Description
This investigation explores the effect of sales framing on the attractiveness of mixed promotions. The
authors propose that the “focused-discount” promotion format (e.g., 50% off the purchase of a second
item) is more attractive to consumers than the “all-discount” promotion format (e.g., 25% off the purchase
of two items) due to the effect that different frames have on the mental accounting processes of
consumers. In Experiments 1 and 2, participants confronted by focused-discount (vs. all-discount)
promotion assigned a relatively higher discount value to the final product and perceived greater savings.
In Experiment 3, priming participants to focus on the first item reduced this framing effect of mixed
promotions. The results show that participants perceived greater savings in a focused-discount framing
(vs. all-discount) due to the differences in how discounts were assigned for each item. Finally, this paper
suggests that companies may frame promotions to increase their relative attractiveness to consumers.
Sales framing, mental accounting, and discount assignments
Hsin-Hsien Liu
a
, Yu-Yeh Chiu
b, *
a
Department of Asia-Paci?c Industrial and Business Management, National University of Kaohsiung, Taiwan, ROC
b
Department of Marketing and Distribution Management, Chien Hsin University of Science and Technology, Taiwan, ROC
a r t i c l e i n f o
Article history:
Received 1 March 2013
Accepted 13 January 2015
Available online 21 July 2015
Keywords:
Framing
Mental accounting
Sales promotions
a b s t r a c t
This investigation explores the effect of sales framing on the attractiveness of mixed promotions. The
authors propose that the “focused-discount” promotion format (e.g., 50% off the purchase of a second
item) is more attractive to consumers than the “all-discount” promotion format (e.g., 25% off the pur-
chase of two items) due to the effect that different frames have on the mental accounting processes of
consumers. In Experiments 1 and 2, participants confronted by focused-discount (vs. all-discount)
promotion assigned a relatively higher discount value to the ?nal product and perceived greater sav-
ings. In Experiment 3, priming participants to focus on the ?rst item reduced this framing effect of mixed
promotions. The results show that participants perceived greater savings in a focused-discount framing
(vs. all-discount) due to the differences in how discounts were assigned for each item. Finally, this paper
suggests that companies may frame promotions to increase their relative attractiveness to consumers.
© 2015 College of Management, National Cheng Kung University. Production and hosting by Elsevier
Taiwan LLC. All rights reserved.
1. Introduction
Marketing managers often use promotions to enhance the
purchasing intentions of consumers. These promotions may focus
on changing the price of target items (e.g., discounts, coupons),
offering extra product (e.g., freebies), increasing the volume of
target items (e.g., buy one get one free, bonus pack), or pursuing a
mixed promotion (e.g., buy two, get 25% off; Chen, Monroe, & Lou,
1998; Diamond, 1992; Sinha & Smith, 2000). Previous research has
shown that the frames used in sales promotions in?uence the value
perceptions of consumers and purchasing intention (Barnes, 1975;
Chen et al., 1998; Diamond, 1992; Diamond & Sanyal, 1990;
Hardesty & Bearden, 2003; Lowe, 2010; Palazon & Delgado-
Ballester, 2009; Sinha & Smith, 2000; Yi & Yoo, 2011). Several au-
thors have also addressed the difference between monetary (e.g.,
discounts) and nonmonetary (e.g., extra products) promotions and
compared the relative attractiveness of different types of pro-
motions (Diamond, 1992; Diamond & Sanyal, 1990; Hardesty &
Bearden, 2003; Sinha & Smith, 2000; Yi & Yoo, 2011). However,
few studies have explored the effects of framing within the mixed
promotion approach or the relative attractiveness of these pro-
motions (e.g., Diamond & Sanyal, 1990; Sinha & Smith, 2000).
Mental accounting theory holds that individual consumers and
households track their ?nancial activities in a manner similar to
enterprises (Thaler, 1985, 1999). Therefore, consumers and house-
holds assign ?nancial events to speci?c accounts and categorize or
label them by expenditures and sources. However, different people
may categorize particular ?nancial events differently (Sha?r &
Thaler, 2006; Soman & Gourville, 2001). This paper proposes that
how a mixed promotion is framed in?uences the accounting pro-
cesses and discount assignments of target consumers. We consider
two framing types: an all-discount framing such as “buy N, get x%
off” and a focused-discount framing, which draws attention to the
last element, as in “buy a second item for y% off.”
The feature-matching model (Tversky, 1977) suggests that in-
dividuals map the features of a subject onto the features of the
referent/target as a starting point for comparison when comparing
two objects. Houston, Sherman, and Baker (1989) further suggest
that when making preference judgments between two items, in-
dividuals tend to use the last item as the subject because the
comparison process does not start until the second object appears.
Therefore, the features of the last object provide the starting point
of the comparison process and are weighted more heavily.
In mental accounting, the use of different semantic cues to
frame the sales promotion (e.g., “buy two, get 25% off” versus “buy a
* Corresponding author. Department of Marketing and Distribution Management,
Chien Hsin University of Science and Technology, No.229, Jianxing Rd., Zhongli Dist.,
Taoyuan City, 32097, Taiwan, ROC.
E-mail address: [email protected] (Y.-Y. Chiu).
Peer review under responsibility of College of Management, National Cheng
Kung University.
HOSTED BY
Contents lists available at ScienceDirect
Asia Paci?c Management Review
j ournal homepage: www. el sevi er. com/ l ocat e/ apmrvhttp://dx.doi.org/10.1016/j.apmrv.2015.01.002
1029-3132/© 2015 College of Management, National Cheng Kung University. Production and hosting by Elsevier Taiwan LLC. All rights reserved.
Asia Paci?c Management Review 20 (2015) 201e209
second item for 50% off”) is known to affect how consumers assign
discounts to each item, even when the ultimate cost of these items
or value of the discounts is the same. Moreover, people tend to
assign greater importance on the last (or immediately preceding)
itemwhen considering mixed promotions because this item acts as
the subject of the comparison. Conversely, the original price rep-
resents a referent (Tversky, 1977) that consumers use as a com-
parison against the last item in a mixed promotion, which is most
salient. Therefore, a focused discount promotion may be presumed
to be more attractive than an all-discount promotion because the
last item in the former frame appears to offer a greater discount.
This paper conducts three experiments to con?rm these pre-
dictions. This study offers several contributions to the existing
literature: (1) a new and more comprehensive perspective for
explaining the relative attractiveness of a mixed promotion across
different framings is suggested; (2) the differences in mental ac-
counting processes and discount assignments between the two
framings are explored; (3) the moderating in?uence of focus
priming on the relative attractiveness of different frames is
demonstrated; and (4) key implications for practice and avenues
for further research are introduced.
2. Theoretical background
2.1. Framing effect of promotions
Retailers often vary the amount and level of discounting and the
way in which promotions are framed in order to in?uence con-
sumer perceptions of transaction value and purchase intentions.
Previous research con?rms that how alternatives are framed in-
?uences the relative preference judgments and choices of con-
sumers (Kahneman & Tversky, 1979; Sinha & Smith, 2000).
This framing effect is well demonstrated in research into mar-
keting promotions. For example, Barnes (1975) found that, when
faced with the three price promotion formats of “special - $75”;
“25% off - $75”; and “regular price $100, sale price $75”, consumers
responded most positively to the latter version. Furthermore, Chen
et al. (1998) demonstrated that a price promotion framed as a dollar
reduction is more attractive in the context of high-price products,
whereas one framed as a percentage reduction is more attractive
for low-price products. Some researchers have employed
assimilation-contrast theory (Gabor & Granger, 1964; Kalyanaram
& Winer, 1995; Monroe, 1971) and Weber's law (Harlam, Krishna,
Lehmann, & Mela, 1995) to explain the effect of price framing. Ac-
cording to the assimilation-contrast theory, there is a price range
that is internal to consumers, known as latitude of acceptance,
which de?nes the range of acceptable prices. A reduced price that is
within a consumer's latitude of acceptance is assimilated into the
acceptable range and becomes acceptable. A reduced price that is
outside the latitude of acceptable will be contrasted with the
acceptable range and becomes noticeable. This concept relates
closely to Weber's law (Monroe, 1971). Weber's law suggests that
the difference threshold (or “just noticeable difference”), which
describes the minimum difference in a stimulus that is detectible
by an individual. A larger change in price is more likely to be
detected or perceived by consumers. Thus, the focus-discount
frame (e.g., “get the second one 50% off”) is more noticeable than
the all-discount frame (e.g., “buy two, get 25% off.”) and more likely
to raise purchase intention (Harlam et al., 1995).
In contrast, other authors have turned to the silver lining prin-
ciple of mental accounting theory. Based in the prospect theory, the
silver lining principle proposes that utility functions are concave for
gains and convex for losses and that the curve for losses is steeper
than that for gains. When consumers confront a small gain (e.g.,
price discount) relative to a large loss (e.g., purchasing cost), they
prefer to separate them psychologically into two different accounts
rather than integrate them. Diamond and Campbell (1989) and
Diamond and Sanyal (1990) also proposed that nonmonetary pro-
motions (e.g., freebies, “buy one, get one free”) appear segregated
from the purchasing cost and framed as gains because the units for
the promotions differ from the price (i.e., money). In contrast,
monetary promotions (e.g., price discounts) should be integrated
into the purchasing cost and framed as reduced losses because the
units are the same, which facilitates their integration. Consumers
then have a preference for extra-product or volume promotions
over price discounts, assuming that the size of the promotion is
small relative to the purchase cost. Sinha and Smith (2000) and
Smith and Sinha (2000) show that consumers prefer stores that
offer extra-product promotions (e.g., buy one, get one free) to stores
offering mixed promotions (e.g., buy two, get 50% off), especially if
the products being offered may be hoarded. However, the silver
lining principle also requires that the amount of the gain should be
relatively small compared with the losses. In Sinha and Smith's
studies, the percentage of discount relative to the cost may be too
large (50%). Further adding to the confusion, it seems that, in
comparison with assimilation-contrast theory and Weber's law, the
silver lining principle suggests that a different discount level is
necessary to exert the framing effects.
We propose another explanation for the variation in attrac-
tiveness of mixed promotions across different frames: Promotion
frames may in?uence consumers' mental accounting processes and
assignments of discounts to each product item such that their
perceptions of their savings differ for each product item in the
various framing conditions.
2.2. Mental accounting
Mental accounting is a cognitive rule that consumers use to
organize, evaluate, and record ?nancial activities. Consumers keep
track of their ?nancial activities, with accounting rules in?uencing
their ?nancial decisions explicitly and implicitly. Mental account-
ing assumes that people value a single event in multiple ways
because they have the ?exibility to categorize. For example, Sha?r
and Thaler (2006) proposed that individuals may evaluate a
transaction in multiple ways if the time of consumption is
different than the time of purchase. An example of this is when
theatergoers purchase tickets months in advance of a perfor-
mance. People tend to regard their early purchases as an invest-
ment rather than an expenditure, and then by the time they
consume the item, they have forgotten the price, so the product
seems free or a savings (e.g., spending $20 for a bottle of wine that
appreciates over time). If they lose a previously purchased product
(e.g., breaking the bottle of wine), the replacement costs appear
more salient.
Soman and Gourville (2001) demonstrated that consumers may
?exibly assign the cost across each item when the connection be-
tween costs and bene?ts is ambiguous, as with price bundling (e.g.,
$1000 for a travel package that includes transportation, hotel, and
food). Similarly, Prelec and Loewenstein (1998) found that paying
with a credit card (versus cash) diminishes the salience of the
payments and reduces the connection between costs and bene?ts.
Together, these studies show that the characteristics of a
transaction affect how consumers assign the cost of that trans-
action. If the discount assigned to each product item varies across
the frames used, does the speci?c classi?cation in?uence con-
sumers' perceptions of savings? The frames of a mixed promotion
(all-discount versus focused-discount) may cause consumers to
assign discounts to items in varying ways, and the relative attrac-
tiveness of the different frames may depend on the focus that
consumers put on the last item.
H.-H. Liu, Y.-Y. Chiu / Asia Paci?c Management Review 20 (2015) 201e209 202
2.3. Feature matching model
In comparing two objects, the subject (i.e., starting point in the
comparison process) is mapped onto the referent (i.e., target of the
comparison process). Thus, the characteristics of the subject are the
checklist used to evaluate the referent. The referent is usually the
more prominent of the two items. For example, most people will
consider the similarity of their friend (subject) to Marilyn Monroe
(referent) rather than the other way around because Marilyn
Monroe is the more famous and prominent of the two (Tversky,
1977). Therefore, Houston et al. (1989) suggested that in making
preference judgments, consumers use the second item as the
subject of their comparison because the comparison process cannot
start until this second item appears.
Based on the feature-matching model and the ?ndings of
Houston et al. (1989), we predict that people focus more on the last
itemwhen considering mixed promotions because that item serves
as the subject of the comparison. Moreover, we predict that con-
sumers use a regular, prominent price as the referent. Therefore,
consumers exposed to a focused-discount framing should perceive
more savings than consumers exposed to an all-discount framing
because the last itemin a focused discount frame accounts for more
of the discount than the last item in the all-discount framing.
