(Get Answer)

business writing question and need guidance to help me learn. For this discussion assignments, I am requesting you to identify a video or program link from the internet that describes a networked approach to conflict resolution (you can decide the setting/context) and briefly critique 2 ways in which the approach is grounded (or not) in the network literature these first two weeks. The literature you from the first two weeks is attached. Requirements: 250 2Understanding Social Network Research•Distinctive features of the social network approach17•Major concepts in social network research25•Summary33•Recommended further reading34The network concept is one of the defining paradigms of the modern era. Infields as different as physics, biology, linguistics, anthropology, sociology andpsychotherapy, network ideas have been repeatedly invoked over the last hundredyears. The network approach allows researchers to capture the interactions ofany individual unit within the larger field of activity to which the unit belongs.The multiple origins of network approaches for the social sciences con-tribute to the eclecticism that characterizes current work. Briefly stated, net-work ideas flowed into the social sciences from three main sources. First,German researchers (such as Kurt Lewin, Fritz Heider and Jacob Moreno), influ-enced by developments in field theory in physics, transferred the network ideato the examination of social interaction. These scientists brought their distinc-tive new approach to the USA during the 1920s and 1930s. Network researchon cognition and interpersonal influence originates with the influential tradi-tions of Lewin and Heider.Secondly, the influence of a mathematical approach to social interaction,evident in Kurt Lewin’s work, was taken up in the USA first by researchers work-ing with graph theory (e.g., Cartwright and Harary, 1956), and later by a Harvardgroup working with Harrison White. This emphasis on mathematics helpedtransform the study of social networks from description to analysis. With theadvent of powerful computers, the promise of the network approach began to berealized: individual units within social fields could be simultaneously analysed todiscover new insights concerning social structure and interaction.CHAPTER CONTENTS The third main source of network ideas in the social sciences derived notfrom mathematically-inclined sociologists but from anthropologically-inclinedorganizational fieldworkers. In the USA, a group based in the Harvard BusinessSchool began in the 1920s a ten-year series of anthropological investigations offactory life in the Hawthorne works of the Western Electric Company ofChicago. The famous Hawthorne Studies were the first to use sociograms todiagram the structure of freely-chosen social interactions. Thus, from the verybeginning, social network analysis had its roots in organizational settings. The Hawthorne researchers were not the only anthropologically-inclinedresearchers who contributed to the developing science of social networks inorganizational settings. A British tradition, centred around the Department ofSocial Anthropology at Manchester University, inspired innovative examinationsof organizational conflict from a social network perspective. In particular, BruceKapferer’s analyses of social interaction, change, and conflict in African work-places such as a garment factory (1972) advanced the practice and the science ofsocial network research. Kapferer, following the innovative social communityresearch of his mentors, such as Barnes, Mitchell and Bott, collected data on theinteractions of every employee of an Indian-owned clothing factory in theZambian town of Kabwe. He tested a series of hypotheses derived from exchangetheory (Blau, 1964) rather than resting content with a purely descriptive accountof factory life. He examined how the social networks of interaction changed overtime in relation to significant events occurring in the factory. Thus, Kapferer wasable to examine a complete network of interaction over time and relate it to sub-stantively interesting organizational issues. It’s worth spending a little time onKapferer’s exemplary research because it illustrates both the history of the net-work approach and also the contemporary possibilities for analysis. To assistreaders in following the technical terms used in this discussion and in the rest ofthe book, we provide a glossary of terms at the end of the book.Kapferer emphasized that the social composition of the factory was an emer-gent property of choices and decisions made by interacting individuals. Hecharted the changes in social networks by collecting network data at three pointsin time. From these data he computed network measures of the extent to whichemployees achieved organizational power and influence through being able toaccess and mobilize people in the factory, anticipating current theoretical workon how individuals’ networks can span across social divides (e.g., Burt, 1992).The most dramatic innovation in Kapferer’s work was his use of socialnetwork data to predict strike activity by the workers. Figures 2.1 and 2.2 depictthe instrumental network in the factory at two points in time. The instrumen-tal network was defined as including such transactional activities as ‘lending orgiving money, assistance at times of personal crisis and help at work’ (Kapferer,1972: 164). He excluded activities that were mandated by