This article was downloaded by: [128.32.74.70] On: 24 August 2017, At: 14:22Publisher: Institute for Operations Research and the Management Sciences (INFORMS)INFORMS is located in Maryland, USA

Organization Science

Publication details, including instructions for authors and subscription information:http://pubsonline.informs.org

An Intraorganizational Ecology of Individual AttainmentChristopher C. Liu, Sameer B. Srivastava, Toby E. Stuart

To cite this article:Christopher C. Liu, Sameer B. Srivastava, Toby E. Stuart (2016) An Intraorganizational Ecology of Individual Attainment.Organization Science 27(1):90-105. https://doi.org/10.1287/orsc.2015.1020

Full terms and conditions of use: http://pubsonline.informs.org/page/terms-and-conditions

This article may be used only for the purposes of research, teaching, and/or private study. Commercial useor systematic downloading (by robots or other automatic processes) is prohibited without explicit Publisherapproval, unless otherwise noted. For more information, contact permissions@informs.org.

The Publisher does not warrant or guarantee the article’s accuracy, completeness, merchantability, fitnessfor a particular purpose, or non-infringement. Descriptions of, or references to, products or publications, orinclusion of an advertisement in this article, neither constitutes nor implies a guarantee, endorsement, orsupport of claims made of that product, publication, or service.

Copyright © 2016, INFORMS

Please scroll down for article—it is on subsequent pages

INFORMS is the largest professional society in the world for professionals in the fields of operations research, managementscience, and analytics.For more information on INFORMS, its publications, membership, or meetings visit http://www.informs.org

OrganizationScienceVol. 27, No. 1, January–February 2016, pp. 90–105ISSN 1047-7039 (print) � ISSN 1526-5455 (online) http://dx.doi.org/10.1287/orsc.2015.1020

© 2016 INFORMS

An Intraorganizational Ecology of Individual Attainment

Christopher C. LiuRotman School of Management-University of Toronto, Toronto, Ontario M5S 3E6 Canada, chris.liu@rotman.utoronto.ca

Sameer B. Srivastava, Toby E. StuartHaas School of Business, University of California, Berkeley, Berkeley, California 94720

{srivastava@haas.berkeley.edu, tstuart@haas.berkeley.edu}

This paper extends niche theory to develop an intraorganizational conceptualization of the niche that is grounded inthe activities of organizational members. We construe niches as positions in a mapping of individuals to formal and

informal activities within organizations. We posit that positional characteristics in this activity-based system are criticaldeterminants of members’ access to information and relationships—two of the vital resources for advancement in organi-zations. Because activities are difficult to observe, we propose a novel empirical strategy to depict niches: we exploit acensus of memberships in electronic mailing lists. We assess three niche dimensions—competitive crowding, status, anddiversity—and show that these attributes affect the allocation of rewards to employees. Propositions are tested in twoempirical settings: an information services firm and the R&D division of a biopharmaceutical company. Results indicatethat people in competitively crowded niches had lower levels of attainment, whereas those in high status and diverse nichesenjoyed higher attainment levels. We conclude with a discussion of email distribution lists as a tool for organizationalresearch.

Keywords : ecology; niche; crowding; status; diversity; formal, semiformal, and informal organization; attainment;distribution lists

History : Published online in Articles in Advance December 18, 2015.

1. IntroductionCareers in organizations often are described with allu-sions to ladders, tracks, and ceilings. These analogiesreflect a widely documented fact in organizational life:wittingly or not, organizations structure careers and areagents of stratification. For instance, specific job titleswithin organizations often are completely or nearly seg-regated by sex, with many implications for individu-als’ compensation and career prospects (Baron 1984,Barnett et al. 2000). Likewise, job ladders may reach todifferent heights based on the sociodemographic back-grounds of their climbers (Kanter 1977), and informalinteractions in organizations may be particularly subjectto exclusionary processes (e.g., Turco 2010, Kleinbaumet al. 2013). Indeed, the stratification of opportunitiesmay begin well before any actual employment rela-tionship is underway: exclusionary hiring practices andimplicit restrictions on access to the social networks overwhich recruitment takes place differentially sort indi-viduals with certain, non-merit-based characteristics intospecific job vacancies (Fernandez and Friedrich 2011,Fernandez and Fernandez-Mateo 2006).

This paper extends the literature on social structurewithin organizations and its effects on individual attain-ment. However, the theoretical lens we ecological theo-ries of the niche (Freeman and Hannan 1983, McPherson1983, Hannan et al. 2003, Popielarz and Neal 2007)to study how different structural features of realized

“niches” inside organizations contour the rewards thatemployees garner. Although organizational ecologistshave a decades-long interest in the intersections of eco-logical reasoning and labor market phenomena (e.g.,Haveman and Cohen 1994, Sørensen 1999, Baron 2004),little of this work elaborates implications of ecologicalreasoning within organizations. Here, we consider howniche properties can help us to understand the variationin discretionary rewards managers allocate to employees.

Theories of the niche have been influential in the liter-ature on interorganizational dynamics. Although a recentformalization of niche theory has highlighted some dis-crepancies in logics across branches of the literature(Hannan et al. 2003), the idea of a niche as a posi-tion in a multidimensional resource space has animateda substantial body of empirical work in organizationalstudies. This research shows that population dynamics,including organizational births (Hannan and Freeman1987), growth and mortality rates (Barron et al. 1994),resource partitioning processes (Carroll 1985, Dobrevet al. 2001), and status differences (Podolny et al. 1996),depend on multiple aspects of the resource spaces thathost organizational populations. Moreover, niche the-ory has been extended to a range of social phenom-ena that broadly can be framed in terms of markets,including occupations vying for professional jurisdic-tion (Abbott 1988), the emergence of forms in an insti-tutional identity space (Ruef 2000), and even musicaltastes (Mark 2003).

90

Dow

nloa

ded

from

info

rms.

org

by [

128.

32.7

4.70

] on

24

Aug

ust 2

017,

at 1

4:22

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Liu, Srivastava, and Stuart: An Intraorganizational Ecology of Individual AttainmentOrganization Science 27(1), pp. 90–105, © 2016 INFORMS 91

In this paper, we bring niche theory to intraorgan-izational analysis. We find many points of correspon-dence between inter and intraorganizational ecologies.At the broadest level, just as the organizations in apopulation experience competitive and symbiotic inter-actions in a confined resource space, employees in anintraorganizational ecology compete and cooperate toobtain scarce resources. In population ecology, orga-nizations vie for customers, employees, financial capi-tal, and legitimacy. Analogously, inside an organization,individuals compete to obtain information, social capi-tal, budget allocations, advancement opportunities, andso on. Likewise, just as the finiteness of resources avail-able to support any given organizational form creates acarrying capacity that shapes vital rates, resource con-straints force many tradeoffs among employees in orga-nizations. Finally, one can draw close parallels betweenintraorganizational processes and recent reformulationsof the niche in terms of form-defining identity codes andthe lenses of audience engagement and appeal (Hannanet al. 2003).

Although niche theory has not been extensively ap-plied to intraorganizational dynamics, we believe thislacuna in the literature has been caused by data(un)availability, rather than the inapplicability of thetheory to pertinent phenomena. The operative question,therefore, is how can the researcher observe and mea-sure employees’ niches in an intraorganizational ecol-ogy? We believe that any useful approach to measuringintraorganizational niches will need to consider individ-uals’ positions in both the formal and informal structureof an organization. There is simply too much theoryand evidence that informal structure matters to solelyrely on the formal structure as the empirical scaffoldto the niche space (Blau and Scott 1962, Allen 1977,Krackhardt and Hanson 1993; most recently, Biancaniet al. 2014, McEvily et al. 2014).

In light of this, one reasonable approach would be toconsider the resource space defined by the multitude ofrecurring “activities” (broadly considered) in organiza-tions. A focus on activities appeals to us because the setof them bridges the formal, semiformal, and informalstructures or organizations. Some activities (e.g., depart-mental meetings) parallel the formal reporting struc-ture; however, many others assemble the organization’sinformal social and interest groups, or members of itsmyriad project teams. In addition to spanning the contin-uum from formal-to-informal modes of organization, anemphasis on activities is consistent with classic defini-tions of niche. For instance, Elton (1927, p. 63) defined“niche” as, “a term to describe the status of an animalin its community, to indicate what it is doing 0 0 0 0” Sim-ilarly, in our framework, intraorganizational niches arisefrom activities that reveal what employees are doing andwith whom they are interacting while performing thoseactivities.

Specifically, we construct niches from a dual-modenetwork that maps individuals to the activities in whichthey participate. Much like Feld’s (1981) observationthat the intersection of people in common interests givesrise to clusters of interaction, we witness the (presumed)networks and information flows that occur when groupsof individuals participate in the same activities. Con-ceptually, this activity-focused affiliation network encap-sulates multiple dimensions of an organization’s socialstructure, but pragmatically, it is difficult to observe. Thesolution we implement relies on a novel data source thathas potentially broad application in organizational anal-ysis: the census of electronic mailing lists in an organi-zation. If we construe each mailing list as a membershiproster for a distinct “activity,” the full set of mailing listsis a dual-mode network with disjoint sets of elements:employees and activities. In the two organizations westudy, email distribution lists provide an extraordinar-ily detailed window into the complex ways in whichwork actually gets done. We find them for everythingfrom office locations to function memberships to stand-ing cross-functional teams, to ad hoc task forces, the“kitchen cabinets” of organizational leaders, professionalinterest groups, as well as social groups, such as thesoftball league, or employees in common ethnic groups.It is in this sense that the activities revealed by mailinglists run the gamut from formal organization structures,to semiformal team structures, to informal social groups.

In the analysis to follow, we demonstrate the utilityof the framework and the data source. First, we measurethree properties of niches—competitive crowding, status,and niche diversity—and show that these characteristicsaffect employee attainment in the directions theory sug-gests. Second, we construct intraorganizational ecologiesin two very different organizations: a private-sector bio-pharmaceutical laboratory and a large information ser-vices company. The findings are remarkably consistentacross the two settings, which bolsters the externalvalidity of the results. Third, we test our hypothesesexploiting two complementary measures of individualattainment: annual bonus and performance rating. Wefind strong concordance in the results between thesetwo outcome variables. Taken together, the findings areconsistent with the proposed conceptualization and mea-surement of intraorganizational niches.

