The relevance-accessibility model of advertising effectiveness

MISSING IMAGE

Material Information

Title:
The relevance-accessibility model of advertising effectiveness
Physical Description:
vii, 338 leaves : ; 28 cm.
Language:
English
Creator:
Baker, William Edward, 1955-
Publication Date:

Subjects

Subjects / Keywords:
Advertising -- Mathematical models   ( lcsh )
Advertising -- Research   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1991.
Bibliography:
Includes bibliographical references (leaves 315-337).
Statement of Responsibility:
by William Edward Baker.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001701182
notis - AJB3368
oclc - 25350808
System ID:
AA00003713:00001

Full Text








THE RELEVANCE-ACCESSIBILITY MODEL OF
ADVERTISING EFFECTIVENESS










By
WILLIAM EDWARD BAKER


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1991

























Copyright 1991

by
William Edward Baker


















ACKNOWLEDGEMENTS


I would like to thank the members of my supervisory committee for working
with me on this dissertation. Most of all I would like to thank Rich Lutz for
being my mentor and friend through this long, but very worthwhile ordeal that
began almost five years ago. Rich has spent countless hours working on the
RAM's conceptual foundation, formal theoretical structure and experimental
design. He has spent an equal number of hours editing manuscripts. This project
would not have been possible without his wisdom, insights and guidance.











TABLE OF CONTENTS


page

ACKNOWLEDGEMENTS.......................................... ................... iii

A BSTRA CT................................................................................................... vi

CHAPTERS

1 THE RELEVANCE-ACCESSIBILITY MODEL OF
ADVERTISING EFFECTIVENESS ...................................... ...... 1

Introduction.................................................................................. 1
Axiom 1: Advertising Information Must be Accessible to be
Effective........................................................................................ 5
Axiom 2: Advertising Must be Relevant to be Effective............. 8
Proposition 1: Advertising Effectiveness Occurs at the Point of
Brand C hoice....................................................................... .... 10
Proposition 2: Consumers Use Three Levels of Information to
Make Brand Choices.................................................................. 15
Proposition 3: Involvement Determines Consumers' Preferred
Level of Information................................................................... 21
Proposition 4: Involvement Determines the Level of Information
Learned......................................................... ................................ 31
Advertising Effectiveness........................ .................. ........ 40
Sum m ary......... .................................... .................................. .... 46

2 LITERATURE REVIEW................................................................ 49

Introduction........................................................ .. ......... 49
Attitude-Behavior Consistency..................................................... 50
The Impact of Consumer Elaboration and Involvement on
Information Accessibility.......................................................... 68
Affect and Automaticity................................................................ 95
Intervening Theoretical Constructs....... ..................... 120

3 THE FORMAL THEORETICAL STRUCTURE....................... 139

A xiom s............................................................... ........ ....... 139
Principles........................................................................... ................ 147
Experimental Hypotheses.............................................. .................. 158






4 M ETHO D..........................................................................................


Experimental Rationale..................................................... ................ 162
Experimental Design............................................................................. 163
Experimental Procedure.................................................................. 167
Independent Variables.............................................. 169
Primary Dependent Variable......................................................... 178
Experimental Pretests................................. ................................... 179

5 EXPERIMENTAL RESULTS AND DISCUSSION................. 206

Manipulation Checks..................................................................... 206
Experimental Results....................................... 211
Demand Effects....................................... ......................................... 223
Summary....................................................................................... 225

6 GENERAL DISCUSSION............................................................ 227

Theoretical Contribution................................ ................. ........ 227
Empirical Contribution.................................................................. 230
Applied Contribution...........................................................232
Future Research........................... ........................... 235

APPENDICES

A MESSAGE APPEAL COPY.................................... 240

B EXPERIMENTAL QUESTIONNAIRE........................................ 243

C PRETEST INSTRUCTIONS ............................................................ 272

D SUPPLEMENTAL ANALYSES.................................................... 309

REFEREN CES........................................... ..................... .......................... 315

BIOGRAPHICAL SKETCH................................... 338


162










Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

THE RELEVANCE-ACCESSIBILITY MODEL OF
ADVERTISING EFFECTIVENESS

By

William Edward Baker

August, 1991

Chairman: Richard J. Lutz
Major Department: Marketing

The Relevance-Accessibility Model (RAM) is a framework to study the issue
of advertising effectiveness. Fundamentally, it proposes that advertising effects
must be both accessible and relevant at the time of brand choice to be effective.
Effect accessibility is a function of consumer behavior at the advertising exposure
occasion (AEO). Effect relevance is a function of consumer behavior at the brand
response occasion (BRO).
The RAM proposes that specific message information recalled at the brand
response occasion, not brand attitudes formed at the advertising exposure, most
typically mediates advertising effectiveness. The model identifies three levels of
information that can mediate brand choice, affect, quality cues and relative
performance information, and links them to three macro brand choice processes:
indifference, satisficing and optimizing, respectively. At the AEO, advertising





message involvement (AMI) determines the level of information that consumers
are most likely to efficiently encode in memory. At the BRO, brand response
involvement (BRI) determines the level of information that consumers are most
likely to seek to discriminate brand alternatives. The Principle of Optimal
Advertising Contribution states that advertising effectiveness is most likely to
be maximized when message content, AMI and BRI correspond. When they do
not correspond, breakdowns in advertising effect accessibility, relevance or both
occur.
An experiment tested the model's fundamental assertions. A 3x3 between
subjects design manipulated AMI and BRI. Three test advertisements were
embedded in a radio show; each communicated one of the three levels of
information. The primary dependent variable was brand choice (i.e., message
appeal choice). Results confirmed the general thesis that advertising
effectiveness is determined at both the AEO and the BRO; there were strong,
significant effects of both AMI and BRI on message appeal choice. The Principle
of Optimal Advertising Contribution was also supported, each message appeal
was dominant in its respective involvement correspondence cell.











CHAPTER 1
THE RELEVANCE-ACCESSIBILITY MODEL
OF ADVERTISING EFFECTIVENESS



Introduction


A beer advertising executive kicks back in his chair. The new campaign is a
winner. Sales for their premium beer are up. He knows why the campaign
succeeded. It communicated the beer's superior quality. The creative director for

the company's agency also knows why it worked. Men relate to the outdoorsy,
rugged lifestyle portrayed in the campaign. They want to drink the beer because
it makes a statement about "who they are." The marketing research director also

knows why it worked. It increased top of mind awareness of the brand relative to

its competitors. "In an undifferentiated product category," he says, "people choose

the first beer that comes to mind." The academic consultant to the company also

knows why it worked. It classically conditioned strong positive feelings to the

brand. People are just naturally "drawn" to it when they see it in on the shelf.

Any one of the reasons are viable. Perhaps each one of them influenced some

people. But which reason influenced the most people in the brand's target

market?

How can advertisers be confident that they have isolated the key message
factor that made their campaign successful or unsuccessful? If they guess wrong,
their next campaign could flop. As exemplified above, a typical beer

advertisement may succeed for several reasons. Most beer advertisements tout




2

the quality of the beer, flaunt active engaging people having the best time of their
lives drinking it, repeat the beer's brand name and display the bottle several times
and feature popular music and striking visual effects. Obviously, it's important
for advertisers to know the real rather than just a good reason why their
advertisement worked or didn't work with their consumer target market. Then,

ads can be designed to emphasize the persuasive elements of the advertising and
omit the superfluous ones.
For decades, advertising researchers have chronicled the many ways

advertising can influence consumer preferences for products and services. Most of

these means of persuasion make intuitive sense and almost certainly would not
surprise advertising professionals or consumers. For example, we have learned
that advertising sometimes persuades simply by associating "good feelings" to
brands by being funny, by having beautiful imagery, by weaving sensual moods,
etc. (Silk & Vavra, 1974; Mitchell & Olson, 1981). We've also learned that

advertising sometimes persuades by making people "think" about product benefits;
for example, by highlighting a brand's unique benefits or by convincing consumers

that one brand has a relative advantage over alternative brands on one or more

delivered benefits (Lavidge & Steiner, 1961, Lutz & Swasy, 1977). And, we have

learned that advertising can work by linking brands with quality "cues" such as
an expert spokesperson, reputable manufacturer, or peer popularity (Chaiken,

1980, Petty & Cacioppo, 1981). Eventually, the sum of this research will produce

a complete "menu" of ways in which advertising messages can persuade
consumers.

What is less obvious is when each of these ways advertising can work, will
work. This latter issue is of primary importance to advertising practitioners.
Cataloguing all the ways that advertising can work is only half the battle.

Strategists also need to know when each of a series of potential strategies is likely







to be most effective. Ultimately, a list of rules is needed to select the optimal
message strategy out of the entire menu of available strategies for any given
situation. For example, the rules must determine when the goal of advertising
should be to make people feel good, or when its goal need be to inform people why a
brand is better than another brand by providing specific, comparative information.
The Relevance-Accessibility Model of Advertising Effectiveness (RAM) is a
theory that confronts the issue of maximizing advertising message effectiveness.
The RAM presumes that the primary purpose of advertising is to present
information that will give a brand a "relative advantage" over competing brand
alternatives at the time of brand choice. This relative advantage serves to
motivate purchase behavior (See Baker & Lutz, 1987). The RAM establishes
guidelines for picking the most effective advertising message strategy from a
menu of strategic options. More specifically, since the ultimate measure of
advertising effectiveness is brand choice, the RAM is a model concerned with
predicting when a given message strategy is most and least likely to influence
brand choice.
The purpose of this chapter is to present two major assumptions or axioms
upon which the Relevance-Accessibility Model is built and four more specific
propositions that facilitate the transformation of these assumptions into a theory
of advertising effectiveness. The specific "rules" or principles that the
Relevance-Accessibility Model offers to pick or predict the best message strategy
in a given advertising scenario come from these propositions. This reader should
not consider this framework to be the formal theoretical structure of the
Relevance-Accessibility Model; this will be presented in Chapter 3.
The first of these axioms states that elements of the advertising message
must be accessible at the time of brand choice to be effective. The key implication
is that advertising effectiveness is not typically determined by brand attitude







formation at the advertising exposure occasion, but by consumer perceptions of
accessible elements of the advertising message at the time of brand choice. The
second axiom states that advertising information must be relevant to be effective.
Breakdowns in advertising effectiveness often occur because advertising is
communicating the wrong level of information.
Four propositions flesh out these two axioms. The first proposition states
that advertising effectiveness is ultimately determined at the point of brand
choice. Attempts to study advertising effectiveness at or near the time of
advertising exposure are likely to greatly overstate effectiveness. The second
proposition states that consumers use three levels of information to make brand
purchase decisions. These levels are relative performance information, quality
cues and pure affect (feelings). The third proposition states that brand response
involvement determines the level of information that consumers prefer to use to
make their brand choice. When brand response involvement is low, consumers
react to the affect evoking qualities of brands. When involvement is moderate,
consumers tend to seek quality cues to reassure themselves they are buying a
"good" brand. When involvement is high, consumers are motivated to optimize
their brand choice by using specific brand information that provides direct
evidence of a brand's relative quality. Proposition four states that advertising
message involvement predicts the level of information that consumers are most
likely to efficiently encode at the advertising exposure occasion. Low involvement
is associated with the automatic encoding of affect; moderate involvement is
linked to the encoding of absolute quality cues and high involvement is
characterized by the processing of detailed relative performance information.
The major implication to draw from these propositions is that advertising
effectiveness is most likely to be successful when advertising messages
communicate at the level of information consumers are most likely to use to







discriminate brands at the purchase occasion. Given this, advertising
effectiveness will be optimized when consumer advertising message involvement
corresponds to brand response involvement. Only then can advertising exposure
typically lead to the efficient encoding of the level of information that consumers
are motivated to use to make their brand choice. Unfortunately, there are many
situations when advertising message involvement is likely to be logwe than brand
response involvement. When this occurs, breakdowns in advertising effectiveness
are likely to occur.



Axiom 1:
Advertising Information Must be Accessible to be Effective


The RAM assumes that advertising can be effective only when the
impressions it creates in a consumer's memory are retrieved (accessed) from
memory when brand evaluations are made. These impressions may be visual,
verbal or emotional.

The idea that a persuasive communication's effect on evaluation is dependent
on the memorability of elements of that communication, but not necessarily an
attitude or judgement of that communication made at the time of exposure, is not
new (Tversky & Kahneman,1973; Lichtenstein & Srull, 1983; Feldman & Lynch,
1988). However, a great deal of current advertising research assumes, sometimes
implicitly, that advertising works by influencing brand attitudes or judgements
at the time of advertising exposure, which in turn influence choice at a later
time. The RAM views brand attitudes, judgments or preferences formed at the
time of advertising exposure as potential mediators of brand choice but no more or
less likely to actually mediate brand choice than memory of the actual or







mentally transformed advertising message content (e.g., benefit claims,
spokespeople endorsements, emotion evoking music and visuals, etc.).
This assumption is simple, but as the brief discussion that follows
illustrates, predicting the accessibility of advertising information is a very
complex issue; there are a multitude of factors to consider.
When a consumer is exposed to advertising, that advertising leaves an
impression in memory. Depending on the level of attention a consumer directs at
an advertisement, the memory impression it leaves may be very weak and difficult
to access at a later time, or it may be very strong and simple to access (Greenwald
& Leavitt, 1984). The direction of focus of a consumer's attention influences the
elements of the advertising that are most likely to be retrievable. For example, all
things equal, if a consumer focuses on the executional aspects of an advertisement
such as background scenery or an actor, then that is what is most likely to be
remembered. If, on the other hand, the consumer focuses on the message, then
that is what is most likely to be remembered.
The impressions left in memory may bear little resemblance to the actual
content of advertising. Consumers may misunderstand the contents of a message
(See Jacoby, 1973). For example, they may believe a message claimed that a pain
reliever lasts six hours, when in fact the message claimed twelve hours. They also
may mentally "summarize" a message. For example, while viewing television
commercials, consumers may make make mental abstractions such as, "my
detergent doesn't get my clothes this white," "Honda really handles the curves" or
"this ad is stupid." These mental summaries and evaluative reactions are more
likely to leave a highly accessible impression than the literal elements of the
advertising (Alba & Hasher, 1983; Carlston, 1980; Wright, 1980).
Advertising information that consumer perceive to be meaningful is more
likely to be heavily attended than information that is not so perceived







(Burnkrant & Sawyer, 1983). For example, advertising that conveys brand
benefits that are important to consumers, depicts brand usage situations that
make sense to consumers and portrays lifestyles and living situations with which
consumers can relate, is more likely to be perceived as meaningful. Since
meaningful information is attended and elaborated more than non-meaningful
information, it is more likely to be accessible at a later time.
There are qualities to information that make it intrinsically more or less
accessible. A picture may not always be worth a thousand words, but visual
information is typically easier to encode and retrieve from memory than verbal
information (See Anderson & Bower, 1980). Unique, distinctive verbal

expressions are easier to remember than common expressions (See Berlyne, 1970).
Highly arousing, mood-evoking information, something frightening, erotic, or
beautiful, is inherently more memorable than flat, non-stimulating information
(See Kroeber-Riel, 1979).
"Environmental" factors also exert a major influence on the accessibility of
advertising information. Accessibility increases with repetition. If repeated
enough, some types of information can be retrieved automatically without effort
when a person sees a brand name or package (Schiffrin & Schneider, 1977; Bargh,
1984). Accessibility increases in the presence of memory cues (Bettman & Sujan,

1987). A brand name, a brand package and scenes or characters from an
advertisement can all cue the retrieval of other advertising information.
Accessibility decreases as the delay between advertising exposure and brand
evaluation increases (See Wyer & Srull, 1986). Some qualitative types of
information decay in memory at a faster rate than other information (Moore &

Hutchinson, 1985; Baker, 1985).
Due to the many motivational (i.e., desire to pay attention to advertising),
structural (e.g., advertising copy and execution elements) and environmental







(e.g, repetition and delay) factors that influence the accessibility of advertising
information, it is difficult at best for any theory to reliably predict one
individual's ability to access specific elements of a given advertisement. This is
not the intent of the RAM. The intent of the RAM is to predict the relative
accessibility of the information inherent in one type of advertising strategy
versus other types of advertising strategies in a given situation across a large
number of individuals. The model as presented in this paper will emphasize
motivational antecedents.



Axiom 2:
Advertising Information Must be Relevant to be Effective


The RAM further assumes that accessed advertising information must
explicitly or implicitly be perceived by the consumer to be relevant at the time of
brand choice if it is to have a direct effect on brand choice. Explicit recognition of
relevance refers to those situations when consumers consciously, deliberately seek
out specific information because they believe it to be a more reliable indicator of
quality than other information. Information that is explicitly judged to be
irrelevant will be discounted in favor of other more relevant information. Implicit
recognition of relevance refers to the more subtle influences of information that is
not sought out, but that can automatically drive behavior unless it is deliberately
discounted, influences such as the effects of familiarity (Baker et al., 1986) and
classical conditioning (Gorn, 1983).

