Corporate social responsibility and consumers’attributions and
brand evaluations in a product–harm crisis
Jill Klein a,1,2,Niraj Dawar b ,*,2
a
INSEAD,Fontainebleau,France
b
Richard Ivey School of Business,University of Western Ontario,London,Ontario,Canada N6A 3K7Received 2April 2003;received in revid form 16August 2003;accepted 5December 2003
Abstract
A growing body of rearch on Corporate Social Responsibility (CSR)in marketing has shown that (1)CSR plays a role in consumers’brand and product evaluations,over and above economic or d rational T considerations such as product attributes;and (2)CSR has a spillover or d halo effect T on oth
erwi unrelated consumer judgments,such as the evaluation of new products.Yet CSR’s halo on consumer behavior may extend beyond product evaluations,into nonroutine types of judgments such as attributions.We examine the possibility that the CSR halo affects consumers’attributions in a product–harm crisis situation.In two studies that employ experimental manipulations of prior CSR on a sample of consumers,we examine whether attributions that are influenced by CSR mediate the impact of product–harm cris on consumers’brand evaluations.The results of Study 1support the hypothesis.Study 2introduces a boundary condition on the results of Study 1,showing that mediation effects are only found for consumers that are CSR-nsitive.The findings point to a role of CSR in consumer behavior that is more complex than previously conceptualized.D 2004Elvier B.V .All rights rerved.
Keywords:Corporate social responsibility;Product–harm cris;Attributions;Brand evaluationschief是什么意思
Corporate Social Responsibility (CSR)has emerged in recent years as both an important
academic construct and a pressing corporate agenda item (Colvin,2001;Harrison &Freeman,1999;Sen &Bhattacharya,2001;Waddock &Smith,2000).Firms have been found to engage in socially responsible behaviors not only to fulfill external obligations such as regulatory compliance
and stake-holder demands,but also due to enlightened-lf-interest considerations such as incread competitive-ness and improved stock market performance (Bansal &Roth,2000;Drumwright,1994,1996;Klasn &
0167-8116/$-e front matter D 2004Elvier B.V .All rights rerved.doi:10.1016/j.ijresmar.2003.12.003
*Corresponding author.Tel.:+15196614187;fax:+15196613959.
E-mail address:jill.klein@inad.edu (J.Klein)8ndawar@ivey.uwo.ca (N.Dawar).
1
Tel.:+33160712645.2
The authors contributed equally to this manuscript;order of authorship is arbitrary.
Intern.J.of Rearch in Marketing 21(2004)203
–217
Mclaughlin,1996;Russo&Fouts,1997;Waddock& Smith,2000).From a marketing perspective,the firm’s economic benefits from CSR have been documented in its link to consumers’positive product and brand evaluations,brand choice,and brand recommendations(Brown&Dacin,1997;Drum-wright,1994;Handelman&Arnold,1999;Osterhus, 1997;Sen&Bhattacharya,2001).Through a variety of theoretical lens,the important contributions of the studies have been to demonstrate that(1)CSR plays a role in routine consumer behavior,over and above economic or d rational T considerations such as product attributes;and(2)CSR has a spillover or d halo effect T on otherwi unrelated routine consumer judgments,such as the evaluation of new products.
