TURNING GOOD IDEAS INTO BAD NEWS: THE EFFECT OF NEGATIVE AND POSITIVE SPONSORSHIP INFORMATION ON SPONSORS' BRAND IMAGE
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Sponsorship InformationBidding and Brand Image Strategies Oliver Schnittka/Henrik Sattler/Mario Farsky* Turning Good Ideas into Bad News: The Effect of Negative and Positive Sponsorship Information on Sponsors` Brand Image** A bstract Through sponsorships, companies primarily expect to establish, strengthen, or change their brand image. We analyze the effects of negative and positive sponsorship infor- mation on the favorability and structure of sponsors´ brand images. We adopt a new, unconventional approach, the Brand Concept Map, which analyzes sponsors´ brand associative networks. We find that negative sponsorship information unfavorably influ- ences the favorability of sponsors´ brand image and the structure of a sponsoring brand’s associative network, but positive sponsorship stimuli have no influence. Fur- thermore, we identify boundary conditions under which the effect of negative sponsor- ship information on sponsors´ brand image is strengthened or diminished. JEL-Classification: M31. Keywords: Brand Concept Maps; Brand Image; Negative and Positive Sponsorship Infor- mation; Sponsorship. 1 I ntroduction Various trends have encouraged greater spending on sponsorships (IEG (2009); Poon and Prendergast (2006)). These trends include the diminishing impact of traditional media; government restrictions, such as those on tobacco advertising and alcohol advertising; and the popularity of sponsored events, such as the FIFA World Cup or Olympic Games (Quester and Thompson (2001)). By using sponsoring opportunities, companies primarily hope for positive spillover effects (Gwinner (1997)). Their thinking is that if they can enhance or strengthen their brand image, they can generate a competitive advantage (Aaker (1991); Keller (1993)). * Oliver Schnittka, PhD, Postdoctoral Researcher, University of Hamburg, Institute of Marketing and Media, Welckerstr. 8, D-20354 Hamburg (Germany), Phone: +49 40 42838 8711, Fax +49 40 42838- 8715, E-mail: schnittka@econ.uni-hamburg.de. Henrik Sattler, Professor of Marketing and Branding, University of Hamburg, Institute of Marketing and Media, Welckerstr. 8, D-20354 Hamburg (Germany), Phone: + 49 40 42838-8714, Fax: +49 40 42838-8715, E-mail: henriksattler@googlemail.com. Mario Farsky, PhD, Ipsos Marketing, Sachsen- str. 6, D-20097 Hamburg (Germany), Phone: + 49 40 80096-4476, e-mail: mario.farsky@ipsos.com. ** The authors thank two anonymous reviewers for their helpful comments on previous versions of this manuscript. The authors also thank Ann-Christin Wilcke for their help in collecting the data. sbr 65 July 2013 227-248 227
O. Schnittka/H. Sattler/M. Farsky In line with these real-world developments, many recent studies have confirmed the posi- tive effects of sponsorship announcements on the favorability of sponsors’ brand image (e.g., Grohs, Wagner, and Vsetecka (2004); Javalgi, Traylor, Gross, and (1994); Simmons and Becker-Olsen (2006); Speed and Thompson (2000)). But negative sponsorship stimuli (i.e., unfavorable information about the sponsee after the sponsorship announcement, such as a sponsee’s corporate failure or scandals) might also have effects. With the excep- tion of a study by Pope, Voges, and Brown (2009), we do not know of any prior research that has analyzed if and how negative sponsorship information might harm sponsors’ brand image, despite the importance of such questions. Imagine a hypothetical case: Volkswagen signs a contract to sponsor the FIFA World Cup 2010 in South Africa, but after committing to this sponsorship, media reports suggest negative information about the sponsee, implying the World Cup is at risk of being cancelled if the stadium construction is not finished in time and that tourists’ safety cannot be guaranteed. If such negative information harms Volkswagen’s brand image, it should cancel its sponsorship, but then the company risks the loss of its sponsorship investments, with no return. Accordingly, sponsors have to evaluate sponsees’ potential to create negative spillover effects before they commit to a sponsorship contract. Whether negative spillover effects through sponsorships occur or not is a challenging question. In related fields, such as celebrity endorsements, there are many studies that investigate empirically the effects of negative stimuli on brand image. However, the find- ings are mixed. Although several of these studies suggest that negative information about one entity (e.g., negative effects of a celebrity) may unfavorably affect the partners´ brand image (e.g., Bailey (2007)), other studies find that negative spillover effects either do not occur (e.g., Money, Shimp, and Sakano (2006)) or else occur only under certain boundary conditions (e.g., Till and Shimp (1998)). Thus, it remains unclear if and under which conditions the unfavorable effects of negative sponsorship information on the sponsors´ brand image might occur. Further, prior findings on negative information through celeb- rity endorsements cannot be automatically translated into the area of sponsorships, since advertising is an overt intent to persuade, but sponsorships are characterized by a more subtle, indirect persuasion (e.g., Meenaghan (2001)). Prior research in related research fields such as celebrity endorsements (e.g., Till and Shimp (1998)) or brand alliances (e.g., Votolato and Unnava (2006)) has not analyzed how nega- tive and positive stimuli affect the structure of a brand’s associative network. Prior studies used semantic differential or Likert scales to measure the effects of negative stimuli on the favorability of a brand’s image, but do not show how negative stimuli affect the structure of a brand’s associative network. The structure of association networks has significant managerial relevance, since it identifies the most important set of brand associations that drive the brand´s image. If they wish to build, leverage, and protect their brands (Aaker (1996); John, Loken, Kim, and Monga (2006)), managers should make this structure their number one target. Against this background, we extend the marketing literature in two ways. First, we analyze the effects of negative and positive sponsorship information on sponsors’ brand image, 228 sbr 65 July 2013 227-248
