Heuristic Search Processes in Decision Making

ABSTRACT - Research in psychology indicates that individuals employ heuristics in performing a variety of complex tasks. Evidence is presented that consumers also utilize heuristics to reduce the amount of information searched in making a decision. Aspects of theory and methodology of cognitive psychology which should prove valuable to the study of consumer behavior are illustrated.


John W. Payne (1976) ,"Heuristic Search Processes in Decision Making", in NA - Advances in Consumer Research Volume 03, eds. Beverlee B. Anderson, Cincinnati, OH : Association for Consumer Research, Pages: 321-327.

Advances in Consumer Research Volume 3, 1976      Pages 321-327


John W. Payne, University of Chicago

[This research was supported by Research Grant MH25788 from the National Institute of Mental Health.]


Research in psychology indicates that individuals employ heuristics in performing a variety of complex tasks. Evidence is presented that consumers also utilize heuristics to reduce the amount of information searched in making a decision. Aspects of theory and methodology of cognitive psychology which should prove valuable to the study of consumer behavior are illustrated.

The purpose of this paper is to present evidence that consumers employ heuristic processes as a way of reducing the amount of information they have to search and evaluate in making a decision. The use of heuristic processes in the performance of complex tasks has been found in studies of such different areas of psychology as human problem solving, visual perception, and decision making under risk. In the first part of this paper, i will briefly review some of that research. The results of a recent study that examined the use of heuristic search strategies in a more consumer-oriented decision task will then be presented. Finally, I will present an example of a contingent information process-inn model based on the behavior revealed in that study. An underlying theme of this paper is the need for a closer integration of research in consumer behavior with the theory and methodology of more traditional areas of cognitive psychology.


Perhaps the most important generalization to come out of efforts to study human information processing is that an individual is a limited information processing system (Newell and Simon, 1972). In particular, the active processing of information occurs in a memory of limited capacity, duration, and ability to place information in more permanent storage. As a result, people appear to utilize heuristics that serve to keep the information processing demands of a task within the bounds of their limited cognitive capacity. Heuristic processes can be defined as

problem-solving methods which tend to produce efficient solutions to difficult problems by restricting the search through the space of possible solutions. The restriction on search is based on evaluation of the structure of the problem (Braunstein, 1976).

The same restriction on search which increases efficiency may, at times, result in individuals ignoring or misusing information in reaching a judgment or achieving a solution to a problem. Heuristic processes, in other words, are procedures used by individuals which sacrifice the certainty of a correct judgment for increased efficiency in the process.

Empirical support for the concept of heuristic processing has been found in studies in a wide variety of areas cf cognitive psychology. For example, clear illustrations of the use of heuristic processes by individuals can be found in studies of human problem solving (Newell & Simon, 1972), pattern recognition, depth perception (Braunstein, 1976), probability estimation, and risky decision making. In the next two sections, I will briefly review some of the evidence for heuristic processing in two of these areas of cognitive psychology. More intensive discussions of the use of heuristics by individuals may be found in Newell and Simon (1972) and Braunstein (1976).

Probabilistic Information Processing

Some of the most striking evidence for the use of heuristics by individuals has been provided by Kahneman and Tversky in their studies of probabilistic information processing. In the case of probability estimation, several recent studies have demonstrated that humans seem to follow a judgmental heuristic called "representativeness'' when faced with tasks involving intuitive prediction (1972, 1973). The idea is that people predict the outcome that appears most representative of the evidence. While this heuristic often leads to correct judgments, it can lead to large and consistent biases that are quite difficult t o eliminate. The reason is that there are factors such as the reliability of evidence and the prior probability of the outcome which affect the likelihood of the outcome but not its representativeness. For example, Kahneman and Tversky (1973) have given subjects a brief personality description allegedly sampled at random from a group of 100 professionals--engineers and lawyers. In contrast to the normative expectation, the subjects rated the probability of that person being an engineer rather than a lawyer the same regardless of whether they were told that the group consisted of 70 engineers and 30 lawyers, or 30 engineers and 70 lawyers. Apparently, subjects considered the degree to which the descriptions were representative of the two stereotypes, lawyers or engineers, with relatively little regard for prior probabilities in making their judgments.

