Cognitive Reference Points in Consumer Decision Making

ABSTRACT - Recent work in decision theory demonstrates the impact of reference points on consumers' decision making processes and outcomes. This paper suggests a series of conceptual and methodological issues that should be addressed by further research in this area.

Citation:

Noreen H. Klein and Janet E. Oglethorpe (1987) ,"Cognitive Reference Points in Consumer Decision Making", in NA - Advances in Consumer Research Volume 14, eds. Melanie Wallendorf and Paul Anderson, Provo, UT : Association for Consumer Research, Pages: 183-187.

Advances in Consumer Research Volume 14, 1987      Pages 183-187

COGNITIVE REFERENCE POINTS IN CONSUMER DECISION MAKING

Noreen H. Klein, Virginia Polytechnic Institute and State University

Janet E. Oglethorpe, Virginia Polytechnic Institute and State University

[The authors would like to thank Kent B. Monroe and Ruth A. Smith for their comments on an earlier draft of this paper.]

ABSTRACT -

Recent work in decision theory demonstrates the impact of reference points on consumers' decision making processes and outcomes. This paper suggests a series of conceptual and methodological issues that should be addressed by further research in this area.

INTRODUCTION

Some of the most exciting recent research in decision theory deals with the effect of perceptual processes on evaluation and choice. In particular, Kahneman and Tversky's (1979) prospect theory shows that the evaluation of outcomes is strongly dependant on a reference point. A reference point determines the position of choice outcomes on a value function, thereby determining their desirability. Host research on reference point effects on evaluation and choice has been done in the context of risky choice, where an alternative is presented as an array of outcomes with associated probabilities of occurrence. However, Thaler (1985) has also demonstrated that prospect theory's value function can explain consumer choice behavior that is contrary to traditional economic theory.

The purpose of this paper is to explore some implications of reference points for consumer decisions. We propose a series of conceptual and methodological issues that should be pursued in this area. The discussion of these issues incorporates the ideas and results of Thaler (1985) and others with liberal speculation about possible extensions. Our immediate objective is to stimulate thinking about the role of reference points in consumer choice, to help to create an agenda for research in the area. We contend that reference points deserve this attention because of their potential consequences for consumer evaluation and choice, information search, and the organization of consumer knowledge.

TYPES OF REFERENCE POINTS

Reference points have generally been defined as any stimulus which "other stimuli are seen in relation to" (Rosch 1975). This definition's simplicity is deceptive; it encompasses a variety of reference points that may differ in terms of how they're derived and what they represent to the consumer. For instance, one reference point may be the consumer's perception of what is typical in the marketplace, another may be an aspiration level for a purchase, and a third may be the alternative currently owned. The variety of possible reference points is reflected in the list of reference points for price shown in Table 1. This list was compiled from discussions of possible reference prices in the literature (cf. Della Bitta and Monroe 1973), and from personal interviews with students about reference prices.

Some implications of such a broad definition need to be explored. First, it should be noted that "a stimulus that other stimuli are seen in relation to" may be external as well as internal. This contrasts with alternative definitions, such as Thaler's (1985) definition of a reference price as the "expected or fair' price, which is an internal representation. In Table 1, "a particular price I've seen recently" may in some instance be the price of an alternative in the evoked set, to which other prices are compared during a choice process. The reference price need not be internalized as an expectation, aspiration or typical value. Second, some of the internalized reference prices in Table 1 are also incompatible with Thaler's definition. For example, the price I would like to pay, the most I would pay, or a good buy may all be distinct from a realistic expectation. Although a more restricted definition may be appropriate in a particular theory, it is important to recognize the existence of alternative reference points and account for their possible use as part of the theory's boundaries. We return to this point later in discussing transaction utility.

TABLE 1

PROPOSED TYPES OF REFERENCE PRICES

The list of distinct reference points in Table 1 is presented with some reservations. Consider the three categories into which they are grouped. Aspiration level reference points represent consciously established goals or targets for an attribute. Market reference points are based on the perceptions of attribute levels in the marketplace. Historical reference points represent a consumer's purchase experience, which may incorporate only a subgroup of market levels. These categories are not mutually exclusive, either empirically or conceptually. For example, in Table 1 the usual price you pay or the typical market price may also be your aspiration level for a particular purchase. The last price you paid may have been determined by a previous goal.

