University of Southern California Dissertations and Theses

Consumers' subjective knowledge influences evaluative extremity and product differentiation

(USC Thesis Other)

Consumers' subjective knowledge influences evaluative extremity and product differentiation

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content CONSUMERS’ SUBJECTIVE KNOWLEDGE INFLUENCES EVALUATIVE

EXTREMITY AND PRODUCT DIFFERENTIATION

by

Kachat Andrew Wong

________________________________________________________________________

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BUSINESS ADMINISTRATION)

December 2010

Copyright 2010 Kachat Andrew Wong

ii
TABLE OF CONTENTS

List of Tables iii

Abstract iv

Chapter 1: Introduction 1

Chapter 2: Literature Review 5
Distinction Between Objective and Subjective Knowledge 5
Factors that Influence Subjective Knowledge 6
The Effects of Objective and Subjective Knowledge 8

Chapter 3: Why Subjective Knowledge Should Influence Differentiation 13
Subjective Knowledge Affects the Processing of Valenced
Information
13
Subjective Knowledge Influences Motivation to Process 15
Subjective Knowledge Increases Confidence in Judgment 17

Chapter 4: Overview of the Experimental Studies 18

Chapter 5: Study 1 23

Chapter 6: Study 2 33

Chapter 7: Study 3 44

Chapter 8: Study 4 54

Chapter 9: Study 5 66

Chapter 10: Conclusion 74

References 85

Appendices: 93
Appendix A: Photo Samples Used in Study 1 93
Appendix B: Study 1 - Transcript from the Computer Study 94
Appendix C: Study 2 - Transcript from the Computer Study 102
Appendix D: Study 3 - Transcript from the Computer Study 111
Appendix E: Study 4 - Transcript from the Computer Study 121
Appendix F: Study 5 - Transcript from the Computer Study 129

iii
LIST OF TABLES

Table 1: Overview of Studies 1 – 5 21

Table 2: Extremity as a Function of Subjective Knowledge in Study 1 30

Table 3: Evaluations as a Function of Subjective Knowledge and Product
Quality in Study 2
41

Table 4: Evaluations as a Function of Subjective Knowledge and Motivation to
Process in Study 3
51

Table 5: Evaluations as a Function of Subjective Knowledge and Public
Accountability in Study 5
64

Table 6: Evaluations as a Function of Subjective Knowledge and Product
Quality in Study 6
71

iv
ABSTRACT
The ability to differentiate among product alternatives is generally attributed to
the actual product knowledge that consumers possess (Alba and Hutchinson 1987). My
research proposes a subjective knowledge model as an alternative to account for when
consumers give extreme evaluations and discriminate among products. Specifically, I
propose that high subjective knowledge leads to the differential processing of valenced
product information. Whereas people high in subjective knowledge process both positive
and negative information, those who are low in subjective knowledge are less likely to
seek out negative information. As a result, high subjective knowledge individuals are
more likely to give extreme product evaluations and demonstrate greater differentiation
than their low subjective knowledge counterparts. Five experiments demonstrate the
predicted effects and boundary conditions.

1
Chapter 1: Introduction
Evaluating and differentiating among product alternatives are fundamental to
making an optimal choice (Hoegg and Alba 2007). To identify the best option out of a
choice set, consumers need to tell the differences among available alternatives, spreading
them apart to an extent that a favorite emerges (Brownstein 2003; Svenson 1992). In
particular, brand choice often requires identification of the best quality option, or at least
ruling out options that are poorer in quality than others. This is the focus of my research –
detecting when consumers fail to differentiate among options, and especially when they
fail to differentiate a high quality option from a low quality option.
I am not the first to investigate this issue. It is well accepted that factors such as
product attribute information (Carpenter, Glazer, and Nakamoto 1994) and the way in
which product information is structured and represented (Diehl, Kornish, and Lynch
2003; Simonson 1989) influence how consumers differentiate among alternatives. When
consumers do have product information available to them, a critical factor influencing
differentiation is a consumer’s knowledge. The ability to process information improves as
one’s knowledge and expertise in a domain increases (Alba and Hutchinson 1987).
Highly knowledgeable consumers’ well-established cognitive structures enable them to

2
refine decision rules and evaluation criteria, to better categorize product alternatives and
to attend to relevant and important attributes, which, in turn, help them better
discriminate. In brief, the ability to differentiate among product alternatives is generally
attributed to the actual knowledge of consumers.
Whereas past research emphasizes the importance of objective knowledge for
differentiation based on theorizing about cognitive structures, I propose a new theoretical
account that explains when consumers demonstrate greater differentiation among product
choices. Specifically, I propose that one’s subjective perception of product knowledge
exerts effects on product evaluations that are independent from the effects of actual or
objective knowledge. Subjective knowledge refers to a person’s self-assessed knowledge
for a domain (Park, Mothersbaugh, and Feick 1994). In my research, I operationally
define subjective knowledge as the consumer’s perceived product knowledge relative to
referent others (Burson 2007). I propose that people who are high in subjective
knowledge (i.e., people who believe that they are relatively more knowledgeable than
others) have a greater propensity to evaluate product alternatives more extremely than
those who are low in subjective knowledge. As a consequence, high subjective
knowledge individuals discriminate among the alternatives to a greater extent than do low

3
subjective knowledge individuals. Moreover, I investigate various mechanisms by which
subjective knowledge influence evaluations. These include the possibility that those low
in subjective knowledge as compared to those high in subjective knowledge (1) do not
seek out negative information as much, (2) are not as motivated and involved in the
judgment task and (3) are not as confident about their judgments. My research contributes
to the fundamentally important topic of consumer knowledge by showing critical
consequences of subjective knowledge that previous researchers have attributed to
objective knowledge. Furthermore, my research has practical implications for marketers.
Marketers can often influence consumers’ sense of subjective knowledge. Marketing
communications can strategically bolster or diminish customers’ subjective knowledge
through message content and execution. Influencing customers’ subjective knowledge
may lead to customers’ behaviors that are advantageous to marketers, such as greater
product differentiation and quicker purchase decisions.
My dissertation is organized as follows. First, I review the literature on
knowledge, focusing on distinctions between subjective and objective knowledge. Then I
explain why subjective knowledge should increase differentiation among alternatives.
This is followed by five empirical studies that tested the hypothesized relationship

4
between subjective knowledge and product differentiation. I provide empirical evidence
that high subjective knowledge affects the processing of valenced information, which in
turn increases extremity in product ratings, resulting in greater differentiation among
alternatives. I also show that those low in subjective knowledge do not search for
negative information and consequently do not differentiate between high and low quality
product as well as do those high in subjective knowledge. My studies indicate that
motivation and confidence do not account for the differentiation effect. I conclude my
dissertation by discussing the theoretical contributions and managerial implications of the
proposed subjective knowledge effect in the final section.

5
Chapter 2: Literature Review
Two literatures – the consumer knowledge literature and the social psychology
literature identifying when consumers make extreme evaluations – are particularly
relevant to my research on the effects of subjective knowledge on consumers’ product
evaluations and differentiation. This section provides conceptual distinctions between
objective and subjective knowledge, and discusses the effects of objective and subjective
knowledge on behaviors.

Distinctions Between Objective and Subjective Knowledge
Typically, knowledge refers to information stored in long-term memory (Brucks
1985; Park et al. 1994). Alba and Hutchinson (1987) noted that consumer knowledge
includes familiarity and expertise. Familiarity refers to “the number of product-related
experiences that have been accumulated by the consumer”. Expertise is defined as “the
ability to perform product-related tasks successfully” (p. 411). Hence, consumer
knowledge subsumes both product experience and product-related skills and abilities
(Shanteau 1992).

6
Most important for my purpose is that product-related knowledge is of two types
– objective knowledge and subjective knowledge (Brucks 1985; Trafimow and Sniezek
1994). Objective knowledge concerns the amount, type, or organization of what an
individual actually has stored in memory. In contrast, self-assessed knowledge or
subjective knowledge reflects people’s perceptions of what or how much they know. In
simple terms, whereas subjective knowledge represents what individuals perceive that
they know, objective knowledge indicates what they actually know. It seems logical that
subjective knowledge should be derived from objective knowledge such that they should
be highly correlated. However, many studies reveal that consumers’ knowledge
calibration (the correspondence between subjective and objective knowledge) is not
always high (Alba and Hutchinson 2000; Park et al. 1994; Radecki and Jaccard 1994;
Mitchell and Dacin 1996).

Factors that Influence Subjective Knowledge
Apart from differences in conceptual definitions, malleability is one of the critical
distinctions between subjective and objective knowledge. Compared to an individual’s
objective knowledge, subjective perceptions of knowledge are more prone to contextual

7
and social influences. Past research has shown two main factors that can effectively alter
subjective knowledge. The first is performance-related. People infer their domain
knowledge based on their performance in relevant tasks. For example, Moorman et al.
(2004) asked participants to complete a knowledge quiz that ostensibly measured their
domain knowledge. Fictitious scores were then provided to either bolster or undermine
subjective knowledge. Similarly, Burson (2007) manipulated task difficulty to change
self-perception of knowledge. People who were assigned to a relatively easy task inferred
higher subjective knowledge compared to those who performed a relatively difficult task
in the same domain (see also Trope and Pomerantz 1998).
The second factor influencing subjective knowledge is socially-induced. Social
comparison targets can either heighten or diminish one’s subjective knowledge (Fox and
Weber 2002). Specifically, an upward comparison (i.e., a comparison with someone who
is more knowledgeable) leads to an unfavorable self-assessment whereas a downward
comparison (i.e., a comparison with someone who is less knowledgeable) results in a
more favorable self-assessment of knowledge (See 2009; Vohs and Heatherton 2004).

8
The Effects of Objective and Subjective Knowledge
Some early studies of consumer knowledge incorporated both subjective and
objective knowledge into the operational definition of product knowledge (e.g., Park and
Lessig 1981). Other studies used subjective knowledge as a surrogate for objective
knowledge (Ailawadi, Dant, and Grewal 2004) by assuming a high correlation between
subjective and objective knowledge. However, the magnitude of subjective and objective
knowledge correspondence varies across product categories and consumption contexts
(Carlson et al. 2009). Using subjective and objective knowledge interchangeably can
produce misleading findings. Moreover, the effects of subjective and objective
knowledge can be confounded if these two constructs of consumer knowledge are not
clearly distinguished. Given that subjective knowledge is more malleable than objective
knowledge and the correspondence between them is not always high, it is important to
isolate the unique effects of subjective knowledge from those of objective knowledge.
The effects of objective knowledge on behaviors. Extant literature posits that
product knowledge implicates a well-developed cognitive structure, which helps the
consumer differentiate among various product alternatives (Alba and Hutchinson 1987),
integrate external information with one’s existing knowledge base (Maheswaran and

9
Sternthal 1990), achieve better memory and reasoning (Shanteau 1992), and detect new
information and available alternatives (Mitchell and Dacin 1996; Sujan 1985). These
differences are seen most sharply when comparing experts with novices. Because of their
superior cognitive functioning, experts enjoy ease in information processing, which
facilitates their acquisition of new information (Brucks 1985). Moreover, experts who are
highly familiar with a given product category are likely to engage in automatic processes,
which frees up cognitive resources for other activities, such as learning (Alba and
Hutchinson 1987). Therefore, experts are more capable and expected to search more
product-related information than novices.
Objective knowledge not only facilitates information search, but also enables
experts to better process information and choose appropriate evaluation strategies.
Experts are more capable of analytic processing (Sujan 1985), are more competent in
drawing inferences about unavailable information (Alba and Hutchinson 1987), elaborate
on information more effortfully regardless of the external motivation to do so
(Maheswaran and Sternthal 1990), and better integrate new information with existing
knowledge relative to novices (Maheswaran, Sternthal, and Gürhan 1996). Furthermore,
because of their ability to categorize stimuli to finer degrees and the associated, rich

10
schema-based knowledge, expertise can use more efficient top-down processing as
opposed to the bottom-up, data-driven strategies favored by novices (Spence and Brucks
1997).
Because high objective knowledge helps experts search, process, and organize
information, experts gradually develop an internally consistent cognitive structure that is
characterized by highly correlated attributes (Judd and Brauer 1995). When evaluating a
product, for example, experts consider many attributes that are perceived as correlated,
resulting in polarizing and extreme evaluations (Lusk and Judd 1988). Nonexperts
generally consider fewer, relatively orthogonal attributes. When overall evaluations are
made on the basis of these attributes, the judgments are more moderate than those of
experts (Brauer et al. 2004). In sum, past research suggests that high objective knowledge
leads to extreme evaluations.
The effects of subjective knowledge on behaviors. Empirical evidence in consumer
research suggests that subjective knowledge exerts unique, independent effects, or even
opposite effects to those of objective knowledge (Brucks 1985; Raju, Lonial, and
Mangold 1995). The effects of subjective knowledge on consumer behaviors, reported in
the literature thus far include the impact on product choices (Burson 2007; Moorman et

11
al. 2004), overconfidence (Trafimow and Sniezek 1994), and information search (Brucks
1985; Park and Lessig 1981). Generally, subjective knowledge is negatively related to the
amount of information search (Brucks 1985; Raju et al. 1995). Yet, the lack of subjective
knowledge can motivate increased search by increasing the perceived importance of, and
receptivity to, new product information (Park and Lessig 1981; Park et al. 1994).
Apart from the amount of search, search strategies are also influenced by
subjective knowledge. Moorman et al. (2004) demonstrated that consumers chose the
locations of search based on their perceived domain knowledge to maintain self-
consistency. Specifically, Moorman et al. (2004) showed that those who believed that
they were knowledgeable about health and nutrition were more likely to search for
products in the areas of a store in which healthy foods were placed. Park and Lessig
(1981) also showed that the level of product familiarity, which was comprised of both
subjective and objective knowledge measures, influenced the selectivity of information
inputs. Specifically, high familiarity individuals are relatively more confident in the use
of intrinsic cues for product evaluations (Rao and Monroe 1988). In comparison, those
who are low in familiarity may be less confident about intrinsic cues. Instead they base

12
their evaluations on marketer-supplied, extrinsic product cues, such as price and brand
names.

13
Chapter 3: Why Subjective Knowledge Should Influence Differentiation
Whereas past research identifies the importance of objective knowledge on
differentiation from an information processing perspective, I propose the novel notion
that subjective knowledge can explain when consumers differentiate among product
choices. Specifically, I suggest that subjective knowledge can exert a unique effect on the
way consumers evaluate products. Relative to people who are low in subjective
knowledge, those high in subjective knowledge increase the extremity of product
evaluations, demonstrating greater differentiation among alternatives. This proposed
subjective knowledge effect can be explained by the differential processing of valenced
product information. In particular, I propose that low subjective knowledge reduces the
processing of negative information.

Subjective Knowledge Affects the Processing of Valenced Information
I propose that subjective knowledge affects the processing of valenced product
information. The rationale hinges on their inferred capability in domain-relevant tasks.
Specifically, high subjective knowledge should enhance consumers’ perception of their
capability to process product information. In contrast, low subjective knowledge dampens

14
one’s belief in how well they process (Bandura 2001). Since consumers have broad
expectations that products have good attributes more than bad ones (Herr, Kardes, and
Kim 1991), they may believe that negative product information is harder to identify than
positive information. The diminished self-view on the processing capability thus
discourages low subjective knowledge people from actively looking for negative
information that is believed to be hard to find. In brief, low subjective knowledge can
create a positivity bias in processing because of a perceived lack of capability to identify
negative attributes and information.
Negative information facilitates clearer categorization of products and thus
greater differentiation among product alternatives than positive information (Herr et al.
1991). Provided that people high in subjective knowledge are more willing to process
negative information because of beliefs about their capability, they should give more
extreme evaluations and demonstrate greater differentiation than should those who are
low in subjective knowledge.

15
H1: People who are high in subjective knowledge about a product domain are
more extreme in product ratings than those who are low in subjective
knowledge.
H2: Whereas people low in subjective knowledge about a product domain are
predominantly positive regardless of the product quality, those who are
high in subjective knowledge differentiate between products, evaluating
poor quality products lower.

Subjective Knowledge Influences Motivation to Process
There are two possible alternative explanations that may account for the effects of
subjective knowledge on evaluative extremity and product differentiation. First,
subjective knowledge may increase the motivation to process product information. On the
one hand, when people perceive that they are knowledgeable about a product category,
they become more engaged in domain-related tasks and motivated to scrutinize products
more effortfully (Fishbach 2009). As a result, they attend to the differences among the
available choices that could have been overlooked when the motivation to process is low.

