Journal of Economic Psychology 33 (2012) 1193–1203

Contents lists available at SciVerse ScienceDirect

Journal of Economic Psychology

journal homepage: www.elsevier.com/locate/joep

Advertising, quality, and willingness-to-pay: Experimental examination

of signaling theory
Hsiao-Chien Tsui ⇑
Department of Economics, National Chung Cheng University, Chia-Yi 621, Taiwan

a r t i c l e i n f o a b s t r a c t

Article history: According to the quality signaling theory, firms are motivated to invest more advertising in
Received 4 August 2011 high-quality products. This paper conducts an experiment through a closed-ended double-
Received in revised form 24 July 2012 bounded dichotomous choice of the contingent valuation method to measure consumer
Accepted 27 August 2012
willingness-to-pay for a fictitious cell phone market of varying quality before and after
Available online 4 September 2012
advertising. The results show that advertising effectively influences consumer awareness
of perceived quality and enhance their WTP. The results also suggest that even though
JEL classification:
the high- and low-quality products differ in the investment of advertising, the effect of
L15
M21
advertising on the increase in consumer WTP for low-quality products as well as for
M37 high-quality products.
Ó 2012 Elsevier B.V. All rights reserved.
PsycINFO classification:
2229
3920

Keywords:
Advertising effect
Quality
Willingness-to-pay
Contingent valuation method

1. Introduction

Numerous studies have suggested that firms are motivated to influence consumer perceived value by advertising because
of the asymmetric information between seller and buyer. Therefore, advertising can enhance consumer perceived quality and
alter their willingness-to-pay (WTP), creating a positive relationship between perceived quality and WTP (Kirmani & Rao,
2000; Kirmani & Wright, 1989; Zeithaml, 1988).1 Perceived quality is the overall subjective judgment of quality relative to
the expectation of quality. These expectations are based on personal experiences, in addition to various other sources including
brand reputation, price, and advertising (Boulding & Kirmani, 1993; Dodds, Monroe, & Grevval, 1991; Zeithaml, 1988).
Theoretically, researchers indicate that firms with high quality are willing to invest more into advertising for greater re-
turns over time. They emphasize advertising as the signal of firms to convey quality (Fluet & Garella, 2002; Nelson, 1974),

⇑ Tel.: +886 5 2720411x34109; fax: +886 5 2720816.

E-mail address: ecdhct@ccu.edu.tw
1
‘‘Advertising’’ derives from Latin (advertere) and it means to draw others’ attention or influence the people’s intention. Basically, the advertising effect can
be divided into the communication effect and the sales effect. The communication effect means the level at which the advertising is accepted by the consumers
and influences the consumers’ attitude and behavior; sales effect is measured by the increase in the actual sales after the advertising. Based on this definition,
the advertising effect in this paper refers to the former.

0167-4870/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.joep.2012.08.011
1194 H.-C. Tsui / Journal of Economic Psychology 33 (2012) 1193–1203

also known as quality signaling theory.2 In contrast to perceived quality, the quality referred to here is objective quality. Objec-
tive quality is the aggregate performance of all vector product attributes (Kopalle & Lehmann, 1995). For example, the objective
quality of a personal computer includes processing speed, hard disk capacity, reliability, battery, graphics, and WebCam. In this
paper, objective quality and quality are synonymous.3
This paper addresses whether advertising of high-quality and low-quality product influence consumer WTP differently
when it takes the advertising effect of quality signaling into account. In other words, when firms invest in different types
of advertising for high- and low-quality products, does consumer WTP for high- and low-quality products change in the
same or not? Specifically, firms are motivated to invest more advertising in high-quality products. When advertising influ-
ences consumer perception of product quality, the advertising affect should differ according to varying product qualities; the
change in consumer WTP for products after advertising differs according to varying product quality. Thus, besides probing
the influence of advertising on consumer WTP, this paper also considers the different advertising effects of product quality.
Unlike related studies which measure the relationship between advertising and WTP by regression analysis or structural
equation modeling, this paper is based on the closed-ended double-bounded dichotomous choice of the contingent valuation
method (CV method). The CV method is a simple, flexible nonmarket valuation method based on a questionnaire that offers
the respondents an opportunity to make an economic decision on a good, which no market exists. That is, the valuation is
contingent upon the simulated market presented to the respondents. Nowadays the method is widely used to evaluate con-
sumer WTP for public goods such as environmental goods or natural resources. However, some literatures have examined
private goods using this approach. The concept of the CV method, closed-ended double-bounded dichotomous choice espe-
cially, will be explained in the next section.
In order to compare the advertising effect on WTP, this paper creates targeted products in a hypothetical market. The data
is collected through internet questionnaires. Since most online users are familiar with electronic products, this paper treats
camera cell phones as the targeted products. The cell phones in the experiment are the fictitious product not yet available in
the Taiwan market. The cell phone is further divided into two segments of high and low-quality to compare their advertising
effects.4 For time and cost consideration, the experiment deals with print advertising. Except for pricing, information such as
weight, size, volume, basic functions, extra functions, and even the checking system of the targeted products are introduced
in the advertising. To highlight the advertising difference of high and low-quality products, this experiment provides additional
content describing the advertising of high-quality cell phones. The questionnaire for the experiment design and data involves a
pretest and survey. The ages of respondents in the pretest ranged from 18 to 35 years old, with 40 questionnaires. The final sur-
vey was based on an experiment questionnaire, yielding 500 valid questionnaires.
The paper is organized as follows. Section 2 presents the closed-ended double-bounded dichotomous choice of the CV
method. Section 3 explains the questionnaire design, data collection, and primary statistical results. Estimated results are
presented in Section 4. Besides measuring the change in consumer WTP for high and low-quality products before and after
advertising, this paper further investigates the marginal effect of income on WTP and the influence of element on WTP. In the
conclusion, besides generalizing the research finding and contribution, this paper proposes suggestions with regard to the
analytical approach.