Tversky and Kahneman (1974) have discussed two other heuristics which appear to be important in probabilistic information processing. The first is a heuristic called "availability." According to the availability heuristic, one judges "the frequency of a class or the probability of an event by the ease with which instances or occurrences can be brought to mind" (p. 1127). Slovic, Fischhoff, and Lichtenstein (1976) report an experiment illustrating how the availability heuristic may lead to important biases in people's perceptions of low-probability, high-consequence events. Bias may also occur when people attempt to ease the strain of processing information through the use of a heuristic called "anchoring and adjustment." This heuristic operates by an individual selecting a natural starting point or anchor in the task environment to be used as a first approximation to the judgment. This anchor is then adjusted to accommodate the implications of additional information. Work by Tversky and Kahneman (1974) and Lichtenstein and Slovic (1971) demonstrates that typically the adjustment is crude and imprecise and fails to do justice to the importance of additional information.

Decision Making Under Risk

Research on how people choose among alternative gambles provides additional evidence that individuals tend to use heuristic processes. The use of heuristics in risky decision making is especially interesting since there exist rather straightforward computational algorithms which guarantee an optimal solution. Risky decision behavior is also highly relevant to consumer choice behavior. Both involve decisions about multidimensional alternatives. In addition, there is a substantial literature relating perceived risk to consumer behavior (Ross, 1975 ).

In the typical risky decision making task, an individual is faced with two alternative gambles. The probability of winning, probability of losing, amount to be won, and amount to be lost are displayed for each gamble. In such situations, people appear to choose among gambles on the basis of a direct comparison among these displayed values, rather than on the basis of maximizing expected value. While choices based on the heuristic use of the displayed probabilities and amounts usually correlate well with the choices which would be obtained with expected value maximization, or at least with the maximization of a function of subjective transformations of probability and value, there are situations in which these heuristics lead to preferences which fail to correspond to those expected from either objective or subjective maximization functions. It has been shown, for example, that people will exhibit consistent preferences among gambles which differ in displayed probabilities and amounts, even when the gambles are mathematically identical and the choice of one over the other can have no objective consequence (Payne and Braunstein, 1971). It has also been shown that the use of displayed probabilities and amounts is not merely due to ignorance of alternative procedures or difficulty in making the required computations. These tendencies persist even after subjects are instructed in the meaning of expected value and given the expected value of each alternative gamble (Lichtenstein, Slovic, and Zink, 1969). Additional evidence for the use of heuristics in risky decision ma-king may be found in studies by Lichtenstein and Slovic (1971), Payne (1975), and Tversky (1969, 1972).

An example of a model of risky decision behavior based on heuristic processes was proposed by Payne and Braun-stein (1971). This model is of interest, in part, because it assumes that individuals may search and utilize only part of the information available about the alternative gambles. The hypothesized decision strategy was presented in the form of a process model involving two stages: an evaluation stage followed by a choice stage. The particular choice rule used by the decision maker was assumed to be contingent upon the outcome of the evaluation process. Specifically, the model assumed that the decision maker initially considers the probability relationship within the gambles in a pair. If the probability to win is less than the probability to lose, a choice rule based on selecting the gamble with the lesser probability to lose is evoked. If the probability to win is greater than the probability to lose within each gamble in the pair, then an attempt is made to maximize the amount to Win. In the case where the amounts are equal within a pair, the probability to win may be used as a secondary criterion, and the gamble with the greater probability to win chosen. For pairs of gambles where the probability to win equals the probability to lose, there are alternative paths which can be taken in the model, possibly corresponding to two types of subjects. These types of pairs of gambles may be treated as if they were pairs of gambles in which the probability to win is greater than the probability to lose, or as if they were pairs of gambles in which the probability to win was less than the probability to lose. While the results that this model was developed to explain were obtained in a study using special types of pairs of gambles, the form of the model provides a clear illustration of a heuristic decision strategy. Based upon an evaluation of the structure of the problem, the decision maker is assumed to restrict his search and use of information to just one or two dimensions of the gambles in making his decision.

