Title: Consumers' beliefs in product benefits
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00100793/00001
 Material Information
Title: Consumers' beliefs in product benefits the effect of irrelevant product information
Physical Description: Book
Language: English
Creator: Meyvis, Tom, 1972-
Publisher: University of Florida
Place of Publication: Gainesville Fla
Gainesville, Fla
Publication Date: 2001
Copyright Date: 2001
 Subjects
Subject: Marketing thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Marketing -- UF   ( lcsh )
Genre: government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )
 Notes
Summary: ABSTRACT: When consumers try to assess the performance of a product on a key benefit, they often encounter irrelevant product information while searching for diagnostic information. Normatively, this irrelevant information should not influence consumers' belief in the product's ability to deliver the benefit. However, this dissertation presents evidence that adding irrelevant product information to supportive benefit information systematically weakens consumers' belief that the product will deliver the desired benefit. This dilution effect persists when subjects are forced to acknowledge the irrelevance of the additional information prior to stating their beliefs, when the information is allegedly randomly selected by a computer, and when the irrelevant information increases the similarity of the product description to the typical desired product. Furthermore, the irrelevant information does not dilute the impact of the supportive information, but instead has a direct, independent effect on consumers' beliefs in the benefit. These findings are inconsistent with previous accounts of the dilution effect. The diluting effect of irrelevant information observed in these studies is not caused by an averaging strategy, a misguided reliance on conversational norms, or the use of a representativeness heuristic. Instead, I suggest that this dilution effect is caused by consumers' reliance on a biased hypothesis testing strategy.
Summary: ABSTRACT (cont.): It is proposed that consumers test the hypothesis that the product will deliver the benefit, selectively search for information that confirms this hypothesis, and categorize all other information (be it irrelevant or disconfirming) as "not confirming." When information is classified as "not confirming," it weakens belief in the hypothesis, even when it does not confirm the alternative hypothesis. As a consequence, irrelevant information weakens consumers' beliefs in the product's ability to deliver the benefit. Consistent with this explanation, the dilution effect is shown to disappear when consumers first process the information without the benefit in mind, and when consumers are forced to consider both the focal hypothesis and the alternative hypothesis. Moreover, the dilution effect even reverses when the product description concerns a brand with a poor reputation, suggesting that, in this case, consumers may be testing the hypothesis that the product will not deliver the benefit.
Summary: KEYWORDS: dilution, irrelevant information, consumer decisions, product benefits, hypothesis testing
Thesis: Thesis (Ph. D.)--University of Florida, 2001.
Bibliography: Includes bibliographical references (p. 87-92).
System Details: System requirements: World Wide Web browser and PDF reader.
System Details: Mode of access: World Wide Web.
Statement of Responsibility: by Tom Meyvis.
General Note: Title from first page of PDF file.
General Note: Document formatted into pages; contains vii, 112 p.; also contains graphics.
General Note: Vita.
 Record Information
Bibliographic ID: UF00100793
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 49876760
alephbibnum - 002729355
notis - ANK7119

Downloads

This item has the following downloads:

diss2 ( PDF )


Full Text













CONSUMERS' BELIEFS IN PRODUCT BENEFITS:
THE EFFECT OF IRRELEVANT PRODUCT INFORMATION








By
TOM MEYVIS

















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


2001















ACKNOWLEDGEMENTS


This dissertation has benefited, directly and indirectly, from interactions with

many people. I would like to thank these persons for their discussions, comments, and

seminars, and for supporting and encouraging my research. First of all, I would like to

thank Chris Janiszewski, my dissertation chair, for the many discussions, for his

mentorship in general, and for making this entire process so much fun. This dissertation

has also benefited from the extensive comments I have received from Joe Alba and Alan

Cooke and from discussions with the other members of the dissertation committee: Steve

Shugan and Ira Fischler. I would also like to thank all the students, former students, and

faculty at the University of Florida, both for stimulating discussions and for guiding me

through the Ph.D. experience, especially Amitav Chakravarti, Michel Pham, John

Pracejus, Stijn van Osselear, Joel Cohen, Barry Schlenker, Joffre Swait, Bart Weitz, and

Jinhong Xie. This research has also improved significantly thanks to comments from

participants in interviews at the 2000 AMA conference in Chicago and in seminars at

INSEAD, MIT, Harvard University, UCLA, the University of Colorado, Stanford

University, Duke University, the University of Pennsylvania, the University of Chicago,

Northwestern University, Columbia University, New York University, UC-Berkeley, and

Tilburg University. I especially would like to thank Bob Wyer and John Lynch for their

very helpful comments.










In a less direct, yet not less important, manner, this dissertation has also benefited

from the continuing support I have received from my family and parents. I would

especially like to convey my gratitude to my parents for supporting me throughout my

education and for always giving me the freedom to explore my interests, even though

these interests tended to change rather unpredictably. Finally, I would like to thank Els

for bringing so much fun to my life that I almost forgot to write this dissertation.















TABLE OF CONTENTS

page

A C K N O W LED G EM E N T S ................................................................................ ii

ABSTRACT ......................................................... vi

CHAPTERS

1 IN TR OD U CTION ..................................... ............................. 1

The Effect of Irrelevant Inform ation .................................................................. 2
Overview ........................................ 4

2 THEORETICAL BACKGROUND .................................. .......... 6

Irrelevant Product Information in the Consumer Research Literature ............... 6
The Dilution Effect ................................................ .... ..... 9
Existing Accounts of the Dilution Effect ................ ........... ... ....... 10
A Biased Hypothesis Testing Account ........................... ....... 15

3 EXPERIMENT 1: DEMONSTRATION OF THE DILUTION EFFECT .......... 24

M e th o d .................................................................................................................. 2 6
Results .................................................. 30
D discussion .................................................. 31
E x p erim ent 1A .............................................................................. 3 2

4 EXPERIMENT 2: A TEST OF THE AVERAGING ACCOUNT .................. 36

M e th o d ................................................................................................................. 3 9
R results and D discussion ............................ ................................................... ......... 40

5 EXPERIMENT 3: A TEST OF THE CONVERSATIONAL NORMS
A C C O U N T ........................................................ ...... ........ 44

Method ............................................................. 47
R results and D discussion ............................ ................................................................ 48
Experim ent 3A ....... ....... ............................................................... ........ 49











6 EXPERIMENT 4: A TEST OF THE REPRESENTATIVENESS
A C C O U N T ............................... ........................................ 5 1

M eth o d ................................... ........................................ 5 5
R esu lts ... .................... ......... ................................................. ......... 57
Discussion ... ................................................. 59

7 EXPERIMENT 5: PROCESSING THE INFORMATION WITHOUT THE
B EN E FIT IN M IN D ..................................................... ............................. 60

M ethod .... ......................................................... 62
R results and D discussion ............................ ............................................................. 63
Experim ent 5A .... .................................. ................ .......... 65

8 EXPERIMENT 6: CONSIDERING BOTH THE FOCAL AND ALTERNATIVE
HYPOTHESIS ............................................................ ............. ......... 69

M ethod ..................................... ................. 70
R results ... ............... .......... ....... ................................................ ......... 7 1
Discussion ... ................................ ...... .......... 72

9 EXPERIMENT 7: EXPECTING THAT THE PRODUCT WILL NOT DELIVER
THE BENEFIT .. ....................................... 73

M ethod ..................................... ........ ......... 74
R results and D discussion ............................ ............................................................. 76

10 G EN E R A L D ISC U SSIO N .................................................................................. 79

11 LIMITATIONS AND FUTURE RESEARCH .............................................. 83

REFEREN CES ........................ ............. 87

APPENDICES

A INSTRUCTIONS TO SUBJECTS ............................ .... .................... 93

B PRODUCT INFORM ATION ......................................................... 98

C EXAMPLES OF COMPUTER DISPLAYS ...................................................... 110

BIOGRAPHICAL SKETCH .... ....................... 112















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


CONSUMERS' BELIEFS IN PRODUCT BENEFITS:
THE EFFECT OF IRRELEVANT PRODUCT INFORMATION

By

Tom Meyvis

May 2001

Chairman: Chris Janiszewski
Major Department: Marketing


When consumers try to assess the performance of a product on a key benefit, they

often encounter irrelevant product information while searching for diagnostic

information. Normatively, this irrelevant information should not influence consumers'

belief in the product's ability to deliver the benefit. However, this dissertation presents

evidence that adding irrelevant product information to supportive benefit information

systematically weakens consumers' belief that the product will deliver the desired benefit.

This dilution effect persists when subjects are forced to acknowledge the irrelevance of

the additional information prior to stating their beliefs, when the information is allegedly

randomly selected by a computer, and when the irrelevant information increases the

similarity of the product description to the typical desired product. Furthermore, the

irrelevant information does not dilute the impact of the supportive information, but

instead has a direct, independent effect on consumers' beliefs in the benefit.










These findings are inconsistent with previous accounts of the dilution effect. The

diluting effect of irrelevant information observed in these studies is not caused by an

averaging strategy, a misguided reliance on conversational norms, or the use of a

representativeness heuristic. Instead, I suggest that this dilution effect is caused by

consumers' reliance on a biased hypothesis testing strategy. It is proposed that

consumers test the hypothesis that the product will deliver the benefit, selectively search

for information that confirms this hypothesis, and categorize all other information (be it

irrelevant or disconfirming) as "not confirming." When information is classified as "not

confirming," it weakens belief in the hypothesis, even when it does not confirm the

alternative hypothesis. As a consequence, irrelevant information weakens consumers'

beliefs in the product's ability to deliver the benefit.

Consistent with this explanation, the dilution effect is shown to disappear when

consumers first process the information without the benefit in mind, and when consumers

are forced to consider both the focal hypothesis and the alternative hypothesis.

Moreover, the dilution effect even reverses when the product description concerns a

brand with a poor reputation, suggesting that, in this case, consumers may be testing the

hypothesis that the product will not deliver the benefit.















CHAPTER 1
INTRODUCTION


When consumers evaluate products or services, they often focus on one or more

critical product benefits. To find out if the product under consideration will indeed

deliver these important benefits, consumers will usually search for diagnostic product

information. For instance, a consumer who wonders if a particular computer model is

fast may be primarily interested in the computer's processing speed or in its working

memory capacity. Managers understand this concern and will usually prominently

display these important features.

However, while searching for this diagnostic information, consumers will almost

inevitably also be exposed to obviously irrelevant information, i.e., information which

consumers perceive as obviously irrelevant with respect to the benefit they are looking

for. This irrelevant information may be encountered as a natural consequence of the

search process, or it may be provided as part of a deliberate management strategy. For

example, a consumer who is searching a web site for information on the processing speed

of a computer, may also notice that the computer has high quality speakers or that the

company sponsors an important art exhibition. Managers may have good reasons to

provide this information. The high quality speakers may appeal to those consumers who

value the sound quality of their computer, while the sponsorship of the arts may create

general positive affect. However, for a consumer who is only trying to find out whether










this computer is fast, this information is completely irrelevant: it does not suggest that the

computer is fast, but it does not indicate that it is slow either. Thus, normatively, the

presence of this irrelevant information should not change consumers' assessment of the

product's ability to deliver the desired benefit. A computer with a powerful processor

should not be perceived as faster or slower than a computer with a powerful processor

and high quality speakers, which is manufactured by a company that sponsors art

exhibits.



The Effect of Irrelevant Information

Yet, many studies suggest that objectively irrelevant information may actually

influence consumers' decisions. A first set of studies indicate that irrelevant product

information may polarize consumers' beliefs in the product benefit. Stated differently, a

computer with a powerful processor and high quality speakers may be perceived as faster

than a computer with only a powerful processor. This is consistent with studies on the

configural primacy effect (e.g., Asch 1946; Peterson and DuCharme 1967; Wallsten

1981), which have shown that people interpret additional information as consistent with

information they have received earlier. Thus, the additional, irrelevant information may

be interpreted as consistent with the supportive benefit information. Similarly, the

literature on the hypothesis confirmation bias (e.g., Beyth-Marom and Fischhoff 1983;

Ha and Hoch 1989; Hoch and Ha 1986; Nickerson 1998) has demonstrated that people

often interpret evidence in a manner that favors the hypothesis being tested. Therefore, if

the supportive information creates the hypothesis that the product will deliver the benefit,










the irrelevant information may be interpreted as supportive of the benefit. Finally, the

literatures on belief polarization (e.g., Tesser 1978) and schema-congruent processing

(e.g., Cantor and Mischel 1979) also suggest that the additional irrelevant information

will be interpreted as consistent with the schema created by the supportive information

and will lead to more extreme judgments. In sum, adding irrelevant information to

supportive benefit information may strengthen consumers' beliefs that the product will

deliver the benefit.

Several studies in the previously cited literatures have indeed demonstrated that

beliefs can become more extreme after confrontation with neutral evidence (e.g., Beyth-

Marom and Fischhoff 1983; Wallsten 1981), ambiguous evidence (e.g., Asch 1949; Ha

and Hoch 1989; Hoch and Ha 1986), and, in some cases, mixed evidence (e.g., Lord,

Ross, and Lepper 1979; but see Kuhn and Lao 1996). However, while these studies

demonstrated that objectively irrelevant information can polarize beliefs, they did not

examine the effect of information that is obviously irrelevant to decision makers, which is

the focus of this research project. In fact, those studies that have explicitly addressed the

impact of obviously irrelevant information have observed that adding this information to

diagnostic information actually dilutes people's judgments. Most of the demonstrations

of this dilution effect can be found in the social judgment literature (e.g., De Dreu,

Yzerbyt, and Leyens 1995; Fein and Hilton 1992; Nisbett, Zukier, and Lemley 1981;

Tetlock and Boettger 1989). For instance, Tetlock and Boettger (1989) observed that a

student who studied 31 hours a week received a higher GPA estimate than another

student who studied 31 hours a week and played tennis 3 times a month. A similar










dilution effect has also been observed in some studies using nonsocial stimuli (e.g.,

Shanteau 1975; Troutman and Shanteau 1977). For instance, Troutman and Shanteau

(1977) showed subjects draws of beads from one of two boxes. The boxes contained

different proportions of red and white beads, but an equal proportion of blue beads. They

observed that when an initial draw with predominantly red beads was followed by an

irrelevant draw with only blue beads, subjects reduced their confidence in the assumption

that the red box was being sampled. In sum, these studies suggest that obviously

irrelevant product information may weaken consumers' belief in a product's ability to

deliver a desired benefit. In other words, a computer with a powerful processor and high

quality speakers may be perceived as slower than a computer with only a powerful

processor.



Overview

The results reported in this dissertation are consistent with the findings of the

dilution effect literature. Across ten studies, adding irrelevant product information to

supportive information systematically weakens consumers' beliefs in the product's ability

to deliver the desired benefit. More importantly, these studies also provide insight into

the mechanism behind this dilution effect, and show under which conditions this effect

may disappear or even reverse. Previous studies have argued that the dilution effect may

be caused by an averaging strategy, the use of the representativeness heuristic, or the

reliance on conversational norms. This dissertation proposes a new explanation for the

dilution effect based on the literature on biased hypothesis testing. It is argued that










consumers may be selectively testing the hypothesis that the product will actually deliver

the desired benefit. They may selectively search for evidence that supports this

hypothesis and classify all information with respect to their search goal: information that

is supportive of the benefit is classified as "confirming," while any additional information

is classified as "not confirming." Consumers' belief in the hypothesis will be stronger to

the extent that their search for confirmatory evidence produced more "confirming"

evidence and will be weaker to the extent that this search produced more "not

confirming" evidence. This alternative explanation will be systematically tested against

the existing accounts of the dilution effect. To construct these tests, I examine a number

of important factors which have not yet received attention in the dilution literature, such

as the manner in which the information is processed and consumers' awareness of the

irrelevance of the information. Furthermore, I will explore when irrelevant information

may in fact strengthen consumers' belief in the product benefit, thus effectively reversing

the dilution effect.

Before presenting the studies, I will first discuss previous research on the

influence of irrelevant product information and review prior demonstrations of the

dilution effect as well as the prior accounts of the effect. I will then propose a new

explanation based on the biased hypothesis testing literature, which will be tested in ten

experiments. Finally, I will provide a general discussion of the findings, as well as a

discussion of the limitations of this dissertation and possible avenues for future research.















