A Social Capital Theory of Career Success

A Social Capital Theory of Career SuccessAuthor(s): Scott E. Seibert, Maria L. Kraimer and Robert C. LidenSource: The Academy of Management Journal, Vol. 44, No. 2 (Apr., 2001), pp. 219-237Published by: Academy of ManagementStable URL: http://www.jstor.org/stable/3069452 .

Accessed: 08/10/2013 16:34

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Academy of Management is collaborating with JSTOR to digitize, preserve and extend access to The Academyof Management Journal.

http://www.jstor.org

This content downloaded from 152.78.130.202 on Tue, 8 Oct 2013 16:34:29 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=aom

http://www.jstor.org/stable/3069452?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


c Academy of Management Journal 2001, Vol. 44, No. 2, 219-237.

A SOCIAL CAPITAL THEORY OF CAREER SUCCESS

SCOTT E. SEIBERT MARIA L. KRAIMER

Cleveland State University

ROBERT C. LIDEN University of Illinois at Chicago

A model integrating competing theories of social capital with research on career success was developed and tested in a sample of 448 employees with various occupations and organizations. Social capital was conceptualized in terms of network structure and social resources. Results of structural equation modeling showed that network structure was related to social resources and that the effects of social resources on career success were fully mediated by three network benefits: access to information, access to resources, and career sponsorship.

Organizational researchers have begun to de-

velop increasingly comprehensive models of career success using demographic, human capital, work- family, motivational, organizational, and industry variables (e.g., Dreher & Ash, 1990; Judge & Bretz, 1994; Judge, Cable, Boudreau, & Bretz, 1995; Kirch-

meyer, 1998). Although this work has provided considerable evidence regarding the determinants of career outcomes, the roles of informal interpersonal behaviors have not been fully explored (Judge & Bretz, 1994; Pfeffer, 1989). Popular advice for getting ahead in one's career rarely fails to mention the importance of networking for the achievement of career goals (e.g., Bolles, 1992; Kanter, 1977). Indeed, Luthans, Hodgetts, and Rosenkrantz (1988) found that the most successful managers in their study spent 70 percent more time engaged in networking activities and 10 percent more time engaged in routine communication activities than their less successful counterparts. Recent advances in social capital theory (Coleman, 1990) have begun to provide a finer-grained analysis of the ways individuals' social networks affect their careers in organizations (Burt, 1992, 1997; Ibarra, 1995; Podolny & Baron, 1997; Sparrowe & Popielarz, 1995). This theoretical perspective has the poten-

Data were collected and the manuscript was submitted and processed while Scott E. Seibert was in the Manage- ment Department at the University of Notre Dame and Maria L. Kraimer was a graduate student at the Univer- sity of Illinois at Chicago. Support for this project was provided by the Management Department at the Univer- sity of Notre Dame and the Alumni Office of the Univer- sity of Notre Dame. The current investigation is part of a larger study of career success.

tial to considerably enhance scholars' knowledge of the role of social processes in career success.

The first purpose of the current study was to integrate the current conceptualizations of social

capital as they pertain to career success. Three different theoretical approaches-weak tie theory (Granovetter, 1973), structural hole theory (Burt, 1992), and social resource theory (Lin, 1990)- focus on different network properties as represen- tations of social capital. However, in all these theories, the key explanatory variables for the effect of social capital on career mobility are greater access to information, resources, and sponsorship. To date, these explanatory variables have not been included in empirical tests. The main contribution of the current investigation to the social network literature on careers is to provide a conceptual integration of the three major social capital theories followed by an empirical test that includes the proposed explanatory mechanisms.

A second purpose of this study was to model social capital effects on a full set of career outcomes, with other processes known to be determinants of career success controlled. Social network approaches to career success, growing out of a sociological research tradition, tend to focus on occupational status or job mobility as the primary career outcome (e.g., Burt, 1992, 1997; Granovetter, 1973; Ibarra, 1995; Lin, Ensel, & Vaughn, 1981a; Spar- rowe & Popielarz, 1995). These studies have often been limited to single organizations and small samples, have had few control variables, or have assessed outcomes over short time spans. Research in the organizational literature, on the other hand, tends to have large and diverse samples and models with broad sets of career processes. A model of career success that does not account for determi-

219



Academy of Management Journal

nants other than social capital may provide a bi- ased estimate of the effect of the latter on careers.

The organizational research has also moved toward the use of a set of extrinsic and intrinsic outcomes as measures of career success (e.g., Judge & Bretz, 1994; Judge et al., 1995; Kirchmeyer, 1998; Seibert, Grant, & Kraimer, 1999; Turban & Dough- erty, 1994; Wayne, Liden, Kraimer, & Graf, 1999). Extrinsic career outcomes are objectively observ- able achievements such as salary and promotions, and intrinsic career outcomes refer to individuals'

subjective feelings of accomplishment and satisfaction with their careers (London & Stumpf, 1982). Career scholars have argued that these are related but distinct constructs (Aryee, Chay, & Tan, 1994; Hall, 1976; Wayne et al., 1999) and that both measures are important because together they reflect not only conventional standards of success, but also feelings of success relative to an individual's own goals and expectations (Judge & Bretz, 1994; Judge et al., 1995; London & Stumpf, 1982; Seibert et al., 1999). By linking the social capital literature with this career success literature, we hoped to

provide a rigorous demonstration of the role of social capital in career success.

The third purpose of this study was to integrate research on social network structure with that on

mentoring and careers. Mentoring has been defined as a developmental relationship in which a less

experienced organization member receives help and guidance from a more experienced member whose intent is to improve the career opportunities and growth of the junior person (Kram, 1985). Re- search has explored the origin and progress of men-

toring relationships and the types of activities tak-

ing place within the mentor-prot6ge relationship (e.g., Chao, Walz, & Gardner, 1992; Noe, 1988; Tur- ban & Dougherty, 1994). The literature has also

emphasized the important effect that having a mentor may have on a protege's career success (e.g., Dreher & Ash, 1990). Kram (1985) suggested that it is important to understand the full constellation of

developmental relationships in which a protege can be involved. However, little research has ex-

plored (1) the simultaneous impact of multiple de-

velopmental contacts, (2) the way in which the

organizational location of these developmental contacts affects their contribution to career success, or (3) how the structure of an employee's social network facilitates access to developmental rela-

tionships in different parts of an organization. In this study, we used social capital theories to extend the mentoring literature by specifying the types of network structures that are likely to provide the most career benefits. At the same time, the social capital literature is enhanced through its link with

the extensive literature on developmental relation-

ships and the benefits they bring to careers.

THEORETICAL BACKGROUND

Theories of Social Capital

Coleman (1990) defined social capital as any as-

pect of social structure that creates value and facilitates the actions of the individuals within that social structure. Just as the creation of physical capital involves changes in materials so as to facilitate production, and human capital involves

changes in an individual's skills and capabilities, social capital is created when the relations among people change in ways that facilitate instrumental action (Coleman, 1990).

Social network researchers have taken the lead in

formalizing and empirically testing theories related to the concept of social capital. Social network researchers regard relationships, or ties, as the basic data for analysis. A network can be defined as the pattern of ties linking a defined set of persons or social actors. Each person can be described in terms of his or her links with other people in the network. The focal person in such an analysis (who is usu-

ally the person supplying the data) is referred to as

"ego," and those he or she is tied to are "alters" (Knoke & Kuklinski, 1982).

The first approach to the conceptualization of social capital, weak tie theory (Granovetter, 1973), focuses on the strength of the social tie used by a

person in the process of finding a job. Granovetter

argued that ties among members of a social clique are likely to be strong (defined as emotionally in- tense, frequent, and involving multiple types of

relationships, such as those with friends, advisors, and coworkers). The information possessed by any one member of the clique is likely to be either shared quickly or already redundant with the information possessed by the other members. How- ever, ties that reach outside of one's social clique are likely to be weak (that is, not emotionally in- tense, infrequent, and restricted to one narrow type of relationship) rather than strong. According to Granovetter (1973), weak ties are often a bridge between densely interconnected social cliques and thus provide a source of unique information and resources. Indeed, Granovetter (1973) found that weak ties were more likely than strong ties to have been the source of information about job openings for the sample of job incumbents he interviewed.

Subsequent research has provided mixed support for the weak tie hypothesis (Bridges & Villemez, 1986; McPherson, Popielarz, & Drobnic, 1992; Mur- ray, Rankin, & Magill, 1981).

April 220



Seibert, Kraimer, and Liden

Burt's (1992) structural holes approach to social capital focuses not on the characteristics of ego's direct ties, but on the pattern of relations among the alters in ego's social network. A structural hole is said to exist between two alters who are not connected to each other. According to structural holes theory, it is advantageous for ego to be connected to many alters who are themselves unconnected to the other alters in ego's network. According to Burt's theory (1992, 1997), networks rich in structural holes provide an individual with three primary benefits: more unique and timely access to information, greater bargaining power and thus control over resources and outcomes, and greater visibility and career opportunities throughout the social system. Burt (1992) critiqued weak tie theory, pointing out that the structural hole concept gets at the bridging property of ties more directly than the weak tie concept and therefore provides a "stronger foundation for theory and a clearer guide for empirical research" (1992: 28). Initial empirical evidence has been supportive of structural holes theory but has also provided a number of boundary conditions limiting the range of the theory's appli- cation (Burt, 1992, 1997; Podolny & Baron, 1997; Sparrowe & Popielarz, 1995). To date, the role of the proposed explanatory processes-access to information, bargaining control, and referral-have not been empirically examined.

