UNIVERSITY Of-
ILLINOIS LIBRARY
AT URBANA-CHAMPAIGN
BOOKSTACKS
H
O/^*^ i> ^ :>
m
H
H
,-<w-
Digitized by the Internet Archive
in 2011 with funding from
University of Illinois Urbana-Champaign
http://www.archive.org/details/cognitivebiasins863schw
FACULTY WORKING PAPER NO. 863
Cognitive Bias in Strategic Decision-Making: Some Conjectures
Charles R. Schwenk
itHE
!i?'ye.
"^^ or
^f
'^"'^'V/l..^/ 'i.,
^'^^^P»
College of Comm-rce and Business Adrrinistraiion Bureau ct Economic and -Busirisss Research University of illincis. Urbana-Chanoaign
"vfl.v:
BEBR
FACULTY WORKING PAPER NO. 863 College of Commerce and Business Administration University of Illinois at Urbana-Champaign April 1982
Cognitive Bias in Strategic Decision-Making: Some Conjectures
Charles R. Schwenk, Assistant Professor Department of Business Administration
ABSTRACT
Strategic decision-making can be viewed as a special kind of decision- making under uncertainty. Researchers in a variety of fields have identified a number of cognitive or perceptual mechanisms by which decision makers distort information from the environment to make it appear simpler and reduce the apparent uncertainty in a decision-making task. Within this paper, a four-stage model of strategic decision-making is developed and the possible effects of cognitive biases at each stage are discussed. Finally, techniques for introducing conflict into strategy formulation are proposed as ways of counteracting the effects of these biases.
COGNITIVE BIAS IN STRATEGIC DECISION- MAKING: SOME CONJECTUKES
Strategic decision-making has been defined as a special kind of ill- structured problem-solving process and as a special type of decision- making under uncertainty (Hofer & Schendel, 1978:46). Because of the importance of strategic decision-making, much effort has been devoted to developing procedures for improving the effectiveness of the process by increasing the amount of data considered and the thoroughness with which it is evaluated (Grant & King, 1979:104-122).
A number of recent articles in the strategic management field have developed and validated descriptive models of the strategic decision- making process (Mintzberg, Raisinghanl, & Theoret 1976; Glueck 1976; Mazzolini, 1981; Hofer & Schendel, 1978). These models involve various numbers of stages and are generally similar to earlier models of the organizational decision-making process (Lang, Dittrich, & White, 1978). The decisional activities at each stage have been examined experimentally by cognitive psychologists and behavioral decision theorists and numerous biases have been identified which limit decisional effectiveness. An understanding of these biases is necessary for the design of effective strategic decision-making aids.
Theorists in the field of Strategic Management have pointed out that human cognitive limitations might affect strategic decision-making (Steiner & Miner, 1977:226-231; Mintzberg, 1973:45-46), basing their arguments on Simon's notions of "bounded rationality" and "satisficing. " They have pointed out that strategic decision-makers are subject to bounded rationality and do not optimize in their decisions but have not
-2-
discussed the specific effects of bounded rationality on decision- makers' perceptions or the specific decisional behaviors adopted in preference to optimizing. However, some researchers have discussed the effects of specific cognitive biases on general managerial decision-making (Taylor, 1975; Hogarth, 19 80) and on the use of forecasts (Hogarth & Makridakis, 1981).
This paper develops conjectures about possible biases in strategic decision-making by drawing on literature in the fields of cognitive psychology and behavioral decision theory dealing with problem formu- lation and decision- making under uncertainty. These biases allow for some specific predictions about the types of errors to which decision makers will be subject in various activities or phrases of strategic decision-making. Such information could be helpful to researchers in attempting to explain decisionad failures and to practitioners inter- ested in reducing the effects of these biases through the use of deci- sion aids. This paper does not, of course, represent an exhaustive review of the literature in these fields. Rather, attention is focused on some of the more widely discussed cognitive biases.
Research on these biases has dealt with them individually and has not focused on the interaction between then. For this reason, and the fact that most of the research was conducted in laboratory settings using relatively simple judgement and decision tasks, statements about the effects of these biases on strategic decision-making must remain speculative.
In this paper, a general model of the strategic decision-making process will be synthesized from existing models. Cognitive biases
-3-
which may operate at each stage of the process will then be discussed. Finally, suggestions for future research on cognitive biases and for improving strategic decisions will be offered.
A DESCRIPTIVE MODEL OF THE STRATEGIC DECISION- MAKING PROCESS The strategic management literature contains a number of strategic decision- making models. Since cognitive biases will be discussed in terms of their effects on various processes in strategy formulation, it is necessary to generate a model which represents the essential features of the most prominent models in the field. Hofer & Schendel (1978, p. 47) have developed a model which builds on the major ana- lytical models of Andrews (1965, 1980), Ansoff (1965), and others. It includes the seven steps of strategy identification, environmental analysis, resource analysis, gap analysis (problem identification), generation of strategic alternatives, strategy evaluation, and stra- tegic choice.
The fact that such processes actually occur in organizational decision-making has been confirmed by Mintzberg, Raisinghani, & Theoret (1977). Several of the stages dealt with in Hofer & Schendel's model are also covered by Mintzberg, et al. However, their model begins with the identification and diagnosis of a problem. This is followed by the search for alternatives and information related to the alternatives. This stage, in turn, is followed by the evaluation and selection of an alternative. This is an iterative model involving numerous feedback loops which allow decisional activity to cycle from later to earlier stages as the strategic problem definition is progressively refined.
-4-
It may be that the two rcdels actually describe two different but equally legitimate types of strategic decision-making. Leotiades (1979:96-102) distinguished between strategy formulation which is part of a regular strategic planning cycle and strategy formulation which occurs in response to a particular problem (for example, a need to make a particular acquisition decision). Ihe former begins with at- tempts to systematically collect information about the environment and the company's resources and is best described by Hofer & Schendel's model. The latter, which begins in response to a particular problem, is probably best described by Mintzberg's model.
Glueck (1976) proposed a model which included the stages of Appraisal (including an analysis of environmental opportunities and threats and company resources). Choice (generation and consideration of alternative strategies and choice among the alternatives), Implementation, and Evaluation. Finally, Mazzolini (1981) developed a nodel involving five major activities based on his own research and literature review. The five activities included Decision-need identification. Search for alternatives for action. Investigation of courses of action. Review and approval, and Implementation.
These models are built on earlier and more general decision-making
models and are consistent with at least the first three stages of
Simon's four-stage description of the management decision-making
process (1960:40-44).
Decision making comprises four principal phases: finding occasions for making a decision, finding pos- sible courses of action, choosing among courses of action, and evaluating past choices.
-5-
The first phase of the decision-niaking process — searching the environment for conditions calling for decision — I shall call intelligence activity (bor- rowing the military ireaning of intelligence). The second phase — inventing, developing, and analyzing possible courses of action — I shall call design activity. The third phase — selecting a particular course of action from those available — I shall call choice activity. The fourth phase, assessing past choices, I shall call re\dew activity.
(Simon, 1960:40-41)
As a basis for the discussion of simplification mechanisms, a simplified four-stage model of the strategy formulation process has been developed based on earlier models. The four stages in the present model are strategic problem identification, strategic alternatives generation, evaluation and selection, and strategy implementation. There seems to be a good deal of agreement among researchers in this field (represented by those whose model we discussed) that at_ least these basic activities characterize strategic decision-making. How- ever, this model is not intended to represent the full complexity of strategic decision-making.
Any model which is representative of the more popular models in the field would have to include the notion of feedback loops. Thus in this derived model, decisional activity may cycle from later to earlier stages as in the Mintzberg et al. model.
Problem identification includes attempts to identify the com- pany's current implicit and explicit goals, objectives, and strategies as well as an assessment of the significant opportunities and threats in the environment, and the company's current resources. Assessment of opportunities and threats in the environment is a preliminary stage to gap analysis and often requires forecasting. Problem identification
-6-
also includes a recognition that a problem exists. This presumably takes the form of a gap between current or projected future perfor- mance and the explicit and implicit goals and objectives of manage- ment. The problem is then defined and clarified as causes of the gap are identified. Alternatives generation involves the identifica- tion of gap-dosing options and their development to a state of re- finement in which they can be evaluated against each other. The evaluation and selection stage involves the screening and evaluation of alternatives and the selection of the alternative which best solves the problem defined in the preceding stage. Finally, the implementation stage involves carrying out the strategy chosen. This model, along with the Mintzberg et al., Glueck, Hofer & Schendel, and Mazzolini models, is presented in Table 1.