Imagine that Mr. M and Mr. Z both enjoy two cups of coffee
(US$2/each) every morning. This morning, Mr. Z notices his coffee
shop has a “buy two, get 25% off” promotion (i.e., all-discount
framing) and takes two cups to the of?ce. Mr. M notices that his
coffee shop has a “50% off a second cup” promotion (i.e., focused-
discount framing) and also buys two cups. Although both Mr. M
and Mr. Z pay US$3 for their two cups of coffee, they may not
perceive the same savings. They both focus on the second cup of
coffee and compare its price with the original price (US$2).
Therefore, Mr. Z likely perceives that he has saved 25%, whereas Mr.
Mperceives that he has saved 50%. As a result, the focused-discount
framing should be more attractive than the all-discount framing:
H 1. Consumers exposed to focused-discount framing perceive
more savings than consumers exposed to all-discount framing.
Different framing-related discount assignments for each prod-
uct itemshould also result in consumers perceiving pain differently
in the event of a loss. In the coffee purchase example, Mr. M in the
focused-discount framing condition perceives a 50% savings on the
second item but no discount for the ?rst item whereas Mr. Z in the
all-discount framing condition perceives a 25% discount on both
items. Accordingly, if they each drop one of their cups on the way to
the of?ce, Mr. Z should feel more pain than Mr. M. Therefore, we
propose:
H 2. When faced with the loss of the ?nal item, consumers in the
all-discount framing condition feel more pain than those in the
focused-discount framing.
H 3. Participants in the all-discount framing condition divide the
discount equally across each item, whereas participants in the
focused-discount framing condition assign the total discount to the
?nal item.
3. Experiment 1: promotional frames and discount
allocations
3.1. Experiment 1A
To test Hypotheses 1e3, we examined whether consumers
exposed to a focused-discount framing perceive more savings than
their all-discount framing peers, as well as whether different
perceptions emerge because consumers focus on the last item and
assign the discounts differently across the two framing conditions.
3.1.1. Participants and procedures
Ninety-eight Taiwanese college students participated in the
experiment as a regular assignment and without the incentive of
additional rewards or credit. As the purpose of the experiment was
to explore the perception processes of consumers, there was no
“right” or “wrong” answer. The participants were asked to answer a
two-page questionnaire that included a judgment task, described
as follows:
Decision scenario: Every morning, Mr. M and Mr. Z both spend
NT$120 to purchase two cups of coffee (unit price NT$60), after
which they take their cups to the of?ce. This morning, when Mr. M
was buying the coffee in the store, he noticed that the coffee shop
was holding a promotion: “buy one cup of coffee, get a second for
half price.” He accepted the promotion and paid for two cups. At the
same time, Mr. Z noticed that his coffee shop was holding a pro-
motion: “25% off the purchase of two cups of coffee.” He also
accepted the promotion and paid for two cups.
After reading the description, participants answered the
following question: “Although Mr. M and Mr. Z both paid NT$90 for
two cups of coffee, who do you think perceived the greater sav-
ings?” Participants chose among three options (Mr. M, the same, or
Mr. Z).
Next, they were asked to imagine that “after Mr. Mand Mr. Z had
?nished their ?rst cups of coffee, they both spilled their second cup
carelessly.” The participants were then asked: “Who do you think
felt greater pain over his loss?” Participants again chose among the
same three-option sets (Mr. M, the same, or Mr. Z). Finally, the
participants were asked to imagine which one of ?ve statements
best captured Mr. M's and Mr. Z's feeling about his loss: (a) a cup of
coffee costs NT$60, so I lost NT$60; (b) I spent NT$90 for two cups
of coffee, so I lost NT$45; (c) I have ?nished a cup of coffee that costs
NT$60, so I lost NT$30; (d) I paid for the coffee this morning, so my
loss costs me nothing; or (e) other (participants checking this item
were asked to describe their feeling). After they ?nished the
questionnaire, the participants were thanked and debriefed.
3.1.2. Results
For the ?rst question, most participants (56.1%) believed that Mr.
M and Mr. Z would perceive the same savings. Among those who
believed savings would be perceived differently (43/98), 30 par-
ticipants believed that Mr. M would perceive a greater savings than
Mr. Z and only 13 participants believed Mr. Z would perceive a
greater savings. After factoring out the group that believed that Mr.
M and Mr. Z would perceive the same savings, the difference was
statistically signi?cant (c
2
(1; N ¼ 43) ¼ 6.72, p < .05) in support of
Hypothesis 1. This result is not an artifact of the inability of par-
ticipants to calculate the costs correctly, as they were aware that
Mr. M and Mr. Z had paid the same amount.
For the second question, 45.9% of the participants (45/98)
indicated that Mr. M and Mr. Z would perceive the same degree of
pain in spilling their second cup of coffee; 10.2% believed that Mr. M
would feel more pain; and 43.9% of the participants believed that
Mr. Z would feel more pain (c
2
(2; N¼98) ¼23.65, p < .05). In other
words, among those participants who believed Mr. M and Mr. Z
would feel differing amounts of pain over their loss, a signi?cantly
larger number believed that Mr. Z would feel more pain than Mr. M
(c
2
(1; N ¼ 53) ¼ 20.55, p < .05). This result supports Hypothesis 2.
For the third question, while 41.8% of participants believed that
Mr. M lost NT$30, only 13.3% of participants believed that Mr. Z had
lost NT$30. In contrast, 63.3% of participants believed that Mr. Z lost
NT$45, with only 32.7% of participants believing the same of Mr. M
(Table 1). These results match our expectations that the all-
H.-H. Liu, Y.-Y. Chiu / Asia Paci?c Management Review 20 (2015) 201e209 203
discount framing scenario encourages consumers to divide the
discount equally across two items and that the focused-discount
framing scenario encourages consumers to assign the entire dis-
counted value to the second item.
This analysis applied CATMOD modeling to the repeated-
measures categorical data (Stokes, Davis, & Koch, 1995), with par-
ticipants' thoughts about Mr. M's and Mr. Z's losses as the depen-
dent variables and the promotion frames as the independent
variables. The difference was signi?cant: c
2
(4; N ¼ 98) ¼ 30.93,
p < .01. Ignoring those participants who chose options (a), (d), or (e)
for either Mr. M or Mr. Z and focusing only on those who chose
options (b) and (c), we again obtained signi?cant results, with c
2
(1;
N ¼ 64) ¼ 24.64, and p < .01, in support of Hypothesis 3.
3.1.3. Discussion
The ?ndings of Experiment 1A support the proposition that
consumers prefer mixed promotions with a focused-discount
framing rather than an all-discount framing. This difference
mainly re?ects the tendency of individuals to assign discounts
differently to each product item and focus on the last item of the
mixed promotions. Those exposed to the focused-discount framing
thus perceive more savings, because the last item in their frame
accounts for more of the discount. Furthermore, participants
believed that a consumer who loses this last item perceives less
pain than a consumer in the all-discount framing scenario.
According to Sha?r and Thaler (2006), an actual loss event
causes the replacement cost to become more salient. However, the
replacement cost effect was not very salient in Experiment 1A (i.e.,
only 18% participants chose option (a)). Perhaps these participants
assigned the discount to the ?rst and second items, whereas par-
ticipants did not need to assign discounts separately in Sha?r and
Thaler's study. This distinction may explain why the effect of
replacement cost is so weak in the present study.
3.2. Experiment 1B
Although the results in Experiments 1A are consistent with our
propositions, the design used for Experiment 1A may have been
overly arti?cial as few people buy two cups of coffee at once. To
overcome this potentially confounding in?uence, we adopted
laundry detergent as the target product for Experiment 1B. Most
consumers use laundry detergent over longer periods of time and
are able to store bulk amounts. In addition, the participants in
Experiment 1B were not asked to indicate which options better
capture Mr. M's and Mr. Z's losses but rather were asked to write
down their perceptions of the loss.
3.2.1. Participants and procedures
Ninety-six Taiwanese college students participated in the
experiment as a regular assignment and without the incentive of
additional rewards or credit. Participants responded to a two-page
questionnaire that included three judgment tasks. The design was
similar to Experiment 1A, as follows:
Decision scenario: M and Z are college students. Both regularly
wash their clothes using X brand laundry detergent. Each time they
use up all their laundry detergent, they go to the supermarket and
spend NT$120 for a box of X brand laundry detergent (2 kg). This
morning, M and Z both noticed they were out of laundry detergent,
so they went to nearby supermarkets. As M was shopping, he
noticed that the supermarket was running a promotion: “Get a
second box of X brand laundry detergent for half price.” Because he
had room to store the laundry detergent and had used X brand to
wash his clothes, M accepted the promotion and paid NT$180 for
two boxes. At the same time, Z noticed that his supermarket was
also running a promotion: “Buy two boxes of X brand laundry
detergent, get 25% off.” He also had room to store the extra deter-
gent, accepted the promotion, and paid NT$180 for two boxes.”
After reading the scenario, participants ?rst completed a judg-
ment task in which they were asked which buyer they thought
would perceive more savings (M, the same, or Z). Next, they were
asked to imagine that “After M and Z used up their ?rst box of
laundry detergent, they opened the second box. However, they
both noticed that their laundry detergent had hardened and
couldn't be used any more.” Participants then answered the
following question: “Which one do you think would feel more
pain?” Participants again chose among the same three-option sets.
Finally, the participants were asked to answer the question: “When
M (Z) noticed his laundry detergent was useless, how much money
do you think M (Z) lost?”
3.2.2. Results
Even with a different target product, the results were similar to
those from Experiment 1A. In the ?rst task, most participants
(70.8%) believed that Mand Z would perceive the same savings, but
among those who did not (28/96), 23 participants believed that Mr.
M would perceive more savings than Z and only 5 participants
believed that Z would perceive more savings. After factoring out
those who believed the two would perceive the same savings, the
difference was statistically signi?cant, c
2
(1; N ¼28) ¼11.57, p < .01,
in support of Hypothesis 1.
In the subsequent task, 60.4% of participants (58/96) indicated
that Mand Z would feel the same degree of pain over the loss of the
second box of laundry detergent; 8.3% thought that M would feel
more pain, and 31.3% thought that Z would suffer more, c
2
(2;
N ¼ 96) ¼ 39.25, p < .01. Thus, most of the participants who felt the
two would feel differing amounts of pain believed that it would be
Z who would feel the most pain, c
2
(1; N ¼ 38) ¼ 12.74, p < .01.
These results supported Hypothesis 2 as well.
Finally, the average perceived loss was NT$66.40 for M and
NT$76.19 for Z. A new variable, “DiffCost,” which is equal to the
within-subject difference in the perceived loss for M and Z, differed
signi?cantly from 0 (t(N ¼ 96) ¼ 5.07), p < .05. Thus, participants
perceived the loss for Z to be greater than the loss for M, in support
of Hypothesis 2.
To explore the discount assignments of participants across the
two frames, we classi?ed the perceived loss for M and Z into ?ve
groups based on the differences between the perceived loss and
Table 1
Participant perceptions of Mr. M's and Mr. Z's losses, Experiment 1A.
Mr. M (%) Mr. Z (%)
(a) A cup of coffee costs NT$60, so I lost NT$60. 17.4 18.4
(b) I spent NT$90 for two cups of coffee, so I lost NT$45. 32.7 63.3
(c) I have ?nished a cup of coffee that costs NT$60, so I lost NT$30. 41.8 13.3
(d) I paid for the coffee this morning, so my loss costs me nothing. 2.0 1.0
(e) Other. 6.1 4.0
Notes: The number in each cell represents the relative share for that opinion.
H.-H. Liu, Y.-Y. Chiu / Asia Paci?c Management Review 20 (2015) 201e209 204
two reference prices, NT$60 and NT$90. The outcomes were labeled
“< 60” if the stated perceived loss was less than NT$60; “ ¼ 60” if it
equaled NT$60; “60e90” if it was greater than NT$60 but less than
NT$90; “ ¼ 90” if the stated perceived loss equaled NT$90; and “>
90” if it was greater than NT$90.
As shown in Table 2, 59.4% of the participants believed that M
had lost NT$60, while only 27.1% believed the same of Z. In contrast,
50% of the participants believed that Z had lost NT$90, while only
22.9% believed the same of M. These results again match our ex-
pectations: Participants in the all-discount framing scenario divide
the discount equally across the two items, whereas those in the
focused-discount framing scenario assign the entire discount to the
second item.
We again applied CATMOD modeling to the repeated-measures
categorical data, with participants' thoughts about M's and Z's
losses as the dependent variables and the promotion frames as the
independent variables. We ignored those participants who chose
the more extreme options for either M or Z because the perceived
losses equal to NT$60 and NT$90 provide the primary means to
examine Hypothesis 3. We again found signi?cant results, with
c
2
(1; N ¼ 82) ¼ 39.33, and p < .01.
3.2.3. Discussion
The results of this study were consistent across the two target
products used. Although the perceived losses for M and Z in
Experiment 1B were based on the subjective statements of partic-
ipants rather than on a choice from among ?ve options as in
Experiment 1A, the results still support Hypothesis 3.
