With the growing recognition of the informational basis of entrepreneurial discovery (Fiet, 1996, 2000; Gaglio, 1997; Gilad, Kaish & Ronen, 1988; Shane, forthcoming; Shane & Venkatraman, 2000; Venkatraman, 1997), knowledge-based competencies are becoming an increasingly important research topic. This shift is occurring partly because the study of context variables by themselves does not seem to offer as much normative promise to aspiring entrepreneurs. Because most new organizations discontinue while emerging (Katz & Gartner, 1988; Vesper, 1983), the study of how competence may influence venture continuity is critically important. There is some evidence that these discontinuances may be due to their being launched by individuals who have fewer employment options (Amit, Muller & Cockburn, 1995. If scholars were actually studying incompetent efforts to launch new ventures, this would confound their research findings and lead to misleading advice for aspirants.

In this research, we focus on knowledge-based competencies that we hypothesize can be used to form new firms. We define entrepreneurial competence as the knowledge to discover and exploit new ways to create wealth. A discovery is valuable information about a potential innovation, such as the founding of a firm, the creation of a new product line, the development of a new technology, the satisfaction of an ephemeral market need through arbitrage, or the like. Knowledge is the capacity to store and systematically recall informational inputs, in this case as they relate to the discovery and exploitation of a valuable venture idea. Exploitation consists a series of actions intended to realize a venture idea’s commercial value, usually through the creation of an organization to provide complementary assets (Teece, Pisano & Shuen, 1997). The provision of complementary assets typically requires both entrepreneurial competence and resources.

It is not enough for entrepreneurs to formulate strategies based on the possession of valuable and rare resources; any platform for competition must also be costly to imitate. Entrepreneurs can use private, knowledge-based competencies to implement strategies that will be costly for others to imitate (Barney, 1997). We are interested in how potentially rare, inimitable characteristics of the underlying knowledge may serve as the basis for forming new firms. Our interest stems from the supposition that the capacity of an entrepreneur to implement certain strategies depends on his or her underlying competencies, which suggests that entrepreneurs are not equally equipped to take advantage of every venture idea. Thus, entrepreneurs differ in their competencies and our understanding of these knowledge differences will likely affect our ability as educators to improve student performance.

The Specificity and Quantity of Information Needed to Be Competent

Fiet (1996) explored the informational basis of entrepreneurial discovery. Building on Hayek’s (1945) seminal work, he argued that the most valuable type of information to evaluate a venture idea is specific information. Specific information only pertains to people, places, timing, technology and special circumstances. We expect that those who know specific information will be more competent to discover and exploit valuable venture ideas than those who do not know it. However, according to Hayek (1945), because entrepreneurs can only use it to assess one venture idea, its acquisition can become a sunk cost. Entrepreneurs possessing specific information hold an option on its exploitation because its inalienability makes it costly to convey to others who might want to appropriate its value as if it were a public good (McGrath, 1999; Jensen & Meckling, 1992). Thus, entrepreneurs making discoveries may be able to keep them secret during their exploitation, which would extend the period during which they can earn above normal returns (Rumelt, 1988).

If the acquisition of competencies based on specific knowledge were not costly and risky, one would expect that all entrepreneurs would be competent. According to Fiet (1996), “both the expected returns and risks from investing in specific information are dependent upon the quantity and specificity of the information that is acquired” (p. 426), which calls for entrepreneurs to make trade-offs. Theoretically, there should be an optimal combination of the quantity and specificity of information upon which knowledge-based competencies depend. Some scholars have correctly noted that it is impossible to conduct an optimal search if we assume that the search domain is unknown (Kirzner, 1997; Shane & Venkatraman, 2000). However, serial entrepreneurs report that they actually narrow their search of the environment after scanning widely (Gaglio, 1997; Kaish & Gilad, 1991), which Fiet, Piskounov & Gustavsson (2000) suggest could be to certain, known-to-them information channels (Marshak, 1971; Marschak & Miyasaw, 1968). Limiting entrepreneurial scanning to particular information channels resolves the theoretical and mathematical difficulties with optimization (Fiet, Piskounov & Gustavsson, 2000). Nevertheless, the empirical questions remain as to whether the quantity and specificity of information acquired by entrepreneurs influence their success in forming new ventures. Based on the above assumptions from informational economics and information theory, we suspect that they do.