This recognition leads to a central conceptual question. Are the items listed in Table 1 themselves reference points for price or are they simply factors that together influence the reference price? In some sense, it seems that both are true. The last price paid may be retrieved by a consumer and used alone to evaluate current prices. However, an aspiration level reference point probably adapts to information such as the last price paid, shifts in typical market prices, specific prices that are currently salient, etc. One research priority should be to determine how consumers themselves conceptualize their reference points, if indeed they do, and to discover the factors that affect that representation.

There is an additional complication in that the reference points in Table 1 may not really exist as points, but rather as a range of values. For instance, two perceptually distinct levels may both be typical. An aspiration level may include several equally satisfactory values. e. R. an acceptable range of prices (Monroe 1984).

In such a case, the comparison of a stimulus to a referent entails a judgment about its inclusion within a range of values, rather than a comparison to a specific value. The "fair price" definition of reference price used in Thaler's theory of transaction utility also seems likely to encompass a range of prices.

Despite the fuzziness of the categories and types of reference points proposed in Table 1, we suggest that the classification highlights some useful distinctions among alternative reference points, such as the information used in their formation, the importance of goals, and other factors that are discussed later. Additional reference points, or alternate groupings that create better theoretical or pragmatic classifications, may become apparent AS more research is done.

FACTORS INFLUENCING REFERENCE POINT SELECTION

Many of the reference points discussed above may be used at various times by a decision maker; the question arises as to what factors influence the use of one type of referent over another. A second issue pertains to influences of the level of the reference point; e.g. given that an aspiration level reference point is used, what determines how high or low that price is? Three areas of influence seem likely: dimensions of the stimulus and its context, the task, and individual differences.

Stimulus and Context Characteristics

First, the value that the stimulus possesses on the attribute being evaluated may influence its initial classification, thereby eliciting a particular reference point. This seems most likely to occur when there are distinct subgroups within a product class, which possess different reference points. For example, different quality/price lines of a product are common. When an attribute is evaluated that is diagnostic of membership in a quality line, the relevant classification is made and the appropriate reference point is selected. Thus a consumer looking at a pair of $60 Vuarnet sunglasses may initially classify them as designer or high performance sunglasses due to the price or brand name, which elicits the reference price for that type of sunglasses. A $7 pair of Foster Grants could elicit the much lower reference price for drugstore sunglasses. Not only the specific reference price but the type of reference price may be determined by the stimulus value. For instance, a price within the normal range for sunglasses may be compared to the typical price paid, whereas an extremely high price could elicit a comparison to an aspiration level, such as "the most I would ever pay".

Characteristics of the choice context should also influence reference point selection. The heterogeneity of the levels of a particular attribute should affect the decision maker's incentive to set goals for that attribute. Homogeneous offerings curtail the decision maker's ability to improve an outcome, therefore the use of aspiration levels may be less frequent. The general attractiveness of an alternative set with regard to an attribute may influence the level at which an aspiration reference point is set, with more attractive alternatives eliciting more ambitious goals.

Task Characteristics and Frames

Evaluation tasks, which call for an absolute judgment, may invoke internal standards as reference points ore readily than choice tasks. Choice heuristics seem likely to involve the comparison of alternatives under consideration to each other with greater frequency, implying the use of external reference points. In addition, the choice objectives of optimizing versus satisficing could be influential, with optimization associated with higher goals, and therefore eliciting more favorable reference points.

