The discount is unfair: Egocentric fairness in risky discounts

Journal of Economic Psychology 39 (2013) 32–43

Contents lists available at SciVerse ScienceDirect

Journal of Economic Psychology journal homepage: www.elsevier.com/locate/joep

The discount is unfair: Egocentric fairness in risky discounts Sungchul Choi a,⇑, Sang-June Park b, Chun (Martin) Qiu c, Mike Stanyer a a b c

School of Business, University of Northern British Columbia, 3333 University Way, Prince George V29 4Z9, Canada School of Business, Chonbuk National University, Jeonju 561-156, South Korea Desautels Faculty of Management, McGill University, Bronfman Building, 1001 Sherbrooke Street West, Montreal H3A 1G5, Canada

a r t i c l e

i n f o

Article history: Received 9 August 2012 Received in revised form 20 June 2013 Accepted 30 June 2013 Available online 11 July 2013 JEL classification: C91 D80 PsycINFO classification: 2229 3940 Keywords: Promotional methods Pricing Risky discounts Scratch and save Egocentric fairness Ambiguous price promotions

a b s t r a c t This paper examines how consumers perceive fairness and enjoy the outcome of an emerging risky discount: the retail industry’s gambling or lottery type ‘‘scratch and save’’ (SAS) price promotions, in which the actual discount is determined by chance at the checkout. Risky discounts such as Scratch-and-Save promotions are relatively new retail tools with very little existing research. Previous work in discount claims focuses on perceptions of the offer, whereas this study focuses on perceptions of the outcome and provides managers with information for implementing risky discounts. Across four studies, experiments with a variety of discount levels and reference prices are used to gather consumers’ responses to disparities between suggested and actual discounts. This paper finds that consumers perceive a promotional offering, which is smaller than the reference discount, as more unfair and less enjoyable, even when they still get a deal. However, consumers perceive a discount greater than the reference discount only as fair and enjoyable as the reference discount. Furthermore, poor outcomes are evaluated with severe negativity in terms of perceived fairness and enjoyment, whereas beneficial outcomes garner only moderate positivity, which is consistent with the combination of prospect theory and egocentric fairness. Nevertheless, perceptions of increased unfairness arising from risky discounts have no negative bearing on future purchase intentions, but rather result in positive attitudes and optimism for future risky discounts. Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction Retailers frequently use price promotions to boost store traffic and sales (Gupta & Cooper, 1992). Adding to this practice is a new format of price promotion – risky discounts, which are usually offered via ‘‘Scratch and Save’’ (hereafter SAS) cards that work in a similar way as lottery cards. Retailers offer ambiguous discounts on the face of SAS cards such as ‘‘Scratch and Save Up to 50% Off’’; the actual discount can only be determined by chance when cashiers scratch the card upon checkout (Kamleitner, Mandel, & Dhami, 2011). As adding uncertainty to price promotions is considered a cost effective promotional tool (Goldsmith & Amir, 2010), risky discounts are gaining popularity among retailers (Choi & Kim, 2007; Kamleitner et al., 2011). For example, department stores such as Kohl’s and Sears often offer SAS promotions; category specialists such as Rona and Toys ‘‘R’’ Us frequently hold SAS events; and extreme-value retailers such as GUESS Factory Stores implement SAS promotions from time to time. When consumers are offered price promotions, their perceived fairness of the offers may be influenced by the promotion outcome (distributive fairness) and/or the procedure that leads to the outcome (procedural fairness) (e.g., Kukar-Kinney, Xia, ⇑ Corresponding author. Tel.: +1 250 960 5107. E-mail addresses: [email protected] (S. Choi), [email protected] (S.-J. Park), [email protected] (C.(M.) Qiu), [email protected] (M. Stanyer). 0167-4870/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.joep.2013.06.007

S. Choi et al. / Journal of Economic Psychology 39 (2013) 32–43

33

& Monroe, 2007; Xia, Monroe, & Cox, 2004). In many situations, both the distributive fairness and procedural fairness matter. For example, when consumers evaluate a price-matching guarantee policy, consumers’ price fairness perceptions are influenced by the final price as well as the perceived fairness of the policy (Kukar-Kinney et al., 2007). In the context of risky discounts, however, because consumers know the outcome is determined by chance, they are less concerned with procedural fairness (Kamleitner et al., 2011) and more concerned with the outcome. In addition, risky discounts usually lead consumers to implicitly expect to receive some level of discount and accordingly to form a reference point for discount. Consequently, the revealed discounts will generate various degrees of deviations from the reference discount. Thus, risky discounts offer an opportunity to study the effect of the distributive fairness caused by variations of reference on consumers’ responses. This paper intends to investigate consumers’ perceived (un)fairness in risky discounts in the specific format of SAS promotions. Price unfairness has been shown to have negative consequences on retailers (e.g., Blakely, 2007; Campbell, 1999). SAS promotions, if perceived less fair, could dilute the surprises in the promotions and the enjoyment of lottery-like outcomes, which are the reason why retailers implement SAS promotions in the first place. In addition, SAS promotions may backfire if the perception of low fairness undermines consumers’ purchase intentions in the future. It is thus important to examine the extent to which the magnitude of differences between reference discounts and actual discounts affects fairness perceptions in these risky discounts, and future purchase intentions. This paper contributes to the literature on consumers’ psychological responses to price promotions. Literature has examined how various ambiguous price promotions influence consumers’ perceptions. For example, much prior research has documented consumers’ responses to ambiguity inducing ‘‘tensile price claims’’ (Biswas & Burton, 1993, 1994; Dhar, Gonzalez-Vallejo, & Soman, 1999; Mobley, Bearden, & Teel, 1988). A few recent papers have looked at SAS promotions (Choi & Kim, 2007; Choi, Stanyer, & Kim, 2010; Kamleitner et al., 2011). Furthermore, Goldsmith and Amir (2010) have acknowledged the relative effectiveness of uncertain promotions compared to certain promotions. However, previous studies only focus on the effect of the price promotion claims in consumers’ responses. In comparison, this paper focuses on the outcome of price promotions and examines the consequences of risky discounts. Furthermore, this paper identifies conditions when risky discounts, despite providing a deal to consumers, are perceived as less fair and less enjoyable. The paper proceeds as follows. In the next section, we review the relevant literature and develop the hypotheses. In the subsequent sections, we introduce four experimental studies and present the results. We conclude the paper in Section 7. 2. Theory and hypotheses development Much of the existing research on fairness and pricing deals with price changes. A subtle difference in the case of SAS promotions is that the regular price remains unchanged, though there will always be some variation in the discount. Despite SAS promotions being a novel tool with little formal examination, existing literature on fairness of pricing practice still provides a basis to work from. According to previous research, price reductions are often perceived to be more fair and price increases to be less fair; and consumers evaluate fairness in a subjective manner (e.g., Bolton & Alba, 2006; Bolton, Warlop, & Alba, 2003; Campbell, 1999, 2007; Kahneman, Knetsch, & Thaler, 1986; Kukar-Kinney et al., 2007; Xia et al., 2004). The study of SAS promotions departs from other work on the topic of pricing and fairness which has examined deviations from a reference price. In comparison, the present study examines deviations from a suggested discount within a range: Attention is placed on the discount amount instead of on the final price. The reference discount provides a basis for evaluating the outcome of the deal that measures consumers’ reactions to higher or lower than expected results. It also serves as a mechanism for prompting deviations: the reference discount is expected to prompt fairness judgments by providing an expected discount that might be different than the actual outcome, thus operating somewhat like a reference price (e.g., Maxwell, 2002). Given the possibility of discount deviations, this paper intends to answer the following questions: Will consumers perceive the outcome of SAS promotions to be homogeneously fair regardless of the discount depth? Or instead, will consumers perceive the outcome to be more or less fair depending on whether the final discount is better or worse than the reference discount? In an SAS promotion, consumers are aware of the discount range before making the transaction. In absolute terms, buyers cannot suffer a loss, regardless of the discount outcome, because they will, at worst, receive the regular price. In the game proposed by the discount range, however, buyers could consider a-lower-than-reference outcome to be less fair. In that case, a worse-than-reference deal would be considered a loss despite providing a discount, whereas a better-than-reference deal would be considered a gain. Thus, we will consider such a fairness perception to be biased toward buyers’ own interests, which is egocentric. 2.1. Egocentric fairness Past research has demonstrated that consumers perceive fairness with a preferential or egocentric motive: Maxwell, Nye, and Maxwell (1999) find that consumers who are prompted to consider social fairness make large concessions in price negotiation, but without prompting for fairness, they pursue only the most personally advantageous price. In a study on cheating in a contest, Krehbiel and Cropanzano (2000) find that less unfairness is reported so long as the outcome benefits the respondents. Furthermore, consumers may deviate from the socially fair model of balanced inputs and outputs if they perceive the