Hypothesis 1: The specificity of information used by entrepreneurs (the specificity of their knowledge-based competencies) will be positively related to firm formation.

Hypothesis 2: The quantity of specific information used by entrepreneurs (the quantity of their specific knowledge-based competencies) will be positively related to firm formation.

Procedural and Declarative Knowledge

In addition to its specificity, other knowledge differences may influence an entrepreneur’s competence to form new firms. Cognitive psychologists subdivide knowledge into procedural and declarative knowledge (Anderson, 1990). Procedural knowledge consists of knowledge about how to do things. Declarative knowledge consists of general rules about the world. We suspect that these dimensions are less important than whether the procedural and declarative knowledge is itself specific. Any type of knowledge that is general in nature is also alienable and cannot serve as a basis for competitive advantage. It is important for competitive information both to apply narrowly so that many others do not possess it and for it to relate specifically to the sources of a venture’s competitive advantage.

Forming a new firm entails knowledge of how to do things. Katz and Gartner (1988) and Reynolds (1997) inferred that certain activities were necessary components of firm emergence. We derived the start-up activities listed in Table 1 from the work of these earlier researchers. We use them as indicators of procedural competence. We assume that as entrepreneurs perform more of these start-up activities that they will become more procedurally competent, which will enable them to form more new firms.

Hypothesis 3: Competence based on procedural knowledge will be positively related to firm formation.

Higher education may be the most common source of declarative knowledge, which suggests that its possession should not be a source of competitive advantage nor successful firm formation. In addition, as policy makers have asked so many times before, education for what? Ideally, declarative knowledge would be valuable because it consists of basic scientific principles that entrepreneurs would use to develop new technologies. It is more difficult to see how higher education would convey procedural knowledge, except in a laboratory or internship setting.

The linkage is clearer when we think about an engineering or scientific graduate school education because aspiring entrepreneurs could use it as a basis for developing new technology. In the case of a non-technical liberal arts education, for example, how would it provide entrepreneurs with a competitive advantage? It may be that an advanced education, regardless of the discipline, prepares aspiring entrepreneurs to process the paperwork connected with forming a firm and generating initial sales. So if we focus on establishing a firm, it seems likely that those with higher education would possess a temporary competitive advantage in processing paper work. Another way to interpret higher education would be as a complementary investment in new knowledge that might be helpful in marshalling needed resources for start-up. Formally, stated hypothesis 4 is:

Hypothesis 4: Competence based on declarative knowledge (typically acquired through higher education) will be positively related to firm formation.

Competence and High Technology Ventures

One reason that high technology ventures have created so much wealth recently may be that entrepreneurs can compete using specific procedural and declarative knowledge related to their innovations. Thus, the introduction of new technology can shift the balance of competitive power within an industry by generating a rare or unique source of competitive advantage (Barney, 1997; Schumpeter, 1934).

For a while, those who attempt to commercialize high-tech ideas may be the only competitors who possess the requisite knowledge. Others may obsolete an innovation by introducing an improvement (Schumpeter, 1934). However, meanwhile, an innovation can serve as an effective competitive platform and may lead to the creation of new wealth (Teece, Pisano & Shuen, 1997). Again, the knowledge related to a high technology venture may serve as the basis for a successful venture, not solely, because its informational basis is valuable, but because its informational basis is rare and inalienable (Jensen Meckling, 1992), which may endow its possessors with a temporary, quasi monopoly over its exploitation (Barney, 1997).

Hypothesis 5: Competence based on knowledge of high technology ventures will be positively related to firm formation.

The population for this research consisted of all Swedish households. Our research objective was to track a random sample of pre-launch venture founders as they attempted to establish new firms. Following the approach developed by the Entrepreneurship Research Consortium in the U.S.A., through screening interviews, we identified an unbiased estimate of Swedish households in a sample of 30,437 Swedish households.

We conducted the household screening interviews during the summer of 1998. These interviews identified a sample of 407 independent start-up efforts. Subsequently, every 6 months, we conducted a telephone interview with most of the initial respondents who also provided follow-up information through two mailed surveys. We now have 18 months of data and a resulting sample consisting of 213 independent, on-going start-up attempts. The sample shrinkage occurred as a result of missing data in the follow-up surveys. Comparing those who answered the mail surveys with non-respondents showed no significant differences. We obtained information about firm formation by consulting with venture champions. A champion was someone who had simultaneously initiated or completed two of the start-up activities listed in Table 1 (c.f., Reynolds, 1997). From the beginning, we suspected that champions might switch organizations.

