Consumer Assessments of Responsibility For Product-Related Injuries: the Impact of Regulations, Warnings, and Promotional Policies

Mitch Griffin, Bradley University
Barry J. Babin, University of Southern Mississippi
William R. Darden, Louisiana State University
ABSTRACT - This paper examines the impact of regulations, warning labels and promotional policies on consumers' attributional processes. The results of an experiment manipulating these characteristics within a product-injury scenario show that the manner in which these managerial factors are treated can affect assignments of responsibility. In addition, the degree to which an injury is unanticipated is also shown to affect consumer attributions.
[ to cite ]:
Mitch Griffin, Barry J. Babin, and William R. Darden (1992) ,"Consumer Assessments of Responsibility For Product-Related Injuries: the Impact of Regulations, Warnings, and Promotional Policies", in NA - Advances in Consumer Research Volume 19, eds. John F. Sherry, Jr. and Brian Sternthal, Provo, UT : Association for Consumer Research, Pages: 870-878.

Advances in Consumer Research Volume 19, 1992      Pages 870-878

CONSUMER ASSESSMENTS OF RESPONSIBILITY FOR PRODUCT-RELATED INJURIES: THE IMPACT OF REGULATIONS, WARNINGS, AND PROMOTIONAL POLICIES

Mitch Griffin, Bradley University

Barry J. Babin, University of Southern Mississippi

William R. Darden, Louisiana State University

ABSTRACT -

This paper examines the impact of regulations, warning labels and promotional policies on consumers' attributional processes. The results of an experiment manipulating these characteristics within a product-injury scenario show that the manner in which these managerial factors are treated can affect assignments of responsibility. In addition, the degree to which an injury is unanticipated is also shown to affect consumer attributions.

INTRODUCTION

A wide range of marketing decisions, including pricing policies, distribution systems, and advertising claims are governed by public policy regulations (Werner 1982). The area of consumer safety, however, may be the most heavily regulated aspect of the marketing discipline. Beginning with the Pure Food and Drug Act (1906), a steady stream of government agencies and legislation has been enacted in an effort to protect consumers from unsafe products. The marketing academic community has responded by investigating attitudes and opinions of various parties involved with public policy and by keeping abreast of developments in product safety legislation. As examples, Busch (1976) surveyed bicycle manufacturers to determine their evaluation of Consumer Product Safety Commission (CPSC) safety regulations. Later, Busch and Hair (1980) compared the attitudes of three different parties, manufacturing executives, insurance executives, and state insurance commissioners, regarding a variety of product safety issues. Taking a different approach, Morgan (1982; 1986; 1988; Morgan and Avrunin 1982) has extensively reviewed product liability legislation and court cases, discussing policy implications as they relate to consumers and policy makers. Each of these studies has provided valuable insight into a diversity of public policy issues.

The present study compliments previous research on product safety issues by offering perceptions and reactions from a consumer's perspective. More specifically, the objective of this research is to assess the effect of three dimensions of product safety policy on consumer attitudes and behavior: (1) product safety regulations; (2) warning labels; and (3) advertisements stressing product safety. To accomplish this goal, this study briefly reviews the present product liability situation, presents a theory-based model of consumer assessment of responsibility for a product-related injury, develops the intervening role of "unanticipated consequences" of product usage on consumer evaluations of these mishaps, and empirically tests relationships implied by the model utilizing experimental scenarios.

BACKGROUND

Unquestionably, the product safety arena is of great importance to consumers and decision makers. To illustrate this point, consider the plight of one industry, general aviation, hit particularly hard by escalating costs. The production of light aircraft has been sharply curtailed, from 17,811 units in 1978 to 2,438 in 1984 (North 1985) and continues to fall. This rapid decrease has been directly attributed to the reported $60,000 to $100,000 per unit liability insurance expense (Gatty 1987).

Likewise, many other industries, including pharmaceutical and cosmetic producers and distributors (Friend 1990), health care providers (consider the plight of OB/GYN physicians), and even volunteer coaches of children's sports teams (Mihoces 1990) have been hard hit by the potential of liability. For consumers, the liability crisis has greatly increased prices (Private Pilot 1990), limited the selection of products available in the market (Olsen 1989), and stifled technological innovation (Martin 1989). With few exceptions, those marketing consumer goods must be better aware of the costs and consequences of product liability, as well as consumer perceptions of product safety issues, to better serve their customers and secure a competitive position.

