Image Theory: an Alternative to Normative Decision Theory

Lee Roy Beach, University of Arizona
ABSTRACT - Research and theory about how people make decisions has changed markedly over the years. The present article describes the course of that change, with focus on a new descriptive theory called Image Theory. The theory's basic structure is described and the research that it has motivated is discussed. Implications for consumer research are examined.
[ to cite ]:
Lee Roy Beach (1993) ,"Image Theory: an Alternative to Normative Decision Theory", in NA - Advances in Consumer Research Volume 20, eds. Leigh McAlister and Michael L. Rothschild, Provo, UT : Association for Consumer Research, Pages: 235-238.

Advances in Consumer Research Volume 20, 1993      Pages 235-238


Lee Roy Beach, University of Arizona


Research and theory about how people make decisions has changed markedly over the years. The present article describes the course of that change, with focus on a new descriptive theory called Image Theory. The theory's basic structure is described and the research that it has motivated is discussed. Implications for consumer research are examined.

For the past forty years, research on individual decision making has been guided by normative models of choice, largely borrowed from economics. Half of that time has been spent demonstrating the failure of these models as credible descriptions of how people actually make decisions (Beach & Lipschitz, in press). In spite of the mounting evidence that normative models fail at the level of individual decision making, there has been a reticence to relinquish them.

There are, however, fresh winds blowing. Alternatives to the economic model of decision making are being developed (see Klein, Orasanu, Calderwood & Zsambok, in press). The purpose of the present paper is to present the alternative that has thus far received the greatest degree of theoretical development and the largest amount of empirical research, image theory (Beach, 1990; Beach & Mitchell, 1987, 1990; Beach, Mitchell, Paluchowski & van Zee, 1992).



To begin, image theory assumes that decision makers use three different schematic knowledge structures to organize their thinking about decisions. These structures are called images, in deference to Miller, Galanter, and Pribram (1960), whose work inspired image theory. The first of the three is the value image, the constituents of which are the decision maker's principles. These are the imperatives for his or her behavior or the behavior of the organization of which he or she is a member and serve as rigid criteria for the rightness or wrongness of any particular decision about a goal or plan. Principles serve to internally generate candidate goals and plans for possible adoption, and they guide decisions about externally generated candidate goals and plans.

The second image is the trajectory image, the constituents of which are previously adopted goals. This image represents what the decision maker hopes he, she or the organization will become and achieve. Goals can be concrete, specific events (getting the money to buy a new Honda Accord DX) or abstract states (achieving a successful career). The goal agendum is called the trajectory image to convey the idea of extension, the decision maker's vision of the ideal future.

The third image is the strategic image, the constituents of which are the various plans that have been adopted for achieving the goals on the trajectory image. Each plan is an abstract sequence of potential activities beginning with goal adoption and ending with goal attainment. One aspect of plans, their concrete behavioral components, are tactics. Tactics are specific, palpable actions that are intended to facilitate implementation of an abstract plan to further progress toward a goal. The second aspect of plans is forecasts. A plan is inherently an anticipation of the future, a forecast about what will happen if certain classes of tactics are executed in the course of plan implementation. However, it need not be inflexibleCit can change in light of information about the changing environment in which implementation is (or might be) taking place. Therefore, it serves both to guide behavior and to forecast the results of that behavior. By monitoring these forecasts in relation to the goals on the trajectory image, the decision maker can evaluate his or her progress toward realization of the ideal agendum on the trajectory image.

Two Kinds of Decisions, Two Decision Tests

There are two kinds of decisions, adoption decisions and progress decisions. These decisions are made using either or both of two kinds of decision tests, the compatibility test or the profitability test.

Adoption decisions also can be divided into two different kinds, screening decisions and choice decisions. Adoption decisions are about adoption or rejection of candidate goals or plans as constituents of the trajectory or strategic images. Screening consists of eliminating unacceptable candidates. Choice consists of selecting the most promising from among the survivors of screening.

Progress decisions consist of assaying the fit between the forecasted future if implementation of a given plan is continued (or if a particular candidate plan were to be adopted and implemented) and the ideal future as defined by the trajectory image. Incompatibility triggers rejection of the plan and adoption of a substitute (often merely a revision of the old plan that takes into consideration feedback about the environment). Failure to find a promising substitute prompts reconsideration of the plan's goal.