2. Research method

This paper gauges the change in the consumers’ WTP before and after advertising using the CV method to construct a
contingent or fictitious market. The questionnaire survey is based on price inquiry to capture respondents’ WTP and after-
ward estimates the subjective price of WTP or willingness-to-accept.5 The CV method includes four types of price inquiry:
open-ended format, sequential bidding format, payment card format, and closed-ended format.6 Hoehn and Randall (1987) sug-
gested that the closed-ended price inquiry is similar to consumption behavior and could thus avoid the strategic bias.
The closed-ended format is also termed a dichotomous choice and is further divided into a single bounded dichotomous
choice and a double-bounded dichotomous choice. Previous studies indicate that statistical efficiency of the parameter esti-

2
Basically, there is positive relationship between advertising and quality. However, some studies suggest there is a reverse ‘‘U’’ relationship between
advertising expense and perception quality. Extremely high advertising expense will reduce perception quality (Kirmani, 1990; Kirmani, 1997). Moreover,
Schmalensee (1978) suggests in the experience goods market with the advertising effect of different periods, the advertising expense of high-quality products is
not necessarily more than low-quality products. Zhao (2000) demonstrates with incomplete information of the product quality, firms which produce high-
quality products are motivated to increase the price and reduce advertising expense. Thus, he suggests the products with high advertising expense are not
necessarily high-quality.
3
Besides advertising, the signals which convey quality also include price, brand and assurance. These allow the consumers to evaluate or have purchase
decision without being familiar with the characteristics of the products. The related literatures are, for example, Azar (2011), Caves and Greene (1996),
Grossman (1979), Milgrom and Roberts (1986), Rao, Qu, and Ruekert (1999), Wolinsky (1983).
4
The targeted product of camera cell phone is currently not on the market of Taiwan. For the concern of the consumers’ prejudice toward the brand and
nationality, the experiment eliminates these two factors in the questionnaire.
5
The CV method is also called ‘‘contingent market evaluation.’’ For the introduction and reorganization of the theoretical base and empirical study related to
the CV method, please find Cummings, Brookshire, and Schulze (1986).
6
‘‘Open-ended format’’ means to directly inquire into the targets’ WTP or price accepted without any hints in advance. The ‘‘gradual auction’’ is similar to
auction. The interviewer suggests a specific price under certain conditions and inquires into the targets’ willingness to accept to continue the adjustment until
the targets accept the prices. A ‘‘payment card’’ means the interviewer sets up different prices according to the primary data and selects the most proper WTP
according to the targets.
 H.-C. Tsui / Journal of Economic Psychology 33 (2012) 1193–1203 1195

Accept WTPi ≥ Ti H

Second price Ti H
Accept

Reject
First price Ti 0
Ti H > WTPi ≥ Ti 0

Accept Ti 0 > WTPi ≥ Ti L

Reject

L
Second price Ti

Reject
Ti L ≥ WTPi

Fig. 1. Price inquiry of CV method of the double-bounded dichotomous choice.

mate of the double-bounded dichotomous choice is more significant than that of the single bounded dichotomous choice.7
The variance and covariance of the parameter estimate of the double-bounded dichotomous choice are also lower than the sin-
gle bounded dichotomous choice (Hanemann, Loomis, & Kanninen, 1991). Thus, this paper analyzes WTP using the closed-
ended double-bounded dichotomous choice. The measurement of WTP also includes the random-utility approach proposed
by Hanemann (1984) and the expenditure function approach suggested by Cameron and James (1987). The estimation results
of the random utility approach and the expenditure function approach are similar. Compared with the random-utility approach,
the analytical process of the expenditure function approach is less complicated. Thus, Cameron (1988) suggested using the logit
model to probe the expenditure function of WTP.8
In the price inquiry process, the double-bounded dichotomous choice successively questions respondents’ WTP for the
targeted product two times in the questionnaire. The price of the second price inquiry depends on the first price (T 0i ). Spe-
cifically, when respondents are willing to pay the first time, the second price (T Hi ) would be higher than the first one
(T Hi > T 0i ), and is typically twice as much as the first one; when respondents are not willing to pay for the first price, the sec-
ond price (T Li ) would be less than the first one (T 0i > T Li ) and the second price is half of the first one. However, when respon-
dents are willing or not willing to pay for two prices, the study could only realize that the respondents’ WTP was more than
T Hi or less than T Li . Thus, the double-bounded dichotomous choice may not result in the upper limit or lower limit of respon-
dents’ WTP.
The price inquiry of the CV method of the double-bounded dichotomous choice is shown in Fig. 1. As I1i and I2i indicate, for
the respondents’ reactions in the first and second times, the result of price inquiry is shown below:
8
>
> ð1; 1Þ; if WTPi P T Hi
>
>
< ð1; 0Þ; if T Hi > WTPi P T 0i
ðI1i ; I2i Þ ¼
> ð0; 1Þ;
> if T 0i > WTPi P T Li
>
>
:
ð0; 0Þ; if T Li P WTPi
where I1i and I2i denote the respondents WTP for the first and second prices. When it is positive, it is shown by the dummy
variable 1; on the contrary, it refers to 0. The upper and lower limits are between ðT Hi ; 1Þ, ðT 0i ; T Hi Þ, ðT Li ; T 0i Þ, and ð0; T Li Þ. v is
the standard error, and the probability of four combinations are as follows:
!
T Hi  WTPi
p11 ¼ ProbðI1i ¼ 1; I2i ¼ 1Þ ¼ 1  F ð1Þ
v

7
‘‘Single bounded dichotomous choice’’ means to directly inquire into the targets for the willingness to purchase the goods. The targets only show their
acceptance or rejection of the random prices. ‘‘Double-bounded dichotomous choice’’ means after the targets accept or reject the first price, the interviewer
inquire after them for the second price. In other words, when the targets accept the first prices, the interviewer will increase the price and inquire after them
again; on the contrary, when the targets reject the price, the interviewer will reduce the amount of money for further inquiry.
8
Based on duality theory, Cameron (1988) analyzed the consumers’ random selection of environmental goods using the minimum expenditure function.
1196 H.-C. Tsui / Journal of Economic Psychology 33 (2012) 1193–1203