Three actions which may be taken by marketers which may affect consumer attitudes toward product safety concern (1) the application of available safety standards, (2) the use of appropriate warning labels on products, and (3) an increased emphasis of product safety in their promotional campaigns. While these three factors do not exhaust the possible actions which might impact consumer perceptions of product-related injuries, they are considered here for three main reasons. First, each is a variable which is highly controllable by policy makers and/or management. Second, under the legal doctrine of strict liability, these actions should not influence assignment of responsibility for a product-related injury. Third, these three factors were revealed as highly salient in qualitative research designed to explore consumers' product safety attitudes. The question of just how these factors might influence consumers' causal attributions of responsibility toward manufacturers, consumers, and situations for product-related injuries is explored by experimentally manipulating these three factors.

Theoretical Development

As consumers, individuals are frequently exposed to media sources reporting stories of people injured by products they themselves use. Further, when consumers become jurors, they are forced to ponder causes of injuries and determine precisely when and how much compensation should be awarded. In both roles, we feel that an individual pursues his/her natural motivation "to attain a cognitive mastery of the causal structure of his environment" (Kelley 1967, p. 109). The model shown in the Figure depicts a number of relationships proposed to represent the manner in which consumers ascribe responsibility for a product-related injury.

FIGURE 1

HYPOTHESIZED RELATIONS BETWEEN PRODUCT CHARACTERIZATION AND ATTRIBUTIONS OF BLAME

A central component of the model is a construct termed unanticipated consequences (UC) which intervenes between two of the manipulations and assignment of responsibility. Clearly, no consumer anticipates being injured or killed by a product they purchase. Other than those prone to suicide, no consumer would make such a purchase. Further, not only are product-related injuries unexpected events, but, they are unexpected negative events. In a recent study of unexpected favorable events (Howard and Barry 1990), the extent to which an event is unanticipated was shown to alter consumer product evaluations due to heightened affect. Considering the negative nature of product injuries, it could be argued that any effects due to the unexpectedness of an event would be even greater than for positive events (Tversky and Kahneman 1981) and be even more likely to evoke "spontaneous attributional activity" (Weiner 1985). Thus, the extent to which a product-related injury is unanticipated (UC) is presumed to alter consumers' attributions of responsibility for a mishap.

The unanticipated consequences (UC) variable shares similarities to the disconfirmation paradigm of product satisfaction (Olshavsky and Miller 1972; Tse and Wilton 1988). Using this paradigm, Churchill and Suprenant (1982, p. 493) define satisfaction as "an outcome of purchase and use resulting from the buyer's comparison of the rewards and costs of the purchase in relation to the anticipated consequences." In this study however, the consequences are largely unanticipated; that is beyond the typical consumer's range of considered outcomes. Rather than merely experiencing a product performance worse than expected, UC may produce extreme levels of negative satisfaction and very strong levels of attribution of blame toward a product's producer. Consider a brief example as illustration. When purchasing a new lawn mower, a consumer has a set of expectations regarding the mower's ease of starting, durability, cutting ability, maintenance requirements and so forth. Negative disconfirmation would occur if the mower's performance on these attributes failed to meet prior expectations, or even if the mower failed to operate at all. However, if a largely unanticipated negative outcome (such as a product-related injury) were to occur, a traditional disconfirmation model may be insufficient to capture the consumer's reaction. More specific support for each of the relationships implied by the model follows.

Research Hypotheses

Indirect Effects. As can be seen in the figure, UC intervenes between two product characteristics and various attributions of blame. A number of hypotheses result. The hypothesized relationship between the experimental manipulations and UC is based on the above discussion as well as the concept of search, experience, and credence properties (Darby and Karni 1973; Nelson 1974). To make the consumer aware of the danger, the manipulations must represent search properties. In other words, we propose only those factors which a consumer is made aware of prior to actually using a product would serve to affect the level of UC (Darby and Karni 1973; Nelson 1974). Therefore, how a product compares to governmental regulations, which is very difficult to determine prior to purchase, would have no affect on UC. On the other hand, proper warnings attached to the product should make the consumer aware of the dangers involved. Conversely, advertisements that stress the safe nature of the product (e.g., Michelin, Volvo, Mercedes Benz, etc.) may heighten UC. For instance, advertisements which depict a product as "totally safe" or "absolutely harmless" may create a latent sense of security regarding product safety. Thus we hypothesize:

H1: Those products which include a safety warning are associated with lower levels of UC than are products without safety warnings.

