Analysis of the Consumer Expectation Test
The Consumer Expectation Test—A Concept in Search of Meaning
By Bill Masters
Wallace, Klor & Mann, P.C.
“Here I would like to say: a wheel that can be turned though
nothing else moves with it, is not part of the mechanism.”
- Wittgenstein, PI part I, § 271
Metaphors of a Magnifico
Twenty men crossing a bridge
Into a village
Are twenty men crossing twenty bridges,
Into twenty villages,
Or one man
Crossing a single bridge into a village
This is old song
That will not declare itself ….
In Oregon, the consumer expectation test (“CET”) is the exclusive standard of strict products liability. McCathern v Toyota Motor Corp., 332 Or 59, 23 P3d 320 (2001). That test has been roundly criticized as ambiguous, vague and overbroad. As Professor Keeton noted, “the test can be utilized to explain most any result that a court or jury chooses to reach.” W. Page Keeton, et al. Prosser and Keeton on the Law of Torts, p. 699 (5th ed. 1984); see Restatement (Third) of Torts: Products Liability § 2(b).
As a purported hedge against that imprecision, consumer expectations are held, for the most part, to be provable empirically. E.g., McCathern v Toyota Am. Inc., 332 Or 59, 78, 23 P3d 320 (2001) (The court has stated that the CET is subject to empirical proof, but has failed to delineate the nature of that proof). For if a party could not offer evidence to establish consumer expectations, the CET would be ontologically akin to the standard in negligence cases of the “reasonably prudent man.” That is, it would be merely a placeholder for the discretionary judgment of the jury.
A number of kinds of information have been postulated as providing adequate proof of consumer expectations, to include: what the plaintiff says; what the jury thinks; evidence of risk-utility or alternate designs; marketing and advertising data; sales or demand data; safety statutes or regulations; industry safety standards; industry custom; surveys; focus group findings; expert opinion; product instructions; and ex cathedra judicial declarations of what consumers expect. See B. Hamlin, Admissible Evidence of “Consumer Expectations” (OSB Products Liability Newsletter, Winter 2003-2004). Yet none of these kinds of information, upon scrutiny, is satisfactory proof of what the community of consumers expects from use of the product.
If consumer expectations are provable empirically, as is the purported case in Oregon, what is it a party needs to prove? The answer to that question must entail some sense of the “intension” of the concept (the defining criteria under which the predicate picks out the class of objects which it describes) in order to gather and present the proof at trial required to instantiate the “extension” of that concept (the class of objects that is described by the predicate). Oxford Dictionary of Philosophy, p. 132 (1996).
This article explores the “intension” of the concept of CE—its conceptual coherence–in the framework of classic choice or rational decision-making theory. See generally, M. Allingham. Choice Theory (2002). The basic premise of this article is that a conceptual analysis of the CET, at bedrock, is a matter of rational decision-making at the level of individual consumers and then at the level of groups of consumers. The thesis of this article is that the CET, in the strong light of decision-making theory, is conceptually incoherent. It is a concept akin to the concept of the “four-sided triangle.” At first glance, it may seem to have a sense, but on further examination it is revealed to have none.
- Is The Consumer Expectation Test Conceptually Coherent?
The first step in assessing the conceptual coherence of the CET, then, is to provide an analytical definition of that standard. The unanalyzed standard, as stated in the Restatement (Second) of Torts, is as follows:
[A] product in a defective condition is “unreasonably dangerous” if it is “dangerous to an extent beyond that which would be contemplated by the ordinary consumer who purchases it, with the ordinary knowledge common to the community as to its characteristics.” Restatement (Second) Torts §402A, Comment i.
In this test, four key words or phrases need to be parsed: (A) What does “dangerous” mean? (B) Who is the “ordinary consumer”? (C) What is “ordinary knowledge”? (D) What is meant by the phrase “common to the community”?
- What Is Dangerousness?
“Dangerousness” is the propensity of a product, in use, to produce an outcome that the user disprefers because the outcome is bodily harm or property damage. See Phillips v Kimwood Machine Company, 269 Or 485, 525 P2d 1033, 1037-38 (1974). But before that harm is actually caused, the user or purchaser has only a subjective sense of the likelihood that the use of the product will result in harm. “Dangerousness” (“D”) (or its inverse “safety”), then, is a function of the “probability” (“Pr”) of certain outcomes upon use of the product and the user’s “preference” or utility (“U”) for those outcomes. In symbolic terms: D(Pr x U).