16
Choices are thus discriminated to a greater extent when motivation is high than when
low.
On the other hand, past research shows that people low in subjective knowledge
may increase their motivation to search and process (Park and Lessig 1981), particularly
when the product domain is high in personal relevance (Radecki and Jaccard 1994).
Besides, highly effortful processing does not necessarily increase extremity in evaluating
a product. It can elevate the complexity of the cognitive representation (i.e., a product
represented by more attributes and features), which in turn reduces evaluation extremity
(Linville 1982).
In brief, subjective knowledge may not have a unidirectional influence on
motivation to process. Even if high subjective knowledge does motivate highly effortful
processing, evaluative extremity does not necessarily increase. Although there is
ambiguity in the aforementioned causal chain that may render the motivation explanation
improbable, I empirically test this potential mechanism.

17
Subjective Knowledge Increases Confidence in Judgment
The second alternative explanation pertains to confidence in judgment.
Confidence in the present context refers to the extent to which a consumer feels capable
and assured with respect to domain-relevant decisions and behaviors (Bearden, Hardesty,
and Rose 2001). Since subjective knowledge implicates domain-relevant capability,
confidence should be a function of subjective knowledge (Park and Lessig 1981). Past
research suggests that high confidence is associated with strong attitude certainty and
extremity (Krosnick et al. 1993). Feeling confident, high subjective knowledge
individuals should be more willing to commit to extreme evaluations and greater
differentiation than should their low subjective knowledge counterparts. In my empirical
studies, I measured confidence in one’s product evaluations and tested this possible
explanation.

18
Chapter 4: Overview of the Experimental Studies
I conducted five laboratory experiments to examine the effects of subjective
knowledge on evaluative extremity and product differentiation. In all studies, participants
experienced products, either aurally or visually, and then were asked to evaluate the
products. The basic procedure was identical in all experiments. Prior to product trial,
participant’s subjective knowledge was manipulated either (1) via false feedback about
their performance on a quiz or (2) by social comparison information. This was followed
by product trial. After product trial, participants gave their product evaluations and wrote
product reviews, as well as completing items used as control variables. Table 1 provides a
summary of the independent and dependent variables in each study.
Study 1 demonstrated the effects of subjective knowledge on evaluative extremity
by showing that high subjective knowledge participants gave more extreme evaluations
than did their low subjective knowledge counterparts.
Study 2 examined the nature of the extremity and showed that the direction of
extremity reflects actual differences in product quality. High subjective knowledge
people showed greater differentiation between high and low-quality products than did
low subjective knowledge people. In addition, low subjective knowledge people

19
demonstrated positivity in written reviews compared to their high subjective knowledge
counterparts.
Study 3 provided a direct test of motivation as an explanation to the subjective
knowledge effect by manipulating motivation to process. Highly motivated people who
were low in subjective knowledge were not as extreme as those who were high in
subjective knowledge. The result refuted motivation to process as a rival explanation to
the proposed subjective knowledge effect. In addition, a direct instruction to focus on the
negative information attempted to increase the extremity among low subjective
knowledge individuals. When the low subjective knowledge group was told to process
negative product information, they were able to differentiate by rating an inferior product
lowly.
Study 4 examined the boundary factor of the subjective knowledge effect by
manipulating public accountability. The effect of subjective knowledge on extremity was
reduced when consumers believed that their evaluations were subject to scrutiny by
others.

20
Finally, study 5 employed a different manipulation of accountability to
demonstrate its effects on positivity in reviews. When the high subjective knowledge
group expected a discussion on their product evaluations with another person, they
became positive in their reviews regardless of product quality.

21
Table 1: Overview of Studies 1 – 5

Study

Hypothesis Key IVs Key DVs Findings
1 Subjective
knowledge
influences extremity
in product ratings.

Subjective
knowledge
Product
ratings
1) SK influenced extremity
(i.e., variance),
2) No objective knowledge
effect,
3) No confidence effect,
4) No effort effect.

2 1) Subjective
knowledge
influences extremity
in product ratings.

2) Subjective
knowledge
influences positivity
in written reviews.

1) Subjective
knowledge,

2) Product
quality
1) Product
ratings,

2) Written
reviews
1) SK influenced
extremity,
2) High SK group
differentiated between
superior and inferior
products whereas low SK
group did not.
3) Low SK group gave
predominantly positive
reviews,
4) No confidence effect,
5) No effort effect.

3 1) Encouraging low
SK group to process
negative
information
increases extremity.

2) Motivate to
process as an
alternative
explanation is ruled
out.
1) Subjective
knowledge,

2) Incentive to
be accurate,

3) Incentive to
be negative-
focused

1) Product
ratings,

2) Positive &
negative
mood
1) SK effect as in study 2
& 4 was replicated when
there was no incentive,
2) Incentive to be accurate
did not make low SK group
differentiate,
3) Incentive to be negative-
focused made low SK
group differentiate,
4) No mood effect.

22
Table 1, Continued

Study

Hypothesis Key IVs Key DVs Findings
4 Public
accountability
moderates the
subjective
knowledge effect.
1) Subjective
knowledge,

2) Public
accountability
Product
ratings
1) SK effect as in study 1
was replicated in the low
accountability condition,
2) High SK group reduced
extremity when made
accountable,
3) Public accountability
reduced confidence,
4) High SK group was
more confident than low
SK group,
5) People spent longer time
to inspect photo when
accountable, but the
inspection time did not
correlate with extremity.

5 Public
accountability
moderates the
subjective
knowledge effect.
1) Subjective
knowledge,

2) Product
quality,

3) Public
accountability

1) Product
ratings,

2) Written
reviews
1) SK effect as in study 2
was replicated,
2) Public Accountability
made review predominantly
positive regardless of
product quality and SK.

23
Chapter 5: Study 1
Method
Design and procedure. The present study was a single-factor design with three
levels of subjective knowledge (high, average, and low). One hundred and nine
undergraduate students participated in the study for course credit. The stimuli were
digital photos that supposedly were produced by different software brands. Participants
were individually seated in partitioned cubicles equipped with personal computers, and
randomly assigned to one of the 3 subjective knowledge conditions (i.e., high, average
and low subjective knowledge). All participants first completed an 18-question
knowledge quiz, which purportedly measured their knowledge and expertise about digital
photography. Upon completion of the quiz, they were informed that the system
automatically calculated their test scores and compared them with those of the 300 past
participants who did the same quiz. About one-third of the participants (in the high
subjective knowledge condition) were told that their scores were in the top 88
th

percentile. Another third (in the average subjective knowledge condition) were told that
their scores were in the 54
th
percentile. The remaining third (in the low subjective
knowledge condition) were told that their scores were in the bottom 12
th
percentile. Their

24
actual quiz scores, which were not available to the participants until the end of the study,
served as the measure of objective knowledge. After learning their bogus quiz
performance, all participants completed items that assessed subjective knowledge.
Next, they were asked to inspect 4 photo samples ostensibly enhanced by the
different software brands, simply named A, B, C, and D (see Appendix A for photo
samples). There was no time limit on the inspection of individual photos. After viewing a
photo sample, participants gave evaluations of the corresponding photo software.
Therefore, each participant gave 4 product evaluations in total. In fact, all photo samples
were digitally manipulated based on the same stock photo by the same professional photo
management software (Adobe Lightroom 1.3.1). Specifically, visual parameters, such as
white balance, contrast, hue, and color depth, were randomly assigned to the same image
to create qualitatively different outputs.
In addition, unbeknownst to the participants, two photo samples, B and D, were
actually identical. The inclusion of two identical samples provided an opportunity to
measure tendencies to differentiate when there was no basis for doing so (further details
are in the following section). After providing product evaluations, participants responded

25
to items that measured their confidence and effort in the task, as well as demographics.
Upon completion, they received their actual quiz scores and were thoroughly debriefed.
Measures. Subjective knowledge about digital photography was assessed by 4
seven-point items adapted from Brucks (1985) and Park et al. (1994): “I think of myself
as an expert on digital photography” (1 = strongly disagree; 7 = strongly agree), “rate
how much you feel you know about digital photography compared to average college
students” (1 = much less; 7 = much more), “how familiar are you with digital
photography?” (1 = very unfamiliar; 7 = very familiar), and “how much experience do
you have with digital photography compared to your friends?” (1 = very inexperienced; 7
= very experienced). These items were averaged into a subjective knowledge index (α =
.82). Objective knowledge was measured based on the number of quiz items each
participant correctly answered, with a possible maximum of 18 and a minimum of zero.
Participants evaluated each photo using three bi-polar 100-point scales, with
anchors positive vs. negative, favorable vs. unfavorable, and high quality vs. low quality.
These items were high in internal consistency (α > .90) and averaged into four product
ratings for the four software products respectively. Evaluative extremity was measured in
two different ways. First, the standard deviation of an individual’s ratings across the set

26
of four target software programs was computed. Second, the range of ratings, which
equaled the difference between the lowest and highest ratings among the four software
programs, was calculated for each participant (Linville 1982). To rule out the possibility
that those high in SK perceived product differences where they did not exist, the
difference in ratings of product B and D, which in fact were identical, was obtained from
each participant.
Confidence in the product evaluations was assessed by two items: “How
confident are you in your evaluation of the digital photo software?” (1 = Not at all
confident; 7 = Very confident), and “How certain are you about the quality of the digital
photo software?” (1 = Very uncertain; 7 = Very certain). They were averaged into a
composite index of confidence (α = .89). To investigate whether participants’ motivation
to process was influenced by their subjective knowledge, they were asked to report their
effort in the evaluation with two seven-point scales with the end points “strongly
disagree” to “strongly agree”: “During the listening, I concentrated very hard on the
photo samples.”, “I paid a lot of effort when I evaluated the photo softwares.”. They were
averaged into a composite index of effort (α = .88).

27
Results
Manipulation checks. The subjective knowledge index was analyzed by a one-
way ANOVA. As predicted, participants assigned to the high SK condition reported the
highest subjective knowledge (M = 3.49). Those in the low SK condition reported the
lowest subjective knowledge (M = 2.75). Subjective knowledge reported by the average
SK group was in the middle (M = 3.09). A linear contrast analysis confirmed that
participants’ subjective knowledge was a linear function of the experimental conditions
(F(1, 106) = 8.50, p = .004).
Evaluative extremity. The subjective knowledge manipulation produced the
expected main effect on evaluative extremity, which was measured by the variance and
the range of product ratings provided by individual participants. First, the linear contrast
analysis for the variance was significant (F(1, 106) = 6.74, p = .01), such that high SK
showed the largest (M = 16.50) and low SK had the smallest standard deviation across
their four product ratings (M = 12.44). The standard deviation produced by the average
SK group was in the middle (M = 14.01). The range of product ratings provides
convergent evidence to the relationship between subjective knowledge and extremity
(M
high SK
= 37.56 vs. M
average SK
= 31.53 vs. M
low SK
= 26.83), (F(1, 106) = 9.24, p = .003).

28
The difference in ratings of two identical products was examined to rule out the
possibility that those high in SK perceived differences where they did not exist (i.e., did
not rely on their subjective experience that the items were identical but instead
differentiated among items on some other basis). The linear contrast analysis was not
significant (M
high SK
= 17.31 vs. M
average SK
= 11.05 vs. M
low SK
= 13.69), (F(1, 106) = 1.85,
NS).. In sum, these results lend support to H1 that the higher the subjective knowledge
the greater the evaluative extremity.
Objective knowledge. Despite random assignment of respondents, a possible
confounding variable was participants’ objective knowledge, which might have partially
contributed to the proposed effects on extremity. Further analyses suggest otherwise.
First, there was no difference in the quiz scores among the three experimental groups
(9.39 vs. 9.17 vs. 9.08), (F < 1). Second, a median split on quiz scores was performed and
entered into analyses as an independent variable. Objective knowledge did not yield any
significant main or interaction effect (Fs < 1). Therefore the subjective knowledge effects
cannot be attributed to the differences in objective knowledge.
Confidence and effort. Since the status of high knowledge might have been
particularly motivating and confidence-bolstering, I explored potential differences in the

29
confidence in judgments and effort expended in the task by examining self-reported
confidence in evaluating the product. The oneway ANOVA revealed that there was no
group difference in confidence (M
high SK
= 3.57 vs. M
average SK
= 3.42 vs. M
low SK
= 2.99),
(F(2, 106) = 1.50, NS), The comparison between the high and low SK groups showed
only a marginal difference (t(71) = 1.67, p = .10). Apparently, the manipulation of
subjective knowledge did not affect how confident the participants were in their
evaluations. Moreover, confidence did not reliably predict extremity (F(1, 107) = 2.14,
NS). It is therefore unlikely that the effect of subjective knowledge on evaluative
extremity was driven by the confidence in judgment (table 2).
Self-reported effort revealed neither group difference (M
high SK
= 4.54 vs. M
average
SK
= 4.36 vs. M
low SK
= 4.20), (F (1, 106) = 1.10, NS), nor an effect on evaluative
extremity (F < 1, NS), (table 2).

30
Table 2: Extremity as a Function of Subjective Knowledge in Study 1

Dependent Measure
High
Subjective
knowledge
Average
Subjective
knowledge
Low
Subjective
knowledge F
Manipulation Check

Subjective knowledge
3.49
a
3.09
a,b
2.75
b
8.50
Product Evaluations

Product rating variance
16.50
a
14.01
a,b
12.44
b
6.74
Range
37.56
b
31.53
a,b
26.83
b
9.24
Difference in ratings of
two identical products
17.3
a
11.05
b
13.69
a,b
NS
Control Variables

Confidence
3.57 3.42 2.99 NS
Effort
4.54 4.36 4.20 NS
Interest in photography
3.56 3.44 3.16 NS
Familiarity
3.78 3.56 3.24 NS
Perceived quiz
difficulty
5.53
a
5.75
a
6.11
b
NS
Predicted scores
7.75 6.83 6.49 NS
Objective knowledge*
9.39 9.17 9.08 NS
N
36 36 37
Note - Means across the rows for a given dependent variable with different superscripts are different at p <
.05. Higher means indicate greater subjective knowledge, greater variance in ratings, larger range, larger
difference in ratings, greater confidence, greater effort, greater interest, higher familiarity, greater perceived
quiz difficulty, higher predicted scores, and higher objective knowledge.
* Objective knowledge is measured by the number of correct answers to an 18-question knowledge quiz
about digital photography.

31
Discussion
Incorporating two different measures of evaluative extremity, namely the variance
and range of product ratings, the results support the notion that higher subjective
knowledge is associated with greater evaluative extremity. Specifically, ratings of four
visual stimuli provided by the high SK group were more dispersed than those provided by
low SK group. The former also showed a larger range in their ratings than did the latter.
On both measures of dispersion, people high in subjective knowledge were more extreme
in the evaluative ratings than their lower subjective knowledge counterparts.
Alternative explanations, such as confidence in the overall product judgment and
motivation to process, cannot account for the observed extremity effect. First, the
manipulation of subjective knowledge did not influence confidence and effort. Second,
neither confidence nor effort predicted extremity. Objective knowledge as a rival
explanation is also ruled out because there was no significant objective knowledge effect
on extremity.
Two limitations of study 1 deserve further attention. First, although the photo
samples were qualitatively different from each other, quality differences cannot be
objectively determined because of the random combinations of visual parameters in

32
producing the stimuli. Differentiation among these products was thus idiosyncratic. It is
therefore difficult to determine the accuracy of the evaluations and the nature of the
observed extremity. In study 2, the quality difference between target products was
objectively manipulated such that one target product was objectively superior to the
other. Whether the direction of the extremity depends on the favorability of product
quality was examined in the following study.
Second, the within-subject design in the present study may have encouraged
participants to either increase or reduce the dispersion of ratings across a set of products.
As mentioned earlier, the premise of evaluating four software programs might have
demanded participants to differentiate, therefore producing a greater extremity than
warranted in a natural environment where consumers focus on fewer choices. In study 2,
the subjective knowledge effect was examined in a between-subjects setting, such that
each participant evaluated either an objectively superior or inferior product. The range of
ratings between the superior and inferior products thus serves as an indicator of
extremity. Also, to enhance the generalizability of the findings, a product category
involving different sensory input (music) was used in the next study.

33
Chapter 6: Study 2
Method
Design and procedure. Ninety-four undergraduate students were randomly
assigned to one of four conditions in a 2 (subjective knowledge: high vs. low) x 2
(product quality: superior vs. inferior) between-subject design. Participants were asked to
evaluate a music software program that was supposed to convert music from analogue
into digital formats.
As in study 1, all participants first completed a 20-question knowledge quiz,
which purportedly measured their knowledge and expertise about digital music. Upon
completion of the quiz, half of the participants (in the high subjective knowledge
condition) were told that their scores were in the top 88
th
percentile whereas the other half
(in the low subjective knowledge condition) were told that their scores were in the bottom
12
th
percentile. Their actual quiz scores, which were not available to the participants until
the end of the study, served as the measure of objective knowledge. After learning their
bogus quiz performance, all participants completed the same manipulation check items as
in study 1.

