You are on page 1of 13

Journal of Retailing 81 (1, 2005) 35–47

Effects of pricing and promotion on consumer perceptions:


it depends on how you frame it
Peter R. Darke a,∗ , Cindy M.Y. Chung b,1
a Marketing Division, Sauder School of Business, University of British Columbia, 2053 Main Mall, Vancouver, BC, Canada V6T 1Z2
b Marketing and International Business Division, Nanyang Business School, Nanyang Technological University,

Nanyang Avenue, Singapore 639798, Singapore

Abstract

This research examined the extent to which different promotional frames increased perceptions of deal value. Price discounts dominate the
sales promotions employed by marketers. The framing literature suggests discounts have robust positive effects on consumer perceptions of
deal value. However, the current research showed that negative quality inferences moderated discount framing effects and undermined deal
value, particularly when no assurance of product quality was provided. Every-day-low-price offers were also vulnerable to negative quality
inferences, while free gift frames maintained quality perceptions and increased deal value. Product trial acted to further magnify promotional
framing effects, according to a confirmation bias. These findings were consistent with an attribute framing mechanism.
© 2005 by New York University. Published by Elsevier Inc.

Keywords: Pricing; Promotion; Consumer perception

Price discounts are by far the most common form of sales day-low-price claims (EDLPs) are said to provide a straight-
promotion employed by marketers, and their use has steadily forward means of conveying value to consumers, which is
increased in recent years. For instance, the number of price less likely to undermine product quality (Bond 1997; Liesse
reductions offered in department stores grew from six per- 1992; Morgan 1994; Ortmeyer et al. 1991). Another alter-
cent of total sales to nineteen percent between 1963 and native is to offer premiums or free gifts. Anecdotal evidence
1986 (Blattberg & Neslin 1991), and the number of coupons suggests consumers often view such offers as highly valuable
distributed by manufacturers more than quadrupled between (e.g., Lobb 1997).
1976 and 2000 (Promotion Marketing Association Coupon Despite a good deal of speculation, very little empirical
Council 2001). In addition, numerous empirical studies have research has directly examined the advantages and disadvan-
found that discount prices lead to a reliable increase in per- tages of discounts compared to other promotional strategies
ceptions of value (Compeau & Grewal 1998). (see Hoch, Dreze, & Purk 1994). This is surprising given
However, price discounts have also received their share that EDLPs and premiums serve as common alternatives to
of criticism (e.g., Ortmeyer, Quelch, & Salmon 1991). Some discounts, and managers often have the option of choosing
critics suggest consumers are skeptical of sale prices, in that between these strategies when deciding how to promote their
they view the lower selling price, rather than the initial price, products. The current studies addressed the need for research
as the “true price” of the item. Discounts may also undermine in this area by identifying conditions under which discounts
perceptions of product quality. This has led at least some ex- have limited effects on deal value, and examining whether
perts to advocate alternative strategies. For instance, every- EDLP or free gift offers provide any advantages over dis-
counts under such conditions. In particular, we show the nor-

mally robust effects of discount framing on perceptions of
Corresponding author. Tel.: +1 604 822 8362.
E-mail addresses: peter.darke@sauder.ubc.ca (P.R. Darke),
value can be moderated by negative quality inferences. This
acchung@ntu.edu.sg (C.M.Y. Chung). idea is developed by comparing the discount framing liter-
1 Tel.: +65 6790 4980. ature to other research suggesting that discounts can have

0022-4359/$ – see front matter © 2005 by New York University. Published by Elsevier Inc.
doi:10.1016/j.jretai.2005.01.002
36 P.R. Darke, C.M.Y. Chung / Journal of Retailing 81 (1, 2005) 35–47

negative effects on perceptions of quality. We also examine ing the probability of future purchases (Dodson, Tybout, &
whether EDLPs or free gifts are any less sensitive to negative Sternthal 1978; Doob, Carlsmith, Freedman, Landauer, &
quality inferences. The results show that unrelated free gift Saleng 1969). None of these studies were included in the
offers, but not EDLPs, maintained quality perceptions and Compeau and Grewal (1998) meta-analysis of discount fram-
communicated greater deal value. Finally, attribute framing ing, presumably because they did not meet the criteria for in-
(Levin, Schneider, & Gaeth 1998) was identified as the psy- clusion. Self-perception theory (Scott & Tybout 1979; Scott
chological mechanism for these promotion effects. & Yalch 1980) offers one explanation for the negative effects
of discounts. It suggests consumers attribute their purchase
to the incentive provided by the discount itself (e.g., “I chose
Discounts, framing and transaction utility theory it because it was on sale”), rather than to the inherent quality
of the product (e.g., “I chose it because it’s a good product”),
A large number of studies show that discount offers can with the net result that quality perceptions are lower when
have positive effects on consumer perceptions in terms of the items are discounted.
value associated with the offer (e.g., Darke & Dahl 2003; The current investigation adopted a slightly different ex-
Inman, Peter, Anil, & Raghubir 1997; Urbany, Bearden, & planation for these negative discount effects, which is that
Weilbaker 1988). Transaction utility theory (TUT; Thaler discount prices produce negative price-quality inferences
1985) suggests that two different types of value are derived (Garbor & Granger 1961; Raghubir & Corfman 1999). Ac-
from such purchase transactions. Discounts provide acqui- cording to this view, consumers use the lower selling price
sition utility, or standard economic value, by lowering the to infer the quality of the product is also lower. The well-
amount paid to receive the same product benefits. Discounts known price-quality literature shows that price levels have
also provide transaction utility, which relates to the perceived reliable effects on perceptions of quality (Rao & Monroe
merits of the deal and goes beyond the economic outcomes. 1989). Price-quality inferences are most likely when it is dif-
The mental calculus for evaluating transaction utility involves ficult to directly judge the quality of the product (Jacoby,
comparing the current price offer (p) to an internal standard Olson, & Haddock 1971), or when consumers are unfamil-
for the expected price of the item (p* ), where transaction iar with the brand (Raju 1977; Wheatley & Chiu 1977) or
utility increases if the price offer meets or falls below the ex- retailer (Stafford & Enis 1969). These findings suggest that
pected reference price. Discounts increase transaction utility consumers tend to use price as an indication of quality unless
by raising consumer expectations of the usual price of the no additional assurances of quality are provided. Interest-
item (i.e., by raising estimates of p* ). ingly, marketers frequently offer discounts in precisely the
A great deal of research has examined discounts using situations where price-quality inferences are most likely to
TUT’s framing perspective. In the majority of these studies occur. Discounts are commonly used to sell generic goods,
(e.g., Lichtenstein & Bearden 1989; Urbany et al. 1988), dis- to induce consumers to try new products, or to switch to less
count size is manipulated by holding the selling price of the familiar brands and retailers (Blattberg & Neslin 1991). Un-
item constant, while varying the discount frame using the der these circumstances, doubt about the quality of promoted
initial or suggested retail price (e.g., $39.99 vs. “was $49.99 items is likely to exist, and consumers are therefore more
. . . now $39.99”). Manipulating the discount in this fashion likely to make negative price-quality inferences.
varies the transaction utility associated with the deal, while
holding acquisition utility constant (i.e., the same amount
of money is paid for the same product regardless of whether Price-quality inferences moderate positive discount
the price is discounted or a regular price). A meta-analysis by effects
Compeau and Grewal (1998) found that framing prices in this
way reliably increased estimates of the reference price and Despite a substantial amount of research concerning pos-
led to more positive evaluations of the deal (see also Krishna, itive discount framing effects and research that examines
Briesch, Lehmann, & Yuan 2002). However, Compeau and consumers’ propensity to make price-quality inferences, rel-
Grewal (1998) emphasized a need to identify important mod- atively little attention has been paid to the implications
erators for these discount framing effects, and suggested neg- these two research literatures have for each other. Following
ative quality inferences are particularly likely to moderate Compeau and Grewal (1998), we suggest that negative price-
discount framing effects. quality inferences are likely to moderate positive discount
framing effects on deal value. This can occur in two ways.
The most obvious possibility is that negative quality infer-
Negative quality inferences and discounts ences will undermine the perceived acquisition utility of the
deal. Lowering a price using a discount offer not only lowers
A number of studies show discounts can also lead to the cost to the consumer, but also threatens to lower percep-
more negative consumer perceptions by undermining the per- tions of product quality through negative price-quality infer-
ceived quality of the discounted item (Raghubir & Corfman ences relating to the lower selling price. Such price-quality
1999; Scott & Yalch 1980; Tybout & Scott 1983), and lower- inferences should attenuate the effects of the lower cost on
P.R. Darke, C.M.Y. Chung / Journal of Retailing 81 (1, 2005) 35–47 37

