ASSOCIATION FOR CONSUMER RESEARCH
Labovitz School of Business & Economics, University of Minnesota Duluth, 11 E. Superior Street, Suite 210, Duluth, MN 55802
Consumer Choices Based on Signals: the Case of Mobile Phone Services in Vietnam
Tho Nguyen, University of Technology, Sydney and University of Economics, Ho Chi Minh City
Trang Nguyen, University of Technology, Sydney and Vietnam National University, Ho Chi Minh City
Nigel Barrett, School of Marketing, University of Technology, Sydney
Although it is well known that firms can use signals to inform consumers about the unobservable of their products or services in a
market where asymmetric information exists, research on the role of signal characteristics in consumer choices of service brands
is largely ignored. Using a choice experiment with best-worst scaling, the authors found that signal credibility, consistency, and
clarity have positive impacts on consumer choices. Further, the interaction between signal clarity and credibility also affects the
choice of consumers. These findings suggest that service providers should send clear, consistent, and credible signals to assist
consumers in making purchase decisions.
[to cite]:
Tho Nguyen, Trang Nguyen, and Nigel Barrett (2006) ,"Consumer Choices Based on Signals: the Case of Mobile Phone
Services in Vietnam", in AP - Asia-Pacific Advances in Consumer Research Volume 7, eds. Margaret Craig Lees, Teresa
Davis, and Gary Gregory, Sydney, Australia : Association for Consumer Research, Pages: 396-400.
[url]:
http://www.acrwebsite.org/volumes/13021/volumes/ap07/AP-07
[copyright notice]:
This work is copyrighted by The Association for Consumer Research. For permission to copy or use this work in whole or in
part, please contact the Copyright Clearance Center at http://www.copyright.com/.
CONSUMER CHOICES BASED ON SIGNALS: THE CASE OF MOBILE PHONE SERVICES IN VIETNAM
Tho D. Nguyen, University of Technology, Sydney and University of Economics
Trang T. M. Nguyen, University of Technology, Sydney and Vietnam National University,
Nigel J. Barrett, University of Technology, Sydney
ABSTRACT
Although it is well known that firms can use signals
to inform consumers about the unobservable of their
products or services in a market where asymmetric
information exists, research on the role of signal
characteristics in consumer choices of service brands is
largely ignored. Using a choice experiment with best-worst
scaling, the authors found that signal credibility,
consistency, and clarity have positive impacts on consumer
choices. Further, the interaction between signal clarity and
credibility also affects the choice of consumers. These
findings suggest that service providers should send clear,
consistent, and credible signals to assist consumers in
making purchase decisions.
INTRODUCTION
The mobile phone service market has just appeared in
Vietnam recently. In the last few years, there were only two
players in the market–Vinaphone and Mobifone. Recently,
the entry of S-phone and Viettel has made the market more
competitive (Ngo 2005). Accordingly, several sales
promotion tools have been launched by service providers to
attract consumers. However, the quality of mobile phone
services is still in question. Several complaints about
quality have been lodged by consumers because they have
not known about the quality of the service before
consuming it (VietnamNet 2005).
Service providers know more about the quality of
their services than do consumers, that is, consumers are not
fully informed about the quality of the services. This is
characterised by the asymmetry of information (Boulding
and Kirmani 1993). In addition, most services comprise
experience quality, that is, consumers are generally unable
to evaluate the quality of a service prior to consumption
(Zeithaml and Bitner 2000). This requires service providers
to use signals to inform the quality of their services to
consumers, and signaling theory is useful (Kirmani and
Rao 2000).
Signaling theory, which is derived from the
information economics literature under the condition of
asymmetric information (Spence 1973), has been widely
employed by researchers in marketing such as in studies of
brand equity (Erdem and Swait 1998; Erdem, Swait, and
Louviere 2002), warranties, product quality (Boulding and
Kirmani 1993; Rao, Qu, and Ruekert 1999; Soberman
2003), price (Simester 1995; Srivastava and Lurie 2004),
and advertising (Caves and Greene 1996; Kirmani and
Wright 1989). However, little research has been devoted to
explore the usefulness of signaling theory in services,
especially, in transition markets like Vietnam. To bridge
this gap, this study aims to apply signaling theory to
investigate consumer choices of services. Specifically, it
explores the impacts of signal characteristics–clarity,
consistency, and credibility–on the choice of mobile
services by consumers. The article is organised around the
following key points: We first review the literature and
propose the hypotheses. We then present the method and
data analysis. We conclude the article by discussing the
results, implications and directions for future research.
