Expanding Entrepreneurial Cognition's Toolbox: Potential Contributions from the Field of Cognitive Science

Do the Cognitions of Entrepreneurs Differ from those of other Persons?

This is a key question because as Mayer, Gartner, and Venkataraman (2000) have suggested, entrepreneurs’ cognitions may be viewed as independent variables that are related, in various ways, to important outcomes in the entrepreneurial process (e.g., the decision to become an entrepreneur; recognition of opportunities, and so forth; Mitchell, Smith, Seawright, & Morse, 2000).

One important line of research on this topic has focused on the specific cognitions shown by entrepreneurs (e.g., Mitchell et al., 2000; Mitchell et al., 2002). This research suggests that entrepreneurs generally have arrangement cognitions—thoughts and mental frameworks concerning the resources, relationships, and assets needed to engage in entrepreneurial activity; willingness cognitions—thoughts and mental frameworks that support commitment to starting a new venture; and ability cognitions—thoughts and mental frameworks concerning the skills, knowledge and capacities needed to create a new venture. Recent findings (Mitchell et al., 2002) suggest that entrepreneurs who are high in certain aspects of all three dimensions make better choices in hypothetical business situations than nonentrepreneurs—that is, they show evidence of having better developed and richer cognitive scripts for such situations.

Additional research concerning the question of whether entrepreneurs think differently from other persons has focused on entrepreneurial intentions—the cognitive state that precedes the decision to act (i.e., form a new venture; e.g., Krueger, 2000, 2003). This research suggests that such intentions are strongly influenced by potential entrepreneurs’ beliefs that it is desirable to start a new venture that it is feasible to do so (e.g., Krueger, 1993, 2000). The key measures in such research have been adapted from models of intentional behavior proposed by Ajzen and Fishbein and others (e.g., Azjen, 1991), in which individuals report their beliefs and intentions, and this information is then related to their decisions to start new ventures.

Additional studies have examined other aspects of entrepreneurs’ thought, such as their attributions (e.g., Gatewood, Shaver, & Gartner, 1995, 2002) and their tendency to engage in counterfactual thinking (imagining “what might have been,”Baron, 2000; Gaglio, 2002). Research methods employed in these investigations have been adapted from those used in the study of cognitive processes in other fields, so they will not be described in detail here. However, they have added greatly to our understanding of important aspects of entrepreneurial cognition

Are Entrepreneurs Subject to the Same Cognitive Biases and Errors as other Persons?

Decades of research on human cognition point to the somewhat unsettling conclusion that we are far from totally rational information processors (e.g., Kunda, 1999). On the contrary, our thinking is often strongly affected by a wide array of errors and biases—“cognitive tilts” that can lead us to faulty decisions, erroneous inferences, and unrealistic expectations. Research in the field of entrepreneurial cognition suggests that entrepreneurs are definitely not immune to such errors. They, too, appear to be subject to many forms of bias. For example, consider overconfidence—an unrealistically high belief in the accuracy of one's judgments. Several studies suggest that entrepreneurs may be especially susceptible to this error, and that this is one reason that they choose to become entrepreneurs in the first place. For instance, Busenitz and Barney (1997) compared 124 firm founders and 74 managers in large organizations and found that firm founders were more overconfident than managers. Similarly, Amit et al. (2000) interviewed 51 firm founders and 28 senior managers who investigated but did not found high technology companies in Canada. They found that the firm founders rated their chances of achieving 11 different goals significantly higher than the managers did.

Additional studies suggest that entrepreneurs are subject to other forms of cognitive bias, such as the illusion of control—unjustified belief in the capacity to influence one's outcomes. Simon, Houghton, & Aquino (2000), for instance, found that for a large sample of MBA students, the stronger the students’illusion of control (belief that their skill could increase performance even in situations where chance plays a large role) and the stronger the students’ belief in the law of small numbers (the tendency to use a small sample of information as a basis for firm conclusions), the greater their tendency to start a new venture.