2. Theory: Niches in anIntraorganizational Ecology

Ecological theories begin with a distinction betweenthe fundamental and the realized niche. Drawing onHutchinson (1957), Hannan and Freeman (1989) definethe fundamental niche of an organizational population asthe region of a resource space in which the populationwill experience a nonnegative rate of growth. The real-ized niche of an organizational population is the subset

Dow

nloa

ded

from

info

rms.

org

by [

128.

32.7

4.70

] on

24

Aug

ust 2

017,

at 1

4:22

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Liu, Srivastava, and Stuart: An Intraorganizational Ecology of Individual Attainment92 Organization Science 27(1), pp. 90–105, © 2016 INFORMS

of the fundamental niche that is actually occupied, giventhe presence of competitive interactions with rivalrouspopulations.

Much of the empirical work on ecological dynamicshas examined the structure of realized niches. We buildon this vein of the literature. Two conceptual aspects ofthis work guide our development of a theory of intraor-ganizational niches. First, as in classic theories of socialstructure (Simmel 1902), ours rests on an analytical dis-tinction between actor and position—positions or nichescan be characterized in general terms; they are inde-pendent of any specific occupant. Put differently, socialstructure is an abstraction based on ongoing, recurrentrelationships among actors (cf. Breiger 1974, Breigerand Mohr 2004). Of course, this assumption underliesecological analyses of all sorts. For instance, the notionthat niche width moderates the strength of audienceappeal is a proposed relationship between an abstraction(niche width) derived from positional characteristics (thespan of an organization across segments of a market)and an outcome of interest (audience appeal).

Second, implicit in the concept of a realized nicheis the relational nature of its definition: realized nichesare elaborated through ongoing interactions among theinhabitants of a resource space. Therefore, intraorgani-zational niches arise from a process akin to endogenouspopulation structuring (Hannan and Freeman 1977).Like Carroll’s (1985) resource partitioning theory or asin models of size-localized competition (e.g., Baum andMezias 1992), we argue that positions in the intraorga-nizational niche space, in part, emerge from endogenousinteractions among employees inside organizations.1

Positions are endogenous to the day-to-day jockeyingthat is routine in organizational life.

Many studies have linked organizations’ niches totheir life chances. McPherson’s (1983) development ofa competition matrix for a group of organizations rep-resents a seminal contribution to this area. McPherson(1983) defined niches of voluntary organizations interms of their locations in a resource space compris-ing the population distribution of possible joiners in theranges of sociodemographic variables targeted by theorganizations under study (see also, Baum and Singh1994, Popielarz and McPherson 1995). Much of thefollow-on work in the ecology-of-affiliation tradition,however, implicitly has examined niches in some formof an activity space. For instance, Podolny et al. (1996)demarcated technological niches of organizations basedon the overall patent citation network in the semiconduc-tor industry. This creates an affiliation network definedby firms’ choices to become active in some technicalareas, but not others. The properties of organization-specific niches are then derived from the relationalstructure of the affiliation network: two firms competeinsofar as their participation in technical activities over-lap. Therefore, the network is defined by the mapping of

companies to fine-grained areas of technology. A sim-ilar conception of niche is presented in Dobrev et al.(2001). In their papere, technological niches are con-structed from the range of engine sizes that automo-bile manufacturers choose to produce (see also, Dobrevet al. 2002).

In this paper, we distill properties of intraorganiza-tional niches in the recurrent activities of the organiza-tion. We focus on employee niches in a broadly definedactivities space because activities are the stage on whichcollaborative and competitive interactions translate intoheterogeneous career outcomes. Put simply, activitiesare the “what and where” of the allocation of individ-uals’ time in organizations. If this is indeed the case,then how—and with whom—employees are embed-ded in the activities that occur in an organization willstrongly influence individuals’ access to two criticalresources that are known to affect career outcomes.These resources are information and relationships.

2.1. The Resource Space: Information andRelationships

Niches are positions in a resource space. Consider, forinstance, resource partitioning theories. In these, often-geometric conceptions of the niche, organizations aremetaphorically assumed to be presences (i.e., shapes) in aEuclidean space (e.g., Carroll 1985, Peli and Nooteboom1999). In empirical studies, of course, the specificationof the resource space depends on the nexus of the pop-ulation under study and the availability of data. Forinstance, Baum and Singh (1994) defined niches of daycare centers in terms of the age ranges of the childrenthey admit, because this defines the group of would-bemembers for whom centers compete. What, then, arethe pertinent resources in an intraorganizational ecol-ogy? We posit that two, related categories of resourcesare vital to employees as they maneuver to advance inorganizations. First, individuals may gain an advantageif they cultivate unique positions in the flow of informa-tion in an organization. Second, advancement hinges onsocial capital: time and again, studies have shown thatpossessing the right relationships and sponsors is criticalto rapid career progression.

A great deal of work, which spans a dozen or moresubfields in organization theory, considers the role ofinformation in social organization. The diversity of thisresearch extends from information processing theoriesof organizations themselves (Weick 1979) to typologiesof coordination when information is distributed acrossindividuals or units in organizations (Thompson 1967)to the challenges of moving information across organi-zational boundaries (Allen 1977). One of the streams ofthis work also considers the role of information in theadvancement of individuals’ careers in organizations.

In fact, ideas about information-based advantages incareer outcomes connect to the voluminous literature

Dow

nloa

ded

from

info

rms.

org

by [

128.

32.7

4.70

] on

24

Aug

ust 2

017,

at 1

4:22

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Liu, Srivastava, and Stuart: An Intraorganizational Ecology of Individual AttainmentOrganization Science 27(1), pp. 90–105, © 2016 INFORMS 93

on social capital. Most notably, Ronald Burt’s (1992)work develops the idea that actors in brokerage networksare ideally situated to gain access to information thatprovides them with an advantage in competitive situa-tions. An “information advantage” may exist for individ-uals in network structures that convey prompt or broadaccess to strategic information. This is one social mecha-nism thought to underlie the many empirical associationsbetween egocentric networks and career outcomes.

Of course, the fact that certain networks facilitatebeing “in the know” is only one potential benefit ofa deep social network. Decades of research shows thathaving the right relationships is necessary for careeradvancement: mentoring, task advice, the developmentof a coherent identity, buy-in from leaders, and polit-ical support all are network-dependent resources (e.g.,Kanter 1977, Ibarra 1995, Podolny and Baron 1997,Srivastava 2015a). In fact, many studies in the job searchliterature establish the value of the right network beforethe employment relationship even begins: often, the rightcontacts are vital to securing jobs, in the first instance.

Abstracting from the many nuances in these liter-atures, we argue that a distillation of the propertiesof employee niches in the activities of an organiza-tion amounts to identifying individuals’ positions in aresource space comprising information and relationships.The social networks and the exchange of tacit knowl-edge that occurs in and emerge from these activitiesinform levels of access to these critical resources, andtherefore individual attainment levels. In short, we focuson employee niches in the activities space because webelieve that positions in this space will strongly influ-ence employees’ access to information and ability toform important, career-enhancing alliances.

2.2. HypothesesWe describe intraorganizational niches in three dimen-sions: competitive crowding, status, and diversity. Com-petitive crowding refers to the density of actors whooccupy similar niche positions. The status of a nichevaries with the extent to which it offers access to col-leagues who are in positions of influence in the orga-nizational hierarchy. Finally, a niche’s diversity gaugesthe extent to which it provides access to employees indifferent functional areas and hierarchical levels of theorganization.

Niche Crowding. In the empirical literature on therealized niche, ecologists have gauged variation in com-petitive crowding across niches in terms of “overlap den-sities” (McPherson 1983), or the count of organizationsthat participate in a given niche (Baum and Singh 1994).When overlap density is high, organizational life chanceshave been shown to be compromised. Beginning withDiMaggio (1986), social networks researchers noted theresemblance between measures of structural equivalence

in a network and of ecological niche overlap (Burt 1987,Podolny et al. 1996). This parallel arises because struc-tural equivalence itself is a measure of overlap in a rela-tional structure; two perfectly equivalent elements in anetwork are, by construction, substitutes: one node mayreplace the other without consequence to the network’sshape. Thus, network-based similarity measures are ameans to quantify the competitive intensity of niches.A literature has since evolved in which the intensity ofcompetition between actors is a function of the similaritybetween them in a resource space, such as recruitmentpatterns in a labor market (Sørensen 1999), a supplier-buyer network (Burt 1982), a geographic area (Lomi1995, Baum and Haveman 1997, Sorenson and Audia2000, Audia and Kurkoski 2012), a technology or scien-tific space (Podolny et al. 1996, Stuart and Ding 2006),or many different formulations of a product featuresspace (Dobrev et al. 2001, Reis et al. 2013).

The general idea that competition has consequencesfor individual attainment is also a long-running theme inresearch on careers. Studies of workforce demography,for instance, have found that employees who enter orga-nizations in large cohorts may face greater competitionfor senior-level jobs (Stewman and Konda 1983, Barnettand Miner 1992) and lower rates of mobility. Likewise,crowding may reduce individuals’ ability to realize thepotential gains from opportunities for brokerage (Burt1997) or increase their proclivity to pursue risky careerstrategies (Bothner et al. 2007). Human capital theo-rists have noted that women frequently enter occupationsrequiring general skills that do not atrophy amid tempo-rary exits from the labor market. The result of this pro-cess is a crowding of female workers into occupationswith this feature and a reduction in the wages attachedto such jobs (Bergmann 1986, Barnett et al. 2000).

Competitive crowding thus has been linked to adverseoutcomes in both relational ecologies of affiliation andin labor market settings. Why do we anticipate the samein an assessment of the competitive crowding of posi-tions in the intraorganizational activity space? We arguethat occupants of crowded niches will be less effective inaccessing valuable information or building effective net-works, ceteris paribus. The literature on informationadvantages from social networks hinges on distinctionsof access and timing. Advantage comes from knowingwhat others do not, or at least knowing things beforethe pertinent information diffuses to a broader audi-ence, at which point its ubiquity eliminates its strate-gic value. From the standpoint of information access,crowded positions imply redundancy in exposure. Whatone person knows, so too do a number of structuralequivalents. In a similar vein, relationships are based onmutual benefit, and occupants of crowded niches willhave less distinctive resources to bring to bear in forg-ing new relationships. Put differently, in facing many

Dow

nloa

ded

from

info

rms.

org

by [

128.