396
LITERATURE REVIEW AND HYPOTHESES
When information asymmetry exists and service
providers know more about their service quality than do
buyers, several marketing signals can serve as signals of
quality of service brands (see Kirmani and Rao 2000 for an
extensive review). Herbig and Milewicz (1996, 35) define
“a marketing signal is a marketing activity which provides
information beyond the activity itself and which reveals
insights into the unobservable.” Marketing signals convey
information about service brands to consumers such as
service attributes, prices, and warranties (Boulding and
Kirmani 1993; Herbig and Milewicz 1994).
Under the condition of asymmetric information, a
service consumer is generally confronted with a problem of
distinguishing between high- and low-quality service
providers. Also, a service provider may face difficulties in
positioning itself against low-quality service providers in
the mind of the consumer (Mishra, Heide, and Cort 1998).
The inability of consumers to assess the quality of the
service provider can create two problems–adverse selection
and moral hazard. The adverse selection problem involves
certain fixed characteristics of the service provider that
have the potential to influence the level of quality
delivered, however, they are unobservable to the consumer.
The moral hazard problem relates to the ability and
motivation of the service provider to cheat the consumer,
such as to change the levels of quality provided for each
transaction (Mishra et al. 1998). Under such conditions, the
service provider can use a number of marketing signals
such as service quality, price, warranties, advertising, and
brand names to show its ability to meet the need of the
consumer and to differentiate it from other service
providers (i.e., to influence the consumer’s choice and to
help the consumer to distinguish it from less-qualified
service providers; Boulding and Kirmani 1993; Rao and
Ruekert 1994). Hence, it is important that the service
provider’s signaling scenarios should not be easily imitated
by less qualified service providers (Koku 1995).
A signal is not part of a product or service itself. It is
a piece of information about the product or service that
assists consumers in making inferences about the quality
and value of the product or service (Herbig and Milewicz
1994). As signaling is a learning process in which
consumers receive the signal, read and interpret it in the
light of experience, and react accordingly (Heil and
Robertson 1991; Herbig and Milewicz 1994), the
characteristics of a signal plays a crucial role in facilitating
this learning process. Three important characteristics of a
signal have been found in the literature: signal clarity;
credibility; and, consistency (Erdem and Swait 1998; Heil
and Robertson 1991). Signal clarity refers to “the absence
of ambiguity in the information conveyed by the brand’s
past and present marketing mix strategies and associated
activities”. Signal consistency is “the degree to which each
marketing mix component or decision reflects the intended
whole”, and signal credibility “underlies consumer
confidence in a firm’s product claims” (Erdem and Swait
1998, 137).
Signal clarity assists consumers in identifying what a
service provider would like to communicate with them,
such as service attributes and position (Heil and Robertson
1991). To make a signal clear, every marketing-mix
element should be consistent (i.e., reflecting the same
attributes, objectives, and position), and stable over time
(Erdem and Swait 1998). A clear signal enables consumers
to quickly understand and interpret it, preventing any
reaction delays. Together with signal clarity, signal
credibility is another key characteristic of the signal
because it determines the effectiveness of information
conveyed (Erdem and Swait 1998; Tirole 1988). Any
effects of the signal will happen only if its credibility has
been perceived (Sternthal, Dholakia, and Leavitt 1978). As
discussed previously, signaling is a learning process and
consumers interpret the signal and adjust their
interpretation based on the history of the sender and
previous transactions (Herbig and Milewicz 1994). To
assess the meaning as well as the credibility of the signal,
consumers may carry out consistency checks. Therefore,
signal consistency will make the interpretation process
faster and more precise (Heil and Robertson 1991). In
addition, signaling theory suggests that most rational firms
are unlikely to send false signals if the signals increase
costs in terms of immediate profits, future profits, and
reputation (Tirole 1988). Accordingly, consumers may
believe that only high quality service providers would send
clear, consistent, and credible signals to their consumers.
As a result, these signal characteristics are vital to the
evaluation of the quality and choice of a service by
consumers. Hence,
H1: Higher signal clarity of a service brand will
increase the probability of consumer choices of that brand.