Dependent Variable

The dependent variable was new firm formation. We classified an event as a firm formation if money was exchanged (Katz & Gartner, 1988). Surely, monetary exchange is not the only or a perfect measure of firm formation. In fact, a consensus on a single measure or a list of measures is difficult to reach. Katz & Gartner (1988) suggest the following multiple indicators of firm formation: intentionality, resources, boundary and exchange. Perhaps, the most measurable of their four indicators is exchange or the generation of income. Reynolds and Miller (1992) concur with the use of income as the most suitable exchange measure of firm formation, if only one indicator of exchange is used.

We operationalized firm formation using the question, “Have you received any money, income or fees from the sale of goods or services?” If so, we also asked that respondents provide the year and month that money was first exchanged. Of the 213 start-ups in the sample, 102 earned income and 54 start-ups earned no income during the period. Those reporting no income were treated as censored data.

Covariates

We measured information specificity with an evenly weighted scale question: “Would you say that serving those missed by others is insignificant, marginally important, important, or critical for a new firm to be an effective competitor?” Market information may be the most important type of specific information for a successful start-up (Fiet, 1995; Von Hippel, 1986), yet if it reveals that a venture start-up cannot survive, the acquisition of the information itself can become a sunk cost. We measured the quantity of specific information with the question: “What percentage of this information would be worthless if the start-up fails?” Using a continuous measure enabled us to control for size effects. We measured procedural competence as the sum of the number of start-up activities competed by a venture. These activities are listed in Table 1. We measured declarative competence with the following question: “Which is your highest level of completed education? Possible responses ranged from pre-school to Ph.D. We also collected responses on the type of higher education. Possible types of higher education were humanistic/social, natural sciences, engineering, economics/business, and general education. If education were a source of declarative competence, it would be useful to understand which type (s) was/were the best preparation to form a new firm. We measured knowledge of technology by asking respondents the following question: “Would you regard the company as hi-tech?”

Controls

We included several control variables that might have influenced the incidence of firm formation. The first control variables were for general procedural knowledge, which should not be a source of competitive advantage for a new firm founder. We included two controls for this type of knowledge (1) knowledge of previous work and (2) knowledge of previous start-ups.

We concur with Shane and Venkatraman (2000) and Shane (forthcoming) that most discoveries originate from an entrepreneur’s prior knowledge. Nevertheless, we view previous work experience and previous start-up experience as control variables because we do not expect them to relate specifically to the formation of a particular new firm. We measured knowledge of previous start-ups with the following question: “How many other businesses have your helped to start?” We posed this question to all new venture team members and summed their collective experience. We measured knowledge of previous work with the following question: “How many years of work experience have you had in this industry?—The one where your new business competes” Again, we posed this question to all new venture team members and summed their collective experience.

Industry concentration is a measure of the market share controlled by a small number of large firms. When industry concentration is high, firms with the largest market share can collude to drive out actual or potential competitors, which they do by erecting barriers to entry (Jacobson, 1992). High industry concentration can also reduce industry dynamism, which creates fewer opportunities for entrepreneurs to take advantage of competitive imbalances (Schumpeter, 1934). There are also fewer arbitrage opportunities when industry concentration is high. Thus, we suspect that opportunities for the formation of new firms will be inversely related to the degree of industry concentration. Of course, if these major firms did exercise their market power, it could be that these larger firms would leave many underserved market niches for start-ups. Because we do not expect these market leaders to neglect opportunities for profit, we predict that industry concentration will negatively affect firm formation (Bain, 1956; Spender, 1989). We measure industry concentration with a single question: “Do you expect the competition to be low, moderate or strong for this new business?” Responses included: zero for none, one for low, two for moderate and three for strong.

There is some evidence that entrepreneurs avoid competition by competing within protected niches (Bain, 1956; Rumelt, 1991). In such circumstances, industry concentration would not be an effective measure of the level of competitive rivalry. If competitive conditions were favorable within a protected market niche, it is possible that these conditions could overshadow the effects of knowledge-based competencies. We measured local market dynamics using the following question: “Would you describe your local economy as getting weaker, being stable or getting stronger?”