perceptions of value. Rather than thinking they are getting this would suggest negative quality inferences are capable of
the same quality item at a lower price, consumers may use moderating positive discount framing.
the lower discount price to infer they are actually getting a
lower quality item at a lower price.
It is also possible for negative quality inferences to under- Attribute framing
mine the value of the discount via transaction utility. TUT
proposed that transaction utility is a function of the expected The process by which price-quality inferences are pre-
price (p* ) relative to the selling price (p). However, if con- dicted to influence judgments of discounted products is best
sumers infer that product quality is consistent with the lower described in terms of an attribute framing mechanism (Levin
discount price, the internal reference price for this product et al. 1998). This form of framing is distinct from the gain/loss
might also be adjusted to match the lower perceptions of framing that underlies mental accounting, and TUT. Attribute
quality (p* would approach p), and thereby undermine trans- framing occurs when superficial variations in the description
action utility. This is consistent with previous criticisms that of a single aspect of the offer leads consumers to make either
discounts are ineffective because consumers assume the sale additional positive or negative associations about the product,
price is the “true” price of the item (Ortmeyer et al. 1991). which then influence the overall valence of further product
We have seen there is a great deal of evidence that discount evaluations. For instance, a study by Levin and Gaeth (1988)
frames tend to produce a reliable increase in perceptions of showed that framing the description of ground beef as 75
deal value. This conclusion seems to contradict the sugges- percent lean rather than 25 percent fat led to more favorable
tion that discounts can also induce negative quality inferences evaluations of the product, despite the fact that the quality
that threaten to undermine perceptions of deal value. How- of the product was identical in both cases. The associative
ever, careful examination of the procedures used in studies of process is analogous to priming effects on price judgment
discount framing suggests that some of the procedures may (e.g., Herr 1989), where subtle variations in the information
have artificially limited the extent to which consumers were presented cause positive or negative associations to become
likely to make negative price-quality inferences. For instance, more salient when making evaluative judgments. In the cur-
many discount framing studies provide consumers with an rent context, different sales promotions are assumed to focus
additional assurance of quality by using name brand prod- consumers on different prices, which alter the associations
ucts (e.g., Darke & Freedman 1993, 1995; Grewal, Monroe, made for the quality of the product in a direction consistent
& Krishnan 1998; Urbany et al. 1988), or by including direct with the most salient price.
instructions that assure consumers of the product’s quality Furthermore, attribute framing may be aided by a con-
(e.g., Darke & Freedman 1993, 1995; Urbany et al. 1988). firmatory bias, where an initial shift in expectations due to
As mentioned, such assurances would be expected to limit priming eventually produces much larger shifts in evaluation
price-quality inferences. through the operation of cognitive biases in processing fur-
In addition, and more importantly, the manner in which ther evidence relevant to product quality (Levin et al. 1998).
discounts are typically manipulated in framing studies makes The effects of confirmatory biases are well known in other
it virtually impossible for negative price-quality inferences areas of research (Snyder 1984), but such effects have not
to occur. As mentioned, studies that produce positive fram- been previously demonstrated in the price/promotion litera-
ing effects tend to manipulate discount size by varying the ture. The effects of promotion on consumer perceptions are
initial price of the item, while holding the selling price con- typically examined either pre- or post-consumption, but we
stant (e.g. comparing an item sold at $39.99 with an item know of no existing studies that have compared the effects of
sold at the same price but said to originally cost $59.99). promotion both before and after product trial.
In this case, getting a discount does not result in a lower
selling price, which means there is simply no way that neg-
ative price-quality inferences could undermine perceptions Negative quality inferences: discounts versus alternative
of value. However, under normal circumstances, the selling strategies
price is lowered when consumers receive a discount, which
means there is greater possibility for negative price-quality From an attribute framing perspective, discounts provide
inferences to undermine deal value. In fact, all the studies that consumers with both a higher initial price (was $59.99) and
find discounts lead to lower quality perceptions have manip- a lower selling price (now $39.99), either of which might be
ulated discounts by lowering the selling price, rather than used to infer quality. We initially assumed that consumers
by raising the initial price (e.g., Raghubir & Corfman 1999; would consider both the initial price and the final price of
Scott & Yalch 1980). the discount when trying to determine the quality of the pro-
Overall, we predicted that negative quality inferences moted item. This meant that, while the lower selling price of
would limit the extent to which discount frames increase deal the discount would be expected to lower quality perceptions
value when the discount is created by lowering the selling (compared to an item sold at the higher original price), the
price rather than by raising the initial price, and when no presence of the higher initial price would also be expected to
additional assurances of product quality are provided. If so, maintain quality perceptions to some extent (compared to an
38 P.R. Darke, C.M.Y. Chung / Journal of Retailing 81 (1, 2005) 35–47