Yet the CSR halo may also play a very different role in consumer behavior.The CSR halo may influence consumer judgments in nonroutine ttings that are rarely studied.In the ttings,CSR may operate for the firm as an insurance policy against the negative impact of untoward events such as product–harm cris.In the nonroutine ttings,the most common consumer cognitive process is attri-bution(Folkes,1984).Yet the impact of CSR on attributions remains unrearched.Attributions are important becau they form the basis of revision and updating of enduring and central consumer judgments such as brand evaluations(Folkes, 1988).The objective of the prent rearch is to uncover this potentially broader scope of CSR in consumer behavior,and to map its impact on c
onsumer attributions of blame in a product–harm crisis tting,and through them on to brand evaluations and purcha intentions.The general hypothesis tested is that CSR affects the attribution process itlf,and that attributions in turn influence brand evaluations.Thus,CSR associations are cast as a moderator of attribution.The attributions are conceptualized as a mediator of the impact of product–harm crisis on brand evaluations.The product–harm crisis tting is a particularly oppor-tune one for the study of brand evaluations,becau not only does it activate corporate associations such as CSR,but the sudden and often large changes in brand evaluation that accompany product–harm cris also make such a tting an invaluable opportunity for rearchers to understand influences on brand evaluations.1.Conceptual background
1.1.The corporate social responsibility halo
A halo effect is the d bias T due to a measure that spills over to another measure(Thorndike,1920).For example,a strong consumer belief about the perform-ance characteristics of a Porsche may spill over onto beliefs about its reliability;or a consumers’overall attitude toward a brand might spillover onto their asssment of specific attributes of that brand(Beck-with&Lehmann,1975).In this study,we are specifically interested in the halo effect of consumers’prior beliefs about the company’s position on CSR, onto attributions about a product–har
m crisis involv-ing that company.The halo effect is most conspicuous when the affected measures are unrelated to the source of the halo,becau when the affected measures are related,it is not possible to parate the halo effect from a nomological effect.Previous studies in marketing have shown how CSR affects unrelated product evaluations.In this study,we examine how CSR beliefs spillover onto attributions made about a product–harm crisis.
Brown and Dacin’s(1997)definition of CSR as the organization’s status and activities with respect to its perceived societal obligations,provides a uful starting point that has also been adopted in subquent rearch on CSR in ,Sen&Bhatta-charya,2001).As in the previous studies,we adhere to this definition but focus on consumers’perceptions of CSR as drivers of consumer behavior.This general definition of CSR allows for many different oper-ationalizations,but the corporate record on the environment is one of the most frequently ud manifestations of the construct(Bansal&Roth, 2000;Klasn&Mclaughlin,1996;Osterhus, 1997),and is one of the six key dimensions on which SOCRATES,the influential CSR index published by KLD Rearch,rates companies.Accordingly,our lens on CSR is that of consumers’perceptions of corporate environmental responsibility.
CSR associations are an element of overall d corporate associations,T distinct from attribute-level inf
ormation about products(Aaker,1996).Product associations,or stored information about product attributes,are generally ud as input for consumers to make everyday product-related judgments such as
J.Klein,N.Dawar/Intern.J.of Rearch in Marketing21(2004)203–217 204
evaluations of quality,comparisons with other prod-ucts,and asssments of value for money.In the interest of cognitive economy,consumers are likely to u information beyond product associations only sparingly,and mainly when existing information is insufficiently diagnostic to make the judgment at hand (Feldman&Lynch,1988).Even so,CSR associations have been shown to have a spillover effect on product and brand evaluations.In a nonroutine tting such as a product–harm crisis,corporate associations includ-ing CSR are all the more likely to be activated,and therefore,we expect the CSR halo to exert an influence on judgments such as attributions(Brown &Dacin,1997).
1.2.Product–harm cris and attributions
Product–harm cris have been defined as well-publicized instances of defective or dangerous prod-ucts(Dawar&Pillutla,2000).The increasing com-plexity of products,more demanding customers,and more vigilant media are making product–harm cris an ever more visible occurrence.Recent cris i
nvolving brands such as Coca Cola in Europe, Firestone in the United States,and Snow milk in Japan,have created consumer and media awareness and nsitivity to such cris.Product–harm cris can imperil long-standing favorable customer impressions about the brand.Firms often institute expensive voluntary product recalls to minimize damage to their brands from product–harm cris.Even so,much of the loss of market capitalization that is associated with product recalls is due to the damage to intangibles such as consumers’perceptions of the brand,rather than by the costs of the recall itlf(Davidson& Worrell,1992;Pruitt&Peterson,1986).The rious-ness and frequency of product–harm cris contrasts with the relative paucity of rearch in this area.
We know that consumers spontaneously construct attributions of blame for faulty or harmful products (Folkes,1984;Folkes&Kotsos,1986).The attributions are important from a marketing perspec-tive becau they form the basis of consumers’brand judgments and behavior.In constructing the attri-butions consumers rely on information,including corporate associations,that goes beyond the product attributes that are normally the basis of evaluation or purcha decisions(Aaker,1996;Folkes,Koletsky,&Graham,1987).The impact of the associations on attributions may depend on their relevance to con-sumers(Crocker,1980;Metalsky&Abramson, 1981);high importance placed on CSR issues may lead consumers to utilize such information in forming their attributions.