The frame used to represent the purchase task has also been shown to elicit different reference points. (See Thaler (1985) for a complete discussion of the effects of these frames, or budgetary accounts.) A calculator being purchased as a gift might be framed as either a calculator purchase or as a gift purchase. When the reference points for these two classifications diverge, the different frames could lead to different choices. A price for a particular offering may be evaluated as unacceptably high for a calculator, but reasonable for the gift occasion (or vice-versa). Another example would be framing a discretionary expenditure as a splurge or reward as opposed to an ordinary purchase within the relevant product category. Both splurges and gifts are interesting frames, because they cut across product categories, but intuitively seem likely to have distinctive reference points with respect to attributes such as price, style, quality, etc.

Individual Differences

The reference points in Table 1 differ in their degree of abstraction; different information and different amounts of processing are required for their formation. Reference points derived directly from a particular alternative (last one bought, one recently seen) are very concrete, require minimal specific information and no elaborate processing of that information. However, reference points involving typicality (typical market level, the typical attribute level I buy) generally require the acquisition and assessment of information about a distribution of attribute levels before the reference point is derived. Aspiration level reference points require that preferences be determined in order to form goals. Since the information required and the demands of its processing vary with the reference point used, individual differences such as the degree of involvement or interest in the product, knowledge of product attributes, and attribute importance may be influential in their selection. When involvement, knowledge, and attribute importance are low, no reference point may even be formed. High involvement and high attribute importance seem most likely to result in conscious goal setting, thus invoking an aspiration level. Greater knowledge of the attribute levels and their distribution in the market environment should be associated with reference Points based on typical values.

The market environment itself may limit such knowledge in some cases. For example, prices of professional services (e.g. legal, medical) are usually not promoted. Therefore historical reference prices, based on personal experience rather than market information, should be used. Zeithaml and Graham (1983) found that students' reference prices for medical and legal services were generally quite inaccurate, and that for medical services little prepurchase price evaluation is done.

EVOLUTION AND ADAPTATION OF REFERENCE POINTS

Reference points are dynamic, and the process by which they are formed and changed deserves examination. Two kinds of change are possible: a change in the type of reference point used and a change in the specific value of the reference point. Consider first the reference point type. A novice buyer of a particular product should initially have little knowledge about the distribution of attribute levels in the market and some uncertainty about preferences for those levels. The use of reference points based on typicality or aspiration level: may therefore be limited. A concrete alternative should be an easier reference point to use: for instance a well known brand or the one that is first encountered. The type of reference point used may evolve as buyers pass through stages of increasing familiarity with the product. For example, a novice PC buyer may acquire information about the IBM PC first because it's the best known exemplar of the product category. Information about attributes of other personal computers is then compared to the levels possessed by an IBM pc. As familiarity with the product class increases and preferences are formed, reference points involving typicality and/or aspirations are possible.

Reference points can therefore serve somewhat different functions as they evolve. An initial reference point based on a single alternative may be used as a way of organizing information; i.e., acquiring a knowledge structure about the product class and alternative brands. Evaluation should play a larger role as knowledge develops, and preferences for attribute levels are formed.

Marketers also need to be concerned with the dynamics of both long term and short term changes in the value of the reference point. These changes have usually been modeled as an anchoring and adjustment process, with the initial reference point forming the anchor which is adjusted by new information. The nature of this adjustment process is uncertain, and probably varies situationally. Helson (1964) hypothesizes that an adaptation level is the weighted geometric mean of all previous stimuli. The characteristics of new information received by consumers, such as its importance, credibility, vividness, degree of discrepancy from the reference point, and accessibility should help to determine the weight it is given, hence determining the degree of adjustment.

Marketers sometimes attempt to simply present a new reference point for a consumer's evaluation, through direct comparisons to specific competitors or to the typical offering in the market. These tactics seem most likely to succeed with consumers whose previous reference point is based on little or uncertain information. Consumers are likely to perceive promotional information with some skepticism. Liefeld and Heslop (1985) gave respondents advertisements with both the sale and regular prices of 8 variety of products and asked them for estimates of the ordinary selling prices. Average estimates ranged up to 18 percent below the true ordinary selling price, indicating that the regular prices presented in the advertisements as reference points were not believed.

