|
The role of involvement and a conceptual model
The Role of Involvement and A Conceptual Model
for Optimal Health Communication Strategies
by
Yungwook Kim
Doctoral Student
College of Journalism and Communications
University of Florida, Gainesville, FL 32611
(O)352- 392-6557
(H)352- 846-5152
[log in to unmask]
Submitted to the Communication Theory and Methodology Division
The AEJMC Conference, Chicago
April 1, 1997
ABSTRACT
The Role of Involvement and A Conceptual Model
for Optimal Health Communication Strategies
This study is the incipient stage for formulating a respondent-oriented
conceptual
model for optimal health communication strategies. As the prevailing theory in
involvement research, the Elaboration Likelihood Model (ELM) is discussed. Also
weaknesses of the ELM and other perspectives for involvement research are
investigated. As moderators of the health communication process, involvement is
divided into enduring and situational involvement. Based on the review of the
literature, a
new health communication model and optimal strategies are proposed. The new
model
has two dimensions (enduring involvement and situational involvement) and four
strategies: an affect-evoking, an information-oriented, a cue-emphasizing, and a
balanced-
argument strategy. For the empirical research, a 2 x 2 experiment with 143
undergraduate
students was conducted. Generally, the moderating effects of both enduring
involvement
and situational involvement are supported. Regardless of some deviations from
the
proposed model, optimal strategies fit into the designated involvement level.
Compared to
the ELM, this proposed model accounts for all cases of involvement and also
explains the
inconsistent cases of enduring and situational involvement. As for the health
behavior, this
proposed model has more explanatory power than the ELM.
Title: The Role of Involvement and A Conceptual Model
for Optimal Health Communication Strategies
Introduction
Research has been conducted to clarify the relationship between
motivation and
preventive health behavior (Moorman and Matulich, 1993) and to enhance the use
of
psychological and behavioral variables in health marketing (Zaltman & Vertinsky,
1971).
Involvement has been a critical variable in many advertising, social psychology,
and
consumer behavior studies (Petty & Cacioppo, 1981, 1986). However, little has
been done
with the role of involvement in health communication Health communication
research
typically has focused on single demographic or psychographic variable such as
age,
gender, minority status, and health beliefs (Moorman & Matulich, 1993).
Respondents'
decision-making is critical in understanding effective health communication;
without
understanding respondents' decision-making processes concerning health
communication,
it is impossible to map out an effective health communication strategy.
The general definition of the target group for a communication campaign
has
developed from the passive target group to the active user. But in health
communication,
differences in the relevance of health messages to individuals has not received
much
discussion. Depending on the different levels of involvement in health messages,
optimal
strategies in health communication can be changed drastically. Many moderators
can affect
respondents' motivation for behavior change, such as involvement, need for
cognition,
mood, and personal responsibility (Inman, McAlister, & Hoyer, 1990).
Involvement is one of the most important factors in behavioral change
studies.
The concept of involvement originated from the analysis of different
respondents. Through
the study of involvement, optimal strategies for reaching diverse groups of
respondents in
health communication can be devised. But the previous discussion of involvement
in the
Elaboration Likelihood Model (ELM) is too narrow to explain respondents' diverse
health
communication processes. Petty and Cacioppo (1983, 1986) differentiated
involvement
only as either high or low, and many researchers have criticized the simple
division of the
peripheral and central route (Baker and Lutz, 1996; Cole, Ettenson, Reinke, &
Schrader,
1990; Bitner & Obermiller, 1985). Thus, the purpose of this paper is to propose
a new
involvement-related health communication model for accommodating diverse groups
of
respondents after investigating antecedents of involvement as moderators of
health
communication.
Some health communication models explain both preventive-health behavior
and sick-role behavior. However, many of the health models focus on predicting
behavior
to curb practices that damage one's health. Examples include smoking behavior,
overexposure to the sun, dietary causes of heart disease or cancer, drug abuse,
drunk
driving, safe sex for reducing HIV infection, and mammography. Health
communication
itself has preventive characteristics. The aim of preventive-health
communication is
changing respondents' behavior to prevent disease. This is totally different
from product or
service campaigns. Researchers have been mired in the differences between "to
do"
promotion and "not to do" promotion because health communication is designed to
prevent the risky behavior, rather than to make respondents buy products. Health
communication is more dependent on respondents' judgment, rather than the
strength of
message. Indeed, a better understanding of the respondents is imperative to
solve this
difference.
This paper starts with (1) the discussion of the ELM framework and the
weakness of
the ELM. Then, the paper describes (2) the antecedents of involvement and
research
streams for clarifying varying levels of involvement and finding out moderators
of the
health communication process. With new moderators (enduring and situational
involvement), the paper then (3) describe a new model that differentiates
between
individual and situational relevance and offers four optimal communication
strategies. As a
further discussion of the health communication process, (4) the idea of
subjective norms
is introduced to explain unconvincing cases of a new involvement-related model
in overall
process. This paper concludes by clarifying the attitudinal control of a new
health
communication model and offering suggestions for future research.
From Brinberg and Mcgrath's (1985) the validity network schema (VNS),
this
research can contribute to the substantive domain for the respondents' health
behavior
with a new proposed model.
Literature Review
Involvement in the Elaboration Likelihood Model (ELM)
Elaboration Likelihood Model (ELM)
The ELM can be applied well to individuals' health behavior as a more
general model.
The ELM describes the reception of persuasive health communication (Petty &
Cacioppo,
1981, 1983, 1986). Depending on the level of involvement, respondents follow the
High High Message Elaboration
Motivation Involvement <Central route>
Attitude Behavior
Ability Low change
Involvement Low Message Elaboration
<Peripheral route>
Figure 1.