34
Next, they were asked to try the focal software by listening to music samples
ostensibly encoded by the software. I tried to minimize attribution of sound quality to the
hardware (i.e., headphones, soundcards) by providing a reference sample, which was
supposed to be a MP3 file encoded by a commonly-used software, before they listened to
the target sample. Specifically, every participant listened to two samples in succession.
The first sample was believed to be a standard MP3 file as a point of reference. The
second sample was the same music segment purportedly encoded by the focal software.
To manipulate the sound quality, I produced two 90-second music segments (excerpt of
the Brahms Symphony No. 1) using an editing software “Audacity 1.3.3”. The two
segments were identical except that the good-quality version was encoded in a standard
bit-rate (192 kbps) whereas the poor-quality version was encoded in a lower bit-rate (96
kbps). In addition, I used a 24-track equalizer to suppress the high and low frequency
ranges in the latter version to further reduce its quality (Shapiro and Spence 2002).
These two versions were randomly assigned as the reference and target samples
respectively. Specifically, participants in the superior quality condition first listened to
the mediocre version as the reference sample and then the good version as the target
sample. In the inferior quality condition, the order was reversed such that the good

35
version became the reference sample whereas the mediocre version became the target
sample. After listening to the two samples, participants wrote a review of the focal
software, evaluated the product and completed items that measured their confidence and
effort in the task and demographics. Upon completion, they received their actual quiz
scores and were thoroughly debriefed.
Measures. Manipulation checks of subjective knowledge, confidence and effort
measures were the same as those used in study 1. The focal music product was evaluated
by three bi-polar 100-point scales, with anchors positive vs. negative, favorable vs.
unfavorable, and high quality vs. low quality. These items were high in internal
consistency (α > .90) and averaged into a composite index of product ratings. Unlike
study 1 in which four photo software programs were evaluated, this study let participants
evaluate only one music software program. In addition to product ratings, open-ended,
written reviews were provided by participants.

Results
Manipulation checks. Subjective knowledge was assessed by four 7-point scales
(α = .85). They were averaged into a subjective knowledge composite and submitted to a

36
2 (subjective knowledge: high vs. low) x 2 (product quality: superior vs. inferior)
ANOVA. Only a significant main effect of subjective knowledge manipulation emerged.
The high SK group reported higher subjective knowledge about digital music (M = 4.52)
than did the low SK group, (M = 3.41), (F(1, 90) = 18.89, p < .001). Apart from the
expected main effect of the subjective knowledge manipulation, no other main or
interaction effects were found.
Product evaluations and extremity. Participants were given three 100-point scales
(“positive”, “favorable”, “of good quality”) to evaluate the focal music software. They
demonstrated high internal consistency (α = .97) and were averaged into a product
evaluation index, which was then analyzed by a 2 (subjective knowledge: high vs. low) x
2 (product quality: superior vs. inferior) ANOVA. The results reveal that there was a
significant main effect of sound quality on the evaluation. Specifically, the superior
version of the target music segment was evaluated more favorably than was the inferior
version (70.21 vs. 53.74), (F(1, 90) = 10.52, p = .002). More importantly, the main effect
of sound quality was qualified by a significant interaction between subjective knowledge
and sound quality (F(1, 90) = 12.65, p = .001).

37
Planned contrasts suggest that the high SK group gave lower evaluation ratings
than did the low SK group when the sound quality was inferior (42.39 vs. 65.08), (t(45) =
-2.87, p = .006). However, when the sound quality was superior, high SK group gave
higher evaluation ratings than did low SK group (76.92 vs. 63.49), (t(45) = 2.10, p = .04).
The range of ratings between objective superior and inferior products was greater among
high SK (34.53) than low SK individuals (1.59).
The results support the hypothesis that product differentiation depends on
subjective knowledge. Specifically, high SK group demonstrated greater differentiation
by giving relatively more extreme evaluations than did low SK group. Conversely, the
low SK group failed to discriminate between objectively superior and inferior quality by
giving relatively more moderate evaluations than did the high SK group.
Written reviews. Participants provided written reviews after listening to the music
samples and before giving evaluation ratings. An independent coder who was blind to the
hypotheses and the conditions to which the participants were assigned coded the reviews
into the numbers of positive and negative statements about the focal software.
The numbers of positive and negative statements were analyzed by a 2 (subjective
knowledge: high vs. low) x 2 (product quality: superior vs. inferior) x 2 (valence of

38
statements: positive vs. positive) mixed-design ANCOVA with the first two as between-
subject and the last as within-subject factors, and the total number of statements in a
review as a covariate. As expected, a significant two-way interaction between sound
quality and valence of statements (F(1, 87) = 14.07, p = .001), and a three-way
interaction emerged (F(1, 87) = 14.69, p = .001).
To further analyze the three-way interaction, I examined the two-way interaction
between sound quality and valence of statements separately for high and low SK groups.
For the high SK group, the two-way interaction was significant (F(1, 44) = 22.30, p <
.001). When the sound quality was superior, positive statements outnumbered negative
statements (1.54 vs. .07), (t(23) = 5.00, p < .001). However, when the sound quality was
inferior, the reverse emerged such that negative statements outnumbered positive
statements (1.10 vs. .32), (t(22) = 2.38, p =.03). The high SK group demonstrated again
their ability to differentiate by writing reviews that reflected the objective quality of the
focal product.
For the low SK group, however, only a significant main effect of valence emerged
such that there were more positive than negative statements regardless of sound quality
(.97 vs. .41), (F(1, 44) = 6.48, p = .01). The two-way interaction was not significant (F <

39
1, NS). Planned contrasts showed that when the quality was superior, the low SK group
gave more positive than negative statements in their reviews (.97 vs. .41), (t(22) = 2.66, p
= .02). When the quality was inferior, they still provided more positive than negative
statements (1.14 vs. .55), (t(23) = 2.70, p = .01). Hence, the low SK group failed to
differentiate (gave predominantly positive reviews) regardless of the objective quality
difference. Together, these results based on written reviews lend support to H2.
Objective knowledge, confidence, and effort. As in study 1, objective knowledge
was measured by the scores on a knowledge quiz. To examine the confounding effect, if
any, of objective knowledge, the data was analyzed in a 2 (sound quality) x 2 (subjective
knowledge) x 2 (objective knowledge) ANOVA. The last independent variable was
created by a median split based on the actual quiz’s scores. Results show no main effect
or high-order interaction that involve objective knowledge. The effects of the interaction
between subjective knowledge and sound quality on product evaluations, positive and
negative statements in written reviews, and the thought valence remained significant (ps
< .005).
The manipulation of subjective knowledge did not influence participants’
confidence and effort in the evaluations. Analyzing confidence in a 2 (sound quality) x 2

40
(subjective knowledge) ANOVA did not produce any main or interaction effects. People
in the high SK conditions were as confident as those in the low SK conditions (3.92 vs.
3.91), (F < 1, NS). A similar analysis of effort gave the same pattern. People in the high
SK conditions reported as much effort as did those in the low SK conditions (4.99 vs.
4.91), (F < 1, NS)

41
Table 3: Evaluations as a Function of Subjective Knowledge and Product Quality in
Study 2

High Subjective Knowledge Low Subjective Knowledge
Dependent Measure
Superior
Quality
Inferior
Quality
Superior
Quality
Inferior
Quality
Manipulation Check
Subjective knowledge 4.42
a
4.59
a
3.63
b
3.21
b

Product Evaluations
Product ratings 76.92
a,c
42.39
b
63.49
c
65.08
c

Positive statements in
written reviews
*

1.54
a
.32
b
.97
c
1.14
c

Negative statements in
written reviews
*

.07
a,c
1.10
b
.41
c
.55
b,c

Control Variables
Confidence 4.42 3.50 3.87 3.96
Effort 5.06 4.91 4.85 4.98
Interest in digital
music
5.00
a
5.28
a
4.07
b
3.65
b

Perceived quiz
difficulty
4.54 3.70 4.78 4.71
Predicted scores 11.38 12.57 10.83 10.88
Objective knowledge 13.67 13.78 12.96 13.42
N 24 23 23 24
Note - Means across the rows for a given dependent variable with different superscripts are different at p <
.05. Higher means indicate greater subjective knowledge, higher product ratings, higher number of
statements, greater confidence, greater effort, greater interest, higher familiarity, greater perceived quiz
difficulty, higher predicted scores, and higher objective knowledge.
*Adjusted means with total number of statements as a covariate.

42
Discussion
Replicating the results of study 1, study 2 showed that product evaluations were
influenced by subjective perceptions of domain knowledge. Specifically, high subjective
knowledge individuals gave more extreme evaluations than did their low subjective
knowledge counterparts, resulting in greater differentiation between high and low-quality
products.
In addition, the contents of written product reviews provided convergent evidence
of the subjective knowledge effect on evaluations. Open-ended written reviews afford
more opportunities to look into how valenced information is processed. If high subjective
knowledge people process both positive and negative information, the valence of their
reviews should be consistent with the product quality. However, if low subjective people
are less willing to process negative information, their reviews should be positive
regardless of whether the product is good or bad. Consistent with their product ratings,
written reviews provided by the high SK group reflected the quality of the music
products, showing the ability to differentiate between objectively good and mediocre
product alternatives. Low SK group not only failed to differentiate by giving

43
approximately the same product ratings but also became predominantly positive in the
reviews regardless of the disparity in product quality, lending support to H2.
The positivity in the reviews provided by the low SK group suggests that people
low in subjective knowledge appeared to have overlooked the negative product
information. Based on the objective knowledge measures, they were as knowledgeable
about digital music as their high SK counterparts, and hence both high and low SK group
should be equally capable of differentiating between products of different quality.
However, they appeared to infer a lack of processing capability from their impoverished
perception of knowledge. As a result, they did not search for the negative information as
much as the high SK group did. Should that be the case, a direct instruction to facilitate
their search for negative product information should reduce the observed positivity in
their reviews, and in turn lead to greater evaluative extremity and product differentiation.
The next study provided a direct empirical test by giving such an instruction. In addition,
motivation to process was also manipulated to rule out this alternative explanation.

44
Chapter 7: Study 3
The main purpose of this study is to rule out motivation to process as an
explanation for the proposed subjective knowledge effect on extremity and
differentiation. An alternative explanation for the findings in previous studies is that
those low in subjective knowledge lacked the motivation to pay attention to differences
among stimuli and so provided less extreme evaluations. Although self-reported effort
did not differ between high and low subjective knowledge participants in previous
studies, the measures might be subject to social desirability. Experimentally manipulating
motivation and objectively measuring how motivated low subjective knowledge
participants are (in terms of the length of their reviews and total time they spent in the
study) can help rule out the effects of low motivation on product evaluations of
differentiation.
Another important goal is to investigate whether the processing of negative
information affects evaluative extremity. The results in study 2 shows that the low SK
group did not discriminate between products of good and mediocre quality. Their reviews
were predominantly positive regardless of product quality, suggesting that they did not
process the negative information as much as they processed the positive information.

45
Therefore, a direct instruction to ask them to focus on negative product information
should reduce such positivity bias. As a result, the low SK group should be able to
differentiate as well as the high SK group.
In addition, this study also measures incidental positive and negative mood after
the manipulation of subjective knowledge to rule out the potential effects of mood.

Method
One hundred and twenty-two undergraduate students were randomly assigned to
one of the six conditions of a 2 (subjective knowledge: high vs. low) x 3 (incentive: no
incentive, accuracy, negative-focused) between-subject factorial. As in study 1 and 2,
subjective knowledge about digital music was manipulated by bogus feedback on a
knowledge quiz. To test whether the motivation to process explains the effects of
subjective knowledge on evaluations, those who were assigned to the accuracy conditions
were told that they could receive a monetary reward of $5 if they accurately evaluated
the target product. Participants in the negative-focused conditions were instructed that
they could earn $5 by focusing on the negative information about the product (e.g.,
something they do not like about the product). The no-incentive conditions did not

46
include any monetary incentive and served as control conditions. Unlike study 2 in which
product quality varied, this study exposed participants only to an inferior music product.
Key measures included manipulation checks of subjective knowledge, product
evaluations and written reviews. All scale items were the same as in study 2. In addition,
ten items that measured incidental positive and negative mood (i.e., happy, excited,
proud, pleased with myself, calm, distressed, sad ashamed, humiliated, incompetent)
were also included to examine the possibility that mood influences evaluations. To assess
the effect of the monetary incentive on the motivation to process, the total time spent in
the study was recorded for each participant individually.

Results
Manipulation checks. The manipulation of subjective knowledge was successful.
The subjective knowledge composite index was submitted to a 2 (subjective knowledge:
high vs. low) x 3 (types of incentive: no incentive, accuracy, negative-focused) ANOVA
and revealed only a significant main effect of subjective knowledge. Participants who
received favorable quiz feedback indicated greater subjective knowledge (M = 4.45) than
did those who received unfavorable quiz feedback (M = 3.61), (F(1, 118) = 9.21, p =

47
.003). To assess the effectiveness of the manipulation of incentive, the totally time spent
in the study was analyzed by the 2 (subjective knowledge) x 3 (types of incentive)
ANOVA. Only the main effect of incentive was significant (F(1, 118) = 13.22, p < .001).
A planned contrast test showed that participants spent longer time in the study when they
received an incentive to be accurate (M = 1003.12 seconds) or an incentive to be
negative-focused (M = 980.56 seconds) than when there was no incentive (M = 713.53
seconds). Having an incentive apparently motivated participants to work harder.
Product evaluations and extremity. Product ratings of the focal music software
were also analyzed in a 2 (subjective knowledge) x 3 (types of incentive) ANOVA. A
significant main effect of subjective knowledge emerged such that high SK people gave
lower ratings (M = 40.92) than did low SK people (M = 57.80), (F(1, 117) = 10.01, p =
.002). There was also a marginally significant interaction between subjective knowledge
and types of incentive (F(2, 117) = 2.33, p = .10). To further understand this interaction,
pairwise comparisons were performed separately for each incentive condition.
In the no-incentive control condition, high SK people gave lower ratings (M =
36.92) than did low SK people (M = 61.08), (t(38) = -2.77, p = .009). Recall that only the
objectively inferior music product was used in the current study. The relatively high

48
ratings provided by the low SK participants suggested that they failed to accurately
evaluate the product.
If low SK participants’ failure to differentiate is due to a lack of motivation, an
incentive to reward accuracy in ratings should have motivated them to scrutinize the
product more effortfully. Their ratings should then reflect the inferiority in the product
quality. In the “incentive to be accurate” condition, although participants did spend more
effort in the product evaluation, the ratings provided by the low SK group were still
higher than those provided by the high SK group (70.06 vs. 44.05), (t(41) = 2.75, p =
.009). The extra incentive did not help the low SK people become more accurate. Hence,
motivation, or the lack thereof, does not seem convincing as an explanation for the effects
of subjective knowledge on evaluative extremity.
The low SK group was accurate only when they were encouraged to process
negative information. Specifically, when there was an incentive to focus on the negative
aspects of the product, both high and low SK people gave similarly low ratings to the
product of inferior quality (41.77 vs. 42.25), (t(39) = .05, NS).

49
Written reviews. An independent coder who was blind to the hypotheses and the
conditions to which the participants were assigned coded the reviews into the numbers of
positive and negative statements about the focal software.
The numbers of positive and negative statements were analyzed by a 2 (subjective
knowledge) x 3 (types of incentive) x 2 (valence of statements) mixed-design ANCOVA
with the first two as between-subject and the last as within-subject factors, and the total
number of statements in a review as a covariate. As predicted, a significant three-way
interaction emerged (F(2, 116) = 8.89, p = .01).
To further analyze the three-way interaction, I examined the two-way interaction
between types of incentive and valence of statements separately for high and low SK
groups. For the high SK group, the two-way interaction was not significant (F < 1, NS).
Only a significant main effect of valence emerged. Negative statements always
outnumbered positive statements (1.90 vs. .56), (F(1, 63) = 5.63, p = .001). High SK
group appeared to identify the inferiority in the product and expressed their unfavorable
opinions in the reviews regardless of whether or not there was an incentive. Among those
low in subjective knowledge, the interaction between types of incentive and valence was
significant (F(2, 56) = 3.68, p = .03). Pairwise comparisons were conducted to further

50
look into the effects. In the no-incentive condition, the numbers of positive and negative
statements did not differ (1.16 vs. .89), (t < 1, NS), demonstrating a positivity bias for an
objectively inferior product. When the low SK group was motivated by an incentive, the
numbers of positive and negative statements did not differ either (1.59 vs. .94), (t < 1,
NS). Only when the low SK group was instructed to processing negative information,
positive statements was outnumbered by negative statements (.64 vs. 2.36), (F(1, 18) = -
3.23, p = .006).
Confidence, effort, and mood. As in previous studies, there was no confidence or
effort effect on evaluations or written reviews. The manipulations of subjective
knowledge and incentive did not influence mood. Participants’ positive and negative
mood had no effect on product evaluations.