CHAPTER 2
THEORETICAL BACKGROUND


Irrelevant Product Information in the Consumer Research Literature

Several studies in the consumer behavior literature have demonstrated that

objectively irrelevant product information can influence consumer decisions. For

instance, Hoch and Ha (1986) observed that a nondiagnostic, ambiguous product

experience can increase the perceived quality of an advertised brand. Similarly,

Carpenter, Glazer, and Nakomoto (1994) report that a brand with a distinguishing, but

irrelevant attribute receives a higher preference rating than the same brand without the

attribute. On the other hand, Simonson, Carmon, and O'Curry (1994) have shown that

consumers are less likely to choose a brand that offers a promotion or feature that has no

value to them. Similarly, Simonson, Nowlis, and Simonson (1993) report that consumers

who notice that others prefer a product for a reason which they find irrelevant are less

likely to choose that product. Finally, Brown and Carpenter (2000) replicated the

findings of Carpenter et al. (1994) and Simonson et al. (1994) and demonstrated that the

direction of the effect partially depends on the size of the choice set.

All these studies demonstrate that irrelevant information influences consumer

decisions. Yet, they may not tell us much about the influence of obviously irrelevant

information on consumers' beliefs in product benefits, because the situations examined in

these studies differ in two important ways from the subject of this dissertation. First,










while the information presented in the preceding studies was objectively irrelevant,

subjects did not perceive the information as irrelevant for their decision and knowingly

relied on it. Subjects in Hoch and Ha (1986) perceived the ambiguous product

experience as diagnostic information and interpreted the experiences as confirming the

advertising claims. On the other hand, while subjects in Carpenter et al. (1994) were

given information from which they could infer that the differentiating attribute did not

influence the quality of the product, they did seem to make quality inferences based on

this information and thus perceived it as relevant for their decision. Consistent with this

assertion, the irrelevant information only increases preference ratings when the target

brand is offered at a high price (study 2), and the effect of the irrelevant information even

reverses when the labels have a negative rather than positive connotation (Broniarczyk

and Gershoff 1997). Similarly, in the experiments by Simonson and colleagues (1993,

1994), subjects are probably aware that an unwanted feature does not change the quality

of the product, but they do perceive it as unattractive and state it as a reason for not

choosing the target brand. Likewise, the majority of subjects in Brown and Carpenter

(2000) refer to the trivial information to justify their choices. Thus, while these previous

studies indicate that consumers can be influenced by objectively irrelevant product

information, they do not inform us whether consumers can be influenced by product

information that they perceive to be irrelevant for their decision.

A second important distinction concerns the type of consumer decisions. None of

the previous studies examined specific product judgments, such as beliefs in a product's

ability to deliver a particular benefit. With the exception of Hoch and Ha (1986), all










these studies examine the influence of trivial information in a choice or preference

context in which consumers have to make comparisons between brands. While this

distinction may seem trivial, it is in fact of great importance, as the choice context is

essential for the explanation of the effects. The irrelevant information influences

consumer decisions because it differentiates the target brand from the other brands

(Carpenter et al. 1994) and thus provides consumers with a justification for their decision

and a way to resolve the choice conflict (Brown and Carpenter 2000; Simonson et al.

1994). Subjects' choice protocols (Brown and Carpenter 2000; Simonson et al. 1993,

1994) and the moderating effect of the choice context (Brown and Carpenter 2000)

provide strong support for these interpretations. However, these mechanisms certainly do

not operate in situations in which consumers make independent assessments of a brand's

ability to deliver a specific benefit. In sum, while these previous findings clearly

demonstrate that objectively irrelevant product information can influence consumer

choices and preferences, they do not inform us whether specific benefit judgments can be

influenced by information which consumers perceive as obviously irrelevant with respect

to these benefits. However, the psychological literature on the dilution effect does look at

the influence of obviously irrelevant information on specific predictions and may thus

offer a more promising perspective on this research problem. I will first review the

general findings from this research stream, followed by an overview of the existing

accounts of the effect.










The Dilution Effect

The dilution effect has been found using social stimuli with natural judgments and

nonsocial stimuli with probability judgments. Studies using social stimuli with natural

judgments always use a person as target stimulus, manipulate the presence of irrelevant

information between subjects, and ask subjects to make a single judgment about a

stereotypic trait (De Dreu et al. 1995; Fein and Hilton 1992; Nisbett et al. 1981; Tetlock,

Lerner, and Boettger 1996; Tetlock and Boettger 1989; Zukier 1982; Zukier and Jennings

1983). For example, in Zukier and Jennings (1983), students were asked to act as jurors

in the trial of a man accused of murdering his aunt. In the control condition, subjects

only received evidence indicative of guilt (e.g., "he was known to have argued with his

aunt" and "he had no alibi"). In the treatment condition, subjects received the same

diagnostic information, as well as additional information that had no implication for the

defendant's guilt (e.g., "the defendant is of average height and vision"). Zukier and

Jennings (1983) observed that subjects in the treatment condition were less likely to

conclude that the suspect committed the murder than were subjects in the control

condition.

Studies using nonsocial stimuli with probability judgments use commodity-like

objects as target stimuli, manipulate the presence of irrelevant information within-subject,

and ask subjects to make multiple, explicit probability judgments (Birnbaum and Mellers

1983; Lichtenstein, Earle, and Slovic 1975; Shanteau 1975; Troutman and Shanteau

1977). For example, Troutman and Shanteau (1977) showed subjects two boxes

containing different proportions of red and white beads (one box had more red beads and










one box had more white beads), but equal proportions of blue beads. Subjects were then

presented with successive samples of beads and, after each sample, indicated the

probability that the predominantly white box was being sampled. Troutman and

Shanteau (1977) found that when a diagnostic draw (a red or white bead) was followed

by an irrelevant draw (a blue bead) or neutral draw (an equal number of red and white

beads), subjects moderated their estimates.



Existing Accounts of the Dilution Effect

Three major explanations of the dilution effect have been proposed in the

literature. The representativeness heuristic and conversational norm explanations have

been proposed in the social judgment literature, while the averaging hypothesis is the

most prominent explanation in the nonsocial judgment literature. I will review these

existing explanations and then propose a new explanation based on the biased hypothesis

testing literature.

Representativeness. The most popular account of the dilution effect within the

social judgment literature relies on the representativeness heuristic (Fein and Hilton

1992; Hilton and Fein 1989; Locksley, Hepburn, and Ortiz 1982; Nisbett et al. 1981;

Tetlock and Boettger 1989; Zukier 1982). The representativeness heuristic is a strategy

by which subjects use the similarity between the available information and the typical to-

be-predicted outcome to estimate the probability that the outcome will occur (Kahneman

and Tversky 1972, 1973). It is assumed that diagnostic information is highly

representative of the to-be-predicted behavior and that irrelevant information is not










representative of this behavior. Therefore, adding irrelevant information to diagnostic

information makes the individual less representative of the behavior and attenuates the

judgment.

For example, a student who studies 31 hours a week is more similar to the

stereotypical high GPA student" than a student who studies 31 hours a week and plays

tennis 3 times a month. Therefore, if people base their estimate of a student's GPA on

the student's similarity to the stereotypical "high GPA student," then the student who also

plays tennis 3 times a month, should receive a lower GPA rating than the student who

does not. This is indeed what Tetlock and Boettger (1996) observed. Similarly, a

computer with a powerful processor will be perceived as more similar to a typical "fast

computer" than a computer with a powerful processor manufactured by a company that

sponsors an arts exhibit. Therefore, if consumers rely on a representativeness heuristic to

assess product benefits, irrelevant product information should weaken consumers' belief

in the product's ability to deliver the benefit.

Averaging. The averaging explanation is the most popular account of the dilution

effect within the nonsocial judgment literature (Birnbaum and Mellers 1983; Lichtenstein

et al. 1975; Shanteau 1975; Troutman and Shanteau 1977). A first group of algebraic

averaging models assumes that the weights of the attributes are being adjusted according

to the weights of the other attributes (e.g., Anderson 1967, 1971, 1974; Birnbaum and

Mellers 1983):

R = C+ w,s, /Xw, +E,
i=0 1=0










Therefore, if the irrelevant information receives a nonzero weight, adding irrelevant

information can weaken the impact of the diagnostic information, thus diluting people's

responses. A second set of averaging models suggest that people use an adjustment

model in which they make separate predictions based on each piece of information, and

then average the predicted outcomes (e.g., Lichtenstein et al. 1975; Lopes 1987; Shanteau

1975). An example of the latter is Shanteau's (1975) study in which subjects were shown

samples from one of two boxes with different proportions of red and white beads. A

single red bead indicated that there was a 60 percent chance that the red box was being

sampled; a draw with equal numbers of red and white beads implied that there was a 50

percent chance that either box was being sampled. Subjects who saw the red bead sample

followed by the neutral sample adjusted their predictions downward, resulting in a

prediction between 50 and 60 percent, consistent with the use of an averaging rule.

Troutman and Shanteau (1977) replicated this finding, but added an equally large number

of blue beads to both boxes. They demonstrated that a draw consisting of only blue

beads had the same diluting effect as a draw consisting of equal numbers of red and white

beads. It could be argued that, since both boxes contained a large proportion of blue

beads, the blue beads draw was very representative of each box. Therefore, Troutman

and Shanteau's (1977) findings suggest that reliance on the representativeness heuristic

cannot explain all occurrences of the dilution effect.

However, there is considerable debate regarding the status of the averaging

explanation. Anderson (1974) argues that the averaging model is an as-if model, a data

model, rather than a process model: "The work on the averaging hypothesis only allows










the conclusion that the subject acts as if he were averaging. ... It seems unlikely that the

subject carries out the steps indicated by equation 1." (Anderson 1974, p 253). Likewise,

Lichtenstein et al. (1975) propose their averaging model as an adequate description of

individuals' observed behavior, but suggest that the representativeness heuristic may be a

more accurate description of the actual decision process. On the other hand, Shanteau

(1975) and Troutman and Shanteau (1977) present their averaging model as a process

explanation and an alternative to the representativeness account. Finally, Lopes (1987)

distinguishes between the algebraic averaging models, which she perceives as data

models, and the adjustment models, which she presents as a possible description of the

decision process.

Conversational norms. A third account of the dilution effect argues that dilution

is an experimental artifact resulting from subjects' mistaken reliance on conversational

norms. This account has recently become very prominent in the social judgment

literature (e.g., Schwarz, Strack, Hilton, and Naderer 1991, Slugoski and Wilson 1998;

Tetlock et al. 1996). In intentional communication, a number of conversational norms

are assumed to be respected, one of which is that all information that is being provided is

relevant for the goal of the conversation. This has been referred to as the maxim of

relation (Grice 1975) or the principle of relevance (Sperber and Wilson 1986). However,

this norm is violated in the experimental context of the dilution studies, thereby leading

subjects to erroneous inferences ("The experimenter provides this information, so it must

be relevant. "). For example, Tetlock et al. (1996) asked subjects to predict the GPA of a

hypothetical student and manipulated subjects' accountability as well as the activation of










conversational norms. They observed that, for accountable subjects, the dilution effect

disappeared when conversational norms were deactivated by mentioning that the

computer had randomly selected the information that was presented. However, non-

accountable subjects still showed a dilution effect when conversational norms had been

deactivated, suggesting that conversational norms may contribute to the dilution effect,

but are not necessary for the effect to occur.

While the conversational norms account explains why subjects relied on the

information despite its irrelevance, this account does not explain why adding the

irrelevant information results in less extreme rather than more extreme judgments.

However, the conversational norms explanation can easily be combined with the

representativeness or averaging accounts to address this problem. These latter

explanations describe decision mechanisms that can lead to less extreme responses,

whereas the conversational norms account explains why the irrelevant information is

actually incorporated into these decision mechanisms. If subjects would not assume that

all the information provided by the experimenter has to be relevant, they may not

incorporate this information in their similarity judgments nor use it as a basis for separate

predictions which will be averaged with predictions based on the diagnostic information.

In any case, regardless of whether the reliance on conversational norms is a

sufficient cause of the dilution effect, this perspective suggests that the assumption that

the information provided is relevant is a necessary condition for the dilution effect to

occur. Does this imply that irrelevant product information may only dilute product

judgments in an experimental setting? Surely not. When consumers encounter product










information in an advertisement, they are also likely to make certain assumptions. They

may, for instance, assume that the advertiser will only release information that will

encourage consumers to purchase the product. Thus, consumers may assume that all the

information provided is relevant. However, an advertisement does not constitute an

intentional face-to-face communication. The advertiser may have good reasons to

assume that the information will appeal to a certain consumer segment, but this does not

imply that the information is relevant for all consumers, regardless of the benefits they

are personally interested in. Thus, consumers may assume that the irrelevant product

information is intended for a different consumer segment and may therefore not rely on

this information. In sum, if the reliance on the relevance principle is a necessary

condition for the dilution effect to occur, the dilution effect may not apply to many

realistic advertising situations.



A Biased Hypothesis Testing Account

All three preceding mechanisms can account for a diluting effect of irrelevant

product information on consumers' beliefs in product benefits. However, I propose an

alternative, fourth account which will be tested against the three existing explanations

within a product judgment context. It is proposed that consumers who are assessing

product benefits, as well as subjects in the dilution experiments, are faced with a

hypothesis testing task. For instance, consumers are testing the hypothesis that the

product will deliver the benefit, while subjects in Zukier and Jennings (1983) are testing

the hypothesis that the defendant is guilty. Thus, the extensive literature on biased










hypothesis testing may provide additional insight into the effect of irrelevant information

in this context. The proposed biased hypothesis testing explanation is based on four

assumptions which will be discussed next.

First, it is assumed that consumers are more likely to test the hypothesis that the

product will deliver the benefit, rather than the hypothesis that the product will not deliver

the benefit. Consumers are more likely to benefit from finding out which products

deliver the desired benefit than by identifying all the products that will not. Moreover,

advertising claims can explicitly create the hypothesis that the product will indeed deliver

the benefit (Ha and Hoch 1989; Hoch and Ha 1986). Furthermore, this preference for

positively framed hypotheses may generalize beyond consumers' search for product

benefits. Evidence suggests that the acceptance of an idea is part of the automatic

comprehension of that idea, occurs before the rejection of the idea, and is also less

effortful than the rejection (Gilbert 1991; Gilbert, Tafarodi, and Malone 1993). In sum,

consumers may start out with assuming that the product will deliver the benefit, and set

out to test this hypothesis rather than its complement.

Second, it is assumed that, when testing this hypothesis, consumers will search for

evidence in a biased fashion. Consumers will selectively search for confirming evidence,

i.e., evidence that suggests that the product will indeed deliver the benefit. There have

been many demonstrations of a biased search for confirming evidence (e.g., Shaklee and

Fischhoff 1982; Snyder and Cantor 1979; Snyder and Swann 1978 ) or, in rule testing,

positive test cases (e.g., Klayman and Ha 1987). While early research demonstrated that

this biased search strategy will often increase the likelihood that the hypothesis will be










accepted (e.g., Snyder and Swann 1978; Wason 1960), later studies indicated that a

biased strategy does not necessarily lead to a biased outcome (e.g., Klayman and Ha

1987). Consistent with this latter perspective, I propose that consumers' biased search

for product information that confirms the hypothesis that the product will deliver the

benefit does not necessarily strengthen consumers' belief in the product benefit.

Third, consistent with the literature on goal-based classification (e.g., Barsalou

1983; Ratneshwar, Pechmann, and Shocker 1996), it is assumed that consumers will

classify the product information / i/h respect to their search goal. In other words,

consumers will classify information as either "confirming" (i.e., the type of information

they were searching for) or "not confirming" (i.e., not the type of information they were

searching for). While ambiguous and obviously supportive information can be

interpreted as supportive and will therefore be classified as "confirming,"

counterdiagnostic and obviously irrelevant information can not be interpreted as

supportive and will be classified as "not confirming."

