I ‘1

run.)
3.5., w

Cu... .. I

i.

: 3‘. Q
292;.

‘ I ,
v . ,
I
33.x.
y.1r

 

!
4ft:

RSITY LIBRARIES

llllllllllWilllllllllllllll Hill ll

 

 

 

 

 

 

This is to certify that the

dissertation entitled

THE ROLES OF MESSAGE COMPREHENSIBILITY AND
COGNITIVE SCHEMATA IN THE PROCESSING OF TELEVISED
POLITICAL ADVERTISEMENTS: WHEN FUZZY WUZZY WAS THE MESSAGE

presented by

MICHAEL E . STEELE

has been accepted towards fulﬁllment
of the requirements for

Ph.D. degree in Mass Media Ph.D. Program

 

 

7 ﬂ):- professor

Date

 

4/5/95

MSU i: an Affirmative Action/Equal Opportunity Institution 0-12771

 

LIBRARY
Michigan State
Unlverslty

 

 

 

PLACE ll RETURN BOXtonmwoﬂTbmﬁom ywrncord.
TO AVOID FINES Mum on or before date duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU loAn Afﬁrmative Action/Equal Oppommy mum

WM!

 

THE ROLES OF MESSAGE COMPREHENSIBILITY AND
COGNITIVE SCHEMATA IN THE PROCESSING OF TELEVISED
POLITICAL ADVERTISEMENTS: WHEN FUZZY WUZZY HAS THE MESSAGE
BY

Michaei E. Steeie

A DISSERTATION

Submitted to Michigan State University in partiai
fuifiliment of the requirements for the degree of

DOCTOR OF PHILOSOPHY

Mass Media Ph.D. Program
1993

ABSTRACT
THE ROLES OF MESSAGE COMPREHENSIBILITY AND
COGNITIVE SCHEMATA IN THE PROCESSING OF TELEVISED
POLITICAL ADVERTISEMENTS: WHEN FUZZY WUZZY WAS THE MESSAGE
BY

Michael E. Steele

A stream of research in political communication has
focused on the active processing of political messages.
The present research continues this line of inquiry. The
experiment was designed to examine how audience cognitive
schemata and message comprehensibility affect message cue
utilization and media message induced attitude change.

The study employs a 2 (cognitive schemata: image vs.
issue) X 2 (message comprehensibility: high vs. low)
factorial design. Subjects’ default processing strategies
were determined in Stage I of the study by recording their
responses to a series of extant political broadcast
commercials. In Stage II of the study, message
comprehensibility was manipulated in political broadcast
commercials designed especially for the study. The video
track for the commercial was crossed with a high
comprehensibility audio track and a low comprehensibility
audio track, producing two versions of the stimulus
commercial.

As predicted: Both image and issue processors encoded

more visual information under low comprehensibility

message conditions, rated low comprehensibility messages
less favorably and as less pleasant. Furthermore, issue
processors rated low comprehensibility messages less
favorably and as more incomprehensible than did image
processors.

Several predictions concerning the effects of message
comprehensibility and audience cognitive schemata on
candidate and message evaluation and on vote likelihood
were not supported. However, the patterns of means were
in the predicted directions. In light of the non-robust
nature of the comprehensibility manipulation, the failure
to reach significance was not surprising. Theoretical and

practical implications of these findings are explored.

Dedicated to
the Fight for Animal Rights and
the Memory of Sweet Pea

iv

ACKNOWLEDGEMENTS

I would like to express my sincerest appreciation and
gratitude to the members of my dissertation committee, Dr.
Gina Garramone, Dr. James Gaudino, Dr. E. Lincoln James
and Dr. Stephen Lacy, for their guidance, encouragement
and patience.

To Dr. Garramone, my dissertation director and mentor,
I extend my deepest appreciation for her intellectual
inspiration, guidance and friendship. Without her
unflagging support, this dissertation never would have
seen the light of day.

For their intellectual insights, critical analyses and
occasional, thoroughly enjoyable beer quaffing episodes, I
extend my deepest regards to my remaining committee
members. It made the long journey a whole lot more

pleasant.

TABLE OF CONTENTS

Page

LIST OF TABLES ...................................... Vlll

LIST OF EXHIBITS .................................... x
CHAPTER I

INTRODUCTION .......................... ....... .. 1

REVIEW OF RELEVANT LITERATURE .................. 4
Political Information and its Processing .. 4
Research on Affect ........................ 6
Information Processing and Cognitive Schemata 9
Message Comprehension ..................... 19

CHAPTER II

METHODOLOGY ... ...... ........................... 28
Overview .. ..... ........................... 28
Subjects .................................. 28
Stage I ..... ...... ........................ 29

Procedure ............................ 29
Stimulus Material .................... 30
Content Analysis of Post-Message
Responses ......... ................... 31
Stage II .................................. 32
Dependent Measures ................... 35
Cognitive Responses ............. 37
Candidate Evaluation ............ 37
Vote Likelihood ................. 38
Message Evaluation .............. 38
Processing Effort ............... 40
Message Pleasantness ............ 40
Manipulation Check .............. 41

CHAPTER III

RESULTS ........................................ 43
Manipulation Check ........................ 43
Hypothesis 1A ............................. 43
Hypothesis 1B ............................. 45
Hypothesis 2 .............................. 45
Hypothesis 3A ............................. 48
Hypothesis 38 ............................. 51
Hypothesis 4A ............................. 55
Hypothesis 48 ............... ....... ....... 55
Hypothesis 5A ............................. 58

vi

Page

Hypothesis 58 ............................. 63
Hyp°th6818 6A I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 63
Hypothesis 68 ..... .. ............. ......... 63
Additional Analyses ..... ........ . ......... 66

CHAPTER IV
DISCUSSION I I I IIIIII I I I I I I I I I I I IIIIIIIIIIIIIII I I 71

Limitations ............................... 72
Interpretation of Results ................. 74
Orientation Consistency .............. 74
Message Evaluation and Pleasantness .. 76
Candidate Evaluation and the Vote .... 77
Processing Effort .................... 79

Additional Analyses ..... ............. 80

summary IIIIIIIII I I I I I I I I I I I I I I I I I I I I I 81

Future Research Priorities .... ..... ....... 81
REFERENCES .. ...... ....... ................ ...... ..... 85
APPENDIX ............................................ 94

vii

TABLE

10.

11.

12.

13.
14.

15.

16.

17.

LIST OF TABLES

PAGE
Candidate Evaluation Scale .... ............... . 39
Message Evaluation Scale ................ ...... 41
Manipulation Check of Message
Comprehensibility ........................ ..... 44
Visually-Oriented Cognitive Responses
by SChema IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 46
Issue-Oriented Cognitive Responses by Schema .. 47
Visually-Oriented Cognitive Responses by
comprehenSibi11ty IIIIIIIIIIIIIIIIIIIIIIIIIIIII 49
Message Pleasantness by Schema/
Comprehensibility ............ ................. 50
Message Pleasantness by Schema/
Comprehensibility ............................. 52
Message Evaluation by Schema/
Comprehensibility ................ ..... ........ 53
Message Evaluation by Schema/
Comprehensibility .... ...... .. ................. 54
Candidate Evaluation by Schema/
Comprehensibility ............................. 56
Candidate Evaluation by Schema/
Comprehensibility ...... .............. ......... 57
Vote Likelihood by Schema/Comprehensibility ... 59
Vote Likelihood by Schema/Comprehensibility ... 60
Low Comprehensibility Message Ratings by
SCth& IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 61
Candidate Evaluation by Schema/
Comprehensibility ..................... ........ 62
Vote Likelihood by Schema/Comprehensibility ... 64

viii

TABLE
18.

19.

20.

21.

PAGE
Processing Effort by Schema/
Comprehensibility ................. ......... ... 65
Message Comprehensibility by Schema/
Comprehensibility .......... ......... ..... ..... 67
Recall by Schema/Comprehensibility .... ........ 68
Recognition by Schema/Comprehensibility ....... 7O

ix

LIST OF EXHIBITS

EXHIBIT PAGE
1. Low Comprehensibility Message ................. 34

2. High Comprehensibility Message ........ . ....... 36

CHAPTER I

INTRODUCTION

As a process, voter decision-making has both puzzled
and fascinated researchers for several decades. At times
it has been conceived to be complex and calculated while
at others it has been viewed as unsophisticated or even
capricious. Various models have been furthered to explain
the process. Each one differs by either explaining the

process ”better,‘ exploring another aspect of the process,
or incorporating new or more variables into the conceptual
framework. Recently, work has focused on advertising’s
role in the political process. Critics contend that
advertising merchandises political candidates much like
bars of soap. They suggest that creative and monetary
resources are channeled into creating winning candidate
images at the expense of discourse about substantive
issues. This criticism has gained momentum (and some
credence) due to the increased proportion of campaign
dollars earmarked for televised political advertising and
the proliferation of and reliance on media pundits. While

the concern has heuristic merit for researchers, it

addresses the very foundations of our democratic way of life.

2

In light of these lofty considerations, political
communication researchers have begun to focus on the
cognitive processes that mediate political communication
effects. One line of inquiry has focused on the role of
both cognitive schemata and media message characteristics
in determining political communication effects. The
present research continues and extends this line of
inquiry. Specifically, it examines the role of cognitive
schemata in the processing of information from televised
political advertisements with high and low levels of audio
comprehensibility. Several broad questions are addressed:
(1) How do cognitive schemata affect message processing?
(2) What is the impact of message comprehensibility on the
acquisition of information? (3) How do message
comprehensibility and cognitive schemata affect memory for
candidate and issues, and evaluation of candidate and
message?

Theory and research from various domains are used as a
framework for formulating hypotheses that follow from
these questions. Relevant literature will be summarized
in the following order. First, an examination of
political information (i.e., issue and image information)
and its processing is undertaken. The next section
addresses the role of cognitive schemata in the processing
of information--in particular, the processing of political

information. Finally, a review of literature dealing with

3
the effect of message comprehensibility on memory and

information processing is performed.

REVIEW OF RELEVANT LITERATURE

Political Information and Its Processing

Today’s voter has access to a myriad of sources for
obtaining information concerning political issues,
candidate issue stands and candidate personal qualities.
One venue, televised political advertising, has received
increasing scrutiny from mass communication and political
scholars (Garramone, 1983, 1984, 1985,1986; Garramone,
Steele, Hogan 3 Rifon, 1987; Garramone, Steele &
Pinkelton, 1991; Garramone, Steele a Hogan, 1988; Geiger a
Reeves, 1981; Kaid & Sanders, 1978; Keeter, 1987; Kraus a
Perloff, 1985; Lau a Sears, 1986) and a disproportionate
increase in campaign dollars (Shyles, 1984).

Researchers have divided information contained in
political advertisements into two broad categories--image
and issue information. Image information reflects
attributes of the candidate. It circumscribes candidate’s
character, personality, appearance and behavior
(Garramone, 1986; Kinder, Peters, Abelson, a Fiske, 1980).
Issue information encompasses candidates’ stands on
issues, their past political performance, and their role
in the political process (Garramone, 1986). Additionally,
issue information can contribute to candidate image. For

example, a candidate who favors increased funding for the

5
homeless, a political issue, might be thought of as a
compassionate person, an image attribution. Content
analyses of televised political advertisements confirm the
existence of both issue and image information (Joslyn,
1980; Hofstetter & Zurkin, 1979).

Issue and image information are utilized by voters in
the decision-making process. For example, voters claim to
use televised political advertisements to learn about
candidate issue stands, qualifications and personal
qualities (Atkin, Bowen a Sheinkopf, 1973; Mendelson &
O’Keefe, 1976). Some research even suggests that
individuals learn more from political advertisements than
they do from news broadcasts (McLure a Patterson, 1976).
However, data indicate that voters are primarily
interested in extracting nonverbal and image information
from political advertisements (Katz & Feldman, 1962;
O’Keefe & Sheinkopf, 1974; Rosenburg & McCafferty, 1987).
Research further suggests that image information and
nonverbal communication can affect the political process.
For example, candidate physical attractiveness has been
demonstrated to affect vote likelihood and evaluation of
candidate and message (Garramone, Steele a Pinkleton,
1991; Rosenburg, Bohan, McCafferty a Harris, 1986;
Rosenburg a McCafferty, 1987; Steele, Garramone a Hogan,
1988; Stokes, 1966).

Results such as these suggest that affective elements

6
of a message (e.g., physical attractiveness) may influence
the processing of the message and, quite probably, its
subsequent effects. The role that affect plays in
information processing has received increased interest and

research energy from several disparate, academic domains.

Research on Affect

Much of the psychological research from the 1960s
through the early 19808 focused on the person as
information processor. During this cognitive revolution,
scholars pursued the impact of message content on various
cognitive variables (e.g., recall and recognition). The
past decade, however, has witnessed a renewed interest in
the effects of persons’ affective reactions to messages
(Abelson, Kinder, Peters a Fiske, 1982; Burke & Edell,
1989; Fiske, 1981; Higgins, Kuiper, a Olson, 1981;
Holbrook a O’Shaugnessy, 1989; Madden, Allen a Twible,
1988; Mitchell a Olson, 1981; Shimp, 1981; Lutz, 1985).
Leaders in cognitive psychology have suggested that
conceptualizing thought as being free from emotional
influence is short-sighted and are now incorporating
affect into their models (Bower, 1981; Norman, 1980,
Piaget, 1981). It even has been suggested that cognitive
and affective systems may operate independently of each
other (Zajonc, 1980).

Fiske (1982) suggests that some categories stored in

7
memory may be affectively charged. This affect may be
transferred to stimulus objects that activate the
category. For example, persons are represented as
instances of a category, such as politician. To the
extent that a person is perceived to match this stereotype
(e.g., politician), the affect associated with the
category will be transferred to the person. Someone
perceived to be a politician by an individual with
negative category affect will be regarded poorly unless
positive, individuating information becomes available to
temper the category affect.

In consumer behavior literature, one important
construct that has emerged from this renewed focus on
affect is attitude-toward—the-ad (AN). A“ is defined as
“a predisposition to respond in a favorable or unfavorable
manner to a particular advertising stimulus during a
particular exposure occasion" (MacKenzie a Lutz, 1989, p.
49). This conceptualization aligns with Fishbein &
Ajzen’s (1975) view of attitude in that it rests solely on
evaluative reactions to a stimulus and excludes cognitive
and behavioral responses. Advocates of A“ contend that a
positive attitude toward an advertisement can translate to
a positive attitude toward the product. This appears to
be the raison d’etre behind image-oriented political
advertisements. In empirical studies, A“ has been shown

to be an important mediator of brand attitude and inten—

8
tion to purchase (Mitchell a Olsen, 1981; Shimp, 1981).

Empirical research in social cognition parallels these
findings. The results suggest that affective as well as
cognitive responses to a message may exert considerable
influence on message evaluation and behavioral reactions.
For example, Snyder & DeBono (1985) demonstrated
differences in how high and low self-monitoring
individuals react to advertising strategies. Low self-
monitors reacted more positively to quality-oriented
advertising whereas high self-monitors reacted more
postively to image-oriented advertisements. High self-
monitors were willing to pay a higher price for products
advertised with an image-orientation, and were more
willing to try products advertised with an image appeal.
It appears affective elements of the image advertisement
were capable of invoking a positive attitude toward both
the advertisement and the product.

Geiger & Reeves (1991) suggest that the distinction
between affective and cognitive message elements may
account for the differential processing of televised
political information. Issue information is viewed as
cognitive and is conveyed primarily through the audio
channel. A voice-over, or the candidate, may express
his/her stand on issues and detail past accomplishments.
The processing of this information is primarily a

cognitive activity as voters must comprehend the meaning

9

of the information if it is to be successfully utilized in
voter decision-making. Due to this increased cognitive
effort, issue information should be processed at a deeper,
semantic level, producing a stronger memory trace (Craik a
Lockart, 1972). Image information is affective and is
conveyed through both the audio and video channels.
Image-oriented advertisements are intended to create
positive affect for the candidate. These messages often
depict the candidate involved in "worthy” activities and
may be accompanied by music and other affectively charged
visuals (e.g., sunsets, wildlife, children playing, etc.).