Insert Table 1 about here
Obviously, all of these stages do not occur in all decisions in exactly this sequence. In fact, Witte (1972:179) found from examining 233 organizational decision processes that the activities related to the "stages" of gathering information, developing alternatives, eval- uating alternatives , and making choices were carried out in approximately the same proportion to each other throughout the decision process. The phase theorem (i.e., the theorem that decision-making occurs in dis- tinct phases or stages) postulates that activities associated with gathering information should predominate early in a decision process and that activities related to choice should dominate toward the end of the decision process. In contrast, Witte found that there was a
-7-
relatively high level of the activities associated with all four "stages" at the beginning of the decision process, a lull toward the middle of the process, and a very high level of activity toward the end of the process. Witte concluded "we believe that human beings cannot gather information without in some way simultaneously devel- oping alternatives. They cannot avoid evaluating these alternatives immediately, and in doing this they are forced to a decision" (Witte, 1972:180).
Therefore, the four activities in the derived model may be seen simply as decisional processes which may or may not occur as stages. It may be that the structured process which often characterizes cyclical, formal strategic planning makes it more likely that the phases will be executed in this order while in strategy formulation guided by a par- ticular problem, the activities are less likely to be executed sequen- tially. Normative work in strategic management suggests that problem identification should be based on detailed data which reveals gaps be- tween performance and expectations, numerous strategic alternatives should be generated, and these should be thoroughly and objectively evaluated prior to choice (Hofer & Schendel, 1978). However, the complexity and uncertainty involved in strategic decision-making makes it unlikely that these normative prescriptions will be carried out.
The next section of the paper will discuss cognitive biases which may operate to reduce the apparent complexity and uncertainty in a decision situation and may simultaneously reduce the quality of the strategic decision.
-8-
COGNITIVE BIASES IN STRATEGIC DECISION-MAKING Research in cognitive psychology and behavioral decision theory has identified numerous cognitive biases which may operate in strategic decision-making. These biases may not operate in all strategic deci- sions. However, their effect may help to explain failure in strategic decision-making.
In the next section, the biases identified in cognitive psychology and behavioral decision theory research have been classified according to strategic decision-making stage they seem roost likely to affect.
Table 2 represents the stages of the strategy formulation process and the cognitive biases which may operate at the first three stages. The biases which operate at each stage will be discussed, as well as their probable effects on each stage.
Insert Table 2 about here
Some biases nay operate to reduce decisional effectiveness at more than one of the stages.
Problem Identification
In the strategic decision-making models previously discussed, the major purpose of information gathering in the beginning of the process is to identify gaps between objectives and performance. However, such gaps may indicate either random fluctuations or changes requiring revi- sions in strategy. Decision-makers* expectations may determine how such gaps are interpreted or even whether information on such gaps will be accepted and used. The following have been identified as cognitive biases which may affect problem identification: prior hypothesis bias.
-9-
adjustment and anchoring, escalating conmitinent, the illusion of con- trol, reasoning by analogy, salience, and misguided parsimony (see Table 2).
Prior Hypothesis Bids
Researchers have identified a number of biases which lead decision- makers to ignore or misinterpret information. Levine (1971), Pruitt (1961) and Kason (1960) showed that individuals who formed erroneous beliefs or hypotheses about the relationship between variables tended to make decisions on the basis of these beliefs despite abundant evi- dence over numerous trials that they were wrong. Further, they sought and used information consistent with these hypotheses rather than dis- confirming information. Jervis (1976:143-181) has also provided num- erous examples from international relations of peoples' tendency to accept information which conforms to existing expectations and beliefs. Under the effects of this bias, decision-makers who wish to believe that the company's current strategy is working may ignore information suggesting gaps between performance and expectation. Conversely, those who wish to believe it is failing may overweight Information on such gaps.
Adjustment and Anchoring
Tversky & Kahneman (1974) discuss another bias which helps decision- makers deny gaps. They call this the anchoring bias. In strategic decision-making individuals must often make initial judgements about values of variables critical in particular decisions and revise these judgements as new data comes in. However, the adjustments are typically
-10-
insuf ficient. Final estimates of values are biased toward the initial values. Individuals involved in the ongoing process of strategy formu- lation may attend to negative information about the success of present strategy but they will tend not to make full use of it in revising their predictions of company performance under the present strategy. These revisions will be smaller than are justified by the new infor- mation.
Escalating Commitment
If these initial biases do not come into play, and the gap is recognized, some research suggests that decision-makers deny the significance of the gap and the need for the revision of strategy. That is, once they have recognized the gap, they may define the problem indicated by the gap as one which does not require a change in strategy. Staw (1976), Staw & Fox (1977), Staw & Ross (1978), and Fox & Staw (1979), in laboratory studies using undergraduates as subjects with simulated investment tasks, demonstrated escalating commitment to a chosen alternative despite negative feedback. They found that once an individual commits a significant amount of money to an investment project, he will tend to allocate more funds to the project if he receives feedback indicating that the project is failing than if he receives feedback indicating that the project is succeeding. The feeling of personal responsibility for the project apparently induces decision-makers to remain with their chosen project in spite of evi- dence that it is not paying off. Staw (1976) found a much weaker tendency to escalate commitment in subjects who had not made the
-11-
initial commitir.ent but were dealing with a commitrQent made by an earlier decision-maker.
Staw & Fox (1979) showed that when decision-makers face a series of commitment decisions the escalating commitment effect is strongest in the early decisions and may not persist over time, which suggests that decision-makers may decrease their assessment of the probability of recovering losses with repeated failure. Thus, decision-makers' perceptions of the causes of failure seem to be important determinants of escalating commitment. Staw & Ross (1978) examined the effects of information indicating either exogenous or endogenous causes of failure and found that subjects invested more resources in a failing project when information pointed to an exogenous rather than an endogenous cavise of the setback.
Finally, Fox & Staw (1979) found that escalating commitment was most likely to occur when decision-makers were vulnerable to job loss and when there was strong organizational resistance to the chosen course of action. An excellent review of these studies is found in Staw (1981). In these studies, it is obvious that the decision-makers perceive the discrepancy indicating a project's failure since they allocate more money to projects which appeared to be failing. However, they did not use this perceived discrepancy to alert them to the need to change their strategy. Rather, they seemed to interpret the nega- tive feedback as a signal that they should commit more funds to save the project.
Other research indicates a possible hypothesis which decision-makers may adopt to explain a perceived discrepancy in such a way that it appears
-12-
to require no change in strategy. They may explain it as a result of chance factors rather than a result of a flaw in initial strategy. If they adopt this interpretation, they are likely to persist in the cur- rent course of action and escalate conmitment to it. Staw & Ross (1979) found that subjects committed significantly more funds to a failing project when the reason for the failure was extrinsic (a chance event which could not have been foreseen) than when the reason was seen to be intrinsic.
Illusion of Control
It may be that decision-makers tend to overestimate the role of chance in their failures. This speculation is supported by the work of Langer and Roth (1975) which shows that decision-makers tend to attribute unsuccessful outcomes to chance while attributing successful outcomes to their own skill.
Lefcourt (1973), Langer (1975) and Larwood & Whittaker (1977) have conducted research which deals with decision-makers' judgements about the role of chance in the outcome of a decision and have defined a bias they call the illusion of control. Among other things, the illusion of control leads decision-makers to attribute desirable out- comes to internal factors (such as their ovn skill, intelligence, in- sight, etc.) but to blame such external factors as luck for failures. Decision-makers who note gaps between performance and expectation may tend to attribute these to chance if the gaps tend to reflect badly on the strategies they were responsible for designing earlier. This would cause them to resist changing strategies which have led to per- formance which is below expectation, strategies which are failing.