This assumption of relevance is completely consistent with the focus of
applied marketing research and many academic models of persuasion and
decision-making. It is standard marketing research practice to identify the
product benefits that are most and least important to consumers when they are







making purchase decisions. Techniques to identify the most relevant benefits in a
product category range from standard focus group and survey research
methodologies to more sophisticated technologies like conjoint analysis and multi-
dimensional scaling. Many academic models of persuasion and decision-making
also act on the assumption that consumer choice is a function of perceived
relative brand performance on one or more product benefits (Lutz, 1975; Lynch,
Marmorstein & Weigold, 1988).
One commonality among most applied and academic models of decision-
making is that they tend to look at relevance from the perspective of relative
brand performance across a series of explicit product benefits or attributes. Using
these approaches, for example, cars may be evaluated on the basis of styling, speed,
comfort and depreciation. Microwaves may be evaluated on the basis of cooking
speed, price and durability.
The purpose of the RAM is Ino to predict which of a set of specific
performance or value features are most important, most relevant to consumers.
The RAM examines relevance from a more general perspective. Its purpose is to
predict whether specific relative performance information is likely to be used at
all to make brand choices. Like the Elaboration Likelihood Model (Petty &
Cacioppo, 1981) and the Percentage Contribution Model (Batra & Ray, 1985), for
example, the RAM proposes that consumers base brand choices on different
qualitative categories of information, each of which is perceived to be most
relevant in different situations.
The RAM presumes that the perceived relevance of various types of
information is determined primarily by motivational factors. However, the model
also recognizes that there are a number of factors that may make it difficult or
impossible for consumers to use the type of information that they may prefer to
use in a situation. There are environmental factors such as the availability of







the preferred information either externally (e.g., in a store) or internally (e.g., in
memory), experienced based factors such as expertise (e.g., the ability to
discriminate brands using the desired information) and opportunity based factors
such as time (e.g., it may take too much time to choose a brand using the
information a consumer would prefer to use) (See Alba & Hutchinson, 1987). A
focus of this paper will be on proposing how consumer motivation at the time of
brand choice influences perceived informational relevance and how this relates to
the issue of maximizing advertising effectiveness.
The bulk of the remaining portion of this paper will be devoted to presenting
four major propositions of the RAM. The final part of the paper will outline the
implications of these propositions for maximizing advertising effectiveness.




Proposition 1:
Advertising Effectiveness Occurs at the Point of Brand Choice


The RAM presumes that brand purchase decisions typically are not made
when people are watching, reading or listening to advertisements. Given this,
the effectiveness of an advertising message strategy must be judged at the brand

response occasion (i.e., the point of brand choice), not the advertising exposure
occasion (i.e., the point of advertising exposure). Usually, these two occasions are
separated by time and space. Typically, advertising message exposure occurs in
the home or car, while brand choice decisions occur at the point of purchase,
usually a store.
Two implications follow directly. First, any persuasive effects of advertising
at the advertising exposure occasion may no longer be accessible at the brand

response occasion. Second, advertising information that is perceived as relevant






at the advertising exposure occasion may not be so perceived at the brand response
occasion.


Memory Decay

Any theory of advertising effectiveness that does not consider the effects of

the decay of advertising information in memory is likely to overstate the impact

of advertising because it exaggerates the memorability of advertising. Consumers

are exposed to hundreds of advertisements every day. Their ability to remember
advertising information over time is limited. As the time between advertising
message exposure and brand choice increases, the ability to remember advertising

information deteriorates rapidly (Ward & Sawyer, 1981).

Similarly, any theory of advertising effectiveness that does not consider the

memorability of advertising relative to other sources of information at the brand

response occasion is likely to overstate the impact of advertising effects on brand

choice at the brand response occasion. Advertising does not operate in a vacuum.

Consumers can base their brand choice decisions on many sources of information
other than advertising such as prior brand usage experiences, package

information and point of purchase display information. Advertising must

compete with these other sources of information at the brand response occasion. If

advertising effectiveness is judged at the time of advertising exposure, then it is

likely to be exaggerated because it is being given a contrived advantage over these

other sources of information.

Since consumers often make purchase decisions quickly with little deliberate
mental effort (Krugman, 1965, Ray, 1973; Hoyer, 1984), in many choice

situations easy to remember, "top of mind" information is more likely to be used to

make a brand choice than less accessible, more difficult to retrieve information,

even if this more difficult to retrieve information is of equal or greater relevance







(Hoch, 1984; Wyer & Srull, 1986; Feldman & Lynch, 1988). As the time between
the advertising exposure occasion and the brand response occasion increases, the
amount of effort it takes to remember and use advertising information relative to
other sources of information increases (See Anderson, 1983). So, even if
advertising information is not forgotten, if it takes considerable effort to
remember it, other more readily accessible information is likely to take its place at
the brand response occasion.
Consider an example. Assume that a shampoo advertisement imparts a
relative advantage for the shampoo over its competitors, perhaps by a discussion of
a critical control ingredient, a professional endorsement of the product or other
aspects of the advertisement. If choice took place at that moment, the consumers
it persuaded would choose it.
However, over time, the memory of the commercial will fade considerably.
After two weeks, many of the consumers originally persuaded may no longer
associate any claims in the advertisement to the brand. In this scenario, the
advertisement can no longer influence choice unless the consumers made a
mental note at the time of advertising exposure to buy the brand, and remembered
it (Lingle et al., 1979; Carlston, 1980). Some of the consumers may be able to
remember the message with considerable effort, but may not be willing to make
that effort at the store. In this case, even if the consumers intend to choose a
brand that gives extra control, or that is endorsed by professionals, the
advertisement is unlikely to influence choice because other information that is
more easily accessed such as package information or usage experience memories
will be used.







Changes in Informational Relevance
The "pool" of easily accessible information from which to draw to make brand
choice decisions is likely to change significantly from the advertising exposure
occasion to the brand response occasion because the two environments are very
different. At the advertising exposure occasion for a given brand, advertising
information for this brand is likely to dominate the "pool" of easily accessible
information because it is an environment where other sources of information are
at an extreme disadvantage. There are no packages or point of purchase displays
that can provide information about competing brands and remind people of prior
usage experiences. Advertising information for competing brands is less
accessible because advertisements for these brands were not just seen, read or
heard.
Given this, advertising information is more likely to be perceived to be
relevant at the advertising exposure occasion than at the brand response occasion.
There is less information with which it must compete. There is less information
that can render the advertising information irrelevant to the choice decision for
consumers. Any theory of advertising effectiveness that does not consider the
potential that the perceived relevance of advertising information may be
artificially inflated at the advertising exposure occasion because other potentially
more relevant information to the consumer is absent or less accessible, is likely to
overstate the impact of advertising messages on brand choice.

Consider the shampoo advertisement example again. At the advertising
exposure occasion, women may report very high purchase intentions for the
shampoo because the information on the critical control ingredient or the
professional endorser seem relevant and, hence, motivates them to want the
brand. At the brand response occasion, however, new information may make this

information less relevant and, hence, no longer motivating. They may learn that







the shampoo costs twice as much as other brands or has a scent they strongly
dislike. They may notice other shampoos that they know through experience
provide excellent control, or recall that other brands have control ingredients just
as effective as the other advertised brand; or out of the biased context of the
advertising exposure occasion, they may remember an advertisement for another
brand that is endorsed by someone whose opinion they respect more than the
endorser for the other advertisement.


Summary
The RAM proposes that any theory whose purpose it is to predict the
effectiveness of advertising on choice behavior must evaluate advertising
effectiveness in the context of the brand response occasion, not the advertising
exposure occasion. Both the accessibility and the relevance of advertising
information are likely to be overstated in the context of the advertising exposure
occasion, which means the impact of advertising on brand choice decisions is
likely to be overstated. Depending on the evaluation of the advertising message,
the impact may be overstated positively (i.e., advertising motivates acceptance of
the brand) or negatively (i.e., advertising motivates rejection of the brand).
The accessibility of advertising information is biased in the context of the
advertising exposure occasion because its natural tendency to decay in memory
between the time of advertising exposure and brand choice is not considered. Also,
through the process of output interference, the high accessibility of brand
advertising information at the advertising exposure occasion inhibits the ability
of consumers to retrieve other information (Alba & Chattopadhyay, 1985).
The relevance of advertising information is overstated in the context of the
advertising exposure occasion because other potentially more relevant sources of







information that are typically present or highly accessible at the time of brand
choice are not so at the advertising exposure occasion.



Proposition 2:
Consumers Use Three Levels of Information to Make Brand Choices


The RAM proposes that advertising influences brand choice through the
effects of three general levels of information: Pure affect (feelings), brand quality
cues, and relative brand performance information.
Relative brand performance information is the most detailed, most direct
information that is available about a brand. It provides explicit, specific "apples
to apples" comparisons among competing brands to identify the best of a set of
brands. For example, a consumer may decide to compare cars on the basis of
performance features such as gas mileage and acceleration, or on the basis of
image features such as styling and color before making a purchase decision.
Much of the persuasion research in the 1960's (See Fishbein, 1967) and the
1970's (See Lutz & Swasy, 1977) operated on the assumption that consumers
explicitly compare brand alternatives across multiple relevant utilitarian and/or

hedonic benefits. Brand choice was expected to be the outcome of a series of
complex trade-offs among brand alternatives. The brand with the greatest
relative advantage across all product benefits was expected to be chosen. Other
approaches assumed that consumers make explicit comparisons between brand
alternatives, but only across the most important benefit they provide, or only
until a brand emerged as a clear winner on one key benefit (See Bettman, 1979).

The common thread through many of these popular models of persuasion and
advertising response is they viewed the consumer as a "rational man," as an
information processing machine similar to a computer. The input to this







information processing machine was assumed to be specific performance
information that allows direct comparisons to be made across brands.
Brand quality cues provide descriptive, but indirect information that can be
used to make inferences about a brand's quality, but they contain no definitive
evidence of a brand's performance relative to other brands. For example, rather
than take the time and expend the effort to compare cars on specific features, an
individual may take a short-cut and buy a car because it's a top seller, produced by
a manufacturer that stands for quality or status or luxury, or simply because it
possesses one or more attributes that are symbolic of quality (e.g., the door "thuds"
when slammed, front wheel drive, fuel injected engine). Although the reliability
of many quality cues may be unshakeable, one might say they provide no more
than "circumstantial evidence" of relative quality and performance.
Almost everybody can personally relate to making brand selections using
quality cues rather than more specific performance information. The specific
persuasive effects of commonly used quality cues such as brand name prestige
(Wheatley, Walton & Chiu, 1982), price (Monroe, 1973)), expert endorsement
(Sternthal, Dholakia, 1978 & Leavitt, 1978) and the sheer number of brand
benefits (Alba, 1987) are intuitively obvious and have been demonstrated
empirically. Theoretical models predict their usage when consumers do not have
the motivation, ability or opportunity to use more specific information (Chaiken,
1980; Petty & Cacioppo, 1981).
Pure affect refers to "free-floating" feelings and emotional responses that are
consciously unlinked to any specific brand attributes, benefits, past promotional
(e.g., advertising) or usage experiences, at least at the time of brand choice. Such
affect provides no real evidence of either absolute or relative product quality, but
consumers may interpret it as such. When individuals make decisions based on







pure affect, they are reacting solely to a feeling, not on the information that led to
that feeling.
Positive feelings intrinsically motivate approach behavior, negative feelings
motivate avoidance behavior (Zajonc, 1980; Skinner, 1974). These naturally
reinforcing properties of affect can influence brand choice. The sources of accessed
affect may be brand name familiarity (Harrison, 1977), the classical conditioning
of affect evoking stimuli like music or beautiful visuals to a brand through
advertising (Gorn, 1983), the affective residue from past usage experiences or a
combination of these and other affective experiences that have come to be
associated to a brand over time (Kiselius & Sternthal, 1984).
In the economists' world of the "rational man" making decisions with perfect
information, consumers would always use relative performance information to
make their brand choices because it is the most reliable, most direct brand
information available. Most marketing academics and practitioners agree that
this does not happen. In fact, everybody but economists agree that this does not
happen. Even a lot of economists agree that this does not happen.
It is certainly not "irrational" for consumers to make purchase decisions with
less than perfect information. We live in an era where brands competing in many
product categories are so equal that few, if any, substantive differences exist
between them. When product innovations do occur, they are copied quickly. In
such product categories many consumers believe that it is a waste of their time to
compare brands on the basis of specific performance features. When consumers
perceive they do not need specific information to make a brand choice, the
motivation to take the time and energy to meticulously compare brand
alternatives is low (Burnkrant & Sawyer, 1983; Robertson, 1976).
Another implication of the similarity among brand offerings is that
consumers may attempt but fail to find differences among brand alternatives on







specific features. For example, consumers may compare two sports cars across
several performance features and conclude they are equal. When this happens,
they may have no recourse but to use other types of information such as brand
quality cues or simple affective reactions to discriminate brands. In the end, one

sports car may be chosen over the other because it it is from a more prestigious
manufacturer (i.e., a brand quality cue), or perhaps more simply because, for some
reason unknown to the consumer at the time of choice, he/she is "drawn" to one
car over the other (i.e., a purely affective response).
In product classes where real differences do exist, the differences are
sometimes difficult to observe. Some consumers do not have the ability to compare
brands on specific information, others have the ability but do not have the time to
make these comparisons. Many consumers would like to compare automobiles on

the basis of specific engine performance features, but do not have the ability. If
grocery shoppers took the time to specifically compare every brand alternative on
the shelf before making every purchase, they would never get out of the store. In
both cases, circumstances force them to make decisions on the basis of other types
of information.


An Extended Example

The potential influence of all three levels of information on brand choice can
be illustrated through the following example. Imagine the following TV
commercial for a brand of shampoo. A young, alluring woman brushes her long

beautiful hair in a luxurious penthouse suite filled with classical music. The
woman explains that her shampoo has made her hair thick and manageable. A

graphic demonstrates that the shampoo has more of a critical control ingredient
than other shampoos. An announcer states that the shampoo is preferred by







professional hairdressers three to one over the leading shampoo. As the
commercial ends, a handsome young gentleman arrives at the penthouse.
All three types of information are included in this commercial. The
commercial may be persuasive for many different reasons. Some motivated
viewers may form specific relative performance beliefs about the shampoo and use
them at the point of purchase to choose the brand. For example, the graphics
display may persuade them that the shampoo provides more control than other
brands, or the theme of the commercial may persuade them that using this
shampoo will improve their social life more than other shampoos.
Other viewers may be persuaded on the basis of the brand quality cues it
communicates. One brand quality cue is the credibility of the woman. Viewers
may believe that a beautiful woman with beautiful hair is a credible source whose
recommendations should be considered. Another brand quality cue is the fact
that 75% of surveyed professional hairdressers prefer the shampoo. Many viewers
may find this to be reassuring evidence that the shampoo is of high quality.
Finally, some viewers may not remember any specific information from the
commercial, but may be persuaded by the affective "residue" generated by the

beauty, atmosphere and music in the commercial. The beautiful woman,
luxurious penthouse, sensuous music, and handsome man may have combined to
associate a very strong, positive feeling to the brand that influences brand choice
when they see the brand on the shelf in the supermarket.


A Brand Loyalty Caveat
The RAM operates under the assumption that consumers do not have
strongly formed preferences for a brand at the time of choice. In these cases, prior
preferences are likely to override the usage of raw information (i.e., relative
performance beliefs, quality cues and affect) and be the primary determinant of







purchase behavior. Situations where prior preferences are expected to dominate
brand choice are not expected to be common for the following reasons. First,
consumer durables (e.g., automobiles, major appliances) are modified frequently,
but purchased infrequently. Due to the changes in brand characteristics and the
potential presence of new brands, prior preferences become "obsolete" (Biehal &
Chakravarti, 1986). Second, there are so few differences between many consumer
nondurables (e.g., packaged foods, household cleaners, etc.) that there is little or
no basis on which to form strong preferences (See Lastovicka & Bonfield, 1982).
The ability of couponing, rebates, and other discounting strategies to constantly
"switch" consumers from brand to brand in these product categories indicates the

absence of strong performance-based preferences.


Summary

A central proposition of the RAM is that consumers can discriminate brands
by using three types of information, relative performance information, quality
cues and pure affect.
The key implication of this proposition is that advertising strategists need to

know which type of information their consumer target market is most likely to
use when making purchase decisions in a given product category. In any given
product category, advertising strategists need to know if their target is most

likely to deliberately compare brand alternatives across performance dimensions,

employ brand quality cues or rely on simple affective reactions to base their brand
choice decision.
The information type that is most likely to be used by the targeted consumers
for a given brand in a given product category needs to be emphasized in
advertising messages for that product category. If most targeted consumers for a
given brand of shampoo buy shampoos on the basis of simple affective reactions,







then the goal of advertising should be to associate positive feelings to the brand.
Communicating relative performance information is a waste of advertising copy
if that type of information is not expected to be used by consumers at the point of
purchase and is only likely to inhibit advertising effectiveness. Conversely, if
targeted consumers buy shampoos on the basis of relative performance
information, then the commercial should stress information capable of persuading
consumers that the brand is superior on one or performance dimensions and de-
emphasize elements of the ad that generate only positive feelings.