the productionprocess itself. At the end of time 1 some senior workers organized walkouts totry to secure wage and work improvements, but their efforts failed to gain thesupport of many of the skilled and unskilled workers, and thus ultimately wereSocial Networks and Organizations14 FIGURE 2.1ZakeyoKalundweSignMubangaChristianJosephChilufyaMabangeHenryAngelIbrahimAbrahamKalambaKalongaChisokoneEnochJohnMukubwaChilwaHastingsBenNyirendaDonaldKamwefuSeamsChipataLyashiLwangaZuluNkoloyaNkumbulaWilliamLmpunduChobeMeshakdeemed a failure. Note the relatively dispersed leadership structure evident inFigure 2.1: the graph has a relatively low degree centralization index of .28,indicating the absence of informal leaders around whom the other employeesare organized (Scott, 2000: 89). Degree centralization is a measure that variesbetween 0 and 1 with higher values indicating a greater degree of centraliza-tion around a central point or points. Figure 2.2 shows that degree centraliza-tion increased to .45 at the time of the second data collection, seven monthslater, indicating a much greater influence of leaders on followers.Simply put, between time 1 and time 2, Kapferer’s data show that thefactory workers were ‘more linked into a common set of interactional relation-ships’ (1972: 180). Relative to time 1, ties at time 2 tended to cross-cut theUnderstanding Social Network Research15Instrumental relationships at time 1 in an African factory different clusters in the factory, tended to be multiplex, and tended to show thesenior workers in the factory exerting a greater degree of power and influence.This greater solidarity among the different factions allowed the factory work-ers to take the decisive action to go on strike in February 1965 in support of aclaim for a £1 increase in wages. Kapferer’s work is exemplary in its combination of network data andethnographic detail. In particular, Kapferer places the interactions within thefactory in a richly-observed context of recreational activities, kinship, marriageand local politics. He interprets the meaning of his quantitative data matricesthrough his detailed knowledge of each specific person in the matrix, and theiractivities. For example, one central person in Kapferer’s analysis is Lyashi, atailor, who attempted to achieve a position of power and influence in the infor-mal network of relationships. Figures 2.1 and 2.2 show that Lyashi succeededSocial Networks and Organizations16FIGURE 2.2WilliamChristianKalundweZakeyoMeshakDonaldMpunduSignChipataNkumbulaSeamsKamwefuChilwaKalambaEnochKalongaHastingsZuluMubangaHenryIbrahimJosephChisokoneAbrahamLyashiLwangaNkoloyaJohnChobeMabangeAngelBenNyrendaMukubwaChilufyaInstrumental relationships at time 2 in an African factory over the seven-month period in moving from a relatively peripheral positionin the instrumental network to the most central position. The book presentsmany details concerning Lyashi and his daily life, including such relevantextra-curricular information as this: ‘Although the Lumpa Church [of which hewas once a Deacon] is banned, … he maintains a vast network of ties withother “erstwhile” members of the movement’ (Kapferer, 1972: 214).Around the time Kapferer’s book was published, articles were beginning toappear in organizational journals containing analyses of communication flows.One of the first of these was a description of communication in a research anddevelopment laboratory showing diagrams of social contacts and some simplestatistics (Allen and Cohen, 1969). The focus in this article and similar articlesby others (e.g., Pettigrew, 1972) was on the role of sociometric stars and gate-keepers in brokering information, a focus that has been rediscovered by therecent literature on structural holes (e.g., Burt, 1992).Surprisingly, these interesting and innovative analyses by Kapferer andothers made no use of the technical developments in graph theory applied toorganizational settings decades earlier. George Homans (1950) had illustratedthe usefulness of rearranging the rows and columns of data matrices to revealunderlying structure. Systematic applications of matrix algebra to sociometricdata had been described in the social science literature (e.g., Festinger, 1949;Forsyth and Katz, 1946). One of the earliest applications of these new matrixtechniques to an organizational data set was published in the AmericanSociological Review(Weiss and Jacobson, 1955). The authors collected data from196 members of a government agency in interviews lasting from one to threehours. The sociometric questions related to the workflow network – that is, thepeople each individual had worked with over the past few months. As part ofthe structural analysis, the authors reordered rows and columns to produceseparate blocks of highly-connected workers. The final analysis allowed theidentification of work groups, liaison persons between groups, and people withno work contacts – the isolates. The authors mention the possibility of relatingstructural indicators such as individual centrality in the network to outcomevariables such as organizational identification, but no data are reported. Thewholesale application of the new social network methods to organizationaldata had to await the availability of relatively cheap computing power. DISTINCTIVE FEATURES OF THE SOCIALNETWORK APPROACHOne of the attractive features of the social network approach to organizationsis the potential to analyse network