-13-
Ihis tendency appears to be stronger in individuals who have expe- rienced a string of successes and may therefore be especially strong in upper- level managers involved in strategy formulation. Having risen to the top in their organizations, they would tend to view themselves as successful decision makers and good managers. This would increase their tendency to attribute performance gaps to chance rather than the failure of their strategies. This bias also has a very strong effect on the evaluation of alternatives, as will be dis- cussed in the section on alternatives evaluation.
Reasoning by Analogy
Decision-makers may admit that the gap does exist and that it indicates a need to change current strategy. If this happens, there is at least one mechanism which helps to determine the manner in which the problem will be defined. Steinbruner (1974) has called this mech- anism reasoning by analogy. Reasoning by analogy involves the applica- tion of analogies and images from one problem situation to another. In strategic decision-making, it typically involves the application of analogies from simpler situations to complex strategic problems. This mechanism helps to reduce the aversive uncertainty perceived in the environment. Reasoning by analogy has been shown to be effective in generating creative solutions to problems (Gordon, 1961; Huff, 1980). However in strategic decisions, which involve a great deal of uncer- tainty and complexity, the use of simple analogies is likely to mis- lead the decision-maker into an overly simplistic view of the situa- tion and an incorrect definition of the problem (Steinbruner, 1974, p. 115). -
-14-
A major problem with arguments from analogy is that they are sub- ject to a bias which Tversky & Kahneman (1974) call availability. Ac- cording to Tversky & Kahneman, decision-makers assess the probability of an event by the ease with which instances or occurrences can be brought to mind. In any strategic decision situation there are poten- tially many analogous situations which may occur to decision-makers. Which analogy will decision-makers choose? It may be that they will choose the analogy which most readily comes to mind. Thus for example, the analogy chosen may be influenced by a decision-maker's functional specialization. Further, recent experiences may provide the most readily available analogies.
Salience and Misguided Parsimony
If decision-makers do not use simple analogies to prematurely define the problem, but rather attempt to locate the real causes of the present problem, there are two possible biases which wovild make it less likely that they will be successful. These biases are dis- cussed by Nisbett & Foss (1980:115-130) under the headings of salience and misguided parsimony. According to these authors, highly visible or salient events or variables are most likely to be taken as causes, leading decision-makers into a post hoc, ergo propter hoc fallacy. The availability bias will cause these events to be most readily re- called. Citing research by Pryor and Kriss (1977), and Taylor and Fiske (1978), they argued that when decision-makers are given verbal information about events, characteristics of the message can deter- mine which aspects of the events are seen as causal factors. When decision-makers observe events directly, accidental features of the
-15-
environment or their own position in it can be important in deter- mining causal interpretations.
Regarding misguided parsimony, Nisbett & Ross suggest that decision- makers tend to believe that events have unitary causes. Because of this, they may fixate on the first plausible cause which occurs to them rather than seeking the multiple causes. Essentially, this is a satisficing approach to determining causality. This bias may have been identified first by John Stewart Mill in his discussion of "the prejudice that a phenomenon cannot have more than one cause" (1843/1974, p. 763). Nisbett and Ross also cite a statement by Kanouse which summarizes the bias:
individuals may be primarily motivated to seek a single sufficient explanation for any event, rather than one that is the best of all possible explana- tions. That is, individuals may exert more cog- nitive effort in seeking an adequate explanation when none has yet come to mind than they do in seeking for further (and possibly better) explana- tions when an adequate one is already available. This bias may reflect a tendency to think of uni- tary events and actions as having unitary (rather than multiple) causes; individuals may assume, in effect, that no more than one sufficient explana- tion is likely to exist for a single phenomenon (Kanouse, 1972, p. 131).
Alternatives Generation
After the strategic problem has been defined, the next stage in the normative model involves the generation of strategic alternatives for dealing with the problem. As Alexander (1979) points out, alter- natives may either be created or "found" through a search process. According to normative theory, a large scale search for alternatives should be undertaken at this point. This search should produce a
-16-
large number of alternatives which are then evaluated in order to select the best. However, Cyert & March (1963) and Lindblom (1959) indicate that the search for solutions to organizational problems does not meet these demands. Rather, very few alternatives are evaluated in any depth. Alexander (1979) found support for this assertion in his study of three top-level decisions. He concluded "[the three decisions'] most salient common feature is the rapid convergence of options, both in number and in range, before the formal evaluation process ever began" (1979: 396). In some cases cognitive biases may lead to a situation in which there is no_ search for alternatives.
Since the biases discussed in this section tend to eliminate the search for alternatives, they could be considered biases in the eval- uations stage. However, since alternatives generation is part of most normative models, these biases are discussed in terms of their limiting effects on alternatives generation. These biases include: single outcome calculations, inferences of impossibility, denying value tradeoffs, and problem sets (see Table 2).
Single Outcome Calculation
Steinbruner (1974) elaborates on Cyert & March's notions of problemistic search and applies it to individuals as well as organi- zations with his discussion of single outcome calculations and re- lated mechanisms identified in behavioral decision theory research. Rather than attempting to specify all relevant values and goals and all alternative courses of action as normative decision theory would suggest, decision-makers may focus on a single one of their goals'or values and a single alternative course of action for achieving it.
-17-
Steinbruner argues that, contrary to normative models of organiza- tional decision-making, uncertainty is not resolved in most instances by probabilistic calculations of the outcomes of alternatives. Rather, favorable outcomes are inferred for preferred alternatives while un- favorable outcomes are projected for non-preferred alternatives. Thus strategic decision-making involves a single-valued problem and a. single-preferred alternative to which the decision-maker is committed from the outset of the decision process (1974, pp. 122-123). This is an extremely powerful simplification bias and is more likely to be used in highly complex and uncertain decision environments. Since this bias allows decision-makers to deny the unpleasant value trade- offs which are always present in a choice between alternatives it significantly reduces the stress associated with ill-structured decision-making.
Inferences of Impossibility
Steinbruner suggests that decision-makers deal with non-preferred alternatives through inferences of impossibility. In contrast to the suggestions from normative decision theory, Steinbruner points out that decision-makers may devote a good deal of effort to identifying the negative aspects of non-preferred alternatives and attempting to convince themselves that they are not possible to implement (1974:119). Since this forces premature rejection of alternatives, it may have disastrous consequences for decision-makers who use it. They will achieve a premature closure at the possible cost of rejecting the most feasible alternative.
-18-
Denying Value Tradeoffs
In addition, as both Steinbruner and Jervis point out, decision- makers over-value their favored alternative by denying value tradeoffs (Jervis, 1976:128-136). That is, they attempt to interpret facts in such a way that the favored alternative appears to serve several values simultaneously and appears to have no costs associated with it. They attempt to deny that there are tradeoffs and that there are some values which may not be served by their favored alternative.
Problem Sets
Problem set is another decisioneil bias which has been demonstrated in laboratory research in cognitive psychology (Anderson & Johnson, 1966; Newell & Simon, 1972). Basically, problem set is demonstrated when perceiving an object serving one function makes it more difficult to perceive it as capable of serving some other function or when re- peated use of one problem-solving strategy makes it more difficult to develop other strategies (Anderson & Johnson, 1966, p. 851). Though this bias has only been experimentally demonstrated using relatively simple and discrete problem-solving tasks, a similar process may be operating when strategic decision-makers become strongly committed to a set of assumptions about the nature of their business and appropriate solutions to its problems. Mason & MLtroff (1981) have identified this as a persistent problem in corporate strategy formulation.
Evaluation and Selection
The evaluation of strategic alternatives is the phase of the decision-making process in which the limitations of Simon's "Admin- istrative Man" are most evident. As Simon (1976) points out, the
-19-
evaluation and selection stage of administrative decision-making falls short of objective rationality in at least three ways.
(1) Rationality requires a complete knowledge and anticipation of the consequences that will fol- low on each choice. In fact, knowledge of consequences is always fragmentary.
(2) Since these consequences lie in the future, imagination must supply the lack of experienced feeling in attaching value to them. But values can be only imperfectly anticipated.
(3) Rationality requires a choice among all possible alternative behaviors. In actual behavior, only
a very few of all these possible alternatives ever come to mind.