Sponsorship Information and Brand Image and we identify the boundary conditions for negative image effects. Second, we apply the Brand Concept Maps (BCM) approach (John et al. (2006)), which identifies consumers´ underlying brand associative networks. We use the BCM to assess the effects of negative (and positive) sponsorship stimuli on the favorability and structure of the sponsor´s brand associative network (e.g., regarding the number and position of associations in the network as well as the number and strength of the corresponding association linkages). Our applica- tion of the BCM approach identifies which associations are primarily devalued by negative sponsorship information and how the composition of the association network changes. The paper proceeds as follows. In Section 2, we derive the theoretical framework on the effect of negative and positive sponsorship information on sponsors´ brand image, and we develop the corresponding hypotheses for our study. We describe the research design and present the main results in Section 3, followed by a discussion of our findings in Section 4. 2 Transferring N egative and P ositive S ponsorship I nformation on S ponsors´ B rand I mage : A Theoretical P erspective In the spreading activation model, consumers´ cognitive structures around a stimulus are conceptualized as information nodes in consumers´ memory. These nodes are directly or indi- rectly linked to the stimulus or with each other (Alba, Hutchinson, and Lynch (1991); Joiner (1998)). Thus, an activation of one information node in memory is expected to spread to other information nodes that are structurally linked (Collins and Loftus (1975)). Consequently, consumers are expected to store brand information in the form of associative networks. A promising procedure for measuring the structure of brand association networks is the consumer mapping technique, such as Brand Concept Maps (BCMs). Contrary to analytical mapping techniques such as network analysis, BCMs elicit brand associative networks directly from consumers. That is, respondents show how their brand associations relate to the brand and one another by constructing their own network of associations. With these individual maps, researchers can aggregate the information to produce a consensus brand association network (John et al. (2006)). Brand association networks identify which associations are directly or indirectly linked to the brand and how these brand associations are connected to one another (Anderson (1983b); Keller (1993); John et al. (2006)). Within the network, associations may vary primarily in the extent to which they are asso- ciated with the brand node. According to information processing theory, which assumes a hierarchical storage of information in consumer memory, more relevant information (e.g., first-order associations) should be linked more directly to the brand and therefore should be easier to retrieve than subordinate information (e.g., second-order associations, Bettmann (1979); Miller (1956)). Although consumers may link many associations with a brand, it is mainly the first-order associations that should be the focus of management efforts to build, leverage, and protect a brand (John et al. (2006)). Moreover, associations not only vary in the length of their associative pathway to the brand node, but also in the strength of the association link to the brand or other associations in consumer memory (Reder and Anderson (1980); John et al. (2006)). sbr 65 July 2013 227-248 229
O. Schnittka/H. Sattler/M. Farsky According to Keller (1993), a brand’s image not only comprises the structure but also the favorability of an associative network. The favorability of that network can, in turn, be illus- trated by three elements: the favorability of brand attributes, brand benefits, and consumers´ general brand attitudes. Attributes represent those descriptive product-related and non- product-related features that characterize a product. Functional, experiential, and symbolic benefits represent the personal value consumers attach to the product attributes (Keller (1993)). Consumers´ general attitudes are a function of the evaluative judgement of all salient attributes and benefits a consumer has about a product (Fishbein and Ajzen (1975)). To address the favorability of a sponsor´s associative network, associative learning theory can explain why negative (positive) sponsorship information after the sponsorship announce- ment might unfavorably (favorably) influence consumers´ attitudes toward both the sponsor and the underlying associations in terms of attributes and benefits. When consumers think about a sponsor, the link with the sponsee is animated through spreading activation (Anderson (1983a)). This simultaneous activation of the sponsor and sponsee nodes provides an opportunity to transfer consumers´ evaluation of the sponsee to the sponsoring brand. Consequently, if negative (positive) information about one entity (e.g., the sponsee) after the sponsorship announcement results in a less (more) favorable evaluation of that entity, the effect may reflect on another related entity (e.g., the sponsor) through the associative link established between them (Anderson (1976); Till and Shimp (1998)). Thus, negative (positive) sponsorship information is expected to unfavorably (favorably) affect consumers´ general attitudes toward both the sponsoring brand and the underlying brand attributes and benefits attributable to the sponsor. Therefore, we hypothesize: H1. Negative sponsorship information has a negative impact on (a) consumers’ attitudes toward the sponsor and (b) the favorability of attributes and benefits about the sponsor. H2. Positive sponsorship information has a positive impact on (a) consumers’ attitudes toward the sponsor and (b) the favorability of attributes and benefits about the sponsor. Pham (1992) shows that greater involvement with the sponsee prompts consumers to devote more resources to processing sponsorship information in general. Such active, detailed processing should enable consumers to learn about the sponsor/sponsee connec- tion. According to associative learning theory (Anderson (1983a)), this additional favor- able (unfavorable) knowledge may favorably (unfavorably) influence consumers’ attitudes toward the sponsor. For example, d’Astous and Bitz (1995) show empirically that greater involvement with the sponsee strengthens the impact of positive sponsorship scenarios (i.e., sponsorship announcements) on consumers´ attitudes. We offer a parallel prediction for negative sponsorship information scenarios, since highly involved consumers are also more likely to devote resources to processing negative sponsorship information: H3. Greater involvement with the sponsee strengthens the impact of negative sponsorship information on consumers’ attitudes toward the sponsor. 230 sbr 65 July 2013 227-248
Sponsorship Information and Brand Image Prior research on information processing assumes that the cognitive structures associated with familiar stimuli are more rigid than are those for less familiar objects, since brand- related associations and experiences are well established. Thus, cognitive structures, and hence consumers´ attitudes toward a brand, are more difficult to influence (Bettman and Sujan (1987); Fazio (1986); Simonin and Ruth (1998)). Therefore, sponsorship messages should have a greater impact on consumer attitudes toward unfamiliar brands compared with familiar brands. Carrillat, Lafferty, and Harris (2005) empirically support this assumption for positive sponsorship scenarios (i.e., sponsorship announcements). We offer a parallel prediction for negative sponsorship information scenarios: H4. Greater familiarity with the sponsoring brand diminishes the impact of negative sponsor- ship information on consumers’ attitudes toward the sponsor. To address the influence of negative (positive) sponsorship information on the struc- ture of the associative network of the sponsoring brand, we build on Krishnan (1996). Krishnan shows empirically that high equity brands (i.e., brands with a more favorable brand image) are characterized by having a greater number of associations included in their brand associative networks compared to low equity brands (i.e., brands with a less favorable brand image). Furthermore, we