ABSTRACT - Following a negative company-related event, such as an airline crash or a food poisoning incident, consumers are likely to be exposed to one or more causal explanations for the event. This paper investigates the extent to which single versus multiple explanations affect consumers' acceptance of the explanation and consumers' anger and blame toward the company. Although the experiment reported here does not demonstrate an effect of single versus multiple explanations on the likelihood or acceptability of an explanation, reactions to single and multiple explanations are shown to operate as predicted by attributional theory.||On December 21, 1988, Pan Am flight 103 crashed into the town of Lockerbie, Scotland, killing all 258 passengers and crew members aboard, as well as a number of people on the ground. All who heard about the disaster wondered how it could have occurred. Was it weather conditions, terrorism, poor security? Who was to blame? Would or could it happen again? As company, government, and media representatives searched for answers, the public tried to make sense of the wide array of causal speculations, explanations, and excuses that filtered through to it.

Citation:

Brian K. Jorgensen (1994) ,"", in NA - Advances in Consumer Research Volume 21, eds. Chris T. Allen and Deborah Roedder John, Provo, UT : Association for Consumer Research, Pages: 348-352.

Advances in Consumer Research Volume 21, 1994      Pages 348-352

CONSUMER REACTION TO COMPANY-RELATED DISASTERS: THE EFFECT OF MULTIPLE VERSUS SINGLE EXPLANATIONS

Brian K. Jorgensen, University of California, Los Angeles

ABSTRACT -

Following a negative company-related event, such as an airline crash or a food poisoning incident, consumers are likely to be exposed to one or more causal explanations for the event. This paper investigates the extent to which single versus multiple explanations affect consumers' acceptance of the explanation and consumers' anger and blame toward the company. Although the experiment reported here does not demonstrate an effect of single versus multiple explanations on the likelihood or acceptability of an explanation, reactions to single and multiple explanations are shown to operate as predicted by attributional theory.

INTRODUCTION

On December 21, 1988, Pan Am flight 103 crashed into the town of Lockerbie, Scotland, killing all 258 passengers and crew members aboard, as well as a number of people on the ground. All who heard about the disaster wondered how it could have occurred. Was it weather conditions, terrorism, poor security? Who was to blame? Would or could it happen again? As company, government, and media representatives searched for answers, the public tried to make sense of the wide array of causal speculations, explanations, and excuses that filtered through to it.

When large-scale negative company-related incidents occur they can be very important because they have the potential of directly and indirectly affecting large numbers of people. Recent incidents involving bad meat from Jack in the Box Restaurant and syringes in Pepsi cans attest to the high level of publicity that these kinds of incidents can generate. From a marketing perspective, company-related disasters and crises can be particularly damaging when a company's products bear the company name. Brand equity and customer loyalty may drop rapidly if a company is blamed for a serious negative incident.

Consumer behavior scholars have studied consumer attributions following relatively small-scale company-related annoyances, such as defective merchandise (Folkes 1984) and delayed airline flights (Folkes, Koletsky, and Graham 1987). However, consumer reactions to major company-related disasters, such as jet crashes, oil spills, or drug-related scares, have been largely neglected. Further, the tendency of attribution theorists to focus their study on single, as opposed to multiple, causal explanations (Leddo, Abelson, and Gross 1984) seems to have carried over into the consumer behavior literature.

This paper investigates consumer reaction to multiple versus single explanations for image-threatening company-related disasters within a framework of attributional theory and conjunctive explanations. Following a review of the literature, the results of a preliminary experiment in this area are presented. Opportunities for future research are then addressed.

ATTRIBUTIONAL THEORY AND COMPANY-RELATED DISASTERS

Because company-related disasters are highly negative and unexpected, consumers are likely to try to understand the cause of these types of incidents (Bucher 1957; Veltfort and Lee 1943; Weiner 1986). However, members of the general public usually do not have first hand information regarding company-related incidents from which to develop their own causal attributions. Rather, in these situations third party sources, such as the media or company representatives, must generally supply possible or probable causes. The important question then becomes how consumers react to the explanation or explanations that have been offered. Of the various available approaches to the study of attributions, Weiner's (1986) attributional theory is particularly well-suited to the study of consumer reactions to company-related events because Weiner's theory focuses on the consequences of attributions, rather than on the process by which the attributions are made.