The research reviewed here in probabilistic information processing and risky decision making is consistent with the generalization reached by Slovic and Lichtenstein (1971) in their comprehensive review of Bayesian and regression studies of human judgment:

We find that judges have a very difficult time weighting and combining information--be it probabilistic or deterministic in nature. To reduce cognitive strain, they resort to simplified decision strategies, many of which lead them to ignore or misuse relevant information (p. 115 ).

In terms of problem solving, Simon and Newell (1971) have proposed that the central process in human problem solving is the use of heuristic methods to carry out highly selective searches of problem spaces.


The processing of information by consumers has recently received increased attention. One finding of this research is that "the amount of information sought is typically small relative to the amount of information available" (cf. Jacoby, 1975 ). Haines (1974) has advanced a "Principle of Information Parsimony." The idea is that "consumers seek to process as little data as is necessary in order to make rational decisions" (p. 96). This principle seems entirely consistent with results obtained in other areas of cognitive psychology. Although the work of Simon (1957) suggests that the term "rational" be replaced by the term "satisfactory," the interesting question then becomes "how do decision makers go about the process of reducing the amount of information sought in order to make a decision?" Haines (1974) suggests a partial answer when he argues that "it is important to understand that people do take advantage of patterns in the task environment to reduce information processing"(p. 96). The definition of a heuristic procedure in problem solving given earlier incorporates the similar idea that the restriction of search through the space of possible solutions will be based on an evaluation of the structure of the problem (Braunstein, 1976). It is the central thesis of this paper that consumers utilize heuristic strategies in making decisions as a way of reducing the amount of information they must search and evaluate.

In the next section of the paper, I will briefly review the results of a recent study which show that individuals, when faced with the problem of choosing a preferred alternative in a sufficiently complex situation, will adopt heuristic decision strategies which produce highly selective searches among the sets of possible alternatives. An example of a computer-based process model that attempts to simulate the behavior revealed in the study will then be presented. The model illustrates the use of a heuristic decision procedure to simplify a decision task.


Research on problem solving has shown individual behavior to be highly adaptive to the demands of the task. This suggested that the information processing procedures used by decision makers might be systematically related to certain characteristics of the decision situation. The study discussed below used two process tracing techniques: (1) explicit information search and (2) verbal protocols, to make clear the effects of two variations in the complexity of a decision task; (1) number of alternatives and (2) number of dimensions, on the information processing strategies subjects use in reaching a preferential choice. It was hypothesized that increases in the complexity of a decision situation would result in decision makers resorting to choice heuristics in an effort to reduce cognitive strain. Two experiments were conducted. (The interested reader is referred to Payne, 1976 for a complete report on this study. )


The method in both experiments was very similar. The second experiment was an extension and replication of the first experiment. Six subjects who were paid college students participated in the first experiment (12 subjects in the second). The stimuli were "information boards" representing different one bedroom furnished apartments. An information board consisted of a number of envelopes containing cards labeled with the name of a dimension (attribute), e.g., "noise level." To obtain the value of that dimension for a particular alternative, the decision maker had to pull the card out of the envelope, turn it around and place it back into the envelope. The information about the value of the dimension was on the back of the card, e.g., "noise level--low." Once a card was turned over, the value of the dimension on the particular alternative was clearly displayed for the remainder of the choice problem.

A decision situation for a subject involved a number of alternatives, either 2, 6, or 12 (2, 4, 8, or 12 in the second experiment), and a number of dimensions of information available per alternative, either 4, 8, or 12. There were three levels of value on such dimension. The dimensional values of an alternative were selected so that each alternative would have a priori both good and poor qualities.