assume, based on the findings by John et al. (2006), that brands with more favorable attitudes are not only characterized by a greater number of brand associations, but also have knowledge structures that are generally more complex and that are characterized by more first-order associations, more brand association links, and stronger brand association links. John et al. (2006) empirically identify highly (less well) integrated knowledge structures for consumers being more (less) familiar with a brand (see Novak and Gowin (1984)). Since mere exposure theory assumes that higher levels of brand familiarity improve consumers’ attitudes toward the stimulus (Anand, Holbrook, and Stephens (1988); Zajonc (1968)), we expect more favorable brand association networks to have a more complex knowledge structure, for instance, in terms of more first-order associations (i.e., those directly linked to the brand) and more brand association links. In line with these findings and based on associative learning theory (Anderson (1983a)), we expect that negative (positive) sponsorship information consequently has a negative (positive) impact on the structure of the associative network: H5. Negative sponsorship information has a negative impact on the structure of the associative network of the sponsoring brand by reducing the number of (a) brand associations, (b) first-order brand associations, (c) brand association links, (d) moderate brand association links, and (e) strong brand association links. H6. Positive sponsorship information has a positive impact on the structure of the associative network of the sponsoring brand by increasing the number of (a) brand associations, (b) first-order brand associations, (c) brand association links, (d) moderate brand association links, and (e) strong brand association links. sbr 65 July 2013 227-248 231
O. Schnittka/H. Sattler/M. Farsky 3 E mpirical Study 3.1 M ethod 3.1.1 D esign We recruited N = 216 respondents (average age: 28.46 years, 56.7% male) for an experi- mental lab survey in Germany conducted in July 2008. All but six respondents completed the entire Brand Concept Maps procedure. To take part in the survey, we required that respondents have a minimum level of involvement with automobiles in the relevant compact car segment (MInvolvement compact class = 4.37, SD = 1.39). We randomly assigned each of the 210 participants in the final sample to one out of four conditions in a between-subjects design: the control condition with no sponsorship information (N = 50); the sponsorship announcement condition (N = 52); the negative sponsorship information condition (N = 56); and the positive sponsorship information condi- tion (N = 52). For our study setting, we select automobiles as the sponsoring product category and Volkswagen as the sponsoring brand. We used the Volkswagen Golf (the most familiar product brand in the German/European compact car segment) instead of the corporate brand Volkswagen as object of investigation because heterogeneous percep- tions of a corporate brand are more likely than they are for a product brand, which is important for the BCM procedure. The FIFA World Cup 2010 in South Africa, one of the largest sports events worldwide, served as the sponsored event. To generate meaningful BCMs, it is important that our participants have high levels of familiarity. Furthermore, we implemented a hypothetical sponsorship to avoid respondent information biases; Volkswagen has never sponsored the FIFA World Cup in the past. 3.1.2 Stimuli Every respondent in the three sponsorship conditions initially received a press release that announced that Volkswagen would be sponsoring the FIFA World Cup 2010 (see Appendix A). Respondents in the control group received no such priming stimulus. After several filler questions, we provided three additional fictitious press releases containing hypothetical negative or positive information about the World Cup 2010 for every respon- dent within the negative or positive sponsorship information condition. These press releases were presented in random order. The negative information stated that the event might be cancelled because the stadium construction would not be finished in time, tourists’ safety could not be guaranteed, and that much of South Africa’s population would be system- atically excluded from the World Cup benefits by local government (see Appendix B). The positive information stated that the construction of all host stadiums would be finished in time, tourists’ safety could be guaranteed during the event, and that South Africa’s economy would profit greatly from the World Cup 2010 (see Appendix C). We made several efforts to keep the sponsorship information scenarios realistic. 232 sbr 65 July 2013 227-248
Sponsorship Information and Brand Image 3.1.3 P rocedure Following the priming stage and after a substantial time interval, we implemented the BCM procedure as described by John et al. (2006). We began by introducing respon- dents to how the BCM task would operate. We then asked them to develop an individual brand map with the Volkswagen Golf emblem in the center and 25 predetermined brand associations, along with five additional blank cards that they could use to write down further relevant associations. Respondents also indicated their evaluations of each brand association for the Volkswagen Golf on a seven-point bipolar semantic differential scale. We asked them to use weak, moderate, or strong association links to connect the selected brand associations, or their own, with the Volkswagen Golf or with each other. We identified the 25 predetermined brand associations in an in-depth interview pretest with N = 35 respondents who were very familiar with compact cars. The pretest gener- ated 121 relevant brand associations, from which we selected the top 25 to create a rele- vant association set for the mapping procedure. Two coders (ϕ = 0.75, p < 0.01) independently classified these associations as attributes or benefits, using the definition by Keller (1993). The resulting list of attributes included comfort, commodiousness, design, durability, environmental sustainability, fuel consumption, groundedness, inno- vation, price, price–performance ratio, prominence, quality, resale value, usualness, and youthfulness; the list of benefits featured driving pleasure, ease of handling, mobility, practicality, prestige, reliability, safety, satisfaction, sympathy, and trustworthiness. We used a binary scale (yes/no) to measure respondents’ prior usage of the Volkswagen Golf (54.8% users) as a proxy for their brand familiarity. We further measured attitude toward the sponsor, involvement with the sponsee, and the product category on seven-point bipolar semantic differential scales and coded each item such that a higher score indicated a more favorable rating. To measure respondents’ attitudes toward the Volkswagen Golf, we used three items from Osgood, Suci, and Tannenbaum (1957): bad/good, negative/positive, and unfavorable/favorable. The measure of the favorability of each brand association only used one of these items: bad/good. The measure of involvement with the FIFA World Cup and with automobiles in the relevant compact car segment used three items, adapted from Mittal (1995): unimportant/important, not concerned/concerned, and not care about/care about. All multi-item measures achieved an acceptable level of reliability, according to their Cronbach’s alphas (each α > 0.80). Furthermore, we conducted a confirmatory factor analysis for each multi-item scale and found an acceptable level of convergent validity; for both scales, item variation explained by the pertinent factor exceeded 75%. 3.2 R esults Before testing our hypotheses, we considered the structural equality of the four experi- mental conditions. An analysis of variance (ANOVA) and chi-square tests did not indicate any significant differences across the four conditions in terms of the age of the respondents In cases of discrepancy, we used a third coder to classify the appropriate associations. sbr 65 July 2013 227-248 233
