Modelling entrepreneurial career progressions: concepts and considerations.
Studying entrepreneurs longitudinally is becoming more prevalent as a research desiderata (Bygrave, 1989) and as a practical possibility (Carroll & Mosakowski, 1987; Katz, 1992). But what remains largely unasked is what set of theories should be used to study entrepreneurs over time. Whether commonsense approaches, sometimes with sophisticated statistical underpinnings, are used to study firm and job creation (Birch, 1987; Reynolds, 1988), or economic modelling is used (Evans & Leighton, 1989), these are not approaches grounded in the actual study of the internal processes and decisions of the entrepreneurs themselves. What are needed to complement the growing capacity for primary and secondary analysis of entrepreneurs longitudinally are theories to complement the datasets and statistical methods.
There are several variants in the social sciences used to study people over time. The specific field concerned with the longitudinal study of people in work situations is called vocational or career theory (Ornstein & Isabella, 1993; Osipow, 1973; Schein, 1978). To date, however, little work has been reported in career theory related to entrepreneurs. Covering 1989-1992, Ornstein and Isabella (1993) found only two such papers, one of which was the precursor of this manuscript. What is needed is the adaptation of career theory to better consider the people and processes of entrepreneurship.
Fortunately one major career theory does provide an explicit consideration of types much like both the "entrepreneurs" and "small business owners" of longstanding debate (Carland, Hoy, Boulton, & Carland, 1984; Gartner, 1988)--the career anchor theory of Edgar Schein (1978). Schein's model provides both a means to conceptualize different types of self-employeds in a manner consistent with existing entrepreneurship research, and a widely recognized and accepted set of variables for describing the nature of an individual's career and the individual's movement through it, called the career cone.
Using Schein's model as a base, this paper will pursue two goals. One goal is to adapt Schein's vocational model to analyze in finer detail a greater range of entrepreneurial career progressions and subsidiary analytic issues. To do this definitions of the three major career cone variables of hierarchy, function, and centrality will be expanded to fit with career-relevant findings from entrepreneurship research, and three additional variables--employment duration, job multiplicity, and self-employment emergence--will be introduced to better account for the nature of entrepreneurship in the workforce. The other goal will be to demonstrate how these variables can be operationalized to facilitate theory-based secondary analyses of entrepreneurial careers using the growing collection of longitudinal datasets of the workforce. The effort starts with introducing and adapting Schein's theory.
ADAPTING SCHEIN'S CAREER THEORY
The Schein model is uniquely suited for research in entrepreneurial careers because it is one of the few models that explicitly considers self-employed individuals. In fact, Schein's two types of career anchors for the self-employed, entrepreneurial endeavors and autonomy, closely follow the conventional entrepreneur/small business owner dichotomy common in entrepreneurship research.
Schein's general model for career cycles forms the basis for developing a theory-driven model for analyzing longitudinal datasets. Schein (1978) characterizes the individual's career progression using what is called a Career Cycle Cone (Figure 1) reflecting movement along one or more of three dimensions: Hierarchy, which reflects the total control or power available to the person in the job; Functional/Technical Area, which reflects the type of activity performed in the job; and Centrality/Inclusion, which reflects movement toward or away from the "core" of the organization or profession.
[FIGURE 1 OMITTED]
The first two variables have a long-established history in vocational and economic theory. For example, in one of the most detailed recent economic models of career mobility using the Panel Study of Income Dynamics (Sicherman & Galor, 1990), change was measured using differences across time in occupational code (a technical measure (1)) and a derived measure of "level" of the occupation based on the education necessary for employment in that occupation.
In addition to the three variables making up the Career Cone, Schein found that all the self-employed fell into one of two Career Anchors. A career anchor is "the pattern of self-perceived talents, motives, and values [which] serves to guide, constrain, stabilize and integrate the person's career" (Schein, 1978, p. 127). Uniquely among personality-based characteristics, career anchors cannot be successfully identified in students, since stable anchors emerge only with work experience. The two anchors favored by the self-employed are: Autonomy/Independence, which is a desire for freedom from rules and the control of others; and Entrepreneurship (originally called Creativity), which focuses on the creation of "something new, involving the motivation to overcome obstacles, the willingness to run risks, and the desire for personal prominence in whatever is accomplished" (Schein, 1985a, p. 30).
These two anchors closely follow respectively the "Craftsman" and "Opportunistic" entrepreneurial orientations originally developed by Smith (1967) and refined in the works of Smith and Miner (1983). The distinction also closely follows the small business owner vs. entrepreneur distinction articulated by Hoy and others (Carland, Hoy, Boulton, & Carland, 1984).
This section considers the dimensions of hierarchy, function, and centrality and the career anchors from the perspective of fine tuning their adaptation to the self-employed, and operationalizes the measures to make possible analysis using established longitudinal surveys of individuals.
Self-employment by definition places the individual at the top of the firm's hierarchy. In one-person firms (Katz, 1984; Star, 1979), which account for over half of all business firms, growth occurs by adding levels below the top of the hierarchy.
This situation rooted in the extreme case offers an approach for adapting hierarchy to the study of the self-employed. In studying the self-employed, hierarchy is best measured by growth in organizational size. This approach measures the number of people over whom the self-employed exercises leadership or power. The owner of a one-person firm would be seen as having power over no one. Hence the owner of a one-person firm has less hierarchy than the owner of a 100-employee firm.