34

S. Choi et al. / Journal of Economic Psychology 39 (2013) 32–43

exchange to be a competition, such as the case of negotiating the price of a car (Oliver & Swan, 1989) or a ‘‘winner takes all’’ gamble, which can be contingently acceptable (Larwood, Kavanagh, & Levine, 1978). This finding is particularly relevant to our study. This is because an SAS promotion employs elements of gambling: It presents participants with risk and a variable outcome; losses in games of chance are known to arouse feelings of unfairness (Ward & Hill, 1991), and gains are perceived positively. In addition, prospect theory states asymmetric perceptions of positive and negative outcomes, with unfavorable outcomes being evaluated more negatively than favorable outcomes being evaluated positively (Kahneman & Tversky, 1979). Thus, if the reference outcome is 15% off, a 10% discount (5% negative deviation) would generate a disproportionately adverse reaction compared to the positive reaction garnered by a 20% discount (5% positive deviation), even though the deviation in both cases is 5%. In other words, the loss function is steeper that the gain function (Kahneman & Tversky, 1979; Thaler, 1985). Accordingly, we hypothesize H1. A greater discount leads to a perception of greater fairness, although the slope of the positive discounts is less steep than the negative discounts. 2.2. Enjoyment By its similarity to a gamble, SAS promotions also evoke the more general hedonic value of enjoyment. Given that intrinsic value is thought to be a motivator for participating in promotional games (e.g., Klingemann, 1995), it is interesting to examine how the outcome of the risky promotion influences the enjoyment of the ‘‘gamble’’ in the promotion. Finding and, particularly, receiving a deal can be enjoyable, even if the amount saved is trivial. While monetary benefits are a major motivation for some gamblers, a winning result in itself can provide enjoyment. For example, Neighbors, Larimer, Lostutter, and Cronce (2001) conducted a comprehensive study to examine motivation for gambling among college student gamblers and found that winning money was the primary motivation. Thus, less favorable game outcomes will diminish the level of enjoyment. In addition, a psychological source of value, derived from having received a deal, is said to arise primarily from perceptions of the fairness of the offer (Darke & Dahl, 2003). Therefore, receiving a lower than reference deal (i.e., a loss) leads to a decrease in enjoyment, whereas receiving a greater than reference deal (i.e., a win) leads to an increase in enjoyment as fairness. Additionally, as prospect theory is applied to fairness perceptions, the slopes of the loss and gain functions will show asymmetrical patterns. Accordingly, we hypothesize H2. A greater discount leads to a perception of greater enjoyment, although the slope of the positive discounts is less steep than the negative discounts. 2.3. Purchase intentions The end application of testing H1 and H2 is to examine how SAS promotions’ outcomes influence post purchase behavior. Expectancy disconfirmation has been found to influence post purchase behavior regarding product and service performance (i.e., Homberg, Koschate, & Hoyer, 2006; Smith & Bolton, 2002; Zeelenberg & Pieters, 2004). In an SAS promotion, the reference discount sets up an expectation which might be disconfirmed positively or negatively by the actual outcome. Thus, we expect that below-reference deal outcomes (negative disconfirmation) decrease future purchase intentions, whereas abovereference deal outcomes (positive disconfirmation) increase future purchase intentions. Therefore, with an expression of prospect theory as in perceived fairness and enjoyment, we hypothesize H3. A greater discount leads to a higher future purchase intention at the store, although the slope of the positive discounts is less steep than the negative discounts.

3. Study 1: Perceptions of fairness and enjoyment We conduct three experiments to test hypotheses H1 and H2 on consumers’ perceptions of fairness and enjoyment. 3.1. Experiment 1A: Method Seventy-nine undergraduate business students at a North American university participated in Experiment 1A in exchange for partial course credit and entry to win cash prizes. Of the participants, 33 (41.8%) were female and 46 (58.2%) were male with an average age of 22.3 years. They were provided with a web address and asked to complete an online questionnaire. The questionnaire presented a scenario in which participants were asked to imagine that they were purchasing a digital camera at a store running the following SAS promotion: ‘‘Scratch and Save up to 50% off on all merchandise in the store.’’ The ad further indicated that consumers would get, on average, 15% off. The experimental design is a single between-subjects factor with three levels of discounts (lower than the advertised average discount, the same as the advertised average discount, and

35

S. Choi et al. / Journal of Economic Psychology 39 (2013) 32–43 Table 1 Means of major variables in Study 1 Experiment 1 A.a Low discount (10%) (Condition 1) Perceived discount Perceived fairness Before enjoyment After enjoyment a b

2.19 3.58 4.59 3.69

(1.02)b (1.72) (1.52) (1.24)

Average discount (15%) (Condition 2)

High discount (20%) (Condition 3)

3.92 5.26 4.20 4.42

4.84 5.12 4.54 4.88

(.69) (1.16) (1.69) (1.35)

(1.28) (1.21) (1.21) (1.16)

This experiment offers up to 50% off and the stated average discount is 15% off. Standard deviations in parentheses.