The Elaboration Likelihood Model of Attitude Change (Modified from Petty &
Cacioppo, 1983)
peripheral route or central route to process information (Figure 1). For
example, a person
who has a family disease history will pay greater attention to a pertinent
health campaign
message (central route), but a person who has no direct experience with a
disease will be
more affected by non-message cues, such as the expertise or attractiveness of a
health
message source rather than message contents (peripheral route). The function of
message
and recipient characteristics has been well documented by this model.
It is not reasonable to assume that everyone would focus on the same
message. Also,
the majority of a given message is not relevant to a given individual. ELM
explains both
issue-relevant and non-issue-relevant attitude change. The central route is
relatively
enduring and predictive of behavior change. The peripheral routes can explain
temporary
and unpredictive of behavior change (Petty & Cacioppo, 1981, 1986; Petty,
Cacioppo, &
Schuman, 1983).
Criticism of the ELM
Petty and Cacioppo's (1981, 1983, 1986) two routes to persuasion do not
explain all attitude change. The peripheral route is an overly broad and vague
definition
High Involvement Elaboration
Motivation
Moderate Involvement Comprehension
Attitude Behavior
Ability change
Low Involvement Focal Attention
Figure 2.
Diagrammatic Depiction of the Relevance-Accessibility Model in Case of
Advertising Message Involvement (Modified from Baker and Lutz, 1996)
of behavioral change because the ELM cannot explain all the cases, such as
moderate level
of involvement (Baker & Lutz, 1988, 1996). The cognitive levels between pure
emotional
response and consideration of simple outer stimuli, such as celebrity reference
and
message credibility, contain different cognitive efforts (In the ELM, all of
these situations
are considered as low involvement).
In the Relevance Accessibility Model (RAM) (Baker and Lutz, 1988, 1996),
the
persuasion process has three route heuristics (Figure 2). Baker and Lutz argue
that the
ELM oversimplifies the cognitive process by failing to describe other possible
approaches
such as moderate level of involvement or inconsistent involvement between
enduring and
situational involvement.
The ELM does not predict how processing will occur when enduring
relevance and
situational relevance are inconsistent (for example, in the case of low personal
interest but
high situational risk). In health communication, types of antecedents of
involvement such
as personal and situational relevance should be considered as moderators of
health
communication because diseases and health messages have different levels of
personal and
situational relevance to different respondents.
The ELM does not take into account the possibility that the function of
cues depends
on respondents' characteristics and situational involvement, rather than whether
or not
cues act centrally or peripherally (Cole, Ettenson, Reinke, & Schrader, 1990).
Also, when
no other central cue is presented, a peripheral cue may act as a central cue
(Bitner &
Obermiller, 1985). Thus, the analysis of central and peripheral routes is
somewhat
inapplicable to health communication because of different levels of enduring
relevance of
health messages.
Other Perspectives on Involvement Research
Involvement has been a critical variable in many persuasion, consumer
analysis, and
advertising theories (Celsi & Olson, 1988; Krugman, 1965; Petty & Cacioppo,
1979,
1981, 1983, 1986). Involvement has been conceptualized in diverse ways. It can
include specific personal characteristics and situational settings (Celsi &
Olson, 1988;
Show, Celsi & Abel, 1990; Haugtvedt & Strathman, 1990).
Consumer behavior research developed the concept of involvement in
advertising
research (Krugman, 1965). Krugman noted that different modes of consumer
information-
processing should affect the effectiveness of advertising. With low involvement,
repetitive
advertisements might be more effective than a dramatic, one-shot advertisement.
With
high involvement, advertising that employs a "conflict of ideas at the level of
conscious
opinion and attitude" might be more effective than repetitive advertising
(Krugman,
1965; Rothschild & Ray, 1974).
In tune with Krugman's definition of the personal "bridging experience,"
involvement is
generally defined as mediating personal relevance in determining the
effectiveness of
communication. But in light of recent studies of the antecedents of involvement,
involvement also can be explained by its other definitions, such as intrinsic
sources of
personal relevance and situational sources of personal relevance, enduring
involvement
and situational involvement (Celsi & Olson, 1988; Houston & Rothschild, 1978;
Zaichkowsky, 1986).
Antecedents of Involvement
In advertising and consumer analysis research, antecedents of involvement
have been
widely discussed to better conceptualize involvement. In conceptualizing
involvement,
Zaichkowsky (1986) proposed three types of antecedents of involvement: personal
characteristics, object characteristics, and situational characteristics.
Personal relevance
related to these three factors is a necessary condition for attention and
attitude change.
Also, respondents use different cues depending on the personal relevance of the
communication or the persuasion process (Petty & Cacioppo, 1979, 1981, 1983,
1986).
With situational manipulations and differences in stimuli, the enduring
personal source
is critical in the communication process. Intrinsic factors in personal
relevance play an
important role in the communication process (Chow, Celsi, and Abel, 1990;
Zaichkowsky,
1986). The introduction of intrinsic personal relevance is a brilliant idea
countering the
uniformity of a communication campaign (for example, personal interest for
safety and
concern about disease are different over all kinds of health messages).
Situational involvement originates in the evaluation of external
conditions. Message
contents or respondents' situations are the primary focus of this research
(Burnkrant &
Sawyer, 1983; Celsi & Olson, 1988; Petty & Cacioppo, 1981, 1986). Situational
antecedents of involvement (situational relevance) and other moderators for
persuasion
(need for cognition, mood, and personal responsibility) are closely related to
this research
stream. (For example, gay men whose sex partners died of AIDS are more
interested in
the HIV/AIDS prevention message compared to people not affected by HIV/AIDS.
