Vulnerable customer

Ethics and Target Marketing: The Role of Product Harm and Consumer Vulnerability Author(s): N. Craig Smith and Elizabeth Cooper-Martin Reviewed work(s): Source: Journal of Marketing, Vol. 61, No. 3 (Jul., 1997), pp. 1-20 Published by: American Marketing Association Stable URL: http://www. jstor. org/stable/1251786 . Accessed: 12/03/2013 16: 10 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www. jstor. org/page/info/about/policies/terms. jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor. org. . American Marketing Association is collaborating with JSTOR to digitize, preserve and extend access to Journal of Marketing. http://www. jstor. org This content downloaded on Tue, 12 Mar 2013 16: 10: 46 PM All use subject to JSTOR Terms and Conditions N. CraigSmith& Elizabeth Cooper-Martin Ethics and Target Marketing: of Product Harm and The Role Consumer Vulnerability Target marketing might be the epitome of the marketing concept. However, in certain instances it has been criticized as unethical. The authors identify explanations for the ethical concern and controversy that can arise over targeting. An empirical study confirms public disquiet over consumer vulnerability and product harmfulness, identifies which targeting strategies are evaluated as less ethical, and highlights the likelihood of consumer boycotts and other disapproving behaviors. Evidence of ethical concern arises when both ” sin” and ” non-sin” products are involved, and it increases for consumers perceived to be more vulnerable. The authors discuss implications for marketing managers, researchers, and public policy. It is not surprisingto find that Tedlow’s (1990) historical account of marketingin America is a history of market segmentation. Marketsegmentation, with its concomitant target marketing (targeting), is one of the most important concepts in marketing. The essence of market segmentation-recognizing the differences among customers and choosing to target a segment of them with similar needshas reached its zenith in the late 20th century. Many consumer markets have fragmented, increasing the need for aided by inforsharplyfocused targetmarketing. Marketers, mation technology, have respondedwith strategiesaimed at smaller and hence more elusive groups of consumers, even to the point of programsdirectedat the individualconsumer. The sophisticationof targetmarketingand recognitionof its importance as a means of achieving efficiency and effectiveness have never been greater. But despite its role in identifying and serving customer needs, more focused target marketinghas been accompaniedby increasedcriticism. In extensive media attentionhas been devoted to the particular, of adult consumer segments viewed as ” vulneratargeting which is the focal ble,” with productsconsidered ” harmful,” issue addressed here. This criticism of targeting has included products such as lottery tickets, fast food, weightloss products, contraceptives, rental furnitureand electrical equipment, food supplements, and financial services, such as auto insurance and credit cards. l Most extensive, howand 1SeeClotfelter Cook 1989; Freedman 1990, 1993; Hwang 1994; Jacobs1992; Keats1994; NYDCA1992; Smith1995. School of Professor Marketing, N. Craig Smith Associate is Georgetown of is President, Elizabeth Business, Cooper-Martin University. Georgetown Sandra Burke, G. J. Jill thank R. Andreasen, Alan Inc. CM, Theauthors AlexanDonald Robin, P. R. Debra Donald Lichtenstein, J. Ringold, Klein, for reviewers and Robert Thomas, threeanonymous J. derSimonson, on drafts thisarticle. of comments earlier by Funding theGeorgehelpful and Relations by the for townUniversity Center Business-Government Fund Provost is gratefully acknowledged. Georgetown University Journal of Marketing Vol. 61 (July 1997), 1-20 ever, has been the criticism of the targetingof alcohol and tobacco products, notably Uptown and Dakota cigarettes and PowerMastermalt liquor. In many respects, targeting epitomizes the marketing concept. Nonetheless, on occasion it has resultedin controversy and even has been criticized as unethical. This seemingly paradoxicaloutcome has received little theoreticalor empirical scrutiny; yet it is clearly of importanceto marketand ing theory and practice. Certainly, marketers public polmust respond if there is public disquiet over taricymakers geting and therefore should be interestedin its causes and consequences. Marketers might need to be especially responsiveif theirpracticesresultin a diminishedreputation for the firm, lost sales, and potentially, the regulationof tarit to geting. Froma theoreticalstandpoint, is important determine whetherthere are boundaryconditionsto the assumed benefits of the targetingconcept; in other words, the potential for controversy and ethical concern might suggest that for targetingis inappropriate some productsand markets. More broadly, there well could be legitimate concern about the ethics of targeting vulnerable consumers with harmful products. Marketersmight respond to public disquiet over targeting because of the possibility of adverse economic consequences. However, they also are expected to make ethical marketing decisions (Laczniak and Murphy 1993) and ” haverespect and concernfor the welfareof those affected by their decisions” (Smith and Quelch 1993, p. 9). We introduce two cases and review the literature to reveal the extent of the criticism of targeting. We then use these materialsto illustrateour conceptualframework. Next, we report two empirical studies that test our hypotheses about the conditions under which criticism of targeting is more likely to arise, particularlythe characteristicsof the marketing strategy. We also investigate possible consequences of this criticism, such as consumer boycotts and negative word of mouth. We conclude with a discussion of the findings and their implications for marketers, researchers, and public policymakers. / EthicsandTargetMarketing 1 This content downloaded on Tue, 12 Mar 2013 16: 10: 46 PM All use subject to JSTOR Terms and Conditions Criticism of TargetMarketing The new productintroductions Uptown and PowerMaster of were terminated prematurely. This was not because they failed to gain acceptance with their target markets, as is more typical, but because of controversyover their targeting, as described in the two cases (Smith 1996) discussed subsequently. In each case, we show first how the logic underlying the strategies of the companies involved was compelling and the execution of these strategiesapparently sound. We then provide the outcomes of the strategies and the explanationsmade for the controversyand forjudgments of the strategiesas unethical. By presentingwhat appearsto be both ” good marketing” and ” badethics” in these cases, we illustratethe paradoxof the criticism of targeting. Case 1: RJR’s Uptown and Dakota Cigarettes In December 1989, the R. J. Reynolds Tobacco Company (RJR) announcedplans for Uptown, a cigarette designed to appeal to black smokers. With cigarette sales declining, tobacco companies were aggressively seeking new customers. Whereas 29% of the adult U. S. populationsmoked, the figure was 34% for blacks. Marketresearchshowed 69% of African-Americansmokers preferredmenthol (compared with 27% for all smokers)and thatmany blacks would favor Uptown’s lightermenthol. Advertisementssuggesting glamour, high fashion, and night life were plannedfor black-oriented media. Moreover, the cigaretteswere to be packedfilter down, another response to research on preferences of black smokers. The attack on Uptown by the black Health and Human In Services SecretaryLouis W. Sullivan was unprecedented. January1990, he charged, toward This brandis cynicallyand deliberately targeted black Americans … when our people desperatelyneed the Case 2: Heileman’s PowerMaster Malt Liquor Alcohol producersalso were facing declining consumption and increasingly were targeting heavy users. In 1990, G. Heileman Brewing Company had seen its sales volume decline for the seventh year in a row and was desperatefor successful new products. Malt liquor, a productdisproportionately consumed by blacks and in low-income neighborhoods, was one of the few growth categories. An industry commentator noted, ” Thecategory was developedfor a consumer who wanted a fast buzz, so the advertisingplays that up” (Freedman 1991a, p. B4). In June 1991, Heileman announcedplans for a new malt liquor called PowerMaster. At 5. 