! !
T Hi  WTPi T 0i  WTPi
p10 ¼ ProbðI1i ¼ 1; I2i ¼ 0Þ ¼ F F ð2Þ
v v
! !
T 0i  WTPi T Li  WTPi
p01 ¼ ProbðI1i ¼ 0; I2i ¼ 1Þ ¼ F F ð3Þ
v v
!
T Li  WTPi
p00 ¼ ProbðI1i ¼ 0; I2i ¼ 0Þ ¼ F ð4Þ
v
Based on the above, this paper compares WTP using the expenditure function. Thus, consumer maximum WTP of individual
expenditure before and after advertising is shown below:
WTP ¼ eðp; q1 ; v 0 ðp; y0 ; q0 ÞÞ  eðp; q0 ; v 0 ðp; y0 ; q0 ÞÞ
ð5Þ
¼ mðp; q1 ; p; y0 ; q0 Þ  mðp; q0 ; p; y0 ; q0 Þ:
where p denotes prices, q0 and q1 are the changes in purchase intention before and after advertising, v0() is the indirect util-
ity before the change in purchase intention (before advertising), e() is consumer expenditure function, and m() is indirect
compensation function. Eq. (5) suggests that when consumers’ utility of some goods increases because of advertising, the
same price of goods would lead to negative compensation for consumers. In other words, consumers are willing to pay
for the compensation besides the original price to purchase goods. With the censored maximum likelihood, Eq. (5) is rewrit-
ten as:
ln WTPi ¼ f ðxi ; bÞ þ ui : ð6Þ
where WTPi is the real WTP of the i person, xi is the factor vector of the i person, b is the factor vector, ui is the residual, and it
meets the logistic distribution: ui/N(0, m2). Although the WTP of each respondent is uncertain, according to the model, it is
between WTPLi and WTPRi , more than WTPRi , or less than WTPLi .
After substituting Eqs. (1)–(4) to acquire the likelihood function of the logarithm, the maximum likelihood function of
WTPi is obtained:
( " !# " ! !#
X
N
1 2 T Hi  WTPi 1 2 T Hi  WTPi T 0i  WTPi
ln L ¼ Ii Ii ln 1  F þ Ii ð1  Ii Þ ln F F þ I2i ð1  I1i Þ
i¼1
m m m
" ! !# " !#)
T 0i  WTPi T Li  WTPi T Li  WTPi
 ln F F þ ð1  I1i Þð1  I2i Þ ln F : ð7Þ
m m m
where F is the accumulated logistic distribution. The advantage of the maximum likelihood function is that the function can
be applied to any function pattern. The starting value of the maximum likelihood estimate is the key to success of the esti-
mation result. However, it is also the most significant obstacle for estimating the maximum likelihood function, because the
starting value is not based on a fixed measurement or standard and typically requires continuous attempts. Thus, according
to Kanninen (1995) and Alberini (1995), this paper sets the starting value between the 10th and the 90th percentile of the
WTP result in the pretest, and constructs four prices to avoid the extreme value of WTP.

3. Experiment design and data statistics

This paper recognizes the influence of advertising on consumer WTP and further explores if consumer WTP for high and
low-quality products varies with different advertising effects. The influence of advertising on WTP is difficult to measure.
However, this work conducts a primary study on the relationship using statistics that are more precise and carefully man-
ages the experiment and questionnaire design to avoid any subjective impressions.
First, as I have mentioned previously, the targeted product in the experiment is a fictitious product not yet available in the
Taiwan market. To eliminate the influence of brand on WTP, the brand name in the advertising was avoided. Based on the
statement that companies are more motivated to invest in advertising for high-quality products, and to emphasize the dif-
ference in advertising for high and low-quality products in the questionnaire, this study segmented the advertising design
for high and low-quality products in terms of information, color of the layout, and even the page.9 Thus, each questionnaire
included two types of cell phone advertising (low-quality and high-quality), and each type of product involved four starting
point prices, resulting in four types of questionnaires.

9
Generally, the investment in advertising is shown by advertising frequency. Due to the limitation of the questionnaire pattern, the advertising dealing in
this paper cannot totally replace advertising frequency. A high frequency of advertising aims to make the consumers be more impressed with the repetitive
influence on their perception quality. According to the result of pretests, advertising designed in this study did show the effect above. Thus, the author suggests
the general result will not be affected by replacing advertising frequency by advertising content and quality.
 H.-C. Tsui / Journal of Economic Psychology 33 (2012) 1193–1203 1197

Table 1
Willingness-to-pay and starting prices.

Variable Definition Coding

ilb1 Low-quality/before advertising/buy it or not for the first stage price 1 = buy; 0 = not
ilb2 Low-quality/before advertising/buy it or not for the second stage price 1 = buy; 0 = not
ila1 Low-quality/after advertising/buy it or not for the first stage price 1 = buy; 0 = not
ila2 Low-quality/after advertising/buy it or not for the second stage price 1 = buy; 0 = not
ihb1 High-quality/before advertising/buy it or not for the first stage price 1 = buy; 0 = not
ihb2 High-quality/before advertising/buy it or not for the second stage price 1 = buy; 0 = not
iha1 High-quality/after advertising/buy it or not for the first stage price 1 = buy; 0 = not
iha2 High-quality/after advertising/buy it or not for the second stage price 1 = buy; 0 = not
plb($) Low-quality/before advertising/starting price 8000, 9000, 9500 and 10,600
pla($) Low-quality/after advertising/starting price 12,200, 13,000, 14,500 and 15,000
phb($) High-quality/before advertising/starting price 11,400, 13,500, 14,000 and 14,600
pha($) High-quality/after advertising/starting price 15,600, 18,000, 19,840 and 20,600

Second, the questionnaire targets were all online users in Taiwan. According to the result of the pretest questionnaire, this
study modified the starting price and items and formally placed the questionnaire on the website (http://www.my3q.com)
to collect the data, which is a new trend to collect figures by the CV method through an internet questionnaire (Berrens,
Bohara, Smith, Silva, & Weimer, 2004).10 Because of the 68.75% online population in Taiwan and the penetration rate of online
families upwards to 72.82%, it was less controversial to conduct an online questionnaire survey in Taiwan, compared with other
nations.11 All questionnaires were displayed on the website with detailed descriptions to enable respondents to answer the
questions. The approach avoided the problems of misleading interviewer attitudes and effectively reduced interviewer errors
in the CV method. Price inquiry of the double-bounded dichotomous choice tends to involve starting point error; therefore, this
study further tested this aspect.
Finally, the experiment included the pretest and survey and the result of the pretest experiment was regarded as the
starting point price. This paper then modified the questionnaire design and expanded the experiment samples for empirical
experimentation. Each questionnaire in the advertising design included low-quality and high-quality cell phone advertising.
Considering the prices upon four different starting points in the pretest, there were four types of questionnaires. The price
inquiry process and the symbols and definitions of the variables in Fig. 1 were reorganized as in Table 1. Other than inquiring
on respondents’ WTP for high and low-quality cell phones before and after advertising by a closed-ended double-bounded
dichotomous choice, this study also investigated the respondents on their current use of cell phones (their purchase of uni-
cellphone, frequency to change cell phones, monthly expenditure, and so on). The final section of the questionnaire collected
respondents’ personal social and economic background, including their gender, age, educational level, occupation, and
monthly disposable income. The definitions of the variables are shown in Table 2. For the details of the empirical question-
naire, please refer to Appendix A.
Four types of questionnaires were used according to the starting prices. After eliminating invalid samples, there were 500
remaining valid samples. The first group had 123 samples, the second had 120 samples, the third had 135 samples, and the
fourth had 122 samples. Using the questionnaire on the low-quality cell phone before advertising as an example, when start-
ing prices (the price in the first stage, T1) were NTD8000, 9000, 9500 and 10,600 (NTD means New Taiwan Dollar, $ here-
after), the percentages that consumers do not purchase were 47.97%, 50.83%, 32.59%, and 31.97%. When prices in the
second stage (T2) were $7000, 8000, 9000, and 9500, the percentages of consumers’ negative WTP were (percentage among
those with negative WTP in the first stage) 81.36%, 70.49%, 36.36%, and 30.77%; those with positive WTP were 18.64%,
29.51%, 63.64%, and 69.23%, respectively. For the details, please refer to Appendix A Table A2.
For inquiry price and purchase intention above, subjects’ willingness to purchase in the first stage referred to 1 and
unwillingness to purchase was 0; in the second stage, willingness to purchase was 1 and unwillingness to purchase
was 0. Thus, there were four possible combinations for the consumers: (0, 0), (0, 1), (1, 0), and (1, 1). Four consumption
combinations, four types of questionnaires (different prices, high and low-quality cell phones, and before/after advertising)
and the related upper and lower limits of WTP are shown in Table 3.
Besides the starting prices, the questionnaire also included respondents’ consumption characteristics of cell phones, so-
cial, and economic background information. According to the data, most respondents have purchased uni-cellphones (372,
74.4%). Cell phone use frequency was measured based on two variables (average interval in months to changing cell phones
(cha) and average monthly expenditure (exp)). Respondents’ changing frequency of cell phones (88.8%) occurred more than