REFERENCE POINT EFFECTS

Transaction Utility

In prospect theory, the reference point determines whether outcomes are characterized as gains or losses; a shift in reference point can alter the comparative value of outcomes and therefore alter choice. Thaler (1985) explains several instances of noneconomic consumer behavior through the application of prospect theory's value function. Thaler (1985) asserts that the total utility of a market exchange is composed of two elements: acquisition utility and transaction utility, whose determinants are defined as follows:

p_a = the price of alternative a

p^*_a = the reference price of alternative a (the expected or fair price)

^-p_a = the value equivalent of alternative a (the sum of money which the individual would be indifferent to receiving as a gift instead of a)

Acquisition utility is a function (v) of the value equivalent of an alternative and its actual price: v(pa,-pa). Transaction utility is derived from comparing the price paid to the expected or fair price on the same value function: v(-pa,-pa). Paying less than expected generally contributes the positive transaction utility of a bargain. The value of buying alternative a at price pa is therefore the sum of its acquisition and transaction utilities:

v(^-p_a,-p_a) + v(-p_a,-p^*_a).

Thaler (1985) has detailed instances in which different frames create different reference points, with subsequent effects on transaction utility and ultimately on willingness to buy or sell. One often quoted illustration of the effect of transaction utility is that of the hypothetical man on the beach, who respondents think will be willing to pay more for a bottle of his favorite brand of beer when it is purchased for him from a resort hotel than when it is purchased from a grocery. In this case the beers are identical, therefore the value equivalent of each beer is the same (ph=pg). However, people expect to pay more for beer at a fancy hotel, to the reference price of the hotel beer is higher (ph>pg). Thus when ph=pg, the transaction utility of purchasing the hotel beer is greater than that of purchasing the grocery beer: v(-ph,-ph) > v(-pg,-pg). The higher reservation price for the hotel beer is the result of this added transaction utility. Some implications of transaction utility are explored below.

Single Versus Multi-alternative Choices

The scenario given above demonstrates the role of transaction utility in the context of a buy-no buy decision for one alternative. Does transaction utility play an analogous role in a choice made among several alternatives? Consider the man on the beach whose friend offers to bring back his favorite brand of beer from either the grocery store or the hotel. Suppose that the prices of the two alternatives are known to be equal (ph=pg), and all additional costs of obtaining the beers are equivalent. The acquisition utility of the two beers is therefore equal. However the decision maker should not be indifferent between these two identical beers at the same price; the higher transaction utility for the hotel beer implies that it should be chosen over the grocery beer.

Suppose that you extend the example by increasing the price of the hotel beer. As p increases, both the acquisition utility (v (ph,-ph)) and the transaction utility (v (-ph,-ph) of the hotel beer decrease. At some price (ph>pg) The total loss of value due to the increased price of the hotel beer should make the decision maker indifferent between it and the lower priced grocery beer. Until the hotel beer reaches this price it has higher value, and should be chosen even when its price exceeds that of the identical grocery beer.

The problem is that it doesn't seem intuitively right that someone will choose a higher priced alternative over an identical lower priced alternative, regardless of initial expectations about prices. Can the choice of the cheaper beer be reconciled with Thaler's (1985) transaction theory? One possibility is that the reference price in a transaction is not always the fair or expected price. In a multi-alternative choice, the salience of some competitive alternative, particularly a leading contender, may result in the use of its attribute levels as reference points; in the example above, ph=pg. Transaction utility in this case depends on the price of the purchase compared to the contender's price; the transaction utility for the hotel beer would be V(-ph,-pg). This would imply that transaction utility may sometimes accrue from the particular set of alternatives in which a choice is embedded, as well as from an initial expectation about its price.

A second possible rationale for the choice of the cheaper beer is that when two alternatives with different expected prices are considered together in a choice, a reference point for a category that includes them both is elicited. In the example above, when presented with the choice between the hotel beer and the grocery beer, the reference price for beer - without a particular purchase location, or in its most common purchase location - may be used. Using a common reference point results in the cheaper beer having greater transaction utility.

