to laziness; however,
it would be extremely good if the person was busy
caring for an ill family member because BSE was
of a lower priority. For some, the choices of
good
or bad implied a general judgment that they were
not willing to make. This concern is consistent
with the findings of Gilligan
(1982)
in
her work
on the moral development of women. She presented
one component of the concept of morality as a
recognition of the limitation of judgment in the
absolute sense and an awareness of psychological
and sociological determinants of behavior. This
morality is contextual and is concerned with all
the particulars of the situation.
Third, the Belief and Attitudes Questionnaire
can pose a conceptual challenge to participants.
The acceptability of TRA questions to respondents
has been addressed by other researchers. Mullen,
Hersey, and Iverson
(1987)
reported that partic-
ipants responded to the evaluation items by saying
that they had already answered that question. The
items are similar enough that respondents might
perceive the question to be the same. In our lon-
gitudinal study, the similarity of items often elicited
responses of frustration and sometimes led the
participants to doubt their own abilities in answering
the questions. This perception that the items are

THEORY
OF
PLANNED BEHAVIOR
/
YOUNG ET AL.
143
the same threatens instrument reliability as
it
may
lead to less careful consideration of the item and
might frustrate the participants into falling back
on response sets.
Finally,
in
a study conducted by health profes-
sionals, the effect of social desirability cannot be
underestimated. Studying a health behavior that
is perceived by the woman as important can
in-
troduce a source of bias that may not be captured
by the prescriptions of either the TRA
or
the TPB.
This bias can be amplified when participants are
recruited through health care providers. Despite
the fact that participants were randomly selected,
many mentioned to the interviewers that they
thought their physicians chose them because of
poor performance of BSE. Many women saw the
study as an opportunity to improve their BSE
performance. This expectation that participation
in
the study itself would serve as motivation to
persist in a health promoting activity is not suf-
ficiently accounted for by behavioral, normative,
or perceived control beliefs.
The potential influence of social desirability in
a health promotion study was illustrated
in
a recent
analysis of the usefulness of the TRA
in
prediction
of BSE (Powell-Cope et al.,
in
press). These au-
thors found that the TRA was more predictive of
intention and behavior in a nonintervention study
than in an intervention study designed to influence
attitudes and behavior. It is possible that measures
of beliefs and intentions of women
in
the inter-
vention study reflected
expected
outcomes of the
educational program,
or
what women thought the
researchers believed, and not their true beliefs.
For
example, although the samples were similar,
intention scores were higher
in
the intervention
group. Such distortion could increase error
in
the
dependent variable, result in a less valid measure
of
beliefs and intentions, and therefore a less re-
liable test of the model. In the nonintervention
study
it
is more likely that attitude and intention
measures were more valid and therefore stable
because women responded without anticipating
attitudinal, intentional,
or
behavioral changes.
General principles for conducting research can
enhance the reliability of the results. Accuracy of
responses can be improved by providing thorough
explanations, by ensuring that adequate time is
available, by promoting a relaxed and accepting
interview environment, and by using a visual scale
at the time of verbal questioning. Interviewer
training is most important
in
order to anticipate
technical problems with the administration of the
questionnaire and
to
ensure that adequate and con-
sistent instructions are available for participants.
Ongoing meetings with the research team can im-
prove the effectiveness of the interviewers as this
practice provides an opportunity for trouble-
shooting and enhances awareness of potential
administration problems. Strategies to minimize
social desirability include explanation about the
aggregate nature of data analysis, assurances of
confidentiality, and clear separation of the data
collectors from clinical roles.
Despite the methodological concerns discussed
in this article, the TPB is a useful model for
prediction of
BSE
behavior when carefully op-
erationalized (Powell-Cope et al., in press). The
instrument development process must be repeated
for each different sample, behavior, or social set-
ting. Although this task can be cumbersome,
it
enhances the validity and power of the measures
and is consistent
with
a view of the uniqueness
of humans
in
their specific contexts.