9% alcohol, it was 31% strongerthan Heileman’s Colt 45, the marketleader, and had 65% more alcohol than regular beer. PowerMaster caused an uproar among anti-alcohol groupsand black leaders. The Centerfor Science in the Public Interest(CSPI)-having earlier reportedthat black men had a 40% higher death rate from cirrhosis of the liver than did whites-asked the brewer to stop distribution, stating, ” higher octane alcoholic beverages have no place on the market, especially in communities where residents already suffer disproportionately from alcohol and other drug problems” (Bureau of National Affairs 1991, p. 41). Boycotts were planned. On June 20, the Bureau of Alcohol, Tobacco and Firearmsannouncedthat its approvalof the PowerMaster label was a mistake and requiredHeileman to drop the word ” power.” PowerMasterbecame ” a magnet of controversy from the moment it rearedits alcohol-enhancedhead. Federalofficials, industryleaders, black activists, and media types weighed in with protests that PowerMaster… was an example of a bad product, bad marketing, and, essentially, a bad idea”(Farnham1992, p. 82). On July 3, Heileman withdrew PowerMaster, because ” the brandname was the product” (Freedman199lb, p. B1). While an anti-alcoholgroup suggested that brewers ” will think twice before targeting vulnerable, inner-city groups again” (Freedman 1991b, p. B1), the Beer Institute accused such critics of patronizing blacks and Hispanics. Fortune described PowerMasteras one of the biggest business goofs of 1991, noting that ” targeting black consumers with anything less wholesome than farinahas become politically risky”(Farnham1992, p. 82). Relevant Literature The research literatureon targeting-relatedethical issues includes studies of direct marketingbecause of privacyconcerns (e. g., Smith 1994) and of the targeting of children (e. g., Pollay 1993) and the elderly (e. g., Benet, Pitts, and LaTour1993). Disquiet over the targetingof both young and elderly consumers rests on the well-established vulnerability of these consumers; for example, Mazis and colleagues (1992, p. 22) write that ” childrenor young adults … [are] vulnerableconsumers … not in a position to make mature, rationaljudgments.” Indeed, it is this vulnerabilityof children that underlies Food and Drug Administration(FDA) restrictionson tobacco marketing, which took effect February 1997. These regulations include a ban on all outdoor advertisingwithin 1000 feet of schools and a ” tombstone” formatfor all other advertisingaccessible to children (Hernandez 1996). However, at issue in the cases presentedhere Uptown’s messageis more messageof healthpromotion, for and death a group disease, moresuffering more already illnessand morethanits shareof smoking-related bearing 1990, p. B8). (Schiffman mortality Given extensive media criticism, RJRcanceled plans for Uptown, noting, ” We regret that a small coalition of believes that black smokers anti-smokingzealots apparently are somehow different from others who choose to smoke” (Specter 1990, p. A3). A smoking policy institute spokeswoman argued, ” Targeting… is a standardprocedure for marketing…. This is a productthat is deadly when used as intended; that’s the real issue” (Specter and Farhi 1990, p. A4). Soon after, RJR also changed the strategy for its Dakota cigarette targetedat white, 18-24-year-old, ” virile” females. RJRhad found itself ” underheavy fire for a plan to marketthe new brandto one of the industry’smost vulnerable segments: young, poorly-educated, blue-collar women” (Freedmanand McCarthy 1990, p. B ). RJR expanded the target to include males, but Dakota failed in test markets. Referring to the confrontation over both Uptown and Dakota, one analyst noted, ” The well-to-do and well-educated … have quit smoking. Those who remain are the disadvantaged. It’s logical to targetthem, except you are sending a message society can’t accept” (Freedman and McCarthy 1990, p. B ). 2/ Journalof Marketing, 1997 July This content downloaded on Tue, 12 Mar 2013 16: 10: 46 PM All use subject to JSTOR Terms and Conditions and in this study as a whole is the targetingof nonelderly, adultconsumers, a grouppresumedhithertoto be fully capable of consumerdecision making. Only recently-and controversially-has it been suggested that a subset of these The criticism adult consumers be considered ” vulnerable.” of targeting vulnerable adult consumers has received little research attention. This is notwithstandingthe cases presented here, the ensuing mediadebate(e. g., Bromberg1990; Calfee 1991; Pomeroy et al. 1992; Zuckoff 1992), and even calls for legislation to restricttargeting(Schlossberg 1990). The literature on targeting adult consumer segments with productsconsidered ” harmful” viewed as ” vulnerable” was promptedlargelyby the events describedin Cases 1 and 2. It is almost exclusively specific to women and minorities and tobacco and alcohol products. Pollay, Lee, and CarterWhitney’s (1992) content analysis of advertisementsfinds that ethnic segmentationof the cigarette marketis not new, dating back to 1950 or earlier. They maintainthat criticism of segmentationhas occurred only with problematicproducts and, in such cases, segmentation’sefficiency ” delivers more death and disease, not more benefits, and provides a disservice, not a service” (p. 46). Pollay (1993) finds that cigarettecompanies have targetedwomen (and youth) since the 1920s, though a study by Rifon, Vanden Bergh, and Katrak(1994) indicates that women have not been disproportionatelytargetedrelative to men. Spratlen (1993) suggests that cigarette advertisingtargetingblacks is unethical because of the vulnerabilityof this marketand the disproportionalityof negativeeffects in this group’s consumption. In a review of alcohol promotion to ethnic minorities, Williams and Mulher (1993) observe that pending legislation would limit alcohol advertisingaimed at ” greaterrisk” consumers, such as heavy drinkers, young people, ethnic minority groups, and women. They differentiate between good ethnic targeting and targeting ” that may have detrimental consequences on consumerwell-being” (p. 69), suggesting thatthe criterionto be appliedis whetheradvertising increases alcohol consumption. However, they found little conclusive evidence of a causal relationshipin a review of empiricalstudies. Moore, Williams, and Quails (1996) agree (though we note that any review of the impact of targeting on