The third major theoretical approach to the conceptualization of social capital is social resources theory (e.g., Lin, Ensel, & Vaughn, 1981a, 1981b). Social resources theory focuses on the nature of the resources embedded within a network. Lin and coauthors (1981a) argued that it is not the weakness of a tie per se that conveys advantage (nor, by extension, is it the bridging property of weak ties), but the fact that such ties are more likely to reach someone with the type of resource required for ego to fulfill his or her instrumental objectives. An alter who possesses characteristics or controls resources useful for the attainment of the ego's goals can be considered a social resource. For example, alters who provide career development advice and support are the relevant social resource when considering an ego's pursuit of instrumental career goals. Lin's research showed that tie strength was negatively related to the occupational prestige of the alter contacted (that is, weak ties reach higher- status alters) and that the alter's occupational prestige was in turn positively related to the prestige of the job secured by ego (Lin et al., 1981a, 1981b; see also De Graaf & Flap, 1988; Marsden & Hurlbert, 1988).

Integration of Social Capital Theories

As the literature review above reveals, controversy exists regarding the proper conceptualization of social capital. Weak tie theory focuses on the nature of ties; structural holes theory focuses on the pattern of the ties among alters; and social resource theory focuses on the characteristics of the alters contacted. The two later theories are each claimed to supersede the earlier theory (e.g., Burt, 1992; Lin et al., 1981a), and competitive model testing has been performed in an effort to empirically determine the best conceptualization (Sparrowe & Pop- ielarz, 1995).

Despite this controversy, a fruitful integration of the differing conceptualizations of social capital is possible. The key to this integration is to recognize an analytical distinction between the structural properties of networks and the nature of the social resources embedded in networks and to thus draw a distinction between their form and their content (see Lin, 1999). Weak tie theory and structural hole theory each focuses on the structure of a network. Social resources theory focuses on the content of a network. These theories are not mutually exclu- sive, as competitive model testing implies, but can function together because they focus on different points in the process of accumulating social capital. According to our conceptual integration, the overarching social capital construct is best thought of as both the different network structures that facilitate (or impede) access to social resources and the nature of the social resources embedded in the network. The key empirical question then becomes, "What network structures lead ego to have more (or less) access to important social resources?"

A MODEL FOR A SOCIAL CAPITAL THEORY OF CAREER SUCCESS

Figure 1 presents the model of social capital and career success tested in the current study. Accord- ing to the model, two measures of social network structure, weak ties and structural holes, are related to two forms of social resources, the number of contacts in other functions and the number of contacts at higher organizational levels. Social resources effects on career success are in turn mediated by three types of network benefits: access to information, access to resources, and career sponsorship. Career success is assessed in terms of current salary, the number of promotions received over the entire career, and career satisfaction. A full set of variables drawn from the organizational literature were used as controls in model tests but are not shown in Figure 1.

2001 221




FIGURE 1

Hypothesized Model of Social Capital Effects on Career Successa

Network Structure Social Resources Network Benefits Career Success

a Control variables and their paths are not shown for the sake of clarity. Current salary was a logarithm.

We examined the network of career developmental relationships maintained by ego within his or her organization in order to capture the social resources relevant for the instrumental ob-

jective of career success. Maintaining consis-

tency with previous research on social networks and promotions (Burt, 1992, 1997; Podolny & Baron, 1997), we focused on intraorganizational ties because contacts within an employee's own

organization were the ones we expected would

provide the kinds of benefits discussed in social

capital theories (information, resources, and

sponsorship) and would thus influence her or his success within the organization. We conceptualized social resources as developmental contacts in other functions and at higher organizational levels because, within formal organizations, functional or technical specialty and hierarchical level are likely to impose salient social boundaries between organization members (Burt, 1992; Ibarra, 1993). These organizational identity groups are likely to form their own interaction

cliques based on shared interests, values, train-

ing, socialization, or worldviews (Ibarra, 1995; Kanter, 1977; Lincoln & Miller, 1979). Develop- mental contacts who are members of social iden-

tity groups other than ego's are likely to provide unique information, resources, and influence useful to ego.

Social Network Structure and Social Resources

The fundamental tenet of weak tie theory is that a weak tie is valuable because it is more likely than a strong tie to act as a bridge between social cliques, providing an organization member with contact with people of different social groups or statuses (Granovetter, 1973, 1982). An assumption of social network approaches to social capital theory is that a person has a finite amount of time and energy to invest in social relationships. Given that, by defi- nition, strong ties require a greater investment of time and energy, the individual must make a stra-

tegic decision to invest his or her social energy in either maintaining a relatively small number of

strong ties or in developing a relatively large number of weak ties (Podolny & Baron, 1997). Our ar-

gument is not that a weak tie to a given individual is better than a strong tie to the same person. Rather, we seek to relate the number of weak ties, a structural property of an ego's network, to the number of valuable social contacts in ego's network. A social network characterized by many weak ties is more likely to provide access to critical social resources. That is, people who choose to invest their social energy in developing large numbers of weak ties will have greater access to social groups other than their own. Although some empirical evidence has suggested that bridging is more likely with weak ties than with strong ties (Friedkin, 1980), surprisingly little empirical research has specifi-

April 222




cally examined the extent to which weak ties bridge relatively unconnected social groups. Hy- potheses la and lb state that, within a career de-

velopment network, the number of weak ties will be related to the number of contacts in other social identity groups defined by the formal properties of an organization:

Hypothesis la. The number of weak ties in a network will be positively related to the number of ego's contacts in other organizational functions.

Hypothesis lb. The number of weak ties in a network will be positively related to the number of ego's contacts at higher organizational levels.

A structural hole is said to ,exist between two alters when they are unconnected to each other (Burt, 1992). An ego who is connected to two alters who are not connected to each other is, by defini- tion, a bridge between those alters. According to Burt (1992), this structural position conveys certain advantages to ego, in that he or she may be able to trade information gathered from one alter to the other. Ego provides added value to the organization through his or her ability to provide information and coordinate activities among separated alters (Burt, 1997; Galbraith, 1977). Further, to the extent that the two alters possess similar resources, ego can play them off against each other, seeking to set up social (or material) exchanges with the alter who offers the best return.

According to structural hole theory, an alter who is already connected to other alters within ego's network is "redundant" (Burt, 1992) and does not convey the kinds of benefits to ego that a nonredundant alter would. Parallel to our argument regarding weak ties is the argument that ego must make the strategic choice either to invest in maintaining a relationship with a redundant alter or to invest in developing a relationship with an alter who is not redundant with other alters in his or her network. The addition of a nonredundant alter in- troduces another set of structural holes into ego's network. Because members of the same social clique or social identity group are likely to be strongly connected to each other (e.g., Festinger, Schacter, & Back, 1950; Homans, 1950; Popielarz, 1994), structural holes are likely to be found between alters who are members of different social groups, such as those defined by functional and hierarchical boundaries within organizations (Burt, 1992; Ibarra, 1993). The fact that ego is acting as a bridge between two unconnected social groups am- plifies the benefits derived from acting as a bridge

between two unconnected individuals. Alters who are members of unconnected social groups are therefore uniquely beneficial to an ego. An ego whose network is rich in structural holes is therefore likely to have greater access to social resources as defined here-contacts in other functions and at higher organizational levels. Hypotheses 2a and 2b allow tests of this proposition:

Hypothesis 2a. The extent of structural holes in a network will be positively related to ego's number of contacts in other organizational functions.

Hypothesis 2b. The extent of structural holes in a network will be positively related to ego's number of contacts at higher organizational levels.

Social Resources and Network Benefits

The notion that social resources embedded in networks will provide benefits to actors is central to all three network approaches to social capital. In general, these benefits include greater and more timely access to information, greater access to fi- nancial or material resources, and greater visibility, legitimacy, or sponsorship within a social system. However, previous researchers have not incorpo- rated these constructs in their models. The overall purpose of the remaining set of hypotheses was to allow us to demonstrate that the effects of social capital on career success are through the three explanatory mechanisms.

The notion that different functional units within an organization have differing perspectives, con- trasting worldviews, and unique information is ax- iomatic to the information-processing perspective on organizations (e.g., Galbraith, 1977; March & Simon, 1958). The literatures on matrix or lateral organizational designs (Davis & Lawrence, 1977; Galbraith, 1994) and overlapping or cross-functional teams (Clark & Fujimoto, 1991) also contain the assumption that different functional groups need to share information across functional boundaries because the groups hold different information and views. This work suggests that an employee's contact with members of other organizational functions will provide access to information not available within his or her own functional group. Al- though other functions may also possess unique resources, these resources are less likely than is information to be available for transfer and of use across functional boundaries. And, unless they are at higher organizational levels, developmental contacts in other functions are not likely to have the

2001 223




status and influence to provide sponsorship to an

employee's career. Accordingly,

Hypothesis 3. The number of contacts in other

functions in a network will be positively related to ego's access to organizational information.