32.7

4.70

] on

24

Aug

ust 2

017,

at 1

4:22

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Liu, Srivastava, and Stuart: An Intraorganizational Ecology of Individual Attainment94 Organization Science 27(1), pp. 90–105, © 2016 INFORMS

structural equivalents in the activities of the organiza-tion, individuals in competitively crowded niches willhave fewer opportunities to carve out positions of uniquevalue to the organization. We hypothesize the following:

Hypothesis H1. Employees in competitively crowdedintraorganizational niches will experience lower levelsof attainment than employees in niches that are lesscrowded.

Niche Status. Crowding, therefore, concerns the levelof competitive differentiation between the actors in asocial system. Status, conversely, references their socialstanding in hierarchical orderings. The sociological liter-ature demonstrates many benefits of status. For instance,high status actors garner greater recognition for a givenquality of product (Merton 1968); they obtain the broad-est range of choice among potential partners (Stuart1998); and status facilitates entry into new market seg-ments (Jensen 2003).

One of the defining features of status is that it “leaks”(Podolny 2005). Within an organization, for example,an individual’s, a work group’s, or even a full depart-ment’s reputation is a function of those who associatewith that person or collectivity. In this sense, status isalways rooted in the relationships that embed actorsinto the rank and expertise distribution in an organiza-tion, as these relationships form the tributaries of statusleakage. This occurs because other community membersinfer status from affiliations: a focal actor’s social sta-tus depends on the statuses of his or her affiliates. Asa few among many examples in the literature, academi-cians derive status from their affiliations with particulardepartments and universities (Merton 1968); law firmsaccrue status from the prestige of the universities theirstaff attended (Phillips and Zuckerman 2001); and youngcompanies derive status from their investors and strate-gic alliance partners (Stuart et al. 1999). Likewise, in anintraorganizational ecology, certain niches are high sta-tus because the activities that define them confer privi-leged access to those in positions of power and influencein the organization.

A central conclusion of this literature is that statusbegets access to socially constructed and socially trans-mitted resources. We argue that occupants of high statusniches are more likely to form relationships with highstatus alters because their niche sets the stage for mul-tiple opportunities to interact with organizational elites.Furthermore, once an individual has—or even is justperceived to have—relationships with high status alters,that person is more likely to gain exposure to valuableinformation because of the positive status spillovers theyreceive from these affiliations. We therefore hypothesizethe following:

Hypothesis H2. Employees in niches that conferaccess to high status actors will experience higher lev-els of attainment than employees in niches that do notconfer access to high status actors.

Niche Diversity. Much of the seminal work in thesocial networks literature concerns how the shape ofa network influences the dissemination of information.The arcs of the social network in a community determinethe distribution of information within it. This is particu-larly true for tacit or proprietary information that is noteasily or willingly transferred through public broadcastchannels.

There are a few types of arguments about the poten-tial advantages that accrue to occupants of niches thatprovide access to a diverse array of contacts. First, strate-gic theories of control in social networks describe thepotential gains from intermediating transactions amongdisconnected alters, including the option to broker rela-tionships among one’s contacts. Second, information-based theories of opportunity posit that individuals withbroad contact networks gain exposure to more variedstreams of ideas and information, and therefore theyare ideally positioned to identify untapped opportuni-ties or to experience the creative spark that comesfrom the recombination of synergistic knowledge inputsarriving from alters in a diverse contact network (e.g.,Brass 1995, Burt 2004, Fleming et al. 2007, Tortorielloet al. 2012).

We argue that the breadth of access to diverse con-tacts also distinguishes intraorganizational niches. Someniches limit their occupants to relatively closed networks,in the sense that homogeneous groups of individualsengage in the activities that define the niche. Otherstypes of activities, by contrast, unite a diverse array ofparticipants and therefore serve as gateways to a rangeof information and relationships with others of differentrank, expertise, or social capital profiles. Occupants ofdiverse niches are exposed to and therefore more likelyto form relationships with people from functional areasand vertical levels other than their own. These relation-ships are more likely to serve as conduits to informationand ideas that are unknown in the focal actor’s organiza-tional unit. Likewise, diverse niches are useful for build-ing a focal employee’s reputation across the organizationand for facilitating opportunity recognition beyond theconfines of a specific job role. Because broader nichesprovide exposure to a range of information and relation-ships, we hypothesize the following:

Hypothesis H3. Employees in niches that conferaccess to a diverse range of contacts will experiencehigher levels of attainment than employees in niches thatdo not confer access to a diverse range of contacts.

3. Data and MethodsWe test these three hypotheses in two, quite differentorganizational settings: an information services provider,which we label ISCO, and a biopharmaceutical com-pany, which we call BTCO. At the time of data collec-tion, there were 4,661 employees in the ISCO sample

Dow

nloa

ded

from

info

rms.

org

by [

128.

32.7

4.70

] on

24

Aug

ust 2

017,

at 1

4:22

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Liu, Srivastava, and Stuart: An Intraorganizational Ecology of Individual AttainmentOrganization Science 27(1), pp. 90–105, © 2016 INFORMS 95

Table 1 Comparison of ISCO and BTCO

ISCO BTCO

Industry Information services BiotechnologyFunction All functional areas R&D onlyOrganizational

structureMany hierarchical

layers; promotionsrelatively frequent

Relatively flat;promotionsrelatively infrequentand tend to occurwithin lab units

Educationalbackground

Wide range ofeducationalbackgrounds

Mostly researchscientists; highconcentration ofPh.D.s

Geography Multiple sites acrossthe United States

Single site

Career mobility(horizontal)

Extensive: careerpaths tend tocross functionallines

Limited: scientiststypically advancewithin a given lab

Available dataAttainment Performance rating

onlyPerformance rating

and annual bonusEmail list names Encrypted Identified

and 916 employees in the BTCO sample. These orga-nizations varied in many dimensions, including size,geographic dispersion, workforce demographics, orga-nizational structure, and internal mobility patterns. Byassessing the hypotheses in two, disparate settings, weseek to enhance the external validity of the findings.

The two organizations slightly differed in the data theymade available to us. For example, ISCO provided usaccess to employee performance ratings but not bonuses,whereas BTCO granted access to both ratings and bonusamounts. Table 1 summarizes the differences betweenthe two organizations and the nature of the data we wereable to acquire from each.

The global information services provider, ISCO, wasthe largest business unit of a conglomerate. At the time ofdata collection, ISCO had a workforce of nearly 10,000in over 100 offices and generated $4 billion in revenue.Our study population included all of its nearly 5,000U.S.-based employees. The company’s domestic oper-ations were organized along functional lines includingproduct development, marketing, sales, finance, legal,and human resources. In addition, the company had asmall number of integrated organizational units that com-bined functional resources and were accountable for theprofitability of an entire line of products and associatedservices.

The study population at BTCO included all membersof the nearly 1,000-person Research Division. Theseemployees were located in spatially proximate buildingsin one metropolitan area. This unit of the company con-ducted basic and applied scientific research to supplythe company’s drug development pipeline. Employees inR&D worked across a range of biological disciplines and

methods. The division was modeled after an academicresearch center, with senior scientists directing the workof junior researchers and staff in a decentralized, labo-ratory setting. Given the scientific nature of its purpose,the workforce at BTCO was very highly educated, withmany doctoral degree holders on staff.

From both organizations, we collected email distri-bution lists and human resource records. Each distribu-tion list is a collection of associated email addresses. Inboth organizations, email addresses were encrypted topreserve employee privacy. In addition, ISCO encrypteddistribution list names before sharing the data with us.Therefore, we could identify the names of all email dis-tribution lists at BTCO, but not at ISCO. We collectedemail lists from both organizations at frequent intervalsfor several months. Because there was not a great dealof variation in the list composition or members overthis short window of time, we chose to analyze the dataas a cross-section.2 In both organizations, distributionlists served to facilitate communication among groups ofemployees who have occasion to interact frequently.

Based on interviews and our own review of the listnames in BTCO, we identified three general categoriesof lists: (a) formal organizational or geographic workunits, such as departments and office locations, (b) socialgroups, such as the company softball team; and (c) work-related teams or professional interest groups, such asgroups for specific molecules under investigation, or spe-cific disease areas. As we describe subsequently, themajority of email distribution lists at BTCO were con-structed for work-related activities and prescribed jobfunctions, but it is clear too that these lists also mapindividuals to a wide variety of informal work and socialgroups.

In addition to the distribution lists, we collectedencrypted employee records from the human resourcesystems of both companies. We used the same encryp-tion algorithm across both data sets (distribution listsand human resource records) so the data sets couldbe merged at the person-level via a common, hashedidentifier. We obtained information on employees’ sex,tenure, rank, organizational subunit, supervisor, annualperformance rating, and target bonus paid (the latteronly at BTCO). Because annual performance ratings andbonuses reflect contributions made in the prior year ofservice, we collected the human resource data in the yearafter the one from which we drew distribution list data.

Dependent Variables. At ISCO, our measure of attain-ment is based on the performance rating given to anemployee by his or her supervisor. Ratings range from 1(does not meet expectations) to 5 (exceeds all expecta-tions). We then created an indicator, high performance,which was set to 1 for employees receiving a 4 or 5rating (48% of employees in the sample).3

At BTCO, we obtained both performance ratings andbonuses, which are highly correlated. We chose annual

Dow

nloa

ded

from

info

rms.

org

by [

128.

32.7

4.70

] on

24

Aug

ust 2

017,

at 1

4:22

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Liu, Srivastava, and Stuart: An Intraorganizational Ecology of Individual Attainment96 Organization Science 27(1), pp. 90–105, © 2016 INFORMS

bonus as the dependent variable because it is moredirectly tied to individual attainment. Specifically, inBTCO’s compensation plan, the HR department used aformula to provide supervisors with a target bonus foreach of the manager’s direct report. Supervisors thenhad discretion to adjust this target bonus based on theirevaluation of the employee’s performance in the prioryear. We report the ratio of actual bonus divided bytarget bonus. This was formulaically centered near onebased on the firm’s compensation policy: the “aver-age” performer achieved the target under the bonus planguidelines.