relations with an ever-expanding rangeof algorithms, programs, and procedures that map closely on to importantUnderstanding Social Network Research17 orienting concepts and characteristics of networks. Comprehensive reviews ofnetwork methods are available (e.g., Wasserman and Faust, 1994), as are intro-ductory handbooks (e.g., Degenne and Forse, 1999; Schensul et al., 1999; Scott,2000). In this section, we offer examples to illustrate commonly used networkmethods and we discuss distinctive features of the social network approach.For the sake of simplicity we will assume that we want to gather informa-tion on friendship ties between individuals in a small organization of 33 people.How do we do this? There are at least three ways of proceeding. First, we couldcollect whole networkdata using a roster of the names of all 33 people in theorganization. We could list the names on a sheet of paper with instructions tothe respondent to tick the names of those individuals whom the respondentconsiders to be his or her personal friends. From these data we could then pre-pare a ‘whole’ network of relations that indicated for each pair of individualswhether one or both of the individuals considered the other to be a friend. Thedata could be arranged in a 33 by 33 asymmetric matrix.But what if we were unable to gain access to all 33 people? An alternativewould be to collect egocentric data from each person available to us. Thiswould entail a significant sacrifice in data quality. Each individual that agreedto participate could be prompted to give us the names of his or her friends inthe organization. This prompting could take the form of a complete roster of33 names. Or we could prompt the individual to remember by providing cuessuch as different roles (‘Do you have any friends who are managers?’). Once therespondent provided a list of names, the respondent could then be asked toindicate the relations between the friends: Which of the respondent’s friendswere friends of each other? From these data, each respondent’s position in theegocentric network could be estimated. This approach seems particularly suit-able for very large organizations where it is impossible to gain data from allorganizational members.A third approach bypasses the individual members of the organizationcompletely and relies on archival records. Personnel records, for example, oftencontain a wealth of information concerning whom job applicants know in theorganization, who is kin with whom, who recommended who for employ-ment, and so on (see Burt and Ronchi, 1990, for a brilliant example of thisstrategy). Records of relationships such as friendship and kinship are collectedfor a range of different purposes and often form the basis of pioneering workon social networks. See, for example, Uzzi’s (1996) work on the garment manu-facturing industry, utilizing records collected by the Ladies Garment Workers’Union, and Padgett and Ansell’s (1993) work on the Medici family, utilizingrecords maintained for hundreds of years. Archival records are particularly use-ful in cases where it may be dangerous to approach respondents, or whererespondents are unlikely to respond to questionnaires.Each approach, therefore, has its uses. What can be done with these dataonce they are arranged in matrices? In the case of whole network data, col-lected by the use of roster-type questionnaires or through archival records, anSocial Networks and Organizations18 almost unlimited range of analytical techniques can be employed. The centralityof each actor in the network can be analysed on several indices (e.g., Brass,1984) including indegree (i.e., how popular the actor is), betweenness (i.e., theextent to which the actor functions as a go-between for others not directly con-nected), and eigenvector (i.e., the extent which the actor is connected toothers who are highly central). The network can be analysed to see how manyand what kinds of clique exist, whether these cliques overlap, and the extentto which each dyadic pair in the network belongs to the same cliques (e.g.,Krackhardt, 1999). The network can be analysed into blocks of actors similaron the basis of their ties to other actors. And of course, the whole network itselfcan be correlated with another matrix of information about these actors suchas a matrix of correlations showing how similar each pair of actors is withrespect to attitudes or behaviours (e.g., Kilduff, 1992).At its best, network research has several distinctive features that differenti-ate it from traditional approaches in the social sciences: (1) Network researchfocuses on relations and the patterns of relations rather than on attributes ofactors; (2) Network research is amenable to multiple levels of analysis, and canthus provide micro–macro linkages; (3) Network research can integrate quanti-tative, qualitative and graphical data, allowing more thorough and in-depthanalysis. None of these features is well established in traditional approaches inthe social sciences.Relations and Patterns of RelationsThe network approach can test whether the pattern of network ties in a parti-cular social world is related to other important patterns such as the pattern ofdecision-making. Let’s look at one simplified example from the research litera-ture. The research question was whether individuals, in making importantdecisions, tended to be influenced by their friends. The author collected datafrom 170 MBA students at Cornell