H2: Those products associated with advertisements stressing product safety are associated with higher levels of UC than products without such ads.

As mentioned above, strong casual attributions can be expected as an injury increases in UC, or is more unexpected. According to the "Just World Hypothesis" (Lerner and Miller 1978), the world is, in general, orderly and an individual's pursuit will not be blocked by environmental interference. Thus, unusual events "require for their occurrence a greater causal role by the victim or perpetrator" (Kelley and Michela 1980, p. 476). Therefore the respondents will be highly motivated to assign the responsibility for such an incident to someone or something. Furthermore, Weiner (1982) has posited that many negative events, such as a product injury, are likely to lead to external, rather than internal, attributions. In this case, higher levels of UC should increase attributions of blame toward the manufacturer (i.e., the manufacturer must have failed in their duty to provide a safe product or to make the consumer aware of the danger) and reduced attributions to the usage situation (i.e., a highly unexpected event would not occur by chance in a just world). We pose the following hypotheses:

H3: A positive relationship exists between UC and assignment of blame to the manufacturer.

H4: A negative relationship exists between UC and assignment of blame to the situation.

The result of this series of hypotheses is an indirect effect of both warnings and safety promotions on attributions of blame. The impact is felt through the intervening variable, UC.

Direct Effects. Aside from the indirect effects, a series of hypotheses represent direct effects of the experimental manipulations on attributions of responsibility. Is there a benefit from going beyond safety standards established by governmental agencies? Under the doctrine of strict liability, simply meeting the required standards is all that is necessary and anything in addition (if an injury actually does result) is wasted. In making attributions, however, the "controllability" (Rosenbaum 1972; Weiner, Russell and Lerman 1979) of the causal factor is relevant information. In practice, safety standards are not under volitional control of a manufacturer, but a willingness to exceed these standards in an effort to make a safer product is. Assuming that government standards are in place to insure the safety of products, any goods exceeding these standards must be even "safer than necessary." Consequently, accidents resulting from one of these products may be perceived of as carelessness on the part of the consumer. Further, such actions by a manufacturer may create perceptions of a company which went out of its way to protect its product users and thus may be less likely to be blamed for a product mishap.

H5: Products having safety standards which exceed government standards are associated with lower levels of attributional blame toward manufacturers than are products that simply meet those standards.

H6: Products having safety standards which exceed government standards are associated with lower levels of attributional blame toward the product-use situation than are products that simply meet those standards.

A similar argument can be made regarding the presence of safety warnings. The manufacturer has fulfilled his responsibility by providing a warning concerning the danger of the product, reducing his blame for any ensuing mishap. While making a safer product (exceeding the safety standards) can reduce accidents due to chance, no amount of safety warnings can prevent a true accident from occurring. We anticipate:

H7: Those products which include a safety warning are associated with lower levels of blame toward the manufacturer than are products without safety warnings.

RESEARCH APPROACH

Focus Groups

As mentioned earlier, the manipulations in the Figure represent three characteristics of product safety identified in a series of focus group interviews concerning attitudes toward product safety and the jurisprudence system. In total, six focus groups comprised of 61 participants were conducted. Focus group members were selected to represent a wide cross-section of demographic characteristics including sex, age, education, income, occupation, and knowledge and experience with the legal system. Throughout each focus group interview, the three factors (warnings, regulations, and advertisements stressing safety) consistently emerged as important factors to consider in judging responsibility for product injuries. These three variables, therefore, were thought to be relevant for further analysis based on qualitative analysis, the level of managerial controllability and supposed courtroom irrelevance.

Experimental Design

Having determined three relevant managerial variables which might determine attributions of responsibility for product-related injuries, a 2 X 2 X 2 between-subjects design was employed to test the various research hypotheses. Each experimental variable was manipulated over two levels: (1) warning labels were positioned as either being present or not present, (2) the manufacturer was said to have either far exceeded or merely met all government safety standards, and (3) the product ads were described as either stressing safety or other product features. A product-related injury scenario was built around these manipulations which was patterned after an actual liability case involving an injury resulting from the use of a claw hammer. In Chappuis v. Sears Roebuck & Co. (1978), a young man lost the vision in his left eye when a chip flew off the head of a hammer and struck him in the eye. Aspects other than the manipulations (i.e., type of injury, plaintiff characteristics, and the defendant's reaction, etc.) were held constant across all subjects.