A subthesis of this article is that dangerousness is not necessarily an assessment invariant across individuals. It is rather an assessment relative to each individual. Just as beauty is in the eye of the beholder, danger in use of a product is in the eye of the consumer or user. So what is dangerous is a function of the individual’s preferences for the outcomes of use and the “subjective probability” that that use will result in that outcome. M.D. Resnik. Choices, pps 68-80 (1987); B.S. Everitt, The Cambridge Dictionary of Statistics. p. 250 (1998) (“[subjective] probability represents a degree of belief in a proposition, based on all the information….[t]he only constraint is that a single person’s probabilities should not be inconsistent”). Moreover, outcomes of bodily injury would be expected to occur to varying degrees of severity. As a result, the user would need to establish his or her preference for those varying degrees of injury.
- Who Is The “Ordinary Consumer”?
The phrase “the ordinary consumer” is designed to signify that the consumer is not using the product in an extraordinary way. That is, the ordinary consumer is different from the extraordinary consumer who may use the product unusually. See, e.g., Gaines-Tabb v. ICI Explosives, USA, Inc., 160 F.3d 613, 624 (10th Cir 1998) (“The ‘ordinary consumer’ is ‘one who would be foreseeably expected to purchase the product…’.” “…the ordinary consumer of [ammonium nitrate] branded as fertilizer is a farmer [not a terrorist”); Brown v. Sears, Roebuck & Co., 328 F3d 1274, 1281 (10th Cir 2003).
- What Is “Ordinary Knowledge”?
The term “knowledge” signifies something more than mere unsupported speculation about the product in use. “Knowledge” signifies, at least, true belief and, at best, “justified true belief.” R. Audi. Epistemology, pps. 220-222 (2003). A false belief is not knowledge. (For example, a Venetian in the 12th century CE believes that the earth is flat.) Nor is a belief based on a lucky guess. (For example, a Minoan in the 10th century BCE believes the earth is spherical.) Knowledge is a justified true belief. (For example, an American in the 21st century CE believes, based on images taken from the space shuttle while in space, that the earth is spherical.) The phrase “ordinary knowledge” signifies the level of knowledge about the product of the non-expert.
- What Is the Meaning of the Phrase “Common to the Community?”
The phrase “common to the community” signifies that the members of the community “share” this ordinary knowledge about the product. See Moss v Crosman Corp., 136 F3d 1169, 1175 (7th Cir 1998) (“The law … creates an objective standard, under which the court must take into account ‘the reasonably anticipated knowledge, perception, appreciation, circumstances, and behavior of expected users'”). This criterion is designed to make the CET objective. E.g., Pruitt v. General Motors Corp., 74 Ohio App3d 520, 599 NE2d 723, 726 (10th Dist Ct of App 1991) (“The test utilizes an objective standard, not the subjective expectations of a particular user or consumer); Greene v. Boddie-Noell Enterprises, Inc., 966 F Supp 416, 419 (WD Va 1997); Redman v John D. Brush & Co., 111 F3d 1174, 1181 (4th Cir 1997)(“[The plaintiff’s] subjective expectations are insufficient to establish what degree of protection … society expects from [the product]”).
It is tempting to conclude that what is signified by the phrase “common to the community” is that the information about the product from which expectations can be formed is merely in the public realm. But the phrase is better interpreted as requiring more than that–namely that the information be not only in the public realm but also be “commonly known.”
But this intuition poses the first serious problem in understanding the CET–the notion of “common knowledge.” Decision theorists (and philosophers and economists) distinguish “common knowledge” from “mutual knowledge.” “Common knowledge” is such that each individual knows A, knows that the other individuals know A, knows that each individual knows that each knows A, and so on. M. Allingham. Rational Choice, ch. 5 (1999); D. Lewis. Convention, p. 52 (2002). “Mutual knowledge,” by contrast, is knowledge such that each individual merely knows A. In the context of the CET, the phrase “ordinary knowledge common to the community” probably means something less than “common knowledge.” If it were otherwise, such a condition would likely not be empirically satisfied. So it is assumed here that that phrase means something akin to “mutual knowledge.”
What knowledge is “common to the community” is a function of the sum of what each individual, as an ordinary consumer with ordinary knowledge, knows. This inference is not a big leap. Knowledge is not floating around in the air like ether—it cannot exist apart from what each individual knows.