higher than the advertised average discount). Participants were randomly assigned to one of three conditions: (1) low discount: 5% less discount (i.e., 10%) than the advertised average discount (15%); (2) average discount: the same discount (i.e., 15%) as the advertised average discount; and (3) high discount: 5% more discount (i.e., 20%) than the advertised average discount. The advertised average discount level, 15%, was chosen based on Della Bitta and Monroe (1980) and Gupta and Cooper (1992), which was enough to attract consumers to a sale. The manipulation of 5% difference across conditions is consistent with previous studies (e.g., Niedrich, Sharma, & Wedell, 2001; Rao, Arjunji, & Murthi, 1995), and is very close to the 6% reported by Kalwani and Yim (1992, p. 96), which is a ‘‘just noticeable difference.’’ The questionnaire included measures of enjoyment of the promotion (before and after scratching the SAS card) and the perceived fairness of the discount outcome. Enjoyment was measured before (i.e., How much do you think you would enjoy participating in the Scratch and Save promotion?) and after revealing the discount level (i.e., How much do you think you would enjoy participating in the Scratch and Save promotion after seeing your savings outcome?), respectively, as an average of three 7-point scale items: ‘‘enjoyable’’ (1 = not at all enjoyable, 7 = very enjoyable); ‘‘fun’’ (1 = not at all fun, 7 = very fun); and ‘‘like’’ (1 = dislike, 7 = like). This measure was modified from Raghunathan and Corfman (2006) and showed a sufficiently high coefficient alpha of 0.94 for both before and after enjoyment, respectively. The perceived fairness of the discount was measured via a 7-point scale item (i.e., I think the SAS savings outcome I received is: 1 = very unfair, 7 = very fair) by adapting a scale from Campbell (1999). Participants then completed a manipulation check for the level of discount they received from the SAS promotion (i.e., How much did you actually get from the Scratch and Save promotion in the experiment? 1 = less than the average, 7 = greater than the average), which was modified from Campbell (2007). 3.2. Experiment 1A: Results The manipulation of the perceived discount level was significant and in the expected direction, with the high discount condition receiving a rating of 4.84; the average discount condition receiving a rating of 3.92; and the low discount condition receiving a rating of 2.19 (F(2, 76) = 44.34, p < .01, g2 = .55).1 Contrasts among the three levels were significant: High discounts were perceived higher than low discounts (p < .01); high discounts were perceived higher than average discounts (p < .01); and low discounts were perceived lower than average discounts (p < .01). Means of major variables appear in Table 1. 3.2.1. Perceived fairness A one-way ANOVA showed that the perceived fairness was significantly influenced by the discount outcome (F(2, 76) = 11.85, p < .01, g2 = .25). Follow-up planned comparisons showed that ‘‘low’’ discounts were perceived as less fair than ‘‘average discounts’’ (Mlow = 3.58 vs. Maverage = 5.26; t(51) = 4.18, p < .01) and ‘‘high discounts’’ (Mlow = 3.58 vs. Mhigh = 5.12; t(50) = 3.72, p < .01). The difference in the perceived fairness between ‘‘average’’ and ‘‘high’’ conditions was not significant (Maverage = 5.26 vs. Mhigh = 5.12; t(51) = .44, p = .66). Hence, H1 is supported only in the negative comparison. Yet, the result demonstrates an asymmetric effect as predicted in H1. 3.2.2. Enjoyment A 3 (discount level: low, average, and high)  2 (enjoyment: before and after) mixed ANOVA, with discount level representing a between-subjects factor and enjoyment a within-subjects variable, showed a significant interaction effect between discount level and enjoyment (F(2, 76) = 6.13, p < .01, g2 = .14). A follow-up comparison showed that, when the discount level was revealed, the enjoyment level was significantly different across discount conditions (F(2, 76) = 5.97, p < .01, g2 = .14). Planned comparisons revealed that ‘‘low’’ discounts were perceived as less enjoyable than ‘‘average’’ (Mlow = 3.69 vs. Maverage = 4.42; t(51) = 2.04, p < .05) and ‘‘high’’ (Mlow = 3.69 vs. Mhigh = 4.88; t(50) = 3.58, p < .01) discounts. Furthermore, comparisons across the two enjoyment measures within each discount condition showed that the enjoyment level dropped significantly after the discount level was revealed only for the low-discount condition (Mbefore = 4.59 vs. Mafter = 3.69; t(51) = 2.51, p < .05), showing an asymmetric pattern. These results partly support H2.

1 We also compared actual perceived discount values that participants chose to the scale midpoint to check the robustness of the manipulation checks. We found that only two participants in condition 3 chose below 4, which is not consistent with our manipulation. This confirms the successful manipulation.

36

S. Choi et al. / Journal of Economic Psychology 39 (2013) 32–43

Table 2 Means of major variables in Study 1 Experiment 1 B.a

Perceived fairness Before enjoyment After enjoyment a b

10% Discount (Condition 1)

15% Discount (Condition 2)

20% Discount (Condition 3)

3.82 (1.65)b 3.97 (1.22) 4.40 (1.66)

3.38 (1.50) 4.21 (1.29) 4.81 (1.09)

3.41 (1.28) 4.28 (.91) 5.33 (.96)

This experiment offers up to 50% off. Standard deviations in parentheses.

3.3. Experiments 1B and 1C Experiment 1A showed how the level of risky discount outcomes influences perceived fairness and enjoyment; however, the advertisement used in Study 1A provided two possible reference discounts: ‘‘Scratch and Save up to 50% off ‘‘and ‘‘the average discount of 15% off’’. Thus, it remains a question as to which discount is the reference: the ‘‘up to 50% off’’ or ‘‘the average of 15% off’’ discount. To answer this, we separately examined the impact of each reference in consumer responses with two additional experiments: 1B and 1C. In Experiment 1B, we considered only ‘‘up to 50% off,’’ whereas, in Experiment 1C, we considered only ‘‘the average of 15% off’’. The procedure for Experiments 1B and 1C was the same as for Experiment 1A. In Experiment 1B, the questionnaire presented a scenario in which participants were asked to imagine that they were purchasing a digital camera at a store running the following scratch and save promotion: ‘‘Scratch and Save up to 50% off on all merchandise in the store.’’ Unlike Study 1, the ad was silent on the average discount (15% off). Eighty-four undergraduate business students were randomly assigned to one of three conditions: (1) low discount: 10% discount; (2) moderate discount: 15% discount; and (3) high discount: 20% discount. Of the participants, 31 (36.9%) were female and 53 (63.1%) were male. Their average age was 22.1 years. The measures of before (a = .87) and after enjoyment (a = .93) and perceived fairness of the discount outcome were the same as in Experiment 1A. We found that the perceived fairness was not different across the three conditions (see Table 2). A 3 (discount level: low, moderate, and high)  2 (enjoyment: before and after) mixed ANOVA, with discount level representing a between-subjects factor and enjoyment a within-subjects variable, showed no significant interaction effect (F(2, 81) = 1.28, p = .28, g2 = .03). Follow-up comparisons showed that the enjoyment level increased significantly after the discount level was revealed only for the high-discount condition (Mbefore = 4.28 vs. Mafter = 5.33; t(52) = 4.12, p < .01). These results are different from those of Study 1 and indicate that ‘‘Up to 50% off’’ is too vague for consumers to use as a reference point. Experiment 1C provided only the average discount information (15% off) of the SAS promotion. Seventy-two undergraduate business students were randomly assigned to one of three conditions: (1) low discount: 5% less discount (i.e., 10%) than the advertised average discount (15%); (2) average discount: the same discount (i.e., 15%) as the advertised average discount; and (3) high discount: 5% more discount (i.e., 20%) than the advertised average discount. Of the participants, 30 (41.7%) were female and 42 (58.3%) were male. Their average age was 22.59 years. Before (a = .87) and after enjoyment (a = .96), perceived fairness of the discount outcome, and perceived discount were measured as in Experiment 1A. Unlike Experiment 1B, Experiment 1C yielded the same results as Experiment 1A (see Table 3). The manipulation of the perceived discount level (compared to the advertised average discount) was significant and showed the same pattern as in Experiment 1A.2 A one-way ANOVA showed that the perceived fairness was strongly influenced by the discount outcome (F(2, 69) = 19.39, p < .01, g2 = .36). Furthermore, ‘‘low’’ discounts were perceived as less fair than ‘‘average discounts’’ (Mlow = 3.54 vs. Maverage = 5.55; t(46) = 4.44, p < .01) and ‘‘high discounts’’ (Mlow = 3.54 vs. Mhigh = 5.83; t(48) = 5.68, p < .01). However, the difference in the perceived fairness between ‘‘average’’ and ‘‘high’’ conditions was not significant (Maverage = 5.55 vs. Mhigh = 5.83; t(44) = .78, p = .44). In addition, a 3 (discount level: low, average, and high)  2 (enjoyment: before and after) mixed ANOVA, with discount level representing a between-subjects factor and enjoyment a within-subjects variable, showed a significant interaction effect between discount level and enjoyment (F(2, 69) = 6.06, p < .01, g2 = .15). Follow-up comparisons showed that the enjoyment level dropped significantly after the discount level was revealed only in the low-discount condition (Mbefore = 4.22 vs. Mafter = 3.38; t(50) = 3.21, p < .01), suggesting that it is the advertised average discount (i.e., 15% off) that acts as the reference point. 3.4. Discussion Study 1 provides an initial demonstration of how a deviation from the reference discount affects perceptions of fairness and enjoyment. That is, participants perceived discounts below the reference discount as less fair and less enjoyable, but perceived discounts above the reference discount as much fair and as enjoyable as the reference discount.