Proposition 3:
Involvement Determines Consumers' Preferred Level of Information


Brand Response Involvement

Simply put, involvement refers to the intensity of mental effort (Cohen,
1982). Brand Response Involvement is the degree of mental effort expended by the
consumer while making a brand choice. When brand response involvement (BRI)
is high, consumers are motivated to think intently about their brand choice; they
are motivated to explicitly search for information and compare brand
alternatives. At the other extreme, when BRI is very low, consumers are not
motivated to think about their brand choice; they have little desire to either seek
information or compare brand alternatives.
Major antecedents of brand response involvement are two dimensions of
perceived product risk: perceived product differentiation and the perceived
consequences of a nonoptimal brand choice decision (Robertson, 1976; Kapferer &
Laurent, 1985). Perceived product differentiation refers to the degree consumers
believe there are performance differences between the brand alternatives in a







product category. As perceived product differentiation increases, the probability
of making a "nonoptimal" brand choice increases. For this reason, increased
differentiation is expected to lead to increased brand response involvement. For
example, if a consumer believes there are performance differences between the
brands of shampoo on a supermarket shelf, he or she is more likely to spend time
seeking information and considering the brand alternatives than if he or she
believes that all shampoos are the same.
The probability that a "nonoptimal" brand choice has negative consequences
is also expected to influence brand response involvement. Negative consequences
may be economic, perhaps the loss of thousands of dollars that result from
choosing the wrong automobile. They may be social, perhaps the peer ridicule that
results from buying the wrong style of clothing. Or, they may be utilitarian,
perhaps from buying a microwave that is missing a useful feature such as
automatic cooking programs.
The highest level of brand response involvement is only expected when
consumers perceive both major product differentiation and serious consequences of
a nonoptimal brand choice. Perceived high product differentiation does not
necessarily imply that serious negative consequences will result from a
nonoptimal brand choice. Consumers may perceive that many differences exits
between brand alternatives and, as a result, may believe there is a high probability
of making a nonoptimal brand choice. They may also believe, however, that there
are no serious consequences of making a nonoptimal choice, no serious economic,
social or utilitarian injury. For example, some consumers may perceive there are
many differences between shampoos' color, smell, control, price, etc., but they may
also believe that a less than ideal choice will have little consequence because even
the worst considered shampoo will perform adequately. In this case, perceived







high product differentiation is not expected to lead to the highest level of brand
response involvement.
Also, the perception that there are serious negative consequences associated
with a non-optimal brand choice does not necessarily imply high perceived product
differentiation. Some product categories trigger perceptions of risk because of
their social, utilitarian and/or economic importance. However, brands within
these product categories may not be significantly differentiated. Choosing the
wrong pain reliever, for example, may have serious utilitarian consequences, but
people are not likely to meticulously choose a pain reliever unless they also believe
there are real performance differences between the alternatives in their
consideration set.


BRI Influences How People Make Brand Choice Decisions.
Brand response involvement is a cornerstone of the RAM because it is
presumed that as brand response involvement changes, different levels of
information are sought by consumers at the brand response occasion, and others
are avoided or ignored. The RAM proposes that brand response involvement is
the primary determinant of the level of information that consumers seek to make
brand choices. What follows is a discussion of three levels of brand response
involvement and how the RAM predicts they correspond to the three levels of
information described in Proposition Two.


High BRI leads to "optimization." The highest level of brand response

involvement is most likely to occur when consumers perceive both high product
differentiation and high product risk. That is, they believe there are major
differences between brand alternatives and they believe an nonoptimal brand

choice will have serious negative consequences. This situation motivates them to







want to optimize their brand choice; they want to make the best decision possible
and are willing to expend the mental effort required to make the best choice.
Optimization is characterized by consumers choosing what they believe to be
the best of a considered set of brands. Assuming consumers have the ability and
opportunity to explicitly compare brand alternatives, it makes sense that if
consumers wish to optimize their brand choice they will seek the most specific
information available information that permits direct "apples to apples"
comparisons between brand alternatives. When consumers are highly involved in
the purchase decision, the RAM presumes they will seek the highest level of
information, relative performance information, to make their brand choice.
Optimizing consumers are willing to spend the time and effort to read package
information, recall brand information from memory, talk to salespeople, read
Consumer Reports, etc.
Highly differentiated product categories with considerable economic, social
and utilitarian risk to many consumers are likely to include automobiles, homes,
durable appliances and clothing. Purchase decisions made to support hobby
(coins, stamps, etc.) and sporting interests (skiing, tennis, golf, etc.) are also
likely to be made in a state of high BRI.
Within a given product category, consumers may be motivated to optimize for
different reasons. Referring to the shampoo example, some consumers may be
highly involved because they want to buy the shampoo with the best control,
others may want the shampoo that is least likely to sting their eyes, others may
want the best-smelling shampoo. Whatever the cause, a person that is
optimizing his/her shampoo brand choice will explicitly compare shampoo
alternatives on the specific performance dimensions that he or she believes to be
important.







Much of the consumer persuasion and decision-making literature in the
1960's and 1970's explicitly or implicitly assumed that consumers optimize
purchase decisions. The plethora of multi-attribute expectancy value models
experimented with during this period cast consumers as sophisticated decision
makers calculating complex trade-offs across brand alternatives on several
utilitarian and/or social attributes (Ryan & Bonfield, 1980; Toy, 1982). These
approaches lost popularity as "all purpose" models of decision-making as the
concept of involvement driven decision strategies grew in popularity (Ray et al.,
1973) and because empirical tests have demonstrated that much simpler
predictors of behavior often perform as well as compensatory models (Kraft,
Granbois & Summers, 1973).
However, advertising research in consumer involvement has found that
consumers do tend to use more specific performance or product related advertising
message information when consumer involvement is high. Such research
supports the idea that consumers do not always optimize, but they are most likely
to attempt to optimize when they are in a state of high involvement. That is,
consumers are most likely to evaluate brands on the basis of specific product
related information when consumer involvement is high (See Ray et al., 1973;

Robertson, 1976; Wright, 1980; MacKenzie, Lutz, & Belch, 1986).
Moderate BRI leads to "satisficing." When consumers perceive significant
product differentiation, but little product risk, or perceive significant product
risk, but little product differentiation, brand response involvement is likely to be
moderate.

In the first case, consumers perceive a high probability, but no serious

negative consequences of an nonoptimal brand choice. For example, a man may
believe that microwaves vary in how quickly they cook food, but he may also believe

they all cook food fast enough for his purposes. Or, a woman may believe that







some conditioners leave hair more manageable than others, but she may also
believe the three or four brands in her consideration set all leave her hair
sufficiently manageable. Assuming that the man is buying a microwave
primarily to cook food fast and the woman is buying conditioner primarily to leave
her hair manageable, there is no real reason for the man or the woman to optimize
their brand choices. Both do not perceive any serious consequences of an
nonoptimal purchase. As a result, the marginal effort required to optimize is not
worth the marginal gain in product performance.
In the second case, consumers perceive serious consequences of a wrong choice,
but little chance that a wrong chance will be made. For example, the purchase of
a sports car has very high social risk to a young man that wants to improve his
image among his peers. However, the three cars that remain in his consideration
may all be very popular among his peers. Or, a woman buying a mink coat

perceives serious economic consequences if she buys the "wrong" mink coat simply
because they cost thousands of dollars. However, the three coats she has to choose
from all cost about the same. Assuming that in both examples there are no other
major perceived differences across the alternatives on risky dimensions, there is

no reason for the man or the woman to optimize their brand choice. Again, the
marginal effort required to optimize is not worth the marginal gain in
performance.

When consumers believe there is no benefit to be gained from putting extreme
mental energy into their purchase decision, BRI is likely to be moderate and
optimization is not likely to occur. However, due to the high perceived risk Qr
high perceived differentiation associated with the purchase, consumers need to
feel comfortable that they are making a satisfactory brand choice. In this scenario
consumers are expected to seek information to reassure themselves that they are
making a good selection.







When BRI is moderate, the RAM presumes that people make satisficing
brand decisions. Satisficing is characterized by consumers selecting the fist
acceptable brand, not necessarily the best of a considered set of brands. The RAM
expects that people will typically use quality cues to select a brand when they are
satisficing. Quality cues include a prior satisfactory usage experience, a credible
spokesperson's endorsement, a friend's recommendation, the knowledge that a
brand is a top seller or the professional image generated by high quality
advertising. The knowledge that a brand possesses a key desired ingredient (e.g.,
aceteminophen, no salt, no sugar), or feature (e.g., automatic cooking programs,
self-cleaning, permanent press, etc.), or that a brand is linked to a particular
lifestyle (e.g., BMWs and "yuppies") or usage situation (e.g., Michelob and "the
night") can also be quality cues. Whatever the quality cue, it must provide quick
reassurance or reinforcement that the brand is good, that the brand is a
satisfactory choice.
Different individuals are likely to use different quality cues to determine
acceptability. In the case of shampoo, some consumers may buy a brand name that
they trust, others may take the advice of a spokesperson, others may buy the most
expensive brand and others may take the first brand they see that has a special
ingredient or performance claim they desire.
Whatever the case, a satisficing brand decision typically takes significantly
less time and effort to make than an optimizing brand decision, primarily because
multiple brands are not explicitly compared across one or more performance
features. Consumers are not motivated to thoroughly search through their
memory or the purchase environment to find quality cues. They are "cognitive
misers" and do not want to expend much mental effort to make their choice (Wyer
& Srull, 1986). Consumers are likely to choose the first acceptable brand that
comes to mind. This means that brands linked with easy to remember quality







cues are more likely to be chosen than brands whose quality cues are less
accessible. When a consumer target market is expected to be satisficing, then,
marketers must not only associate persuasive quality cues to their brand, they
must make their quality cues more memorable than other brands' quality cues
(See Wright & Rip, 1980; Hoch, 1984).
The link between the use of quality cue type information and lower levels of
involvement is well established in both the advertising and social psychology
literature. Experiments involving the Elaboration Likelihood Model and similar
approaches have repeatedly shown that people less involved with an issue will
tend to use peripheral cues such as the credibility of a spokesperson to evaluate an
issue rather than specific message arguments (Petty & Cacioppo, 1981, Chaiken,
1980). Other research has shown that consumers less involved with advertising
tend to evaluate products on the basis of the appeal of the advertising execution
rather than the product related message arguments (MacKenzie & Lutz, 1986;
Baker, 1985).


Low BRI leads to "indifference." When BRI is very low, buyers perceive little
or no product differentiation and little or no product risk. There is a low perceived
probability of a "wrong" brand choice and trivial consequences if a wrong brand
choice is made.
If there are no serious perceived economic, social or utilitarian risks in a
product category and no perceived performance differences between brand
alternatives in the product category, there is little reason for a consumer to
expend any significant deliberate mental effort when making the brand choice.
It's as if there are no real choices, no real alternatives from which to choose.
In this situation consumers are expected to be "indifferent." Indifferent
purchase decisions are characterized by consumers selecting the first brand that







comes to mind. If there are multiple salient brand alternatives as in a
supermarket shopping situation, indifference purchase decisions are characterized
by consumers selecting the brand in the set with which they feel most
comfortable. Indifferent decisions are expected to typically be driven by simple
affective reactions towards brands that come to mind effortlessly at the time of
brand choice.
For example, if a woman believes there are no differences among shampoos and
no serious negative consequences associated with the choice of a "nonoptimal"
brand of shampoo, then her purchase choice is likely to be influenced by simple
affective reactions. Few, if any, deliberate brand comparisons or evaluation of
quality cues accompany the decision process. While scanning a shelf of brand
alternatives, she is likely to pick the shampoo that makes her feel most
comfortable, the shampoo that evokes the most positive emotional reaction. These
feelings may be created by the look of the package, by the familiarity of the brand
name or by images that advertising has associated to the brand. She doesn't
deliberately bring these feelings to mind, they automatically come to mind when
she sees the brand.
When BRI is very low, decisions are made as quickly as possible. As a result,
factors like shelf position, POP displays and package designs are predicted to have
their strongest influence on purchase.
The RAM predicts that the free-floating affect that drives purchase
decisions when involvement is very low can result from effects of affective
classical conditioning (Gor, 1983) and the exposure effect (Zajonc, 1980). Other
research predicts that these types of effects are most likely to influence brand
choice when consumer involvement in the purchase process is at its lowest level
(Nord & Peter, 1980; Krober-Riehl, 1979, 1984; Obermiller, 1985).







Summary
Three types of purchase behavior have been identified and linked to brand
response involvement: optimizing, satisficing and indifference. Each type of
purchase behavior corresponds to a different decision-making process. Each
decision-making process is associated with a different level of information:
relative performance information, quality cues and affective reactions.
The advertising implications are straightforward. For any given product
category, strategists need to identify the expected level of BRI of their consumer
target market and produce advertising that communicates that level of
information. When consumers optimize, present relative performance
information about key product benefits. When consumers satisfice, present
memorable, credible quality cues. When consumers are indifferent, build package
and brand awareness and link positive feelings to the brand.
In the case of shampoo, if consumers are expected to optimize, then an
advertisement presenting specific performance information that allows direct
comparisons with other brands has the best chance of directly influencing brand
choice. If consumers are satisficing, then an advertisement that uses a credible
quality cue such as a famous model is most likely to be effective. If consumers are
making "indifferent" decisions, then an advertisement that builds brand name
familiarity and associates good feelings to the brand is most likely to be
effective.lf advertising is produced that is not communicating the "correct" level
of information, then the RAM predicts that the impact of advertising at the
point of purchase will suffer. If shampoo purchasers are optimizing at the point of
brand choice and advertising presents simple quality cue type information, or if
shampoo purchasers are operating indifferently and advertising presents relative
performance information, then advertising is less likely to be effective because the
advertising is communicating an inappropriate level of information; it is







communicating information that is not likely to be used at the time purchase
decisions are made.



Proposition 4:
Involvement Determines the Level of Information Learned


Advertising Message Involvement
Advertising messages that communicate the "correct" level of information,
the level of information consumers are expected to use at the brand response
occasion, are not necessarily effective. Simply because an advertisement contains
relative performance information, for example, does not mean that a consumer
exposed to the advertisement will remember that information or the implications
of that information at the time of brand purchase.
One reason a consumer may not remember the relative performance
information in an advertising message is because the message may not have been
"processed" in a manner that could lead to the efficient "storage" of the relative

performance information in memory. For example, simply because an automobile
commercial contains the results of a braking test among three cars does not mean
that those results will be remembered by a consumer exposed to the advertisement.

The consumer may not have paid sufficient attention to the relative performance
information to store it in memory in a manner that would make it readily
"accessible" when he or she is in the process of purchasing a car. Generally,

consumers must mentally rehearse and perhaps simplify complex information in
order to remember its meaning at a later time (See Alba & Hutchinson, 1987).
The RAM predicts that consumer motivation to attend, comprehend and
integrate advertising message information into memory at the time of

advertising exposure is a primary determinant of the level of information that







consumers will "take away" from the advertisement and be likely to remember at
the time brand decisions are made (Greenwald & Leavitt, 1984).
Advertising message involvement (AMI) refers to consumer motivation to
attend, comprehend and integrate advertising message information into memory
at the time of advertising exposure. As advertising message involvement
increases, the cognitive effort expended by the consumer to process the contents of
an advertising message increases (Gardner, Mitchell & Russo, 1985).


Advertising Message Involvement Antecedents
There are many factors that are likely to influence a consumer's level of
advertising message involvement. Advertisers are not helpless. There are tactics
they can employ to increase consumers' motivation to attend to advertising. One
critical factor is information meaningfulness (Burnkrant & Sawyer, 1983). If
advertisers are going to include specific performance information in advertising,
then it needs to be information about benefits that are most important to
consumers. If the message is going to include quality cues, then it needs to
discuss quality cues that people perceive to be reliable predictors of quality. If the
message goal is to make people feel good, then the visual (e.g., people, situations,
scenery) and audio (e.g., music) tactics used to create positive emotions must be
ones that are most likely to touch chords among the advertiser's customer target
market.

Advertisers can also employ certain executional tactics to increase
advertising message involvement. Unique, vivid visuals, sounds and dialogue are
more likely to capture and hold attention than common ordinary executions
(Berlyne, 1970; Reyes, Thompson & Bower, 1980). And executions that are
emotionally intense are more likely to grab attention than emotionally flat
executions (Silk & Vavra, 1974).







Despite these tactics, there are several factors that influence advertising
message involvement that are at least partially out of the control of advertisers.
As a result, advertisers cannot necessarily influence the level of AMI at which
an advertisement will be processed. This has strong implications for advertising
effectiveness. Four relatively uncontrollable antecedents of AMI include (a)
anticipated brand response involvement, (b) the extent to which the consumer
perceives a need for additional information before making purchases in the
product category, (c) the consumer's interest in the advertised product category,
and (d) consumer distraction at the time of advertising message exposure.
If consumers typically behave "indifferently" at the point of purchase because
they don't perceive any differentiation or risk associated with purchases in that
product class, then there is no logical reason why those people would bother
deliberately attending to quality cue or relative performance type advertising
information in that product category. Unless an advertiser can somehow increase
a consumer's belief that there are differences between brands or that there is
social, utilitarian or economic risk associated with purchases in the product
category, it is likely to be very difficult to get people to attend to and learn
information at a higher level than what they expect to use when they make their
brand choice. Take the shampoo example. People that don't believe there are any
differences between shampoos may enjoy looking at the beautiful actress in the
advertisement and they may listen to the background music, but they are less
likely to pay much attention to the specific message information. People that do
believe there are differences, people that do associate risk with purchases in the
product category are more likely to focus their attention on the message.
Many people may be likely to make their brand choices in a product category
by either "optimizing" or "satisficing," but they still may not be likely to expend
much energy attending to specific performance or quality cue type information in







advertising if they do not believe they need additional information before they
make their brand choice (Burnkrant & Sawyer, 1983). Some people may believe
they have learned all the information they need to make a brand choice through
their prior experience with using brands in the product category. Other people
may simply believe that they don't need advertising information before they make
their brand choice, perhaps because they find it to be less credible than other
information (Smith & Swinyard, 1983). These people may believe they will need
information before they make their choice, but that they can get it at the point of
purchase through package information, salesmen or some means other than
advertising. Unless advertisers unleash messages that convince the former group
that they do need additional information or convince the latter group that they
are giving them valuable information that they cannot get elsewhere, then these
people are not likely to deliberately attend and learn advertising message
information.
Another factor that may prevent people who are likely to optimize or satisfice
at the time of brand choice from learning relevant information in advertising is
lack of interest in the product category. Although interest in any given product
category changes from one individual to the next, some product categories tend to
be inherently interesting to people, others are not. Interest alone can influence
the degree of attention paid to advertising (Muhlbacher, 1985). Automobile,
home electronics and clothing advertisers benefit from the high interest many
people have in their categories. This high interest increases the likelihood that
people will pay attention to their advertising. Battery, paper towels and
condiment advertisers are likely to be hurt by the intrinsically low interest many
people have in their product categories. These advertisers must work harder to
develop creative executions to draw consumers' attention towards their
advertising messages.







