Some Perspectives on Risk Acceptance

Jacob Jacoby, Purdue University
ABSTRACT - The theme of this session -- namely, that any hazard management system developed to protect consumers from the risks associated with products must take into consideration subjective (psychological) as well as technical factors -- was illustrated by considering the hazardous consequences that potentially could result from implementation of the U.S. Department of Agriculture's proposed rule on providing uncured frankfurters, bologna, and bacon to consumers. Several positive and negative aspects of the empirical research described at this session were then discussed. The paper concluded by presenting a framework for conceptualizing hazard management that is emerging from currently ongoing research into consumer awareness, comprehension and usage of health and safety information.
[ to cite ]:
Jacob Jacoby (1981) ,"Some Perspectives on Risk Acceptance", in NA - Advances in Consumer Research Volume 08, eds. Kent B. Monroe, Ann Abor, MI : Association for Consumer Research, Pages: 511-516.

Advances in Consumer Research Volume 8, 1981      Pages 511-516

SOME PERSPECTIVES ON RISK ACCEPTANCE

Jacob Jacoby, Purdue University

[Preparation of this paper was supported, in part, by a grant from the National Science Foundation (PRA 7920585); see Jacoby and Jaccard, in progress.]

ABSTRACT -

The theme of this session -- namely, that any hazard management system developed to protect consumers from the risks associated with products must take into consideration subjective (psychological) as well as technical factors -- was illustrated by considering the hazardous consequences that potentially could result from implementation of the U.S. Department of Agriculture's proposed rule on providing uncured frankfurters, bologna, and bacon to consumers. Several positive and negative aspects of the empirical research described at this session were then discussed. The paper concluded by presenting a framework for conceptualizing hazard management that is emerging from currently ongoing research into consumer awareness, comprehension and usage of health and safety information.

As many of us who teach consumer behavior tell our students, the modern era of consumer protection can be traced to President John F. Kennedy's Declaration of Consumer Rights in 1962. This Declaration described four rights: the right to be informed, the right to choose, the right to safety, and the right to be heard (or the right of redress). Partly because of its intriguing nature, to date, the lion's share of consumer research directed toward public policy issues has focused on the right to be informed (including the necessity for information to be free of deceptive or misleading connotations and the necessity for the information to be sufficiently complete so that the consumer is able to make an informed decision). It is time that a commensurate research effort be mounted in regard to product safety. As Feldman (1980, p. 73) points out: "Compared with other consumer protection issues, product safety is paramount because unlike other issues such as deceptive advertising or the failure to make good on warranties, product safety directly involves injury and loss of life."

Our objectives in this commentary are two-fold. The first is to consider selected aspects of the empirical papers presented at this session. The second is to indicate directions for future work.

COMMENTS ON THE PAPERS PRESENTED

The fundamental tenet which underlies these papers is that any hazard management system developed to protect consumers from the risks associated with products must take into consideration consumer psychological (that is, internal, subjective) factors as well as technical factors. As eloquently noted by Slovic, Fischhoff and Lichtenstein; "People respond to the hazards they perceive. If their perceptions are faulty, efforts at public and environmental protection [which do not consider these perceptions] are likely to be misdirected." Further: "Our basic premises are that both the public and the experts are necessary participants in [the risk assessment] process, that assessment is inevitably subjective, and that understanding public policy perceptions is crucial to effective decision making." We fully concur with this set of premises. Indeed, this point needs to be underscored; it cannot be emphasized enough! It especially needs to be brought to the attention of policymakers and regulators who sometimes end up generating misguided and incorrect public policies through their failure to understand and take into consideration the relevant consumer psychological factors which my be operating.

On the Risk of Ignoring Consumer Psychological Factors

As a concrete illustration of how the disregard of consumer psychological factors might result in seriously increasing rather than decreasing the safety hazards associated with consumer products, consider the U. S. Department of Agriculture's efforts in regard to providing consumers with uncured (i.e., nitrite and nitrate free) meat products. Stimulated by research (which has since been discredited) that seemed to indicate that the ingestion of nitrites and nitrates was associated with unreasonably high levels of cancer, and also stimulated by consumer activists, the USDA developed a proposed rule on the subject. The proposed rule, published in the Federal Register on August 21, 1979 and scheduled to become effective on September 20, 1979, provided that meat and meat products that were not cured with nitrates, nitrites, or other preservatives might be sold under their traditional name, so long as the word "uncured" appeared on the label as part of the product name and the label contained the following cautionary language: "No nitrate, or nitrite added; not preserved, keep refrigerated below 40 Fahrenheit at all times." (See Figure 1) The regulation also required that the uncured products be similar in size, shape, flavor, color, consistency and general appearance to the product tacitly prepared with nitrite or nitrite. In other words, the consumer would have no other cues to the fact that the frankfurter (or bacon, or balogna) she was purchasing was somehow different than frankfurters (bacon, balogna) she had purchased before --except for the cautionary labeling prescribed by the rule.