item regularly sold at the lower price). Overall, we expected ment 3) would prove to be any less vulnerable to negative
quality judgments for the discounted price to fall somewhere price-quality inferences. Finally, direct tests of the attribute
between quality inferences based on the initial price alone framing process, as well as the confirmatory bias hypothesis
and inferences based on the lower selling price alone. were conducted in the final experiment.
The current research also examined whether other forms
of promotion/pricing would prove to be less vulnerable to
negative quality inferences and better convey the value of Experiment 1: price-quality inferences and discount
the offer. Surprisingly, no existing empirical research has framing
explicitly examined the effects EDLPs have on consumer
perceptions. As mentioned, some experts have speculated Method
that EDLPs are less likely to undermine product quality, and
are better able to communicate value to consumers than dis- Procedure
counts (e.g., Ortmeyer et al. 1991). However, attribute fram- Eighty-two subjects were randomly given a scenario cor-
ing would seem to predict just the opposite. More specifically, responding to one of the conditions in the 2 (Price: full-price
lowering prices in the form of EDLPs provides only the lower vs. discount) × 2 (Quality: assured vs. not assured) between-
selling price as a basis for inferring quality (EDLP of $39.99). subject design. The dependent measures were perceptions of
This suggests that lowering prices in this way should actually quality and deal value.
be more vulnerable to negative quality inferences compared The scenario asked consumers to imagine they needed
to discounts with the same selling price but also including a a haircut for a job interview the next day, but found their
higher initial reference price. regular hairstylist was not available. This meant they had
In contrast, attribute framing suggests that free gift offers to find another stylist in the phone book. When they called,
should be better able to maintain perceptions of the target those in the full-price condition were told the price at the new
product’s quality than discounts of the same monetary value. hairstylist was the same as their regular stylist. Subjects in
This is because consumers are not likely to correct for the the discount condition were told the new stylist offered a 50
value of the free gift when inferring the quality of the target percent discount. In addition, subjects in the quality assured
item (Campbell & Diamond 1990). This correction would condition were told the quality of the haircut at the new place
involve the additional mental operation of subtracting the was the same as their regular stylist, while those in the no-
value of the free gift from the overall price of the item before assurance condition were not given any information about
making any quality inferences. However, attribute framing the service quality.
studies consistently show that individuals are unlikely to re- After reading the scenario, participants rated from 1 (not at
frame information in this way (see Levin et al. 1998, for a all) to 7 (very) whether the new hairstylist seemed qualified,
review). Instead, consumers are more likely to consider the reliable, and good. Responses were averaged into an overall
full price of the item when making price-quality inferences, quality index (α = .75). Subjects also rated the value of the
as this is the most readily available price. For example, if a deal using seven point scales anchored as follows: unattrac-
product priced at $59.99 was offered with a free gift worth tive/attractive, bad-buy/excellent-buy, extremely unfair/fair,
$20, quality inferences should be based on the full price of no-savings-at-all/extremely large savings, extremely worth-
$59.99 rather than the corrected price of $39.99. According less/valuable (Inman et al. 1997; Lichtenstein & Bearden
to the attribute framing perspective, this is because the full 1989; Urbany et al. 1988). An overall index was computed
price of the item is more salient to consumers than the price by averaging these items (α = .75). Higher scores indicated
that corrects for the value of the gift. greater deal value. Finally, participants indicated whether the
price had been discounted or not as a manipulation check.
The quality ratings acted as a check on the quality assurance
Summary manipulation.

With this background, we undertook a series of exper- Results


iments designed to examine the role that negative quality
inferences play in limiting the positive effects of discounts Eighty-four percent of those in the discount condition
on consumer perceptions of value. Experiment 1 examined correctly recognized the deal, whereas all participants in
our suggestion that discounts should have positive effects the full-price condition indicated they paid the full price
on perceptions of value primarily when consumers are pro- [χ2 (1) = 59.14, p < .001]. A 2 × 2 ANOVA was computed for
vided with some additional assurance that the quality of the each dependent measure. Follow-up contrasts used the full
promoted item has not changed. In contrast, when no such as- MSE. For the perceived quality measure, the ANOVA indi-
surances are provided, discounted products should be vulner- cated significant Discount and Assurance main effects [F’s(1,
able to negative quality inferences, which undermine percep- 78) = 16.10 and 5.07, p’s < .05], and a marginal interaction
tions of value. Subsequent experiments examined whether [F(1, 78) = 2.91, p < .10]. When no quality assurance was pro-
EDLP (Experiments 2a and 2b) or free-gift offers (Experi- vided, the discounted haircut was perceived as lower in qual-
P.R. Darke, C.M.Y. Chung / Journal of Retailing 81 (1, 2005) 35–47 39

ity than the full price haircut [M’s = 3.91. vs. 4.89, p < .001], ity of the product described. The context was closely
whereas the discount effect was not significant when qual- based on past studies that found discounts have positive
ity was assured [M’s = 4.59 vs. 4.98, p > .10]. The analysis framing effects on price perceptions and deal value (e.g.,
for deal value revealed significant Discount and Assurance Urbany et al. 1988). In the experimental conditions, prices
main effects [F’s(1, 78) = 26.19 and 7.43, p’s < .01], as well as were lowered using either an EDLP or discount claim,
a significant interaction [F(1, 78) = 6.26, p < .01]. The price while the price was varied without any extra promotional
discount significantly increased the perceived value of the of- claim in the control conditions. Standard price perception
fer when quality was assured [M’s = 5.90 vs. 4.71, p < .001], measures were used as an additional means of assessing
but had only a marginal effect on value when the quality was any framing effects of discounts. Experiment 2b was a
in question [M’s = 5.08 vs. 4.67, p < .10.]. replication of Experiment 2a, with the difference that the
discounts were explicitly time limited in Experiment 2b
Discussion (Inman et al. 1997).

The results of this initial experiment were consistent with Method (Experiment 2a)
our suggestion that discounts should give rise to negative
quality inferences primarily when there is some doubt as to Participants and design
the quality of the product. The findings also showed the dis- One-hundred and seventy-six undergraduate students
count had a positive effect on perceptions of deal value when completed the study. The experimental design was a 3 (Offer-
additional assurance was provided, whereas these discount Type: Control, EDLP, Discount) × 3 (Selling-Price: $59.99,
effects were marginal when there was no quality assurance. $39.99, $19.99) between-subjects factorial. Perceptions of
The findings were therefore consistent with our prediction product quality and price perceptions served as the main de-
that negative price-quality inferences can undermine the per- pendent measures.
ceived value of a discount frame.
Procedure
Each person examined an ad for a pair of headphones. The
Experiments 2a and 2b: discounts versus EDLPs ad was said to be from the current catalogue for a local store,
and included pricing information for the relevant experimen-
The remaining experiments examined the effects discount tal condition, a picture of the headphones, and a description
offers had on perceptions of quality and value in an adver- of the item. The description read, “Full bass response. Clear
tising context where negative quality inferences were likely sound. Foam pad inside ear cups for comfortable fit.” No
to occur. In particular, these studies investigated whether the brand information was provided.
alternative strategy of lowering prices in the form of EDLPs
(Experiment 2a/2b), or providing an unrelated free gift (Ex- Independent variables
periment 3) might prove any less vulnerable to negative qual-
ity inferences, and better at communicating the value of the Pricing information was varied using the final Selling-
offer. Price and the Type of offer made. See Table 1. Levels for
Importantly, all of the remaining studies included a set the Selling-Price factor were determined in a pretest by ask-
of control conditions where the regular price of the target ing participants to indicate the average price for a pair of
item was varied without making any special promotional headphones that were of low, average and high quality. Sell-
claims. These control conditions allowed us to determine ing prices of $19.99, $39.99, and $59.99 were selected. The
whether negative quality inferences or positive framing ef- Offer-Type was varied according to the additional claim at-
fects would dominate perceptions of the deal. For instance, tached to the selling price. In the EDLP condition, the claim
by comparing an item discounted from $59.99 to $39.99 of “every day low price” preceded the selling price. Discount
with a control condition where the item is regularly sold at offers framed the selling price relative to a higher initial price
$59.99, we can determine whether the lower selling price
has any negative effects on quality perceptions. In addition,
by comparing the same discount offer to a control condition Table 1
where the item is regularly sold at $39.99, we can determine Conditions for Experiment 2a
whether the higher initial price of the discount exerts any Offer-Type Selling-Price
positive framing effects on deal value. Therefore, this design $59.99 $39.99 $19.99
allowed us to observe any significant negative price qual- Control Price $59.99 Price: $39.99 Price: $19.99
ity inferences, as well was any significant positive framing
Every-day-low EDLP: $59.99 EDLP: $39.99 EDLP: $19.99
effects.
price (EDLP)
Experiments 2a and 2b examined both EDLPs and dis-
count offers. Participants were given an ad that included Discount Was: $79.99 Was: $59.99 Was: $39.99
Now: $59.99 Now: $39.99 Now: $19.99
pricing information and were asked to estimate the qual-
40 P.R. Darke, C.M.Y. Chung / Journal of Retailing 81 (1, 2005) 35–47