Weiner’s(1980)widely ud attribution model conceptualizes three causal dimensions of attribution that lead to an overall judgment of responsibility or blame:(1)the locus of the behavior(the event that triggers the crisis),which can be internal or external to the actor(in our ca,the firm);(2)the stability of the behavior,which can be unchanging or temporary;and (3)the controllability of the behavior,which can be within or outside the control of the actor.If the locus is internal,and the behavior is stable and controllable, obrvers(in our ca,consumers)tend to attribute responsibility to the actor,and subquent consumer behavior such as blame or anger,is directed toward the actor.If on the other hand,the locus is external, and the behavior is temporary and uncontrollable, attributions will tend to be made to external factors (Folkes,1984).The recent product–harm crisis involving Firestone tire blowouts that allegedly caud consumer deaths,and the subquent recall of millions of tires,helps illustrate this model of the attribution process.If consumers believe that the tires were poorly made,that Firestone has had a history of product defects,and that they could have averted the problems with better quality control,they will be likely to attribute responsibility to Firestone.In contrast,if they believe the problems were caud by harsh driving or vehicle conditions,that this is the first time Firestone’s tires have been implicated as the cau of accidents,and that driving and vehicle conditions are in fact outside the control of the tire manufacturer,they will be more likely to attribute responsibility to external factors,such as to the vehic
le manufacturer or to driving conditions.Our intention in studying attributions in a product–harm crisis tting is to establish whether CSR influences locus,stability,and controllability attributions.
way1.3.Attributions and consumer behaviorodl
Kelley and Michela(1980)in their broad review of attribution theory characterized the field as consisting of studies of the antecedents or determinants of
J.Klein,N.Dawar/Intern.J.of Rearch in Marketing21(2004)203–217205
attributions,such as information,beliefs and motiva-tion,and the conquences,such as affect,behavior, and expectancy.Marketing studies of attribution have tended to focus on conquences rather than informa-tional antecedents.Folkes(1984),in a pioneering study,demonstrated that consumer reactions to product failure are a function of the values on the three causal dimensions of locus,stability,and controllability.Variance in consumers’desire for a refund or exchange,their expectation of an apology, and even their desire for revenge were predicted on the basis of the values on the three causal dimensions. In a ries of subquent studies,Folkes et al.(1987) and Folkes and Kotsos(1986)showed how attribu-tions in the context of rvice delays led to the subquent desire to complain,and affected repurch-a behavior.Jorgenn(1994)applied
Weiner’s model to consumers’attributions in the context of rious company disasters(a fatal airliner crash),and found that consumers’attributions of the cau of the incident changed their affect and attitudes toward the company.Taken together,the studies provide compelling evidence for the effects of attributions on consumers’attitudes and behavior.But less is known about the determinants of attributions.
Studies of informational antecedents have found mixed ,Sparkman&Locander,1980; Yalch&Yoshida,1983).Sparkman and Locander (1980)reported that the context changed consumers’attributions about celebrity product endorment. However,despite manipulating a number of contex-tual cues,they found that only the extent to which other celebrities also endord the advertid product was a significant predictor of attributions.Yalch and Yoshida(1983)also found little evidence of con-textual antecedents on attributions.One explanation suggested for the mixed results is that the studies employed realistic stimuli to increa external validity, but that the stimuli introduced strong prior beliefs that swamped the effects of experimental manipula-tions(Folkes,1988).Additionally,the antecedents examined in the studies are in fact contextual cues that are cloly related to the behavior or actor,and the studies did not examine locus,stability,or controllability as contributing factors in the attribu-tion.In contrast,our interest lies in
understanding the effects of information such as CSR associations on attributions of stability,locus,and controllability.Furthermore,unlike the previous studies,we ek to systematically vary prior CSR beliefs to determine their impact on attributions on all three causal dimensions.Finally,to ensure that we accounted for any dampening effects of realistic stimuli,we ran two studies,one using a fictitious firm as the actor,and a cond a real firm.