The advantage of accepting such a size measure means that differences between large- and small-business CEOs become matters of degree along a clearly operationalizable dimension, instead of requiring different sets of theory for the two classes of CEOs. Additionally, such an approach fits well with wealth-centered (Stevenson, Roberts, & Grousbeck, 1989; Timmons, Muzyka, Stevenson, & Bygrave, 1987) and political economy (Birch, 1987) definitions of entrepreneurship. The measure is also included in most surveys of the self-employed and is relatively unambiguous in its meaning, measurement, and interpretation.
There are, however, three problems related to the use of such a size measure. First, organizations, especially small ones, can grow substantially in ways that do not add employees. For example, small businesses often grow through the use of independent contractors or through the use of automation such as answering machines or computers. Second, particularly in new firms, number of employees may be an inferior size measure compared to sales. As new firms establish themselves in the market, a fixed number of employees often produce substantially greater sales with additional experience and customer awareness. Finding such firms in samples would require having information on the age of the firm in addition to its size. Third, in most organizational theory research, hierarchy is measured directly or inferred using some rule of thumb about the span of control. This approach is likely to be impractical for application in smaller enterprises for three reasons. First, the self-employed remains at the first level of hierarchy regardless of the number of levels below. Second, few of the longitudinal surveys of individuals contain information about number of hierarchical levels in firms, although most of these surveys do ask the self-employed for the number of employees in their firm. Third, while some established norms exist for converting number of employees into number of levels, those norms are based on large, often publicly held and professionally managed firms. When this approach to organizational design is contrasted to the models common in small business, it is commonly found that smaller, owner-managed firms tend to have exceptionally flat (Mintzberg, 1979) or simple (Jaques, 1976) structures, with few levels of hierarchy.
Therefore, unless new, separate norms can be established for smaller, owner-managed firms, the least ambiguous way to measure hierarchy of the self-employed is to use the number of employees.
Function is perhaps the weakest of the three dimensions because the rules of thumb underlying research differ by country. For example, in some countries occupational classification schemes differentiate between the activities of salaried and self-employed managers in the same type of vocation (e.g., Carroll & Mosakowski, 1987), while the American system, based on census classifications, does not. A self-employed plumber in America becomes a "proprietor," even though the vast bulk of the individual's work is no different from that of a wage-or-salaried plumber without managerial duties. Does employment status make a difference in the functions performed by the individual?
Looking at American vocational theory, one might guess the answer to be "no." Little conceptual work has been done to document or analyze differences between wage-or-salaried and self-employed individuals in what are ostensibly the same occupation. And in practice, the continuing reliance on 2-digit Census Occupational Codes in most American surveys means that there is little opportunity to compare functional differences between wage-or-salaried and self-employed individuals. The major known problem stemming from the use of the 2-digit classification relates to those individuals who are classified as proprietors rather than as self-employeds practicing a particular occupation, as in the case of the plumbers mentioned above. Worse yet, while coding of this variable appears consistent within organizations (e.g., ISR, NORC), it may differ between organizations.
Despite the oversight in theory and census coding, the little research to date suggests that employment status does alter function. Eden (1975), using data from the Quality of Employment Surveys, has shown that salaried and self-employed work do have some systematic differences. In Eden's parlance, wage-or-salaried work can be referred to as "maintenance-free employment." This means that while a self-employed individual must worry about maintaining the firm, meeting payrolls, paying licenses and firm fees, the employee does not. The burden of maintaining the workplace as an ongoing institution falls on the self-employed, not the wage-or-salaried.
Centrality, like hierarchy, loses some meaning when the owner is the central figure in the firm. From an inter-firm or longitudinal perspective, however, the solution follows the logic applied to hierarchy. Over time the defining characteristic of centrality of the self-employed is their relation to their expertise. If centrality is proximity to the critical expertise of the firm or the career, the adaptation becomes clearer. Entrepreneurs' careers can be measured in terms of deviations from their central expertise.
The adaptation is clearest when studying entrepreneurs in the nonbusiness professions. A dentist who opens a dental practice has directly applied expertise. In taking on and supervising subordinate dentists, the entrepreneur-dentist is doing some dentistry, but is also doing supervision, requiring some generalization of the lessons learned in dental school. A dentist who creates a multimember dental plan (the dental equivalent of an HMO) has generalized on the core expertise even more, to the extent that some may see the dentist now more as a manager than as a practicing dentist.
Dentists may also branch out from their profession. For example, for almost two decades Holiday Inns' largest franchisee was a Memphis dentist. Another Memphis dentist, Winfield Dunn, served as governor of the state. In both cases the business owner and the politician reported that their analytic abilities and people-oriented skills were directly based on what they learned as dentists.
Centrality information is variably available in longitudinal surveys. For example, in the work detail supplements of the Panel Study of Income Dynamics (PSID) and the National Longitudinal Studies (NLS), there are detailed entries about the type of work done, permitting the possibility of evaluating centrality. In life history datasets such as the German one described by Carroll and Mosakowski (1987), such details are available for coding. However, most longitudinal studies, such as those listed by Katz (1992) as particularly relevant for entrepreneurship researchers, would not permit centrality analyses. Overall, the effectiveness of this measure will depend on the biographical and qualitative nature of the data being used.
Career anchors have been assessed through two methods. Originally, career anchors were assessed only through the use of an extensive interview or self-study protocol (Schein, 1978). Later, a somewhat simplified dyadic interview protocol was developed (Schein, 1985a) and included a 41-item questionnaire, the Career Orientations Inventory (COI). The COI, however, is a poor method for assessing career anchors when using secondary data analyses; psychometric information on the COI is scarce; its relation to other personality or vocational instruments is not known; and it has not been used in major national surveys. For these reasons, the interview protocol and its scoring system will be discussed here.