We expect that contacts at higher levels of the

organization will also be beneficial to an individual in a number of ways. A fundamental principle of rational organization design is that higher-level positions convey more authority than lower-level positions (March & Simon, 1958; Massie, 1965; We- ber, 1946). Ideally, higher positions also convey a broader perspective on issues relevant to an organization and greater access to information upon which incumbents can base decisions (Galbraith, 1977; March & Simon, 1958). According to the clas- sical principles of formal organizations, those in

positions at higher levels also have greater formal

decision-making authority over the allocation of resources than do those in lower-level positions. This legitimate basis of social power (French & Raven, 1968) means that higher-level individuals have greater formal power, influence, and control over resources. They may also be members of dom- inant coalitions within the organization and therefore enjoy informal bases of power, influence, and control over resources as well (Thompson, 1967). Positive relations with developmental contacts at

higher levels should therefore provide an actor with greater access to these benefits.

Hypothesis 4a. The number of an individual's contacts at higher organizational levels will be

positively related to access to organizational information.

Hypothesis 4b. The number of an individual's contacts at higher organizational levels will be

positively related to access to resources.

We expect contacts at higher organizational levels to also be related to the third network benefit, career sponsorship. Social network theorists have

variously referred to visibility, legitimacy, social

credentialing, and inclusion in career opportunities (Burt, 1992; Lin, 1999) as a benefit of specific network structures and social resources. Social network researchers have, however, overlooked the

conceptual overlap of this construct with the notion of career sponsorship developed in the literature on mentoring and careers (Kram, 1985; Noe, 1988). Career sponsorship, one of the primary functions fulfilled by a mentor (Kram, 1985; Noe, 1988), involves providing a protege with favorable and timely exposure, opportunities to engage in chal-

lenging assignments, and career advice and coach- ing. Because mentors are found among individuals at higher organizational levels (Kram, 1985), we

expect relations with developmental contacts at higher organizational levels to be related to receiv-

ing career sponsorship.

Hypothesis 4c. The number of an individual's contacts at higher organizational levels will be

positively related to career sponsorship.

Network Benefits and Career Success

There are two reasons to expect access to information and access to resources to each be related to

objective career success. First, greater access to information and resources should enhance individual work performance. Information and resources have been noted as contextual factors that empower employees, leading to higher levels of motivation and performance (e.g., Hackman & Oldham, 1980; Spreitzer, 1996). Burt (1992, 1997) argued that individuals able to use their network positions to fill a broker or boundary spanner role within an organization add greater value to the organization. In fact, centrality in the advice network has been linked to job performance (Sparrowe, Liden, Wayne, & Kraimer, 2001). The information-processing approach to organizational design also emphasizes the value added to organizations by individuals who play key boundary-spanning, liaison, conflict resolution, and coordination roles (e.g., Galbraith, 1977). Improved work performance and adding value should enhance an individual's objective career outcomes (Burt, 1992, 1997; London & Stumpf, 1983; Medoff & Abraham, 1981). Second, information and resources are fundamental bases of social

power (French & Raven, 1968). Greater access to information and resources will increase an individual's organizational reputation (Kilduff & Krack- hardt, 1994; Tsui, 1984), and the individual will be perceived as more powerful or influential in the organization (Brass, 1984; Brass & Burkhardt, 1993). These perceptions should make the individual better able to secure valuable organizational rewards independent of her or his actual performance (Ferris & Judge, 1991; Luthans et al., 1988).

We also expect access to information and access to resources to be positively related to career satisfaction. Having access to relevant organizational information and to resources such as funds, materials, and space should increase feelings both of control and competence at work (Gist & Mitchell, 1992) and of psychological empowerment (Spre- itzer, 1996). Psychological empowerment in general and self-determination and competence in par-

April 224




ticular are extensions of job design theory (Kraimer, Seibert, & Liden, 1999; Spreitzer, 1995, 1996), according to which enriched jobs are more satisfying to individuals (Hackman & Oldham, 1980; Spre- itzer, 1996). Thus, those who feel greater psychological empowerment with respect to their careers should be more satisfied with their career progress. Theory supports the existence of a relationship between access to information and career satisfaction and between access to resources and career satisfaction, but we are not aware of empirical research

directly testing these propositions. The above discussion leads to the following hy-

potheses regarding access to information:

Hypothesis 5a. An employee's access to information will be positively related to current salary, independent of other career outcomes.

Hypothesis 5b. An employee's access to information will be positively related to the number of promotions received over his or her career, independent of other career outcomes.

Hypothesis 5c. An employee's access to information will be positively related to his or her level of career satisfaction, independent of other career outcomes.

Our discussion also leads to these hypotheses regarding access to resources:

Hypothesis 6a. An employee's access to resources will be positively related to current salary, independent of other career outcomes.

Hypothesis 6b. An employee's access to resources will be positively related to the number of promotions received over her or his career, independent of other career outcomes.

Hypothesis 6c. An employee's access to resources will be positively related to her or his level of career satisfaction, independent of other career outcomes.

Although the relationships between access to information and career success and between access to resources and career success are thought to work through similar processes (work performance, power, reputation, and empowerment), these are considered to be independent constructs having unique effects on outcomes (Spreitzer, 1996). How- ever, it is also likely that access to information has a direct influence on access to resources. Informa- tion is a basis for social power (French & Raven, 1968), and access to and control over information can be translated into power and influence in organizations (Hickson, Hinings, Lee, Schneck, & Pen- nings, 1971; Pfeffer, Salancik, & Leblibici, 1978).

Possessing relevant organizational information would give individuals the ability to develop means of acquiring necessary resources other than

going through formal channels. In essence, information expands one's options for acquiring resources.

Hypothesis 7. An employee's access to information will be positively related to his or her access to resources.

The positive effects of mentoring in general and career sponsorship in particular on career outcomes have been amply demonstrated in the careers literature (Chao, 1997; Chao et al., 1992; Dreher & Ash, 1990; Wayne et al., 1999; Whitely, Dougherty, & Dreher, 1991). For example, Dreher and Ash (1990) found that, after they controlled for several demographic, human capital, and organizational variables, individuals who reported more extensive mentoring received more promotions, had higher incomes, and were more satisfied with their total compensation. Chao and colleagues (1992) performed a canonical correlation analysis and found career mentoring to be most strongly associ- ated with employees' intrinsic job satisfaction. Ac- cordingly, the following hypotheses were formu- lated to allow replication of previous research on career mentoring within an integrated model of social capital and career success:

Hypothesis 8a. The level of an employee's career sponsorship will be positively related to current salary, independent of other career outcomes.

Hypothesis 8b. The level of an employee's career sponsorship will be positively related to the number of promotions received over her or his entire career, independent of other career outcomes.

Hypothesis 8c. The level of an employee's career sponsorship will be positively related to her or his career satisfaction, independent of other career outcomes.

Control Variables

Organizational researchers have tested relatively comprehensive models of career success. Some specific findings are that human capital variables (education, years in the workforce, experience in multiple organizations, and career interruptions), demographic variables (gender, marital status, and spouse's employment status), and organizational characteristics (organization size, metropolitan location, and industry sector) have effects on salary,

225 2001




promotions, and/or career satisfaction (Judge & Bretz, 1994; Judge et al., 1995; Kirchmeyer, 1998; Seibert et al., 1999; Wayne et al., 1999). In addition, career satisfaction and salary have been found to

vary by occupation. Specific theoretical links between each category of variables and the career outcomes discussed above were beyond the scope of this study (see Judge et al., 1995) but these variables were included as controls.

METHODS

Sample and Procedures

A diverse sample of 2,781 randomly selected un-

dergraduate business, master of business administration (M.B.A.), and engineering school alumni of a large, private, midwestern university received

surveys at their home addresses. The alumni, who had graduated from 3 to 30 years prior to the date of the study, were instructed to complete the survey and return it directly to the first author in a postage- paid envelope. Confidentiality of survey responses was ensured to all alumni. To encourage responses, we entered all alumni respondents into a drawing for three prizes of approximately $50 in value. A reminder postcard was mailed three weeks after the initial mailing.

A total of 773 surveys were returned (a 28 percent response rate). Respondents who were not cur-

rently working (n = 73), working only part-time (n = 42), or self-employed (n = 116) were eliminated from the analyses because their career outcomes would not be comparable to those of alumni with current full-time employment. After cases with missing data had been eliminated, the final

sample consisted of 448 alumni. T-tests revealed that respondents and nonrespondents did not significantly differ with respect to gender, race, or

major (business versus engineering). However, the

percentage of responses from M.B.A. graduates (13%) was lower than the percentage they represented in the target sample (25%).