51
Table 4: Evaluations as a Function of Subjective Knowledge and Motivation to
Process in Study 3

No Incentive
Incentive to be
Accurate
Incentive to Process
Negative Info
Dependent Measure High SK Low SK High SK Low SK High SK Low SK
Manipulation Checks
Subjective knowledge 4.85 3.68 4.21 3.44 4.46 3.70
Total time (seconds) 681.05 753.65 845.63 1003.12 880.79 980.56
Product Evaluations
Product ratings 36.92 61.08 44.05 70.06 41.77 42.25
Positive statements in
reviews
*

.41 1.16 .62 1.59 .58 .64
Negative statements in
reviews
*

1.65 .89 2.00 .94 1.74 2.36
Control Variables
Confidence 4.35 4.21 4.62 4.68 4.63 4.53
Effort 5.82 5.29 5.95 5.41 5.80 5.69
Positive mood 4.02 4.25 4.13 3.90 4.32 4.61
Negative mood 2.22 2.40 1.76 2.04 1.97 1.84
Interest in digital music 5.21 4.13 5.03 4.15 4.88 4.38
Familiarity 5.21 4.40 4.96 3.80 4.92 4.22
Perceived quiz
difficulty
3.94 4.29 4.36 4.59 4.30 4.25
Predicted scores 13.00 12.48 13.20 10.41 12.44 13.88
Objective knowledge
**
13.59 14.00 14.05 13.12 14.69 14.63
N 20 19 23 20 21 19
Note - Higher means indicate greater subjective knowledge, higher product ratings, higher number of
statements, greater confidence, greater effort, greater interest, higher familiarity, greater perceived quiz
difficulty, higher predicted scores, and higher objective knowledge.
** Objective knowledge is measured by the number of correct answers to a 20-question knowledge quiz
about digital music.

52
Discussion
Motivation to process does not explain the effects of subjective knowledge on
evaluative extremity. Compared to people in the no-incentive condition, those who
received a monetary incentive to accurately evaluate products did work harder on the
product evaluations, as shown by the longer time they spent. However, the accuracy
incentive did not make the low SK group more accurate. High SK people, regardless of
whether or not an incentive was present, gave product ratings that reflected the objective
quality. In comparison, low SK people who either received an accuracy incentive or had
no incentive were less extreme because they failed to give low ratings to the target
product. These results are consistent with the proposed theory that subjective knowledge
influences extremity and help rule out motivation to process as an alternative explanation.
If the low SK individuals’ failure to provide unfavorable ratings was due to their
reluctance to processing negative information, a direct instruction that encouraged
processing of negative information should make them more willing to give unfavorable
reviews and low ratings to an objectively inferior product. Data in the negative-focused
condition supported this hypothesis, showing that those who were low in subjective

53
knowledge did have the ability to accurately identify the negative attributes when the
quality was inferior.

54
Chapter 8: Study 4
The primary goal of study 4 is to examine public accountability as a boundary
condition for the subjective knowledge effect. Public accountability refers to the “implicit
or explicit expectation that one may be called on to justify one’s beliefs, feelings and
actions to others” (Lerner and Tetlock 1999; p. 255). It is proposed that high subjective
knowledge increases consumers’ propensity to differentiate. Performing behaviors that
are consistent with high knowledge, such as spreading alternatives apart by increasing the
extremity in ratings, should further reinforce one’s favorable perception of knowledge
(Goldstein and Cialdini 2007). However, when the product evaluations are subject to
judgments by others, potential disagreement can lead to disconfirmation of their
perception of knowledge. As a result, the favorable self-perception cannot be maintained.
Therefore, when people are subject to public scrutiny, they will adopt a conservative
strategy such that they give less extreme product ratings because people generally believe
that extreme attitudes are harder to justify or defend (Cialdini et al. 1973). Therefore
public accountability should reduce extremity in product ratings.
To ensure the generalizability of the subjective knowledge, I used a different
manipulation of subjective knowledge in the present study. Specifically, social

55
comparison information was provided at the beginning of the study to alter participants’
perception of knowledge (See 2009). In addition, public accountability was manipulated
by a verbal instruction after the manipulation of subjective knowledge.

Method
Design and procedure. One hundred and seventy-three undergraduate students
took part in the study for course credit. They were randomly assigned to one of the four
conditions in a 2 (subjective knowledge: high vs. low) x 2 (public accountability: high vs.
low) between-subject factorials. As in study 1, they were purportedly asked to evaluate a
set of 4 anonymous photo software programs. Before they inspected the photos
supposedly produced by these programs, they received the manipulation of subjective
knowledge by social comparison information (See 2009). Specifically, the manipulation
of subjective knowledge was achieved by introducing either a knowledgeable or
unknowledgeable referent group in the beginning of the study. Participants assigned to
the high-SK condition were told that the same study would also be administered to “a
group of high school students who have no experience in digital photography”. Those
who were assigned to the low-SK condition were told that the same study would also be

56
administered to “a group of professional photographers and graphic designers who work
with digital images”.
Immediately after the manipulation of subjective knowledge, people indicated
their subjective knowledge about digital photography in four 7-point scale manipulation
check items (same as previous studies). Next, they encountered the manipulation of
public accountability in the form of a written instruction (Diehl and Stroebe 1987). In the
high-accountability condition, people read the instruction as follows: “Evaluating the
quality of photographic products requires knowledge and critical ability. There are
objectively right and wrong answers. Your product reviews and evaluation ratings will be
analyzed to see how accurate you are.” Whereas in the low-accountability condition,
participants read a different version of instruction: “Evaluating the quality of
photographic products is a matter of personal preference and taste. There is no
objectively right or wrong answer. Your product reviews and evaluation ratings will NOT
be subject to any judgment or comparison.”
After receiving the manipulation of public accountability, participants inspected 4
photo samples ostensibly enhanced by the 4 anonymous programs in question. As in
studies 1, all photo samples were digitally modified based on the same stock photo and

57
given different combinations of visual characteristics. Unlike study 1 in which 2 samples
(B and D) were in fact identical, the current study presented 4 different samples labeled
A, B, C, and D respectively. After inspecting a photo sample, people provided product
ratings, and then moved on to the next photo until all 4 photos were inspected and
evaluated.
Upon the completion of product evaluations, participants reported their
confidence in the product evaluations, the concerns for public evaluations as a
manipulation check of public accountability, the amount of effort they expended in the
task, and completed a 18-question quiz that examined their knowledge about digital
photography. Note that in the studies 1, 2, and 3, the quiz as well as the bogus quiz scores
were given before the product evaluation as a manipulation of subjective knowledge. In
the present study, the quiz was placed at the end of the study as a measure of objective
knowledge. Finally, participants provided their demographic data, thoroughly debriefed
and thanked.
Measures. Subjective knowledge about digital photography was assessed by 3
seven-point items adapted from Brucks (1985) and Park et al. (1994): “I think of myself
as an expert on digital photography” (1 = strongly disagree; 7 = strongly agree), “I am

58
clueless about” (1 = strongly disagree; 7 = strongly agree; reversed coded), “how familiar
are you with digital photography?” (1 = very unfamiliar; 7 = very familiar), and “how
much experience do you have with digital photography compared to your friends?” (1 =
very inexperienced; 7 = very experienced). These items were averaged into a subjective
knowledge index (α = .81).
Each photo program was evaluated by three bi-polar 100 points scales (positive
vs. negative, favorable vs. unfavorable, and high quality vs. low quality). They were
averaged into four product ratings for the four programs respectively (αs > .90). As in
study 1, the extremity in product ratings was obtained by calculating the standard
deviation of the four ratings and the range of the ratings for each participant (Linville
1982). In addition, the time people spent on inspecting each photo was recorded. The four
time measures were summed into the total time of photo inspection (in seconds). After
people provided their product ratings, they answered questions that measured their
concerns for being evaluated by others. Specifically, three items adapted from Leary et al.
(1987) were used: “How concerned are you with performing well on this product
evaluation task?” (0 = not at all concerned; 100 = very concerned), “Did you feel anxious
about evaluating the photo software products?” (0 = not at all anxious; 100 = very

59
anxious), and “Did you feel any pressure in evaluating the photo software products?” (0 =
no pressure at all; 100 = a lot of pressure). These items demonstrated good internal
consistency and averaged into a composite index of evaluation concern (α = .75).
The items used to assess confidence and effort were the same as previous studies.
They formed an index of confidence and an index of effort respectively (αs > .90).
Objective knowledge was measured based on the number of quiz items each participant
correctly answered, with a possible maximum of 18 and a minimum of zero.

Results
Manipulation checks. The subjective knowledge index was analyzed by a 2
(subjective knowledge: high vs. low) x 2 (public accountability: high vs. low) ANOVA.
Only a significant main effect of subjective knowledge emerged such that people in the
high SK condition reported higher subjective knowledge (M = 4.14) than did those who
in the low SK condition (M = 3.57), (F(1, 169) = 12.34, p = .001). Similarly, the
evaluation concern index was submitted to the same 2 (subjective knowledge) x 2 (public
accountability) ANOVA. The only significant effect was the main effect of public
accountability. Those who were in the high accountability condition had a greater

60
concern about being judged (M = 34.68) than did those who were in the low
accountability condition (M = 28.80), (F(1, 169) = 4.38, p = .04). No other main or
interaction effect was in these analyses, confirming the success of the manipulations of
subjective knowledge and public accountability respectively.
Objective knowledge (i.e., the knowledge quiz scores) was examined in the same
2 x 2 ANOVA. Unexpectedly, there was a significant main effect of subjective
knowledge such that high SK participants performed better (M = 9.34) than their low SK
counterparts (M = 8.63), (F(1, 169) = 3.65, p = .001). To control for the possible
confounding effects of objective knowledge, all the dependent variables were analyzed
with objective knowledge as either an independent variable or a covariate. There was no
significant effect that involved objective knowledge. Therefore all data analyses excluded
objective knowledge as a factor.
Product evaluations and extremity. The two indicators of extremity, namely the
variance and the range of product ratings, were separately analyzed. First, the 2
(subjective knowledge) x 2 (public accountability) ANOVA revealed a significant main
effect of public accountability on the variance of ratings (F(1, 169) = 6.79, p = .01).
Specifically, those who were in the high accountability condition produced a smaller

61
variance (M = 15.38) than did those who were in low accountability condition (M =
18.52). This significant main effect was consistent with the notion that when made
accountable, people tend to be moderate in evaluating products. More importantly, this
main effect was qualified by an interaction between subjective knowledge and
accountability (F(1, 169) = 5.02, p = .03). To further understand the interaction effect,
pairwise comparisons were done separately for high and low accountability conditions. In
the low accountability condition, high SK participants were more extreme in their ratings
(M = 20.52) than low SK participants (M = 16.53), (t(87) = 2.15, p = .03), replicating the
subjective knowledge effect unveiled in previous studies. In the high accountability
conditions, however, the evaluative extremity exhibited by high and low SK was not
different (14.66 vs. 16.09), (t(82) = .95, NS). Among those who were high in subjective
knowledge, they appeared to reduce the extremity in ratings when they expected to be
accountable (M = 14.66) than when they did not expect public accountability (M =
20.52), (t(83) = 3.43, p = .001).
Confidence and effort. The 2 subjective knowledge x 2 accountability ANOVA
revealed two significant main effects. High SK participants were more confident in their
product evaluations (M = 3.67) than their low SK counterparts (M = 3.24), (F(1, 169) =

62
4.54, p = .04). People in the low accountability condition (M = 3.74) were also more
confident than those who were in the high accountability condition (M = 3.16), (F(1, 169)
= 6.39, p = .004).
Self-reported effort did not reveal any difference among experimental conditions.
As discussed in previous studies, explicit measures of effort might be influenced by
social desirability concerns. To examine the effects of subjective knowledge and
accountability on how much effort people expended, the total time spent on inspecting
photo samples, which was recorded unbeknownst to the participants, was analyzed by the
2 x 2 ANOVA. The main effect of subjective knowledge approached significance (F(1,
169) = 3.60, p = .06). High SK individuals spent more time scrutinizing the photo
samples (M = 42.7 seconds) than did low SK people (M = 36.83 seconds). There was a
significant main effect of accountability such that high accountability participants spent
longer time to inspect the photos (M = 43.69 seconds) than did low accountability people
(M = 35.31 seconds), (F(1, 169) = 8.86, p = .003). No interaction between subjective
knowledge and accountability was found. Although subjective knowledge affected the
amount of effort people expended in this study, effort did not appear to be a viable
alternative explanation. The inspection time and variance in ratings were not correlated

63
(r(173) = -.09, NS), which offers no support to the alternative explanation that extremity
is determined by participants’ effort.

64
Table 5: Evaluations as a Function of Subjective Knowledge and Public
Accountability in Study 4

High Public Accountability Low Public Accountability
Dependent Measure High SK Low SK High SK Low SK
Manipulation Checks
Subjective knowledge 3.97
a,b
3.60
a
4.30
b
3.54
a

Accountability concern
35.32
a
34.04
a,b
30.60
a,b
27.01
b

Product Evaluations
Product rating variance 14.66
a
16.09
a
20.52
b
16.53
a

Range 33.02
a
36.12
a
46.58
b
36.26
a

Control Variables
Confidence 3.28
a
3.05
a
4.06
b
3.43
a

Effort 4.65 4.77 4.92 4.51
Photo inspection time
(seconds)
48.23
a
39.15
b
36.11
b
34.51
b

Interest in photography 3.41
a,b
2.99
a
3.62
b
2.84
a

Perceived quiz difficulty
5.65
a,b
5.86
a
5.22
b
5.91
a

Predicted scores
6.63
a
5.86
a,b
5.22
b
5.91
a

Objective knowledge
**
9.18 8.68 9.51 8.57
N 40 44 45 44
Note - Means across the rows for a given dependent variable with different superscripts are different at p <
.05. Higher means indicate greater subjective knowledge, greater evaluation apprehension, greater variance
in ratings, greater range, greater confidence, greater effort, greater interest, greater perceived quiz
difficulty, higher predicted scores, and higher objective knowledge.
** Objective knowledge is measured by the number of correct answers to an 18-question knowledge quiz
about digital music.

65
Discussion
The present study empirically supports the notion that public accountability
moderates the effect of subjective knowledge. When participants believed that their
evaluations were subject to public scrutiny, evaluative extremity was lower than when
people believed that they were not accountable for their evaluations.
The behaviors in association with subjective knowledge can produce a mutually
reinforcing effect on self-perception of knowledge. When high subjective knowledge
people believe that they are capable of differentiating, they give extreme evaluations and
discriminate among product alternatives, which in turn can fortify their subjective belief
of knowledge. However, this intra-personal loop can be disrupted by social reality.
Individuals are cognizant that others can potentially invalidate one’s evaluations. For a
high SK person, this means that another may invalidate one’s favorable perception of
oneself. To shield the favorable self-perception of knowledge from the confrontation by
others, high SK people therefore reduce the extremity of their evaluations.

66
Chapter 9: Study 5
Study 5 introduced a different method to induce public accountability from study
4. Past research suggests that public accountability increases when people anticipate
public discussion of their opinions (Cialdini et al. 1973; Lerner and Tetlock 1999). In the
present study, participants expected to exchange their written product reviews with
another participants at the end of the session (Konrath, Bushman, and Campbell 2006).

Method
One hundred and ten undergraduate students were randomly assigned to one of
the four conditions in a 2 (subjective knowledge: high vs. low) x 2 (product quality:
superior vs. inferior) between-subject design. In the beginning of the study, participants
were instructed that they would listen to a music product and then give product ratings
and a written review. At the end of the session, they would exchange their written
reviews with another participant. The procedure for the rest of the study was similar to
that of study 2. People completed a knowledge quiz about digital music and received
either favorable or unfavorable scores. Then they indicated their subjective knowledge.
Next, they listened to a music sample of either good or mediocre sound quality.

67
Afterwards, they gave ratings and wrote a review of the music product. The items that
measured subjective knowledge, objective knowledge, product ratings, and the open-
ended product review were the same as in study 2.