(Simon, 1976:81)
The following discussion deals with biases which may affect the activities of evaluation and selection. These include: representa- tiveness, the illusion of control, certainty effects, restriction of evaluation criteria, focus on negative criteria, and devaluation of partially described alternatives.
Representativeness
Tversky & Hahnemann (1974) have pointed out a number of individual cognitive biases which may distort judgements. The first they call the representativeness bias. This causes a decision-maker to over- estimate the extent to which a situation or sample is representative of the situation or population to which he wishes to generalize. This bias may be responsible for the fact that decision-makers tend to view strategic decisions in terms of simple analogies. It also causes them to overestimate the extent to which the past is representative of the
-20-
present and the extent to which solutions offered for problems in the past will be of value in the present problem.
Part of this representativeness bias involves insensitivity to predictability. In making predictions of the effects of various courses of action decision-makers do not take into account the extent to which the evidence for the predictions is reliable, or the extent to which the criterion is related to the cues which they use to predict it (Tversky & Kahnemann, 1974:1125),
Decision-makers are also insensitive to sample size in making predictions. Though a large number of observations are necessary in order to make generalizations to a population, strategic decision- makers are often unable to collect data on a large number of past strategies and are quite willing to generalize from a small data base. Further, they have too much confidence in their predictions from small amounts of data, feeling that these data are representative of the population as a whole. Tversky & Kahnemann call this a belief in "law of small numbers" (Tversky & Kahnemann, 1974:1125). Nisbett & Ross (1980:55-59) have suggested that decision-makers are especially susceptible to the law of small numbers when considering one or a few very vividly described cases. A single vivid description of a new ven- ture's failure in a particular industry may influence the decision about entering the industry more than volumes of statistical data indicating high success rates in the industry.
Finally, in decision-making tasks which involve high levels of uncertainty, decision-makers should be aware of this uncertainty and of their inability to accurately forecast events in the decision environment.
-21-
Their lack of ability to forecast outcomes should make them wary of be- coming overcommltted to particular courses of action and should encourage them to develop contingency plans. However, decision-makers tend to be overly confident in their ability to predict outcomes (Tversky & Kahnemann, 1974:1129) which perpetuates the illusion that they do not need to develop formal contingency plans. Einhorn & Hogarth (1978) call this over- confidence the illusion of validity.
Illusion of Control
Langer's illusion of control may also affect decision-makers' perceptions of the need for contingency plans. They will overestimate the extent to which the outcomes of a strategy are under their personal control and may assume that through additional effort they can make their strategy succeed should problems arise. Langer (1975) conducted six studies which showed that subjects making a variety of decisions expressed an expectancy of personal success inappropriately higher than the objective probability would warrant. They tended to overestimate the impact of their skill on the outcome or to overestimate their skill. Larwood & I'Jhittaker (1977) comparing management students' and executives' performance on a marketing problem found further evidence for the illusion of control. The management students tended to overestimate their abilities and the performance of the hypothetical firm of which they were sales managers. Managers also overestimated performance in this exercise but showed less tendency to do this if they had experienced unsatisfactory results in earlier planning experiences.
-22-
Certainty Effects
Another decisional bias which may be related to the desire for control over the environment is the certainty effect (Kahnemann & Tversky, 1979; Tversky & Kahnemann, 1981). Kahnemann & Tversky sum- marize a number of studies which show that people exhibit decisional biases when they evaluate outcomes which are considered certain against to outcomes which are merely probable. When given a choice between a certain outcome with a given expected value and a probablistic outcome with a slightly higher expected value, decision-makers will often choose the certain outcome. This result alone might simply be evidence of risk aversion. However, when given a choice between a certain loss of a small amount and the probability of a higher expected loss, they will tend to choose the probabilistic loss. As Kahnemann & Tversky point out (1979:269), these results are incompatible with the concept of risk aversion which holds that certainty is always desirable. Rather, it appears that cer- tainty increases the aversiveness of losses as well as the desirability of gains.
These results may be explained in terms of the desire for control over the environment if we asstime that control over the environment involves the ability to avoid negative outcomes. People will choose certainty when faced with a potential gain because the certain outcome minimizes the possibility of a zero gain which is aversive. On the other hand, when people are faced with a choice of alternatives, one involving a potential loss and one involving a certain loss, they will choose the alternative which, though rasre risky, at least gives them a chance of avoiding the loss. In other words, they are risk averse
-23-
with respect to gains and risk seeking with respect to losses. It is often possible to formulate a problem as either a choice between los- ses or a choice between gains. Thus, these authors have shown that different formulation of a problem, different reference points against which outcomes are evaluated, may result in different choices.
Criteria Restriction and Negative Forces
Wright (1974) has conducted laboratory research which reveals two additional biases which come into play under time pressure or stress. Business undergraduates made a choice among a number of car models based on a number of cues with time pressure and distractions varied. The re- sults support the claim that decision-makers restrict the number of cues they use in evaluating the quality of various alternatives. Further, they tend to pay most attention to and accentuate negative cues, those which provide evidence which will lead to the rejection of alternatives. In strategic decision-making, this could lead to the rejection of alter- natives which have important positive features but which have perhaps one negative feature (Wright, 1974:588-559).
Devaluation of Partially Described Alternatives
Finally, Yates, Jagacinski, & Faber (1978) demonstrated a pref- erence for con^leteness of information which biases decision-makers' evaluation of alternatives. Among a group of strategic alternatives, it is likely that the probable consequences of some of the alterna- tives will be more completely described than others. Yates, et al. fovmd that decision makers tend to devalue the alternative that is partially described. Since partially described alternatives involve
-24-
uncertainty for decision-makers, they tend to negatively evaluate these against alternatives which are better described and therefore, resolve more uncertainties.
CONJECTURES The biases discussed in the preceeding sections form the basis for some conjectures about probable errors in the strategic decision- making process. These conjectures, which are numbered below, suggest the specific ways in which strategic decision-making may depart from normative prescriptions.
Strategic Problem Identification:
1) Decision-makers will tend to perceive fewer gaps than their data indicate due to the prior hypothesis and adjustment and anchoring biases (Levine, 1971; Pruitt, 1961; Wason, 1960; Jervis , 1976; Tversky & Kahnemann, 1974),
2) Decision-makers will minimize the significance of gaps and not use the gaps as a basis for strategy revision due to the escalating commitment (Staw, 1976 & 1981; Staw & Fox, 1977; Staw & Ross, 1978; Fox & Staw, 1979).
3) Decision-makers will tend to attribute unfavorable gaps to chance due to the illusion of control (Langer, 1975; Langer & Roth, 1974; Lefcourt, 1973; Larwood & Whittaker, 1977).
4) If the significance of a gap is recognized, decision-makers will tend to define the problem causing the gap through an analogy to a simpler situation. Recent experience is most likely to provide the analogy (Steinbrvmer, 1974; Tversky & Kahnemann, 1974),
-25-
5) Because of the effects of salience and misguided parsimony, decision-makers who do seek the causes of a strategic problem will tend to identify a single, highly visible cause (Nisbett & Ross, 1980; Pryor & Kriss, 1977; Taylor & Fiske, 1978; Kanouse, 1972; Wilson & Nisbett, 1978).
Strategic Alternatives Generation
6) In searching for a solution to a strategic problem, decision- makers will tend to generate a single alternative rather than several alternatives due to the effects of the bias toward single outcome cal- culation (Steinbnmer, 19 74).
7) Decision-makers will tend to deal with non-preferred alterna- tives by denying that they serve any values better than the preferred alternative and by overestimating the difficulty in implementing them. This is due to the biases toward denying value tradeoffs and infer- ences of impossibility (Steinbruner, 1974),
8) Because of unchallenged assumptions and problem sets, decision- makers who attempt to generate more than one alternative will tend to generate very few (Anderson & Johnson, 1966; Newell & Simon, 1972; Mason & Mitroff, 1981).
Evaluation and Selection
9) Decision-makers will tend to over-estimate the accuracy of their predictions of the consequences of alternatives because of the representativeness bias (Tversky & Hahnemann, 1974; Nisbett & Ross, 1980).
-26-
10) Decision-makers will tend to overestimate the importance of their own actions in assuming the success of strategic alternatives due to the illusion of control (Langer, 1975; Langer & Roth, 1974; Lefcourt, 1973; Larwood & I^ittaker, 1977).