The impact of information on a person’s memory and
behavior is likely dependent on that person’s information
processing behavior. Since the amount and variety of
information contained in any political advertisement is
likely to be greater than a person possibly could attend
to effectively, individuals’ processing behavior is
necessarily selective in what is noticed, attended to and
encoded. This selective process doesn’t appear random but
rather appears dictated by cognitive structures that have

been labeled cognitive schemata.

Information Processing and Cognitive Schemata
Social psychology has been dominated by three
conceptual strains in recent history: man as consistency

seeker, as naive scientist and as cognitive miser (Lau a

10
Sears, 1986; Taylor, 1981). As consistency seeker,
several theories suggest that people are uncomfortable
when they experience inconsistency (Festinger, 1957;
Heider, 1946). The inconsistency may be between beliefs
and attitudes, attitudes and behavior or among attitudes.
People are driven to resolve the discomfort of
inconsistency by revising their attitudes, beliefs or
behaviors. Although some data support this view, evidence
also suggests that people are quite capable of living with
the inconsistency (Taylor, 1981).

The naive scientist approach suggests that people are
rational thinkers and problem solvers. For example,
attribution theory (Jones & Davis, 1965) suggests that
people ascribe causality for others’ behavior by
methodically and logically analyzing the person/situation
context. The rational man approach was dealt a telling
blow, however, as research emerged demonstrating that
human reasoning is replete with ”faulty" logic and
heuristic processes (Cohen, 1979; Kahneman a Tversky,
1982; Nisbett a Ross, 1980).

Based on these limitations, a third strain of thinking
has developed which views man as a cognitive miser,
economizing with respect to information selectivity,
attention and active memory. Central to much of this work
in social and political cognition is the notion of

cognitive schemata (Cantor a Kihlstrom, 1981; Conover a

11
Feldman, 1984; Fiske a Kinder, 1981; Fiske & Taylor, 1984;
Lau a Sear, 1986; Hamill a Lodge, 1986).

Cognitive schemata represent organized knowledge
structures about a particular concept (Bartlett, 193?).
They include both the defining features and relevant
attributes that constitute the concept and the
relationships among the attributes (Rumelhart a Ortony,
1977; Crocker, Fiske a Taylor, 1984). They aid
information processing by determining what information is
attended to, how it is encoded and stored, and how it is
utilized (Crocker et al., 1984). People use schemata for
perceiving visual arrays, for understanding meaningful
press, for processing information about the natural world,
for understanding and perceiving other persons and for
perceiving and guiding their own behavior (Teaser, 1978).

Schemata are hierarchically organized with more
abstract or general knowledge at the top and categories of
more specific information nested within the general
categories (Taylor a Crocker, 1981). For example, within
the dominant schema for politician is likely to be
subschemata for Republican and Democratic politicians.
This allows processing of information at an abstract level
for a candidate with an unknown political affiliation.
However, once political affiliation is discerned, the more
concrete subschema is called upon to process additional

information.

12

Schemata are composed of variables. Values for the
variables are determined by the environment. A schema for
politician, for example, probably would include the
variable political affiliation. Democrat and Republican
would be constraint values; these are values that the
variable can assume. For a politically unsophisticated
person, these may be the only constraint values. However,
as the person becomes more politically conversant,
additional constraint values can be added for party
affiliation (e.g., Libertarian, Green, American
Socialist). Variables and values can be added or deleted
as the schema develops. During this process, schema
structure becomes more complex as cohesive linkages are
formed between a variety of variables based on experience.

One way schemata facilitate information processing is
by filling-in variables with default values if no values
are specified (Rumelhart & Ortony, 1977; Crocker et al.,
1984). The default value is a best guess based on the
values of associated variables. For example, if a person
listens to a speech by an unrecognized politician, s/he
may fill-in the value Republican if the politician
espouses pro-military and anti-abortion stands.

Thus, perceptual input may be shaped by schemata. If
a person encounters a new piece of information and
interprets it to be a type for which there is an existing

schema, the schema becomes instantiated (activated) and

13
the information is processed. Since the schema contains
information on what relationships should exist among
variables, the input may be assimilated (shaped) to fit
the schema (i.e., values may be filled-in for variables).
Schemata are not immutable, however. If instances arise
indicating that a schema is no longer representative of a
particular domain, it may be accommodated (revised) to
better fit the data (Crocker et al., 1984, Rumelhart a
Ortony, 1977; Tesser, 1978).

People develop schemata for their various domains of
experience, including the political domain (Garramone et
al., 1991; Lau, 1986). Schemata evolve, grow and mature
to fit the variety of situations that a person encounters.
For example, a person knowledgeable of political issues
and the political process (schematic) would be adept at
processing new information about a political topic. On
the other hand, a person with little political
sophistication (aschematic) might find the same
information to be so much political babble (Converse,
1975; Fiske, Kinder a Larter, 1983). There is little
chance that the aschematic would encode much of the
information and any that was encoded would be hard to
retrieve due to the unstructured, disorganized way it was
processed.

Based on experience, people have varying levels of

schema development. The degree of schema development

14
affects the quantity, organization, compactness and use of
schematic knowledge (Crocker et al., 1984). An expert’s
schema contains more information and more dimensions than
a novice’s. For example, in one study (Chase a Simon,
1973), master chess players and non—chess players were
shown chessboards in various stages of play and
chessboards with pieces randomly scattered over the board.
Master chess players were able to recall all the in-play
boards but were no better than the novices at recalling
the random boards. Non-chess players found both boards
equally difficult to recall. Whereas master chess
players’ schemata allowed them to chunk in-play boards
into meaningful units, the non-chess players had no
schemata to rely on for organizing the daunting array of
chess pieces. Markus a Smith (1981) found similar results
concerning experience-based differences in self knowledge.
Individuals expert at being independent or dependent
(self-schematic) were better able to process information
about self and had readier access to self-descriptive
behaviors than did aschematics.

Since people have limited cognitive capacity for
processing information, schemata function to direct
attention to information relevant to a particular task
(Crockere, et al., 1984; Tesser, 1978; Valenti and Tesser,
1981). For example, if a voter desires to form an image

for a political candidate, the cognitive schema for

15
politician image formation presumably would be invoked.
This schema would include traits and their relationships
that are deemed relevant for politician image formation.
These might include traits such as power, compassion,
honesty, fluency, and intelligence (Kinder et al., 1980).

Garramone (1983) investigated the impact of image
motivation (i.e., to attend to a political message to form
an impression of the candidate’s personality) on
information processing. Hypothesizing that a person who
is motivated to process a political message to form an
impression would attend to candidate physiognomic
characteristics as well as verbal information, she found
that image motivated persons paid greater attention to the
video component of a commercial than did issue-motivated
persons.

Lau (1986) suggests peoples’ political schemata
emphasize one or more of the following factors (within
cognitive limitations): (1) political issues, (2) group
relations, (3) party identification and (4) candidate
personality factors. Moreover, individuals are relatively
consistent over time in their use of political schemata.
Johnston (1986) found that voters maintain either a
predominantly issue or image processing orientation
irrespective of advertising content. They consistently
encoded either high image, high issue, or moderate image-

moderate issue responses to either high image or high

16
issue commercials. Thus, the information a person
extracts from a political message is a joint function of
message content and cognitive schemata. Meaning is not in
the message alone, but is a result of the interaction of
the schema with the message.

Depending on individuals’ schemata, they consistently
process, store and, ultimately, use for evaluation
specific types of information. Peoples’ information
processing behavior, strongly influenced by the schemata,
governs the information channels and message
characteristics which will be attended to and encoded.

It is expected that people using issue schemata will focus
primarily on the audio channel, which is where issue
information is most likely contained. Since image
information is conveyed by both the audio and video
channels, image processors would be expected to attend to
both channels and encode more visual information than
issue processors (Garramone, 1983, 1984; Garramone et al.,
1991; Steele et al., 1988). If recall and the prominence
of visual and verbal information in memory are means for
assessing encoding, then the following predictions are
offered:

FH.: Image processors will encode more visual

information than issue processors.
H1b: Issue processors will encode more issue

information than image processors.

17

An extensive literature in psychology suggests verbal
and visual information processing and memory may be quite
different (Paivio, 1975, 1978). Paivio & Begg (1974)
propose a dual-coding model in which visual stimuli tend
to be processed and stored simultaneously in an imagery
system, whereas verbal information is handled sequentially
and is stored in an independent verbal system. Another
body of literature, although not incompatible, suggests
the differences in processing may be due to hemispheric
lateralization (Anderson, Garrison a Anderson, 1979;
Geschwind, 1979). Evidence suggests that the left brain
is responsible for analytic, logical thought processes,
which would characterize the verbal channel, whereas the
right hemisphere specializes in gestalt processing, which
would characterize the video channel. Exactly how this
research may tie in with the notion of schemata is beyond
the scape of this work. However, it may suggest that
persons’ schemata ultimately may be driven by hemispheric
dominance. For example, a person with an image schema for
political information processing may be driven by left
hemispheric specialization and thus attend to and encode a
greater amount of visual information than right hemisphere
driven issue processors.

Although people are relatively consistent in their
schema choice within a domain, occasionally they may use

different schema to think about the same object (Tesser,

18

1978). Indeed, even though a person has greater facility
using one schema does not mean that the schema will always
be used. As situations arise and objectives change,
individuals may choose to “tune-in" different schemata
(Jeffery & Mischel, 1979; Tesser, 1978; Yalch a Elmore-
Yalch, 1984; Zajonc, 1968). For example, early in a
presidential campaign people who are predominantly issue-
oriented processors may choose to winnow the crowded field
based on image information. It may be easier to decide
initially who seems presidential and to concentrate effort
on their issue positions rather than processing all
candidates’ issue stands. Schema choice, then, can be a
function of objectives and circumstances. Choosing one
schema over another has important implications, however,
since it would call attention to different aspects of the
stimulus (Garramone, 1983, 1984, 1986; Garramone et al.,
1991; Steele et al., 1986; Teaser a Durheiser, 1978).

Cognitive schemata exert a strong influence on
information processing; they direct attention to channels
relevant to the current processing objective. For
example, as noted, image motivated people would more
likely attend to the visual channel than would issue
motivated people (Garramone, 1983, 1984; Steele et al.,
1988; Garramone et al., 1991). Moreover, various message
characteristics should come into and out of prominence

depending on the processing objective. Thus, the impact

19
of these message characteristics would vary according to
persons’ processing orientation. One message
characteristic that might be influenced by processing
objectives, and ultimately cognitive schemata, is message

comprehensibility.

Message Comprehension

With one stroke of the pen, Lasswell (1948) set much
of the agenda for communication and persuasion researchers
for the latter half of this century. Research spurred by
his oft cited question “Who says what in which channel to
whom with what effect?“ has dominated the empirical
landscape since the 19508 (p. 36). For example, the Yale
Group (Hovland, Janis & Kelley, 1953) organized its
research program around the question and studied the
effect (attitude change and message retention) of who
(source) says what (message) to whom (receiver) through
which channel (medium). They clearly recognized that
comprehension was a major factor in this communication
process, labeling it an internal mediating device.

More recently McGuire (1968, 1972) introduced one of
the first models of information processing. Its
conceptual underpinnings are evidently rooted in the work
of the Yale Group, and message comprehension plays a
central role here as well. The stochastic model has five

steps: (1) attention, (2) comprehension, (3) yielding,

20
(4) retention and (5) action. Being probabilistic in
nature, failure at any one stage precludes reaching the
succeeding stage(s). For example, if a person does not
comprehend the arguments in a message, s/he will not yield
to the position advocated.

Comprehension generally has been considered one of the
most important results of communication (Jacoby a Hoyer,
1982). It is viewed as critically important since it is
antecedent to other effects--memory retention, belief
formation and change, attitude formation and change, and
behavioral intention (Eagly, 1974; Eagly a Chaiken, 1984;
Jacoby a Hoyer, 1982; Lavidge a Steiner, 1961; McGuire,
1981; Petty a Cacioppo, 1981). Despite the critical role
that comprehension has been accorded, there has been a
dearth of research effort addressing the variable (Chaiken
& Eagly, 1984; Fishbein a Ajzen, 1981; Jacoby a Hoyer,
1982, 1987, 1989; Ratneshwar a Chaiken, 1991). Much of
the communication and persuasion research that does
address comprehension combines it with attention into a
single step (reception) because of the difficulty of
measuring attention and comprehension separately. The
existing research seldom even directly measures or
manipulates the reception process. Instead, investigators
typically manipulate independent variables (e.g.,
credibility) that may influence reception. They then

measure recall and recognition (i.e., measures of

21
retention) as surrogate measures of reception. Finally,
they determine whether the independent variables exert
parallel effects on retention and attitude change.

In an attempt to manipulate comprehension directly,
Eagly (1974) told subjects they were about to hear a
lecture by a physiological psychologist on sleep research.
Subjects were told the lecturer would suggest that people
need less sleep than they normally get. The first group
heard a clear tape of the lecture, which contained six
arguments to justify the less sleep recommendation. The
second group heard the same lecture and six arguments, but
the recording was of low quality with plenty of background
noise. Subjects in the good comprehensibility condition
were more likely to agree with the lecturer’s
recommendation and recalled more of the support arguments
than the low comprehensibility group. Such data provide
further evidence of the importance comprehension plays in
communication and persuasion processes.

Comprehension may be described as the successful
interaction of the perceiver with the message (Rumelhart a
Ortony, 1977). Since comprehension is a learning process
in which prior knowledge plays an important role, schemata
are central to the process of comprehension. Anderson
(1977) and Schank a Abelson (1977) describe the
constructive role of schemata, indicating that schemata

serve as organizers for new material and without them the

22

very notion of comprehension and learning would be
impossible. Information processing is directed at
choosing schemata which best account for incoming
information (Petty & Cacioppo, 1981). Once sufficient
accounting is achieved, comprehension is said to have
occurred (Rumelhart a Ortony, 1977). What gets stored in
memory is not a veridical copy of the input but rather the
interpretation of the information provided by the
instantiated schema. Thus, comprehension can be
considered the selection of appropriate schema and binding
variables to account for a situation, and then verifying
that the schemata do indeed account for the situation.

Political communication scholars have become
interested in the process by which messages exert their
effect rather than focusing solely on the effect
(Garramone et al., 1991). Research dealing with cognitive
schema and comprehension suggests that people bring to
messages schemata appropriate for processing the
particular type of information (Crocker et al., 1984).
For example, within the political domain, a person with an
issue schema would approach a political advertisement with
the objective of comprehending issue stands for the
particular candidate. Image processors would approach the
same advertisement with the primary objective of forming
an image of the candidate.

Despite these objectives, there is still the

23

possibility that the channel(s) to be relied on for
obtaining schema-specific information could be
incomprehensible. For example, a political message
discussing national fiscal policy and debt reduction could
easily become mired in economic jargon and quantitative
minutia, reducing the message to a low comprehensibility
level. This has interesting implications when coupled
with research on the channel selectivity of issue and
image processors. For example, if the audio channel is
incomprehensible, how would this affect image and issue
processors? Would issue processors shift their attention
to the video channel? Would image processors devote even
more of their processing energy to the video channel? A
few studies may help shed some light on these questions.