Results
Manipulation checks. Subjective knowledge was assessed by four 7-point scales
(α = .84). They were averaged into a subjective knowledge composite and submitted to a
2 (subjective knowledge: high vs. low) x 2 (product quality: superior vs. inferior)
ANOVA. Only a significant main effect of subjective knowledge manipulation emerged.
The high SK group reported higher subjective knowledge about digital music (M = 4.49)
than did the low SK group, (M = 3.52), (F(1, 106) = 21.19, p < .001). Apart from the
expected main effect of the subjective knowledge manipulation, no other main or
interaction effects were found.
Product evaluations and extremity. Three 100-point scales (“positive”,
“favorable”, “of good quality”) to evaluate the focal music software were averaged into a
product evaluation index, which was then analyzed by a 2 (subjective knowledge: high
vs. low) x 2 (product quality: superior vs. inferior) ANOVA. The results reveal that there

68
was a significant main effect of sound quality on the evaluation. The superior version of
the target music segment was evaluated more favorably than was the inferior version
(71.17 vs. 57.51), (F(1, 106) = 12.25, p = .001), demonstrating the effectiveness of the
product quality manipulation. In addition, there was a marginally significant interaction
between subjective knowledge and sound quality (F(1, 106) = 3.41, p = .07). Planned
contrasts show that high SK group gave more favorable ratings to the superior product
than to the inferior product (73.22 vs. 52.36), (t(54) = 3.48, p = .001). However, low SK
group failed to differentiate between superior and inferior quality by giving similar
ratings (69.12 vs. 62.67), (t(52) = 1.30, NS). These results were consistent with those of
the previous studies, giving further support to the proposed effect of subjective
knowledge on extremity.
Written reviews. An independent coder who was blind to the hypotheses and the
conditions to which the participants were assigned coded the product reviews into the
numbers of positive and negative statements about the focal software, which were then
submitted to a 2 (subjective knowledge: high vs. low) x 2 (product quality: superior vs.
inferior) x 2 (valence of statements: positive vs. positive) mixed-design ANCOVA with
the first two as between-subject and the last as within-subject factors, and the total

69
number of statements in a review as a covariate. There was as significant main effect of
the valence of statements such that positive statements outnumbered negative statements
regardless of product quality (1.36 vs. .40), (F(1, 105) = 19.32, p < .001). A significant
interaction between the valence of statements and the product quality also emerged (F(1,
102) = 7.81, p = .006). The number of positive statements was greater when the quality
was superior (M = 1.78) than when the quality was inferior (M = .96), (t(105) = 3.49, p =
.001). However, the number of negative statements did not differ between the superior
and inferior products (.29 vs. .50), (t(105) = - 1.38, NS). It is intriguing to observe that the
inferior product did not generate more negative statements than did the superior product.
Together with the main effect that positive statements outnumbered negative statements,
these results suggest an apparent positivity bias in the reviews, which were believed to be
subject to public scrutiny.
No significant effect that involved subjective knowledge was found. Unlike in
study 2 in which high SK group wrote more negative than positive statements when the
product was inferior (1.10 vs. .32), in the current study when high SK people were made
public accountable for their reviews, the number of their positive and negative statements

70
did not differ (t(28) = .46, NS). Directionally, the number of positive statements was
greater than that of negative statements for an objective inferior product (.72 vs. .58).

71
Table 6: Evaluations as a Function of Subjective Knowledge and Product Quality in
Study 5

High Subjective knowledge Low Subjective knowledge
Dependent Measure
Superior
Quality
Inferior
Quality
Superior
Quality
Inferior
Quality
Manipulation Check
Subjective knowledge 4.51
a
4.48
a
3.72
b
3.32
b

Product evaluation
Product ratings 73.22
a
52.36
c
69.12
a,b
62.67
b

Positive statements in
written reviews
*

1.84
a
.91
b
1.54
a,c
1.20
b,c

Negative statements in
written reviews
*

.23 .62 .32 .40
Control Variables
Confidence 4.58
a
3.40
b
3.61
b
3.63
b

Effort 5.29 4.95 4.65 4.70
Interest in digital music 4.90
a
4.53
a,b
4.57
a,b
4.04
b

Familiarity 4.56
a,b
4.73
a
4.19
a,b
3.98
b

Perceived quiz difficulty
4.35
a,b
4.53
a,b
3.96
a
4.89
b

Predicted scores
12.50
a
12.80
a
12.67
a
10.59
b

Objective knowledge
**
13.85
a,b
13.97
a,b
14.07
a
12.89
b

N 26 30 27 27
Note - Means across the rows for a given dependent variable with different superscripts are different at p <
.05. Higher means indicate greater subjective knowledge, higher product ratings, higher number of
statements, greater confidence, greater effort, greater interest, higher familiarity, greater perceived quiz
difficulty, higher predicted scores, and higher objective knowledge.
*Adjusted means with total number of statements as a covariate.

72
Discussion
When the behavior of the participants (in this case, the written reviews) was
anticipated to be subject to public scrutiny, the concerns for justification appears to
increase (Lerner and Tetlock 1999). Participants became strategically positive in their
reviews regardless of the product quality perhaps because giving predominantly positive
evaluations was expected to project favorable traits such as generous and good-natured
(Folkes and Sears 1977). When there is an anticipation of public accountability, high SK
people behaved like their low SK counterparts. They gave less extreme ratings (study 4)
and became predominantly positive in their reviews (study 5).
It is worth noting that the public accountability applied only to the written reviews
and not to product ratings provided by the participants. This may explain the different
patterns observed in the product ratings and written reviews. Whereas high and low SK
participants differed in the way they rated the product, replicating the subject knowledge
effects on extremity found in previous studies, they both demonstrated a positivity bias in
written reviews. Perhaps they believed that product ratings would remain private,
therefore the need to manage a positive image was not as strong as in the written reviews,
which were anticipated to be exchanged in the end of the experimental session. The

73
specificity of accountability is a novel idea that has not been examined, and may warrant
further research in the future.

74
Chapter 10: Conclusion
Past research generally attributes the ability to differentiate among product
alternatives to the actual knowledge that consumers possess (Alba and Hutchinson 1987;
Lusk and Judd 1988). I propose that the self-perception of domain knowledge also exerts
an independent effect on the way consumers evaluate products. My five empirical studies
consistently support this novel idea. Compared to people who were low in subjective
knowledge, those who were high in subjective knowledge gave more extreme product
evaluations (studies 1 and 4), and better distinguished between objectively superior and
inferior products (studies 2, 3, and 5).
The data in studies 2, 3, and 5 indicate that the failure to differentiate among low
SK individuals is related to the lack of processing of negative product information.
Consumers generally expect that products have more positive attributes more than bad
ones (Herr et al. 1991); they believe negative product information is harder to identify
than positive information. The low expectancy to perform in association with low
subjective knowledge, coupled with the expectation that negative information is
relatively more difficult to find, make low SK people process less negative information
relative to high SK people. The predominantly positive product reviews given by the low

75
SK people, particularly when product quality was inferior, point to the proposed
difference in information processing between high and low SK people. Past research has
identified a variety of factors that leads to differential processing of valenced
information, such as framing of a decision (e.g., choosing versus rejecting; Shafer 1993),
priming of information search process (Shen and Wyer 2008), and regulatory focus
(Crowe and Higgins 1997). My research contributes to the information processing
literature by suggesting a new factor, namely subjective knowledge, which influences the
processing of negative information.
Motivation to process is ruled out as an alternative explanation. Self-reported
effort showed no difference between high and low SK groups. Although it is plausible
that people felt compelled to claim that they worked hard in the study because of social
desirability concerns, study 4’s objective measure of motivation and effort showed that
effort does not account for the effects. The high SK group did spend more time
examining the photo samples than did the low SK group. However, the photo inspection
time was not related to evaluative extremity. Together, these results suggest that the
subjective knowledge effect cannot be explained by the difference in the motivation to
process.

76
I found mixed results pertaining to confidence in judgment. All but one study
(study 4) revealed no difference in confidence between high and low SK groups.
Confidence did not reliably predict evaluative extremity or product differentiation. The
self-report nature of confidence could have induced the concern for social desirability,
which might reduce the reliability of the measures of confidence. Future research should
consider both direct and indirect measures of confidence to capture the relationship
between subjective knowledge, confidence, and domain-relevant performance (Petty,
Brinol, and Tormala 2002).

The Malleability of Subjective Knowledge
In my research, I operationally defined subjective knowledge as the consumer’s
perceived product knowledge relative to referent others (Burson 2007). I used two
different methods to manipulate subjective knowledge and obtained consistent results.
First, in studies 1, 2, 3, and 5, participants first completed a knowledge quiz and received
bogus feedback on the quiz performance prior to the product evaluations. They were told
that their performance was either better or worse than that of previous participants.
Participants thus inferred their subjective knowledge from their quiz performance relative

77
to a group of nonexistent participants. Second, in study 4, a referent group was
introduced at the beginning of the study. Participants were led to believe that either a
group of highly knowledgeable or a group of unknowledgeable people also took part in
the same study (See 2009). The mere mention of a social comparison target seemed to be
sufficient to either bolster (when the referent group was seemingly less knowledgeable)
or diminish one’s subjective knowledge (when the referent group was apparently more
knowledgeable).
Using two different subjective knowledge manipulations, either a relatively subtle
(i.e., referent group) or direct approach (i.e., bogus quiz feedback), not only gives
confidence to the consistent findings of the subjective knowledge effect, but also speaks
volumes about the malleability of subjective knowledge. Outside the laboratory, social
comparison targets are prevalent in the real world consumption situations. For example,
salespeople, acquaintances, or even total strangers in the same consumption environment
can exert social influences that change self-perceptions and behaviors (Argo, Dahl, and
Manchanda 2005). The changeability of subjective knowledge, coupled with the effects
of subjective knowledge on evaluations, suggest that the social environment in which
consumers engage may influence the way they evaluate and discriminate product choices.

78
Social presence in a consumption context can become a standard for comparison
and produce a contrast effect on subjective perception of knowledge. It is also plausible
that people assimilate their subjective knowledge assessment toward someone with
whom they are closely affiliated. According to the vicarious self-perception model,
people may infer their own attributes from others with whom they share a common
identity (Goldstein and Cialdini 2007). A close friend who is an expert on a certain
product may thus enhance an individual’s subjective knowledge should they be
psychologically merged (i.e., their self-identities are closely overlapped; Radecki and
Jaccard 1995) The assimilation of subjective knowledge should lead to corresponding
changes in product evaluations and discrimination. The presence of an expert who is also
a close friend should therefore make an individual become more extreme in rating
products. In sum, the relationship with a comparison target can determine whether
subjective knowledge is assimilated toward or contrasted away from the comparison, and
in turn influence how products are evaluated. This represents an interesting social
phenomenon that deserves further research efforts.

79
The Interplay between Subjective and Objective Knowledge
It is worth noting that the effect of subjective knowledge on product evaluation is
independent from that of objective knowledge. Objective knowledge as an independent
variable neither produced a main effect nor interacted with subjective knowledge. The
pronounced main effect of subjective knowledge is striking, particularly when low SK
was crossed with high OK and when high SK was crossed with low OK. When the
subjective knowledge of high OK people was diminished by a highly knowledgeable
referent group or unfavorable quiz feedback, they reduced the extremity in their product
ratings and even failed to discriminate between products of good and mediocre quality.
The opposite was observed when the subjective knowledge of the low OK individuals
was bolstered.
The absence of an objective knowledge effect raises concerns about the student
population in which experimental studies were conducted. Those who were classified as
high in objective knowledge might not be genuine experts in the product categories of
interest. Instead, they merely fared better in the knowledge quiz than others. In other
words, the homogeneity of the student participants may have restricted the range of their
objective knowledge, which potentially explains the lack of an objective knowledge

80
effect. However, in the real world situations, the majority of consumers are of average
knowledge at best. College students are in fact one of the largest segments at which
digital photography and music products target. Therefore students should be
representative of typical consumers of these products. The absence of an objective
knowledge effect in the student samples is probably indicative of the patterns in the real
world context. To better understand the interplay between subjective and objective
knowledge, further research should consider a direct manipulation of objective
knowledge (such as giving a tutorial; Wood and Lynch 2002) such that subjective and
objective knowledge are manipulated orthogonally.

Future Research
My research argues that consumers infer their capability to differentiate from
their subjective knowledge. When they perceive that they can differentiate, their
behaviors are in line with their beliefs. It is also possible that their self-perception of
knowledge prescribes what they should do. Experts represent an important social
category that fulfills advisory and problem-solving functions (Brauer et al. 2004).
Although expertise is domain-specific, experts across different discipline often display a

81
common set of traits and cognitive skills (Shanteau 1988, 1992). Over time, these
common traits and skills are generalized and closely tied with the category as a whole.
From a role theory perspective, these common characteristics become part of the role
expectations of experts (Biddle 1986). In the context of product evaluations, experts are
expected to differentiate among product alternatives (Alba and Hutchinson 1987). When
a consumer is high in subjective knowledge, typical traits and behavioral expectations of
experts become relevant and applicable to the self. As a result, those who are high in
subjective knowledge believe that they should differentiate, and thus behave in
accordance with the role expectation.
The role fulfillment explanation to the subjective knowledge effect warrants
further research efforts. For future research, it is advisable to identify any existing
expectations about experts pertaining to the way knowledgeable individuals give
evaluations. The manipulation of subjective knowledge should affect the self-application
of these role expectations. For example, those who are high in subjective knowledge
should have a greater expectation to differentiate among product choices. The heightened
expectations should in turn lead to role fulfillment by performing expectation-consistent

82
behaviors. In other words, the role expectations about experts should mediate the effect
of subjective knowledge on product evaluations and differentiation.

Managerial Implications
The fact that subjective knowledge can influence evaluative extremity, coupled
with the possibility of manipulating subjective knowledge regardless of how much one
actually knows, opens up new opportunities for managers. Greater extremity and
differentiation can advance preference formation and purchase decisions among high
subjective knowledge individuals. Past research demonstrates that consumers may defer
choices when they experience trade-off difficulty with similarly attractive alternatives
(Dhar 1997; Dhar and Nowlis 1999). Should enhanced subjective knowledge lead to
more extreme evaluations, the greater differentiation among product choices may reduce
the difficulty in choosing and speed up purchase decisions. For example, salespeople in
retail stores can strategically elevate customers’ subjective knowledge such that they are
apt to differentiate and more likely to commit to the choice that they identify as the best.
This technique can be particularly effective in selling skill-based products, such as golf
clubs and cameras (Burson 2007).

83
Apart from the potential effects on joint evaluation, in which multiple options are
presented simultaneously and evaluated comparatively, subjective knowledge can also
influence decision-making in single option evaluations, in which options are presented in
isolation and evaluated separately (Hsee et al. 1999). Buying a home or a car are typical
examples of single or isolated evaluation, in which case two options are difficult to
physically compare side-by-side. Extremity demonstrated by those who are high in
subjective knowledge can make a favorable option even more favorable, making the
buyer commit early rather than “shopping around”.
High subjective knowledge encourages greater differentiation whereas low
subjective knowledge reduces the propensity to differentiate. As shown in study 2, people
low in subjective knowledge failed to tell the difference between high and low quality
products. They rated the inferior product as favorably as the superior product. Marketers
whose products are on par with or inferior to their competitors’ products may take
advantage of consumers’ deficient differentiation by eliciting a low subjective
knowledge.
Contextual factors, such as the presence of others, have been shown to effectively
influence an individual’s subjective knowledge. Managers who want to gain accurate

84
information from consumers should be aware of the social environment in which the
requested information is provided. For example, in a focus group setting, a highly
knowledgeable respondent can induce an inferior sense of subjective knowledge among
other respondents, causing them to become more moderate than when there is no social
comparison. To avoid the influence of the instantaneous change in subjective knowledge,
managers and market research companies can consider grouping respondents based on
their self-reported knowledge.
In sum, consumer’s subjective knowledge about a product can be changed by
contextual factors and social presence such as salespeople and other consumers. Elevated
or diminished subjective knowledge can have instantaneous effects on the way
consumers evaluate products and communicate their product experience. Hence,
marketers should take the effect of subjective knowledge into consideration when they
make decisions on marketing variables such as the physical retail environments,
salespeople training, or even the composition of target customers.

85
REFERENCES

Ailawadi, Kusum L., Rajiv P. Dant, and Dhruv Grewal (2004), “The Difference between
Perceptual and Objective Performance Measures: An Empirical Analysis,” MSI
Report No. 04 – 103, Marketing Science Institute, Cambridge, MA.

Alba, Joseph W. and J. Wesley Hutchinson (1987), “Dimensions of Consumer
Expertise,” Journal of Consumer Research, 13 (March), 411 – 442.

______ and ______ (2000), “Knowledge Calibration: What Consumers Know and What
They Think They Know,” Journal of Consumer Research, 27 (September), 123 –
156.

Argo, Jennifer J., Darren W. Dahl, and Rajesh V. Manchanda (2005), “The Influence of a
Mere Social Presence in a Retail Context,” Journal of Consumer Research, 32
(September), 207 – 212.