11) Because of certainty effects, decision-makers will tend to choose more certain outcomes when attempting to maximize gain and riskier alternatives when attempting to minimize loss (Kahnemann & Tversky, 1979).
12) Decision-makers will exhibit a bias toward restricting the number of evaluation criteria used and focusing on negative evaluation criteria (Wright, 1974).
13) Decision-makers will exhibit a preference for alternatives described in greater detail, even though partially described alterna- tives may score higher on the decision-maker's evaluation criteria (Yates et al., 1978).
Though most of these biases were identified in research on indi- vidual decision-makers, it is assumed that they will also operate in group and organizational decisions. This assumption is by no means universally held. Indeed, Nisbett & Ross (1980, pp. 249-254) have suggested that collective decision-making may be a way of reducing the effects of some cognitive biases. For this reason, among others, the preceding statements about the effects of the biases are stated in the form of conjectures.
However, there is some support for the assumption that these biases may actually be aggravated by group and organizational processes which serve to restrict information reaching decision-makers.
-27-
Those responsible for making the top-level strategic decisions may not be those who collect the information required for the decision. Organizational structures and processes distort the information reaching upper-level decision-makers. Crozier (1963:51) summed up the problem in the following way, "Those who have the necessary information do not have the power to decide, and those who have the power to decide cannot get the necessary information." Information passing from "experts" to top level decision-makers is subject to hierarchical distortion in both quantity and quality. This much is well-known and intuitively plausible. However, there is less discussion in the literature about the direction this bias could be expected to take.
First, since experts may act as mindguards, protecting decision- makers from potentially threatening information (Janis, 1972; Janis & Mann, 1977) information threatening to top-level decision-makers or information which reflected negatively on their past decisions may be distorted or omitted. Second, information presented by experts is often presented as part of a proposed solution to a problem or strategy. In this case, information might be distorted in favor of the proposed solution.
Carter (1971) documented the effects of the use of experts in strategy formulation when he attempted to apply Cyert & March's (1963) organizational decision-making framework to top level corporate deci- sions. He examined six top level strategic decisions and found that these decisions differed from the operational decisions described by Cyert and March in two ways. First, they tended to involve more levels of the organization and second, they involved people of more vairying
-28-
backgrounds. For these reasons, proposals for solutions to top level problems were often presented by coalitions to top-level corporate decision-makers and supported by staff experts' analysis.
Carter suggests that bias will be added to the appraisal of proposed problem solution by technical or functional staff experts to the extent that:
a) The success of the project depends on the top-level decision- makers' acceptance of the staff's representation of the relevant issues in the problem.
b) There is uncertainty in data relevant to the problem.
c) The top-level decision-makers possess much less knowledge about the problem than do the staff people.
d) The top-level decision-makers perceive a great deal of un- certainty in the problem and a need for the expertise of the staff.
All four of these conditions are likely to hold in strategic decisions.
Further, Carter suggests that the amount of bias added to data provided for the evaluation of a given decision and the amount and type of data provided will depend on the following factors:
a) UTiich data are perceived as desired by higher levels of manage- ment,
b) The amount of data necessary to gain a favorable decision.
c) The ease of developing data.
d) The extent to which the staff people or their departments will be held accountable for the consequences of decisions which were based on the data.
-29-
The actions of experts or mindguards may reinforce biases through the restriction of information necessary to adequately fonmilate the problem, to generate feasible alternatives, and to evaluate those al- ternatives,
CONCLUSIONS AND IMPLICATIONS In this paper, research on selected cognitive biases has been summarized and conjectures regarding the possible effects of these biases on strategic decision-making have been developed. In this final section of the paper, some implications for research and prac- tice in strategic management will be drawn.
Fesearch Implications
Future research should focus on documenting the presence of these biases in strategic decision-making and assessing their effects.
Uiere are those who argue that we cannot draw inferences about executives' performance at real world decision-making from students and laboratory decision-making tasks (L\igson, Braunstein, & Hall, 1981). They hold that decision-making abilities required to rise to a position of strategic responsibility and extensive experience with a variety of complex decisions produce generally high-quality decision- making performance in executives which is relatively free from bias. In support of this claim, there is some evidence to show that some groups of professional decision-makers such as weather forecasters make good use of statistical information in forecasting do not exhibit a high degree of decisional bias (Hogarth, 1975, pp. 277-278). However, these decision-makers were able to learn from their mistakes and improve
-30-
their performance over time because they are required to make numerous predictions based on clearly identified data and receive continuous relatively unambiguous feedback soon after they make their predictions (Hogarth, 1975, p. 278; Nisbett & Koss, 1980, p. 265). It could be argued that none of these conditions hold in strategic decision-making; that such decisions occur infrequently and Involve ambiguous data and possibly a disagreement about which data are relevant. Further, the feedback about the success of the strategy is often ambiguous since there may be multiple evaluation criteria applied to the company's performance, some of which may yield contradictory results and may not be available for years after the implementation.
These considerations make the existence of these cognitive biases in organizational decision-making more plausible. Further support for their existence and effects comes from the fact that several of these biases have been found to operate in private and public sector decision-makers (Steinbruner, 19 74; Staw, 1981; Larwood & Whittaker, 1977; Einhom & Hogarth, 1981),
At this point, it is tempting to suggest that researchers attempt to identify possible examples of each bias in the literature describing well-known business decision-making failures. Indeed, books such as Hartley's Marketing Mistakes (1976) and Smith.' s Corporations in Crisis (1963) offer numeroios potential examples of some of these biases. How- ever, there is a consideration which suggests that such examples may be of little value. Since most of the biases have been identified in laboratory settings very different from businesses, their existence in strategic decision-making is still to some extent a matter of
-31-
conjecture. This being the case, effort should be made to document the existence of these biases in managerial decision-making through more detailed data including interviews with managers or records of meetings. Records of business decisions such as those found in the above-mentioned books merely demonstrate behavior on the part of the managers which appears consistent with the biases and which may (or may not be) the result of these biases. These reports are insuffi- cient, by themselves, to demonstrate the biases' existence.
Future research on these potential biases should take two direc- tions. First, since many have been examined exclusively in laboratory research, an attempt should be made to document their existence and effects in field settings. Researchers may be able to identify the biases in executives' detailed descriptions of problem solving pro- cesses such as those collected by Mintzberg, et al. (1976), Field observation of decision processes may also provide insights into the effects of these biases.
A second approach would involve further laboratory research in- vestigating these biases in laboratory tasks more representative of the ill-structured problems encountered in strategic decision-making. Such concurrent laboratory and field research has been advocated in the most fruitful approach to research for several questions In strategic management (Schwenk, 1982).
Implications for Practice
It is necessary to establish that these biases do in fact exist in strategic decision-making before making strong recommendations to managers regarding techniques for avoiding them. However, practitioners
-32-
who suspect that any of these biases may be adversely affecting their decisions have a wide variety of techniques available for reducing their effects. Schwenk and Thomas (1982) have summarized research on a number of techniques for reducing cognitive biases, three examples of which will be discussed here. For improving the collec- tion and use of information in problem identification a technique called Strategic Assumption Analysis has been recommended (Emshoff & Finnel, 1978; Emshoff & Mitroff, 1978; Mason, 1969; Mason & Mitroff, 1981; Mitroff & Emshoff, 1979; Mitroff, Emshoff, & Kilmann, 1979). This technique involves the structured presentation and analysis of divergent assumptions about data relevant to a problem.
There are a variety of creativity stimulants which have been proven useful In stimulating the generation of alternative solutions (Gordon, 1961; Huff, 1980; Stanford Research Institute, 1969; Warfleld, 1975), The focus of such techniques is typically the suspension of critical processes and the encouragement of unusual associations. Finally, for improving the evaluation of alternatives, the use of the devil's advocate technique has been proposed (Cosier, 1978, 1980; Cosier & Aplin, 1980; Cosier & Rose, 1977; Cosier, Ruble, & Aplin, 1978; Herbert & Estes, 1977; Janls, 1972; Jervls, 1976; Schwenk & Cosier, 1980), This technique involves the development of critiques questioning the wisdom of a preferred alternative and the challenging of data and analysis supporting this alternative.