Finally, distractions at the time of advertising message exposure can direct
consumers' attention away from advertising. People do not watch television or
listen to the radio in isolation. Commercial time can also be conversation time,
reading time, snack time, etc.


Advertising Message Involvement Effects

The RAM predicts that advertising message involvement is a primary
determinant of the level of information that consumers most effectively encode in
memory at the time of advertising message exposure and, hence, remember at the
time brand purchase decisions are made.


High advertising message involvement. High product class interest, high
expected brand response involvement and a strong perceived need for information
are all expected to contribute to high advertising message involvement. When
AMI is high, consumers are expected to focus their attention on advertising
message information that facilitates the specific comparison of the brand relative
to its competitors. Message information relating to specific product attributes
and benefits is likely to be attended. Lower level information is likely to be used to
judge the credibility of the specific performance information, but not to directly

evaluate the brand. Consumers are expected to process the advertisement in a
manner that will allow the formation of relative performance beliefs either at the
time of advertising exposure or at the time of brand choice (e.g., better feature,

better product).
If consumers believe that no information in the advertising message is
pertinent to the formation of relative performance beliefs, then consumer

attention to the message is likely to "switch off' (Greenwald & Leavitt, 1984), or

worse, consumers may have a strong negative reaction that may transfer to the







brand (Petty & Cacioppo, 1981). Think about a commercial for a sports car. The
RAM predicts that highly involved consumers will focus their attention on those
aspects of the advertising message that will help them to form specific
performance beliefs about the car. If a consumer is most concerned with the social
implications of car ownership, then he will seek out information in the
advertisement that will help him to gauge the social status of this sports car
relative to others. If he is most concerned with the utilitarian aspects of sports
car ownership, he will seek out message information on, for example, the relative
acceleration or handling of this car versus other sports cars. If the advertisement
has no information that can be used to form relative performance beliefs, then the
person is likely to either withdraw his/her attention from the ad, or worse, become
frustrated and react negatively to the ad, and perhaps the car. The
advertisements for Nissan's Infiniti that did not show the automobile are likely
to have frustrated many people.


Moderate advertising message involvement. Lower product interest, expected
brand response involvement and/or perceived need for information are expected to
contribute to moderate AMI. When AMI is low, consumers are likely to be
motivated to comprehend the advertising message, but not to analyze and
integrate the implications of specific message claims with other brand-related or

competitive information in memory (Krugman, 1965). Consumers are not
expected to seek out specific relative performance information or expend the
cognitive effort to transform message information into relative performance
beliefs. Instead they are more likely to attend to quality cue type information and
make absolute mental judgements or summaries about the ad (e.g., good ad, good
spokesperson), the brand's benefits (e.g., the brand has this feature, this feature is

good) and the product in general (e.g., good brand). They are not expected to make







relative judgements about the advertisement (e.g., better ad, better spokesperson)
or the brand (e.g., better brand, better feature) because they are not expected to
deliberately relate or integrate brand information in the advertisement with
information about other brands (See Fiske, 1982; Alba & Hasher, 1983).
Recall the shampoo example. The RAM asserts that consumers exhibiting a
low level of AMI are most likely to attend to quality cues. They are neither
motivated to learn "details" about the brand because they are not interested in
such information nor do they expect to ever use it. It is easier to draw a quick
evaluative inference from this type of information than it is from more specific
performance information. Assume that Jacqueline Bissett is the actress in the
ad. Both Jacqueline Bissett's credibility and the hairdresser endorsement are
easily understood and have direct evaluative implications. If a consumer wants to
get a "quick read" of the brand from the advertising message, it makes more sense
to attend to the quality cues.
Quality cue type information is also more likely to be remembered than
specific attribute claims when one is not motivated to deeply process and
integrate the contents of an advertisement into memory because simple, novel
information is easier to remember than complex, common information (Alba &
Hutchinson, 1987). Quality cue information is often more novel than specific
performance information. Many shampoos have hair control ingredients, but only

one brand can be endorsed by Jaqueline Bissett. Consider product classes that
debate ceaselessly about who has the most of a given benefit, for example, the
amount of pain reliever in aspirin, the fiber content of cereals, the acceleration
rate of sports cars. It is difficult even for a motivated consumer to remember each
brand's specific claims on these common benefits. The barrage of claims interfere
with each other in memory and make it difficult for consumers to learn and access
it.







(Hoch, 1984; Keller, 1987). It is easier to remember the general themes of
advertisements in these product classes than it is to remember the values of the
debated benefits (e.g., the milligrams of pain relief offered by Bayer, Excedrin,
Tylenol, etc.) because the themes are generally more novel, more arousing and,
hence, easier to efficiently encode into brand memory than the factual content
(Anderson & Bower, 1980).


Very low advertising message involvement. Very low product interest,
anticipated brand response involvement and/or little or no perceived need for
brand information are likely to lead to very low AMI. Strong distractions at the
time of advertising exposure can also lead to very low AMI even if these other
conditions are not present. The RAM presumes that when advertising message
involvement is very low, there is no motivation to attend to brand information in
the advertising message. Consumers are not expected to expend any mental effort

to comprehend the message nor are they expected to intentionally generate any
absolute or relative mental summaries or judgements of the brand or any of its
benefits.
Consumer attention to any executional themes is expected to be purely a
function of their novelty and affective qualities. As a result, the only type of
brand information that is likely to be remembered is simple affective reactions to
executional elements of the advertising that became effortlessly associated to the
brand name through repeated exposure (Zajonc, 1980; Bargh, 1984).
Consumers exhibiting very low AMI while watching the shampoo commercial
are not expected to attend deliberately to any message elements. As a result,
memory for specific claims, whether they be quality cues or relative performance
information is expected to be poor. If, however, viewers found the music, luxurious
setting, or the model's beauty highly arousing, or if they were constantly exposed







to the brand name, it is possible that positive feelings could become linked to the
brand automatically. Consumers may not be able to explain accurately the
reasons) why their perceptions became more positive.


Summary

The RAM proposes that consumer advertising message involvement is a
strong predictor of the level of information that is most likely to be attended,
comprehended and integrated into brand memory at the time of advertising
message exposure and, hence, accessible to the consumer at the brand response
occasion.

When AMI is high, consumers prefer to attend to relative performance type
information. As a result, it is the level of information that is likely to be most
accessible at the brand response occasion. When AMI is low, quality cue type

information is most likely to be attended and efficiently encoded into memory at
the advertising exposure occasion and thus, be more accessible at the brand
response occasion. And, when AMI is very low, inherently interesting
executional elements of advertising are all that is likely to be attended at the

advertising exposure occasion and, as a result, remembered at the time of brand

choice.

The major implication of the effects of AMI is that advertising effectiveness

cannot be completely controlled by advertisers. An advertising message may
contain the correct level of information, information that is more relevant than

anything the consumer could find at the point of purchase, information that the

consumer would seek at the point of purchase; but if the consumer is not

attending and integrating this information into memory at the time of
advertising message exposure, the advertisement is not likely to influence brand







choice because the "correct" level of information is not likely to be efficiently
stored into memory.
Well designed advertisements can fail because advertising message
involvement and brand response involvement have different antecedents. Two of
AMI's "uncontrollable" antecedents, product category interest and the perceived
need for information, create the potential that advertising message involvement
will be different, usually lower, than brand response involvement. When this
happens, advertising is less likely to be effective because the "correct" level of
information is less likely to be attended to and stored into memory at the
advertising exposure occasion. For example, a consumer may use a satisficing
strategy at the point of purchase, but due to a low perceived need for information
and/or low product category interest, that consumer may have not paid sufficient
attention to a well designed ad containing reliable quality cues when he/she was
exposed to it. As a result, the ad is ineffective. Other information will drive the
brand choice decision at the point of purchase.



Advertising Effectiveness


Four Prerequisites to Effective Advertising
The RAM identifies four necessary and mutually sufficient conditions that
advertising must meet to have the highest probability of directly influencing
brand choice.
The first condition is information availability. Before advertising can have
any opportunity to influence brand choice, it must leave an imprint in consumer
memory (Kisielius & Sternthal, 1984). Advertisements must be presented at
times and in places where consumers will be exposed to their contents.







Advertisements must also affect consumer attention in a way that the message,
not just the execution, will be attended, comprehended and stored into memory. If
a consumer is not exposed to an advertisement, or if a consumer is exposed to an
advertisement but neither attends or comprehends the message, then advertising
has next to no chance of being effective.
The second condition is level relevance. Advertising effects are level relevant
if there is a one-to-one correspondence between the level of information sought by
the consumer to discriminate brand alternatives at the point of purchase (i.e.,
pure affect, quality cues or relative performance information) and the level of
information encoded into memory at the time of advertising exposure. If, for
example, consumers are seeking quality cues at the point of purchase and quality
cues were encoded into memory at the time of advertising exposure, then
advertising effects are level relevant. If, on the other hand, consumers are
making brand selections on the basis of relative performance information, but
encoded advertising effects are limited to simple affective information, then
advertising is not level relevant.
Two requirements must be met to maximize the likelihood that advertising
effects in memory will be level relevant. First, advertisements need to be designed
to emphasize the "correct" level of information as determined by expected
consumer brand response involvement. Second, consumer advertising message
involvement needs to correspond to brand response involvement.

Using the shampoo commercial example, if an advertisers' consumer target
market is expected to satisfice at the time of brand choice, then advertising effects
are most likely to be level relevant if the shampoo advertising emphasizes quality
cues rather than pure affective information or specific performance information.
The quality cue information may be an expert endorsement or survey research
evidence that three out of four hairdressers recommend its critical control







ingredient. Level relevance will be maximized if this commercial is processed at
the middle level of advertising message involvement. At this level of
involvement, consumers are expected to be most likely to attend to the quality cue
type information in a manner that it will be accessible at the brand response
occasion.
The third condition is relative accessibility. Relative accessibility refers to
the probability that level relevant advertising effects will be accessed at the time
of brand choice. Level relevant advertising effects may be available in memory,
but the amount of effort required to retrieve them relative to other sources of
information such as past usage experiences or package information can inhibit
their retrieval at the point of purchase.
The relative accessibility of level relevant information is a direct function of
how easily that information can be retrieved at the point of purchase relative to
other accessible level relevant information such as package information or past
usage experiences. Relative accessibility is influenced both by non-motivational
(e.g., message design factors, repetition, package or POP tie-ins) and
motivational (e.g., consumer effort to attend, comprehend and integrate the
advertising message into memory) factors.
The relative accessibility of advertising information becomes increasingly
important as brand response involvement decreases. As BRI decreases, consumers
are expected to expend less effort to search their memories for relevant
information. When BRI is high, consumers are motivated to deliberately search
for relevant information regarding every brand in their consideration set. When
BRI is low, consumers are only expected to search until they access a quality cue
that allows them to confidently select a brand. When BRI is very low, consumers
are not expected to deliberately try to recall any information from memory.
According to the RAM, then, relative accessibility is a more critical issue to







advertisers promoting brands to consumers that are not involved in the purchase
process.
The third condition is relative relevance. Assuming that an advertising
effect in memory is level relevant, its relative relevance is its perceived ability to
better, more reliably discriminate brand alternatives than other level relevant
information such as package information, salesperson advice, usage experience
memories, etc. Within each of the three levels of information, the RAM supposes
that consumers encode, access and use information they perceive to be most
reliable and to avoid and discount implications of less important information.
The relative relevance of advertising information becomes less important as
brand response involvement decreases. Optimizing consumers are expected to
take the effort to locate and use the information they believe to be the most
reliable discriminator of brand alternatives; relative relevance is important.

Satisficing consumers seek only a reliable brand discriminator, not necessarily
the best discriminator; relative relevance is less important. Indifferent
consumers do not deliberately seek or use any information to base brand choice,
relative relevance is not considered.


Maximizing Advertising Effectiveness
Within the context of the relevance-accessibility model, several steps can be
taken to meet the conditions outlined above. The first two steps are unique to the
RAM, but also consistent with much of the consumer involvement literature.
The next three steps are not unique to the RAM but are cast within the context
of the RAM's major propositions.
First, advertising strategists need to correctly predict the brand response
involvement of the consumers they intend to target. This knowledge will reveal
which level of information needs to be emphasized in advertising. Brand response







involvement can be predicted by studying its primary antecedents, particularly
perceived product risk and product differentiation. For example, in the case of
paper towels, if strategists learn that their target consumers perceive little or no
substantive differences between brand alternatives and no serious negative
consequences associated with a nonoptimal purchase, then they can conclude that
brand response involvement is very low.
Second, advertising messages must communicate level relevant information.
If brand response involvement is expected to be high, then the advertising
message should communicate relative performance information. If it is expected
to be very low, then it should build brand and package awareness and effectively
associate pure affect to the brand. For example, if the consumer target for a new
sports car is expected to be in a state of high brand response involvement during
the automobile purchase process, then to impact directly on brand choice,
advertising needs to contain information on factors such as acceleration, braking,
handling, styling, comfort, status, etc., relative to other sports car alternatives.
If the consumer target for paper towels is expected to be in a state of very low BRI
during the purchase process, then to impact directly on brand choice, advertising

need only increase brand name and package familiarity and associate good feelings
to the brand.
Third, advertising messages must communicate relatively relevant
information. Once the level of information the target market is expected to use to
make the brand choice is identified, research is needed to identify the best quality
cue, the most important performance benefit and/or the most arousing, compelling
execution to associate affect to the brand. In the case of the sports car, if BRI is
expected to be high, then research needs to determine which performance
dimension is the most important to communicate through advertising.







Fourth, the advertising message must be designed to maximize its relative
accessibility to the consumer target market. Tactics strategists can use to
improve relative accessibility include communicating the message in a novel and
arousing manner, repeating the message throughout the advertisement,
communicating the message through multiple sensual modes (words, sounds and
pictures), focusing consumer attention on the message, not the execution, and
heavy repetition schedules.
Maximizing the relative accessibility of advertising information is not
something that advertisers can control completely. Advertisers can influence it
by designing relevant advertising messages with compelling executions, but they
cannot control consumer advertising message involvement at the time of
advertising exposure.


The Principle of Optimal Advertising Contribution

The RAM presumes that advertising effectiveness is most likely to be
maximized when steps one through four above are accomplished and consumer
advertising message involvement corresponds to consumer brand response
involvement. When consumer advertising message involvement and brand

response involvement correspond, advertising has the best chance of being

effective because the level of information that is sought by the consumer at the
time of brand choice is most likely to correspond to the most accessible advertising
effects in memory.
For example, if a consumer's brand response involvement is moderate when
selecting a shampoo, then the RAM predicts that the person is most likely to seek

a quality cue to select a brand at the point of purchase (e.g., top selling brand
name). If the consumer also experienced moderate advertising message
involvement when exposed to the shampoo commercial, then the persuasive effect







of the quality cues presented in the advertisement are most likely to be highly
accessible at the brand response occasion (e.g., spokesperson credibility). When
both AMI and BRI are moderate, quality cues are most likely to be sought at the
brand response occasion and the effects of quality cues in the advertising message
are most likely to be highly accessible at the brand response occasion.
If, on the other hand, the consumer experienced very low AMI when exposed
to the advertisement, then the advertisement is expected to have a significantly
lower chance to influence consumer choice because the quality cue information in
the message is expected to be less accessible at the point of purchase. The
consumer is not expected to have deliberately attended to the quality cue
information in the advertising. The advertising effect that is more likely to be
accessible in this situation, free-floating affect, is expected to be perceived as less
relevant to the consumer. Rather that base a choice on advertising information
in this situation, the consumer is expected to use a quality cue from some other
information source.