consumption cannot be limited to advertising alone because targetingencompasses all elements of the marketing mix). They suggest that the targeting of alcohol and tobacco to ethnic minorities is ” good business” ratherthan conscious racism but propose that there could be a perception of a racist motivationundercertain conditions, including whethertargetingtakes advantageof consumerswho are more vulnerabledue to income, education, knowledge, age, maturity, life’s circumstances, and so on. Ringold (1995), in a review of social criticisms of the targeting of cigarettes and alcohol, observes that targeting in that involves ” equalparticipants” transactionstypically is regardedas acceptable, whereas ” objectionsare almost ceror tain if targetingentails ‘disadvantaged’ ‘vulnerable’conin sumers participating transactionsinvolving productssuch as alcohol and cigarettes”(p. 579). She proposes that proponents of targeting subscribe to a ” competent consumer model,” in which consumersare generally skepticalof commercial information, recognizing its limitations and useful- ness. By contrast, critics of targetingsubscribeto a ” vulnerable consumer model,” in which vulnerability is due to a diminished capacity to understand advertising, product effects, or both. Ringold questions claims of consumervulnerabilityand argues that there is consumer skepticism and of substantialunderstanding the harmfuleffects of tobacco and alcohol, even among young consumers. However, Smith and Quelch (1993, p. 193) submitthattargeting” createsvictims of market segments” when harmful products are involved. Finally, Treise and colleagues (1994), in a survey of consumer perceptions of a variety of advertising practices, find that consumers agreed that liquor and cigarette advertisingto inner-citymarketsis unethical, as are lottery advertisementsto low-income consumers. In summary, the literatureexamines the social disquiet over targetingbut is largely specific to alcohol and tobacco productstargetedat ethnic minorityconsumers. Moreover, it offers only limited explanationsfor this criticism and gives little consideration to the prospect of concern about other productsor targets. Our cases and the literatureprovide the basis for a conceptual frameworkthat informs understandingof ethical concern and controversyover targetingand guides our empirical investigation of the conditions under which criticism could arise. The key components of targetingstrategiesare the product and the target. The cases described previously indicate that criticism of targetingis relatedto two key factors:(1) the perceivedharmfulnessof the productand (2) the perceived vulnerabilityof the target. As Ringold (1995, p. 579) suggests, ” the social acceptability of targeting is largely a function of individual commentators’judgments about particularconsumers and specific products.” In the RJRcase, criticism focused on the harmfulnessof cigarettes coupled with the perceived vulnerability of blacks and young, poorly educated, white women. In the Heileman case, criticism was based on the harmfulnessof alcohol, particularly in inner-cityneighborhoods, coupled with the perceived vulnerability of low-income minority consumers. Accordingly, if productsare considered to be more or less harmfuland targets viewed as high or low in vulnerability, we can conceive of four generic types of targetingstrategies, as shown with illustrative examples in Figure 1. Product harmfulness, target vulnerability, and their roles in the process of ethical evaluationof targeting strategiesare discussed more fully in the following sections. Product Harmfulness The Code of Ethics of the American MarketingAssociation should conform to the basic rule of prostates that marketers fessional ethics not to do harm knowingly, and they should offer products and services that are safe and fit for their intendeduses. Hence, targetingcould be criticized and evaluated as unethical when it involves products perceived as harmfulbecause of the marketer’sobligation to avoid causing harm. Much of the discussion of productsafety in the literaturerelates to physical harm(e. g., Laczniakand Murphy 1993, pp. 84-85). However, we also would include eco/ Ethics andTargetMarketing3 Conceptual Framework This content downloaded on Tue, 12 Mar 2013 16: 10: 46 PM All use subject to JSTOR Terms and Conditions FIGURE 1 Types of Targeting Strategies1 PRODUCT Less Harmful Low MoreHarmful Vulnerability e. g., Low-fat e. g., High-interest to rate creditcard hamburger to suburban above-average income consumer consumer target target TARGET Strategy 1 Strategy2 Low-nicotine e. g., High-alcohol e. g., content malt cigaretteto blackconsumer liquorto less than hightarget2 school educated consumer target Strategy3 Strategy4 High Vulnerability are and to harmfulness targetvulnerability to 1References product couldbe conthe perceptions these factors. of Also, bothin reality each is for ceivedas having continuum; a however, ourpurposes, into divided twocategories. to consider would 2Somepeoplenever cigarettes be ina less harmnature evidenceof the harmful becauseof the strong fulcategory in levcan we of smoking. However, notethatcigarettes differ their the in a els of harm(e. g., through reduction benzo(a)pyrene, canin and cerouscompound tobacco smoke) hencesometypescanbe considered ” lessharmful.” nomic harm (e. g., overpaying tax because of a faulty tax software package) and psychological harm(e. g., embarrassment from a hair coloring productthat results in an unnatural color). If perceived productharmfulnesswere the sole explanation for criticism of targeting, the implicationsfor managers and policymakers would be well established and the threat to a core tenet of marketingobviated. However, though it is true that there has been disquiet only about targetingthat involves ” harmful” products(Pollay, Lee, and Carter-Whitthis criticism has been expressed only when tarney 1992), geting also involves ” vulnerable” populations. Target Vulnerability refers to a susceptibility to injury or to being Vulnerability taken advantage of by another person. Benet, Pitts, and LaTour(1993, p. 46) referto the perceptionof the elderly as ” a vulnerablegroup, more susceptibleto unscrupulous business practices than younger groups.” Andreasenand Manning (1990, p. 13) refer to ” those who are at a disadvantage in exchange relationshipswhere that disadvantageis attributable to characteristicsthat are largely not controllableby them at the time of transaction.” special issue of the JourA nal of Public Policy & Marketing (Spring 1995), which was vulnerabilitiesof the medically underserved, African-Americans, women, rural residents, Mexican immigrants, drug addicts and alcoholics, children, and the recently bereaved. Although consumervulnerabilityis not defined specifically, these articles provide some indicationof who is considered to be a vulnerable consumer and how this vulnerability influences consumptionbehaviors. In a review of their legal status, Morgan, Schuler, and Stoltman(1995, p. 267) suggest vulnerableconsumersoriginally were conceived as ” unusually susceptible … small groups of consumers who have idiosyncratic reactions to products that are otherwise harmless when used by most people.” They propose a broaderview of vulnerabilityconsistent with recent litigation and statutes that goes beyond physical hypersensitivityto include persons ” incapableof making informed decisions at the time of