O. Schnittka/H. Sattler/M. Farsky (F = 2.39, p = 0.07), their involvement with compact automobiles (F = 1.26, p = 0.29), gender (χ2 = 3.99, p = 0.26), or their use of a Volkswagen Golf (χ2 = 3.49, p = 0.32). However, we do find a significant main effect of the sponsorship scenario on consumers’ attitudes toward the Volkswagen Golf (F = 4.61, p = 0.01). We use a nonparametric Kruskal-Wallis test to analyze the differences for each association between the experimental treatments, because we can not assume a normal distribution of the data for most associa- tions. The test shows a significant main effect of the sponsorship scenario on consumers´ evaluations of the commodiousness attribute (χ2 = 9.82, p = 0.02) and of the following benefits: safety (χ2 = 8.41, p = 0.04), satisfaction (χ2 = 8.23, p = 0.04), sympathy (χ2 = 8.07, p = 0.04), and trustworthiness (χ2 = 14.31, p = 0.01) of the Volkswagen Golf. For all other attributes and benefits, we find no significant main effect (each p > 0.05; see Table 1, p. 235). Initially, we confirm the evidence in prior research (e.g., Grohs, Wagner, and Vsetecka (2004); Javalgi et al. (1994); Simmons and Becker-Olsen (2006); Speed and Thompson (2000)) on the positive effect of sponsorship announcements on consumers´ attitudes toward the sponsor (t = 2.00, p = 0.04). Additionally, the sponsorship announcement enhances consumers’ evaluations of specific attributes of the Volkswagen Golf, such as its commodiousness (z = 2.78, p = 0.01) as well as specific benefits such as trustworthiness (z = 2.23, p = 0.03). Because we cannot assume a normal distribution of the data for most associations, we also use a nonparametric Mann-Whitney test to analyze the mean difference for each associa- tion between the experimental treatments. In support of H1a, a test of contrasts regarding consumers’ attitudes shows that negative information about the FIFA World Cup 2010 has a negative impact on consumers´ atti- tudes toward the Volkswagen Golf compared to the sponsorship announcement condi- tion (t = 3.22, p = 0.01). Like our H1a results, we find partial support for H1b. Thus, negative information about the FIFA World Cup 2010 has a negative impact on the favorability of specific attributes and benefits of the Volkswagen Golf in relation to the commodiousness attribute (z = 2.05, p = 0.04) and the following benefits: safety (z = 2.31, p= 0.02), satisfaction (z = 2.04, p = 0.04), sympathy (z = 2.31, p = 0.02), and trustworthiness (z = 2.62, p = 0.01). Contrary to our expectations (H2a), a test of contrasts also shows that positive informa- tion about the FIFA World Cup 2010 released after the sponsorship announcement has no impact on consumers´ attitudes toward the Volkswagen Golf compared to the initial sponsorship announcement condition (t = 0.37, p = 0.71). We also find that positive performance of the FIFA World Cup 2010 does not enhance consumers’ evaluations of the attribute commodiousness (z = 0.61, p = 0.54) or the following benefits: safety We note that we test all hypotheses by considering Bonferroni correction for multiple significance tests. 234 sbr 65 July 2013 227-248
Sponsorship Information and Brand Image Table 1: Consumers´ Evaluations of Brand Attitude toward the Sponsor and the Favorability of Single Brand Associations Control Announcement Negative Positive scenario scenario scenario scenario Brand attitude M SD N M SD N M SD N M SD N Fa Volkswagen Golf 4.65 1.15 50 5.10 1.15 52 4.29 1.44 56 5.01 1.31 52 4.61** Brand association M SD N M SD N M SD N M SD N χ2 b Comfort c 4.46 1.36 26 4.90 1.09 21 4.46 1.13 13 4.84 .96 19 2.45 Commodiousness 3.78 1.53 27 5.21 .98 14 4.09 1.38 11 4.77 1.30 13 9.82* Design 3.71 1.33 28 4.26 1.68 27 3.60 1.54 20 3.74 1.61 27 2.86 Durability 5.96 .80 28 5.84 .90 25 6.08 .75 26 6.07 .70 29 .98 Environmental 4.46 1.33 13 4.62 1.33 13 3.60 2.01 10 4.00 1.72 24 2.95 sustainability Fuel consumption 4.56 1.21 16 4.33 1.50 15 4.38 1.46 16 4.95 1.02 21 2.62 Groundedness 5.79 .88 34 5.84 1.34 25 5.76 1.58 21 5.84 1.18 25 1.02 Innovation 3.32 1.34 19 3.53 1.65 19 4.20 2.15 10 3.26 1.60 23 1.83 Price 4.70 1.15 30 4.79 1.32 29 4.30 1.65 33 5.33 1.16 30 7.66 Price-performance 3.07 1.46 28 3.41 1.28 27 3.31 1.67 26 4.07 1.67 29 6.67 ratio Prominence 6.83 .44 46 6.63 .85 46 6.53 .67 51 6.62 .88 42 7.23 Quality 5.49 1.00 39 5.63 .84 35 5.62 1.01 37 5.71 .74 42 1.01 Resale value 4.92 1.47 26 5.50 1.14 24 5.36 1.34 28 4.88 1.51 25 4.01 Usualness 5.62 1.36 26 6.00 1.07 15 6.45 .74 22 6.38 .65 24 7.57 Youthfulness 4.39 2.06 23 4.83 1.44 23 4.60 1.81 15 4.75 1.82 24 .41 Driving pleasure 4.73 1.57 30 4.68 1.25 25 4.78 1.29 9 4.22 1.54 23 1.89 Easy of handling 5.22 1.11 18 6.00 .89 11 5.81 .83 16 5.87 .69 23 5.77 Mobility 5.83 1.04 18 5.85 .90 13 5.88 .89 16 6.08 .86 13 .66 Practicality 6.05 .84 22 5.73 .80 15 5.50 .93 8 6.08 .76 13 3.58 Prestige 3.92 1.38 26 4.12 1.66 26 4.33 1.95 15 3.24 1.73 21 4.35 Reliability 5.90 .61 30 5.71 .94 34 5.83 .92 35 5.79 .73 34 .74 Safety 5.29 .86 31 5.82 .83 34 5.06 1.50 32 5.43 .97 30 8.41* Satisfaction 5.24 .94 21 5.92 .80 26 5.31 .95 13 5.74 .87 19 8.23* Sympathy 4.68 1.41 25 5.59 1.02 29 4.58 1.63 26 5.03 1.33 30 8.07* Trustworthiness 5.41 .97 27 5.97 .75 35 5.07 1.41 27 5.91 1.38 32 14.31** Notes: a F-value teststatistic; b Kruskal-Wallis test statistic; c associations in italics represent attributes while all other associations represent benefits; ** p < 0.01; * p < 0.05. sbr 65 July 2013 227-248 235
O. Schnittka/H. Sattler/M. Farsky (z = 1.94, p= 0.06), satisfaction (z = 0.79, p = 0.43), sympathy (z = 1.61, p = 0.11), and trustworthiness (z = 0.9, p = 0.37). Thus, H2b is rejected. By using a moderation analysis (Aiken and West (1991)), we find that the effect of unfavorable information of the FIFA World Cup 2010 is moderated by consumers’ use of the sponsor’s products, which is the measure we use as proxy for brand familiarity. The effect of negative sponsorship information on consumers’ attitudes toward the Volkswagen Golf is smaller if respondents have used a Volkswagen Golf (β = -0.25, ΔR2 = 0.06, F = 13.48, p = 0.01), in support of H4. However, the effect of negative sponsorship information is stronger when consumers are highly involved with the sponsee, in our example, the FIFA World Cup 2010 (β = 0.22, ΔR2 = 0.05, F = 8.56, p = 0.02). These results support H3. For our results for H5 and H6, we provide Figures 1–4, pp. 237-238, which represent consensus maps for our four experimental treatments. We develop the maps according to the aggregation rules provided by John et al. (2006). In these figures, the solid circles represent core associations of Volkswagen’s brand