2 As in Experiment 1A, we compared actual perceived discount values that participants chose to the scale midpoint to verify the robustness of the manipulation checks. We found that only one participant in conditions 1 and 3, respectively, evaluated the perceived discount as inconsistent with the manipulation.

37

S. Choi et al. / Journal of Economic Psychology 39 (2013) 32–43 Table 3 Means of major variables in Study 1 Experiment 1 C.a 10% Discount (Condition 1) Perceived discount Perceived fairness Before enjoyment After enjoyment a b

2.77 3.54 4.22 3.38

(1.90)b (1.68) (.78) (1.07)

15% Discount (Condition 2)

20% Discount (Condition 3)

3.95 5.55 4.39 4.54

5.08 5.83 4.42 4.89

(.58) (1.41) (1.25) (1.24)

(1.93) (1.09) (1.05) (1.74)

This experiment provides a stated average discount of 15%. Standard deviations in parentheses.

In Study 1, we considered only one (dis)advantageous deviation. Thus, in Study 2, we consider multiple levels of (dis)advantageous discount outcomes to further examine this issue. In addition, we examine how (un)fair discount outcomes influence consumers’ future purchase intentions. 4. Study 2: Multiple outcome levels of SAS promotions 4.1. Method The procedure and buying situation of Study 2 were the same as in Study 1. Two hundred and twenty-four undergraduate business students were randomly assigned to the conditions of a 2 (reference discount: low [when the advertised average discount is 15%] vs. high [when the advertised average discount is 25%])  2 (deviation: advantageous discount vs. disadvantageous discount)  2 (deviation level: low [5%] vs. high [10%]) between-subjects design. 104 participants (46.4%) were female and 120 (53.6%) were male. Their average age was 21.67 years. In addition to the 15% advertised average discount used in Study 1, a 25% average discount was chosen as an additional reference discount level. We did this to determine whether consumers based their fairness judgments on the deviation that the test introduces (subjective), or on the actual price outcome (objective). By choosing these two reference discounts, the consumers’ subjectivity and objectivity are made mutually exclusive: The lower reference discount’s positive deviation overlaps the higher reference discount’s negative deviation. For example, a 20% discount was expected to be more fair with a 15% reference discount (a 5% gain) but less fair with a 25% reference discount (a 5% loss). As we expect consumers to be more concerned with the increment of gain or loss than the overall outcome, we do not expect their perceived level of fairness to differ between same increments of deviation (i.e., -5%) across different reference discounts. We conducted a pretest of 32 participants to calibrate the manipulation of the reference discount levels. Participants assessed their perceptions toward the four advertised average discount levels, such as 15%, 20%, 25%, and 30%, using a 7-point scale item (1 = low, 7 = high). We found a significant difference between the 20% and 25% discounts (M25% advertised = 4.47 vs. M20% advertised = 3.66; t(61) = 2.29, p < .05). However, no significant difference was observed between the two adjacent cases (i.e., 15% vs. 20%, and 25% vs. 30%). This indicates that a 25% discount could be perceived significantly higher than a 15% discount. With respect to the discount deviation manipulation, participants were told they received (less) more discounts than the advertised average discount in a (dis) advantageous discount condition. Unlike Study 1, however, the average discount condition, where the revealed discount level was the same as the advertised average discount level, was not considered in order to stress the role of discount deviations in consumer responses after revealing the actual discount. With respect to the deviation level manipulation, participants in low-deviation conditions were told they received a discount of 5% more or less than the advertised average discount, whereas participants in high-deviation conditions were told they received a discount of 10% more or less than the advertised average discount. Before (a = .93) and after enjoyment (a = .96), perceived fairness of the discount outcome, and perceived discount (for manipulation checks) were measured, as in Study 1. Participants also indicated the likelihood (How likely is it that you would. . .? 1 = most likely, 7 = least likely) of their intention not to shop at the store (in the future) and not to participate in such promotions (at this store in the future), modified from Singh (1988) (r = .67). We also measured the perceived advertised discount using a 7-point scale item (1 = low, 7 = high), as we did in the pretest. 4.2. Results We checked the manipulation of perceived discount using a 2 (reference discount: low vs. high)  2 (deviation: advantageous discount vs. disadvantageous discount)  2 (deviation-level: low vs. high) between-subjects design.3 As anticipated, we observed a main effect of deviation, with higher levels of perceived discounts in the advantageous (M = 4.74) than in the disadvantageous (M = 1.98) conditions (F(1, 216) = 258.32, p < .01, g2 = .548). In addition, the interaction of anticipated deviation and deviation level was significant (F(1, 216) = 8.38, p < .01, g2 = .04), indicating a significant decrease in the perceived 3 As in Study 1, we compared actual perceived discount values that participants chose to the scale midpoint to verify the robustness of the manipulation checks. We found that only 9 out of 224 participants evaluated the perceived discount as inconsistent with the manipulation. This confirms the successful manipulation in this study.