Data Collection

The data collection procedure follows the approach suggested by Abramson and Mosher (1975; Abramson, Goldberg, Abromson, and Gottesdiener 1975) and utilized in a psycholegal study by Feild (1978). Abramson and Mosher (1975) suggest using a large number of interviewers with varied demographic and personality traits in the data collection process in an effort to eliminate any interviewer bias. Feild (1978, p. 160) applied this procedure in a study of attitudes toward rape. In Feild's experiment, students were familiarized with the survey instrument and trained in the administration of its measures. The students were then used as field interviewers in questionnaire administration.

Following these procedures, students enrolled in marketing research courses at a large midwestern university were thoroughly familiarized with the survey instrument, and received training in its administration. Each student completed interviews with four members of the urban community, excluding students and university personnel, in exchange for course credit. Telephone surveys were conducted with a subsample of the respondents to control the quality of the survey. Following the telephone verification process and deleting questionnaires with incomplete responses, 91% of the questionnaires distributed were retained for analysis. The resulting sample consisted of 204 subjects. The demographic characteristics of the sample are representative of the community as a whole based on age distribution, male-female ratio, and ethnic origin, and slightly above average on education and income.

This procedure provides several benefits. First, interviewers of both sexes, a mixture of ethnic backgrounds, and varied styles of interaction will be utilized to administer the survey instrument, thus avoiding the pitfall pointed out by Abramson et al. (1975). Second, the variance in characteristics of the interviewers is likely to be reflected in the respondents. Third, interviewers will be able to clarify any questions the respondents might have while completing the questionnaire. Finally, the nonresponse bias often associated with mail surveys can be avoided; Feild (1978) reported a usable response rate of 82%.

Manipulation/Confounding Checks

In experimental designs, subjects are exposed to various conditions and expected to react differently to those conditions. To insure the internal validity of this experiment, manipulations were tested and revised during pretesting as suggested by Aronson and Carlsmith (1968) and Wetzel (1977). The effectiveness of the experimental manipulations in the final scenarios was then assessed by asking the respondents if: (1) the product's warning labels were sufficient (warnings); (2) the manufacturer's advertisements stressed product safety (advertisements); and (3) the product exceeded government safety regulations (regulations).

To examine the manipulations the subjects were classified into high and low groups for each of the manipulations and then tested for a difference in the mean response on the three check questions. With this approach, both the effectiveness of the manipulations and any potential confounding effects between manipulations can be tested (Wetzel 1977). Or, as presented by Perdue and Summers (1986, p. 318), "the experimenter would like to be able to demonstrate that (1) the treatment manipulations are related to "direct" measures of the latent variables they were designed to alter and (2) the manipulations did not produce changes in measures of related but different constructs." In each case, a significant difference was found between mean scores on the appropriate measure, but not the others, suggesting that the manipulations were effective and independent. The results are presented in Table 1.

Measurement Results

Assignment of Blame/Responsibility. In measuring the assignment of blame/responsibility, two questions had to be answered. First, the appropriate bases for the causal attributions had to be identified. That is, who or what is the accident attributed to? Second, what was the suitable term(s) to use to capture causality?

In regard to the first question, Kelley's (1967, p. 194) principle of covariance (Kelley Cube) established three dimensions of causal inferences: (1) the stimulus object or entities; (2) the observer of the event or person; and (3) the context or time in which the effect occurs. In a frequently cited study of causal attribution, McArthur (1972, p. 175) operationalized these dimensions by asking respondents to assign the cause of an event to either (1) "something about the person," (2) "something about Stimulus X," (3) "something about the particular circumstances" or (4) a combination of these factors. Bettman (1979) has suggested that when applied to the study of consumer behavior, the corresponding causal agents would be the product, the person, and the situation. Folkes (1984) successfully utilized this categorization in her study of causal attributions of product failure. It is important to note, however, that Folkes (1984, p. 75) expanded the "product" category to include the retailer which sold the product (pants) as well as the manufacturer. Based on these studies, the consumer, the product/manufacturer, and the situation were determined to be the appropriate bases of causal attribution for the study.