The demographic breakdown of the respondents was as follows: Their average age was 35.6 years; the average time since their graduation from the

university was 13.0 years; 65 percent were male; 72

percent were married; and 96 percent were Cauca- sian. Forty-three percent had a bachelor's degree as their highest degree attained, 9 percent had a master's degree other than an M.B.A., 41 percent had an M.B.A., 5 percent had a law degree, and 2 percent had a Ph.D. Average tenure in their current organization was 6.16 years.

Measures

Social capital. Respondents were asked to list (by initials) "the people who have acted to help your career by speaking on your behalf, providing you with information, career opportunities, advice or psychological support or with whom you have regularly spoken regarding difficulties at work, alternative job opportunities, or long-term career goals." We chose this description of a network relation on the basis of theoretical considerations, as it directly addresses the social resources mobilized in pursuit of instrumental career goals. Network size was the total number of alters indicated by a respondent.

For alters who were past or current members of a respondent's organization, the respondent also indicated the alter's organizational function relative to him or herself ("same function" or "different function") and organizational level ("lower," "same," or "higher"). Contacts in other functions is the number of individuals who were identified as members of a different function. The number of individuals who were identified as higher in organizational level formed the contacts at higher levels variable. Addi- tionally, we asked respondents to indicate how close they felt to the alter and how close the alters were to each other on a scale where 2 was "especially close," 1 was "less close," and 0 was "distant." Weak ties was the sum of all ties between the respondent and his/ her alters coded as 0 or 1. Following Burt (1992, 1997), we calculated structural holes, using the ego network data, as 1 minus the calculated value of the constraint posed by an individual alter j: cij = (Pij +

piqpqj)2, for q + i, j, where Pi is the proportion of i's relations invested in contact j, Piq is the proportion of i's relations invested in q, and pq is the proportion of q's relations invested in j. The total in parentheses is the proportion of i's relations that are directly or in-

directly invested in the connection with contact j. The sum of squared proportions, cij, is the constraint

posed by the entire network. One minus constraint (1 - cij) is thus the lack of constraint, or the degree of structural holes present in the ego network.

Network research typically relies on single-item sociometric questions, which by themselves do not provide information regarding measurement reliability. To ensure reliability, we pretested the survey instrument and allowed respondents to list a

large enough number of individuals to generate an accurate representation of their networks (Mars- den, 1990; Rogers & Kincaid, 1981). Sociometric questions were also designed to be as specific as

possible, to enhance accuracy of recall. Finally, questions were focused on typical, long-term rela-

tionships rather than on brief, transient, or episodic interactions. Research has shown that people's re-

April 226




call of brief, episodic interactions is highly inaccu- rate (Bernard, Killworth, Kronenfeld, & Sailer, 1984) but that people are remarkably accurate in

recalling typical interactions and long-term relationships (Freeman, Romney, & Freeman, 1987) like the developmental relationships analyzed in the current study.

Access to information and resources. The access variables were measured using Spreitzer's (1996) six-item scale. Three items were designed to measure access to information, and three were de-

signed to measure access to resources (1 = "strongly disagree" to 7 = "strongly agree"). An explor- atory factor analysis specifying a varimax rotation indicated that five of the six items formed two clear factors representing access to resources (a = .89) and access to information (a = .87). One item intended to measure access to information, "I have access to the strategic information I need to do my job well," cross-loaded and was therefore eliminated from further analyses.

Career sponsorship. The eight career sponsorship items from Dreher and Ash's (1990) global mentoring scale were summed to form a composite (a = .90). This scale assesses the extent to which senior colleagues have provided sponsorship, ex-

posure and visibility, challenging assignments, and

protection. Responses were made with a five-point scale (1 = "very little" to 5 = "a great deal"). An example item is, "Gone out of his/her way to pro- mote your career interests."

Career success. Career success was measured in terms of promotions, salary, and career satisfaction.

Respondents reported the number of promotions they had received over their entire careers. Promo- tions were defined as "any increases in level and/or any significant increases in job responsibilities or job scope." They also indicated their current annual salaries (including bonuses and other direct income). Self-reports of income have been shown to correlate highly with archival company records (Judge et al., 1995; Turban & Dougherty, 1994). Because a Z-test on the skewness statistic indicated a nonnormal distribution for salary (Z = 36.5, p < .001), we followed Gerhart and Milkovich's (1989) recommendation and used a natural logarithmic transformation of salary for all analyses. Career satisfaction was measured using Greenhaus, Parasura- man, and Wormley's (1990) five items (1 = "very dissatisfied" to 5 = "very satisfied"), which were summed to form a composite (a = .83). An example item is, "The progress I have made toward meeting my goals for advancement."

Control variables. These variables were gender (0 = "male," 1 = "female"); marital status (0 = "not married," 1 = "married"); spouse employment sta-

tus (0 = "spouse not employed," 1 = "spouse employed"); having an M.B.A. degree (1 = "yes"), having had an employment gap and, if so, the number of weeks it lasted; residence in a major metropolitan city, defined as one with a population over 1 million (1 = "yes"); the number of organizations worked for over the career; and having an occupation categorized as general management (1 =

"yes").1 The number of employees in a respondent's firm was scaled from 1, "less than 50 employees," to 5, "more than 1,000 employees." Num- ber of years in the workforce was measured by subtracting the year of graduation from the year the study was conducted. We also used network size as a control variable in order to examine the unique effects of developmental contacts in other functions and at higher levels as distinct from the sheer number of developmental contacts.

Analyses

The hypothesized structural equations model was tested using LISREL 8 (Joreskog & S6rbom, 1993), with the covariance matrix as input. Because no conventional estimate of error was available for the single-item measures, we assumed no error in the measurement model for the four social capital variables, the single-item control variables, and the two single-item dependent variables. The assumption of no error provided a conservative test of the

1 As previous research has shown differences in reported salary, promotions, and/or career satisfaction that are based on different occupations and industries, we asked respondents to indicate which of 11 categories best described their occupations and which of 12 categories best described their industries. We dummy-coded each of those variables so that manufacturing was the comparison category for all industries and accounting was the comparison category for all occupations. Rather than include every dummy-coded variable in the hypothesized model, we ran multivariate analyses of variance (MANOVAs) using career satisfaction, promotions, and salary as the dependent variables and industry type and occupation type as the independent variables. We then included only the statistically significant industry and occupation types in the LISREL analysis. The results of the MANOVA revealed that the only statistically significant occupation category was general management. Thus, the dummy code for general management is included in the hypothesized model as a control variable. Although two industry categories (nonprofit and finance) were significant in the MANOVA, their paths were not statistically significant in the initial LISREL analysis of the hypothesized model; thus, we deleted them from the hypothesized model before making further model comparisons.

2001 227




model. In order to increase sample size relative to the parameter estimates, we used single-scale score indicators to measure the other latent constructs in the structural model. For access to information, access to resources, career sponsorship, and career satisfaction, the measurement path estimates were set equal to 1 in order to scale the latent variables (Bollen, 1989), and the error variance was set equal to the scale variance times 1 minus the reliability in order to account for measurement error (Hayduk, 1987). We also allowed the error terms for the three

endogenous latent constructs for salary, promotions, and career satisfaction to be correlated in order to account for the correlations among these three constructs.

In testing the theoretical framework, we fitted several nested models to the data, each incorporating different assumptions about the model parameters. Comparisons with reasonable alternative models are recommended as a means of showing that a hypothesized model is the best representation of the data and are considered to be an important part of assessing model fit (Bollen, 1989; Kelloway, 1998). The first alternative model specifies only the direct paths from the control variables to the career outcome variables. This control-variables-only model provided a base- line fit for an assessment of the incremental contribution of the additional paths in the theoretical model. The relationships tested in the control-variables-only model were based on prior research findings. The hypothesized model specifies both the control variable paths and the set of paths hypothesized in this

study. With the remaining alternative models, we com-

pared the fully mediated hypothesized model with the partially mediated and nonmediated models, as recommended by Kelloway (1998). We examined three partially mediated models that assessed both the direct and indirect effects between our sets of constructs. The partially mediated models are tests of whether hypothesized mediated effects are better

represented as direct and indirect effects. We also examined two nonmediated models in order to assess whether the effects of social capital on career success were only direct, with no indirect effects through the network benefits. All of these alternative models included the control variable paths.

RESULTS

Preliminary Analyses

To assess the degree to which common method bias might present a problem, we subjected all the scale items for the variables used in this study to a principal components analysis using varimax

rotation (Harman, 1967). From this analysis, six clear factors representing the expected constructs emerged: career sponsorship, career satisfaction, access to resources, social capital, objective career success, and access to information. The average item loading on the intended construct was .75 and, of the 120 potential cross-loadings, only 3 were above .30, with the largest equal to .40. The absence of cross-loadings among the items for social capital, network benefits, and career success provides confidence that common method bias was not a problem for this study. Additionally, the fact that the access to resources, access to information, and career sponsorship items loaded on three sep- arate factors provides evidence of discriminant validity among our network benefits constructs.