According to Weiner's theory, once a person has made a causal attribution, he is expected to experience particular affects or be motivated to perform particular behaviors based on where he determines that the cause falls along three distinct dimensional continuua. These three dimensions are locus, controllability, and stability. The locus dimension addresses the degree to which the cause is internal to or external to the target of the attribution. Thus, for example, an airline accident that is attributed to an improper instruction by an air traffic controller may be considered external to the company, while an accident due to pilot error would probably be considered internal. Controllability concerns the extent to which a cause is within the control of the target entity. Therefore, using the previous example, the air traffic controller problem would probably be considered uncontrollable by the company, while the pilot error problem might be seen as more controllable. Stability reflects the degree to which a cause is something unchanging as opposed to fluctuating or changing over time. Given the infrequency with which they occur, most company-related disasters should tend to be caused by relatively unstable causes, and, therefore, stability is not further considered here.

Weiner's causal dimensions are theorized as influencing various emotional reactions and, ultimately, behavior (Weiner 1986). With particular reference to the company disaster situation, the controllability and locus dimensions, which are somewhat overlapping in most cases, are expected to influence the emotions of anger and pity (Folkes 1984). Behavior is thought to be indirectly affected through the attributionally-induced affective states and resulting attitudes.

EXPLANATIONS OF COMPANY-RELATED DISASTERS

Impression Management

The way in which consumers react to a company-related disaster situation may depend not only on the attributional circumstances underlying the incident but also on the way in which management responds. "Impression management" describes the process by which people, or in this case companies, control others' impressions of them (Leary and Kowalski 1990; Russ 1991). Impression management can be particularly important in response to serious negative events (Schlenker 1980).

One of many approaches that management might take to manage impressions following a company-related disaster is referred to in the crisis management literature as "telling one's own story" (Meyers and Holusha 1986). In other words, management is advised to give its own explanation for what caused the incident. Often, more than one factor may be potentially responsible for a particular incident. For instance, the deaths linked to Jack in the Box hamburgers were eventually traced to both bad meat (an external cause) and improper cooking temperatures (an internal cause). In other cases, until the actual cause or causes are isolated, a number of potential explanations may be available, some of which may be better for the company than others.

The Conjunctive Fallacy and Conjunctive Explanations

Little, if any, of the research stemming from Weiner's (1986) attributional theory has addressed the consequences of multiple attributions. However, in the judgment and inference literature, Tversky and Kahneman (1983) and others have investigated the different effects of multiple or "conjunctive" versus single statements and explanations on people's judgments and predictions. These investigations have centered around what has become known as the "conjunctive fallacy."

The Conjunctive Fallacy. The conjunctive fallacy, or conjunctive error, describes people's tendency to estimate the joint probability of "A and B" to be greater than the probability of "A" or "B" individually, where A and B are descriptions of a person or thing or explanations for an event or action (Tversky and Kahneman 1983). Thus, for example, when asked to judge the likelihood of possible causes of an airline accident, people might consider the conjunction of "sunspot activity and pilot error," to be more likely than "sunspot activity." This tendency is regarded as a fallacy or error because it is statistically impossible for a conjunction of two items to be more likely than either of the items that make up the conjunction.

The robustness of findings of conjunctive errors in various contexts has led to a number of investigations into why people make these types of errors (e.g., Leddo, Abelson, and Gross 1984; Locksley and Stangor 1983; Tversky and Kahneman 1983; Wells 1985; Zuckerman, Eghrari, and Lambrecht 1986). Initially, the conjunctive fallacy was linked to the "representativeness" of the component items (Tversky and Kahneman 1983, Wells 1985). Thus, for example, if pilot error is considered to be a more representative cause of airline accidents than is sunspot activity, then the combination of the two explanations may appear more representative, and thus more probable, than the sunspot cause. Wells (1985) found some support for a representativeness explanation by showing that the combination of representative and unrepresentative statements produced strong conjunctive error effects, while the combination of two unrepresentative or two representative statements produced lower error rates.