Research suggests that children are sensitive to
television program comprehensibility, paying attention to
those portions of the program that are most understandable
(Lorch, 1979; Pezdek & Hartman, 1983; Anderson, Lorch,
Smith, Bradford a Levin, 1981). Marketing research
suggests that low-comprehensibility messages affect
persuasion (Yalch & Elmore-Yalch, 1984), reduce interest
in and attention to the message (Anderson a Jolson, 1980),
and reduce motivation to process (Witt, 1976). For
example, Yalch & Elmore-Yalch (1984) exposed subjects to
commercial messages with varying amounts of quantitative

information under conditions of high and low source

24
expertise. (Messages high in quantitative information are
considered more complex than their less quantitative
counterparts and, thus, are deemed less comprehensible.)
Subjects exposed to the high quantitative message became
disinterested in pursuing issues in the message. Instead,
they shifted from processing central cues (issues) to
processing peripheral cues (source expertise). Thus,
messages high in quantitativeness that employed high
expertise sources (i.e, a peripheral cue) were more
persuasive than high quantitative messages coupled with
low expertise sources. However, source expertise had no
effect when information was presented in a nonquantitative
manner. The few studies in marketing that found
persuasion enhanced by quantitativeness were those in
which audiences were motivated and capable of processing
the quantitative messages (Anderson & Jolson, 1980;
Holbrook, 1978).

In the present research, it is suggested that if the
audio channel is low in comprehensibility, issue
processors will devote more cognitive energy to the video
channel than they would if the message was higher in
comprehensibility. That is, they will tune-in the
alternative schema (image) in an attempt to comprehend the
material in the message. Furthermore, image processors
will devote even more of their attention to the video

channel, as the the audio channel will offer little in the

25
way of comprehensible image information. Based on these
predictions, the following hypothesis is offered:
H2: Issue and image processors will encode more
visual information under low audio
comprehensibility conditions than under high

audio comprehensibility conditions.

Several studies suggest that people find the
processing of low comprehensibility messages less than
favorable. They rate message processing difficult (Witt,
1976) and unpleasant (Anderson a Jolson, 1980; Chaiken a
Stangor, 1987; Eagly, 1976; Eagly a Warren, 1974; Schmidt
& Sherman, 1984). One would expect both image and issue
processors to rate low audio comprehensibility messages as
unpleasant, since both would devote some processing energy
to the low comprehensibility audio channel. Also, they
should both rate low comprehensibility messages as less
favorable than high comprehensibility messages. Thus, the
following predictions are offered:

H3.:

Both issue and image processors will rate low
audio comprehensibility messages more unpleasant
than higher audio comprehensibility messages.
H3h‘ Both issue and image processors will rate low
audio comprehensibility messages less favorably

than higher audio comprehensibility messages.

26

Based on research on affect and comprehension, one
would expect unfavorable "feelings” engendered by a low
comprehensibility message to be transferred to the
candidate (Burke a Edell, 1989; Chaiken a Stangor, 1987;
Eagly, 1974; Eagly & Warren, 1976; Gardner, 1985; Higgins
1981; Schmidt a Sherman, 1984). According to Higgin’s
(1981) communication game rule, communicators are expected
to be cogent and understandable. If these expectations
are violated, the source, the message and the topic are
likely to be judged adversely. In the present context,
this negativity is likely to influence overall candidate
evaluation and vote likelihood. This line of argument
suggests the following:

H4- Beth issue and image processors will evaluate

a'
the candidate less favorably in low audio
comprehensibility conditions.

H4b‘ Both issue and image processors will be less
likely to vote for the candidate in low audio

comprehensibility conditions.

Since issue processors will find little comprehensible
issue information in a low comprehensibility audio channel
(i.e., their primary information channel), they will be
forced to redirect (i.e., tune-in another schema) their
processing energy to the video channel. This message-

induced schema switching should prove frustrating and will

27
transfer to lower ratings of message, candidate and vote
likelihood. Formally:

H5.:

Issue processors will rate message and

candidate less favorably than will image
processors under low audio comprehensibility
conditions.

H5b: Issue processors will be less likely to vote for
candidates than image processors under low

audio comprehensibility conditions.

Furthermore, issue processors will feel they expend
more effort processing the low comprehensibility message
due to the message-induced schema switching and the
necessity of using a less familiar processing orientation.
Additionally, the schema switching, use of a less familiar
schema, and the ensuing perceived effort expended, should
make issue processors feel low comprehensibility messages
are more incomprehensible than would image processors.
Thus, the following hypotheses are suggested:

H6.: Issue processors will perceive higher amounts of
energy expended in attempting to understand low
audio comprehensibility messages than will image
processors.

H6b: Issue processors will rate low audio

comprehensibility messages as less

comprehensible than will image processors.

Chapter II

METHODOLOGY

Overview

The study was conducted in two stages. Stage I and
Stage II data collection were separated by approximately
two weeks. In Stage I subjects’ default political
schemata were determined by content analyzing their post-
message cognitive responses to a series of televised
political commercials. In Stage II, subjects viewed
televised political commercials either high in audio
comprehensibility or low in audio comprehensibility.
After viewing the commercials, subjects were asked to list
their thoughts, to rate the message and candidate, and to
recall and recognize issue information from the

commercials.

Subjects

Subjects were 249 students solicited from
undergraduate communication classes to participate in
Stage I and Stage II of the study. Subjects received
extra course credit or fulfilled course requirements for

their participation. Those participating as part of

28

29
course requirements were given the option of doing an

alternate, nonresearch-based class assignment.

Stage I

Procedure

Subjects participated in Stage I in groups ranging in

size from 25 to 38. Booklets were distributed before data

collection. The following instructions were printed on

the front of the booklet and were read aloud by the

experimenter:

We are conducting a study to learn how people process
information. There are three parts to the study:
First, you will view a series of political
commercials. You then will complete a short
questionnaire regarding the commercials. Finally,
you will be debriefed regarding the study’s theory,
methodology, and expected findings. The entire study
will take approximately 15 minutes of your time.

You may decline to participate or withdraw from the
study at any time. All results will be treated with
strict confidentiality. You may request a copy of
the study’s findings by contacting Michael Steele in
the Department of Advertising.

If you are willing to participate, please indicate so
by both signing AND printing your name in the spaces
below.

Subjects were then exposed to a series of extant

televised political commercials. The videotaped

commercials were shown on a 25” television in full-view of

all subjects. After viewing the commercials, subjects

were told to turn to the first page of the booklet.

30
Directions were written at the top of page 1 and were read

aloud by the experimenter:

Now we’d like you to list all the thoughts that you
had while viewing the political commercials. Please
write one and only one thought in each box. Use as
many boxes as you need; two pages have been provided.
Remember write only one thought per box.

Students were given 3 minutes to complete the task. The
booklets were then collected and subjects were debriefed
and dismissed. To protect the integrity of Stage 2,
subjects were debriefed as to the essence of the study but
not the particulars. They were read the following

statement as a debriefing:

We are examining how commercial type affects the
manner in which people process the information from
commercials. This will be assessed through a content
analysis of the thoughts that you have generated.

Stimulus Material

The stimulus material was a series of nine
videorecorded political commercials from a 1984 senatorial
race in Mississippi. The nine commercials were chosen
from a pool of commercials for the two opposing candidates
in the senatorial race. Since subjects’ processing
schemata (i.e., image vs. issue) would be determined by
their post-message cognitive responses to these

commercials, it was essential that the commercials contain

31
both image and issue information. Thus, commercials were
selected that contained both references to candidates’
issue stands (issue information) and to their behavior,

character and personality (image information).

Content Analysis of Post-Message Responses

Each thought was coded as either “image," "issue,“ or
"other” by two trained, independent coders. Image
thoughts were defined as those referring to candidates’
enduring character, personality, appearance or behavior
(Garramone et al., 1991; Johnston, 1986). This category
also included evaluations of the candidate based on issue
positions (e.g., Winters is crazy to be pro-nuclear).
Also, thoughts were coded as "image" if they referred to
the commercial’s image-orientation (e.g., too much flag
waiving stuff).

Thoughts were coded as "issue" if they restated issue
information from the commercial, referred to a candidate’s
role in the political process or referred to a candidate’s
political performance (e.g., Why does "he" support nuclear
build-up?"). Also, thoughts were coded as ”issue" if they
referred to issue-oriented elements of the commercial
(e.g., Winter’s stand on education was not spelled-out).

Thoughts were coded as "other" if they did not fit
into the "issue" or ”image" categories. In cases when

several thoughts were strung together into one statement,

32
coders were instructed to divide the statements’ content
into individual thoughts and to code each thought
according to the protocol detailed above.

To control for chance agreement, intercoder
reliability was determined using Scott’s 21, which
indicated a level of 84.36%. Disagreements were resolved
through discussion.

Processing schemata were determined by calculating the
percentage of “issue" and "image" thoughts for each
subject. (Thoughts coded as "other" were not included in
the percentage determination.) Subjects with 60% or
greater thoughts in either the "image" or "issue" category
were designated as possessing that schemata. For example,
a person whose thoughts were 63% "issue“ and 37% "image"
would be regarded as an issue processor. Subjects with
less than 60% in either category (e.g., 45% "image," 55%
“issue") were excluded from the analysis for Stage II.

The effective sample after Stage I was 219: 122 issue

processors and 97 image DI’OCGSSOI’S.

Stage II
Procedure
Two hundred and two students returned from Stage I to
participate in Stage II of the study. They participated
in groups of 22 to 36. Subjects were randomly assigned to

one of two conditions: either low audio comprehensibility

33
or higher audio comprehensibility. Before data
collection, subjects were distributed a booklet of
dependent measures. On the front of the booklet were the

following directions, which the experimenter read aloud:

We are conducting a study to determine how people
process the information contained in political
commercials. You will be asked to view a political
commercial and then answer a few questions pertaining
to the commercial. The entire process should take
about 15 minutes.

Your participation is voluntary and you may terminate
your participation at any time. All results will be
treated with the strictest confidence. You may
request a copy of the research findings by contacting
Michael Steele in the Department of Advertising.

If you agree to participate, please indicate so by

both signing and printing your name in the spaces
below.

Subjects sat in full-view of a 25“ television on which

the videorecorded political commercials were played.

Stimulus Material

Political broadcast commercials were designed for the
study. A video track was created from newscasts and movie
clips. Visual items were chosen to parallel audio scripts
that were written to simulate a congressional race in
”another” state. For example, when the audio track
discussed hydropower, a scene with turbines and running
water was shown. The video track was crossed with a low

comprehensibility audio track (Exhibit 1) and a higher

34
Exhibit 1
Low Comprehensibility Message

Hello, this is Congressman Bob Moran. I’d like to
take a few moments to address criticisms of the projected
Sonoma Hydroelectric Plant. As you know I’ve undertaken
extensive studies which provide evidence that Sonoma Hydro
is a safe, monetarily-wise alternative to the Haskee-Sharm
Nuclear Power Plant already under construction. I will
give you four reasons why Sonoma Hydro is a fiscally sound
alternative to Haskee-Sharm.

First, admittedly there will be a large direct fiscal
shortfall commensurate with the curtailment of Haskee-
Sharm construction. But this will be ameliorated since
the state fiscal revenues increased, in total, by an
unprecedented $109.9 million or 6.8% from 1990-1991 to
1991-1992. This represents the net effect of increased
petro-chemical taxation, and a reduction in the
appropriation base resulting from the State’s Negative
Supplemental Bill.

Second, $27.3 million or 6% of fiscal 1993’s revenues
will not be required to cover previously approved state-
wide commitments and externally controlled factors.
Examples include: no need to annualize prior salary
commitments, no increase in state employee fringe benefit
packages and formula driven research allocations to
private, and academic institutions did not increase as
projected.

Third, the increased raw material costs anticipated
for the Haskee site through fiscal 1993 have not been
realized. In general, costs paralleled the rate of
inflation, not the 8% rate of increase as projected by the
House Budget Appropriations Committee.

Fourth, contract negotiations between Calcon
Construction (site contractors for Haskee-Sharm) and its
employees did not provide the compensation escalations
anticipated by the House Appropriations Committee.
Therefore, projected budgetary reductions and
reallocations deemed necessary to offset the results of
the collective bargaining will not be required. However,
since the reductions and reallocations have already been
prescribed and implemented, it would be fiscally prudent
to use these funds to reach contractual settlement when
the current contract between Calcon and the State is
abrogated in favor of the Sonoma Hydro Project.

I hope you will support me in my bid for Congress and
that you will support Proposition A -- construction of
Sonoma Hydro and curtailment of Haskee-Sharm Nuclear.
Yes, we need power, but it’s safe power we deserve. Vote
Bob Moran for Congress and YES on Proposition A.

35

comprehensibility audio track (Exhibit 2). The low
comprehensibility audio track was complex (e.g., large
words and data). The high comprehensibility track was a
"translation” of the low comprehensibility track,
substituting simpler language and words in place of data.

Message comprehensibility was assessed by application
of Flesch’s "Reading Ease" formula (Flesch, 1951).
Results of the analysis indicated that the less complex
message requires an 8th-9th grade reading level for
comprehension, whereas the high complexity message

requires a 13th-16th grade (i.e., college) reading level.

Dependent Measures

After viewing the political commercial, subjects were
instructed to open their booklets to page 1. They read
the following directions printed at the top of the page as

the experimenter read them aloud:

Now we’d like you to list all the thoughts that you
had while viewing the commercial. Please write one
and only one thought in each box. Use as many boxes
as you need; several pages have been provided.
Remember write only one thought per box.

Subjects were given 3 minutes to complete the task.
Subjects then completed the remainder of the dependent
measures: candidate evaluation, vote likelihood, message

measures (evaluation, comprehensibility, perceived

36
Exhibit 2
High Comprehensibility Message

Hello, this is Congressman Bob Moran. I’d like to
take a few moments to address criticisms of the projected
Sonoma Hydroelectric Plant. As you know I’ve undertaken
extensive studies which provide evidence that Sonoma Hydro
is a safe, smart money alternative to the Haskee-Sharm
Nuclear Power Plant already under construction. I will
give you four reasons why Sonoma Hydro is a financially
sound alternative to Haskee-Sharm.

First, admittedly there will be a large short range
loss of money if the Haskee project is ended. But this
loss can be handled since state revenues have increased at
a record rate of $109.9 million or a 6.8% increase over
1991-1992. Increased gasoline taxes and a smaller expense
budget are the reasons. The smaller expense budget was a
result of the State’s Negative Supplemental Bill.

Second, $27.3 million or 6% of 1993’s revenues will
not be needed to cover several items already approved.

For example, neither yearly wages or benefits increased
for state employees as expected, and research budgets for
private and state institutions have not increased.

Third, raw material costs at the Haskee site did not
increase for 1992 as expected. Costs remained at the
level of inflation, not at 8% as the House Budget
Committee expected.

Fourth, Calcon Construction (site contractors at
Haskee) was able to keep employee wage increases lower
than the House Appropriations Subcommittee projected.
Therefore, the amount of money that the state set aside to
meet the expected wage increases is not entirely needed.
Yet the money has already been diverted from other
projects. It would be wise and possible to use this money
to fund a settlement with Calcon if the present contract
is ended early in favor of the Sonoma Hydro Project.

I hope you will support me in my bid for Congress and
that you will support Proposition A -- construction of
Sonoma Hydro and stopping construction of Haskee-Sharm.
Yes, we need power, but it’s safe power we deserve. Vote
Bob Moran for Congress and YES on Proposition A.

37

processing effort, and pleasantness), and issue recall and

recognition.
Qegnjtive :essenses were coded by two independent

coders as either "issue, visual," or “other." The
coding protocol for "image” was the same as that employed
in Stage I. Responses were coded as "visual" if they
contained information that was available only from the
video channel (e.g., Cochran’s family is goofy looking.”).
Thoughts were coded as ”other” if they did not fit into
the "issue" or “visual" categories. If several thoughts
were strung together, coders were instructed to divide the
content into separate thoughts and code each according to
the protocol outlined above. Intercoder reliability using
Scott’s Q1 was 86.21%. Disagreements were resolved
through discussion.

