Consideration of Inhibition As a Mechanism in Decision Processes

Mary Frances Luce, Duke University
ABSTRACT - This paper explores the implications for choice which result from abandoning the capacity model of attention for the selection for action model. Specifically, the likely role of the process of inhibition in choice is discussed. Functions which inhibition may fulfill during choice are proposed. For instance, inhibition may operate upon representations of attractive but nonchosen alternatives, allowing the decision maker to move forward with a minimum of cognitive conflict. Predictions regarding the amount and direction of inhibition during choice are also offered.
[ to cite ]:
Mary Frances Luce (1992) ,"Consideration of Inhibition As a Mechanism in Decision Processes", in NA - Advances in Consumer Research Volume 19, eds. John F. Sherry, Jr. and Brian Sternthal, Provo, UT : Association for Consumer Research, Pages: 161-165.

Advances in Consumer Research Volume 19, 1992      Pages 161-165

CONSIDERATION OF INHIBITION AS A MECHANISM IN DECISION PROCESSES

Mary Frances Luce, Duke University

ABSTRACT -

This paper explores the implications for choice which result from abandoning the capacity model of attention for the selection for action model. Specifically, the likely role of the process of inhibition in choice is discussed. Functions which inhibition may fulfill during choice are proposed. For instance, inhibition may operate upon representations of attractive but nonchosen alternatives, allowing the decision maker to move forward with a minimum of cognitive conflict. Predictions regarding the amount and direction of inhibition during choice are also offered.

The fields of decision theory and marketing have traditionally borrowed from cognitive psychology in order to predict and explain decision making. The capacity model of attention, which has been dominant in cognitive psychology for some time, has often been utilized to explain choice behavior. Criticisms of, as well as alternatives to, the capacity model have recently emerged. The selection for action model, a dominant alternative, stresses the importance of inhibitory processes to cognitive functioning. This paper will consider the choice process within a selection for action framework, concentrating upon inhibition's possible role in choice.

MODELS OF ATTENTION

General capacity or resource views of attention cite a limited supply of cognitive resources as the reason for limitations in cognitive functioning. Several compelling criticisms of the capacity view, which in total leave much doubt as to its viability, have been formulated (e.g., Navon 1984; Neumann 1987). The capacity model is likely unfalsifiable and methods which are purported to test this model have alternative explanations (Navon 1984). Resources defy definition, and methods used to determine resource utilization or availability usually yield conflicting results (e.g., Navon 1984). For instance, it is often proposed that one can deduce resource requirements by measuring interference found between concurrently performed task pairs. However, results in these dual task studies are highly dependent upon the specific structure of each task. These conflicting results led to the suggestion that distinct resource pools exist and that tasks will interfere with each other to the extent that they draw upon identical pool(s). However, even multiple resource views cannot with any degree of parsimony account for both the extremely specific (i.e., based upon task similarity) and the general (i.e., present with any task combination) interference found among dual task combinations (Hasher and Zacks 1988; Neumann 1987). Finally, it is unclear at what stage in information processing the resource "bottleneck" manifests itself, for humans are quite capable of parallel information processing prior to overt response (e.g., Neill 1989).

An alternative to capacity views, selection for action models generally seem to account for data better than and avoid the many theoretical difficulties of capacity models. These models take a more functional approach, searching for sources of attentional limitations. The most obvious source of limitation in information processing results directly from motor limitations. Simply, some mechanism must prevent physically incompatible actions from being attempted, for such attempts would preclude coherent physical or mental activity (e.g., Neumann 1987). A second source of limitation is inferred from the fact that information streams may apparently become confused during processing. For instance, in the Stroop paradigm, subjects required to name the ink color of a word are slowed when the word spells a different color. In other words, multiple, often conflicting representations of stimuli must be narrowed down so that (ideally) only task relevant information is sustained in consciousness (Neill 1989). These motor- and information-based limitations are often discussed separately; however, interference among information streams may result from the fact that such streams often have incompatible implications for action. For instance, in the Stroop paradigm, incompatible tendencies to speak two different color names may arise. Thus, all cognitive limitations may ultimately be traceable to the motor system.