Bandura, Albert (2001), “Social Cognitive Theory: An Agentic Perspective,” Annual
Review of Psychology, 52, 1 – 26.

Bearden, William O., David M. Hardesty, and Randall L. Rose (2001), “Consumer Self-
Confidence: Refinements in Conceptualization and Measurement,” Journal of
Consumer Research, 28 (June), 121 – 134.

Biddle, B. J. (1986), “Recent Developments in Role Theory,” Annual Review of
Sociology, 12, 67 – 92.

86
Brauer, Markus, Patrick Chambres, Paula M. Niedenthal, and Angelique Chatard-
Pannetier (2004), “The Relationship Between Expertise and Evaluative
Extremity: The Moderating Role of Experts’ Task Characteristics,” Journal of
Personality and Social Psychology, 86 (1), 5 – 18.

Brownstein, Aaron L. (2003), “Biased Predecision Processing,” Psychological Bulletin,
129 (4), 545 – 568.

Brucks, Merrie (1985). “The Effects of Product Class Knowledge on Information Search
Behavior,” Journal of Consumer Research, 12 (June), 1 – 16.

Burson, Katherine A. (2007), “Consumer-Product Skill Matching: The Effects of
Difficulty on Relative Self-Assessment and Choice,” Journal of Consumer
Research, 34 (June), 104 – 110.

Carlson, Jay P., Leslie H. Vincent, David M. Hardesty, and William O. Bearden (2009),
“Objective and Subjective Knowledge Relationships: A Quantitative Analysis of
Consumer Research Findings,” Journal of Consumer Research, 35 (October), 864
– 876.

Carpenter, Gregory S., Rashi Glazer, and Kent Nakamoto (1994), “Meaningful Brands
from Meaningless Differentiation: The Dependence on Irrelevant Attributes,”
Journal of Marketing Research, 31 (August), 339 – 350.

Cialdini, Robert B., Alan Levy, C. Peter Herman, and Scott Evenbeck (1973), “Attitude
Politics: The Strategy of Moderation,” Journal of Personality and Social
Psychology, 25 (1), 100 – 108.

87
Crowe, Ellen and E. Tory Higgins (1997), “Regulatory Focus and Strategic Inclinations:
Promotion and Prevention in Decision-Making,” Organizational Behavior and
Human Decision Processes, 69 (2), 117 – 132.

Dhar, Ravi (1997), “Consumer Preference for a No-Choice Option,” Journal of
Consumer Research, 24 (September), 215 – 231.

______ and Stephen M. Nowlis (1999), “The Effect of Time Pressure on Consumer
Choice Deferral,” Journal of Consumer Research, 25(March), 369 – 384.

Diehl, Kristin, Laura J. Kornish, and John G. Lynch (2003), “Smart Agents: When Lower
Search Costs for Quality Information Increase Price Sensitivity,” Journal of
Consumer Research, 30 (June), 56 – 71.

Diehl, Michael and Wolfgang Stroebe (1987), “Productivity Loss in Brainstorming
Groups: Toward the Solution of a Riddle,” Journal of Personality and Social
Psychology, 53 (3), 497 – 509.

Fishbach, Ayelet (2009), “The Function of Value in Self-Regulation,” Journal of
Consumer Psychology, 19, 129 – 133.

Folkes, Valerie S. and David O. Sears (1977), “Does Everybody Like a Liker?” Journal
of Experimental Social Psychology, 13, 505 – 519.

Fox, Craig R. and Martin Weber (2002), “Ambiguity Aversion, Comparative Ignorance,
and Decision Context,” Organization Behavior and Human Decision Processes,
88 (May), 479 – 498.

88
Goldstein, Noah J. and Robert B. Cialdini (2007), “The Spyglass Self: A Model of
Vicarious Self-Perception,” Journal of Personality and Social Psychology, 92 (3),
402 – 417.

Herr, Paul M., Frank Kardes, and John Kim (1991), “Effects of Word-of-Mouth and
Product-Attribute Information on Persuasion: An Accessibility-Diagnosticity
Perspective,” Journal of Consumer Research, 17 (March), 454 – 462.

Hoegg, JoAndrea and Joseph W. Alba (2007), “Taste Perception: More than Meets the
Tongue,” Journal of Consumer Research, 33 (March), 490 – 498.

Hsee, Christopher K., George F. Loewenstein, Sally Blount, and Max H. Bazerman
(1999), “Preference Reversals Between Joint and Separate Evaluations of
Options: A Review and Theoretical Analysis,” Psychological Bulletin, 125 (5),
576 – 590.

Judd, Charles M. and Markus Brauer (1995), “Repetition and Evaluative Extremity,”
Attitude Strength: Antecedents and Consequences, ed. R. E. Petty and J. A.
Krosnick, Hillsdale, NJ: Erlbaum, 43 – 71.

Konrath, S., B. J. Bushman, and K. Campbell (2006), “Attenuating the Link between
Threatened Egotism and Aggression,” Psychological Science, 17, 995 – 1001.

Krosnick, Jon A., David S. Boninger, Yao C. Chuang, Matthew K. Berent, and Catherine
G. Carnot (1993), “Attitude Strength: One Construct or Many Related
Constructs?” Journal of Personality and Social Psychology, 65 (6), 1132 – 1151.

89
Leary, Mark R., Byron D. Barnes, Carol Griebel, Eileen Mason, and Don McCormack,
Jr. (1987), “The Impact of Conjoint Threats to Social- and Self-Esteem on
Evaluation Apprehension,” Social Psychology Quarterly, 50 (4), 304 – 311

Lerner, Jennifer S. and Philip E. Tetlock (1999), “Accounting for the Effects of
Accountability,” Psychological Bulletin, 125 (2), 255 – 275.

Linville, Patricia W. (1982), “The Complexity-Extremity Effect and Age-Based
Stereotyping,” Journal of Personality and Social Psychology, 42 (2), 193 – 211.

Lusk, Cynthia M. and Charles M. Judd (1988), “Political Expertise and the Structural
Mediators of Candidate Evaluations,” Journal of Experimental Social Psychology,
24, 105 – 126.

Maheswaran, Durairaj and Brian Sternthal(1990), “The Effects of Knowledge,
Motivation, and Type of Message on Ad Processing and Product Judgments,”
Journal of Consumer Research, 17 (June), 66 – 73.

______ , Brian Sternthal, and Zeynep Gürhan (1996), “Acquisition and Impact of
Consumer Expertise,” Journal of Consumer Psychology, 5 (2), 115 – 133.

Miller, Dale T. and William Turnbull (1986), “Expectancies and Interpersonal
Processes,” Annual Review of Psychology, 37, 233 – 256.

Mitchell, Andrew A. and Peter A. Dacin (1996), “The Assessment of Alternative
Measures of Consumer Expertise,” Journal of Consumer Research, 23
(December), 219 – 239.

90
Moorman, Christine, Kristin Diehl, David Brinberg, and Blair Kidwell (2004),
“Subjective Knowledge, Search Locations, and Consumer Choice,” Journal of
Consumer Research, 31 (December), 673 – 680.

Park, C. Whan and V. Parker Lessig (1981), “Familiarity and Its Impact on Consumer
Decision Biases and Heuristics,” Journal of Consumer Research, 8 (September),
223 – 231.

______ , David L. Mothersbaugh, and Lawrence Feick (1994), “Consumer Knowledge
Assessment,” Journal of Consumer Research, 21 (June), 71 – 82.

Petty, Richard E., Pablo Brinol, and Zakary L. Tormala (2002), “Thought Confidence as
a Determinant of Persuasion: The Self-Validation Hypothesis,” Journal of
Personality and Social Psychology, 82 (5), 722 – 741.

Radecki, Carmen M. and James Jaccard (1995), “Perceptions of Knowledge, Actual
Knowledge, and Information Search Behavior,” Journal of Experimental Social
Psychology, 31 (March), 107–138.

Raju, P. S., Subhash C. Lonial, and W. Glynn Mangold (1995), “Differential Effects of
Subjective Knowledge, Objective Knowledge, and Usage Experience on Decision
Making: An Exploratory Investigation,” Journal of Consumer Psychology, 4 (2),
153 – 180.

Rao, Akshay R. and Kent B. Monroe (1988), “The Moderating Effect of Prior
Knowledge on Cue Utilization in Product Evaluations,” Journal of Consumer
Research, 15 (September), 253 – 264.

91
See, Kelly E. (2009), “Reactions to Decisions With Uncertain Consequences: Reliance on
Perceived Fairness Versus Predicted Outcomes Depends on Knowledge,” Journal
of Personality and Social Psychology, 96 (1), 104 – 118.

Shafer, Eldar (1993), “Choosing Versus Rejecting: Why Some Options Are Both Better
and Worse Than Others,” Memory and Cognition, 21 (4), 546 – 556.

Shanteau, James (1988), “Psychological Characteristics and Strategies of Expert Decision
Makers,” Acta Psychologica, 68, 203 – 215.

______ (1992), “Competence in Experts: The Role of Task Characteristics,”
Organizational Behavior and Human Decision Processes, 53 (November), 252 –
266.

Shapiro, Stewart and Mark T. Spence (2002), “Factors Affecting Encoding, Retrieval,
and Alignment of Sensory Attributes in a Memory-Based Brand Choice Task,”
Journal of Consumer Research, 28 (March), 603 – 617.

Shen, Hao and Robert S. Wyer (2008), “Procedural Priming and Consumer Judgments:
Effects on the Impact of Positively and Negatively Valenced Information,”
Journal of Consumer Research, 18 (December), 325 – 344.

Simonson, Itamar (1989), “Choice Based on Reasons: The Case of Attraction and
Compromise Effects,” Journal of Consumer Research, 16 (September), 158 –
174.

92
Spence, Mark T. and Merrie Brucks (1997), “The Moderating Effects of Problem
Characteristics on Experts’ and Novices’Judgments,” Journal of Marketing
Research, 34 (May), 233 – 247.

Sujan, Mita (1985), “Consumer Knowledge Effects on Evaluation Strategies Mediating
Consumer Judgments,” Journal of Consumer Research, 12 (June), 31 – 46.

Svenson, Ola (1992), “Differentiation and Consolidation Theory of Human Decision
Making: A Frame of Reference for the Study of Pre- and Postdecision Processes,”
Acta Psychologica, 80, 143 – 168.

Trafimow, David and Janet A. Sniezek (1994), “Perceived Expertise and Its Effect on
Confidence,” Organizational Behavior and Human Decision Processes, 57
(February), 290–302.

Trope, Yaacov and Eva M. Pomerantz (1998), “Resolving Conflicts Among Self-
Evaluative Motives: Positive Experiences as a Resource for Overcoming
Defensiveness,” Motivation and Emotion, 22, 53 – 72.

Wood, Stacy L. and John G. Lynch (2002), “Prior Knowledge and Complacency in New
Product Learning” Journal of Consumer Research, 29 (December), 416 – 426.

Vohs, Kathleen D. and Tood F. Heatherton (2004), “Ego Threat Elicits Different Social
Comparison Processes Among High And Low Self-esteem People: Implications
For Interpersonal Perceptions,” Social Cognition, 22, 168 – 191.

93
Appendix A: Photo Samples Used in Study 1

Enhanced by Product A

Enhanced by Product B (or D)

Enhanced by Product D

94
Appendix B: Study 1 - Transcript from the Computer Study

Digital Photo Software Study
The purpose of this study is to compare experts and nonexperts evaluations of a product.
Later on, you will review and evaluate four digital photo softwares based on the images
enhanced by these products.
This study is divided into two parts:
Part 1 asks about your interest and experience with digtial photography. There is also a
knowledge quiz to assess your expertise on digital photography.
Part 2 requires you to inspect image samples enhanced by four different softwares. You
will provide your comments and evaluations afterwards.

Part 1: Knowledge quiz and interest in digital photography

Digital Photography Quiz
Please answer the following questions about digital photography knowledge and
expertise and be as accurate as possible.
Your expertise score will be calculated as soon as you finish the quiz.
With a 3-megapixel camera, you can take a higher resolution picture than most computer
screens can display.
1. With a 3-megapixel camera, you can take a higher resolution picture than most
computer screens can display.
a. True
b. False

95
2. Which of the following is NOT a Japanese camera manufacturer?
a. Canon
b. Nikon
c. Pentax
d. Leica
3. An over-exposed image will:
a. Be too light
b. Be too dark
c. Be orange colored
d. Have a pronounced blue cast
4. "Red-eye reduction" usually involves multiple flashes in a single exposure.
a. True
b. False
5. Film speed refers to:
a. How long it takes to develop film
b. How fast film moves through film-transport system
c. How sensitive the film is to light
6. DSLR is the abbreviation for:
a. Digital Single Lens Reflex
b. Digital Standard Lens Refraction
c. Digital Simple Light Recorder
7. What is backlighting?
a. The main light is in front of the subject iTunes
b. The main light is behind the subject
c. The main light is on top of the subjects
8. CCD and CMOS are two types of image sensor used in a digital camera.
a. True
b. False
9. On a digital camera, what is Macro Mode?
a. A type of far away caption
b. A type of super close-up
c. A type of big file-size format
10. Low ISO speeds make images more grainy.
a. True
b. False

96
11. A raw image is the unprocessed data directly from the camera's sensor, whose size
therefore is the smallest.
a. True
b. False
12. What type of lens would give a wider depth of field?
a. 35mm
b. 50mm
c. 100mm
13. A polarizing filter may be used outdoors to do what?
a. Increase contrast
b. Minimize reflections
c. Darken blue tones
14. Some digital cameras have sensors that make educated guesses when detecting
colors. This is a process called:
a. Interpolation
b. Guesstimation
c. Extrapolation
15. CompactFlash cards are one type of media that store digital images.
a. True
b. False
16. Pixels on an image sensor can capture ______, and not _______ .
a. Brightness, color
b. Color, brightness
c. Both color and brightness
17. The small opening that allows light to pass through the lens of a digital camera is
called:
a. The aperture
b. The focal point
c. The lens pass
d. The viewfinder
18. The bigger the lens opening, the greater the depth of field.
a. True
b. False

97
How easy or difficult is this Digital Photography Expert Quiz to you?
Very easy 1 2 3 4 5 6 7 Very difficult

How many out of the 18 questions do you think you have answered correctly?
___________

We are now calculating your score.

Your Digital Photography Expertise Score:
Your weighted average score is compared with those of the past 300 participants at USC.
Your digital photography expertise score is in the 88
th
/ 54
th
/ 12
th
percentile, meaning
that 88% / 54% / 12%) of all people scored lower than did you.

Please answer the following questions about yourself.

[Subjective knowledge manipulation checks]
I think of myself as an expert on digital photography.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

I am clueless about digital photography.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

98
Rate how much you feel you know about digital photography compared to your friends.
Much less 1 2 3 4 5 6 7 Much more

Rate how much you feel you know about digital photography compared to average
college students.
Much less 1 2 3 4 5 6 7 Much more

How familiar are you with digital photography?
Very unfamiliar 1 2 3 4 5 6 7 Very familiar

How much experience do you have with digital photography?
Very inexperienced 1 2 3 4 5 6 7 Very
experienced

Part 2: Product Trial
High-end digitial cameras allow users to capture images in uncompressed formats, which
require further processing by a computer.
There are many softwares that enables users to adjust the quality of digtial photos. Some
of them feature a "One-Click" function - by clicking just one button, users can let the
software automatically adjust exposure, white balance, contrast, sharpness, color depth
and other parameters.
Since different softwares use different algorithms in their automatic enhancement
processes, results may therefore differ. We would like you to inspect and evaluate photos
that have been "One-Click" enhanced by four different softwares.

99
We will first present the unprocessed photo directly retrieved from the camera for your
reference. To help you arrive at unbiased judgments, all brand names are omitted. These
products are simply called "A", "B", "C", and "D"
[Unprocessed photo]
You can spend as much time as you like to inspect the photo. Click "Continue" when you
are ready to proceed.
Photos enhanced by softwares "A", "B", "C", and "D" will be shown in sequence.
How do you rate the digital photo software based on the image you just viewed?
(Three 100-point sliders with anchors: Negative vs. Positive; Unfavorable vs. Favorable;
Low quality vs. High quality)

[Measures of confidence]
How certain are you about the quality of the digital photo softwares?
Very uncertain 1 2 3 4 5 6 7 Very certain

How confident are you in your evaluation of the digital photo softwares?
Not at all confident 1 2 3 4 5 6 7 Very
confident

[Measures of effort]
During the photo reviews, I concentrated very hard on the photo samples.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree
I paid a lot of effort when I evaluated the photo softwares.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

100

[Interest in digital music measures]
I have a strong interest in digital photography.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

Digital photography is an important part of my life.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

Demographic Information
Gender: _____
Age: _____
Ethnicity: _____
Grade: _____

Debriefing

The purpose of this experiment was to examine how subjective expertise influences
consumer behaviors such as product evaluation. Since we need to compare consumers
with different levels of expertise, we have to vary people's perceptions of their expertise.
Therefore, we included a deception in this study. We gave false feedback on the quiz - to
either bolster or undermine people's subjective expertise about digital photography.
Earlier in the experiment, you completed a digital photography quiz and received
feedback on your score (in the 12th, 54th or 88th percentile). The feedback, which was
randomly assigned to you, did not reflect your true performance.