Many other excellent studies have examined the role of cognitive biases in many aspects of the entrepreneurial process—especially, in the initial decision to become an entrepreneur (see, e.g., Stewart & Roth, 2001). The methods used in these studies have generally been adapted from those employed to study these phenomena in cognitive science and other fields, so there are no strong grounds for describing them in detail here.

What Cognitive Factors Play a Role in Opportunity Recognition?

Opportunity recognition is often viewed as a central aspect of entrepreneurship—and of entrepreneurial cognition. For instance, Krueger (2003, p. 105) notes that “…an orientation toward seeing opportunities” can be viewed as the “heart” of entrepreneurship. It is far from surprising, then, that many studies have sought to examine the factors that play a role in opportunity recognition, including cognitive variables (e.g., Baron, forthcoming; Krueger, 2003).

Noteworthy among this work are recent articles by Gaglio and Katz (e.g., 2001), suggesting that entrepreneurs, and especially successful entrepreneurs, may possess a schema they term entrepreneurial alertness. This is a cognitive framework that assists such persons in being alert to opportunities. More specifically, Gaglio and Katz (2001) hypothesize that persons who possess such a schema show a tendency to search for and notice change and market disequilibria, to respond to information that does not match their current schemas, and to adjust existing schemas on the basis of such nonmatching information. In addition, persons possessing well–developed alertness schemas seek to be objectively accurate, and possess more complex information concerning the nature of change, the nature of specific industries, and specific social environments. In sum, because of their complex and adaptive mental frameworks, entrepreneurially alert individuals will be more able to “think outside the box” than persons lower in alertness. These suggestions strike us as being insightful and valuable, but as yet, have not been investigated in empirical research. Methods developed in cognitive science, described later in this article may, well be useful in conducting such research.

Additional research by Shane (2000, 2001) and others has examined the role of information in opportunity recognition. This work points to the conclusion that some individuals are better able to recognize opportunities than others because they have better access to pertinent information or they are better able to utilize the information they have (e.g., Shane, 2003). For instance, in one revealing study, Shane (2000) studied eight entrepreneurs who had discovered entrepreneurial opportunities. He found that prior knowledge of a particular market increased the likelihood of discovering an opportunity in that market. Several other studies point to the same conclusion: access to information or better use of available information play an important role in opportunity recognition.

Additional findings suggest that other cognitive factors such as creativity (Hills et al., 2002) and certain aspects of learning (Corbett, 2002) are also important. Again, the methods used in these studies have generally mirrored those employed in cognitive science, so it is not necessary to describe them in detail here.

Widely Used Dependent Measures: Reaction Time and he Number and Pattern of Correct Responses and Errors

While many different dependent measures are employed in cognitive research, two are by far the most ubiquitous: reaction time and overall pattern of correct responses and errors. Reaction time is defined as the amount of time that passes between a stimulus event and an individual's overt response to the event. Depending on the researcher's interests, the stimulus could be a word, a picture, an object, or something more complex (e.g., technical specifications for a new product), and the response could be a verbal utterance, a button press, or virtually any other overt reaction that can be noted, and whose speed can be measured.

A general assumption underlying the use of reaction time is that it is an indicator of the time required to complete given mental operations, thereby making it a window through which the otherwise obscure workings of the mind can be examined. That is, although cognitive processes and the conceptual structures on which they operate may not be open to direct observation, reaction time can serve as an indirect measure of the cognitive functioning involved, providing answers to such questions as how readily a given stimulus can be comprehended, how quickly particular types of information can be processed or accessed, and how easily computations can be made and problems solved.

Reaction time data can be used to test simple hypotheses regarding the relative speed of processing of information across different groups of participants, stimulus materials, and situations (e.g., do persons with certain characteristics or experiences process certain kinds of information more quickly than others?). Alternatively, it can be used to perform more elaborate tests of the quantitative and qualitative predictions derived from complex process models (see e.g., Townsend, 2003).