Finally, it is assumed that when consumers evaluate the product information, they

will only consider its implications for the focal hypothesis and ignore the consequences

for the alternative hypothesis. Thus, consumers' hypothesis testing is not only

confirmatory, but it is also selective. Several studies on selective or pseudo-diagnostic

hypothesis testing have demonstrated that people indeed only consider the focal

hypothesis when interpreting evidence (Beyth-Marom and Fischhoff 1983; Sanbonmatsu,

Posavac, Kardes, and Mantel 1998; Sanbonmatsu, Posavac, and Stasney 1997; Trope and

Liberman 1996). Thus, when consumers classify ambiguous information as "confirming"










that the product will deliver the benefit, they will strengthen their belief in the product

benefit, even though the information is equally supportive of the hypothesis that the

product will not deliver the benefit. On the other hand, when consumers classify

obviously irrelevant information as "not confirming" that the product will deliver the

benefit, they will weaken their belief in the product benefit, even though the information

also does not support the hypothesis that the product will not deliver the benefit. While

the first process leads to a confirmation bias, the latter results in dilution.

Similar to the literature on biased information search, most demonstrations of

selective hypothesis testing have shown that it can lead to an unwarranted increase in

confidence in the focal hypothesis when the evidence is very likely given either

hypothesis (e.g., Beyth-Marom and Fischhoff 1983). However, several researchers

(Beyth-Marom and Fischhoff 1983; Sanbonmatsu et al. 1997, 1998) have indicated that

the same strategy may actually lead to an unwarranted decrease in confidence in the

hypothesis when the evidence is unlikely given any of the hypotheses. Beyth-Marom and

Fischhoff (1983) have even suggested that this strategy may explain the dilution findings

reported by Nisbett et al. (1981). They propose that people only consider the likelihood

of the evidence given the focal hypothesis, i.e., P(DIH), and do not consider the

likelihood of the evidence given the alternative hypothesis, i.e., P(Dlnot H), thus ignoring

the denominator of the likelihood ratio in Bayes' theorem. Beyth-Marom and Fischhoff

(1983) argue that, because Nisbett et al. (1981) designed their stimuli so that the

irrelevant information did not fit with either of the possible classification categories, a

selective focus on P(DIH) may have decreased subjects' confidence in the focal










hypothesis. For example, some subjects in Nisbett et al. were asked to estimate the

probability that a student, who was either a music or an engineering major, was in fact a

music major. The irrelevant, highly personal information about the student was not very

likely given that the student was a music major (i.e., P[DIH] was low), but was also not

very likely given that the student was an engineering major (i.e., P[Dlnot H] was also

low). Thus, the personal information may have decreased subjects' belief that the student

was a music major despite its irrelevance, because students only considered the low

likelihood of this information given that the student was a music major.

Yet, while this process, by itself, can indeed explain the dilution effect when the

irrelevant information is unlikely given either hypothesis, it cannot account for the many

demonstrations of the dilution effect using irrelevant information that does not fit this

restriction. In fact, Zukier and Jennings (1983) even demonstrated that the dilution effect

was more pronounced when the irrelevant information was typical rather than atypical.

Therefore, while the selective focus on a single hypothesis is a necessary assumption for

a biased hypothesis testing account of the dilution effect, it is not sufficient. However,

when it is combined with the previous three assumptions, it may indeed account for a

wide range of dilution effects.

In sum, it is proposed that consumers are testing the hypothesis that the product

will deliver the benefit. They selectively search for information that suggests that the

product will indeed deliver the benefit. Consumers will then classify product information

with respect to this search goal. Obviously supportive and ambiguous information will

be classified as "confirming," while counterdiagnostic and obviously irrelevant










information will be classified as "not confirming." When information is classified as

"confirming," it will strengthen consumers' belief in the benefit, even when it also

supports that the product will not deliver the benefit. On the other hand, when

information is classified as "not confirming," it will weaken consumers' belief in the

benefit, even when it also does not support that the product will not deliver the benefit.

For instance, suppose that a consumer is testing an advertising claim and is

confronted with an ambiguous product experience. According to the process described

above, the consumer will search for information that confirms the advertising claim and,

since the ambiguous experience can easily be interpreted as consistent with the claim, she

will classify it as "confirming" and strengthen her belief in the claim. This is indeed what

Hoch and Ha observed in their studies (Ha and Hoch 1989; Hoch and Ha 1986).

However, what would happen if this consumer would encounter obviously irrelevant

instead of ambiguous product information? In that case, she cannot interpret this

information as confirming the claim and she will classify it as "not confirming."

Moreover, her belief in the advertising claim will be weakened, even though the

information does not confirm that the claim is not true either. Thus, while this biased

hypothesis testing process may produce a confirmation bias when the product

information is ambiguous, the same process may produce a dilution effect when the

product information is obviously irrelevant.

The generality of the biased hypothesis testing explanation. While the biased

hypothesis testing explanation was formulated in a product judgment context, this same

mechanism may also apply to prior demonstrations of the dilution effect in other domains










(although I do not argue that one single mechanism is responsible for all observations of

the dilution effect). Consider, for instance, Shanteau's (1975) experiment in which

subjects draw from two boxes with different proportions of red and white beads. If

someone first draws predominantly red beads, this may create the hypothesis that the

beads are being sampled from the "red box." The person will now be looking for

information that confirms this hypothesis, i.e., draws with predominantly red beads. If

the second sample contains equal numbers of red and white beads, the sample will be

perceived as "not confirming" and confidence in the hypothesis will be reduced.

The four assumptions on which this alternative explanation is based will not hold

in all situations. For instance, consumers may not start out with the hypothesis that the

product will deliver the benefit, when they have a very negative prior opinion of that

product. In that case, they may test the hypothesis that the product will not deliver the

benefit. Similarly, consumers may sometimes consider the implications of the product

information for both the focal and alternative hypotheses (i.e., "Does this suggest that the

product will deliver the benefit, or does this suggest that the product will not deliver the

benefit?"). As will be demonstrated, the dilution effect will disappear, or even reverse, in

these situations.

Comparison with existing explanations. The proposed biased hypothesis testing

explanation is not incompatible with previous accounts of the dilution effect. As was

mentioned earlier, the averaging account may provide a description of people's behavior

rather than a description of their decision process (e.g., Lichtenstein et al. 1975). Thus, it

is plausible that when people engage in biased hypothesis testing, they may generate










responses that can often be perfectly described by an averaging model of information

integration. However, there are situations in which the predictions of the hypothesis

testing account can not be adequately represented by an averaging model. One such

situation will be examined in experiment 2.

The representativeness explanation, on the other hand, is more similar to the

biased hypothesis testing account. In fact, one could interpret the representativeness

heuristic as a heuristic hypothesis testing strategy: people may test hypotheses by relying

on the similarity between the evidence and the typical hypothesized outcome. However,

there are two important distinctions between the proposed explanation and the

representativeness account. On the one hand, the representativeness explanation is less

specific than the biased hypothesis testing account. The biased hypothesis testing

explanation explicitly states that, for the dilution effect to occur, the product information

needs to be processed with the hypothesis (i.e., the desired benefit) in mind. While this

constraint can be added to the representativeness explanation (i.e., the information needs

to be processed with the desired prototype in mind), it is not essential. If consumers first

form an impression of the described product and only then compare the product to the

desired prototype, the dilution effect could still occur, since the irrelevant product

information would still make consumers' impression of the product less similar to the

typical desired product. On the other hand, the representativeness explanation is also

more specific than the biased hypothesis testing account. The biased hypothesis testing

account does not specify how consumers decide whether to classify a piece of product

information as "confirming" or "not confirming." They may make inferences based on










prior theories, or they may rely on heuristics such as similarity judgments. However,

while the representativeness account states that this reliance on similarity judgments is

essential for the explanation of the dilution effect, it is not a necessary requirement for the

biased hypothesis testing account.

The following chapters will present ten experiments in which the biased

hypothesis testing account is tested against each of the three existing explanations. First,

experiments 1 and 1A demonstrate the effect, show its robustness across product

categories, presentation orders, and belief measures, and provide evidence against a

distraction of attention account. Then, experiment 2 demonstrates that the diluting effect

of irrelevant product information can not be accounted for by an averaging model, while

experiments 3 and 3A provide evidence against a conversational norms explanation of the

effect, and experiment 4 shows evidence that is inconsistent with the representativeness

account. While the findings of these first experiments are all consistent with the biased

hypothesis testing explanation, the remaining experiments will more systematically

manipulate the assumptions of the proposed mechanism. Experiments 5 and 5A

demonstrate that the dilution effect only occurs when the irrelevant information is

processed with the desired benefit in mind, whereas experiment 6 shows that the effect

also disappears when consumers consider the implications of the evidence for both the

focal and alternative hypothesis. Finally, experiment 7 demonstrates that the effect even

reverses when consumers set out to test the hypothesis that the product will not deliver

the benefit.















CHAPTER 3
EXPERIMENT 1: DEMONSTRATION OF THE DILUTION EFFECT


The objective of this first experiment is to demonstrate the diluting effect of

irrelevant product information and provide an initial test of a distraction explanation of

this effect. Although there is considerable evidence for dilution in social and non-social

judgments, these demonstrations will not necessarily generalize to a product judgment

context. Unlike subjects in social judgment experiments, consumers who make product

judgments can not rely on easily accessible stereotypes, which may be an essential

requirement for some decision mechanisms to occur (e.g., representativeness).

Similarly, the abstract cues and within-subject manipulations, typical of the non-social

demonstrations of the dilution effect, seem to have few parallels in product judgment

contexts. Furthermore, many studies have shown that adding nondiagnostic neutral or

ambiguous information can polarize rather than dilute judgments (e.g., Asch 1949;

Beyth-Marom and Fischhoff 1983; Ha and Hoch 1989; Hoch and Ha 1986; Wallsten

1981). Therefore, given the limited overlap between the product judgment context and

contexts in which dilution effects have been observed, and given the evidence that

nondiagnostic information can also lead to polarization, experiment 1 provides an initial

test of the dilution effect in a product judgment context.

A second objective of this first experiment is to test whether the irrelevant

information is diluting consumers' beliefs by distracting resources otherwise allocated to










the diagnostic information. While this distraction explanation has not been tested

previously, it can account for the dilution effect by assuming that people extract less

information from the supportive evidence when their attention is distracted by the

irrelevant evidence. To test this explanation, I measured subjects' recognition of the

diagnostic information at the end of the experiment. If the irrelevant information did

indeed cause subjects to elaborate less on the supportive information, then subjects who

had been exposed to this irrelevant information should recognize the supportive

information less easily than those who had only received the supportive information.

Each of the ten experiments reported in this dissertation used a very similar

procedure in which subjects received descriptions of different products or services and

were asked to indicate their belief that this product or service would deliver a specific

benefit. In experiment 1, subjects were presented with eight different products or

services. For each category, they were first given a specific desirable benefit (e.g., "You

are looking for a fast computer"). Subjects then received the product description. In the

baseline condition, the product description only contained one piece of supportive

information that strongly suggested that the product would deliver the benefit (e.g., "Very

Powerful Processor"), whereas in the treatment condition, the description also contained

three pieces of irrelevant information (e.g., "Assembled in the USA," "Airs Commercials

on NBC & CBS," and "Can Be Ordered On-Line"). After subjects had received the

entire product description, they were asked to indicate whether the product would deliver

the desired benefit. When all descriptions had been shown, subjects were given a brief

reaction time training, which provided a baseline measure of subjects' response latencies.










Finally, subjects were given a recognition task, in which they were presented with the

eight pieces of supportive information, as well as eight new pieces of information. They

were asked to indicate, as fast as possible, if the information had been presented earlier in

the experiment.

First, it was expected that the addition of irrelevant information would weaken

beliefs in the product benefit, thus demonstrating the dilution effect. In other words,

subjects who had received both supportive and irrelevant information were expected to

provide lower belief ratings than subjects who had only received supportive information.

Second, if this dilution effect was due to a distraction of resources by the irrelevant

information, the addition of irrelevant information should impair the recognition of the

supportive information.



Method

Subjects and design. Subjects were 36 undergraduate students who participated in

return for class credit. The design was a 2 (type of information) by 8 (product replicates)

mixed design. Each subject was presented with descriptions of eight different products.

For each of the product replicates, subjects were randomly assigned to either the baseline

condition or the treatment condition.

Stimuli. Eight different products or services were selected and a desirable benefit

was specified for each category: apartments (safe), package delivery service (fast), frozen

dinners (healthy), airlines (superior service), toothpaste (fights cavities), car (sportive),

stereo system (reliable), and computers (fast). For each replicate, three irrelevant










attributes had to be selected, as well as one attribute that strongly suggested that the

product would deliver the benefit, i.e., the supportive attribute (see Appendix B for a

complete list of the product information presented in this and other experiments). A

pretest (n = 30) was conducted to select these attributes. The pretest listed a wide range

of facts for each product or service. Subjects were asked to allocate these facts to one of

three categories: "suggests not [benefit]", "is not helpful for my decision", or "suggests

[benefit]". The 24 irrelevant facts selected for the main experiment were classified as

"not helpful" by an average of 90 % of pretest subjects, as supportive of the benefit by 6

% of the subjects, and as counterdiagnostic by only 4 % of the subjects. Each of the

irrelevant facts was classified as "not helpful" by at least 80 % of pretest subjects. The

irrelevant information included package information (e.g., a toothpaste that comes in 6

oz. tubes), product attributes (e.g., a computer that can be ordered on-line), marketing

information (e.g., an airline that sponsors the NYC Marathon), and product availability

(e.g., a frozen dinner brand that is available at most grocery stores). The eight supportive

facts selected for the main experiment were classified as suggesting the benefit by an

average of 94% of pretest subjects.

While the first pretest demonstrated that subjects perceived the information as

irrelevant in isolation, it was possible that this apparently irrelevant information became

relevant in the context of the complete product description. Therefore, a second pretest

was conducted to measure subjects' perception of the information in the context of the

complete description. Thirty subjects were presented with the full product descriptions

and asked to indicate the relevance of each piece of information. The 24 irrelevant facts










were classified as irrelevant by an average of 93 % of the subjects, as diagnostic of the

benefit by 6 % of the subjects and as counterdiagnostic by only 1 % of the subjects. The

eight supportive facts were classified as suggesting the benefit by an average of 93 % of

pretest subjects, as irrelevant by 6 % of the subjects, and as counterdiagnostic by less

than 1 % of the subjects.

A final pretest examined the possibility that even though pretest subjects indicated

that the irrelevant facts were not diagnostic, subjects in the actual experiment may still

use these facts as a basis for forming beliefs about the product benefit. For example,

although a subject may classify a fact as irrelevant, the fact may still be informative

because it is positively or negatively correlated with unstated facts that are relevant. To

examine the possible direction of such an effect, 18 subjects were presented with the

irrelevant facts and were asked to rate them on a 6-point scale (ranging from 1 = "Will

probably not [deliver benefit]" to 6 = "Will probably [deliver benefit]"), thus forcing

them to classify the information as either diagnostic or counterdiagnostic. The irrelevant

facts were classified as diagnostic of the benefit by an average of 65% of the subjects and

as counterdiagnostic by an average of 35% of the subjects. The average rating was 3.94

which was significantly higher than 3.50, the midpoint of the scale (t(432) = 7.22, p <

.001). Thus, if the irrelevant information would indeed lead to inferences in the

experimental context, these inferences would support subjects' beliefs in the product

benefit, rather than counteract them.

Procedure. The entire experiment was administered by personal computer.

Subjects entered the experimental lab and were asked to sit at a computer. All










instructions were provided by the computer program. Subjects were first informed that

they would receive information about eight different products (services) and that they

would have to indicate their belief that the described product would deliver a particular

benefit. Subjects were told that the information they would receive "may or may not be

helpful for the decision that you have to make" (see Appendix A for the complete

instructions used in this and other experiments). Subjects then received the information

for the first replicate. First, subjects were informed of the benefit they were looking for

(e.g., "You are looking for a fast computer."). Immediately after this statement, subjects

were provided with the first piece of information, which was always the supportive

attribute (e.g., "Very Powerful Processor"). In the treatment condition, this information

was followed by three additional pieces of irrelevant information that were presented

sequentially (e.g., "Assembled in the USA", "Airs Commercials on NBC and CBS," and

"Can Be Ordered On-Line"). Finally, while the entire product description remained on

the screen, subjects were asked to specify their belief that the product would deliver the

benefit (e.g., "Is this computer fast?"). Responses were made on a 9-point scale (e.g.,

anchored by 1 = "Definitely Not Fast" and 9 = "Definitely Fast"). Examples of the

complete screen displays in the baseline and treatment conditions can be found in

Appendix C. After subjects had rated their belief in the benefit, they received the

information for the next replicate, until all eight product categories had been presented.