Also,
if those men are uneducated, the level of literacy on the message plays a role
in the
attitude change).
Enduring involvement represents the pre-existing experience and knowledge
structure
of the respondents (Celsi & Olson, 1988; Houston & Rothschild, 1978). Product
involvement also is included in this category. In regards to health
communication, personal
concern about risk is deeply related to this concept.
Enduring and Situational Involvement as Moderators of a New Model
Zaichkowsky (1986) explained that personal factors (needs, importance,
interests, and
values), and situational factors (purchase/use and occasion) can serve as
antecedents of
involvement and affect the possible results of involvement. Celsi and Olson
(1988) divided
personal relevance into intrinsic sources of personal relevance (ISPR) and
situational
sources of personal relevance (SSPR). For example, if somebody became interested
in
buying products because of rebates and coupons, he/she is in high SSPR. If
somebody
enjoys shopping, he/she is in high ISPR. ISPR is an "enduring and stable"
structure, but
SSPR is an "dynamic and changeable" structure.
Houston and Rothschild (1978) pioneered these concepts of enduring and
situational
involvement. They found that enduring involvement and situational involvement
affect
"response involvement." In health communication, these research streams can be
treated as
the antecedents of involvement because respondents have different enduring and
situational relevance to the same health communication stimuli.
In conceptualizing this study, enduring and situational involvement can
be defined as
the antecedents of involvement. These antecedents are the moderators of the
health
communication process. Celsi and Olson's (1988) ISPR and SSPR provide similar
definitions for this study. Enduring involvement is the pre-existing
relationship between a
respondent and the stimuli (Houston & Rothschild, 1978; Zaichkowsky, 1985, 1986)
or
enduring personal relevance and stable predisposition in long-term memory (Celsi
&
Olson, 1988). Situational involvement is non-enduring characteristics that give
momentum
to respondents with concern for their behavior (Houston & Rothschild, 1978) or
"dynamic
and changeable felt involvement" (Celsi & Olson, 1988).
Proposing a Health Communication Model and Optimal Strategies
Issues of Involvement as a Moderator of Health Communication
Health behavior research also has some inherent problems in applying
theories related
to involvement. Health communication research typically has focused on how to
send
messages, not on whom to persuade.
Most of the applicable health communication models do not consider the
level of
involvement among respondents. Also, the weakness of health marketing, including
the
social marketing approach, is that it is too dependent on demographic and
psychographic
segmentation. Without considering the level of involvement, a simple
segmentation
strategy would not work.
In previous communication research, two relatively distinct routes to
persuasion have
been the main focus, rather than the function of involvement. Involvement is
defined as the
moderator of motivation and ability in persuasion. But in health communication,
intrinsic
personal relevance of health issues and situational relevance seem more
important than in
other communication situations. For example, an individual whose family members
have
been seriously injured by a drunk driver or one who generally is more
safety-oriented will
be more interested in anti-drunk-driving health communication messages, than an
individual who does not have any related experience or is a carefree person.
Typically, in health campaigns respondents are not involved with health
issues.
As previous involvement research has indicated, enhancing respondents'
involvement is an
effective strategy in health campaigns (Chaffee and Roser, 1986). But this task
cannot be
systematic unless the antecedents of involvement are clarified. For an exact
diagnosis, an
in-depth analysis of involvement is needed.
There has not been much research that directly investigates the role of
involvement in
health communication. The purpose of the new health communication model is to
overcome the limitations of other health communication research. Consumer
behavior and
advertising researchers already have tried to investigate the sources of
involvement and
have studied the relationship between involvement and consumer behavior in a
more
detailed way.
Proposed New Health Communication Model and Optimal Strategies
The biggest problem of previous health models is their extremely
straightforward
approach to health communication. The message-sender-oriented approach does not
always account for the diversity of respondents. Also, Petty and Cacioppo's
(1981, 1983)
two heuristics to persuasion are not enough to provide for optimal health
communication.
In fact, health communication is deeply related to personal and situational
involvement.
The proposed health communication model has two dimensions of involvement:
enduring and situational involvement (Figure 3). Depending on the degree of each
type of
Situational Involvement
Low High
E
I
n n Cell
1 Cell 2
d v Low
Affect-evoking Cue-
u o
Strategy emphasizing
r l
Strategy
i v
n e Cell 3 Cell 4
g m Information- Balanced-
e High oriented argument
n Strategy Strategy
t
t
Figure 3.
Diagrammatic Depiction of Proposed Health Communication Model and Optimal
Strategies for each cell.
involvement, there are four different optimal strategies that correspond to the
four
involvement situations: low-enduring involvement and low-situational
involvement, high-
enduring involvement and low-situational involvement, low-enduring involvement
and
high-situational involvement, high-enduring involvement and high-situational
involvement.
The different levels of involvement and optimal strategies were concerned with
enhancing
preventive-health behavior. Thus, sick-role behavior is not discussed in this
model.
H1: Enduring involvement and situational involvement moderate the effect
of
strategies in health communication.
Cell 1. An Affect-Evoking Strategy.