purchase” (p. 272). Theirexpandedtypology of vulnerability(pp. 273-74) includes ” physical competency” (e. g., hearing disabilities associated with aging), ” mental competency” (e. g., dyslexia), and ” level of sophistication”(e. g., low socioeconomic background), as well as physical hypersensitivities. In keeping with these examples and definitions, we define vulnerableconsumersas those who are more susceptible to economic, physical, or psychological harmin, or as a result of, economic transactionsbecause of characteristics that limit their ability to maximize their utility and wellbeing. Limiting characteristicsamong adult consumers can include a low level of education or income. These characteristics, in addition, can be associated with ethnicity and domicile (e. g., inner-city residents frequently have lower levels of education and income). In additionto these demographicfactors, a varietyof othervariablesalso can limit the consumer, such as low cognitive ability, asymmetryof information, and restrictedmobility. Consumer vulnerabilityhas not been researchedextensively and, as yet, is inadequately understood. However, prior criticism of targeting has emphasized certain demographic characteristicsgenerallyperceived to be associated with vulnerability. Accordingly, in our investigation, we focus on ethnicity, domicile, and low levels of educationand income. Ethical Evaluation of Targeting Strategies The media, special interestgroups, and some public officials have criticized the targetingof certain products. Smith and Quelch (1993, p. 193) highlight the role of organized and vocal interestgroups in pressuringcompanies to stop targeting. As we note in the first case, RJRblamed” a small coalition of anti-smoking zealots.” Moreover, the extent of the debate over PowerMasterwas due in part to a well-orchestrated campaign by CSPI that resulted in Heileman being ” sandbagged” by the media. 2 However, it has not been establishedwhetherthereis a broadersocietal concernabout targeting. Indeed, Calfee (1991, p. 18), refers to the ” astonishing degree of political suspicion that has descended on the practice of targeting” and suggests that the media got it wrong. Accordingly, do the ethical evaluationsof targeting withGeorgeHacker, Director of 2Interview the firstauthor by theAlcoholPolicyProject CSPI, January at 1995. devoted to vulnerable populations, included articles on the 4 / Journalof Marketing, 1997 July This content downloaded on Tue, 12 Mar 2013 16: 10: 46 PM All use subject to JSTOR Terms and Conditions strategiesby observersreportedin the media reflect those of a broader group of public observers (” publics”), be they consumers of the product, members of the media, the government or other organizations(e. g., churches, groups that representthe targetedconsumer), or the public at large? We can identify a process of ethical evaluationby these publics, incorporatingproduct harmfulnessand target vulnerability. This process begins with a marketerdeveloping a are targetingstrategy; its key characteristics the productand the target. Perceptions of product harmfulness and target vulnerabilityaffect publics’ judgments of the ethics of the as strategy(” ethicalevaluations,” operationalizedin the following section), which in turn influence any behavioral responses. Approving and disapprovingbehaviors provide feedback to the marketerand can affect subsequentmarketing strategies. We state the differentpossible ethical evaluations of targeting strategies and their consequences more formally in the following section, as hypotheses. To isolate the effect of productharm, in H2 we hold vulnerability constant and compare ethical evaluations for productsdiffering in perceivedharmthat are targetedat the same high-vulnerability segment. Hence, in comparing strategies3 and 4, we expect that ethical evaluationswill be lower when productharm is greater: to of a product a highH2: A strategy targeting moreharmful 4) vulnerability segment (strategy will receivelowerethicalevaluations willa strategy targeting lessharmthan a of ful product a high-vulnerability to 3). (strategy segment If ethical concern over targetingwere only a function of the harmfulnessof the productinvolved, as some commentators have suggested, we would not anticipateany difference in ethical evaluationsfor strategiesinvolving the same productstargetedat different segments. However, our conceptual frameworksuggests that target vulnerability does influence ethical evaluations of targeting strategies. In H3, We we isolate the effect of targetvulnerability. hold product harmconstantand compareethical evaluationsfor products of high perceivedharmthatare targetedat segmentsthatdiffer in their levels of perceived vulnerability: to A of a product a highH3: strategy targeting moreharmful lower ethical 4) (strategy willreceive vulnerability segment a of than evaluations will a strategy targeting moreharmful to 2). (strategy segment product a low-vulnerability Marketersare likely to consider whetherethical evaluations lead to action, especially expressions of criticism or consumer boycotts. Such disapprovingbehaviors can have powerful effects, as RJR and Heileman discovered. But, conceivably, publics could respond positively by praisinga company’s actions, in recognitionof segmentation’sbeneficial effects. Our conceptual frameworksuggests that a consumer who evaluatesa strategyas less ethical is more likely to engage in a disapproving behavior and less likely to engage in an approvingbehaviorthan is a consumer with a more ethical evaluation. Because of the difficulties in eliciting behaviorin a controlled study, the following hypotheses specify behavioralintentions: to (approving) H4a (H4b): Intentions engage in disapproving will to behaviors be related negatively (positively) theethicalevaluations a strategy that receives. These hypotheses identify the conditions under which ethical concernand criticismcould arise. In additionto these considerations, we investigatedwhat types of consumers, in terms of demographicdescriptors, are most critical of targeting and most likely to take disapprovingor approving actions. Hypotheses Our empiricalstudies assess whetherCases 1 and 2 are generalizable. We attemptto establish the likelihood of public disquiet over targeting that involves a variety of possible ” vulnerable” products. We also investargetsand ” harmful” tigate the basis for this concern: Is it due to product harmfulness, targetvulnerability, or both? We doubt that vulnerabilityalone is the basis for criticism, because it appearsto have resultedonly over targeting that also involves ” harmful” products. Indeed, though targeting minorities has resulted in controversy because of their perceived vulnerability, these segments historically have been undertargeted, and some targeting of minority marketshas been praised, such as Mattel’s African-American doll, Shani, and Estee Lauder’s ” All Skins” makeup (Zuckoff 1992). Moreover, our definition of vulnerability indicates that susceptibility to harm is key to vulnerability, suggesting a role for productharmfulness. Many commentators (e. g., Bromberg 1990; Schlossberg 1990; Zuckoff 1992) maintainthatthe questioningof Uptown, Dakota, and PowerMasterwas entirely due to productharmfulness:” It’s not the targetingthat’s the enemy. It’s the product”(Moore 1990, p. 5D). However, given evidence in the cases and the literaturethat perceived consumer vulnerability is also