Hypothesized Model

Table 1 reports the means, standard deviations, and correlations among the study variables. Our hypothesized model fit the data well (X = 191.11, df = 88, p < .01; root mean square error of approximation [RMSEA] = .05; adjusted goodness-of-fit index [AGFI] = .90; normed fit index [NFI] = .91; comparative fit index [CFI] = .95). Utilizing the change in

chi-square test (Bentler & Bonett, 1980), we compared our hypothesized model with a number of nested models (Table 2). The first comparison showed that the hypothesized model provided a significantly better fit than did the control-variables-only model (AX2 = 498.83, Adf = 18, p < .01). The second comparison was between the hypothesized model and the first partially mediated model (partially mediated model 1). Partially mediated model 1 specified the

paths in the hypothesized model as well as the direct

paths from the network structure constructs (weak ties and structural holes) to the network benefits (access variables and sponsorship). The change in chi-

square test showed that this alternative model was

significantly better than the hypothesized model (AX2 = 25.82, Adf = 6, p < .01).

Partially mediated model 1 was therefore retained as the best-fitting model and was then com-

pared with the second partially mediated model (partially mediated model 2) and the third partially mediated model (partially mediated model 3). Par-

tially mediated model 2 specifies the same paths as

partially mediated model 1 and also the direct paths from the social resources constructs (contacts across functions and at higher levels) to career outcomes. Partially mediated model 3 includes all direct paths from network structure to network benefits, from network structure to career outcomes, and from social resources to career outcomes, in addition to the hypothesized mediated effects. The

228 April



TABLE 1 Means, Standard Deviations, and Correlationsa

Variable Mean s.d. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

1. Contacts in other 0.99 1.33 functions

2. Contacts at higher 2.83 1.70 .27**

levels 3. Access to information 5.40 1.39 .16** .11*

4. Access to resources 4.88 1.33 .03 .09 .49** 5. Career sponsorship 2.91 0.86 .03 .13** .25** .24** 6. Salaryb 4.31 0.54 .09 .10* .21** .18** .16** 7. Promotions 4.76 2.88 .18** .14** .20** .13** .18** .49** 8. Career satisfaction 3.74 0.71 .08 .12** .35** .38** .38** .32** .23** 9. Weak ties 1.92 1.52 .17** .54** -.03 .01 -.12* .05 .06 -.05

10. Structural holes 0.35 0.25 .14** .45** .05 .05 .04 .02 .10* .08 .31**

11. Network size 5.32 1.99 .25** .48** .07 -.02 .05 -.10* -.01 .11* .37** .42**

12. Years since graduation 13.02 7.07 .09* .02 .04 -.02 -.03 .47** .41** -.03 .09 .07 -.16**

13. Female 0.35 0.48 -.06 .05 -.01 -.02 .07 -.22** -.16** .13** .02 .05 .02 -.30**

14. Married 0.72 0.45 .06 .01 .08 .04 .01 .31** .28** .05 .06 -.05 -.08 .33** -.20**

15. Spouse employment 0.49 0.50 -.02 -.05 -.01 -.02 -.01 -.02 .03 .06 .01 -.03 -.08 .04 .14** .61**

16. M.B.A. 0.41 0.49 .13** .04 .11* .01 -.05 .25** .19** -.01 .02 .00 -.07 .03 -.11* .06 -.03

17. Employment gap in 16.83 48.13 .04 -.06 .03 -.04 -.03 -.14** -.10* -.02 -.04 -.00 -.08 .14** .09* -.09 -.01 .08

weeks 18. Number of employees 4.16 1.28 -.07 .03 .11* .06 -.02 .19** .04 .13** -.02 .07 -.03 -.03 .09 .01 .00 .14** -.09

19. Metropolitan area 0.69 0.46 -.11* .03 -.04 -.04 .01 .15** -.01 .02 -.02 -.00 -.02 -.04 .02 -.07 -.04 .01 -.04 .16**

20. Number of employers 3.18 2.13 .12* -.01 -.06 -.09 -.11* -.00 .19** -.09 .01 .10* -.01 .26** -.09* .04 -.00 .10* .29** -.23** -.06

21. General management 0.14 0.35 .13** .05 .11* .13** .14** .30** .26** .16** .02 .03 .01 .20** -.12** .04 -.09 .03 .03 -.01 -.11* .03

an = 448. b Natural logarithm.

* p < .05

**p < .01




TABLE 2 Nested Model Comparisons

Model 2 (df) AX2 (Adjf RMSEA AGFI CFI NFI Model Comparisons

Hypothesized 191.11** ( 88) .05 .90 .95 .91

Control variables only 689.94** (106) -498.83** (18) .11 .74 .69 .67 Control variables compared to hypothesized model

Partially mediated 1 165.29** ( 82) 25.82** ( 6) .05 .91 .96 .92 Partially mediated model 1 compared to

hypothesized model

Partially mediated 2 158.41** ( 76) 6.88 ( 6) .05 .90 .96 .92 Partially mediated model 2 compared to partially mediated model 1

Partially mediated 3 154.46** ( 70) 10.83 (12) .05 .90 .96 .93 Partially mediated model 3 compared to partially mediated model 1

Nonmediated 1 297.90** ( 85) -143.44** ( 5) .08 .84 .89 .86 Nonmediated model 1 compared to partially mediated model 3 (nested models)

Nonmediated 2 301.25** ( 85) -146.79** ( 5) .08 .84 .89 .86 Nonmediated model 2 compared to partially mediated model 3 (nested models)

** p < .01

change in chi-square tests revealed that partially mediated models 2 and 3 were not significantly better than the first partially mediated model and were less parsimonious.

Lastly, the nonmediated model test comparisons were conducted. In the first nonmediated model, the paths from network benefits to career outcomes were constrained to zero, but the paths from social resources to career outcomes were freely estimated. The second nonmediated model also constrained the paths from network benefits to career outcomes to zero but allowed the paths from network structure to career outcomes to be estimated. Because the nonmediated models were not nested within

partially mediated model 1 but were nested within

partially mediated model 3, the nonmediated models were compared to partially mediated model 3. Both nonmediated models fit significantly worse than the third partially mediated model. Table 2

reports the change in chi-square for all nested model comparisons and other fit indexes for all seven comparative models.

The nested model comparisons indicated that the model including the hypothesized effects and direct effects from network structure to network benefits (partially mediated model 1) was the best-

fitting, most parsimonious model. Thus, we retained this model as the best-fitting model and interpret it below in order to examine the hypothesized relationships.

Examination of the standardized parameter estimates indicated that 14 of the 17 hypothesized relationships were significant and in the predicted

directions (see Figure 2) when the control variables were accounted for. Specifically, Hypotheses la and lb positively relate weak ties to contacts in other functions (la) and at higher levels (lb). The statistically significant parameter estimates (b = .14 and .44, respectively; p < .01) indicated support for Hypotheses la and lb. Hypotheses 2a and 2b positively relate structural holes to contacts in other functions (2a) and at higher levels (2b). A

statistically significant parameter estimate was found for the path between structural holes and

higher-level contacts (b = .31, p < .01) providing support for Hypothesis 2b, but the estimated path between structural holes and cross-function contacts was only marginally significant (b = .09, t =

1.94). Overall, the hypotheses relating network structure to social resources were supported.

Hypothesis 3 was supported, as a statistically significant parameter estimate was found for the

path between contacts in other functions and access to information (b = .15, p < .01). Hypotheses 4a, 4b, and 4c relate contacts at higher levels to access to information (4a), access to resources (4b), and career sponsorship (4c). A statistically significant parameter estimate was found for the path from contacts at higher levels to access to information (b = .14, p < .05) and from these contacts to career sponsorship (b = .29, p < .01), but the path between these contacts and access to resources was not significant. Thus, support was indicated for

Hypotheses 4a and 4c, but not for Hypothesis 4b.

Hypotheses 5a, 5b, and 5c positively relate access to information to salary (5a), promotions (5b), and

April 230




FIGURE 2 Results of Structural Equation Modela

Career -W Career \ Sponsorship e .32 o Satisfaction J

a Fit: X2 = 165.25, df = 82, p < .01; AGFI = .91, NFI = .92, CFI = .96. Parameter estimates are from the completely standardized solution

and are significant at p < .05. Hypothesized relationships are represented by bold arrows, and relationships that were not hypothesized are represented by light arrows. Hypothesized paths that were not significant were eliminated from the model. Control variables and their paths are not shown for the sake of clarity. Current salary was a logarithm.

career satisfaction (5c). The results indicated support for Hypotheses 5b (b = .11, p < .05) and 5c (b = .14, p < .05) but failed to support Hypothesis 5a. Respondents who indicated greater access to information reported more promotions and greater career satisfaction. Hypotheses 6a, 6b, and 6c positively relate access to resources to salary (6a), promotions (6b), and career satisfaction (6c). The results indicated support for Hypotheses 6a (b = .11, p < .05) and 6c (b = .28, p < .01). Respondents who indicated greater access to resources reported higher salaries and greater career satisfaction. Hy- pothesis 7 was supported as a significant, positive parameter estimate was found for the path from access to information to access to resources (b = .56, p < .01). Fully supporting Hypotheses 8a, 8b, and 8c, respondents who reported greater career sponsorship reported higher salaries, more promotions, and greater career satisfaction, as all three parameter estimates were statistically significant and positive (b = .12, .17, and .32, respectively; all p < .01).