The order of the replicates was randomized.

After all replicates had been presented, subjects were given a reaction time

training task. Subjects were shown six simple statements unrelated to the experiment










(e.g., "Paris is the capital of France") and had to answer "True" or "False" as fast as

possible by pressing the "1" or "0" key on their keyboard. Subjects received feedback

about their response time and accuracy after each statement. For each subject, the

average reaction time of the accurate responses was recorded for use as a response

latency covariate in the analysis. Finally, subjects were shown the actual reaction time

task. They were told that they would receive statements about the product descriptions

they had received earlier and would have to indicate whether they were true or false using

the same technique as in the training task. For each statement, subjects were first shown

a warning screen, which mentioned the product for which they would receive a statement

(e.g., "The next statement concerns: The Computer"), and asked subjects to place their

fingers on the "0" and "1" keys. The screen then displayed a true statement, such as "The

computer had a very powerful processor," or a false statement, such as "The computer

was loaded with games." For each replicate, consumers received one true and one false

statement. The true statements always concerned the supportive piece of information.

The false statements contained randomly chosen positive features that were unrelated to

the desired benefit. The presentation order of the 16 statements was randomized.



Results

The addition of the irrelevant information significantly weakened subjects' beliefs

in the product's ability to deliver the desired benefit (F(1,272) = 6.56, p = .01). Subjects

who only received the supportive information reported more extreme judgments ( X=

6.28) than those who also received the irrelevant information (X= 5.83). The effect of










irrelevant information did not depend on the specific product or service that people were

judging (F(7,272)= 1.47, ns)1. These results indicate that the diluting effect of irrelevant

information does indeed generalize to product judgments.

Subjects' recognition of the supportive information at the end of the experiment

was very high overall. Interestingly, subjects who only received supportive information

were slightly less likely to recognize the supportive information (2 = 93%) than were

subjects who also received the irrelevant information (j = 97%; X2 = 2.91, df= 1, p <

.1). Those subjects who recognized the supportive information did not show any

differences in reaction time depending on the product information they had received

earlier (RTsupp= 1.53, RTSUPP+IRREL= 1.57; F< 1, ns).



Discussion

The results of experiment 1 demonstrate that the diluting effect of irrelevant

information generalizes to product judgments. Across eight different products and

services, adding irrelevant product information to supportive benefit information

systematically weakened subjects' beliefs in the product's ability to deliver the desired

benefit. Furthermore, the findings of the recognition task indicate that this dilution effect

is not due to the fact that the irrelevant information diverts resources which otherwise

would have been used to process the supportive information. In fact, those subjects who

only received supportive information were marginally less likely to recognize the


1 For the sake of clarity and conciseness, non-significant interactions with the replicate
factor will not be discussed for the remaining experiments. All interactions are non-
significant, unless explicitly stated otherwise.










supportive information than were those who also received the irrelevant information.

Furthermore, there were no differences between the average reaction times in the two

information conditions. These results suggest that the supportive information was not

processed to a lesser degree when additional irrelevant information was present, which is

inconsistent with a distraction of attention account of the dilution effect.



Experiment 1A

While the first experiment demonstrated the diluting effect of irrelevant

information and provided evidence against a distraction of resources explanation,

experiment 1A tested the robustness of this phenomenon. First, the experiment examined

whether the dilution effect also occurs when consumers are comparing multiple products

and select the product that is most likely to deliver the benefit. In experiment 1, the

addition of irrelevant information could not only have influenced consumers' perception

of the described product, but could also have changed consumers' perception of the belief

measure. In other words, similar scores on the 9-point belief scale may have reflected

different perceptions depending on the information condition. This problem was

addressed in experiment 1A by using a selection task rather than a belief measure.

Furthermore, this experiment also manipulated the order in which the information

was presented. Consistent with most previous demonstrations of the dilution effect (e.g.,

Fein and Hilton 1992; Hilton and Fein 1989; Nisbett et al. 1981), the supportive

information was always presented first in experiment 1. However, one could argue that

inserting the irrelevant information between the supportive information and the belief










measure reduced the salience of the supportive information at the time of the belief

elicitation. While the first experiment demonstrated that, at the end of the experiment,

the accessibility of the supportive information was not reduced by the addition of

irrelevant information, it is still possible that the specific ordering of the information

temporarily reduced its accessibility when the belief measure was presented. To test this

possibility, experiment 1A manipulated the order in which the information was presented.

In the supportive-first condition, the supportive information was always presented before

the irrelevant information, while in the supportive-last condition, the supportive

information was always presented after the irrelevant information. If the dilution effect

was due to the reduced salience of the supportive information at the time of the belief

elicitation, then the irrelevant information should not influence subjects' beliefs when it

is presented before the supportive information.

Subjects and design. Subjects were 131 undergraduate students who participated

in return for class credit. The design was a 2 (type of information) by 2 (order of

information) by 2 (counterbalancing factor) by 8 (product replicates) mixed design. Each

subject was presented with descriptions of eight different products. For each of the

product replicates, subjects were randomly assigned to either the baseline condition or the

treatment condition. The order of information and the counterbalancing factor were

manipulated between subjects.

Stimuli and procedure. The product information used was the same information as

used in experiment 1, with the exception of eight additional pieces of supportive

information. These pieces of supportive information were pretested with 30










undergraduate students. The eight additional pieces of information were classified as

supportive (rather than irrelevant or counterdiagnostic) by an average of 93% of the

pretest subjects. The presentation of the target product was identical to the procedure

used in experiment 1. For each product category, subjects were first informed of the

desired benefit, and then received the description of the target product ("product A"),

which contained either only one piece of supportive information or also three pieces of

irrelevant information. The information for the alternative product ("product B"), which

always consisted of one piece of supportive information, was then displayed below the

description of the target product. The supportive information used for the two products

was counterbalanced. For half of the subjects, the target product descriptions listed the

supportive information used in experiment 1, while the descriptions of the alternative

product used the eight new pieces of supportive information. For the other half of the

subjects, this assignment was reversed. When the information for both products had been

displayed, subjects were asked to indicated which of the two products was more likely to

deliver the desired benefit. Subjects indicated their choice by clicking on a button with

the corresponding product label.

Results and discussion. When both product descriptions only contained

supportive information, 53% of subjects indicated that the target product was more likely

to deliver the benefit, compared to 47% who selected the alternative product. This

difference was not significant (Z= 1.59, ns), as would be expected since the supportive

information was counterbalanced. However, when the irrelevant information was added

to the description of the target product, the proportion of subjects who selected the target










product dropped significantly to less than 38% (2 = 26.26, df= 1, p < .001). While

subjects were, on average, indifferent between the target product and the alternative

product when both were described using only supportive information, the majority of

subjects perceived the alternative product as more likely to deliver the benefit when the

target product description contained additional irrelevant information (Z = 5.56, p <

.001). The effect of the irrelevant information did not depend on the order in which the

product information was being presented (x2 = 1.81, df= 1, ns). Adding the irrelevant

information decreased the proportion of subjects selecting the target product both when

the supportive information was presented first (2 = 5.38, df= 1, p < .05), and when it

was presented last (2 = 24.13, df = 1,p < .001).

Experiment 1A thus demonstrates that adding irrelevant product information not

only weakens beliefs about isolated products, but also reduces the likelihood that a

product is perceived as more likely to deliver a benefit than an alternative product.

Moreover, the dilution effect does not depend on the order in which the product

information is presented, thus refuting that the effect is caused by the reduced salience of

the supportive information at the time of the decision.















CHAPTER 4
EXPERIMENT 2: A TEST OF THE AVERAGING ACCOUNT


The first experiments demonstrated the robustness of the dilution effect and

provided evidence against explanations of the effect based on the distraction of attention

by the irrelevant information. The following experiments will examine the biased

hypothesis testing account of the effect by systematically testing it against each of the

existing explanations. Experiment 2 provides such a test between the biased hypothesis

testing account and the averaging explanation.

A first way to test between these two explanations is to examine the effect of

adding less supportive, rather than irrelevant information. Less supportive information is

information that suggests that the product will deliver the benefit, but is not as strong as

the original information. For example, the fact that a computer has a well-known brand

name suggests that it may be fast, but it is not as convincing as the fact that it has a

powerful processor. According to the averaging model proposed by Lichtenstein et al.

(1975), adding less supportive product information to strongly supportive information

should also dilute product beliefs. The less supportive information by itself should result

in judgments that are less extreme than judgments based on the strongly supportive

information. Averaging these separate judgments should result in an overall judgment

that lies between these two judgments and is less extreme than a judgment based only on

the supportive information.












HI: Averaging: Adding less supportive information to strongly supportive benefit

information will weaken consumers' beliefs in the product benefit.



On the other hand, the biased hypothesis testing account predicts that adding less

supportive information to strongly supportive information should lead to more extreme

judgments. This account assumes that consumers are looking for information that

confirms the hypothesis that the product will deliver the benefit. Since the less

supportive information suggests that the product will deliver the benefit, it will be

classified as "confirming," rather than "not confirming," and it will strengthen

consumers' belief in the hypothesis that the product will deliver the benefit.



H2: Biased Hypothesis Testing: Adding less supportive information to strongly

supportive benefit information will strengthen consumers' beliefs in the product

benefit.



The averaging explanation can also be tested by examining how the addition of

irrelevant information affects consumers' sensitivity to the diagnostic information.

According to the averaging model, "... since the weights must sum to one, adding a new

relevant stimulus to a set will cause the weights of the old stimuli to decrease"(Anderson

1974, p. 239). Since the irrelevant information must have a weight that is significantly

greater than zero to account for the dilution effect, adding irrelevant information should










reduce consumers' sensitivity to the supportive information. Or, in other words, the

irrelevant information dilutes the impact of the diagnostic information. This can be tested

by comparing the difference in consumers' belief ratings based on "supportive" versus

"less supportive" information to the difference in belief ratings based on "supportive +

irrelevant" versus "less supportive + irrelevant" product information. If consumers rely

on an averaging strategy, the manipulation of the degree of support should have less

influence on consumers' belief in the benefit when the irrelevant information is present

than when it is not.



H3: Averaging: The difference in consumers' responses to strongly supportive versus

less supportive benefit information will be less pronounced when irrelevant

information is added to the product descriptions.



On the other hand, the biased hypothesis testing account predicts that, instead of

diluting the impact of the diagnostic information, the irrelevant information will have an

independent diluting effect on consumers' beliefs. According to this explanation, the

irrelevant information is classified as "not confirming," and this classification directly

weakens consumers' belief in the hypothesis that the product will deliver the benefit.

Therefore, the irrelevant information will not decrease consumers' sensitivity to the

supportive information.










H4: Biased Hypothesis Testing: The difference in consumers' responses to strongly

supportive versus less supportive benefit information will not be affected by the

addition of irrelevant information to the product descriptions.



Method

Subjects and design. Subjects were 58 undergraduate students who participated in

return for class credit. The design was a 5 (type of information) by 8 (product replicates)

mixed design. Each subject was presented with descriptions of eight different products or

services. For each of the product replicates, subjects were randomly assigned to the

supportive information condition (i.e., the baseline condition in experiment 1), the

supportive + irrelevant information condition (i.e., the treatment condition in experiment

1), the less supportive information condition, the less supportive + irrelevant information

condition, or the supportive + less supportive information condition.

Stimuli and procedure. The stimulus set used in experiment 2 included all stimuli

used in experiment 1 and added three pieces of less supportive information for each

replicate. A pretest was conducted to select the 24 less supportive facts. Thirty subjects

were asked to classify different pieces of information for each replicate as either

supportive of the benefit, counterdiagnostic of the benefit, or not helpful for the decision.

The objective was to select information that was perceived as supportive of the benefit by

the majority of subjects, yet was not as strong as the original supportive information. On

average, the less supportive information was classified as suggesting the benefit by 71%

of the pretest subjects. The design of the actual experiment included a manipulation










check to test whether the less supportive information was indeed weaker than the original

supportive information.

The procedure was identical to the one used in experiment 1, with the exception

that three information conditions were added to the design. In the supportive + less

supportive condition, the product description contained one piece of supportive

information (e.g., "Very Powerful Processor") followed by three pieces of less supportive

information (e.g., "Well-Known Brand Name", "64 Mbyte Working Memory", "32-Speed

CD-ROM"). In the less supportive condition, the product description only consisted of a

single piece of less supportive information, which was randomly drawn from the three

pieces of less supportive information selected for that replicate. This randomization was

performed independently for each subject. Finally, in the less supportive + irrelevant

information condition, the product description contained one randomly selected piece of

less supportive information, which was followed by three pieces of irrelevant

information.



Results and Discussion

The results are summarized in Figure 1. First, subjects' beliefs in the product

benefit were weakened when irrelevant information was added to either supportive

information (X sP = 5.76, X SUPP+,= 5.17; F(1,424) = 6.23, p < .05) or less supportive

information (X LESSPP = 4.76, X LESS SUPP = 4.00; F(1,424)= 5.87,p < .05). Thus, the

dilution effect observed in the first experiments was replicated in experiment 2.

Moreover, these results demonstrate that irrelevant product information also dilutes










consumers' belief when the original product information is only weakly supportive of the

product benefit.

Second, subjects' beliefs in the product benefit were stronger when the product

information they had received was supportive ( X= 5.76) rather than less supportive

(X= 4.76; F(1,424) = 12.91, p < .001). This confirms that the manipulation of the

degree of support was successful.




7.5 7.26

S6.5
5.76
S5.5 5.17
4.76
4.5 4.00

3.5
Supportive+ Supportive Supportive+ Less Supportive Less Supportive
Less Supportive Irrelevant + Irrelevant
Type of Information


FIGURE 1
EXPERIMENT 2: EFFECT OF THE TYPE OF INFORMATION
ON SUBJECTS' BELIEF IN THE PRODUCT BENEFIT


Third, subjects receiving both supportive and less supportive information reported

more extreme judgments (X= 7.26) than subjects receiving only supportive information

(X= 5.76; F(1,424)= 22.42, p < .001). While adding irrelevant information to

supportive information weakened subjects' belief in the product benefit, adding less

supportive information to supportive information actually strengthened subjects' belief in










the benefit. This polarization effect is inconsistent with the averaging model (H1), but

consistent with the biased hypothesis testing explanation (H2), which predicts that

consumers will strengthen their belief in the benefit after classifying the less supportive

information as "confirming" the hypothesis. However, while this result clearly indicates

that subjects are not simply averaging separate predictions based on each piece of

information, it can be accounted for by averaging models that include an initial

impression (e.g., Anderson 1967; Lopes 1987). These latter models propose that people's

averaging calculus does not only include separate predictions based on each piece of

explicitly provided information, but also includes a "baseline prediction" which reflects

the decision maker's initial impression prior to receiving any information. If subjects'

initial impression is very low, then averaging this impression and the prediction based on

the supportive evidence could indeed result in weaker beliefs than would averaging this

impression, the prediction based on the supportive evidence, and the three predictions

based on the pieces of less supportive evidence.

To address this alternative account, I conducted a second test of the averaging

explanation by examining how the addition of irrelevant information influenced subjects'

sensitivity to the supportive information. Even averaging models that include an initial

impression (e.g., Anderson 1967) predict that adding information that receives a non-zero

weight in the decision should reduce people's sensitivity to the original information.