38

S. Choi et al. / Journal of Economic Psychology 39 (2013) 32–43

Table 4 Means of dependent variables in Study 2. 5% Discount when 15% advertised Perceived discount Perceived fairness Before enjoyment After enjoyment Post behavior a

1.75 3.50 4.62 3.48 2.96

(.75)a (1.27) (1.36) (1.67) (1.69)

10% Discount when 15% advertised

20% Discount when 15% advertised

25% Discount when 15% advertised

15% Discount when 25% advertised

20% Discount when 25% advertised

30% Discount when 25% advertised

35% Discount when 25% advertised

2.24 4.24 4.72 4.05 2.88

4.46 5.15 4.69 5.47 2.58

5.00 5.64 4.64 5.59 2.80

1.73 4.04 4.62 3.83 2.62

2.19 4.54 4.59 4.72 2.85

4.50 4.88 4.59 5.26 2.73

5.00 5.19 4.56 5.79 2.71

(1.36) (1.41) (1.15) (1.24) (1.71)

(.91) (1.06) (1.20) (.99) (1.43)

(1.41) (1.18) (1.46) (1.37) (1.65)

(.60) (1.46) (1.38) (1.53) (1.52)

(1.50) (1.29) (1.47) (1.32) (1.68)

(1.57) (1.59) (1.41) (1.37) (1.67)

(1.47) (1.51) (1.49) (.98) (1.37)

Standard deviations in parentheses.

discount level as disadvantageous deviations become bigger. No other significant main and interaction effects were observed. These results suggest the effectiveness of the manipulation. Means of major variables appear in Table 4. 4.2.1. Perceived fairness and enjoyment Replicating the findings of Study 1, we found that the perceived fairness and enjoyment levels were strongly influenced by the actual discount. An ANOVA on perceived fairness toward the actual discount revealed the only main effect of deviation, with higher levels of perceived fairness in the advantageous (M = 5.22) than in the disadvantageous (M = 4.08) conditions (F(1, 216) = 29.97, p < .01, g2 = .12). The interaction of deviation and deviation-level was significant (F(1, 216) = 6.01, p < .05, g2 = .03). In particular, a 10% less discount than the advertised average discount was perceived as less fair than a 5% less discount (M10% less = 3.76 vs. M5% less = 4.39; t(103) = -2.05, p < .05). However, when the actual discount was advantageous, there was no difference in perceived fairness between ‘‘10% more’’ and ‘‘5% more’’ discounts (M10% more = 5.41 vs. M5% more = 4.99; t(117) = 1.55, p = .13). Hence, these results demonstrate the asymmetric effect of SAS promotion outcomes on fairness perception, and partially support H1. A 2 (reference discount: low and high)  2 (deviation: advantageous and disadvantageous discounts)  2 (deviation level: low and high)  2 (enjoyment: before and after) mixed ANOVA, with reference discount, deviation, and deviation level representing a between-subjects factor and enjoyment a within-subjects variable, showed significant interaction effects of deviation and enjoyment (F(1, 216) = 51.21, p < .01, g2 = .19) as well as deviation, deviation level, and enjoyment (F(1, 216) = 6.12, p < .05, g2 = .03). A follow-up analysis on after enjoyment revealed the main effect of deviation, with higher levels of after enjoyment in the advantageous (M = 5.49) than in the disadvantageous (M = 4.01) conditions (F(1, 216) = 69.75, p < .01, g2 = .24). In addition, we observed a significant interaction of deviation and deviation-level (F(1, 216) = 8.50, p < .01, g2 = .04). Further comparisons suggested that, like perceived fairness, a 10% less discount was perceived less enjoyable than a 5% less discount (M10% less = 3.65 vs. M5% less = 4.39; t(103) = -2.60, p < .05). When the actual discount was advantageous, the difference in enjoyment between ‘‘10% more’’ and ‘‘5% more’’ discounts was insignificant (M10% more = 5.69 vs. M5% more = 5.34; t(117) = 1.56, p = .12). We thus found a similar asymmetric effect, which partially support H2. 4.2.2. Future purchase intentions An ANOVA on future purchase intentions showed neither main effects nor interaction effects, which is inconsistent with H3. We further conducted separate ANOVAs on each variable of future purchase intentions, such as intention not to shop at the store and not to participate in such promotions at this store in the future, and found no significant effects. This result indicates that the actual discount level does not influence consumers’ intentions to buy from the same retailer in the future. 4.3. Content analysis We analyzed open-ended questions to further examine the reasons for the lack of effect from discounts below the advertised average discount on future purchase intentions. Two independent coders categorized the open-ended thoughts into the following four categories for the disadvantageous discounts: (1) responses indicating that the discount was a bonus because it was more than nothing (Positive 1: Better than Nothing); (2) responses indicating that a better discount is expected in the future (Positive 2: Better in the Future); (3) responses indicating that the discount was disappointing (Negative 1: Disappointment); and (4) responses uninformative of the experimental manipulation (Uninformative). The following four categories were used for the advantageous discounts: (1) responses indicating that the participants would be satisfied with the discount, thereby they would be more inclined to return (Positive 3: Luck Again); (2) responses indicating that participants felt lucky but do not expect it to happen every time (Negative 2: No More Luck); (3) responses indicating that the discount does not really matter (Indifferent); and (4) responses uninformative of the experimental manipulation (Uninformative). The percent agreement between the two judges was 93% and 95% for disadvantageous and advantageous discounts, respectively, and differences in coding were resolved in subsequent discussion. A series of chi-square tests indicate that the distribution of the response frequencies in the four categories is not significantly different across both the 5% (and 10%) discount conditions and 15% (and 25%) discount conditions. Thus, we aggregate the four (dis)advantageous discount conditions into one (dis)advantageous case.