10
The survey was from February 16, 2007 to March 31, 2007. Basically, on the questionnaire website, one person cannot answer the questions repetitively. At
the time of the publication, the questionnaire was deleted and due to the limited space in the publication, for the questionnaire content in detail, please contact
with the author.
11
The advantages of online questionnaire refer to low cost and rapid return. According to ‘‘survey on Taiwan internet use’’ suggested by TWNIC at the
beginning of 2008, the online population in Taiwan is about 15.80 million, the online users are 68.75%; online families are 5.18 million (penetration rate is
72.82%). Thus, the error caused by the sampling will be reduced to the least. However, since most of the online users are the youth, this research finding can
only demonstrate change of the young people (instead of the public) in WTP before and after advertising.
1198 H.-C. Tsui / Journal of Economic Psychology 33 (2012) 1193–1203

Table 2
Variable definitions and sample mean.

Variable Definition and coding Mean S.D.

uni 1 if ever purchased cell phone; 0 otherwise 0.943 0.163
cha 1 if frequency of cell phone changing 5 6 months; 0 otherwise 0.399 0.462
exp 1 if monthly telephone charge of cell phone 5 $1000; 0 otherwise 0.508 0.539
sex 1 if male; 0 otherwise 0.641 0.350
age Age of respondent as of 2008 28.098 10.117
mar 1 if marriage; 0 otherwise 0.440 0.508
income Monthly income of each respondent  1000 23.117 44.053
imp I if respondent who bears family finance; 0 otherwise 0.325 0.592
edu1 1 if elementary school; 0 otherwise 0.027 0.141
edu2 1 if completed high school; 0 otherwise 0.677 0.281
edu3 1 if =college; 0 otherwise 0.486 0.319
job1 1 if student; 0 otherwise 0.679 0.466
job2 1 if public employee or military man; 0 otherwise 0.172 0.340
job3 1 if industry, commerce and service employee; 0.201 0.47
0 otherwise
p1 Starting price

every 6 months, and 49% of respondents’ monthly cell phone expenditure was more than $1000. Gender has been an impor-
tant variable in market segmentation. Among 500 respondents, 301 were men (60.2%), mostly 20–29 years old (63.2%). For
educational level, most were university students and graduates (47.4%). For marital status (mar), 18.4% of respondents were
married and others (81.6%) were unmarried, cohabitating, separated, divorced, widowers, or widows. For monthly dispos-
able income, 27% of respondents earned less than $10,000 and 22% earned $10,000–20,000. Their income tended to be
low, which might have been because most were students. A total of 14.4% of respondents were the major financial support-
ers of families (imp). Finally, occupation included students (177, over 35% of respondents), military and civil servants (65),
industrial and service industry (160), and others (98), as measured by dummy variables.

4. Results and findings

Table 4 shows the result of consumer WTP for cell phones before and after advertising. Without starting prices (p1), the
average WTP for the low-quality cell phone before advertising was $9905, and after advertising was $13,276. In regard to
the low-quality product, the WTP after advertising is statistically significantly higher than the WTP before advertising at
the confidence level of 95.5%. Advertising increased consumer WTP by $3371 and increased consumers’ average WTP for
low-quality cell phone by 34%. On the other hand, consumers’ average WTP for a high-quality cell phone before advertising
was $13,663 and after advertising was $17,170. It also shows that the WTP after advertising is statistically significantly high-
er than the WTP before advertising for high-quality product. Consumer WTP increased by $3507, with a percentage increase
of approximately 26%. For absolute price, consumer increase in WTP for the high-quality cell phone after advertising was
more than for the low-quality cell phone (3507 > 3371). However, the increased percentage of advertising effect of the
low-quality cell phone was more significant than for the high-quality cell phone (34% > 26%) at the 10% level.12
Table 5 shows the consumer average WTP for starting prices (p1) of the low-quality cell phone before advertising at
$9806 and after advertising at $13,258; the increase in WTP is $3452 and the increased percentage is 35% of the average
WTP before advertising. The average WTP of high-quality products is $13,609 before advertising and $17,412 after advertis-
ing, a percentage increase of 28%. The above result is the same as that without starting prices. Although the increase in con-
sumer WTP for the high-quality cell phone is more significant (3803 > 3452), with regard to percentage increase, the low-
quality cell phone is still more significant (35% > 28%).13 Thus, the advertising effect of the low-quality cell phone is better than
the high-quality one.
This study next probes the influences of other variables on WTP. Without starting prices, in Table 4, high and low-quality
cell phones, before and after advertising, purchasing or without purchasing uni-cellphones (uni), change frequency (cha),
and monthly cell phone expenditure (exp) all reveal significant and positive effects. In other words, consumers who pur-
chased uni-cellphones, change of cell phones within less than six months, or had average monthly cell phone expenditure
more than $1000 have higher WTP for cell phones. In addition, for the consumers’ educational level (edu) before advertising,
the effects of high or low-quality products are significantly negative. This indicates that consumers with a higher educational
level have lower WTP for cell phones. The dummy variable (that is, the student) (job1) also reveals significant and positive
effect in all conditions. Compared with others, students have a higher WTP for cell phones. According to the analytical result

12
At the confidence level of 90%, the confidence interval for the increased percentage of low-quality cell phone (34%) is the range of [0.305, 0.375], high-
quality cell phone (26%) is the range of [0.228, 0.292].
13
At the confidence level of 90%, the confidence interval for the increased percentage of low-quality cell phone (35%) is the range of [0.315, 0.385], high-
quality cell phone (28%) is the range of [0.247, 0.313].
 H.-C. Tsui / Journal of Economic Psychology 33 (2012) 1193–1203 1199

Table 3
Four consumption combinations, four kinds of questionnaire and the related upper and lower limits of WTP.