Analysis of the interview protocols has always proceeded in a holistic, qualitative fashion. The goal has been to identify major themes in the history, and to seek to establish whether these themes form any consistent patterns across the person's working life. These patterns, where they exist, are weighed, and the dominant one is labelled the individual's career anchor.
This approach has the potential for adaptation to existing datasets through pattern matching. In pattern matching, profiles of typical behaviors for Autonomy- and Entrepreneurship-oriented individuals are developed, and individuals' career histories are compared to the two profiles. With a few exceptions, the major longitudinal studies do not code answers verbatim or ask the types of questions that would permit a direct application of career anchor thematic analysis approaches. However, given sufficient points of comparison, classification into one of two categories should be possible. Such taxonomic methods have been applied in organizational studies by McKelvey (1982). An example of such patterns based on Schein's work (1978, 1985a, 1985b) is given below in Table 1.
Table 1 Prototype for Career Anchor Pattern Asssessment Autonomy Entrepreneurship Status Changes in Little Changes to Higher Wage or-salaried Status Frequent Employment Moves Firm Sales Show Typically No Typically Yes Substantial Growth (esp. after start-up) Income Low-to-Moderate High (for industry) (for industry) Firm Sole Typically Yes Typically No Proprietorship Number of Employees Low-to-Moderate High (for industry) (for industry) Habitual Entrepreneur Owns Firms Owns Multiple Firms Pattern Serially Simultaneouly Multiple Business Typically No Typically Yes Sites
To explain the reasoning behind the entries, the prototypical autonomy-oriented individual would tend to emphasize freedom rather than financial or status achievement. Therefore, such a person, when employed by others, is less likely to change jobs to increase his or her level of autonomy. As Smith (1967) pointed out, to control a business, its size must be kept manageable. This means one site, where the owner is always present, only enough employees to be directed personally, and growth limited to what is needed to provide a consistent living, and this is likely to "top out" at moderate levels. If the individual has multiple episodes of self-employment, they are handled one-at-a-time, in order to maximize control over the firm and the demands placed on the owner by work. Rather than share control with partners or directors, the autonomy-oriented person seeks control through a sole proprietorship.
Those with an entrepreneurship anchor are likely to change salaried work for greater pay or status, and are more likely to pursue opportunity. Their firms are more often designed for growth, for example being incorporated to facilitate raising capital. Their firms are likely to show growth in sales, number of sites and employees, and where successful are likely to show high levels of each. When entrepreneurs have multiple episodes of self-employment, they are more likely than others to have more than one firm going at the same time.
The problem in this approach is determining the degree of correspondence necessary to classify the career anchor as autonomy or entrepreneurship, and not one of the other six anchors, or even the unclassifiable 4.6% who appear to have no clear anchor (Schein, 1985b, p. 7). One demanding but workable approach is to develop classification schemes for all anchors to assure the discriminant quality of the two anchors of interest.
A final concern around Schein's anchors and mainstream entrepreneurship research centers on the incidence of autonomy and entrepreneurship anchors. Based on 14 studies with a total of 348 individuals, Schein (1985b) reported an incidence of entrepreneurship anchors of 7.7% across the groups, while autonomy anchors appeared in 11.1% of participants. The incidence of entrepreneurship in this sample is roughly twice the incidence rate projected for the population in large-scale studies such as Reynolds (1988) or Cooper and Dunkelberg (1981).
Sample differences most probably account for much of this difference. Schein's samples are generally MIT graduates, managers, or professionals. Virtually all have college degrees, and many have even more education. These individuals are drawn from the wage-and-salaried as well as the self-employed. Cooper and Dunkelberg (1981) surveyed only self-employeds, including many self-employeds with minimal education, who were further preselected as NFIB members, while Reynolds surveyed owners or managers of small businesses identified through Dun and Bradstreet reports. Among these studies, there emerges a demonstrated need to compare methods for determining the incidence of entrepreneurship from different approaches to improve the accuracy of self-employment projections. The differences identified between the career anchor and other approaches suggest the form such a comparison might take.
ADDITIONS TO THE SCHEIN MODEL
While the three cone variables and career anchors explain much about career dynamics, there remain untouched but important elements of careers, particularly those of the self-employed. Three key variables of entrepreneurial careers, derived from the literature below, serve as a starting point. They are duration, multiplicity, and emergence.
The literature underlying these three additional variables is eclectic, drawing from virtually all areas of the social sciences, as well as from a number of research efforts in entrepreneurship itself. The data sources used to identify the additional variables needed to analyze the nature of organizational careers for the self-employed fall into two main categories: research works and anecdotal works. Research works included empirical mainstream entrepreneurship models (Dyer, 1992; Katz, 1992; Katz & Gartner, 1988; Roberts, 1991; Ronstadt, 1984b, 1985, 1988; Savage, 1979; Schein, 1978; Scherer, Bordzinski, & Weibe, 1991; Smith, 1967); models from labor economics (Evans & Leighton, 1989); and models from organizational theory on organizational formation and self-employment (Carroll & Mosakowski, 1987; Perkins, Nieva, & Lawler, 1978; Sarason, 1972; Van deVen, Hudson, & Schroeder, 1984). Anecdotal works included biographic compilations (Fucini & Fucini, 1985; Gervitz, 1984; Silver, 1985); mainstream historical biographies (Livesay, 1979); business cases (Ronstadt, 1984a; Stevenson, Roberts, & Grousbeck, 1989); and journalistic essays from Inc. and Entrepreneur magazines.