QQDQIQBI§_QXBIUBLIQD was determined by subjects’
responses to ”rate Senator Bob Moran" on thirteen 7-point
semantic differential scales. Scale items were chosen
using two guiding principles: (1) Items must be
appropriate for evaluating a candidate’s personality,
character and performance. (2) Items must address the
dimensions (i.e., evaluative, activity and potency) of
attitude (Osgood, Suci a Tannenbaum, 1957). The scales
were composed of the following bipolar adjective items:
evaluative (dishonest-honest, unintelligent-intelligent,

insincere-sincere, friendly-unfriendly, attractive-

38
unattractive, trustworthy-untrustworthy), activity and
potency (decisive-indecisive, weak-strong, warm-cold,
knowledgeable-unknowledgeable, unlikeable-likeable,
consistent-inconsistent, competent-incompetent). Items
were reversed at intervals to prevent response bias. To
determine internal consistency of the candidate evaluation
scale, Cronbach’s alpha was calculated (a = .8058). It
was determined that all items contributed to the scale, as
alpha did not increase when any item was deleted
(Table 1). Hence, a composite candidate evaluation score
was calculated for each subject by summing all candidate
evaluation items.

¥Q&§_L15211DQQQ was assessed by subjects’ responses to
"how likely is it that you would vote for Bob Moran in an
election if you had the opportunity" on a 7-point semantic
differential scale ranging from "1" very unlikely to "7"
very likely (M = 3.135, 80 = 1.434).

Messege_exeleesisn was measured by subjects’ responses
to ”rate the message" on five 7-point semantic
differential scales selected to tap the evaluative
dimension of attitude (Osgood, et al.). The scales were
composed of the following adjective sets: reasonable-
unreasonable, illogical-logical, well-written-not well
written, sensible-not sensible, poor job—good job.
Cronbach’s alpha was calculated for the message evaluation

scale to determine internal consistency (a = .8613). It

39
Table 1

Candidate Evaluation Scale

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[ 13 55. 4539 90.9457 9.9970 .5059]
..xmmma
Item 1 7928
Item 2 8016
Item 3 7943
Item 4 7924
Item 5 7962
Item 6 7946
Item 7 7825
Item 8 7906
Item 9 7924
Item 10 8027
Item 11 7906
Item 12 7872
Item 13 .7891

 

 

 

 

40
was determined that all items contributed to the message
evaluation scale, since alpha did not increase when any
item was deleted from the scale (Table 2). Therefore, a
general measure of message evaluation was formed for each
subject by summing all message evaluation items.

Pcecessjng effect was determined by subjects’
responses to "how much effort did you put into
understanding issues in the message" on a 7-point semantic
differential scale ranging from "1" very little effort to
“7” very much effort (M = 3.536, SD = 1.724).

Messege_eleesseseess was measured by having subjects
indicate “how pleasant was the commercial" on a 7-point
scale, ranging from "1" very unpleasant to ”7" very
pleasant (M = 2.6492, SD = 1.246).

A menipglssien QDQQK was administered to determine if
the manipulation was successful (i.e., high audio
comprehensibility versus low audio comprehensibility).
Subjects were asked to rate “how comprehensible did you
feel the message was" on a 7-point scale with "1"
indicating incomprehensible and “7” indicating
comprehensible.

Finally, subjects were asked to resell and LQQQSDIZB
information from the commercial (Appendix). Separate

scores were derived for items recalled correctly (M =

.7765, SD .3034) and items recognized correctly (M =

3.163, SD 1.237). Intercoder agreement for the recall

41
Table 2

Message Evaluation Scale

 

......

......
.....

......

. - u

e
e

.......
......
......

......

.......
uuuuuuuu
.......
........
n

.......................................

..........................................
uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu
..........................................
...........................................

.............

v t v

..........
..........
..........

 

......

 

 

 

 

 

 

 

 

 

 

 

 

,,,,,, 16.3665 36.8860
Item 8333
Item .8415
Item .8395
Item .8199
Item .8279

 

 

 

 

42
measure was calculated using Scott’s pi. It yielded a
reliability coefficient of 86.6%. Disagreements were
resolved through discussion.

One hundred ninety-two subjects completed all
dependent measures and were categorized as either issue
processors or image processors in Stage I. (As indicated,
subjects not reaching the 60% issue or image criterion in
Stage I were excluded from data analyses.)

Data were analyzed in a 2 (Comprehensibility) X 2
(Schema) factorial design. T-tests, two-way analyses of
variance and oneway analyses of variance with e ereri

contrasts were employed.

Chapter III

RESULTS

This chapter presents the results of analyses of the
data collected according to the concepts outlined in
Chapter 1 and the methods described in Chapter 2. First,
results of the manipulation check are presented. Next,
results of hypothesis testing will be addressed in the
order hypotheses appeared in Chapter 1. Finally,
additional data analyses for which no hypotheses were

proposed are presented.

Manipulation Check:

A paired t-test demonstrated that subjects in the low
audio comprehensibility condition indeed perceived the
message as less comprehensible (M = 1.955) than subjects
in the higher audio comprehensibility condition (M =

2.721), §(190) = 3.96, Q.<.05. (See Table 3.)

Hypothesis 1A:
Hypothesis 1A suggests that image processors will
encode more visual information than issue processors.

A paired t-test suggested that subjects identified in

43

44

Table 3

Manipulation Check Message Comprehensibility

Comprehensibility

 

- a v o a
e I a e o .
o . c u e
n - c e - .
e n . e .
e . u . n
u . a e .
e e . o c .
e . u e u
e n u I e .
. a u e I
w o . e u o
e . u o I A
n e c e e a
n . e u e
u . p u u n
. a . e u
a o n e 0 c
e - u e e A
I I I I I e

.
o
u
.
e
5
u

e e
e e
e .

a'riis

e
. o
.

 

EMA

5 2
”AP 4 1
BU. 5 2
“to“ 9 7
ﬁru e I
5.6.“. 1 2

 

 

..... ..... .....
...... ..... ......
..... . .........
...... ..... ......
Ole-3‘ IaIvIIe.IIv
...... ..... ......
.....A .... u.....
T..... ..... ......
...-e urea Yen-v.
...... ..... ......
..... .... r.....
...... ..... ......
.....A .... .......
.. ... ..... ......
..... .... ......
...... ..... ......
..... .... ......

...... ..... .....,
noun. one! You...

...... ..... ......
...... e... ......
...... .. .. ......
.....A 0.1. 1e....

'11... IIIII eI-Iv.
..... .... ......

...... ..... .. ...
...... .... ...m..

...... .. .. ... .
... .... ... .

e. Ate-l In. -

... . ..... ... .
...Mu. .... ...nH.

... . ..... .. .

no. Av. 4.
...AU. ..MW ..AUF
... . .. v.

... .. .. . ......
.. ... +.| .

...r. 20.....1:
-- Au. - T. .

.. . .... ... ..
.. ..e 7. .

...... ..... ......
.....I..... .....

r>.. ..rk» r.rrr

 

 

 

 

 

 

T Table

.000

 

 

3.96

 

 

e
u
.

5.551255

.-
e

no
a
A

u e e
e e e
. e u
. n u

a
e
e

 

 

 

 

e
e n
u
. .
n
a

. .

1 1

III]

...
IAIAL

 

45
Stage 1 as image processors produced more visually-
oriented cognitive responses (M = 2.812) than subjects
identified as issue processors (M = 1.667), t(190) = 4.52,

Q.<.05. (See Table 4.)

Hypothesis 1B:

This hypothesis predicted that subjects identified in
Stage 1 as issue processors would produce more issue-
oriented cognitive responses than subjects identified as
image processors. As predicted, subjects identified in
Stage 1 as issue processors produced significantly more
issue-oriented cognitive responses (M = 2.253) than
subjects identified as image processors (M = 1.667),

3(190) = 3.03, Q.(.O5. (See Table 5.)

Hypothesis 2:

Hypothesis 2 predicted that both issue and image
processors would encode more visual information under low
audio comprehensibility conditions than under higher audio
comprehensibility conditions. There is partial support
for the impact of comprehensibility on visually-oriented
cognitive responses. While issue processors produced more
visually-oriented thoughts under low audio
comprehensibility conditions (versus higher) (M = 2.040
vs. 1.327), 1(188) = -2.163, g.<.05 as predicted, image

processors did not produce significantly more visually-

46

Table 4: Hypothesis 1A

Visually-Oriented Cognitive Responses by Schema

C

 

:-: t 243?;
2.812

I

Grou

1.667

 

 

....
....
......
......
......
'I-Iee
......
Incl-I
......
Feel.-
......
......
1..
.....L
.....A
1.....
.....I
.....r
'la n
.. L

 

II:
:m. 3
..
.... . 6
..u
.. .4 8 el
IIIIC I-III'IIIII
..on. ....vue-e.
.I-‘I. ‘eII IleO
...-. ...-.71....
IIIII III-ﬁollln
acct-A .ee.T-.adu
...U; .9...
.A .3123
...r. . :8
T
I

 

 

 

 

 

T Table

 

.000

 

 

 

 

 

 

 

 

 

47

Hypothesis 18

Table 5

Issue-Oriented Cognitive Responses by Schema

Schema

 

 

1
e
1

1'5.

rT—s r
.-
.
..
1
.
1
'l

a
1
o

 

 

I
1
1
O
u

n
e
1

iéuﬁ
1.667
2.253

I
a
1
1..
1
a
1
1
111

 

.101-

  

 

 

 

 

 

 

 

  

 

 

T Table

 

[.003

3.03

 

 

 

 

 

5

a

 

 

1
1

1 e
1 1
o

1
1

1
a
1
1

i

1
1
1

BE

A

11

A_LLJ FA

1
1
1
1

 

 

48
oriented thoughts under low audio comprehensibility
conditions (H = 2.974) than under higher audio
comprehensibility conditions (H = 2.633), 1(188) = -.936,
9.).05. (See Table 6.)

Hypothesis 3A:

Hypothesis 3A predicted that image and issue
processors would rate low audio comprehensibility messages
as more unpleasant than higher audio comprehensibility
messages. An anova on message pleasantness ratings
revealed a main effect for schema. Overall, issue

processors rated messages as less pleasant (H = 2.413)

compared to image processors (H 3.062). E(1,188) =
10.649, p.<.05. Furthermore, a main effect for
comprehensibility emerged. Low audio comprehensibility
messages were rated less pleasant (M = 2.464) than higher
audio comprehensibility messages (M = 3.061), E(1,188) =
4.891, p.<.05 (Table 7). However, a oneway analysis with
contrasts revealed only partial support for the
hypothesis. As predicted, issue processors rated low
audio comprehensibility messages as less pleasant (M =
2.020) than higher audio comprehensibility messages (M =
2.727). 3(187) = 2.869, p.<.05. Although image
processors’ mean pleasantness ratings were lower under low

audio comprehensibility conditions (H = 2.838) than under

higher audio comprehensibility conditions (H = 3.061), the

49
Table 6: Hypothesis 2

Visually-Oriented Cognitive Responses by Comprehensibility

 

 

 

 

 

 

ANOVA TABLE
”5:223:-:3:1:3:1:3:3:3:3:3:3:1:3:3:3:?:3:3:521:3:1:T:-:3:3:37:1:3:3:33:323:323:5:3:3:23:31::3t3r723:3:1:3:1:3:T:1f——
, $061306?§5§3§5§E§3§5§3§5§P§3§3§P liﬂﬂﬁﬁihiféff ..DE;.;.:.;. ”can é3§3§5§3§3F§3§5§3§335§3§ §5E3§3§325§5§3¥§5§
g ''''' 50313161366: 2 ''''' 9706:5631 11111361553365: fag-:6
Egéﬁéﬁiﬁﬁﬁﬁﬁiiﬁi 64.260 1 64.260 22.564 .000
Comprehension 13.030 1 13.030 4.575 .034
”i8i3?§é£1§ﬁ§§§é 2.247 1 2.247 .769 .376
ﬂﬁégiﬁﬁgiiéfiigi 526.653 185 2.646
ﬁiéﬁiiﬁ §§§§§5 § 604.519 166 3.216

 

 

 

 

 

 

 

 

A Priori Contrasts

01: Image Proc./Low Comp. vs. Image Proc./Higher Comp.
02: Issue Proc./Low Comp. vs. Issue Proc./Higher Comp.

 

 

 

 

 

 

 

 

 

 

 

 

 

{gaggagggggag§gigggTﬁhﬁﬁéiﬁﬁﬁéﬁﬁéiﬁﬁﬁiTéééﬁafﬁéaﬁéﬁiéﬁéﬁﬁ
”fﬁiﬁeiibwigﬁﬁﬁiﬁﬁﬁ 38 2.9737
gim§§§7ﬁiéﬁégiiiﬁﬁi 49 2.6327
Vigéﬁéyiéﬁiﬁiiﬁéﬁiﬁ 50 2.0400
Its-119111190" 55 1.3273
1 vatuaSError Tvalue TProb
iéiiiﬁﬁﬁiiﬁi -.3410 .3644 -0.936 .351
f§§§§§§§§§§ -.7127 .3294 -2.163 .032

 

 

 

 

 

 

 

50

Hypothesis 3A

Table 1

Message Pleasantness by Schema/Comprehensibility

Schema

 

 

 

-
.

36

E

  

-90
3.062
2.413

.906

 

 

 

 

 

 

 

 

 

 

Comprehensibility

 

   

65

A

- n

n
u
4
o
0
u
a
2

a...

A

o
0
A

2.464
3.061

6663

u
A

 

I
e

n
o

a
a

a
a

 

 

 

 

 

 

 

 

ANOVA TABLE

 

.001
.028

.129

 

649

10.

4.891
2

320

 

18.924

8.692
4.123

1.777
1.922

 

1
1
1

185

188

 

 

18.924

328.768
361.407

 

 

 

 

 

00'

u n

a

a
O
.

603

 

66

 

5m

.
1‘:
E"

_.

 

é

.

n

6

b

 

 

o
u
n

 

 

12.5

4.9

U3

664

   

 

 

 

 

 

51
difference was not statistically significant, 1(187) =
.831, 9.).05. (See Table 8.)

Hypothesis 3B:

This hypothesis posited that issue and image
processors would rate low audio comprehensibility messages
less favorably than higher audio comprehensibility
messages. Mean message favorability scores indicated a
main effect for both schema and comprehensibility. For

the schema main effect, issue processors rated messages

less favorably (M 15.211) than image processors (M =
17.784), E(1,188) = 8.647, p.<.05. The comprehensibility
main effect indicated that low audio comprehensibility
messages were rated less favorably (H = 14.813) than
higher audio comprehensibility messages (M = 17.706),
£(1,188) = 11.186, p.<.05 (Table 9). The hypothesis
received marginal support. Consistent with the
hypothesis, issue processors rated low audio
comprehensibility messages less favorably (M = 13.140)
than higher audio comprehensibility messages (M = 17.091),
t(187) = 3.499, p.<.05. Although in the predicted
direction, image processors did not rate low audio
comprehensibility messages significantly less favorably (M

= 17.000) than higher audio comprehensibility messages (M
= 18.396), 1(187) = 1.112, p.>.05 ( Table 10).