Summary


The Relevance-Accessibility Model of Advertising effectiveness has several
key implications for the planning and implementation of advertising strategy.
First, it stresses the importance for advertisers to understand more "global"
aspects of consumer decision-making, not just tactical methodologies such as
benefit segmentation. In order to develop a sound advertising strategy advertisers
must first be able to ascertain what level of information consumers are likely to be
using when they make purchase decisions. Second, it asserts that advertisers
need to focus on brand response involvement as the critical input to determine







whether to pursue an affective, quality cue or attribute based advertising
strategy.
By studying the antecedents of advertising message involvement and brand
response involvement, the RAM offers a means to predict those situations when
advertising is most and least likely to be able to directly influence brand choice.
In situations, where AMI and BRI can be expected to be relatively equal,
advertising has the greatest opportunity to directly impact on choice. In those
situations when AMI is expected to be less than BRI, advertising has less
opportunity to directly influence brand choice. Effective advertising in this latter
situation, requires more expert manipulation of the controllable antecedents of
AMI (e.g., highly relevant message content and unique, arousing executional
elements). In this situation it may be prudent to "downgrade" advertising's role
to influence consideration rather than choice. It may be best advised for other
elements of the marketing mix such as packaging, POP, sales promotion,
distribution intensity and shelf space dominance to take greater priority than
advertising.



What is Ahead in the Dissertation


The goal of this chapter was to provide a general overview of the RAM's
major propositions. Chapter Two will provide a comprehensive review of the

advertising effects literature that influenced the conceptualization of the RAM.
Chapter Three will present the RAM's specific theoretical framework. Chapter
Three will also isolate the central hypothesis to be tested in the dissertation.

Among others, the following hypothesis will be tested: All things equal, level

relevant advertising will be most effective when AMI corresponds to BRI. For




48

example, when BRI is high, advertising will be most effective when it emphasizes
relative performance information and when AMI is high. Similarly, when BRI is
very low, advertising will be most effective when it emphasizes pure affective
information and when AMI is very low.
Chapter Four explains the experimental design and methodology to test this
hypothesis. Chapter Five details the experimental results and their immediate
implications. Chapter Six offers a more general discussion of the implications of
the research and the model.











CHAPTER 2
LITERATURE REVIEW



Introduction


Supporting evidence for the RAM's axioms and propositions can be found in
several streams of research. This chapter will review the elements of these
streams that influenced the development of the Relevance-Accessibility Model of
Advertising Effectiveness (RAM). In order of discussion, this chapter will
review attitude-behavior consistency, the role of consumer elaboration and
involvement on evaluation, the role of automatic processing and affect on
evaluation and hypothetical construct perspectives on advertising effects (e.g.,
attitude toward the ad).
Familiar perspectives with few conflicting points of view, at least at the level
of generality at which they will be discussed, will be reviewed more briefly than
more recent, controversial evidence. In these latter cases, conflicting research
results will attempt to be reconciled as they pertain to the RAM by examining
boundary conditions and methodological limitations.
With a few notable exceptions, most of the research reviewed in this chapter
focuses on consumer activities at the advertising exposure occasion (AEO), not
the brand response occasion (BRO); most commonly it does not separate the
advertising exposure occasion from the brand response occasion. Also, most

research on communication effects has used measures of brand attitudes or
purchase intentions rather than choice as the primary dependent variable. As a







result, support for the RAM's propositions about consumer choice activity at the
BRO can often only be inferred, not directly empirically demonstrated.



Attitude-Behavior Consistency


Introduction
Attitudes play a central role in models of consumer behavior. Theoretically,
often studied and emulated approaches such as the classic hierarchy of effects
(Lavidge & Steiner, 1961; Palda, 1966; McGuire, 1974), the Fishbein model
(1965), the Theory of Reasoned Action (Fishbein & Ajzen, 1975) and the
Elaboration Likelihood Model (Petty & Cacioppo, 1981) focus on the process of
changing or modifying brand (or object) attitudes. Methodologically, the vast
majority of communication effects research in the past few decades has relied on
paper and pencil measures of brand attitudes and purchase intentions to test the
persuasive impact of independent variables. Based on the literature, it is fair to
say that an explicit (i.e., theoretical) or implicit (i.e., methodological)
assumption of almost all models of communication effects is that advertising
ultimately affects behavior by modifying brand attitudes.
The RAM departs from this philosophy. The RAM does not espouse the
position that advertising-generated changes in cognitive structure necessarily
cause or reflect a change in attitude. Instead, it advocates that advertising more
typically changes the informational or affective content of brand memory that
may or may not later be deliberately or undeliberately invoked to make a brand
purchase decision.
Ultimately, consumer behavior entails choice and consumption, not brand
attitudes. There is no compelling theoretical reason for brand attitudes to be the







focus of study in consumer research. Methodologically, brand attitudes are
convenient substitutes for measures of real or simulated choice, particularly since
their interval structure permits the use of ANOVA and multivariate statistical
packages to analyze empirical research results. But the only real justification for
the use of brand attitude measures en lieu of choice measures in research is if
they are reliable surrogates of choice. However, there is no strong, consistent
evidence that this is the case. In fact, brand attitudes rarely explain more than
15% of the variance in consumer choice decisions (See Wicker, 1969; Smith &
Swinyard, 1983). The weight of the evidence suggests that far too much credence
has been placed on the notion that brand attitudes measured at time A reliably
predict behavior at time B.
Based on the evidence, the position of this thesis is that research on
communication effects needs to shift away from the traditional perspective that
advertising influences brand attitude which in turn influences choice to the
perspective that advertising influences memory structure which influences
choice. This latter perspective accepts the potential that a previously generated
brand attitude may predict choice, but it also allows for the probability that raw
information in the form of pure affect, choice cues or relative performance
information may be more predictive of choice.
Several reasons for this perspective will be presented. First, it is questionable
whether stable brand attitudes naturally exist in many product categories

(Lastovicka & Bonfield, 1982). Second, in those product categories where stable
brand attitudes do exist, direct brand experience rather than persuasive
communication is more likely to mediate their formation. Third, in the event
that a persuasive communication does mediate the formation or modification in
brand attitude, the necessary conditions for that attitude to directly mediate
subsequent behavior severely constrains the likelihood of that event.






For these reasons, the RAM proposes that it is more important to look at
advertising in the context of its ultimate contribution to choice processes that
discriminate brands rather than its influence on attitude formation, which may
be unrelated to choice. In other words, it is a more important goal for advertisers
to communicate the specific information that will be used to discriminate brand
alternatives than it is to communicate information that will elicit a positive
attitudinal response towards the brand.
A third perspective on the role of brand attitudes in the choice process is
what might be termed the "bias perspective". Fazio (1986) describes a chain of
events whereby salient brand attitudes bias the accessibility and interpretation of
information used in the choice process, thus biasing choice in the direction of the
salient brand attitude. A considerable body of research in addition to his own
supports Fazio's proposition (Isen, 1984; Owens, Bower, & Black, 1979; Bargh,
1984; Alba & Ofir, 1984; Hastorf & Cantril, 1954).

Table 1
Communication Effects Perspectives

Traditional Perspective: Advertising-->Brand Attitude-->Choice
RAM Perspective: Advertising-->Memory Structure-->Choice
Bias Perspective: Advertising-->Brand Attitude-->Perceptual Structure-->Choice

Defining Brand Attitudes
The importance of the attitude concept dates back to the 1920s when
behavioral scientists began to search for "black box" factors mediating behavior.
Behaviorism and genetics were deemed unsatisfactory. As Smith & Swinyard
(1983) observed, one problem with early definitions is that they presume a direct
effect on behavior. Allport (1935) defined an attitude as, "A mental and neural
state of readiness exerting a direct and dynamic influence upon the individual's






response." Similarly, Doob (1947) defined attitudes as, "An implicit response
which affects subsequent overt responses." Certainly, a workable definition of
attitude cannot presume attitude-behavior consistency.
There are two other requirements of a brand attitude definition within the
context of a model of choice. First, the definition cannot presume salience at the
time of choice. Attitude availability, let alone accessibility, is not a given at the
time of choice (Fazio, 1981; Bargh, 1984; Kisielius & Sternthal, 1984). Second,
given the renewed focus on the role of affect in decision-making (Gardner, 1986;
Lutz, 1985), the differences between brand attitudes and other affective responses
must be explicated clearly.
Here, a brand attitude is defined as a "deliberately formed brand evaluation."
It is both conscious and cognitive; it is not free-floating affect as described by
Fazio (1986). Linking brand attitudes with affect is not workable in the context
of the RAM because there would be a serious confounding of attitudes with affect
as information, the third level of information described in the model. However,
the source of the attitude may be "hot" affective responses (e.g., a strong
emotional response) or "cold" information (e.g., quality cues, specific performance
information).
This definition assumes that attitudes are formed through a conscious

cognitive process; they are products of deliberate expressions of summary
evaluations. This definition not only distinguishes an attitude from the three
levels of information described by the RAM, particularly free-floating affect; it
also recognizes and conforms to the methodological realities of attitude
measurement. Whenever an attitude is measured, that measurement by its
nature forces a cognitive component and deliberate expression, whether it existed
prior to the measurement or not (Lastovicka & Bonfield, 1982).







Do Brand Attitudes Typically Exist?
Brand attitudes cannot mediate behavior unless they exist. Most
communications effects modelers assume they are tapping not creating brand
attitudes in their research, but there are other points of view. Lastovicka and
Bonfield (1982) argue that people are likely to store firm attitudes towards social
issues, but not typically towards branded products. As a result, responses to
attribute belief and importance scales in survey research may often be misleading
and reflect temporary tendencies, created on the spur of the moment, simply
because a question was asked (See also Feldman & Lynch, 1988). Lastovicka and
Bonfield asserted that minor choice behavior can be best explained through
principles of behavioral learning theory because low involvement decisions, which
most brand purchases are thought to be, are unlikely to be governed by specific
attribute beliefs.
There is theoretical support for this position. According to functional theory
(Katz, 1960), in order for attitudes to be held toward a given object, the object
must be related to the concept of self, linked to an individual's important values,
relevant to an important social group and perceived to give differential need
satisfaction relative to other related objects.
Using these criteria, it may be argued that consumers are unlikely to hold
enduring attitudes towards brands in a great many product categories. The final
criterion above adds particular insight into the problem of attitude development
for many classes of branded products. It implies that consumers may have stable
attitudes toward product categories (e.g., peanut butter, paper napkins), but not
towards specific brands within categories (e.g., Skippy vs. Peter Pan peanut
butter). The lack of perceived interbrand differences in many product categories
minimizes the potential for differential need satisfaction among brand
alternatives within the product class (i.e., ketchup, paper towels, salt, etc.).







According to concept formation theory (Rhine, 1958), people acquire
denotative meaning about objects and issue, but evaluative and connotative
meaning do not follow unless the brand carries significant personal importance..
If one takes takes the position that most brands do not carry significant personal
importance, concept formation theory can used to argue that people are likely to
acquire knowledge of a brand that allows them to identify (e.g., "Crest is a brand
of toothpaste") and describe it ("It's made by Proctor & Gamble," "It's the
number one brand." "It has fluoride."), but not necessarily to evaluate ("It's an
excellent brand." "It's the best brand.") or prefer it ("I'm going to buy it next time
I need toothpaste."). Emotive and primitive evaluative responses, such as those
postulated to be created by advertising, may result in purchase behavior, but do
not necessarily lead to cognitively based attitudes. For example, repeated
exposure to the brand name and package can lead to exposure effect induced
purchase (Zajonc, 1980), emotional associations "conditioned" to the brand by
advertising can motivate buying (Kroeber-Riel, 1979, 1984), as can the operant
effects of price discounting (Nord & Peter, 1980). All of these effects, however,
neither require belief formation or attitude formation prior to purchase or after
purchase.

Table 2
Concept Formation Theory
Denotative Cognition-->Conation--> Attitude (maybe, but not often)

An experiment conducted by Lastovicka and Bonfield (1982) supported the
basic propositions of functional theory and concept formation theory as they relate
to branded products. Their results also supported the notion that close-ended
attitude measurement scales overestimate the presence of brand attitudes. The
authors found that 92% of subjects displayed attitudes towards twelve popular







brands of bread, toothpaste, toilet tissue and automobiles on a close-ended attitude
measurement task (i.e., semantic differential scale), but only 15% revealed
attitudes on an open ended (i.e., free association) task. In contrast, 96% and 45%
of respondents displayed attitudes on political issues on close-ended and open-
ended measures, respectively. Despite the study's small sample size, these results
differ significantly. As concept formation theory might predict, responses on the
open end measure of attitudes tended to be strictly denotative for the branded
products, but both denotative and evaluative for topical political issues.
The perspective that brand attitudes may be nothing more than
measurement artifacts, particularly in low involvement decision situations is not
new. Warnings of an over emphasis on deliberate information processing
perspectives to explain decision-making behavior have been sounded from other
sources as well (Kassarjian, 1977; Olshavsky & Granbois, 1978). Dissatisfaction
with the explanatory power of models driven by hypothetical constructs has
rekindled interest in behavioristic perspectives on consumer decision-making
which do not even acknowledge the role of attitudes in communication effects and
choice behavior The concept of persuasion without active consumer involvement
is consistent with behaviorism and the principles of behavior modification
(BMP). These principles act as an alternative, more parsimonious explanation
than information processing, belief-based interpretations of attitude formation
and change because there are fewer posited intervening steps between advertising
exposure and choice. Marketing objectives can be accomplished by studying
environmental conditions and manipulating them to influence consumer behavior
(See Nord & Peter, 1980).







Does Advertising Typically Motivate Brand Attitude Modification?
Assuming that brand attitudes do exist in a given product category, it is still
questionable whether advertising impacts behavior by mediating a shift in
attitudes. Evidence suggests that brand experience, and not advertising,
generally mediates the formation of stable brand attitudes. There are two reasons
why advertising may not reliably mediate changes in brand attitudes. First,
consumers often process advertising at a level of involvement that inhibits the
elaboration of message content (Krugman, 1965, 1971). Without elaboration,
there can be no change in brand attitudes (assuming the definition of brand
attitude given above which states that attitudes cannot be formed or modified
without the production of deliberate, evaluative cognitive abstractions). Second,
advertising is not a credible source of information relative to direct experience.
Even if processed at a level of involvement sufficient to motivate message content
elaboration, the level of counterargumentation caused by advertising often
inhibits the formation of strong attitudes (Smith & Swinyard, 1983).
The pioneering work of Krugman (1965) and Ray et al.. (1973) on low
involvement communication processing paved the way for the wave of research on
less deliberate, less involved communication response and decision-making

processes. Their research questioned the notion that communications necessarily
work by affecting brand attitudes through a classic hierarchy of effects type
process.
When consumers are highly involved, Krugman (1965, 1971) affirms a
classic hierarchy of effects type persuasion sequence. But when consumers are in
a state of "low involvement," he claims that they react passively to advertising
messages, generating few or no cognitive responses (Greenwald, 1968; Wright,

1973), or what he termed "conscious bridging experiences" from the message to the
self. Krugman was among the first researchers to claim that a large proportion of







consumer advertising processing is characterized by far less deliberate cognitive
activity than is assumed by hierarchies of effects perspectives. He asserted that
under low involvement, attitude change results from "a gradual shift in
perceptual structure," not deliberately induced changes in brand attitudes or its
antecedents (e.g., attribute beliefs). "In short, with low involvement product
choices, we might look for product adoption through gradual shifts in perceptual
structure, aided by repetition, activated by behavioral choices and followed at some
time by attitude change" (Krugman, 1965, pg 225).
The precise meaning of a "gradual shift in perceptual structure" was never
clarified, but the essential point is that the behavioral effects of advertising in
low involvement are not mediated by deliberate, conscious shifts in brand
attitudes. This point of view is consistent with the RAM's perspective that
advertising influences purchase behavior through other means than a direct
mediational link through brand attitudes stored in memory.
The Ray et al.. (1973) research program at Stanford University provided
strong empirical support for Krugman's gradual shift in perceptual structure
proposition. This research stream uncovered evidence for the existence of three
hierarchies of communication effects. The Learning Hierarchy is roughly
equivalent to a classic hierarchy conceptualization. The effects sequence is
cognitions (e.g., message comprehension, belief formation), followed by affect
formation (e.g., attitude formation); and eventually conation (e.g., purchase
behavior). This hierarchy is assumed to be operative when consumers are involved
in the purchase, presumably because clear, significant differences are perceived to
exist among brand alternatives.
The Dissonance-Attribution Hierarchy is the reverse of the learning
hierarchy. The persuasion sequence is behavior, affect, and cognitions. The
authors presume its occurence in situations where consumers are involved in the






choice process, but for whatever reason (e.g., ability or opportunity constraints)
are unable to distinguish the differences among brand alternatives. The
uncertainty created by the purchase choice is postulated to motivate consumers to
bolster their choice decisions through rationalizing processes that may include
the search for supportive information. The output of this search is hypothesized
to be attitude polarization and the formation of strong attribute beliefs supportive
of the chosen brand.
The Low Involvement hierarchy follows the sequence cognitions, behavior, and
affect. As predicted by Krugman (1965), brand choice is thought to occur prior to
attitude change, presumably because advertising-generated shifts in brand
cognitions occur imperceptibly and thus, do not result in conscious, deliberate
prepurchase attitude shifts. This hierarchy is expected to be dominant when
consumers are uninvolved with the purchase and/or perceive minimal differences
to exist among brand alternatives (Ray et al. 1973; Robertson, 1976).
A substantial literature supports the proposition that direct object experience
increases attitude-behavior consistency relative to vicarious experience (Smith &
Swinyard, 1982, 1983; Abelson, 1976; Fazio & Zanna, 1981; Fazio, Powell & Herr,
1983). Smith and Swinyard (1982, 1983) propose that vicarious experience
cannot typically lead to the formation of stable, belief-based brand attitudes
because it provides less tangible evidence on which to base attitude formation. In
the specific case of advertising, low involvement processing behavior and low
credibility further suppress the formation of belief-based attitudes. Advertising
can lead to positive approach tendencies that may initiate trial, but not
committed purchase behavior. Only direct experience can lead to the formation of
belief-based attitudes that can reliably predict behavior. This interpretation is
consistent with the findings of Fazio and Zanna (1981) that an individual with
an attitude derived from direct behavioral experience will hold that attitude with







greater confidence than an individual whose attitude is formed from indirect
experience. It is also consistent with the behavioral perspective that vicarious
reinforcement is not as effective as direct, personal reinforcement.
Smith and Swinyard's (1983) empirical work involving a new snack product
supported their position. The correlation between attitudes and choice was .32
when attitudes were generated by advertising and .60 when attitudes were
generated by direct product experience.
Attitudes formed through direct experience may also be more accessible than
attitudes formed through vicarious experiences such as advertising. Fazio (1986)
asserts that a stronger object-evaluation link is created when attitudes are formed
on the basis of direct experience. Several experiments support this proposition
(Fazio & Powell, 1984; Fazio, 1986). However, it is unclear whether the attitude
behavior consistency enhanced by direct experience is due to the greater
accessibility of a prior attitude, or greater accessibility of a prior behavior, which
consumers then use to construct rather than access an attitude as self-perception
theory would predict (Bem, 1972). So, while evidence strongly supports the
proposition that direct experience leads to stronger, more confidently held
attitudes relative to advertising, more research is needed to determine if direct
experience enhances the strength of the attitude-object link in memory rather
than a behavior-object link in memory that is used to infer attitudes.