S. Choi et al. / Journal of Economic Psychology 39 (2013) 32–43

39

Overall, for the disadvantageous discount conditions, 25.2% of participants were classified with Positive 1: Better than Nothing; 10.3% were classified with Positive 2: Better in the Future; 15.0% were classified with Negative 1: Disappointment. These results indicate that a significant number of participants (35.5%) interpret the less fair discount outcomes positively, grateful for a no worse outcome and with expectations of a better outcome next time. For the advantageous discount conditions, 23.1% of participants were classified with Positive 3: Luck Again; 9.4% were classified with Negative 2: No More Luck; and 4.3% were classified with Indifferent. These results indicate that only 23.1% of participants optimistically expected to have a continuous advantageous discount in the next shopping trip. 4.4. Discussion Study 2 shows the same pattern as Study 1 with different severities of high and low discounts: the higher the magnitude of negative discount outcomes, the lower the perceived fairness and enjoyment. Nevertheless, the magnitude of positive discount outcomes does not influence consumers’ perceived fairness and enjoyment. In addition, we find that the negative outcomes of SAS promotions do not affect participants’ future purchase intentions. On the contrary, we find through a content analysis that a significant number of participants who received disadvantageous discounts hold positive expectations for future SAS promotions. So far, we have focused on the role of external reference discounts. Next, we turn our attention to the internal reference discount to check the robustness of our findings. We investigate a scenario where consumers may easily ignore the odds of each possible discount rate and rely on internal reference discounts. 5. Study 3: Internal reference discount 5.1. Method The procedure and buying situation were essentially the same as in Study 2. One hundred undergraduate business students were randomly assigned to the conditions of a 2 (deviation: advantageous discount vs. disadvantageous discount)  2 (deviation-level: low [±5%] vs. high [±10%]) between-subjects design. Of the participants, 46 (46%) were female and 54 (54%) were male. Their average age was 21.78 years. Unlike Studies 1 and 2, we did not provide an external reference discount. Instead, we measured participants expected discount toward the SAS promotion before revealing the discount outcome to them. Participants were free to respond to the internal reference discount with either a point estimate or a range. However, we found that only 4 (4%) of the 100 responses were expressed as range estimates, which is consistent with previous research (e.g., Dhar et al., 1999). We repeated our analyses with and without the four range estimates and found no differences in the basic pattern of results. Thus, we reported results including the range estimates. The average expected discount was 17% off. Before we measured their expected discount, participants were asked to indicate their enjoyment level toward the SAS promotion (i.e., before enjoyment) by using the same three 7-point scale items used in Studies 1 and 2 (a = .91). Once participants indicated their expected savings level, they were informed of the actual discount level, which was 5% (or 10%) lower (or greater) than their expected discounts. For example, if a participant expected to have a 20% discount, the revealed discount was 25% (15%) in the low (dis)advantageous deviation condition. Then, participants were asked to evaluate their enjoyment (a = .93), perceived fairness, perceived discount level of the actual discount, and future purchase intention (r = .68), as was done in Study 2. 5.2. Results We first checked whether participants’ expected discounts provided unreasonable discount outcomes. For example, if a participant expected a 5% discount in the highly disadvantageous discount condition, the revealed discount would be 5% off, which is not reasonable. We found 13 such responses and dropped them from the subsequent analyses. We assessed the manipulation checks using a 2 (deviation: advantageous discount vs. disadvantageous discount)  2 (deviation-level: low vs. high) ANCOVA, using the expected discount as a covariate. We observed the deviation main effect, such that participants showed higher levels of perceived discounts in the advantageous (M = 4.63) than in the disadvantageous (M = 2.39) conditions (F(1, 82) = 63.38, p < .01, g2 = .44). In addition, the deviation  deviation-level interaction was observed (F(1, 82) = 7.42, p < .01, g2 = .08), indicating a significant increase (decrease) in the perceived discount level as the level of (dis) advantageous deviation increases. These results indicate the effectiveness of the manipulation (see Table 5 for means of major variables in Study 3).4 A two-way ANCOVA on perceived fairness toward the actual discount, using the expected discount as a covariate, showed the main effect of deviation, with higher levels of perceived fairness in the advantageous (M = 4.89) than in the disadvantageous (M = 3.71) conditions (F(1, 82) = 17.86, p < .01, g2 = .31). The covariate of the expected discount was also significant

4 As in the previous studies in the paper, we compared actual perceived discount values to the scale midpoint to check the robustness of the manipulation checks. We found that only 2 out of 100 participants evaluated the perceived discount as inconsistent with the manipulation. This confirms the successful manipulation in this study.

40

S. Choi et al. / Journal of Economic Psychology 39 (2013) 32–43

Table 5 Means of dependent variables in Study 3. 10% Less discount than expected Perceived discount Perceived fairness Before enjoyment After enjoyment Post behavior a

2.06 3.61 5.18 3.98 2.89

(0.94)a (1.34) (1.10) (1.57) (1.37)

5% Less discount than expected

5% More discount than expected

10% More discount than expected

2.65 3.78 4.22 3.54 3.20

4.17 4.92 4.90 4.76 3.15

5.14 4.86 4.82 5.20 3.32

(1.53) (1.62) (1.53) (1.46) (1.62)

(1.55) (1.44) (1.55) (1.42) (2.10)

(1.04) (1.39) (1.39) (1.39) (1.77)

Standard deviations in parentheses.

(F(1, 82) = 6.38, p < .05, g2 = .10). However, unlike Study 2, the interaction of deviation and deviation-level was not observed. This severity difference between internal and external reference discounts is consistent with Kumar et al.’s (1998) finding that consumers are less sensitive to internal reference prices than to external reference prices. These results together indicate that, although the perceived fairness was strongly influenced by the direction of discount deviations, the difference between two (dis)advantageous discount outcomes did not significantly affect perceived fairness. A 2 (deviation: advantageous and disadvantageous discounts)  2 (deviation level: low and high)  2 (enjoyment: before and after) mixed ANCOVA, with deviation and deviation level representing a between-subjects factor and enjoyment a within-subjects variable, showed a significant interaction of deviation and enjoyment (F(1, 82) = 12.89, p < .01, g2 = .13). A followup ANCOVA on after enjoyment, similar to the one on perceived fairness, revealed the significant main effect of deviation, with higher levels of after enjoyment in the advantageous (M = 4.97) than in the disadvantageous (M = 3.73) conditions (F(1, 82) = 19.55, p < .01, g2 = .19). In addition, the expected discount covariate was significant (F(1, 82) = 8.15, p < .01, g2 = .01). These results are consistent with the findings on perceived fairness. Consistent with Study 2, an ANCOVA on future purchase intentions showed neither main effects nor interaction effects; neither did separated ANCOVAs on each future purchase intention variable reveal any significant effects. We also analyzed the open-ended thought data for Study 3 using the same procedure and categories as in Study 2. The two judges agreed on 93% and 95% of the cases for disadvantageous and advantageous discounts, respectively, and differences in coding were resolved in subsequent discussion. A series of chi-square tests indicate that the distribution of the response frequencies in the four categories is not significantly different across the 5% and 10% discount conditions. Thus, we combined the 5% and 10% more (less) discount conditions to form the (dis)advantageous condition. Overall, for the disadvantageous discount conditions, 26% of participants were classified with Positive 1: Better than Nothing; 8% were classified with Positive 2: Better in the Future; and 16% were classified with Negative 1: Disappointment. These results indicate that a significant number of participants (34%) interpret the less fair discount outcomes positively, grateful for a no worse outcome and with expectations of a better outcome next time. For the advantageous discount conditions, 34% of participants were classified with Positive 3: Luck Again; 8% were classified with Negative 2: No More Luck; and 8% were classified with Indifferent. These findings are consistent with Study 2. 5.3. Discussion Study 3 demonstrates that consumers have internal reference discounts, which operate in a manner complementary to the external reference discount observed in previous studies in the paper. Unlike the case of the external reference discount, however, the internal reference discount did not present sensitivity to levels of deviations. Furthermore, consistent with the external reference discount in Study 2, the internal reference discount in Study 3 showed no effect of deal outcome on future purchase intentions. So far we have focused on the reference average discount. However, retailers frequently provide the minimum guaranteed discount when they use uncertain price promotions (e.g., Kohl’s). Unlike the referent average discount, the advertised minimum discount claim does not create a negative deviation in the actual discount outcome because consumers get at least the guaranteed minimum discount. Thus, we might expect consistent findings with advantageous discount conditions in the paper. In the next study, we examine the role of the guaranteed minimum as the reference discount. 6. Study 4: Guaranteed minimum discounts 6.1. Method The procedure and buying situation were essentially the same as in Study 1. The questionnaire presented a scenario in which participants were asked to imagine that they were purchasing a digital camera at a store running the following scratch and save promotion: ‘‘Scratch and Save, At Least 15% Off!!’’ Eighty-three undergraduate business students were randomly assigned to one of three conditions: (1) 15% discount (receiving the guaranteed minimum discount); (2) 20% discount (receiving 5% more than the guaranteed minimum discount); and (3) 25% discount (receiving 10% more than the guaranteed minimum discount). Of the participants, 32 (38.6%) were female and 51 (61.4%) were male. Their average age was

41

S. Choi et al. / Journal of Economic Psychology 39 (2013) 32–43 Table 6 Means of major variables in Study 4.a 15% Discount (Condition 1) Perceived fairness Before enjoyment After enjoyment Post behavior a b

5.33 4.63 4.72 2.65

(1.97)b (1.28) (1.59) (1.80)

20% Discount (Condition 2)

25% Discount (Condition 3)

5.52 4.74 5.12 2.43

5.94 4.84 5.66 2.45

(1.31) (1.09) (1.03) (1.70)

(.67) (1.56) (1.53) (2.05)

This experiment states a minimum discount of 15%. Standard deviations in parentheses.