Firm founders recruit team members to marshal resources and to compensate for missing competencies (Aldrich & Zimmer, 1986; Hansen, 1995; Larsson & Starr, 1993). In fact, the existence of a start-up team is a reliable predictor of firm survival. Thus, we controlled for team size because we expect that the size of a venture team will be positively related to firm formation. To control for team size, we summed the number of legal owners and team members into a single item. We collected data for this summation by asking respondents to identify owners and team members.

Analysis

We performed an event history analysis using Cox regression (Allison, 1985; Yamaguchi, 1991). This approach controls for censoring problems by analyzing when a particular event commences and ends. To analyze the data in this way it was necessary to organize it into spells. A spell for a firm formation was a period from the start of reported business activities, which dated from 1985, to when a new venture signaled its formation by exchanging money. Data gathered back to 1985 was retrospective. We selected Cox regression because of its suitability for modeling time-to-event data in the presence of potentially censored cases. A Cox regression incorporates predictor variables as covariates and provides estimated coefficients for each of the covariates, which facilitates the assessment of the impact of multiple covariates in the same model. A Cox regression also permits the use of categorical and continuous variables in the same equation.

We expended considerable effort on telephone and mail follow-ups at each stage in the data collection to ensure high response rates because event history analysis cannot accommodate missing data. If a case contained missing data, we dropped it from this study. Fortunately, we found no significant differences between the demographic data for the whole sample (N = 623) and a control group (N = 608). In the end, we were able to infer whether a firm had been formed for each of the ventures in the initial sample. If a firm had been formed, we were able to date its formation on a monthly basis from 1985 to 1999.

Table 2 contains bivariate correlations and descriptive statistics for the covariates and control variables. Knowledge of previous work and knowledge of start-ups were both skewed. To normalize these distributions, we calculated their natural logarithms using a constant of one for zero values for each variable. These procedures produced minimal bivariate correlations—the highest being 0.318, which was between knowledge of previous work and knowledge of previous start-ups. Because all other correlations were lower, we consider this to be an independent set of covariates and control variables.

Table 3 tabulates the effects of the covariates and control variables on firm formation at different points in time. Model 1 shows that two controls for general knowledge—knowledge of previous work and knowledge of previous start-ups—are not significantly related to firm formation (p < .05). Nor did industry concentration or local market dynamics have any predictive value (p < .05). The only control variable that might have had a significant influence on firm formation was team size (Exp(B) = 1.106, p < 0.05). Moreover, team size was robust in all three models. The robustness of team size does not detract from any significant results for hypothesis testing, provided that the hypothesized results are statistically significant. The effect of a significant control variable is to partial away some variance between the predictor variables and firm formation that might otherwise flow in the hypothesized direction.

Model 2 in Table 3 incorporates the effects of information specificity, quantity of specific information and procedural competence on firm formation. The inclusion of these three covariates significantly improves model fit (D C² = 51.191, p<0.001), suggesting their inclusion improves our understanding of the effects of knowledge-based competencies on firm formation. As predicted, these results failed to disconfirm hypothesis 1 that information specificity is positively related to firm formation (Exp(B) = 0.994, p < 0.1). Likewise, Model 2 failed to disconfirm hypothesis 2 that the quantity of specific information was related to firm formation (Exp(B) = 0.767, p < 0.5). Likewise, Model 2 failed to disconfirm hypothesis 3 that procedural competence is related to firm formation (Exp(B) = 1.254, p < 0.1).

Model 3 failed to disconfirm the relationship between firm formation and declarative competence, as suggested in hypothesis 4 (Exp(B) = .893, p < .05). In addition, it failed to disconfirm hypothesis 5 (Exp(B) = 0.522, p < 0.05), which suggested that hi-tech competence was positively related to firm formation. Figure 1 graphically illustrates the differential influence of hi tech competence on the formation of hi-tech and low-tech firms. It is noteworthy that high competence does not increase the formation of low-tech ventures as much as it does high-tech ventures. We infer that differences in rates of formation relate to the specificity of the requisite competence required for each type of start-up.

Because we failed to disconfirm the influence of declarative competence on firm formation, we tested for the influence of different types of higher education on firm formation. It seemed appropriate to test these relationships because we assume that higher education is a primary source of declarative competence. Although declarative competence was related to firm formation, none of the individual types of higher education was related to firm formation. Thus, we can say that we could not disconfirm the relationship between higher education and firm formation, however the relationship between individual types of higher education was not statistically significant.