Although not hypothesized, there were two other significant path estimates pertaining to social capital in partially mediated model 1: significant, neg- ative parameter estimates were found for the path from weak ties to access to information (b = -.14, p < .05) and for the path from weak ties to career sponsorship (b = -.28, p < .01). None of the paths from structural holes to the network benefit variables were statistically significant.

Finally, consistent with previous findings in the careers literature, several of the control variables were significantly (p < .05) related to the career success outcomes. Salary was predicted by years since graduation (b = .41), marital status (b = .22), spouse employment (b = -.14), M.B.A. degree (b = .22), employment gaps (b = -.18), size of employing organization (b = .12), metropolitan area (b = .16), and management occupation (b = .16). The number of promotions was predicted by years since graduation (b = .33), marital status (b = .11), M.B.A. degree (b = .17), employment gaps (-.20), number of employing organizations (b = .18), and management occupation (b = .14). Career satisfaction was predicted by network size (b = .11), gender (.11), and the size of the employing organization (b = .11). The social capital and control variables together explained 47 percent of the variance in current salary, 34 percent of the variance in promotions, and 36 percent of the variance in career satisfaction. The explained variance in the career outcomes was greater in the alternative hypothesized model than in the control-variables- only model: the latter explained 42 percent of salary, 29 percent of promotions, and 5 percent of career satisfaction.

DISCUSSION

Support of our hypothesized model revealed the importance of social capital to career success. Spe-

2001 231




cifically, our results demonstrate that two measures of network structure, weak ties and structural holes, positively relate to the level of social resources embedded in a person's network, measured as the number of developmental contacts in functional areas of an organization other than their own and at higher levels in the organization. Social resources were in turn positively related to current

salary, number of promotions over the career, and career satisfaction through their positive relation-

ships with three measures of network benefits- access to information, access to resources, and career sponsorship. The veridicality of these results was strengthened by the control of a diverse set of

demographic, human capital, and organizational variables as well as network size. This was clearly demonstrated by the significantly better fit of the model including the social capital and control variables over the alternative model containing only the control variables as predictors. Our findings have implications both for the social capital and careers literatures.

Social Capital

The current results provide support for the model of social capital used in the current study and help to resolve a growing controversy in the literature regarding the conceptualization of social

capital. Various scholars have tended to focus on

only one aspect of the social network as the defin-

ing element of social capital, but the current study demonstrates the analytic utility of separately de-

fining social resources and social network structure and empirically examining the ways in which network structure influences the level of social resources embedded in a network.

Specifically, the results of this study show that the number of weak ties and the level of structural holes in an actor's network each has independent effects on the level of social resources. As indicated

by the larger parameter estimates, the weak ties measure appears to have the stronger and more robust effect on social resources. However, weak ties were significantly, negatively related to two of the network benefit measures, access to information and career sponsorship. These results supply additional support for the traditional emphasis placed on the value of strong ties in providing information and social support (Festinger et al., 1950; Krackhardt, 1992). That is, people with whom one has a strong relationship are likely to

provide one with more information and assistance. Overall, these results redress the overemphasis in weak tie theory on the strength of the tie per se, rather than on its function as a bridge for social

benefits. Our results suggest that, from a practical standpoint, it may be best for a person to invest in the development of weak ties to increase the level of social resources embedded in his or her network, but then to invest (perhaps selectively) in strength- ening those ties to increase the benefits actually mobilized on his or her behalf.

Previous research has demonstrated direct relations between network structure and several organizational outcomes, such as promotions (Burt, 1992, 1997), influence (Brass & Burkhardt, 1993), and turnover (Krackhardt & Porter, 1986). The effects of social capital on these outcomes have often been theoretically explained (but not empirically tested) as occurring because of the access to information, resources, and sponsorship opportunities resulting from social contacts (e.g., Blau & Alba, 1982; Burt, 1997). The current study was unique for its inclusion of these mediating variables that aid in

understanding why social capital affects outcomes.

Specifically, we found support for the role of access to information, resources, and career sponsorship as full mediators of the relations between social

capital and career success. Support for the mediated rather than direct paths from social network variables to career outcomes suggests that mere

"schmoozing" with individuals outside of one's work unit will not affect career outcomes unless one is able to reap resources and sponsorship from these contacts (cf. Wayne & Liden, 1995).

Findings of the current investigation also suggest that the value of social capital may vary with re-

spect to the nature of the contacts that comprise a social network. Specifically, we found that devel-

opmental contacts in other functions were related to access to information and indirectly to access to resources but were not significantly related to career sponsorship. Developmental contacts at higher organizational levels were related to access to information and career sponsorship and, indirectly, to access to resources. Thus, the location of devel-

opmental contacts appears to be differentially related to the nature of the network benefits provided. Consistent with traditional mentoring research, higher-level contacts provided more career sponsorship than did those across functions. With respect to resource access, it is likely that contacts in other functions did not have formal

authority to directly provide resources to those in other units. It was, however, surprising that contacts at higher organizational levels did not directly relate to access to resources, but only indirectly provided resources by providing more information to their junior colleagues. Future research is needed to determine if the location of exchange contacts affects access to information, resources,

April 232




and sponsorship opportunities differently at each stage in a career. For example, close contacts with the immediate superior and peers may be most beneficial during an employee's early socialization in an organization (Liden, Wayne, & Stilwell, 1993; Major, Kozlowski, Chao, & Gardner, 1995), but after she or he has been assimilated into the organization, contacts with individuals at higher levels and other functions may prove most useful.

The varying effects of the mediating variables on career outcomes also invites additional research. Although simple correlations between the three mediating variables and the three career outcomes were all positive and significant, path estimates within the model showed that career sponsorship was the most important network benefit in terms of career success. The importance of career sponsorship relative to resource and information access should be verified in future research. Future research should also examine the effects of other types of networks, such as those composed of friendship and interorganizational ties, on career success.

Careers

The results of the current investigation strongly suggest the relevance of integrating social capital theory with research on careers. Previous research demonstrating the influence of social capital on career success variables such as promotions (Brass, 1984; Burt, 1992) has not been linked to the literature on careers. Similarly, most studies in the careers literature have not included social capital as an antecedent of career success (Judge & Bretz, 1994; Judge et al., 1995; Wayne et al., 1999). Inves- tigations of career success that have acknowledged the importance of social capital have typically not used social network methods (e.g., Gould & Penley, 1984; O'Hara, Beehr, & Colarelli, 1994). To our knowledge, no previous attempt has been made to examine social capital within the context of the careers literature.

In the current study, an explicit attempt was made to merge the literatures on social capital and careers. Specifically, many of the variables included in comprehensive investigations of career success (Judge et al., 1995) were tested as controls. Thus, the added contribution of social capital, op- erating through the mediating variables, was clearly demonstrated. Additionally, most of the research on networks and careers has focused on promotion rates or mobility. By incorporating both objective and subjective measures of career success, we were able to more readily compare the social capital effects on career success found here

with findings from other career success research in the organizational literature. These results provide a strong incentive for future researchers to consider social capital as a key variable in relation to both objective and subjective career success.

The findings of the current study also have important implications for a key subfield of the careers research, mentoring. It is assumed in some mentoring studies that proteges have only one mentor at a time (Chao et al., 1992; Judge & Bretz, 1994; Tepper, 1995). In other studies (Dreher & Ash, 1990; Turban & Dougherty, 1994), researchers have assessed the amount of mentoring received without differentiating between respondents with single and those with multiple mentors. Our results show that individuals with multiple mentors reap greater career benefits than those having only one mentor. Furthermore, the results indicate that the more contacts that people establish at high organizational levels, the more they gain from career sponsorship. Kram's (1985; Kram & Isabella, 1985) interviews suggested that peers play an important role in career development, but the current investigation is unique because it relates a full set of developmental relationships involving superiors and peers situ- ated in different parts of organizations to level of career success. In essence, we have expanded the study of mentorship to include a network of mentors who should be located in various functions of and at higher levels in the focal employee's organization.

This notion of a "mentoring network" suggests many topics for future research. For example, superiors may be beneficial not only for the mentoring that they personally provide, but also for intro- ducing subordinates to influential organization members who subsequently play a role in mentoring the subordinates (Sparrowe & Liden, 1997). Longitudinal research could assess the way in which a newcomer to an organization develops a mentoring network. Future research might also examine the type of mentoring provided by each mentor. For example, one mentor might provide emo- tional support and spiritual guidance. Another mentor might help the protege with task and technical advice. Yet another mentor might assist the protege in coping with organizational politics. Dis- covering that each mentor does indeed provide a unique type of advice or assistance would explain our finding that the larger the mentoring network, the more beneficial it is to the protege. In essence, the more mentors, the greater the coverage of different types of advice and support. Conversely, our results demonstrate the constraining weakness of multiplex ties in which an individual is in the precarious position of relying on the same tie(s) for

2001 233




multiple types of assistance (Podolny & Barron, 1997).