TABLE 1

RESULTS OF MANIPULATION CHECKS

TABLE 2

FACTOR ANALYSIS OF BLAME AND RESPONSIBILITY ITEMS

The second question arises from the conceptual work of Shaver (1985), who argues persuasively that attributions of blame and responsibility are not identical and the different dimensions might relate more to one attribution than to another. Based on Shaver's supposition, pretests have been conducted using attributions of both blame and responsibility toward the person, product, and situation. The results revealed a high correlation between measures of blame and responsibility for each of the three bases of attribution. A similar finding has recently been reported by McCaul, Veltum, Boyechko, and Crawford (1990) who employed measures of attributions of both blame and responsibility in a study of rape victims. Their results (McCaul et al. 1990, p. 13) also indicated a high correlation between the two measures (average correlation = .62), which they consequently summed for further analysis. Therefore, despite Shaver's arguments, empirical evidence suggests that respondents have a difficult time distinguishing between the concepts of blame and responsibility.

Furthermore, principal components analysis of the six items indicates that a two factor solution is appropriate. As can be seen in Table 2, the four items measuring attributions to the consumer and manufacturer loaded on a single factor (but with opposite loadings) and exhibit a high degree of internal consistency (Cronbach a = .92). At the same time, the responsibility and blame items for the situation loaded on a separate factor. Thus, scales utilizing both terms are included in the survey instrument to provide a multi-item measure, but will be summed to create a measure of blame/responsibility toward (1) the manufacturer and (2) the situation.

TABLE 3

MODEL STATISTICS - PREDICTING ATTRIBUTIONS OF BLAME FOR PRODUCT MISHAPS

Unanticipated Consequences. UC is designed to measure the respondent's judgement of the plaintiff's expectations regarding product safety. In other words, does the respondent think the user of the product could or should have anticipated the danger associated with its use? Five items were devised to capture UC and refined and modified in pretesting. Three items were retained that both load highly on a single factor (.76 to .82) and display acceptable internal consistency (a = .76), providing a parsimonious yet reliable scale. The items reflects the extent to which an incident should be (or should have been) expected given the nature of the product.

MODEL ANALYSIS AND RESULTS

The Figure depicts a three-equation path model employed to test each hypothesized relationship. Each hypothesis is illustrated by a single path in the proposed structure, and the strength and nature of each relationship can be represented by its standardized regression coefficient (b). Given that the model is recursive and overidentified, ordinary least squares can be used to estimate its parameters and assess the level of predictive validity associated with each endogenous construct (Blalock 1961; Asher 1976; Darden 1981). Consequently, Table 2 contains the results of estimating each equation using a series of OLS regressions, from which, a test of the previously discussed hypotheses can be inferred.

Overall Fit

Although our intent was not to provide a test of a fully specified model, an analysis of any regression model should begin by assessing its general level of significance. The overall fit of each equation comprising the path model, as indicated by its coefficient of determination (R2) and corresponding F-ratio, achieved an acceptable level of statistical significance. On closer inspection, the model was much more successful in predicting attributions of blame toward the manufacturer/consumer (p < .001), than it did attributions of blame toward a situation (p < .02) or UC (p < .01). However, given that a significant amount of variance in each dependent variable was explained by the proposed model, some general level of support is provided for its hypothesized relationships, and more specifically, justification is provided for interpreting each independent variable's predictive ability (Neter, Wasserman, and Kutner 1985). Specific results, by independent variable, for each hypothesis follow.

Hypothesis Testing

Safety Regulations. Two hypotheses (5-6) concerned attributional effects of a manufacturer meeting (coded 0) or exceeding (coded 1) government imposed safety regulations. Essentially, these hypotheses predict that manufacturers perceived as exceeding safety standards will experience (1) lower levels of blame directed toward themselves and higher levels directed at the product user (H5), as well as (2) lower levels of blame directed toward a situation in which an injury occurred (H6), than will firms merely meeting those standards. The first row of path coefficients shown in Table 3 supports H5 (-.27) but fails to support H6 (-.03). Thus, while exceeding safety standards may help alleviate some blame from manufacturers at the expense of consumers, it does not appear to influence attributions toward the usage situation.

Warning Labels. A second set of hypotheses involves the influence of warning labels. As predicted by H1 a negative relationship (-.21) is found between use of warning labels and unanticipated consequences (UC). More simply, this relationship indicates that warning labels do provide consumers with information concerning potential results from misuse. In contrast, consumers may not consider these results when warning labels do not accompany a purchase. Additionally, as predicted by H7, a negative relationship (-.21) is observed between use of warning labels and attributions of blame toward a manufacturer. Thus, by employing warning labels, marketers may be perceived as somewhat less responsible for a product-related injury.