The issue underlying all of these scenarios is whether two alternatives thrown together in a choice set are still evaluated with respect to their individual expected or fair prices. What variables would promote the use of their original reference prices instead of the price of a competing alternative, or a new, common reference price?

factor may be the degree of similarity between the two alternatives. With identical items (like the beers) the most natural choice heuristic may be a direct comparison of the two on their sole distinguishing characteristic: price. The more dissimilar the two are, the greater may be the tendency to segregate their evaluations, evaluating each ones price with respect to its fair or expected price.

When two alternatives are so distinct that they are categorized into two separate product subclasses, the use of a common reference point or a direct price comparison should seem less appropriate. Consider two choices: the first is between two brands of a high performance stereo component systems; one brand commands a slight price premium. The second choice is between a high performance and 8 low performance system. Although both choice sets involve two different reference prices, the alternatives in the first choice set are more directly comparable. It seems more likely that their prices will be seen in relation to each other or to a common reference price for high performance stereos. The disparity of the two offerings in the second choice may work against such a direct comparison, leading the decision maker to refer to the fair price for each.

A second potential factor in the use of a fair price to evaluate an alternative is how far the price of that alternative diverges from the fair price. The fair price may not be especially salient unless an offered price falls strikingly above or below it, signaling a bargain to pursue or an overcharge to avoid. This would be especially likely if the fair price is actually a range of prices.

A final factor may be the ambiguity of information with which the decision maker is working. Expectations about fair prices should be held with more confidence and used more often when the decision maker understands the product class attributes and how to evaluate them. The greater the ambiguity of the stimulus being evaluated, either from lack of information about it or from the decision maker's inability to assess it, the harder it should be to confidently establish a fair or expected price. One solution would be to evaluate attributes by inter-alternative comparisons, which require fewer absolute judgments. Another strategy would be to use the offering price as a diagnostic cue for quality, which means that the offered price is implicitly evaluated as fair. In this situation, increasing the price of an alternative actually increases the value equivalent of the product.

Multi-attribute Alternatives

In the transaction theory previously stated, the product is seen as a unidimensional entity, whose value equivalent is compared to its price. Marketers are concerned with the attributes that create that value equivalent. Are reference point effects important for attributes other than price? Reference points on other attributes are plausible; Table 2 shows a more general statement of the reference points proposed for price in Table 1. What role might such reference Points play in evaluation and choice?

TABLE 2

PROPOSED TYPES OF REFERENCE POINTS

Let us assume that any attribute level better than its reference point is considered a gain, while any below the reference point is considered a loss. An alternative can therefore be represented as a set of gains and/or losses. These gains and losses can often be either segregated (presented separately: e.g. $100 and $50) or integrated (presented in merged form: e.g. $150). Thaler (1985) proposes that the shape of the value function implies that the evaluation of an alternative is improved if losses are integrated with each other or with larger gains, and gains are segregated from each other and from greater losses. The sticking point in applying this principle to the presentation of multiple attributes is that different attributes are generally expressed in different metrics, which limits the possibility of integration. Dollar losses or gains can be added; four-wheel drive and low mileage cannot be. It has also been suggested that the advisability of segregating or integrating mixed outcomes is affected by the order in which the outcomes are presented (see Ross and Simonson, this session). This would imply that the order in which (necessarily) segregated attributes of an alternative are presented would affect its ultimate evaluation.

Another important aspect-of reference points for multiple attributes is their effect on the choice heuristics used. Siegel (1957) proposed that an examination of the utility function for a set of outcomes (in this case, an attribute's levels) would show that a decision maker's aspiration level is positioned as the upper bound of the part of the function with the steepest slope. (This proposition agrees with prospect theory's value function being steeper for losses than for gains.) Siegel (1957) characterized all outcomes below this aspiration level as unacceptable. Logically, it should be these unacceptable levels that are eliminated in noncompensatory cutoffs. Slein (1983) also presents some evidence that noncompensatory cutoffs are positioned at the point above the steepest slope of the utility function. Changes in the reference point may therefore change the location of noncompensatory cutoffs, which can affect the final choice.