Another possible explanation for the conjunctive fallacy is that subjects are misinterpreting the single statements as though they were meant to exclude the conjunctive statements. So, for example, subjects might be interpreting the "pilot error" explanation for an airline accident as "pilot error without sunspot activity." Leddo, Abelson, and Gross (1984) have discounted this explanation, noting that the errors have been found in a between-subjects study, where subjects saw only the single or the conjunctive explanation (Pennington 1984), and that in other studies the statement "A" was replaced with "A, whether or not B," without a weakening of conjunction effects (Locksley and Stangor 1984; Tversky and Kahneman 1983). On the other hand, Morier and Borgida (1984) have found that certain task features in conjunctive problems, such as ranking versus rating the probability of simple and conjunctive statements, can reduce, but not eliminate, conjunctive errors. They have also found that some conjunction problems can be debiased by clarifying wording.

The conjunctive fallacy has also been attributed to improper combination procedures on the part of subjects. Abelson, Leddo, and Gross (1987), have shown that the likelihood of a conjunction can often be approximated by a geometric mean of each of the conjunction's components. Also, Yates and Carlson (1986) have shown that a procedure for summing the likelihoods of conjunction components can predict incidence of errors in some instances.

Conjunctive Explanations. In comparing conjunctive effects across a wide array of studies, Abelson, Leddo, and Gross (1987) find much stronger conjunctive effects in explanation tasks than in other tasks. Further, Locksley and Stangor (1984) find that rare events or outcomes are much more likely to bring about conjunctive errors than common events or outcomes. Presumably, more common events can be more easily explained by single causes.

Since company-related disasters are generally accompanied by explanations, and since these are relatively uncommon events, the conjunctive explanations literature seems highly relevant to the company disaster situation. However, this connection should still be made with some caution. First, the array of possible causes for company-related disasters is much more complex than for the outcomes described in the conjunctive explanation studies. For example, some causes may be insufficient in and of themselves to cause a disaster but may be able to contribute to a disaster in conjunction with another cause. This possibility is not addressed in the conjunctive explanations literature.

Further, the conjunctive explanations literature addresses the effects of conjunctive versus single explanations on judgments of the likelihoods of the explanations. Effects of multiple versus single explanations on affective responses, attitudes, and behaviors are not addressed. Although a more likely explanation may be presumed to be better explanation, and may, therefore, be expected to lead to stronger reactions, this line of reasoning is not yet supported in the literature.

STUDY AND HYPOTHESES

As a preliminary study of single versus multiple explanations, an experiment was designed to study consumer reactions to a company-related disaster, where the disaster is attributed to either an internal/controllable factor, an external/uncontrollable factor, or both an internal/controllable factor and an external/uncontrollable factor (hereafter referred to as "mixed" or "mixed/ambiguous"). Given the robustness of the conjunctive explanation findings, the conjunctive error effect was expected to be observed:

H1: A multiple cause for a negative company-related incident that includes both internal/controllable and external/uncontrollable factors will be judged as more likely than at least one of the single factor causes making up the multiple cause.

Weiner's (1986) attributional theory proposes that when a negative outcome is controllable by the attributional target, greater blame and anger and less sympathy should be expressed toward the target than if the outcome is uncontrollable. Anger and sympathy should, in turn, influence general attitudes toward the attributional target, such that consumers should express more negative attitudes and purchase intentions toward companies connected with negative events where the cause appears to be internal/controllable, as opposed to external/uncontrollable. Therefore:

H2: Consumers will express greater anger, less sympathy, poorer attitudes and purchase intentions, and higher levels of blame toward a company involved in a negative incident to the extent that the cause of the incident is perceived as more internal to and more controllable by company management.