Variable Definition Coding Questionnaire 1 Questionnaire 2 Questionnaire 3 Questionnaire 4

tlb1 tlb1 = low-quality/before advertising/first (ilb1, ilb2) Lower Upper Lower Upper Lower Upper Lower Upper
stage price (starting price) (0, 0) – 7000 – 8000 – 9000 – 9500
(0, 1) 7000 8000 8000 9000 9000 9500 9500 10,600
(1, 0) 8000 9000 9000 9500 9500 10,600 10,600 16,000
tlb2 tlb2 = second stage price (1, 1) 9000 – 9500 – 10,600 – 16,000 –
tla1 tla1 = low-quality/after advertising/first stage (ila1, ila2) Lower Upper Lower Upper Lower Upper Lower Upper
price (starting price) (0, 0) – 10,000 – 12,200 – 13,000 – 14,000
(0, 1) 10,000 12,200 12,200 13,000 13,000 14,000 14,000 15,000
(1, 0) 12,200 13,000 13,000 14,000 14,000 15,000 15,000 17,000
tla2 tla2 = second stage price (1, 1) 13,000 – 14,000 – 15,000 – 17,000 –
thb1 thb1 = high-quality/before advertising/first (ihb1, ihb2) Lower Upper Lower Upper Lower Upper Lower Upper
stage price (starting price) (0, 0) – 10,000 – 11,400 – 13,500 – 14,000
(0, 1) 10,000 11,400 11,400 13,500 13,500 14,000 14,000 14,600
(1, 0) 11,400 13,500 13,500 14,000 14,000 14,600 14,600 20,000
thb2 thb2 = second stage price (1, 1) 13,500 – 14,000 – 14,600 – 20,000 –
tha1 tha1 = high-quality/after advertising/first (iha1, iha2) Lower Upper Lower Upper Lower Upper Lower Upper
stage price (starting price) (0, 0) – 14,340 – 15,600 – 18,000 – 19,840
(0, 1) 14,340 15,600 15,600 18,000 18,000 19,840 19,840 20,600
(1, 0) 15,600 18,000 18,000 19,840 19,840 20,600 20,600 23,000
tha2 tha2 = second stage price (1, 1) 18,000 – 19,840 – 20,600 – 23,000 –

Table 4
Regression results on consumers’ WTP (without starting prices).

Variable Low-quality High-quality

Before advertising After advertising Before advertising After advertising
Intercept 7.9167⁄⁄⁄ 8.4240⁄⁄⁄ 8.1856⁄⁄⁄ 8.5179⁄⁄⁄
(0.2248) (0.1656) (0.1992) (0.1634)
uni 0.1380⁄⁄⁄ 0.0187⁄⁄⁄ 0.1179⁄⁄⁄ 0.0783⁄⁄⁄
(0.0317) (0.0243) (0.029) (0.0241)
cha 0.0878⁄⁄ 0.0713⁄⁄⁄ 0.0838⁄⁄ 0.0735⁄⁄⁄
(0.0413) (0.0267) (0.034) (0.0259)
exp 0.0566⁄⁄ 0.0448⁄⁄ 0.0458⁄⁄ 0.0661⁄⁄⁄
(0.0270) (0.0187) (0.0230) (0.0188)
sex 0.0496⁄⁄ 0.0498⁄⁄ 0.0668⁄⁄⁄ 0.0266
(0.0250) (0.0179) (0.0215) (0.0178)
age 0.0018 0.0024⁄ 0.0008 0.0012
(0.0021) (0.0014) (0.0018) (0.0015)
edu 0.0155⁄⁄ 0.0043 0.0146⁄⁄⁄ 0.0005
(0.0061) (0.0042) (0.0054) (0.0042)
mar 0.0460 0.04176 0.0546 0.0217
(0.0451) (0.0298) (0.0382) (0.0289)
ln(income) 0.1265⁄⁄⁄ 0.0939⁄⁄⁄ 0.1346⁄⁄⁄ 0.1080⁄⁄⁄
(0.0232) (0.017) (0.0203) (0.0166)
imp 0.0053 0.0666⁄⁄ 0.0757⁄⁄ 0.0480⁄
(0.0448) (0.0294) (0.0374) (0.0279)
job1 0.5229⁄⁄⁄ 0.5211⁄⁄⁄ 0.2986⁄⁄ 0.3138⁄⁄
(0.1692) (0.1213) (0.1481) (0.1240)
job2 0.0047 0.0040 0.0304 0.0290
(0.0445) (0.0303) (0.0372) (0.0292)
job3 0.0137 0.0254 0.0336 0.0109
(0.0343) (0.0240) (0.0295) (0.0237)
edu  job1 0.0289 0.320⁄⁄⁄ 0.0141 0.0190⁄⁄
(0.0108) (0.0079) (0.0096) (0.0080)
Log likelihood 580.1135 701.2913 675.9977 601.4683
WTP 9905 13,276 13,663 17,170
(1693.20) (1721.49) (2190.79) (2341.68)
t = 31.2 > t(0.995, 1) = 2.576 t = 24.6 > t(0.995, 1) = 2.576
Accept 13,276 > 9905 Accept 17,170 > 13,663
⁄⁄⁄ ⁄⁄ ⁄
, and indicate that the variables are significant at the 1%, 5% and 10% levels, respectively. Numbers in Parentheses are estimated standard errors.
1200 H.-C. Tsui / Journal of Economic Psychology 33 (2012) 1193–1203

Table 5
Regression results on consumers’ WTP (with starting prices).