Niche Characteristics—Crowding. To measure char-acteristics of intraorganizational niches, we first created,for each organization, a person-by-distribution list, two-mode matrix, Ai1k:

ai1 k = 1 if i ∈ k3 0 otherwise1

where i indexes actors and k indexes lists. To defineniche crowding, we converted the two-mode matrix,Ai1k, into a one-mode matrix, Oi1 j, of overlapping distri-bution lists between actors, i, and alters, j . The entriesof Oi1 j are given by the following:

oi1 j =∑

k

ai1 k ∗ aj1 k0

Next we defined a matrix of supervisor overlap, Si1 j,between i and j . The entries of Si1 j are given by thefollowing:

si1 j = 1 if i1 j share same supervisor1 z3 0 otherwise0

Competitive crowding of the niche occupied by actor iis then defined as:

Competitive crowdingi

=

∑

j 6=i44oi1 j∗si1 j5/4∑

k ai1 k +∑

k aj1 k55∑

j 6=i si1 j0

In this equation, the numerator represents the numberof email lists shared between actor i and a given alter j ,conditional on both reporting to a common supervisor, z.This measure is then weighted by the sum of distribu-tion lists to which i and j each belong, and summedacross all alters, j . Finally, we weight the resulting mea-sure by the number of alters, j , with whom i shares asupervisor. Translating this equation to words, competi-tive crowding rises for a focal actor i when colleagues jwho report to actor i’s supervisor are structurally equiv-alent to actor i in the overall activity space. To flexiblyidentify people occupying highly crowded niches, wecreated an indicator, set to 1 for people in the top quar-tile of the distribution of competitive crowding and to 0otherwise.4

Niche Characteristics—Status. We measured nichestatus based on an individual’s exposure, through listcomemberships, to high-ranking colleagues. These indi-viduals, who were all in well-compensated senior man-agement or executive positions, comprised fewer than10% of the employees in each organization.5 We defineda vector, E:

ei = 1 if i is high ranking3 0 otherwise0

Next, we calculated for each list, k, the proportion, pk,of high-ranking individuals on the list:

pk =

∑

i ai1 k∗ei∑

i ai1 k

0

The status of the niche occupied by actor, i, is then givenby the following:

Statusi =

∑

k pk∑

k ai1 k

0

That is, the status of an actor’s niche in the activity spaceincreases in the mean proportion of high-ranking col-leagues across all lists to which the actor belongs. Weagain created an indicator, which is set to 1 for people inthe top quartile of this distribution, and to 0 otherwise.

Niche Diversity. Activities vary in the extent towhich they bring together individuals who are otherwiseunlikely to interact. We constructed two, separate mea-sures of the diversity of employees’ niches. The firstreflects an individual’s exposure, through list comember-ship, to colleagues from different organizational units.The second gauges exposure to colleagues at differenthierarchical levels of the organization, regardless of thelevel. In other work on intraorganizational communica-tion patterns, it has been clearly shown that there isvery limited, direct communication between individualsin different divisions, functions, and organizational lev-els (e.g., Han 1996, Hinds and Kiesler 1995). Therefore,both measures reflect the extent to which individuals’activities in the company expose them to broad crosssections of organizational members, which they are oth-erwise unlikely to be in communication with (Kleinbaumand Stuart 2014, Kleinbaum et al. 2013, Srivastava andBanaji 2011).

The measures we use are based on the “Blau index”of heterogeneity (Blau 1977). For each list, k, we calcu-late the proportion of members, pu, across all U , whichindexes either organizational units or hierarchical levels.The resulting measure is as follows:

“Blau” diversityi =

∑

k441 −∑

u p2u5/4

∑

i ai1 k55∑

k ai1 k

That is, for each distribution list, we sum the squaresof proportions of members from each organizational unit

Dow

nloa

ded

from

info

rms.

org

by [

128.

32.7

4.70

] on

24

Aug

ust 2

017,

at 1

4:22

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Liu, Srivastava, and Stuart: An Intraorganizational Ecology of Individual AttainmentOrganization Science 27(1), pp. 90–105, © 2016 INFORMS 97

Table 2 Descriptive Statistics

ISCO BTCO

Mean Std. dev. Min Max Mean Std. dev. Min Max

Attainment a 30502 00715 1 5 10056 00265 0 205Competitive crowding—Same supervisor 00062 00098 0 1 00087 00095 0 00467Status—Seniority 00229 00131 0 1 00094 00051 00027 00406Diversity—Function 00051 00111 0 20843 00142 00223 00001 20001Diversity—Level 00178 00249 0 50664 00241 00314 00001 20723Female 00532 00499 0 1 00510 00500 0 1Number of lists 120220 60786 2 65 120244 60484 2 47Tenure 80549 60611 00523 4203 60759 60042 1 31

Note. N = 41661 for ISCO and 916 for BTCO.aAttainment was based on performance rating at ISCO and target bonus payments at BTCO.

or hierarchical level, which we then subtract from one.We then divide this quantity by the size of the distri-bution list. We chose to denominate by the size of thelist for substantive reasons: larger lists, such as onesthat encompass all employees or an entire division ofthe organization, represent less meaningful social groupsthan smaller ones. Insofar as the mechanisms of actioninvolve exposure to information and relationships fromcoparticipation in organizational activities, we believethat particularly when diverse actors are involved, mean-ingful communication and relationship building is likelyto occur only in small groups.6 Finally, we compute themean of this measure across all lists of which person iwas a member. We again create indicators based on thesemeasures, set to 1 for people in the top quartile of thedistribution, and to 0 otherwise.

Control Variables. To interpret the effects of individ-uals’ niches, it was necessary to control for employees’overall level of engagement with the activities in thefirm. To account for the skewed distribution of this mea-sure, we created indicators for individuals in each quar-tile of the distribution of list memberships. We includedthis highly flexible specification to guard against the pos-sibility that the coefficients for any of the other nichecharacteristics pick-up a misspecification of the func-tional form of the relationship between the volume ofwork in which employees participated and the outcomevariables.

In addition, we include an indicator for sex (set to 1for females) and multiple dummy variables for employeerank and for job function. We also control for employeetenure using linear and quadratic terms, following con-vention in the estimation of earnings equations. Finally,in all regressions using the BTCO data, we includeindicators for employees’ highest educational degreeattained. (Educational attainment was not available foremployees of ISCO.)

For analyses of attainment at ISCO, where the depen-dent variable was a binary indicator of whether or not anemployee received a high performance rating, we esti-mated logit regressions with robust standard errors. For

analyses of attainment in BTCO, the dependent variable,the proportion of target bonus actually paid, is nor-mally distributed. We conducted these estimations usingordinary least squares (OLS) regressions with robuststandard errors, which we clustered at the laboratorylevel because the lab head was the primary decisionmaker in setting individual bonuses.

4. ResultsTable 2 reports descriptive statistics. Of the 4,661 em-ployees in the data from ISCO, 53.2% were female.The mean tenure in the organization was 8.5 years. Themean performance rating was 3.5 on a 5-point scale. Ofthe 916 employees in BTCO’s research division, 51%were female. The mean tenure at BTCO, 6.8 years, wasslightly lower than at ISCO. Despite the many differ-ences between the two organizations, we found surpris-ing consistency in employees’ levels of participation inthe activities space. At ISCO, the average employeewas a member of 12.2 distribution lists. Remarkably,at BTCO, the typical staff member also belonged to12.2 lists.

Because these data have not been previously usedin organizational analysis, we describe in Table 3 thecomposition of the mailing lists in BTCO in somedetail. Based on a review of list names and our knowl-edge of the BTCO organization, we identified threecategories of distribution lists: social, organizational,and workflow.7 Social lists included a broad range ofrecreational activities and pragmatic interest groups,such as the running club and carpools. Organizationallists included same-building occupants and membersof formal organizational units. Given our focus onthe R&D organization, workflow lists often assembledpeople into science-based specializations and interests,such as lists for specific drug targets and molecularpathways. On average, organizational and social listmembership rosters were considerably larger than thoseof workflow lists. Not surprisingly, organizational lists

Dow

nloa

ded

from

info

rms.

org

by [

128.

32.7

4.70

] on

24

Aug

ust 2

017,

at 1

4:22

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Liu, Srivastava, and Stuart: An Intraorganizational Ecology of Individual Attainment98 Organization Science 27(1), pp. 90–105, © 2016 INFORMS

Table 3 Distribution List Descriptive Statistics—BTCO

Social lists Organizational lists Workflow lists

(e.g., sports, (e.g., departments, (e.g., molecules,commuting) labs, buildings) molecular pathways)

Avg. no. of list members 2008 (31.2) 2103 (29.1) 809 (12.7)Mean age 3600 (4.6) 3808 (3.7) 4203 (5.3)Mean tenure 407 (2.4) 505 (2.7) 704 (5.3)Mean % married 48 (24.2) 64 (22.7) 74 (27.3)Mean % female 50 (17.4) 53 (23.4) 35 (29.8)Mean % white 47 (30.4) 54 (23.9) 64 (32.9)Mean % Ph.D. 33 (19.7) 51 (31.0) 74 (26.3)Mean status—Senior colleagues 00006 (0.013) 00006 (0.016) 00060 (0.105)Mean functional diversity 00065 (0.045) 00062 (0.093) 00085 (0.091)Mean rank diversity 00068 (0.048) 00084 (0.078) 00135 (0.090)Total no. of unique individuals 278 802 342Total no. of lists 18 84 71

Notes. Where indicated, list means are shown, with standard deviations in parentheses. There are no statistics for competitive crowding,as there is not an obvious, list-level analogue to this.

matched the overall demographics of the research divi-sion. In contrast, social lists comprised younger, eth-nically more diverse, and less-well-educated members.Workflow lists, by contrast, were male- and White-dominated, with an older membership, a higher fractionof doctoral degree holders and a high degree of func-tional and rank diversity. Thus, we observe a first-ordercorrespondence between workplace demography and theorganization’s activities space, as revealed by distribu-tion lists.

We report a separate regression table for each of thetwo organizations: Table 4 reports results correspondingto ISCO, and Table 5 presents an analogous table forBTCO. In Model 1 in Table 4, the baseline regression,Female is positive and statistically significant, as are theindicator variables representing the two largest quartilesof distribution list membership. Female employees andthose deeply engaged in the organization’s many activi-ties were more likely to receive high performance ratingsthan males and those with limited involvement in inter-nal activities. In Model 2, consistent with Hypothesis H1,the indicator variable for an employee’s occupancy of acrowded niche at ISCO is negative and significant, sug-gesting that individuals in crowded regions of the activityspace experienced lower performance ratings. Model 3shows that the indicator variable for occupying an espe-cially high status niche is positive and statistically sig-nificant, which accords with Hypothesis H2. In Models4 and 5, in support of Hypothesis H3, the indicator vari-ables for occupying diverse niches, based on the mea-sures of functional and rank diversity, respectively, alsoare positive and statistically significant.