University by asking them to look carefullydown a list of their classmates and check off the names of those they consid-ered to be personal friends. From these data a square matrix (known as an adja-cency matrix) was constructed showing for each pair of people in the samplewhether one considered the other to be a friend. One row in the matrix showedfor an individual all of those he or she had chosen as friends. A section of sucha matrix is illustrated in Figure 2.3. In the first row we can see that Dana hasreported that she is friends with Bill and Cy. If we look at the rows for thesetwo, we see that Bill fails to reciprocate Dana’s friendship nomination, but thatCy does reciprocate. Thus, the matrix contains asymmetricdata.Each individual in the sample signed up for one or more interviews withthe 120 companies recruiting at the school. The number of interviews eachperson could sign up for was restricted by a points system. All of the sign-upinformation was available for public inspection on bulletin boards, and wasUnderstanding Social Network Research19 also archived on the computer. Thus it was possible to create another matrix(known as an incidence matrix) showing for all 120 companies which oneseach individual had signed up with. The matrix had 170 rows (one for eachperson) and 120 columns (one for each company). The methodological ques-tion was: How could we compare each person with every other person in termsof how similar their decision-making was? The answer was to correlate eachrow with every other row so that for any two people in the sample we knewprecisely how similar their decision-making was in this one important arena oforganizational choice.Figure 2.4 illustrates the similarity matrix that resulted from this proce-dure. For each pair of people in the sample there is a correlation expressinghow similar their choices were across the 120 recruiting companies. Note thatthis matrix is necessarily symmetric. Once this matrix was created it was possi-ble to ask whether individuals’ patterns of behaviours were similar to those oftheir friends. In the research article this question was answered in the affirma-tive through the use of a non-parametric regression analysis that controlled foralternative explanations, and took into account the non-independence of thedata (Kilduff, 1990). The basis of this analysis, however, derives from whetherthe ‘1s’ in Figure 2.3 are in the same cells as the high correlations in Figure 2.4.Social Networks and Organizations20FIGURE 2.3BillRedEdSueJoeRed11Sue11Bill1DanaJuneEd1111Cy11CyJune11Dana11Adjacency matrix showing who is friends with whom Much network research explores social structures, defined as ‘patterns ofconnectivity and cleavage within social systems’ (Wellman, 1988b: 26). Socialstructures are abstract representations of patterns of relationships betweenactors (Nadel, 1957: 12). Studying social structures helps us to understand theways in which groups of actors cluster together in social space (e.g., Burt, 1978).Emergent structures can be compared with other structural depictions of thesame actors to determine, for example, the degree of overlap between observedstructure and a structure derived from theory (e.g., Barley, 1990). Networkresearch can analyse both the whole system of relations and parts of the systemsimultaneously: ‘Analysts are therefore able to trace lateral and vertical flows ofinformation, identify sources and targets and detect structural constraints oper-ating on flows of resources’ (Wellman, 1988b: 26). This ability to capture thestructure of the whole interacting system and its constituent parts in oneanalysis makes the social network approach particularly attractive to studentsof organizations.Let’s look at a specific example of how the network approach reveals socialstructure. As the examples in Figures 2.3 and 2.4 illustrate, individuals con-nected to the same organizations are, in a sense, connected to each otherthrough those organizations. In the same way, organizations are connected toeach other through the people they attract as members. Thus, social structureUnderstanding Social Network Research21FIGURE 2.4DanaBillEdJuneCyRedSueJoeBill.49.10-.05.03.07.09.37Ed.13.10.56-.07.33.25.06.01June-.05.56.22.10-.08.15.22Cy.03-.07.22.09.04.12Red-.04.07.33.10.09-.09.19Sue-.06.09.25-.08.04-.09.44Dana.49.13.01.22-.04-.06.09Similarity matrix showing the correlation of each pair’s interview selections involving people and organizations has a dual quality: people are connected toeach other through organizations and organizations are connected to eachother through people (Breiger, 1974). To illustrate this point we turn to a cele-brated data set collected by Galaskiewicz (1985) showing the affiliations of 26Minneapolis area CEOs to 15 clubs and corporate boards. Figure 2.5 models both the CEOs’ and their organizational affiliations in thesame analytical space using a technique called correspondence analysis thatuses an objective criterion to display optimally the correlations among two setsof entities. (For more information on this type of analysis and its application tothese particular data, see Wasserman and Faust, 1994: 334–42.) The display inFigure 2.5 shows what appears to be a core set of CEOs who meet each other ata core set of clubs and boards, forming an elite structure, with other CEOs dis-persed around the periphery. The core people and clubs (contained within theheart-shaped dotted line) are