Variable Low-quality High-quality

Before advertising After advertising Before advertising After advertising
Intercept 7.3607⁄⁄⁄ 7.6919⁄⁄⁄ 7.7809⁄⁄⁄ 8.2236⁄⁄⁄
(0.1977) (0.1449) (0.1915) (0.1339)
uni 0.0879⁄⁄⁄ 0.0412⁄⁄ 0.0881⁄⁄ 0.0429⁄⁄
(0.0259) (0.0188) (0.0257) (0.0190)
cha 0.0621⁄ 0.0543⁄⁄⁄ 0.0733⁄⁄ 0.0562⁄⁄⁄
(0.0334) (0.0201) (0.0301) (0.0202)
exp 0.0405⁄⁄ 0.0261⁄ 0.0379⁄ 0.0403⁄⁄⁄
(0.0218) (0.0143) (0.0203) (0.0148)
sex 0.0380⁄ 0.0399⁄⁄⁄ 0.0563⁄⁄⁄ 0.0209
(0.0203) (0.0136) (0.0190) (0.0138)
age 0.0021 0.0024⁄⁄ 0.0006 0.0008
(0.0018) (0.0011) (0.0016) (0.0012)
edu 0.0084⁄ 0.0016 0.0099⁄⁄ 0.0036
(0.0051) (0.0032) (0.0083) (0.0033)
mar 0.0334 0.0246 0.0463 0.0126
(0.0365) (0.0228) (0.0339) (0.0228)
income 0.0789⁄⁄⁄ 0.0557⁄⁄⁄ 0.1037⁄⁄⁄ 0.0718⁄⁄⁄
(0.0189) (0.0130) (0.0182) (0.0133)
imp 0.0083 0.0444⁄⁄ 0.0596⁄ 0.0383⁄
(0.0357) (0.0226) (0.0329) (0.0218)
job1 0.3950⁄⁄⁄ 0.3834⁄⁄⁄ 0.2302⁄ 0.2240⁄⁄
(0.1370) (0.0912) (0.1292) (0.0956)
job2 0.0068 0.0098 0.0259 0.0217
(0.0358) (0.0228) (0.0328) (0.0225)
job3 0.0029 0.0119 0.0263 0.0087
(0.0275) (0.0181) (0.0258) (0.0182)
edu  job1 0.0214⁄⁄ 0.0239⁄⁄⁄ 0.0101 0.0137
(0.0087) (0.0059) (0.0083) (0.0062)
p1 0.0001⁄⁄⁄ 0.0001⁄⁄⁄ 0.0001⁄⁄⁄ 0.0001⁄⁄⁄
(0.0000) (0.0000) (0.0000) (0.0000)
Log likelihood 548.5521 652.1507 661.3138 570.8737
WTP 9806 13,258 13,609 17,412
(1718.75) (1806.32) (2132.90) (2310.15)
t = 30.9 > t(0.995,1) = 2.576 t = 27.0 > t(0.995,1) = 2.576
Accept 13,258 > 9806 Accept 17,412 > 13,609
⁄⁄⁄ ⁄⁄ ⁄
, and indicate that the variables are significant at the 1%, 5% and 10% levels, respectively. Numbers in Parentheses are estimated standard errors.

(Table 5), the significant effect after including starting prices does not differ from that without including starting prices.
Thus, this paper does not elaborate on this point.14
Income has been an important factor for WTP. In the research finding, income reveals a significant and positive effect for
any products before or after advertising. Consumers with higher incomes tend to have a higher WTP. In terms of the marginal
effect of income, the increase in one unit of income leads to an increase in WTP ($72.73) for the low-quality cell phone before
advertising. After advertising, WTP increases by $101.10. Before and after advertising, the marginal WTP of income is $28.37.
When the income increases by one unit, WTP increases for the high-quality cell phone before advertising by $70.75. After
advertising, it increases to $90.42. Thus, after advertising, marginal WTP increases by $19.67. Thus, marginal WTP of income
for the high-quality cell phone is lower than for the low-quality one; consumers’ marginal WTP of income after advertising
for the high-quality cell phone is higher than for the low-quality cell phone, shown in Table 6.
Gender has a positive influence on WTP, indicating that males are willing to pay more for cell phones. However, some
studies suggest that consumption of female consumers tends to be influenced by discounts, promotions, and advertising
(Dittmar & Beattie, 1995). This paper further analyzes the change in WTP according to gender before and after advertising,
and demonstrates that, regardless of the quality of cell phones, the percentage increase in female respondents’ WTP after
advertising is higher than for male respondents (for low-quality products, the percentage increase in females is 36.83%
and 33.79% for males; for high-quality products, the percentage increase in females is 31.21% and 25.96% for males). This
finding shows that although male WTP is higher than females, the percentage increase in WTP for females with the adver-

14
Table 5 shows the starting prices are significant, indicating the error of the starting point in this study. In other words, the subjects’ WTP for higher starting
prices is more significant. However, the starting prices in this paper are significant, they are close to 0 and lower than the factor of WTP by the CV method in
most of the researches. Thus, this paper will not modify this error.
 H.-C. Tsui / Journal of Economic Psychology 33 (2012) 1193–1203 1201

Table 6
The marginal WTP of income (without starting prices).

Advertising WTP ln(WTP) ln(income) Marginal WTP

Low-quality Before 9904.92 9.0201 0.1265 72.7335
After 13276.43 9.4937 0.0939 101.1049
High-quality Before 13663.16 9.5225 0.1346 70.7464
After 17169.82 9.7509 0.1080 90.2862

tising effect is more significant. This is consistent with the finding that females tend to be influenced more by the external
atmosphere, such as advertising.