Schwenk and Thomas point out that dec is ion- makers must be aware of the nature of the strategic decison-making process in order to make effective use of these aids. Since the process is iterative and
-33- cyclical, techniques designed to improve performance at one stage may have impacts at other stages as well. For example, decision-makers using the devil's advocate approach to improve the evaluation of alter- natives may find that its use leads to the generation of new alter- natives and identification of new strategic problems. Of course, this can greatly increase the time and effort involved in reaching a deci- sion. Thus, the use of any of these decision aids may reduce biases at all stages of the strategic decision-making process as well as in- creasing decision time. Practitioners should be aware of this trade- off relationship in making use of them.
-34-
References
Alexander, E. R., The design of alternatives in organizational contexts. A pilot study. Administrative Science Quarterly. 1979, 24_, 382-404,
Anderson, B. F. & W. Johnson, Two kinds of set in problem solving. Psychological Reports. 1966, 19_, 851-858.
Andrews, K,, The Concept of Corporate Strategy. Revised Edition, Homewood, Illinois; Irwin, 1980,
Andrews, K. , E. Learned, C. R. Christensen, & W. Guth, Business Policy; Texts and Cases. Homewood, Illinois: Irwin, 1965
Ansoff, H. I., Corporate Strategy. New York; McGraw-Hill, 1965,
Carter, E. E. , The behavioral theory of the firm and top-level corporate decisions. Administrative Science Quarterly, 1971, 16, 418-428,
Cosier, R. A. , The effects of three potential aids for making strategic decisions on prediction accuracy. Organizational Behavior and Human Performance, 1978, 22^ 295-306.
Cosier, R. A., Inquiry method, goal difficulty, and context effects on performance and satisfaction. Decision Sciences. 1980, 11, 1-16.
Cosier, R. A. & J. C. Aplin, The application of dialectical inquiry
to strategic planning: A critical view. To appear in Strategic Management, 1980, 1_, 343-356,
Cosier, R. A, & G. L. Rose, Cognitive conflict and goal conflict
effects on task performance. Organizational Behavior and Human Performance. 1977, 19, 378-391.
Cosier, R. A., T. L. Ruble, & J. C. Aplin, An evaluation of the
effectiveness of dialectical inquiry systems. Management Science, 1978, 24^ 1483-1490.
Crozier, M. , The Bureaucratic Phenomenon. Chicago: University of Chicago Press, 1964.
Cyert, R. M. & J. G. March, A Behavioral Theory of the Firm. Englewood Cliffs, New Jersey: Prentice-Hall, 1963.
Einhorn, H. J. & R. M. Hogarth, Confidence in judgement: Persistence of the illusion of validity. Psychological Review, 1978, 85, 395-416.
Emshoff, J. R. & A. Finnel, Defining corporate strategy: A case study using strategic assumption analysis. Wharton Applied Research Center. Working paper. No. 8-78, October, 1978.
-35-
Emshoff, J. R. & I. I. Mitroff, Improving the effectiveness of corporate planning. Business Horizons, 1978, 21, 49-60.
Fox, F. V. & B. M. Staw, The trapped administrator: Effects of job insecurity and policy resistance upon comm.itment to a course of action. Administrative Science Quarterly, 1979, 24, 449-471.
Glueck, W. F., Business Policy: Strategy Formulation and Management Action. New York: McGraw-Hill, 1976,
Gordon, W. J. J., Synectics. New York: Harper & Row, 1961.
Grant, J. & W. King, Strategy formulation: Analytical and normative models . In Strategic Management; A New View of Business Policy and Planning, pp. 104-122. Edited by D. Schendel and C. Hofer. Boston: Little, Brown and Co., 1979.
Hartley, R. F. , Marketing Mistakes. Columbus, Ohio: Grind, Inc., 1976.
Herbert, T. T. & R. W. Estes, Improving executive decisions by
formalizing dissent: The corporate devil's advocate. Academy of Management Review, 1977, _2, 662-667.
Hofer, C. W. & D. Schendel, Strategy Formulation: Analytical Concepts. St. Paul, Minnesota: West, 1978.
Hogarth, R. M. , Cognitive processes and the assessment of subjective probability distributions. Journal of the American Statistical Association. 1975, 70_, 271-289.
Hogarth, R. M., Judgement and Choice: The Psychology of Decision. Chichester, England: Wiley, 1980.
Hogarth, R. M. & S. Makridakis , Forecasting and planning: An evaluation. Management Science. 1981, 22, 115-138.
Huff, A. S. , Evocative metaphors. Human Systems Management. 1980, _1, 1-10.
Janis, I. L., Victims of Groupthink. Boston, Houghton-Mifflin, 1972.
Janis, I. L. & L. Mann, Decision-Making: A Psychological Analysis of
Conflict, Choice, and Commitment. New York: The Free Press, 1977.
Jervis , R. , Perception and Misperception in International Politics. Princeton, New Jersey: Princeton University Press, 1976.
Kahneman, D. & A. Tversky, Prospect theory: An analysis of decision under risk. Econometrica, 1979, 47, 263-291.
-36-
Kanouse, D. E. Language, labeling, and attribution. In E. E. Jones, et al. (Eds.). Attribution; Perceiving the Causes of Behavior. Morristown, New Jersey: General Learning Press, 19 72,
Lang, J. R., J. E. Dittrich, & S. E. I^ite, Managerial problem-solving models: A review and a proposal. Academy of >!anagement Review, 1978, 3, 354-865,
Langer, E. J., The illusion of control. Journal of Personality and Social Psychology. 1975, 32^ 311-328,
Langer, E, J, & J, Roth, The effect of sequence of outcomes in a chance task on the illusion of control. Journal of Personality and Social Psychology. 1975, 32, 951-955,
Larwood, L. & W, I-Zhittaker, Managerial myopia: Self-serving biases
in organizational planning. Journal of Applied Psychology, 1977, 67_, 194-198,
Lefcourt, H. M,, The function of the illusions of control and freedom, American Psychologist. 1973, 28, 417-425,
Leontiades, M., Strategies for Diversification and Change, Boston: Little, Brown, & Co., 1980,
Levine, M., Hypothesis theory and nonleamlng despite ideal S-R re- inforcement contingencies. Psychological Review. 1971, 78, 130-140.
Llndblom, C. E. , The science of "muddling through". Public Administra- tion Review, Spring, 1959,
March, J. G. & H. A. Simon, Organizations. New York: Wiley, 1958,
Mason, R, 0., A dialectical approach to strategic planning. Management Science. 1969, 15, B403-B414,
Mason, R. 0, & I, I, Mitroff, Challenging Strategic Planning A-ssumptlons, New York: Wiley, 1981.
Mazzollni, R., How strategic decisions are made. Long Range Planning, 1981, 14, 85-96,
Mill, J. S, A System of Logic Ratioclnatlve and Inductive, Toronto: University of Toronto Press, 1974 (originally published in 1843).
Mlntzberg, H., Strategy-making In three modes. California Manager:ent Review. 1973, Winter, 44-53,
Mlntzberg, H., P. Raisinghani, & A. Theoret, The structure of
"unstructured" decision processes. Administrative Science Quarterly. 19 76, 21, 246-275.
-37-
Mltroff, I. I. & J. R. Emshoff, On strategic assumption-making: A
dialectical approach to policy and planning. Academy of >fenagem-ent Review, 1979, 4_, 1-12.
Mitroff, I. I., J. R, Emshoff, & R. H. Kilmann, Assumptional analysis: A methodology for strategic problem-solving. Management Science, 1979, 25_, 583-593.
Newell, A. & H. A. Simon, Human Problem Solving. Englewood Cliffs, New Jersey: Prentice-Hall, 19 72.
Nisbett, R. & L. Ross, Human Inference. Englewood Cliffs, New Jersey: Prentice-Hall, 1980.
Pruitt, D. G., Informational requirements in making decisions. American Journal of Psychology. 1961, 74_, 433-439.
Pryer, J. B. & M. Kriss. Cognitive dynamics of salience in the attribu- tion process. Journal of Personality and Social Psychology, 1977, 35., 49-55.