In Weiner’s(1980)model,the three causal dimensions of attribution are not objective facts bad on unbiad data,but rather judgments formulated on the basis of information about the event or behavior available to the obrver.In the context of product–harm cris,such information may be gleaned from the media,from the company,and from other sources. Attributions are potentially prone to influence becau they are constructed through an interaction of event-related information with the obrver’s prior beliefs (Folkes,1988).Indeed,bias due to prior beliefs have been shown to influence consumer judgments in product harm cris.Dawar and Pillutla’s(2000)data showed that consumers’interpretation of a firm’s respon to a product–harm crisis was subject to their prior expectations about the firm.Expectations were defined in terms of consumers’accumulated experi-ence with the company and information about its past behavior toward customers.Their rearch focus on the interaction of firm respon with consumers’prior expectations.While it provides a bas
is for positing directional effects,it does not touch on the effects of specific corporate associations such as CSR,nor does it address the potentially important mediating role of attributions.
Our hypothes are drawn from an integration of the strands of rearch on the impact of CSR on consumer behavior,on consumers’attributions,and the impact of product–harm cris on brand evalua-tions.We propo that CSR beliefs will be activated in respon to a product–harm crisis,as part of the activation of corporate associations that occurs becau consumers engage in making attributions about the crisis.This activation enhances the like-lihood of the CSR halo having a spillover effect on attributional judgments.Specifically,we hypothesize that CSR beliefs moderate consumers’perception of the locus of the crisis event as internal or external, whether they e it as stable or temporary,and whether they believe it to have been controllable or not.CSR beliefs are a key element of the knowledge
J.Klein,N.Dawar/Intern.J.of Rearch in Marketing21(2004)203–217 206
of the company that an individual holds and draws upon to make the kinds of judgments(Brown& Dacin,1997).As with other types of halo effects, we expect that information related to the crisis will be interpreted in a confirmatory ,Darley &Gross,1983;Snyder&Swann,1978).In particular,
we propo that for firms that enjoy positive prior CSR,the trigger event for the product–harm crisis will be judged as more external, less stable and less controllable,relative to firms that do not enjoy positive CSR.Specifically,we hypothesize that:
H1a.The locus of the crisis will be perceived as external rather than internal when prior CSR is positive vs.when prior CSR is negative;
H1b.The crisis event will be perceived as unstable rather than stable when prior CSR is positive vs.when prior CSR is negative;and
H1c.The crisis event will be perceived as uncontrol-lable rather than controllable when prior CSR is positive vs.when prior CSR is negative.
In addition,as in previous studies of attribution,we anticipate that blame for the event will be influenced by attributions(Folkes&Kotsos,1986),but becau we view CSR as a moderator of attributions,we suggest that attributions will mediate the relationship between CSR and blame for the firm.
H2.Attributions will mediate the relationship between CSR and company blame.
Blame,in turn,will affect consumers’brand evaluations,over and above the direct effects of CSR on brand evaluation.That is,positive or negative CSR information is likely to directly affect brand evaluations,but blame is expected to also affect brand evaluations,even controlling for the direct effect of CSR.
Following the logical conquences,we know brand evaluations affect consumers’purcha inten-tions.In other words,we expect blame to be inverly correlated with brand evaluations,and brand evalua-tions to be positively related to purcha intentions. This hypothesis is intended as a validation of the knock-on effects of the attributions.The existence of the effects will indicate that the effects on attribu-tion do indeed lead to potentially enduring changes in brand evaluations,and that the in turn have an effect
desirableon behavioral intentions.Specifically:
H3a.Blame will be inverly related to brand
evaluations.
H3b.Brand evaluations will predict buying intentions.
2.Study1
2.1.Design and measures
The design involved three between-subjects con-
ditions(positive CSR,negative CSR,and a control
condition in which no information about CSR was
provided),and accordingly,three versions of the
questionnaire.In the prent study,a fictitious firm
name was ud.At the outt respondents were
provided with background information about the firm
that included the following description:
The following information is about a real,well-
known oil company that has been in business for
many decades.For the purpos of this study we
will call the company OilCo.