Since the irrelevant information weakens subjects' belief, it should receive a non-zero

weight in the averaging model, and thus reduce the weight of the other, diagnostic

information. However, the results demonstrate that the addition of irrelevant information










does not reduce the impact of the degree of support manipulation. Interestingly, the

difference between the supportive information condition and the less supportive

information condition becomes even slightly more pronounced when the irrelevant

information is also present (DNO IRR = 1.00, DIRR = 1.17; F< 1, ns). In sum, these

results are inconsistent with an averaging account of the dilution effect, according to

which the irrelevant information reduces the impact of the diagnostic information (H3),

but is consistent with the biased hypothesis testing account which states that the

irrelevant information directly affects the belief measure by being classified as "not

confirming" (H4).

This last result not only provides evidence against an averaging account of the

dilution effect, it also provides additional evidence against the distraction of resources

account, and, more importantly, provides some insight into the nature of the dilution

effect. While many researchers have assumed that the irrelevant information dilutes the

impact of the diagnostic information (e.g., Fein and Hilton 1992; Hilton and Fein 1989),

these findings suggest that the irrelevant information has an independent, direct effect on

consumers' beliefs. The irrelevant product information does not change the impact of the

diagnostic information, but directly reduces consumers' belief that the product will

deliver the desired benefit.















CHAPTER 5
EXPERIMENT 3: A TEST OF THE CONVERSATIONAL NORMS ACCOUNT


The previous experiments demonstrated the robustness of the dilution effect and

provided evidence that was inconsistent with distraction and averaging accounts of the

effect, but consistent with the biased hypothesis testing explanation. Experiments 3 and

3A will pit the biased hypothesis testing explanation against the conversational norms

account of the dilution effect. According to the conversational norms account, the

dilution effect may result from subjects' misguided reliance on the maxim of relation,

which states that communications have to be relevant for the goal of the conversation

(Grice 1975). Subjects may assume that all the information provided by the experimenter

has to be relevant for the judgment they are asked to make and may therefore use all the

available information to arrive at a decision. Thus far, evidence for this alternative

hypothesis is mixed. First, conversational norms explain why people rely on the

irrelevant information, but do not explain why they use it to dilute rather than polarize

their judgments. In fact, the third pretest of the first experiment showed that the majority

of subjects classified the irrelevant information as diagnostic of the benefit when forced

to regard it as relevant. Thus, if subjects in the previous experiments perceived the

irrelevant information as relevant, they were more likely to perceive it as supportive than

counterdiagnostic. Since experiment 2 demonstrated that the addition of less supportive

information tends to result in more extreme judgments, this suggests that treating










the irrelevant information as relevant should have lead to polarization rather than

dilution. Furthermore, subjects in each of the preceding experiments were explicitly told

that the information they received "may or may not be helpful for the decision that you

have to make". This should have indicated to subjects that the common principle of

relevance did not apply to the experimental setting. While these arguments are far from

conclusive, they do cast doubt on a conversational norms account of the observed dilution

effects.

Nevertheless, it is important to provide a direct test of the conversational norms

explanation in this product judgment context. First, many researchers have endorsed a

conversational norms explanation of the dilution effects observed in the social judgment

literature (e.g., Schwarz et al. 1991; Simonson et al. 1994; Slugoski and Wilson 1998;

Tetlock and Boettger 1989; Tetlock et al. 1996). Moreover, there is evidence that

subjects in these experiments indeed rely on conversational norms and that this

contributes to the dilution effect (e.g., Schwarz et al. 1991; Tetlock and Boettger 1989;

Tetlock et al. 1996). Finally, and most importantly, no research to date has directly tested

this explanation by examining whether subjects in dilution experiments perceive the

additional information as irrelevant. Experiment 3 will provide such a test by asking

subjects whether the additional information is relevant before they state their belief in the

product benefit.

On the one hand, one could argue that, if subjects in the dilution studies rely on

the relevance principle, they should classify the additional information as supportive or

counterdiagnostic. However, by asking subjects to specify the relevance of the










information, the procedure also reveals that the traditional guarantee of relevance does

not apply. Hence, subjects are likely to behave as subjects in the protests and classify the

information as irrelevant. How would the dilution effect be influenced by subjects'

acknowledgement of the irrelevance of the information immediately before stating their

belief in the benefit? According to the conversational norms explanation, subjects who

have just acknowledged the irrelevance of the information should realize that the

relevance principle does not hold and simply ignore the irrelevant information, thus

eliminating the dilution effect.



H5: Conversational Norms: Adding irrelevant product information to supportive

benefit information will not affect consumers' belief in the benefit when

consumersacknowledge the irrelevance of the information.



According to the biased hypothesis testing account, on the other hand, subjects in

the earlier experiments were also aware that the information was irrelevant, but

automatically classified it as "not confirming" in their search for hypothesis confirming

evidence. Thus, having consumers acknowledge the irrelevance of the information

should not influence the dilution effect.



H6: Biased Hypothesis Testing: Adding irrelevant product information to supportive

benefit information will weaken consumers' beliefs in the benefit, even when

consumers acknowledge the irrelevance of the information.










Method

Subjects and design. Subjects were 47 undergraduate students who participated in

return for class credit. The design was a 2 (type of information) by 12 (product

replicates) mixed design. Each subject was presented with six different products or

services. For each of the product replicates, subjects were randomly assigned to one of

two information conditions, either the supportive information condition or the supportive

+irrelevant information condition.

Stimuli and procedure. The stimulus set used in experiments 1 and 2 was

expanded with 4 additional replicates: hotel (luxurious), mountain bike (sturdy), movie

(action-packed), and printer (high graphic quality). A pretest (n = 30) was conducted to

select the product information for each additional replicate. The twelve irrelevant facts

were classified as "not helpful" by an average of 87% of the pretest subjects. The four

supportive facts were classified as suggesting the benefit by an average of 96% of the

pretest subjects. Each subject was only exposed to six randomly selected replicates. Two

filler descriptions were added to these six target replicates. The fillers were inserted in

the second and fifth position and served to make the pre-programmed structure of the

stimuli less obvious.

Subjects received the same instructions as those who participated in experiment 1.

The procedure was identical to the procedure followed in the first experiment, except for

the presence of intermediate questions. After each piece of information had been

presented, the following question was displayed: "What does this particular piece of

information tell you about this [product]?" Subjects could respond by clicking on one of










three buttons labeled "That it is [benefit]," "That it is NOT [benefit]," and "This

information is not helpful here." After all pieces of information had been presented and

evaluated, subjects rated their belief in the product's ability to deliver the benefit on the

same 9-point scale that was used in experiments 1 and 2.



Results and Discussion

Consider first the classification of the product information. The irrelevant

information was classified as "not helpful" on 360 of 390 occasions (92%). On 15

occasions (4%) it was classified as diagnostic and on 15 other occasions (4%) it was

classified as counterdiagnostic. This indicates that, on average, subjects clearly perceived

the additional information as irrelevant. While this result does not necessarily imply that

the additional information was also perceived as irrelevant in the previous dilution

studies, it does allow us to determine the effect of the irrelevant information when its

irrelevance is made very explicit immediately prior to the belief ratings.

The analysis of the belief ratings reveals that, even after subjects have

acknowledged the irrelevance of the additional information, the irrelevant information

still dilutes their belief in the product benefit. Subjects' belief ratings were more extreme

when the product description only contained supportive information (X= 6.84) than

when it also contained irrelevant information (X= 5.54, F(1,252) = 51.89, p < .01),

consistent with the biased hypothesis testing account (H6), but inconsistent with the

conversational norms explanation (H5). Most importantly, even when the analysis was

restricted to trials on which all three pieces of information were classified as irrelevant,










the additional information still weakened subjects' beliefs in the product's ability to

deliver the benefit (X SUPPORTIVE = 6.84, XIRRELEVANT = 5.49, F(1,228) = 50.79, p < .01).



Experiment 3A

The results of experiment 3 clearly indicate that subjects perceive the additional

information as irrelevant and that they still display the dilution effect when this

irrelevance is highly salient. Although these findings cast doubt on the conversational

norms explanation of the observed dilution effect, they cannot rule it out completely.

Indeed, the fact that someone chooses to express an irrelevant assumption may itself be

highly relevant (Sperber and Wilson 1986, p.121). Subjects may assume that the

experimenter must have provided the irrelevant information for some reason and

therefore use the information despite its apparent irrelevance. To address this

interpretation, an additional study (n = 21) was conducted in which subjects were

informed that the information was randomly selected by a computer. This procedure is

similar to manipulations used in previous studies that have found support for the

conversational norms explanation (e.g., Tetlock et al. 1996). Subjects were told that the

information was being randomly sampled by the computer and that, consequentially,

some information would be helpful, while other information would not. Moreover,

before each piece of information was retrieved, the message "Randomly Drawing

Information" was displayed, as well as a rapidly filling clock, thus reminding subjects

that the computer was sampling information. Otherwise, the procedure was identical to

the procedure used in experiment 1, but using the expanded stimulus set of experiment 3.










The results showed that adding irrelevant product information to supportive information

again weakened subjects' beliefs in the product's ability to deliver the benefit (X SUPP =

6.95, XSUPP+IRR = 5.94, F(1,321) = 8.83,p < .01). Together, the results from studies 3

and 3A rule out a conversational norms explanation of the dilution effects observed in the

current research. The dilution effect still occurs when subjects are explicitly aware of the

irrelevance of the additional information and when they assume that the information is

being randomly sampled by a computer. This does not only indicate that the dilution

effect is not caused by a reliance on conversational norms, but also informs us about the

range of situations in which the dilution effect may be observed. These results suggest

that irrelevant product information may still weaken consumers' product beliefs when it

is encountered in a mass advertising situation, in which consumers realize that the

information is not necessarily intended for them and that the maxims of intentional

communication may therefore not apply.















CHAPTER 6
EXPERIMENT 4: A TEST OF THE REPRESENTATIVENESS ACCOUNT


While the results of the preceding experiments are inconsistent with averaging

and conversational norms accounts of the dilution effect, they do confirm the predictions

of the biased hypothesis testing explanation. However, these results can also be

accounted for by consumers' reliance on a representativeness heuristic. Consumers may

assess their belief in the product benefit by relying on the similarity between the

described product and the typical desired product. Irrelevant information may reduce the

similarity with this prototype, while less supportive information may enhance it.

Furthermore, similar to biased hypothesis testing, a reliance on the representativeness

heuristic also predicts that the dilution effect will persist when subjects are very aware of

the irrelevance of the information, and when the information seems to be randomly

selected. Moreover, the addition of the irrelevant information directly reduces the

similarity of the product description to the typical desired product. Therefore, the

representativeness account also predicts that the irrelevant information directly affects

consumers' beliefs in the product benefit, without reducing consumers' sensitivity to the

diagnostic information. Since both explanations can account for this entire pattern of

results, experiment 4 was designed to test between these two competing accounts.










To test between these explanations, the product information for experiment 4 was

selected so that adding the irrelevant information to the supportive information would

actually increase the similarity to the typical desired product, rather than reduce it. In the

previous studies, the addition of irrelevant information tended to reduce the similarity

between the product description and the typical desired product. For instance, an

apartment with 24 hour on-site security is more similar to the typical safe apartment, than

an apartment with 24 hour on-site security, a 40-year old manager, 1 and 2 bedroom

apartments, and named "Haywood Park." However, in this experiment, the product

information was selected so that adding irrelevant information would increase the

similarity with the typical desired benefit. For instance, an apartment with surveillance

cameras above all doors is less similar to the typical safe apartment than an apartment

with surveillance cameras above all doors, a fully equipped club house, a great fitness

center, and monthly pest control included in the rent. If adding this irrelevant product

information indeed increased the perceived typicality of the product description, and if

consumers are relying on a representativeness heuristic to predict the product's ability to

deliver the desired benefit, then adding irrelevant information should strengthen rather

than weaken consumers' beliefs in the product benefit.

Furthermore, the experiment also contained two conditions in which subjects

were encouraged to either follow a biased hypothesis testing strategy or a

representativeness strategy. By comparing these conditions to a condition in which

subjects were not directed to one particular strategy, it could be tested whether subjects

spontaneously behaved as if they were engaging in biased hypothesis testing, or if they










spontaneously behaved as if they were relying on a representativeness heuristic. This

resulted in four strategy conditions: a typicality condition, afree strategy condition, a

biased hypothesis testing condition, and a representativeness condition.

In the typicality condition, subjects did not have to rate their belief in the product

benefit, but were instead asked to rate the similarity between the product description and

the typical desired product. It was assumed that adding the specially selected irrelevant

information would increase the perceived similarity with the typical desired product. In

thefree strategy condition, subjects were presented with the same product descriptions

and, as in previous experiments, asked to state their belief in the product benefit. In the

biased hypothesis testing condition, subjects were first asked to classify each piece of

information as either "confirming" or "not confirming" that the product will deliver the

benefit, similar to the process proposed by the biased hypothesis testing account. After

classifying all pieces of information, subjects rated their belief in the product benefit.

Finally, in the representativeness condition, subjects were first presented with the entire

product description, and were then asked to rate the similarity between this description

and the typical desired product. Following the similarity rating, these subjects also rated

their belief in the product benefit.

If subjects are relying on a representativeness heuristic, and if the irrelevant

information does indeed increase the perceived typicality of the product, then adding

irrelevant information in thefree strategy condition should lead to polarization instead of

dilution. Moreover, when subjects are encouraged to rely on their typicality judgments to

assess their belief in the product benefit, they should produce belief ratings that are










similar to those observed in thefree strategy condition. Thus, if subjects are

spontaneously relying on a representativeness heuristic, there should be no difference

between the effect of irrelevant information in thefree strategy condition and its effect in

the representativeness condition.



H7: Representativeness: Adding irrelevant product information that enhances the

product's similarity to the typical desired benefit will strengthen consumers'

beliefs in the product benefit, regardless of whether consumers are encouraged to

use similarity ratings as a basis for their belief in the product's ability to deliver

the benefit.



However, if the proposed biased hypothesis testing mechanism applies, adding

irrelevant information should dilute product beliefs, regardless of its effect on the

perceived typicality of the product description. Even the highly typical irrelevant

information will be classified as "not confirming," resulting in weaker beliefs in the

product benefit. Furthermore, when subjects are asked to classify all information as

either "confirming" or "not confirming," they should produce belief ratings that are

similar to those observed in thefree strategy condition. Thus, if subjects are

spontaneously using a biased hypothesis testing strategy, there should be no difference

between the effect of irrelevant information in thefree strategy condition and its effect in

the biased hypothesis testing condition.










H8: Biased Hypothesis Testing: Adding irrelevant product information that enhances

the product's similarity to the typical desired benefit will weaken consumers'

beliefs in the product benefit, regardless of whether consumers are encouraged to

use a biased hypothesis testing strategy to assess the product's ability to deliver

the benefit.



Method

Subjects and design. Subjects were 83 undergraduate students who participated in

return for class credit. The design was a 2 (type of information) by 4 (strategy

conditions) by 8 (product replicates) mixed design. Each subject was presented with

eight different products or services. For each of the product replicates, subjects were

randomly assigned to either the supportive information condition or the supportive +

irrelevant information condition. The strategy factor was manipulated between subjects.

Stimuli and procedure. To ensure that adding the irrelevant product information

would increase the perceived typicality of the product description, I selected atypical

supportive information (e.g., a computer with "revolutionary triple processors") and

typical irrelevant information (e.g., "includes DVD player"). While the triple processors

suggest that the computer is fast, they are not part of the representation of a typical fast

computer. On the other hand, the DVD player does not affect the speed of the computer,

but does enhance the similarity to the typical fast computer. Subjects' perception of the

new information was measured using the same pretest as used in experiment 1 (n = 40).

The eight supportive facts were classified as suggesting the benefit by an average of 93%










of pretest subjects. The 24 irrelevant facts were classified as "not helpful" by an average

of 93% of the subjects, as suggesting the benefit by 5% of subjects, and as

counterdiagnostic of the benefit by only 2% of subjects. The typicality condition in the

actual experiment tested whether this irrelevant information did indeed increase the

perceived similarity to the typical desired product.

The procedure in thefree strategy condition was identical to the one used in

experiment 1. For each replicate, subjects were told that they were looking for a specific

benefit, followed by the product information, which was sequentially presented on the

screen. When the entire product description had been displayed, subjects were asked to

indicate their belief in the benefit on a 9-point scale. In the typicality condition, subjects

received the same product information, but before the product description appeared, the

statement "try to imagine a typical [desired product]" appeared at the top of the screen.