using the phrase “Was (initial price), Now (selling price),” ac- showed that decreases in the selling price significantly de-
cording to the format from previous studies (Lichtenstein & creased price perceptions at each level [M’s = $40.17, $32.39,
Bearden 1989). For comparison purposes, the initial price for and $23.89, p’s < .01]. However, there were no significant
each condition was set at the level of the next highest selling Offer-Type effects [p’s > .10]. Overall, this meant that a de-
price. For the high selling price condition, the initial price crease in the selling price lowered price perceptions, but dis-
of the discount was set at $79.99, which a pilot study found counts did not frame price perceptions at a higher level.
was the highest plausible price for high quality headphones.
In the control conditions, no special claims were made (price Method (Experiment 2b)
$X).
Ninety-six undergraduate students participated. The main
Measures change from Experiment 2a was the discount was made ex-
Subjects first completed items relating to the cover-story. plicitly time limited and the design was reduced in size. The
After this, they completed the perceived quality measure by discount condition included the information that the sale was
rating how well-made the headphones were, the sound qual- over at the end of the month (namely May 31), which was
ity, and the overall quality of the headphones. Seven-point approximately two weeks after the study was completed. The
scales were used (−3 to +3), and responses were averaged to ad was said to be from the April issue of a locally distributed
form an index of perceived quality (α = .88). Participants then catalogue. A 3 (Offer-Type: Control, EDLP, or Temporary
completed a standard price perception measure (Lichtenstein Discount) × 2 (Selling-Price: $39.99 or $19.99) between-
& Bearden 1989; Urbany et al. 1988), where they estimated subjects design was used. A high price control condition
the average price and the lowest price for the item else- was also included (price: $59.99). Finally, a measure of per-
where. Responses were averaged to form a price perception ceived value was added to the questionnaire. Participants
index (r = .83, p < .001). As a manipulation check, partici- were asked to consider both the quality and the price of the
pants were asked to recall the selling price of the item, and item, and rate the overall value of the offer from −3 (low) to
identify whether the item was sold at a discount, an every- +3 (high).
day-low-price, or a regular price.
Results (Experiment 2b)
Results (Experiment 2a)
Manipulation checks
Manipulation checks
The majority of participants correctly identified the type
The checks showed participants knew the selling prices
of offer made [regular price = 86 percent, EDLP = 76 percent,
[M’s = $23.63, $40.71, and $59.14, F(2, 169) = 98.98,
and discount = 86 percent, χ2 (12) = 133.41, p < .001].
p < .001], and the Offer-Type [93 percent, 75 percent, 95 per-
cent correct responses for regular price, EDLP and discount
conditions, χ2 (6) = 243.01, p < .001]. Perceived quality
A 2 (Offer-Type) × 2 (Selling-Price) ANOVA was com-
Perceived quality puted for the index of perceived quality, with additional con-
A 3 (Offer-Type) × 3 (Selling-Price) between-subjects trasts used to make comparisons with the high priced control
ANOVA was used to analyze all dependent measures. Follow group. All contrasts used the full MSE. Overall, the find-
up contrasts used the full MSE. The ANOVA showed only the ings conceptually replicated Experiment 2a. The ANOVA
Selling-Price main effect was significant [F(2, 163) = 5.92, revealed no significant effects [F’s < 1, p’s > .40]. The con-
p < .01]. Contrasts showed the moderate and low priced con- trasts showed that perceived quality was higher in the high
ditions did not differ [M’s = −0.24 and −0.32], but both of priced control condition than all the other conditions. Sup-
these were significantly lower in perceived quality than the porting the price-quality prediction, both moderate and low
high priced condition [M = 0.33, p’s < .01]. These findings priced items [M’s = −0.36 and −0.64] were lower in per-
suggest consumers used price to infer quality, although the ceived quality than the high priced control item [M = 0.45,
price-quality effects were observed between high and mod- p’s < .05]. EDLPs [M’s = −0.22 and −0.53] and temporary
erate price levels only. Also, Offer-Type had no significant discounts [M’s = −0.49 and −0.40] were also lower than the
effects [p’s > .25 for both the main effect and interaction]. high priced controls [p’s < .10]. The findings again suggested
This implies that both discount and EDLP strategies were that discounts and EDLPs provided no significant protection
vulnerable to negative quality inferences, and neither strat- against negative quality inferences.
egy was effective in attenuating negative quality inferences
relative to controls. Price perceptions
An ANOVA for price perceptions showed these decreased
Price perceptions with the selling price [M’s = $19.36 vs. $33.30, for low vs.
There was also a significant main effect of Selling-Price moderate selling prices; F(1, 72) = 57.31, p < .001]. Again,
on price perceptions [F(2, 163) = 15.41, p < .001]. Contrasts there were no significant effects of Offer-Type [p’s > .15],
P.R. Darke, C.M.Y. Chung / Journal of Retailing 81 (1, 2005) 35–47 41