image, i.e., associations that are included on at least 50% of the individual maps; dashed-line circles represent non-core associations, i.e., associations that are included on less than 50% of the individual maps; and bold lines indicate moderate association links, i.e., the average strength of association link across all respondents in the treatment. Thin lines signify weak association links. To test the reliability of consumers’ maps, we use a split-half reliability procedure according to John et al. (2006). For the control condition, we find acceptable levels of reliability for the presence of brand associations (ϕ = 0.45, p = 0.01), the presence of first-order brand associations (ϕ = 0.48, p = 0.01), and the presence of specific brand association links (ϕ = 0.45, p = 0.01) in both split-halfs. The known-groups approach enables us to assess nomological validity (John et al. (2006)), such that we compare the consensus brand maps produced by respondents with differing Volkswagen Golf usage levels. Again, we find acceptable levels of nomological validity similar to those by John et al. (2006). Table 2 shows these levels. For example, in the negative sponsorship information scenario, users of compact class automobiles formed brand maps with more brand associations (t = 3.12, p = 0.01), first-order associations (t = 2.27, p = 0.02), brand association links (t = 1.88, p = 0.03), and moderate brand association links (t = 1.76, p = 0.04). However, the number of strong brand association links showed no significant differences (t = 0.55, p = 0.29). We find a significant main effect of sponsorship on consumers’ structures of their brand association networks of the Volkswagen Golf in terms of number of brand associations (F = 6.22, p < .001), number of first-order-associations (F = 2.74, p = 0.04), number of brand association links (F = 3.36, p = 0.02), and number of moderate brand association links (F = 3.73, p = 0.01). 236 sbr 65 July 2013 227-248
Sponsorship Information and Brand Image Figure 1: Consensus Map (Control Scenario, N = 50) reliability trustworthiness sympathy durability satisfaction safety prominence prestige ease of handling quality usualness VW Golf design practicality youthfulness innovation groundedness price environmental sustainability resale value price- performance fuel consumption ratio Figure 2: Consensus Map (Sponsorship Announcement Scenario, N = 52) groundedness prestige internationality innovation comfort trustworthiness usualness prominence driving pleasure commodiousness satisfaction youthfulness design practicality reliability VW Golf sympathy durability mobility price quality resale value price- performance fuel consumption ratio safety environmental sustainability sbr 65 July 2013 227-248 237
O. Schnittka/H. Sattler/M. Farsky Figure 3: Consensus Map (Negative Sponsorship Information Scenario, N = 56) durability trustworthiness usualness safety reliability groundedness practicality resale value satisfaction quality prominence VW Golf comfort sympathy fuel consumption price youthfulness environmental sustainability commodiousness price- performance prestige ratio Figure 4: Consensus Map (Positive Sponsorship Information Scenario, N = 52) reliability trustworthiness durability sympathy satisfaction safety prominence prestige ease of handling quality usualness VW Golf design practicality youthfulness innovation groundedness price environmental sustainability price- resale value performance fuel consumption ratio 238 sbr 65 July 2013 227-248
Sponsorship Information and Brand Image Table 3: Analysis of Consumers’ Associative Network Structures Control Announcement Negative Positive scenario scenario scenario scenario (N = 50) (N = 52) (N = 56) (N = 52) Measure M SD M SD M SD M SD Fa Number of brand 13.00 4.85 12.00 3.99 9.70 3.67 12.58 4.85 6.22** associations Number of first-order 6.20 2.07 6.06 2.14 5.25 1.94 6.17 1.91 2.74* brand associations Number of brand 8.34 5.95 7.06 4.85 5.18 4.37 7.00 5.50 3.36* association links Number of moderate 3.40 2.69 2.88 2.71 1.79 2.08 2.52 2.76 3.73* brand association links Number of strong 1.54 1.94 1.15 1.84 1.36 1.97 1.73 2.32 .78 brand association links Notes: a F-value test statistic; ** p < 0.01, * p < 0.05. In line with H5a, H5b, H5c, and H5d, negative sponsorship information unfavorably affects the structure of the Volkswagen Golf ’s associative network, such that it contains fewer brand associations (t = 3.13, p = 0.01); first-order brand associations (t = 2.06, p = 0.04); brand association links (t = 2.12, p = 0.04); and moderate brand association links (t = 2.37, p = 0.02; see Table 3). However, we find no significant main effect in terms of strong brand association links (F = 0.78, p = 0.51). Thus, H5e is rejected. Our results suggest that the negative effect of unfavorable sponsorship information about the FIFA World Cup 2010 released after the sponsorship announcement actually worsens the associative network structure of the Volkswagen Golf compared to the control condi- tion of no sponsorship announcement. For instance, individual brand maps within the negative sponsorship information condition are characterized by fewer brand associations (t = 3.98, p = 0), fewer first-order-associations (t = 2.44, p = 0.02) and fewer brand association links (t = 3.14, p = 0.01) compared to the control condition. Thus, we present what we believe is first evidence that negative sponsorship information unfavor- ably overcompensate the positive effect we initially expected would follow a sponsorship announcement within the negative information condition. Although Volkswagen main- tains its sponsoring of the FIFA World Cup 2010, consumers in the negative information condition evaluate the Volkswagen Golf less favorably compared to a control condition in which no sponsorship information has been announced. In contrast to our expectations, positive information from the FIFA World Cup 2010 has no positive impact on the structure of the associative network of the Volkswagen Golf compared to the sponsorship announcement condition in terms of the number of brand sbr 65 July 2013 227-248 239
O. Schnittka/H. Sattler/M. Farsky associations (t = 0.66, p = 0.51), first-order brand associations (t = 0.29, p = 0.77), brand association links (t = –0.06, p = 0.96), and moderate brand association links (t = –0.68, p = 0.5), Thus, positive sponsorship information leaves the associative network of the sponsoring brand unchanged, and we must reject H6a, H6b, H6c, and H6d. H6e is also rejected, since we do not find a significant main effect in terms of strong brand association links. 