Promoting Safety. One hypothesis concerns the use of safety as an appeal in ads and promotions. Hypothesis 2 indicates that safety appeals may give consumers a sense of security making injuries more unforeseeable. The positive path coefficient (.18) between the safety manipulation and UC supports this hypothesis. In turn, safety as an appeal may affect attributions of blame through its ability to make an injury more unanticipated.

Unanticipated Consequences (UC). Two hypotheses concern the direct effect of UC on attributions of blame. Hypothesis 3 predicts higher levels of blame toward manufacturers when an injury is more unanticipated and is strongly supported in this analysis. A positive path coefficient of .63 is observed between UC and attributions of blame toward manufacturers indicating that when a product-related injury occurs, manufacturers (consumers) are more likely to be blamed under high (low) levels of UC. Further, this significant path accentuates the effects of warning labels on attributions toward a manufacturer/consumer. The indirect, or intervening, effect of UC on this relationship contributes an additional influence of -.13 or (-.21 X .64) to the effect of warning labels on attributions of blame toward a manufacturer/consumer yielding a total contribution of -.34 to the total zero order correlation between these two variables. Thus the theoretical relationship between warning labels and attributions of blame on the manufacturing/consumer construct is sensitive to the intervening presence of unanticipated consequences.

Additionally, H4 proposes lower levels of attribution of blame toward the situation when an injury is highly unanticipated. Consistent with this hypothesis, a significant negative relationship is found between UC and situational blame. That is, consumers tend to blame the situation less when a product-related injury is unanticipated. Further implications of these results follow.

DISCUSSION

This study demonstrates that managerial actions do affect consumers' assignment of responsibility for product mishaps. More interestingly, since the types of activities considered in the study should have little effect on due process in liability trials according to the doctrine of strict liability, it appears that the mental arithmetic of consumers when making personal causal attributions of blame does not resemble the legal process of strict liability. That is, the reasoning used by consumers does not coincide with the reasoning of the court. If additional research were to prove consistent with this result, it may well spark a debate among policy makers over the reasonableness of recent product liability legislation.

A number of implications for corporate policy makers flow more directly from the experimental manipulations. For example, using warning labels to make consumers more aware of potential dangers appears one way to alleviate some of the negative results of product liability suits. The results of this study showed a marked reduction in blame attributed to manufacturers when a warning label was involved in the product-injury scenario. And while in some cases the use of warning labels is predicated upon previous court findings, the findings with regard to safety regulation indicate that going over and above legislative guidelines can reduce negative repercussions of product injuries. In the present study, firms portrayed as incorporating safety features above and beyond what is required legally experienced lower levels of causal attributions of blame than did firms simply meeting government standards. In addition, these findings may be indicative of how severe the impact of negative public publicity surrounding product injury situations will be (Griffin, Babin and Attaway 1991). Publicizing its use of warning labels and extra safety precautions may alleviate some of the fall out from these unfortunate situations.

Further, some evidence is presented which suggests that stressing safety as a feature of a product can create a sense of security among consumers and reduce consumer perceptions of unanticipated negative consequences from use of a product. Left for further study is the potential impact that steps taken to deal with product safety have on consumer purchase intentions and behavior. It is quite possible that tactics useful for dealing with the high costs of product liability (e.g., warnings) have a negative impact on sales. If so marketers are faced with trading off potential liability for potential sales.

Another interesting issue investigated in this paper deals with how an unanticipated event affects relationships between marketer actions and consumer attitudes. Evidence of this effect is demonstrated by the secondary influence of warnings on unanticipated consequences of product use, which in turn acts as an additional causal factor on attributions of blame toward the manufacturer. Practically speaking, further steps taken by marketers to increase consumer awareness of hazardous product uses can allay some attributions of blame toward the manufacturer. However a similar caveat as mentioned above, with respect to these actions impact on purchase intentions, holds here.

Additionally, the study of unanticipated negative events has some relevance for basic marketing research. The impact of UC demonstrates that like unexpected favorable events (Howard and Barry 1990), unexpected negative events also influence consumer decision processes. Further, although no direct comparisons are made in this study, an interesting issue related to this research concerns direct comparisons of the impact of unexpected positive versus unexpected negative events. As discussed earlier, a substantial theoretical base (e.g., Tversky and Kahneman 1981) supports a more severe effect of losses compared to gains. Also, in terms of consumers' post-purchase product evaluations, unanticipated negative events may create consumer reactions which are beyond the scope of a traditional disconfirmation paradigm. Further studies are needed which directly compare the validity of the disconfirmation model to a model based on unanticipated rather than expected events.