5. Conclusions

Due to the asymmetric information, advertising has been one of the measures for firms to convey product information to
consumers. According to the theory of quality signaling, firms invest more advertising in high-quality products. If advertising
can change consumer perception of product quality and further influence their WTP, do the advertising effects of high and
low-quality products differ? This paper treats a fictitious market of camera cell phone as the target, and conducts the exper-
iment using a questionnaire survey through the closed-ended double-bounded dichotomous choice of the CV method to
measure and compare consumer WTP for high and low-quality cell phones before and after advertising. This paper first de-
signs a high and low-quality cell phone and provides two different types of advertising. The advertising content of high-qual-
ity products is better than that for low-quality products in product information, pictures, color, and pages. Valid samples
collected by internet total 500.
The result suggests that regardless of product quality, advertising can effectively influence the perception of consumer
product quality and increase their WTP. Advertising influences consumer WTP differently regarding different types of prod-
uct quality; after advertising, the percentage increase in consumer WTP for low-quality products is higher than for high-
quality products. Income is also a critical factor affecting consumer WTP. The finding indicates that consumers with higher
income have higher WTP. However, consumers’ marginal WTP effect of income for high-quality products is lower than for
low-quality products. This paper also reveals significant gender influences on WTP. Before and after advertising, male WTP is
higher than for females for low or high-quality cell phones. However, after advertising, the percentage increase in female
WTP is higher than for males.
The advertising strategy of firm is not only an approach to convey product information, but also an important measure to
promote products or construct a reputation. Does advertising influence consumer WTP? The result of this paper supports
their causal relationship. If firm advertising can increase consumer perception of quality and further enhance consumer
WTP, which advertising effect is better? According to the results, for absolute price, the advertising effect of high-quality
products is superior; however, for the percentage increase in WTP, the advertising effect of low-quality products is better.
However, this paper does not consider advertising costs and only evaluates advertising effect according to the level of
WTP. If more investment in advertising refers to greater costs, the advertising effect of high or low-quality products depends
on the situation.
Based on the above, past studies have tried to find how advertising changes consumer WTP. Differing from the existing
studies, this paper not only carefully analyzes the relationship between advertising and WTP using an experiment question-
naire, but also considers product quality to probe the influences of different qualities on advertising effect. Although this pa-
per targets one fictitious market using an experiment questionnaire and a more precise analytical approach, due to the
limitation of time and funds, the study is still incomplete. The experiment attempts to meet the hypothesis that firms invest
more in high-quality products using advertising. However, the sampling is based on a questionnaire experiment over a short
period and the participants responded to questions immediately after seeing the advertisement, it differs from real market
transactions. A longer period should be used in the future.

Acknowledgements

I am grateful to the Editor, anonymous Referees and Yuji Honjo for their very useful comments and suggestions. I also
thank Chia-Yu Wu for assistance with the database and acknowledge financial support form the National Science Council
of Taiwan (NSC 96-2415-H-194-005). All remaining errors are solely my responsibility.

Appendix A. Experiment pretest

Kanninen (1995) suggested that a starting price of WTP in the single bounded model should be between the 15th and
85th percentile; the starting price of WTP in the double-bounded model should be between the 10th and 90th percentile.
The pretest subjects are 20–35 years old and there are 40 questionnaires in total. Apart from basic information and the
use of cell phones, this pretest changes the questionnaire assumptions into an open-ended inquiry. For instance, after read-
1202 H.-C. Tsui / Journal of Economic Psychology 33 (2012) 1193–1203

Table A1
Price percentile in the pretest.

Percentile (%) Low-quality High-quality

Before advertising After advertising Before advertising After advertising
2 7000 10,000 10,000 14,340
20 8000 12,200 11,400 15,600
42 9000 13,000 13,500 18,000
60 9500 14,000 14,000 19,840
80 10,600 15,000 14,600 20,600
100 16,000 17,000 20,000 23,000

Table A2
Frequency distribution of willingness-to-pay.

Low-quality/before advertising Low-quality/after advertising

T1 Frequency (%) T2 Frequency (%) T1 Frequency (%) T2 Frequency (%)
Questionnaire 1 No 59 (47.97) No 48 (81.36) No 76 (61.79) No 44 (57.89)
n = 123 Yes 11 (18.64) Yes 32 (42.11)
Yes 64 (52.03) No 21 (32.81) Yes 47 (38.21) No 17 (36.17)
Yes 43 (67.19) Yes 30 (63.83)
Questionnaire 2 No 61 (50.83) No 43 (70.49) No 74 (61.67) No 57 (77.03)
n = 120 Yes 18 (29.51) Yes 17 (22.97)
Yes 59 (49.17) No 13 (22.03) Yes 46 (38.33) No 18 (39.13)
Yes 46 (77.97) Yes 28 (60.87)
Questionnaire 3 No 44 (32.59) No 16 (36.36) No 73 (54.07) No 17 (23.29)
n = 135 Yes 28 (63.64) Yes 56 (76.71)
Yes 91 (67.41) No 40 (43.96) Yes 62 (45.93) No 30 (48.39)
Yes 51 (56.04) Yes 32 (51.61)
Questionnaire 4 No 39 (31.97) No 12 (30.77) No 73 (59.84) No 20 (27.40)
n = 122 Yes 27 (69.23) Yes 53 (72.60)
Yes 83 (68.03) No 74 (89.16) Yes 49 (40.16) No 21 (42.86)
Yes 9 (10.84) Yes 28 (57.14)
High-quality/before advertising High-quality/after advertising
Questionnaire 1 No 65 (52.85) No 47 (72.31) No 79 (64.23) No 59 (81.36)
n = 123 Yes 18 (27.69) Yes 20 (18.64)
Yes 58 (47.15) No 18 (31.03) Yes 44 (35.77) No 29 (32.81)
Yes 40 (68.97) Yes 15 (67.19)
Questionnaire 2 No 62 (51.67) No 48 (77.42) No 79 (65.83) No 50 (70.49)
n = 120 Yes 14 (22.58) Yes 29 (29.51)
Yes 58 (48.33) No 22 (37.93) Yes 41 (34.17) No 27 (22.03)
Yes 36 (62.07) Yes 14 (77.97)
Questionnaire 3 No 66 (48.89) No 26 (39.40) No 103 (76.30) No 43 (36.36)
n = 135 Yes 40 (60.60) Yes 60 (63.64)
Yes 69 (51.11) No 12 (17.39) Yes 32 (23.70) No 11 (43.96)
Yes 57 (82.61) Yes 21 (56.04)
Questionnaire 4 No 52 (42.62) No 16 (30.77) No 80 (65.57) No 63 (30.77)
n = 122 Yes 36 (69.23) Yes 17 (69.23)
Yes 70 (57.38) No 60 (85.71) Yes 42 (34.43) No 25 (89.16)
Yes 10 (14.29) Yes 17 (10.84)

ing the picture above (before and after advertising), how much would you be willing to pay for L (H) cell phones of A brand?
After reading the advertising, how much would you be willing to pay for L (H) cell phones of A brand? After deducting the
first and last 5% of the starting price before and after advertising of high and low-quality products, I arrange the order from
low to high levels.
To reduce starting point error, this experiment overlaps the combinations of different prices or uses a lower price com-
bination to increase efficiency of the median estimate of WTP. According to four types of prices (the least price combination
in the paper) suggested by Alberini (1995), this paper arranges the order of price acquired in an open-ended pretest ques-
tionnaire by first eliminating the maximum and minimum (the first and last 5%) and then regarding the 20th, 42nd, 60th, and
80th percentiles as first stage prices. For subjects ‘‘willing to purchase,’’ this paper treats the 42nd, 60th, 80th and 100th per-
centiles as second stage prices. For targets ‘‘not willing to purchase,’’ this study treats the 2nd, 20th, 42nd, and 60th percen-
tiles as second stage prices.
 H.-C. Tsui / Journal of Economic Psychology 33 (2012) 1193–1203 1203