H2: Higher signal credibility of a service brand will
increase the probability of consumer choices of that brand.
H3: Higher signal consistency of a service brand will
increase the probability of consumer choices of that brand.
The effectiveness of information received by
consumers depends heavily on whether signals sent are
credible or not (Erdem and Swait 1998; Tirole 1988).
Therefore, the credibility of brand signals is perhaps the
most important characteristic. For that reason, it is
important for a service provider to send credible signals to
consumers. In so doing, signals will create greater
consumers’ confidence in the firm’s product or service
claims (Erdem and Swait 1998). However, the confidence
of consumers depends on what they receive and make
inferences overtime, that is, signal credibility is sensitive to
time (Herbig and Milewicz 1993). Clear and consistent
signals facilitate consumers to interpret the content signaled
by the firm, reinforcing consumers’ belief that the firm’s
willingness and ability to offer the promised service
(Erdem and Swait 1998). Consequently, the clarity and
consistency of signals will increase consumers’ confidence
in the claims. In other words, the interaction effects
between signal credibility and clarity, between signal
credibility and consistency, and between signal consistency
and signal clarity are likely to exist. Thus,
H4: The impact of signal credibility on the
probability of consumer choices will be stronger with high
signal clarity than with low signal clarity.
397
H5: The impact of signal credibility on the
probability of consumer choices will be stronger with high
signal consistency than with low signal consistency.
METHODOLOGY
Experimental Design
A choice experiment with a best-worst scaling
method was employed in this study. There were three
attributes, that is, clarity, consistency, and credibility. Each
attribute had two levels: low and high. Three options,
coded as A, B, and C, were used for each choice set.
Consequently, the total profiles for the full factorial design
were 512 (23x3). Several strategies to construct such a
design can be found in the literature (Street, Burgess, and
Louviere 2005). For example, we can utilise a software
package such as SAS to generate an OMEP (Orthorgonal
Main Effects Plan) and then, using a search algorithm to
construct choice sets (Kuhfeld 2004). Another way is to use
a LMA approach (Louviere, Hensher, and Swait 2000). The
first strategy gives us an efficient design, however, we do
not know if a better design is available. The second
approach requires the number of choice sets much larger
than needed to estimate the effects of interest (Street et al.
2005).
In order to minimise the number of choice sets and to
estimate all main and two-way interaction effects, we
utilised a method of optimal designs developed by Street et
al. (2005). The construction of the choice sets in this study
is as follows: For the first option (option A), a full factorial
design was used (23 = 8 profiles). The profiles in option B
were obtained by leaving the levels of the second attribute
unchanged and systematically changing the levels of the
first and third attributes by adding 1 (mod 2) to the
respective attributes of the profiles in option A. For
example, the levels of signal clarity for choice sets 1 and 4
(option B) were obtained by adding 1 to those of in option
A: 0 + 1 = 1 and 1 + 1 = 0 (see column 5, table 1).
Similarly, the profiles in option C were obtained by leaving
the levels of the third attribute unchanged and
systematically changing the levels of the first and second
attributes by adding 1 (mod 2) to the respective attributes
of the profiles in option A. This design is 100% efficient
and is shown table 1 (see Street et al. (2005) for detailed
explanation).
The Sample and Measures
A sample of 279 in-service training students, both
undergraduate and post-graduate, in the University of
Economics, Ho Chi Minh City and Vietnam’s Fulbright
Economics Teaching Program was invited to participate in
the experiment. The sample was comprised of 103 female
and 176 male participants. In terms of age, 217 participants
were less than or equal to 30 years of age and 62 were more
than 30 years of age.
A best-worst scaling method (Finn and Louviere 1992;
Marley and Louviere 2005) was used to obtain the data.