Furthermore, the TPB holds promise in cross-
cultural research, as indicated by Davidson, Jac-
card, Triandis, Morales, and Diaz-Guerrero
(1976).
These authors propose that the model offers an
etic (or universal) structure for the conduct of
research and lends itself to emic operationalization.
In
other words, the process of instrument devel-
opment always is grounded
in
the population of
interest and has relevance
in
diverse groups. The
TPB, therefore, is a potentially important model
in nursing research because
it
allows the inves-
tigator to incorporate individual and cultural dif-
ferences into the design of the study in a systematic
way that is amenable to testing.
An understanding of the factors that influence
a health-related behavior is of potential use
in
the
development of preventive nursing approaches.
Once salient beliefs about a behavior and the most
important social influences have been identified,
interventions can
be
designed to target these factors.
The Theory of Planned Behavior offers a pre-
scriptive approach for the identification of salient
behavioral influences, a necessary first step in the
design of relevant interventions (Lierman et al.,
1990).
REFERENCES
Ajzen,
I.,
&
Fishbein, M.
(1980).
Understanding
attitudes and predicting behavior.
Englewood
Cliffs: Prentice-Hall.
Ajzen,
I.
(1985). From intentions
to
actions:
A
theory
of planned behavior. In
J.
Kuhl
&
J.
Beckmann
(Eds.),
Action conrrol
(pp.
11-39). Berlin:
Springer-Verlag.
Ajzen,
I.,
&
Madden,
T.J.
(1986). Prediction of
goal-directed behavior: Attitudes, intentions, and

144
RESEARCH
IN
NURSING
8
HEALTH
perceived behavioral control.
Journal
of
Exper-
imental Social Psychology, 22,
453-474.
Ajzen,
I.
(1987). Attitudes, traits, and actions:
Dispositional prediction of behavior in personality
and social psychology. In L. Berkowitz (Ed.),
Advances in experimental socialpsychology
(pp.
1-63). New York: Academic Press.
Bandura, A. (1977). Self-efficacy: Toward a uni-
fying theory of behavioral change.
Psychological
Review, 84,
191-215.
Davidson, A.R., Jaccard, J.J., Triandis, H.C.,
Morales, M.L.,
&
Diaz-Guerrero, R. (1976).
Cross-cultural model testing: Toward a solution
of
the etic-emic dilemma.
International Journal
Gilligan, C. (1982).
In
a different voice.
Cambridge:
Harvard University Press.
Kuhl,
J.
(1985). Volitional aspect of achievement
motivation and learned helplessness: Toward a
comprehensive theory of action control. In B.A.
Maher (Ed.),
Progress
in
experimental personality
research
(pp. 99-171). New York: Academic
Press.
Lierman,
L.M.,
Young,
H.M.,
Powell-Cope, G.M.,
OfPsychology,
II,
1-13.
Georgiadou,
F.,
Benoliel, J.Q. (1990). Influencing
breast self-examination in older women.
Com-
municating Nursing Research, 23,
76.
Lierman, L., Young, H.M., Kasprzyk, D.,
&
Ben-
oliel, J.Q. (1990). Measuring Beliefs and Atti-
tudes Predicting BSE.
Nursing Research,
39,
Lierman,
L.,
Kasprzyk, D., Benoliel, J.Q. (in
press). Understanding adherence to breast self-
examination in older women.
Western Journal
of
Nursing Research.
Montano, D.E. (1986). Predicting and understanding
influenza vaccination behavior.
Medical Care,
Mullen,
P.D.,
Hersey, J.C.,
&
Iverson, D.C. (1987).
Health behavior models compared.
Social Science
&
Medicine, 24,
973-981.
Powell-Cope, G.M., Lierman, L.M., Kasprzyk, D.,
Young, H.M., Benoliel,
J.Q.
(in
press). The
Theory of Reasoned Action in prediction of breast
self-examination: A comparison of two studies.
Health Care for Women International.
Triandis, H.C. (1977).
Interpersonal behavior.
Monterey, CA: BrookslCole.
97- 101.
24,
438-453.