FIGURE 1

USDA'S PROPOSED PRECAUTIONARY LANGUAGE AS IT WOULD BE APPLIED TO FRANKFURTERS

On September 20, 1979, the National Pork Producers Council, together with three members of the U. S. House of Representatives, filed a lawsuit challenging the proposed rule. These plaintiffs were subsequently joined by the National Independent Meat Packers Association. Their argument was heard in Federal Court in Des Moines, Iowa, in November of 1979. In presenting their case, these plaintiffs brought forth two expert witnesses.

The first witness, an internationally acknowledged authority on the subject of food toxins, testified that eight ounces of botulism spore were sufficient to wipe out the population of the entire planet. He went on to state that under certain conditions (such as might happen if a high school student took an uncured bologna sandwich to school, left it in his locker on a hot day, chose not to eat it for lunch but took it to his after school game instead, and then ate it at 5:00 p.m.) deadly botulism spore could develop in as little as 8 hours.

Participating as the second expert witness, the present author testified regarding likely consumer attention to, comprehension of, and tree of the cautionary labeling information. In brief, and based upon evidence obtained across more chin a score of consumer information usage investigations conducted to dace, the testimony pointed out that: (1) Many consumers (perhaps as many as 20%) do not read any information on the labels for frequency purchased consumer products [Remember: the uncured product had to be similar in size, shape, flavor, color, and general appearance to the cured product. The only distinguishing characteristic would be the cautionary language.] beyond the brand name. (2) The 80% who do read package information rarely read all the available information but tend to focus on a very small percentage of this information (e.g., price). (3) Some of those who happen to read the proposed cautionary labeling would not understand its significance. From informal research conducted by this author (e.g., asking at least 500 people in his classes and in audiences to whom he has spoken whether they understood what the word "cured" meant when it was applied to a food product) and others, it would appear that fewer than 5% of the adult population comprehends that the term "cured" refers to the inclusion of preservatives. The overwhelming majority of people seem to have no idea as to what "cured" stands for in this context and a much smaller proportion (in the order of 5% to 15%) seem to believe that cured means "smoked." Further, even though most individuals would be expected to understand the literal meaning of "Not preserved -- keep refrigerated below 40 Fahrenheit at all times," this language provides no understanding of the dire consequences that might ensue from not keeping the product refrigerated below 40 at all times. (4) Finally, even if the consumer did comprehend the full significance of this statement, once the meat had been taken from the refrigerator(and, for example, been placed in a sandwich), subsequent conditions might be out of her control and it would still be possible for the meat to be handled unsafely.

After hearing the case and considering the record, on February 12, 1980 the District Court in Iowa entered a final order enjoining the government from enforcing or applying the challenged regulations. This decision was appealed by the USDA and reversed just a few days ago, on September 23, 1980. In noting the reasons for its reversal, the Circuit Court cited the fact that warning labels have been generally shown to be effective (how come, then, people still smoke cigarettes? etc.), and that many of the consents submitted by consumers in response to the proposed rule "demonstrated consumer awareness of the potential dangerous consequences of marketing and consuming meat products that do not contain nitrites or nitrates." The Court apparently did not recognize the danger of extrapolating from such a small and unrepresentative sample as the group of consumers who chose to send in written comments on the proposed rule. The Court further argued that "the cautionary labeling will adequately inform the consumer of how to maintain such products in a wholesome condition until consumed." Preposterous! At the present time, the National Pork Producers are considering appealing this reversal to the next highest level of the Judiciary which, in this case, is the Supreme Court.

To my way of thinking, the proposed rule and subsequent reversal by the Circuit Court of the District Court's decision provides a frightening example of how public policy can be misguided when only technical factors are considered and subjective, consumer psychological factors are ignored. Having served as an expert witness in more than a dozen matters, I can honestly say that I have never felt as strongly (from my perspective as a consumer researcher) that the courts and policy makers acted in complete disregard for what is known regarding human behavior -- and that this might result in tragedy of catastrophic proportion for American consumers, True, if all consumers did indeed read and understand that uncured meat had to be refrigerated at all times and acted in accord with this belief, the risk would be reduced substantially. [Risk would still be associated with the stages of product preparation and handling prior to its reaching the consumer.] However, it is highly unlikely that all consumers will read and understand the cautionary language and behave as expected. By not taking these subjective considerations into account, the government is, in my opinion, opening up a Pandora's Box and substantially increasing the health and safety hazards to which the American populace will be subjected.