Best-worst scaling has been found to have several
advantages over traditional scaling methods such as
containing less respondent error, more discriminating way
to measure attribute importance than either rating scales or
the method of pair comparisons (Charzan 2005). The
format of the best-worst scaling employed in this study is
shown in table 2. The questionnaire included eight choice
sets. Subjects were presented 8 choice sets. In each choice
set, there were three options: A; B; and, C. For each choice
TABLE 1
CHOICE SET DESIGN
Choice
set
Option (brand) A
Clarity
1
Option (brand) B
Credibility
0
Clarity
0
Consistency
0
2
0
0
3
0
1
4
0
5
1
6
7
8
Option (brand) C
Credibility
1
Clarity
1
Consistency
0
1
Consistency
1
Credibility
0
1
1
0
0
1
1
0
1
1
1
1
1
0
0
1
1
1
0
0
0
1
0
1
0
1
0
1
0
1
0
1
0
1
1
1
0
0
0
0
0
1
1
0
1
1
0
0
0
1
1
1
0
1
0
0
0
1
set, they were requested to choose the best option and the
worst option among these three options, based on their
judgment about the combination of signal characteristic
levels. It is also noted that the subjects were asked to make
their choices based on three latent constructs, that is, signal
clarity, consistency, and credibility. Therefore, prior to
participating in the experiment, the participants were
instructed about the meaning of these constructs. For
example, signal clarity was understood that a service firm
provides clear and sufficient information about its services
for consumers. This information makes consumers free
from trouble in figuring out what the service firm is trying
to communicate with its customers. Signal consistency
means that the service firm provides the information about
its services for consumers is very consistent for several
years and this information matches its overall image. Signal
credibility is understood that the service firm delivers what
it promises to consumers. It does not pretend to be
something it is not, which makes consumers fully
believable.
TABLE 2
THE BEST-WORST MEASURE (CHOICE SET 1)
Signal characteristics
Clarity
Consistency
Credibility
Which option do you prefer best
Which option do you prefer least
Option (brand) A
low
low
low


Data Coding and Entry
The data were coded and inputted using the following
procedure. There were three options for each choice set: A;
B; and, C. Therefore, three pair comparisons were made for
each choice set option, that is, AB, AC, and BC. For
example, if the best option chosen was A and the least
option chosen was C, the data for this respondent for choice
set 1 would be coded as follows: AB: 10; AC:10; and,
BC:10 (1: if a respondent chose the option and 0: if he or
she did not choose the option). As a result, each respondent
had six rows for each choice set in the data matrix. There
were eight choice sets, resulting in 48 rows for each
respondent. With a sample sise of 279, the data matrix had
a total of 13392 (279x48) cases.
DATA ANALYSIS AND RESULTS
Binary logistic regression was utilised to analyse the
data. We also aimed to investigate the effects of two
respondent characteristics, that is, gender (female
respondents, coded as 0, and male respondents, coded as 1)
and age (equal or less than 30 years of age, coded as 0, and
more than 30 years of age, coded as 1). The results indicate
398
Option (brand) B
high
low
high


Option (brand) C
high
high
low


that the model fitted the data well: χ2(7) = 6323.13 (p <
.001); -2Loglikelihood = 12242.13; Cox & Snell R2 = .376;
Nagelkerke R2 = .502; and, overall correct choice predicted
= 78.9%.
The results indicate that all main effects were
significant (p < .001), supporting hypotheses 1, 2, and 3.
Among these characteristics of a brand signal, that is,
clarity, consistency, and credibility, signal credibility had
the greatest impact on consumer choices (βcredibility = 2.819,
compared to βclarity = 1.236 and βconsistency = 1.024). Further,
the interaction effect between signal clarity and credibility
on the brand choice was also significant (βclarity x credibility =
.479, p < .001), supporting hypothesis 4. However, the
interaction effect between signal consistency and
credibility on brand choice was not significant (βconsistency x
credibility = .084, p > .14). Consequently, hypothesis 5 was
not supported (see table 3). In addition, the effects of signal
clarity, consistency, and credibility on consumer choices
were the same for young and older, as well as for male and
female respondents.
TABLE 3
BINARY LOGISTIC REGRESSION RESULTS
Factors
Estimate
Standard error
Wald
Clarity
1.236
.067
341.393
Consistency
1.024
.065
244.712
Credibility
2.819
.085
1108.966
Consistency x Credibility
.084
.094
0.806
Clarity x Credibility
.479
.099
23.452
Gender
.000
.047
.000
Age
.000
.055
.000
Constant
-2.628
.083
1005.885
Statistics: -2Loglikelihood = 12242.13; Cox & Snell R2 = .376; Nagelkerke R2 = .502
DISCUSSION AND CONCLUSIONS
The purpose of this study is to explore the role of
marketing signals in the choice of mobile phone services by
consumers. Based on the assumption that consumers are
often uncertain about the quality of services, this study
examines whether the clarity, consistency, and credibility
of a brand signal have any effects on consumer choices.