Research works tend to be grounded in theoretical or empirical efforts by academics, while anecdotal works can be developed by academics, as cases usually are, or by journalists or even the entrepreneurs themselves.
Duration is the time that the person is self-employed. For example, Ross Perot or An Wang can be seen as long-duration self-employeds, while Donald Burr, who started People Express, was self-employed for a much shorter period. Since short-lived efforts may well represent failure, or lessons that could be learned only by direct experience, these short-duration episodes are likely to have a substantial impact on career decision making, and need to be covered wherever possible.
However, for secondary analysts duration poses a special problem related to the between-interview interval. An enduring issue in the analysis of firm formation or organizational emergence is the ability to monitor all efforts to create firms, to accurately gauge which organizational forms or mutations have been tried, and which succeed and which fail (Katz & Gartner, 1988). In developing entrepreneurial sampling frames, a key problem is gathering information on firms that were attempted, or were created but died, within a between-interview interval. For example, Aldrich (1990) shows that telephone books and D&B lists may miss firms that are very short-lived.
One of the few empirically based efforts to identify the rate of short-lived business is seen in a footnote to Evans and Leighton (1989, p. 525). Using the Current Population Survey data on the unincorporated self-employed white males, they estimated a 41.4% exit rate within 12 to 21 months. Conversely, Carroll and Mosakowski (1987) reported that the self-employed were generally among the most stable workers in the sample, staying in their firm longer than typical wage-or-salaried individuals stay in their jobs. The difference between these types of 'findings most probably represents the "liability of newness" argument of Singh, Tucker, and House (1986).
Another duration-related area which has received little focus but where theory may be usefully applied is in the seasonality of self-employment. For example, many short-duration yard-maintenance people work primarily in spring and summer. In contrast, tax preparers may work primarily in winter and early spring. What distinguishes these types of work is that they are of limited duration and the likely durations can be specified in advance. To date, little effort has been made to develop seasonal adjustments for self-employed work, although such adjustments for wage-or-salaried estimates are common.
Multiplicity is the number of jobs concurrently held by the self-employed individual. About one in five self-employeds also works at a wage-or-salaried job. A variant of this pattern is called by Jennifer Starr (1992) the "habitual entrepreneur," the person who is self-employed in more than one business at a time.
Multiplicity can occur with a mixture of wage-or-salaried and self-employment, as well as with a mix of simultaneously held self-employed positions. Autonomy- and entrepreneurship-anchored individuals both appear likely to "try out" self-employment while at work in a wage-or-salaried job. Likewise, there does not appear to be any theory-drive argument suggesting differences in the incidence of non-moonlighting self-employment.
Where the greatest likelihood for a difference to exist between entrepreneurship and autonomy-anchored individuals is in the pattern for owning a number of firms through a career span. Studies such as Katz (1990) suggest that most self-employeds operate only one business at a time, although there is a contingent of people who return to self-employment repeatedly. However, Starr's habitual entrepreneurs, who arguably represent the extreme case of entrepreneurship-anchored individuals, often have more than one personally owned business operating at a time.
From the career anchor perspective, it is unlikely that an autonomy-oriented individual would be interested in owning a number of firms simultaneously. Control of diverse firms, particularly if they are in different locations, would be difficult. Additionally, with more firms to operate, each probably requiring some minimum additional increment of the owners' time, autonomy-oriented individuals would find less of their time under their control. This type of person might own several firms across a lifetime, but they would be owned one-at-a-time, to maximize control over the firm and one's own life. When the autonomy-anchored person is involved in two organizations simultaneously, it commonly represents the person using self-employment as an income supplement, through "moonlighting," or the individual's testing out self-employment by dabbling in it part-time.
For the entrepreneurship-oriented person, multiplicity may simply reflect the strong drive to create firms. Since here the creation is important, the time costs of multiplicity would not be a consideration, although concerns about the survival of the different firms would remain an issue. For these people, the potential for simultaneous multiple firm ownership is viable. This analysis is supplemented by available information that virtually all of Starr's habitual entrepreneurs demonstrate strong identification with the entrepreneurship anchor (Starr, 1992).
Emergence in careers means actual entry into self-employment. It represents a more general case of the concept of organizational emergence offered by Katz and Gartner (1988), and like that earlier version is characterized by the presence of four characteristics: intention to enter self-employment, boundary between the role of self-employed person and other life roles, resource dedication to the self-employment activity, and goods or service or financial exchange across the self-employment boundary. The organizing process (Weick, 1979) for a new firm requires the four characteristics to come together. With one or more element missing, the individual is in the process of organizing a firm.
The emergence variable is important because its four characteristics help conceptually differentiate incomplete and failed efforts at achieving self-employment from efforts that go through to completion. While several studies have noted the "gestational" period prior to self-employment entry (Perkins, Nieva, & Lawler, 1978; Van deVen, Hudson, & Schroeder, 1984), little work has been done looking at the pre-entry period. Two exceptions, which look at the start-up process in very limited fashion, are studies of venture-capital-backed efforts, and business planning.
Katz and Gartner (1988) have suggested that a useful means of analyzing emergence comes from the sequence in which the four properties come into the process. Using the autonomy and entrepreneurship career anchors, an autonomy-oriented individual is more likely to be driven by the desire to have freedom from control by others. This type of intention often is at the forefront in histories of small business owners. A variant of this is the assertion of or desire for territoriality, "a place of my own," which reflects a boundary consideration. For the autonomy-oriented, issues of market opportunity (exchange) or asset accumulation (resources) are less likely to be their first considerations in starting a firm.