In Experiment 1, statistical analyses were conducted after
excluding the data that indicated no difference between Mand Z, as
the within-subject design format poses the risk of reducing the
framing effect (Levin, Johnson, & Davis, 1987; Levin, Gaeth,
Schreiber, & Lauriola, 2002). A framing effect that exists in a
between-subject design vanishes in a within-subject design (Levin
et al., 1987). Similarly, the goal-framing effect decreases more
signi?cantly in the within-subject design than in the between-
subject design (Levin et al., 2002). To control for individual vari-
ance and to address consumer preference, a within-subject design
may be a better choice, even if participants may desire to respond in
a consistent manner (Wiseman & Levin, 1996). Authors who have
adopted within-subject designs typically summarize the results
after excluding data that indicate no difference between frames or
formats (e.g., LeBoeuf & Sha?r, 2003). Experiment 2 adopts a
between-subject design that obviates the need to either support or
exclude alternative explanations.
4. Experiment 2: framing of sales promotions and discount
allocations (between-subjects)
Although the results of Experiment 1 support the hypotheses,
we conducted Experiment 2 with a between-subjects design to
exclude the alternative explanations for these results. Additionally,
we used a different judgment scenario (concert ticket) and a
different discount level (e.g., buy three, get one free) to enhance the
generalizability of our ?ndings. The price of the target (concert
ticket) in this scenario was much higher than those in Experiment 1
(coffee, laundry detergent). Support for our hypotheses in the
context of high priced targets would further bolster con?dence that
these hypotheses apply not only to consumer but also to higher-
priced retail goods.
4.1. Experimental design
Experiment 2 uses a between-subjects design: The all-discount
frame indicates “buy four, get 25% off,” and the focused-discount
frame notes, “buy three, get one free.” The scenario, modi?ed
from Sha?r and Thaler (2006), is as follows:
Decision scenario (All-discount/focused-discount): Mr. Chang
spends NT$2000 monthly on buying concert tickets. Today, when
he went to buy his concert ticket for the following week, he noticed
that the concert hall was conducting a “buy four tickets, get 25% off
[buy three tickets, get one free]” promotion. The four tickets could
be used separately over six months. Although the regular price for
four tickets would be NT$8000, Mr. Chang only spends NT$6000 to
buy them. He plans to use one ticket each month.”
4.2. Participants and procedure
One hundred and ?fty-two Taiwanese college students partici-
pated in the experiment as a regular assignment and without the
incentive of additional rewards or extra credit. Participants were
randomly assigned to either the focused discount (N ¼ 78) or all-
discount (N ¼ 74) group. After reading the scenario, all partici-
pants answered three questions. The ?rst question “Which of the
following descriptions do you think best captures Mr. Chang's
feeling when he spent NT$6000 for the four tickets?” had the
following four potential answers: (a) The NT$6000 spent to buy
four tickets is no different fromspending the same amount on other
retail items; (b) I made a NT$6000 investment in four tickets, which
I will gradually consume in the future; (c) I spent NT$6000 to buy
four tickets that normally cost NT$8000, so I just saved NT$2000;
and (d) other (please explain). The second question “After months
when Mr. Chang is ready to use his ?nal ticket, which of the
following descriptions do you think best capture his feeling about
the cost of this ?nal ticket?” had the following four potential an-
swers: (a) This ticket cost NT$2000, the same as the current price of
a single ticket; (b) This ticket cost NT$1500 because I spent a total of
NT$6000 for the four tickets; (c) This ticket is free, because the
NT$6000 I paid covered the three previous tickets; and (d) other
(please explain). The third question “Imagine that when Mr. Chang
is getting ready for the fourth concert, he found that he had lost his
ticket and had no way of purchasing another as a replacement.
Which of the following descriptions do you think best captures Mr.
Chang's feeling about losing the ?nal ticket?” prompted the par-
ticipants to choose one of the four options posited for the second
question.
4.3. Results
As shown in Table 3, most participants (about 90%) believed that
Mr. Chang either saved NT$2000 or made an investment when he
paid NT$6000 for the four concert tickets regardless of framing
condition (c
2
(3; N ¼ 152) ¼ .86, p > .1). This ?nding is similar to
Sha?r and Thaler (2006).
When considering the use of the ?nal ticket, 46.2% of partici-
pants in the focused-discount framing group believed that the last
ticket was free, while only 31.1% of those in the all-discount framing
group believed the same. In contrast, 52.7% of the participants in
the all-discount framing group believed that the cost of the last
ticket was NT$1500 (i.e., 25% off), while only 41.0% of those in the
focused-discount framing group believed the same. If we ignore the
Table 2
Participants' feelings about M's and Z's losses, Experiment 1B.
90
M (%) 11.5 59.4 0 22.9 6.2
Z (%) 10.4 27.1 2.1 50 10.4
Notes: The number in each cell represents the relative share of that thought.
H.-H. Liu, Y.-Y. Chiu / Asia Paci?c Management Review 20 (2015) 201e209 205
data from those who chose options (a) or (d), the result is still
marginally signi?cant, c
2
(1; N ¼ 130) ¼ 3.28, p ¼ .07.
With regard to the lost ticket scenario, 48.6% of the participants
in the all-discount framing group considered the lost ticket equal to
a loss of NT$1500, while only 29.5% of those in the focused-discount
framing group believed the same. In contrast, 46.2% of the focused-
discount framing group believed that Mr. Chang had lost NT$0,
while only 29.7% of those in the all-discount framing group
believed the same. If we ignore the data from those who chose
options (a) and (d), the difference is statistically signi?cant, c
2
(1;
N ¼ 117) ¼ 6.24, and p ¼ .01.
Thus, the present study found overall support for Hypothesis 3:
Participants in the focused-discount framing group are more likely
to assign the entire discount to the ?nal ticket and regard it as free,
while participants in the all-discount group tend to assign the
discount equally to each ticket.
4.4. Discussion
Hypothesis 3 continues to be supported even though Experi-
ment 2 involves a different experimental design (between-sub-
jects), a different judgment scenario (concert ticket), and a different
discount level (e.g., buy three, get one free). We thus consider our
proposition generalizable across conditions.
At the moment that the four tickets were purchased, the par-
ticipants differed little in their perceptions of the purchase. As in
Sha?r and Thaler's (2006) study, most participants regarded the
payment as an investment or form of savings. However, in
considering the ?nal ticket under the different framing conditions,
the participants held varying cost perceptions for that ticket when
it was both used and lost.
As previously noted, the loss of the ?nal ticket did not make the
replacement cost effect very salient (i.e., 21e22% of participants
chose option (a)). This may imply that the assigned discount
weakens the effect of the replacement cost, although the loss of the
?nal ticket increases the effect of the replacement cost slightly in
comparison with its use (from 12e14% to 21e22%), which is
consistent with the result posited by Sha?r and Thaler (2006).
5. Experiment 3: the moderator of framing effects
Combining mental accounting and feature matching (Tversky,
1977) may support an inference that consumers will prefer
focused-discount framing to all-discount framing. However, might
the framing effects decrease or even reverse if consumers are
primed to focus on the ?rst rather than the ?nal item? In other
words, while we have established that consumers assign a greater
discount to the last item in the focused-discount framing, the
framing effect of mixed promotions may decline or even reverse if
consumers place greater focus on the ?rst item.
H 4. The framing effect of a mixed promotion is moderated by the
focus priming of consumers, with a focus on the ?rst item resulting
in a declining or even reverse effect rather than a focus on the
second item.
5.1. Experimental design and procedures
Experiment 3 was conducted similarly to Experiment 1A, except
that participants were prompted to focus on either the ?rst or
second (?nal) product item. Thus, 195 Taiwanese college students
participated in the experiment as a regular assignment and without
the incentive of additional rewards or extra credit. These partici-
pants were randomly assigned to one of two priming conditions
(?rst versus second item) and responded to a two-page question-
naire with three judgment tasks. The ?rst task was the same as the
?rst task in Experiment 1A. For the second task, participants were
primed to focus on the ?rst or second item by imagining either that
“As Mr. M and Mr. Z are enjoying their ?rst cups of coffee, they each
spill it carelessly” (N ¼ 99) or that “After Mr. M and Mr. Z have
enjoyed the ?rst cup of coffee, they start to drink the second cup of
coffee and spill it carelessly” (N ¼ 96). The remaining tasks were
similar to the ?nal two items in Experiment 1A.
5.2. Results
The results for the ?rst task were similar those from Experiment
1A: Most (61.5%) participants believed that Mr. M and Mr. Z would
perceive the same savings; 28.7% believed that Mr. M would
perceive a greater savings; and only 9.7% believed that Mr. Z would
perceive a greater savings, c
2
(2; N ¼ 195) ¼ 7.13, p ¼ .03. In
analyzing the data of only those participants who believed Mr. M
and Mr. Z would perceive a difference in savings, we identi?ed a
statistically signi?cant result, c
2
(1; N ¼ 75) ¼ 18.25, p < .01.
When participants focus on the spilled second cup, 57.3%
believed that Mr. Mand Mr. Z would suffer the same perceived loss,
29.2% believed that Mr. Z would suffer more, and only 13.5%
believed that Mr. M would suffer more. This result is similar to that
of Experiment 1A. However, when we primed participants to focus
on the ?rst spilled cup, the results reversed in that 51.5% believed
that Mr. M and Mr. Z would suffer the same perceived loss, 29.3%
believed that Mr. Mwould suffer more, and only 19.2% believed that
Mr. Z would suffer more, c
2
(2; N ¼ 195) ¼ 7.93, p ¼ .02. After
factoring out those participants who perceived the same pain
Table 3
Results for three tasks in Experiment 2.
Buy 3, get 1 free Buy 4, get 25% off
Task 1: At the time you make the purchase:
(a) The NT$6000 spent to buy four tickets is no different from spending the same amount on other retail items. 9.0 10.8
(b) I made an NT$6000 investment in four tickets, which I will gradually consume in the future. 32.1 27.0
(c) I spent NT$6000 to buy four tickets that normally cost NT$8000, so I just saved NT$2000. 57.7 59.5
(d) Other. 1.2 2.7
Task 2: The perceived cost of the ?nal ticket:
(a) This ticket cost NT$2000, the same as the current price of a single ticket. 11.5 13.5
(b) This ticket cost NT$1500 because I spent a total of NT$6000 for the four tickets. 41.0 52.7
(c) This ticket is free, because the NT$6000 I paid covered the three previous tickets. 46.2 31.1
(d) Other. 1.3 2.7
Task 3: The perceived cost of losing the ?nal ticket:
(a) This ticket cost NT$2000, the same as the current price of a single ticket. 20.5 21.7
(b) This ticket cost NT$1500 because I spent a total of NT$6000 for the four tickets. 29.5 48.6
(c) This ticket is free, because the NT$6000 I paid covered the three previous tickets. 46.2 29.7
(d) Other. 3.8 0
Notes. The number in each cell represents the relative percentage (%) for that thought.
H.-H. Liu, Y.-Y. Chiu / Asia Paci?c Management Review 20 (2015) 201e209 206
across both priming conditions, the result was even more pro-
nounced in favor of our hypothesis, c
2
(1; N ¼ 89) ¼ 7.31, p < .01.
Furthermore, when participants considered Mr. M's and Mr. Z's
losses separately (see Table 4), they differed signi?cantly between
the two priming conditions. When focusing on the second cup,
45.8% believed that Mr. M lost NT$30, while only 14.6% believed
that Mr. Z lost NT$30. In contrast, 54.2% believed that Mr. Z lost
NT$45, while only 28.1% believed that Mr. M lost NT$45. These
?ndings are similar to those from Experiment 1A and support
Hypothesis 3. For the participants that focused on the ?rst cup,
70.7% believed that Mr. Z lost NT$45, 41.4% believed that Mr. M lost
NT$45, 24.2% believed that Mr. M's loss equaled NT$60, and only
8.1% believed that Mr. Z's loss equaled NT$60.
Thus a focus on the ?rst rather than the second cup caused most
participants to imagine a higher perceived cost for Mr. M. For
example, 24.2% and 41.4% of the participants who focused on the
?rst cup thought Mr. M had lost NT$60 and NT$45, respectively,
whereas only 10.4% and 28.1% of this group assigned these amounts
when they focused on the second cup. In contrast, 24.2% of the ?rst-
cup-focused participants believed that Mr. M lost NT$30, while
45.8% of the second-cup-focused participants believed the same.
This ?nding supports the proposition that the focused-discount
framing scenario encourages consumers to assign the entire dis-
count to the last item.
The highest costs perceived with regard to Mr. Z do not differ
signi?cantly between the two priming conditions. Although the
ratio of those choosing NT$45 increases from 54.2% for the second-
cup-focused group to 70.7% for the ?rst-cup-focused group, the
ratio of those choosing NT$60 and NT$30 both decrease when the
focus is placed on the ?rst cup.
Regardless of which cup spills, most participants (70.7% ?rst cup
and 54.2% second cup) indicate that Mr. Z has lost NT$45. These
results suggest that participants distributed the discount equally
across the ?rst and second cups when using an all-discount
framing, in support of Hypothesis 3.
Finally, ignoring those participants who chose options (d) or (e)
for either Mr. M or Mr. Z and again applying CATMOD, with Mr. M's
and Mr. Z's presumed losses as the dependent variables and pro-
motion framing, participant focus (?rst versus second cup), and
their interaction as independent variables, we discover that pro-
motion framing (c
2
(2; N ¼ 167) ¼ 40.33, p < .01), participant focus
(c
2
(2; N ¼ 167) ¼ 10.95, p < .01), and their interaction (c
2
(2;
N¼167) ¼4.58, p < .01) are all signi?cant. In support of Hypothesis
4, the framing effect of promotions thus reverses when participants
focus on the ?rst item.