suggesting temporary discounts did not frame price percep- $20) should cause consumers to make their quality estimates
tions at a higher level. on the basis of the full price ($59.99), rather than on the basis
of the price minus the value of the gift ($39.99). In terms of
Perceptions of value attribute framing, the full price of the offer should be more
The ANOVA for the measure of perceived value showed salient to consumers when a free gift is offered, compared
no significant effects of either selling price or offer-type to the corrected price, which should therefore help maintain
[F’s < 1, p’s > .50]. The contrasts for the high priced con- perceptions of quality for the target product, and commu-
trol group were all nonsignificant [p’s > .25]. These findings nicate greater deal value. The attribute framing mechanism
implied that none of the offer-types were effective in com- was also directly tested by including measures of the extent to
municating the objective value of the lower price offer. which each price level was thought to provide a reliable indi-
cation of product quality. If consumers fail to correct for the
Discussion value of the gift as predicted, perceptions concerning the re-
liability of the full price ($59.99) should reflect the pattern of
Experiments 2a and 2b suggest that both discounts and quality judgments, and the reliability for the corrected price
EDLPs were highly vulnerable to negative quality inferences. ($39.99 = $59.99 minus the $20 gift) should be unrelated to
Lower prices caused a significant drop in perceptions of qual- quality effects.
ity regardless of whether they were in the form of an EDLP Finally, this study examined the Levin et al. (1998) sug-
or a discount (compared to higher priced control conditions). gestion that the attribute framing may lead to a confirma-
Furthermore, neither EDLPs nor discounts provided any sig- tion bias, where initial expectations are magnified when con-
nificant protection against negative quality inferences (these sumers make further evaluations. Confirmation bias is par-
conditions did not differ from lower priced controls). The ticularly likely to occur when evaluations are subjective or
result was that consumers failed to recognize the objective ambiguous (Snyder 1984), such as when estimating the qual-
increase in value relating to the lower selling price of the ity of headphones. To examine this hypothesis, participants
headphones. Finally, the results of both studies also showed rated product quality immediately after seeing the ad, and
that negative quality inferences occurred between high and again after directly examining the product.
moderate price levels but not between moderate and low price
levels. This finding is consistent with Rao and Monroe (1989)
Method
observation that price-quality inferences are more likely to
occur for items that are initially relatively expensive.
Participants and design
From an attribute framing perspective, it is not surpris-
Participants were 85 undergraduate students who were
ing that the EDLP claim failed to improve perceived quality,
randomly assigned to one of the experimental conditions. The
since it does not provide any specific reference price that
conditions included high price ($59.99) and moderate price
might be used as an alternative cue for making quality in-
($39.99) controls, a standard discount offer (was $59.99, now
ferences. However, it was somewhat surprising that discount
$39.99), and a free-gift offer (price: $59.99, plus free $20 gift).
frames did not provide any real protection against negative
A repeated measure for perceptions of product quality served
quality inferences. While we had expected the lower selling
as the main dependent variable.
price to lead to some slippage in perceived quality, we also
expected positive inferences based on the higher initial price
of the discount to counteract this slippage to at least some Procedure
extent. In addition, discount offers failed to raise price per- Participants completed the study in individual sessions.
ceptions to any significant degree, and had no effect on the They first examined the ad for the headphones, which in-
perceived value of the offer, in contrast to past findings in the cluded the picture and product description used in Exper-
discount framing literature (Lichtenstein & Bearden 1989; iment 2a/2b. We also specified the actual brand name and
Urbany et al. 1988). Overall, Experiments 2a and 2b showed model number of the product in the current study (Gamma
a consistent lack of discount framing across quality, price, LH036). Participants were largely unfamiliar with the brand
and value perceptions. (96 percent). In the free gift condition, the ad stated the price
of the headphones was $59.99, and included a free $20 mu-
sic CD of their choice. The prices in the control and discount
Experiment 3: free gifts as an alternative to discounts conditions were varied as before. (The actual cost of the head-
phones was over $90.)
Given the observed vulnerability of EDLP and discount Product ads were presented online using a simulation of
offers to negative quality inferences, it seemed important an internet shopping environment. Participants sat at a com-
to identify alternative forms of promotion that might better puter that displayed the ad on the webpage of a (fictitious)
maintain perceptions of quality, and convey the true value of retailer. After seeing the ad, participants completed the ini-
the deal. We initially proposed that framing the value of an of- tial set of quality ratings (see below). Next, subjects directly
fer in terms of a free gift (price: $59.99, plus a free gift worth examined the headphones described in the ad by listening to
42 P.R. Darke, C.M.Y. Chung / Journal of Retailing 81 (1, 2005) 35–47

a music tape for 30 s. Subjects then completed the same qual- percent), moderate price (95 percent), discount (95 percent),
ity ratings for a second time, as well as the other dependent and free gift (96 percent), χ2 (9) = 231.98, p < .001].
measures.
Main analyses
Dependent measures
The same perceived quality ratings used in Experiment Perceived quality
2 were completed, plus an additional rating of the extent to An ANOVA with experimental condition as a between-
which the headphones were stylish (α = .84). The measure subject factor and the two sets of quality ratings as a repeated
of deal value described in Experiment 1 was also completed measure (ad-only and product trial ratings) revealed a signifi-
(α = .85). In addition, participants indicated their price per- cant experimental effect [F(3, 81) = 3.78, p < .05], a repeated
ceptions (α = .74) by reporting the average price they would measure effect [F(1, 81) = 31.38, p < .001], and a significant
expect to pay elsewhere, their perceptions of the retailer’s interaction [F(3, 81) = 3.25, p < .05]. Cell means are reported
usual price, and the price if the retailer sold the item at cost in Table 2. Contrasts using the full MSE indicated the per-
(Darke & Freedman 1993; Lichtenstein & Bearden 1989). ceived quality for the free gift condition was higher than the
As a manipulation check, participants identified the pric- moderate-control condition, both after the ad [M’s = 0.83 vs.
ing information using a multiple-choice format. Finally, sub- 0.54, p < .10], and after product trial [M’s = 1.64 vs. 0.61,
jects completed a measure of cue reliability. Those in the p < .05]. Moreover, the free gift condition was similar in
high-control, discount, and free-gift conditions evaluated the quality to the high-control condition for both sets of ratings
reliability of the $59.99 price using three rating scales, rang- [M’s = 0.71 and 1.24, for post ad and trial ratings]. In con-
ing from not at all (0) to very (6). Specifically, participants trast, the discounted product was lower in perceived quality
were asked whether the higher price was: a good indication than the high-control and free-gift conditions, both after the
of the item’s quality, a reliable indication of the quality of ad [M = 0.13], and after product trial [M = 0.70, p’s < .05].
the item, and reflects the true quality of the item. Those in the Finally, the pattern of means for the significant interaction
control moderate, discount and free-gift conditions made the showed that product trial acted to further increase the initial
same ratings for the $39.99 price. In the free gift condition, differences in quality ratings between conditions (see Fig. 1).
we explicitly pointed out that $39.99 was the price of the Overall, the free gift promotion was effective in maintaining
headphones with the value of the free music CD subtracted quality perceptions, whereas the discount offer again proved
from the overall price. These were averaged to form reliabil- vulnerable to negative quality inferences. The observed in-
ity indices for the $59.99 price (α = .89) and the $39.99 price teraction suggested the effects of promotional framing were
(α = .91). aided by a confirmation bias.

Results Price perceptions


The experimental effect was also significant for price
Manipulation checks perceptions [F(3, 79) = 10.24, p < .001]. These perceptions
Almost all participants (96 percent) correctly identified were higher for the control-high and free gift conditions
the pricing information [correct responses: high price (100 [M’s = $43.76 and $45.21] than the control-moderate and dis-

Table 2
Means for conditions in Experiment 3
Measures Experimental conditions

High control (A) Moderate control (B) Regular discount (C) Free gift (D) Post hoc contrasts
Perceived quality
Ad-only 0.71 0.54 0.13 0.83 D > B, p < .10
A, B, D > C, p < .05
Product trial 1.24 0.61 0.70 1.64 A, D > B, C, p < .05
Value of deal 3.97 3.67 4.08 4.45 D > A, p < .10
D > B, p < .05
Reliability of $59.99 3.38 NA 2.54 3.64 A > C, p < .10
D > C, p < .05
Reliability of $39.99 NA 3.44 3.40 3.30
Price perceptions $43.76 $33.43 $35.00 $45.21 A, D > B, C, p < .05
N 20 22 20 23
NA = not applicable (measure not included in this condition).
P.R. Darke, C.M.Y. Chung / Journal of Retailing 81 (1, 2005) 35–47 43