We estimated a final, full specification, which simul-taneously entered all covariates (Model 6). When theeffects are estimated jointly, the indicator variables forcompetitive crowding, status, functional diversity, andrank diversity indicator all are significant and all have

the hypothesized signs. Using the parameter estimatesin Model 6, the effects also are consequential in magni-tude. The odds that employees in crowded niches receivea stellar performance rating are 14% lower than theodds for employees in less crowded niches. By contrast,the odds that individuals in high status niches receive ahigh performance rating are 35% higher than the oddsof those in lower status niches. Likewise, occupyingpositions of high functional and rank diversity corre-lates with 30% and 21% higher odds of a stellar perfor-mance rating, respectively, for employees in functional-and rank-diverse niches. In sum, the results reported inTable 4 support the three hypotheses and indicate sub-stantively meaningful effect sizes.

Table 5 turns our attention to BTCO. In a baselinemodel, just as we found at ISCO, we observe that indi-viduals who were more engaged in intraorganizationalactivities, as measured by membership in a greater num-ber of distribution lists, receive higher bonuses.8 Greatereducational attainment also is positively associated withhigher year-end bonus (Model 1), which comes as littlesurprise in an R&D organization. Model 2 indicates sup-port for Hypothesis H1. Individuals who were in espe-cially crowded niches received bonus payouts that wereless than those received by their peers. Consistent withHypothesis H2, individuals in high status niches earnedbonus payouts that were higher than those of employ-ees in lower status niches (Model 3). In accord withHypothesis H3, Models 4 and 5 indicate that occupyingespecially functional- or rank-diverse niches is positivelyassociated with bonus payments.

With the exception of rank diversity, the full modelfor BTCO (6) yields similar coefficients to the sepa-rately estimated effects, although the indicators for bothhigh status niche and rank diversity are less preciselyestimated in Model 6. In subsequent investigations, wefind that the imprecision in the coefficient estimates in

Dow

nloa

ded

from

info

rms.

org

by [

128.

32.7

4.70

] on

24

Aug

ust 2

017,

at 1

4:22

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Liu, Srivastava, and Stuart: An Intraorganizational Ecology of Individual AttainmentOrganization Science 27(1), pp. 90–105, © 2016 INFORMS 99

Table 4 Logit Regressions of High Performance Rating on Covariates and Robustness Check (ISCO)

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model R-7

Crowding—Same supv. (top 25%) −00217∗∗ −00150∗ −000924000725 4000745 4000785

Status—Senior colleagues (top 25%) 00418∗∗∗ 00303∗∗∗ 00328∗∗∗

4000765 4000805 4000995Diversity—Function (top 25%) 00412∗∗∗ 00259∗∗ 00291∗∗

4000775 4000835 4001115Diversity—Level (top 25%) 00302∗∗∗ 00192∗ 00175∗

4000825 4000875 4000805Female 00305∗∗∗ 00309∗∗∗ 00348∗∗∗ 00326∗∗∗ 00317∗∗∗ 00360∗∗∗ 00356∗∗∗

4000605 4000605 4000615 4000615 4000615 4000615 4000615Tenure 00016 00010 00018 00020 00022 00019 00023

4000155 4000155 4000155 4000155 4000155 4000155 4000155Tenure2 −00001 −00001 −00001∗ −00001∗ −00001∗ −00001∗ −00001∗

4000005 4000005 4000005 4000005 4000005 4000015 4000015No. of lists—25%–50% 00057 00070 00035 00025 00025 00009 −00103

4000935 4000935 4000935 4000935 4000935 4000935 4000975No. of lists—50%–75% 00271∗∗ 00271∗∗ 00214∗ 00196 00192 00131 −00015

4001005 4001005 4001015 4001015 4001025 4001035 4001125No. of lists—75%–100% 00574∗∗∗ 00562∗∗∗ 00396∗∗∗ 00362∗∗ 00370∗∗ 00174 00035

4001095 4001095 4001145 4001165 4001225 4001275 4001415Constant −00519∗∗∗ −00414∗∗∗ −00558∗∗∗ −00553∗∗∗ −00564∗∗∗ −00524∗∗∗ 00795∗∗

4000895 4000965 4000895 4000895 4000905 4000985 4002885

X2 8704 9602 11507 11309 9907 13702 13800Prob.>X2 00000 00000 00000 00000 00000 00000 00000N 4,661 4,661 4,661 4,661 4,661 4,661 4,661

Notes. Robust standard errors. Coefficients for function indicators not reported. Model R-7 represents a robustness check with loggedcontinuous niche covariates rather than the top 25% spline.

∗p < 0005; ∗∗p < 0001; ∗∗∗p < 00001 (two-tailed tests).

the full model is introduced because of a high correla-tion between the two niche diversity measures at BTCO,coupled with lower statistical power in this data set.In Models 7 and 8, we present regressions that simul-taneously include niche crowding, status, and eitherrank (Model 7) or function (Model 8) diversity, but notboth. Without the introduction of correlation betweenthese two variables, all findings are as hypothesized, andall effects are statistically significant. Using the resultsin Model 7 to illustrate magnitudes, crowding reducedbonus payouts by 5.4%, status increased bonus payoutsby 15.5%, and level diversity increased payouts by 5.7%.We conclude that the results at ISCO and BTCO arehighly compatible with one another and that they cor-roborate the hypotheses.

Finally, because the spline specification we have im-plemented requires judgment in the selection of cutpoints (we selected the 75th percentile of each distri-bution as the cut point), we have re-estimated the fullmodels with logged, continuous measures of each nicheattribute. We chose a log specification because the nichevariables exhibit right skew. Table 4, Model R-7, andTable 5, Models R-9 and R-10, report results with loggedcontinuous covariates. We find the results to be relativelyunaffected by these functional form choices, with oneprimary exception. In the full ISCO model, the effect of

Crowding is negative but not statistically significant, inthe log specification. Also, like in the full BTCO modelwith the 75th percentile spline, the two log (continu-ous) niche diversity measures (which are highly corre-lated with one another) are estimated more cleanly whenintroduced separately. The diversity effects are stronglystatistically significant when entered separately, and shyof statistical significance when entered jointly.

5. Concerns About Causality andEconometric Identification

Although the results are consistent with the theory, thelimitations of the data create a number of empiricalconcerns. Most significantly, because the number andmemberships of distribution lists at both research siteschanged only modestly during the observation window,the regressions are run in the cross-section. Therefore,we cannot eliminate the possibility that unobserved indi-vidual differences influence both the niches individualsoccupied in the emergent activity space and their attain-ment levels. In short, standard concerns about omittedvariables apply here. Likewise, the possibility exists thatniche positions also correlate with unobserved points oforganizational priority. It is conceivable, for instance,that mailing lists that incorporate diverse members of an

Dow

nloa

ded

from

info

rms.

org

by [

128.

32.7

4.70

] on

24

Aug

ust 2

017,

at 1

4:22

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Liu, Srivastava, and Stuart: An Intraorganizational Ecology of Individual Attainment100 Organization Science 27(1), pp. 90–105, © 2016 INFORMS

Table 5 OLS Regressions of Bonus on Covariates and Robustness Checks (BTCO)

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model R-9 Model R-10

Crowding—Same supv. −00035∗ −00025 −00053∗ −00054∗ −00020∗∗ −00021∗∗

(top 25%) 4000165 4000175 4000255 4000255 4000075 4000075Status—Senior colleagues 00079∗∗ 00074∗∗ 00150∗∗ 00155∗∗ 00141∗∗∗ 00140∗∗∗

(top 25%) 4000295 4000285 4000555 4000555 4000375 4000355Diversity—Function 00082∗∗ 00067∗ 00072∗ 00039∗∗∗

(top 25%) 4000275 4000295 4000275 4000095Diversity—Level 00061∗ 00028 00057∗ 00040∗∗∗

(top 25%) 4000275 4000295 400285 4000095Female −00004 −00004 −00004 −00004 −00005 −00004 −00004 −00004 −00004 −00004

4000175 4000175 4000175 4000175 4000175 4000175 4000165 4000175 4000175 4000175No. of lists—25%–50% 00026 00030† 00029 00023 00022 00026 00022 00020 −00016 −00030

4000185 4000185 4000185 4000185 4000185 4000185 4000185 4000185 4000215 4000215No. of lists—50%–75% 00045∗∗ 00045∗∗ 00041∗∗ 00031∗ 00033∗ 00024 00025 00026 −00033 −00045∗

4000155 4000155 4000165 4000155 4000175 4000175 4000165 4000175 4000225 4000225No. of lists—75%–100% 00175∗∗ 00172∗∗ 00163∗∗ 00118∗∗ 00126∗∗ 00093∗∗ 00101∗∗ 00104∗∗ 00038 00029

4000245 4000245 4000245 4000255 4000305 4000295 4000265 4000315 4000315 4000315Tenure 00007 00006 00005 00005 00006 00003 00002 00002 −00003 −00003

4000055 4000055 4000055 4000055 4000055 4000055 4000055 4000055 4000055 4000055Tenure2 −00000 −00000 −00000 −00000 −00000 −00000 −00000 −00000 −00000 −00000

4000005 4000005 4000005 4000005 4000005 4000005 4000005 4000005 4000005 4000005Highest degree—M.A. 00052∗ 00051∗ 00043† 00046† 00048† 00036 00038 00039 00033 00034

4000255 4000255 4000255 4000245 4000255 4000255 4000255 4000255 4000255 4000255Highest degree—Ph.D. 00120∗∗ 00112∗∗ 00089∗∗ 00116∗∗ 00116∗∗ 00081∗∗ 00086∗∗ 00085∗∗ 00051∗ 00055∗

4000225 4000225 4000235 4000225 4000225 4000235 4000235 4000225 4000235 4000235Constant 00923∗∗ 00942∗∗ 00929∗∗ 00929∗∗ 00928∗∗ 00950∗∗ 00999∗∗ 10000∗∗ 10465∗∗∗ 10442∗∗∗

4000305 4000305 4000305 4000305 4000305 4000315 4000405 4000395 4001135 4001095

No. of observations 916 916 916 916 916 916 916 916 903 903R2 0017 0017 0018 0018 0017 0019 0020 0019 0021 0021No. of lab clusters 148 148 148 148 148 148 148 148 146 146

Notes. Robust standard errors (clustered by laboratory for BTCO). Models 9 and 10 represent robustness checks with logged continuousniche covariates rather than the top 25% spline. Because they are highly correlated, neither Diversity—Function nor Diversity—Level issignificant when both are entered simultaneously as logged continuous niche covariates.