an issue, the alternateexplanationis that it is the combination of perceived product harmfulness and target vulnerability that results in criticism of targeting. In many respects, this is a more serious matterfor marketersbecause targetingitself is being challenged. The cases suggest that the least ethical strategy among the four types proposedwithin the conceptual frameworkis targeting a more harmful product to a high-vulnerability segment. If the levels of both productharmfulnessand target vulnerabilityare elevated, we would expect there to be substantialethical concern. This leads to our first hypothesis (see Figure 1 for labeling of strategies): a to of product a highHi: A strategy targeting moreharmful 4) (strategy will receivelowerethivulnerability segment a of than cal evaluations will a strategy targeting lessharmto ful product a low-vulnerability 1). (strategy segment Method We tested our hypotheses in two largely similar studies that differed by product type and target demographics. A between-subjects design was used to minimize demand effects. The design was a 2 (more/less harmful) x 2 (high/low vulnerability) full factorial, in which each cell one represented of the four strategiesin Figure 1. Each study included two target characteristics (e. g., race, education) and two productclasses (e. g., cigarettes, hamburgers), using / EthicsandTargetMarketing5 This content downloaded on Tue, 12 Mar 2013 16: 10: 46 PM All use subject to JSTOR Terms and Conditions all possible combinations (i. e., each target characteristic and each product class was representedin all four cells of the design). There was a total of 16 scenarios in each study. Independent Variables We operationalizedtarget vulnerability as a demographic characteristicgenerally perceived to limit the consumer’s ability to maximize utility and well-being in economic transactions. We operationalizedproductharm as a product attribute perceivedto be more or less harmful, dependingon the amount in the product(e. g., nicotine in cigarettes). The high level equaled the highest amountavailable in the market. The low level equaledthe lowest amountavailablein the market for all products, except for malt liquor in Study II (because there is a legal minimum in some states for malt liquor, we used the lowest amount allowed by law: 5. 1%). We used the same productsacross all four strategiesto prevent variations between product classes from affecting results and avoided all brand names to prevent influence from previous publicity or opinions. Pretest. Ourchoice of targetingstrategieswas guided by a pretest that examined perceptions of less versus more A harmfulproductsand low versus high targetvulnerability. rated eight productand four sample of 59 college students targetdescriptions. We chose the productsand targetstested from those that have elicited some criticism in the media (not limited to Cases 1 and 2). Scenarios. Each scenario included a targetdescribed in terms of a vulnerability characteristic and a product To described by its harmfulproductattribute. make the raw values of the harmful attributes more meaningful, we included the mean amountacross all major brandsas a reference. For example: introto A largecompany, well-known thepublic, recently to is This duceda newcigarette. cigarette intended appeal to consumers who are college graduates. The cigarettehas of amount nicotine The of . 05 milligrams nicotine. average is acrossall companies . 81 milligrams. percigarette Dependent Variables Ethical evaluations. To measurethe ethical evaluationof the scenarios, we adopted Reidenbachand Robin’s (1990) multidimensionalethics scale (MES). 3Previousstudies find alphas of . 71 to . 92, for each of the three MES subscales, plus moderateto good convergent, discriminant, construct, and predictive validity (Reidenbach and Robin 1990; Reidenbach, Robin, and Dawson 1991). Althoughcriticisms have been raised (Hansen 1992; Skipperand Hyman 1993; also see replies by Reidenbachand Robin 1993, 1995), the scale is used widely to measure ethical judgments (e. g., Flory et al. 1992; LaTourand Henthorne 1994; Robertson and Ross 1995; Tansey, Hyman, and Brown 1992). Furthermore, our search of the literaturedid not reveal anothervalto idated, multidimensionalscale appropriate evaluatingthe ethics of marketing strategies. Therefore, we adopted the MES scale. However, we modified the instructionsfor the scale to explain ” unspokenpromise” and ” unwrittencontract,” because respondents reported difficulty in understandingthese items in the pretest. The three dimensions or subscales are ” moral equity” (i. e., just/unjust, fair/unfair, morally right/not morally right, and acceptable/not acceptable to my family), ” relativistic” (i. e., culturally acceptable/unacceptableand traditionally acceptable/unacceptable), and ” contractualism” (i. e., violates/does not violate an unspoken promise and violates/does not violate an unwrittencontract). The moral equity dimension is broad-based. Reidenbach and Robin (1990, p. 646) suggest it ” relies heavily on lessons from our early training that we receive in the home regarding fairness, right and wrong as communicated throughchildhood lessons of sharing, religious training, morals from fairy tales, and fables.” The relativism dimension, they suggest (1990, p. 646) is ” more concerned with the guidelines, requirements, and parameters inherent in the social/cultural system than with individual considerations” and can be acquiredlater in life. The contractualism dimension is in keeping with the notion of a social contract between business and society (Donaldson and Dunfee 1994). If support for hypotheses about ethical evaluations is more evident on one dimension than another, this could allow some speculation as to the rationale for ethical judgments of targeting. For example, differences in the evaluation of targeting strategies on the contractualism dimension might suggest that respondents are concerned about the violation of the ” ethics of exchange” (Reidenbach and Robin 1990, p. 647). This could have implications for responses by marketers. Conceivably, they might argue that the ” terms” of the social contract are misunderstood and that targeted consumers’ freedom to choose is paramount. Behavioral intentions. The second set of dependentvariables measured the likelihood of performingeach of five disapproving behaviors (e. g., stop buying the company’s products)and two approvingbehaviors (e. g., tell friends to buy the company’s products). The items were based on descriptions of consumer activism (Berry 1977; Vogel 1978). Suitable preexistingscales were not available. Reliability and validity of the MES. We examined the reliability and validity in our study of the MES scale, the more critical dependentvariable. The coefficient alphas for the three dimensions of the MES scale indicate satisfactory internal reliability as follows: moral equity dimension . 93 . 86 (Study I) and . 91 (Study II), contractualism and . 85, and relativism . 70 and . 69 (these alphas are consistent with previous studies). To test for constructvalidity, we analyzedthe correlations between each ethical dimension and each behavior. As expected, each ethical dimension correlates negatively and significantly with each disapprovingbehavior, and there are significant, positive correlations between 3TheMES providesa measureof respondents’evaluationsof the ethics of our scenarios and provides some indicationof their rationale for these evaluations. It cannot provide definitive valuejudgments on the targetingstrategies described. This is a task of normative marketingethics, as laterexplained. ethical dimensions and approvingbehaviors (see Table 1). 