Selection for action authors propose that attention operates to suppress representations which potentially conflict with one's ongoing stream of action (e.g., Neill 1989; Neumann 1987). This is necessary to avoid attempts to complete physically incompatible actions. Of course, physically compatible actions are also frequently prevented. This may happen because the brain operates by a simple blocking mechanism, with one ongoing action or set of actions typically suppressing all others (e.g., Neumann 1987). This apparent inefficiency in attention is likely necessary because there is no "central executive" with sufficient knowledge to ascertain which actions are incompatible. Thus, while capacity models explain most cognitive limitations as resulting from an inadequate resource supply, selection for action models view these same limitations as resulting from active inhibition. This emphasis upon inhibition is likely the most significant departure of the selection for action model from the capacity model. Empirical evidence for active inhibitory processes may be taken as support for the selection for action view; this evidence is reviewed below.

Empirical Demonstrations of Inhibition

There is abundant evidence that inhibition operates in lower level cognitive processes. In selective attention tasks, an ignored stimulus is generally disadvantaged when it is the stimulus to which the subject must respond on the subsequent trial (e.g., Hasher and Zacks 1988). This indicates that an active inhibitory process acts upon the representations of distractors. For example, if during a Stroop task the ignored word on trial one is identical to the ink color one must name on trial two, response times for the second trial lengthen (e.g., Neill and Westberry 1987). This general pattern of results has been found numerous times using many attention tasks (e.g., Driver and Tipper 1989; Neill, Lissner, and Beck 1990).

More important to my hypotheses concerning the relationship between inhibition and choice, inhibition has been implicated in higher level cognitive processing. Bjork (1989) proposes that a process of retrieval inhibition operates, beyond effects of differential rehearsal, to eliminate access to items which subjects are instructed to forget. Retrieval inhibition may also be inadvertently produced; for example, rehearsal may result in a loss of access to related information (Bjork 1989). Inhibition may operate within comprehension, suppressing words' alternative meanings and irrelevant associations (e.g., Hasher and Zacks 1988; Simpson and Kellas 1989). Simpson and Kellas (1989) find slower responses to a word related to one meaning of an ambiguous prime when the prime has previously been paired with a word related to a different meaning. For example, their subjects are slower to respond to the word "money" after the prime "bank" when they have previously responded to "river" after "bank." This and other empirical work indicates that inhibition is an active, goal directed process which decouples representations from consciousness and/or from action control.

CHOICE

Decision theorists are painfully aware of man's limitations as an information processor. For instance, individuals do not seem capable of trading off or otherwise combining more than a few attributes during decision making (e.g., Shepard 1964). Capacity limits have, at least implicitly, been accepted as the explanation for observed limitations in decision making. Given the compelling criticisms of capacity models, it may be useful to explore alternative causes of these limitations, such as those implied by the selection for action model.

It is not immediately apparent what abandoning the capacity model for the selection for action model implies for choice theories. This new model uniquely predicts that inhibition will play a central role in cognitive functioning, but what understanding or new hypotheses does the process of inhibition add to choice? This question is addressed below.

Inhibition in Choice

Alternatives in choice almost invariably elicit conflicting action tendencies, or the desire to approach (i.e., to buy) more than one alternative simultaneously. It will very likely cause conflict to give up any attractive alternative(s) when one is not absolutely certain that one's choice is optimal. Such certainty in choice seems unlikely, especially given that people often do not know what attribute levels they do or will prefer (e.g., Shepard 1964). Hence, choice is inherently laden with conflict. Limits apparently exist on the amount of conflicting information one can comfortably sustain in consciousness (e.g., Neill 1989). In choice, individuals may be limited by this inability to sustain cognitive conflict rather than by resource supplies.

Processes which allow a decision maker to deal with conflict in choice have been proposed. Individuals may selectively attend to goal related information, encoding only items which will bolster or protect their current intentions or decisions (Montgomery 1989; Kunda 1990). Montgomery (1989) describes decision making as dominance structuring, or the search for a cognitive representation in which the chosen alternative is best on every relevant attribute. Such a representation may be possible only if the decision maker restricts her attention to a subset of alternatives and attributes, ignoring those which violate dominance (Montgomery 1989).