52
Table 8: Hypothesis 3A

Message Pleasantness by Schema/Comprehensibility

A Priori Contrasts

01: Image Proc./Low Comp. vs. Image Proc./Higher Comp.
02: Issue Proc./Low Comp. vs. Issue Proc./Higher Comp.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

................................................................................................................ 5
EGroups ‘ 1101110618?233?;5;52§3§§§=§:§3§1§3353355355? ' 66666“11.763.11.395?322553223553
3iﬁ§§§¥§§i§i§f§§§fﬁ 37 3.0612
”16606761ﬁhéiiiiiif 49 2.6376
fig.déitaugisigiiza 50 2.020.
“iégﬁéiﬁiéhgéiiiiii 55 2.7273
§5§§5§§5§§§§§316353§£§5973666665”¥$§§§1§3§¥§§i§$§$ﬁ$§§§
E§$§§§§5§3§§.2234 .2746 0.613 .417
E62§§§§§§§§? .7073 .2466 2.669 .005

 

 

 

 

 

 

 

53
Table 9: Hypothesis 38

Message Evaluation by Schema/Comprehensibility

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Schema
iéibuﬁﬁﬁﬁ¥ﬁéﬁﬁﬁéﬁﬁﬁfNumbéﬁﬁﬁﬁﬁiﬁiﬁﬁEEEIGﬁéQEVMéénéﬁﬁﬁiﬁﬁﬁ
{IMéééﬁéﬁﬁﬁﬁﬁﬁﬁiiﬁﬁ 86 17.764
ﬁiéédéﬁﬁiiﬁﬁiaﬁﬁﬁéﬁ 105 15.211

Comprehensibility
argupﬁﬁﬂﬁﬁﬁrégggﬁkr'N9m06ﬁiﬁﬁﬁﬁﬁﬁéﬁﬁgqg Gr600“Meansmaéﬁﬁﬁi
HL6w£¢§36536£§£5525 66 14.613
.H1ghaazxuaggsggggg 103 17.706

ANOVA TABLE

Source Sumwf _ iéiDLELéiéiél ...... -§iéiii%i§1+'§iiﬁéiéiiféii F
F3z3:525:3212323332323523131212323:izizisisi: ..... 5911917955 Iéiiiiliiiiiliii . . .5999??? 4.4169“? 1891991..
{schemaiﬁﬁﬁﬁgﬁéi 266.634 ‘1 266.634. 6.647 .004
{Cbmbtéhénéjbhf, 373.641 1 373.641 11.166 .001
”ihiéfﬁéiiéﬁﬁﬁii 76.507 1 76.507 2.291 .132
Léeéiduiiﬁﬁﬁﬁﬁﬁé 6246.045 167 33.401
TD‘blL 7008.346 190 36.886

 

 

54

Table 10: Hypothesis 38

Message Evaluation by Schema/Comprehensibility

A Priori Contrasts

 

 

 

 

 

 

 

 

 

 

 

 

01: Image Proc./Low Comp. vs. Image Proc./Higher Comp.

02: Issue Proc./Low Comp. vs. Issue Proc./Higher Comp.
iéiéﬁﬁgiiiiﬁiiiiiﬁiiﬁﬁ5$§§§§§§§§§ﬁﬁ§5§iéigﬁﬁiﬁgiﬁéiiﬁiﬁiﬁ
“i.§§§7;§.§§§§££;3§ 36 17.0000
"$66067Hiéhéﬁﬁliiﬁﬁ 46 16.3956
“iééﬁéitéﬁéiéﬁiﬁéii 50 13.1400
“I36667ﬁ13héiiiﬁiiﬁ 55 17.0909
55711175065
iéiﬁiﬁﬁﬁﬁﬁih 1.3956 1.2549 1.112 .267
3§§§§§353§§3 3.9509 1.1293 3.499 .001

 

 

 

 

 

 

 

 

55
Hypothesis 4A:

Hypothesis 4A predicted that both issue and image
processors would evaluate the candidate less favorably in
low audio comprehensibility conditions versus higher audio
comprehensibility conditions. There was no main effect
for schema on candidate evaluation. However, a main
effect emerged for comprehensibility. Subjects rated
candidates less favorably under low audio
comprehensibility conditions (M = 54.023) than under
higher audio comprehensibility conditions (M = 56.794),
£(1,188) = 4.376, p.<.05 (Table 11). The hypothesis was
not supported, however. Although the patterns of means
were in the predicted directions, they were not
statistically significant. The mean candidate evaluation
by issue processors under low audio comprehensibility
conditions (M = 53.245) was not significantly lower than
under higher audio comprehensibility conditions (M =
55.852), 1(186) = 1.479, 9.).05. Furthermore, image
processors’ mean candidate evaluation under low audio
comprehensibility conditions (M = 55.026) was not
significantly lower than under higher audio
comprehensibility conditions (M = 57.490), t(186) = 1.276,
2.).05 (Table 12).

Hypothesis 4B:

This hypothesis predicted that both image and issue

56

Table 11: Hypothesis 4A

Candidate Evaluation by Schema/Comprehensibility

 

......
.....A
......
.....
......
Ciel.
......
...-o
i.....
.....
......
.....
......

a
n
.
.

.
a
6

e858

Tun.
On-
I...
...
- '-
...
.

a.

.

...
...

56.603
54.611

”Mao.”
Mar“

 

 

 

66.1

9
4
.
.
-

 

 

 

 

 

 

Comprehensibility

 

54.023
56.794

 

 

H

 

 

 

 

 

 

 

ANOVA TABLE

 

.144
.038
.972

 

 

2.155
4.376

.007

 

 

169.235
343.628

I

.567
78.534
80.102

 

|
. .
5
c
.
.
e
I
A

01.:

4
u
e

a
.
.
.

1
1

1

185

188

 

e
.
9
.
o
I

..
.-

‘6

 

.

.
.
.
o
4

169.235
343.628

567

14528.829

15059.185

 

 

 

 

.-.... ........... ...... ..... ......
......... ..... ..... ............ .....
....... . . ..g......w....- .....A......
......... ..... ..... ............ .....
u... .... ...-.... .v...-. .....A......
......... ..... ..n. ............ .....
. ....... ........ . ..... ............
......... ..... ... ............ ......
. -..... . ......Q. ..... ............
........q .... ... ............ ......
6... g. . ... 1......vo. .....A......
......... .-.. .31 ...n....... .....
‘IIIO.C.I Int-1.4: .I II .....AOI-ICI
....... . ..... ... ....... .....
......... . .. ...s. ..O ..... ......
......... ..... ... ....... ......
. .. .... .... .7. ... ............
...-.94 a 4.6-. I. I4 90.1!IIe-O- no...
......... ........n.v.. .. ............
......... ..... .... ....... .....
.......,. . ...... .... ..1.......
.... .... ... ... ... .....
. .....- ... .v.. . ... +......
......... .n. ...c....a. .....
llvlvwl I I... I! vVlI I III 4.....-
.. .. . .. . ... .... . .....
.. a ...e.f..a ...u 2......
.. ..... .. ,... ... .... . .....
.. .... .. .... ..v.. .. .. ..,...
.. .... .. ...r ...'_.:.d. ..1.
.. .. .... ... ... ...
.. ...... .. ... . ... ....... ..
...r.... . ....pv..e ..01 ...l.
.. ..... ., .. .. .. ..

5. bin! 0. .ar aveotu ... .0 I
... ..... .. . .. .. .... . ..

.. ..... .. . .... .

.. .... ..c ...Q ..n ..e

.. ..... . ... .

I. v... n 4". nth-vi. on

. ..... ..s ..c ..I...R.

. .... . .. .f.. . ..
......... ..... ............ .....
. ....... ...........v..... ..... ......
.....n. ..F tpbi .b b ... .....

 

 

 

 

 

 

 

 

57
Table 12: Hypothesis 4A

Candidate Evaluation by Schema/Comprehensibility

A Priori Contrasts

C1: Image Proc./Low Comp. vs. Image Proc./Higher Comp.
02: Issue Proc./Low Comp. vs. Issue Proc./Higher Comp.

 

 

 

 

 

 

 

 

 

 

 

 

 

Térﬁﬁééﬁiiﬁﬁiiiﬁﬁﬁﬂiiﬁﬁmﬁéﬁﬁiiﬁiﬁﬁiiﬁﬁﬁﬁféiéﬁﬁfﬁéiﬁéﬁiiﬂiﬁi
Viaaaéitaaiaﬁﬁﬁiiiﬁ 36 55.02.
"18336761EEEE§§¥§§§ 49 57.490
”iéébéiiéﬁﬁ§§§§§§5§ 49 53.245
ﬁié§aéiﬁiéﬁééiiiééﬁ 54 55.652
tWilma$151673!“Tl'alue TProb
Eé$§§§§§§§§32.4635 1.9306 1.276 .204
§§§§§§§§§§§12.6070 1.7624 1.479 .141

 

 

 

 

 

 

 

58
processors would be less likely to vote for a candidate in
low (vs. higher) audio comprehensibility conditions. The
data indicated no main effects for schema or compre—
hensibility on vote likelihood (Table 13). Furthermore,
the data did not support the hypothesis, although vote
likelihood means were in the predicted directions. Issue
processors’ vote likelihood ratings were not significantly
lower under low (vs. higher) audio comprehensibility
conditions (M = 2.840 vs. 3.146), 5(188) = 1.163, p.>.05.
Additionally, vote likelihood ratings for image processors
under low audio comprehensibility conditions (M = 3.237)
were not significantly lower than under higher audio
comprehensibility conditions (M = 3.347), 3(188) = .339,
Q. >.05 (Table 14).

Hypothesis 5A:

Hypothesis 5A furthered that issue processors would
rate the message and candidate less favorably under low
audio comprehensibility conditions than would image
processors. There was some support for this premise.
Issue processors rated the message less favorably (M =
13.140) than image processors (M = 17.000) under low audio
comprehensibility conditions, t(187) = 3.103, p.<.05
(Table 15). However, issue processors did not evaluate
the candidate significantly lower (M = 53.245) than image
processors (M = 55.026) under low audio comprehensibility

conditions, 3(186) = .923, p.>.05 (Table 16).

59

Hypothesis 48

Table 13

Vote Likelihood by Schema/Comprehensibility

Schema

 

    

e. o
...-0
60.0..
...-o

666

 

 

 

 

 

 

 

 

Comprehensibility

 

 

 

 

 

 

 

 

 

ANOVA TABLE

 

64
O 296

.1

640

 

.952
1.099

1

.219

 

4.003
2.254

.449
2.051
2.055

 

1
1
1

185

 

if
uares

 

o
.
a
9

150
q

 

385.527

 

 

 

 

 

0
.
u

on

:1
5n

.
v

.
n

4
a .
e
. .

a
a

.-
e

.-
.

anréﬁ
é

 

b
.
>0
p.
.
y.
.
.

 

 

o
a, .a
.
4*;

 

ﬁt

I

 

24.“..-

L

d

 

. .
I'.
.-
. .
- .
.

o
.
-

UK.

a

n

. . 4
n

. .

 

.

59-

 

u
v

o

u
4

o
.

u
.

.

btuPD.

 

        

 

 

 

60

Table 14: Hypothesis 48

Vote Likelihood by Schema/Comprehensibility

Ci: Image Proc./Low Comp. vs.
Issue Proc./Low Comp. vs.

C2:

A Priori Contrasts

Image Proc./Higher Comp.
Issue Proc./Higher Comp.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

'-:-:;:-:;:-:;:;:;:::;:;:-3:37. :;:;:-:;:;:?:-:?:;:;:;:T:f:;';:;::.::;:::;:;:;:;:;2::;:;:;:32;:jzfzfz?:;:;::r;:;:3:;:;::t;:;:;:;:;:;: 2?:3:7:T§?:;:;:g:;:;:;:;:;:;:
Ei§??§?iéié??§?§3§3i?i?§?§9iéﬁéiéé iéiéiéééééiﬁT-EEEWJFHC' 2.23252335rrtojr: 1€?§?§?§?§32‘|?S§5¥§§Inleié 721.31?be I
c1 .1101 . 3245 . 339 . 722
102 . 3055 . 2627 1 .163 . 276

 

 

 

61
Table 15: Hypothesis 5A
Low Comprehensibility Message Ratings by Schema

 

 

 

 

ANOVA TABLE
59W“ 0” SUméf MW! FRatW FPWb
ﬁﬁ 5§3§5§5§3§5§3§55 §5§3§3§3§3§§99§r§§f ........ §HU§Y§Q 3?55333333533553???if? 53E3§3§3§3§3§3§3§3§5§3§5§3§
1665110920,? 3 762.3009 254.1003 7.6075 .0001

$551453...

:fﬁfiiﬂiﬁfff; 187 6246.0046 33.4013
,§f¢H§$iﬁ

natal 190 7008.3455

 

 

 

 

 

 

 

 

A Priori Contrast

C1: Issue Proc./Low Comp. vs. Image Proc./Low Comp.

 

 

 

 

 

 

 

 

17""d‘YVVVVIIUjerVVY‘VVr'I IV'VY‘"VVYY'V‘VV‘V‘V'I'VI'V V'VVYVVTva'VIVIIIIYV'VVVV‘V
.....................................................................................
....................................................................................
.....................................................................................
...................................................................................
...........................................................................
................................................
...............................................
................................................
...............................................
.. . .................................................
LA _A_ LA ‘__L_A__A__A _A_‘__A_I_.A _A AA IAAAA‘AAAJAAIAAA-LAA‘AI‘AALL‘AAA-AAA L+LLA“J AAAAAAAAAAAA
............................
............................
............................
............................
.........................
.............
..............
.............
.............. .
.............
........................
............................
.............................
............................
............................
........................
..............
.............
..............
............. .
.................
...........................
v v v v- ‘ --1vvvvv-v-rvvvv --- v v----- vsv----v-----v-~
....................................................................................
...................................................................
.....................................................................................
..................................................................
....................................
..........................
..........................
..........................
OOOOOOOOOOOOOOOOO .llhee
.........................
...................................................................

 

3.8600 1.2438 l3.103 .002

 

 

 

 

 

 

Table 16: Hypothesis 5A

62

Candidate Evaluation by Schema/Comprehensibility

 

 

 

 

 

 

 

ANOVA TABLE
:Zji;I;I;IjI;ZjIjI;IjZ:IjI; :I:I:Z:.’;I:I;I:I I;3:3;2:2:::I;I:Ij37:2;ijf323231:Z:E;fji;3fijl;lji:1;5{Iﬁjfjijljﬂﬂﬁfijfﬁ:Iﬂzijiﬁﬁfqi: ;.:.:.:.:.j.:.:.;4:773}:
HWHWQX. .... 95:25:22i?§§§§§§5§§§5§5§§§HH¥5H§.:;.§.Z.EEH§H§§§§§§§§§§§533535391?“ .... 5.95993...
ﬁﬁﬁﬁﬁgﬁﬁﬁiiﬁﬁﬁi¥ﬁ§£§$§H§t§§ ........ 5666653;3633533552253333332
Eféﬁéﬁééﬁffj 3 457.6475 152.6492 1.9119 .1291
“9.523‘193'3'3:
1111mm 166 14641.0946 79.7906
,sr6Qp§:;
Total 169 15296.7421

 

 

 

 

 

C1: Issue Proc./Low

A Priori Contrast

Comp. vs. Image Proc./Low Comp.

 

 

 

 

 

 

 

 

 

.v‘vvvvvTrrrrvvwirvvvwvrv. wvvv.....rvvuvvvv-uvvavvrr- viv‘v'vv‘v vvvvvvvvvvvvvvvvvv
.....................................................................................
....................................................................................
...................................................................................
........................................................
.....................................................................
...............................................
y ...............................................
................................................
L . ...............................................
.. ....................................................
.A.-..L. LL. .. ...“...A..A4 ................ ..
............................
............................
.............................
............................
..................
.............
..............
.............
.............. .
.............
.........................
2-224 ..............
............................
.............................
............................
............................
........................
...............
.............
..............
............. .
...............
.....................
vv~~ v v vv—fvv vvv v 1 -- vverV,
IIIIIIIIIIIIIIII quOIIIICIBBIOIIOD .....WOICOIIIBIII IIICBIO....I ..IIIOIUCIIIUOBII
...............................................................................
................................................................................
...............................................................................
................. - ............-. ......W.... ......... ... .......
................ .. ......u.. .. ..... .....-
................. . ........ . ..... ......
......................... .. ..... . ....-
................. . ........ 3‘23 0.... b.....
................................... .. . .....
...............................................................................

 

................

 

[01 ......

1.9308 .923

 

 

 

 

63
Hypothesis 58:

The proposition that issue processors will be less
likely to vote for the candidate than image processors
under low audio comprehensibility conditions was tested by
this hypothesis. The data did not support the
proposition, although vote likelihood means were in the
predicted direction. Issue processors’ vote likelihood
ratings under low audio comprehensibility conditions (M =
2.840) were not significantly lower than image processors’

(M = 3.237), §(188) = 1.29, 9.).05. (See Table 17.)