6 / Journalof Marketing, 1997 July This content downloaded on Tue, 12 Mar 2013 16: 10: 46 PM All use subject to JSTOR Terms and Conditions TABLE 1 Correlations Between Ethical Evaluations and Behavioral Intentions Ethical Dimensions Behavioral Intentions Disapproving Behaviors Stop buying the company’s products Tell your friends to stop buying the company’s products Complain to a newspaper or radio or television station Call or write company to complain about action Write member of Congress to criticize company Approving Behaviors Tell friends to buy company’s products Call or write company to praise the action ap < . 06. *p < . 05. **p < . 01. ***p < . 005. Moral Equity Study I -. 68*** -. 70*** -. 37*** -. 40*** -. 36*** Study II -. 59*** -. 57*** -. 35*** -. 36*** -. 39*** Relativism Study I -. 52*** -. 53*** -. 32*** -. 33*** -. 32*** Study II -. 23*** -. 23*** -. 18*** -. 17** -. 19*** Contractualism Study I Study II -. 41 ** -. 27*** -. 43*** -. 30*** -. 29*** -. 29*** -. 30*** -. 23*** -. 29*** -. 33*** . 14*** -. 07 . 02 . 00 . 28*** . 09* . 15** . 10a . 18*** . 02 . 02 . 04 Overall, these results supportthe reliability and validity of the MES scale dimensions. 4 Sample and Procedure We recruited convenience samples of adult respondents while they were waiting to see one of two historical sites or to board a train at a station in a South Atlantic city. These locations were selected to include adults (i. e., nonstudent respondents), to achieve both geographicaland racial diversity (and obtain a sample close to being nationallyrepresentative), and to ensure the cooperationof respondentswilling to take the 20 minutes requiredto complete the questionnaire. Trained assistants (uninformed about the study's hypotheses) collected the data from people waiting for relatively long times, either to gain admission (a 60-90 minute wait) or to boarda train(at least 30 minutesfrom departure). These assistants asked potential subjects to answer questions individually on a self-administered, written survey. Each subject read one scenario and then answeredthe items for ethical evaluation followed by those for disapproving and approvingbehaviors (in a randomorder). Next, responchecks for perceivedproddents answeredthe manipulation uct harmfulness(three items) and perceived targetvulnerability (two items). Last, they answered demographicquestions. In addition, respondentsansweredquestions relatedto ongoing research on consumer vulnerability, including items (before the scenario) involving individualpersonality differences. In Study I only, they also answered questions 4As further evidence of the validity of the MES, we tested whether its dimensions measure ethical evaluations and the value of adding a dimension on virtue ethics (given recent interest in As virtueethics applied to marketing). expected, each dimension of the MES correlatessignificantly, in the appropriate direction, with each of two univariatemeasuresof ethics, which supportsthe constructvalidity of the scale. A virtue ethics scale, developed for this test, correlatessignificantly and positively with each MES dimension, and the evidence of supportfor Hi-H3 does not change when tested with the virtue ethics scale. These results suggest that the virtue scale is redundant of the existing MES dimensions and thereforenot a necessary addition. (immediately before the demographic items) about their own vulnerabilitywhen buying a varietyof products. Study I Sample In recruitingrespondentsfor this study, we made a specific effort to ask blacks as well as whites to respondbecause the survey included targets described as white or black. Our goal was to have the sample's mix of blacks and whites approximatethat of the U. S. population. Of 720 people approached, 94 refused to participate, which resulted in 626 surveys collected (87% participation rate); 522 surveys (83%) were complete on the dependent variables and used for analysis. The sample was 52% female, 79% white, and 10% black. Median age category was 38-47 years, median household income category was $45, 000-$59, 999, and median education was a college degree. The U. S. residents (84%) were from all parts of the country; the largest group (19%) was from the South Atlantic region, in which the data were collected. Comparedwith the U. S. adult population, the domestic portion of the sample was closely representative on gender and race but had higher levels of education and household income. The distribution of age groups and geographicalregion within the domestic sample approximatedthe U. S. adult population, except that people 38-47 years of age were overrepresented, and those 68 the years of age and older were underrepresented; Midand Atlantic region was underrepresented, West NorthCentral was overrepresented. Target Characteristics and Product Classes The target characteristicsfor vulnerability were education and race. The productattributesfor harmfulnesswere nicotine in cigarettes and fat in fast-food hamburgers. Because tobacco is a " sin" productand might be a special case, we also used fast-food hamburgers, a relatively benign (" nonsin") product. Although targetingstrategiesfor this product Ethics andTargetMarketing 7 / This content downloaded on Tue, 12 Mar 2013 16: 10: 46 PM All use subject to JSTOR Terms and Conditions have elicited criticism (Farham 1992; Freedman 1990), it has been less vigorous than that for cigarettes. All manipulations were successful. Targets with less than a high school education were rated significantly more vulnerable(at 5. 04 on a seven-point scale) than were targets with a college education (3. 10) (t = 12. 0, df = 259, p < . 0001), as were blacks (4. 39) versus whites (3. 33) (t = 6. 81, df = 242, p < . 0001). Cigaretteswith 2. 1 milligramsof nicotine were rated as significantly more harmful(at 5. 94 on a seven-point scale) than were cigarettes with . 5 milligrams (5. 11) (t = 3. 95, df = 251, p < . 0001), as were hamburgers with 63 gramsof fat (5. 75) versus those with 9 grams (3. 04) (t = 15. 2, df = 250, p < . 0001). The average amountsof the harmful attributes(as found in the marketplace)were . 81 milligramsof nicotine and 29 grams of fat, and these values were included in the scenarios. It should be noted that to lessen demand effects, each respondent answered the manipulationquestions on product harm for three product attributes, including the one in his or her scenario; we used the same approach for target vulnerability. However, manipulationcheck results are based only on subjects who saw the particular productor segment in his or her scenario. Tests of Hypotheses Priorto testing the hypotheses, we determinedthe appropriateness of combining scenarios that differed on targetcharacteristicor productclass. The test was an ANOVA with the main effect and all two-way interactionsfor the harm, vulnerability, productclass, and targetcharacteristicterms. We conducteda separatetest for each dimension of ethical eval- uations. There was a significant interactionbetween target characteristic(educationor race) and vulnerability(high or low), for each of the three dimensions: moral equity (F(1, 520) = 5. 92, p < . 05), relativism(F(1, 520) = 6. 13, p < . 