The compatibility test describes adoption decisions as deriving from the compatibility (or incompatibility) between the candidate and standards that are defined by the three images. Actually, the focus is upon lack of compatibility in that a candidate's compatibility decreases as a function of the weighted sum of the number of its violations of the standards that derive from the images, where the weights reflect the importance of the standard (Beach, Smith, Lundell & Mitchell, 1988; Beach & Strom, 1989; van Zee, Paluchowski & Beach, 1992). Violations are defined as negations, contradictions, contraventions, preventions, retardations, or any similar form of interference with the actualization of one of the images' constituents. Each violation is all-or-none (-1 or 0). The decision rule is that if the weighted sum of the violations exceeds some absolute rejection threshold, the candidate is rejected, otherwise it is adopted. The rejection threshold is that weighted sum above which the decision maker regards the candidate as incompatible with his, her, or the organization's principles, goals, and ongoing plans.

The formal statement of the compatibility test is:


where Vcs = 0 or -1, and .00_W_1.0.

That is, incompatibility, I, is zero when an option violates no principles and increases (becomes increasingly negative) as the number of violations increases. In the equation, c is a characteristic of an option, s is a standard, V is a violation of a standard s by characteristic c and counts as a -1, and W is the weight for each standard; W is between and including .00 and 1.00. Because violations count only as -1, incompatibility is measured as the negative sum of the importance weights of the violated standards; a particular characteristic can violate more than one standard (Beach, 1990; Beach and Mitchell, 1987; 1990).

The decision maker's tolerance of incompatibility is described theoretically as the rejection threshold. This is the critical value of I that the decision maker will tolerate before screening out an option. Because it is influenced by many factors (Beach, 1990), the critical value of I may vary from one decision to the next. Therefore, in experiments the critical value of I, the rejection threshold, must be inferred from data. Nonetheless, as we shall see, I is a theoretically viable construct, the crucial role of which in screening can be empirically demonstrated.

The compatibility test describes progress decisions as deriving from the compatibility (or incompatibility) between the trajectory and strategic images. In this case violations are of the trajectory image's constituents by the strategic image's constituents (its forecasts). The decision rule is that when the weighted sum of violations exceeds the rejection threshold, reevaluation of the plan that generated the forecast is undertaken and the faulty plan is replaced. Note that the compatibility test serves both adoption and progress decisions.

The profitability test describes choices from among the survivors of screening by the compatibility test. Unlike the compatibility test, the profitability test is not a single mechanism. Instead, it is a short-hand term for the unique repertory of choice strategies (Beach & Mitchell, 1978) that the individual decision maker possesses for adopting the potentially most profitable candidate from among a set of two or more candidates, all of which are at least minimally acceptable. The minimal acceptability of the adoption candidates from among which the choice is to be made is assured by the prior application of the compatibility test. In short, the profitability test is a 'tie breaker' when more than one adoption candidate passes the compatibility test's screening. The compatibility test describes how wholly unacceptable candidates are eliminated from further consideration and the profitability test describes how the best is chosen from among the survivors. Of course, if only one candidate survives the compatibility test there is no need to apply the profitability testCthe candidate simply is adopted on the basis of compatibility. The profitability test serves adoption decisions but does not serve progress decisions.


Image theory draws upon a broad and varied conceptual and empirical literature (see Beach, 1990). However, because the theory is so new, the research that it has generated has necessarily been somewhat narrow in focus. Thus far the emphasis has been upon the compatibility test in screening and the profitability test in choice.

Compatibility in Screening

The study of screening had its beginning with research by Payne (1976). Based upon the way in which subjects searched for information about the characteristics of the options, Payne concluded that they began by screening out the least acceptable options using simple, noncompensatory strategies, and then switched to more complicated, compensatory strategies to choose the best option from among the survivors of screening. The most commonly accepted interpretation of these results is that subjects strive to reduce the processing demands of the task by using simple strategies to eliminate obviously flawed options, reserving their cognitive resources for the more complicated strategies by which they make choices among the survivors. In short, screening is merely preliminary to the main event, choice.

Image theory resurrects the question of screening. It proposes that decision making indeed has two steps, screening and choice, but that because screening determines the choice set from which the best will be chosen, it clearly is as central to the process as choice is. Indeed, whenever screening results in a choice set that contains only one option, a common circumstance, there is no need for choice at all and the decision is wholly determined by screening.

Image theory research on screening has followed Payne's (1976) lead in using decision options (entry level jobs, rental rooms, or condominiums) that are easily understood and evaluated by the college students who serve as research subjects. The general strategy has been to present an array of options to the subjects with instructions to examine the options' characteristics and form a 'short list' (i.e., to screen and to form a choice set). Then they are instructed to choose the best (most promising) option from the short list. Depending upon the experiment, the dependent variables are the way in which information is examined, ratings of the options' attractiveness, and subjects' selections for the short list or for the best option.