This is the most effective model for cases of low-enduring involvement
and low-
situational involvement. When both types of involvement are low, the
individual's
motivation to process health communication messages is very low. For example, a
health
organization is preparing a health campaign to inform the public about the risk
of
overexposure to the sun. When the respondents of the health-communication
message are
in low enduring and situational involvement groups, an advertising campaign that
includes
a horrible scene of skin cancer patients and gruesome background music and
focuses
on pure emotion will be more effective than an message-oriented advertising or
celebrity
advertising campaigns. But the level of affect-evoking should not be too strong.
H2a: An affect-evoking strategy is more effective in the low enduring and
situational
involvement than in other involvement levels.
Cell 2. A Cue-Emphasizing Strategy
This category is needed to enhance the effectiveness of health
communication when
respondents are low in enduring involvement but high in situational involvement.
When a
person generally does not care about his health and does not concern about
safety, health
communication should focus on effective presentation of the health risk. By
using celebrity
advertising or endorsement campaign, health communicators effectively inform
these
individuals about the risk of unhealthy behavior. If they start to perceive the
risks of
unhealthy behavior, the effect of health communication will be greatly enhanced.
Source
credibility and endorsement are the most important factors. For example, if a
tourist in
Florida for only a few days would be low-enduring involvement and
high-situational
involvement in the risk of overexposure to the sun, he would not be concerned
about
getting a sunburn. This tourist will be more likely to respond to a
cue-emphasized
campaign than a heavy information appeal. He does not care a lot about health
issues in
general. In this case, a celebrity who has had a skin cancer experience can help
him
understand the risk. Celebrity campaign give a attention-grabbing function to
only
situationally involved respondents. This strategy will be more effective than
merely
presenting detailed information or showing a horrible scene of skin cancer
patients.
H2b: A cue-emphasizing strategy is more effective in the low enduring and
high
situational involvement than in other involvement levels.
Cell 3. An Information-Oriented Strategy
This strategy is very effective in cases of high-enduring involvement
and low-
situational involvement. This strategy is very important in health communication
because
other marketing campaigns rarely use the information-oriented campaign. An
information-
oriented strategy is focused on the distribution of information to people who
need the
information at that moment, rather than on message quality and interest.
For example, consider an individual who originally did care about health
issues
because of cancer history in his family. But he is very healthy even though he
concerns
about health issues. The man would be classified as high in enduring involvement
but
low in situational involvement. In this case, providing as much information as
possible
about skin cancer is more important than the message quality because this person
wants to
satisfy his specific need for health information. Only respondents in high
enduring
involvement would be readily equipped to absorb an information message. In this
case, a
informative approach is more effective than horrible scenes or celebrity
campaigns.
H2c: An information-oriented strategy is more effective in the high
enduring and
low situational involvement than in other involvement levels.
Cell 4. A Balanced-Argument Strategy
This strategy should be very effective for cases of high-enduring
involvement and
high-situational involvement. This is in line with the central route of the ELM
in cases of
high involvement (Petty & Cacioppo, 1981, 1983). Also, this cell is related to
the
founding that the central and peripheral routes intertwine under high
involvement (Lutz,
MacKenzie, Belch, 1983).
Message cues are still important in reaching high involvement respondents
because
they help to get attention. The most important part of this strategy is the
balanced
structure of message cues and argument. Both information and cues can grab the
attention
of highly involved respondents, but a one-way approach that emphasizes only one
of these
strategies will lessen interest in a health message because a person with high
enduring
and situational involvement reacts to both a cue and information. A balanced
message
argument that uses strong arguments buttressed with statistical comparison and
some
external cues optimally fits this kind of involvement. For example, if a
health-concerned
person finds out that his neighborhood has several cases of skin cancer, he is
in a state of
high enduring and situational involvement. In this case, health campaign
advertising which
emphasizes a strong relationship between overexposure to the sun and skin cancer
what is
endorsed by doctors, will be more effective than other strategies.
H2d: A balanced-argument strategy is more effective in the high enduring
and high
situational involvement than in other involvement levels.
Methodology
Subjects and Design
A total of 143 men and women, enrolled in the undergraduate class of
"Television
and American Family" participated in the study to earn extra credits. Experiment
was a
2 (enduring involvement: high or low) * 2 (situational involvement: high or low)
design.
The enduring involvement and situational involvement groups were between-group
subjects.
Experiment Procedure
To test the research hypotheses, safe sex for preventing HIV/AIDS was
selected as
the domain of involvement. This issue has been discussed in diverse ways among
college
students. Two research processes were conducted. In the first process, a
screening test
with 143 college students was executed. The screening questionnaire measured
subjects'
enduring involvement about safe sex. 124 out of 143 students, who represented
high and
low enduring involvement levels, was selected to conduct the experiment. Before
testing
stimulus materials, subjects received one of two manipulations of situational
involvement
by the different kinds of information.
Stimulus materials and Pretest
Four different advertising strategies were created: an affect-evoking
strategy, an
information-oriented strategy, a cue-emphasizing strategy, and a balanced
argument
strategy. The affect-evoking strategy was designed to evoke subjects' emotional
response.
An advertisement contained mostly a horrible scene of AIDS patients and a
headline. This
attempted to facilitate affective response by a visual scene.
The information-oriented strategy was designed to transfer information to
subjects as
much as possible about HIV/AIDS. An advertisement focused only on the
distribution of
information which contained a headline and a lengthy copy.
The cue-emphasizing strategy was using celebrity for enhancing the risk
of
HIV/AIDS. This advertisement used little information about HIV/AIDS, but
introduced
a college football starplayer to reinforce subjects' perception about the
seriousness of
HIV/AIDS.
The balanced-argument strategy was designed to balance the message cue
and
information. An advertisement contained endorsement of an expert and detailed
information about HIV/AIDS.