4 D iscussion In this research, we analyze the effects of negative – and at the same time positive – sponsorship information that is released after the sponsorship announcement on the favorability and the structure of sponsors´ brand image. To measure such effects on sponsors´ brand image, we apply the Brand Concept Maps approach, which identifies consumers´ underlying brand associative networks. Our findings show that negative sponsorship information released after the sponsor- ship announcement negatively affects the favorability of sponsors´ brand image and the specific brand associations in terms of attributes and benefits. Thus, we provide what we believe is first evidence that the recent termination strategies of sponsors who become involved in negative sponsorship information (e.g., Deutsche Telekom terminating their Tour de France engagement; The Wall Street Journal–Eastern Edition (2007)) are legiti- mate. Moreover, the negative effect of unfavorable sponsorship information released after the sponsorship announcement actually worsens consumers´ attitudes toward the sponsor compared to the control condition by trend. Thus, if negative sponsorship information is released, then to safeguard their brand image against further harm, spon- sors should seriously consider terminating their sponsorships. We also show that negative sponsorship information primarily affects single benefits of the brand. For example, dimensions of customers´ satisfaction with the sponsoring brand in terms of general satisfaction, sympathy, and trustworthiness. Following attribu- tion theory, if a sponsor does not abandon the sponsee, then consumers might perceive that the sponsor legitimates and supports the misbehavior of the sponsee to a certain degree. This eventually reduces such attributes as the perceived trustworthiness of the sponsor (Kelley (1973); Rifon et al. (2004)). Similarly, unfavorable associations with the sponsee, such as not guaranteeing tourists´ safety during the FIFA World Cup 2010 within the priming stage, unfavorably attach to the sponsoring brand and its products (e.g., less favorable evaluations of the perceived safety of a Volkswagen Golf ) as well. However, negative sponsorship information usually affects only a limited number of benefits and brand associations. Considering that brand image is a long-term construct, but we surveyed respondents’ evaluations promptly after the priming stage (Keller (1993)), we believe it is reasonable that H1b attains only partial support across all these many specific brand associations. Moreover, we do not find positive effects for any association. Finally, with only one exception, negative sponsorship information has no effect on single attributes of the brand. This finding is not surprising, since we do not 240 sbr 65 July 2013 227-248
Sponsorship Information and Brand Image expect either product-related or non-product-related features to be influenced by (nega- tive) sponsorship information contrary to benefits representing customers` satisfaction such as trustworthiness. In addition, we identify boundary conditions for the effect of negative information on spon- sors´ brand image. Our results suggest a dilemma for sponsors. Sponsors primarily want to enhance their brand image by transferring associations of the sponsee to the sponsor. Therefore, more attractive sponsees should promise more favorable brand image transfers to the sponsor. However, our results show that a highly attractive sponsee increases the risk of a negative spillover effect if it has unfavorable associations for the public. Another key goal of sponsorships is the targeting and acquisition of new and promising consumers through increased levels of brand awareness and recognition (Crowley (1991); Cornwell, Roy, and Steinard II (2001)). However, we find that negative information has a greater influence on consumers with lower levels of brand awareness, which means the impact of negative information is greatest on sponsors’ primary target groups, its potential customers with low levels of brand familiarity and awareness. Further research should consider other moderating variables, such as sponsor–sponsee fit (see Till and Shimp (1998)), that might reduce the potential negative effects of unfavorable information on a sponsor’s brand image. The consideration of additional drivers in the sponsorship selection process should help sponsors to reduce the risk of negative spillover effects, even in cases of a scandal. Additionally, our findings show that negative sponsorship information unfavorably influ- ences the structure of a sponsoring brand’s associative network such that they include fewer brand associations and fewer brand association links. For instance, the consensus map in the negative sponsorship information scenario does not include benefits such as driving pleasure, contrary to the sponsorship announcement scenario. Following Krishnan (1996), a brand with fewer unique brand associations compared to those of competing brands will have lower brand equity, since unique brand associations are essential if a sponsor is to stand out from the competitors within the product category (Keller (1993)). Similarly, fewer association links are likely to reduce the density of the network, which should result in lower brand equity as well. Individual brand associations are not connected in a way that they might quickly activate each other (Teichert and Schöntag (2010)). Our qualitative analysis of Brand Concept Maps further shows that associations such as satisfaction, sympathy, and trustworthiness that could suffer from negative sponsorship information will fall from core to non-core associations, such that their association link with the Volkswagen Golf weakens and their position within the network changes. Our analysis suggests that Brand Concept Maps are a promising tool to specify possible effects on sponsor´s brand image. Thus, research in related fields that investigate empiri- cally the effects of negative stimuli on brand image, such as celebrity endorsements (e.g., Till and Shimp (1998)), brand alliances (e.g., Votolato and Unnava (2006)), or brand extensions (e.g., Keller and Aaker (1992)), should consider the use of Brand Concept Maps. This implication seems to be particularly important, since associative network structures represent a key driver of brand equity (Krishnan (1996)). sbr 65 July 2013 227-248 241
O. Schnittka/H. Sattler/M. Farsky Our results further suggest that (at least under the experimental conditions explained above) positive sponsorship information that is released after the sponsorship announce- ment neither affects the favorability of sponsors´ brand image nor the favorability of specific brand associations in terms of attributes and benefits. We provide what we believe is first evidence that positive sponsorship information keeps the sponsors´ asso- ciative networks unchanged in terms of number of brand associations or brand associa- tion links. One explanation might be that favorable associations with a sponsee (e.g., a highly successful sports team) are primarily transferred to sponsors´ brand image when the sponsorship cooperation is announced (see Dalakas and Levin (2005); Grohs, Wagner, and Vstecka (2004)). Therefore, additional positive information about the sponsee after the sponsorship announcement might be perceived as nondiagnostic for sponsors´ brand image (Pope, Voges, and Brown (2009)). Furthermore, we have made considerable efforts to keep the sponsorship information scenarios realistic. However, our scenario – that the construction of all host stadiums would be finished in time or tourists’ safety could be guaranteed during the event – might be perceived as more self-evident than positive, since most respondents might a priori associate the FIFA World Cup with finished stadium constructions or high safety stan- dards for tourists. Finally, our finding that negative sponsorship information seems to affect the sponsors’ brand image more than does positive sponsorship information, in terms of both brand image and the structure of the brand associative network, coincide