As the authors of these manuscripts have noted, the consumer element in hazard amazement is of critical concern and cannot be neglected. If the consular fails to recognize that a risk exists, then his behavior can not be shaped accordingly. In other words, these authors are barking up a very important tree and their efforts need to be encouraged, applauded, and hopefully supplemented by considerable research from others. It is also commendable that their efforts, particularly in the cause of Slovic, Fischhoff, and Lichtenstein, are clearly long term and programmatic.

On The Empirical Work Presented

Having acknowledged the significance of this work and the genuine debt that consumer researchers owe to these individuals, we still find several shortcomings associated with the papers presented. These tend to be shortcomings which come with the territory when researchers embark on journeys over relatively unexplored terrain. Fortunately, none of these problems are insurmountable.

One entry on the negative side of the ledger is in regard to the relatively small and unrepresentative samples employed. It is conceivable that this might hive been a major contributing factor in Slovic et al's failure to replicate the findings of dimensional structure that had been obtained in the earlier investigation by Fischhoff. A second problem would be the controversial and somewhat questionable application of such high powered analytical techniques as factor analysis (Slovic et al. 1981) and stepwise multiple regression (Rethans and Albaum 1981) to such enroll samples, particularly given such large numbers of variables.

However, our major concern with these studies is not the issue of sample size or the questionable application of high powered statistics to samples of such size, but the operationalizations of the core concept of risk acceptance. The measures of risk acceptance used by Slovic et al. and Rethans and Albaum are provided in Figures 2 and 3, respectively. Consideration of these measures suggests that, while the researchers may have had a very clear idea in mind of what risk acceptance was, and while their operationalizations may be technically accurate reflections of these conceptualizations, it is highly questionable whether many of their subjects adequately understood these operationalizations. Each of these operationalizations is long, complicated, contains several difficult concepts and are the kinds of concepts with which most lay people can be expected to be unfamiliar. Interpretation of these responses necessarily rests on the assumption that the respondent comprehends both the concepts and vocabulary being used. Yet how certain can we be that the respondents satisfactorily comprehend such sentences as: "On the other hand, there may be some activities or technologies that you believe are currently safer than the acceptable level of risk."? And just how confident can we be that the typical respondent has enough understanding of the issues to be able to assess whether "the advantages of increased safety are not worth the cost to society of reducing the risks by altering the product or restricting its use."? To interpret a response to these questions as being valid assumes that the respondent understands just what the "advantages of increased safety" are, knows what "the costs to society" are, has some idea of what these costs and benefits are "worth," has an understanding of what kinds of "alterations" of the product are possible, or what kinds of "restrictions on its use" may be imposed, etc. The questions further assume that the respondents can keep all these relatively unfamiliar and complicated considerations simultaneously in mind while answering the question. They further assume at certain parts that the respondent's thinking will be unidirectional, that is, if he thinks the product shouldn't be altered, he also thinks its use shouldn't be restricted. To my way of thinking, without any attempt to confirm and validate these numerous key assumptions, none of them are tenable.

FIGURE 2

SLOVIC, FISCHHOFF AND LICHTENSTEIN'S OPERATIONALIZATION OF RISK ACCEPTANCE

FIGURE 3

RETHAN'S OPERATIONALIZATION OF RISK ACCEPTANCE

In other words, I doubt whether the majority of subjects fully and accurately comprehended the questions they were responding to. Substantial evidence in the literature on questionnaire construction exists to indicate that even simple questions are miscomprehended by large proportions of the populace. Indeed, in a set of studies described at this Conference yesterday (cf. Jacoby, Nelson and Hoyer 1981), evidence was provided to indicate that 70 to 85% of subjects trying to understand a remedial advertising statement which was only one sentence in length either incorrectly understood that statement or extracted a confused meaning from it. How much greater, then, can we expect the levels of miscomprehension to be for a passage containing ten sentences referring to complicated and unfamiliar concepts? Until such time as evidence is supplied to indicate that the respondents accurately understood the operationalizations, we must entertain the possibility that the data reflect GIGO.

FUTURE DIRECTIONS

Clearly, work needs to be directed toward developing operationalizations which are something more than a collection of words which the investigators assume are capable of adequately assessing the concept of interest. It first needs to be established that these words are properly understood by the respondents before data obtained using these measures can be interpreted as being meaningful and programmatic work using these measures is undertaken.