Schwenk, C. R. Why sacrifice rigor for relevance? A proposal for
combining laboratory and field research in strategic management. Strategic Management Journal (forthcoming).
Schwenk, C. R. & R. A. Cosier, Effects of the expert, devil's advocate, and dialectical inquiry methods on prediction performance. Organizational Behavior and Human Performance, 1980, 26, 409-423,
Schwenk, C. R. and H. Thomas. Fonnulating the mess: The role of
decision analysis and other decision aids in problem formulation. Unpublished manuscript. University of Illinois, 1982.
Simon, H. A., Administrative Behavior (Third Edition). New York: The Free Press, 1976.
Simon, H. A. , The New Science of Management Decision. Englewood Qiffs, New Jersey: Prentice-Hall, 1960.
Smith, R. A. Corporations in Crisis. Garden City, New York: Doubleday, 1963.
Stanford Research Institute. Structured Approaches to Creativity. Menlo Park, California, 1969.
Staw, B. M., Knee deep in the Big Itaddy: A study of escalating coo- mitment to a chosen course of action. Organizational Behavior and Human Performance, 1976, 16, 27-44.
Staw, B. M., The escalation of commitment to a course of action. Academy of Management Review, 1981, 6_, 577-587.
-38-
Staw, B. M. & F. V. Fox, Escalation: The determinants of commitment to a course of action. Human Relations, 1977, 30, 431-450.
Staw, B. M, & J. Ross, Commitment to a policy decision: A multi- theoretical perspective. Administrative Science Quarterly, 1978, 23, 40-64.
Steinbruner, J. D., The Cybernetic Theory of Decision. Princeton, New Jersey: Princeton University Press, 1974.
Steiner, G. A. & J. B. Miner, Management Policy and Strategy. New York: McMillan, 1977.
Taylor, R. N., Psychological determinants of bounded rationality:
Implications for decision-making. Decision-Sciences , 1975, 6_, 409-429.
Taylor, S. E. & S. T. Fiske. Salience, attention and attribution: Top of the head phenomena. In L. Berkbwltz (Ed.) Advances in Experimental Social Psychology. New York: Academic Press, 1978.
Tversky, A. & D. Kahneman, Judgement under uncertainty: Heuristics and Biases. Science. 1974, 185. 1124-1131.
Tversky, A. & D. Kahneman, The framing of decisions and the psychology of choice. Science. 1981, 211. 453-458.
Ungson, G. R., D. N. Braunstein, & P. D. Hall, Managerial information processing: A research review. Administrative Science Quarterly, 1981, 26, 116-134.
Warfleld, J. N., Methods of Idea Management. Academy for Contemporary Problems, 1975.
Wason, P. C, On the failure to eliminate hypotheses in a conceptual
task. Quarterly Journal Experimental Psychology, 1960, l^., 129-140.
Wilson, T. D. & R. E. Nisbett. The accuracy of verbal reports about the effects of stimuli on evaluations and behavior. Social Psychology, 1978, 4^, 118-131.
Witte, E, , Field research on complex decision-making processes — The
phase theorem. International Studies of Management and Organization. 1972, 59, 555-561.
Wright, P., The harassed decision maker: Time pressures, distractions, and the use of evidence. Journal of Applied Psychology. 1974, 59, 555-561.
Yates, J. R., Evaluation of partially described multiattribute options. Organizational Behavior and Human Performance. 1978, 21, 240-251.
M/C/265
01
-5
u
<u
Q
CO CO
u
w
?
§
M M U U O
U
H CO
o
CO
1-3
Q
•H
O 00
N <J\
Sd.
CO ^-^
3 ox r-} H
4-1
(U 4-1
c s
0)
•a
d
(U o ^-s
CO 00
r^
<^ ox
.H H ^-^ <U U-l
o
pa
o
CO
(3)
eg
<*-i
o u
fl ^
I
a o
•H CO
•H O (U
Q
0)
•a
o
ctt O •H
T3 C 0) (U
3 M
03
(U
CO
c o
•H 4J
n3
CJ •H «4-l •H
4J
<u
T3 1-1
<u a
a g
a o ca (u }4
H 4J -H
a. (u >
D.'O c q
00
o a <u u
CO
CO
4J
o
a. a. o
0)
.5
4J
eg u n]
i* o u
^ CO
|
>. |
CO |
|
G |
+J |
|
CO |
c |
|
CU |
CO |
|
e |
> |
|
o |
T3 |
|
y |
CO |
CJ CO Q Q
o
CO
u
to u
a
(Si
B
to
CO
CO •H CO
CO (U rH
CO M-l C -H o to
OJ
o
CO u >> 3
CO C -H iH O
^J <U > to to O.
4J -a c fi (ji to ' W < PS o
ca
CM
CO »J-
CO
>
CJ -H •H 4-)
CO to
Q) c:
4J n
to <u <u
!-i W fl
CO 5
CO 0)
>
•H 4J
Id
a o
•H
4J
CO U
(U CJ)
(J
4-1 to l4
<u o _
en VH IH
CS
c o
•H
4-1
o
<:
V4
o
<0 CO
PU
0)
• • Tj
0) -H CJ CO •H C
2 o CJ v_/ ca
CO
(U
cj.^
00 CO
(u a
4-> )4
to (U
u u
0) CO
to
Xi
4J
c:
0)
E a o i-i
(U
>
(U Q
|
^ |
C3 |
|
CJ |
00 |
|
u |
•H |
|
to |
CO |
|
<u |
<U |
|
CO |
Q |
n ,y
|
m |
|
|
(U |
|
|
> |
|
|
CJ |
•H |
|
•H |
4J |
|
00 CU |
s |
|
4-1 |
U |
|
to |
<U |
|
u |
4J |
CO ^
a o
o
rH 0)
> C 0) W CO CO
O
U CO
to (U
00 CO
•H H
4J 3
CO O
>
o u
c o
•H
4J .. - CJ _ (U CJ ■< -H > > T3
c <+-! u-i a) c -1 o o p4 a
en
M M
(U CO
to
JZ (U
Pi ^
4J ^~«
•• 0) 00
oj CO 0)
CJ o u
•H O CO
o x: 1^
^ CJ w
CJ v-' CO
5
0)
s
0)
rH
C3 O •H 4-1 CO 4J
c
M
m
o
CO
4J C
(U
c o
CO 3 i-l
|
c |
|||
|
o |
|||
|
■H |
|||
|
ti |
U |
||
|
<u |
o |
to |
|
|
M |
•H |
N |
|
|
CO |
4J |
•H |
|
|
x: |
a |
3 |
M |
|
(U |
0) |
O |
|
|
0) |
T-l |
42 |
|
|
c |
h |
CO |
4J |
|
o |
CJ |
> |
3 |
|
•H |
CO |
W |
<: |
|
4-1 |
|||
|
CJ |
••s |
•-\ |
rf—N |
|
<U |
• |
• |
• |
|
rH |
m |
vO |
r* |
|
(U |
|||
|
CO |
c:
CO u
sg
to rH
U to CO W
>x
00
4J O
CO -H V" o
*J J=
CO CJ
<
H M
>
a
M
o
M CO l-l
u
Q
O H
H CO
O
u
u
a:
CO CO
<
O 4J
•H a
4J (1)
0) U-l
0) CO
M T3
M C
0)
00 c
CO O
W T-l
CO 4J
rH CO
cd CO
> -H
H PQ
o u
•W CJ
4J (1)
CO U-i
U <J-I
<u w
c
CO
<u
>
u
M C
M U
(U
OOr-l
to < u
CO CJ
00 <u
4-1
CO CO 1.1 CO
CO m
c
O i-t
•H a
U 0)
to >*-i
CJ IM
T-l w <4H
|
01 |
|
|
•o |
|
|
M |
M |
|
0) |
s |
|
00 |
(U |
|
to |
rH |
|
iJ |
J3 |
|
CO |
O |
|
W |
|
|
PU |
00
<u u
to en
U CO
4J -H
CO eq
AJ O
to -H
3 CJ
U -H
O 'O
to (U
M C^ O
<u
CJ
c <u
3
cr
(U
CO
c o
CJ
t-l U '
^ 9i Q) c -a
o 4-1 m j3 o uj
rH Q) • e -H to
(UAJtO <4-i> iU3iJ3
4-> c o 1-1 a c: CO
to OJ iJ-i . 4J C 3
00 CO to a) 0} r-l
^ ^ >-i <u to
Q) (J >
3 CU
CO
0 o m
to CO
to to to
"tO.HtOMiH"4-i;ja)t*^VJ
a to
CJ
1 a;
4J cO 4-1 '+-I O* T3
1-icO^iCf^tOtOD.PiOO
4J
CM
|
1 to 4-1 |
1 |
|
|
c |
CO |
to |
|
Q) |
to |
c |
|
CO |
<a |
(U |
|
c 3* |
s |
|
|
a |
> |
•H |
K 4J x^ ^^
o
4-1 CO m 0) -H CO <U 3 to > C -H O
03
&a ^
^^
|
to |
|
|
Q) |
|
|
> |
|
|
to |
•H |
|
4J |
4J |
|
CJ |
CO |
|
•H |
g |
|
4J |
<U |
|
CO |
4J |
|
tS |
to |
S fi
4J O
3 tH O 4J
to
Q) rH M) O
C rH •H to CO O
|
•a |
|
|
T3 |
(U |
|
<U |
> |
|
Vl |
iH |
|
o |
(U |
|
s |
u u |
|
•H |
9i |
|
a. |
|
|
tu |
|
|
CJ |
4J |
|
c: |
o |
|
<u |
c |
|
•a |
|
|
•H |
CO |
|
^ |
o. CO |
|
oo |
I
o
^^ pa u
O CO •H -H W CO
Pd a»
<4H
o
a
O iH
•w o
CO U
3 4J
rH C
ri 8
CM
<4-l 0)
o (U
^ >
Vj O 4J
3 -H to
4J 4J C
gU )H
« (U
<U -ri 4J
>l fl) rH
ill >-l CO
|
<w |
>» |
|
o |
4J |
|
•H |
|
|
CO |
rH |
|
(U |
•H |
|
o |
J3 |
|
c |
•H |
|
01 |
CO |
|
J-! |
to |
|
0) |
O |
|
VH |
o. |
|
•s |
•a |
M
CO
CO U
4J 0)
>-l <4H
(U >4-l
U 01
c o
C -H O U
•H CO •
4J 3 CO
O rH -H
•H Id M
H > 0)
4J 0) 4J
OB -H
O; MH >H
« O O
e»i
o
0) Id
> 3
4J «
0 > .