The results of the experiment show that signal clarity,
consistency, and credibility have positive impacts on
consumer choice of service brands. Among these
characteristics of a signal, signal credibility has the greatest
impact on consumer choices. In addition, the interaction
between signal clarity and credibility also affects consumer
choices. However, this effect is very modest compared to
the main effects. Further, our additional analysis indicates
that consumers’ gender and age do not moderate consumer
choices. These findings are consistent with research in the
signaling theory perspective on brands (Erdem and Swait
1998; Rao, Qu, and Ruekert 1999). Thus, other things
being equal, signal credibility is the key element in
building a strong brand based on signaling theory.
The findings of this study suggest several
implications for mobile phone service marketers. When a
market is characterised by information asymmetry, signals
are an appropriate mechanism to convey the unobservable
of services to consumers. Nevertheless, not all signals are
efficient. Only clear, consistent, and credible signals
perceived by consumers can direct their choices. In order to
have such signals, service providers should invest in their
brand signals and use marketing-mix strategies to
communicate the signal with consumers. This will facilitate
the development of consistent brand images, fostering the
relationship between brands and their consumers (Farquhar
1989). Previous research has shown that the credibility of a
brand signal will be higher when the brand receives higher
investments (Erdem and Swait 1998). This will benefit
service firms in terms of profits and reputation (brand
image) if truthful signals are received by consumers. As
signaling theory suggests, consumers evaluate the quality
of services based on the signals provided by service firms
because they believe that only quality service providers
would send clear, consistent, and credible signals to
consumers. However, if discrepancies between promised
and actual offerings are found to exist, that is, false signals
have been sent, costs in terms of immediate profits, future
profits, and reputation are likely to occur (Kirmani and Rao
2000; Tirole 1988). Therefore, service providers should
invest in brand signals in order to achieve more effective
399
df
1
1
1
1
1
1
1
1
p-value
.000
.000
.000
.369
.000
1.000
1.000
.000
signals and should keep in mind the costs associated with
their false signals. In sum, signals are an appropriate and
effective mechanism that service providers can employ
where asymmetric information exists. The effectiveness of
a signal can only be achieved when it is clear, consistent,
and credible.
This study has several limitations. Firstly, it is the use
of a student sample. Although research has shown that parttime students can be used as a surrogate for consumers
(James and Sonners 2001), a more representative sample is
of interest in future research. Secondly, this study examines
only one service, that is, the mobile phone service.
Different types of services should be examined to capture a
wider picture of the usefulness of signaling theory. Thirdly,
this study examines the choice of consumers based only on
signal characteristics. Future research should incorporate
other service attributes to the profile in order to make a
comparison between their effects on consumer choices.
Fourthly, this study employs a simple method of analysis–
binary logistic regression–to analyse best-worst data. Other
methods of analysis such as multinomial logit models will
also be suitable for the analysis of such a data set, which
can be employed in future research. Finally, signaling
theory offers explanations of consumer choices under the
condition of asymmetry of information, that is, how they
respond to firms’ actions. Psychological research into
consumers helps understand how consumers make
inferences, that is, how they process signals (Boulding and
Kirmani 1993). Therefore, an integration of these two
approaches in future research will enhance our
understanding of customers’ beliefs in and reactions to
signaling.
REFERENCES
Boulding, W. and A. Kirmani (1993), “A Consumer-Side
Experimental Examination of Signaling Theory: Do
Consumers Perceive Warranties as Signals of
Quality?”, Journal of Consumer Research, 20 (June),
111-23.
Caves, R. E. and D. P. Greene (1996), “Brand’s Quality
Levels, Prices, and Advertising Outlays: Empirical
Evidence on Signals and Information Costs”,
International Journal of Industrial Organisation, 14
(1), 29-52.
Charzan, K. (2005), “The Options Pricing Model: An
Application
of
Best-Worst
Measurement”,
http://www.sawtoothsoftware.com.
Erdem, T. and J. Swait (1998), “Brand Equity as Signaling
Phenomenon”, Journal of Consumer Psychology, 7
(2), 131-57.
Erdem, T., J. Swait, and J. J. Louviere (2002), “The Impact
of Brand Credibility on Consumer Price Sensitivity”,
International Journal of Research in Marketing, 19,
1-19.