The procedure was very simple. Each subject was run individually in one hour sessions. The subjects were told that they would be presented with a number of alternatives to choose among and a certain amount of information about each alternative. The subject was not instructed in how much of the information he or she had to use in reaching a decision. No time constraints were placed on the subjects. They were instructed to work at their own pace and that they should have plenty of time to finish.

In addition to instructions regarding the use of the information boards, the subjects were also instructed to "think aloud" while making their decision. Verbal protocols have proved useful as the basis for models of human behavior in the performance of a variety of tasks (cf. Newell & Simon, 1972).

Results (First Experiment)

A complete transcript of the verbal reports given by each subject was made. Consistent with the procedure suggested by Newell and Simon (1972), the protocols were broken up into short phrases. During the analysis of the results, excerpts from the protocols of the six subjects will be presented. The complete protocols may be obtained from the author.

The search data for each subject was primarily organized in terms of amount of available information search per alternative in a choice set and the pattern of search, interdimensional or intradimensional. The pattern of search was determined by examining the alternative and dimension associated with the nth + 1 piece of information searched by a subject as a function of the nth piece of information search. Details on how the pattern of search was determined are available in Payne (1976). Similar analyses of single-step transition search data can be found in Bettman and Jacoby (1975).

These two characteristics of the search data, amount of information searched per alternative and pattern of search, proved useful in discriminating between four alternative models of decision making which were of particular interest in this study. The four models were: additive, additive difference, conjunctive, and elimination-by-aspects. Each of these models implies, at least in its most common forms, different information search processes.

An additive decision strategy implies an interdimensional pattern of search and a constant amount of search per alternative. The constant pattern of search is implied because the additive model assumes that each alternative in a choice set is evaluated separately and that the comparison process involves only the overall values that have been determined for each alternative. A decision maker following a strict additive difference model would also have to search a constant amount of information per alternative, but would search in an intradimensional fashion. Both the conjunctive and elimination-by-aspects models, on the other hand, imply the possibility of a decision maker using a variable search pattern. For example, the elimination-by-aspects model (Tversky, 1972), describes choice as a covert elimination process. In choosing among multidimensional alternatives, the individual is assumed to proceed in the following manner. A dimension or aspect is selected. Then all the alternatives that do not possess that dimension or aspect are eliminated. Consequently, those alternatives eliminated early would have only a few (one) dimension searched. The elimination procedure is repeated until all but one of the alternatives are eliminated. Those alternatives eliminated late in the choice process would have more dimensions searched. A conjunctive decision strategy would also result in a pattern of search involving a minimum of one dimension examined on some alternatives (those unsatisfactory on the initial dimension searched) up to an examination of all the relevant dimensions of the preferred alternative. Both the conjunctive and elimination-by-aspects processes are examples of heuristics by which people seek to reduce the amount of information processing involved in complex decision making. The two models differ, however, in whether they imply an interdimensional (conjunctive) or intra-dimensional (elimination-by-aspects) search strategy.

The information search pattern exhibited by each subject was determined for each of the different decision situations. The results indicated that the information processing leading to a preferential choice will vary as a function of task complexity. The most important determinant of complexity examined was clearly the number of alternatives available. Table i shows the classification of the search patterns for each subject as a function of the number of alternatives available.

It is clear from Table I that when faced with a more complex decision task, either six or twelve alternatives, subjects employed decision strategies which re-suited in a variable amount of information searched across alternatives. This supports the hypothesis that increases in the complexity of a decision situation will result in decision makers resorting to choice heuristics in an effort to reduce cognitive strain. As mentioned, the use of an elimination-by-aspects choice process by a decision maker represents an example of such a heuristic.



The verbal protocols obtained provide further evidence that subjects tended to adopt decision strategies which would eliminate some of the available alternatives as quickly as possible and on the basis of a limited amount of information search and evaluation. For example, consider this excerpt from the protocol of subject 1:

A162: Apartment E.

A163: The rent for apartment E is $140,

A164: which is a good note.

A165: The noise level for this apartment is high.

A166: That would almost deter me right there.

A167: Ah, I don't like a lot of noise.

A168: And, if it's high, it must be pretty bad.