The “ordinary consumer” with “ordinary knowledge” “common to the community” is a surrogate for the community of non-expert consumers (purchasers) of the product. That is, “the ordinary consumer” is a phrase designed to differentiate the ordinary from the extraordinary user of the product. The phrase “ordinary knowledge” is designed to differentiate the knowledge of the product of the non-expert from that of the expert. Moreover, the ordinary consumer is not making an assessment on the basis of his or her individual ordinary knowledge, but on the basis of what the community of ordinary or non-expert consumers mutually know about the product.
The Consumer Expectation Test Restated
The CET restated to reflect these foregoing analytical theses is as follows:
The product has a greater degree of danger as evaluated by experts than the expected degree of danger as evaluated by the community of non-expert purchasers or consumers of the product.
This phrasing of the CET, in opposition to the original phrasing, captures the impending decision problem more concisely.
III. Decision Theory
The pending question here is, does the community of non-expert purchasers or consumers of the product exist? This question is analyzed within the context of decision theory, the study of quantitative models of the judgments involved in and leading to rational deliberative choice. S. French & D. Rios Insua. Statistical Decision Theory, p. 3 (2000); see generally, M. Allingham. Choice Theory (2002). This theory indicates how rational decision makers should make decisions, not how people who may or may not be rational actually make decisions. In that light, it is a premise of this article that no methodological difference exists between having a consumer identify his or her expectations about the product and having the consumer rationally decide, using decision theory, what are the outcomes of various uses of the product. See D.J. White. Decision Theory, p. 85 (1969).
What follows will assess first (A) the nature of individual assessments of expectations of the danger posed by the use of the product and then (B) the nature of group or community assessments of expectations of the danger posed by the use of the product, with the group or community expectation being a function of individual expectations.
- Individual Expectation
A consumer would use the product in a variety of ways and those uses in various circumstances (or states of nature) would result in a variety of outcomes, some of which will be desirable and some undesirable. And so, an individual’s expectation is a function of the four basic variables of decision theory:
(1) An action (“A”)—or intended use of the product;
(2) States of nature (“SN”) that may become instantiated generating particular outcomes (O) from various uses of the product;
(3) A “subjective probability” (“SPr”) assessment of the likelihood of the occurrence of various states of nature or outcomes resulting from use of the product; and
(4) The “utility” (“U(O)”) or preference ranking for the various outcomes. M.D. Resnik. Choices, chpt. 1 (1987)
More particularly, each action (A) or use of the product is a gamble with several possible outcomes (O), each with its own subjective probability (SPr) of occurring and its own index of preference (U(O)). (In the following discussion, the CET is defined in terms of subjective probabilities and utility; but not in terms of the risk-utility of alternative designs. This omission insures that the definiendum, “consumer expectations,” is stated in the definiens in terms other than the risk-utility test.)
Menu of Expected Outcomes
Each consumer would establish his or her expectations about the dangerousness (or safety) of using the product by considering a menu of all possible individual expectations about use of the product—that is, the universe of possible individual expectations. The individual non-expert purchaser or user would select from this menu only those outcomes (expectations) for which he or she has justified [true] beliefs, through a process of contemplative thought based on knowledge about the product before the accident occurs. This contemplated series of performances or outcomes would be imagined prospectively to avoid “hindsight bias.” “Hindsight bias” is the tendency of people to assign probabilities to the occurrence of an event higher in retrospect than they would in prospect. J. Baron. Thinking and Deciding, p. 145 (2000). The tendency would be to assign, in retrospect rather than in prospect, a higher probability to an unsafe use than to a safe use.
For each individual, each expected outcome of use of the product would have a particular utility. For example, if the “individual non-expert purchaser (or user)” #1 decides to drive his car at the speed limit and the state of nature of heavy rain occurs, he may drive his car into a heavy puddle of water on the roadway, and as a result his car may leave the roadway, crash into a tree, breaking his leg. To this outcome, the “individual non-expert purchaser (or user) #1, may assign a low utility. (The outcomes involving bodily injury may involve different degrees of injury and hence different assignments of utility or preference.) Each individual’s selections from the menu of all possible expectations can be then arranged in a “utility matrix.” So, for various uses of his car, the “individual non-expert purchaser (or user) #1 may have a utility matrix as follows:
|Action (A) /State of Nature (SPr(SN))||SPr (SN1)
(Drive at speed limit)
- Collective or Community Expectations
Once all the “individual non-expert purchasers (or users) [“INEP’s”] have identified their individual expectations in the use of the product, these individual expectations [or their components: SPr(O š SN) and U(O)] must be the basis for a collective or community expectation pursuant to a “group preference profile.”