22.16 years. The average buying intention to purchase a digital camera within a one-year period was 4.40. Before (a = .90) and after enjoyment (a = .96), perceived fairness of the discount outcome, and future purchase intention (r = .75) were measured. In addition, they were asked to answer two questions for the manipulation checks: the advertised discount level and actual discount level they received. Eighty-one of eighty-three participants correctly answered the two manipulation check questions. 6.2. Results The perceived fairness showed no difference across the three conditions (see Table 6). A 3 (discount level: 15%, 20%, and 25%)  2 (enjoyment: before and after) mixed ANOVA, with discount level representing a between-subjects factor and enjoyment a within-subjects variable, showed no significant interaction effect (F(2, 80) = 1.82, p = .17, g2 = .04). Follow-up comparisons showed that the enjoyment levels of the promotion before/after knowing the actual discount level were not different across the three conditions, although the enjoyment level of the promotion after knowing the actual discount level showed an increasing pattern as the actual discount increased. However, the enjoyment level increased significantly only for the ‘‘25% off’’ condition (Mbefore = 4.84 vs. Mafter = 5.66; t(62) = 2.08, p < .05). In addition, future purchase intentions showed no difference across the three conditions. These results are consistent with findings of advantageous discount conditions in the paper. 6.3. Discussion Study 4 demonstrates the role of advantageous discounts in consumer responses to uncertain price promotions when the reference is a minimum guaranteed discount. Regardless of information type (overt average discount or minimum discount), consumers are less sensitive to positive discount outcomes. Consistent with other studies presented in the paper, positive discount outcome deviations did not affect perceived fairness and future purchase intentions. 7. Conclusion This paper studies risky discounts. In particular, it examines consumers’ perceived fairness and enjoyment of gamblinglike SAS promotions. We find that despite receiving a deal, consumers perceive a discount lower than the reference discount to be less fair and less enjoyable than the reference discount. Yet, consumers perceived a discount above the reference discount to be as fair and enjoyable as the reference discount (manipulated by an average discount [Studies 1 and 2], an expected discount [Study 3], and a minimum guaranteed discount [Study 4]). The results from Study 2 show participants reacting significantly more critically to a large loss of discount than to a small one, but the results also show no difference in their reaction to the two levels of gained discount. Such asymmetry disappears when the internal reference discount was considered in Study 3, which suggests that the internal reference discount is less precise than the external reference discount. The two content analyses conducted in Studies 2 and 3, respectively, provide evidence that the perceptions of less fairness caused by SAS promotions do not result in any damage to future purchase intentions; instead, they might generate positive expectations for further SAS promotions. One explanation for this counter-intuitive finding is that many consumers understand the nature of uncertain SAS promotions: (1) the worse possible outcome was possible; and (2) a better and luckier outcome may be possible next time. Hence, the outcome will not be upsetting enough to elicit anger, which is a prerequisite for undesirable future purchase intentions (Bougie, Pieters, & Zeelenberg, 2003; Zeelenberg, & Pieters, 2004). Indeed, Goldsmith and Amir (2010) show that uncertain promotions could be an effective promotional strategy because consumers might respond positively to uncertainty. Consistently, the lack of effects that less fair discounts have on future purchase intentions supports the effectiveness of promotions with uncertain components. Our findings offer several implications to retail practitioners who consider employing SAS promotions to stimulate store traffic and sales. First, we find that risky promotions are generally looked upon favorably by customers and that the external reference discount has a greater impact than an internal reference. Thus, retailers should use mass communications such as advertisements to promote SAS activities. They should pitch the money-saving nature of discounts by highlighting the external reference discount in terms of advertising execution. Second, we identify the asymmetric pattern of consumer responses

42

S. Choi et al. / Journal of Economic Psychology 39 (2013) 32–43

to discount levels of risky promotions, and detect that the marginal impact of positive discounts on consumers is trivial. The size of the discount is not important to the customers’ perceptions of fairness or pleasure (but only the size of the discount relative to the reference discount). Retailers can therefore provide primarily smaller discounts as long as they are greater than the external reference discount or the percentage necessary for a sale (e.g., 17%). Finally, we compare different forms of phrasing SAS promotions (e.g., maximum discounts to be received, minimum discounts to be received, or average discounts to be received) in terms of consumer perceptions. Our findings suggest that it might be best for retailers to indicate the minimum saving level plus, perhaps, a mention of the maximum discount. Hence, retailers should present the external reference discount that is either the minimum that a customer can get (e.g., ‘‘at least 20%’’) or a very high maximum amount (e.g., ‘‘up to 50%’’). Due to the scope of this paper and the nature of our methodology, this paper has several limitations. First, this paper investigated the shopping decisions only in a lab setting with limited sample sizes. Thus, future research in a more realistic setting (e.g., controlled field experiments) with larger sample sizes is needed to verify external validity of the findings. It would be particularly interesting to examine whether SAS promotions could be an effective way to bring people into a store or to get them to spend more money once they are there. Furthermore, as relatively small sample sizes might be a reason for the lack of sensitivity to discount size in our experiments, future research would be necessary to investigate if the observed sensitivity could be held with bigger sample sizes. Second, this paper only considered a single SAS promotion. SAS promotions may have more obvious benefits when repeated over time rather than considered in a single instance. Basic learning theory suggests that this kind of random reinforcement leads to relatively persistent behavior at a relatively low cost compared to other forms of reinforcement. This reinforcement may cause consumers to continue trying to get the discount in the future, even if they are dissatisfied with the last outcome, and perhaps even more so. If consumers are repeatedly treated with less fairness, the appeal of prospective winnings might be diminished; in turn, additional less fair discounts could significantly lower purchase intentions. Therefore, this paper’s findings should not be taken to mean that consumers will grudgingly accept disappointing promotions, but instead prompt the corresponding research question worth investigation: Under what conditions is future purchase behavior no longer insulated from less fair discount outcomes? Third, this paper established and rationalized observations of internal and external reference discounts. With a beginning understanding of reference discounts in place, including consumers’ fairness and enjoyment responses, further work can be done to elaborate on these findings. In particular, it would be desirable in future studies to measure emotional responses, which would be useful in explaining any unforeseen absence of future purchase intention. Finally, it is not entirely impossible for participants to think of unluckiness when they receive a lower discount than expected or advertised. They may believe that they got the small discount because of a ‘‘chance,’’ as in a lottery – that they had had a fair chance at a larger discount and that luck had just not come their way. If so, they might interpret their outcome as ‘bad luck’, as the seller does not intervene to determine their personal outcome. Thus, future research should consider the possibility of (un)luckiness. Acknowledgments The authors would like to thank Erik Hoelzl and two anonymous reviewers, whose comments and suggestions greatly helped improve this manuscript. The authors also gratefully acknowledge the financial support of the Social Sciences and Humanities Research Council of Canada (864-2007-0318) to the first author. The authors are listed alphabetically. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.joep.2013.06.007. References Biswas, A., & Burton, S. (1993). Consumer perceptions of tensile price claims in advertisements: An assessment of claim types across different discount levels. Journal of Academy of Marketing Science, 21(3), 217–229. Biswas, A., & Burton, S. (1994). An experimental assessment of effects associated with alternative tensile price claims. Journal of Business Research, 29, 65–73. Blakely, R. (2007). Apple sued for $1m over $200 iPhone discount. Times, 2 October. . Bolton, L. E., & Alba, J. W. (2006). Price fairness: Good and service differences and vendor costs. Journal of Consumer Research, 33(September), 258–265. Bolton, L. E., Warlop, L., & Alba, J. W. (2003). Consumer perceptions of price (Un) fairness. Journal of Consumer Research, 29, 474–491. Bougie, R., Pieters, R., & Zeelenberg, M. (2003). Angry customers don’t come back, they get back: The experience and behavioral implications of anger and dissatisfaction in services. Journal of the Academy of Marketing Science, 3(4), 377–393. Campbell, M. C. (1999). Perceptions of price unfairness: Antecedents and consequences. Journal of Marketing Research, 36(May), 187–199. Campbell, M. C. (2007). ‘Says Who?!’ How the source of price information and affect influence perceived price (Un) fairness. Journal of Marketing Research, 44(May), 261–271. Choi, S., & Kim, M. (2007). The effectiveness of ‘Scratch and Save’ promotions: The moderating roles of price consciousness and expected savings. Journal of Product and Brand Management, 16(7), 469–480. Choi, S., Stanyer, M., & Kim, M. (2010). Consumer responses to the depth and minimum claimed savings of ‘Scratch and Save (SAS)’ promotions. Psychology and Marketing, 27(8), 796–809. Darke, P. R., & Dahl, D. W. (2003). Fairness and discounts: The subjective value of a bargain. Journal of Consumer Psychology, 13(3), 328–338. Della Bitta, A. J., & Monroe, K. B. (1980). A multivariate analysis of the perception of value from retail price advertisements. In K. B. Monroe (Ed.). Advances in consumer research (Vol. 8, pp. 161–165). Anne Arbor, MI: Association for Consumer Research.