Beyond these empirical considerations, I believe that the research thrust outlined by these authors would benefit considerably from a closer integration with the two decades old body of literature on the issue of perceived risk, This literature began with Bauer's (1960) seminal conceptualization and has since been summarized by Cox (1967) and Rose (1975), among others. At least one hundred studies have been conducted using this concept and a large body of findings have emerged. Among these findings is the fact that at least six different types of perceived risk (including perceived safety risk) have been identified and cross validated, and that perceived safety risk correlates with overall perceived risk .67 but contributes the least amount of explained variance in overall perceived risk scores (cf. Jacoby and Kaplan 1972; Kaplan, Szybillo and Jacoby 1974). Consideration of this body of literature further reveals that the first three conclusions offered by Slovic et al. -- namely, perceived risk is quantifiable and predictable; lay people sometimes differ systematically in their perceptions of risk; and the greater the perceived risk, the greater the desired risk production -- already have a well-established history in the consumer literature. One would hope that a merging and integration of this earlier body of research with the current approach would enable us to go beyond what is already known.

Finally, I believe that the work that is conducted needs to be placed into a broader, conceptually richer framework. One emerging approach [We thank Rethans for bringing this to our attention.] is to consider the subject of hazard management as involving two major components: technical and social issues (see Figure 4). The technical side of the ledger consists of the identification of hazards and the measurement of these hazards in terms of their probability of occurrence and the severity of their consequences. The social side of the ledger includes risk perception (that is the probabilities and consequences as perceived by the respondents) and risk acceptance.

FIGURE 4

A PERSPECTIVE ON HAZARD MANAGEMENT

Our own work on consumer awareness, comprehension, and usage of health and safety information (Jacoby and Jaccard, in progress) has involved developing an elaborated perspective based upon the earlier framework. This framework has continued to evolve and the current version is offered here for critique and use by others (see Figure 5). The perspective retains the notion of technical and social (or objective and subjective) components as being fundamental.

On the technical side, the framework begins with the actual identification of hazardous conditions and hazardous consequences. Hazardous conditions are assumed to be of two main types -- those associated with the product and those which occur with the interaction between the product and the environment (where the environment is meant to include the consumer as well). Table 1 illustrates these distinctions for insecticides. Note that various hazardous conditions are believed to be associated with the product and the manner in which it is being provided for consumer use. For example, the product may (or may not) contain ingredients which are toxic and pose a safety problem for humans. The product may be packaged in a container which does not have a childproof cap and/or it may contain inadequate labeling and usage instructions. In addition, many insecticides tend to be safe in terms of their contents and packaging but will become hazardous when certain environmental conditions are present. For example, some insecticides are safe except when used in the presence of exposed flames. Other insecticides are safe except when the environment in which they are used contains either pregnant women, asthmatics, pollen sensitive people, infants, pets, people seriously ill, or people on prescribed medications. Yet other insecticides (particularly of the spray variety) only become dangerous to humans when used in the presence of exposed food stuffs.

TABLE 1

ILLUSTRATING THE DISTINCTION BETWEEN HAZARDOUS CONDITIONS AND HAZARDOUS CONSEQUENCES FOR THE PRODUCT CATEGORY "INSECTICIDES"

The actual presence of a hazardous condition (e.g. a bald tire on a car) does not necessarily lead to a hazardous consequence (e.g., an accident). The hazardous consequences which evidence suggests are associated with insecticide usage include poisoning, lasting nerve damage, miscarriages, birth defects, a deleterious impact on the production of liver enzymes, and cancer.

While the first major aspect of technical issues involves the identification of hazardous conditions and consequences, the second aspect involves assessment of: the probability of occurrence of each hazardous condition; the probability of each hazardous condition resulting in a hazardous consequence; and the severity of the consequences should they occur.

The third element on the technical issue side involves the identification of appropriate coping measures. These may be divided into those which are preventative in nature and those which are corrective. Preventative coping measures are those which enable the individual either to reduce or completely eliminate the existence of the hazardous condition or the hazardous consequence. (An example would be the use of safety goggles when using electrical tools). Corrective coping measures are those which should be employed to correct the existence of a hazardous condition or the occurrence of a hazardous consequence, e.g., adding air to an underinflated tire (preventative) or turning the wheel of a skidding automobile in the direction of the skid while pumping the brakes rapidly (corrective).

FIGURE 5

TOWARD AN ELABORATED PERSPECTIVE ON HAZARD MANAGEMENT

FIGURE 6

TOWARD INTEGRATING OBJECTIVE AND SUBJECTIVE COMPONENTS.