Pretests were conducted to test these four sets of advertisements with
separate 40
subjects. The differences of four advertisements were confirmed by pretests.
Measuring and Manipulation of Involvement
Zaichkowsky's (1985) Personal Involvement Inventory (PII) was used to
measure
subjects' enduring involvement to safe sex. Her personal category of involvement
represents "inherent interests, values or needs that motivate one toward the
object (p.
342)." A semantic differential scale was developed to capture the level of
involvement. PII
includes these 20 items such as important-unimportant, irrelevant-relevant,
valuable-
worthless, trivial-fundamental, and significant-insignificant. Subjects'
response to these
items for safe sex were summed, producing a minimum score of 20 and maximum
score of
140. Subjects were divided into the low, moderate, and high level by scoring.
Only the low
and high level of enduring involvement subjects were selected for next steps.
The subjects in the low situational involvement were informed that they
would be
asked to choose the most persuasive advertisement from four different
advertisements.
But no information and situational stimuli were given. The subjects in the high
situational
involvement were manipulated by three steps. In the first step, the subjects
read the article
which inform the fact that 10 percent of students in their college were infected
HIV. In the
second step, the experimenter make them assume that their close friend got
infected with
HIV recently. In the third step, they were informed that their choice of
advertising material
would be used in the campus safe sex campaign.
Measures
Selection of stimulus materials. Subjects in each group received the four
advertising
materials which were manipulated by the level of enduring and situational
involvement.
After exposed to all advertisements, subjects were asked to select the most
persuasive
advertisement.
Evaluation of each advertisement. After selecting the most persuasive
advertisement,
subjects were asked to evaluate each advertisement. 7-point scale (1= absolutely
not
persuasive, 7= very persuasive) was used.
Analysis
The parameters of model (H1) was analyzed with a simple chi-square
method and a
logit loglinear model were used for investigating the relationship between the
level of
involvement and optimal strategies, and moderating effects of enduring and
situational
involvement. The logit loglinear model is used for formulating a model between
dependent
categorical variables and independent categorical variables. In this research,
an
independent variable is the levels of involvement and a dependent variable is
optimal
strategies. Measures of association also were used to test the model adequacy.
For finding
the optimal strategy for each cell (H2), ANOVAs with Multiple Classification
Analysis
(MCA) were used.
Outcomes
Stability of Enduring involvement and Manipulation Check
The stability of the enduring involvement measure was tested by t-tests
for differences
between the high and low enduring involvement.
For the manipulation check of subjects' level of situational involvement,
their felt
involvement were measured by two seven-point scales after the experiment. Two
scales
were constructed by importance and involvement.
After the stability and manipulation check, both enduring and situational
involvement
showed the significant difference between high and low subjects.
Hypothesis Tests
H1: Enduring involvement and situational involvement moderate the effect of
strategies
on a health behavior model.
To test H1, a chi-square test was used to measure the strength and nature
of
relationships between two categorical variables. To test the association between
involvement variables and optimal strategies, Pearson chi-square statistic was
used. From
the crosstabulation (Table 1), the null hypothesis that involvement variables
and optimal
strategies are independent can be rejected (Pearson (2 =17.12, p = 0.04). But in
the low
enduring involvement and high situational involvement, the affect-evoking
strategy is more
selected than a cue-emphasizing strategy. In the high enduring involvement and
low
situational involvement, the balanced-argument strategy has the same selections
with the
information-oriented strategy. Also we do not know the strength of the
association. To
quantifying the strength of the relationship between two categorical variables,
The logit
loglinear model (Table 2) was used (Norusis, 1994). While the loglinear models
can
examine the interrelationships among categorical variables, the logit loglinear
analysis can
model the relationship between one of more dependent categorical variables and
independent categorical variables. In this research, each level of involvement
is an
independent variable and each strategy is a dependent variable.
H1 was generally supported. The chi-square test shows the significant
association
((2 =17.12, p = 0.04). Also enduring involvement and situational involvement as
independent variables showed relative strength respectively on optimal
strategies. Table 2
shows the parameter estimate summary. Numbers are parameter estimate Lambda's
and
numbers in parentheses are parameter estimates.
The logit loglinear model shows moderating effects of involvement in the
proposed
health behavior model. the parameters involving high enduring and situational
involvement
are set to 0, so they serve as the basis for comparison. For example, the value
of e( in low
enduring involvement and low situational involvement and the affect-evoking
strategy is
3.06. The affect-evoking strategy was selected almost 3.06 times more likely in
the low
enduring and low situational involvement than in the high enduring and
situational
involvement (e( =1). The information-oriented strategy in the high enduring
involvement
and low situational involvement (e( =2.08) is almost three times more selected
than in the
Table 1. Crosstabulation and Chi-square Analysis for Involvement and Optimal
Strategies
Low EI Low EI High EI
High EI Total
High SI High SI Low SI
High SI
Affect-evoking 14 14 8
8 44
Cue-emphasizing 6 11 5
2 24
strategy
Information-oriented 2 5 9 5
21
Strategy
Balanced-argument 7 6 9 13
35
Strategy
Total 29 36 31 28 124
Note. Pearson (2 =17.12, p = 0.04
Table 2. Parameter Estimate Summary in a Logit Loglinear Model
Affect-evoking Cue Information
Balanced
Strategy -emphasizing -oriented
-argument
Low EI 1.12 (3.06) 1.54 (4.66) -.20 (0.82) 0 (1)
Low SI
Low EI 1.27 (3.56) 2.26 (9.58) .73 (2.08) 0 (1)
High SI
High EI .35 (1.42) 1.14 (3.13) .89 (2.44)
0 (1)
Low SI
High EI 0 (1) 0 (1) 0
(1) 0 (1)
High SI
Note. EI = Enduring involvement
SI = Situational involvement
* Parameter estimate lambda (()
** Parameter estimate (e()
low enduring and low situational involvement (e( = 0.82). From the parameter
estimate
summary (Table 2), we can assume the strength of relationship between two
categorical
variables. Also the moderating effects of involvement on optimal strategies can
be
interpreted.