After all the product information had been displayed, subjects were asked to rate the

similarity between the product description and the typical desired product on a 9-point

scale anchored by "not similar at all" and "very similar". The representativeness

condition was identical to the typicality condition, with the exception that subjects were

also asked to indicate their belief in the product benefit after providing the similarity

rating. Finally, in the biased hypothesis testing condition, subjects were asked, for each

piece of information that appeared on the screen, whether "this particular piece of

information indicates that this [product] is [benefit]." They could respond by clicking on

buttons labeled "yes" and "no". After the entire product description had been displayed,










subjects indicated their belief in the product benefit, but did not make any similarity

judgments.



Results

The results are summarized in Figure 2. As intended, adding the irrelevant

product information to the supportive information increased the perceived similarity

between the product description and the typical desired product (X SUpp = 5.16,

X SUPP+IRR = 6.51; F(1,600) = 15.83, p < .001). However, while adding the irrelevant

information increased the perceived typicality of the product description in the typicality

condition, it still weakened subjects' beliefs in the product benefit in thefree strategy

condition (X supp = 6.59, X SUPP+IRR = 5.55; F(1,600) = 9.34, p < .01). This is

inconsistent with the prediction of the representativeness account (H7), but does confirm

the prediction of the biased hypothesis testing perspective (H8). An additional test

between these two theories is provided by comparing the forced strategy conditions to the

free strategy condition.











7.5
o- 6.51 6.59
o 6.5 6.24 6.3 6.18
co5.55
.J 55.32
0- w 5.5 5.16

4.5 I I,
Typicality Free Strategy Represent. Hypothesis
Testing
Strategy Condition

OSupportive OSupportive + Irrelevant




FIGURE 2
EXPERIMENT 4: EFFECT OF STRATEGY AND TYPE OF INFORMATION
ON SUBJECTS' BELIEF IN THE PRODUCT BENEFIT OR THE PERCEIVED
TYPICALITY OF THE PRODUCT DESCRIPTION


If subjects in thefree strategy condition are indeed using the proposed biased

hypothesis testing strategy, then they should behave similarly to subjects who are

explicitly encouraged to follow this strategy. On the other hand, if subjects in thefree

strategy condition are relying on the representativeness heuristic, they should behave

similarly to those subjects who are making explicit similarity judgments before rating

their belief in the benefit. The results show that the addition of irrelevant product

information weakened subjects' beliefs ratings in the biased hypothesis testing condition

(X SUPP = 6.18, X SUPP+IRR = 5.32; F(1,600) = 7.56, p < .01), and that this dilution

effect was not significantly different from the effect observed in thefree strategy

condition (F < 1, ns). However, the addition of irrelevant information did not affect

subjects' belief ratings in the representativeness condition ( X SUPP = 6.24, X SUPP+IRR










= 6.30; F < 1, ns), resulting in an effect that was significantly different from the dilution

effect observed in thefree strategy condition (F(1,600) = 4.50, p < .05). Thus, subjects

in thefree strategy condition seemed to react to the irrelevant information in a similar

way as those in the biased hypothesis testing condition, but differently from those in the

representativeness condition, indicating that, while these subjects may have been using a

biased hypothesis testing strategy, they were most likely not relying on the

representativeness heuristic.



Discussion

The findings of experiment 4 demonstrate that adding irrelevant information still

weakens consumers' belief in the product benefit when this information actually

increases the perceived similarity between the product description and the typical desired

product. Moreover, consumers who choose their own strategy for assessing whether a

product will deliver a benefit behave similarly to those who are encouraged to follow a

biased hypothesis testing strategy, but different from those who are encouraged to rely on

similarity ratings to determine their beliefs in the product benefit. These findings

therefore indicate that the diluting effect of irrelevant product information reported in

these studies is not due to subjects' reliance on a representativeness heuristic. Instead,

these findings are consistent with the biased hypothesis testing account of the dilution

effect.















CHAPTER 7
EXPERIMENT 5: PROCESSING THE INFORMATION
WITHOUT THE BENEFIT IN MIND


The results observed in the previous experiments are inconsistent with distraction,

averaging, conversational norms, and representativeness accounts of the dilution effect,

while they are consistent with the biased hypothesis testing explanation. On the other

hand, the previous experiments failed to directly test some of the essential characteristics

of the biased hypothesis testing mechanism. The following experiments will therefore

examine some direct implications of the proposed explanation. In experiments 5 and 5A,

I will manipulate whether consumers initially process the information with the hypothesis

in mind. Experiment 6 will examine whether the dilution effect disappears when

consumers consider the implications of the product information for both the focal and the

alternative hypothesis. Finally, experiment 7 will test whether the effect reverses when

consumers start out with the hypothesis that the product will not deliver the desired

benefit, instead of the hypothesis that the product will deliver the benefit.

In the previous studies, as well as in virtually all previous demonstrations of the

dilution effect, subjects knew the outcome that had to be predicted prior to processing the

evidence. They could therefore engage in goal-oriented, top-down processing of the

information. The fifth experiment will replicate this top-down scenario, but will also add

a condition in which the desired benefit is only revealed after subjects have read the

product description, forcing subjects to first process the product information without the










benefit in mind, i.e., in a bottom-up fashion. According to the biased hypothesis testing

explanation, subjects in the top-down condition will (1) process the product description

while searching for information that supports the hypothesis that the product will deliver

the benefit, (2) classify the irrelevant information with regard to their search goal as "not

confirming", and (3) weaken their belief in the hypothesis. However, subjects in the

bottom-up condition will (1) process the product description without a specific search

goal, (2) learn about the hypothesis, and (3) redirect their attention to those pieces of

information that may confirm the hypothesis. Because subjects in the bottom-up

condition have already processed the product information, they can engage in a more

efficient search for supportive evidence. They can immediately focus on the supportive

evidence and ignore information that is obviously irrelevant. While they have still

processed the irrelevant information, they have not used it to evaluate any hypothesis,

hence dilution should not occur.



H9: Biased Hypothesis Testing: Adding irrelevant product information to supportive

benefit information will weaken consumers' beliefs in the product benefit when

the information is first processed with the benefit in mind, but not when it is first

processed without the benefit in mind.



A second objective of experiment 5 was to examine the robustness of the

polarizing effect of the less supportive information observed in the second experiment.

To this end, the experimental design also contained conditions with less supportive










information. It was expected that the polarization effect would be replicated in both the

top-down and bottom-up conditions. Even when the less supportive information has first

been processed without the benefit in mind, it should be revisited in the search for

confirming evidence once the desired benefit has become known. It will then be

classified as "confirming," and strengthen consumers' belief in the hypothesis that the

product will deliver the desired benefit.



Method

Subjects and design1. Subjects were 57 undergraduate students who participated

in return for class credit. The design was a 4 (type of information) by 2 (processing

mode) by 12 (product replicates) mixed design. Each subject was presented with 8

different product descriptions out of a total of 12 product replicates. For each of the

product replicates, subjects were randomly assigned to either the supportive, supportive +

irrelevant, less supportive, or supportive + less supportive information condition. The

processing strategy (bottom-up or top-down) was manipulated between subjects.

Stimuli and procedure. The same 12 replicates used in experiment 3 were also

used in experiment 5. However, an additional pretest (n = 30) was conducted to select

less supportive information for the four replicates not used in experiment 2. The twelve

additional pieces of less supportive information were perceived as suggesting the benefit


1 The original design of the experiment also contained a top-down reversed condition.
This condition was identical to the top-down condition, with the exception that the
irrelevant information was presented first, followed by the supportive information. The
results obtained in this condition did not differ from those in the top-down condition,
demonstrating that the effects do not depend on the order in which the information is
presented.










by 87% of respondents. The procedure in the top-down condition was identical to the

procedure used in experiment 1, while the procedure in the bottom-up condition differed

in that the desired benefit was not displayed before the product description appeared.

Instead, the instructions at the beginning of the experiment informed subjects that they

would have to evaluate each product "on a certain dimension". As in the previous

experiments, the separate pieces of product information appeared sequentially and all

information remained on the screen when the belief measure appeared.



Results and Discussion

The results are summarized in Figure 3. First, it can be seen that, across both

processing modes, product descriptions consisting only of supportive information lead to

stronger beliefs than descriptions consisting only of less supportive information (F(1,408)

= 15.68,p < .001). This effect was significant in the top-down condition (F(1,408) =

15.87, p < .001), and marginally significant in the bottom-up condition (F(1,408) = 2.24,

p = .13). Thus, the manipulation of the degree of support was successful. Second, the

addition of less supportive information increased the strength of subjects' product beliefs,

regardless of the manner in which the information was presented (F(1,408) = 13.00, p <

.001). There was a significant polarization effect of the less supportive information in

both the top-down condition (F(1,408) = 4.42, p < .05) and the bottom-up condition

(F(1,408) = 8.83, p < .01). The strength of this polarization effect did not depend on the

manner in which the information was being processed (F < 1, ns). These results attest to










the robustness of the polarizing effect of less supportive information observed in

experiment 2.



5 8 7.32 7.34
7 6.56 6.45
6.18
5.68 5.76
S6 5.37
5 5
4
Top-down Bottom-Up
Processing Mode
E Supportive l Supportive + Irrelevant
U Supportive + Less Supportive 0 Less Supportive

FIGURE 3
EXPERIMENT 5: EFFECT OF PROCESSING MODE AND
TYPE OF INFORMATION ON SUBJECTS' BELIEF IN THE PRODUCT BENEFIT


However, the main objective of this experiment is to examine how the dilution

effect is affected by the manner in which the product information is being processed. A

general dilution effect was observed when irrelevant information was added to the

supportive information (F(1,408) = 4.38, p < .05). However, this effect did depend on

the manner in which the information had been processed (F(1,408) = 4.46, p < .05). The

irrelevant information significantly diluted subjects' beliefs in the top-down condition

(F(1,408) = 7.42, p < .01), but not in the bottom-up condition (F < 1, ns). Thus,

consistent with the prediction of the biased hypothesis testing account, irrelevant

information only affects product beliefs when it is initially processed with the desired

benefit in mind. When consumers have already processed the irrelevant information










without the benefit in mind, they can selectively focus on the supportive information

when testing the hypothesis. Thus, these consumers never classify the irrelevant

information as "not confirming," and consequentially do not weaken their belief in the

hypothesis that the product will deliver the desired benefit.



Experiment 5A

The results from experiment 5 suggest that the dilution effect in the bottom-up

condition failed to occur because subjects' pre-processing of the information allowed

them to ignore the irrelevant information when searching for information that supported

the hypothesis. To directly test this assumption, experiment 5A examines a situation in

which the initial processing of the information would not inform subjects that the

information could not possibly be supportive for the to-be-revealed benefit. This would

force subjects to reconsider the information after the benefit was revealed, and thus

classify this information as "not confirming". In this case, the irrelevant information

should still weaken consumers' beliefs in the product benefit, even when the information

has initially been processed without the benefit in mind.

To create such a situation, the "obviously irrelevant" information used in the

previous experiments was replaced with information that had a high degree of "typical

diagnosticity." In other words, the selected information was often relevant in similar

decisions, but was clearly irrelevant for the benefit desired in this specific case. For

example, the fact that a computer has been assembled in the United States, can be ordered

online, and airs commercials on NBC is not only irrelevant for assessing the speed of the










computer, but it does not inform consumers about any other important benefits either. On

the other hand, the fact that a computer has high quality speakers, is loaded with games,

and has a flat screen monitor may not be relevant for assessing the speed of the computer,

but it is often relevant in typical computer purchase decisions. This latter type of

information can be called "pseudo-relevant" information or "typically diagnostic"

information (Hilton and Fein 1989; Yzerbyt et al. 1997). When consumers first process

this pseudo-relevant information, they cannot rule out that the information may support

the desired benefit. Thus, once they learn about the desired benefit, they will have to go

back to this information and evaluate whether or not it implies that the product will

deliver the benefit.

Subjects and design. Subjects were 51 undergraduate students who participated in

return for class credit. The design was a 2 (type of information) by 2 (processing mode)

by 8 (product replicates) mixed design. Each subject was presented with 8 different

product descriptions. For each replicate, subjects were randomly assigned to either the

supportive information condition or the supportive + pseudo-relevant information

condition. The processing strategy was manipulated between subjects.

Stimuli and procedure. Experiment 5A used a subset of the replicates used in

experiment 5: hotel (luxurious), movie (action-packed), car (sportive), apartments (safe),

package delivery service (fast)), frozen dinners (healthy), toothpaste (fights cavities), and

computers (fast). A first pretest (n = 29) was conducted to select the "pseudo-relevant"

information. The 24 pseudo-relevant facts were classified as "not helpful" for the benefit

judgment by an average of 90% of the pretest subjects. A second pretest (n = 24) was










conducted to compare the pseudo-relevant information used in this study to the irrelevant

information used in experiment 3. Subjects were asked to indicate to what extent each

piece of information was typically helpful for evaluating the product endpointss 1 = "Not

Helpful at All", to 7 = "Very Helpful"). The results showed that, across product

categories, the pseudo-relevant information was perceived as more helpful for evaluating

the products (X= 5.65) than was the obviously irrelevant information (X= 4.19;

F(1,1040) = 202.74, p < .01). The procedure was identical to the procedure used in

experiment 5, with the exception that the product descriptions either contained one piece

of supportive information, or one piece of supportive information and three pieces of

pseudo-relevant information.

Results and discussion. The results are summarized in Figure 4. There is a

significant main effect of the information manipulation. Adding pseudo-relevant

information diluted product beliefs (F(1,390) = 20.20, p < .01). This effect did not

interact with the manner in which the information was being processed (F < 1, ns). The

irrelevant information weakened product beliefs both when subjects processed the

information with the benefit in mind (F(1,390) = 5.27, p < .05) or when they processed

the information without the benefit in mind (F(1,390) = 11.72, p < .01).











S6.82
1 7.00
c 6.49
6.09
S5.87
6.00


5.00
Top-down Bottom-Up

Processing Mode

l Supportive Supportive + Pseudo-relevant


FIGURE 4
EXPERIMENT 5A: EFFECT OF PROCESSING MODE AND TYPE OF
INFORMATION ON SUBJECTS' BELIEF IN THE PRODUCT BENEFIT


The results from experiments 5 and 5A suggest that processing product

information prior to having a benefit in mind may inhibit dilution, but only under certain

conditions. When the additional information is obviously irrelevant, people know that

the irrelevant information is not possibly supportive when they learn about the desired

benefit, and they can restrict their search for supportive evidence to the actual supportive

information. In contrast, when additional information is "pseudo-relevant", subjects must

reconsider it when they learn about the desired benefit, since the information may be

supportive. Even though this "pseudo-relevant" information is not diagnostic with respect

to the benefit, the simple act of considering this information results in dilution. These

findings indicate that the dilution effect is the result of goal-directed processing and

provide further evidence for the biased hypothesis testing account of the dilution effect.















CHAPTER 8
EXPERIMENT 6: CONSIDERING BOTH THE FOCAL
AND ALTERNATIVE HYPOTHESES


Experiment 6 examines a second implication of the proposed biased hypothesis

testing mechanism. One assumption of this account is that irrelevant information dilutes

product beliefs because consumers only consider whether the information supports the

focal hypothesis, while ignoring whether the information supports the alternative

hypothesis. Whereas obviously irrelevant information does not confirm that the product

will deliver the benefit, it also does not confirm that the product will not deliver the

benefit. Therefore, the dilution effect should not occur when consumers consider the

implications of the irrelevant information for both hypotheses. Experiment 6 tests this

prediction by manipulating the number of questions subjects are asked about each

product description. In the single hypothesis condition, they are only asked to rate their

belief in the benefit, whereas in the dual 1iypo1,the, condition, subjects are also asked to

rate their belief that the product will not deliver the benefit. For example, subjects who

are looking for a fast computer do not only have to rate their belief that the described

model is fast, but also have to indicate their belief that the described model is slow.