Inhibition may operate within dominance structuring or other cognitive processes to eliminate conflict in choice. Decisions have been proposed to result from temporary states of mind which allow one alternative to seem best (Shepard 1964). These temporary states of mind may be attained by inhibiting representations of alternatives, attributes, and/or specific alternative/attribute values. Inhibition may operate to decouple all but the chosen alternative from the control of action. The inhibition of problematic attributes may allow choice to proceed without the completion of difficult trade offs. For instance, a car buyer who is unable to decide how much he is willing to pay for an additional "unit" of safety may simply inhibit safety considerations. People may inhibit disadvantages of their chosen alternative and/or advantages of other alternatives. In summary, inhibition may selectively restrict the activation of information, producing a cognitive representation in which one's chosen alternative seems more like a dominating alternative than it is in any objective sense.

P1: Immediately after choice, a individual should display:

a.) inhibition of items inconsistent with dominance.

b.) more difficulty retrieving these items than retrieving items consistent with dominance.

The inhibition proposed above, and discussed in several later propositions, would be indicated by a longer latency in responding to (i.e., naming) an item than in responding to a control item.

Long Term Effects of Inhibition. An inhibition framework implies ways in which successive choice outcomes may be related; capacity models, in contrast, do not directly address these relationships. Inhibition may have enduring effects upon thoughts which an individual has reason to escape. Relatively permanent inhibition may operate upon items such as the disadvantages of one's chosen alternative. Inhibitory pathways analogous to the associative links developed when two thoughts are simultaneously rehearsed may cause one node to automatically inhibit another. If so, inhibition during choice should make consideration of a chosen alternative's disadvantages less likely and therefore should correlate negatively with post decisional regret or cognitive dissonance. Further, long term inhibition acting upon nonchosen brands could cause resistance to their reconsideration. This resistance may hold even after preferences and/or product attributes have changed, perhaps making an inhibited brand optimal. This process could operate to create and maintain brand loyalty. Inhibition may also underlie the demonstrated tendency for chosen alternatives to be remembered best (Johnson and Russo 1981) and the low probability of choosing a previously rejected brand even when new information makes that brand more attractive (Biehal and Chakravarti 1983).

Factors Influencing Inhibition

The following sections will focus upon the influence which situational factors and goals may have upon the direction and magnitude of inhibition in choice. Two situational factors are discussed: the overall nature of the evoked set and agenda effects. The discussion of goals focuses upon the effect of effort/accuracy trade-offs upon inhibition. The relationships hypothesized below may not be apparent when using a capacity framework to understand decision making.

Situational Factors. Inhibition seems to operate rather selectively upon representations which have the potential to produce confusion or conflicting response tendencies (e.g., Neill 1989). Therefore, one may be able to predict the target(s) of inhibition by determining the likelihood of conflict. For example, conflict in choice seems more likely to the extent that alternatives are close in expected value. Therefore, when alternatives' values are close, inhibition, especially inhibition of nonchosen alternatives with relatively high expected values, should rise. Attributes on which the chosen alternative is weak may also show increased inhibition in this case.

Inhibition may dampen activation of material that is extraneous to the choice process as well as information causing conflict. The functions of eliminating conflict and of suppressing irrelevant information may have conflicting implications for the inhibition of any one item. Material which is completely irrelevant to choice should not produce conflict, and vice versa. For example, dominated alternatives may produce no conflict, but they may be clearly extraneous to the decision process. Therefore, it is difficult to determine whether a dominated alternative will be inhibited. However, situational factors may allow for such prediction. For instance, it seems that alternatives dominated by the chosen alternative will not be inhibited because they can be profitably incorporated into a dominance structure, and therefore are not truly irrelevant to choice.

P2: Inhibition should operate upon:

a.) information which produces cognitive conflict.

b.) extraneous information activated during choice.