Hypothesis 6A:

Hypothesis 6A proposed that issue processors would
perceive expending higher amounts of energy attempting to
process low audio comprehensibility messages than would
image processors. A oneway analysis with contrasts did
not confirm this hypothesis as the mean scores were nearly

identical. In low comprehensibility conditions, issue

processors did not perceive expending more effort (M

3.620) than did image processors (M = 3.632), 1(188)
.031, p.>.05. (See Table 18.)

Hypothesis 68:
This hypothesis furthered the notion that issue
processors would rate low audio comprehensibility messages

as 1688 COMOPBhOﬂS‘ibTB than 1111896 DI’OCBBSOI’S. Data

64
Table 17: Hypothesis 58
Vote Likelihood by Schema/Comprehensibility

 

 

 

 

 

 

ANOVA TABLE
mﬁﬁﬁﬁfmﬁﬁSﬂﬁﬂﬁﬁﬁﬁﬁ$§$ﬁ$ﬁ$ﬁﬁﬁﬁﬁﬂﬂﬁﬁﬁﬁ$ﬁz
59145991123..HE;:;:;:;:33333555555335.1591???5.95%iiggz§.}.§.I¢I!HSESEEES§EEE
.... :Esngﬂézgii §2§s§e§s§3§s§ssguajr§esf:=:- , , , . 6500816692
Between 3 6.9623 2.3174 1.1301 .3362
{95999941-
within 166 365.5266 2.0607
ﬁGtéHbﬁjﬁé
Total 191 392.4792

 

 

 

 

 

 

 

 

A Priori Contrast

C1: 18806 PFOC./LOW Comp. VB. Image Proc./Low Comp.

 

vvvvrvrvvvvvvv‘vvuvvvvuvrv ‘vvvvvvvvvvvvvrrTw—v—rv-vvrvvv vrvrvvvvvwvwrvvrv v —v vvvvvvvvv

 

 

 

 

 

 

 

 

r .....................................................................................
....................................................................................
....................................................................................
. ............................................................................
............................ . ................................... ...........----.
.................. . .................. . ..........-
p ................... . . ................... . . ..........
.................. - ................... .. ...........
. ................. .. ................... . ..........
... .. . .........
..................... 141.111.22J LL‘I .. .-.
............................
............................
.............................
............................
.. ......... . ........
. ............
.. ..............
. .............
y. ..............
. ............. I
.........................
#AA. AAAA#Q
............................
.............................
............................
.............................
. ...............
. asue 21°"; ....... ' .....
. ..............
. .............
. nnnnnnnnnnnnnnnnn .
. ..... .....
...................................................................................
...................................................................................
...................................................................................
.......................................
...............................................................
........................................
........................................
....................................
.......................................
I .....
...................................................................................

 

 

 

 

 

...............
.....

 

65

Table 18: Hypothesis 6A

Processing Effort by Schema/Comprehensibility

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ANOVA TABLE
25991563151EDIE;sumaf iiMeah .L.Q.F.I:5§1=ji'éf PM: if
Prawaxamanaiﬁﬁisawaxaquar55 ..... $9H§r9§.ﬁﬁﬁﬁﬁﬁﬁﬁﬁﬁﬂﬁﬂﬁmﬁﬁﬁi
{61632566651 3 1.4960 4967 .1667 .9167
615921115333?

[315117.30in 166 562.2466 2.9907
VGﬁﬁHP§.L
137123532:-;:;:;:;:;:;
EfTOpﬂjfzfﬁzfzfr 191 563.7448

A Priori Contrast
Cl: 188116 PFOC./LOW Comp. V8. Image PPOC./LOW Comp.
'91'500055gigsgégigéiégéfi§§§éiégigégégiiii Number“ ‘G'r'ou'p"'HeansﬁézésftSéféfzitisfe
Image/mug. 38 3-6316
’j15511167Léwiéisisisisiiéééiéiéiéiééé? 50 3-6200

 

 

 

 

 

 

 

 

 

 

 

 

 

66
supported the hypothesis. Issue processors rated the
message as less comprehensible (M = 1.720) than image
processors (M = 2.263) under low audio comprehensibility

conditions, t(188) = 1.906, p.<.05. (See Table 19.)

Additional Analyses

Although no specific hypotheses were offered,
additional analyses were performed to address the impact
of schema and comprehensibility on recall and recognition.
A 2 (Comprehensibility) x 2 (Schema) analysis of variance
was conducted on recall and recognition. Main effects
were revealed for both schema and comprehensibility on
recall. The main effect for schema indicated that issue
processors recalled more information (M = .910) than image
processors (M = .643), £(1,188) = 10.478, p.<.05.
Furthermore, the main effect for comprehensibility
revealed that more information was recalled from higher
audio comprehensibility messages (M = .987) than from
lower audio comprehensibility messages (M = ..566),
E(1,188) = 4.371, p.<.05. No interaction effect emerged.
(See Table 20.) Similarly, main effects were revealed for
both schema and comprehensibility on recognition. The
main effect for schema indicated that issue processors
recognized more information (M = 3.507) than image
processors (M = 2.821), E(1,188) = 23.742, p.<.05.

Furthermore, the main effect for comprehensibility

67
Table 19: Hypothesis 68

Message Comprehensibility by Schema/Comprehensibility

ANOVA TABLE

 

 

 

 

59§$5$5353393QWﬁﬁ$§$ﬁﬂﬁﬁf$ﬂﬂﬁﬁﬁﬂﬁ$ﬁ$§$ﬁﬁ$ﬁ3$ﬁ$ﬁ$iﬂﬂﬁﬁﬁﬂﬁﬁr
6994:99 ......... 0F F5ﬁiéiiéisésésﬁumsisgf.. ggiggiﬁaénéiséséﬁés 2.2.5625331219 . ..E...F’.r9b.-. .....
33333333555335?¥S§Q§$§Q§f§§ ..... 3§Q§f§§§§§§$Q$Q¥5¥Q$Q§Q§E§¥
Efﬁéiﬁééﬁf 3 37.2793 12.4264 7.0906 .0002
.Gﬁwbﬁifi

:fﬁiiﬁﬁgﬁﬂff; 166 329.4655 1.7525

KGMQPEESI

{Total 191 366.7446

 

 

 

 

 

 

 

 

A Priori Contrast

C1: Issue PPOC./LOW Comp. VB. Image PFOC./LOW Comp.

 

 

 

 

 

 

 

 

 

rv:vv.-.vvv:uvrr11w—rv—r vvvvvv r ‘l’v’Y‘ﬁf rﬁj r fwﬁ 1 fr TYﬁ
’- ....................................................................................
....................................................................................
.....................................................................................
...................................................................................
..........................................................................
...............................................
...............................................
................................................
..............................................
.........................................
."""""........LLA;4....ALA..L.JAA...16;..‘.A‘........L.A.A.AL...JLL ALAH'M ....... .. AAAAAA
............................
............................
t ----------------------------
............................
............
............
..............
.............
..............
............. C
............ L
............................
b ............................
............................
............................
...............
...............
.............
..............
............. .
.................
............ .... ..
L--. A-
'7- vvvvvv' v'vw rvﬁv ~ a vv w vvvvv wt Vf v+vv f
....................................................................................
...................................................................................
---------------------------------- bee-uo-an.a---u-e .n-aeeII'e-Ia-a-cn-eeaaee~aIIu-e-u
..................................................................................
........................................................
....................................
.........................................
....................................
......................... ‘ ..... ......
..........................................
.....................................................................................

 

 

 

 

 

 

68
Table 20

Reca11 by Schema/Comprehensib111ty

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Schema
ﬁéﬁéubﬁﬁﬁﬁﬂﬁﬁﬁﬁﬁﬁﬁﬁ'NﬂmbéﬁﬁﬁﬁﬁﬁﬁﬁﬁﬁﬁﬁﬁT636U67Mééﬁéﬁﬁﬁﬁ%ﬁﬁ
Piméééﬁﬁﬁﬁﬁﬁﬁﬁﬁﬁﬁﬁx 97 .643
itééuéﬁﬁﬁgﬁﬁggﬁgiﬁﬁ 105 .910

Comprehensibility
2161:6961i:I435$;ézé:2is§s§s:é:éié:é:i:s:~:is LNuimﬁérigéaé:%zéifzizéséié:éziaiiiszésgégiii Group Means
Mﬁéwfﬁﬁﬁﬁﬁfﬁﬁﬁﬁﬁﬁiﬁ 88 .566

‘H1ghara¢agggm¢ggae 104 .997

ANOVA TABLE
ﬁﬁéﬁﬁ¢§§§¥§¥ﬁﬁﬁ¥...ﬁﬂﬁCfo$9593:Ifﬁéﬁﬁﬁﬁﬁiiifﬁﬁﬁﬁﬁf5¥§5555
Fiﬁﬂﬁﬁﬁﬁﬁﬁﬁﬂﬁﬁﬁi..59Q§r§§.¥£ﬁ¥ﬁi..SQQEFQQ.,héﬂﬂkiguiﬁﬁé'l"
yschemayggggggag 39.612 1 39.612 10.479 .000
{Céhbféhéhiibhf_ 16.110 1 16.110 4.371 .041
”iﬁiéfiéﬁiéﬁﬁﬁﬁi 5.223 1 5.223 1.417 .203
ﬁﬁééiﬂﬁﬁiﬁﬁﬁﬁﬁﬁi 693.932 199 3.695
EEiiétgﬁﬁiéiiééiiséséségééégi@3523; 752.779 191 3.941

 

 

 

 

 

 

69
revealed that more information was recognized from higher
audio comprehensibility messages (M = 3.479) than from
lower audio comprehensibility messages (M = 2.847.),
E(1,188) = 4.305, Q.<.O5. The scheme by comprehensibility

interaction on recognition was not significant (Table 21).

70
Table 21

Recognition by Schema/Comprehensibility

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1' llllllllllllllllllllll D IIIIIIIIIIIIIIIIII 0 I I 3 o l ‘ IIIIIIIIIIIIIIIIIIIIII I ..... i
............................................. ‘........... ............................
.....................................................................................
...................................................................................................................................................................
. . ........................................ . ................
G .................... t' ................... r‘ ..........
t r p ................... .‘mba .................. .‘srb. ‘p ans ...........
4:... .................... " ................... a ..........
' ............................................................................ I i 1" I .......................
............................
, ............................
............................
.............................
L . .........................................
. .....................
>- mage‘-'.'u'.'.:-‘.'.'-'-:.'-'.'.'.:. . .:.: .
...............
. . 1 . w—rT .T vvvvvvvvvvvvvvvvvvv
.............................
............................
.............................
1" ........................................... 7
. . ...................
............................
E SSUG ............................... 105 3050
............................ ... 1' . VvT v—vwv‘rVﬁrw.....-...vvvvvrvvrvv.v.rjr.vvv
....................................................................................
r .........................................................................................................................................................
, ....................................................................................................
. ............................................................................
, GrOUP .................... Numb'r .................... Group ELMean'U ..........
. . ...............................................
. . .................... . .............................
. ............................................
, . ..................................................... .
....... 4 #. .L.-.. . .44L._L-L .....
............................
y ----------------------------
h ........................................................
I. 1 n ‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘
L‘ .0” ............................................. 84
,7 7 . ...................... e
I' lllll V lllllll 'l'O'I O I 07" IIIIIIII
............................
.............................
_.. .......................
L . . . '.;.'.'.:.'.:.'.'-‘.'.'.'.:.'.'. llll
‘ ‘ ' ' . . . .':'.':':'.‘I'.‘i'I‘I'Z':':‘.':‘I‘.‘:‘ O
................................ ‘l . . . . . . r . .. . . . . . .
’ ........................................................................ . ..........................................................................
........................ , ...
. . ............ . . . . ................. . ........
‘ SOL" Ge """"""""""""" sum 0' ' .'° ' Mean """""""" ' """"""" ' """"
............ . °, ‘ ' ',' ' - .n . ... ...... . ",'.'.'.'.'.'.'.'
......................................................................
ﬁ ................................................................................................................ _ ............... . .
. ........... s ‘arﬁs ................. aras Rafi p ' '
........................ |.' . . . . ., " . . .<:> . .
.................................................................... . - -, Q .
..................................... J.‘ , ""-------------------.------...‘.4.....
..r......r1.—rv—v .......
..............................................
...........................................
Echama ........................ 66 336 1 6 3
:.. . . f . . . . ."""'.":'..‘:'f‘t':‘.":' ' ' ' '
.. J . - .4 ...... A; .
.......................
.......................
.......................
.......................
'7.'. .f5ﬁ.fuvoﬁ.v. .ﬁ ..........
.......................
......................
.......................
.....
.En.‘ 6:: ..........
..... . . . .
.........
...................
.......................
......................
.......................
...................
..........
1 ...........
.......... 2 7
EEERBS dun .............. 5160841 188 e 94
.....................
.......................
.......................
.......................
ETW1 ............................. 597 337 191 3 77
................
....................
Qt! ......................... o a |

 

 

 

 

Chapter Iv

DISCUSSION

Research evidence is accumulating which suggests that
political schemata may interact with source, message and
channel to produce differential effects. By specifying
the nature of these processes, political communication
scholars hope to gain a keener insight into the nature of
political message effects and ultimately the political
process. In the present study an information processing
approach to an understanding of political message effects
was undertaken. The study was designed to examine the
mediating role of audience cognitive schemata on mass
media message characteristics. Since scholars have
decried the dearth of research energy devoted to message
comprehension, a variable that has been accorded a central
role in communication and persuasion processes, the
present research was directed at examining message
comprehension within a political information processing
framework. Specifically. the focus of the experiment
centered on how political cognitive schemata may mediate
the effects of message comprehensibility.

Results of the experiment are consistent with the

71

72
position that political audience schemata play an
important role in the processing of political messages.
Furthermore, it is apparent from the current findings that
mediating variables, like cognitive schemata, may interact
with message, source, and channel variables to produce
differential effects. Before discussing the findings of
this experiment, the limitations of the study’s methods

and findings will be considered.

Limitations

A number of limitations detract from the study’s
contribution value. First, the usual restrictions of a
laboratory setting may apply even more in a study of this
nature. Subjects were placed in an environment that may
have implicitly demanded and facilitated an analytical
approach. The control imposed by an experimental setting
has the potential to distort and magnify some effects
while minimizing others. Therefore, subjects may have
judged the messages differently than they would have in a
natural context. Furthermore, the sample is homogeneous,
college students, with above average cognitive
capabilities that probably do not mirror the general
electorate. This may limit the generalizability of the
research findings.

Moreover, seldom are voters required to judge

candidates and make vote decisions based on exposure to a

73
single political message. The political process normally
takes place over a longer period of time with people
having the ability to judge candidates based on numerous
issue stands, various political messages and debates, and
in the context of alternative candidates.

Another probable limitation inherent in the
experiment’s design makes some of the results even more
interesting. Even though there was a statistically
significant difference in subjects’ perception of the
higher and lower comprehensibility messages, the actual
difference on a 7-point scale was quite small (1.955 for
low vs. 2.721 for high). Despite this limitation, there
was still a discernable impact for comprehensibility on
numerous variables of interest.

Comprehensibility has been manipulated in various ways
by other researchers. For example, low comprehensibility
messages have been created by making ”dirty" xerox copies,
by distorting audio tapes, and by randomly mixing sentence
halves. It was felt that the manipulation in the present
study came closer to achieving ecological validity.
However, the comprehensibility effect might have been even
greater had the higher comprehensibility message been
perceived as more comprehensible in real terms. The fact
that there was some quantification in the higher
comprehensibility message may have been enough to lower

subjects’ comprehensibility ratings and thus the

74
manipulation’s effect.

Many of the hypotheses proposed in this research
received some degree of support. In the discussion that
follows, main effects for schema and comprehensibility on
each variable will be dealt with first and then further
analysis for the variables will be addressed. When
necessary, additional data analyses were conducted to

clarify and elaborate relationships.