For those with an entrepreneurship anchor, the opportunity-recognition process (Gilad & Levine, 1986) or wealth-creation process (Stevenson, Roberts, & Grousbeck, 1989) is foremost in their consideration. Often these people report themselves "pulled" into ownership by the external pressures of market (exchange) or wealth (resources). Even those entrepreneurs who become self-employed as a reaction to rejection of the firm's service or product by the individual's former employer, such as Apple founders Jobs and Wozniak, often describe the need to see their idea go forth into the market as their driving force.
Taken together, duration, multiplicity, and emergence offer substantial additional information to the career cone variables of hierarchy, function, and centrality, and the career anchors of autonomy and entrepreneurship. Table 2 summarizes the comparison of autonomy-anchored to entrepreneurship-anchored individuals using the career cone variables and the additional variables introduced in this analysis.
Table 2 Summary of Career Anchor Differences on Key Variables Autonomy Entrepreneurship "Career Cone" Variables Hierarchy At Top of Smaller At Top of Larger Firm, Firm, Simple Simple Stracture Structure Likely Likely More Complex Structure Possible Function Managerial/Proprietor Managerial/Proprietor Less Likely More Likely Centrality Less Variation from Greater Variation from Core Occupation Core Occupation Additional Variables Multiplicity Less Likely. Person More Likely. Person Usually Owns Firms Usually Owns Multiple Serially Firms Simultaneously Duration Longer Duration Shorter Duration in Within Self-Owned Self-Owned Firm (or Firm until multiplicity occurs) Emergence Intention and Resources or Exchange Boundary Likely Likely Before Before Resources or Boundary or Intention Exchange
Collectively, these variables make possible the analysis of career dynamics of the self-employed at a more sophisticated level. Fundamental analyses of self-employment entry have been reported by researchers interested in occupational mobility (Markey & Parks, 1989; Sicherman & Galor, 1990), and in more focused analyses by Carroll and Mosakowski (1987) and Katz (1990). However, occupational mobility analysts tend to consider only emergence and functional changes, while self-employment entry analyses have looked only at duration of episodes, and in the case of the Katz analysis considered partially emergent cases.
One useful integration of these variables is in the extension of Schein's career cycle concept into the specific processes of career trajectory (DiPrete, 1987; Juhasz, 1989). Among the wage-or-salaried, the concept of career progressions is well understood, and relatively easy to describe. In most of these models, the progression is made easy because of a succession of positions of greater hierarchy and/or centrality. For the self-employed, the reliance on centrality and hierarchy minimizes the interpretability of the progression or trajectory, since self-employeds are by definition at the top and most central position in the firm from its inception. Because of the reduction in variance, the inclusion of other variables, such as function, duration, multiplicity, and emergence, is needed to produce an analysis of the self-employeds' trajectories that matches the detail of conventional analyses of the wage-or-salaried.
In addition, the lesson of the two career anchors predominating among the self-employed suggests substantially different types of career trajectories. As is best seen in Smith's (1967) work, autonomy-oriented individuals are inclined to grow a business to a psychologically comfortable level (best defined as "just as much as I can personally control, and not one bit more") and then stop growth. Entrepreneurship-oriented individuals, or opportunistic entrepreneurs in Smith's sample, will continue to grow the firm, using delegation, advertising, and other techniques. In Smith's analysis of manufacturers, the average sales for the two types differed by a factor of ten, with opportunistic types, Schein's entrepreneur-anchored people, reporting the larger sales.
From the Starr (1992) work on habitual entrepreneurs, it is evident that those entrepreneurs sampled were inclined toward an entrepreneurship anchor. Their use of multiple, simultaneous firm ownership reflects one special type of trajectory. To differentiate this approach from the autonomy-anchored individual who "wears many hats" such as mechanic, real estate agent, and farmer, the size of the firm (a hierarchical measure) would be needed in analysis.
There are other likely trajectories.
Intermittent. Individuals alternate between wage-or-salaried and self-employed work, because (1) they may be marginally capable of work as often described in sociological and psychological literature (Shapero & Sokol, 1982) or (2) their work is strongly seasonal (e.g., tax preparers, lawn services, fruit stands) and they intersperse the two forms of employment.
Helical. Work-related changes are affected by family or personal factors, altering their timing or size and making the individual's career trajectory appear erratic (Juhasz, 1989; Schein, 1978, p. 24).
Downward. The business fails or is downscaled (Ronstadt, 1985).
Ambiguous. The individual is unsure of his or her self-employment.
Obscured. The individual is either unsure of self-employment or is pursuing it as part of the "underground economy" and is resistant to public admission of self-employment.
Jump-started. The individual enters self-employment in an existing and substantial business through purchase or inheritance.
These approaches contrast with the traditional models of entrepreneurial entry (Cooper & Dunkelberg, 1981) by focusing on the total career process. The analysis of trajectory and the authentication of these potential trajectories defined by cross-sectional or small sample models depend on the ability to usefully create the Schein variables and their extensions described above. The potential for doing that is explored in the next section.
The goal in the above efforts was to develop a theoretically grounded model for analyzing the careers of self-employeds which has potential for analysis using existing longitudinal datasets. The theory chosen is that of Edgar Schein, specifically his model for career mobility using the variables of hierarchy, function, and centrality, and his concept of career anchors. To this were added three variables found useful in dynamic analyses of entrepreneurship: duration, multiplicity, and emergence. In developing these variables for analysis of the self-employed, several distinctions were made between autonomy-anchored and entrepreneurship-anchored individuals, and these serve as hypotheses for test in the future. Similarly, several cautions regarding the measurement or conceptualization of the variables were identified, and where possible, suggestions for how problems can be evaluated or resolved were given.