Brand Attitude Accessibility
Assuming that brand attitudes are held in a given product category and that
advertising is able to influence these brand attitudes, the attitude must still be
accessed before it can influence brand choice (Kiselius & Stemthal, 1984; Biehal
& Chakravarti, 1985; Fazio, 1986). This is no trivial matter. People may not
retrieve their attitude towards a stimulus before they act. There is no guarantee







that there is anything, but the weakest of links between the representation of the
object and the representation of the attitude in memory (Langer, 1978; Fazio,
Powell & Herr, 1983; Bargh, 1984).
The likelihood that an advertising-generated brand attitude or any other
attitude will mediate choice is a function of its salience at the time a decision is
made. A study by Snyder and Swan (1976) demonstrates this point. Subjects
were asked to act as judges in a simulated sex discrimination case. In an earlier
task, subjects were asked to communicate their attitudes towards affirmative
action. Half the subjects in the experiment were asked to recall their attitude
towards affirmative action immediately prior to reading the case; the other half
were not. The correlation between subjects' views on affirmative action and their
case decisions was significantly higher among the subjects asked to review their
attitude towards affirmative action prior to reading the court case.
A primary mediator of attitude accessibility is repeated attitudinal
expression (Fazio, 1986). In one series of experiments, attitude accessibility was
demonstrated to be a direct function of the number of times subjects were asked to
verbalize an existing object attitude (Fazio, Chen, McDonel & Sherman, 1982).
Accessibility was measured using a response latency measure. In subsequent
experiments, Fazio found support for the potential of repeated attitudinal
expression to lead to the spontaneous activation of attitudes when the attitude
object is encountered (Fazio & Powell, 1984).
However, experiments by Fazio and Herr (1984) demonstrated that even
highly accessible attitudes do not necessarily directly or indirectly (e.g., bias
consumer decision processes) mediate evaluation. In one experiment, subjects
were told they were about to read a legal case concerning affirmative action. In a
2 x 2 design, half the subjects were asked to prepare to reach a verdict as they read
the case; half were told to memorize all the factual details, but were not asked to







reach a verdict. Also, half the subjects were asked to review their personal
attitudes towards affirmative action prior to reading the case, and half were not.
Attitudes towards affirmative action correlated strongly to the verdict only in
the cell in which prior attitudes were rehearsed and subjects were asked to prepare
to reach a verdict. In the memory condition, attitude-behavior correlations were
low regardless of attitude salience. Apparently, subjects found the specific
information more relevant to their decision than their attitudes on affirmative
action. The findings suggest that conscious or subconscious attitudinal
mediation of behavior is not a necessary outcome of attitude salience during the
decision process. This finding is highly consistent with the RAM perspective
that attitude, like any other element of a brand memory structure will only
mediate choice when it is both accessible and relevant.


Brand Attitude Relevance
Even if consumers hold brand attitudes in a product class, and even if these
attitudes can be influenced by advertising, and even if they are highly accessible
in memory, there is still no guarantee they will mediate choice at the brand
response occasion. The question thus remains when will attitudes formed at time
A mediate choice at time B? Assuming that the conditions above are met,
evidence indicates that the answer is when the attitude formed at time A is more
relevant, i.e., more diagnostic than other salient information at time B.
Much of the available research on the effects of prior judgments and attribute
information on evaluation or choice is not designed in such a manner that it can
specifically address the above question. A large body of evidence demonstrates
that memory-based judgments can dominate the evaluative implications of
recalled attribute information on subsequent evaluation (Lichtenstein & Srull,
1983; Lingle & Ostrom, 1979; Dreben, Fiske & Hastie, 1979). An equally large







body of evidence reveals a dominance of attribute information over previously
formed judgments (Reyes, Thompson & Bower, 1980; Biehal & Chakravarti, 1986;
Chattopadhyay, 1986). The research successfully demonstrates effects of
judgment and effects of attribute information, but does not specify clear boundary
conditions for these effects other than accessibility and does not offer a theoretical
framework to explain or predict when each effect is likely to occur.
Feldman and Lynch (1988) built a framework that is able to reconcile much of
the conflicting evidence in the literature. In the context of brand choice, Feldman
and Lynch propose that the likelihood that any cognition about a brand will be
used to choose that brand is a function of three factors: the accessibility of the
input in memory, the accessibility of alternative brand cognitions in memory and
the relative diagnosticities of the inputs. These conditions parallel the two
general axioms of the RAM regarding the potential of advertising generated
cognitions to mediate choice: superior accessibility and superior relevance.
Brand attitudes that are the output of deliberately formed summary
evaluations (i.e., cognitive elaboration) are likely to have an accessibility
advantage in memory over more specific brand information because they are likely
to be more efficiently stored in brand memory (Wright, 1980; Anderson, 1983;
Alba & Hutchinson, 1987). For this reason, they are more likely to become the
product of automatic processing (Bargh, 1984) and less likely to become the
victim of output interference (Hoch, 1984; Alba & Chattopadhyay, 1986). This
gives prior judgments a ceteris paribus accessibility advantage over other
information. In Lichtenstein and Srull (1983), both attribute information and
prior judgements were given an equal chance of mediating evaluation. In this
scenario, judgments formed at time A were accessed and used in favor of the
attribute information at time B. However, there was no reason for subjects to
make the effort to use attribute information in their evaluation at time B. The







evaluation environment was exactly the same at time A and time B; there were
no brands, no new brand information, no changes in the purported usage occasion,
etc. In reality, especially when comparing an advertising exposure occasion and a
brand choice occasion, there will almost always be differences. In an advertising
exposure occasion, memory for other brands is likely to be limited relative to the
brand choice occasion. Competitive information is also much more likely to be
highly accessible at the brand choice occasion.
If there is a change in the evaluation environment between the time of
attitude formation and the time of subsequent evaluation, evidence suggests that
specific performance information is more diagnostic than prior summary
evaluations and, hence, will be used in favor of prior judgments. Evidence
suggests that consumers perceive prior judgments to be non-diagnostic (or in
RAM terms, irrelevant) if new brands, attributes or usage situations are
introduced into subsequent evaluation contexts (Biehal & Chakravarti, 1986;
Lynch & Srull, 1982). Experiments designed to test Lynch and Feldman's basic
propositions have supported the framework (Lynch, Marmorstein & Weingold,
1988). In one experiment, subjects were put in "mixed choice" tasks where they
were forced to choose between a previously evaluated brand and a new brand.
Subjects relied on attribute information, and not the prior judgment to make the
new evaluation if the attribute information was accessible, ostensibly because the
prior judgment had no diagnostic implications in the new evaluation setting. A
follow-up study demonstrated that even when specific attribute information is
not accessible, subjects will use brand names as choice cues rather than rely on a
prior judgments perceived to have little diagnostic utility to the choice situation
(Lynch, Buzas & Marmorstein, 1987).







Brand Attitude Measures as Surrogates for Choice
Up to now, the discussion has focused on the likelihood and circumstances of
advertising generated brand attitudes mediating choice through their retrieval
and use as a choice cue. The discussion was largely theoretical. One may also
examine attitude behavior consistency from a more methodological point of view.
That is, one may agree that brand attitudes are not typically used as choice cues,
but one may also argue that measuring brand attitudes at time A can be highly
predictive of behavior at time B because a brand attitude measure simulates the
choice process.
The assertion that brand attitudes measured at time A can reliably predict
behavior at time B requires that two key assumptions be accepted. First, one must
assume that the attitude formation process and the brand choice process are
identical. The pool of information and the rules used to transform this
information into evaluations must be identical for the outcomes to be reliably
identical. Second, one must assume that the pool of available information, the
accessibility and relevance of this information, and the motivation to use that
information are unchanged from time A to time B. Evidence suggests that the
validity of these two assumptions is highly questionable.
Bettman (1979) demonstrated that many identified choice-based strategies
such as lexicographic, conjunctive and elimination by aspects do not require an
overall evaluation of brand alternatives during the choice process, but rather only
evaluation or recognition of specific performance attributes or quality cues.
Other research affirms the position that different informational sets and
evaluation strategies appear to drive attitude formation and choice decisions
(Johnson & Russo, 1981; Bettman & Zins, 1977). If this is true, there is no
reason to trust the ability of attitude measures to predict choice outcomes.







One major reason choice and judgment processes are likely to differ is a
decision involvement gap. One antecedent of decision-making involvement is the
perceived risk associated with the purchase (Robertson, 1976). There is little risk
associated with forming a "wrong" attitude when asked because there are no real
consequences associated with the evaluation. In many product classes, however,
there is considerable economic, social and utilitarian risk associated to a "wrong"
choice. This involvement gap between attitude and choice formation processes is
likely to lead to evaluations based on different informational inputs and different
decision rules. When this occurs attitude-behavior consistency should be low. For
this reason, survey measurement procedures that simulate the true "state of
mind" (e.g., involvement) at the time of choice are more predictive of choice
(Wright & Kriewall, 1980).
Assuming equal levels of involvement at the time of attitude formation and
choice and even identical evaluation processes, the correlation between brand
attitude measures at time A and brand behavior at time B is still likely to be low
unless involvement is very high. Momentarily activated cues have a
disproportionate influence over judgments made about an object or related
behaviors performed shortly after their activation (Wyer & Srull, 1986; Kiselius
& Sternthal, 1984). Environmental influences have a tremendous effect on
salience. Cognitions that influence evaluation are cued to memory by contextual
factors and only a small subset of the available information will be accessed at any
one evaluation episode (Feldman & Lynch, 1988). Since accessibility is a function
of the time since the most recent activation of that cognition (Wyer & Srull,
1986; Anderson, 1983), attitudes measured at or near the time of advertising
exposure are likely to be heavily biased by advertising content. Attitudes
measured in another environment are likely to be driven by the most salient
information in that environment.






Attitude-behavior consistency is likely to be particularly low when consumers
engage in low involvement decision-making behavior. In such a scenario, the
latitude of acceptance of alternative courses of action is likely to be high (Sherif,
Sherif, Nebergall, 1973) and evaluations are likely to be a function of highly
salient contextual cues. Assuming attitude measurement immediately follows
exposure to advertising, the salient contextual cues are likely to be advertising
effects. Assuming choice takes place in a store, the salient contextual cues are
likely to be packages and POP information for both the "target" brand and

competing brands. When involvement is high, attitude-behavior consistency
should be higher because evaluations are more dependent on deliberately accessed
rather than cued information.
Early critiques of attitude behavior consistency recognize that attitudinal
measures often do not capture the effects of "other" factors that influence choice
(Wicker, 1969; Ehrlich, 1969; Vinson et al., 1977). Although not explicitly
stated, their observations imply inconsistencies between attitude and choice
formation contexts and processes. Their remedy was to improve attitudinal
measures by better simulating choice processes at the time of measurement.


Summary

Evidence has been presented that questions the theoretical and
methodological bases for the primary role of brand attitudes in the study of
communication effects. There are several barriers to the consistent use of brand

attitudes as choice cues. First, brand attitudes may not be held towards brands in
many product categories. Second, in categories where brand attitudes are held,
they are more likely to be based on tangible brand experience rather than vicarious
experiences such as advertising. Third, if an advertising generated brand

attitude exists, it must be salient at the time of brand choice. Fourth, a salient







brand attitude must be perceived at least as relevant as other salient information
at the brand response occasion. The Feldman and Lynch (1988) framework
reconciles much of the conflicting research on the role of prior judgments in a
choice context. Its stated conditions for an effect of prior judgments on choice,
accessibility and diagnosticity, are quite consistent with the RAM perspective.
The ability of measures of brand attitude to accurately simulate future choice
and thus act as reliable measures of communication effectiveness is also
questionable. First, evidence suggests that attitude formation and choice
processes are not identical. Informational inputs and evaluation rules are likely
to differ. Second, unless involvement is very high, the effects of varying
environmental cues and time delays (or lack thereof) on information accessibility
in the communication context and the choice context are likely to lead to different
informational inputs to the brand attitude formation process and the choice
process. If this occurs, attitude-behavior consistency can be expected to be low.
These findings support two primary assertions of the RAM. First, affecting
brand attitude should be neither a primary theoretical nor a methodological goal
of advertising research. Second, advertising effectiveness needs to be measured at
the brand response occasion, not the advertising exposure occasion.




The Impact of Consumer Elaboration and Involvement on Information
Accessibility


Introduction
The preceding section of this chapter proposed that advertising typically may

not influence brand choice by modifying brand attitudes at the advertising
exposure occasion. Instead, advertising is presumed to work by affecting the







informational content of brand memory. Brand attitude formation is one possible
outcome of advertising exposure, but other outcomes such as the encoding of
affect, choice cues or performance beliefs can also lead to advertising effects at the
brand response occasion. In other words, brand attitude formation at the

advertising exposure occasion may be sufficient to mediate brand choice, but it is

not necessary.
Both the perspective that advertising functions by influencing brand
attitudes and the more general perspective that advertising functions by
affecting the structure of brand memory assume that information accessibility
moderates evaluation. The accessibility perspective forwarded in this thesis

builds on the availability-valence (AV) hypothesis (Kisielius & Sternthal,

1984). The AV hypothesis presumes that limited processing capacity inhibits
consumers from basing brand evaluations on all available information in brand
memory. Instead, brand evaluation is a function of that subset of information in
brand memory that is accessed at the time of brand evaluation. The net valence of
accessed information determines the favorableness of the judgement.
Two key issues concerning this perspective are the focus of this section. First,

what determines which information in brand memory will be most accessible at

the time an evaluation is made? Second, which highly accessible information will

be accessed and used by the consumer in making a brand choice? The first
question raises the issue of accessibility. The second question raises the issue of
relevance.

First, this section will discuss the effect of cognitive elaboration on
information accessibility. A large body of literature in both cognitive psychology
and consumer research establishes a direct relationship between elaboration and

accessibility, and between elaborated information and evaluation. Next, the issue
of consumer involvement will be raised. Two involvement constructs will be







discussed, i.e., advertising message involvement (AMI) and brand response
involvement (BRI). Advertising message involvement is expected to moderate
both the extent to which advertising information is elaborated and the type of
information that is attended. Brand response involvement will be introduced as a
construct that both moderates advertising message involvement and drives the
perceived relevance of advertising message information at the time of brand
evaluation. It will be proposed that both the relative accessibility and the relative
perceived relevance of a piece of information in memory affect whether it is used
or discounted at the time of evaluation. Finally, since anticipated brand response
involvement is only one of a number of antecedents of advertising message
involvement, the likelihood that a consumer's advertising message involvement
and brand response involvement do not always correspond will be discussed.
The implications of this section for the Relevance-Accessibility Model of
advertising effectiveness are as follows. First, advertising message involvement
is a primary predictor of the type of information that is most likely to be encoded
into memory at the advertising exposure occasion. Second, brand response
involvement is a primary predictor of the level of information consumers find
relevant and, hence, seek to discriminate brands at the brand response occasion.
Third, since the advertising exposure occasion and the brand response occasion are
usually distinct in time and space, and since AMI and BRI do not necessarily
correspond, models of consumer behavior must clearly delineate each of their roles.