TABLE

Although mixed/ambiguous causal attributions, have begun to be used in attributional studies (e.g., Weiner, Graham, Peter, and Zmuidinas 1991), systematic theoretical or empirical analyses of their properties and effects have not yet been carried out. However, Weiner, et al. (1991) suggest that affective and behavioral reactions to mixed attributions should fall within the range between the reactions to each single cause that makes up the mixed cause. How closely the reactions to the mixed attribution mirror those of one or the other of the single attributions may depend to a large degree on the circumstances of the situation. In the present study, the mixed attribution presents a situation where the company that is connected with the negative event is the only volitional entity that can be blamed for the incident. Since blame is generally accorded only to volitional entities (Anderson 1991), the level of blame, as well as the levels of other affective and behavioral responses, should, in this case, be more similar between mixed and internal/controllable attributions than between mixed and external/uncontrollable attributions.

H3: Where the external cause of a company-related disaster is not controlled by a person or entity, the reported levels of affects, attitudes,and blame for a mixed set of explanations should be more comparable to levels for the internal/controllable attribution than to levels for the external/uncontrollable attribution.

Methodology

An experiment was conducted with 36 subjects consisting of 18 adult undergraduate students and 18 adult members of a church group. The subjects were run in two groups on the same day. A vignette approach was chosen because this approach has been widely used in attribution studies of this type and also because company disaster information generally reaches the consumer in the form of a story, such as a news story or a conversation. Following the description of a fatal airliner crash, one of three alternatives was suggested as the possible cause of the crash: (1) bad weather (external/uncontrollable), (2) poor aircraft maintenance (internal/controllable), or (3) bad weather and poor maintenance (ambiguous/mixed). The study design was completely between-subject, with 12 subjects viewing each distinct vignette.

On the experimental cover sheet, subjects were instructed to read the company-related vignette and imagine that they were reading a current news story. They were further instructed not to turn back to the story after they had finished reading it. On the page following the cover sheet, subjects were presented with the vignette. The vignette was followed by a questionnaire, which included a number of dependent measures and manipulation checks. The first question measured subjects' overall reaction to the airline referenced in the vignette on four seven-point semantic differential scales anchored by very unfavorable-very favorable, bad-good, negative-positive, dislike very much-like very much. These scales were averaged for the attitude measure. Next, three questions measured, on seven-point scales, how much anger and sympathy subjects felt toward the airline, if any, and the extent to which subjects felt that the airline was to blame, if at all. A measure of purchase intention was then presented, in which subjects were asked to rate the likelihood that they would choose to fly with this particular airline as opposed to other comparable airlines. This measure was taken on a seven-point scale anchored by very unlikely-very likely.

Finally, subjects responded to manipulation checks designed to test whether the causal dimension manipulations had successfully presented causes that were perceived as either internal/controllable or external/uncontrollable. A six item set of scales was modified from Russell (1982). Two each of the six seven-point items measured the extent to which the subject found the cause to be controllable/uncontrollable, internal/external, and stable/unstable, respectively. Each pair of scales was averaged to arrive at a single scale score for each dimension. The stability measures were taken to assure that the manipulations did not differentially affect stability.

RESULTS AND DISCUSSION

The data were analyzed by one-way ANOVA for each of the manipulation checks and dependent variables. Cell means for each dependent variable are presented in the Table. The results of specific statistical analyses are presented in the Table.

Manipulation Checks

As expected, the causal attribution given had a significant effect on controllability (F(2,33)=12.25, p<.0001) and locus (F(2,33)=14.76, p<.0001). Paired tests using the .05 level indicated that for both of these manipulation check variables the external/uncontrollable attribution differed from the internal/controllable and mixed/ambiguous attributions, which did not differ from one another. However, in directional terms, the mixed attribution was situated substantially more midway between internal and external with regard to locus than with regard to controllability. Also as expected, the judged stability of a cause was not significantly affected by causal attribution.

Hypothesis 1

Contrary to expectations, whether the airliner crash seemed due to an internal/controllable, an external/uncontrollable, or a mixed cause had no effect on the judged likelihood that the cause given was indeed the actual cause. Thus, a conjunctive error was not demonstrated in this case. A number of possible reasons for this result can be suggested.