Fig. 1. Effects of promotional frames on quality ratings in Experiment 3.

count conditions [M’s = $33.43 and $35.00, p’s < .05]. This is provided by comparing the experimental effects on qual-
suggests the free gift was framed in terms of the higher refer- ity before and after the perceived reliability of the higher
ence price, whereas there was no positive framing effect for price is covaried (Baron & Kenny 1986). An initial ANOVA
discounts on price perceptions. showed a significant experimental effect on perceived quality
after product trial [F(2, 60) = 4.82, p < .01]. However, the ex-
Value of the deal perimental effect was no longer significant [F(2, 57) = 1.21,
The experimental manipulation also had a significant ef- p > .30] when the reliability of the higher price was added as
fect on deal value [F(3, 81) = 3.39, p < .05]. The free gift a covariate [r = .40, F(1, 57) = 11.42, p < .001 for covariate].
promotion [M = 4.45] was seen as more valuable than the These findings suggest that the superior quality associated
regularly priced items [M’s = 3.97 and 3.67, p’s < .10]. The with the free gift condition was fully mediated by the per-
discount condition was not significantly different from any ceived reliability of the higher price.
of the other conditions.
Discussion
Perceived reliability
A oneway ANOVA computed for the measure of whether This study showed the free gift offer maintained quality
the higher price point ($59.99) was perceived as a re- perceptions, whereas the discounted product was vulnerable
liable indication of quality also proved to be significant to negative quality inferences (replicating Experiments 1 and
[F(2, 58) = 3.20, p < .05]. The higher price was perceived 2a/2b). The results also showed that the initial quality dif-
as more reliable in the control-high and free gift condi- ferences became further exaggerated following product trial.
tions [M’s = 3.38 and 3.64] than in the discount condition The free gift offer also significantly increased the value of
[M = 2.54, p’s < .10]. Finally, the ANOVA for the perceived the deal relative to the control conditions, while the discount
reliability of the moderate price ($39.99) was not significant frame did not. These findings confirmed the prediction that
[F(2, 61) = 0.076, p > .90]. framing an offer as a separate free gift should be particularly
good at communicating value to consumers. Discounts were
Testing mediation less effective in this respect. Overall, negative quality effects
The findings for the perceived quality and reliability mea- undermined the subjective value of the discount, while the
sures were consistent with the idea that product quality was fact quality perceptions were maintained for free gift promo-
perceived as relatively high for the free gift offer because tions meant greater value was ultimately conveyed.
the higher price acted as the frame, rather than the price cor- There was also direct support for the idea that attribute
rected for the gift. A direct test of this mediational hypothesis framing played a part in maintaining higher perceptions of
44 P.R. Darke, C.M.Y. Chung / Journal of Retailing 81 (1, 2005) 35–47

quality for the free gift offer. The effects of the free-gift studies within the discount framing paradigm, since these
on perceived quality were mediated by the perception that studies often provide additional quality assurances, and typi-
the higher price was the best indication of the item’s true cally manipulate the size of the discount by varying the initial
quality. Therefore, consistent with the attribute framing pro- price while holding the final price constant. As noted, it is
cess specified by Levin et al. (1998), the free gift promotion unlikely consumers would make negative price-quality infer-
seemed to be effective in maintaining perceptions of qual- ences under such conditions. The current studies show that
ity because consumers made quality inferences based on the in the absence of quality assurances, and when discounts are
original price rather than the price corrected for the value of manipulated in the normal everyday manner by lowering the
the free gift. selling price of the item, negative price-quality effects mod-
One alternative explanation for the effectiveness of the erated discount framing. This was true for measures of both
free gift promotion in maintaining quality perceptions is that deal value and price perception. Overall, negative quality in-
consumers may have appreciated the fact that the gift would ferences dominated consumer perception and suppressed any
cost the marketer less than the full $20 value of the CD. If signs of positive discount framing.
so, participants may have corrected for only a fraction of The evidence is consistent with the idea that negative qual-
the CD’s value when considering the quality of the head- ity inferences undermined both acquisition utility and trans-
phones, rather than correcting for the full $20. Although, action utility for discounts. In terms of acquisition utility,
this seemed unlikely, we included a measure of the perceived negative quality inferences led consumers to conclude they
costs of the promotion to the seller to investigate. After indi- were receiving a lower quality item at a lower price, rather
cating their responses to the reliability items, participants in than the same quality item at a lower price. In addition, the
the two promotion conditions rated how costly and expensive fact that price perceptions reflected the selling price but not
the promotion was for the retailer, from 0 (not at all) to 6 (ex- the initial price of the discount offer is consistent with the idea
tremely). However, the average of these ratings did not reveal that consumers matched their reference price to the quality
any significant differences [M’s = 2.44 and 2.52 for discount implied by the lower selling price, which would limit trans-
and free gift, t(40) = 0.19, p > .80], suggesting participants action utility since p* matched p. Once consumers inferred
did not view the value of the gift from the perspective of the the quality of the discounted item was lower, they seemed
marketer. In addition, any tendency for consumers to assume unwilling to adjust price expectations upwards to match the
the free gift was worth less than $20 would act against our initial price claim.
ability to show the observed increase in the value of the free Our studies also examined alternatives to discounts when
gift offer. negative quality inferences were likely to occur, and we
found that free gift frames were effective in increasing deal
value under such conditions (Experiment 3). When a free
General discussion gift offer was made, consumers based their quality judg-
ments on the full price of the item without accounting for
As predicted, negative quality inferences were shown to the value of the gift, which maintained perceptions of qual-
limit the extent to which discount frames were capable of in- ity, and led to higher perceptions of deal value. As men-
creasing perceptions of deal value (all experiments). Positive tioned negative price-quality inferences are most likely to
discount framing was observed only when additional qual- occur when the brand name or seller are not well-known
ity assurances were provided, and not when consumers were (Rao & Monroe 1989), or when intrinsic cues for product
less certain about product quality (Experiment 1). In fact, quality are ambiguous or unavailable (Jacoby et al. 1971).
discount framing proved to be somewhat more vulnerable to Under such circumstances, our findings suggest marketers
negative quality inferences than we had expected. The initial should consider offering free gifts instead of discounts. Al-
price claim provided little help in maintaining perceptions ternatively, Experiment 1 suggested marketers could pro-
of quality (Experiments 2 and 3). Furthermore, measures of vide additional quality assurances when offering discounted
price expectations suggested that consumers largely disre- prices.
garded the initial reference price claim. Overall, these find- Lowering prices in the form of EDLPs proved to be highly
ings were generally consistent with the critics’ suggestion vulnerable to negative quality inferences (Experiment 2).
that consumers often consider the discounted price, rather While this finding was predicted by attribute framing, it
than the initial price claim, to be the ‘true’ price of the item clearly contradicts previous speculation that EDLPs should
(Ortmeyer et al. 1991). be better at communicating value to consumers than discounts
In their review of the literature, Compeau and Grewal (e.g., Ortmeyer et al. 1991). Our finding that EDLPs are vul-
(1998) speculated that it was possible a number of factors nerable to negative quality inferences adds to existing criti-
might moderate the effects of discount frames, however they cisms of this strategy. These include the suggestion that con-
were unable to specifically test for these moderators in their sumers find EDLPs boring (DeNitto 1993; Weinstein 1992),
meta-analysis. Our findings provide initial evidence for one that EDLPs are symptomatic of a lack of creativity (Hastings
such moderator, namely negative price-quality inferences. 1993), and that they often reduce margins to a level at which
Such inferences were essentially controlled for in previous sales are no longer profitable (Hoch et al. 1994; Reda 1994).
P.R. Darke, C.M.Y. Chung / Journal of Retailing 81 (1, 2005) 35–47 45