Farquhar, P. H. (1989), “Managing Brand Equity”,
Marketing Research, 1 (September), 24-33.
Finn, A. and J. Louviere (1992), “Determining the
Appropriate Response to Evidence of Public
Concern: The Case of Food Safety”, Journal of
Public Policy and Marketing, 11 (1), 231-39.
Heil, O. and T. Robertson (1991), “Toward a Theory of
Competitive Market Signaling: A Research Agenda”,
Strategic Management Journal, 12 (6), 403-18.
Herbig, P. and J. Milewicz, (1993), “The Relationship of
Reputation and Credibility to Brand Success”,
Journal of Consumer Marketing, 10 (3), 18-24.
____ (1994), “Marketing Signals in Service Industries –
Implications for Strategic Decision Making and
Profitability”, Journal of Services Marketing, 8 (2),
19-35.
____ (1996), “Market signalling: A Review”, Management
Decision, 34 (1), 35-45.
James, W. L. and B. S. Sonner (2001), “Just Say No to
Traditional Student Samples”, Journal of Advertising
Research, 41 (September-October), 63-71.
Kirmani, A. and A. R. Rao (2000), “No Pain, No Gain: A
Critical Review of the Literature on Signaling
Unobservable Product Quality”, Journal of
Marketing, 66 (April), 66-79.
Kirmani, A. and P. Wright (1989), “Money Talks:
Perceived Advertising Expense and Expected Product
Quality”, Journal of Consumer Research, 16
(December), 160-71.
Koku, P. S. (1995), “Price Signaling: Does It Ever Work?”,
Journal of Consumer Marketing, 12 (1), 45-9.
Kuhfeld, W. F. (2005), Marketing Research Methods in
SAS, SAS Institute Inc.
Louviere, J. J., D. A. Hensher, and J. D. Swait (2000),
Stated Choice Methods: Analysis and Applications,
Cambridge, UK: Cambridge University Press.
Marley, A. A. J., Louviere, J. J. (2005), “Some
Probabilistic Models of Best, Worst, and Best–Worst
Choices”, Journal of Mathematical Psychology, 49
(6), 464-80.
Mishra, D. P., J. B. Heide, and S. G. Cort (1998),
“Information Asymmetry and Levels of Agency
Relationships”, Journal of Marketing Research, 35
(August), 277-95.
Ngo, T. T. (2005), “Viettel: Competition by its own way”,
The Vietnam Manager, November, 28 (in
Vietnamese).
Rao, A. R. and R. W. Ruekert, “Brand Alliances as Signals
of Product Quality”, Sloan Management Review, 36
(1), 87-97.
Rao, A. R., L. Qu, and R. W. Ruekert (1999), “Signaling
Unobservable Product Quality through a Brand
Ally”, Journal of Marketing Research, 36 (May),
258-68.
Simester, D. (1995), “Signaling Price Image Using
Advertised Prices”, Marketing Science, 14 (Spring),
166-89.
Soberman, D. A. (2003), “Simultaneous Signaling and
Screening with Warranties”, Journal of Marketing
Research, 40 (May), 176-92.
Spence, M. (1973), “Job Market Signaling”, Quarterly
400
Journal of Economics, 87 (3), 355-74.
Srivastava, J. and N. H. Lurie (2004), “Price-Matching
Guarantees as Signals of Low Store Prices: Survey
and Experimental Evidence”, Journal of Retailing,
80, 117-28.
Sternthal, B., R. Dholakia, and C. Leavitt (1978), “The
Persuasive Effect of Source Credibility: Tests of
Cognitive Response”, Journal of Consumer
Research, 4 (March), 252-60.
Street, D. J., L. Burgess, and J. J. Louviere (2005), “Quick
and Easy Choice Sets: Constructing Optimal and
Nearly Optimal Stated Choice Experiments”,
International Journal of Research in Marketing, 22
(4), 459-70.
Tirole, J. (1988), The Theory of Industrial Organisation,
Cambridge, CA: MIT Press.
VietnamNet (2005), “We Need Quality, Not Sales
Promotions”,
http://www.vnn.vn/cntt/
2005/10/498549.
Zeithaml, V. A. and M. J. Bitner (2000), Services
Marketing, Boston: McGraw-Hill.