Group Preference Profile
|SPr(O š SN)||UTILITY|
|INEP#1||SPr(O š SN)||U(O)|
|INEP#2||SPr(O š SN)||U(O)|
|INEP#3||SPr(O š SN)||U(O)|
|INEP#4||SPr(O š SN)||U(O)|
|3 ?||3 ?|
Three potential models are candidates for group decision-making: (1) aggregation of individual preferences (or matrices); (2) sole decision-maker with advisors; and (3) group interaction to consensus. S. French & D. Rios Insua, Statistical Decision Theory, ch. 4 (2000). Of these three models, the first is most appropriate in the context of analysis of the CET. The second model is inconsistent with the CET in that the decision-maker does not treat the advisors as providing information that can be disregarded if the decision-maker is so inclined. (This model also does not square with the surrogate model of the CET discussed earlier at II D.) The third model is inconsistent with the CET in that the members of the community do not interact as would, say, a jury in reaching a consensus about what should be the group decision.
If aggregation of individual matrices is analytically possible, and if the user determines that it is rational to use the product in a particular use (A), and that use results in an accident, the post accident outcome will entail the following information: (1) Action (A); (2) state of nature (SN); (3) outcome (O); (4) utility of outcome [U(O)]. This information can then be plugged into the appropriate cell in the “aggregated” matrix to determine what the expectations were of the community of consumers as to that particular A, SN and U(O).
With an aggregation model, what would be aggregated would be the components of the individual matrices: (1) the individual utilities [U(O)] and (2) the individual subjective probability assessments [SPr (O š SN)]. The question is, is the aggregation of these components of the individual matrices analytically possible?
(1) Aggregation of Utilities
The first logical obstacle to the conceptual coherence of the CET is this: Individual utilities cannot be aggregated into a group preference in any meaningfully rational way. K. Arrow. Social Choice and Individual Values (3ed 1961); FS. Kelly. Arrow’s Impossibility Theorems (1978); M. Bacharach. Group Decisions In The Face of Differences of Opinion. Management Science, 22: 182-191 (1975); Jones, B., et al. Condorcet Winners and the Paradox of Voting: Probability Calculations for Weak Preference Orders. Am. Political Science Review, 89: 137-144 (1995). This impossibility is demonstrated by “Arrow’s Impossibility Theorem.”
Arrow’s General Possibility Theorem (Impossibility Theorem)
It is logically impossible to add up the choices of the “individual non-expert purchasers (or users) [INEPs] into an unambiguous collective or community choice, not just under a “constitution” based on the principle of majority rule but under every conceivable constitution, except a dictatorship. [A constitution is the way in which individuals’ preferences are taken into account in determining the group’s choices.]
As stated by Nobel prize winner Kenneth Arrow himself, “If we exclude the possibility of interpersonal comparisons of utility, then the only methods of passing from individual tastes [preferences] to social preferences [collective or community preferences] which will be satisfactory and which will be defined for a wide range of sets of individual [preference] orderings are either imposed or dictatorial.” KJ. Arrow. Social Choice and Individual Values, p. 59. (1951).
The problem of aggregating utilities would be less intractable if utilities could be compared interpersonally. But most decision theorists believe such comparisons are unfeasible, at best, and illogical, at worst. M.D. Resnik. Choices (1987)(“…the problem of the interpersonal comparison of utilities remains unresolved”).
(2) Aggregation of Subjective Probability Distributions
Not only are individual preferences not subject to being aggregated, but also individual subjective probabilities cannot be aggregated into a group probability distribution. N. Dalkey, An Impossibility Theorem for Group Probability Functions. Rand Paper p-5683, Santa Monica, CA.
- The Consumer Expectation Test Is a Rhetorical Device
If the CET were conceptually incoherent, then its purpose would seem to be, by default, that of a rhetorical device–a trope to be used to move the jury emotionally and to excuse proof of safer alternative designs. As the foregoing has demonstrated, the CET is indeed conceptually incoherent. And so, it follows, CET is merely a rhetorical device. As a result, a search for its “extension” is futile.
This result is worrisome: Vacuous legal concepts such as the CET (concepts with no coherent intension and no extension) can be manipulated by juries to justify results on the basis of variables other than on the basis of the announced variable–in this case the standard of consumer expectations. That is, such vacuous concepts can act as a screen for the jury to base its decisions on bias and prejudice rather than on the rule of law. In effect, the court would be more candid with the jury if it were to change UCJI No. 48.03 to read as follows:
“A product is unreasonably dangerous when it is dangerous to an extent [jurors, please use whatever criteria you wish].”