Interpretation of Results

Consistency in Processing Orientation. Hypothesis 1A
and 18 should be considered more than manipulation checks.
Indeed, the results suggest that people’s processing
orientation are consistent even when faced with messages
ranging in levels of comprehensibility. For example,
regardless of level of comprehensibility issue processors
continued to produce more issue—oriented cognitive
responses than image processors, and image processors
produced more visually-oriented cognitive responses than
issue processors. This places greater emphasis on the
receiver as an active information processor, a notion
running counter to much traditional communication and
persuasion research which considers the receiver a passive'
message recipient. The results obtained here further
broaden the notion of audiences as active in their

approach to the mass media (Blumler, 1979; Garramone,

75
1984; 1985; 1986; Levy a Winhal, 1989). This does not
suggest, however, that manipulating message variables will
not affect cognitive response production. For example, in
the present study, the absolute number of cognitive
responses varied according to level of comprehensibility.
Thus, comprehensibility should still be regarded as an
important variable in the case of cognitive response
generation.

It was interesting to note that only issue processors
produced significantly more visually-oriented responses in
low versus higher comprehensibility conditions (Hypothesis
2). It was expected that issue processors would find
difficulty extracting issue information from low
comprehensibility messages and would rely more on visual
information and, subsequently, produce more visually-
oriented cognitive responses. The data support this
assumption. However, a similar argument was made for
image processors too, but the data did not support this
prediction.

While the data for image processors was initially
puzzling, the results may be attributable to a
methodological artifact. Since the higher
comprehensibility message was still low in
comprehensibility, image processors may have been
producing near maximum visually-oriented thoughts in the

higher comprehensibility condition (i.e., there is a

76

ceiling effect). A greater perceptual difference in
comprehensibility may be required to produce a
statistically significant difference in visual responses.

The fact that more issue-oriented thoughts were
produced in the low comprehensibility condition may not be
so intuitive (Hypothesis 18). However, this too may be
explained by virtue of the nature of the cognitive
response measures. When calculating total issue thoughts,
no differentiation was made between two thoughts about one
issue or two thoughts about two independent issues. For
example, the fact that a subject may have generated
several thoughts complaining about an issue not being
clearly articulated would be counted as several issue
thoughts. Therefore, a high issue score doesn’t
necessarily mean subjects were processing more issue
information but rather they could be dwelling on a small
number of issues.

Message Evaluation and Pleasantness Ratings.
Hypotheses 3A and 38 found a main effect for both schema
and comprehensibility on message evaluation and
pleasantness ratings. Issue processors rated messages
lower than image processors. One suggestion for this
phenomenon may be attributable to the fact that both
messages were low in comprehensibility. Since issue
processors are by definition more motivated to process the

audio track than image processors, the two messages, both

77

low in audio comprehensibility, would be expected to
engender lower ratings for them than for image processors.
The anticipated main effect for comprehensibility showed
that lower comprehensibility messages were evaluated lower
and were rated as less pleasant. However, this
relationship held only for issue processors. Data for
image processors were in the same direction although not
significant. Again, this may be due to the low
comprehensibility of the higher comprehensibility message.
Both messages may have been low enough in
comprehensibility to turn image processors’ attention away
from the audio track thus making the audio
comprehensibility manipulation less discernable for them.
Additional support for this contention may be found in the
data for hypothesis BB. Image processors did not perceive
the low comprehensibility message as incomprehensible as
did issue processors. A better manipulation of
comprehensibility would help answer this question.

Hypothesis 5A was partially confirmed. Issue
processors rated the lower comprehensibility messages as
less pleasant than image processors. However, they did
not evaluate the candidate lower than image processors in
the low comprehensibility condition, although means were
in the anticipated direction.

Candidate Evaluation and Vote. It was interesting

that even though subjects evaluated low comprehensibility

78
messages lower and as less pleasant, they did not appear
to transfer this negative affect to the candidate
(Hypothesis 4A and 5A) or the vote (Hypothesis 48 and 58).
However, since there was a main effect for
comprehensibility on candidate evaluation, the fact that
neither the issue nor image subgroup achieved significance
for comprehensibility on candidate evaluation is suspect.
At least two important factors may have contributed to the
failure to achieve significance. First is the sample
size. Once the overall sample was broken into subgroups,
the effective sample size was reduced significantly. Due
to the smaller sample size, the power of the statistical
test was lowered. Secondly, the mean differences would
have had to be quite pronounced for the effects to reach
significance with the diminished power. This does not
appear to be the case, although means were in the intended
direction.

Two factors may help explain why the effects of
comprehensibility were attenuated for candidate evaluation
and vote likelihood: (1) people realize that candidates
often don’t write their own political messages. Media
stories allude to the use of professional speech writers
and advertising and public relations firms for speech and
message creation. Therefore, the candidate may not have
been held fully culpable for the comprehensibility

failure. (2) Also, people may intimate that their own

79
lack of knowledge in a subject area may make the message
appear less comprehensible than it actually is. Thus,
they may attribute the lack of comprehensibility more to
themselves than to the candidate.

Qualitative analysis of cognitive responses lends
support for these suppositions. For example, many
thoughts were actually questions which asked who the
intended audience for the messages was, suggesting the
messages were better suited for an educated audience.
Furthermore, responses were often directed at a message
"creator," not the candidate. Many of the subjects
suggested firing the person who wrote the message for the
candidate.

Contemporary politics may shed even more insight on
the issue. Recently, Ross Perot was chided by critics for
his remarks to a black audience at an NAACP meeting. He
made allusions to "you people" and "your people" when
referring to blacks. Many of these same critics,
however, said Perot’s problem was not a lack of
sensitivity but rather a lack of political sophistication.
They suggest he should have relied on the services of a
professional speech writer, who would have avoided such
inflammatory faux pas.

Processing Effort. There are two potential reasons
for image and issue processors reporting similar amounts

of effort processing the low comprehensibility message

80

(Hypothesis 6A). First, there may be a ceiling effect for
processing effort. For example, both groups’ effort
ratings suggest a large amount of effort was expended.
There may be a point at which people, regardless of
processing orientation, will not (or cannot) devote any
more energy to processing the message. A second
possibility is that since issue processors find low
comprehensibility messages less pleasant, and evaluate
them lower and as less comprehensible than image
processors, they may have only been willing to allocate so
much energy to the current processing task (i.e., the
ability was there but the motivation was not). This would
suggest that an audience may only go so far in its attempt
to decipher a message. Message creators should be aware
of this limitation and design messages to stay within the
motivational and ability boundaries of their audience.

Additional Analyses. The findings are consistent with
the notion of schemata and the theoretical arguments
regarding comprehension. That is, issue processors
recalled and recognized more information than image
processors, and less information was recalled and
recognized from low comprehensibilty messages. Although
these findings appear intuitively simple, they still have
interesting implications. For example, if a politician
feels it is imperative for constituents to remember

certain information, it might be possible to "prime"

81
voters to use their issue schemata. This would increase
the likelihood that the information would be retained.
Moreover, if a politician must relate information that
could be injurious (e.g., admission of guilt or misdeed),
it might be possible to obfuscate the information with a
layer of complexity, thus reducing its retrievability in a
later context (e.g., the voting booth).

Summary. Taken as a whole, the results of this study
provide additional support for the belief that audience
cognitive schemata may affect the processing and
acquisition of information from political messages. The
research reported here furthers our understanding of how
schemata may moderate the effects of message
characteristics such as comprehensibility. It would
appear from the results that audience schemata are
important characteristics to be considered when
formulating messages for various media audiences. Much
more work needs to be done, however, before this
prescription can be readily applied on a work-a-day basis,
but the findings do suggest that the approach is

conceptually and methodologically viable.

Future Research Priorities

A major task for future research is to articulate a

theoretical framework that explains how differences in

82
people’s information processing orientations affect
political message and candidate evaluation, the content
and organization of political information in memory, and
the retrieval and utilization of the information. Broad
questions will surely emerge as this conceptual foundation
develops. How do processing orientations develOp? Can
processing orientations be manipulated (i.e., can the
audience be tailored or "primed“ for the message as well
as the message tailored for the audience)? Moreover, as
the framework evolves, better operational procedures must
be developed and refined for identifying and measuring
audience cognitive schemata. Ultimately, this will better
enable message creators to tailor their fare for the
motivations of their audience.

Certainly more research energy should be devoted to
message comprehensibility in the political domain and in
other domains. Work should be directed at identifying the
dimensions of comprehensibility. For example, levels of
quantification as well as verbal difficulty can influence
message comprehensibility. It stands to reason that these
various dimensions may have substantially different
effects on some variables. Moreover, from an experimental
design perspective, researchers need to consider whether
they are truly manipulating comprehensibility (i.e., low
comprehensibility version vs. higher comprehensibility

version of the same message) in research designs or

83
whether they are creating two distinct messages.
Differences in reactions to two different messages should
not be construed as the result of a comprehensibility
manipulation. Furthermore, more work should be directed
at achieving higher levels of ecological validity. For
example, seldom will communicators have to worry about
their sentence halves being randomized.

The fact that subjects did not rate the candidate or
vote lower in low comprehensibility conditions has
interesting implications. Further research should address
the inferential processes in the political process. For
example, on what occasions is the candidate held
accountable for what is said? In the present case it may
be that subjects blamed themselves (i.e., lack of
knowledge) or the message for comprehension failure, but
it is doubtful that audiences are always so forgiving.
However, there may even be contexts in which low message
comprehensibility could be positive. If voters equate
message complexity with the candidate’s intellectual
acumen, it could prove to be an asset rather than a
liability. To this end, research that identifies the
influence of prior knowledge and political involvement on
the comprehension process would be a benefit to various
disciplines that study message effects.

Research should also address how other variables

affect the political communication process. For example,

84
how would the inclusion of a source attractiveness or
expertise manipulation affect the role of processing
orientation and the effect of message comprehensibility.
Would image processors be influenced more than issue
processors by attractive sources in low comprehensibility
conditions? Would issue processors be more effected by
experts under low comprehensibility conditions than higher
comprehensibility conditions? The answers to these
questions would extend knowledge in the political domain
and would extend the utility of current persuasion
theories (see, for example, Chaiken, Liberman & Eagly,
1989; Petty and Cacioppo, 1986). Clearly, only through
programmatic research will answers to these questions be

addressed in a timely and systematic fashion.

 

REFERENCES

Abelson, R. P., Kinder, D. R., Peters, M. D., & Fiske, S.
T. (1982). Affective and semantic components in
political person perception. r P r i

and_§22121_E§1§b21991. 52 (4). 619-630.

Anderson, D. R., Lorch, E. P., Smith, R., Bradford, R., &
Levin, S. R. (1981). Effects of peer presence on
preschool children’s television-viewing behavior.

ngglogmgnﬁgl Eﬁxgnglggx, 11,(4), 446-453.

Anderson, R. C. (1977). The notion of schema and the
educational enterprise. In R. C. Anderson, R. J.

Shapiro, a W- Montasue (Eds.), §9n9211n9_and_tba
nngnigition of knowlgngn. Hillsdale, NJ: Erlbaum.

Anderson, P. A., Garrison, J. P., a Anderson, J. F.
(1979). Implications of a neurophysiological approach
for the study of a nonverbal communication. ﬂnnnn

ngmunication Regngngn, Q, 74-89.

Atkin, C. K., Bowen, L., Nayman, O. 8., & Sheinkopf, K. G.
(1973). Quality versus quantity in televised political
ads. Enblic Opinion Qngntgnly, 31, 209-224.

Bartlett, F. A. (1932). A u i x ri l o i l
gsynnn1_gy. NY: Cambridge University Press.

Bower, G. H. (1981). Mood and memory. Angricnn
Egyghologist, §§, 129-148.

Burke, M. C., & Edell, J. A. (1989, February). The impact
of feelings on ad-based affect and cognition. Jgnrnnl

2£_Mackatins_8a§aancb. 25. 69-83.
Cantor, N., & Kihlstrom, J. F. (1981). Personality,
i ' l ' r i . Hillsdale, NJ:
Erlbaum.

Chaiken, S., Liberman, A., & Eagly, A. H. (1989).
Heuristic and systematic information processing within
and beyond the persuasion context. In J. S. Uleman &
J. A. Bargh (Eds.), n - i i of

ar n ' n i n r . NY: Guilford Press.

Chaiken, 8., & Stangor, C. (1987). Attitudes and attitude
change. Annngl ngjgn nf Esygnnlggy, 38, 575-630.

85

86

Chase, W. G., & Simon, H. A. (1973). The mind’s eye in

chess. In W. G. Chase (Ed.), Miguel infonmetien
nrocessing. NY: Academic Press.

Cohen, L. J. (1979). On the psychology of prediction:
Whose is the fallacy? Qognitien, 1, 385-407.

Conover, P. J., & Feldman, S. (1986). The role of
inference in the perception of political candidates.
In R. R. Lau & D. 0. Sears (Eds.), Eelitieel_cegnjtjen;

The 19th enndel Carnegie Symposium on cegnitjen.
Hillsdale, NJ: Erlbaum.

Converse, P. E. (1975). Public opinion and voter
behavior. In F. I. Greenstein & N. W. Polsby (Eds.),

Handbook of dolitieel eeience, (Vol. 4). Reading, MA:
Addison-Wesley.

Craik, F. I., & Lockhart, R. S. (1972). Levels of
processing: A framework for memory research. dednnal

of Verbal Leerning end Vernel Beheyjer, 11, 671-684.

Crocker, J., Fiske, S. T., a Taylor, S. E. (1984).
Schematic bases of attitude change. In J. R. Eiser

(Ed.), Attitudine] jndgment. NY: Springer-Verlag.

Eagly, A. H. (1974). Comprehensibility of persuasive
arguments as a determinant of opinion change. dedrnel

of Pereonality end Seeial Peycnelogy, 22 (6), 758-773.

Eagly, A. H., & Chaiken, S. (1984). Cognitive theories of
persuasion. In L. Berkowitz (Ed.), Advenees in

exgecimentel sociel nsycholegy, (Vol. 11). NY: Academic
Press.

Eagly, A. H. & Warren, R. (1975). Intelligence,
comprehension, and opinion change. dounnel ef
P r n lit , AA (2), 226-242.

Festinger, L. (1957). A theory of cognitiye_dieeenenee.
Stanford, CA: Stanford University Press.

Fishbein, M., & Ajzen, I. (1975). i i e
n = '01 an- othe - ‘ '

, I ,
reeeereh. Reading, MA: Addison-Wesley.

‘ O. .0 O ‘2. .A‘O

Fishbein, M., & Ajzen, I. (1981). Acceptance, yielding,
and impact: Cognitive processes in persuasion. In R.
E. Petty, T. M. Ostrom, a T. C. Brock (Eds.), Qegnitiye
resnonees jn nereueeje . Hillsdale, NJ: Erlbaum.

 

87

Fiske, S. T. (1981). Social conflict and affect.' In J.
Harvey (Ed.), Qegnition.teociel behevior end the
envirenment. Hillsdale, NJ: Erlbaum.

Fiske, S. T. (1982). Schema-triggered affect:
Applications to social perception. In M. 8. Clarke &

S. T. Fiske (Eds. ), Af_eet_end_eegnition: Thgfllth
n u Carn m s m o n i . Hillsdale, NJ:

Erlbaum.

 

Fiske, S. T., & Kinder, D. R. (1981). Involvement,
expertise, and schema use: Evidence from political
cognition. In N. Cantor & J. F. Kihlstrom (Eds.),

rsonali c ' i i l r .
Hillsdale, NJ: Erlbaum.

Fiske, S. T., Kinder, D. R. & Larter, M. (1983). The
novice and the expert: Knowledge-based strategies in

political cognition JouLnal_o£_§znanimantal_§ocial
Eaxcbolosx 19. 381- 400.

Fiske, S. T., a Taylor, S. E. (1984). i l ' n.
Reading, MA: Addison-Wesley.

Fleecn. R. (1951). ﬂow_to_te§t_raadabilitx. New York:
Harper.

Gardner, M. (1985, December). Mood states and consumer
behavior: A critical review. r l r

Bﬁﬁﬁéﬁﬁh: 124 261-300.