There remains a practical question--can the variables be operationalized successfully using existing longitudinal datasets? In this section, the variables introduced above will be operationalized using variables from the Panel Study of Income Dynamics (PSID).
The PSID (Morgan, 1972) is a longitudinal survey conducted yearly by the Survey Research Center of the University of Michigan since 1968. Its subjects are a panel of 3426 households that were representative of the American population in 1968 (with low-income households oversampled). Spinoffs of the original households are kept in the panel, and the data can be weighted to provide population estimates. (2) The surveys, which are conducted by phone, have a core set of questions that remains consistent year-to-year, and another set of questions that varies each year. Fortunately, the information needed to operationalize the majority of the variables comes from the core question set of the study. In addition to the standard questions, a special PSID work history supplement was gathered in 1987, and serves as a separate dataset.
Building on the adapted and new variables introduced in this paper and summarized in Table 2, Table 3 demonstrates both the potential and the shortcomings of secondary analysis done from an a priori theoretical framework. The best measured variables in Table 3 are intention and duration, for which direct measures are available, and are consistent with the theory or established methods of prior research. However, key variables such as hierarchy cannot be gleaned from the information available in the PSID. Other variables, such as resources or exchange, can only be studied in the most superficial way, or in the case of boundary, only for established businesses.
Table 3 Summary of Variable Operationalizations for Key Variables Description of Operationalization in the PSID dataset "Career Cone" Variables Hierarchy None Function Occupation (Census 2-digit Code), Industry (Census 3-digit code) (1) Centrality Prior occupations can be compared to present occupation Additional Variables Multiplicity Each income source is separately listed Duration Interview date--When current job started (2) = Time in current job Current job start data--prior end date = Time between current job and last job Prior job start date--Prior job end date = Time in prior job Emergence Intention Job plans for the next year, Part-time self-employment Boundary Self-employment (full or part-time), Whether business is incorporated Exchange Income from self-employment, Non-cash transfers (in and out of business) Resources General assets available, Income (1) The PSID also asks respondents "What are your most important activities or duties?" but does not code the responses. (2) This measure is of course subject to right censoring for survival type analyses.
Variables such as function or centrality represent approximations of the desired variables. In the case of function, occupation is a longstanding, surrogate measure. Centrality requires developing a novel measure. Since the codes are based on the 1970 Census of Occupation, the codes reflect two characteristics simultaneously. Most importantly, the numerical classification generally reflects a Census and Department of Labor classification of job families, where similar jobs are grouped together. (3) A second classification is a job status hierarchy, with the highest status workers (professionals and technical workers) at the top and the lowest status workers (private household workers) at the bottom. Either way, centrality can be measured as differences in the ordinal scale values for current and prior jobs, in terms of job family difference scores, or status differences. The smaller the difference, the more central the current job.
Although not covered in Table 3, the PSID offers considerable opportunity for indirectly classifying the career anchors of autonomy and entrepreneurship from attitudinal information. The study includes raw item results for items included in an Atkinson model achievement motivation questionnaire, as well as an additional set of attitude questions about economic decision making.
Other operational measures introduced in Table 3 that can be directly measured from existing PSID variables are status changes in employment moves, income, and habitual entrepreneurship. Where an individual's income growth can be accepted as a surrogate for a firm level measure, firm sales growth can be estimated, and incorporated firms can be distinguished from sole-proprietorship and partnerships. Given national statistics indicating the small percentage of partnerships in the population of firms, this measure can be adapted to reflect an estimated number of sole proprietorships. Of the potential variables listed in Table 3, only number of employees and multiple business sites could not be measured or derived from the PSID.
These problems of lack of variable coverage are the common problems of secondary analysis in general. What makes such compromise palatable is the faster access to longitudinal data. Depending on the type of information sought, having information of a few key variables for a 20-plus-year period may provide useful information.
In this specific instance, is the fit of Schein's career dynamics model to secondary analysis likely to be a useful one? The value of theory in general, and Schein's theory of career dynamics in particular, is that it provides coherence to the structuring of research. The potential threat to the secondary analyst is the slide into "dustbowl empiricism," the search for statistical significance rather than conceptual meaningfulness. To date, most career analyses of the self-employed have relied solely on studies of entry. Once in self-employment, few of the existing theories have explanatory power. Ronstadt's (1988) observation of the corridor principle (4) is an exception to this, however it describes only one component of the career development process.
As was shown above with PSID data, some meaningful analyses of models inspired by Schein's work can be generated with existing data, and without using elaborate proxies based on chains of suppositions. Other datasets, such as those described by Katz (1992), offer opportunities, especially in cross-sectional analyses, for more detailed testing of the Schein model.
The extension of Schein's three career cone variables to include multiplicity, duration, emergence, and the integrated concept of trajectory is grounded in the idea of operationalizing the actions assumed in the Schein model, for example entry in the case of emergence, and duration in the positions held. The concept of trajectory within a career, and the impact of life and family elements on career issues, are already considered by Schein, although the specific implications for the self-employed are not elaborated in his existing work.
It is in this elaboration process that findings from mainstream entrepreneurship research, and the related works of labor economists, sociologists, and organizational ecologists become the basis for specific content descriptions of the potential forms the career processes might take. In the example of potential trajectory forms, virtually all the career cone variables, the extended variables, the career anchors, and the findings of other researchers come into play to identify empirically and theoretically derived trajectories. In this case, the Schein model serves as the unifying element. With the extensions elaborated here, it possesses the requisite variables to accommodate the piecemeal contributions of the different disciplines and still provide an overarching, integrated, and consistent model that permits the creation of new knowledge from the works of others. From the standpoint of a secondary analysis approach, that is all that one can ask.