Finally, Yates and Jagacinski (1979) have proposed that the general attractiveness of an alternative's attribute levels may create a reference for the evaluation of any one of those levels. For example, consider two alternatives: A has the worst level on all but one attribute (x); A has the best level on all but x. In their study, varying the value of x had more impact on the evaluation of A than it did on A . This might be explained in terms of prospect theory's steeper value function for losses than for gains. The other attributes of A create a high reference point for x, so that any value on x is perceived as either neutral or as a loss, while for A values of x are perceived as either neutral or as gains. It is an empirical question as to whether the less consistent and extreme attributes of most products in the marketplace create reference effects in the same way.

Sequential Decisions

Decisions involving whether or not to buy a single alternative and choices of one alternative from a set have been discussed. A third type of decision presents an interesting scenario for the observation of reference point effects: one in which alternatives are presented sequentially, with decisions made on one alternative before the next is available for consideration. Examples of these decisions may include house purchases, job decisions, or buying antique collector's items. While similar in many respects to a single alternative decision, sequential decisions accentuate some characteristics. The most striking of these is the uncertainty about the quality (or even existence) of future options, or the future availability of a current option. The reference point used for evaluation may determine the decision maker's willingness to forgo moderately good alternatives in order to achieve potentially greater gains; as in prospect theory, a more ambitious (i.e. more favorable) reference point should lead to greater risk seeking (i.e. longer search).

Another intriguing aspect of this type of sequential decision making is that the generally longer time frame permits the reference point to fully adapt to the levels of the alternatives being presented. The degree and direction of this adaptation should depend on several variables. The first is the content of previously presented alternatives, which should influence the decision maker's expectations about future choices by indicating the general availability of various attribute levels. Trends in the content of alternatives should also affect expectations; one would hypothesize that an improving sequence of alternatives will lead to more optimistic reference points (and longer search) than will a deteriorating sequence. The decision maker's inferences about the causal mechanism underlying the trend would be a major factor moderating this effect.

Controlling the sequence of alternatives might successfully manipulate the reference point in a certain direction. Consider a realtor's strategy in showing houses to 8 new prospect with a relatively low (and invariably unrealistic) reference price for their prospective purchase. Presenting a series of houses in ascending order of price, which allows gradual adaptation to occur, may be more likely to result in adaptation to a higher reference price than if the order was descending or mixed.

Measurement of Reference Points

Studies on framing and reference points have seldom tried to directly measure naturally occurring reference points. In risky choice decisions, zero loss or gain creates a natural reference point for most decision makers. Outcomes can be translated about this reference point in order to study reference point effects (cf. Payne, Laughhunn and Crum 1980). Alternatively, goals can be assigned as reference points, as Payne, Laughhunn and Crum (1981) did. Puto (1986) and Puto, Patton and King (1985) tried a third approach by creating scenarios with several plausible reference points; decision makers identified the reference point they had used from a list of descriptions provided.

The same type of manipulations and measures are possible for studying reference points in consumer choice. For price, frames involving an original versus a discounted price, or an offered versus a manufacturer's suggested retail price are plausible (Liefeld and Heslop 1985; Monroe and Chapman, this session). However, marketers wishing to know what reference point the consumer ordinarily brings into the decision process needs to eventually focus on direct measurement. Consumers' awareness of reference points in decisions probably varies with its use in the decision process. Deliberate comparisons should be more memorable while the use of reference points in an attribute evaluation may be relatively automatic, and less accurately reported. The feasibility of direct measurement needs to be resolved empirically.

SUMMARY

This paper evaluates possible reference point effects on riskless choice outcomes and processes. While reference points have already been shown to have dramatic effects on evaluation and choice (Thaler 1985), more work is needed to assess the strength of their impact within various choice contexts. In particular, one research priority should be to assess reference point effects when a choice is made from a multi-alternative set, a context that may be more typical of consumer behavior than that of a buy-no buy decision for one alternative. It is proposed that a reference point used in such a situation may not always be some previously evolved standard that is retained throughout the choice process. Adaptation of the reference point during the choice process, and the use of a reference point supplied by the choice set itself deserve further investigation.