01 01 cd
T-) -H
0 o a>
1 '<-■
d 0) -H
O to h
3 3 O
en m
<4H 14H
o
0)
(d
00 H
(3 •H <U
5> =^ C rH
0) cd
Q >
CI
|
"o) |
^ |
|
|
(4-1 |
0) |
|
|
Cd |
s |
a |
|
_tn |
OD |
|
|
CO |
(0 |
|
|
D. |
> |
|
|
tw |
to |
■H |
|
o |
J2 |
4J |
|
M |
Cd |
|
|
0) o |
0) a. |
g |
|
•H |
<u |
|
|
O |
4J |
u |
|
6 |
3 |
Cd |
0)
u
c •
<U CO
to 0)
- 0) >
C 3 ^1 -rl
td
o
C rH C
O VI 0)
•n >H O 4J
0) 4J O rH
pj CO cx Id
in
vo
|
to |
>» • |
||
|
0) • |
C rH <U |
||
|
> T3 |
B |
O rH > |
|
|
•H 0) |
o • |
tH Cd "O -H |
|
|
W *J |
c |
lU -H Cd |
4J -H a> 4J |
|
Id u |
o |
> 4J iH |
rt 4-1 j3 g 3 M -H C |
|
a -H |
•H cd n |
||
|
VI Vi |
la |
4J 3 0> |
rH cd Vl Vl |
|
0) 4J |
3 |
td rH 4J |
<d o. u 01 |
|
4-1 a |
o |
00 Cd -rl |
> to u |
|
:^2 |
o |
0) > VI |
0) VH OJ rH |
|
p^ |
C 0) o |
Q O 13 cd |
|
a |
|
|
0> •— 1 |
|
|
ca |
|
|
o |
4J |
|
Vl |
0> |
|
CU |
CO |
9
|
• |
rH |
13 Id |
|||||||||
|
* |
c |
U |
0) Vl |
||||||||
|
13 |
o |
O |
•H 4J |
||||||||
|
1 |
0) |
«k |
•H |
0) |
t4H MH ID |
||||||
|
Vl m |
• |
N |
13 |
1 |
to |
Vl |
O iH |
||||
|
flj D.'a |
0,1 |
a. |
01 |
Vl 0> |
•^ |
VI |
*J 0, |
||||
|
T3 td |
J |
4J |
Cd |
•H |
4J |
1 0> 4-1 |
o |
0) B U |
|||
|
C oo |
> |
CJ -H |
o |
00 (M |
o |
CO > Id |
0) |
u |
to 0, cd |
||
|
3 |
•H |
§.s |
a |
•H |
B |
•HO « •rl |
Vl |
13 |
3 13 •H |
||
|
» |
OJ |
<4H |
4J |
g ^-'^-v Vl |
4J |
0) |
to -H Vl • |
||||
|
(U -o |
o |
o 6 |
gs |
• o |
c |
53 13* |
13 a |
g3 |
o |
•H |
y _ & B |
|
U 0) |
Vl |
•H |
13 |
0) |
E T3 0) O |
c |
<4H |
S o o |
|||
|
C N |
« |
<4H D. |
u |
Qi CO |
13 |
0) 0> |
01 0, -rl Vl |
0) |
•H |
0) 0) Vl •rl |
|
|
a -H |
o. |
•rl CO |
4-1 |
CO 0, |
•H |
4J to |
rH C rH O. |
4J |
0) |
4J |
rH rH O- a OO^ CV ^H |
|
13 rH |
a'* |
to |
•H oa |
2-^ |
j3 -H a. o. |
td |
m |
B |
|||
|
•H -H |
4J |
u |
> 3 |
4-1 |
o <*H 6 td |
Vl |
3 |
0) |
B O d > •H Vl B a> |
||
|
^^ |
o |
■H <4H |
4-1 |
0) td |
o |
4J OJ |
Vl 0, -rl C |
4-1 |
c3 |
13 |
|
|
c: |
CO O |
CO |
M (L> |
c |
to Vl |
Ph 13 to -rl |
CO |
■H |
to O. •rl Vl |
||
|
rf'^ |
^-> |
•-N |
^\ |
•-s |
^^ |
||||||
|
CM |
tn |
(0 |
60 a |
to C |
in 4J |
vO |
r*. |
||||
|
00 |
cd |
•H |
0) o |
•rl |
|||||||
|
d |
•H |
13 |
CJ -H |
rH |
|||||||
|
•H |
P3 |
Vl UH |
3 4J |
•rl |
|||||||
|
4J VI |
OO 4-1 |
Id o |
CO -H |
>^ ^ >» |
|||||||
|
C O |
a c |
4J |
Q) d |
4J |
00 00 Cd y |
13 >» |
|||||
|
0) J= |
•rl 0) |
Vl |
Vl 0) |
C O rH C |
01 |
0) B |
|||||
|
e CJ |
'V) 5 |
0) |
13 o |
rH |
O D--H rH -H 0) |
o |
Si |
||||
|
(D ■< 3 |
CO iJ |
CU |
C -H |
O |
•rl S |
C td Cd o |
c |
||||
|
rH -H to i |
3 |
g 3 |
Vl 4J |
V* o P-1 u |
§.§^S |
0) •H |
ml CO |
||||
|
-rro |
CJ B |
rH |
a |
cd |
rH |
to Vl |
|||||
|
•o c |
^8 |
^*^ |
"-^ rH |
o |
•-S ^*\ |
0) >»^-s^-s |
Cd |
g£ |
|||
|
«3^ Cd |
to |
ja M |
u |
td ^ |
Di ^ td ^ |
CO |
CM
CO
in
•iiftJi'ii'': '!,?■■(;';■
1,'^mi'i
mmmm&^mm
ECKMAN
NDERV INC.
JUN 95 ,