First, unlike nearly all conjunctive explanation and conjunctive error studies, this study was a completely between-subjects study. Thus, while subjects in traditional conjunctive error studies are exposed to both the individual causes and the conjunctive cause, the subjects in this study saw only one or the other. Those studies showing that conjunctive errors are attributable to the way tests are constructed and worded (e.g., Morier and Borgida 1984) support the idea that through debiasing, which can be effected through a between-subjects study, the conjunctive error may be greatly reduced or eliminated.

The results here may also be partly attributable to the fact that the causes making up the conjunctive explanation were judged to be fairly equally likely. Conjunctive errors have usually been found to be more pronounced in situations where one component cause is considered much more likely than the other. Further, studies that trace conjunctive errors to statistically incorrect averaging processes (e.g., Abelson, Leddo, and Gross 1987; Yates and Carlson 1986) also suggest that the errors should be more likely in cases of unequal component causes.

The failure to observe conjunctive error effects may also stem from the complexity of company crisis situations. Consumers may feel ill-equipped to judge for themselves and may, instead, defer to the trustworthiness of the source of the information to determine the likelihood that what is reported is true. If this is the case, future studies should show that both common and unusual causes reported by the same source should be considered equally likely candidates for the true cause.

Hypotheses 2 and 3

Affective measures and blame. According to Weiner's (1986) attributional theory, a negative event that is due to controllable causes should lead to greater anger and blame and to less sympathy than an event that is due to uncontrollable causes. The effect of attribution on anger was significant (F(2,33)=3.62, p<.05). However, pairwise tests at the .05 level found a difference only between the external/uncontrollable attribution and the mixed attribution. Still, directional results supported both hypotheses 2 and 3 in that both internal/controllable and mixed attributions led to greater anger than did the external/uncontrollable attribution. Although not significant, the results for sympathy were also in the predicted direction.

As expected, the effect of attribution on blame was significant (F(2,33)=10.52, p<.001). In addition, pairwise tests at the .05 level supported hypothesis 3 in that the higher levels of blame for internal/controllable and mixed attributions were significantly different from the level of blame for external/uncontrollable attributions, although not significantly different from one another.

Attitude and purchase intentions. The effect of attribution on attitude approached significance (F(2,33)=2.73, p<.10), with a directional indication that attitudes were poorer toward companies connected with negative incidents when the causal attribution was internal/controllable or mixed than when the attribution was external/uncontrollable. The effect of attribution on purchase intention was significant in the expected direction (F(2,33)=6.23, p<.01). Pairwise tests indicated that external/uncontrollable attributions led to higher purchase intentions than did internal/controllable or mixed attributions.

With regard to purchase intentions, the scenario in the vignettes used here may have resulted in stronger results than would a scenario in which the company-related disaster does not have an effect on the product itself. Whereas an oil spill or chemical leak will not affect the product that is sold at the gas pump or the drug store, an airline crash may signal that the airline is offering an inferior product.

Summary and General Discussion

The results of this study do not lend support to the hypothesis that conjunctive errors play a role in consumer reaction to company disasters, although the study does not rule out the possibility that in other disaster settings conjunctive errors could play a role. The study does, however, support the expectation that consumers' affects, attitudes, and, perhaps, behaviors following a negative company-related incident may be influenced by their understanding of the cause or causes underlying the incident. Also, under the particular circumstances of the vignette presented in this study, consumers' reactions to a mixture of internal/controllable and external/uncontrollable causes were similar to their reactions to a sole internal/uncontrollable cause. However, this result may not necessarily generalize to every case. If, for example, the external cause had been an attribution to a volitional entity, such as a terrorist, the result may have been different, since more blame may have been focussed externally. Also, a greater difference between the likelihoods of the two component explanations may have affected the relative likelihood of the conjunctive explanation, and perhaps other variables as well.

FUTURE DIRECTIONS AND CONCLUSIONS

Consumer reaction to various explanations for company-related disasters is an important area of research that has received little attention. This preliminary research into the area of conjunctive explanations suggests that differences in a company's approach to a disaster situation can have important consequences. The number and types of explanations for a particular incident that reach the consumer may influence emotions, attitudes, and behaviors toward the company.