†p < 0010; ∗p < 0005; ∗∗p < 0001; ∗∗∗p < 00001 (two-tailed tests).

organization are more likely to be assembled to pursuesignificant corporate objectives, and employees are morelikely to be rewarded when they are involved in thesesalient teams. In short, do niche characteristics have acausal effect, or do they simply reflect the social pro-cesses that underlie how individuals match to lists?

We undertook two supplemental analyses to partiallyaddress this question. First, we re-estimated regressionmodels on two subsamples of employees, (i) those withlimited organizational tenure and, separately, (ii) thosewho held nonleadership positions in the formal hier-archy. We reasoned that relative to more senior andlonger-tenured employees, individuals in these subsam-ples would have had fewer opportunities and less discre-tion to proactively maneuver themselves into desirableniche positions. These employees were more likely tobe channeled into specific niche positions based on theirspecific roles in the organization. Table 6 reports resultsfrom these analyses for both research sites. For ease ofcomparison, we reproduce in this table the full modelsbased on the full samples from both companies. The pre-

vious, full-sample results for ISCO are in column (1),which is included to compare to columns (2) and (3).Column (2) is a regression in which the data are subseton short-tenured employees at ISCO, and columns (3)subsets on lower rank employees. Columns (2) and (3)largely confirm the results from the full sample modelin column (1)—although in the restricted samples someeffects drop to p-values < 0010.

The results for BTCO, which appear in columns (5)–(8)in Table 6, are once again less precisely estimated be-cause of small sample sizes, especially when we subseton low tenure and lower rank employees. The impreci-sion in the estimates notwithstanding, comparing the fullmodel using the complete sample from BTCO, whichis reproduced as Model 5 in Table 6, to columns (6)and (7), we see that the coefficients broadly are in line.The signs are identical in the restricted samples, even ifsome of the effects fall shy of statistical significance.

As we consider alternative interpretations for the find-ings in this paper, we believe that the most uncertainresult concerns the effect of niche status. This finding

Dow

nloa

ded

from

info

rms.

org

by [

128.

32.7

4.70

] on

24

Aug

ust 2

017,

at 1

4:22

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Liu, Srivastava, and Stuart: An Intraorganizational Ecology of Individual AttainmentOrganization Science 27(1), pp. 90–105, © 2016 INFORMS 101

Table 6 Robustness Checks—OLS Regressions of Bonus on Covariates and Logit Regressions of High Performance onCovariates—Subset of Employees with <3 yrs. of Tenure and in Nonleadership Roles; Alternative Measure of Status

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8Setting ISCO ISCO ISCO ISCO BTCO BTCO BTCO BTCOData set Int. model Tenure < 3 yrs. Nonleader All Int. model Tenure < 3 yrs. Nonleader All

Status—Rising status of 00418∗∗∗ 00064∗∗

alters (top 25%) 4000715 4000245Crowding—same supervisor −00150∗ −00520∗∗∗ −00435∗∗∗ −00141† −00025 −00077† −00038 −00038

(top 25%) 4000745 4001475 4001045 4000735 4000175 4000415 4000265 4000265Status—Senior colleagues 00303∗∗∗ 00367† 00479† 00254∗∗ 00074∗∗ 00099 00097∗∗ 00218∗∗

(top 25%) 4000805 4001975 4002565 4000805 4000285 4000705 4000315 4000715Diversity—Function (top 25%) 00259∗∗ 00689∗∗ 00585∗∗∗ 00223∗∗ 00067∗ 00094 00064∗ 00058∗

4000835 4002175 4001645 4000845 4000295 4000705 4000315 4000295Diversity—Level (top 25%) 00192∗ 00489† 00444∗∗ 00220∗∗ 00028 00113 00019 00015

4000875 4002585 4001645 4000865 4000295 4000875 4000305 4000305Female 00360∗∗∗ 00349∗∗ 00437∗∗∗ 00355∗∗∗ −00004 −00011 00016 00002

4000615 4001325 4001005 4006135 4000175 4000425 4000235 4000165No. of lists—25%–50% 00009 00285† −00171 −00132 00026 00013 −00014 00016

4000935 4001465 4001215 4000945 4000185 4000515 4000315 4000185No. of lists—50%–75% 00131 00504∗ −00248† 00099 00024 00017 −00011 00020

4001035 4002265 4001375 4001035 4000175 4000475 4000275 4000175No. of lists—75%–100% 00174 00275 −00412∗ 00160 00093∗∗ 00112 00079† 00076∗

4001275 4005585 4001085 4001285 4000295 4000955 4000415 4000305Constant −00524∗∗∗ −00804∗∗∗ −00351∗∗ −00715∗∗∗ 00950∗∗ 00946∗∗ 00997∗∗ 00991∗∗∗

4000985 4001555 4001115 4000935 4000315 4001515 4000565 4000415

R2 0019 0037 0019 0023X2 13702 8404 7504 16506Prob.>X2 00000 00000 00000 00000No. of observations 4,661 1,051 1,891 4,661 916 107 550 916No. of lab clusters 148 52 110 148

Notes. Status—Rising status of alters is the proportion of list-members who, in the prior year, received a promotion and salary increase (forISCO) and a high performance rating (for BTCO). Robust standard errors (clustered by laboratory for BTCO). Coefficients for rank, tenure,education, and function indicators not reported.

†p < 0010; ∗p < 0005; ∗∗p < 0001; ∗∗∗p < 0001 (two-tailed tests).

is drawn into question because we know that high sta-tus groups preserve their status by maintaining exclusivemembership criteria. In the typical case, only individualswho stand in the good graces of the elites in the organi-zation are asked to participate in the activities that createhigh status niches. In fact, it would be entirely consis-tent with our general understanding of status processesif much of the effect of memberships in high status listsis due to an underlying, unobserved assignment processthat matches individuals with certain, desirable charac-teristics to activities that include high status members.In short, the issue is one of reverse causation: do indi-viduals accelerate onto the fast track because they moveinto high status niches, or do they earn their place insuch activities because they are on the fast track?

In our measurement of the status of niches, changesin person-specific niche characteristics result from a fewdifferent processes. They occur when new distributionlists are created and no-longer-used ones are culled;when there are changes to the membership rosters ofexisting lists; and when there are changes to the employ-ment statuses 4promotions, departmental reassignments,etc.5 of the members of pre-existing lists. We believe that

empirical support for a causal effect of niche status isstronger if the econometric identification of the statuscoefficient for a focal employee is solely based on vari-ation in niche status that occurs when already existingmembers of that employee’s distribution lists experiencepromotions, are singled out for bonuses, or experiencesome other form of an increase in status. Estimationsbased on this empirical strategy exclude the variationthat arises when individuals are assigned to high statuslists for reasons we cannot observe and thus provide amore conservative test.

The challenge in implementing this empirical strat-egy is, once again, the short time frame spanned by thedistribution list data, and therefore the limited numberof promotions that occur within the observation win-dow. Nonetheless, when we conducted this analysis, wefound in both organizations that when promotions topositions of high status are earned by coparticipants inthe activities in which a focal individual engages, thatperson experiences higher performance appraisals. Thesefindings appear in Models 4 (ISCO) and 8 (BTCO) inTable 6, in which we add a covariate labeled, “Status,

Dow

nloa

ded

from

info

rms.

org

by [

128.

32.7

4.70

] on

24

Aug

ust 2

017,

at 1

4:22

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Liu, Srivastava, and Stuart: An Intraorganizational Ecology of Individual Attainment102 Organization Science 27(1), pp. 90–105, © 2016 INFORMS

Rising Star of Alters.” This result lends credence to thetest of Hypothesis H2.

Although these robustness checks corroborate ourmain findings, we acknowledge that we cannot cleanlyidentify causal effects with these data. In fact, the coef-ficients in many of the regressions are large, and itwould not be a surprise to discover that the magnitudesare partly attributable to unobserved correlates of nichepositions. In thinking through this issue, it is useful tomake a general distinction in the literature in estimat-ing treatment effects in observational data. Let us dis-tinguish between, (i) the process by which a focal actor(employee or organization) comes to occupy a particu-lar niche; that is, how he or she is assigned to a spe-cific treatment condition, and (ii) the causal effect of aniche characteristic on relevant outcomes. In develop-ing the arguments in the paper, we have focused on (ii)and ignored (i). We hope that we have convinced read-ers that, in theory, ecological conceptions of competitionin a finite resource space, and the specific measures ofniche characteristics that organizational ecologists havedeveloped in their past work, have potential explanatorypower in an intraorganizational context.

In ignoring the process of assignment of individu-als to niches, however, we have a weak claim to cleanidentification. What are the prospects for addressing thisweakness in subsequent work? Future research designscan improve upon our suggestive results in severalrespects. For instance, longitudinal distribution list datamay enable the researcher to model the matches of indi-viduals to mailing lists, and to exploit the understandingsthat emerge to estimate a causal effect of list member-ships on subsequent outcomes (cf. Azoulay et al. 2014).

Another, compelling identification strategy would beto exploit the information available in list type. Weassume that membership in many workflow lists andall social lists are voluntary. Conversely, membershipson organizational lists are prescribed. If this is cor-rect, in the short term, niche positions in organizationallists are exogenous, whereas positions in workflow listsare self-selected. For example, let us assume that everyemployee is assigned to exactly one physical office loca-tion; that there is a distribution list for each office; andthat office locations are assigned randomly, conditionalon rank and organizational unit. In this scenario, peoplewill be quasi-randomly assigned to offices that lead todifferential access to information and social capital. Forexample, if an individual happens to be assigned to anoffice that is coinhabited with high status actors, he/sheis more likely to develop relationships that provide valuefor career advancement. This is exactly the type of par-titioning of variance that future projects can use to morecleanly identify effects; one can imagine separately com-puting niche measures based on exogenous (employer-determined) and endogenous (self-selected) activities.