with prior evidence pertaining to attribution theory (Kelley (1973)) and prospect theory (Kahneman and Tversky (1979)), which similarly posit that negative information has a stronger impact on product evaluations and behavior than does positive information. Our study has several limitations. First, further research should generalize our findings across different kinds of products and cultures, and across positive and negative sponsor- ship scenarios. Second, we used a scenario technique to measure sponsorship information effects. While such a technique has the advantage of simulating the alternative effects of negative sponsee information on sponsors´ brand image before an actual sponsor- ship contract is signed, it suffers from hypothetical experimental conditions within the priming stage. For instance, information about the sponsee in reality is expected to occur not shortly after the sponsorship announcement, like in our experiment. A longitudinal field experiment could address this issue and would result in a higher external validity. Third, we measured only the short-term effects of sponsorship information on sponsors´ brand image after the sponsorship information has been released. Therefore, further research should validate our findings by analyzing long-term effects, especially when the sponsee´s failure is nonrecurring. Doing so might, over time, reduce the unfavo- rable effect of negative sponsorship information. Finally, we adapted Keller´s (1993) categorization of brand associations and distinguish between attributes and benefits. Although we pretested our classification with coders who independently classified these associations as attributes or benefits, using Keller’s (1993) definition, we cannot exclude the possibility that respondents might perceive the classification of some associations as attributes or benefits differently. 242 sbr 65 July 2013 227-248
Sponsorship Information and Brand Image Our findings show that negative sponsorship information released after the sponsorship announcement unfavorably affects the sponsor´s brand image. Indeed, terminating a spon- sorship after the occurrence of negative sponsorship information seems to be a justified strategy. However, doing so merely protects the sponsoring brand from further harm, since the occurrence of negative sponsorship information might immediately spill over to the sponsoring brand. Therefore, sponsors should carefully select their sponsoring engage- ments for the potential for scandals of the sponsee a priori. Doing so would prevent the occurrence of negative sponsorship information. A ppendix A: P riming Stimulus S ponsorship A nnouncement VW A nnounces S ponsorship of the FIFA World C up 2010 Zurich – The FIFA has officially announced that automotive manufacturer Volkswagen will be the new premium sponsor of the FIFA World Cup 2010 in South Africa. Martin Winterkorn, CEO of Volkswagen, illustrated that “the FIFA World Cup 2010 is promising to be a spec- tacular sports event with millions of participants all over the world. Simultaneously, the FIFA World Cup entails a unique possibility for the African continent and its population to set a milestone for a better future. By sponsoring the FIFA World Cup 2010, we want to support this development of the African continent and especially the South African population”. B: A dditional P riming Stimuli N egative S ponsorship I nformation Stadium C haos : FIFA C onsiders R elocation of the World C up 2010 Johannesburg – The FIFA has delivered a damning judgement after an inspection of the ten host cities of the FIFA World Cup 2010 and its stadiums. FIFA delegation guide Ron Delomont announced that the delegation “is very shocked about the current state of almost all visited stadiums. At the moment we are very pessimistic that the construction of all stadiums will be finished till the beginning of the World Cup in June 2010.” Thus, as expected by many critics in the past, the probability that South Africa is being deprived of the World Cup 2010 is continuously increasing. Danny Jordaan, president of the World Cup Organization Committee, apprehends that “a relocation of the mega event to other candidates as Australia or the US would be a slap in the face of the whole African continent. The industrial nations would turn their backs on the black continent once again.” N ot Participating , J ust Watching : W hy P otential World C up Tourists S hould R ather Stay at H ome Pretoria – Despite the increased and massive security precautionary measures, the South African government does not seem to resolve the sustainable problem of prevailing sbr 65 July 2013 227-248 243
O. Schnittka/H. Sattler/M. Farsky crime two years before the opening game of the World Cup 2010. Latest surveys show that the crime rate, which is disproportionately high within the international comparison, anyway, increases by 15% on average in all ten host cities. For the first time, the FIFA remarked sustainable concerns about the absolute security for foreign tourists during the mega event. FIFA spokesman Andreas Herren confirmed that “the security concept of the organization committee is more than fragmentary. Thus, we are extremely concerned about the security of all participants. The brutal criminality, foreign tourists are confronted with, is not acceptable for us. The organizers promised to combat this massive problem shortly after the nomination of South Africa as host for the World Cup 2010. Unfortunately, we have not observed an improvement of the situation until now. Thus, we would not recommend foreign tourists to visit South Africa during the World Cup 2010.” C up of the P oor H opes Cape Town – One of the biggest challenges for the South African government, making the FIFA World Cup 2010 become a spectacle for the whole South African population seems to have failed already two years before the beginning of the mega event. FIFA’s gen- eral secretary Jerome Valcke recognized that “the South African government agencies can not guarantee the stadium participation of the poor population during the World Cup games. The tickets are too expensive and the former plan to distribute free tickets among the poor population has now being rejected by the organization committee.” However, the participation of the poor population in the whole event including stadium visits was one of FIFA’s main reasons to nominate South Africa for the World Cup 2010. Valcke disappointedly admitted that “the World Cup should has been a unique chance for South Africa to continue its successful development process enabling benefits for all parts of the population. This chance has been missed by a highly inefficient and unsuccessful event organization although the World Cup has not started yet.” C: A dditional P riming Stimuli P ositive S ponsorship I nformation L et the Party Start : C onstruction of S uperlative Stadium has been Finished Johannesburg – The construction of the most prestigious stadium project of the South African World Cup Committee has been already finished two years before the opening ceremony of the FIFA World Cup 2010. The “Durban Moses Mabhida Stadium” will be officially opened in a few weeks when the South African national team is playing a friendly match against Brazil. The multifunctional stadium corresponds to modern European safety standards and was assigned as “five stars” stadium by FIFA delegation guide Ron Delmont. “We are very delighted about the current state of the Durban Moses Mabhida stadium as well as all the other visited stadiums. We can guarantee that the construction of all other stadiums will be finished several months before the beginning of the World Cup in June 2010.” 244 sbr 65 July 2013 227-248