A169: Which means, you couldn't sleep.

A170: I would just put this aside right there. I wouldn't look any further than that.

Another explicit example of the elimination of alternatives after a limited search and evaluation is provided by this example from the protocol of subject h:

D289: Since we have a whole bunch here,

D290: I'm going to go across the top and

D291: see which noise levels are high

D292: If there are any high ones,

D293: I'll reject them immediately.

. . .

D297: Go to D.

D298: It has a high noise level.

D299: So, we'll automatically eliminate D.

Results (Second Experiment)

The results obtained in the second experiment were similar to those obtained in the first experiment. The conclusion that as the number of available alternatives increases, decision makers shift from decision strategies involving a constant amount of search per alternative, e.g., compensatory procedures, to decision strategies which involved eliminating some alternatives on the basis of only a few dimensions, e.g., conjunctive or elimination-by-aspects procedures, was supported. A three-way analysis of variance (number of alternatives available, amount of information available per alternative, and subjects) showed that the amount of variation in percentage of available information searched increased as the number of alternatives increased, F(3,33) = 12.36, (p < .01). The main effect of amount of information per alternative, F(2, 22 ) = .78, and the interaction, F(6,66) = 1.10, were not significant.

Stronger support for the use of a strict elimination-by-aspect process by decision makers was obtained by calculating the number of search patterns that showed not only a variable and intradimensional pattern but also a pattern where all (remaining) alternatives would be completely searched on a dimension before some subset of those alternatives would be searched on another dimension. Thirty-three of the 108 multi-alternative search patterns were consistent with a strict elimination-by-aspects decision process. For 99 of the 108 multi-alternative choice situations (nine for each subject), the amount of available information searched was as great as or greater for the alternative chosen than for any other alternative in the choice set. In the remaining nine situations, the preferred alternative had next to the maximum amount of information searched.


It is clear from the results of the two experiments just discussed that individuals will employ decision strategies resulting in a restricted pattern of information search when the choice task becomes sufficiently complex. These decision strategies can be characterized as heuristic processes similar to those found in studies of human problem solving. Evidence for the use of heuristics has been found in another recent study of information seeking behavior in multidimensional choice (Russo & Dosher, 1975).

Important individual differences in information processing also were shown to exist in both the experiments. In particular, some of the subjects tended to search for information in an interdimensional fashion, others in an intradimensional fashion. Bettman and Jacoby (1975) also found such individual differences.

It was suggested in Payne (1976) that one possible explanation for differences in information search may be in how the decision maker represents the knowledge he acquires about the alternatives in the decision task. For example, a decision maker might store information about the decision alternatives in the following form: Apartment A (rent, $140). This object (property, value) representation would suggest that an individual might find it easier to search and evaluate (store) information within an alternative and across dimensions. On the other hand, a decision maker might choose to code information in terms of rent (Apartment A, $140). This property (object, value) representation would suggest that an intradimensional form of processing might be easier for an individual. This type of knowledge representation has been widely used in building theories of human cognitive processes (Simon and Newell, 1974). In the context of consumer choice, Calder (1975) has also pointed out that the structure of decision processes may depend upon how information is coded in memory. Bettman and Jacoby (1975) present a further discussion of the implications of individual differences in memory storage for consumer information processing. Research aimed at relating an individual's pattern of search to any preferences that individual might have for encoding information in memory is currently being undertaken.


The results of the information search experiments presented above indicate: (1) the same individual will process information in different ways as a function of simple task variations; and (2), that different individuals will process information in different ways in the same decision task. Both these findings support conclusions reached by Haines (1974). However, there was also evidence of different processes being used by the same individual within a single decision situation. Table 2a presents excerpts from the protocol of subject 2 in the first experiment when faced with a 12 alternative choice situation.