The promotional effects observed here were generally pre- of other brands regularly priced at $39.99. None of these
dicted by the attribute framing view. According to this view, discount formats increased perceived quality [M’s = 0.45,
EDLP, discount, and free gift frames caused consumers to 0.54, 0.44, and 0.71] compared to a $39.99 control price
focus on different prices, which led them to make different [M = 0.09; F(4, 141) = 1.29, ns]. The current findings were
quality inferences. Discounts and EDLPs focused consumers therefore relatively robust with respect to the format of the
on the lower price, leading them to make lower quality in- discount.
ferences, thereby undermining the value associated with the It might seem possible to explain the negative effects of
deal. In contrast, free gift frames focused consumers on the discounts observed here in terms of self-perception theory,
higher original price of the item, which helped maintain the which suggests discounts lower quality perceptions because
level of quality they inferred, and conveyed greater value. consumers attribute their product choice to the reward pro-
This mechanism was directly confirmed in Experiment 3. It vided by the discount itself, rather than to the quality of the
is important to note that mental accounting (Thaler 1985) can product (Scott & Yalch 1980). However, there are a number
also explain the greater value consumers attached to free gift of reasons to doubt this explanation in the present case. First,
offers in terms of its silver-lining principle. However, this per- self-perception studies tend to make very salient the idea
spective cannot account for the negative effects of discount that the promotion is specifically offered as an inducement to
frames, and mental accounting makes no explicit predictions choose one product over another. This was not the case here.
for the effects of EDLPs. Moreover, mental accounting does Also, self-perception effects should occur only after a choice
not make any explicit predictions for the effects of promotion is made (Blattberg & Neslin 1991). However, our participants
on perceptions of quality, whereas these perceptions were were not asked to make any choices before indicating their
central to the attribution framing mechanism. Our findings quality perceptions, and negative quality inferences were ob-
add to the attribute framing literature by showing that this served in evaluations made both before (all studies) and af-
mechanism is also relevant to understanding the effects dif- ter product trial (Experiment 3). Self-perception theory also
ferent forms of sales promotion have on quality inferences has difficulty predicting other aspects of the current findings.
and perceptions of deal value. For instance, this view would predict free gift offers should
Experiment 3 also provided evidence for a confirmation also induce negative quality judgments (Forehand 2000), but
bias in the effects of promotion on perceived quality. Product Experiment 3 showed that quality perceptions were actu-
trial served to further exaggerate the initial promotion effects ally maintained for free gift offers. Finally, self-perception
based on the ad alone. This study is the first to demonstrate does not make clear predictions for EDLPs, whereas nega-
the effects of a confirmation bias in the context of sales pro- tive price-quality inferences clearly account for the EDLP
motion. A confirmation bias is most likely to occur in situa- findings.
tions where judgments of product quality are malleable, in the It is important to note that certain factors may limit the
sense that they are either experiential or ambiguous in nature effects of a free gift offer. For instance, Simonson, Carmon,
(Snyder 1984). We should further note that this finding also and O’Curry (1994) showed that consumer preference de-
lends additional support for the attribute framing mechanism. creased when unattractive gifts were offered because con-
Our studies employed a number of different measures and sumers thought they were paying extra for a gift they did not
manipulations that were closely based on past studies that re- want. In contrast, the gift used in Experiment 3 was likely to
liably demonstrate positive discount framing effects. These be popular (free music CD of the participant’s choice). In ad-
procedures were sensitive enough to identify the positive ef- dition, the monetary value of the gift was explicitly stated in
fects of a discount frame when quality was assured in Experi- Experiment 3. This was important because Raghubir (2004)
ment 1, to identify the effects of the price manipulation in the shows consumers can underestimate the true value of a free
control conditions of Experiments 2 and 3, and to identify the gift when it is not explicitly stated, and thereby limit the
effects of the free gift offer in the final experiment. There- overall value of the offer. It is therefore important to provide
fore, the failure to observe positive discount effects in the consumers with desirable free gifts, and to directly state the
predicted conditions does not seem likely to be due a lack of value of the gift in the ad itself.
power, nor to some other artifact of the procedures used here. The current findings seem to contradict the success that
It might still seem possible that a somewhat different type at least some retailers have had using EDLPs and discounts.
of discount format would be better able to maintain quality We would attribute their success to additional factors in their
perceptions (Krishna et al. 2002). However, additional re- marketing strategy which help avoid negative price-quality
search we conducted found that even substantial alterations inferences. For instance, Wal-mart has established a positive
in the format of the discount made little difference to qual- reputation of service and quality with its customers, and they
ity perceptions. Specifically, a pair of headphones was dis- carry many name brand products. EDLPs may have been
counted from $59.99 to $39.99 according to a number of more successful for Wal-mart because consumers are less
different formats: (1) a very temporary discount (this week- willing to use price as an indication of quality when store
end only); (2) a discount available in limited quantities; (3) and brand quality are high. Discounts are also more likely to
a discount available as $20 cash back in a scratch-and-save be successful in conveying value when additional assurances
contest; and (4) a standard discount presented in the context are provided for the quality of the discounted item. However,
46 P.R. Darke, C.M.Y. Chung / Journal of Retailing 81 (1, 2005) 35–47