Garramone, G. M. (1983). Issue versus image orientation
and effects of political advertising. genndnieetien
BQSQQLQD. 19. 59-76-

Garramone, G. M. (1984). Audience motivation effects:

More evidence. Communication_8a_aancb .1. 79- 96.

Garramone, G. M. (1985). Motivation and political
information processing: Extending the gratifications
approach.. In S. Kraus & R. Perloff (Eds. ), neee_medie

! III' 1 I! I! ! i f r l' _ .
ADDLQQQD- Beverly Hills, CA: Sage.

Garramone, G. M. (1986). Candidate image formation: The
role of information processing. In L. L. Kaid, D.
Nimmo, & K. Sanders (Eds.), Nen eernectiyee en
gelitieel_edyentieing. Carbondale, IL: Southern

Illinois University Press.

68

Garramone. G. M., Steele, M. E., Hogan, P. & Rifon, N.
(1987, August). Gratifications sought and the
processing of print messages. Paper presented to the
Association for Education in Journalism and Mass
Communication, San Antonio, TX.

Garramone, G. M., Steele, M. E., & Pinkelton, B. (1991).
The role of cognitive schemata in determining candidate
characteristic effects. In Biocca, F. (Ed.),

Ia1ax1a12n_and_nolitica1_adxantlaing. (Vol. 1.).
Hillsdale, NJ: Erlbaum.

Geiger, S. F., & Reeves, 8. (1991). The effects of visual
structure and content emphasis on the evaluation and
memory for political candidates. In F. Biocca (Ed.),

vi 1 n ' i l , (Vol 1.).
Hillsdale, NJ: Erlbaum.

Geschwind, N. (1979). Specializations of the human brain.
Scientific Americen, 24] (3), 180-199.

Hamill, R., a Lodge, M. (1986). Cognitive consequences of
political sophistication. In R. R. Lau & D. 0. Sears
(Eds.), li i l ' ' - T n '
Symnoeium en eognitjen. Hil sdale, NJ: Erlbaum.

Heider, F. (1946). Attitudes and cognitive organization.
dodcnel of Eeycholegy, 21, 112-127.

Higgins, E. T. (1981). The ”Communication Game”:
Implications for social cognition and persuasion. In
E.T. Higgins, C.P. Herman, & M.P. Zanna (Eds.), Seeiel

cognjtien; Ine Qnterie Symneejdn, (Vol. 1). Hillsdale,
NJ: Erlbaum.

Higgins, E. T., Kuiper, N. A., & Olson, J. (1981). Social
cognition: A need to get personal. In E.T. Higgins,
C.P. Herman, & M.P. Zanna (Eds.), §Q£181_22981L1281_Ih8
Qntenie_§ynne§1dm, (Vol. 1). Hillsdale, NJ: Erlbaum.

Hofstetter, C. R., & Zurkin, C. (1979). TV network news
and advertising in the Nixon and McGovern campaign.

Jedrneliem Qderterly, 5e, 106-115, 152.

Holbrook, M. 8. & O’Shaugnessy, J. (1984). The role of
emotion in advertising. Eaxcnolosx_1_uankatins. 1 (2).
45-64.

Hovland, C. 1., Janis, I. L., & Kelley, H. H. (1953).
Cemmdnjeetien end neredesjen. New Haven, CT: Yale

University Press.

89

Jacoby, J., & Hoyer, D. H. (1982). Viewer
miscomprehension of televised communication: Selected

findings. dedrnel ef nerketing, 16 (Fall), 12-26.

Jacoby, J., & Hoyer, D. H. (1989). The
comprehension/Miscomprehension of print communication:

Selected findings. 12uLna1_21_92n§umat_ﬂa§aazcb. 15
(4), 434-443.

Jeffery, K. M., a Mischel, W. (1979). Effects of purpose
on the organization and recall of information in person

perception. dednne] ef Pereeneljty, 51, 397-418.

Johnston, D. D. (1986, May). Image and issue political
information processing. Paper presented to the
International Communication Association, Chicago, IL.

Jones, E. E. & Davis, K. E. (1965). From acts to
dispositions: The attribution process in person
perception. In L. Berkowitz (Ed.), Adxeneee_1n

exnerimentel seeie] geyeholegy, (Vol. 2) NY: Acaemic
Press.

Joslyn, R. A. (1980). The content of political ad spots.
dourneleim Quertecly, 7, 92-98.

Kahneman, D., Slovic, P., & Tversky, A. (Eds.). (1982).
u m n n r r in r ic i e . NY:
Cambridge University Press.

Kaid, L. L., & Sanders, K. R. (1978). Political
television commercials: An experimental study of type

and length. Communieetion Beeeerch, d (1), 57-70.

Katz, E., & Feldman, J. (1962). The debates in the light
of research: A survey of surveys. In S. Kraus (Ed.),

Ine_gneet_denetee. Bloomington, IL: Indiana University
Press.

Keeter, S. (1987). The illusion of intimacy: Television
and the role of candidate personal qualities in voter

choice. Bu911c_921n19n_guastanlx. 51. 344-358.

Kinder, D. R., Peters, M. D., Abelson, R. P., a Fiske, S.
T. (1980). Presidential prototypes. EEﬂjjﬂjﬁﬂ_

Behevior, 2 (4), 315- 337.

Kraus, S., & Perloff, R. M. (1985). neee_ned1e_end

221111221_tn229ht11An information_osoca§a1n9
er c iv . Beverly Hills, CA: Sage.

 

90

Lasswell, H. D. (1948). The structure and function of
communication in society. In L. Bryson (Ed.), Ine
eommunieetion of ideee. NY: Harper.

Lavidge, R. G., & Steiner, G. A. (1961). A model for
predictive measurements of advertising effectiveness.
Journel of Merketing, 5, 59-62.

Lau, R. R. (1986). Political schemata, candidate
evaluations, and voting behavior. In R. R. Lau & D. 0.
Sears, (Eds.), Politicel cognition; Ine_1§tn_dnndgl
Carne i S o ium on i . Hillsdale, NJ:
Erlbaum.

Lau, R. R., & Erber, R. (1985). Political sophistication:
An information processing perspective. In S. Kraus &

R. Perloff (Eds. ). Maaa_med1a_and_2911112a1_tnou2b11_1n

intendstionzonoce§§1n2_apntoecb- Beverly Hills. CA:
Sage.

Lau, R. R. & Sears, D. O. (Eds.). (1986). Eelitieel
c n ' - nnual ' m '
c hi i n. Hillsdale, NJ: Erlbaum.

Lorch, (1979). The relationship between visual attention
and children’s comprehension of television. Qnild

Qexalonmant. 59. 722-727.

Lutz, R. J. (1985). Affective and cognitive antecedents
toward the ad: A conceptual framework. In L. F. Alwitt

& A. A. Mitchell, 8 l i r
adxett1§1n2_eiiect§: Theory. reaaatcn_and_aoolication-

 

Hillsdale, NJ: Erlbaum.

MacKenzie, S. 8., & Lutz, R. J. (1989, April). An
empirical examination of the structural antecedents of
attitude toward the ad in an advertising pretesting

context. dourne] ef Merketjng, 53, 48-65.

Madden, T. J., Allen, C. T., a Twible, J. L. (1988,
August). Attitude toward the ad: An assessment of
diverse measurement indices under different processing

Meets - J_HLn£__Q__MQLKEL_QE_B§§§§LED 2§s 242‘ 252-

Markus, H., a Smith, J. (1981). The influence of self-
schemata on the perception of others. In N. Cantor &
J. F. Kihlstrom (Eds.), i c

eeeiel_1nteneetien. Hillsdale, NJ: Erlbaum.

McClure, R. D., 5 Patterson, T. E. (1974). Television
news and political advertising: The impact of exposure

0” VOter bellafe- QQME!D1§§LIQE.B§§2§LQD1 l (1): 3‘31-

91

McGuire, W. J. (1968). Personality and suseptibility to
social influence. In E. F. Borgatta & W. W. Lambert
(Eds.). ﬂandpook_o1_naLa2na1111.1neo£x_and_te§aasch-
Chicago, IL: Rand McNally.

McGuire, W. J. (1981). The probabilogical model of
cognitive structure and attitude change. In R. E.
Petty, T. M. Ostrom, & T. C. Brock (Eds.), Qegnitiye

L9§228§§§.iﬂ.2§£§28§i§ﬂ- Hilledele. NJ: Erlbaum.
Mendelsohn, H., a 0’ Keefe, G. L. (1976). Ine_deenle_eneee
r i en 8 . NY:
Praeger.

Mitchell, A. A., & Olson, J. C. (1981, August). Are
product attribute beliefs the only mediator of
advertising effects on brand attitude? dednne1_ef

Marketing Reeeereh, 18, 318-332.

Nisbett, R., 3 Ross, L. (1980). ﬁnnen_1nﬁenenee1
Stretegjee and.§hQLIQQminaﬁ.21_§921§l_iﬂgﬂmgﬂ5'
Englewood, Cliffs, NJ: Prentice-Hall.

Norman, D. A. (1980). Twelve issues for cognitive

science. In D. A. Norman (Ed. ). E§_§D§£&ix2§_98
ni iv k l .

Hillsdale, NJ: Erlbaum.

O’Keefe, T. a Sheinkopf, K. (1974). The voter decides:

Image or camoaisn issue? Jousnal_91_ﬁtoadcaat1ns. 18.
403-411.

Osgood, C. E., Suci, G. J. a Tannenbaum, P. H. (1957).

Ibe_Mea§uLament_21_Maan1ns- Urbano. IL: University of
Illinois Press.

Paivio, A. (1975). Perceptual comparison’s through the

mind’s eye. Memotx_1_992n111on. 3. 635-647.

Paivio, A. (1978). A dual coding approach to perception
and cognition. In H. L. Pick., a E. Saltzman (Eds.),
l f r ' ' .
Hillsdale, NJ: Erlbaum.

Paivio, A., & Begg, I. (1974). Pictures and words in
visual search. Memenx_1_gegn151en, 2, 515-521.

Petty, R. E., & Cacioppo, J. T. (1981). Att1tndee_end
r i - l r r .
Dubuque, IA: Wm. C. Brown.

92

Petty, R. E., & Cacioppo, J. T. (1986). The elaboration
likelihood model of persuasion. In L. Berkowitz (Ed.),
A v nc x r l c l ch , (Vol. 19).
NY: Academic Press.

Pezdek, K., & Hartman, E. (1983). Children’s television
viewing: Attention and comprehension of auditory versus

visual information. §n11d_Qeyelennent, 51, 1015-1023.

Piaget, J. (1981). n li n 1 ° ir
rel ti hi hil . Palo Alto, CA: Annual
Reviews.

Ratneshwar, S., a Chaiken, S. (1991). Comprehension’s
role in persuasion: The case of its moderating effect
on the persuasive impact of source cues. dedrnel ef

Censumen Reeeeren, 18, 52-62.

Rosenburg, S. W., Bohan, L., McCafferty, P., & Harris, K.
(1986). The image and the vote: The effect of
candidate presentation on voter preference. £29119

Oninien Quertenly, 29, 108-127.

Rosenburg, S. W., & McCafferty, P. (1987). The image and
the vote: Manipulating voters’ preferences. Peelie

Opinion Qderterly, §1. 31-47.

Rumelhart, D. E., & Ortony, A. (1977). The representation
of knowledge in memory. In R. C. Anderson, R. J.

Shapiro, & W. E. Montague (Eds.), Scheeljng end tne

ecgdisition ef knenledge. Hillsdale, NJ: Erlbaum.
Schank, R. C. & Abelson, R. (1977). Serinte, elene,

goals, and dnderetending. Hillsdale, NJ: Erlbaum.

Schmidt, 0. F. & Sherman, R. C. (1984). Memory for
persuasive messages: A test of a schema-copy-plus-tag
model. rn r o i n ,
11 (1), 17-25.

Shimp, T. A. (1981). Attitude toward the ad as a mediator

of consumer brand choice. dednne1_eﬁ_ndyent1e1ng, 1n
(2), 9- 15.

Shyles, L. (1986). The televised political spot: Its
structure, content, and role in the political system.
In L. L. Kaid, D. Nimmo, & K. Sanders (Eds.), Hen

nernectives en eeljtjeel edvertjejng. Carbondale, IL:

Southern Illinois University Press.

93

Snyder, M., & DeBono, K. G. (1985). Appeals to image and
claims about quality: Understanding the psychology of
advertising. rn f P r l

P§¥Cbglogx, 1e (3), 586-597.

Steele, M. E., Garramone, G. M., & Hogan, P. (1988, May).
Attractiveness and persuasion: The roles of processing
motivation and involvement. Paper presented to the
International Communication Association, New Orleans.

Stokes, D. E. (1966). Some dynamic elements of contests

for the presidency. American Political_§cience_ﬂaxiew.
en, 19-23.

Taylor, S. E. (1981). The interface of cognitive and
social psychology. In J. Harvey (Ed. ), gegn1t1en1

secial behavior and tne envinenm ment. Hillsdale, NJ:
Erlbaum.

Taylor, S. E., & Crocker, J. (1981). Schematic bases of
social information processing. In E. T. Higgins, C. P.

Herman & M- P- Zanna (Eds.). Social_cogn1119n1_lba
Onterio Symgosium, (Vol. 1). Hillsdale, NJ: Erlbaum.

Tesser, A. (1978). Self-generated attitude change. In L.

Berkowitz (Ed.). Adxgnca§_1n_e_2a£1mental_§ocial
gsyehology, (Vol. 11). NY: Academic Press.

Tesser, A., & Danheiser, P. (1978). Anticipated
relationship, salience of partner and attitude change.

Eersoneljty end Seeiel Peyenelegy Bulletin, 1, 35-38.

Valenti, A. C., & Tesser, A. (1981). On the mechanism of

thought-induced attitude change. §9§i§l_ﬂ§h§¥i§£_nnd
P r o i . a (1). 17-22.

Witt, W. (1976). Effects of quantification in science
writing. Joutnal_21_92mmunicatlon. 29 (Winter). 67-69-

Yalch, R. F., & Elmore-Yalch, R. (1984). The effects of
numbers on the route to persuasion. dednne1_ef

Consumet_ﬂe§eatcb. 11 (June). 522-527-

Zajonc, R. B. (1968). Cognitive theories in social
psychology. In G. Lindzey & E. Aronson (Eds.),

uandoook_oﬁ_aocigl_naxcnologx. (Vol.1). Reading. MA:
Addison-Wesley.

Zajonc, R. B. (1980). Feeling and thinking: Preferences

need no inferences. Ame:1can.£axcb21921§t..35. 151-
175.

APPENDIX

94
Please do your best to answer the following questions:
1. Bob Moran was
a. In favor of Haskee-Sharm Nuclear
b. In favor of Sononma Hydro
c. In favor of Sonoma Nuclear
d. In favor of Haskee-Sharm Hydro

2. Bob Moran presented how many reasons to support his

stand?
a. 3
b. 4
c. 5
d. 6

3. Bob Moran is running for

Congress
Senate
Mayor
. Governor

0.009

4. The construction company for the Haskee project is

a. Corona
b. Alcor
c. Coran
d. Calcon

5. State revenues increased by what percent over 1986-

1987?

a. 3.2%
b. 406*
Ca 601‘
d. 608*

6. Raw material costs at Haskee increased at

a. rate of inflation
b. at

0. 5x

d. $27.3 million

7. To support Bob Moran’s position you’d vote

a. Yes on Proposition A
b. Yes on Proposition 8
c. No on Proposition A
d. No on Proposition 8

95

Please do not turn back to the last page of questions.

8. State revenues for 1987-1988 were

a.
b.
c.
d.

$109.9 million
$23.7 million
$104.6 million
$27.3 million

9. List the two reasons state revenues increased by
record rates.

1.

2.

 

 

10. If the Sonoma project is adopted, money set aside to
fund the anticipated wage increases for Calcon
employees could be used to

 

11. List two previously approved areas revenues will not
have to cover

1.
2.