It has also been shown that the reference point for price can represent a variety of concepts to the consumer. The role of knowledge about the attribute appears to play a significant role in this representation. More powerful tests of the effects of the individual and market characteristics discussed here are needed. Decision makers who vary in the basis for their reference points may also vary in their susceptibility to manipulation of that reference point by marketers. The question of how easy reference points are to manipulate deserves some attention. Changing a reference point based on a long process of assimilating information and forming goals should be a non-trivial proposition. Understanding what the reference point is, and how it affects the evaluation of your offering, may be the highest priority for a marketer.

REFERENCES

Della Bitta, Albert A and Kent B. Monroe (1973) "The Influence of Adaptation Levels on Subjective Price Perceptions", in Ward, Scott and Peter Wright, Eds. Advances in Consumer Research, Volume I, Urbana, IL: Association for Consumer Research, 359-369.

Helson, Harry (1964), AdaPtation-Level Theory, Harper & Row.

Kahneman, Daniel and Amos Tversky (1979) "Prospect Theory: An Analysis of Decision Under Risk", Econometrica, 47, (March) 263-291.

Kahneman, Daniel and Amos Tversky (1984), "Choices, Values, and Frames," American Psychologist, 39 (April), 341-350.

Klein, Noreen M. (1983) "Utility and Decision Strategies: A Second Look at the Rational Decision Maker", Organizational Behavior and Human Performance. 31, 1-25.

Liefeld, John and Louise A. Heslop (1985), "Reference Prices and Deception in Newspaper Advertising," Journal of Consumer Research, 11 (March), 868-876.

Monroe, Kent B. (1984) "Theoretical and Methodological Developments in Pricing", in T. Kinnear (ed.) Advances in Consumer Research. Volume XI. Ann Arbor, MI: Association for Consumer Research, 636-637.

Monroe, Kent B. and Joseph D. Chapman (1987) "Framing Effects on Buyers' Subjective Product Evaluations", in Paul F. Anderson and Melanie Wallendorf, Eds., Advances in Consumer Research. Volume XIV. Ann Arbor, MI: Association for Consumer Research.

Payne, John W., Dan J. Laughhunn and Roy Crum (1980), "Translation of Gambles and Aspiration Level Effects in Risky Choice Behavior", Management Science. 26 (October), 1039-1060.

Payne, John W., Dan J. Laughhunn and Roy Crum (1981), "Further Tests of Aspiration Level Effects in Risky Choice Behavior", Management Science, 27 (August), 953-958.

Puto, Christopher P. (1986, forthcoming) "The Framing of Buying Decisions" Journal of Consumer Research.

Puto, Christopher P., Wesley E. Patton III, and Ronald H. Ring (1986) "Risk Handling Strategies in Industrial Vendor Selection Decisions", Journal of Marketing, 49, (Winter), 89-98.

Rosch, Eleanor (1975), "Cognitive Reference Points,' Cognitive Psychology. 7, 532-547.

Ross, William T. and Simonson, Itamar (1987) "Influences on Consumers' Preferences for Multiple Outcomes", in Paul F. Anderson and Melanie Wallendorf, Eds., Advances in Consumer Research, Volume XIV, Ann Arbor, MI: Association for Consumer Research.

Siegel, Sidney (1957) "Level of Aspiration and Decision Making" Psychological Review. 64, 253-262.

Thaler, Richard (1985), "Mental Accounting and Consumer Choice," Marketing Science. 4 (Summer), 199-214.

Yates, J. Frank and Carolyn n. Jagacinski (1979) "Reference Effects in Multiattribute Evaluations", Organizational Behavior and Human Performance. 24, 400-410.

Zeithaml, Valerie A. and Karen L. Graham (1983), "The Accuracy of Reported Reference Prices for Professional Services", in Richard P. Bagozzi and Alice M. Tybout, Eds. Advances in Consumer Research. Volume X. Ann Arbor, MI: Association for Consumer Research, 607-611.

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