it is not a simple matter to create positive store or brand rep- DeNitto, Emily. (1993). The concept of EDLP is changing. Advertising
utations in the first place. This is an expensive and time con- Age, 64, 33–34.
suming strategy. Furthermore, EDLPs (and discounts) may Dodson, Joe A., Tybout, Alice M., & Sternthal, Brian. (1978). Impact of
deals and deal retraction on brand switching. Journal of Marketing
erode brand perceptions in the long-run. Research, 15, 72–81.
The current findings contribute to the existing litera- Doob, Anthony N., Carlsmith, J. Merrill, Freedman, Jonathan L., Lan-
ture in a number of ways. First, they show that negative dauer, Thomas K., & Saleng, Tom, Jr. (1969). Effect of initial selling
quality inferences can serve as an important moderator of price on subsequent sales. Journal of Personality and Social Psychol-
the classic discount framing effects. In addition, free gifts, ogy, 11, 345–350.
Forehand, Mark R. (2000). Extending overjustification: The effect of per-
but not EDLPs, provided a useful alternative for convey- ceived reward giver intention on response to rewards. Journal of Ap-
ing value to consumers when discounts are likely to under- plied Psychology, 85, 919–931.
mine perceptions of quality. The results also supported an Garbor, Andre, & Granger, Clive W. J. (1961). On the price consciousness
attribute framing mechanism, which has not been previously of consumers. Applied Statistics, 10, 170–180.
used to understand the effects of promotion on consumer Grewal, Dhruv, Monroe, Kent B., & Krishnan, R. (1998). The effects
of price-comparison advertising on buyers’ perceptions of acquisition
perceptions. Finally, attribute framing was shown to pro- value, transaction value, and behavioral intentions. Journal of Mar-
duce a confirmation bias that intensified the initial effects keting, 62, 46–59.
of promotion on consumer perceptions following product Hastings, Hunter. (1993). EDLP is a sign only of the lack of ideas in
trial. marketing. Brandweek, 34, 58.
Herr, P. M. (1989). Priming price: Prior knowledge and context effects.
Journal of Consumer Research, 16, 67–75.
Hoch, Stephen J., Dreze, Xavier, & Purk, Mary E. (1994). EDLP, hi-lo,
Acknowledgements and margin arithmetic. Journal of Marketing, 58, 16–27.
Inman, Jeffrey, Peter, Anil C., & Raghubir, Priya. (1997). Framing the
deal: The role of restrictions in accentuating deal value. Journal of
The authors are grateful to Errol Smythe for his help Consumer Research, 24, 68–79.
in programming the internet simulation used in Experiment Jacoby, Jacob, Olson, Jerry C., & Haddock, Rafael A. (1971). Price,
3. We would also like to thank Natalie Lam for her help brand name, and product composition characteristics as determinants
with data collection. This work was supported by means of perceived quality. Journal of Applied Psychology, 55, 570–579.
Krishna, Aradhna, Briesch, Richard, Lehmann, Donald R., & Yuan, Hong.
of a grant from the Direct Selling Education Foundation of
(2002). A Meta-analysis of the impact of price presentation on per-
Canada and the Social Sciences and Humanities Research ceived savings. Journal of Retailing, 78, 101–118.
Council of Canada. Helpful comments on previous drafts of Levin, Irwin P., & Gaeth, Gary J. (1988). How consumers are affected
this paper were provided by Amitava Chattopadhyay, Darren by the framing of attribute info. Journal of Consumer Research, 15,
Dahl, Raymond Darke, Dhruv Grewal, Dale Griffin, Cynthia 374–379.
Levin, Irwin P., Schneider, S. L., & Gaeth, Gary J. (1998). All frames
Holmes, Michael Levy, Leaf Van Boven, Chuck Weinberg,
are not created equal: A typology and critical analysis of framing
and three anonymous reviewers. effects. Organizational Behavior and Human Decision Processes, 76,
149–188.
Liesse, Julie. (1992). How EDLP influences brand equity. Advertising
Age, 63, 17.
References Lichtenstein, Donald R., & Bearden, William O. (1989). Contextual influ-
ences on perceptions of merchant-supplied reference prices. Journal
Baron Reuben, M., & Kenny, A. David. (1986). The moderator-mediator of Consumer Research, 16, 55–66.
variable distinction in social psychological research: Conceptual, Lobb, J. (1997). Stuffing it to Coke: How Pepsi used a promotion to
strategic, and statistical considerations. Journal of Personality and counter its rival’s olympics megablitz. Marketing, 102, 18.
Social Psychology, 51, 1173–1182. Morgan, Bradford W. (1994). It’s the myth of the ‘90s: The value con-
Blattberg, Robert C., & Scott, A. Neslin. (1991). Sales promotion: Con- sumer. Brandweek, 35, 17.
cepts, methods, and strategies. Prentice Hall: Englewood Cliffs, NJ. Ortmeyer, Gwen, Quelch S John, A., & Salmon, Walter. (1991, Fall).
Bond, Cathy. (Feb. 1997). P&G’s gamble: Price or promotion? Marketing, Restoring credibility to retail pricing. Sloan Management Review,
13, 30–33. 55–66.
Campbell, Leland, & Diamond, William D. (1990). Framing and sales Promotion Marketing Association Coupon Council. (2001, August).
promotions: The characteristics of a “Good Deal”. Journal of Con- Promotion Marketing Association Celebrates National Coupon
sumer Marketing, 7, 25–31. Month this September. Coupon Council News. Retrieved from
Compeau, Larry D., & Grewal, Dhruv. (1998). Comparative price adver- http://www.couponmonth.com/pages/article1.htm.
tising: An integrated review. Journal of Public Policy and Marketing, Raghubir, Priya. (2004). Free gift with purchase: Promoting or discount-
17, 257–273. ing the brand? Journal of Consumer Psychology, 14, 181–185.
Darke Peter, R., & Dahl, Darren W. (2003). Fairness and the subjec- Raghubir, Priya, & Corfman, Kim. (1999). When do price promotions
tive value of a bargain. Journal of Consumer Psychology, 13, 328– affect pretrial brand evaluations? Journal of Marketing Research, 36,
338. 211–222.
Darke, Peter R., & Freedman, Jonathan L. (1993). Deciding whether to Raju, P. S. (1977). Product familiarity, brand name and price influences
seek a bargain: Effects of both amount and percentage off. Journal on product evaluation. Advances in Consumer Research, 4, 64–71.
of Applied Psychology, 78, 960–965. Rao, Akshay R., & Monroe, Kent R. (1989). The effect of price, brand
Darke, Peter R., & Freedman, Jonathan L. (1995). Nonfinancial mo- name, and store name on buyer’s perceptions of product quality:
tives and bargain hunting. Journal of Applied Social Psychology, 25, An integrative review. Journal of Marketing Research, 26, 351–
1597–1610. 357.
P.R. Darke, C.M.Y. Chung / Journal of Retailing 81 (1, 2005) 35–47 47

Reda, Susan. (1994). Is E.D.L.P. coming up S.H.O.R.T.? Stores, 76, Thaler, Richard. (1985). Mental accounting and consumer choice. Mar-
22–25. keting Science, 4, 199–214.
Scott, Carol A., & Tybout, Alice M. (1979). Extending the self-perception Tybout, Alice M., & Scott, Carol A. (1983). Availability of well-defined
explanation: The effect of cue salience on behavior. Advances in Con- internal knowledge and the attitude formation process: Information
sumer Research, 6, 50–54. aggregation versus self-perception. Journal of Personality and Social
Scott, Carol A., & Yalch, Richard F. (1980). Consumer response to initial Psychology, 44, 474–479.
product trial: A Bayesian analysis. Journal of Consumer Research, 7, Urbany, Joel E., Bearden, William O., & Weilbaker, Dan C. (1988). The
32–41. effect of plausible and exaggerated reference prices on consumer per-
Simonson, Itamar, Carmon, Ziv, & O’Curry, Suzanne. (1994). Experi- ceptions and price search. Journal of Consumer Research, 15, 95–
mental evidence on the negative effect of product features and sales 110.
promotions on brand choice. Marketing Science, 13, 23–40. Weinstein, Steve. (1992). EDLP: fact and fiction. Progressive Grocer, 77,
Snyder, Mark. (1984). When belief creates reality. Advances in Experi- 50.
mental Social Psychology, 18, 185–208. Wheatley, John, & Chiu, John. (1977). The effects of price, store image,
Stafford, James, & Enis, Ben. (1969). The price-quality relationship: An and product and respondent characteristics on perceptions of quality.
extension. Journal of Marketing Research, 6, 456–458. Journal of Marketing Research, 14, 181–186.

You might also like