The research presented here suggests the need for further investigation into a number of different questions regarding single versus multiple explanations for company-related disaster situations. One important research direction would entail an examination of mixed attribution situations where the external cause is controllable by a person or group outside of the company. Another interesting question concerns disaster situations where there is no effect on the company's product, such as in the case of the Exxon Valdez oil spill or the Union Carbide Bhopal incident. A third direction would be the examination of incidents where multiple explanations consist of two internal/controllable or two external/uncontrollable explanations, rather than a mixture of one of each type.

REFERENCES

Abelson, Robert P., John Leddo, and Paget H. Gross (1987), "The Strength of Conjunctive Explanations," Personality and Social Psychology Bulletin, 13 (June), 141-155.

Anderson, Norman H. (1991), "Psychodynamics of Everyday Life: Blaming and Avoiding Blame," in Contributions to Information Integration Theory, Volume II: Social, ed. Norman H. Anderson, Hillsdale, NJ: Erlbaum, 243-275.

Bucher, Rue (1957), "Blame and Hostility in Disaster," American Journal of Sociology, 62 (March) 467-475.

Folkes, Valerie S. (1984), "Consumer Reactions to Product Failure: An Attributional Approach," Journal of Consumer Research, 10 (March), 398-409.

Folkes, Valerie S., Susan Koletsky, and John L. Graham (1987), "A Field Study of Causal Inferences and Consumer Reaction: The View from the Airport," Journal of Consumer Research, 13 (March), 534-539.

Leary, Mark R. and Robin M. Kowalski (1990), "Impression Management: A Literature Review and Two-Component Model," Psychological Bulletin, 107 (1) 34-47.

Leddo, John, Robert P. Abelson, and Paget H. Gross (1984), "Conjunctive Explanations: When Two Reasons are Better than One," Journal of Personality and Social Psychology, 47 (November), 933-947.

Locksley, Anne and Charles Stangor (1984), "Why versus How Often: Causal Reasoning and the Incidence of Judgmental Bias," Journal of Experimental Social Psychology, 20, 470-483.

Meyers, Gerald C. and John Holusha (1986), When It Hits the Fan: Managing the Nine Crises of Business, New York: Mentor.

Morier, Dean M. and Eugene Borgida (1984), "The Conjunction Fallacy: A Task Specific Phenomenon," Personality and Social Psychology Bulletin, 10 (June), 243-252.

Pennington, N. (1984), Technical Note on Conjunctive Explanations, Unpublished manuscript, Center for Decision Research, University of Chicago.

Russell, Dan (1982), "The Causal Dimension Scale: A Measure of How Individuals Perceive Causes," Journal of Personality and Social Psychology, 42 (June), 1137-1145.

Schlenker, Barry R. (1980), Impression Management, Belmont, CA: Wadsworth, Inc.

Tversky, Amos and Daniel Kahneman (1983), "Extensional Versus Intuitive Reasoning: The Conjunction Fallacy in Probability Judgment," Psychological Review, 90 (October), 293-315.

Veltfort, Helene Rank and George E. Lee (1943), "The Coconut Grove Fire: A Study in Scapegoating," Journal of Abnormal and Social Psychology, 38 (April) 138-154. (Clinical Supplement)

Weiner, Bernard (1986), An Attributional Theory of Motivation and Emotion, New York: Springer-Verlag.

Weiner, Bernard, Sandra Graham, Orli Peter, and Mary Zmuidinas (1991), "Public Confession and Forgiveness," Journal of Personality, 59 (2), 281-312.

Wells, Gary L. (1985), "The Conjunction Error and the Representativeness Heuristic," Social Cognition, 3 (Fall), 266-279.

Yates, J. Frank and Bruce W. Carlson (1986), "Conjunction Errors: Evidence for Multiple Judgment Procedures, Including 'Signed Summation,'" Organizational Behavior and Human Decision Processes, 37 (April), 230-253.

Zuckerman, Miron, Haleh Eghrari, and Mark R. Lambrecht (1986), "Attributions as Inferences and Explanations: Conjunction Effects," Journal of Personality and Social Psychology, 51 (6), 1144-1153.

----------------------------------------