Are such measures useful in terms of understanding the nature of entrepreneurial cognition? We believe that they are. For example, reaction time measures could be used to determine whether entrepreneurs prefer heuristic to systematic processing (Issue #1 above). Heuristic processing is generally faster, so one would predict shorter reaction times to various stimuli (e.g., business case scenarios) among entrepreneurs than others. Similarly, if entrepreneurs possess more fully developed knowledge structures in certain domains, they would be expected to respond more quickly than other persons to stimuli relating to these domains (Issue #2 above), and perhaps to reach decisions concerning them more rapidly (Issue #4). In short, reaction time measures—which have proven to be a very powerful tool in investigating basic questions about cognition—may also prove useful to entrepreneurial cognition researchers, and for much the same reason: such measures provide a quantifiable means for investigating cognitive processes that cannot be directly observed.

Additional ubiquitous dependent measures used in cognitive science research are the number and pattern of correct responses and errors made while performing a given task. Just as shorter reaction times may be taken as indicators of better execution of particular cognitive processes, few errors and many correct responses can serve as indicators of proficiency. Similarly, just as differential reaction times across groups of individuals and kinds of materials may be taken as indicators of differences in abilities or in the ease of processing particular types of information, so too, can differential patterns of errors and correct responses. In short, these measures may be useful in the context of investigating Issues #1 and #2 described earlier in this article (entrepreneur's preferences for heuristic versus systematic thinking, and whether their knowledge structures differ from those of other persons).

We should hasten to note that what constitutes a correct response or an error is specific to the task and materials under consideration. Similarly, a huge number of more specific dependent measures used by cognitive scientists are best understood within the context of the tasks from which they are derived. Consequently our attention now shifts from broad, general measures, such as reaction time and the overall pattern of correct responses and errors, to sets of specialized procedures and their respective dependent measures.

Procedures for Examining Conceptual Structures

The field of cognitive science has long been interested in the nature and structure of people's existing knowledge and the processes by which that knowledge is acquired, organized, and accessed. As noted earlier in this article, this topic has also been of considerable interest to the field of entrepreneurial cognition where, for instance, Mitchell and his colleagues (Mitchell et al., 2002) have measured the arrangements, willingness, and commitment cognitions of entrepreneurs and compared these with the corresponding cognitions of nonentrepreneurs. In cognitive science, many different procedures have been developed to investigate the nature and structure of knowledge possessed by individuals, and several of these may be applicable to examining the conceptual functioning of entrepreneurs. These include relatively simple tasks such as naming, in which participants are asked to name a presented word or picture, lexical decision, in which they are asked to determine whether or not a presented string of letters constitute a word, classification or categorization, in which they assign stimuli to various groupings or categories, and verification, in which they must judge the validity of a statement (e.g., robins are birds) or determine whether a stimulus meets a specified criterion (e.g., whether a pictured item is an instance of some particular category).

Additional types of tasks include listing procedures, in which participants generate lists of exemplars in response to category labels (e.g., instances of the category “fruit”) or lists of properties they believe to be true of particular entities (e.g., characteristic properties that “birds” have in common), and rating procedures in which they assess individually presented items along particular dimensions (e.g., the extent to which specific exemplars are typical of a given category) or compare pairs of items (e.g., the similarity between exemplars).

These tasks are the “tools of the trade” for cognitive scientists interested in examining the structure of individuals’ existing knowledge and the processes through which such knowledge is acquired, organized, and accessed. They are relatively easy to administer, sometimes even allowing for group testing, and each in its own way yields an important piece of the cognitive puzzle. For example, a simple listing task in which people are asked to list as many examples of a category as they can think of in a fixed period of time can be used to reveal striking consistency in the dominant items for that category and thereby point to the most accessible items (see e.g., Battig & Montague, 1969, as a classic example). Highly accessible items can be expected to influence performance in a range of other tasks, such as the way in which people use their imagination (see e.g., Ward et al., 2002). In addition, because listing responses appear to be influenced by previous experience with instances of a given concept (see e.g., Vallee–Tourangeau et al., 1998), listing data can serve as indicator of those experiential differences, and may well reveal important differences between entrepreneurs and nonentrepreneurs in terms of the knowledge they have stored in memory.

As another example, classic and recent findings from simple classification tasks have revealed differences in conceptual organization related to development (e.g., shifting from thematic groupings, such as dog and bone to taxonomic ones, such as dog and cat’ (Smiley & Brown, 1979) and domain expertise (e.g., novices grouping physics problems by surface features, such as use of an inclined plane, and experts grouping by principles, such as conservation of momentum, Chi, Feltovich, & Glaser, 1981).

In sum, several measures for assessing or mapping the knowledge possessed by individuals and, more important, how such knowledge is organized, have been developed by cognitive scientists. These methods are directly relevant to Issue #2 (whether the knowledge structures of entrepreneurs differ from those of other persons). For instance, research could examine the richness of entrepreneurs’ domain knowledge by determining their use of principles versus surface features in classifying problems from their field of endeavor. Similarly, we can envision research in which such methods are used to draw a closer bead on what types of experience are most useful to entrepreneurs at different points in the entrepreneurial process, and whether different ways of organizing existing knowledge increase or decrease entrepreneurs’ susceptibility to various cognitive errors. For instance, consider the planning fallacy—the tendency to assume that we can complete more work in a given period of time than we actually can (e.g., Buehler, Griffin, & MacDonald, 1997). It seems possible that if knowledge based on past experience is organized in terms of two categories—“failures to complete projects on time” and “successes in completing projects on time,” entrepreneurs possessing such categories might find it easier to retrieve information about past failures, and so be less susceptible to the planning fallacy. This, and many other intriguing questions can be examined through the use of measures relating to the knowledge possessed by entrepreneurs and how it is organized or structured.

Insights into these question may also shed important light on Issue #5—the question of whether entrepreneurs, because they possess richer and more fully interconnected knowledge structures in certain domains, are better at perceiving useful patterns in seemingly unrelated changes in technology, market, and demographic factors (better, in a sense, at “connecting the dots”). Methods developed by cognitive science may be useful in exploring this possibility, perhaps by applying them to both new and repeat (habitual) entrepreneurs. Differences between these groups might well be very informative.

Priming: Another Technique for Revealing the Nature of Knowledge Structures

As we noted earlier, reaction time to an individual stimulus in isolation can be informative. However, reaction time to a stimulus in the context of additional items can be even more so, and provides another valuable tool (aside from naming, listing, and verification tasks) for understanding individuals’ knowledge structures. This principle is illustrated by the technique of priming. In a priming task, two stimuli are presented in succession, with the first being referred to as the prime and the second as the target. Priming is said to occur when exposure to the prime results in some measurable change in the participant's response to the target. For example, in a task that requires individuals to indicate whether a string of letters is a word (a lexical decision task), participants might have shorter reaction times to the target word NURSE when it has been preceded by the related prime DOCTOR than when it is preceded by an unrelated word, such as BASKET. Such a result would be referred to as semantic priming because the faster responding to the word NURSE is, presumably, due the fact that DOCTOR and NURSE are related to one another in terms of meaning. As a result, the word DOCTOR primes reactions to the word NURSE (e.g., speeds responses to it).

Priming phenomena are informative because they can reveal the existence of relationships between the prime and the target, as well as the nature of this relationship. This, in turn, helps us to “map” the nature of underlying knowledge structures. Primes and targets can be related to one another semantically (in terms of meaning; e.g., DOCTOR NURSE), associatively (in terms of being associatively linked in memory; e.g., LOCK, KEY), orthographically (in terms of spelling; e.g., POUR, TOUR), and phonologically (in terms of how they sound; PHIAL, FILE; see e.g., Neely, 2003). To the extent that such prime–target relations affect responding, it can be concluded that the property in question is important in the processing or representation of task information for the individual or group being tested. For example, the fact that responses to the word NURSE are faster following exposure to the word DOCTOR than following exposure to the word BASKET suggests that some type of conceptual, meaning–based connection exists between DOCTOR and NURSE whereby access to the former facilitates access to the latter. Because DOCTOR and NURSE are neither orthographically nor phonologically similar, the facilitation could not be attributed to those factors, and would instead indicate a semantic link between these words. In this way, priming effects can be used to reveal the nature of knowledge structures—how, in essence, various kinds of information are related or interconnected in our cognitive systems. For instance, priming techniques could be used to determine whether entrepreneurs differ from other persons in terms of key concepts such as “opportunity,”“risk,” and even “effort.” With respect to effort, for example, this concept may be more closely linked semantically to such concepts as “success” and “personal satisfaction” than is true for nonentreperneurs, thus helping to explain why entrepreneurs are willing to work 100 hours or more each week. Similarly, entrepreneurs—and especially repeat entrepreneurs—may have stronger links between the concept of “opportunity” and such additional concepts as “practicality” or “marketability” than other persons. These and many other intriguing hypotheses can be investigated through appropriate adaptations of priming techniques, and the results of such research might provide useful “maps” of entrepreneurs’ knowledge structures, including the way in which they organize knowledge about specific domains (e.g., particular industries, markets, and so forth). In other words, priming techniques could provide a valuable tool for mapping the knowledge structures of entrepreneurs (Issue #2 above), how they make decisions (Issue #4), and their ability to recognize complex patterns of change in the external world (Issue #5).

Memory Measures

Information we have acquired through life experiences is stored in memory—our cognitive system (or, really, systems) for storing information. In some situations it may be of interest to know how much information people can remember following exposure to it (for instance, what do they remember after hearing a presentation at a business meeting or after reading a business plan?). Cognitive researchers have developed a variety of paradigms for measuring such retention of information. Consider, for example, a situation in which a participant is shown a set of 20 or so individual stimuli, such as words or pictures, one at a time with instructions to try to remember the items for a later memory test. One way to assess the person's memory for the material would be to administer a free recall test in which the person would be given some period of time to recall as many items as he or she could with no restrictions on the order in which the items are recalled.

In addition to using the total number of items recalled as an overall measure of memory, free recall data can also be used to examine variations in performance for items occurring at particular positions on the lists of to–be–remembered items. Such data often reveal better recall for items at the beginning (primacy effect) and end (recency effect) of the lists. An early interpretation of these effects is that the primacy items were being retrieved from a long–term storage bin, whereas more recently presented items were being retrieved from a short–term storage bin. Although this “bin” approach gave way to interpretations based on how the early versus late items are processed (see e.g., Craik & Lockhart, 1972), the distinction between a short–term memory or working memory system (see e.g., Baddeley, 1986), which holds onto information briefly versus a long–term memory system, which holds onto information for longer periods of time is till very much alive.

As we noted earlier, however, working memory is far more than a short–term storage system; it also plays a key role in attention. Specifically, this memory system includes a mechanism that permits some information to enter into long–term storage while other information is largely ignored. This suggests that the more efficiently working memory operates, the better an individual's ability to focus on important information and to ignore extraneous information. One measure of the efficiency or performance of working memory is its capacity. While short–term memory capacity refers to the amount of information people can retain for brief periods of time, working–memory capacity refers to the ability to use attention to maintain or suppress information—in essence, to pay attention to what it is important to remember. One measures of working memory capacity is obtained in the following manner. Individuals read sentences out loud, with each sentence being followed by an unrelated word. After the last sentence–word combination is read, they try to recall the list of unrelated words. The more they can recall, the higher their working memory capacity.

Research findings indicate that the higher individuals’ working memory capacity, the better their performance on complex cognitive tasks such as following directions, understanding complex written passages, writing, reasoning, and even writing computer programs (Engle, 2002). In other words, the ability to focus one's attention on what's important is related to an important aspect of human intelligence—the abilities to think and reason (known as fluid intelligence).

Evidence that working memory capacity is closely related to being able to focus attention is provided by research using a procedure known as the Stroop task. On this task, individuals read the names of various colors (e.g., red, green) printed in ink that is either the same color as the word (red ink for the word “red”) or a different color (green ink for the word “red”). Their task is to name the color of the ink while ignoring the color names. If the word and ink color are different, many errors occur: people say the word rather than name the color of the ink. Performance on this task reflects the ability to focus attention on what's important—in this case, the color of the ink—while ignoring extraneous information (in this case, the color names). Recent studies (e.g., Kane & Engle, 2001) indicate that the higher individuals’ working memory capacity (as measured by procedures such as the ones described above), the better they do on this task. In sum, this and related measures provide an avenue for investigating Issue #3, which relates to entrepreneurs’ capacity to focus their attention on relevant information, and this, in turn, relates to Issue #2 (entrepreneurs’ knowledge structures) and Issue #4, their decision–making abilities. Presumably, the more effectively entrepreneurs can focus their attention on pertinent information, the more readily can such information enter into memory, and the better, ultimately, will be their decisions.

The data from free recall tasks can also be examined for a tendency to organize the information according to particular principles. For example, if the lists contain several items from the same category, do participants tend to recall those items together? Such a tendency would reveal a sensitivity to that categorical information on the part of the participant, and there may be theoretical grounds for expecting entrepreneurs to be more sensitive to different types of organizing principles than other persons. For instance, to the extent entrepreneurs actively search for opportunities, they may tend to organize information to which they are exposed according to the extent to which it suggests such factors as unmet needs in the marketplace, the possibility of new products or services, limited current competition, and so on. In other words, data concerning what types of information entrepreneurs remember best may provide important insights into the way in which information is structured in their cognitive systems (Issue #2 above). This, in turn, can contribute to our understanding of entrepreneurial cognition.

An alternative procedure for testing memory is a recognition task in which the items from the original list are shown again along with new items that were not part of the original list. Participants in such research are asked to identify whether the items are old ones from the original lists (items they have seen before), or new ones. Scoring could make use of the principles of signal detection theory, in which correctly saying “old” to the original items is considered a hit (a correct identification of a stimulus), while saying “old” to a new item is a false alarm. Similarly, saying “new” to an original item is a miss, and saying “new” to a new item is a correct rejection. In addition to revealing overall differences in memory performance, such procedures might also be used to reveal different propensities toward certain types of responding (e.g., a bias to interpret new information as familiar or vice versa). Perhaps successful entrepreneurs are less likely to fall prey to such memory errors than unsuccessful ones—a hypothesis that can be readily investigated in future research using these methods, and which is related to Issue #4 above (decision making by entrepreneurs).

Decision–Making and Reasoning

In addition to their attempts to come to grips with how people comprehend, store, organize, and retrieve information, cognitive scientists have also focused considerable attention on how people use their stored knowledge for making decisions and in reasoning about situations. We will highlight a few examples of this type of research that seem especially pertinent to understanding entrepreneurial cognition.

In decision–making or choice tasks, participants are typically presented with sets of alternatives from which they are to choose the most preferred course of action, such as which car to buy, which apartment to rent, or which gamble to opt for. The alternatives can vary along specified dimensions in such a way that conflict (in choosing between them) is created by making all alternatives relatively attractive on some dimensions and unattractive on other dimensions. For instance, apartment A might be close to campus and relatively large, but expensive and in poor condition, whereas B might be distant and small, but cheaper and in better condition. Moreover, these paradigms can be varied to produce relatively simple or relatively complex decisions by manipulating the number of alternatives that must be considered and the number of dimensions along which those alternatives vary.

By noting what type of information participants seek about the alternatives, the order in which they seek that information, and their ultimate choice, decision–making paradigms can be used to draw conclusions about the extent to which people use particular types of strategies. For instance, do they assign utilities to each level of a given dimension (e.g., large is assigned a higher utility than small), weight the dimensions by their relative importance (e.g., proximity to campus is assigned more weight than cost), sum the resulting information across dimensions for each alternative, and choose the alternative with the highest result, as suggested by multi–attribute utility theory? Or, do they satisfice, considering alternatives one at a time until one that meets a minimal level of acceptability is found? Alternatively, do they engage in elimination by aspects, considering each dimension in turn, eliminating any alternatives that do not reach a minimum level on the dimension until only one option remains? Considerable research suggests that the answer to these questions depends on the complexity of the decision situation. With many alternatives and dimensions, people seem to simplify and drop alternatives through procedures such as elimination by aspects and then use more compensatory, additive type strategies to choose among the remaining smaller set (see e.g., Payne & Bettman, 2003, for a review). Given the complexities of decisions that confront entrepreneurs, it would be of interest to assess the extent to which they adopt these simplifying decision–making strategies—and the effects of such tactics on the quality of their decisions. Such research would employ the tasks and methods outlined above, and would be directly related to Issue #4 (the question of whether entrepreneurs reason or make decisions differently than other persons).

Generative Tasks: The Cognitive Bases of Creativity

A small but growing literature in cognitive science has begun to focus on the generative aspects of cognition—aspects of cognition by which individuals move beyond their existing knowledge to produce novel creations. Such work has clear relevance for understanding entrepreneurial cognition. Procedures that have been used to study this topic include having people engage in fairly typical creative activities, such as writing or telling stories, constructing collages, and making drawings (e.g., Amabile, 1979; Amabile, Hennessey, & Grossman, 1986; Niu & Sternberg, 2001; Pavlik, 2002), as well as having them engage in creative generation tasks to develop ideas for a variety of novel products, including businesses, ads for fictional products, designs for mechanical devices, and imaginary animals, fruit, tools, and toys (e.g., Goel & Pirolli, 1992; Jansson & Smith, 1991; Mumford et al., 2002; Smith, Ward, & Schumacher, 1993; Ward, 1994; Ward et al., 2002). Products from these types of generative tasks can be examined for their originality or practicality, for links to their creators’ existing knowledge structures, or to answering more detailed questions about situational, group or individual differences in creative functioning. Since the generation of new ideas is often considered to be an important aspect of the entrepreneurial process (e.g., Ward, forthcoming), such research, and the methods it has developed, may be directly relevant to the field of entrepreneurship.

Creative generation procedures have been used successfully with a wide range of populations, including young children (e.g., Cacciari, Levorata, & Cicogna, 1997), gifted adolescents (e.g., Ward, Saunders, & Dodds, 1999), college students (e.g., Ward, 1994), and architects and design engineers (e.g., Goel & Pirolli, 1992; Jansson & Smith, 1991), indicating that they are broadly applicable and could readily be used to answer questions about the generative thinking of entrepreneurs. In addition, creative generation studies have begun to identify and provide answers to key questions about creative functioning, which may be directly relevant to understanding entrepreneurial cognition. These include questions about how the structure of people's concepts influences their new ideas and product designs (Issue #2), how strategies of accessing that information affect performance, how exposure to examples of existing ideas (or products) can constrain originality, and how beliefs and cultural factors influence performance.

For example, one of the most striking findings of such research, and one with important implications for entrepreneurial cognition, is that exposure to examples of previous ideas or work can greatly restrict creative thought. Even when participants are told to avoid copying the examples to which they are exposed, properties of those examples tend to be incorporated into their designs, and this holds true for professional designers as well as for college students (Jansson & Smith, 1991; Marsh, Landau, & Hicks, 1996; Marsh, Ward, & Landau, 1999; Smith, Ward, & Schumacher, 1993). Similar studies with entrepreneurs could assess their tendency to either incorporate recently encountered information into the novel ideas they develop or reject it. These findings are related to Issue #1, the question of whether entrepreneurs tend to prefer heuristic to analytic thinking. If prior examples stimulate heuristic thinking (e.g., a tendency to incorporate features of existing products into new ones), they may prevent entrepreneurs from engaging in the analytic thought that may often be necessary for creative cognition.

In addition, research with college samples reveals that accessing information at more abstract, principled levels leads to greater originality in forming new ideas (e.g., Ward et al., 2002), and such studies could be adapted to investigating Issue #2 and the questions of whether entrepreneurs possess richer knowledge structures and use them more effectively. Related studies, in which the originality of students’ ideas has been increased by a simple instructional manipulation to consider abstract principles (Ward, Patterson, & Sifonis, forthcoming), could also be adapted to determine whether or not the kind of entrepreneurial cognition that underlies the development of new ideas can be similarly enhanced. In addition to serving a research purpose, such studies could have far reaching practical implications. For instance, the cognitive science studies provide a hint that encouraging people to use principle–based knowledge to develop more abstract business ideas (e.g., “supplier of entertainment”) might lead to greater long–run success than accessing more specific ideas (e.g., “DVD rentals”) which are more susceptible to the vagaries of technological development. Other cognitive science studies point to the importance of conceptual combination as a source of new ideas (e.g., Estes & Ward, 2002; Wilkenfeld & Ward, 2001) and to analogy and a range of other generative processes that could be adapted for basic and applied research in entrepreneurial cognition (e.g., Ward, Smith, & Vaid, 1997).

Behavioral and Neuropsychological Measures

Thus far, we have concentrated mostly on “indirect” measures of cognition, in which task performance is used to draw conclusions about the nature of underlying cognitive representations and the mental processes that manipulate them. It should be noted that cognitive scientists also use measures, which, if not more direct, nevertheless provide more tangible data. These include recordings of eye direction and eye movements during task performance to provide a concrete indicator of where a participant is directing attention, and a growing number of procedures for assessing brain functioning. One measure of brain activity is electroencephalography (EEG) or the recording of the electrical activity of the brain by way of electrodes placed on the scalp. An alternative is functional magnetic resonance imaging (FMRI) through which an image of the brain is produced while the participant performs a task in an MRI machine, which passes a magnetic field through the brain. Resultant changes in oxygen atoms provide evidence of relatively more or less activity in certain areas because more active areas draw more oxygenated blood. Positron emission tomography (PET) also reveals brain activity, but by way of glucose consumption as evidenced through the rate of metabolizing a radioactive form of glucose.

Together, these methods have revealed much about cognitive processes. For instance, they suggest that emotional information (e.g., the tone of another person's voice or facial expressions) is processed primarily in the right rather than the left cerebral hemisphere (e.g., Heller, Nitschke, & Miller, 1998). In addition, there appear to be important differences between the left and right hemispheres of the brain with respect to two key aspects of emotion: valence—the extent to which an emotion is pleasant or unpleasant, and arousal—its intensity. Activation of the left hemisphere is associated with approach, response to reward, and positive affect (i.e., feelings), while activation of the right hemisphere is associated with avoidance, withdrawal from aversive stimuli, and negative affect (Heller, Nittschke, & Miller, 1998). Further, anterior (frontal) regions of the hemispheres are associated primarily with the valence (pleasant/unpleasant) dimension, while posterior regions are associated primarily with arousal (intensity). These findings have important implications for understanding the neural basis of various psychological disorders, and could also be relevant to efforts to understand the nature of entrepreneurial cognition. For instance, research could be conducted in which entrepreneurs and other persons are asked to read descriptions of opportunities or other business situations. Presumably, the entrepreneurs (and perhaps especially entrepreneurs who are highly experienced) will show greater activity in their left cerebral hemispheres (indicative of positive reactions) than other persons in response to opportunities that have been rated as excellent by panels of experts; further, they may also show signs of more intense reactions to excellent opportunities for new ventures than to poor ones. Such research would be relevant to Issue #5 (recognition of complex patterns). Additional research could examine activation of other regions of the brain in order to assess predictions relating to Issue #2 (differences in knowledge structures between entrepreneurs and others). Due to their specialized nature, studies using these approaches will almost certainly require collaboration between entrepreneurship researchers and researchers in neuroscience who are expert in their use and interpretation.