In the positive CSR condition,subjects read that
OilCo had been ranked1st among14major oil
companies on treatment of the environment,that the英文信格式
company was viewed as very environmentally respon-
sible,had placed at the top of similar environmental
rankings in the past,and had consistently shown that
it cared about the environment.Furthermore,OilCo’s
d recent work on th
e prervation o
f the Peruvian rain forest T was cited.In the negative condition,respond-
ents read that OilCo had been ranked last of14major
oil companies on treatment of the environment,that
the company was viewed as very environmentally
irresponsible,had been placed at the bottom of similar
environmental rankings in the past,and had consis-
mala
tently shown that it did not care about the environ-
ment.Furthermore,OilCo’s d recent harm caud to
the Peruvian rain forest T was cited.In the control
condition,no information was given about OilCo’s
environmental record.
After reading the introduction,respondents(in the
treatment conditions)rated OilCo on environmental
social responsibility,as a manipulation check.They
J.Klein,N.Dawar/Intern.J.of Rearch in Marketing21(2004)203–217207
also indicated how important it is that a company is environmentally responsible.Respondents then read a newspaper article in which an OilCo product failure was described:
OilCo Lubricant Tied to Engine Problems [Dateline]There have been veral hundred reports of vere engine damage linked to OilCo’s ET Synthetic Blend Lubricant,an automotive engine lubricant designed to protect engines and improve performance.
英语面试问题及回答
impressingApparently,ingredients in the oil interact with the plastic jugs in which the lubricant is sold.Over time,the chemicals in the plastic thicken the oil, which caus engine damage.The company that makes the plastic jugs is Carson Plastics.
An OilCo spokesperson said yesterday,b There is no problem with the lubricant if it is sold by stores before the expiration date marked on the bottle.All of the engine problems encountered can be tied to a national grocery store chain that sold the lubricant after the expiration date Q.
After reading the article,respondents completed the questionnaire that included questions concerning the locus,stability and controllability of the prob-lem,company blame,brand evaluations,buying intention,and veral demographic questions.Ques-tions were asked in this order(with the exception that the brand evaluation and buying questions were mixed together).The specific items were consistent with previous rearch on attributions and brands (Agarwal&Rao,1996;Dawar&Pillutla,2000; Weiner,1980).
2.2.Respondents and procedure
One hundred and fifty respondents participated in a mall–intercept study.The mall was located in the suburbs of a large city in a Midwestern state and attracted predominantly middle-class shoppers.The average age of the sample was36.7(ranging from17 to74)and64%were female.Ninety-eight percent had at least a high school diploma,with44%of the sample holding a college degree.Questionnaires were lf-administered and respondents were recruited and randomly assigned to condition by profe
ssional interviewers working for a market rearch firm specializing in mall–intercept interviews.On average the study took15min to complete.On completion, respondents were provided monetary compensation ($2)for their time,plus a bonus($1)if they answered correctly a final quiz that tested their memory of the news story.Respondents were told about this bonus award at the start of the study to insure that the respondents read the stimulus materials carefully. Finally,participants were given a debriefing that explained the purpo of the study and uncovered the d cover story T ud in the manipulation.
3.Results
3.1.Manipulation checks
The manipulations worked as intended.Subjects in the negative condition rated OilCo’s actions toward the environment as more harmful(m=2.2;on a1, harmful to7,helpful scale),wor than other companies(m=2.1),and less socially responsible (m=2.1)than tho in the positive condition(m=4.9, 5.3,5.1,respectively;all p’s b0.001).As expected, there were no differences between the two groups in terms of perceived importance of company environ-mental responsibility,or attitudes toward buying from a company that was not environmentally responsible (both t’s b1.2,n.s.).
3.2.Measures
Two ts of questions tapped the d locus T dimen-sion.First,subjects were asked to rate the likelihood of various parties being a source of the engine lubricant problem(locus)on a1–7-point scale(from not at all likely to very likely).They rated OilCo, retailers,consumers,and the supplier of the plastic container as potential sources of the problem.In a cond t of locus questions,respondents were asked to assign a d percentage of the problem T that might be due to each of the four parties(with totals summing to 100%).Locus was measured using the proportion of likely source judgments given to OilCo relative to all of the parties,and the percent given for the role played by OilCo relative to all of the parties.Both measures were proportions(r=0.48,p b0.001),and an average index was created from the two.
蒙蒙亮
我的大学生活J.Klein,N.Dawar/Intern.J.of Rearch in Marketing21(2004)203–217 208