5.3. Discussion
Experiment 3 reveals the moderating effect of focus priming on
the framing effect of promotions. The results con?rm our propo-
sition, as the framing effect in?uences consumers' assignment of
discounts across items, with a greater focus on the ?nal item,
consumers perceive greater savings in the focused-discount
framing scenario than in the all-discount framing scenario. How-
ever, priming participants to focus on the ?rst item reduced the
framing effect.
While it may seem contradictory that participants perceived Mr.
M's loss as NT$45 rather than NT$60 for a ?rst cup of spilled coffee,
this result dovetails with the observations of Sha?r and Thaler
(2006) that consumers regard an advance purchase as an invest-
ment that they amortize over time. Therefore, many participants
did not consider Mr. M's loss as NT$60 but rather as NT$45 because
they had already assigned a portion of the discount to the ?rst item.
6. Conclusions and general discussion
Our three experiments support our overall proposition that
consumers perceive more savings when presented as a focused-
discount framing than as an all-discount framing. The framing ef-
fects of the mixed promotion result mainly from differential as-
signments of the discounts to each item in the different promotion
frames. Furthermore, we con?rmed the framing effect for a variety
of product types, including those that consumers buy in advance
and store (Experiments 1B and 2) and those that are bought and
consumed nearly simultaneously (Experiments 1A and 3).
Furthermore, the effect was con?rmed across different designs
(between-versus within-subjects) and promotion levels (buy 2, get
25% off versus buy 4, get 25% off). Finally, we demonstrated that
priming the focus of participants moderates the framing effect of
mixed promotions, with ?ndings robust across decision scenarios.
6.1. Comparison with previous studies of promotional framing
Sinha and Smith (2000) and Smith and Sinha (2000) suggested
that consumers prefer the “buy one, get one free” format over the
“buy two, get 50% off” format because the latter implies that con-
sumers must buy two items to get the discount, whereas the former
promises that they receive the discount by buying only one item.
However, their proposition does not explain the different prefer-
ences in Experiments 1 and 3 because our participants were
required to buy two items in both framing conditions. Moreover,
their proposition does not explicate the moderating effect of con-
sumer focus on the framing effect of the mixed promotion
(Experiment 3). In contrast, the results of the present study helps
explain the results of Sinha and Smith because consumers con-
fronted with a “buy two, get 50% off” deal tend naturally to focus on
the last item and perceive a half-off discount on the second item
purchased. In contrast, the second item in the “buy one, get one
free” deal is perceived as free (100% discount), which is much more
attractive.
Some researchers have employed the assimilation-contrast
theory and Weber's law (Harlam et al., 1995) to explain why
Table 4
Participants' thoughts about loss in Experiment 3.
Mr. M Mr. Z
First cup (N ¼ 99) Second cup (N ¼ 96) First cup (N ¼ 99) Second cup (N ¼ 96)
(a) A cup of coffee costs NT$60, so I lost NT$60. 24.2 10.4 8.1 15.6
(b) I spent NT$90 for two cups of coffee, so I lost NT$45. 41.4 28.1 70.7 54.2
(c) I have ?nished a cup of coffee that costs NT$60, so I lost NT$30 (second item focus).
[I still have a cup of coffee that costs NT$60, so I lost NT$30 (?rst item focus).]
24.2 45.8 12.1 14.6
(d) I paid for the coffee this morning, so my loss costs me nothing. 5.1 4.2 4.0 6.3
(e) Other. 5.1 11.5 5.1 9.3
Notes. The number in each cell represents the relative percentage of that thought. The wording of option (c) differed slightly across the two priming conditions, depending on
the focus priming scenarios.
H.-H. Liu, Y.-Y. Chiu / Asia Paci?c Management Review 20 (2015) 201e209 207
consumers prefer to “get the second one 50% off” instead of “buy
two, get 25% off.” They infer a range of acceptable prices that
consumers use to judge a price, such that a larger price change is
more likely to be outside the acceptable range and should be more
noticeable. However, their proposition does not account for the
?nding that priming participants to focus on the ?rst item reduces
the framing effect.
Diamond and Sanyal (1990) adopted Thaler's (1985) silver lining
principle to explain why an extra product (e.g., buy one, get one
free) is preferable to a mixed (buy two, get 50% off) promotion,
positing that the former appears as a gain and the latter is treated as
a reduced loss. However, as before, the silver lining principle does
not explain the results obtained by Experiment 3.
Distinguishing this study fromprevious research is important in
terms of the potential impact of the present study on future hy-
potheses. For example, the silver lining principle would likely
anticipate that the framing effects in relevant hypotheses to decline
or even reverse in line with promotional levels due to its reliance on
comparing gains and losses. In contrast, the assimilation-contrast
theory and Weber's law suggests that promotional levels should
be large enough to reach a threshold at which point the framing
effects should increase with promotional levels. However, the au-
thors note that the framing effect has been shown to exist across
many discount levels: 20.6% (Diamond & Sanyal, 1990), 25% (this
study), and 50% (Smith & Sinha, 2000).
Thus, our proposition appears more comprehensive and
reasonable with regard to how promotion frames in?uence the
mental accounting processes of consumers, especially when con-
sumers assign discounts to each item. This perspective indicates
why perceived savings differ across two different framing condi-
tions while explaining effectively the ?ndings of prior research. In
addition, understanding the assignment of discounts will assist
companies to set the format mixed promotions for maximum
impact, as discussed below.
6.2. Comparison with previous studies of the recency effect
The results in this study may re?ect recency, which is an order
effect. According to Haugtvedt and Wegener (1994), a primacy ef-
fect exists if people who encounter two opposing messages form
judgments more consistent with the ?rst message, while a recency
effect exists if judgments are more consistent with the opposing
message. However, studies into order effects tend to include many
pieces of information in order to explore the persuasive effect of
target information (e.g., Buda & Zhang, 2000; Haugtvedt &
Wegener, 1994). The order effect seems less likely to exist in the
context of a small number of information frames such as the two
presented in this study. Moreover, while individuals typically
receive information in a certain order, participants in this study
were informed that the consumers paid for two items at the same
time (Experiments 1 and 3). Thus we cannot assume a recency ef-
fect in this study. And if an order effect exists, why is it more likely
to be a recency than a primacy effect? Therefore, we ?nd no reason
to believe that the results of the present study are affected by a
recency effect.
6.3. Managerial implications
This study demonstrates the robustness of framing effects. Re-
sults are applicable to marketing practice, especially within the
context of retailers or marketing managers applying a mixed
approach to enhance the attractiveness of a promotional deal.
Moreover, this type of promotion format does not increase costs.
Therefore, the ?ndings of this study suggest that marketing
managers should use focused-discount framing rather than all-
discount framing to maximize the savings perceptions of
consumers.
However, the mixed promotion effect may decline if a company
consistently adopts the same promotion format due to consumers
no longer taking note of new promotions or reducing their refer-
ence price for the target product. Therefore, in general, companies
should vary their promotional formats. However, when adopting an
all-discount promotion after using a focused discount promotion, a
company should prime customers to focus on the ?rst item. An
example advertisement might read: “Enjoy two cakes for 25% off …
taste the delicious ?avor from the very ?rst bite.” This focus may
effectively increase the promotion-related savings perceived by
consumers.
Con?icts of interest
All contributing authors declare no con?icts of interest.
Acknowledgment
The research reported in this article was supported by research
grant NSC 99-2410-H-390-018-MY2 and NSC 101-2410-H-231-005
from the National Science Council of Taiwan. The authors
acknowledge the editor and anonymous reviewers for their helpful
suggestions. They also thank Tsai, Hsiao-Tsung for helpful discus-
sions and for suggesting relevant references.
References
Barnes, J. G. (1975). Factors in?uencing consumer reaction to retail newspaper sale
advertising. In Combined Proceedings of the American Marketing Association,
Chicago (pp. 471e477).
Buda, R., & Zhang, Y. (2000). Consumer product evaluation: the interactive effect of
message framing, presentation order, and source credibility. Journal of Product &
Brand Management, 9(4), 229e242.
Chen, S. S., Monroe, K. B., & Lou, Y.-C. (1998). The effects of framing price promotion
messages on consumers' perception and purchase intentions. Journal of
Retailing, 74(3), 353e372.
Diamond, W. D. (1992). Just what is a dollar's worth? Consumer reactions to price
discounts vs. extra product promotions. Journal of Retailing, 68(3), 254e270.
Diamond, W. D., & Campbell, L. (1989). The framing of sales promotions: effects on
reference price change. Advances in Consumer Research, 16, 241e247.
Diamond, W. D., & Sanyal, A. (1990). The effect of framing on the choice of super-
market coupons. Advances in Consumer Research, 17, 488e493.
Gabor, A., & Granger, C. (1964). Price sensitivity of the consumer. Journal of Adver-
tising Research, 4(4), 40e44.
Hardesty, D. M., & Bearden, W. O. (2003). Consumer evaluations of different pro-
motion types and price presentations: the moderating role of promotional
bene?t level. Journal of Retailing, 79(1), 17e25.
Harlam, B. A., Krishna, A., Lehmann, D. R., & Mela, C. (1995). Impact of bundle type,
price framing and familiarity on purchase intention for the bundle. Journal of
Business Research, 33(1), 57e66.
Haugtvedt, C. P., & Wegener, D. T. (1994). Message order effects in persuasion: an
attitude strength perspective. Journal of Consumer Research, 21(1), 205e218.
Houston, D. A., Sherman, S. J., & Baker, S. M. (1989). The in?uence of unique features
and direction of comparison on preferences. Journal of Experimental Social
Psychology, 25(2), 121e141.
Kahneman, D., & Tversky, A. (1979). Prospect theory: an analysis of decision under
risk. Econometrica, 47(2), 263e291.
Kalyanaram, G., & Winer, R. S. (1995). Empirical generalizations from reference price
research. Marketing Science, 14(3), 161e169.
LeBoeuf, R. A., & Sha?r, E. (2003). Deep thoughts and shallow frames: on the sus-
ceptibility to framing effects. Journal of Behavioral Decision Making, 16(2),
77e92.
Levin, I. P., Gaeth, G. J., Schreiber, J., & Lauriola, M. (2002). A new look at framing
effects: distribution of effect sizes, individual differences, and independence of
types of effects. Organizational Behavior and Human Decision Processes, 88(1),
411e429.
Levin, I. P., Johnson, R. D., & Davis, M. L. (1987). How information frame in?uences
risky decisions: between-subjects and within-subject comparisons. Journal of
Economic Psychology, 8(1), 43e54.
Lowe, B. (2010). Consumer perceptions of extra free product promotions and dis-
counts: the moderating role of perceived performance risk. Journal of Product &
Brand Management, 19(7), 496e503.
H.-H. Liu, Y.-Y. Chiu / Asia Paci?c Management Review 20 (2015) 201e209 208
Monroe, K. B. (1971). Measuring price thresholds by psychophysics and latitudes of
acceptance. Journal of Marketing Research, 8(4), 460e464.
Palazon, M., & Delgado-Ballester, E. (2009). Effectiveness of price discounts and
premium promotions. Psychology & Marketing, 26(12), 1108e1129.
Prelec, D., & Loewenstein, G. (1998). The red and the black: mental accounting of
savings and debt. Marketing Science, 17(1), 4e28.
Sha?r, E., & Thaler, R. H. (2006). Invest now, drink later, spend never: on the mental
accountingof delayedconsumption. Journal of Economic Psychology, 27(5), 694e712.
Sinha, I., & Smith, M. F. (2000). Consumers' perceptions of promotional framing of
price. Psychology & Marketing, 17(3), 257e275.
Smith, M. F., & Sinha, I. (2000). The impact of price and extra product promotions on
store preference. International Journal of Retail & Distribution Management,
28(2), 83e92.
Soman, D., & Gourville, J. T. (2001). Transaction decoupling: how price bundling
affects the decision to consume. Journal of Marketing Research, 38(1), 30e44.
Stokes, M. E., Davis, C. S., & Koch, G. G. (1995). Categorical data analysis using the SAS
system. Cary, NC: SAS Institute Inc.
Thaler, R. H. (1985). Mental accounting and consumer choice. Marketing Science,
4(3), 199e214.
Thaler, R. H. (1999). Mental accounting matters. Journal of Behavioral Decision
Making, 12(3), 183e206.
Tversky, A. (1977). Features of similarity. Psychological Review, 84(4), 327e352.
Wiseman, D. B., & Levin, I. P. (1996). Comparing risky decision making under con-
ditions of real and hypothetical consequences. Organizational Behavior and
Human Decision Process, 66(3), 241e250.
Yi, Y., & Yoo, J. (2011). The long-term effects of sales promotions on brand attitude
across monetary and non-monetary promotions. Psychology & Marketing, 28(9),
879e896.