Finally, as we move from model 1 to model 3, the change in the chi-square fit statistics are statistically significant (p < .05), indicating the addition of the hypothesized variables improves the overall predictive validity of the models.

We can infer that knowledge-based competencies matter to the success of forming new firms. We have also provided evidence using logic from informational economics that the knowledge platform, on which individual competencies depends, must be specifically related the conditions on which a venture depends for its success. General competencies such as knowledge of previous work and previous start-ups may not be valuable as a source of information and knowledge for forming new firms. The relationship between the specificity of competencies and firm formation was statistically significant even after controlling for knowledge of previous work and start-ups, industry concentration, local market dynamics and team size.

One of the interesting results of this study is that after including specific types of higher education in Model 3, the quantity of specific information lost it statistically significant relationship with firm formation. Two explanations for this loss of statistical significance seem possible. First, it may be that individual types of higher education are correlated with the quantity of specific information. Individual types of higher education might be correlated with the quantity of specific information and firm formation if a firm to be formed were actually related in some specialized way to a particular type of higher education. An example, might be studying art history as preparation for starting a business to take tours of an art museum.

The second explanation for the lack of a statistically significant relationship between the quantity of specific information and firm formation is that entrepreneurs are not very good at solving the optimal stopping problem because they search too widely for information. The optimal stopping problem addresses the relationship between understanding that the acquisition of specific information may reduce the risk of start-up failure. However, its acquisition is not cost free and thus entrepreneurs must determine how to trade-off the expected benefit of acquiring specific information against its cost.

Fiet, Piscounov and Gustavasson (2000) address the optimal stopping problem in entrepreneurial search as it relates to the acquisition of specific information. They point out that it is possible to solve the stopping problem using an optimization procedure if entrepreneurs restrict their search to a known consideration set of information channels. Thus, their search would focus on the search of known channels, not on the search for unknown venture ideas that depend on entrepreneurs being lucky (Demsetz, 1983).

This research is important for entrepreneurs because it adds support to the notion that entrepreneurs face a choice. They can either immediately exploit whatever competence they possess in response to opportune venture ideas. If they act without matching their competence with the requirements of the venture idea, they may fail to form a new firm. Alternatively, they can decide in advance what type of competence would be needed to start a particular type of firm. Once they decide which type of competence would be most beneficial, they can suspend any efforts to form a firm in preference to acquiring the competence that they expect to need in the future. The lesson in this trade-off for entrepreneurs is that they should very carefully chose between the classes of start-up venture ideas that they wish to pursue. A consideration set of information channels can be a very useful tool to avoid firm failure if entrepreneurs will prepare themselves in advance. An additional lesson from this research for entrepreneurs is that the acquisition of general information will not set them apart from competitors in their race to acquire competence. Again, the haunting question remains, competence for what. This is not an easy question to answer, so entrepreneurs may wish to rely on scholars for guidance.

This research is important for scholars of entrepreneurial competence, exploitation and discovery because it highlights the usefulness of exploring the implications of informational economics and information theory. The application of information theory and informational economics to firm formation may be more helpful in developing theory than in describing actual entrepreneurial actions. Nevertheless, developing theory to improve entrepreneurial success should be a legitimate goal of scholarship in entrepreneurship.

CONTACT: James O. Fiet, Jonkoping International Business School, P.O. Box 1026  SE-551 11, Jonkoping, Sweden; (T) +46 36 15 62 58; (F) +46 36 16 10 69; james.fiet@jib.hj.se

Aldrich, H., & Zimmer, C. (1986) Entrepreneurship through social networks. In D. Sexton and R. Smilor (Eds.) The Art and Science of Entrepreneurship. Cambridge: Balinger.

Allison, P.D. (1985) Event History Analysis—Regression for Longitudinal Event Data. Sage University Paper series on Quantitative Applications in the Social Sciences, series no 46. Beverly Hills: Sage Publications.

Amit, R., Muller, E., & Cockburn, I. (1995) “Opportunity costs and entrepreneurial activity.” Journal of Business Venturing 10: 95–106.

Anderson, J.R. (1990) Cognitive Psychology and Its Implications. New York: W.H. Freeman and Company.