This discussion of exogeneity also raises an impor-tant difference between intra and interorganizationalcontexts—the different processes by which actors areassigned to niches. We can debate how much discre-tion there is in the classic ecological context: much ofwhat the ecological perspective has taught us is thatentrepreneurs face genuine constraints on new venturecreation because of legitimacy constraints and the needto appeal to predefined audience tastes, though recentformulations have acknowledged that these constraintsare “fuzzier” than once thought (Hannan et al. 2007,Bogaert et al. 2014). For instance, in resource partition-ing theory, the success of organizations depends on acorrespondence between identity claims and consumerperceptions of value and authenticity. We know fromthis work that social and economic resources are fun-neled into particular configurations based on societal val-ues and audience tastes, and this defines which orga-nizations acquire enough resources to thrive. But theseare admittedly invisible hands, and in the intraorgani-zational context we study, there is a heavier, visibleone. The metamechanisms in our analysis are competi-tion and social recognition, similar to ecological studiesof niches, but organizational actors play a more centralrole in the construction of niches than does the invisiblehand of the market. Thus, it is arguably the case thatintraorganizational niches are more exogenously deter-mined than are interorganizational niches, and we mustkeep this distinction and its implications in mind as wethink about the generalization of ecological theory to theintraorganizational context.

6. ConclusionThe goal of this paper is to focus an ecological lenson intraorganizational attainment. Drawing on ecologicalinsights about how niche characteristics script compet-itive and symbiotic dynamics in populations of organi-zations, we argue that individuals within firms occupyniche spaces that influence the resources they obtain.These intraorganizational niches are defined by posi-tioning individuals in recurring organizational activities,which sometimes correspond to the organization’s for-mal structure, but also derive from how employees sit-uate in the cross-functional project teams, task forces,and informal work and social groups that crisscross theformal structure. We identified three niche characteris-tics and theorized about their relationship to individ-ual attainment. These features are the extent to which aperson-specific niche is, (1) competitively crowded, (2) agateway to high status actors in the organization, and(3) a conduit to individuals in possession of dissimilarknowledge, skills, and ranks. We then proposed a novelsolution to the empirical challenge of this line of inquiry:How does the analyst observe the myriad forms of recur-ring activity that occur in organizations? Our approach

Dow

nloa

ded

from

info

rms.

org

by [

128.

32.7

4.70

] on

24

Aug

ust 2

017,

at 1

4:22

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Liu, Srivastava, and Stuart: An Intraorganizational Ecology of Individual AttainmentOrganization Science 27(1), pp. 90–105, © 2016 INFORMS 103

is to characterize niche positions based on an affiliationnetwork derived from the complete roster of electronicmailing lists within organizations.

The findings in the paper are remarkably consistentacross two, quite-different empirical settings, and twodistinct but related measures of attainment. We findthat people who occupied competitively crowded nichesachieved lower levels of attainment, whereas those whooccupied high status niches received more positive per-formance appraisals and bonuses. Likewise, using twodifferent measures of niche diversity, we find that indi-viduals who were exposed to a broader cross sectionof members of their organization’s rank hierarchy andlandscape of organizational units were evaluated morepositively.

This paper makes two primary contributions. First,it builds our understanding of ecological processes inintraorganizational settings. Whereas prior research hasconsidered the interplay of ecology and organizationalchange (e.g., Amburgey et al. 1993), our study demon-strates that the same niche characteristics that influencethe outcomes of resource competition in populationsof organizations also affect the attainment of individu-als within organizations. Second, we introduce a noveldata source that has broad applicability in organizationalresearch. Email distribution lists may provide a meansfor researchers to “dust the fingerprints of informal orga-nization” (Nickerson and Silverman 2009, p. 538) andto assess where people are located within the semifor-mal organization (Srivastava 2015b). In addition to thepotential to use the details of these data, such as a clas-sification of list by types, to hold a magnifying glassover the informal organization, the data promise an effi-cient and unobtrusive means to gather very rich infor-mation about the inner workings of even large, complexorganizations. In sum, this study lays a conceptual andempirical foundation for future research on the effectsof intraorganizational niches and for deeper investigationinto their origins and dynamics.

AcknowledgmentsAll three authors contributed equally. The authors thank lead-ers at ISCO and BTCO for providing access to the data forthis project. The authors also thank Senior Editor Ray Reagansfor constructive feedback throughout the review process. Theusual disclaimers apply.

Endnotes1The idea that niches arise from endogenous population struc-turing does raise thorny theoretical and empirical issues. Inthe context we study, niches are partially endogenous in thatthey are influenced by deliberate choices employees makeabout which groups to join. We will return to this issue inthe discussion section. Here, we simply note that the same isgenerally true of the literature on the realized niche. In thatcase, organizations make strategic choices about which prod-ucts and services to offer, which market segments to enter,

which employees to hire, and so forth. At the same time,their choices are greatly constrained, especially in alreadyresource-partitioned markets, in which pre-existing audiencetastes determine and delimit viable organizational choices.Thus, in most studies of the realized niche, measured nichepositions are aggregates of endogenous choices and exogenousfactors.2In ISCO, there was 88% membership overlap between thedistribution lists in use over a six-month period. In BTCO,the overlap was 82%. Given the limited amount of change inthe lists, we performed our analyses in the cross section. Ofcourse, this will limit our ability to make any causal claimsfrom the analyses to follow.3We have also treated the raw performance ratings as thedependent variable in ordered logistic regressions. In thesespecifications, we found the proportional odds assumptionto be violated. Therefore, we transformed the variable intoa binary indicator. We obtained comparable results to thosereported in Table 4 when using the 1–5 performance rating anda generalized ordered logit (using the “gologit2” command inSTATA) estimator. For ease of presentation, we have opted toreport only the results from the logistic regressions using thebinary indicator.4This simple spline offers greater flexibility than a linear effectand does not introduce the correlation of a quadratic specifi-cation. In all of the analyses, we use the 75th percentile as thecut point. Although it is necessary to make an arbitrary choicefor a cut point, we have replicated all of the regressions usingthe 90th percentile as the cut point for each covariate. Allniche covariates exhibit right skew. Therefore, as a robustnesscheck, we have also estimated models with log (continuous)niche covariates, which we describe below.5In ISCO, a high-ranking employee was defined as someone inan executive-level salary band. These individuals representedthe top 5% of employees in ISCO. In BTCO, a high-rankingemployee was defined as a laboratory-head. 9% of BTCOemployees were in these roles. We also constructed a sec-ond, network-based measure of status. In addition to executiverank, we obtained email data that enables us to calculate eachemployee’s centrality in the internal communication network.We are able to use these data to compute status-weighted mea-sures of the niche. In the equation above, simply replace eiwith an indicator variable that actor i is in the top 5% of theindegree centrality scores in the corporate email network, ver-sus the current, rank-based measure. When we do this, weobtain nearly identical results.6Our operative assumption is that in large groups, there isample opportunity for members with common backgroundsand organizational affiliations to band together, so that rela-tively homogeneous subgroups may form in large activities.When diverse participants actively engage in a small group,however, we anticipate that a greater amount of mixing willoccur.7In many instances, the title of a mailing list did not defini-tively indicate its type. For instance, some lists had uninter-pretable titles like, “6,789-f.” We chose to be conservative inassigning lists to categories: we categorized lists only when we

Dow

nloa

ded

from

info

rms.

org

by [

128.

32.7

4.70

] on

24

Aug

ust 2

017,

at 1

4:22

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Liu, Srivastava, and Stuart: An Intraorganizational Ecology of Individual Attainment104 Organization Science 27(1), pp. 90–105, © 2016 INFORMS

were highly confident in our ability to know their type. Basedon conversations with the company, however, we believe thatthe majority of the unclassified lists would fall in the workflowcategory.8Recall that the dependent variable in Table 4, ISCO, is receiv-ing a high performance evaluation. The dependent variable inTable 5 is an actual measure of the size of each person’s bonus.Hence, coefficient magnitudes between the two tables shouldnot be directly compared.

References

Abbott A (1988) The System of Professions (The University ofChicago Press, Chicago).

Allen TJ (1977) Managing the Flow of Technology: TechnologyTransfer and the Dissemination of Technological InformationWithin the R&D Organization (MIT Press, Cambridge, MA).

Amburgey TL, Kelly D, Barnett WP (1993) Resetting the clock:The dynamics of organizational change and failure. Admin. Sci.Quart. 38(1):51–73.

Audia PG, Kurkoski J (2012) An ecological analysis of competi-tion among U.S. communities. Indust. Corporate Change 21(1):187–215.

Azoulay P, Liu CC, Stuart TE (2014) Social influence given (par-tially) deliberate matching: Career imprints in the creation ofacademic entrepreneurs. Working paper, MIT Sloan School ofManagement, Cambridge, MA.

Barnett WP, Miner AS (1992) Standing on the shoulders of oth-ers: Career interdependence in job mobility. Admin. Sci. Quart.37(2):262–281.

Barnett WP, Baron JN, Stuart TE (2000) Avenues of attainment:Occupational demography and organizational careers in the Cal-ifornia civil service. Amer. J. Sociol. 106(1):88–144.

Baron JN (1984) Organizational perspectives on stratification. AnnualRev. Sociol. 10(August):37–69.

Baron JN (2004) Employing identities in organizational ecology.Indust. Corporate Change 13(1):3–32.

Barron DD, West E, Hannan MT (1994) A time to grow and a timeto die: Growth and mortality of credit unions, 1914–1990. Amer.J. Sociol. 100(2):381–421.

Baum JAC, Haveman HA (1997) Love thy neighbor? Differentiationand agglomeration in the manhattan hotel industry, 1898–1990.Admin. Sci. Quart. 42(2):304–338.

Baum JAC, Mezias SJ (1992) Localized competition and organiza-tional failure in the manhattan hotel industry, 1898–1990. Admin.Sci. Quart. 37(4):580–604.

Baum JAC, Singh JV (1994) Organizational niches and the dynamicsof organizational mortality. Amer. J. Sociol. 100(2):346–380.

Bergmann BR (1986) The Economic Emergence of Women (BasicBooks, New York).

Biancani S, McFarland DA, Dahlander L (2014) The semiformal orga-nization. Organ. Sci. 25(5):1306–1324.

Blau PM (1977) Inequality and Heterogeneity: A Primitive Theory ofSocial Structure (Free Press, New York).

Blau PM, Scott RW (1962) Formal Organizations: A ComparativeApproach (Chandler Publishing Company, San Francisco).

Bogaert S, Boone C, Negro G, van Witteloostuijn A (2014) Organiza-tional form emergence: A meta-analysis of the ecological theoryof legitimation. J. Management, ePub ahead of print March 25,http://dx.doi.org/10.1177/0149206314527129.

Bothner MS, Kang J-H, Stuart TE (2007) Competitive crowding andrisk taking in a tournament: Evidence from NASCAR racing.Admin. Sci. Quart. 52(2):208–247.

Brass DJ (1995) A social network perspective on human resourcesmanagement. Res. Personnel Human Resources Management13(1):39–79.