H.-H. Liu, Y.-Y. Chiu / Asia Paci?c Management Review 20 (2015) 201e209 209
doc_137428966.pdf
This investigation explores the effect of sales framing on the attractiveness of mixed promotions. The
authors propose that the “focused-discount” promotion format (e.g., 50% off the purchase of a second
item) is more attractive to consumers than the “all-discount” promotion format (e.g., 25% off the purchase
of two items) due to the effect that different frames have on the mental accounting processes of
consumers. In Experiments 1 and 2, participants confronted by focused-discount (vs. all-discount)
promotion assigned a relatively higher discount value to the final product and perceived greater savings.
In Experiment 3, priming participants to focus on the first item reduced this framing effect of mixed
promotions. The results show that participants perceived greater savings in a focused-discount framing
(vs. all-discount) due to the differences in how discounts were assigned for each item. Finally, this paper
suggests that companies may frame promotions to increase their relative attractiveness to consumers.
Sales framing, mental accounting, and discount assignments
Hsin-Hsien Liu
a
, Yu-Yeh Chiu
b, *
a
Department of Asia-Paci?c Industrial and Business Management, National University of Kaohsiung, Taiwan, ROC
b
Department of Marketing and Distribution Management, Chien Hsin University of Science and Technology, Taiwan, ROC
a r t i c l e i n f o
Article history:
Received 1 March 2013
Accepted 13 January 2015
Available online 21 July 2015
Keywords:
Framing
Mental accounting
Sales promotions
a b s t r a c t
This investigation explores the effect of sales framing on the attractiveness of mixed promotions. The
authors propose that the “focused-discount” promotion format (e.g., 50% off the purchase of a second
item) is more attractive to consumers than the “all-discount” promotion format (e.g., 25% off the pur-
chase of two items) due to the effect that different frames have on the mental accounting processes of
consumers. In Experiments 1 and 2, participants confronted by focused-discount (vs. all-discount)
promotion assigned a relatively higher discount value to the ?nal product and perceived greater sav-
ings. In Experiment 3, priming participants to focus on the ?rst item reduced this framing effect of mixed
promotions. The results show that participants perceived greater savings in a focused-discount framing
(vs. all-discount) due to the differences in how discounts were assigned for each item. Finally, this paper
suggests that companies may frame promotions to increase their relative attractiveness to consumers.
© 2015 College of Management, National Cheng Kung University. Production and hosting by Elsevier
Taiwan LLC. All rights reserved.
1. Introduction
Marketing managers often use promotions to enhance the
purchasing intentions of consumers. These promotions may focus
on changing the price of target items (e.g., discounts, coupons),
offering extra product (e.g., freebies), increasing the volume of
target items (e.g., buy one get one free, bonus pack), or pursuing a
mixed promotion (e.g., buy two, get 25% off; Chen, Monroe, & Lou,
1998; Diamond, 1992; Sinha & Smith, 2000). Previous research has
shown that the frames used in sales promotions in?uence the value
perceptions of consumers and purchasing intention (Barnes, 1975;
Chen et al., 1998; Diamond, 1992; Diamond & Sanyal, 1990;
Hardesty & Bearden, 2003; Lowe, 2010; Palazon & Delgado-
Ballester, 2009; Sinha & Smith, 2000; Yi & Yoo, 2011). Several au-
thors have also addressed the difference between monetary (e.g.,
discounts) and nonmonetary (e.g., extra products) promotions and
compared the relative attractiveness of different types of pro-
motions (Diamond, 1992; Diamond & Sanyal, 1990; Hardesty &
Bearden, 2003; Sinha & Smith, 2000; Yi & Yoo, 2011). However,
few studies have explored the effects of framing within the mixed
promotion approach or the relative attractiveness of these pro-
motions (e.g., Diamond & Sanyal, 1990; Sinha & Smith, 2000).
Mental accounting theory holds that individual consumers and
households track their ?nancial activities in a manner similar to
enterprises (Thaler, 1985, 1999). Therefore, consumers and house-
holds assign ?nancial events to speci?c accounts and categorize or
label them by expenditures and sources. However, different people
may categorize particular ?nancial events differently (Sha?r &
Thaler, 2006; Soman & Gourville, 2001). This paper proposes that
how a mixed promotion is framed in?uences the accounting pro-
cesses and discount assignments of target consumers. We consider
two framing types: an all-discount framing such as “buy N, get x%
off” and a focused-discount framing, which draws attention to the
last element, as in “buy a second item for y% off.”
The feature-matching model (Tversky, 1977) suggests that in-
dividuals map the features of a subject onto the features of the
referent/target as a starting point for comparison when comparing
two objects. Houston, Sherman, and Baker (1989) further suggest
that when making preference judgments between two items, in-
dividuals tend to use the last item as the subject because the
comparison process does not start until the second object appears.
Therefore, the features of the last object provide the starting point
of the comparison process and are weighted more heavily.
In mental accounting, the use of different semantic cues to
frame the sales promotion (e.g., “buy two, get 25% off” versus “buy a
* Corresponding author. Department of Marketing and Distribution Management,
Chien Hsin University of Science and Technology, No.229, Jianxing Rd., Zhongli Dist.,
Taoyuan City, 32097, Taiwan, ROC.
E-mail address: [email protected] (Y.-Y. Chiu).
Peer review under responsibility of College of Management, National Cheng
Kung University.
HOSTED BY
Contents lists available at ScienceDirect
Asia Paci?c Management Review
j ournal homepage: www. el sevi er. com/ l ocat e/ apmrvhttp://dx.doi.org/10.1016/j.apmrv.2015.01.002
1029-3132/© 2015 College of Management, National Cheng Kung University. Production and hosting by Elsevier Taiwan LLC. All rights reserved.
Asia Paci?c Management Review 20 (2015) 201e209
second item for 50% off”) is known to affect how consumers assign
discounts to each item, even when the ultimate cost of these items
or value of the discounts is the same. Moreover, people tend to
assign greater importance on the last (or immediately preceding)
itemwhen considering mixed promotions because this item acts as
the subject of the comparison. Conversely, the original price rep-
resents a referent (Tversky, 1977) that consumers use as a com-
parison against the last item in a mixed promotion, which is most
salient. Therefore, a focused discount promotion may be presumed
to be more attractive than an all-discount promotion because the
last item in the former frame appears to offer a greater discount.
This paper conducts three experiments to con?rm these pre-
dictions. This study offers several contributions to the existing
literature: (1) a new and more comprehensive perspective for
explaining the relative attractiveness of a mixed promotion across
different framings is suggested; (2) the differences in mental ac-
counting processes and discount assignments between the two
framings are explored; (3) the moderating in?uence of focus
priming on the relative attractiveness of different frames is
demonstrated; and (4) key implications for practice and avenues
for further research are introduced.
2. Theoretical background
2.1. Framing effect of promotions
Retailers often vary the amount and level of discounting and the
way in which promotions are framed in order to in?uence con-
sumer perceptions of transaction value and purchase intentions.
Previous research con?rms that how alternatives are framed in-
?uences the relative preference judgments and choices of con-
sumers (Kahneman & Tversky, 1979; Sinha & Smith, 2000).
This framing effect is well demonstrated in research into mar-
keting promotions. For example, Barnes (1975) found that, when
faced with the three price promotion formats of “special - $75”;
“25% off - $75”; and “regular price $100, sale price $75”, consumers
responded most positively to the latter version. Furthermore, Chen
et al. (1998) demonstrated that a price promotion framed as a dollar
reduction is more attractive in the context of high-price products,
whereas one framed as a percentage reduction is more attractive
for low-price products. Some researchers have employed
assimilation-contrast theory (Gabor & Granger, 1964; Kalyanaram
& Winer, 1995; Monroe, 1971) and Weber's law (Harlam, Krishna,
Lehmann, & Mela, 1995) to explain the effect of price framing. Ac-
cording to the assimilation-contrast theory, there is a price range
that is internal to consumers, known as latitude of acceptance,
which de?nes the range of acceptable prices. A reduced price that is
within a consumer's latitude of acceptance is assimilated into the
acceptable range and becomes acceptable. A reduced price that is
outside the latitude of acceptable will be contrasted with the
acceptable range and becomes noticeable. This concept relates
closely to Weber's law (Monroe, 1971). Weber's law suggests that
the difference threshold (or “just noticeable difference”), which
describes the minimum difference in a stimulus that is detectible
by an individual. A larger change in price is more likely to be
detected or perceived by consumers. Thus, the focus-discount
frame (e.g., “get the second one 50% off”) is more noticeable than
the all-discount frame (e.g., “buy two, get 25% off.”) and more likely
to raise purchase intention (Harlam et al., 1995).
In contrast, other authors have turned to the silver lining prin-
ciple of mental accounting theory. Based in the prospect theory, the
silver lining principle proposes that utility functions are concave for
gains and convex for losses and that the curve for losses is steeper
than that for gains. When consumers confront a small gain (e.g.,
price discount) relative to a large loss (e.g., purchasing cost), they
prefer to separate them psychologically into two different accounts
rather than integrate them. Diamond and Campbell (1989) and
Diamond and Sanyal (1990) also proposed that nonmonetary pro-
motions (e.g., freebies, “buy one, get one free”) appear segregated
from the purchasing cost and framed as gains because the units for
the promotions differ from the price (i.e., money). In contrast,
monetary promotions (e.g., price discounts) should be integrated
into the purchasing cost and framed as reduced losses because the
units are the same, which facilitates their integration. Consumers
then have a preference for extra-product or volume promotions
over price discounts, assuming that the size of the promotion is
small relative to the purchase cost. Sinha and Smith (2000) and
Smith and Sinha (2000) show that consumers prefer stores that
offer extra-product promotions (e.g., buy one, get one free) to stores
offering mixed promotions (e.g., buy two, get 50% off), especially if
the products being offered may be hoarded. However, the silver
lining principle also requires that the amount of the gain should be
relatively small compared with the losses. In Sinha and Smith's
studies, the percentage of discount relative to the cost may be too
large (50%). Further adding to the confusion, it seems that, in
comparison with assimilation-contrast theory and Weber's law, the
silver lining principle suggests that a different discount level is
necessary to exert the framing effects.
We propose another explanation for the variation in attrac-
tiveness of mixed promotions across different frames: Promotion
frames may in?uence consumers' mental accounting processes and
assignments of discounts to each product item such that their
perceptions of their savings differ for each product item in the
various framing conditions.
2.2. Mental accounting
Mental accounting is a cognitive rule that consumers use to
organize, evaluate, and record ?nancial activities. Consumers keep
track of their ?nancial activities, with accounting rules in?uencing
their ?nancial decisions explicitly and implicitly. Mental account-
ing assumes that people value a single event in multiple ways
because they have the ?exibility to categorize. For example, Sha?r
and Thaler (2006) proposed that individuals may evaluate a
transaction in multiple ways if the time of consumption is
different than the time of purchase. An example of this is when
theatergoers purchase tickets months in advance of a perfor-
mance. People tend to regard their early purchases as an invest-
ment rather than an expenditure, and then by the time they
consume the item, they have forgotten the price, so the product
seems free or a savings (e.g., spending $20 for a bottle of wine that
appreciates over time). If they lose a previously purchased product
(e.g., breaking the bottle of wine), the replacement costs appear
more salient.
Soman and Gourville (2001) demonstrated that consumers may
?exibly assign the cost across each item when the connection be-
tween costs and bene?ts is ambiguous, as with price bundling (e.g.,
$1000 for a travel package that includes transportation, hotel, and
food). Similarly, Prelec and Loewenstein (1998) found that paying
with a credit card (versus cash) diminishes the salience of the
payments and reduces the connection between costs and bene?ts.
Together, these studies show that the characteristics of a
transaction affect how consumers assign the cost of that trans-
action. If the discount assigned to each product item varies across
the frames used, does the speci?c classi?cation in?uence con-
sumers' perceptions of savings? The frames of a mixed promotion
(all-discount versus focused-discount) may cause consumers to
assign discounts to items in varying ways, and the relative attrac-
tiveness of the different frames may depend on the focus that
consumers put on the last item.
H.-H. Liu, Y.-Y. Chiu / Asia Paci?c Management Review 20 (2015) 201e209 202
2.3. Feature matching model
In comparing two objects, the subject (i.e., starting point in the
comparison process) is mapped onto the referent (i.e., target of the
comparison process). Thus, the characteristics of the subject are the
checklist used to evaluate the referent. The referent is usually the
more prominent of the two items. For example, most people will
consider the similarity of their friend (subject) to Marilyn Monroe
(referent) rather than the other way around because Marilyn
Monroe is the more famous and prominent of the two (Tversky,
1977). Therefore, Houston et al. (1989) suggested that in making
preference judgments, consumers use the second item as the
subject of their comparison because the comparison process cannot
start until this second item appears.
Based on the feature-matching model and the ?ndings of
Houston et al. (1989), we predict that people focus more on the last
itemwhen considering mixed promotions because that item serves
as the subject of the comparison. Moreover, we predict that con-
sumers use a regular, prominent price as the referent. Therefore,
consumers exposed to a focused-discount framing should perceive
more savings than consumers exposed to an all-discount framing
because the last itemin a focused discount frame accounts for more
of the discount than the last item in the all-discount framing.