Bain, J.S. (1956) Barriers to New Competition. Cambridge, MA: Harvard University Press.

Barney, J.B. (1997) Gaining and Sustaining Competitive Advantage. Menlo Park: Addison Wesley

Demsetz, H. (1983) “The neglect of the entrepreneur.” In Joshua Ronen (Ed.), Entrepreneurship. Lexington, MA: Lexington Books.

Fiet, J.O. (1995) “Risk avoidance strategies in venture capital markets.” Journal of Management Studies, 32 (4): 551–573.

Fiet, J.O. (1996) “The informational basis of entrepreneurial discovery.” Small Business Economics 8: 419–430.

Fiet, J.O., Piskounov, A. & Gustavsson, V. (2000) “How to decide how to search for entrepreneurial discoveries.” Paper presented at the Kaufman Babson Research Conference. Babson College, Wellesley, MA

Gaglio, C.M. “Opportunity identification: Review, critique and suggested research directions.” In J Katz (Ed.) Advances in Entrepreneurship Research, Firm Emergence and Growth 3: 139–202.

Gilad, B., Kaish, S. & Ronen, J. (l988) “The entrepreneurial way with information.” In Sholomo Maital (Ed.) Applied Behavioural Economics, Vol. II. Brighton, England, pp. 480–503.

Hansen, E.L. (1995) Entrepreneurial network and new organization growth. Entrepreneurship Theory and Practice 19 (4): 7–17.

Hayek, F.A. (1945) “The use of knowledge in society.” American Economic Review 35: 519–530.

Jacobson, R. (1992) “The ‘Austrian’ school of strategy.” Academy of Management Review, 17 (4): 782–807.

Jensen, M.C., & Meckling, W.H. (1992) “Specific and general knowledge and organizational structure.” In Lars Werin and Hans Wijkander (Eds.) Contract Economics. Cambridge, MA: Blackwell Publishers.

Katz, J., & Gartner, W.B. (1988) “Properties of Emerging Organizations.” Academy of Management Review 13 (3): 429–441.

Kaish, S., & Gilad, B. (1991) “Characteristics of opportunities search of entrepreneurs versus executives: Sources, interests and general alertness. Journal of Business Venturing (6): 45–61.

Kirzner, I. (1997) “Entrepreneurial discovery and the competitive market process: An Austrian approach.” Journal of Economic Literature (35): 60–85.

Larsson, A., & Starr, J.A. (1993) A network model of organization formation. Entrepreneurship Theory and Practice (Winter): 5–15.

McGrath, R.G. (1999) “Falling forward: Real options reasoning and entrepreneurial failure.” Academy of Management Review 24 (1) 13–30.

Reynolds, P., (1997) National Panel Studies of Business Start-ups: Research Program Status Report and Policy Implications, Working Paper presented to the working party on small and medium enterprises ninth session, Seoul, Republic of Korea.

Reynolds, P., and Miller, B. (1992) “New firm gestation: conception, birth and implications for research,” Journal of Business Venturing,(7): 405–417.

Rumelt, R.P. (1988) “Theory strategy and Entrepreneurship.” In D.J. Teece (Ed.), The Competitive Challenge: Strategies for Industrial Innovation and Renewal. Cambridge, MA: Ballinger, 137–158.

Rumelt, R.P. (1991) How much does industry matter? Strategic Management Journal (12): 167–185.

Schumpeter, J. (1934) Theory of Economic Development. Cambridge, MA: Harvard University Press.

Shane, S. (Forthcoming) “Prior knowledge and the discovery of entrepreneurial opportunities.” Organization Science.

Shane, S. & Venkatraman, S. “The promise of entrepreneurship as a field of research.” Academy of Management Review 25: 217–226.

Spender, J.C. (1989) Industry Recipes: An Enquiry into the Nature and Sources of Managerial Judgment. New York: Blackwell.

Venkatraman, S. “The distinctive domain of entrepreneurship research.” In J. Katz (Ed.), Advances in Entrepreneurship, Firm Emergence and Growth, Vol. 3. New York: JAI Press, 119–138.

Von Hippel, E. (1986) “Lead users: A source of novel product concepts.” Management Science (32): 791–805.

Yamaguchi, K. (1991) Event History Analysis. Sage applied social research methods series no 28. Beverly Hills: Sage Publications.