Cognitive Elaboration
Elaborations that reflect the meaning an individual derives from a single
piece of information are more durable in memory than the message elements from
which these inferences are derived. Empirical research in psychology and
consumer research has demonstrated repeatedly that, over time, specific message






content decays in memory before cognitively elaborated summaries of the
information (Watts & McGuire, 1964; Watts, 1967; Alba & Ofir, 1984; Neisser,
1981; Frederickson, 1975). Frederickson (1975) demonstrated that elaborations
represent a higher proportion of total recalled information over time. Alba and
Ofir (1984) attained similar results in a marketing context. Examining the
temporal stability of message abstractions, they found that the proportion of
abstractions in a recall task increased over a delay of approximately one week.
Apparently, unelaborated message information decays significantly more quickly
than elaborated material.
The positive effects of elaboration on memory are enhanced in "crowded"
memory structures. Keller (1987) and Burke and Srull (1988) showed that recall
of brand claims is significantly reduced by the amount of competitive advertising.
Keller (1987) demonstrated that recall declines as competitive intensity
increases. Recall of a test brand's claims was higher with one as compared to
three advertised brands in the competitive mix. Unlike brand claims, however,
recall of cognitive responses was not affected by competitive advertising intensity.
A follow-up study by Keller (1990), showed interference effects on the recall
of cognitive responses, but only when there was competitive advertising for three
brands, not one or two. Interference effects remained much greater for brand
claims. Finally, when subjects were given a retrieval cue, recall of cognitive

responses did not differ between the zero and the three competitor conditions. The
cues had less impact on the recall of brand claims.
The cognitive response literature in consumer research blossomed after the
work of Wright (1973) and has become a foundation of models of attitude change.
The major premise of this literature is that cognitive responses evoked by a
communication are a key mediator of the direction and degree of attitude change
produced by the communication. Dozens of empirical studies have found







consistently that recipient modified and recipient generated responses to
communications account for significantly more attitude variance than recall of
message arguments. Apparently, message receivers rely on their own descriptive
or evaluative reactions to message content rather than the content itself to form
or modify brand based beliefs and attitudes (Wright 1973; Lutz & Swasy, 1977;
Petty & Cacioppo, 1979; Belch, 1981, 1982; Toy, 1982). As Wright (1980)
summarized, "Variations in people's verbalized thoughts often correlate
significantly with their post message attitude statements, especially when the
thought sampling is time-limited, the thought coding reliable and the subjects
are involved in message processing."
Conversely, researchers have been unable to find any reliable relationship
between message element recall and evaluation (Dreben, Fiske & Hastie, 1979;
Haskins, 1964; Watts, 1967; Ross, 1982). Keller (1990) found that subjects'
inability to recall brand claims did not influence target brand evaluation unless
cognitive response recall was also impaired.
Why do cognitive responses mediate brand evaluation? First, unlike simple
recall, cognitive responses add valence to a memory trace. Simple recall of a

persuasive argument or product benefit cannot reveal if an argument has been
rejected or accepted, or if a product benefit is perceived to add value to the product.
Cognitive responses can be loaded with valence. Second, cognitive responses are
more accessible than unelaborated information and have a greater probability of
becoming salient at the time of brand evaluation. Finally, cognitive responses
reflect an individual's reaction to those elements of the message that were worth
attending at the time of message exposure, ostensibly because they were perceived
to be more relevant to the person's viewing goals than other message elements.







Consumer Involvement
What determines which elements of a message will be attended and
elaborated? The likelihood that any given element of a message will be attended
and elaborated at the advertising exposure occasion and, thus, made more
accessible at the brand response occasion is a function of consumer involvement.
Involvement has been conceptualized as an activation level at a particular
moment, as opposed to a sequence of cognitive activities carried out subsequent to
message reception, or a predisposition to respond to a specific stimulus (Cohen,
1982). The key feature of this conceptualization and others is that involvement
is not confounded with its antecedents (e.g., novelty, intensity, an individual's
goals, beliefs, and interests) or its consequences (e.g., arousal, attention, cognitive
elaboration). By separating involvement from antecedent and consequent
variables, a cleaner conceptualization emerges, a straightforward building block
for theory development without excess baggage (Cohen, 1982).
Involvement per se is non-directional; it can be operationalized in terms of
total activation, but there are no tangible consequences on subsequent cognitive
activity until a directional indicator is added to the construct. It is the
consequences of the involvement by product interaction (i.e., decision-making

involvement) and the involvement by message interaction (e.g., advertising
message involvement) that are of immediate relevance to the Relevance-
Accessibility Model. Advertising message involvement and brand response
involvement are more specific involvement conceptualizations that include
direction. Their impact on advertising effects at the advertising exposure
occasion and brand response occasion can be studied.







Advertising Message Involvement
Early research on the effects of issue involvement on message processing
suggested that increased issue involvement leads to increased resistance to
persuasion (Sherif & Hovland, 1961). The justification for this explanation was
social judgement theory which postulates that high involvement narrows
individuals' "latitude of acceptance" towards persuasive messages thereby
increasing the probability that message arguments are counterargued or ignored
(Sherif, Sherif & Nebergal, 1965). As involvement decreases, social judgement
theory predicts that the latitude of acceptance widens increasing the likelihood of
persuasion.
Subsequent research in social psychology, however, demonstrated that high
issue involvement does not necessarily lead to increased resistance to persuasion.
New evidence indicated it can lead to greater persuasion (Eagly, 1967; Chaiken,
1979; Petty & Cacioppo, 1979). Two major findings emerged from this stream of
research. First, involvement increases the amount of thought consumers direct
towards the communication and its message arguments. Second, as involvement
increases persuasion becomes more dependent on the quality of arguments in the

message. If arguments are strong, positive attitude change occurs; if arguments
are weak, then negative change occurs.
Krugman (1965) first introduced the notion that consumer involvement with
the advertising message is linked to the production of cognitive responses or
elaboration evoked by the advertisement (Krugman, 1971; Mitchell, 1979, 1981;
Petty & Cacioppo, 1979, 1981, 1983; Batra & Ray, 1982, 1985). He conceptualized
involvement as "the number of conscious bridging experiences, connections, or
personal references per minute between the viewers life and the stimulus"
(Krugman, 1965, p. 128) and defined low involvement as the absence or near
absence of these bridging experiences.







Empirical research has demonstrated repeatedly that unless a message
receiver is in a high involvement processing mode, cognitive responses (1) will be
few and (2) will be weakly related or unrelated to attitude towards the message
object. Subjects' message relevant thoughts correlate more with brand attitudes
under conditions of high rather than low involvement (Chaiken, 1979; Wright,
1973, 1980, Cacioppo & Petty, 1979; Petty & Cacioppo, 1981; Petty et al. 1983).
Petty and Cacioppo (1979) demonstrated that persuasive effects are more
related to message argument cogency under conditions of high rather than low
involvement. In the high involvement condition, significantly more
counterarguments than support arguments were generated upon exposure to weak
message arguments; the opposite occurred during exposure to strong arguments.
In the low involvement condition, neither support or counterarguments
production was affected by argument quality. Message relevant thoughts were
significantly better predictors of attitude change in the high involvement
condition; message recall was not related to attitude change in either
involvement condition. Petty and Cacioppo's findings in this and other related
experiments led to the conceptualization of the Elaboration Likelihood Model
(ELM), which will be discussed later in this section.
Mitchell, Gardner and Russo (1980) replicated Petty and Cacioppo's (1979)
findings. Rather than indirectly affect advertising message involvement by
manipulating brand response involvement, however, they specifically instructed
their subjects to either focus their attention on the brand related message
arguments (brand processing condition) or on executional elements of advertising
(nonbrand advertising). More cognitive responses were elicited in the brand
processing condition, and cognitive responses in the nonbrand condition did not
explain attitude differences as well as they did in the brand processing condition.







Oualitative Effects of Advertising Message Involvement. As advertising
message involvement increases, information is better comprehended, better
integrated into brand or product class memory structures, and, is more accessible.
Below is a brief summary of research examining the consequences of differential
involvement during message reception. Some researchers discuss effects in terms
of the information that is stored in memory as a result of processing. Others
speak in terms of the type of advertising information that is subsequently
accessible and potentially able to mediate brand attitudes.
Despite the greater attentional focus on specific message arguments, message
comprehension and information integration associated with high advertising
involvement, low involvement processing can modify brand memory networks
sufficiently to affect brand evaluation. Krugman (1967, 1971, 1980) speculated
that low involvement effects are typically visual and processed predominantly in
the right brain. As a result, they can best be measured through recognition
measures. High involvement effects are verbal and processed predominantly
through the left brain. He also proposed that since consumers are basically
satisficers and act on the basis of familiarity, the subtle perceptual shifts created
by low involvement processing are sufficient to change behavior. Although
Krugman's work lacks a strong theoretical framework, his assertions have
intuitive appeal. Many of his notions on the potential of persuasion without

cognitive elaboration are consistent with subsequent work on automatic
processing in cognitive psychology (Anderson, 1983; Bargh, 1984) and the rebirth
of behaviorism as an appropriate framework for investigating low involvement
behavior (Kroeber-Riel, 1979; Gorn, 1983; Nord & Peter, 1983). Persuasion
without cognitive elaboration will be examined in more detail in the next section

of this chapter.







In their treatise on the structure of knowledge in memory, Alba and
Hutchinson (1987) make a distinction between analytic and holistic processing.
Analytic processing is characterized by highly deliberate learning employing
complex schemas containing specific rules to make informational classifications
and brand evaluations. Holistic processing is characterized by less intentional,
even automatic, processing that makes classifications and brand evaluations
through similarity matching and the use of of category exemplars. Analytic
processing is characteristic of highly involved experts, while holistic processing is
characteristic of novices who do not have the necessary motivation, ability and/or
opportunity to engage in analytic processing.
Others have described low and high involvement effects in terms of (a) the
activation of memory structures during processing or (b) the effect of processing
on memory structures. Mitchell (1981) associated high involvement with
significant schema activation to integrate consumers' subjective interpretation of
external information into internal memory and low involvement with minimal or
no such schema activation. Gardner, Mitchell and Russo (1981) speculated that
information processing under low involvement may be stored in episodic memory
rather than a product class or brand semantic memory network. The inability of
low involvement processing to affect semantic memory was expected to decrease
the accessibility of attended information at a later time. This position is
supported by Thorson (Cites).
In the research introducing the Attitude Toward the Ad construct, Mitchell
& Olson (1981) produced evidence that when processing is impaired, affective
reactions evoked by visual executional elements can lead to brand evaluations as
extreme as attribute oriented message arguments. Given the relative absence of
motivation to integrate information into brand or product class memory
structures under conditions of low involvement, and the potential for incidental







learning of easily processed information, Batra & Ray (1981) prescribed molar
strategic goals based on a high-low involvement dichotomy. Under high
involvement, advertising should contain strong arguments about salient
attributes. Given the deliberate nature of processing, light repetition schedules
are sufficient. In low involvement, advertising should induce awareness-based
affect through heavy repetition.
Leavitt, Greenwald and Obermiller (1981) catalogued a number of reasons
why heavily repeated advertising processed at a low level of involvement can lead
to persuasive effects without cognitive elaboration. First, reduced
counterargumentation increases the likelihood that messages will "seep" into
brand memory without counterargumentation and, hence, carry a positive valence
(See Wright, 1973). Second, heavy repetition schedules can link a brand to a
usage situation to the point where activation of the usage situation leads to the
activation of the brand, thus increasing the probability the brand will be included
in consumers' evoked sets in that situation. Third, heavy repetition can also
change the relative accessibility of brand attributes used to judge brand
alternatives and thus affect the decision rules used to discriminate brands;
Wright and Rip (1980) called this problem framing. Finally, the massive
advertising schedules designed to build brand name or package familiarity can
associate positive positive feelings to brands through the exposure effect (Zajonc,
1980; Obermiller, 1985). Another means by which positive feelings can become
associated to a brand that was not discussed by Leavitt et al. (1981) is affect
transfer or what others call affective classical conditioning (Kroeber-Riel, 1979;
Gor, 1983; Baker, 1985).
Involvement as a Continuum. Until the mid 1980s, most involvement
research dichotomized the construct; involvement effects were characterized as
high or low. Definitions of high advertising involvement tended to be fairly







consistent; however, conceptualizations of low advertising involvement varied.
Some researchers characterized the consequences of low advertising message
involvement as processing devoid of an significant deliberate processing of the
advertising stimulus (Krugman, 1965, 1971; Kroeber-Riel, 1979). Others
characterized it only as an absence of detailed processing of specific message
arguments (Mitchell, Gardner & Russo, 1981; Petty & Cacioppo, 1981, 1983). If
one was to think of involvement as a continuum rather that a dichotomy, it seems
that the former conceptualization would be at the low end of that continuum
while the latter would be somewhere in the middle.
Before the effects of involvement at the time of message reception can be
studied systematically, a precise specification of the consequences of different
levels of advertising message involvement on message processing is needed.
Approaching this issue from the perspective of a network model of memory and
the concept of spreading activation, Greenwald and Leavitt (1984) categorized
advertising message involvement (AMI) into four categories along the
involvement continuum. The categories are defined in terms of the extent of
activation of internal memory structures and the corresponding level of meaning
that is extracted and encoded into the brand concept and/or brand response
schema. Examination of this typology suggests a correspondence between three of
the four categories and the three levels of information identified in the relevance-

accessibility model.
Greenwald and Leavitt (1984) term the lowest involvement category as
preattentive analysis. Preattentive analysis involves the monitoring of
background stimulation for the occurrence of novel, intense or goal-relevant
events. If these events are perceived, then higher levels of processing are invoked.

To date, there is no strong evidence that encoding at preattentive analysis can
lead to any type of persuasive effect. However, advocates of subliminal persuasion







or the "mere exposure effect" (Zajonc, 1968, Moreland & Zajonc, 1977, 1979;
Wilson, 1979) contest this position. Evidence for both of these effects is currently
tenuous (Moore, 1982; Bimbaum & Mellers, 1979; Obermiller, 1985), but
certainly cannot be ruled out. It is conceivable that affective reactions can be
influenced by information that impinges on the nervous system, but does not
penetrate the cognitive system. Inconsistencies between avowed attitudes and
behavior, discrepancies between verbal and nonverbal information emanating from
the same individual (Ekman & Friedman, 1969) and influences on human
behavior that cannot be reported accurately by the cognitive system (Nisbett &
Wilson, 1977) have all been offered as evidence for this non-cognitive path. Until
evidence becomes more concrete, however, this investigation will be limited to the
cognitive system.
The second category of advertising message involvement, focal attention,
entails sufficient deliberate attention to identify the stimulus at the sensory
level. In essence, focal attention leads to the activation of a stimulus' identifying
node and other nodes that are automatically associated to that node in memory
(Schiffrin & Scheider, 1977); however, it implies no deliberate cognitive
activation of memory networks to comprehend the stimulus nor deliberate
cognitive elaboration to describe the stimulus. Familiar objects are semantically
identified and associates are activated automatically. The level of cognitive
activity required to generate propositional facts about the stimulus is not present.
Therefore, unfamiliar stimuli establish sensory traces, but no semantic
elaboration Since stored affective responses to stimuli are efficiently encoded in
memory and, as a result, are more subject to the effects of automatic encoding and
retrieval than other levels of information (Bower, 1981; Bargh, 1984), it follows
that focal attention processing can lead to the sufficient encoding of pure affect







either through a familiarity based mechanism (i.e., the exposure effect) or an
associative mechanism (i.e., affective conditioning).
The third level of advertising message involvement is termed comprehension.
Comprehension involves the deliberate activation of memory structures required
to comprehend the global meaning of the advertising message as opposed to simply
identifying its elements. Semantic analysis occurs, absolute abstractions of
meaning may be generated, and evaluative inferences may be made on the basis of
message content or absolute abstractions. There is, however, no deliberate
integration of message-based propositions with brand associates currently in the
brand concept or with brand associates in other brand concepts. As a result,
absolute rather than relative brand meaning is extracted from the message unless
the message explicitly is communicating relative information. Since the content
of the message is not deliberately integrated with either the brand concept or the
brand response schema, heavy repetition is required to facilitate sufficient
semantic encoding.
Comprehension level processing is expected to lead to the encoding of absolute
brand beliefs and message based quality cues such as source credibility, the
number of message arguments, the global evaluation of the stimulus (e.g.,

Attitude Toward the Ad), etc. It is not expected to typically be sufficient to
generate relative performance beliefs, because this form of associate requires
integration at both the brand concept and brand response schema levels.
Comprehension level associates are expected to be sufficient to derive absolute
evaluations of the brand from a single inference or to provide input into non-

compensatory decision strategies (Bettman & Park, 1980; Biehal & Chakravarti,
1982, 1985), but insufficient to act as reliable input into sophisticated
compensatory decision strategies involving direct competitive comparisons along







multiple performance dimensions (Lutz, 1975, 1977; Bettman, Capon, & Lutz,
1974).
The final advertising message involvement category is cognitive elaboration.
It is at this level that message information is explicitly integrated with semantic
brand memory and evaluative schemas. Message information is not only
processed, it is integrated with prior knowledge. Since information is integrated,
memory traces are more durable (Alba & Hutchinson, 1987) and do not require
massive repetition to become accessible (Anderson & Rivolli, 1984). It is at this
level of involvement that message arguments must be perceived to be relevant and
strong relative to prior knowledge to be persuasive (Petty & Cacioppo, 1981).
Relative performance information is expected to be the primary encoding goal
because consumers are seeking the most reliable, concrete evidence available. This
information is sought and, if present and meaningful, it is expected to receive the
most processing during exposure.