clustered around members such as R4 and R14 andSocial Networks and Organizations22FIGURE 2.5-2.54-2.05-1.56-1.07-0.57-0.080.410.901.391.892.38-2.54-1.56-0.570.411.392.38R1R2+R3+R4+R5R6R7+R8+R9+R10+R11+R12++R13R14R15R16+R17+R18R19+R20R21+R22+R23R24+R25R26+C1+C2+C3C4+C5+C6+C7+C8+C9+C10+C11C12+C13C14+C15++++++++++++Rs indicateCEOsCs indicatethe clubs and boards to whichCEOs belongThe social structure of CEOs and their clubs they attend clubs such as C3 and C15. We could perform many other structuralanalyses on these data to check how the CEOs cluster together, and whether thedominance structure suggested by Figure 2.5 was supported.Micro–Macro LinkagesThis example of social structural analysis illustrates how the social networkapproach helps us understand micro–macro linkages in organizations. A hypo-thetical example of such linkages is illustrated in Figure 2.6 that borrows fromthe so-called bathtub model developed by James Coleman (1990: 8). In thefigure, the overall social network of relationships involving CEOs and theirclubs is hypothesized to influence the individual connections that CEOs make(link 1),and these individual connections are predicted to affect the actionsthat individual CEOs take (link 2). These actions, in turn, may contribute to thedominance by an elite group of CEOs over the distribution of resources withinthe community (link 3). Also shown is the direct link between the whole net-work of CEOs and elite dominance across the top of the figure. Thus, the net-work approach can help us understand the ways in which individuals affectinstitutional outcomes and how larger social structures affect individuals. (SeeHuber, 1991, for a general treatment of this issue.) Note also that networkanalysis helps delineate such structural features of organizational contexts asthe density of social ties (Alba, 1982: 40). Structural features such as densityderive from interactions among individuals, and these emergent structures canbe used to interpret individual behaviours.The social structures that emerge from network analyses constitute socialrealities of which the social actors themselves may not be aware (Galaskiewicz,1996: 21). Social structure is often not obvious because it involves a complexmeshing of different types of network ties that may span across different levelsof analysis and may have accumulated over many years. For example, a net-work analysis of the social relationships in one conflict-ridden factory revealedpatterns of kinship and dependence formed over the 30-year life of the factorythat were unknown to many of the key actors involved (Burt and Ronchi,1990). Social network research has an emancipatory potentialin that it caninform actors of non-obvious constraints and opportunities inherent in patternsof social connections. Part of the increasing popular interest in the results ofsocial network research derives from this potential. An Integration of Quantitative, Qualitativeand Graphical DataAnother distinctive aspect of the social network approach that we haveexploited in this chapter is its ability to supplement quantitative analysis withqualitative and graphical data. Traditional social science research tends to focusUnderstanding Social Network Research23 on mean differences between groups. Thus, data analysis proceeds at a highlevel of abstraction. But social network research enables the researcher to stayclose to the data. For example, an ethnographic study of the conversationsbetween radiologists and technologists in two hospitals was supplementedwith extensive analyses and representations of the social ties between occu-pants of different roles (Barley, 1990). Similarly, a quantitative analysis of themarginality of underrepresented groups was visually reinforced by a networkdepiction of friendship ties between individuals (Mehra et al., 1998). In boththese examples, the network pictures added a degree of realism largely lackingin the regression tables of the typical journal article.An excellent example of the power of a network sociogram to supplementquantitative analysis is provided in Figure 2.7 that shows interactions among 14participants and four instructors at a National Science Foundation summer campin 1992 (see Borgatti et al., 1999, for more details). The research question is: Towhat extent do people’s interactions exhibit homophily, specifically a tendency tointeract with similar others such as members of their own sex (see McPhersonet al., 2001, for a review)? This question can be addressed quantitatively, but thediagram also provides clear evidence. The sexes tend to clump together in recog-nizable groups, with the only exception being the female Brazey who hasattached herself to the group that includes the four instructors (Steve, Bert, Russand Gery). Note also how the diagram clearly illustrates the strategic importanceof cutpoints: actors (like John and Holly) who constitute the only links betweendifferent groups. (See Brass, 1985, for a similar example of gender homophily.) In yet another example of how network diagrams serve analytical purposes,the main burden of proof in one research article was carried by a series of depictionsSocial Networks and Organizations24FIGURE 2.61 3 2 Individual CEOconnections Individual CEOactions Network of CEOs Elite control of resources Bathtub model of micro–macro linkages of the various ways in which individuals misperceived their own friendshipnetworks in organizations (Kumbasar