CONCLUSIONS

The area of product safety clearly presents a heavy regulated operating environment for marketers and consumers. These regulations, the current legal doctrines guiding product liability suits, as well as the high cost of product liability, serve to limit alternatives for dealing with these issues. However, results presented in this paper indicate that marketer actions do have an impact on consumer attributions of responsibility. In addition to these results, some interesting theoretical issues are raised concerning the effect of unanticipated negative events. Clearly, unanticipated consequences do affect consumer decision processes, however, more work is needed to understand more closely the nature of this effect.

REFERENCES

Abramson, Paul R. and Donald L. Mosher (1975), "Development of a Measure of Negative Attitudes Toward Masturbation," Journal of Consulting and Clinical Psychology, 43, 485-490.

Abramson, P., P. Goldberg, D. Mosher, L. Abramson, and M. Gottesdiener (1975), "Experimenter Effects on Responses to Explicitly Erotic Stimuli," Journal of Research in Personality 9, 136-146.

Aronson, Elliot and J. Merrill Carlsmith (1968), "Experimentation in Social Psychology," in The Handbook of Social Psychology, 2nd ed., Vol.2, Gardner Lindzey and Elliot Aronson, eds., Reading, MA: Addison-Wesley Publishing Co.: 1-79.

Asher, Herbert B. (1976), Causal Modeling, Sage Publications: Beverly Hills.

Blalock, Hubert H. (1961), Causal Inferences in Nonexperimental Research, The University of North Carolina: Chapel Hill.

Busch, Paul and Joseph F. Hair (1980), "Product Liability and Safety: Prospectives from Business Versus Public Policy Makers," Journal of Business Research 8, 485-499.

Busch, Paul (1976), "A Review and Critical Evaluation of the Consumer Product Safety Commission: Marketing Management Implications," Journal of Business Research 4 (October), 41-49.

Chappuis v. Sears Roebuck & Co., La., 368 So.2d (1978): 926-931.

Churchill, Gilbert A. and Carol Surprenant (1982), "An Investigation into the Determinants of Consumer Satisfaction," Journal of Marketing Research, 24 (November), 491-504.

Darby, M. R., and E. Karni (1973), "Free Competition and the Optimal Amount of Fraud," Journal of Law and Economics, 16 (April), 67-86.

Darden, William R. (1981), "Review of Behavioral Models in Marketing," in Review of Marketing 1981, American Marketing Association: Chicago.

Darden, William R., James DeConinck, Barry J. Babin, and Mitch Griffin (1991), "The Role of Consumer Sympathy in Product Liability Suits: An Experimental Investigation of Loose Coupling," Journal of Business Research, 22 (January), 65-89.

Dickerson, R. (1951), Products Liability and the Food Consumer, Brown and Company: New York.

Escola v. Coca Cola Bottling Company, 24 Cal. 2d. 453, 150 P. 2d. 436. 1944.

Feild, Hubert S. (1976), "Attitudes Toward Rape: A Comparative Analysis of Police, Rapists, Crisis Counselors, and Citizens," Journal of Personality and Social Psychology, 36 (February), 156-179.

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

Friedman, L. (1973), A History of American Law, Simon and Schuster: New York.

Friend, Tim (1989), "Cosmetic Makers Under Fire Over Safety," USA Today (April 10,), 1D.

Gatty, Bob (1987), "Solving the Liability Crisis," Nation's Business (June), 39-41.

Griffin, Mitch, Barry J. Babin and Jill S. Attaway (1991) "An Empirical Investigation of the Impact of Negative Publicity on Consumer Attitudes and Intentions," in Advances in Consumer Research, Vol. 18, ed. Mike Solomon, Ann Arbor, MI: Association for Consumer Research.

Howard, Daniel J. and Thomas E. Barry (1990), "The Evaluative Consequences of Experiencing Unexpected Favorable Events," Journal of Marketing Research 27 (February), 51-60.

Interagency Task Force on Products Liability (1977), Final Report, I-1, Nov.1.

Jury Verdict Research (1988), Personal Injury Valuation Handbook, Jury Verdict Research, Inc: Solon, Ohio.