Thanks are extended to Gibb Dyer, Jennifer Starr, Ed Schein, Norris Krueger, Bill Olbrecht, and Cheryl Nietfeldt for their help in the development of the ideas used in this paper, although the resulting ideas are solely the responsibility of the author. A version of this work was presented as part of the Symposium: New Directions in Research on Entrepreneurial Careers, at the 1992 Annual Meeting of the Academy of Management, Las Vegas, August 11, 1992. Support for this work came from the Jefferson Smurfit Center for Entrepreneurial Studies, the Department of Management and Decision Sciences, the Beaumont Faculty Development Fund, and the Summer Research Grant Fund of the School of Business and Administration, Saint Louis University.
(1.) One problem of such approaches is that they categorize all self-employeds in one group, with equal status. In many cases this underestimates the average status and the variance in status of the self-employed because it fails to consider the higher status of self-employed professionals such as doctors or lawyers in private practice. In some occupations, such as dentistry, the self-employed dominate, while in most professions in the U.S. the self-employed occur near the national average of 12%.
(2.) Among the datasets mentioned in Katz (1992) as applicable to the study of the self-employed, the PSID is one of the two most useful. The other survey, the National Longitudinal Survey (NLS) from Ohio State University (Parnes, 1976) generally has many of the same core variables as the PSID, but has substantially different non-core information. Additionally, the NLS is organized by cohorts of older men, mature women, and young men and women. In general, the findings for variables based on the PSID approximate what could be done using the NLS.
(3.) The result remains an ordinal scale with all its attendant problems. For example, differences between job families are more substantial than differences within a job family. Thus a 10-point difference within a job family would have less import than a 2-point difference that puts the person into another job family. However, the ordinal scale can be weighted to produce an approximation of an interval scale, for example, by separating job families by arbitrary differences of 100 points, while leaving point differences within families unchanged.
(4.) Ronstadt's discovery of the corridor principle demonstrates the typical gulf between entrepreneurship researchers studying entrepreneurial careers and those who make career research their focus. In one of the definitive review texts on theories of career development, Osipow (1973) detailed the opening of opportunity contingent on prior decisions, what Ronstadt calls the corridor principle, as a structural derivative of his own systems approach to career development.
Aldrich, H. E. (1990). Using an ecological perspective to study organizational founding rates. Entrepreneurship Theory and Practice, 14(3), 7-24.
Birch, D. L. (1987). Job creation in America: How our smallest companies put the most people to work. New York: Free Press.
Bygrave, W. D. (1989). The entrepreneurship paradigm (II): Chaos and catastrophes among quantum jumps? Entrepreneurship Theory and Practice, 14(2), 7-30.
Carland, J. W., Hoy, F., Boulton, W. R., & Carland, J. C. (1984). Differentiating entrepreneurs from small business owners: A conceptualization. Academy of Management Review, 9(2), 354-359.
Carroll, G. R., & Mosakowski, E. (1987). The career dynamics of self-employment. Administrative Science Quarterly, 32, 570-589.
Cooper, A., & Dunkelberg, W. C. (1981). A new look at business entry. In K. H. Vesper (Ed.), Frontiers of entrepreneurship research. Wellesley, MA: Babson College.
DiPrete, T. A. (1987). Horizontal and vertical mobility in organizations. Administrative Science Quarterly, 32, 422-444.
Dyer, W. G. (1992). The entrepreneurial experience. San Francisco: Jossey-Bass.
Eden, D. (1975). Organizational membership vs. self-employment: Another blow to the American dream. Organizational Behavior and Human Performance, 13, 79-94.
Evans, D. S., & Leighton, L. S. (1989). Some empirical aspects of entrepreneurship. American Economic Review, 79(3), 519-535.
Fucini, J. J., & Fucini, S. (1985). Entrepreneurs: The men and women behind famous brand names and how they made it. Boston: G. K. Hall.
Gartner, W. B. (1988). "Who is an entrepreneur?" is the wrong question. American Journal of Small Business, 12(1), 11-32.
Gervitz, D. (1984). The new entrepreneurs: Innovation in American business. New York: Penguin.
Gilad, B., & Levine, P. A. (1986). Behavioral model of entrepreneurial supply. Journal of Small Business Management, 24(4), 45-53.
Jaques, E. (1976). A general theory of bureaucracy. New York: Halsted.
Juhasz, A. M. (1989). A role-based approach to adult development: The triple helix model. International Journal of Aging and Human Development, 29(4), 301-315.
Katz, J. A. (1984). One person organizations: A resource for researchers and practitioners. American Journal of Small Business, 18(3), 24-30.
Katz, J. A. (1990). Longitudinal analysis of self-employment follow-through. Entrepreneurship arm Regional Development, 2(1), 15-25.
Katz, J. A. (1992). Secondary analysis in entrepreneurship: An introduction to data bases and data management. Journal of Small Business Management, 30(20), 74-86.
Katz, J., & Gartner, W. B. (1988). Properties of emerging organizations. Academy of Management Review, 13(3), 429-441.
Livesay, H. C. (1979). American made: Men who shaped the American economy. Boston: Little Brown.
McKelvey, W. (1982). Organizational systemics: Taxonomy, evolution, classification. Berkeley: University of California Press.
Markey, J. P., & Parks, W. II (1989). Occupational change: Pursuing a different kind of work. Monthly Labor Review, September, 3-12.