In summary, the current study makes several contributions to the literature: (1) We enhanced research on social networks by integrating contend- ing theories of social capital, testing an analytical distinction between network structure and network resources and testing for their relationship. (2) We extended the social capital and careers literatures

by testing for the effects of social capital on a full set of career outcomes after controlling for other variables related to careers. (3) Our integrated social capital theory of career success was further

supported by the findings that information, resource access, and sponsorship play a mediating role in the relation between social network variables and career outcomes. And (4) we augmented the mentoring literature by examining the influence that multiple developmental relationships have on career success and by assessing the way in which the location of the developmental contact in an organization influenced a protege's access to information, resources, and career sponsorship.

Limitations

Although its large sample and numerous control variables represent strengths of the study, a limitation is that, owing to the cross-sectional design, it was not possible to make inferences concerning causal direction for the paths tested in our model. For example, rather than higher-level contacts providing information to focal individuals, perhaps focal individuals who have gained access to information are more attractive to contacts at higher organizational levels. Another limitation associ- ated with the design is that valid variance in career outcomes explained by differences in organizations, culture, or norms was treated as error. An ideal study would include a large sample of individuals within a large sample of organizations. This would enable the researcher to examine organization type, and accompanying culture and norms, as

multiple-level effects in a comprehensive model of career success.

A second limitation of the study is the possibility of common method bias, which can inflate relation-

ships among variables. We took several steps to minimize the problem, including separating the items for the independent and dependent variables into different sections of the survey instrument and using different question formats for each set of variables. Because the social capital variables are behavioral and the career outcome variables are fac- tual or attitudinal, common method bias should have been minimized.

A third limitation is that although the initial re-

sponse rate of 28 percent is typical for mailed surveys, the effective response rate of 17.3 percent that pertained after we had removed responses from self-employed individuals and part-time employees and those with missing data is somewhat low.

Although nonrespondents did not differ from re-

spondents on demographic characteristics or major in school, it was found that the response rate for M.B.A. graduates was lower than that for under- graduate alumni. It is therefore possible that the sample is not as representative of the population of M.B.A.'s as it is of the population of undergradu- ates. A final limitation of our research is the use of

single-item scales for some of the variables. Al- though single-item scales are common in social network research, it would be preferable in future research to use three or more items for each measure so that reliability can be estimated.

In sum, the path model tested in the current

investigation represents an attempt to test an inte-

grated theory of social capital and career success.

Support for our hybrid model suggests that both the social capital and careers literatures can be enhanced through integration. It follows that future research on career success would benefit from the inclusion of social capital variables.

REFERENCES

Aryee, S., Chay, Y. W., & Tan, H. H. 1994. An examination of the antecedents of subjective career success among a managerial sample in Singapore. Human Relations, 47: 487-509.

Bentler, P. M., & Bonett, D. G. 1980. Significance tests and goodness of fit in the analysis of covariance structures. Psychological Bulletin, 88: 588-606.

Bernard, R. H., Killworth, P., Kronenfeld, D., & Sailer, L. 1984. The problem of informant accuracy: The validity of retrospective data. In B. J. Siegel (Ed.) & A. R. Beals & S. A. Tyler (Assoc. Eds.), Annual review of anthropology, vol. 13: 495-517. Palo Alto, CA: Annual Reviews.

Blau, J. R., & Alba, R. D. 1982. Empowering nets of

participation. Administrative Science Quarterly, 27: 363-379.

Bollen, K. A. 1989. Structural equations with latent variables. New York: Wiley.

Bolles, R. N. 1992. What color is your parachute? Berke-

ley, CA: Ten Speed Press.

Brass, D. J. 1984. Being in the right place: A structural

analysis of individual influence in an organization. Administrative Science Quarterly, 29: 518-539.

Brass, D. J., & Burkhardt, M. E. 1993. Potential power and power use: An investigation of structure and behav-

April 234




ior. Academy of Management Journal, 36: 441- 470.

Bridges, W. P., & Villemez, W. J. 1986. Informal hiring and income in the labor market. American Sociolog- ical Review, 51: 574-582.

Burt, R. S. 1992. Structural holes: The social structure

of competition. Cambridge, MA: Harvard University Press.

Burt, R. S. 1997. The contingent value of social capital. Administrative Science Quarterly, 42: 339-365.

Chao, G. 1997. Mentoring phases and outcomes. Journal

of Vocational Behavior, 51: 15-28.

Chao, G., Walz, P., & Gardner, P. 1992. Formal and informal mentorships: A comparison on mentoring functions and contrast with nonmentored counterparts. Personnel Psychology, 45: 619-636.

Clark, K., & Fujimoto, T. 1991. Product development performance. Boston: Harvard Business School Press.

Coleman, J. S. 1990. Foundations of social theory. Cam-

bridge, MA: Harvard University Press.

Davis, S. M., & Lawrence, P. R. 1977. Matrix. Reading, MA: Addison-Wesley.

De Graaf, N. D., & Flap, H. D. 1988. With a little help from

my friends: Social capital as an explanation of occu-

pational status and income in the Netherlands, the United States and West Germany. Social Forces, 67: 452-472.

Dreher, G. F., & Ash, R. A. 1990. A comparative study of

mentoring among men and women in managerial, professional, and technical positions. Journal of Ap- plied Psychology, 75: 539-546.

Ferris, G. R., & Judge, T. A. 1991. Personnel/human resources management: A political influence perspective. Journal of Management, 17: 447-488.

Festinger, L., Schacter, S., & Back, K. W. 1950. Social

pressures in informal groups. Stanford, CA: Stan- ford University Press.

Freeman, L. C., Romney, A. K., & Freeman, S. C. 1987.

Cognitive structure and informant accuracy. Ameri- can Anthropologist, 89: 310-325.

French, J. R. & Raven, B. 1968. The bases of social power. In D. Cartwright & A. Zander (Eds.), Group dynamics (3rd ed.): 259-269. New York: Harper & Row.

Friedkin, N. E. 1980. A test of the structural features of Granovetter's strength of weak ties theory. Social Networks, 2: 411-422.

Galbraith, J. R. 1977. Organizational design. Reading, MA: Addison-Wesley.

Galbraith, J. R. 1994. Competing with flexible lateral

organizations (2nd ed.). Reading, MA: Addison- Wesley.

Gerhart, B. A., & Milkovich, G. T. 1989. Salaries, salary growth, and promotions of men and women in a

large, private firm. In R. T. Michael, H. I. Hartmann, & B. O'Farrell (Eds.), Pay equity: Empirical inquir- ies: 23-43. Washington, DC: National Academy Press.

Gist, M. E., & Mitchell, T. N. 1992. Self-efficacy: A theoretical analysis of its determinants and malleability. Academy of Management Review, 17: 183-211.

Gould, S., & Penley, L. E. 1984. Career strategies and

salary progression: A study of their relationships in a

municipal bureaucracy. Organizational Behavior and Human Performance, 34: 244-265.

Granovetter, M. S. 1973. The strength of weak ties. Amer- ican Journal of Sociology, 6: 1360-1380.

Granovetter, M. S. 1982. The strength of weak ties: A network theory revisited. In P. V. Marsden & N. Lin (Eds.), Social structure and network analysis: 105- 130. Beverly Hills, CA: Sage.

Greenhaus, J. H., Parasuraman, S., & Wormley, W. M. 1990. Effects of race on organizational experiences, job performance evaluations, and career outcomes.

Academy of Management Journal, 33: 64-86.

Hackman, J. R., & Oldham, G. R. 1980. Work redesign. Readings, MA: Addison-Wesley.

Hall, D. T. 1976. Careers in organizations. Santa Monica, CA: Goodyear.

Harman, H. H. 1967. Modern factor analysis (2nd ed.). Chicago: University of Chicago Press.

Hayduk, L. A. 1987. Structural equation modeling with LISREL: Essentials and advances. Baltimore: John Hopkins University Press.

Hickson, D. J., Hinings, C. A., Lee, C. A., Schneck, R. E., & Pennings, J. M. 1971. A "strategic contingencies" theory of intraorganizational power. Administrative Science Quarterly, 14: 378-397.

Homans, G. C. 1950. The human group. New York: Har- court, Brace & World.

Ibarra, H. 1993. Network centrality, power, and innova- tion involvement: Determinants of technical and administrative roles. Academy of Management Jour- nal, 36: 471-501.

Ibarra, H. 1995. Race, opportunity, and diversity in social circles in managerial networks. Academy of Man-

agement Journal, 38: 673-703.

Joreskog, K. G., & Sorbom, D. 1993. LISREL 8: User's

reference guide. Chicago: Scientific Software Inter- national.