et al., 1994). The network approachenables the analyst to retain the richness of the data rather than havingto sacri-fice richness for statistical power.MAJOR CONCEPTS IN SOCIALNETWORK RESEARCHOrienting ConceptsAs we have alluded to in the discussion up to this point, the social networkapproach to organizations is premised on the importance of several conceptsUnderstanding Social Network Research25FIGURE 2.7BillHarryDonMichaelAn arrow from A to Bindicates that A has chosenB as one of his or her top threeinteraction partnersHollyPatJenniePamAnnCarolPaulineJohnGeryRussSteveBertLeeBrazeyInteractions among attendees at a cultural anthropology summer camp that include embeddedness, social capital, structural holes and centrality.These concepts orient the researcher towards specific aspects of organizationalphenomena that might otherwise be overlooked.According to the embeddednessargument, work-related transactions tend tooverlap with patterns of social relations (Granovetter, 1985). Thus, business isembedded in social networks, and patterns of transactions within and betweenfirms may depart from what might be expected from a pure economic per-spective. People may prefer to do business with contractors and others withwhom they have ties of friendship or kinship rather than find exchange part-ners in the open market (Uzzi, 1996). One example from our own research (Tsaiand Kilduff, 2002) is displayed in Figure 2.8. The figure shows how importantknowledge, such as technological advances, were communicated among the 36business units in a multi-billion dollar food company. What is striking is theextent to which the 14 business units run by members of the family thatfounded the company tended to be the central players in this knowledge trans-fer network. The family-run units tend to cluster in the central area of theknowledge transfer network. These units tended to favour each other with newknowledge, and also tended to receive new knowledge from business units runby non-family members. Knowledge transfer was, in fact, embedded in kinshiprelationships rather than following purely economic logic.Some organizations may suffer from a ‘liability of unconnectedness’(Powell et al., 1996) in the sense that organizational members fail to developstrong bonds of trust to important actors inside and outside the organization.Top management may even punish those who create links across organiza-tional boundaries to potential competitors and other industry players.Organizations such as Digital Equipment Company were notorious fortress-likecultures in which ‘the internal mattered so much’ (Johnson, 1996). At theother extreme, a dynamic social network of interacting individuals may span awhole geographical area threatening the traditional hegemony of organiza-tional boundaries. As one executive of a Silicon Valley company commented:‘There’s far greater loyalty to one’s craft than to one’s company. … A companyis just a vehicle which allows you to work’ (Saxenian, 1990: 97). Resource flowswithin organizations are likely to depart from what a purely economic modelwould predict, according to the embeddedness argument. People are likely tofavour their family and friends with timely information, recommendations,interesting projects and other career-building opportunities. The path toadvancement may be embedded in social relationships. Just having a contactin an organization can enhance the chances of being offered a job (Fernandezand Weinberg, 1997) and can significantly increase salary negotiation outcomes(Seidel et al., 2000).This emphasis on the importance of social relationships is summarized inthe concept of social capital. This concept can be defined, at the individualactor level, as the potential resources inherent in an actor’s set of social ties. Inone of the first uses of the term in the network literature, social capital wasSocial Networks and Organizations26 13361523363511333426223230162919201827242517102128215497831Family unitNon-family unitdescribed as ‘personal investments’ that could be used for economic advantageby the activation of ‘particular links in a social network’ (Mitchell, 1974: 286).Used in this sense of a personal investment, social capital can be traded forother types of capital such as money or cultural capital (Bourdieu, 1980).Understanding Social Network Research27The embeddedness of knowledge transfer relations in the kinship networkin a multi-unit companyFIGURE 2.8 Personal connections can be useful in facilitating access to jobs (Granovetter,1974) and promotions (Brass, 1984; Burt, 1992). The social network approachassumes that different configurations of social ties produce different benefitsfor actors (Burt, 2000). Social capital is often described as different from moneyand other types of capital in that it inheres in the relationships betweenpeople. Actors do not control their social capital in the same way they controltheir money or their human capital. To use social capital, it is necessary to drawupon the cooperation of another actor by, for example, asking for advice orhelp at work.Social capital can also be defined as the benefits that accrue to the collec-tivity as a result of the maintenance of positive relations between differentgroups, organizational units or hierarchical levels (e.g., Burt and Ronchi, 1990;Tsai 2000; Tsai and Ghoshal, 1998). One of the unexplored aspects of socialcapital concerns how the individual use of personal connections to