The protocol clearly shows this decision maker initially using a strict elimination-by-aspects process to eliminate alternatives. Support for this conclusion was also obtained from a detailed analysis of the single-step transition data in the individual's pattern of search. Notice how this decision maker appears to reduce the choice problem from 12 alternatives to eight alternatives, and eventually to just a pair of alternatives. At that point, the protocol shows the decision maker shifting from an elimination-by-aspects procedure to what appears to be an additive difference strategy. Eventually, after directly comparing the final two choice alternatives, the decision maker chose apartment J. Other subjects also give indications of similar combinations of decision processes. Einhorn (1971) suggested a similar explanation of decision making in complex situations. See also Wright (1974).

One way to conceptualize the four decision processes examined in the previous two studies would be as different subroutines in a general choice program. The control conditions under which one of these sets of processes might be called would seem to depend, at least in part, on the characteristics of the decision problem. In that respect, the less cognitively demanding decision procedures, conjunctive and elimination-by-aspects, might be called early in the decision process as a way of simplifying the decision task by quickly eliminating alternatives until only a few alternatives remained as choice possibilities. The decision maker might then employ one of the cognitively more demanding choice procedures, e.g., additive difference model, to make the final evaluations and choice.


A contingent process model that attempts to simulate the behavior revealed by subject 2 was developed. The model is based on assumptions derived from the analysis of the verbal report of subject 2 excerpts of that report are given in Table 2a). The model encompasses three categories of behavior: (1) a control process that selects one of two decision procedures on the oasis of an evaluation of the complexity of the decision task, (2) an elimination of alternatives procedure, and (3) a compensatory comparison procedure:

C0: Evaluate complexity of decision task by determining number of alternatives in List-available alternatives.

C1: If the number of alternatives is greater than two then E0.

C2: If the number of alternatives is exactly two then D0.

C3: If the number of alternatives is less than two then R0.

E0: Eliminate alternatives by dimensions.

E1:Search List-dimensions for first (next) most important dimension (D), and set goal--Eliminate by dimension (D), if end of list then R0.

E2: Search List-available-alternatives for first (next) alternative (A), if end of list then C0.

E3: Search external environment for dimension (D, value) for alternative (A).

E4: Search memory for List-acceptable- dimension (D, values).

E5:Determine if dimension (D, value) for alternative (A) on List-acceptable (D, values). If value on List then E2, else mark alternative (A) eliminated and remove from List-available-alternatives.

E6: Go to E2.

D0: Direct Comparison Process.

D1:Search List-available-alternatives for first alternative (X) and next alternative (y), and set goal--comparison (X) and (Y).

D2: Search List-dimensions for last (next) most important dimension--D, if end of list then DS.

D3:Search memory for dimension (D, value) for alternative (X) and alternative (y), if not available then search external environment for dimension (D, value) for alternative (X) and for alternative (y).

D4:Compare dimension (D, value) for alternative (X) and dimension (D, value) for alternative (y) with ordered List-acceptable dimension (D, values).

If alternative (X) ordered higher than alternative (y) on dimension (D), then respond alternative (X)--better, and increment overall worth of alternative (X) by relative importance value of dimension (D). Else if alternative (y) ordered higher than alternative (X) on dimension (D), then respond alternative (y) better, and increment overall worth of alternative (Y) by relative importance value of dimension (D). Else respond alternatives equal and go to D,.

D5: Compare overall worth of alternative (X) with overall worth of alternative (Y).

If worth of alternative (X) greater, then respond alternative (X) preferred. Else if worth of alternative (Y) greater, then respond alternative (Y) preferred. Else R~.

R0: Respond--No choice possible.

The model has been coded in BASIC. The output (trace) of the program corresponding to the parts of the verbal protocol given in Table 2a is presented in Table 2b.

There are a number of ways of comparing the behavior of an individual with the output of a computer-based model that attempts to simulate that behavior. At the summary level of final choice, the model presented here selected the same alternative out of the 12 possible that the decision maker selected. This is, of course, the first and most necessary property of such a model. However, the real value of a process (computer) model is in the ability of the model to account for the sequential aspects of the decision maker's behavior.