Judge, T. A., & Bretz, R. D. 1994. Political influence behavior and career success. Journal of Manage- ment, 20: 43-65.

Judge, T. A., Cable, D. M., Boudreau, J. W., & Bretz, R. D. 1995. An empirical investigation of the predictors of executive career success. Personnel Psychology, 48: 485-519.

235 2001




Kanter, R. M. 1977. Men and women of the corporation. New York: Basic Books.

Kelloway, E. K. 1998. Using LISREL for structural equation modeling: A researcher's guide. Thou- sand Oaks, CA: Sage.

Kilduff, M., & Krackhardt, D. 1994. Bringing the individual back in: A structural analysis of the internal market for reputation in organizations. Academy of Management Journal, 37: 87-109.

Kirchmeyer, C. 1998. Determinants of managerial career success: Evidence and explanation of male/female differences. Journal of Management, 24: 673-692.

Knoke, D., & Kuklinski, J. H. 1982. Network analysis (Sage university paper series on quantitative appli- cations in the social sciences no. 07-028). Beverly Hills, CA: Sage.

Krackhardt, D. 1992. The strength of strong ties: The

importance of philos in organizations. In R. Eccles & N. Nohria (Eds.), Networks and organizations: Structure, form and action: 216-239. Boston: Har- vard Business School Press.

Krackhardt, D., & Porter, L. W. 1986. The snowball effect: Turnover embedded in communication networks. Journal of Applied Psychology, 71: 50-55.

Kraimer, M. L., Seibert, S., & Liden, R. C. 1999. Psycho- logical empowerment as a multidimensional construct: A test of construct validity. Educational and

Psychological Measurement, 59: 127-142.

Kram, K. E. 1985. Mentoring at work: Developmental relationships in organizational life. Glenview, IL: Scott, Foresman.

Kram, K. E., & Isabella, L. A. 1985. Mentoring alterna- tives: The role of peer relationships in career devel-

opment. Academy of Management Journal, 28: 110-132.

Liden, R. C., Wayne, S. J., & Stilwell, D. 1993. A longitudinal study on the early development of leader- member exchanges. Journal of Applied Psychology, 78: 662-674.

Lin, N. 1990. Social resources and social mobility: A structural theory of status attainment. In R. L. Breiger (Ed.), Social mobility and social structure: 247- 271. New York: Cambridge University Press.

Lin, N. 1999. Building a network theory of social capital. Connections, 22(1): 28-51.

Lin, N., Ensel, W. M., & Vaughn, J. C. 1981a. Social resources and occupational status attainment. Social Forces, 59: 1163-1181.

Lin, N., Ensel, W. M., & Vaughn, J. C. 1981b. Social resources and strength of ties. American Sociologi- cal Review, 46: 393-405.

Lincoln, J. R. & Miller, J. 1979. Work and friendship ties in organizations: A comparative analysis of rela- tional networks. Administrative Science Quarterly, 24: 181-199.

London, M., & Stumpf, S. A. 1982. Managing careers. Reading, MA: Addison-Wesley.

London, M., & Stumpf, S. A. 1983. Effects of candidate characteristics on management promotion decisions: An experimental study. Personnel Psychology, 36: 241-259.

Luthans, F., Hodgetts, R. M., & Rosenkrantz, S. A. 1988. Real managers. Cambridge, MA: Ballinger.

Major, E. A., Kozlowski, S. W., Chao, G. T., & Gardner, P. D. 1995. A longitudinal investigation of newcomer expectations, early socialization outcomes, and the moderating effects of role development factors. Jour- nal of Applied Psychology, 80: 418-429.

March, J. G., & Simon, H. A. 1958. Organizations. New York: Wiley.

Marsden, P. V. 1990. Network data and measurement. In W. R. Scott & J. Blake (Eds.), Annual review of sociology, vol. 16: 435-463. Palo Alto, CA: Annual Reviews.

Marsden, P. V., & Hurlbert, J. S. 1988. Social resources and mobility outcomes: A replication and extension. Social Forces, 66: 1039-1059.

Massie, J. L. 1965. Management theory. In J. G. March (Ed.), Handbook of organizations: 387-422. Chi- cago: Rand McNally.

McPherson, J. M., Popielarz, P. A., Drobnic, S. 1992. Social networks and organizational dynamics. American Sociological Review, 57: 153-170.

Medoff, J. L., & Abraham, K. G. 1981. Are those paid more really more productive? The case of experience. Journal of Human Resources, 16: 186-216.

Murray, S. O., Rankin, J. H., & McGill, D. W. 1981. Strong ties and job information. Sociology of Work and Occupations, 8: 119-136.

Noe, R. A. 1988. An investigation of the determinants of successful assigned mentoring relationships. Per- sonnel Psychology, 41: 457-479.

O'Hara, K. B., Beehr, T. A., & Colarelli, S. M. 1994.

Organizational centrality: A third dimension of in-

traorganizational career movement. Journal of Ap- plied Behavioral Science, 30: 198-216.

Pfeffer, J. 1989. A political perspective on careers: Inter- ests, networks, and environments. In M. G. Arthur, D. T. Hall, & B. S. Lawrence (Eds.), Handbook of career theory: 380-396. New York: Cambridge Uni- versity Press.

Pfeffer, J., Salancik, G. R., & Leblibici, H. 1978. Uncer-

tainty and social influence in organizational decision making. In M. W. Meyer (Ed.) Environments and organizations: 306-332. San Francisco: Jossey- Bass.

Podolny, J. M., & Baron, J. N. 1997. Resources and rela-

tionships: Social networks and mobility in the work-

place. American Sociological Review, 62: 673-693.

Popielarz, P. A. 1994. Do weak ties span longer dis-

236 April




tances? Paper presented at the International Social Networks Conference, New Orleans.

Rogers, E. M., & Kincaid, D. L. 1981. Communication networks. New York: Free Press.

Seibert, S. E., Crant, J. M., & Kraimer, M. L. 1999. Proac- tive personality and career success. Journal of Ap- plied Psychology, 84: 416-427.

Sparrowe, R. T., & Liden, R. C. 1997. Process and structure in leader-member exchange. Academy of Man-

agement Review, 22: 522-552.

Sparrowe, R. T., Liden, R. C., Wayne, S. J., & Kraimer, M. L. 2001. Social exchange processes, social structure, and performance. Academy of Management Journal, 44: 316-325.

Sparrowe, R. T., & Popielarz, P. A. 1995. Weak ties and structural holes: The effects of network structure on careers. Paper presented at the annual meeting of the Academy of Management, Vancouver.

Spreitzer, G. M. 1995. Individual empowerment in the

workplace: Dimensions, measurement, and valida- tion. Academy of Management Journal, 38: 1442- 1465.

Spreitzer, G. M. 1996. Social structural characteristics of psychological empowerment. Academy of Manage- ment Journal, 39: 483-504.

Tepper, B. J. 1995. Upward maintenance tactics in super- visory mentoring and nonmentoring relationships. Academy of Management Journal, 38: 1191-1205.

Thompson, J. D. 1967. Organizations in action. New York: McGraw-Hill.

Tsui, A. S. 1984. A role set analysis of managerial reputation. Organizational Behavior and Human Per- formance, 34: 64-96.

Turban, D. B. & Dougherty, T. W. 1994. Role of protege personality in receipt of mentoring and career success. Academy of Management Journal, 37: 688- 702.

Wayne, S. J., & Liden, R. C. 1995. Effects of impression management on performance ratings: A longitudinal study. Academy of Management Journal, 38: 232- 260.

Wayne, S. J., Liden, R. C., Kraimer, M. L., & Graf, I. K. 1999. The role of human capital, motivation, and supervisor sponsorship to career success. Journal of Organizational Behavior, 20: 577-595.

Weber, M. 1946. From Max Weber: Essays in sociology (H. H. Gerth & C. W. Mills, eds.). New York: Oxford University Press.

Whitely, W., Dougherty, T. W., & Dreher, G. F. 1991. The relationship of career mentoring and socioeconomic origin to managers' and professionals' early career

progress. Academy of Management Journal, 34: 331-351.

Scott E. Seibert (Ph.D., Cornell University) is an assistant professor of management and labor relations at the James J. Nance College of Business Administration, Cleveland State University. His interests in personality, interpersonal processes, and social networks inform his research on group decision making, mentoring, and career success. He will be joining the department of managerial studies at the University of Illinois at Chicago in the fall of 2001.

Maria L. Kraimer (Ph.D., University of Illinois at Chi- cago) is an assistant professor of management and labor relations at the James J. Nance College of Business Ad- ministration, Cleveland State University. Her research interests include career-related issues, employee adjust- ment, and the employee-employer relationship. She will be joining the department of managerial studies at the University of Illinois at Chicago in the fall of 2001.

Robert C. Liden (Ph.D., University of Cincinnati) is a professor of management at the University of Illinois at Chicago. His research focuses on interpersonal processes within the context of such topics as leadership, groups, career progression, and employment interviews.

2001 237



A Social Capital Theory of Career Success

Documents