Method

Subjects and design. Subjects were 112 undergraduate students who participated

in return for class credit. The design was a 2 (type of information) by 2 (number of

hypotheses) by 8 (product replicates) mixed design. Each subject was presented with

descriptions of nine different products, consisting of eight experimental replicates and

one filler replicate. For each product, subjects were randomly assigned to either the

supportive condition or the supportive + irrelevant condition.

Stimuli and procedure. The stimulus set was identical to the information used in

experiment 1, with the exception that one additional "practice category" was included as

the first description for each subject. This practice category was included to ensure that

subjects in the dual lol)pithewe condition would understand that they would have to rate

each product on both dimensions. The practice category consisted of a description of an

MBA program and always contained the same four pieces of information. Subjects'

responses for this replicate were not included in the analyses.

With the exception of the inclusion of this practice replicate, the procedure in the

single hypothesis condition was identical to the procedure used in experiment 1. In the

dual lpolIthew,\ condition, subjects were also first told that they were looking for a

particular benefit (e.g., "You are looking for a fast computer."), followed by a description

of the product, and the measure of subjects' belief in the benefit (e.g., "Is this computer

fast?"). However, unlike in the single hypothesis condition, this measure was followed

by a measure of subjects' belief in the opposite of the benefit (e.g., "Is this computer










slow?") on a similar 9-point scale (e.g., anchored by 1 = "definitely not slow" and 9 =

"definitely slow").



Results

The results are summarized in Figure 5. The addition of irrelevant product

information to supportive information again weakened subjects' belief in the product

benefit (F(1,864) = 6.42, p = .01). However, this dilution effect did depend on the

number of hypotheses subjects were evaluating (F(1,864) = 4.07, p < .05). When

subjects only rated their belief in the benefit, irrelevant information indeed diluted

subjects' beliefs (X SUPP = 6.59, X SUPP+IRR = 6.09; F(1,864) = 10.60, p < .01).

However, when subjects also indicated their belief that the product would not produce the

benefit, the irrelevant information did not influence their belief in the product benefit

(X SUpp = 6.44, X SUPP+IRR = 6.38; F < 1, ns). Furthermore, the irrelevant information

also did not influence their belief that the product would not deliver the benefit (X SUpp


3.13, X SUPP+IRR= 3.27; F< 1, ns).










e-
5 7.00 6.59
6.44 6.38
6.09
6.00


5.00
Single Hypothesis Dual Hypotheses

Number of Hypotheses

l Supportive u Supportive + Irrelevant



FIGURE 5
EXPERIMENT 6: EFFECT OF NUMBER OF HYPOTHESES AND TYPE OF
INFORMATION ON SUBJECTS' BELIEF IN THE PRODUCT BENEFIT


Discussion

Experiment 6 demonstrates that irrelevant product information does not influence

consumers' beliefs in the product benefit when consumers assess the likelihood that the

product will deliver the benefit, as well as the likelihood that the product will not deliver

the benefit. Thus, consistent with the biased hypothesis testing account, consumers'

unique focus on the implications for the hypothesis that the product will deliver the

benefit seems to be a necessary condition for the dilution effect to occur. The dilution

effect disappears when consumers also consider whether the evidence supports the

hypothesis that the product will not deliver the benefit.















CHAPTER 9
EXPERIMENT 7: EXPECTING THAT THE PRODUCT
WILL NOT DELIVER THE BENEFIT


The experiments so far have shown that adding irrelevant product information

will usually weaken consumers' beliefs in the product benefit. Yet, this does not imply

that brands should always avoid the communication of additional irrelevant information

if they want to emphasize a product benefit. If consumers follow the proposed biased

hypothesis testing process, then the confrontation with irrelevant information may

sometimes strengthen consumers' beliefs in the product's ability to deliver the benefit.

The first assumption of the proposed explanation states that consumers will

usually test the hypothesis that the product will deliver the benefit. However, this

assumption may not hold when consumers have a strong reason to suspect that the

product will not deliver the benefit. For instance, when the product carries a brand name

that has a very poor reputation on the critical dimension, consumers may set out to test

the hypothesis that the product will not deliver the benefit. They may then search

information that confirms this hypothesis (i.e., information that is counterdiagnostic of

the benefit) and classify information with regard to this search goal as "confirming" (i.e.,

counterdiagnostic of the benefit) or "not confirming" (i.e., not counterdiagnostic of the

benefit). Irrelevant information will be classified as "not confirming" and weaken

consumers' belief in the hypothesis that the product will not deliver the benefit. In other

words, irrelevant information will strengthen consumers' belief in the benefit.










Consider, for instance, a consumer who is looking for a trendy store and reads an

ad for a K-Mart store. The consumer may assume that this store is unlikely to be trendy

and search for information that confirms that the store is indeed not trendy. She may then

encounter information that is irrelevant with respect to the benefit (e.g., the store accepts

all major credit cards), which is not the type of information she was looking for, will be

classified as "not confirming," and reduce her confidence in the initial hypothesis that the

store was not trendy. In other words, the irrelevant information actually strengthened her

belief that the store is trendy. Experiment 7 will test this prediction by manipulating the

presence of a negatively perceived brand name.



Method

Subjects and design. Subjects were 68 undergraduate students who participated in

return for class credit. The design was a 2 (type of information) by 2 (presence of brand

names) by 7 (product replicates) mixed design with a brand name only baseline condition

that served as a control. Each subject was presented with three different product

descriptions from a total of seven product replicates. For each of the product replicates,

subjects were randomly assigned to either the supportive condition or the supportive +

irrelevant condition. The presence of brand names was manipulated between subjects.

Subjects in the no brand name condition only received product descriptions, while

subjects in the brand name condition received both brand names and product

descriptions. Subjects in the brand name only control condition received only brand

names and no product descriptions.










Stimuli and procedure. The stimulus set consisted of seven target categories

(products or services) and five filler categories. Only two of the target categories were

taken from previous experiments (hotel room and car)1. The other five categories were

either completely new or required changes in the product information: beer (great taste),

apartments (safe), clothing store (trendy), shampoo (high quality hair care), and

restaurant (healthy). A pretest (n = 36) was conducted to select one supportive fact and

three irrelevant facts for each of the new categories. The fifteen irrelevant facts were

classified as "not helpful" by an average of 87% of the pretest subjects, while the five

supportive facts were classified as suggesting the benefit by an average of 92% of

respondents.

The procedure used in the no brand name condition was similar to the one used in

experiment 1. The procedure in the brand name condition differed in some important

ways. First, subjects were asked to rate a set of brands in the seven target categories and

five filler categories2. In each product category, subjects were presented with four to

seven brand names and asked to indicate whether each brand would deliver a particular

benefit on a scale ranging from -3 (definitely not [benefit]) to +3 (definitely [benefit]).


1 The majority of the twelve replicates used in the previous experiments could not be used
here for two reasons. First, many of the categories were not associated with brands for
which subjects held clearly negative priors. Second, product information needed to be
compatible with whatever brand name subjects perceived as least likely to provide the
benefit.

2 This procedure was necessary because protests had shown that subjects not only
differed in their brand beliefs within a category, but also in their brand beliefs between
categories. While some subjects perceived some beer brands very negatively and gave
neutral evaluations to all shampoo brands, other subjects showed an exactly reversed
pattern.










After a filler task, subjects were exposed to information for five filler categories and three

target categories for which they had indicated strong negative beliefs for at least one of

the brands3. Subjects first received the name of the brand for which they had strong

negative priors, followed by the description of the product or service. The target category

descriptions either contained only supportive information or both supportive and

irrelevant information4. For instance, if the pretest indicated that a subject thought K-

Mart was not trendy, the subject would be presented with the instruction, "You are

looking for a trendy store. The store you are considering is K-Mart." This instruction

could be followed by supportive information (e.g., "Has announced the opening of a

Tommy Hilfiger section.") or both supportive information and three pieces of irrelevant

information (e.g., "Closes at 9 PM," "Major credit cards accepted," and "Airs

commercials on CBS and NBC"). In the brand name only condition, subjects did not

receive a product description and had to base their judgment on the brand name.



Results and Discussion

The results are summarized in Figure 6. First, a manipulation check showed that

adding supportive information made the negatively perceived brand more desirable ( X=

3.75) than presenting the brand name by itself (X= 1.61; F(1,129) = 35.29, p < .01).

The remainder of the analyses will concentrate on the 2 X 2 design manipulating the type


3 To maintain comparability, assignments of replicates in the no brand name condition
depended on the replicate selection in the brand name condition. This guaranteed that
the proportion of categories selected did not differ between conditions.










of information and presence of the brand name. First, there was a main effect of brand

name. As expected, the belief ratings were higher when subjects only received the

product description ( X= 5.11), than when they also received the negatively perceived

brand name (X= 4.11, F(1,128) = 9.58, p < .01). Second, there was no main effect of

adding irrelevant information (F(1,128) = 0.14, ns). Third, the effect of the irrelevant

information depended on the presence or absence of the brand name (F(1,128) = 6.52, p

< .05). When subjects only received the product information, the irrelevant information

weakened subjects' beliefs in the product benefit from 5.54 to 4.61 (D = -0.93; F(1,128)

= 5.65, p < .05). However, when subjects were also given the brand name, the irrelevant

product information strengthened product beliefs from 3.75 to 4.44, consistent with H9

(D = 0.69; F(1,128) = 4.22, p < .05).

These results demonstrate that providing irrelevant information in addition to

supportive information will not always hurt product perceptions. When a brand has a

strong, negative image, consumers' beliefs will become more favorable after they

encounter both supportive and irrelevant information rather than only supportive

information. These findings are consistent with a biased hypothesis testing account of the

dilution effect. When consumers process information regarding a negatively perceived

brand, they search for counterdiagnostic information that confirms the brand will not

deliver the benefit. The irrelevant information does not confirm this hypothesis, reduces

confidence in the hypothesis, and results in more favorable product beliefs.


4 The descriptions in the filler product categories confirmed subjects' positive or negative
priors, thereby reducing suspicion about the accuracy of the product information.











6 5.54
a) 5 4.61 4.44
I 4 *3.75
4-
3 3

M 2 1.61
1 -- ---- ^---m-^

No Brand Name Brand Name Brand Name Only

Condition
0 Supportive U Supportive + Irrelevant U Brand Name Only



FIGURE 6
EXPERIMENT 7: EFFECT OF A NEGATIVELY PERCEIVED BRAND NAME AND
TYPE OF INFORMATION ON SUBJECTS' BELIEF IN THE PRODUCT BENEFIT















CHAPTER 10
GENERAL DISCUSSION


Marketing managers are undoubtedly aware that consumers are often looking for

specific product benefits and will try to emphasize product information that suggests that

their brand will deliver those benefits. A website may prominently feature a computer's

powerful processor, suggesting that it is fast. An airline commercial may mention that

the airline was voted #1 in a survey on airline service, suggesting that their service is

superior to that of the competition. In spite of this emphasis, consumers will still

encounter a myriad of product information that is not relevant for the benefit they are

seeking. The website may also list that the computer features high quality speakers and

the airline commercial may also mention that the airline is a proud sponsor of the

Olympic Games. There are undoubtedly good reasons to mention this additional

information. An additional feature may appeal to a niche in the market that highly values

a specific benefit, while other product information may create generalized positive affect

toward the brand. Managers often assume that these actions may only increase brand

equity and cannot possibly hurt it. Indeed, if consumers are behaving normatively, they

should simply ignore this information when assessing whether the product will deliver

the benefit.

However, as research on social judgment (e.g., Nisbett et al. 1981) suggests, and

as this research demonstrates, the irrelevant information can actually have a negative










impact on consumers' product perceptions. In ten different studies, and across seventeen

different products and services, the addition of irrelevant information to supportive

benefit information weakened consumers' beliefs in the product's ability to deliver the

benefit. This dilution effect did not depend on the order in which the information was

presented or the manner in which the belief was measured (experiment 1A). Moreover,

the effect even persisted when consumers acknowledged the irrelevance of the

information prior to stating their beliefs (experiment 3), when they believed that the

information was being randomly sampled by a computer (experiment 3A), and when the

irrelevant information made the product description more similar to the typical desired

product (experiment 4).

While the findings indicate that this dilution effect is a quite robust phenomenon,

they also show that irrelevant information does not always dilute consumers' beliefs in

the benefit. Irrelevant information does not affect consumers' belief in the benefit when

the information is processed without the benefit in mind (experiment 5) or when

consumers also consider whether the information suggests the opposite of the benefit

(experiment 6). In fact, adding irrelevant information to supportive benefit information

may even strengthen consumers' beliefs in the product's ability to deliver the benefit,

when the brand under consideration has a very poor reputation (experiment 7).

By examining the influence of these factors which have not been examined

previously, these studies also provide new insight into the mechanism that underlies the

dilution effect. None of the explanations put forth in earlier research can adequately

explain the entire pattern of results observed across the ten studies. The polarizing effect










of less supportive information (experiments 2 and 3) and the fact that the irrelevant

information does not influence consumers' sensitivity to the supportive information

(experiment 2) provide evidence against an averaging account of the dilution effect. The

conversational norms account, on the other hand, cannot explain why the effect persists

when consumers first acknowledge the irrelevance of the product information

(experiment 3) and when the information is allegedly randomly sampled by a computer

(experiment 3A). Furthermore, the representativeness explanation predicts that irrelevant

information that significantly increases the product's similarity to the typical desired

product should lead to more favorable product judgments, while this information in fact

dilutes consumers' beliefs in the product benefit (experiment 4). Finally, the distraction

of resources account is inconsistent with the observation that the irrelevant information

does not affect consumers' recognition of the supportive information (experiment 1) nor

consumers' sensitivity to the supportive information (experiment 2).

Instead, a new explanation is proposed that can account for the entire pattern of

findings. It is argued that consumers are following a biased hypothesis testing procedure

when assessing the likelihood that a product will deliver a benefit. This proposed

mechanism is based on four assumptions. First, consumers test the hypothesis that the

product will deliver the benefit, rather than the hypothesis that it will not deliver the

benefit. Second, consumers selectively search for information that confirms the

hypothesis, i.e., supports the benefit. Third, consumers classify all information with

regard to their search goal, either as "confirming" or "not confirming". Finally,

consumers rely on this classification to determine their belief in the product's ability to










deliver the benefit. Classifying information as "confirming" strengthens their belief,

while classifying information as "not confirming" weakens their belief. When consumers

encounter irrelevant information, they classify it as "not confirming" and weaken their

belief in the product benefit. They do not consider the fact that the information does not

confirm the alternative hypothesis either.

This mechanism is consistent with the findings of the first six experiments. It

explains the direct diluting effect of irrelevant information on product beliefs, as well as

the polarizing effect of less supportive information. It also predicts that the dilution

effect will persist when consumers are aware of the irrelevance of the information, when

the information is allegedly randomly sampled, and when the information increases the

typicality of the product description. Moreover, the last four experiments confirm the

predictions of the biased hypothesis testing perspective regarding the boundary

conditions of the dilution effect. When the irrelevant information is not processed with

the hypothesis in mind, the information is not classified as "confirming" or "not

confirming" and the dilution effect disappears (experiments 5 and 5A). Similarly, the

dilution effect also does not occur when consumers consider the implications of the

information for both the focal hypothesis (that the product will deliver the benefit) and

the alternative hypothesis (that the product will not deliver the benefit) (experiment 6).

Finally, when consumers have very negative priors regarding the brand under

consideration, they may set out to test the hypothesis that the product will not deliver the

benefit, leading to a reversal of the dilution effect (experiment 7).















CHAPTER 11
LIMITATIONS AND FUTURE RESEARCH


Although these studies indicate that irrelevant information can weaken consumers'

belief in a product's ability to deliver a benefit, they do not inform us about its effect on

the overall product evaluation. Obviously, if a consumer is only interested in a single

product benefit, her decreased confidence in the product's ability to deliver that benefit

will reduce her overall product evaluation. However, if a consumer is interested in

multiple benefits, the effect of additional information about other product benefits on the

product's overall evaluation will depend on the trade-off between reduced confidence in

the original benefit and increased confidence in the other benefit. Similarly, the effect of

irrelevant sponsorship associations on overall product evaluations depends on the trade-

off between the dilution effect and the creation of positive affect.