Imagine that Mr. M and Mr. Z both enjoy two cups of coffee
(US$2/each) every morning. This morning, Mr. Z notices his coffee
shop has a “buy two, get 25% off” promotion (i.e., all-discount
framing) and takes two cups to the of?ce. Mr. M notices that his
coffee shop has a “50% off a second cup” promotion (i.e., focused-
discount framing) and also buys two cups. Although both Mr. M
and Mr. Z pay US$3 for their two cups of coffee, they may not
perceive the same savings. They both focus on the second cup of
coffee and compare its price with the original price (US$2).
Therefore, Mr. Z likely perceives that he has saved 25%, whereas Mr.
Mperceives that he has saved 50%. As a result, the focused-discount
framing should be more attractive than the all-discount framing:
H 1. Consumers exposed to focused-discount framing perceive
more savings than consumers exposed to all-discount framing.
Different framing-related discount assignments for each prod-
uct itemshould also result in consumers perceiving pain differently
in the event of a loss. In the coffee purchase example, Mr. M in the
focused-discount framing condition perceives a 50% savings on the
second item but no discount for the ?rst item whereas Mr. Z in the
all-discount framing condition perceives a 25% discount on both
items. Accordingly, if they each drop one of their cups on the way to
the of?ce, Mr. Z should feel more pain than Mr. M. Therefore, we
propose:
H 2. When faced with the loss of the ?nal item, consumers in the
all-discount framing condition feel more pain than those in the
focused-discount framing.
H 3. Participants in the all-discount framing condition divide the
discount equally across each item, whereas participants in the
focused-discount framing condition assign the total discount to the
?nal item.
3. Experiment 1: promotional frames and discount
allocations
3.1. Experiment 1A
To test Hypotheses 1e3, we examined whether consumers
exposed to a focused-discount framing perceive more savings than
their all-discount framing peers, as well as whether different
perceptions emerge because consumers focus on the last item and
assign the discounts differently across the two framing conditions.
3.1.1. Participants and procedures
Ninety-eight Taiwanese college students participated in the
experiment as a regular assignment and without the incentive of
additional rewards or credit. As the purpose of the experiment was
to explore the perception processes of consumers, there was no
“right” or “wrong” answer. The participants were asked to answer a
two-page questionnaire that included a judgment task, described
as follows:
Decision scenario: Every morning, Mr. M and Mr. Z both spend
NT$120 to purchase two cups of coffee (unit price NT$60), after
which they take their cups to the of?ce. This morning, when Mr. M
was buying the coffee in the store, he noticed that the coffee shop
was holding a promotion: “buy one cup of coffee, get a second for
half price.” He accepted the promotion and paid for two cups. At the
same time, Mr. Z noticed that his coffee shop was holding a pro-
motion: “25% off the purchase of two cups of coffee.” He also
accepted the promotion and paid for two cups.
After reading the description, participants answered the
following question: “Although Mr. M and Mr. Z both paid NT$90 for
two cups of coffee, who do you think perceived the greater sav-
ings?” Participants chose among three options (Mr. M, the same, or
Mr. Z).
Next, they were asked to imagine that “after Mr. Mand Mr. Z had
?nished their ?rst cups of coffee, they both spilled their second cup
carelessly.” The participants were then asked: “Who do you think
felt greater pain over his loss?” Participants again chose among the
same three-option sets (Mr. M, the same, or Mr. Z). Finally, the
participants were asked to imagine which one of ?ve statements
best captured Mr. M's and Mr. Z's feeling about his loss: (a) a cup of
coffee costs NT$60, so I lost NT$60; (b) I spent NT$90 for two cups
of coffee, so I lost NT$45; (c) I have ?nished a cup of coffee that costs
NT$60, so I lost NT$30; (d) I paid for the coffee this morning, so my
loss costs me nothing; or (e) other (participants checking this item
were asked to describe their feeling). After they ?nished the
questionnaire, the participants were thanked and debriefed.
3.1.2. Results
For the ?rst question, most participants (56.1%) believed that Mr.
M and Mr. Z would perceive the same savings. Among those who
believed savings would be perceived differently (43/98), 30 par-
ticipants believed that Mr. M would perceive a greater savings than
Mr. Z and only 13 participants believed Mr. Z would perceive a
greater savings. After factoring out the group that believed that Mr.
M and Mr. Z would perceive the same savings, the difference was
statistically signi?cant (c
2
(1; N ¼ 43) ¼ 6.72, p < .05) in support of
Hypothesis 1. This result is not an artifact of the inability of par-
ticipants to calculate the costs correctly, as they were aware that
Mr. M and Mr. Z had paid the same amount.
For the second question, 45.9% of the participants (45/98)
indicated that Mr. M and Mr. Z would perceive the same degree of
pain in spilling their second cup of coffee; 10.2% believed that Mr. M
would feel more pain; and 43.9% of the participants believed that
Mr. Z would feel more pain (c
2
(2; N¼98) ¼23.65, p < .05). In other
words, among those participants who believed Mr. M and Mr. Z
would feel differing amounts of pain over their loss, a signi?cantly
larger number believed that Mr. Z would feel more pain than Mr. M
(c
2
(1; N ¼ 53) ¼ 20.55, p < .05). This result supports Hypothesis 2.
For the third question, while 41.8% of participants believed that
Mr. M lost NT$30, only 13.3% of participants believed that Mr. Z had
lost NT$30. In contrast, 63.3% of participants believed that Mr. Z lost
NT$45, with only 32.7% of participants believing the same of Mr. M
(Table 1). These results match our expectations that the all-
H.-H. Liu, Y.-Y. Chiu / Asia Paci?c Management Review 20 (2015) 201e209 203
discount framing scenario encourages consumers to divide the
discount equally across two items and that the focused-discount
framing scenario encourages consumers to assign the entire dis-
counted value to the second item.
This analysis applied CATMOD modeling to the repeated-
measures categorical data (Stokes, Davis, & Koch, 1995), with par-
ticipants' thoughts about Mr. M's and Mr. Z's losses as the depen-
dent variables and the promotion frames as the independent
variables. The difference was signi?cant: c
2
(4; N ¼ 98) ¼ 30.93,
p < .01. Ignoring those participants who chose options (a), (d), or (e)
for either Mr. M or Mr. Z and focusing only on those who chose
options (b) and (c), we again obtained signi?cant results, with c
2
(1;
N ¼ 64) ¼ 24.64, and p < .01, in support of Hypothesis 3.
3.1.3. Discussion
The ?ndings of Experiment 1A support the proposition that
consumers prefer mixed promotions with a focused-discount
framing rather than an all-discount framing. This difference
mainly re?ects the tendency of individuals to assign discounts
differently to each product item and focus on the last item of the
mixed promotions. Those exposed to the focused-discount framing
thus perceive more savings, because the last item in their frame
accounts for more of the discount. Furthermore, participants
believed that a consumer who loses this last item perceives less
pain than a consumer in the all-discount framing scenario.
According to Sha?r and Thaler (2006), an actual loss event
causes the replacement cost to become more salient. However, the
replacement cost effect was not very salient in Experiment 1A (i.e.,
only 18% participants chose option (a)). Perhaps these participants
assigned the discount to the ?rst and second items, whereas par-
ticipants did not need to assign discounts separately in Sha?r and
Thaler's study. This distinction may explain why the effect of
replacement cost is so weak in the present study.
3.2. Experiment 1B
Although the results in Experiments 1A are consistent with our
propositions, the design used for Experiment 1A may have been
overly arti?cial as few people buy two cups of coffee at once. To
overcome this potentially confounding in?uence, we adopted
laundry detergent as the target product for Experiment 1B. Most
consumers use laundry detergent over longer periods of time and
are able to store bulk amounts. In addition, the participants in
Experiment 1B were not asked to indicate which options better
capture Mr. M's and Mr. Z's losses but rather were asked to write
down their perceptions of the loss.
3.2.1. Participants and procedures
Ninety-six Taiwanese college students participated in the
experiment as a regular assignment and without the incentive of
additional rewards or credit. Participants responded to a two-page
questionnaire that included three judgment tasks. The design was
similar to Experiment 1A, as follows:
Decision scenario: M and Z are college students. Both regularly
wash their clothes using X brand laundry detergent. Each time they
use up all their laundry detergent, they go to the supermarket and
spend NT$120 for a box of X brand laundry detergent (2 kg). This
morning, M and Z both noticed they were out of laundry detergent,
so they went to nearby supermarkets. As M was shopping, he
noticed that the supermarket was running a promotion: “Get a
second box of X brand laundry detergent for half price.” Because he
had room to store the laundry detergent and had used X brand to
wash his clothes, M accepted the promotion and paid NT$180 for
two boxes. At the same time, Z noticed that his supermarket was
also running a promotion: “Buy two boxes of X brand laundry
detergent, get 25% off.” He also had room to store the extra deter-
gent, accepted the promotion, and paid NT$180 for two boxes.”
After reading the scenario, participants ?rst completed a judg-
ment task in which they were asked which buyer they thought
would perceive more savings (M, the same, or Z). Next, they were
asked to imagine that “After M and Z used up their ?rst box of
laundry detergent, they opened the second box. However, they
both noticed that their laundry detergent had hardened and
couldn't be used any more.” Participants then answered the
following question: “Which one do you think would feel more
pain?” Participants again chose among the same three-option sets.
Finally, the participants were asked to answer the question: “When
M (Z) noticed his laundry detergent was useless, how much money
do you think M (Z) lost?”
3.2.2. Results
Even with a different target product, the results were similar to
those from Experiment 1A. In the ?rst task, most participants
(70.8%) believed that Mand Z would perceive the same savings, but
among those who did not (28/96), 23 participants believed that Mr.
M would perceive more savings than Z and only 5 participants
believed that Z would perceive more savings. After factoring out
those who believed the two would perceive the same savings, the
difference was statistically signi?cant, c
2
(1; N ¼28) ¼11.57, p < .01,
in support of Hypothesis 1.
In the subsequent task, 60.4% of participants (58/96) indicated
that Mand Z would feel the same degree of pain over the loss of the
second box of laundry detergent; 8.3% thought that M would feel
more pain, and 31.3% thought that Z would suffer more, c
2
(2;
N ¼ 96) ¼ 39.25, p < .01. Thus, most of the participants who felt the
two would feel differing amounts of pain believed that it would be
Z who would feel the most pain, c
2
(1; N ¼ 38) ¼ 12.74, p < .01.
These results supported Hypothesis 2 as well.
Finally, the average perceived loss was NT$66.40 for M and
NT$76.19 for Z. A new variable, “DiffCost,” which is equal to the
within-subject difference in the perceived loss for M and Z, differed
signi?cantly from 0 (t(N ¼ 96) ¼ 5.07), p < .05. Thus, participants
perceived the loss for Z to be greater than the loss for M, in support
of Hypothesis 2.
To explore the discount assignments of participants across the
two frames, we classi?ed the perceived loss for M and Z into ?ve
groups based on the differences between the perceived loss and
Table 1
Participant perceptions of Mr. M's and Mr. Z's losses, Experiment 1A.
Mr. M (%) Mr. Z (%)
(a) A cup of coffee costs NT$60, so I lost NT$60. 17.4 18.4
(b) I spent NT$90 for two cups of coffee, so I lost NT$45. 32.7 63.3
(c) I have ?nished a cup of coffee that costs NT$60, so I lost NT$30. 41.8 13.3
(d) I paid for the coffee this morning, so my loss costs me nothing. 2.0 1.0
(e) Other. 6.1 4.0
Notes: The number in each cell represents the relative share for that opinion.
H.-H. Liu, Y.-Y. Chiu / Asia Paci?c Management Review 20 (2015) 201e209 204
two reference prices, NT$60 and NT$90. The outcomes were labeled
“< 60” if the stated perceived loss was less than NT$60; “ ¼ 60” if it
equaled NT$60; “60e90” if it was greater than NT$60 but less than
NT$90; “ ¼ 90” if the stated perceived loss equaled NT$90; and “>
90” if it was greater than NT$90.
As shown in Table 2, 59.4% of the participants believed that M
had lost NT$60, while only 27.1% believed the same of Z. In contrast,
50% of the participants believed that Z had lost NT$90, while only
22.9% believed the same of M. These results again match our ex-
pectations: Participants in the all-discount framing scenario divide
the discount equally across the two items, whereas those in the
focused-discount framing scenario assign the entire discount to the
second item.
We again applied CATMOD modeling to the repeated-measures
categorical data, with participants' thoughts about M's and Z's
losses as the dependent variables and the promotion frames as the
independent variables. We ignored those participants who chose
the more extreme options for either M or Z because the perceived
losses equal to NT$60 and NT$90 provide the primary means to
examine Hypothesis 3. We again found signi?cant results, with
c
2
(1; N ¼ 82) ¼ 39.33, and p < .01.
3.2.3. Discussion
The results of this study were consistent across the two target
products used. Although the perceived losses for M and Z in
Experiment 1B were based on the subjective statements of partic-
ipants rather than on a choice from among ?ve options as in
Experiment 1A, the results still support Hypothesis 3.