H2a: An affect-evoking strategy is more effective in the low enduring and
situational
involvement than in other involvement level.
H2b: A cue-emphasizing strategy is more effective in the low enduring and high
situational involvement than in other involvement levels.
H2c: An information-oriented strategy is more effective in the high enduring and
low
situational involvement than in other involvement levels.
H2d: A balanced-argument strategy is more effective in the high enduring and
high
situational involvement than in other involvement levels.
H1 investigated the moderating effects of both enduring and situational
involvement
in the proposed health communication model through the logit loglinear model. H2
Table 3. ANOVAs for the means of each strategy in each involvement level
Low EI Low EI High EI High
EI F p
High SI High SI Low SI
High SI
Affect-evoking 6.10 4.97 5.38 5.00
2.56 0.05
(H2a)
Cue-emphasizing 4.55a 5.67b 5.03 4.68c
4.01 0.009
Strategy
(H2b)
Information 4.07 4.19 5.26
4.79 3.11 0.03
-oriented
(H2c)
Balanced-argument 4.62d 4.53e 4.61f 5.71g
4.18 0.007 Strategy
(H2d)
Note.- The means between a-b and b-c indicate the significant differences
between two
groups from the modified LSD test.
- The means between d-g, e-g, and f-g indicate the significant
differences between
two groups from the modified LSD test.
- H2a and H2c have no difference between groups.
* 1-7 scale (From 1= Absolutely not persuasive to 7= very persuasive)
predicts the effectiveness of each strategy in each level of involvement using
one-way
ANOVAs.
With each level of involvement as a factor, Table 3 shows the partial
support for H2.
All ANOVA models show significant differences (Table 3). But, in the multiple
classification analysis (MCA, modified LSD test was used). For each ANOVA,
difference
of means between different involvement groups are supported in only two levels
of
involvement. For H2a, ANOVA shows the marginally significant difference between
groups (F= 2.56, p=0.05). But no two different involvement group are
significantly
different at 0.05 significance level. For H2b, ANOVA shows the significant
difference (F=
4.009, p= 0.009). From the result of MCA, three groups show significant
difference
between groups. For H2c, ANOVA shows the significant difference between groups
(F=
3.11, p=0.03). But no two different involvement group are significantly
different at 0.05
significance level. For H2d, ANOVA shows the significant difference (F= 4.18, p=
0.007).
From the result of MCA, all four groups show significant differences between
groups.
Discussion
This research began with recognizing the individual differences in
accepting health
communication. For example, let us assume that somebody is designing an internet
homepage for informing people of the risk of getting the common cold in the
winter. How
can the homepage designer know the involvement level of the expected
respondents? If it
is assumed that most respondents have low enduring and high situational
involvement,
what is the optimal strategy for that involvement category?
Very little research has been conducted to assess the effectiveness of
health
communication from the viewpoint of respondents who are the main subject of
health
communication. This research investigated sources of involvement as moderators
of the
communication process. From two dimensions of involvement, four optimal
strategies are
proposed: the affect-evoking, the information-oriented, the cue-emphasizing, and
the
balanced-argument strategy.
The results of the experiment are generally supportive of the
relationship between the
proposed health behavior model and optimal strategies. H1 test shows the
moderating
effects of involvement in the proposed health behavior model. Chi-square test
shows the
significant association between involvement and optimal strategies. But in the
low
enduring involvement and high situational involvement, the affect-evoking
strategy is more
selected than the cue-emphasizing strategy. The affect-evoking strategy has
overall high
selections over all four involvement levels. These results originate from the
sample
selection. Samples in this study are undergraduate students. They seem more
responsive
to the affect-evoking approach than other respondents do. Also in the high
enduring
involvement and low situational involvement, the balanced-argument strategy has
the same
selections with the information-oriented strategy. In the high enduring
involvement, the
balanced-oriented strategy has an overall effect because information and a cue
are
balanced in this strategy. From the logit loglinear model, we can find the
moderating
effects of involvement on optimal strategies. For example, the cue-emphasizing
strategy is
selected 9.58 times more than the balanced-argument strategy. The logit linear
model does
not show the exact parameters as we expected. But generally, optimal strategies
show
high parameter estimates in their specific situation.
H2 test indicates the effectiveness of each optimal strategy in the
designated
involvement situation. Even though only the cue-emphasizing and the
balanced-argument
strategy indicate the difference between groups from MCA, all four ANOVAs for
each
strategy show the significant difference in the models. These results are in
line with the
previous outcomes of the Chi-square and the logit loglinear model test.
Generally, the moderating effects of both enduring involvement and
situational
involvement are supported. Even though some deviations from the proposed model
are
noticed, optimal strategies fit into the designated involvement level.
Compared to the ELM, this proposed model accounts for all cases of
involvement
and also explains the inconsistent cases of enduring and situational
involvement. As for
health behavior, this proposed model has more explanatory power than the ELM.
Suggestion and Limitation
These optimal strategies are not independent. In reality, these strategies
can be used
collaboratively or collectively to account for diverse respondents and
involvement
situations.