Violations. Beach and Strom (1989) asked subjects to assume the role of a newly graduated student who was looking for a job. Each subject was presented with an array of jobs; the characteristics of each job violated or did not violate the job seeker's standards. The results showed that rejection of options regularly occurred after observation of roughly four violations; this is the rejection threshold. Nonviolations played virtually no role at all in screening except to stop information search when no violations were observed (or else a perfect option never would be accepted because the search for violations would never stop). In short, the Beach and Strom (1989) study supports the hypothesis that screening relies almost exclusively upon on violations of standards. It also demonstrates the existence of a rejection threshold, and suggests that for a specific decision task the rejection threshold may remain fairly constant.

The Beach and Strom (1989) finding about the primacy of violations in screening was confirmed by Rediker, Mitchell, Beach and Beard (in press) in a different context when MBA students made decisions about various computer firms as options for acquisition by a diversifying business. A correlation of -.95 was obtained between the rated acceptability of options during screening and the number of violations by those firms of the customary practices and strategic goals (standards) of the acquiring business.

Two-step process. Van Zee, Paluchowski and Beach (1992) examined what happens to the information that has been used in screening when the time comes to choose the best from among the survivors in the choice set. In the course of the experiment, subjects made ratings of the acceptability of the options (rooms to rent) at various points in the process of screening and choosing. As expected, results showed that ratings made during screening precisely reflected the options' violations, and that violations precisely predicted which options were rejected and which were retained for the choice set. However, when ratings were made during choice, something quite unexpected was found. These ratings reflected only the information that was obtained about the surviving options after they had been screened. That is, the information used to screen the options did not appear to have much impact on evaluations made during choice (correlations between .02 and .07), while information received after screening had a major impact on pre-choice evaluations (correlations between .67 and .86). It was as though, as one subject remarked, the information used in screening had been 'used up' and had nothing to contribute to choice.

That the two steps, screening and choice, are quite different tasks is even more clearly shown by the results of a study by Potter and Beach (in press, b). Here some subjects were given descriptions of six options (time-share condos) and asked to screen them to create a choice set. Other subjects were given the same descriptions of the options and were asked to choose among them. The unique feature of the experiment was that in both cases each description contained information about the probability (.75 or .25) that the option would be available when the decision maker wanted it.

For the screening group, the prediction was that the availability information would be used as an absolute standard, with .25 being regarded as a violation and .75 being regarded as a nonviolation. This means that when the probability is .25 the result simply is an increase in the number of violations. In the data, this increase would be evidenced as an additive combination of the 'availability violation' with the other violations.

For the choice group, the prediction was that availability would be used as a probability in the manner prescribed by the expected utility choice model (Raiffa, 1968); the option's utility would be discounted by the probability, .25 or .75, that it would be available. In the data, discounting would be evidenced as a multiplicative combination of availability and utility.

Results were as predicted. The same information, availability, was used differently in the two tasks. The implication is that screening and choice are distinctly different tasks that rely on distinctly different information processing.

An empty choice set. Potter and Beach (in press, a) examined what happens when subjects screen and then move on to choice only to find that the options in the choice set have become unavailable. For example, suppose the job applicants on one's short list have taken other jobs, or the condos one was going to examine further and choose from among have already been taken.

In the first experiment, 35 subjects screened options (rooms to rent). Then, when they were ready to choose the best option from among the survivors they were told that none of the options in the choice set were available. They were asked whether they would prefer to search for new options or to go back and reconsider the options they had previously rejected. Thirty-one of the 35 (89%) said that they would prefer to start all over with a new set of options.

Other subjects were presented with the same task except that they were told that there were no new options to be had and they would have to rescreen the options that they had rejected and arrive at a new choice set. Two things happened. On the one hand, subjects lowered their standards a little when they rescreened. On the other hand, they also relaxed their rejection thresholds a little when they rescreened. By doing both things they were able to avoid having to make gross compromises in either the standards or in their tolerance of imperfection.

Imperfect information. Finally, Potter and Beach (1992) examined the effects on screening when only imperfect information about options (condos) was available. Each option had eight characteristics, some of which violated standards and some of which did not. Additionally, however, information about various characteristics was missing for different options. Results showed that subjects treated missing information about a characteristic as though it were a violation. The effect of doing this is to make screening increasingly conservativeCthe less that is known about an option, the less the decision maker is inclined to pass it on to the choice set. Then, on top of this initial conservatism is added an even greater disinclination to pass the option as its observed violations increase in number.