05), and contractualism(F(1, 520) = 5. 45, p < . 05). Therefore, tests involving vulnerabilitywere conductedseparately for education versus race. Likewise, to determine whether samples from the three collection locations could be combined, we tested for interactionsbetween location and the manipulatedvariables; none were significant, so we combined locations in the following analyses. We show results for Hi-H3 in Table 2. Hi-H3 predict differences in ethical evaluationsof pairs of targetingstrategies. They were tested with contrastsfrom a between-subjects, one-way (four types of scenarios) ANOVA for each dimension of ethical evaluation. H1 is supportedfor both race and education scenarios, with each dimension of ethical evaluation. As expected, the more harmful/high-vulnerability scenario(strategy4) receives lower ethical evaluations scenario (stratthan does the less harmful/low-vulnerability It is also importantto note that the mean values for egy 1). strategy4 (more harmful/highvulnerability)were generally well below the scale midpoint and in contrastto mean values above the midpointfor strategy 1 (less harmful/lowvulnerability). These absolute reactionsof respondentssuggest that not only is strategy4 judged less ethical than strategy 1, but it also might be considered unethical by respondents, because it was viewed by far as the most negative strategy (see Table 2). TABLE 2 Tests of H1-H3 on Ethical Evaluations in Study I Mean Responses for Ethical Evaluations1 Strategies (HarmfulnessNulnerability) Scenarios with Education Moralequitydimension Relativisticdimension dimension Contractualism Scenarios with Race Moralequity dimension Relativisticdimension dimension Contractualism More/High(4) 2. 34 3. 11 3. 27 2. 34 2. 89 3. 39 Less/Low (1) 4. 43 4. 98 4. 91 3. 50 3. 87 4. 11 Less/High (3) 3. 69 3. 97 4. 45 3. 59 3. 94 4. 31 More/Low(2) 2. 98 3. 76 4. 28 2. 27 3. 08 3. 42 Significance Levels of Contrasts for Ethical Evaluations2 Scenarios with Education H1: Moreharmful/high (1) vulnerability < Less harmful/low (4) vulnerability H2: Moreharmful/high vulnerability (3) vulnerability < Less harmful/high (4) H3: Moreharmful/high vulnerability (2) vulnerability < Moreharmful/low (4) MoralEquity . 0001 . 0001 . 05 Relativism . 0001 . 01 . 05 Contractualism . 0001 . 0001 . 0001 MoralEquity Relativism Contractualism Scenarios with Race . 0005 . 005 . 05 H1: Moreharmful/high vulnerability < Less harmful/low (4) (1) vulnerability . 01 . 0001 . 0005 (3) vulnerability < Less harmful/high (4) vulnerability H2: Moreharmful/high ns ns ns (2) vulnerability < Moreharmful/low (4) vulnerability H3: Moreharmful/high 1-7 1From scales, where7 was moreethical. 2Thestatistical used didnotgivea t-valueforeach contrast, relied critical on t but valuesof t, as follows:= 1. 97forp = . 05, t = 2. 60 package for p = . 01, t = 2. 83 for p = . 005, t = 3. 53 for p = . 0005, and t = 3. 95 for p = . 0001. 8 / Journalof Marketing, 1997 July This content downloaded on Tue, 12 Mar 2013 16: 10: 46 PM All use subject to JSTOR Terms and Conditions H2 also is supportedfor both race and educationscenarios with each dimension of ethical evaluations. Product harmfulnesshas the expected effect on ethical evaluations; the more harmful/high-vulnerability scenario (strategy4) is less ethical than the less harmful/high-vulnerability judged scenario (strategy3). H3 is supported for education scenarios, with each dimension of ethical evaluations, but not for race scenarios. Specifically, strategy4, targetingconsumerswith less thana high-school education, is perceivedas less ethical thanstrategy 2, targetingcollege graduates. As expected, targetvulnerability, as well as product harmfulness, affects ethical evaluations. However, despite manipulationchecks indicating that blacks are perceived as more vulnerableconsumers than whites, targeting blacks is not evaluated differently from targetingwhites. H4a (H4b) predicted that intentions for disapproving (approving) behaviors would be related negatively (positively) to ethical evaluations (see Table 1). The significant, negative correlations for each disapproving behavior and each dimension of ethical evaluation supportH4a. Support for H4bis evident in the significant, positive correlationsfor " tell friends to buy company's products" and each ethical dimension and for " praisethe company's action" and moral equity. The most likely behaviorsare " stopbuying the company's products" and " tell friends to stop buying" (see means in Table 4). Demographic Differences Which consumers view targetingstrategies as less ethical? The answer was ascertainedusing regression analysis (see Table 3). Women and older respondentsjudge the scenarios to be less ethical, on each MES dimension. Nonwhitesjudge the scenarios to be less ethical than whites do, on the moral equity and relativism dimensions (on contractualism, this difference between races approachessignificance; p < . 08). Across both race and educationscenarios, thereis no significant relationshipbetween respondents'levels of education and ethical evaluations; however, for education scenarios only, respondentswith less educationjudge them to be less ethical on the moral equity dimension. Lower-income respondentsjudge the scenarios to be less ethical than do higher-incomerespondentson the relativismdimension; for education scenarios only, lower-income respondentsjudge the scenarios to be less ethical on all three dimensions. We also examined more directly whetherethical evaluations differed if the respondentwas " in-target"(i. e., those as respondentswith the same demographiccharacteristic the consumer in the scenario) versus " non-target" (those target outside the target described in the scenario). For the lowvulnerability conditions (strategies 1 and 2), across both race and education scenarios, in-target respondentsevaluated the scenarios as significantlymore ethical thandid nontargetrespondentson the moral equity dimension (p < . 05). conditions The effect was reversedfor the high-vulnerability (strategies 3 and 4); in-targetrespondentsevaluatedthe scenarios as significantly less ethical than did non-target respondents on the moral equity dimension (p < . 05). Accordingly, respondents who possess the demographic characteristicassociated with greaterconsumer vulnerability view the scenarios as less ethical, including those strategies that targethigh-vulnerabilityconsumers. Which consumers are more likely to take disapproving or approving actions (see Table 4)? Women are more disposed to stop buying the company'sproductsor spreadnegative word of mouth but are less likely to do either approving behavior. Older respondentsare more inclined to stop buying and spread negative word of mouth. Nonwhites are more disposed to doing each of the disapprovingbehaviors. Respondents with less education are more likely to spread positive word of mouth. There are no differences related to income. It should be noted that the positive behaviors are much less likely to occur than the negativeones. Other Findings The tests of our hypotheses indicate that both productharm and targetvulnerabilityaffect ethical evaluations. To assess the relative impact of these two factors, we used a series of regressions with each ethical dimension as the dependent variable and four independent variables, each of which referredto the targeting scenario. These independentvariables were perceivedlevel of harmfulnessof the productand perceived vulnerability of the target (data from questions used for the manipulationchecks) and dummy variablesfor and the productclass (i. e., cigarettesversus hamburgers) the characteristic(i. e., race versuseducation). Using stantarget dardizedbeta coefficients, productharmfulnesshas a larger effect on ethical evaluations than does target vulnerability for the moral equity and relativism dimensions (and the coefficients are significantlydifferentatp < . 