An analysis of the verbal protocol of subject 2 indicated four major types of statements: (1) goal statements, e.g., "Let's just see what the rents are in all the apartments first,"(2) statements reflecting search, e.g., "The rent of A is $140," (3) evaluation statements, e.g., "Um, $170 is too much," and (4) comparison statements, e.g., "In J the rooms are larger." The trace shown in Table 2b also contains four types of statements which correspond to those contained in the subject's protocol. A comparison of the sequence of goal statements, observations, evaluative statements, and comparative statements contained in the protocol and trace presented in Table 2 suggests that the correspondence between the subject's behavior and the model is reasonable. More formally, it is possible to construct Problem (Decision) Behavior Graphs (Newell and Simon, 1972) based on the decision maker's protocol and on the trace of the model. One way to analyze the correspondence between the two Decision Behavior Graphs would be in terms of the sequential dependencies among the various types of statements (behavior). That is, how often was one type of behavior followed by another type of behavior. This analysis is based on a suggestion for the analysis of production systems presented by Newell (1966). Table 3 presents the results of such a sequential analysis. The model appears capable of providing sequential behavior which corresponds closely to the actual sequential behavior of the subject.

Unfortunately, space limitations make it impossible to expand upon this comparison of the model trace with the subject's protocol. A more intensive treatment of the protocol and model presented here would, for example, give a detailed examination of the exact information being searched and evaluated at each point in time in both the protocol and the trace. However, it should be noted that the heuristic processing form of this model is consistent with research in other areas of behavior (cf. Newell and Simon, 1972). Examples of similar types of process models dealing with consumer choice are provided by Bettman (1970) and Haines(1974). For a more complete discussion of the role of verbal protocol analysis in decision making research, see Payne (1974). See also Bettman (1974) for several interesting suggestions regarding procedures for formally analyzing simulation models.




The emphasis in this paper has been on relating the study of consumer information search and decision making to certain theoretical concepts and methodological techniques from areas of cognitive psychology, such as human problem solving. Of particular concern was applying the concept of heuristic processes to explain the limited amount of information search and evaluation consumer decision makers appear to employ when faced with complex choice problems. Evidence supporting the use of heuristic decision strategies was found in a review of studies of complex judgment in two areas of cognitive psychology closely related to consumer behavior, and in the results of two experiments specifically aimed at investigating information search in a complex decision task. It appears that consumers do utilize heuristics to reduce the amount of information they have to process in making a decision.


James R. Bettman, "Information Processing Models of Consumer Behavior," Journal of Marketing Research, 7 (August, 1970), 370-376.

James R. Bettman, "Toward a Statistics for Consumer Decision Net Models," Journal of Consumer Research, (June, 1974), 71-80.

James R. Bettman and Jacob Jacoby, "Patterns of Processing in Consumer Information Acquisition," Working Paper No. 31, Center for Marketing Studies, University of California, Los Angeles, 1975.

Myron L. Braunstein, Depth Perception Through Motion (New York: Academic Press, 1976).

Bobby J. Calder, "The Cognitive Foundations of Attitudes: Some Implications for Multi-Attribute Models," in M. J. Schlinger, ed., Advances in Consumer Research, Volume II (Chicago: Association for Consumer Research, 1975), 241-247.

George H. Haines, Jr., "Process Models of Consumer Decision Making," in G. D. Hughes and M. L. Ray, eds., Buyer/Consumer Information Processing (Chapel Hill, NC: University of North Carolina Press, 1974).

Jacob Jacoby, "Consumer Psychology: An Octennium," Annual Review of Psychology, 27 (1976), in press.

Daniel Kahneman and Amos Tversky, "On the Psychology of Prediction," Psychological Review, 80 (1973), 237-251.

Daniel Kahneman and Amos Tversky, "Subjective Probability: A Judgment of Representativeness," Cognitive Psychology, 3 (1972), 430-454.

Sarah Lichtenstein and Paul Slovic, "Reversals of Preference Between Bids and Choices in Gambling Decisions," Journal of Experimental Psychology, 89 (1971), 46-55.