More specific hypotheses are also possible, such as:

P3: As a non-dominated alternative's expected value becomes closer to that of the chosen alternative, inhibition of that alternative becomes more likely.

P4: Dominated alternatives are likely to be inhibited, unless they are dominated by the chosen alternative.

The above analysis is very different than the one which a capacity viewpoint would yield. The selection for action framework uniquely implies that one's ability to simultaneously consider a number of alternatives will vary as a function of the relationship(s) among the alternatives, rather than strictly of available capacity. For instance, this framework implies that a large dominance structure may be more easily maintained in consciousness than a smaller amount of information relating to severely conflicting alternatives. Stated differently, information overload is no longer strictly a function of the difference between available capacity and externally available "bits" of information. Information overload may occur due to incomplete inhibitory processes and may be indicated by sustained activation of extraneous or conflicting information. This break down may be correlated with both the amount and the type of information available.

P5: Information overload is more likely as:

a.) more information is available.

b.) pieces of information related to a choice conflict with one another more severely.

Again, it is also possible to develop more specific predictions, for example:

P6: Subjects will be better able to process and to recall a given amount of choice information as that information more closely approximates a dominance structure.

A second situational factor which may affect inhibition is the order in which alternatives or attributes are considered. Inhibition in attention tasks is most likely once an individual focuses upon a candidate for response (Neill 1989). During decision making, a tentative choice is analogous to a candidate for response. Much of the choice process seems to involve a building up of one's intention toward such a tentative choice (Montgomery 1989). Once one alternative is tentatively accepted, any information which indicates that this alternative is not best will potentially cause competing action tendencies, and so may be inhibited. It seems likely that there is some arbitrariness to which out of an acceptable set of alternatives becomes the tentative choice, given that preferences are not always well defined; thus, inhibition may be subject to agenda effects.

P7: Items which are processed earlier will have a lower probability of becoming inhibited than items which are processed later will have.

Note that this proposition implicitly assumes that the individual is processing by alternative, at least during the earliest stages of choice. The order in which attributes are considered may likewise influence which alternative first becomes the tentative choice, especially during more attribute-based processing (Montgomery 1989).

Goals. The selection for action framework often emphasizes the apparent goal directedness of inhibition (e.g., Hasher and Zacks 1988; Neill 1989). This framework implies that more complete information processing is possible if the decision maker adopts the appropriate goal(s). On the other hand, the capacity framework implies that information processing is simply possible before and impossible after a decision problem reaches some level of demand for capacity. Again, the selection for action model uniquely implies situational variance in individuals' abilities to process information.

If inhibition is goal directed, the decision maker's goals should directly influence the choice process. Inhibition may be driven by the extent to which the goal of minimizing effort (versus maximizing accuracy) is prevalent. Inhibition may act as a substitute for effort during choice, allowing confidence in a decision although all of the appropriate comparisons and trade-offs are not made. Hence, as one's relative emphasis on minimizing effort rises, more inhibition should be observed in choice. Of course, inhibition may also be necessary when an individual has made the most accurate choice possible, for choice will almost always involve both conflicting and extraneous information. Still, if accuracy is foremost, inhibition should be relatively less prevalent, allowing more complete memory search and decision processes.

Further, specific goals have been hypothesized to enhance the accessibility of information consistent with them (Kunda 1990). Perhaps inhibition is the mechanism through which this occurs. For instance, some motivation to buy one brand may initiate inhibition of that brand's most attractive competitors.

P8a: Within a given level of confidence in a given decision, inhibition and effort will be inversely related.

P8b: Goals of accuracy and effort should have inverse effects upon inhibition's magnitude while more specific goals should influence inhibition's direction.

The goal to minimize effort may be somehow "hardwired," or present unless an individual makes considerable effort to overcome it. Effort minimization, and the concomitant inhibition, may have adaptive value in many decision situations. Neumann (1987) discusses the problem of picking a path of escape from a predator. In such cases, unwavering commitment to a course of action is often more important than which alternative course is actually chosen. Inhibition may serve the crucial function of eliminating doubt, therefore allowing the chosen alternative to be executed as quickly and completely as possible. This is also consistent with Shepard's (1964) assertion that the primary goal in choice is to escape the conflict of the decision situation. Situational factors, such as involvement, will likely influence the extent to which the decision maker desires to make a truly accurate choice.