101
Your actual score was ________ correct answers out of 18.
If you require further information, wish to withdraw your data or if you have any
problems concerning this project, you can contact the experimenter, Andy Wong, at
kacwong@usc.edu.

End of the study.

102
Appendix C: Study 2 - Transcript from the Computer Study

Digital Music Study
The purpose of this study is to compare experts and nonexperts evaluations of a product
(music coder software). You will listen to a music sample encoded by a new software.
This music encoder converts audio CD tracks into digital files (analogous to MP3).
This study is divided into two parts:
Part 1 asks about your interest and experience with digital music. There is also a
knowledge quiz to assess your expertise on digital music.
Part 2 requires you to listen to a music segment encoded by the software. You will
provide your comments and evaluations afterwards.

Part 1: Digital Music Expert Quiz
The term “digital music” refers to music in a digital format (such as MP3, WMA, WAV).
We would like to know how expert you are about digital music. Please answer the
following questions about digital music knowledge and expertise and be as accurate as
possible. Your expertise score will be calculated as soon as you finish the quiz.

1. Encoding a song at higher bit rates (e.g., 192 kbit/s) results in higher sound
quality than at lower bit rates (e.g., 128 kbit/s)
a. True
b. False
2. The portable media player, Zune, is manufactured by:
a. Apple
b. Creative
c. Microsoft
d. Samsung

103
3. You can buy legal music downloads from:
a. Amazon.com
b. iTunes
c. Rhapsody
d. All of the above
4. Generally, the smaller the file size of an MP3, the better the sound quality.
a. True
b. False
5. What kind of frequency range does a subwoofer reproduce?
a. Treble
b. Mid-tone
c. Bass
6. The current models of Apple iPod are compatible with:
a. Mac OS only
b. Microsoft Windows only
c. Both Mac OS and Windows
7. As of August 2008, who is the number one music retailer in the US?
a. Best Buy
b. iTunes
c. Wal-Mart
8. Because an album art is a graphic file, you cannot tag it to an audio file such as an
MP3.
a. True
b. False
9. What does the “shuffle” function in a digital music player do?
a. Conserve energy
b. Reduce distortion
c. Randomize songs
10. “In-ear headphones” are earbuds that are inserted directly into the ear canal.
a. True
b. False
11. In general, Constant Bit Rate (CBR) encoding produces better audio quality than
Variable Bit Rate (VBR) encoding.
a. True
b. False

104
12. When a compressed audio file sounds perceptually the same as the uncompressed
source, the compression is generally referred to as:
a. Transparent
b. Equivalent
c. Psychoacoustic
13. What is the main benefit of a noise-cancelling headphone?
a. Reduce environmental noise
b. Reduce noise in the recording
c. Reduce white noise (static)
14. To convert music on CDs to digital files, the most commonly-used sampling
frequency is:
a. 44.1 kHz
b. 48.2 kHz
c. 96 kHz
15. Playing songs encoded at high bit rates consumes more battery power of a digital
music player.
a. True
b. False
16. The term “DRM” stands for:
a. Digital copyRight Measure
b. Direct Restrictions Mandate
c. Digital Rights Management
17. The number of devices on which you can play your purchased music from iTunes
is:
a. 1 only
b. Up to 5
c. Up to 10
d. Unlimited
18. Which of the following is NOT a lossy compression format?
a. AAC
b. MP3
c. Vorbis
d. FLAC

105
19. The full name of "MP3" is:
a. MPEG-3
b. MPEG-1 Audio Layer 3
c. MusicPlay-3
d. MPEG ISO-3
20. The more often an MP3 is played, the faster the quality degrades (i.e., get worse).
a. True
b. False

How easy or difficult is this Digital Music Expert Quiz to you?
Very easy 1 2 3 4 5 6 7 Very difficult

How many out of the 20 questions do you think you have answered correctly?
___________

(After an intended short delay)

Your Digital Music Expertise Score:
Your weighted average score is compared with those of the past 300 participants at USC.
Your digital music expertise score is in the 88th (12th) percentile, meaning that 88%
(12%) of all people scored lower than did you.

Please answer the following questions about yourself.
I think of myself as an expert on digital music.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

106
I am clueless about digital music.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

Rate how much you feel you know about digital music compared to your friends.
Much less 1 2 3 4 5 6 7 Much more

Rate how much you feel you know about digital music compared to average college
students.
Much less 1 2 3 4 5 6 7 Much more

I have a strong interest in digital music.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

Digital music is an important part of my life.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

How familiar are you with digital music?
Very unfamiliar 1 2 3 4 5 6 7 Very familiar

How much experience do you have with digital music compared to your friends?
Very inexperienced 1 2 3 4 5 6 7 Very
experienced

107
Part 2: Product Trial
Information from the software developer:
“This new audio encoder, VAC, features a different audio compression format. It is
compatible with most of the digital music players (such as Window Media Player and
iTunes) and portable media players (such as iPod and most music cellphones).”
You will listen to a short music sample encoded by this audio software (VAC).
Afterwards you will be asked to evaluate what you have heard. Because sound quality is
affected by the quality of testing equipments (e.g., PC soundcard, headphones, etc), we
provide you a standard MP3 sample to serve as a point of reference.
You will listen to two segments of music in sequence: The first segment is an MP3
encoded by a commonly-used software. The second segment is the same piece of music
encoded by VAC.

(Participants will first listen to the reference and then the target sample)

In the box below, please provide your review of the software, and/or any thoughts that
came to your mind when you were listening to the samples.

How do you rate the digital encoder software (VAC) based on the sample to which you
just listened?
(Three 100-point sliders with anchors: Negative vs. Positive; Unfavorable vs. Favorable;
Low quality vs. High quality)

108
Was the second audio sample (VAC) you just heard better, worse than or the same as the
first sample?
a. VAC is better
b. Both samples are almost the same
c. VAC is worse

Compared to the MP3 format, the sound quality of the new VAC format is:
Definitely inferior 1 2 3 4 5 6 7 Definitely
superior

Between the MP3 and the new VAC, which one do you prefer?
Definitely MP3 1 2 3 4 5 6 7 Definitely
VAC

How certain are you about the quality of the digital encoder software?
Very uncertain 1 2 3 4 5 6 7 Very certain

How confident are you in your evaluation of the digital encoder software?
Not at all confident 1 2 3 4 5 6 7 Very
confident
If you were to buy this software, what would be the maximum price that you are willing
to pay? $________

During the listening, I concentrated very hard on the music samples.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

109

I paid a lot of effort when I evaluated the music samples.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

How much do you like the song (Brahms Symphony No. 1) to which you just listened?
Not at all 1 2 3 4 5 6 7 Very much
How much do you like classical music in general?
Not at all 1 2 3 4 5 6 7 Very much

Overall, how you rate this product trial experience?
(Three 100-point sliders with anchors: Unfavorable vs. Favorable; Negative vs. Positive;
Not at all enjoyable vs. Very enjoyable)

Demographic Information
Gender: _____
Age: _____
Ethnicity: _____
Grade: _____

110
Debriefing

The purpose of this experiment was to examine how subjective expertise influences
consumer behaviors such as product evaluation. Since we need to compare consumers
with different levels of expertise, we have to vary people's perceptions of their expertise.
Therefore, we included a deception in this study. We gave false feedback on the digital
music quiz - to either bolster or undermine people's subjective expertise about digital
music.
Earlier in the experiment, you completed a digital music quiz and received feedback on
your score (either in the 12th or 88th percentile). The feedback, which was randomly
assigned to you, did not reflect your true performance.
Your actual score was ________ correct answers out of 20.
If you require further information, wish to withdraw your data or if you have any
problems concerning this project, you can contact the experimenter, Andy Wong, at
kacwong@usc.edu.

End of the study.

111
Appendix D: Study 3 - Transcript from the Computer Study

Digital Music Study
The purpose of this study is to compare experts and nonexperts evaluations of a product
(music coder software). You will listen to a music sample encoded by a new software.
This music encoder converts audio CD tracks into digital files (analogous to MP3).
This study is divided into two parts:
Part 1 asks about your interest and experience with digital music. There is also a
knowledge quiz to assess your expertise on digital music.
Part 2 requires you to listen to a music segment encoded by the software. You will
provide your comments and evaluations afterwards.

Part 1: Digital Music Expert Quiz
The term “digital music” refers to music in a digital format (such as MP3, WMA, WAV).
We would like to know how expert you are about digital music. Please answer the
following questions about digital music knowledge and expertise and be as accurate as
possible. Your expertise score will be calculated as soon as you finish the quiz.
1. Encoding a song at higher bit rates (e.g., 192 kbit/s) results in higher sound
quality than at lower bit rates (e.g., 128 kbit/s)
a. True
b. False
2. The portable media player, Zune, is manufactured by:
a. Apple
b. Creative
c. Microsoft
d. Samsung

112
3. You can buy legal music downloads from:
a. Amazon.com
b. iTunes
c. Rhapsody
d. All of the above
4. Generally, the smaller the file size of an MP3, the better the sound quality.
a. True
b. False
5. What kind of frequency range does a subwoofer reproduce?
a. Treble
b. Mid-tone
c. Bass
6. The current models of Apple iPod are compatible with:
a. Mac OS only
b. Microsoft Windows only
c. Both Mac OS and Windows
7. As of August 2008, who is the number one music retailer in the US?
a. Best Buy
b. iTunes
c. Wal-Mart
8. Because an album art is a graphic file, you cannot tag it to an audio file such as an
MP3.
a. True
b. False
9. What does the “shuffle” function in a digital music player do?
a. Conserve energy
b. Reduce distortion
c. Randomize songs
10. “In-ear headphones” are earbuds that are inserted directly into the ear canal.
a. True
b. False
11. In general, Constant Bit Rate (CBR) encoding produces better audio quality than
Variable Bit Rate (VBR) encoding.
a. True
b. False

113
12. When a compressed audio file sounds perceptually the same as the uncompressed
source, the compression is generally referred to as:
a. Transparent
b. Equivalent
c. Psychoacoustic
13. What is the main benefit of a noise-cancelling headphone?
a. Reduce environmental noise
b. Reduce noise in the recording
c. Reduce white noise (static)
14. To convert music on CDs to digital files, the most commonly-used sampling
frequency is:
a. 44.1 kHz
b. 48.2 kHz
c. 96 kHz
15. Playing songs encoded at high bit rates consumes more battery power of a digital
music player.
a. True
b. False
16. The term “DRM” stands for:
a. Digital copyRight Measure
b. Direct Restrictions Mandate
c. Digital Rights Management
17. The number of devices on which you can play your purchased music from iTunes
is:
a. 1 only
b. Up to 5
c. Up to 10
d. Unlimited
18. Which of the following is NOT a lossy compression format?
a. AAC
b. MP3
c. Vorbis
d. FLAC

114
19. The full name of "MP3" is:
a. MPEG-3
b. MPEG-1 Audio Layer 3
c. MusicPlay-3
d. MPEG ISO-3
20. The more often an MP3 is played, the faster the quality degrades (i.e., get worse).
a. True
b. False

How easy or difficult is this Digital Music Expert Quiz to you?
Very easy 1 2 3 4 5 6 7 Very difficult

How many out of the 20 questions do you think you have answered correctly?
___________

(After an intended short delay)

Your Digital Music Expertise Score:
Your weighted average score is compared with those of the past 300 participants at USC.
Your digital music expertise score is in the 88th (12th) percentile, meaning that 88%
(12%) of all people scored lower than did you.

Please answer the following questions about yourself.
I think of myself as an expert on digital music.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

115
I am clueless about digital music.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

Rate how much you feel you know about digital music compared to your friends.
Much less 1 2 3 4 5 6 7 Much more

Rate how much you feel you know about digital music compared to average college
students.
Much less 1 2 3 4 5 6 7 Much more

I have a strong interest in digital music.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

Digital music is an important part of my life.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

How familiar are you with digital music?
Very unfamiliar 1 2 3 4 5 6 7 Very familiar

How much experience do you have with digital music compared to your friends?
Very inexperienced 1 2 3 4 5 6 7 Very
experienced

116
Part 2: Product Trial
Information from the software developer:
“This new audio encoder, VAC, features a different audio compression format. It is
compatible with most of the digital music players (such as Window Media Player and
iTunes) and portable media players (such as iPod and most music cellphones).”
You will listen to a short music sample encoded by this audio software (VAC).
Afterwards you will be asked to evaluate what you have heard. Because sound quality is
affected by the quality of testing equipments (e.g., PC soundcard, headphones, etc), we
provide you a standard MP3 sample to serve as a point of reference.
You will listen to two segments of music in sequence: The first segment is an MP3
encoded by a commonly-used software. The second segment is the same piece of music
encoded by VAC.

[Manipulation of motivation to process]
The software company wants to know whether VAC is better than, the same as, or worse
than the MP3 format. To encourage your active participation, the company will offer you
an extra $5 if you evaluate the product accurately. More details can be found at the end
of the study.

(Participants will first listen to the reference and then the target sample)

In the box below, please provide your review of the software, and/or any thoughts that
came to your mind when you were listening to the samples.

117
How do you rate the digital encoder software (VAC) based on the sample to which you
just listened?
(Three 100-point sliders with anchors: Negative vs. Positive; Unfavorable vs. Favorable;
Low quality vs. High quality)

Was the second audio sample (VAC) you just heard better, worse than or the same as the
first sample?
d. VAC is better
e. Both samples are almost the same
f. VAC is worse

Compared to the MP3 format, the sound quality of the new VAC format is:
Definitely inferior 1 2 3 4 5 6 7 Definitely
superior

Between the MP3 and the new VAC, which one do you prefer?
Definitely MP3 1 2 3 4 5 6 7 Definitely
VAC

How certain are you about the quality of the digital encoder software?
Very uncertain 1 2 3 4 5 6 7 Very certain

How confident are you in your evaluation of the digital encoder software?
Not at all confident 1 2 3 4 5 6 7 Very
confident

118
If you were to buy this software, what would be the maximum price that you are willing
to pay? $________

During the listening, I concentrated very hard on the music samples.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

I paid a lot of effort when I evaluated the music samples.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

How much do you like the song (Brahms Symphony No. 1) to which you just listened?
Not at all 1 2 3 4 5 6 7 Very much
How much do you like classical music in general?
Not at all 1 2 3 4 5 6 7 Very much

Overall, how you rate this product trial experience?
(Three 100-point sliders with anchors: Unfavorable vs. Favorable; Negative vs. Positive;
Not at all enjoyable vs. Very enjoyable)
Please indicate how you are feeling right now:
Happy Not at all 1 2 3 4 5 6 7 Very much
Excited Not at all 1 2 3 4 5 6 7 Very much
Ashamed Not at all 1 2 3 4 5 6 7 Very much
Pleased with Not at all 1 2 3 4 5 6 7 Very much
myself

119
Distressed Not at all 1 2 3 4 5 6 7 Very much
Sad Not at all 1 2 3 4 5 6 7 Very much
Proud Not at all 1 2 3 4 5 6 7 Very much
Humiliated Not at all 1 2 3 4 5 6 7 Very much
Incompetent Not at all 1 2 3 4 5 6 7 Very much
Calm Not at all 1 2 3 4 5 6 7 Very much

Demographic Information
Gender: _____
Age: _____
Ethnicity: _____
Grade: _____

Debriefing

The purpose of this experiment was to examine how subjective expertise influences
consumer behaviors such as product evaluation. Since we need to compare consumers
with different levels of expertise, we have to vary people's perceptions of their expertise.
Therefore, we included a deception in this study. We gave false feedback on the digital
music quiz - to either bolster or undermine people's subjective expertise about digital
music.
Earlier in the experiment, you completed a digital music quiz and received feedback on
your score (either in the 12th or 88th percentile). The feedback, which was randomly
assigned to you, did not reflect your true performance.
Your actual score was ________ correct answers out of 20.

120
If you require further information, wish to withdraw your data or if you have any
problems concerning this project, you can contact the experimenter, Andy Wong, at
kacwong@usc.edu.

End of the study.

121
Appendix E: Study 4 - Transcript from the computer study

Digital Photo Software Study
The purpose of this study is to examine how consumers of different levels of knowledge
evaluate products. Later on, you will review and evaluate four digital photo softwares
based on the images enhanced by these products.
[Manipulation of subjective knowledge]
(In the high SK condition)
We will also be administering this study to a group of high school students who have no
experience in digital photography.
(In the low SK condition)
We will also be administering this study to a group of professional photographers and
graphic designers who work with digital images.