Saran Lichtenstein, Paul Slovic, and Donald Zink, "Effect of Instruction in Expected Value on Optimality of Gambling Decisions," Journal of Experimental Psychology, 79 (1969), 236-240.

Allen Newell, "On the Analysis of Human Problem Solving Protocols," Department of Computer Science Paper, Carnegie-Mellon University, Pittsburgh, Pa., 1966.

Allen Newell and Herbert A. Simon, Human Problem Solving (Englewood Cliffs, N. J.: Prentice-Hall, 1972).

John W. Payne, "A Process Tracing Study of Risky Decision Making: Examples of Protocols and Comments," Complex Information Processing Working Paper No. 274, Carnegie-Mellon University, Pittsburgh, Pa., 1974.

John W. Payne, "Relation of Perceived Risk to Preferences Among Gambles," Journal of Experimental Psychology: Human Perception and Performance, 1 (1975), 86-94.

John W. Payne, "Task Complexity and Contingent Processing in Decision Making: An Information Search and Protocol Analysis," Organizational Behavior and Human Performance (1976), in press.

John W. Payne and Myron L. Braunstein, "Preferences Among Gambles with Equal Underlying Distributions," Journal of Experimental Psychology, 87 (1971), 13-18.

Ivan Ross, "Perceived Risk and Consumer Behavior: A Critical Review," in M. J. Schlinger, ed., Advances in Consumer Research, Volume II (Chicago: Association for Consumer Research, 1975), 1-19.

J. Edward Russo and Barbara Anne Dosher, "Dimensional Evaluation: A Heuristic for Binary Choice," unpublished manuscript, Department of Psychology, University of California, San Diego (August, 1975).

Herbert A. Simon, Models of Man (N.Y.: Wiley, 1957).

Herbert A. Simon and Allen Newell, "Human Problem Solving: The State of the Theory in 1970," American Psychologist, 26 (1971), 145-159.

Herbert A. Simon, "Thinking Processes," in D. H. Krantz, R. C. Atkinson, R. D. Luce, and P. Suppes, eds., Contemporary Developments in Mathematical Psychology, Volume 1 (San Francisco: Freeman, 1974).

Paul Slovic, Baruch Fischhoff, and Sarah Lichtenstein, "Cognitive Processes and Societal Risk Taking," in J. S. Carroll and J. W. Payne, eds., Cognition and Social Behavior (Potomac, Md.: Lawrence Erlbaum Associates, 1976).

Paul Slovic and Sarah Lichtenstein "Comparison of Bayesian and Regression Approaches to the Study of Information Processing in Judgment," Organizational Behavior and Human Performance, 6 (1971), 649-74a.

Amos Tversky, "Elimination by Aspects: A Theory of Choice," Psychological Review, 79 (1972), 281-299.

Amos Tversky, "Intransitivity of Preferences," Psychological Review, 76 (!969), 31-48.

Amos Tversky and Daniel Kahneman, "Judgment Under Uncertainty: Heuristics and Biases," Science, 185 (1974), 1124-1131.

Peter Wright, "The Use of Phased, Noncompensatory Strategies in Decisions between Multi-Attribute Products," Research Paper 223, Graduate School of Business, Stanford University, 1974.



John W. Payne, University of Chicago


NA - Advances in Consumer Research Volume 03 | 1976

Share Proceeding

Featured papers

See More


Moral Arguments Are Most Persuasive in Changing Attitudes of Opponents of Genetically Modified Foods

Sydney Scott, Washington University, USA
Yoel Inbar, University of Toronto, Canada
Paul Rozin, University of Pennsylvania, USA

Read More


To Trace is to Trust: From Product Traceability to Brand Trust

Jing Wan, University of Groningen, The Netherlands
Pankaj Aggarwal, University of Toronto, Canada
Min Zhao, Boston College, USA

Read More


The Ritualistic Dimension of Microlending

Domen Bajde, University of Southern Denmark, Denmark
Pilar Silveira Rojas Gaviria, Pontificia Universidad Católica de Chile

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.