05), but the two factors have equal impact on contractualism(see Table 5). The regressionresults on the relativeimpactof vulnerability and harmfulnesssuggest thatin comparingmixed strategies, more harmful/low vulnerability (strategy 2) would be viewed as less ethical than less harmful/highvulnerability (strategy 3). Contrastsbetween these scenarios supportthis post hoc hypothesis on each dimension of ethical evaluations for race scenarios (p < . 05) and on the moral equity dimension for education scenarios (p < . 05) (see means in Table 2). We also examined the effect of product harmfulness when targetinglow-vulnerabilityconsumers, by contrasting the two strategies to these groups (i. e., strategy 1 versus strategy 2). For both education and race scenarios, these strategiesdiffered on moralequity (p < . 001) and relativism (p < . 01). In each contrast, the scenariowith the more harmful product(strategy2) is evaluatedas being less ethical than the scenario with the less harmfulproduct(strategy 1) (see means in Table 2). This suggests there could be concern about targetingharmfulproducts, even to low-vulnerability groups. Finally, we contrastedthe two strategies involving less harmful products targetedat groups that differ in vulnerability (i. e., strategy 1 versus strategy 3). The scenario with the high-vulnerability target(strategy3) is evaluatedas less ethical than the scenario with the low-vulnerabilbeing ity target (strategy 1) for education scenarios on the moral equity and relativism dimensions (p < . 05) (see means in Table 2). / EthicsandTargetMarketing 9 This content downloaded on Tue, 12 Mar 2013 16: 10: 46 PM All use subject to JSTOR Terms and Conditions TABLE 3 Differences Across Demographic Groups in Ethical Evaluations for Study I Race1 Dependent Variable for All Scenarios Moral equity Relativism Contractualism . 14** . 11* . 08 Education Gender2 Income Age F (df) Adjusted R2 . 05 . 04 -. 02 -. 22*** -. 18*** -. 14** . 07 . 13** . 06 -. 17*** -. 22*** -. 11 * 9. 7 (5, 465)*** 9. 8 (5, 465)*** 3. 4 (5, 465)*** . 08 . 09 . 03 Dependent Variable for Education Scenarios Only Moral equity Relativism Contractualism . 07 . 10 . 07 . 13* . 06 . 03 -. 18** -. 16** -. 13* . 15* . 22*** . 13* -. 14* -. 20** -. 12 5. 7 (5, 235)*** 6. 5 (5, 235)*** 2. 6 (5, 235)* . 09 . 10 . 03 Dependent Variable for Race Scenarios Only Moral equity Relativism Contractualism . 21*** . 13* . 09 -. 01 . 03 -. 06 -. 25*** -. 19** -. 15* -. 02 . 03 -. 01 -. 18** -. 23*** -. 09 5. 8 (5, 224)*** 4. 4 (5, 224)*** 1. 5 (5, 224) . 09 . 07 . 01 whites coded as 1 and nonwhitescoded as 0. 1Codedas a dummyvariable; women coded as 1 and men coded as 0. 2Codedas a dummyvariable; *p < . 05. **p < . 01. < ***p . 001. Study II We conducted a second study, within the same two-month period, to assess whether the findings in Study I could be replicated using different target characteristicsand product classes. Our purpose was to establish whether we could make broadergeneralizationsabout ethical evaluationsand the consequences of targeting and thereby improve the robustnessof our findings. Sample Of 421 people approached, 59 refused to participate, which resulted in a total of 362 surveys collected (86% participation rate); we obtained 322 (89%) complete surveys for analysis. The sample was 41% female, 78% white, and 10% black. The median age category was 38-47 years, median household income category was $45, 000-$59, 999, and median education was a college degree. The U. S. residents (96% of the sample) were from all partsof the country; the largest group (29%) was from the South Atlantic region, where the data were collected. Comparedwith the adultU. S. population, the domestic portion of the sample had more males, fewer Hispanics, fewer members from the Western United States, and higher levels of both education and household incomes, and was younger. Target Characteristics and Product Classes The targetcharacteristicsfor vulnerabilitywere income and domicile. The product attributesfor harmfulnesswere alcohol content in malt liquor and annualinterestrate for credit cards. Creditcards, which also have received some criticism 10 / Journalof Marketing, 1997 July for targeting (Keats 1994), were included to broaden the findings beyond " sin" products. All manipulations were successful. Consumers with below-averageincome were perceivedas significantlymore vulnerable (4. 95 on a seven-point scale) than were those with above-average income (3. 17) (t = 7. 97, df = 132, p < . 0001), as were consumerswho live in the inner city (4. 99) versus those in the suburbs (3. 21) (t = 7. 90, df = 128, p < . 0001). Malt liquorwith 14%alcohol content was perceived as being more harmful (5. 51 on a seven-point scale) than was malt liquor with 5. 1% alcohol content (4. 34) (t = 4. 66, df = 131, p < . 0001), as were credit cards with a 22. 5% annual interest rate (5. 82) versus cards with a 6. 5% rate (3. 56) (t = 9. 31, df = 147, p < . 0001). The mean amountsof the harmful attributeswere a 6% alcohol content and a 16. 5%interestrate and were included in the scenarios. Tests of Hypotheses As in Study I, we tested whether it was appropriate colto lapse across target dimensions, product classes, and the threecollection locations. Forthe contractualism dimension, none of the relevantinteractionswere significant. For moral equity, there was a significant interactionbetween product class (credit cards or malt liquor) and vulnerability(high or low) (F(2, 315) = 3. 80, p = . 05). Therefore, tests on moral equity involving vulnerability were done separately for credit cards and malt liquor. For the relativism dimension, there was a significant interactionbetween location and vulnerability(F(2, 315) = 4. 64, p = . 01). Respondentsat location 3 had a lower income This content downloaded on Tue, 12 Mar 2013 16: 10: 46 PM All use subject to JSTOR Terms and Conditions TABLE 4 Results for Behavioral Intentions in Study I Mean Values of Behavioral Intentions by Strategies (HarmfulnessNulnerability)l More/High Less/Low Less/High More/Low Race2 (2) (3) (4) (1) Intentions for Disapproving Behaviors Stop buyingthe company'sproducts Tellyourfriendsto stop buyingthe company'sproducts Complainto a newspaper or radio or televisionstation Call or writecompanyto complain about action Writememberof Congress to criticizecompany 5. 53 5. 62 3. 08 3. 61 2. 95 3. 41 3. 49 2. 17 2. 36 2. 20 3. 87 3. 95 2. 43 2. 60 2. 53 2. 64 1. 86 5. 57 5. 47 2. 69 2. 98 2. 72 1. 96 1. 73 -. 10* -. 16*** -. 24*** -. 27*** -. 18*** -. 08 -. 09 Standardized B Across Dem Education Gender3 A -. 08 -. 06 . 01 . 05 . 05 -. 10* -. 07 . 21** . 19*** -. 02 . 07 -. 01 m r' -I _. Intentions for Approving Behaviors Tellfriendsto buy company's products 2. 67 1. 83 1. 72 1. 93 Call or writecompanyto praise the action 'From 1-7 scale, where7 was verylikely. a codedas 0. as whitescodedas 1 andnonwhites 2Coded a dummy variable; 3Coded a dummy as womencodedas 1 andmencodedas 0. variable; *p < . 05. **p