CONCLUSIONS

This paper represents an attempt to develop the implications for choice realized by utilizing an alternative to the capacity framework of attention. In general, it seems that decision researchers would be well served to abandon the computer metaphor for a brain metaphor acknowledging the parallel processing, and the excitation and inhibition, which underlie cognitive functioning (Bjork 1989). Part of this process will likely involve abandoning the capacity model which has been heavily influenced by the computer metaphor (Bjork 1989; Navon 1984). It is hoped that this work provides a small step toward such a paradigm shift.

Investigation of inhibition in choice is important for a broader reason than its impact upon consumer and decision making research. Inhibition has the potential to be a powerful unifying principle in psychology, illuminating topics from perception (McClelland and Rumelhart 1981) to choice. In fact, the principle of inhibition may have the considerable unifying power often attributed to the capacity metaphor. The finding that inhibition operates in choice would be evidence for such unifying power.

REFERENCES

Bjork, Robert A. (1989), "Retrieval Inhibition as an Adaptive Mechanism in Human Memory," in Varieties of Memory and Consciousness: Essays in Honor of Endel Tulving, eds. H.L. Roediger, III and F.I.M. Craik, Hillsdale N.J.: Erlbaum., 309-330

Biehal, Gabriel and Dipankar Chakravarti (1983), "Information Accessibility as a Moderator of Consumer Choice," Journal of Consumer Research, 10 (June), 1-14.

Driver, Jon and Steven P. Tipper (1989), "On the Nonselectivity of Selective Seeing," Journal of Experimental Psychology: Human Perception and Performance, 15 (2), 304-314.

Hasher, Lynn and Rose T. Zacks (1988), "Working Memory, Comprehension, and Aging: A Review and a New View," in The Psychology of Learning and Motivation, Vol. 22, ed. G.H. Bower, N.Y.: Academic Press, 193-225.

Johnson, Eric J. and J. Edward Russo (1981), "Product Familiarity and Learning New Information. in Advances in Consumer Research, Vol. 8, ed. Kent B. Monroe, Ann Arbor, MI: Association for Consumer Research, 151-155.

Kunda, Ziva (1990), "The Case for Motivated Reasoning," Psychological Bulletin, 108 (3), 480-498.

McClelland, James L. and David E. Rumelhart (1981), "An Interactive Activation Model of Context Effects in Letter Perception," Psychological Review, 88 (5), 375-407.

Montgomery, Henry (1989), "From Cognition to Action: The Search for Dominance in Decision Making," in Processes and Structure in Human Decision Making, eds. H. Montgomery and O. Svenson NY: John Wiley and Sons, 23-49.

Navon, David (1984), "Resources--A Theoretical Soupstone?" Psychological Review, 91, 216-234.

Neill, W. Trammell and Richard L. Westberry (1987), "Selective Action and the Suppression of Cognitive Noise," Journal of Experimental Psychology: Learning Memory, and Cognition, 13 (2), 327-334.

Neill, W. Trammell (1989), "Lexical Ambiguity and Context: An Activation-Suppression Model," in Resolving Semantic Ambiguity, ed. D.S. Gorfein, NY: Springer-Verlag, 63-83.

Neumann, Odmar (1987), "Beyond Capacity: A Functional View of Attention," in Perspectives on Perception and Action, eds. H. Heuer and A. F. Sanders, Hillsdale, NJ: Erlbaum, 361-394.

Shepard, R.N. (1964), "On Subjectively Optimum Selections Among Multi-Attribute Alternatives," in Human Judgments and Optimality, eds. M.W. Shelley and G.L. Bryan, N.Y.: Wiley.

Simpson, Greg B. and George Kellas (1989), "Dynamic Contextual Processes and Lexical Access," in Resolving Semantic Ambiguity, ed. D.S. Gorfein, NY: Springer-Verlag, 40-62.

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