Sponsorship Information and Brand Image N o Threats , J ust Party : A V ision W ill B ecome True Pretoria – Due to the increased and massive security precautionary measures, the South African government seems to increasingly resolve the sustainable problem of prevailing crime two years before the opening ceremony of the FIFA World Cup 2010. These up- to-date preventative measures including 80.000 police officers during the whole World Cup will guarantee tourists´ safety on the streets and within each stadium zone. Vice po- lice president Andre Pruis constitutes that “the World Cup 2010 is a fluke for the South African population. Latest surveys show that the criminate rate decreases by over 10% on average in all ten host cities since the security measures have been implemented. All inhabitants identify themselves with the forthcoming mega-event and therefore want to present South Africa in the best light. Thus, all foreign tourists are recommended to visit South Africa during the World Cup 2010 to join a peaceful mega-event.” DIW- survey R eveals H igher E conomic B enefits for S outh A frica Than I nitially E xpected Cape Town – The German Institute for Economic Research (DIW) predicts that the economic benefits of the FIFA World Cup 2010 for the South African economy will amount to approximately five billion Euros. Thus, the economic benefits of the forth- coming World Cup are actually expected to exceed the income and employment ef- fects of the past World Cup in Germany 2006. DIW president Klaus Zimmermann argues that “the favorable development of the local security development has strength- ened firms´ as well as tourists´ trust in the South African government. This will lead to a higher inrush of tourists as well as foreign firms settling in the host cities.” The DIW survey predicts more than 150.000 additional jobs as well as more than 850 mil- lion Euros additional tax revenues. These additional receipts should be primarily rein- vested in social projects reducing populations’ poverty as well as enhancing the quality of South African drinking water. R eferences Aaker, David. A. (1991), Managing brand equity: Capitalizing on the value of a brand name, New York: The Free Press. Aaker, David A. (1996), Building strong brands, New York: The Free Press. Aiken, Leona S. and Stephen G. West (1991), Multiple regression: Testing and interpreting interactions, Newbury Park, CA: Sage. Alba, Joseph W., J. Wesley Hutchinson, and John G. Lynch Jr. (1991), Memory and decision making, in Thomas S. Robertson and Harlold H. Kassarijian, (eds.): Handbook of consumer behavior, Englewood Cliffs, NJ: Prentice Hall, 1–49. Anand, Punam, Morris B. Holbrook, and Debra Stephens (1988), The formation of affective judgments: The cogni- tive-affective model versus the independence hypothesis, Journal of Consumer Research 15, 386–391. Anderson, John R. (1976), Language, memory, and thought, Hillsdale, NJ: Lawrence Erlbaum Associates. Anderson, John R. (1983a), A spreading activation theory of memory, Journal of Verbal Learning and Verbal Behavior 22, 261–295. sbr 65 July 2013 227-248 245
O. Schnittka/H. Sattler/M. Farsky Anderson, John R. (1983b), The Architecture of Cognition, Cambridge, MA: Harvard University Press. Bailey, Ainsworth A. (2007), Public information and consumer skepticism effects on celebrity endorsements: Studies among young consumers, Journal of Marketing Communications 13, 85–107. Bettmann, James R. (1979), An information processing theory of consumer choice, Reading, MA: Addison-Wesley Publishing Company. Bettman, James R. and Mila Sujan (1987), Effects of framing on evaluation of comparable and noncomparable alter- natives by expert and novice consumers, Journal of Consumer Research 14, 141–154. Carrillat, Francois A., Barbara A. Lafferty, and Eric G. Harris (2005), Investigating sponsorship effectiveness: Do less familiar brands have an advantage over more familiar brands in single and multiple sponsorship arrangements?, Journal of Brand Management 13, 50–64. Collins, Alan M. and Elizabeth F. Loftus (1975), A spreading activation theory of semantic processing, Psychological Review 87, 407–428. Cornwell, Bettina T., Donald P. Roy, and Edward A. Steinard II (2001), Exploring managers’ perceptions of the impact of sponsorship on brand equity, Journal of Advertising 30, 41–51. Crowley, Martin G. (1991), Prioritising the sponsorship audience, European Journal of Marketing, 25, 11–21. d’Astous, Alain and Pierre Bitz (1995), Consumer evaluations of sponsorship programmes, European Journal of Marketing 29, 6–22. Dalakas, Vassilis and Aron M. Levin (2005), The balance theory domino: How sponsorships may elicit negative consumer attitudes, Advances in Consumer Research 32, 91–97. Dean, Dwane H. (1999), Brand Endorsement, popularity, and event sponsorship as advertising cues affecting consumer pre-purchase attitudes, Journal of Advertising 28, 1–13. Fazio, Russell H. (1986), How do attitudes guide behavior?, in Richard M. Sorrentino and E. Tory Higgins (eds.): The Handbook of Motivation and Cognition: Foundations of Social Behavior, New York: Guilford Press, 204–243. Fishbein, Martin and Icek Ajzen (1975), Belief, attitude, intention, and behavior: An introduction to theory and research, Reading, MA: Addison-Wesley Publishing Company. Grohs, Reinhard, Udo Wagner, and Sabine Vsetecka (2004), Assessing the effectiveness of sport sponsorships – An empirical examination, Schmalenbach Business Review 56, 119–138. Gwinner, Kevin (1997), A model of image creation and image transfer in event sponsorship, International Marketing Review 28, 47–57. Heider, Fritz (1958), The psychology of interpersonal relations, New York: John Wiley & Sons. IEG (2009), Turning good ideas into bad news: The effect of positive and negative sponsee performance on sponsors’ brand image, IEG Press Release, December, No. 24, http://www.marketingpower.com/ResourceLibrary/Documents/Content%20Partner%20Documents/IEG/2010/spon- sorship_spending_recedes_for_first_time.pdf Javalgi, Rajshekhar G., Mark B. Traylor, Andrew C. Gross, and Edward Lampman (1994), Awareness of sponsorship and corporate image: An empirical investigation, Journal of Advertising, 23, 47–58. John, Deborah R., Barbara Loken, Kyeongheui Kim, and Alokparna B. Monga, (2006), Brand concept maps: A methodology for identifying brand association networks, Journal of Marketing Research 43, 549–563. Joiner, Christopher (1998), Concept mapping in marketing: A research tool for uncovering consumers´ knowledge structure associations, Advances in Consumer Research 25, 311–322. Kahneman, Daniel and Amos Tversky (1979), Prospect theory: An analysis of decision under risk, Econometrica 47, 263–291. Keller, Kevin L. (1993), Conceptualizing, measuring, managing customer-based brand equity, Journal of Marketing 57, 1–22. Keller, Kevin L. and David A. Aaker (1992), The effects of sequential introduction of brand extensions, Journal of Marketing Research 29, 35–50. 246 sbr 65 July 2013 227-248
You can also read