The Profitability Test and Choice

Research on the profitability test is older than research on the compatibility test because the former was motivated by Beach and Mitchell's earlier (1978) decision strategy selection model. As was stated above, the profitability test is a name for the decision maker's repertory of choice strategies. Image theory incorporates the Beach and Mitchell (1978) strategy selection model as the profitability test. Thus research on the profitability test, strategy selection, draws heavily from existing literature (e.g., Lussier & Olshavsky, 1979; Olshavsky, 1979; Payne, 1976; Payne, Bettman & Johnson, 1988, in press; Svenson, 1979) in addition to work done in our own laboratory (Christensen-Szalanski, 1978, 1980; Huffman, 1978; McAllister, Mitchell and Beach, 1979; Nichols-Hoppe and Beach, 1989; Smith, Mitchell and Beach, 1982; Waller and Mitchell, 1984). Because space does not permit an account of this research in the same depth as that presented above for screening, the interested reader is referred to Beach (1990) for a detailed description. Briefly, the results support the contention that decision makers possess repertories of choice strategies and that their selection of which strategy to use in a particular decision is contingent upon specific characteristics of the decision problem, the decision environment, and of the decision maker himself or herself. In the context of image theory the profitability test specifies some of these contingencies (Beach, 1990).


For consumer research, the major implication of the results reported above is that screening and choice are very different processes. Because screening dictates what is available for choice, it is not merely preliminary to the more important and more interesting process, choice. Sometimes screening is the only process; when only one option is being considered or when only one option survives screening, choice is unnecessary. In fact, it is only when more than one option survives screening (meaning that from the point of view of meeting minimal standards all of the survivors are equal) that choice is required to break the tie. In the broadest sense, choice really is less interesting than screening rather than more interesting.

The second implication is that the transfer of options from the consideration set to the choice set is not based on the degree to which they comply with the decision maker's standards, as is generally assumed. Rather, retention is based on the absence of violations, which sounds like much the same thing, but empirically it is not. Even though decision makers prefer to know about nonviolations first, their screening decisions are almost totally determined by violations. This means that presentation of options to consumers should focus on putting the best possible light on violations (or better yet, correcting them), rather than focussing exclusively on the nonviolations.

The third implication is that decision makers apparently regard screening and choice as two separate tasks, using pre-screening information to screen options and using post-screening information to make choices. The fact that the tasks are separate and use different information suggests that appropriately timing of the presentation of information to consumers may influence their final choices.

The fourth implication is that doubts (low probability) about whether outcomes of options will in fact be forthcoming is treated as just another violation in screening (additively), but may be treated in some other way in choice, among which is discounting the attractiveness of the outcomes (multiplicatively). Similarly, missing information about options is treated as a violation during screening. Therefore, failure to present information about an option's violations will itself be treated as a violation, thereby reducing the acceptability of the option.

The fifth implication is that if the choice set is empty when it is time to make a choice, decision makers will resist having to form a new choice set from among the previously screened out options. Empty choice sets arise when no option survives screening or when the survivors become unavailable before a choice is made. If they are forced to re-screen the rejected options, decision makers will lower their standards and readjust their rejection thresholds. This frequently is observed when consumers buy options that are in limited supply, such as houses. If the house that meets their standards has been sold by the time they are ready to buy it, prospective homeowners will select a less acceptable house by reducing the importance of some characteristics of houses and by being more tolerant of houses that previously were regarded as unacceptable.

Finally, the research on the multiplicity of strategies by which decision makers make choices has its own implications for consumer research. The research results make it abundantly clear that expected value maximization is but one class of choice strategies and that consumer research must explore the alternatives if a thorough understanding of consumer choice is to be achieved. Decision strategy selection is contingent upon the characteristics of the decision makers themselves, upon the characteristics of the decision task, and upon the characteristics of the environment in which the decision arises. The foundations for the necessary research program is provided by the citations given previously in this article and by the citations in a recent comprehensive review of the work by John Payne and his colleagues (Payne, Bettman & Johnson, in press).


Beach, L. R. (1990). Image theory: Decision making in personal and organizational contexts. Chichester, England: Wiley.

Beach, L. R. & Lipshitz, R. (in press). Why classical decision theory is an inappropriate standard for evaluating and aiding most human decision making. In G. A. Klein, J. Orasanu, R. Calderwood & C. Zsambok (Eds.) Decision making in action: Models and methods. New York: Ablex.