This research did not clarify the consistency of attention,
comprehension, attitude,
and behavior. Future research should build up the health behavior model using
LISREL (Joreskog & Sorbom, 1993) to verify relationships between those
variables.
The strengths and weaknesses of enduring and situational involvement were
not
discussed. It is assumed that situational involvement has a greater effect on
attention and
comprehension in health behavior than enduring involvement. Future studies
should clarify
this hypothesis.
The behaviors of gay men are deeply related to the deviant behaviors.
Also subjective
norm described by Fishbein and Ajzen (1980) can be another factor in deciding
their
behavior. But the possible relationship between attitude and subjective norm was
not
discussed in this research.
Also, the relationship between involvement levels and subjective norm was
not
handled. Because this study starts with the ELM framework, subjective norm was
outside
the framework. However, the role of subjective norm seems pretty important in
the
comprehensive health behavior process..
(Proposed Health Model) (Depending on Dominance)
Motivation
Attitude
Attention Comprehension (W1)
Behavior Behavior
-------------- Intention
Ability
Subjective
Norm(W2)
Four involvement levels Situation specific
predictors
1. Low EI, Low SI
2. Low EI, High SI Attitudinal control:
W1>W2
3. High EI, Low SI ------- X-------Normative control: W1<W2
4. High EI, High SI
Figure 4.
The assumed Health Communication Process and the Diagrammatic Depiction of the
Comprehensive Health Behavior Change Process
The whole health behavior process can be determined by either attitudinal
control or
normative control because health behaviors contain deviant behaviors and strong
peer
group pressures (Figure 4). In this study, optimal strategies only cover
attitudinal control.
To make up the weakness of the proposed model, the health communication model
should
consider changing the strength of normative belief and the motivation to comply
under
normative control situation in future study.
Reference
Ajzen, I., & Fishbein, M. (1977). Attitude-behavior relations: A theoretical
analysis and
review of empirical research. Psychological Bulletin, 84, 888-918.
Ajzen, I., & Fishbein, M. (1980). Understanding Attitudes and Predicting Social
Behavior,
Englewood Cliffs, NJ: Prentice-Hall.
Alba, J. W., & Hasher, L. (1983). Is memory schematic?. Psychological Bulletin,
93, 203-
231.
Andrews, J. C., & Shimp, T. A. (1990). Effects of involvement, argument strength
and
source characteristics on central and peripheral processing of advertising.
Psychology
& Marketing, 7, 195-214.
Baker, W. E., & Lutz, R. (1988). The relevance-accessibility model of
advertising
effectiveness. in S. Hecker and D. W. Stewart (Eds.) Nonverbal
Communication in
Advertising (pp. 59-84). Lexington, MA: Lexington Books.
Baker, W. E., & Lutz, R. (1996). An empirical test of the
relevance-accessibility model of
advertising effectiveness. Unpublished manuscript.
Bandura, A. (1986). Social Foundations of Thoughts and Social Cognitive Theory.
Englewood Cliffs, NJ: Prentice-Hall.
_____. (1994). Social cognitive theory of mass communication. In J. Bryant & D.
Zillmann (Eds.) Media Effect (pp. 61-90). Hillsdale, NJ: Lawrence Erlbaum
Associates.
Becker, M. H. (1974). The health belief model and personal health behavior.
Health
Education Monograph, 2, 324-508.
Bitner, M. J., & Obermiller, C. (1985). The Elaboration Likelihood Model:
Limitations and extensions in marketing. Advances in Consumer Research.
12, 420-425.
Bloch, P. H. (1981). An exploration into the scaling of respondents' involvement
with a
product class. Advances in Consumer Research, 6, 61-65.
Boechner, L. S., Kohn, H., and Rockwell, S. K. (1990). A risk-reduction
nutrition course
for adults. Journal of the American Dietetic Association, 90, 260-263.
Bowen, L., & Chaffee, S. H. (1974). Product involvement and pertinent
advertising
appeals, Journalism Quarterly, 51, 613-621.
Brinberg, D. and McGrath, J. E. (1985). Validity and the research process.
Beverly Hills,
CA: Sage.
Burnkraft, R. E., & Sawyer, A. G. (1983). Effects of involvement and message
content on
information-processing intensity. In R. J. Harris (ed.)
Information-Processing Research
in Advertising (pp. 43-64). Hillsdale, NJ: Lawrence Erlbaum Associates.
Cameron, G T., & Schleuder, J. S. (1990, June). The effects of involvement and
knowledge on recall and recognition of persuasive statements. Paper
presented at the
International Communication Association.
Celsi, R. L., & Olson, J. C. (1988). The role of involvement in Attention and
comprehension process. Journal of Consumer Research, 15, 210-224.
Chaffee, S. H., & Roser, C. (1986). Involvement and the consistency of
knowledge,
attitudes, and behaviors. Communication Research, 13, 373-399.
Chow, S., Celsi, R. L., & Abel, R. (1990). The effects of situational and
intrinsic sources
of personal relevance on brand choice decisions. Advances in Consumer
research, 17,
755-760.
Cohen, J. C. (1983). Involvement and you: 1000 great ideas. Advances in Consumer
Research, 10, 325-328.
Cole, C., Ettenson, R., Reinke, S., & Schrader, T. (1990). The Elaboration
Likelihood
Model (ELM): Replication, extentions and some conflicting findings.
Advances in
Consumer Research, 17, 231-236.
Craik, I. M., & Lockhart, R. A. (1972). Levels of processing: A framework of
Memory
research. Journal of Verbal Learning and Verbal Behavior, 11, 671-684.