Kelley, Harold H. (1967), "Attribution Theory in Social Psychology," in Nebraska Symposium on Motivation, David Levine, ed., University of Nebraska Press: Lincoln, NE. 1967.

Kelley, Harold H. and John L. Michela (1980), "Attribution Theory and Research," Annual Review of Psychology, 31, 457-501.

Lerner, M. J. and D. T. Miller (1978), "Just World Research and the Attribution Process: Looking Behind and Ahead," Psychological Bulletin, 85, 1030-1051.

Levy, L. (1957), The Law of the Commonwealth and Chief Justice Shaw, Oxford University Press.

Martin, Robert (1989), "The Defense Never Rests," AOPA Pilot 32 (October), 160-162.

McArthur, Leslie Z. (1972), "The How and What of Why: Some Determinants and Consequences of Causal Attribution," Journal of Personality and Social Psychology, 22 (May), 171-183.

McCaul, Kevin D., Lois G. Veltum, Vivian Boyechko, and Jacquelin J. Crawford (1990), "Understanding Attributions of Victim Blame for Rape: Sex, Violence, and Foreseeability," Journal of Applied Psychology 20 (1), 1-28.

Mihoces, Gary (1990), "Liability Insurance Suggested for Coaches," USA Today (April 26), 14C.

Morgan, Fred W. (1982), "Marketing and Product Liability: A Review and Update," Journal of Marketing 46 (Summer), 69-78.

Morgan, Fred W. (1986), "Product Liability Developments and the Nonmanufacturing Franchisor or Trademark Licensor," Journal of Public Policy and Marketing, 5, 129-41.

Morgan, Fred W. (1988), "Tampered Goods: Legal Developments and Marketing Guidelines," Journal of Marketing, 52 (April): 86-96.

Morgan, Fred W. and Dana I. Avrunin (1982), "Consumer Conduct in Product Liability Awards," Journal of Consumer Research, 9 (June), 47-55.

Nelson, P. (1974), "Advertising as Information," Journal of Political Economy, 81 (July/August), 729-754.

Neter, John, William Wasserman, and Michael H. Kutner (1985), Applied Linear Statistical Models, Irwin: Homewood, IL.

North, David M. (1985), "General Aviation Manufacturers Forecast Static Deliveries," Aviation Week and Space Technology, 12.2 (January 28): 79-80.

Olsen, Walter (1989), "The Case Against Expert Witnesses," Fortune, (September 25), 136-38.

Olshavsky, Richard W. and John A. Miller (1972), "Consumer Expectations, Product Performance, and Perceived Product Quality," Journal of Marketing Research, 9 (February), 19-21.

Perdue, Barbara C. and John O. Summers (1986), "Checking the Success of Manipulations in Marketing Experiments," Journal of Marketing Research, 23 (November), 317-326.

Private Pilot (1990), "Current Production Piston-Engine General Aviation Aircraft," 25 (May), 51.

Rosenbaum, R. M. (1972), A Dimensional Analysis of the Perceived Causes of Success and Failure, Doctoral Dissertation, University of California, Los Angeles.

Settle, Stephen M. and Sharon Spigelmyer (1984), Product Liability: A Multibillion-Dollar Dilemma, American Management Association: New York.

Tse, David K. and Peter C. Wilton (1988), "Models of Consumer Satisfaction Formation : An Extension," Journal of Marketing Research, 25 (May 1988): 204-12.

Tversky, Amos and Daniel Kahneman (1981), "The Framing of Decisions and the Psychology of Choice," Science, 211 (1981), 453-58.

Weiner, Bernard (1982), "The Emotional Consequences of Causal Attributions," in Affect and Emotion, Margaret Clark and Susan Fiske, eds., Lawrence Erlebaum Associates: Hillsdale, NJ. 1982.

Weiner, Bernard (1988), "Spontaneous Causal Thinking," Psychological Bulletin, 97 (1), 74-84.

Weiner, Bernard, Dan Russell, and David Lerman (1979), "The Cognition-Emotion Process in Achievement-Related Contexts," Journal of Personality and Social Psychology 37 (7), 1211-1220.

Werner, Ray O. (1982), "Marketing and the United States Supreme Court," Journal of Marketing, 46 (Spring), 73-81.

Wetzel, Christopher G. (1977), "Manipulation Checks: A Reply to Kidd," Representative Research in Social Psychology, 8 (2), 88-93.

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