Beach, L. R. & Mitchell, T. R. (1978). A contingency model for the selection of decision strategies. Academy of Management Review, 3, 439-449.

Beach, L. R. & Mitchell, T. R. (1987). Image theory: Principles goals, and plans in decision making. Acta Psychologica, 66, 201-220.

Beach, L. R. & Mitchell, T. R. (1990). Image theory: A behavioral theory of decisions in organizations. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational behavior (Vol. 12). Greenwich, CT: JAI.

Beach, L. R., Mitchell, T. R., Paluchowski, T. F. & van Zee, E. H. (1992). Image theory: Decision framing and decision deliberation. In F. Heller (Ed.), Leadership and decision making. Cambridge, UK: Cambridge University Press.

Beach, L. R., Smith, B., Lundell, J. & Mitchell, T. R. (1988). Image theory: Descriptive sufficiency of a simple rule for the compatibility test. Journal of Behavioral Decision Making, 1, 17-28.

Beach, L. R. & Strom, E. (1989). A toadstool among the mushrooms: Screening decisions and Image Theory's compatibility test. Acta Psychologica, 72, 1-12.

Christensen-Szalanski, J. J. J. (1978). Problem-solving strategies: A selection mechanism, some implications, and some data. Organizational Behavior and Human Performance, 22, 307-323.

Christensen-Szalanski, J. J. J. (1980). A further examination of the selection of problem-solving strategies: The effects of deadlines and analytic aptitudes. Organizational Behavior and Human Performance, 25, 107-122.

Huffman, M. D. (1978). The effect of decision task characteristics on decision behavior (Technical Report No. 78-16). University of Washington, Department of Psychology, Seattle.

Klein, G. A., Orasanu, J., Calderwood, R. & Zsambok, C. (Eds.) (in press). Decision making in action: Models and methods. New York: Ablex.

Lussier, D. A. & Olshavsky, R. W. (1979). Task complexity and contingent processing in brand choice. Journal of Consumer Research, 6, 154-165.

McAllister, D., Mitchell, T. R. & Beach, L. R. (1979). The contingency model for selection of decision strategies: An empirical test of the effects of significance, accountability, and reversibility. Organizational Behavior and Human Performance, 24, 228-244.

Miller, G. A., Galanter, E. & Pribram, K. H. (1960). Plans and the structure of behavior. New York: Holt, Rinehart & Winston.

Nichols-Hoppe, K. T. & Beach, L. R. (1990). The effects of test anxiety and task variables on predecisional information search. Journal of Research in Personality, 24, 163-172.

Olshavsky, R. W. (1979). Task complexity and contingent processing in decision making: A replication and extension. Organizational Behavior and Human Performance, 24, 300-316.

Payne, J. W. (1976). Task complexity and contingent processing in decision making: An information search and protocol analysis. Organizational Behavior and Human Performance, 16, 366-387.

Payne, J. W., Bettman, J. R. & Johnson, E. J. (1988). Adaptive strategy selection in decision making. Journal of Experimental Psychology: Learning, Memory and Cognition, 14, 534-552.

Payne, J. W., Bettman, J. R. & Johnson, E. J. (1992). The adaptive decision maker. In press.

Potter, R. E. & Beach, L. R. (in press, a). Decision making when the acceptable options become unavailable. Organizational Behavior and Human Decision Processes.

Potter, R. E. & Beach, L. R. (in press, b). Imperfect information in pre-choice screening of options. Organizational Behavior and Human Decision Processes.

Raiffa, H. (1968). Decision analysis: Introductory lectures on choices under uncertainty. New York: Addison-Wesley.

Rediker, K. J., Mitchell, T. R., Beach, L. R. & Beard, D. W. (in press). The effects of strong belief structures on information processing evaluations and choice. Journal of Behavioral Decision Making.

Smith, J. F., Mitchell, T. R. & Beach, L. R. (1982). A cost-benefit mechanism for selecting problem solving strategies: Some extensions and empirical tests. Organizational Behavior and Human Performance, 29, 370-396.

Svenson, O. (1979). Process descriptions in decision making. Organizational Behavior and Human Performance, 23, 86-112.

van Zee, E. H., Paluchowski, T. F. & Beach, L. R. (1992). The effects of screening and task partitioning upon evaluations of decision options. Journal of Behavioral Decision Making, 5, 1-23.

Waller, W. S & Mitchell, T. R. (1984). The effects of context on the selection of decision strategies for the cost variance investigation. Organizational Behavior and Human Performance, 33, 397-413.