Eagly, A. H., & Chaiken, S. (1993). The Psychology of Attitudes. Fort Worth, TX:
Harcourt Brace Jovanovich College Publishers.
Green, K, Rubin, D. L., & Hale, J. L. (1995). Egocentrism, message explicitness,
and
AIDS message directed toward adolescents: An application of the theory of
reasoned
action. Journal of Social behavior and Personality, 10. 547-570.
Haefner, D. P. (1974). The health belief model and preventive dental behavior.
Health
Education Monographs, 2, 420-432.
Haugtvedt, C. P. & Strathman, A. J. (1990). Situational product relevance and
attitude
persistence. Advances in Consumer Research, 17, 766-769.
Inman, J. J., McAlister, L. M., & Hoyer, W. D. (1990). Promotion Signal: Proxy
for a
price cut?. Journal of Consumer behavior, 17, 74-81.
Janz, N. K., & Becker, M. H. (1989). The health belief model: A decade later.
Health
Education Quarterly, 11, 1-47.
Johnson, D. W., & Johnson, R. T. (1985). Nutrition education: A model for
effectiveness,
a synthesis of research, Journal of Nutrition Education, 17, S1-S44.
Joreskog, K. G., & Sorbom, D. (1993). LISREL 8: A Guide to the Program and
Applications. Chicago: Scientific Software International.
Kirscht, J. P. (1974). The health belief model and illness behavior. Health
Education
Monographs, 387-408.
Kotler, P. (1975). Marketing for Nonprofit Organization. Englewood Cliffs, NJ:
Prentice
Hall.
Kotler, P. & Zaltman, G. (1971). Social marketing: An approach to planned social
change.
Journal of Marketing, 35, 3-12.
Krugman, H. E. (1965). The impact of television advertising: Learning without
involvement. Public Opinion Quarterly. 29, 349-356.
Lastovicka, J. L., & Gardner, D. M. (1979). Components of involvement. In J. E.
Maloney & B. Silverman (Eds.) Attitude Research Play for High Stakes (pp.
53-73).
Chicago, IL: American Marketing Association.
Lefebvre, R. C., & Flora, J. A.(1988). Social Marketing and public health
intervention.
Health education Quarterly, 15, 299-315.
Lutz, R. J. (1973). The role of attitude theory in marketing. in H. Kassarjian
and T.
Robertson (Eds.) Perspectives in Consumer Behavior (pp. 317-339). Englewood
Cliff,
NJ: Prentice Hall.
Lutz, R. J., MacKenzie, S. B., and Belch, G. E. (1983). Attitude toward the ad
as a
mediator of advertising effectiveness: Determinants and consequences.
Advances in
Consumer Research, 10, 532-539.
Maiman, L. A., & Becker, M. H. (1974). The health belief model: Origins and
correlates
in psychological theory. Health Education Monographs, 2, 336-353.
Mitchell, A. A. (1979). Involvement: A potentially important mediator of
consumer
behavior. Advances in Consumer Research, 6, 191-196.
Moorman, C. (1990). The effects of stimulus and consumer characteristics on the
utilization of nutrition information. Journal of Consumer Research, 17,
362-374.
Moorman, C., & Matulich, E. (1993). A model of respondents' preventive health
behaviors:
The role of health motivation and health ability. Journal of Consumer
Research, 20,
208-228.
Petty, R. E., & Cacioppo, J. T. (1979). Issue involvement can increase or
decrease
persuasion by enhancing message-relevant cognitive responses. Journal of
Personality
and Social Psychology, 37, 1915-1926.
_________. (1981). Attitudes and Persuasion: Classic and Contemporary
Approaches.
Dubuque, Iowa; Wm. C. Brown.
_________. (1983). Central and peripheral routes to advertising effectiveness:
The
moderating role of involvement. Journal of Consumer Research, 10, 135-146.
_________.(1986). Communication and Persuasion: Central and peripheral routes to
Attitude Change. New York: Springer-Verlag.
_________, Goldman, R. (1981). Personal involvement as a determinant of
argument-
based persuasion. Journal of Personality and Social Psychology, 5, 847-55.
_________, Schuman, D. (1983). Central and peripheral routes to advertising
effectiveness: The moderating role of involvement. Journal of Consumer
research, 10,
135-46.
Rosenstock, I. M. (1974). The health belief model and preventive behavior.
Health
Education Monographs, 2, 354-386.
Rothschild, M. L., & Ray, M. S. (1974). Involvement and political advertising
effects.
Communication Research, 1, 264-285.
Sutton, S. M., Balch, G. I., and Lefebvre, R. C. (1995). Strategic questions for
consumer-
based health communications. Public Health Reports, 110, 725-733.
Walsh, D. C., Rudd, R. E, Moeykens, B. A., and Moloney, T. W. (1993). Social
marketing for public health. Health Affairs, 12, 104-119.
Wicker, A. W. (1971). An examination of the other variables' explanation of
attitude-
behavior inconsistency, Journal of Personality and Social Psychology, 19,
18-30.
Wilkie, W. L. (1994). Consumer Behavior. New York: John Wiley & Sons.
Zaichkowsky, J. L. (1985). Measuring the involvement construct. Journal of
Consumer
Research, 12, 341-352.
Zaichkowsky, J. L. (1986). Conceptualizing involvement. Journal of Advertising,
15,
4-14.
Zaltman, G., & Vertinsky, I. (1971). Health service Marketing: A suggested
model.
Journal of Marketing. 35, 19-27.
|