Breiger RL (1974) The duality of persons and groups. Soc. Forces53(2):181–190.

Breiger RL, Mohr JW (2004) Institutional logics from the aggregationof organizational networks: Operational procedures for the anal-ysis of counted data. Comput. Math. Organ. Theory 10(1):17–43.

Burt RS (1982) Toward a Structural Theory of Action: Network Mod-els of Social Structure, Perception, and Action (Academic Press,Inc., New York).

Burt RS (1987) Social contagion and innovation: Cohesion versusstructural equivalence. Amer. J. Sociol. 92(6):1287–1335.

Burt RS (1992) Structural Holes: The Social Structure of Competition(Harvard University Press, Cambridge, MA).

Burt RS (1997) The contingent value of social capital. Admin. Sci.Quart. 42(2):339–365.

Burt RS (2004) Structural holes and good ideas. Amer. J. Sociol.110(2):349–399.

Carroll GR (1985) Concentration and specialization: Dynamics ofniche width in populations of organizations. Amer. J. Sociol.90(6):1262–1283.

DiMaggio P (1986) Structural analysis of organizational fields: A block-model approach. Staw BM, Cummings LL, eds. Research inOrganizational Behavior, Vol. 8 (JAI Press, Greenwich, CT),335–370.

Dobrev SD, Kim T-Y, Carroll GR (2002) The evolution of orga-nizational niches: U.S. automobile manufacturers, 1885–1981.Admin. Sci. Quart. 47(2):233–264.

Dobrev SD, Kim T-Y, Hannan MT (2001) Dynamics of niche widthand resource partitioning. Amer. J. Sociol. 106(5):1299–1337.

Elton C (1927) Animal Ecology (Sidgwick & Jackson, Ltd., Lon-don, UK).

Feld SL (1981) The focused organization of social ties. Amer. J.Sociol. 86(5):1015–1035.

Fernandez RM, Fernandez-Mateo I (2006) Networks, race, and hiring.Amer. Sociol. Rev. 71(1):42–71.

Fernandez RM, Friedrich C (2011) Gender sorting at the applicationinterface. Indust. Relations: A J. Econom. Soc. 50(4):591–609.

Fleming L, Mingo S, Chen D (2007) Collaborative brokerage, gener-ative creativity, and creative success. Admin. Sci. Quart. 52(3):443–475.

Freeman J, Hannan MT (1983) Niche width and the dynamics oforganizational populations. Amer. J. Sociol. 88(6):1116–1145.

Han S-K (1996) Structuring relations in on-the-job networks. Soc.Networks 18(1):47–67.

Hannan MT, Freeman J (1977) The population ecology of organiza-tions. Amer. J. Sociol. 82(5):929–964.

Hannan MT, Freeman J (1987) The ecology of organizational found-ing: American labor unions, 1836–1985. Amer. J. Sociol. 92(4):910–943.

Hannan MT, Freeman J (1989) Organizational Ecology (Harvard Uni-versity Press, Cambridge, MA).

Hannan MT, Carroll GR, Pólos L (2003) The organizational niche.Sociol. Theory 21(4):309–340.

Hannan MT, Polos L, Carroll GR (2007) Logics of OrganizationTheory: Audiences, Codes, and Ecologies (Princeton UniversityPress, Princeton, NJ).

Haveman HA, Cohen LE (1994) The ecological dynamics of careers:The impact of organizational founding, dissolution, and mergeron job mobility. Amer. J. Sociol. 100(1):104–152.

Dow

nloa

ded

from

info

rms.

org

by [

128.

32.7

4.70

] on

24

Aug

ust 2

017,

at 1

4:22

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Liu, Srivastava, and Stuart: An Intraorganizational Ecology of Individual AttainmentOrganization Science 27(1), pp. 90–105, © 2016 INFORMS 105

Hinds P, Kiesler S (1995) Communication across boundaries: Work,structure, and use of communication technologies in a large orga-nization. Organ. Sci. 6(4):373–393.

Hutchinson GE (1957) Concluding remarks. Cold Spring HarborSympos. Quant. Biol. 22(2):415–427.

Ibarra H (1995) Race, opportunity, and diversity of social circles inmanagerial networks. Acad. Management J. 38(3):673–703.

Jensen M (2003) The role of network resources in market entry:Commercial banks’ entry into investment banking, 1991–1997.Admin. Sci. Quart. 48(3):466–497.

Kanter RM (1977) Men and Women of the Corporation (Basic Books,New York).

Kleinbaum AM, Stuart TE (2014) Inside the black box of the corpo-rate staff: Social networks and the implementation of corporatestrategy. Strategic Management J. 35(1):24–47.

Kleinbaum A, Stuart TE, Tushman ML (2013) Discretion within con-straint: Homophily and structure in a formal organization. Organ.Sci. 24(5):1316–1336.

Krackhardt D, Hanson JR (1993) Informal networks: The companybehind the chart. Harvard Bus. Rev. 71(4):104–111.

Lomi A (1995) The population ecology of organizational founding-location dependence and unobserved heterogeneity. Admin. Sci.Quart. 40(1):111–144.

Mark NP (2003) Culture and competition: Homophily and distanc-ing explanations for cultural niches. Amer. Sociol. Rev. 68(3):319–345.

McEvily B, Soda G, Tortoriello M (2014) More formally: Rediscov-ering the missing link between formal organization and informalsocial structure. Acad. Management Ann. 8(1):299–345.

McPherson JM (1983) An ecology of affiliation. Amer. Sociol. Rev.48(4):519–532.

Merton RK (1968) Social Theory and Social Structure (Free Press,New York).

Nickerson JA, Silverman BS (2009) New frontiers in strategic man-agement of organizational change. Nickerson JA, SilvermanBS, eds. Economic Institutions of Strategy (Emerald JAI, Bing-ley, UK), 525–542.

Peli G, Nooteboom B (1999) Market partitioning and the geometryof the resource space. Amer. J. Sociol. 104(4):1132–1153.

Phillips DJ, Zuckerman E (2001) Middle status conformity: Theo-retical restatement and empirical demonstration in two markets.Amer. J. Sociol. 107(2):379–429.

Podolny JM (2005) Status Signals: A Sociological Study of MarketCompetition (Princeton University Press, Princeton, NJ).

Podolny JM, Baron JN (1997) Resources and relationships: Socialnetworks and mobility in the workplace. Amer. Sociol. Rev.62(5):673–693.

Podolny JM, Stuart TE, Hannan MT (1996) Networks, knowledge,and niches: Competition in the worldwide semiconductor indus-try, 1984–1991. Amer. J. Sociol. 102(3):659–689.

Popielarz PA, McPherson MJ (1995) On the edge or in between:Niche position, niche overlap, and the duration of voluntary asso-ciation memberships. Amer. J. Sociol. 101(3):698–720.

Popielarz PA, Neal ZP (2007) The niche as a theoretical tool. AnnualRev. Sociol. 33(April):65–84.

Reis S, Negro G, Sorenson O, Perretti F, Lomi A (2013) Resourcepartitioning revisited: Evidence from Italian broadcasting. Indust.Corporate Change 22(2):459–487.

Ruef M (2000) The emergence of organizational forms: A communityecology approach. Amer. J. Sociol. 106(3):658–714.

Simmel G (1902) The number of members as determining the socio-logical forum of the group. Amer. J. Sociol. 8(1):1–46.

Sørensen JB (1999) The ecology of organizational demography:Managerial tenure distributions and organizational competition.Indust. Corporate Change 8(4):713–744.

Sorenson O, Audia P (2000) The social structure of entrepreneurialactivity: Geographic concentration of footwear production in theUnited States, 1940–1989. Amer. J. Sociol. 106(2):424–462.

Srivastava SB (2015a) Network intervention: A field experiment toassess the effects of formal mentoring on workplace networks.Soc. Forces 94(1):427–452.

Srivastava SB (2015b) Intraorganizational network dynamics in timesof ambiguity. Organ. Sci. 26(5):1365–1380.

Srivastava SB, Banaji MR (2011) Culture, cognition, and collaborativenetworks in organizations. Amer. Sociol. Rev. 76(2):207–233.

Stewman S, Konda SL (1983) Careers and organizational labor mar-kets: Demographic models of organizational behavior. Amer. J.Sociol. 88(4):637–685.

Stuart TE (1998) Network positions and propensities to collabo-rate: An investigation of strategic alliance formation in a high-technology industry. Admin. Sci. Quart. 43(3):668–698.

Stuart TE, Ding W (2006) When do scientists become entrepreneurs?The social structural antecedents of commercial activity in theacademic life sciences. Amer. J. Sociol. 112(1):97–144.

Stuart TE, Hoang H, Hybels RC (1999) Interorganizational endorse-ments and the performance of entrepreneurial ventures. Admin.Sci. Quart. 44(2):315–349.

Thompson JD (1967) Organizations in Action: Social Science Basesof Administrative Theory (McGraw-Hill, New York).

Tortoriello M, Reagans R, McEvily B (2012) Bridging the knowledgegap: The influence of strong ties, network cohesion, and networkrange on the transfer of knowledge between organizational units.Organ. Sci. 23(4):1024–1039.

Turco CJ (2010) Cultural foundations of tokenism: Evidence from theleveraged buyout industry. Amer. Sociol. Rev. 75(6):894–913.

Weick KE (1979) The Social Psychology of Organizing (Addison-Wesley Publishing, Reading, MA).

Christopher C. Liu is an assistant professor of strategyat the Rotman School of Management, University of Toronto.He received a DBA in Management from Harvard Universityand a Ph.D. in biology from MIT. His research examines howgeographic and social space, explored across settings rangingfrom scientists to politicians, constrain and shape an individ-ual’s social networks.

Sameer B. Srivastava is an assistant professor at the HaasSchool of Business at University of California, Berkeley. Hereceived a Ph.D. in sociology and organizational behaviorfrom Harvard University. His research examines how the socialstructure within organizations—particularly that manifestedin workplace social networks—affects individuals and theircareers.

Toby E. Stuart is a professor at the Haas School of Busi-ness at University of California, Berkeley. He received a Ph.D.in organizational behavior from the Stanford Graduate Schoolof Business. His research examines how social structure andsocial networks influence entrepreneurship, innovation, andother career and organizational outcomes.

Dow

nloa

ded

from

info

rms.

org

by [

128.

32.7

4.70

] on

24

Aug

ust 2

017,

at 1

4:22

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.