In Experiment 1, statistical analyses were conducted after
excluding the data that indicated no difference between Mand Z, as
the within-subject design format poses the risk of reducing the
framing effect (Levin, Johnson, & Davis, 1987; Levin, Gaeth,
Schreiber, & Lauriola, 2002). A framing effect that exists in a
between-subject design vanishes in a within-subject design (Levin
et al., 1987). Similarly, the goal-framing effect decreases more
signi?cantly in the within-subject design than in the between-
subject design (Levin et al., 2002). To control for individual vari-
ance and to address consumer preference, a within-subject design
may be a better choice, even if participants may desire to respond in
a consistent manner (Wiseman & Levin, 1996). Authors who have
adopted within-subject designs typically summarize the results
after excluding data that indicate no difference between frames or
formats (e.g., LeBoeuf & Sha?r, 2003). Experiment 2 adopts a
between-subject design that obviates the need to either support or
exclude alternative explanations.
4. Experiment 2: framing of sales promotions and discount
allocations (between-subjects)
Although the results of Experiment 1 support the hypotheses,
we conducted Experiment 2 with a between-subjects design to
exclude the alternative explanations for these results. Additionally,
we used a different judgment scenario (concert ticket) and a
different discount level (e.g., buy three, get one free) to enhance the
generalizability of our ?ndings. The price of the target (concert
ticket) in this scenario was much higher than those in Experiment 1
(coffee, laundry detergent). Support for our hypotheses in the
context of high priced targets would further bolster con?dence that
these hypotheses apply not only to consumer but also to higher-
priced retail goods.
4.1. Experimental design
Experiment 2 uses a between-subjects design: The all-discount
frame indicates “buy four, get 25% off,” and the focused-discount
frame notes, “buy three, get one free.” The scenario, modi?ed
from Sha?r and Thaler (2006), is as follows:
Decision scenario (All-discount/focused-discount): Mr. Chang
spends NT$2000 monthly on buying concert tickets. Today, when
he went to buy his concert ticket for the following week, he noticed
that the concert hall was conducting a “buy four tickets, get 25% off
[buy three tickets, get one free]” promotion. The four tickets could
be used separately over six months. Although the regular price for
four tickets would be NT$8000, Mr. Chang only spends NT$6000 to
buy them. He plans to use one ticket each month.”
4.2. Participants and procedure
One hundred and ?fty-two Taiwanese college students partici-
pated in the experiment as a regular assignment and without the
incentive of additional rewards or extra credit. Participants were
randomly assigned to either the focused discount (N ¼ 78) or all-
discount (N ¼ 74) group. After reading the scenario, all partici-
pants answered three questions. The ?rst question “Which of the
following descriptions do you think best captures Mr. Chang's
feeling when he spent NT$6000 for the four tickets?” had the
following four potential answers: (a) The NT$6000 spent to buy
four tickets is no different fromspending the same amount on other
retail items; (b) I made a NT$6000 investment in four tickets, which
I will gradually consume in the future; (c) I spent NT$6000 to buy
four tickets that normally cost NT$8000, so I just saved NT$2000;
and (d) other (please explain). The second question “After months
when Mr. Chang is ready to use his ?nal ticket, which of the
following descriptions do you think best capture his feeling about
the cost of this ?nal ticket?” had the following four potential an-
swers: (a) This ticket cost NT$2000, the same as the current price of
a single ticket; (b) This ticket cost NT$1500 because I spent a total of
NT$6000 for the four tickets; (c) This ticket is free, because the
NT$6000 I paid covered the three previous tickets; and (d) other
(please explain). The third question “Imagine that when Mr. Chang
is getting ready for the fourth concert, he found that he had lost his
ticket and had no way of purchasing another as a replacement.
Which of the following descriptions do you think best captures Mr.
Chang's feeling about losing the ?nal ticket?” prompted the par-
ticipants to choose one of the four options posited for the second
question.
4.3. Results
As shown in Table 3, most participants (about 90%) believed that
Mr. Chang either saved NT$2000 or made an investment when he
paid NT$6000 for the four concert tickets regardless of framing
condition (c
2
(3; N ¼ 152) ¼ .86, p > .1). This ?nding is similar to
Sha?r and Thaler (2006).
When considering the use of the ?nal ticket, 46.2% of partici-
pants in the focused-discount framing group believed that the last
ticket was free, while only 31.1% of those in the all-discount framing
group believed the same. In contrast, 52.7% of the participants in
the all-discount framing group believed that the cost of the last
ticket was NT$1500 (i.e., 25% off), while only 41.0% of those in the
focused-discount framing group believed the same. If we ignore the
Table 2
Participants' feelings about M's and Z's losses, Experiment 1B.
90
M (%) 11.5 59.4 0 22.9 6.2
Z (%) 10.4 27.1 2.1 50 10.4
Notes: The number in each cell represents the relative share of that thought.
H.-H. Liu, Y.-Y. Chiu / Asia Paci?c Management Review 20 (2015) 201e209 205
data from those who chose options (a) or (d), the result is still
marginally signi?cant, c
2
(1; N ¼ 130) ¼ 3.28, p ¼ .07.
With regard to the lost ticket scenario, 48.6% of the participants
in the all-discount framing group considered the lost ticket equal to
a loss of NT$1500, while only 29.5% of those in the focused-discount
framing group believed the same. In contrast, 46.2% of the focused-
discount framing group believed that Mr. Chang had lost NT$0,
while only 29.7% of those in the all-discount framing group
believed the same. If we ignore the data from those who chose
options (a) and (d), the difference is statistically signi?cant, c
2
(1;
N ¼ 117) ¼ 6.24, and p ¼ .01.
Thus, the present study found overall support for Hypothesis 3:
Participants in the focused-discount framing group are more likely
to assign the entire discount to the ?nal ticket and regard it as free,
while participants in the all-discount group tend to assign the
discount equally to each ticket.
4.4. Discussion
Hypothesis 3 continues to be supported even though Experi-
ment 2 involves a different experimental design (between-sub-
jects), a different judgment scenario (concert ticket), and a different
discount level (e.g., buy three, get one free). We thus consider our
proposition generalizable across conditions.
At the moment that the four tickets were purchased, the par-
ticipants differed little in their perceptions of the purchase. As in
Sha?r and Thaler's (2006) study, most participants regarded the
payment as an investment or form of savings. However, in
considering the ?nal ticket under the different framing conditions,
the participants held varying cost perceptions for that ticket when
it was both used and lost.
As previously noted, the loss of the ?nal ticket did not make the
replacement cost effect very salient (i.e., 21e22% of participants
chose option (a)). This may imply that the assigned discount
weakens the effect of the replacement cost, although the loss of the
?nal ticket increases the effect of the replacement cost slightly in
comparison with its use (from 12e14% to 21e22%), which is
consistent with the result posited by Sha?r and Thaler (2006).
5. Experiment 3: the moderator of framing effects
Combining mental accounting and feature matching (Tversky,
1977) may support an inference that consumers will prefer
focused-discount framing to all-discount framing. However, might
the framing effects decrease or even reverse if consumers are
primed to focus on the ?rst rather than the ?nal item? In other
words, while we have established that consumers assign a greater
discount to the last item in the focused-discount framing, the
framing effect of mixed promotions may decline or even reverse if
consumers place greater focus on the ?rst item.
H 4. The framing effect of a mixed promotion is moderated by the
focus priming of consumers, with a focus on the ?rst item resulting
in a declining or even reverse effect rather than a focus on the
second item.
5.1. Experimental design and procedures
Experiment 3 was conducted similarly to Experiment 1A, except
that participants were prompted to focus on either the ?rst or
second (?nal) product item. Thus, 195 Taiwanese college students
participated in the experiment as a regular assignment and without
the incentive of additional rewards or extra credit. These partici-
pants were randomly assigned to one of two priming conditions
(?rst versus second item) and responded to a two-page question-
naire with three judgment tasks. The ?rst task was the same as the
?rst task in Experiment 1A. For the second task, participants were
primed to focus on the ?rst or second item by imagining either that
“As Mr. M and Mr. Z are enjoying their ?rst cups of coffee, they each
spill it carelessly” (N ¼ 99) or that “After Mr. M and Mr. Z have
enjoyed the ?rst cup of coffee, they start to drink the second cup of
coffee and spill it carelessly” (N ¼ 96). The remaining tasks were
similar to the ?nal two items in Experiment 1A.
5.2. Results
The results for the ?rst task were similar those from Experiment
1A: Most (61.5%) participants believed that Mr. M and Mr. Z would
perceive the same savings; 28.7% believed that Mr. M would
perceive a greater savings; and only 9.7% believed that Mr. Z would
perceive a greater savings, c
2
(2; N ¼ 195) ¼ 7.13, p ¼ .03. In
analyzing the data of only those participants who believed Mr. M
and Mr. Z would perceive a difference in savings, we identi?ed a
statistically signi?cant result, c
2
(1; N ¼ 75) ¼ 18.25, p < .01.
When participants focus on the spilled second cup, 57.3%
believed that Mr. Mand Mr. Z would suffer the same perceived loss,
29.2% believed that Mr. Z would suffer more, and only 13.5%
believed that Mr. M would suffer more. This result is similar to that
of Experiment 1A. However, when we primed participants to focus
on the ?rst spilled cup, the results reversed in that 51.5% believed
that Mr. M and Mr. Z would suffer the same perceived loss, 29.3%
believed that Mr. Mwould suffer more, and only 19.2% believed that
Mr. Z would suffer more, c
2
(2; N ¼ 195) ¼ 7.93, p ¼ .02. After
factoring out those participants who perceived the same pain
Table 3
Results for three tasks in Experiment 2.
Buy 3, get 1 free Buy 4, get 25% off
Task 1: At the time you make the purchase:
(a) The NT$6000 spent to buy four tickets is no different from spending the same amount on other retail items. 9.0 10.8
(b) I made an NT$6000 investment in four tickets, which I will gradually consume in the future. 32.1 27.0
(c) I spent NT$6000 to buy four tickets that normally cost NT$8000, so I just saved NT$2000. 57.7 59.5
(d) Other. 1.2 2.7
Task 2: The perceived cost of the ?nal ticket:
(a) This ticket cost NT$2000, the same as the current price of a single ticket. 11.5 13.5
(b) This ticket cost NT$1500 because I spent a total of NT$6000 for the four tickets. 41.0 52.7
(c) This ticket is free, because the NT$6000 I paid covered the three previous tickets. 46.2 31.1
(d) Other. 1.3 2.7
Task 3: The perceived cost of losing the ?nal ticket:
(a) This ticket cost NT$2000, the same as the current price of a single ticket. 20.5 21.7
(b) This ticket cost NT$1500 because I spent a total of NT$6000 for the four tickets. 29.5 48.6
(c) This ticket is free, because the NT$6000 I paid covered the three previous tickets. 46.2 29.7
(d) Other. 3.8 0
Notes. The number in each cell represents the relative percentage (%) for that thought.
H.-H. Liu, Y.-Y. Chiu / Asia Paci?c Management Review 20 (2015) 201e209 206
across both priming conditions, the result was even more pro-
nounced in favor of our hypothesis, c
2
(1; N ¼ 89) ¼ 7.31, p < .01.
Furthermore, when participants considered Mr. M's and Mr. Z's
losses separately (see Table 4), they differed signi?cantly between
the two priming conditions. When focusing on the second cup,
45.8% believed that Mr. M lost NT$30, while only 14.6% believed
that Mr. Z lost NT$30. In contrast, 54.2% believed that Mr. Z lost
NT$45, while only 28.1% believed that Mr. M lost NT$45. These
?ndings are similar to those from Experiment 1A and support
Hypothesis 3. For the participants that focused on the ?rst cup,
70.7% believed that Mr. Z lost NT$45, 41.4% believed that Mr. M lost
NT$45, 24.2% believed that Mr. M's loss equaled NT$60, and only
8.1% believed that Mr. Z's loss equaled NT$60.
Thus a focus on the ?rst rather than the second cup caused most
participants to imagine a higher perceived cost for Mr. M. For
example, 24.2% and 41.4% of the participants who focused on the
?rst cup thought Mr. M had lost NT$60 and NT$45, respectively,
whereas only 10.4% and 28.1% of this group assigned these amounts
when they focused on the second cup. In contrast, 24.2% of the ?rst-
cup-focused participants believed that Mr. M lost NT$30, while
45.8% of the second-cup-focused participants believed the same.
This ?nding supports the proposition that the focused-discount
framing scenario encourages consumers to assign the entire dis-
count to the last item.
The highest costs perceived with regard to Mr. Z do not differ
signi?cantly between the two priming conditions. Although the
ratio of those choosing NT$45 increases from 54.2% for the second-
cup-focused group to 70.7% for the ?rst-cup-focused group, the
ratio of those choosing NT$60 and NT$30 both decrease when the
focus is placed on the ?rst cup.
Regardless of which cup spills, most participants (70.7% ?rst cup
and 54.2% second cup) indicate that Mr. Z has lost NT$45. These
results suggest that participants distributed the discount equally
across the ?rst and second cups when using an all-discount
framing, in support of Hypothesis 3.
Finally, ignoring those participants who chose options (d) or (e)
for either Mr. M or Mr. Z and again applying CATMOD, with Mr. M's
and Mr. Z's presumed losses as the dependent variables and pro-
motion framing, participant focus (?rst versus second cup), and
their interaction as independent variables, we discover that pro-
motion framing (c
2
(2; N ¼ 167) ¼ 40.33, p < .01), participant focus
(c
2
(2; N ¼ 167) ¼ 10.95, p < .01), and their interaction (c
2
(2;
N¼167) ¼4.58, p < .01) are all signi?cant. In support of Hypothesis
4, the framing effect of promotions thus reverses when participants
focus on the ?rst item.