The fourth element in the technical system would involve a consideration of the systems and procedures (media mix, execution strategies, etc.) that could/should be used for communicating the information regarding hazardous conditions, consequences, and coping measures to the consumer.

On the other side of the ledger are the subjective factors. These begin with risk perception and the consumer's awareness of the nature of the hazardous conditions and consequences associated with a particular product's use. In concert with Slovic et al. and Rethans, we firmly believe that if a consumer is unaware of either hazardous conditions or consequences, his behavior will not be affected by such conditions or consequences even though they may be serious and exist at high probability levels. In addition to awareness, one should consider the individual's probability judgments regarding these hazardous conditions and consequences as that individual believes they apply to others in general, and to himself in particular. As a preliminary attempt, Figure 6 outlines a framework for integrating some of these basic subjective factors with their objective counterparts.

Beyond the notion of risk perception, one has to take into consideration the factors affecting risk acceptance. While the consumer's perception of risk hazards and consequences may be accurate, he may still be willing (and even agree) to use a very risky product. In this regard, the authors of the papers presented at this session have identified at least a dozen factors seemingly associated with risk acceptance. To this list we would like to add the notion of "general risk-taking propensity." It appears to us that certain individuals are more inclined to be risk takers than others, and that this factor needs to somehow be taken into consideration in attempts to develop practical hazard management systems.

The subjective side of the ledger also needs to consider the risk elimination, risk reduction, and risk coping strategies that consumers can engage in. These include risk avoidance behavior (e.g., shunning the questionable product entirely), risk reduction behavior (i.e., reducing one's vulnerability to risk while maintaining at least some level of product usage), and continuing full use of the product while acquiring information regarding how to cope with hazards if and when they surface.

In this latter regard, it would be particularly useful for researchers who wish to influence public policy to identify the types of information consumers do use, and the sources to which they turn to in order to obtain this important in formation.

Beyond the behaviors that consumers actually engage in, one needs to assess the extent and accuracy of the beliefs consumers possess in mind regarding preventative measures and corrective measures. Again, to the extent that consumers have nonexistent or erroneous beliefs regarding corrective measures, different means for affecting their behavior will have to be employed.

Finally, any attempt to develop policies which will affect consumers at large needs to consider general individual difference and market segment factors. Certain risks tend to be much more likely, given certain combinations of market and individual difference factors, For example, it seems likely that risks are more probable with the elderly consumer, the disadvantaged consumer, the undereducated consumer, the very young consumer, etc.

CONCLUSION

In conclusion, we are enthused with the efforts engaged in by these researchers and would like to encourage both their efforts and others to embark on similar efforts. However, we recommend considerable caution in accepting their results until we can be more confident that the measures of their key concept (namely, risk acceptance) are meaningful and valid.

REFERENCES

Bauer, R. A. (1960),"Consumer Behavior as Risk Taking." In R. S. Hancock (Ed.) Dynamic Marketing for a Changing World. Chicago: American Marketing Association.

Cox, D. F. (Ed.) (1967), Risk Taking and Information Handling in Consumer Behavior. Cambridge, Mass.: Harvard University Press.

Feldman, L. P. (1980), Consumer Protection: Problems and Prospects (2nd ed.). St. Paul, Minnesota, West Publishing Co.

Jacoby. J. and Jaccard, J. J. "Assessing the Effects of Science Based Information on Consumer Technological Choices." Grant from the Division of Policy Research and Analysis, National Science Foundation. In progress.

Jacoby. J. & Kaplan, L. B. (1972), "The Components of Perceived Risk." In M. Venkatesan (Ed.), Proceedings, Third Annual Conference of the Association for Consumer Research. University of Chicago: pp. 382-393.

Jacoby. J., Nelson, M. C., and Hoyer, W. D. (1981), "Correcting Corrective Advertising." In K. B. Monroe (Ed.) Advances in Consumer Research, Volume 8.

Kaplan, L. B., Szybillo, G. J. and Jacoby, J. (1974),"Components of Perceived Risk in Product Purchase: A Cross-Validation." Journal of Applied Psychology, 59(3), 287-291.

Rethans, A. J. and Albaum, G. S. (1981), "Toward Determinants of Acceptable Risk: The Case of Product Risks." In K. B. Monroe (Ed.), Advances in Consumer Research, Volume 8.

Ross, I. (1975), "Perceived Risk and Consumer Behavior: A Critical Review." In M. J. Schlinger (Ed.), Advances in Consumer Research, 2, 1-20.

Slovic, P., Fischhoff, B. and Lichtenstein, S. (1981), "Facts and Fears: Societal Perception of Risk." In K. B. Monroe (Ed.), Advances in Consumer Research, Volume 8.

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