S. Choi et al. / Journal of Economic Psychology 39 (2013) 32–43

43

Dhar, S. K., Gonzalez-Vallejo, C., & Soman, D. (1999). Moderating the effects of advertised price claims: Tensile versus price claims? Marketing Science, 18(2), 154–177. Goldsmith, K., & Amir, O. (2010). Can uncertainty improve promotions? Journal of Marketing Research, 47(December), 1070–1077. Gupta, S., & Cooper, L. G. (1992). The discounting of discounts and promotional thresholds. Journal of Consumer Research, 19, 401–411. Homberg, C., Koschate, N., & Hoyer, W. D. (2006). The role of cognition and affect in the formation of customer satisfaction: A dynamic perspective. Journal of Marketing, 70(July), 21–31. Kahneman, D., Knetsch, J. L., & Thaler, R. (1986). Fairness as a constraint on profit seeking: Entitlements in the market. The American Economic Review, 76(September), 728–741. Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. The Econometric Society, 47(2), 263–291. Kalwani, M. U., & Yim, C. Y. (1992). Consumer price and promotions expectations: An experimental study. Journal of Marketing Research, 29, 90–100. Kamleitner, B., Mandel, D. R., & Dhami, M. K. (2011). Risky discounts: Do people prefer them on a per-item or per-purchase basis and why? Journal of Economic Psychology, 32, 951–961. Klingemann, H. K. H. (1995). Games, risk and prevention: The rehabilitation of ‘‘Homo ludens’’. Journal of Alcohol and Drug Education, 41(1), 99–123. Krehbiel, P. J., & Cropanzano, R. (2000). Procedural justice, outcome favorability and emotion. Social Justice Research, 13(4), 339–360. Kukar-Kinney, M., Xia, L., & Monroe, K. B. (2007). Consumers’ perceptions of the fairness of price-matching policies. Journal of Retailing, 83(3), 325–337. Kumar, V., Karande, K., & Reinartz, W. J. (1998). The impact of internal and external reference prices on brand choice: The moderating role of contextual variables. Journal of Retailing, 74(3), 401–426. Larwood, L., Kavanagh, M., & Levine, R. (1978). Perceptions of fairness with three alternative economic exchanges. The Academy of Management Journal, 21(March), 69–83. Maxwell, S. (2002). Rule based price fairness and its effect on willingness to purchase. Journal of Economic Psychology, 23, 191–212. Maxwell, S., Nye, P., & Maxwell, N. (1999). Less pain, same gain: the effects of priming fairness in price negotiations. Psychology and Marketing, 16(7), 545–562. Mobley, M. F., Bearden, W. O., & Teel, J. E. (1988). An investigation of individual responses to tensile price claims. Journal of Consumer Research, 15(2), 273–279. Neighbors, C., Larimer, M. E., Lostutter, T. W., & Cronce, J. M. (2001). Exploring college student gambling motives. Journal of Gambling Studies, 18, 361–370. Niedrich, R. W., Sharma, S., & Wedell, D. H. (2001). Reference price and price perceptions: A comparison of alternative models. Journal of Consumer Research, 28, 339–354. Oliver, R. L., & Swan, J. E. (1989). Equity and disconfirmation perceptions as influences on merchant and product satisfaction. Journal of Consumer Research, 16(December), 372–383. Raghunathan, R., & Corfman, K. (2006). Is happiness shared doubled and sadness shared halved? Social influence on enjoyment of hedonic experiences. Journal of Marketing Research, 43, 386–394. Rao, R. C., Arjunji, R. V., & Murthi, B. P. S. (1995). Game theory and empirical generalizations concerning competitive promotions. Marketing Science, 14(2), G89–G100. Singh, J. (1988). Consumer complaint intentions and behavior: Definitional and taxonomical issues. Journal of Marketing, 52(1), 93–107. Smith, A. K., & Bolton, R. N. (2002). The effect of customers’ emotional responses to service failures on their recovery effort evaluations and satisfaction judgments. Journal of the Academy of Marketing Science, 30(1), 5–23. Thaler, R. (1985). Mental accounting and consumer choice. Marketing Science, 4(3), 199–214. Ward, J. C., & Hill, R. P. (1991). Designing effective promotional games: Opportunities and problems. Journal of Advertising, 20(September), 69–81. Xia, L., Monroe, K. B., & Cox, J. L. (2004). The Price is Unfair! A conceptual framework of price fairness perceptions. Journal of Marketing, 68(October), 1–15. Zeelenberg, M., & Pieters, R. (2004). Beyond valence in customer dissatisfaction: A review and new findings on behavioral responses to regret and disappointment in failed services. Journal of Business Research, 57, 445–455.