The Principle of Higher Level Dominance

Pure accessibility theses such as the availability-valence hypothesis
(Kiselius & Sternthal, 1984) predict that all balanced information salient at the
time of brand evaluation affects brand evaluation. Greenwald and Leavitt (1984)
introduced the principle of higher level dominance to predict the level of
advertising information that is most likely to mediate brand attitude formation
or change after an advertising exposure. The principle predicts the domination of
higher level information over lower level on the basis of relative accessibility.
Every successive level of advertising message involvement in the Greenwald
and Leavitt typology is linked to a higher level of information: "....the effects
associated with the highest level of analysis applied to a message should be
dominant. For example, among attitudinal effects, affective conditioning which







is associated with focal attention should be outweighed by message-based
persuasion effects, which should in turn be dominated by cognitive response based
persuasion effects" (Greenwald & Leavitt, 1984, pg. 588). Furthermore, each
level of analysis is momentarily passed through to reach the next level. Before a
consumer reaches comprehension level processing, he/she must pass through
preattentive analysis and focal attention. Effects of the highest level of analysis
applied to a message are expected to dominate any subsequent evaluation because
they are more accessible than any effects encoded at lower levels of analysis.
Effects at higher levels of analysis are also more accessible because each successive
level involves more deliberate processing and integration into brand memory
structure than the preceding level.
Essentially, the principle of higher level dominance presumes that once
encoded, advertising effects are driven by relative accessibility. The most
accessible advertising generated information in memory will mediate any
advertising effect on brand evaluation. The principle allows for relevance effects,
but these are presumed to occur at the time of encoding, not brand evaluation. As
Burnkrant and Sawyer (1981) state, the perceived relevance or meaningfulness of
a piece of advertising information at the advertising exposure occasion is a major
determinant of the level of processing that the information receives. Relevant
information can motivate higher advertising message involvement, while
irrelevance leads to lower involvement.
The principle of higher level dominance examines advertising effects in the
context of the advertising exposure occasion on the basis of a single exposure to
advertising. In this scenario, it makes sense. When advertising effects on brand
evaluation are studied in the context of a brand response occasion distinct in time

and space from the advertising exposure occasion and in scenarios involving
multiple exposures to advertising, the premise begins to break down.







There are two points of contention with the principle of higher level
dominance in the context of the brand response occasion and multiple advertising
exposures. First, there is no strong evidence that higher level information in
memory will always or even typically be more accessible than lower level
information. In certain scenarios, lower level advertising effects are likely to be
more accessible than higher level effects. For example, simple affective
associations evoked by the exposure effect (Zajonc, 1980) or affective conditioning
are potentially products of automatic processing after many exposures. Effects of
higher level information such as relative performance beliefs, due to their more
complex structure in memory, are less likely to become subject to automatic
processing effects (Bower, 1981; Bargh, 1984). Also, when one considers the high
accessibility of lower level non-advertising effects (e.g., packaging, POP, etc.) at
the brand response occasion, it is almost a certainty that higher level advertising
effects must compete with lower level effects more accessible than the advertising.
Second, there is no evidence to suggest that higher level information, if more
accessible than lower level information, is always perceived to be more relevant to
the evaluation task at hand. Highly accessible higher level information will not
always dominate lower level information at the time of brand evaluation because
consumers may believe it is not worth the effort to use higher level information in
the evaluation. Consumers may have the ability and opportunity to meticulously
compare brand alternatives across several performance dimensions, but may only
be motivated to make the evaluation on the basis of an accessible quality cue such
as source credibility (Robertson, 1976; Petty & Cacioppo, 1981).
It is the position of this thesis that accessibility is a necessary but
insufficient prerequisite for advertising effects to mediate brand evaluation.
Highly accessible information can be accessed and discounted at the brand
response occasion if it is not perceived to be relevant (Feldman & Lynch, 1988).






The dominance of higher level information over lower level information is not
possible if it is not accessed, but ultimately its use over lower level information is
also dependent on its perceived relevance relative to highly accessible lower level
information.
Evidence from an experiment testing the relative mediating strength of three
hypothetical constructs supports the notion of informational dominance through
an accessibility and relevance mechanism rather than simply an accessibility
mechanism (Baker, 1985). In this experiment, higher level information was
dominant; however the results suggested that the dominance was not the result of
accessibility, but because accessible lower order information was deliberately
discounted.
In the experiment, subjects were exposed to a series of wrist watch
advertisements for a test brand and competitors. Attitude toward the ad and
brand name familiarity were experimentally manipulated. Each ad also
contained brand attribute information. Brand attitude measures were taken one
day and one week after exposure in a between subjects design. Recall of brand
attribute information and advertising execution information was also measured.
Significant effects of brand name familiarity on purchase intention occurred, but

only after a week's delay from exposure. Interestingly, the significant effects of
relative brand name familiarity (i.e., a measure of subjective familiarity relative

to competitors designed to capture the exposure effect) occurred only when (1) the
accessibility of advertisement execution information and brand attribute
information were at their lowest levels and (2) relative brand name familiarity
was at its highest level. If simple accessibility drove results, brand name
familiarity should have mediated brand attitudes in both the short and long delay

conditions. Instead, it mediated evaluation only when higher level information

(e.g., executional quality cues and relative performance information) was no







longer accessible. The same outcome held true for Attitude Toward the Ad. It
mediated brand evaluation only when executional recall levels were high and
brand attribute recall levels were low.


Brand Response Involvement Mediates Informational Relevance

Most research on advertising message involvement focuses on its
consequences rather than its causes. Message processing depth or direction is
manipulated and consequences are analyzed (Gardner, Mitchell & Russo, 1983;
Batra & Ray, 1985). Other research begins with the brand response occasion and
work backwards to predict how involvement with brand evaluation or choice is
likely to affect advertising message involvement (Petty, Cacioppo & Schumann,

1983). This latter perspective is more complete because it focuses on both causes
and consequences of advertising message involvement. Effects of advertising

message involvement can be better understood if they are studied in the context of
the construct's primary antecedents. One of the major antecedents of advertising
message involvement is expected involvement with the brand evaluation or brand
choice, termed brand response involvement.

Evidence indicates that brand response involvement is a primary

determinant of the level of information that consumers perceive to be relevant at

the time of brand evaluation (Petty & Cacioppo, 1981; Robertson, 1976). That is,

brand response involvement determines the level of information that consumers
seek to discriminate brand alternatives. The key implication for the affect of
brand response involvement on brand choice is that relevant information, i.e.,
information that consumers prefer to use to discriminate brand alternatives, is

not necessarily the highest level of information that is accessible at the time of
evaluation.







The idea that people do not always prefer to make a brand choice by
meticulously comparing brand alternatives across a series of performance
dimensions is both intuitively appealing and heavily supported in the psychology
and consumer research literature. People are cognitive misers; the degree of effort
put into a brand choice decision is a function of the expected benefit of the effort
(Wyer, 1986). If there is no perceived benefit to be gained from a highly deliberate
information search involving specific comparisons of brand alternatives, then the
decision process will be simplified (Bettman, 1973; Kassarjian, 1977; Olshavsky &
Granbois, 1973; Nord & Peter, 1980).
Brand response involvement is a motivational construct embodying the
degree of cognitive effort expended by the consumer in any given brand evaluation
process. Involvement with a brand evaluation is strongly related to the perceived
risk associated with that decision (Bauer, 1967; Assael. 1983; Kapferer &
Laurent, 1984). Cox (1967) identified two components of perceived risk, amount
at stake and uncertainty. Uncertainty or what may be termed as risk probability
is the likelihood of making an incorrect or non-optimal brand choice. It is a direct
function of the perceived similarity between considered brand alternatives on
utilitarian, economic and social dimensions (Bettman, 1973; Ray et al. 1973;
Robertson, 1976; Hugstad & Taylor, 1979). Amount at stake relates to the
consequences that are likely to result from an incorrect or non-optimal brand
decision. Consequences may be social, economic, or utilitarian. Bloch and
Richins (1983) described this dimension as risk importance and noted that this
dimension should not just be thought of as loss potential, but also the potential to
gain by a correct or optimal decision.
Despite the widespread agreement that risk probability and risk importance
are fundamental predictors of brand response involvement, there is no strong
theoretical framework interrelating the various antecedents of brand response







involvement. Some researchers have laid the groundwork for such frameworks
(See Kapferer & Laurent, 1985 & Muhlbacher, 1985). However, these efforts have
done a better job of listing, justifying and empirically relating a set of antecedents
to various types of involvement than they have at specifically modelling their
causal effect on involvement and on one another. Research is needed to develop a
framework that includes the two components of risk and other major antecedents
such as product class interest.
Levels of brand response involvement have been linked explicitly to specific
types of brand decision strategies and advertising processing strategies. That is,
BRI is a primary predictor of both the type of information consumers use at the
brand response occasion to discriminate brands and the level of advertising
message involvement consumers at which consumers will process advertising.
Robertson (1976) used the terms low and high commitment consumer
behavior rather than low and high involvement. He related high commitment
behavior to Ray's learning hierarchy and low commitment behavior to his low
involvement hierarchy. Ray's data indicated that the low involvement hierarchy
was dominant (Ray et al.., 1973); others agreed (Kassarjian, 1977; Olshavsky &
Granbois, 1973; Nord & Peter, 1980).
Robertson argued that commitment is maximized when consumers perceive a
high number of "distinguishing" attributes among brands and a level of "salience"
attached to these attributes. In a high commitment situation, consumers are
active, information seeking and resistant to influence. Strong beliefs are held
about brands, particularly the preferred brand. Advertising is typically a weak
influence; it is received in an active mode, but is typically screened out or
counterargued.
In a low commitment situation, few beliefs are held about specific brands.
There is low cognitive resistance to advertising: "It is not worth the individual's







exercise of cognitive energy to reason on incidental matters of consumption"
(Robertson, 1976, pg 20). The classic hierarchy of effects is collapsed to awareness
followed by trial. Advertising can induce trial, which is the main means of
information evaluation (See Smith & Swinyard, 1983). Since advertising is
received in a passive mode, the weight of advertising rather than its content may
be the key to sales. To be effective, repetition levels must be high. Decision-
making is characterized by minimal cognitive effort.
Rothschild (1979) made a distinction between low involvement and very low
involvement. He characterized low involvement by the use of a few heavily
weighted attributes in a non-compensatory manner with satisficing, not
optimizing, the decision goal. In extreme low involvement, he postulated that
familiarity alone drives evaluation. High involvement decision-making is
characterized by a large set of attributes combined in a compensatory manner.
Information is centrally processed to arrive at a choice.
Brand response involvement is postulated to be a mediator of advertising
message involvement. Empirically, it has been used to manipulate advertising
message involvement (Petty, Cacioppo & Schumann, 1983; Chaiken, 1979). As
stated by Shimp (1981), advertising message receivers are resistant to factual,
attribute oriented ads in product classes perceived to have no real differences
among brand alternatives, presumably because in these cases there should be little
or no motivation to process message information. Advertisements that generate
easily processed affective associations to the brand or which emphasize global
evaluations are likely to be more effective in these cases.
A large body of research concludes that high brand response involvement
leads to high advertising message involvement which in turn leads to
communication based evaluations mediated by the perceived strength of message
arguments. Low brand response involvement invokes low advertising message







involvement and message based evaluations mediated by quality or other cues.
When involvement is low, brand evaluation has been mediated by source
credibility (Chaiken, 1980; Petty, Cacioppo, & Schumann, 1983), Attitude
Toward the Ad (Shimp, 1981; Lutz, MacKenzie, & Belch, 1982), affect transfer
(Gorn, 1982), brand name familiarity (Obermiller, 1985) and recognition of the
number (not content) of message arguments (Petty & Cacioppo, 1984; Alba,
1985). In all of these studies, when message processing involvement was high,
evaluation was increasingly a function of the interpretation of performance
related information.
The Elaboration Likelihood Model (ELM) is perhaps the best known
representation of this perspective. The model identifies two general routes to
persuasion (Petty & Cacioppo, 1981; Petty, Cacioppo, & Schumann, 1983). The
critical "switching mechanism" in the ELM model is issue involvement. When
consumers are involved in the decision process they are active information
processors and employ what is termed the "central route." Persuasion is a
function of the active interpretation of message elements. When consumers are
not involved in the decision process they are passive information processors and
follow the "peripheral route." Persuasion is a function of persuasion cues and
principles of behavioral learning theory.
Under the central route, message recipients attend to advertisement message
arguments, cognitively interpret them and derive their brand evaluations on the
basis of these interpretations. In essence, persuasion is dependent on the
interpretation and integration of these arguments with existing brand knowledge.
If messages are perceived to be strong, persuasion will occur. If messages are
perceived to be weak, persuasion will not occur. This is highly consistent with the
cognitive response literature which predicts that persuasion is a function of the







valence of recipient generated, message-related cognitive responses elicited during
message exposure (Greenwald, 1968; Wright, 1973, 1980; Lutz & Swasy, 1977).
Under the peripheral route, persuasion is determined by persuasion cues.
Persuasion cues are defined as "....any factors or motives inherent in the
persuasion setting that are sufficient to produce an initial attitude change
without any active thinking about the attributes of the issue or object under
consideration" (Petty & Cacioppo, 1981, pg. 158). Peripheral process effects do
not rule out active thinking in the persuasion process, only active thinking about
the object or its attributes. Peripheral effects, then, can range from active
consideration of non-message elements to completely passive effects such as
classical conditioning. Examples of active peripheral effects include evaluative
inferences generated by a message source (Sternthal, Dholokia, & Leavitt, 1978;
Chaiken, 1980) or recall of the number (not interpretation) of message arguments
communicated in the advertisement (Petty & Cacioppo, 1984; Alba, 1985).
Examples of purely passive effects include effects of music (Gorn, 1982; Milliman,
1978) or emotion (Kroeber-Riel, 1984; Ross & Ross, 1975) and are analogous to
classical conditioning.
Some empirical studies do not support the ELM perspective that persuasion
follows either the central or peripheral route. Instead, they tend to support a dual
mediation of both routes when involvement is high and the dominance of the
peripheral route when involvement is low (MacKenzie, Lutz & Belch, 1986; Park
& Young, 1984; Gardner, 1986). In defense of the ELM, most of these studies did
not directly manipulate brand response involvement; they manipulated the
direction of consumer processing through instructions to either attend brand
related message arguments or executional elements of the advertising.
However, there appears to be some problems with the mutual exclusivity of
the central and peripheral routes. A limitation of the ELM is that it does not







clearly discriminate evaluation processes involving deliberate cognitive effort
from those which do not, nor does it clearly link different types of information
with the two persuasion routes. The ELM's peripheral route embraces levels of
cognitive effort ranging from total passivity (e.g., classical conditioning effects)
to the highly deliberate consideration of executional elements of the advertising.
The ELM's definition of peripheral effects presents a real problem in an
advertising context. In advertising, executional (i.e., peripheral) elements are
likely to contain hedonic attribute information (e.g., lifestyle congruency, image
enhancement, reference group appeal). Thus, peripheral processing as defined by
the ELM can include effects of relative performance beliefs, quality cues or simple
affect.


Involvement Types Do Not Necessarily Correspond
Advertising effects research has never explicitly considered brand response
involvement and advertising message involvement separately. There has been an
implicit assumption that the two correspond. However, if one considers that
advertising message involvement's antecedents include, but are not limited, to
brand response involvement, it becomes reasonable to assume that the two do not
necessarily correspond.
Several antecedents of advertising message involvement have been identified
other than expected brand response involvement. The first three antecedents are
likely to lower advertising message involvement relative to brand response
involvement. The first is the need for information in decision-making
(Burnkrant & Sawyer, 1983; Petty & Cacioppo, 1981; Petty, Cacioppo, &
Schumann, 1983; Batra & Ray, 1982, 1985; Bettman, 1979). Consumers may
expect to be highly involved at the time of brand choice, but may believe they do
not need advertising information to aid their decision-making process. The







second is the perceived relevance of information to the specific decision task (Toy,
1982; Brock & Albert, 1978; Ostrom & Lingle, 1979; Loken & Hoverstad, 1985).
Expected brand response involvement may be high at the time of advertising
exposure, but the advertising message may not contain the type of benefit
information consumers perceive at that time to be important to their decision.
The third is the presence of competing processing goals at the time of advertising
exposure (Wright, 1980; Petty & Cacioppo, 1981). Advertising message
involvement is not likely to be high if consumers are engaging in conversation,
reading or snacking during advertising exposure.
The following two antecedents may either facilitate or hinder advertising
message involvement relative to brand response involvement. The first
antecedent is the novelty or distinctiveness of the advertising stimuli. Novel
stimuli are better attended and more deeply processed than familiar stimuli
(Taylor & Fiske, 1978; Eysenck, 1979; Berlyne, 1970). If the novel elements of the
advertising stimulus relate to the message, then message processing should be
facilitated; it they do not, then message processing is likely to be hindered. The
second antecedent is the ability of the communication to evoke a strong emotional
response. Stimuli that evoke extreme emotional responses, either positive or
negative, are likely to be more efficiently encoded into memory than stimuli that
produce little or no arousal (Silk & Vavra, 1974; Bolles, 1988; Mandler, 1984;
Nilsson, 1984; Bower, 1981; Anderson, 1983). Again, affect generating stimuli
relating to the message should facilitate message encoding; affective stimuli
unrelated to the message will distract message processing.


Summary
Advertising message involvement is a primary predictor of the type of
information that will be efficiently encoded into memory at the advertising