The reported studies are also limited in that subjects are always instructed that a

particular benefit is desirable for them. This does not necessarily correspond to situations

in which desirable benefits are spontaneously generated by consumers. Whether a

benefit becomes salient can depend on its habitual salience for a certain consumer or

because it is primed by a certain usage situation (Ratneshwar, Warlop, Mick, and Seeger

1997). In these situations, the activated benefit may lead to similar outcomes as the

explicit instructions used in the experiments reported here. However, consumers can also










derive the desired benefit from the product information itself. While some product

information may seem irrelevant to consumers at first, they may rely on the relevance

principle (Grice 1975, Sperber and Wilson 1986) and infer that the information has to

convey some value. In this manner, the irrelevant information may prime a benefit

consumers had not considered earlier, thus increasing the appeal of the product, rather

than diluting it.

Furthermore, though the reported studies are most consistent with a biased

hypothesis testing explanation, these results do not imply that other processes cannot

contribute to the dilution effect, or even be the unique cause of the effect in situations in

which this proposed mechanism does not apply. For instance, many studies have shown

that people often do rely on a representativeness heuristic when making predictions (e.g.,

Andreassen 1988; Epstein, Donovan, and Denes-Raj 1999; Kahneman and Tversky 1972,

1973). It is plausible that subjects in the social judgment dilution studies (e.g., Nisbett et

al. 1981) indeed relied on the similarity between the described person and the

stereotypical murderer or child abuser. While it may be hard to assess the similarity

between a product description and an abstract benefit, it is a lot easier to assess the

similarity between a person description and an easily accessible stereotype. In fact,

consumers may also rely on a representativeness heuristic when predicting product

benefits. However, either the product information, or the desired benefit, may have to be

connected to a well defined subcategory or a prototypical brand.

Finally, I would like to emphasize that I only examined the influence of one type

of irrelevant information. The concept of irrelevance used in this dissertation has three










essential characteristics. First, information can only be irrelevant with respect to a

context. The information used in these studies was only irrelevant with respect to one

specific benefit. It was not necessarily irrelevant with respect to product choice or with

respect to the overall evaluation of the product. Second, the irrelevance is subjective

rather than objective. The information was labeled as obviously irrelevant because the

great majority of subjects classified the information as "not helpful" rather than

"suggesting [the benefit]" or "suggesting [not the benefit]". Finally, the irrelevance of

the information is not absolute. I assume that there are degrees of (ir)relevance. The

information is usually not classified as irrelevant by all subjects. Moreover, even for

those who classify it as irrelevant, the information may still have a minimal diagnostic

value. Therefore, it was essential to demonstrate that the irrelevant information tends to

be categorized as supportive rather than counterdiagnostic pretestt 3, experiment 1) and

that less supportive information tends to polarize rather than dilute judgments

(experiment 2).

How does this conceptualization compare to other interpretations of relevance? It

is clearly very different from a consequentialist perspective, which would state that

information is irrelevant for a decision when it does not influence the decision. From this

perspective, the additional product information only is irrelevant in those conditions in

which the dilution effect disappears. A more related perspective is that of Sperber and

Wilson (1986), who argue that an assumption is relevant in a context when it has some

(subjective) contextual effect, the amount of which determines the degree of

(ir)relevance. Interestingly, they argue that the degree of irrelevance also depends on the










effort required to process the information and obtain the contextual effect. One could

indeed argue that all product information can have some diagnostic value for any desired

benefit, but that the effort required to extract this value is so great that it makes the

information irrelevant for most consumers. On the other hand, they also indicate that the

context is not determined before processing the utterance, but is selected so as to

maximize the possible relevance of the statement (since people assume it is relevant).

This indicates that the apparently irrelevant information may suggest new benefits to the

consumer, as mentioned earlier in this section. However, this assertion is based on the

relevance principle, which governs intentional communication. This principle may not

hold in many advertising situations, since consumers may assume that the additional

information is actually intended for another consumer segment. The principle governing

advertising communications may be that the communication has to convey value to some

consumers, but not necessarily to the individual processing the message.















REFERENCES


Anderson, Norman H. (1967), "Averaging Model Analysis of Set-Size Effect in
Impression Formation," Journal ofExperimental Psychology, 75 (2), 158-165.

(1971), "Integration Theory and Attitude Change," Psychological Review, 78
(May), 171-206.

(1974), "Information Integration Theory: A Brief Survey," in Contemporary
Developments in Mathematical Psychology: L Learning. Memory and Thinking,
eds. David H. Krantz, Richard C. Atkinson, Duncan R. Luce, and Patrick Suppes,
San Francisco: W.H. Freeman, 236-270.

Andreassen, Paul B. (1988), "Explaining the Price-Volume Relationship: The Difference
Between Price Changes and Changing Prices," Organizational Behavior &
Human Decision Processes, 41 (June), 371-389.

Asch, Solomon E. (1946), "Forming Impressions of Personality," Journal ofAbnormal &
Social Psychology, 41 (July), 258-290.

Barsalou, Lawrence W. (1983), "Ad Hoc Categories," Memory & Cognition, 11 (May),
211-227.

Beyth-Marom, Ruth and Baruch Fischhoff (1983), "Diagnosticity and
Pseudodiagnosticity," Journal ofPersonality and Social Psychology, 45
(December), 1185-1195.

Birnbaum, Michael H. and Barbara A. Mellers (1983), "Bayesian Inference: Combining
Base Rates With Opinions of Sources Who Vary in Credibility," Journal of
Personality and Social Psychology, 45 (October), 792-804.

Broniarczyk, Susan M. and Andrew D. Gershoff (1997), "Meaningless Differentiation
Revisited," Advances in Consumer Research, 24, 223-228.

Brown, Christina L. and Gregory S. Carpenter (2000), "Why Is the Trivial Important? A
Reasons-Based Account for the Effects of Trivial Attributes on Choice," Journal
of Consumer Research, 26 (March), p 372-385.










Cantor, Nancy and Walter Mischel (1979), "Categorization Processes in the Perception of
People," in Advances in Experimental Social Psychology, Vol. 12, ed. L.
Berkowitz, New York: Academic Press, 3-52.

Carpenter, Gregory S., Rashi Glazer, Kent Nakamoto (1994), "Meaningful Brands From
Meaningless Differentiation: The Dependence on Irrelevant Attributes," Journal
of Marketing Research, 31 (August), 339-350.

De Dreu, Carsten K. W., Vincent Y. Yzerbyt, and Jacques-Philippe Leyens (1995),
"Dilution of Stereotype-Based Cooperation in Mixed-Motive Interdependence,"
Journal ofExperimental Social Psychology, 31 (November), 575-593.

Epstein, Seymour, Sean Donovan, and Veronika Denes-Raj (1999), "The Missing Link in
the Paradox of the Linda Conjunction Problem: Beyond Knowing and Thinking
of the Conjunction Rule, the Intrinsic Appeal of Heuristic Processing,"
Personality & Social Psychology Bulletin, 25 (February), 204-214.

Fein, Steven and James L. Hilton (1992), "Attitudes toward Groups and Behavioral
Intentions toward Individual Group Members: The Impact of Nondiagnostic
Information," Journal ofExperimental Social Psychology, 28 (March), 101-124.

Fischhoff, Baruch and Ruth Beyth-Marom (1983), "Hypothesis Evaluation From a
Bayesian Perspective," Psychological Review, 90 (July), 239-260.

Gilbert, Daniel T. (1991), "How Mental Systems Believe," American Psychologist, 46
(February), 107-119.

,Romin W. Tafarodi, and Patrick S. Malone (1993), "You Can't Not Believe
Everything You Read," Journal of Personality & Social Psychology, 65 (August),
221-233.

Grice, H. Paul (1975), "Logic and Conversation," in Syntax and Semantics: 3. Speech
Acts, ed. Peter Cole and Jerry L. Morgan, New York: Academic Press, 41-58.

Ha, Young-Won and Stephen J. Hoch (1989), "Ambiguity, Processing Strategy, and
Advertising-Evidence Interactions," Journal of Consumer Research, 16
(December), 354-360.

Hilton, James L. and Steven Fein (1989), "The Role of Typical Diagnosticity in
Stereotype-Based Judgments," Journal of Personality & Social Psychology, 57
(August), 201-211.










Hoch, Stephen J. and Young-Won Ha (1986), "Consumer Learning: Advertising and the
Ambiguity of Product Experience," Journal of Consumer Research, 13
(September), 221-233.

Kahneman, Daniel and Amos Tversky (1972), "Subjective Probability: A Judgment of
Representativeness," Cognitive Psychology, 3 (July), 430-454.

and Amos Tversky (1973), "On the Psychology of Prediction," Psychological
Review, 80 (July), 237-251.

Klayman, Joshua and Young-Won Ha (1987), "Confirmation, Disconfirmation, and
Information in Hypothesis Testing," Psychological Review, 94 (April), 211-228.

Koriat, Asher, Sarah Lichtenstein, and Baruch Fischhoff (1980), "Reasons for
Confidence," Journal of Experimental Psychology: Human Learning & Memory,
6 (March), 107-118.

Kuhn, Deanna and Joseph Lao (1996), "Effects of Evidence on Attitudes: Is Polarization
the Norm?" Psychological Science, 7 (March), 115-120.

Lichtenstein, Sarah, Timothy C. Earle, and Paul Slovic (1975), "Cue Utilization in a
Numerical Prediction Task," Journal ofExperimental Psychology: Human
Perception and Performance, 104 (February), 77-85.

Locksley, Anne, Christine Hepburn, and Vilma Ortiz (1982), "Social Stereotypes and
Judgments of Individuals: An Instance of the Base-Rate Fallacy," Journal of
Experimental Social Psychology, 18 (January), 23-42.

Lopes, Lola L. (1987) "Procedural Debiasing," Acta Psychologica, 64 (February), 167-
185.

Lord, Charles G., Lee Ross, and Mark R. Lepper (1979), "Biased Assimilation and
Attitude Polarization: The Effects of Prior Theories on Subsequently Considered
Evidence," Journal ofPersonality & Social Psychology, 37 (November), 2098-
2109.

Nickerson, Raymond S. (1998), "Confirmation Bias: A Ubiquitous Phenomenon in Many
Guises," Review of General Psychology, 2 (June), 175-220.

Nisbett, Richard E., Henri Zukier, and Ronald E. Lemley (1981), "The Dilution Effect:
Nondiagnostic Information Weakens the Implications of Diagnostic Information,"
Cognitive Psychology, 13 (April), 248-277.










Peterson, Cameron R. and Wesley M. DuCharme (1967), "A Primacy Effect in
Subjective Probability Revision," Journal ofExperimental Psychology, 73
(January), 61-65.

Pitz, Gordon F., Leslie Downing, and Helen Reinhold (1967), "Sequential Effects in the
Revision of Subjective Probabilities," Canadian Journal ofPsychology, 21
(October), 381-393.

Ratneshwar, S., Cornelia Pechmann, and Allan D. Shocker (1996), "Goal-Derived
Categories and the Antecedents of Across-Category Consideration," Journal of
Consumer Research, 23 (December), 240-250.

,Luk Warlop, David Glen Mick, and Gail Seeger (1997), "Benefit Salience and
Consumers' Selective Attention to Product Features," International Journal of
Research in Marketing, 14 (July), 245-259.

Sanbonmatsu, David M., Steven S. Posavac, Frank R. Kardes, and Susan P. Mantel
(1998), "Selective Hypothesis Testing," Psychonomic Bulletin & Review, 5
(June), 197-220.

Randon Stasney (1997), "The Subjective Beliefs Underlying Probability
Overestimation," Journal ofExperimental Social Psychology, 33 (May), 276-295.

Schwarz, Norbert, Fritz Strack, Denis Hilton, and Gabi Naderer (1991), "Base Rates,
Representativeness and the Logic of Conversation: The Contextual Relevance of
'Irrelevant' Information," Social Cognition, 9 (Spring), 67-84.

Shaklee, Harriet and Baruch Fischhoff (1982), "Strategies of Information Search in
Causal Analysis," Memory & Cognition, 10 (November), 520-530.

Shanteau, James (1975), "Averaging versus Multiplying Combination Rules of Inference
Judgment," Acta Psychologica, 39 (February), 83-89.

Simonson, Itamar, Ziv Carmon, and Suzanne O' Curry (1994), "Experimental Evidence
on the Negative Effect of Product Features and Sales Promotions on Brand
Choice," Marketing Science, 13 (Winter), 23-41.

,Stephen M. Nowlis, and Yael Simonson (1993), "The Effect of Irrelevant
Preference Arguments on Consumer Choice," Journal of Consumer Psychology, 2
(3), p 287-306.

Slugoski, Ben R. and Anne E. Wilson (1998), "Contribution of Conversation Skills to the
Production of Judgmental Errors," European Journal of Social Psychology, 28
(July), 575-601.











Snyder, Mark and Swann, William B. (1978), "Hypothesis-Testing Processes in Social
Interaction," Journal ofPersonality and Social Psychology, 36 (November),
1202-1212.

and Nancy Cantor (1979), "Testing Hypotheses About Other People: The Use of
Historical Knowledge," Journal ofExperimental Social Psychology, 15 (July),
330-342.

Sperber, Dan and Deirdre Wilson (1986), Relevance: Communication and Cognition,
Cambridge, MA: Harvard University Press.

Tetlock, Philip E. and Richard Boettger (1989), "Accountability: A Social Magnifier of
the Dilution Effect," Journal of Personality and Social Psychology, 57
(September), 288-398.

,Jennifer S. Lerner, and Richard Boettger (1996), "The Dilution Effect: Judgmental
Bias, Conversational Convention, or a Bit of Both?," European Journal of Social
Psychology, 26 (November), 915-934.

Tesser, Abraham (1978), "Self-Generated Attitude Change," in Advances in
Experimental Social Psychology, 11, 289-338.

Trope, Yaacov and Akiva Liberman (1996), "Social Hypothesis Testing: Cognitive and
Motivational Mechanisms," in Social Psychology: Handbook ofBasic Principles,
ed. Edward Tory Higgins and Arie W. Kruglanski, New York, NY: The Guilford
Press, 239-270.

Troutman, C. Michael and James Shanteau (1977), "Inferences Based on Nondiagnostic
Information," Organizational Behavior and Human Performance, 19 (June), 43-
55.

Wallsten, Thomas S. (1981), "Physician and Medical Student Bias in Evaluating
Diagnostic Information," Medical Decision Making, 1 (October), 145-164.

Wason, Peter C. (1960), "On the Failure to Eliminate Hypotheses in a Conceptual Task,"
Quarterly Journal of Experimental Psychology, 12 (August), 129-140.

Yzerbyt, Vincent Y., Jacques-Philippe Leyens, and Georges Schadron (1997), "Social
Judgeability and the Dilution of Stereotypes: The Impact of the Nature and
Sequence of Information," Personality and Social Psychology Bulletin, 23
(December), 1312-1322.






92



Zukier, Henri (1982), "The Dilution Effect: The Role of the Correlation and the
Dispersion of Predictor Variables in the Use of Nondiagnostic Information,"
Journal of Personality and Social Psychology, 43 (December), 1163-1174.

and Dennis L. Jennings (1983), "Nondiagnosticity and Typicality Effects in
Prediction," Social Cognition, 2 (July), 187-198.










APPENDIX A
INSTRUCTIONS TO SUBJECTS


Standard Instructions

We will now show you 8 different products. For each of these products, you are looking

for a particular benefit (e.g., you want a fast computer).

For each of these products, we will provide you with some information. This

information may or may not be helpful for the decision you have to make. You should

look at this information as objective information, i.e., information as provided by

Consumer Reports. After each product description, we will ask you to rate the product

on an important dimension.



Instructions Experiment 1A

We will now show you 8 different products categories. In each category, you are

looking for a particular benefit (e.g., you want a fast computer).

For each category, we will provide you with information about two different brands.

The information may or may not be helpful for the decision you have to make. You

should look at this information as objective information, i.e., information as provided by

Consumer Reports. After you have received information for both brands, we will ask

you to indicate which brand is most likely to deliver the benefit.




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs