Prior-knowledge and opportunity identification

Kirzner (2005, p. 76) emphasizes a distinction between information and knowledge, which states that the former is an input in a process of learning that results in knowledge (i.e. subjectively perceived and processed information). By implication, any information an individual possesses is necessarily subjectively perceived and processed in one way or another (i.e. there is no such thing as a “raw datum” in the mind of an individual). While our focus is on how differences in information affect the kind of opportunity one may discover, we recognize that ultimately, we are examining how differences in knowledge matter across individuals for the kind of opportunities they discover. Indeed, entrepreneurship in Kirzner’s work is fundamentally a theory of perceived information (i.e. knowledge) that may lead to opportunity recognition through the process of alertness. In other words, some knowledge may lead to entrepreneurial discovery, while some other may not. In all cases, Kirzner focuses on knowledge (i.e. perceived information), not simply on “raw” information. Given the links between information (i.e. “raw data”) and knowledge (i.e. subjectively perceived and processed information) and our goals in this paper, however, we do not need to emphasize the distinction between the two.

See, for instance, Ardichvili et al. (2003); Baron (2004); Buenstorf (2007); Dimov (2007a, b); Fiet and Patel (2008); Gaglio and Katz (2001); Kaish and Gilad (1991); Plummer et al. (2007); Shane (2000); Shane and Venkatarman (2000); Shepherd and DeTienne (2005); Tang et al. (2008).

See, for instance, Lavoie (1991), Chamlee-Wright (1997), Storr (2004, 2012), Tominc and Rebernik (2007) as well as Storr and Butkevich (2007) for a discussion of how culture affects the opportunity recognition.

Another way of stating it is to say that prior knowledge is about directing one’s gaze not enhancing alertness.

Shepherd and DeTienne (2005, p. 93) give the following example: “For example, Ed Pauls, the creator of NordicTrack, epitomizes how prior knowledge allows individuals to identify opportunities. Ed was a trained mechanical engineer whose passion revolved around cross-country skiing. His passion was unfulfilled when inclement weather prevented him from going skiing. Thus, he invented an indoor cross-country ski machine.”

We should also note that many, if not most, economic experiments involve some degree or component of alertness (entrepreneurship). Unfortunately, since the entrepreneurial process is so rarely studied explicitly, we can infer little by way of causal relationships in those experiments. For example, some of the discovery of specialization and exchange found in Kimbrough et al. (2008) must undoubtedly be entrepreneurial. Additionally, there have been several economic experiments endeavoring to tease out the characteristics of successful entrepreneurs (see, for instance, Burmeister and Schade 2007).

Since the only action of interest to us is discovery, rather than profit maximization or total welfare, details of the cost functions and diminishing marginal valuations employed are not relevant, and have therefore been omitted for the sake of space, focus, and simplicity. Selling fruit in the village did produce the highest profit, but that could not be known until after the discovery had occurred. Similarly, relative earnings depended on how early discovery occurred, if at all. Complete cost and payment schedule details are available by request.

Although we did not screen for differences in K _T, our design eliminates systematic effects resulting from differences in K _T by randomizing treatment assignments.

We use a t test here simply for its familiarity and robustness to non-normal distributions. For a non-parametric alternative, the Wilcoxen Mann–Whitney yields p = 0.049. Table 2 relies on the WMW because the sample sizes there are not large enough to justify the t test. All statistics were generated by STATA 10; code and data are available on request.

We do not provide formal content analysis here. Nor can we correct for the unfortunate fact that surveys done after the experiment are necessarily soliciting retrospection (Gaglio and Katz 2001). We cannot determine whether participants’ ex post discussions of why they behaved as they did actually map to ex ante motivation for their actions or if they are merely ex post rationalizations. As a matter of future research, we would like to have in-process descriptions of activity, though we expect such invasive observation will have strong effects on the observed. The full list of survey responses is available upon request.

Gender, being binary for our sample, is the only category variable with enough observations and variance for us to comment on. We could have added additional controls on group composition differences, but as we discussed earlier, we are not investigating a type identification algorithm, nor do we have any theoretical impetus for what sorts of secondary observations might matter. With hundreds of additional data points, we might get away with some data mining, but we feel such an enterprise is better suited for a different type of study.

The inference may seem unwarranted, but since K _t is a part of K _T, all else equal, we believe that it is logically correct to surmise that what is true for K _t is also true for K _T.

As one reviewer pointed out, neither treatment would be considered a control group in the classic sense, in that both groups are given some sort of prior knowledge. Our experiment is therefore more analogous to a test of health effects from various levels of dietary salt intake. In such an experiment, one might be concerned that forbidding one group from having any salt at all over a prolonged period would likely have adverse effects. Therefore, the difference in salt intake between groups would be the variable of interest. Similarly, after careful consideration, we decided that reading no story at all would confound our results, since both the contextualizing aspect of a story and the prior knowledge itself would be at play. We opted instead for equivalent depths of story emersion, but with opposite vectors of relevant prior knowledge. It’s the difference that matters, though of course the difference is impossible to measure directly.

“Three major dimensions of prior knowledge are important to the process of entrepreneurial discovery: prior knowledge of markets, prior knowledge of ways to serve markets, and prior knowledge of customer problems” (Shane 2000, p. 452). Also Shepherd and DeTienne (2005, pp. 104–105) state that “it is important for entrepreneurship scholars investigating the relationship between prior knowledge and the identification of opportunities to distinguish between types of prior knowledge.”

See Ward (2004, p. 175), for instance, “Sometimes knowledge provides a bridge to the next new development and sometimes it becomes a fence that blocks our path.”

See Dimov (2007a, p. 576) who state that “Individuals’ prior knowledge of the opportunity domain increases their likelihood of acting on their initial opportunity insights only when their style of learning is compatible with the situation at hand.”