5.3. Discussion
Experiment 3 reveals the moderating effect of focus priming on
the framing effect of promotions. The results con?rm our propo-
sition, as the framing effect in?uences consumers' assignment of
discounts across items, with a greater focus on the ?nal item,
consumers perceive greater savings in the focused-discount
framing scenario than in the all-discount framing scenario. How-
ever, priming participants to focus on the ?rst item reduced the
framing effect.
While it may seem contradictory that participants perceived Mr.
M's loss as NT$45 rather than NT$60 for a ?rst cup of spilled coffee,
this result dovetails with the observations of Sha?r and Thaler
(2006) that consumers regard an advance purchase as an invest-
ment that they amortize over time. Therefore, many participants
did not consider Mr. M's loss as NT$60 but rather as NT$45 because
they had already assigned a portion of the discount to the ?rst item.
6. Conclusions and general discussion
Our three experiments support our overall proposition that
consumers perceive more savings when presented as a focused-
discount framing than as an all-discount framing. The framing ef-
fects of the mixed promotion result mainly from differential as-
signments of the discounts to each item in the different promotion
frames. Furthermore, we con?rmed the framing effect for a variety
of product types, including those that consumers buy in advance
and store (Experiments 1B and 2) and those that are bought and
consumed nearly simultaneously (Experiments 1A and 3).
Furthermore, the effect was con?rmed across different designs
(between-versus within-subjects) and promotion levels (buy 2, get
25% off versus buy 4, get 25% off). Finally, we demonstrated that
priming the focus of participants moderates the framing effect of
mixed promotions, with ?ndings robust across decision scenarios.
6.1. Comparison with previous studies of promotional framing
Sinha and Smith (2000) and Smith and Sinha (2000) suggested
that consumers prefer the “buy one, get one free” format over the
“buy two, get 50% off” format because the latter implies that con-
sumers must buy two items to get the discount, whereas the former
promises that they receive the discount by buying only one item.
However, their proposition does not explain the different prefer-
ences in Experiments 1 and 3 because our participants were
required to buy two items in both framing conditions. Moreover,
their proposition does not explicate the moderating effect of con-
sumer focus on the framing effect of the mixed promotion
(Experiment 3). In contrast, the results of the present study helps
explain the results of Sinha and Smith because consumers con-
fronted with a “buy two, get 50% off” deal tend naturally to focus on
the last item and perceive a half-off discount on the second item
purchased. In contrast, the second item in the “buy one, get one
free” deal is perceived as free (100% discount), which is much more
attractive.
Some researchers have employed the assimilation-contrast
theory and Weber's law (Harlam et al., 1995) to explain why
Table 4
Participants' thoughts about loss in Experiment 3.
Mr. M Mr. Z
First cup (N ¼ 99) Second cup (N ¼ 96) First cup (N ¼ 99) Second cup (N ¼ 96)
(a) A cup of coffee costs NT$60, so I lost NT$60. 24.2 10.4 8.1 15.6
(b) I spent NT$90 for two cups of coffee, so I lost NT$45. 41.4 28.1 70.7 54.2
(c) I have ?nished a cup of coffee that costs NT$60, so I lost NT$30 (second item focus).
[I still have a cup of coffee that costs NT$60, so I lost NT$30 (?rst item focus).]
24.2 45.8 12.1 14.6
(d) I paid for the coffee this morning, so my loss costs me nothing. 5.1 4.2 4.0 6.3
(e) Other. 5.1 11.5 5.1 9.3
Notes. The number in each cell represents the relative percentage of that thought. The wording of option (c) differed slightly across the two priming conditions, depending on
the focus priming scenarios.
H.-H. Liu, Y.-Y. Chiu / Asia Paci?c Management Review 20 (2015) 201e209 207
consumers prefer to “get the second one 50% off” instead of “buy
two, get 25% off.” They infer a range of acceptable prices that
consumers use to judge a price, such that a larger price change is
more likely to be outside the acceptable range and should be more
noticeable. However, their proposition does not account for the
?nding that priming participants to focus on the ?rst item reduces
the framing effect.
Diamond and Sanyal (1990) adopted Thaler's (1985) silver lining
principle to explain why an extra product (e.g., buy one, get one
free) is preferable to a mixed (buy two, get 50% off) promotion,
positing that the former appears as a gain and the latter is treated as
a reduced loss. However, as before, the silver lining principle does
not explain the results obtained by Experiment 3.
Distinguishing this study fromprevious research is important in
terms of the potential impact of the present study on future hy-
potheses. For example, the silver lining principle would likely
anticipate that the framing effects in relevant hypotheses to decline
or even reverse in line with promotional levels due to its reliance on
comparing gains and losses. In contrast, the assimilation-contrast
theory and Weber's law suggests that promotional levels should
be large enough to reach a threshold at which point the framing
effects should increase with promotional levels. However, the au-
thors note that the framing effect has been shown to exist across
many discount levels: 20.6% (Diamond & Sanyal, 1990), 25% (this
study), and 50% (Smith & Sinha, 2000).
Thus, our proposition appears more comprehensive and
reasonable with regard to how promotion frames in?uence the
mental accounting processes of consumers, especially when con-
sumers assign discounts to each item. This perspective indicates
why perceived savings differ across two different framing condi-
tions while explaining effectively the ?ndings of prior research. In
addition, understanding the assignment of discounts will assist
companies to set the format mixed promotions for maximum
impact, as discussed below.
6.2. Comparison with previous studies of the recency effect
The results in this study may re?ect recency, which is an order
effect. According to Haugtvedt and Wegener (1994), a primacy ef-
fect exists if people who encounter two opposing messages form
judgments more consistent with the ?rst message, while a recency
effect exists if judgments are more consistent with the opposing
message. However, studies into order effects tend to include many
pieces of information in order to explore the persuasive effect of
target information (e.g., Buda & Zhang, 2000; Haugtvedt &
Wegener, 1994). The order effect seems less likely to exist in the
context of a small number of information frames such as the two
presented in this study. Moreover, while individuals typically
receive information in a certain order, participants in this study
were informed that the consumers paid for two items at the same
time (Experiments 1 and 3). Thus we cannot assume a recency ef-
fect in this study. And if an order effect exists, why is it more likely
to be a recency than a primacy effect? Therefore, we ?nd no reason
to believe that the results of the present study are affected by a
recency effect.
6.3. Managerial implications
This study demonstrates the robustness of framing effects. Re-
sults are applicable to marketing practice, especially within the
context of retailers or marketing managers applying a mixed
approach to enhance the attractiveness of a promotional deal.
Moreover, this type of promotion format does not increase costs.
Therefore, the ?ndings of this study suggest that marketing
managers should use focused-discount framing rather than all-
discount framing to maximize the savings perceptions of
consumers.
However, the mixed promotion effect may decline if a company
consistently adopts the same promotion format due to consumers
no longer taking note of new promotions or reducing their refer-
ence price for the target product. Therefore, in general, companies
should vary their promotional formats. However, when adopting an
all-discount promotion after using a focused discount promotion, a
company should prime customers to focus on the ?rst item. An
example advertisement might read: “Enjoy two cakes for 25% off …
taste the delicious ?avor from the very ?rst bite.” This focus may
effectively increase the promotion-related savings perceived by
consumers.
Con?icts of interest
All contributing authors declare no con?icts of interest.
Acknowledgment
The research reported in this article was supported by research
grant NSC 99-2410-H-390-018-MY2 and NSC 101-2410-H-231-005
from the National Science Council of Taiwan. The authors
acknowledge the editor and anonymous reviewers for their helpful
suggestions. They also thank Tsai, Hsiao-Tsung for helpful discus-
sions and for suggesting relevant references.
References
Barnes, J. G. (1975). Factors in?uencing consumer reaction to retail newspaper sale
advertising. In Combined Proceedings of the American Marketing Association,
Chicago (pp. 471e477).
Buda, R., & Zhang, Y. (2000). Consumer product evaluation: the interactive effect of
message framing, presentation order, and source credibility. Journal of Product &
Brand Management, 9(4), 229e242.
Chen, S. S., Monroe, K. B., & Lou, Y.-C. (1998). The effects of framing price promotion
messages on consumers' perception and purchase intentions. Journal of
Retailing, 74(3), 353e372.
Diamond, W. D. (1992). Just what is a dollar's worth? Consumer reactions to price
discounts vs. extra product promotions. Journal of Retailing, 68(3), 254e270.
Diamond, W. D., & Campbell, L. (1989). The framing of sales promotions: effects on
reference price change. Advances in Consumer Research, 16, 241e247.
Diamond, W. D., & Sanyal, A. (1990). The effect of framing on the choice of super-
market coupons. Advances in Consumer Research, 17, 488e493.
Gabor, A., & Granger, C. (1964). Price sensitivity of the consumer. Journal of Adver-
tising Research, 4(4), 40e44.
Hardesty, D. M., & Bearden, W. O. (2003). Consumer evaluations of different pro-
motion types and price presentations: the moderating role of promotional
bene?t level. Journal of Retailing, 79(1), 17e25.
Harlam, B. A., Krishna, A., Lehmann, D. R., & Mela, C. (1995). Impact of bundle type,
price framing and familiarity on purchase intention for the bundle. Journal of
Business Research, 33(1), 57e66.
Haugtvedt, C. P., & Wegener, D. T. (1994). Message order effects in persuasion: an
attitude strength perspective. Journal of Consumer Research, 21(1), 205e218.
Houston, D. A., Sherman, S. J., & Baker, S. M. (1989). The in?uence of unique features
and direction of comparison on preferences. Journal of Experimental Social
Psychology, 25(2), 121e141.
Kahneman, D., & Tversky, A. (1979). Prospect theory: an analysis of decision under
risk. Econometrica, 47(2), 263e291.
Kalyanaram, G., & Winer, R. S. (1995). Empirical generalizations from reference price
research. Marketing Science, 14(3), 161e169.
LeBoeuf, R. A., & Sha?r, E. (2003). Deep thoughts and shallow frames: on the sus-
ceptibility to framing effects. Journal of Behavioral Decision Making, 16(2),
77e92.
Levin, I. P., Gaeth, G. J., Schreiber, J., & Lauriola, M. (2002). A new look at framing
effects: distribution of effect sizes, individual differences, and independence of
types of effects. Organizational Behavior and Human Decision Processes, 88(1),
411e429.
Levin, I. P., Johnson, R. D., & Davis, M. L. (1987). How information frame in?uences
risky decisions: between-subjects and within-subject comparisons. Journal of
Economic Psychology, 8(1), 43e54.
Lowe, B. (2010). Consumer perceptions of extra free product promotions and dis-
counts: the moderating role of perceived performance risk. Journal of Product &
Brand Management, 19(7), 496e503.
H.-H. Liu, Y.-Y. Chiu / Asia Paci?c Management Review 20 (2015) 201e209 208
Monroe, K. B. (1971). Measuring price thresholds by psychophysics and latitudes of
acceptance. Journal of Marketing Research, 8(4), 460e464.
Palazon, M., & Delgado-Ballester, E. (2009). Effectiveness of price discounts and
premium promotions. Psychology & Marketing, 26(12), 1108e1129.
Prelec, D., & Loewenstein, G. (1998). The red and the black: mental accounting of
savings and debt. Marketing Science, 17(1), 4e28.
Sha?r, E., & Thaler, R. H. (2006). Invest now, drink later, spend never: on the mental
accountingof delayedconsumption. Journal of Economic Psychology, 27(5), 694e712.
Sinha, I., & Smith, M. F. (2000). Consumers' perceptions of promotional framing of
price. Psychology & Marketing, 17(3), 257e275.
Smith, M. F., & Sinha, I. (2000). The impact of price and extra product promotions on
store preference. International Journal of Retail & Distribution Management,
28(2), 83e92.
Soman, D., & Gourville, J. T. (2001). Transaction decoupling: how price bundling
affects the decision to consume. Journal of Marketing Research, 38(1), 30e44.
Stokes, M. E., Davis, C. S., & Koch, G. G. (1995). Categorical data analysis using the SAS
system. Cary, NC: SAS Institute Inc.
Thaler, R. H. (1985). Mental accounting and consumer choice. Marketing Science,
4(3), 199e214.
Thaler, R. H. (1999). Mental accounting matters. Journal of Behavioral Decision
Making, 12(3), 183e206.
Tversky, A. (1977). Features of similarity. Psychological Review, 84(4), 327e352.
Wiseman, D. B., & Levin, I. P. (1996). Comparing risky decision making under con-
ditions of real and hypothetical consequences. Organizational Behavior and
Human Decision Process, 66(3), 241e250.
Yi, Y., & Yoo, J. (2011). The long-term effects of sales promotions on brand attitude
across monetary and non-monetary promotions. Psychology & Marketing, 28(9),
879e896.
H.-H. Liu, Y.-Y. Chiu / Asia Paci?c Management Review 20 (2015) 201e209 209
doc_137428966.pdf