Mintzberg, H. (1979). The structuring of organizations. Englewood Cliffs, NJ: Prentice Hall.
Morgan, J. N. [Principal Investigator] (1972). A panel study of income dynamics: Study, design, available data: 1968-1972 interviewing years (Volumes 1 and 2). Ann Arbor, MI: Survey Research Center, ISR.
Ornstein, S., & Isabella, L. A. (1993). Making sense of careers: A review 1989-1992. Journal of Management, 19(2), 243-267.
Osipow, S. H. (1973). Theories of career development (2nd ed.). Englewood Cliffs, NJ: Prentice Hall.
Parnes, H. S. (1976). Handbook for the national longitudinal surveys. Columbus: Ohio State University.
Perkins, D. N. T., Nieva, V. F., & Lawler, E. E. (1978). Causal forces in the creation of a new organization. Ann Arbor, MI: ISR.
Reynolds, P. D. (1988). Organizational births: Perspectives on the emergence of new firms. In F. Hoy (Ed.), Academy of Management Best Paper Proceedings 1988, pp. 69-73. Athens, GA: Academy of Management.
Roberts, E. B. (1991). Entrepreneurs in high technology. New York: Oxford University Press. Ronstadt, R. C. (1984a). Entrepreneurship: Text, cases and notes. Dover, MA: Lord.
Ronstadt, R. C. (1984b). Ex-entrepreneurs and the decision to start an entrepreneurial career. In J. A. Hornaday, F. Tarpley, Jr., J. A. Timmons, & K. H. Vesper (Eds.), Frontiers of entrepreneurship research, pp. 437-460. Wellesley, MA: Babson College.
Ronstadt, R. C. (1985). Every entrepreneur's nightmare: The decision to become an ex-entrepreneur and work for someone else. In J. A. Hornaday, E. B. Shils, J. A. Timmons, & K. H. Vesper (Eds.), Frontiers of entrepreneurship research, pp. 409-434. Wellesley MA: Babson College.
Ronstadt R. C. (1988). The corridor principle. Journal of Business Venturing, 13(1), 31-40.
Sarason, S. B., and associates (1972). The creation of settings and the future societies. San Francisco: Jossey-Bass.
Savage, D. (1979). Founders, heirs and managers: French industrial leadership in transition. Beverly Hills, CA: Sage.
Schein, E. H. (1978). Career dynamics: Matching individual and organizational needs. Reading, MA: Addison-Wesley.
Schein, E. H. (1985a). Career anchors: Discovering your real values. San Diego, CA: University Associates.
Schein, E. H. (1985b). Career anchors: Trainer's manual. San Diego, CA: University Associates.
Scherer, R. F., Bordzinski, J. D., & Weibe, F. A. (1991). Assessing perception of career role-model performance: The self-employed parent. Perceptual and Motor Skills, 72(2), 555-560.
Shapero, A., & Sokol, L. (1982). The social dimensions of entrepreneurship. In C. Kent, D. Sexton, & K. Vesper (Eds.), Encyclopedia of entrepreneurship. Englewood Cliffs, N J: Prentice Hall.
Sicherman, N., & Galor, O. (1990). A theory of career mobility. Journal of Political Economy, 98(1), 169-192.
Silver, A. D. (1985). Entrepreneurial megabucks: The 100 greatest entrepreneurs of the last twenty-five years. New York: Wiley.
Singh, J. V., Tucker, D. J., & House, R. J. (1986). Organizational legitimacy and the liability of newness. Administrative Science Quarterly, 31, 171-193.
Smith, N. R. (1967). The entrepreneur and his firm. E. Lansing, MI: Bureau of Business and Economic Research, MSU.
Smith, N. R., & Miner, J. B. (1983). Type of entrepreneur, type of firm, and managerial innovation: Implications for organizational life cycle theory. In J. A. Hornaday, J. Timmons, and K. Vesper (Eds.), Frontiers of entrepreneurship research, pp. 51-71. Wellesley, MA: Babson College.
Star, A. D. (1979). Estimates of the number of quasi and small businesses, 1948 to 1972. American Journal of Small Business, 4(2), 44-52.
Starr, J. A. (1992). Climbing the occupational ladder: Career development of habitual entrepreneurs. Paper presented at the annual meeting of the Academy of Management, Las Vegas, Nevada, August 11.
Stevenson, H. H., Roberts, M. J., & Grousbeck, H. I. (1989). New business ventures and the entrepreneur (3rd ed.). Homewood, IL: Irwin.
Timmons, J. A., Muzyka, D. F., Stevenson, H. H., & Bygrave, W. D. (1987). Venture capital and the search for potentially successful ventures: The characteristics of successful ventures. Paper presented at the Babson Entrepreneurship Research Conference.
Van deVen, A. H., Hudson, R., & Schroeder, D. M. (1984). Designing new business startups: Entrepreneurial, organizational, and ecological considerations. Journal of Management, 10(1), 87-107.
Weick, K. E. (1979). The social psychology of organizing (2nd ed.). Reading, MA: Addison-Wesley.
Jerome A. Katz is Professor of Management and Associate Director of the Jefferson Smurfit Center for Entrepreneurial Studies at Saint Louis University.
|Printer friendly Cite/link Email Feedback|
|Author:||Katz, Jerome A.|
|Publication:||Entrepreneurship: Theory and Practice|
|Date:||Dec 22, 1994|
|Previous Article:||Toward a theory of entrepreneurial careers.|
|Next Article:||Second-career entrepreneurs: a multiple case study analysis of entrepreneurial processes and antecedent variables.|