This paper uses four types of questionnaire (price choice). The following is the price choice of the formal questionnaire.
For L cell phones of the A brand, before advertising, the prices of the first stage are $8000, 9000, 9500, and 10,600; the prices
of the second stage of ‘‘willing to purchase’’ increase to $9000, 9500, 10,600, and 16,000; the prices of the second stage of
‘‘not willing to purchase’’ reduce to $7000, 8000, 9000 and 9500; the prices of the first stage of the H product of the A brand,
before advertising are $1400, 13,500, 14,000, and 14,600; the prices of the second stage of ‘‘willing to purchase’’ increase to
$13,500, 14,000, 14,600, and 20,000; the prices of the second stage of ‘‘not willing to purchase’’ reduce to $10,000, 11,400,
13,500, and 14,000; the prices of the first stage of the L product of A brand, after advertising, are $12,200, 13,000, 14,000, and
15,000; the prices of the second stage of ‘‘willing to purchase’’ increase to $13,000, 14,000, 15,000, and 17,000; the prices of
the second stage of ‘‘not willing to purchase’’ reduce to $10,000, 12,200, 13,000, and 14,000; the prices of the first stage of the
H product of A brand after advertising are $15,600, 18,000, 19,840, and 20,600; the prices of the second stage of ‘‘willing to
purchase’’ increase to $14,340, 15,600, 18,000, and 19,840; the prices of the second stage of ‘‘not willing to purchase’’ reduce
to $18,000, 19,840, 20,600, and 23,000. The pretest price percentile for high and low-quality products before and after adver-
tising is shown in Table A1.

References

Alberini, A. (1995). Testing willingness-to-pay models of discrete choice contingent valuation survey data. Land Economics, 71, 83–95.
Azar, O. H. (2011). Do people think about absolute or relative price differences when choosing between substitute goods? Journal of Economic Psychology, 32,
450–457.
Berrens, R. P., Bohara, A. K., Smith, H. C. J., Silva, C. L., & Weimer, D. L. (2004). Information and effort in contingent valuation surveys: Application to global
climate change using national internet samples. Journal of Environmental Economics and Management, 47, 331–363.
Boulding, W., & Kirmani, A. (1993). A consumer-side experimental examination of signaling theory. Journal of Consumer Research, 20, 111–123.
Cameron, T. A. (1988). A new paradigm for valuing non-market goods using referendum data: Maximum likelihood estimation by censored logistic
regression. Journal of Environmental Economics and Management, 15, 355–379.
Cameron, T. A., & James, M. D. (1987). Efficient estimation methods for ‘closed-ended’ contingent valuation surveys. Review of Economics and Statistics, 69,
269–276.
Caves, R. E., & Greene, D. P. (1996). Brands’ quality levels, prices, and advertising outlays: Empirical evidence on signals and information costs. International
Journal of Industrial Organization, 14, 29–52.
Cummings, R. G., Brookshire, D. S., & Schulze, W. S. (1986). Valuing environmental goods: A state of the art assignment of the contingent valuation method.
Totowa, NJ: Rowman and Allenheld.
Dittmar, H., & Beattie, J. (1995). Gender identity and material symbols: Objects and decision considerations in impulse purchases. Journal of Economic
Psychology, 16, 491–511.
Dodds, W. B., Monroe, K. B., & Grevval, D. (1991). Effects of price, brand, and store information on buyers’ product evaluations. Journal of Marketing Research,
28, 307–319.
Fluet, C., & Garella, P. G. (2002). Advertising and prices as signals of quality in a regime of price rivalry. International Journal of Industrial Organization, 20,
907–930.
Grossman, S. J. (1979). The information role of warranties and private disclosure about product quality. Journal of Law and Economics, 24, 461–483.
Hanemann, W. M. (1984). Welfare evaluations in contingent valuation experiments with discrete responses. American Journal of Agricultural Economics, 66,
332–341.
Hanemann, M., Loomis, J., & Kanninen, B. (1991). Statistical efficiency of double-bounded dichotomous choice contingent valuation. American Journal of
Agricultural Economics, 73, 1255–1263.
Hoehn, J. P., & Randall, A. (1987). A satisfactory benefit cost indicator from contingent valuation. Journal of Environmental Economics and Management, 14,
226–247.
Kanninen, B. J. (1995). Bias in discrete response contingent valuation. Journal of Environmental Economics and Management, 28, 114–125.
Kirmani, A. (1990). The effect of perceived advertising costs on brand perceptions. Journal of Consumer Research, 17, 160–171.
Kirmani, A. (1997). Advertising repetition as a signal of quality: If it’s advertised so often, something must be wrong. Journal of Advertising, 26, 77–86.
Kirmani, A., & Rao, A. R. (2000). No pain, no gain: A critical review of the literature on signaling unobservable product quality. Journal of Marketing, 64,
66–79.
Kirmani, A., & Wright, P. (1989). Money talks: Perceived advertising expense and expected product quality. Journal of Consumer Research, 16, 344–353.
Kopalle, P. K., & Lehmann, D. R. (1995). The effects of advertised and observed quality on expectations about new product quality? Journal of Marketing
Research, 32, 280–290.
Milgrom, P., & Roberts, J. (1986). Price and advertising signals of product quality. Journal of Political Economy, 94, 796–821.
Nelson, P. (1974). Advertising as information. Journal of Political Economy, 82, 729–754.
Rao, A. R., Qu, L., & Ruekert, R. W. (1999). Signaling unobservable product quality through a brand ally. Journal of Marketing Research, 36, 258–268.
Schmalensee, R. (1978). A model of advertising and product quality. Journal of Political Economy, 86, 485–503.
Wolinsky, A. (1983). Prices as signals of product quality. Review of Economic Studies, 50, 647–658.
Zhao, H. (2000). Raising awareness and signaling quality to uninformed consumers: A price-advertising model. Marketing Science, 19, 390–396.
Zeithaml, V. A. (1988). Consumer perceptions of price, quality and value: A means-end model and synthesis of evidence. Journal of Marketing, 52, 2–22.