Part 1: Please tell us about your experience and interest in digital photography.
[Subjective knowledge manipulation checks]
I think of myself as an expert on digital photography.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

I am clueless about digital photography.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

How knowledgeable do you feel about digital photography?

122
Not at all 1 2 3 4 5 6 7 Extremely
knowledgeable knowledgeable

How familiar are you with digital photography?
Very unfamiliar 1 2 3 4 5 6 7 Very familiar

How much experience do you have with digital photography?
Very inexperienced 1 2 3 4 5 6 7 Very
experienced

Part 2: Product Trial
High-end digitial cameras allow users to capture images in uncompressed formats, which
require further processing by a computer.
There are many softwares that enables users to adjust the quality of digtial photos. Some
of them feature a "One-Click" function - by clicking just one button, users can let the
software automatically adjust exposure, white balance, contrast, sharpness, color depth
and other parameters.
Since different softwares use different algorithms in their automatic enhancement
processes, results may therefore differ. We would like you to inspect and evaluate photos
that have been "One-Click" enhanced by four different softwares.

[Manipulation of public accountability]
(In the high public accountability condition)
Evaluating the quality of photographic products requires knowledge and critical ability.
There are objectively right and wrong answers. Your product reviews and evaluation
ratings will be analyzed to see how accurate you are.

123
(In the low public accountability condition)
Evaluating the quality of photographic products is a matter of personal preference and
taste. There is no objectively right or wrong answer. Your product reviews and evaluation
ratings will not be subject to any judgment or comparison. Feel free to express what you
like or dislike.
We will first present the unprocessed photo directly retrieved from the camera for your
reference. To help you arrive at unbiased judgments, all brand names are omitted. These
products are simply called "A", "B", "C", and "D"
[Unprocessed photo]
You can spend as much time as you like to inspect the photo. Click "Continue" when you
are ready to proceed.
Photos enhanced by softwares "A", "B", "C", and "D" will be shown in sequence.
How do you rate the digital photo software based on the image you just viewed?
(Three 100-point sliders with anchors: Negative vs. Positive; Unfavorable vs. Favorable;
Low quality vs. High quality)

[Measures of confidence]
How certain are you about the quality of the digital photo softwares?
Very uncertain 1 2 3 4 5 6 7 Very certain

How confident are you in your evaluation of the digital photo softwares?
Not at all confident 1 2 3 4 5 6 7 Very
confident

124
[Measures of public accountability concerns]

How concerned are you with performing well on this product evaluation task?
Not at all concerned 1 2 3 4 5 6 7 Very
concerned
Did you feel anxious about evaluating the photo software products?
Not anxious at all 1 2 3 4 5 6 7 Very anxious

Did you feel any pressure in evaluating the photo software products?
No pressureat all 1 2 3 4 5 6 7 A lot of
pressure

[Measures of effort]
During the photo reviews, I concentrated very hard on the photo samples.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

I paid a lot of effort when I evaluated the photo softwares.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

[Interest in digital music measures]
I have a strong interest in digital photography.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

125
Digital photography is an important part of my life.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

[Quality check #1]
Please choose 2 (the second circle from the left) to show that you have read the question
carefully.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

Digital Photography Quiz
Finally, we would like to know how much you know about digital photography. Please
answer the following questions be as accurate as possible.

With a 3-megapixel camera, you can take a higher resolution picture than most computer
screens can display.
1. With a 3-megapixel camera, you can take a higher resolution picture than most
computer screens can display.
a. True
b. False

2. Which of the following is NOT a Japanese camera manufacturer?
a. Canon
b. Nikon
c. Pentax
d. Leica
3. An over-exposed image will:
a. Be too light
b. Be too dark
c. Be orange colored
d. Have a pronounced blue cast

126
4. "Red-eye reduction" usually involves multiple flashes in a single exposure.
a. True
b. False
5. Film speed refers to:
a. How long it takes to develop film
b. How fast film moves through film-transport system
c. How sensitive the film is to light

6. DSLR is the abbreviation for:
a. Digital Single Lens Reflex
b. Digital Standard Lens Refraction
c. Digital Simple Light Recorder
7. What is backlighting?
a. The main light is in front of the subject iTunes
b. The main light is behind the subject
c. The main light is on top of the subjects
8. CCD and CMOS are two types of image sensor used in a digital camera.
a. True
b. False
9. On a digital camera, what is Macro Mode?
a. A type of far away caption
b. A type of super close-up
c. A type of big file-size format
10. Low ISO speeds make images more grainy.
a. True
b. False
11. A raw image is the unprocessed data directly from the camera's sensor, whose size
therefore is the smallest.
a. True
b. False
12. What type of lens would give a wider depth of field?
a. 35mm
b. 50mm
c. 100mm

127

13. A polarizing filter may be used outdoors to do what?
a. Increase contrast
b. Minimize reflections
c. Darken blue tones
14. Some digital cameras have sensors that make educated guesses when detecting
colors. This is a process called:
a. Interpolation
b. Guesstimation
c. Extrapolation
15. CompactFlash cards are one type of media that store digital images.
a. True
b. False
16. Pixels on an image sensor can capture ______, and not _______ .
a. Brightness, color
b. Color, brightness
c. Both color and brightness
17. The small opening that allows light to pass through the lens of a digital camera is
called:
a. The aperture
b. The focal point
c. The lens pass
d. The viewfinder
18. The bigger the lens opening, the greater the depth of field.
a. True
b. False

How easy or difficult is this Digital Photography Expert Quiz to you?
Very easy 1 2 3 4 5 6 7 Very difficult

How many out of the 18 questions do you think you have answered correctly?
___________

128

129
Demographic Information
Gender: _____
Age: _____
Ethnicity: _____
Grade: _____

Debriefing

The purpose of this experiment was to examine how subjective product knowledge
influences consumer behaviors such as product evaluation. Earlier, you were told that the
same study would also be given to another group of participants - either a group of high
school students or professional photographers. This referent group serves as an implicit
comparison target such that they either bolster or undermine people's subjective
knowledge about digital photography. However, the referent group does not exist and
only the undergraduate students of USC participate in this study.
Your product evaluations will not be judged against any accuracy criteria. There is
actually no right or wrong answer.
If you require further information, wish to withdraw your data or if you have any
problems concerning this project, you can contact the experimenter, Andy Wong, at
kacwong@usc.edu

[End of the study]

130
Appendix F: Study 5 - Transcript from the computer study

Digital Music Study
The purpose of this study is to compare experts and nonexperts evaluations of a product
(music coder software). You will listen to a music sample encoded by a new software.
This music encoder converts audio CD tracks into digital files (analogous to MP3).

[Manipulation of public accountability]
After listening to the music sample, you will be asked to give your comments and
evaluations of the software. At the end of the study, your comments will be sent to other
participants for their reference. You can also read other participants' reviews to know
what they think about the product. We'd appreciate if you can include your first name in
your comments to facilitate the exchange of comments.

This study is divided into two parts:
Part 1 asks about your interest and experience with digital music. There is also a
knowledge quiz to assess your expertise on digital music.
Part 2 requires you to listen to a music segment encoded by the software. You will
provide your comments and evaluations afterwards.

Part 1: Digital Music Expert Quiz
The term “digital music” refers to music in a digital format (such as MP3, WMA, WAV).
We would like to know how expert you are about digital music. Please answer the
following questions about digital music knowledge and expertise and be as accurate as
possible. Your expertise score will be calculated as soon as you finish the quiz.

131
1. Encoding a song at higher bit rates (e.g., 192 kbit/s) results in higher sound
quality than at lower bit rates (e.g., 128 kbit/s)
a. True
b. False
2. The portable media player, Zune, is manufactured by:
a. Apple
b. Creative
c. Microsoft
d. Samsung
3. You can buy legal music downloads from:
a. Amazon.com
b. iTunes
c. Rhapsody
d. All of the above
4. Generally, the smaller the file size of an MP3, the better the sound quality.
a. True
b. False
5. What kind of frequency range does a subwoofer reproduce?
a. Treble
b. Mid-tone
c. Bass
6. The current models of Apple iPod are compatible with:
a. Mac OS only
b. Microsoft Windows only
c. Both Mac OS and Windows
7. As of August 2008, who is the number one music retailer in the US?
a. Best Buy
b. iTunes
c. Wal-Mart
8. Because an album art is a graphic file, you cannot tag it to an audio file such as an
MP3.
a. True
b. False

132
9. What does the “shuffle” function in a digital music player do?
a. Conserve energy
b. Reduce distortion
c. Randomize songs
10. “In-ear headphones” are earbuds that are inserted directly into the ear canal.
a. True
b. False
11. In general, Constant Bit Rate (CBR) encoding produces better audio quality than
Variable Bit Rate (VBR) encoding.
a. True
b. False
12. When a compressed audio file sounds perceptually the same as the uncompressed
source, the compression is generally referred to as:
a. Transparent
b. Equivalent
c. Psychoacoustic
13. What is the main benefit of a noise-cancelling headphone?
a. Reduce environmental noise
b. Reduce noise in the recording
c. Reduce white noise (static)
14. To convert music on CDs to digital files, the most commonly-used sampling
frequency is:
a. 44.1 kHz
b. 48.2 kHz
c. 96 kHz
15. Playing songs encoded at high bit rates consumes more battery power of a digital
music player.
a. True
b. False
16. The term “DRM” stands for:
a. Digital copyRight Measure
b. Direct Restrictions Mandate
c. Digital Rights Management

133
17. The number of devices on which you can play your purchased music from iTunes
is:
a. 1 only
b. Up to 5
c. Up to 10
d. Unlimited
18. Which of the following is NOT a lossy compression format?
a. AAC
b. MP3
c. Vorbis
d. FLAC
19. The full name of "MP3" is:
a. MPEG-3
b. MPEG-1 Audio Layer 3
c. MusicPlay-3
d. MPEG ISO-3
20. The more often an MP3 is played, the faster the quality degrades (i.e., get worse).
a. True
b. False

How easy or difficult is this Digital Music Expert Quiz to you?
Very easy 1 2 3 4 5 6 7 Very difficult

How many out of the 20 questions do you think you have answered correctly?
___________

(After an intended short delay)

134
Your Digital Music Expertise Score:
Your weighted average score is compared with those of the past 300 participants at USC.
Your digital music expertise score is in the 88th (12th) percentile, meaning that 88%
(12%) of all people scored lower than did you.

[Subjective knowledge manipulation checks]
Please answer the following questions about yourself.
I think of myself as an expert on digital music.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

I am clueless about digital music.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

Rate how much you feel you know about digital music compared to your friends.
Much less 1 2 3 4 5 6 7 Much more

Rate how much you feel you know about digital music compared to average college
students.
Much less 1 2 3 4 5 6 7 Much more

[Interest in digital music measures]
I have a strong interest in digital music.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

135

Digital music is an important part of my life.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

[Familiarity and experience measures]
How familiar are you with digital music?
Very unfamiliar 1 2 3 4 5 6 7 Very familiar

How much experience do you have with digital music compared to your friends?
Very inexperienced 1 2 3 4 5 6 7 Very
experienced

Part 2: Product Trial
Information from the software developer:
“This new audio encoder, VAC, features a different audio compression format. It is
compatible with most of the digital music players (such as Window Media Player and
iTunes) and portable media players (such as iPod and most music cellphones).”
You will listen to a short music sample encoded by this audio software (VAC).
Afterwards you will be asked to evaluate what you have heard. Because sound quality is
affected by the quality of testing equipments (e.g., PC soundcard, headphones, etc), we
provide you a standard MP3 sample to serve as a point of reference.
You will listen to two segments of music in sequence: The first segment is an MP3
encoded by a commonly-used software. The second segment is the same piece of music
encoded by VAC.

136
(Participants will first to the reference and then the target sample)

[Written review]
In the box below, please provide your review of the software, and/or any thoughts that
came to your mind when you were listening to the samples.

[Product evaluations]
How do you rate the digital encoder software (VAC) based on the sample to which you
just listened?
(Three 100-point sliders with anchors: Negative vs. Positive; Unfavorable vs. Favorable;
Low quality vs. High quality)
Was the second audio sample (VAC) you just heard better, worse than or the same as the
first sample?
g. VAC is better
h. Both samples are almost the same
i. VAC is worse

Compared to the MP3 format, the sound quality of the new VAC format is:
Definitely inferior 1 2 3 4 5 6 7 Definitely
superior

Between the MP3 and the new VAC, which one do you prefer?
Definitely MP3 1 2 3 4 5 6 7 Definitely
VAC

137

[Confidence in judgment measures]
How certain are you about the quality of the digital encoder software?
Very uncertain 1 2 3 4 5 6 7 Very certain

How confident are you in your evaluation of the digital encoder software?
Not at all confident 1 2 3 4 5 6 7 Very
confident

[Effort measures]
During the listening, I concentrated very hard on the music samples.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

I paid a lot of effort when I evaluated the music samples.
Strongly disagree 1 2 3 4 5 6 7 Strongly agree

[Other covariates]
How much do you like the song (Brahms Symphony No. 1) to which you just listened?
Not at all 1 2 3 4 5 6 7 Very much
How much do you like classical music in general?
Not at all 1 2 3 4 5 6 7 Very much

138

Overall, how you rate this product trial experience?
(Three 100-point sliders with anchors: Unfavorable vs. Favorable; Negative vs. Positive;
Not at all enjoyable vs. Very enjoyable)

Demographic Information
Gender: _____
Age: _____
Ethnicity: _____
Grade: _____

Debriefing
The purpose of this experiment was to examine how subjective expertise influences
consumer behaviors such as product evaluation. Since we need to compare consumers
with different levels of expertise, we have to vary people's perceptions of their expertise.
Therefore, we included a deception in this study. We gave false feedback on the digital
music quiz - to either bolster or undermine people's subjective expertise about digital
music.
Earlier in the experiment, you completed a digital music quiz and received feedback on
your score (either in the 12th or 88th percentile). The feedback, which was randomly
assigned to you, did not reflect your true performance.
Your actual score was ________ correct answers out of 20.
If you require further information, wish to withdraw your data or if you have any
problems concerning this project, you can contact the experimenter, Andy Wong, at
kacwong@usc.edu.
End of the study.

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The motivational power of beauty: how aesthetically appealing products drive purchase effort in consumers

PDF

How sequels seduce: consumers' affective expectations for entertainment experiences

PDF

How product designs and presentations influence consumers’ post-acquisition decisions

PDF

Powerful brand influentials: conceptualization, measurement, and distinctiveness of a brand’s influential consumers

PDF

Why firms should be thankful for thankful customers: incidental gratitude enhances judgments about consumption experiences

PDF

Scale construction and effects of brand authenticity

PDF

Essays on consumer product evaluation and online shopping intermediaries

PDF

Creation and influence of visual verbal communication: antecedents and consequences of photo-text similarity in consumer-generated communication

PDF

The effects of a customer's comparative processing with positive and negative information on product choice

PDF

Savoring future experiences: antecedents and effects on evaluations of consumption experiences

PDF

Novelty versus familiarity: divergent effects of low predictability and low personal influence

PDF

Marketing strategies with superior information on consumer preferences

PDF

Essays on product trade-ins: Implications for consumer demand and retailer behavior

PDF

Manipulating consumer opinion on social media using trolls and influencers

PDF

Essays on the role of entry strategy and quality strategy in market and consumer response

PDF

Essays on revenue management with choice modeling

PDF

Model selection principles and false discovery rate control

PDF

Rapid prototyping and evaluation of dialogue systems for virtual humans

PDF

Shrinkage methods for big and complex data analysis

PDF

Automatic quantification and prediction of human subjective judgments in behavioral signal processing

Asset Metadata

Creator Wong, Kachat Andrew (author)

Core Title Consumers' subjective knowledge influences evaluative extremity and product differentiation

School Marshall School of Business

Degree Doctor of Philosophy

Degree Program Business Administration

Publication Date 08/24/2011

Defense Date 06/30/2010

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag consumer knowledge,evaluative extremity,OAI-PMH Harvest,processing of valenced information,product differentiation,subjective knowledge

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Folkes, Valerie (committee chair), Hollingshead, Andrea (committee member), MacInnis, Deborah (committee member), Park, C. Whan (committee member)

Creator Email kachat@gmail.com,kacwong@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m3410

Unique identifier UC1100160

Identifier etd-Wong-4018 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-385910 (legacy record id),usctheses-m3410 (legacy record id)

Legacy Identifier etd-Wong-4018.pdf

Dmrecord 385910

Document Type Dissertation

Rights Wong, Kachat Andrew

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu