European Management Journal (2011) 29, 335–346

journal homepage: www.elsevier .com/ locate /emj

Cooperation among competitors as status-seekingbehavior: Network ties and status differentiation

Marco Tortoriello a,*, Vincenzo Perrone b, Bill McEvily c

a IESE Business School, Avenida Pearson 21 Barcelona, Spainb SDA, Bocconi University School of Management, via Bocconi 8, Milan, Italyc University of Toronto, Rotman School of Management, 105 S., George St. Toronto, Canada

02do

*

91

KEYWORDSNetworks;Status;Cooperation

63-2373/$ - see front mattei:10.1016/j.emj.2011.02.00

Corresponding author. Tel3.E-mail address: mtortoriel

r ª 2011

.: +34 91

lo@iese.

Summary Status-based affiliation represents a valuable resource in economic exchange.However, affiliation strategies introduce a status ordering paradox: whereas higher statusactors seek to maintain status hierarchies, lower status actors strive to affiliate with moreprestigious actors. In this paper, using original data on the network of social and profes-sional ties among 72 hotel managers clustered in an Italian hotel district, we develop atheory of status-seeking behavior that involves the exchange of status for valuedresources.ª 2011 Elsevier Ltd. All rights reserved.

Introduction

Social status is a valuable resource for market exchange.High status producers are able to charge higher prices or re-duce their costs of production, thereby increasing revenuesand market share (Podolny, 1993). Whenever a market ac-tor�s quality is difficult to asses a priori, market ties conveyimportant informational cues about a producer�s underlyingquality (Baum & Oliver, 1992; Podolny, 2001). As a conse-quence, high status firms tend to behave according to theprinciple of status homophily – i.e., interacting with firmsof similar status (Gould, 2002) – by avoiding entering trans-actions with lower status ones since such affiliations mayreduce their own status. However, as acknowledged by

1 Elsevier Ltd. All rights reserved.

2 113 237; fax: +34 913 572

edu (M. Tortoriello).

Podolny (1993, p. 846), the homophily assumption is only apartial account of status dynamics. While some firms areinterested in the reinforcement of the status quo, othersstrive to enhance their status position (Benjamin & Podolny,1999; Kilduff & Krackhardt, 1994; Latour, 1987; Podolny,1993; Roberts & Khaire, 2009) seeking affiliations with high-er status partners. The co-occurrence of these opposingforces represents a status ordering paradox: while lowerstatus actors strive to be associated with those of higherstatus, higher status actors avoid associating with lower sta-tus ones since this might tarnish their status position.

Conceptualizing social status as a resource that flowsthrough linkages is not new (Podolny, 1993), however, fewstudies have explicitly focused on the relationship betweenstatus dynamics and network structure. Looking at thesystem of relationships defined among different nodes,network scholars often consider the social structure as given

336 M. Tortoriello et al.

and assume, at least implicitly, that all the nodes areequally accessible. However social networks are seldom‘‘flat’’, undifferentiated structures; typically, differentactors within the same network occupy different sociallydefined positions (Podolny, 1993) and this might affect theiropportunities for interaction. Using Podolny�s metaphor,more than just ‘‘pipes’’ through which resources flow, net-works work also as prisms or lenses projecting actors� qual-ities to potential exchange partners (Podolny, 2001). In thispaper, we focus on the tension between status differencesand cooperative behaviors by arguing that status affiliationsare traded for exchange of valuable resources. We do so inthe context of an Italian hotel district, in which differencesin status among hotels and exchange of resources in termsof client referrals are observed.

This setting is ideal because trading status affiliation foreconomic resources requires elements of cooperation andcompetition. Competition, because higher status positionsare associated with greater economic benefits (Benjamin& Podolny, 1999; Podolny, 1993). Cooperation because sta-tus gets ‘‘traded’’ among actors through acts of affiliations(Coleman, 1990; Podolny, 1993). Specifically, we examinethe system of social relationships (i.e., network) among ho-tel managers operating in the same local district in order tounderstand how differences in network positions shapecooperation among competing organizations measured interms of client referrals. Along similar lines, Ingram andRoberts (2000) show that friendship ties among managersat rival hotels improve organizational performance throughcollaboration, mitigation of competition, and informationsharing. What is different in our study is the explicitacknowledgement that status differences among hotels af-fect the flow of resources among different organizationsand the treatment of status as a feature of social networkposition.

We argue that social status represents an important, yetrelatively unexplored, dimension of social networks. Wesuggest that both the social network of relationships andthe social hierarchy that results from these relationshipsinfluence cooperative behaviors. In doing so, we intend tocontribute to the stream of research on social networks byshowing that the perceived social order defined within anetwork is an important complement to social structuraleffects associated with undifferentiated/flat networks.Specifically, we propose a model of interorganizationalcooperation among competitors that is a function of dyadic

Interpersonal Trust - to embedded others - to non-embedded others

Exchange of Information - reciprocal - non reciprocal give - non reciprocal receive

Social Status - Higher - Same - Lower

Figure 1 Theor

ties, network structure, and social status. Using a uniquedataset on the network of social and professional ties among72 hotel managers operating in the resort area studied, weanalyze social status differentiation, exchange of informa-tion, and interpersonal trust, controlling for a number ofmarket-based variables that may predict cooperation (seeFigure 1).

Theory

Interorganizational cooperation: customer referralamong hotels

While there are many different forms of interorganizationalcooperation (development of a new product, joint R&Dinvestments, integrative agreements on commercial ex-changes, common policies on purchases etc.), not all ofthem are directly applicable to the hotel industry. Prior re-search on the hotel industry has identified several commonforms of cooperation, including: joint advertising efforts,jointly run schools of hospitality (Ingram & Inman, 1996),joint purchases of goods and services, chain affiliation(Baum & Ingram, 1998; Ingram & Baum, 1997) and referralsof overflow customers (Ingram & Roberts, 2000). This lastform of cooperation is particularly important. Because ofrivalry among hotel districts, hoteliers belonging to thesame district often prefer to retain customers within theboundaries of the district rather than lose them to nearbydistricts (Ingram & Inman, 1996). Given inter-district com-petition, referral of customers to other hotels with thesame district not only benefits the hotel receiving the refer-ral, but also indirectly benefits the district as a whole. Suchreferrals are also an additional service for the customersince they eliminate the need to search for alternativeaccommodations, thus saving customers� time and energy.Moreover, for the hotel making a referral, sending acustomer to another hotel in the same district retains thecustomer within the district and creates the possibility ofcompeting for this customer again at some point in thefuture.

Consistent with these arguments Ingram and Inman, intheir study of the US and Canadian Niagara Falls hotelpopulations, showed that while a given hotel in one site‘‘competes with the hotel down the block, [. . .] all hotelswithin a given community share rivals’’ (Ingram & Inman,

Likelihood of Customer Referrals

etical model.

Cooperation among competitors as status-seeking behavior: Network ties and status differentiation 337

1996, p. 631). This tension between intra- and inter-districtcompetition rests on the fact that two hotels within thesame district are competitors, but they both prefer to at-tract tourists to their district (so that they can competefor them) rather than lose them to another district. Forthese reasons, we considered customer referrals among ho-tels in the same district as a particularly important form ofinterorganizational cooperation.

At the same time, customer referrals entail potentialrisks. For instance, by sending customers to competing ho-tels the referring hotel might, overtime, erode its customerbase if referrals are not reciprocated. In addition, if cus-tomers are disappointed with the referral they might pro-ject their disappointed onto the referring hotel and choseto go elsewhere in the future. We argue that there are socialmechanisms at work that allow hotels to minimize suchrisks. One such mechanism is represented by the differentpositions that hotels occupy in the status ordering as per-ceived by hotel managers. Hotel managers seek to ensuresome minimal level of comparable service by referring cus-tomers to hotels of equal or higher status. In addition to sta-tus ordering, hotel managers engage in exchange ofinformation as a means of increasing the likelihood of reci-procity in customer referrals. To increase this likelihood,hotel managers also consider the trustworthiness of thecounterparts receiving the referral. In the followingsections we discuss each of these social processes in moredetail and explain how they influence interorganizationalcooperation among hotel managers.

1 An interesting counterargument would suggest that hotel man-agers refer to lower status hotels as a way of demonstrating thatthey are superior so that in the future the referred customers willprefer the referring, to the referred, hotel. However, this hypoth-esis is based on the assumptions that customers perceive thereferring hotels as more exclusive than the referred hotel and thatcustomers will come return to the same district. While we cannotrule out the first assumption, we shall notice that in less than 20 kmof coast-line (Rimini Riviera), tourists can chose between 7different hotel districts (see map below). This is likely to inducesome elements of intra/intergroup competition that would leadhotel managers to think not only in terms of single hotels but also interms of hotel district.

Social status

Podolny defines producer�s status in the market as ‘‘the per-ceived quality of that producer�s product in relation to theperceived quality of that producer�s competitors� products’’(1993, p. 830). Holding quality constant, higher status pro-ducers realize additional revenues either because of greatermarket shares or higher prices (Podolny, 1993). Moreover,since market exchanges can be seen, at least in part, as actsof affiliation, one would expect producers to carefully man-age their strategy of affiliations. For instance, produces mayseek to establish cooperative relationships with higher sta-tus producers and avoid relationships with lower statusones. Being associated with higher status actors enhancesone�s position in the status hierarchy, while being associatedwith lower status actors detracts from it. Yet, research onstatus dynamics has thus far converged on the widely ac-cepted view that strong inertial forces contribute to keep-ing the status ordering relatively stable. Higher statusactors will seek association with similar status partners try-ing to avoid, at the same time, relationships with lower sta-tus ones. While several studies focus on the mechanismsreinforcing status homophily (Benjamin & Podolny, 1999;Gould, 2002; Podolny, 1993) little work has looked at themechanisms through which social actors attempt to modifythe existing status quo. Following Coleman (1990) we pro-pose that affiliations between low and high status actors oc-cur when status is traded for immediately tangible benefits.As Coleman (1990, p. 130) argued ‘‘the awarding of status tobalance unequal transactions or to make possible half-transactions appears to be the most widespread functional

substitute for money in social and political systems.’’Accordingly, we argue that lower status hotels will seek toassociate with higher status hotels by referring customersto them. Higher status hotels, while not reciprocating cus-tomers to lower status ones, will enjoy nonetheless the flowof referrals from lower status hotels as a ‘‘fee’’ to compen-sate for the asymmetric affiliation.

In the context of this study, whenever a hotel managersends customers to another hotel, implicitly she is makinga comparison between the two. She will prefer high quality,reliable hotels that will not leave her customers disap-pointed. Therefore we argue that, when making referraldecisions, hotel managers favor hotels that are at least asgood as their own in terms of quality. In addition to that,referrals to higher status hotels should also be driven bythe tendency for individuals to strive to be associated withpositively perceived others.1 If someone is perceived to beconnected with a positively evaluated other, then he is alsolikely to be positively evaluated and this will eventually leadto increasing flows of resources and opportunities (Podolny,1993). This tendency to seek prestigious affiliation hasbeen described as the ‘‘basking-in-reflected-glory effect’’(Cialdini et al., 1976; Kilduff & Krackhardt, 1994), and itis based on the way individuals extend their perceptionsfrom one subject to another. In our case, when tourists stayat a referred hotel, at least part of their overall experiencewill be associated with the referrer. As Podolny argued ‘‘thepresence (or absence) of a tie between two market actors isan informational cue on which others rely to make infer-ences about the underlying quality of one or both of themarket actors.’’ (Podolny, 2001, p. 34). In this sense, qual-ity of referral can be of strategic importance to the extentthat customers equate the quality of the hotel receiving thereferral with that of the sending hotel. Sending overbookedcustomers to hotels that enjoy higher status can be benefi-cial to the sender�s own status because it proxies for one�sown quality. Accordingly, we predict,

H1:. Customer referrals will be positively associated withthe status of a hotel such that hotel managers will sendreferrals to hotels of higher status.

Sharing of information

Customer referrals, while providing a service to the custom-ers, expose the referring hotel to the risk of eroding itscustomer base. In fact, while reciprocation of customerreferrals could provide a net benefit to both parties, if the

338 M. Tortoriello et al.

referred hotel does not cooperate by sending customers inreturn, the referrer faces the risk of a reduction in his/herbusiness. We argue that preliminary forms of interactioncould be beneficial to limit this risk, thus enhancing the like-lihood of referral decisions. The exchange of informationprovides cues to evaluate others� intentions, while makingthe relationship between the actors involved more salientthan others. For instance, Ingram and Roberts (2000)showed that friendship between hotel managers at compet-ing hotels was beneficial for hotel revenues because of theexchange of relevant market information. Similarly, Uzzidescribed the mechanism of fine-grained information shar-ing as one of the building blocks of embedded relationships(Uzzi, 1997). Information exchange between organizations,especially if they are competing, can be seen as a form ofcooperation per se; however, what we consider particularlyimportant about information sharing is its role in facilitatingmore consequential forms of cooperation.

In this sense, information sharing represents an interme-diate step toward the establishment of more significantforms of cooperation such as customer referrals. In general,socio-economic interactions are based on exchange of infor-mation, and exchange of information facilitates the coordi-nation of economic activity. However, the exchange ofinformation also represents a social process through whichexchange partners get to know each other better and eval-uate each others� motives and intentions. The extent towhich information exchange is reciprocal and mutuallyacknowledged by both parties is a crucial test that deter-mines whether to further pursue the relationship or not. Ifthe information flows in one direction, only the recipientof the information will benefit. The sender will eventuallyrealize that the asymmetric relationship is not worth pursu-ing since the recipient is not reciprocating. To emphasizethe role of reciprocity, we conceptualize the exchange ofinformation between two hotels as a relationship in whichthey both exchange information with one another and theyare both aware of this process. When information is recipro-cated, and reciprocity is mutually acknowledged by bothparties, the mutual confidence that is developed throughinteraction is likely to evolve toward more compelling anddirect forms of cooperation (i.e. customer referrals). Theflow of information, the common construction of reality en-tailed by this process, and the opportunity for repeatedinteractions represent a fertile ground for further develop-ment of mutually rewarding forms of interactions. When acounterpart reciprocates information, she is proving thatshe is not opportunistically taking advantage of the informa-tion received but is willing to return the favor by sharing herown information. This is interpreted as a preliminary steptoward the establishment of more substantial forms ofinteraction such as the sharing of customer overflows.

H2: Customers referrals will be positive associated withconfirmed reciprocal flows of information among hotelmanagers.

Trust

Although the decision to refer customers to a competitorentails considerable risks, hotel managers� expectations

about the propensity for different hotel managers to recip-rocate informs their decision regarding the choice of a par-ticular competitor to which to refer customers. We adoptthe perspective that trust acts as a powerful heuristic thatallows economic agents to conserve on cognitive resources(McEvily, Perrone, & Zaheer, 2003; Zaheer, McEvily, &Perrone, 1998; Uzzi, 1997). Trust is generally defined asthe willingness to be vulnerable (Mayer, Davis, &Schoorman, 1995) under conditions of risk and interdepen-dence (Rousseau, Sitkin, Burt, & Camerer, 1998, p. 395).We argue that there are three main reasons why trust leadsto the development of interorganizational cooperationamong rival hotels: counterparts� willingness to reciprocate,counterparts� ability to satisfy the customer, and reductionof the cognitive load associated with the choice of referral.

The expectation that a counterpart will not take advan-tage of us, even when the possibility of opportunism is pres-ent, is an important precursor to interorganizationalcooperation (Rousseau et al., 1998). Trust in another hotelmanager should facilitate interorganizational cooperationin the form of referrals because it creates the expectationthat the referred hotel would not asymmetrically exploit areferral, but rather would acknowledge it and eventuallyreciprocate in the future. If the hotel receiving the referraldoes reciprocate in the future, customer referrals wouldsum to a mutually beneficial inter-temporal exchange ofresources.

A second reason why trust should influence the referraldecision is due to competence, or the expectation thatthe customer would be served well at the referred hotel.As Ingram and Roberts (2000, p. 391) reported in their studyof the Sydney hotel industry, customer referrals was com-mon between hotel managers who are friends ‘‘also becausethe friendship allowed a manager to trust that the referredcustomers would be well treated.’’ In this way, counterparttrustworthiness can be interpreted as professionalreliability.

A third reason why trust should influence the referraldecision is due to what Uzzi describes as the ‘‘heuristic va-lue of trust’’ (Uzzi, 1997, p. 44). Decision-making processesusually entail the evaluation of alternatives. This requirestime, attention, cognitive resources, and efforts that thedecision-maker might have allocated elsewhere. Heuristicsare decision-making shortcuts ‘‘that economizes on cogni-tive resources, time, and attention without jeopardizingthe quality of decisions’’ (Uzzi, 1997). Relying on a trust-worthy counterpart is one such heuristic and represents auseful way of simplifying the referral decision-makingprocess.

Taken together these arguments lead us to predict thatreferrals are more likely to be directed to managers consid-ered trustworthy by their peers. Accordingly:

H3a: Customer referrals will be positively associated withthe trustworthiness of a hotel manager.

Trust is inherently a relational phenomenon involving, asource (trustor) and a recipient (trustee). However, trustcan also be embedded in the social structure surroundinga dyad. Trusting someone that is trusted by others to whomboth the trustor and trustee are connected is distinct fromtrusting someone where there are no overlapping ties to

Cooperation among competitors as status-seeking behavior: Network ties and status differentiation 339

third parties. Being embedded in a network of common thirdparties means that the trustee has more at stake than atrustee with non-overlapping ties in terms of reputation,and therefore should be more likely to honor its obligationsto the trustor. Trusting a disconnected trustee is more riskybecause the consequences of defecting are lower since noneof the trustee�s contacts observe untrustworthy behavior.This is consistent with Coleman�s (1988, pp. 105–108) no-tion of closure in social networks, according to which multi-ple, redundant ties create trustworthiness in a socialstructure. Based on this argument, the same relationship(trust in this case) could have different consequencesdepending on how social ties are distributed within the net-work. One would be more willing to develop a cooperativerelationship with highly trusted individuals since they havemore at stake in terms of reputation. Outside close-knit net-works, defection could be sanctioned only by the person towhom the obligation is owed, while in a structure with highclosure an eventual defection is open to the sanctioning of abroader set of contacts (Coleman, 1988). Based on thesearguments, we make the following prediction:

H3b:. The positive association between customer referralsand trustworthiness of a hotel will be stronger when thetrustor and the trustee share ties to common third parties.

2 From the questionnaire it turned out that chain affiliation wasvery limited (less than 6% of the hotels in our sample). Also from theinterviews, more than 80% of the hotel managers reported theyusually send customers to other hotels within the district; while notone reported referrals toward other districts. Neither joint pur-chases of goods and services nor common advertising efforts wasreported by any of the hotel managers.3 Hotel managers typically have direct experience with all of their

guests; they greet them upon arrival, and they are always availableshould any problem arise. The typical hotel in this district has alimited number of rooms (average = 32) and being family owned,owners proud themselves with the close relationships they are ableto entertain with all of their customers.

Similarity and other confounding effects

To assess the validity of our hypotheses about the role of so-cial structure as a predictor of cooperation, we control fordifferent hotel features such as hotel category, physical dis-tance, price ranges, number of services offered, hotel size,and chain affiliation. The greater the number of character-istics that two hotels have in common, the higher the like-lihood that a referral might take place regardless of thesocial relationships existing between hotel managers. Onewould expect, for instance, that referrals would not spanacross hotel categories, or that similar prices and servicesshould be offered both by the referrer and the referred ho-tel, or that hotels located close to each other are morelikely to share customers than hotels located in different re-gions. Hotels that belong to the same category usually havea common target in terms of customers; they generally offersimilar services with comparable standards and also thenumber of personnel and their professional skills are sup-posed to be similar.

Physical proximity might also be related to customerreferrals, both for practical and theoretical reasons. Froma practical point of view, relative distance between hotelsmight be salient when redirecting customers; closer hotelsmight be more readily available in the mind of the referrerthan remote ones or customers might just prefer not tomove too far from their original choice because they likea particular area or because they want to get settled as soonas possible. From a theoretical standpoint, physical locationitself represents a measure of similarity between organiza-tions; for instance, hotels similar along some dimensionstend to be located close to each other (Baum & Haveman,1997).

Price could also be an important determinant ofcustomer referrals; in order to meet customers� initial

expectations, the price of the new hotel should be as closeas possible to the one of the referrer. A similar argumentcan be made for the number of services offered; hotelsoffering a range of services similar to the referrer are morelikely to be taken into consideration. We also wanted tomake sure that no confounding effects due to prices (ceterisparibus low price hotels are likely to be filled first); numberof services offered (hotels offering a particularly small num-ber of services might be unable to cope with customer�sexpectation); hotel category (e.g. referrals taking placeonly in given categories); or hotel size (ceteris paribus,smaller hotels are more likely to fill more quickly than largerones, and larger hotels might be considered sooner sincethey are more likely to have vacancies).

Methods

Site and measures

The district studied, is located in Italy in the Marche regionand it is in many ways an extension of the well-known RiminiRiviera to the north. Given the well-established tourist tra-dition of the area, this district is one of the busiest holidayspots in Italy. In only 20 km of coast, it is possible to choosebetween several distinct tourist districts. During a series ofpreliminary field interviews, hotel managers noted that cus-tomer referrals were the most common form of cooperationadopted, and, in addition to that, costumer referrals wereconsistently reported by hotel managers as a common prac-tice within the district.2

We collected data about the hotels studied from twomain sources: questionnaires and archival data. Question-naires were filled out by each hotel manager in the presenceof one member of the research team, and the objective ofeach questionnaire was the collection of information aboutsocial relationships defined between the respondents and allother hotel managers within the district. Archival data werecollected using local and national hotel guides and the infor-mation provided by the Local Hotel Association. Out of the102 hotels located in the district studied, we were eventu-ally able to obtain complete data for 72 of them (71% ofthe population).

The organizations studied have a relatively small numberof employees (11 on average) and given the small scale ofactivities, managers are directly involved in all the relevantaspects of business.3 The hotel district is extremely densesuggesting that many opportunities for social and profes-sional interaction exist. We gathered some of the key vari-ables using relational questionnaires, while for others we

340 M. Tortoriello et al.

used publicly available information (like the Local HotelAssociation or National hotel guides). For social networkvariables, we provided respondents with a list of all themanagers working within the district and asked them tocheck the name of the managers to whom they were tiedby a particular relationship. Next, we describe how eachvariable was operationalized.

Customer referrals. Our dependent variable was definedby asking each hotel manager ‘‘. . . to which hotel do youusually refer customers when you are overbooked or whenyou cannot accommodate particular customers� requests’’.The result was a relational matrix where a value of 1 definedbetween hotel i and hotel j means that hotel i refers cus-tomers to hotel j.

Status. We measured status by asking hotel managersto indicate all the organizations within the district thatthey considered to be ‘‘always a step ahead of the othersbecause of their openness toward the adoption of newtechnology and new techniques, their willingness to inno-vate and their particular ability to satisfy customer�s newneeds’’. Social status is defined here not in terms of cus-tomers� but in terms of competitors� perception of quality.Starting from this relational information, we computedthe Bonacich�s (1987) c(a, b) centrality measure which istypically used to derive status ordering from relationaldata (Podolny, 1993, 1994; Benjamin & Podolny, 1999;Podolny, 2001). Bonacich�s centrality is formally definedas:

cða; bÞ ¼ aX1

k¼1bkRkþ11

where a is a scaling factor, R is the transposed of the origi-nal matrix with the status evaluations of each hotel man-ager, b is a weighting factor and 1 is a column vector ofones. The parameter a affects only the length of the vectorc(a, b) and in the analysis it was set in such a way that:X

k

ciða; bÞ2 ¼ n

where n is the number of actors in the social system. In thedefinition of status measures not all the nodes are equal; infact a node�s centrality in the status network is actually afunction of the centrality of nodes to which it is related(Bonacich, 1987). Status acknowledgement received fromlow status nodes is not the same as status acknowledgementreceived from high status nodes. The formula used here canbe interpreted as both a measure of degree centrality orstatus; the difference is given by the parameter b. When bapproaches zero the formula yields values close to degreecentrality. When b > 0 the formula takes into account othernodes� centrality. Greater values of b assign higher weightsto more central nodes. The weighting factor b in our casewas set equal to the reciprocal of the largest eigenvalueof the relational matrix,4 in this way, the higher the statusof one hotel, the more it is enhancing the status of anotherhotel in the relational matrix. Once we defined the vectorwith the status scores, using the Attribute routine in Ucinet6we projected it into a matrix that took into account all the

4 This is because the reciprocal of the largest eigenvalue repre-sents b�s upper bound. However, even considering a range of valuesfor b, we found no substantial differences in the status scores.

differences in status between each dyad of hotels. Theresulting matrix was divided into three different matricesto consider the three different situations of lower, equal,or higher status. The first matrix defines the relationshipof lower status (i, j = n meaning that hotel i has a statusscore n points lower than hotel j for every i 5 j); the secondmatrix defines the relationship of equal status (i, j = nmeaning that hotel i has a status score of n points that isequal to the status score of hotel j for every i 5 j); the thirdmatrix defines a situation of higher status between hotels (i,j = n meaning that hotel i has a status score n points higherthan hotel j for every i 5 j).

Information exchange. Given the goal of defining anobjective measure for the information-sharing network,we asked two different questions of each manager; oneabout receiving information and the other about sendinginformation. We first asked them to check the managerson the list to whom they usually go when they need someinformation concerning work-related topics. In a separatequestion, we asked them to check which managers usuallycome to them seeking the same kind of information. Thesetwo questions allowed us to map the in-coming and out-going flow of information for any given dyad of hotels andto define three different variables: reciprocal exchange ofinformation, information given but not received and infor-mation received but not given. The first variable was in-tended to measure the actual, reciprocal sharing ofinformation between hotels. It is important to note that thisvariable captures not only that two nodes claim they sendinformation to each other (symmetry in giving information)but also that both of them confirm they receive informationfrom the counterpart (symmetry in giving and receivinginformation). The reciprocal sharing of information was de-fined as follow: in the cell ij a value of 1 would representthe situation in which i sends information to j, j sends infor-mation to i and both i and j acknowledge actually receivinginformation from their counterpart, a value of 0 was im-puted otherwise. This was considered as a measure of theobjective flow of information between two managers; onlyreciprocal and confirmed exchanges of information were re-tained. Similarly we defined two other variables describingthe situation in which: (1) i was giving information to j, jacknowledged he was receiving information from i but j inturn, was not sending information to i; (2) i was receivinginformation from j, j was giving information to i, but i inturn, was not giving information to j. These last two vari-ables described, respectively, the situation in which therewas asymmetry in sending information (given but not re-ceived back) and in receiving information (received butnot given back). Note that restricting our analysis to con-firmed relationships� asymmetry, in this case, describes ac-tual relationships and not eventual/possible perceptionbiases.

Interpersonal trust. Because trust is a complex andmultidimensional concept, it can be difficult to operation-alize. In a sociometric network study such as ours, whererespondents are asked to report on relationships with allother members of the network, measuring trust is evenmore challenging. Rather than attempting to devise a mul-tidimensional measure of trust (McEvily & Tortoriello,2011) we used a single-item measure. We first presentedrespondents with a formal definition of trust to call their

Cooperation among competitors as status-seeking behavior: Network ties and status differentiation 341

attention to the particular content of the question5; thenwe presented a list with all the names of the hotel man-agers in alphabetical order. The question asked was:‘‘according to the definition of trust that has been givenabove, whom do you consider to be trustworthy amongthe following?’’ This variable was defined by a 0/1 matrixin which a value of one in the cell ij means that i considersj to be trustworthy according to the definition given upfrontand a value of zero means that no such relationship is de-fined. From this matrix we then derived two different andmutually exclusive matrices to distinguish between the sit-uations in which the trustor and the trustee shared (or didnot share) ties to common third parties: trust to embeddedothers and trust to non-embedded others. Trust to embed-ded others describes the situation in which the individualsin the trust relationship are not just related with eachother, but they are also related to at least one commonthird party. Starting from the original trust matrix, we firstcalculated the co-membership matrix using the clique rou-tine in UCINET6. The co-membership matrix gives us the ex-tent to which different individuals belong to the sameclique(s). We then recoded the co-membership matrix tokeep only the trust relationships defined between individu-als tied to at least one third party in common. Trust to non-embedded others describes the relationship i considers j tobe trustworthy whenever i and j are not related to a com-mon third party. In other words this measure is given by a0/1 matrix in which the relationship of trust between a sen-der and a receiver occurs in isolation from common thirdparties.

Physical distance. This proximity variable was definedusing a ‘‘block-distance’’ measure to express physicaldistance between each hotel and all the others. Block-distance, derived from local maps, simply considers the sit-uation in which 2 hotels are in the same block, one blockaway, two blocks away etc. up to 4 blocks as the maximumdistance. We considered the log of all the distances amongall the hotels.

Price. To define this variable we collected ‘‘high season’’daily prices charged by each hotel for a person/night. Start-ing from the price data vector, we defined a squared matrixcomposed by the difference between each value in the vec-tor and all the others. The basic assumption here is thatprice similarity should increase the likelihood of referral.Hotel managers are more likely to send a customer to an-other hotel that charges about the same price as they dorather than to a hotel charging (in relative terms) a muchhigher or lower price. However, it is reasonable to assumethat a decrease in this difference will have a small effecton referrals if this difference is already small (that is, ifthe price of the two hotels is roughly the same). But, ifthe price difference is large, even a marginal change in rel-ative price will have a greater effect on referral decisions.For this reason we considered the price squared-differencesbetween each hotel and all the others. Just like for price, toconsider the impact of number of services offered by

5 Trust was defined as the ‘‘willingness to rely on someone inconcrete and important situation because of the positive expecta-tion that: (1) the subject will fulfill his obligations, (2) will bepredictable in his actions and (3) will not take advantage of youeven when external conditions would make it possible’’.

different hotels on customer referrals, we defined a matrixobtained by the squared difference of the number of ser-vices available in each hotel assuming that similarity inthe number of services offered would increase the likeli-hood of customers� referral. Referrals between hotelsshould be positively related to similarity in the number ofservices that they have in common; the higher thediscrepancy in the number of services offered the lowerthe likelihood of referrals.

Hotel category and chain affiliation. These two variableswere defined to take on a value of one each time two hotelsbelonged to the same category (same chain) and a zerootherwise.

Analysis

In this study each variable is represented by a matrixdescribing different relationships between hotel managers.Unlike conventional data, dyads are not independent ofeach other, creating a problem for statistical analysis knownas autocorrelation. Because of the high interdependencebetween observations, standard OLS procedures are likelyto give biased estimates. To overcome this methodologicallimitation, statistical analysis with network data is usuallyperformed through Quadratic Assignment Procedure(QAP). Huber and Shultz introduced this method for bivari-ate cases (Huber & Shultz, 1976) and Krackhardt extendedit to multivariate cases showing that, when the error termsare correlated, QAP provides unbiased estimates while OLSresults in positively biased estimates (Krackhardt, 1987,1988). The QAP procedure is a non-parametric test basedon permutations of values. The algorithm first performs anordinary regression across corresponding cells of dependentand independent variables storing the obtained beta coeffi-cients. In a second step, the procedure defines many differ-ent matrices (10,000 in our case) randomly permuting rowsand columns of the dependent matrix; after each permuta-tion, a new regression is run and the corresponding coeffi-cients are stored. The measure of significance for the firstregression is given by the proportion of times random per-mutations of the dependent variable yield beta coefficientsequal or larger than the ones obtained with the actual val-ues6 and this proportion measures the estimates� level ofsignificance. A value of .05 would represent the number oftimes randomly obtained matrices yield beta coefficientsequal or larger than the regression performed with theactual values for the dependent variable. As our dependentvariable is dichotomous we used the netlogit QAP regressionas implemented in R.

Results

The hypotheses are examined through five different models:the first one introduces all the controls, the second addsstatus and trust measures, the third makes the distinctionbetween trust in embedded others and trust in non-embed-ded others, the fourth includes the variables describing theexchange of information using the baseline measure of

6 This assumes that the beta is positive; if the beta is negative,the algorithm will consider the number of times values are smallerthan the one first obtained.

Table 1 Correlation matrix (1 of 2).

Mean SD 1 2 3 4 5 6 7

1 Same category 0.387 0.4872 Chain affiliation 0.001 0.034 �0.0043 Physical distance (log) 0.348 0.198 �0.010 �0.0124 Services (squared differences) 49.496 64.320 �0.120* �0.014 �0.0175 Price (squared differences) 239.557 390.217 �0.150** �0.002 �0.028 0.389***

6 Rooms (j bigger than i) 0.514 0.500 0.008* �0.001 0.001 �0.006* �0.008**7 Category (j higher than i) 0.306 0.461 �0.528*** 0.002 0.005 0.105*** 0.079** 0.341***

8 Number of services (j greater than i) 2.839 4.079 �0.104*** �0.007 �0.007 0.480*** 0.189*** 0.330*** 0.440***

9 Average status 0.498 0.776 0.025 0.003 �0.010 0.079 0.232** �0.168** �0.238*10 Same status 0.051 0.500 �0.011 �0.012 0.017 �0.186* �0.409*** 0.009* 0.00611 Lower status (j lower than i) 2.057 5.833 �0.001 0.056* �0.026 0.178*** 0.358*** �0.144* �0.185***12 Higher status (j higher than i) 2.057 5.833 �0.001 0.056* �0.024 0.192** 0.356*** 0.153* 0.185***

13 Interpersonal trust 0.017 0.129 0.100 0.040 �0.027* �0.023 �0.003 0.046* 0.01814 Interpersonal trust to non-embedded others 0.001 0.121 0.014 0.046* 0.011 �0.011 �0.048* 0.050** 0.01715 Interpersonal trust to embedded others 0.016 0.124 �0.003 �0.004 �0.038 �0.014 0.044 0.000 0.00216 Exchange of information (give and not receive) 0.001 0.031 �0.012 �0.001 �0.013 �0.006 �0.009 �0.020 �0.02117 Exchange of information (receive and not give) 0.001 0.031 �0.012 �0.001 �0.013 �0.006 �0.009 0.030 0.034*

18 Exchange of information (reciprocal) 0.003 0.056 0.022 �0.002 �0.024 �0.007 �0.014 0.005 0.00719 Customer referrals 0.030 0.169 0.037* 0.028 �0.161*** �0.048** �0.027 �0.013 �0.01620 Hotel category 2.500 0.726 0.172** 0.000 0.009 �0.085 0.065 �0.282** �0.658**21 Price 54.764 10.869 0.112** 0.055* �0.004 0.070 0.365** �0.249** �0.395**22 Number of rooms 36.028 15.505 0.078* 0.000 �0.050* 0.051 0.224** �0.479** �0.248**23 Number of services 13.264 4.936 0.108** 0.029 �0.017 �0.037 0.189** �0.279** �0.376**

8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

�0.280***�0.088* �0.589***�0.227*** 0.552*** �0.356***0.417*** �0.078*** �0.362*** 0.126***

0.008 0.025 0.001 0.012 0.092*

0.014 �0.032 0.037* �0.098* �0.012 0.510***

�0.006 0.058* �0.035 0.108** 0.108** 0.544*** �0.444***�0.022 0.035 0.030* 0.044* �0.012 0.141** 0.103*** 0.0460.018 �0.006 0.030 �0.031 0.043* 0.093* 0.051* 0.046 �0.001**�0.003 0.037 0.055* 0.030 0.040* 0.338** �0.001 0.275*** �0.002** �0.002*�0.026** 0.03 �0.001 0.024 0.039* 0.227** 0.008 0.229*** 0.032** 0.032 0.310**

�0.395** 0.400** �0.249*** 0.265** �0.039** �0.011 �0.012 0.000 �0.012 0.004 0.011 0.000�0.367** 0.600** �0.390*** 0.629** �0.082** 0.027 �0.037 0.063* 0.029 0.010 0.069** 0.026 0.592**

�0.233** 0.235* �0.151* 0.306* �0.038* 0.012 �0.029 0.041 0.016 �0.001 0.036 0.022 0.343** 0.459**

�0.635** 0.517** �0.316*** 0.468** �0.061** 0.012 �0.025 0.037 0.034* 0.001 0.037* 0.027 0.575** 0.654** 0.426**

Correlations performed with 10,000 permutations.* p < .05.** p < .01.*** p < .001.

342M.Torto

riello

etal.

Table 2 Values in columns represent standardized beta coefficients.

Model 1 Model 2 Model 3 Model 4 Model 5

Same category 0.015164 0.011260 0.008488 0.007670 0.003908Chain affiliation 0.023500 0.011954 0.012202 0.013623 0.014216Physical distance (log) �0.160799*** �0.154895*** �0.152042*** �0.150686*** �0.148925***Services (squared differences) �0.063353** �0.046962* �0.047148 �0.04565* �0.04773Price (squared differences) �0.037846* �0.047985* �0.045916* �0.05063* �0.051186*Rooms (j bigger than i) �0.01389 �0.025446 �0.025088 �0.025112 �0.024083Category (j higher than i) �0.013502 �0.017197 �0.020391 �0.015093 �0.019709Number of services (j greater than i) 0.048541 0.004693 0.00832 �0.005652 �0.001612Average status 0.026859 0.023372 0.022728 0.016605 0.01559Same status hotels 0.034846Lower status (j lower than i) 0.009083 0.008121 0.026215 0.026049Higher status (j higher than i) 0.056543* 0.057366* 0.060811* 0.070327**

Interpersonal trust 0.214419*** 0.156786***

Interpersonal trust to non-embedded others 0.129437*** 0.097249***

Interpersonal trust to embedded others 0.274558*** 0.203791***

Exchange of information (give and n/receive) 0.007743 0.005575Exchange of information (receive and n/give) 0.017232 0.019166Exchange of information (reciprocal) 0.213043*** 0.198961***

Hotel category �0.056943 �0.054784 �0.057735 �0.046379 �0.049799Price 0.035607 0.036959 0.038796 0.026215 0.027866Number of rooms 0.003982 �0.00032 �0.000948 0.001065 0.000854Number of services 0.04153 0.012746 0.016693 0.000077 0.00468Adjusted R2 0.031 0.081 0.095 0.122 0.134

Regressions performed with 10,000 permutations.* p < 0.05.** p < 0.01.*** p < 0.001.

Cooperation among competitors as status-seeking behavior: Network ties and status differentiation 343

trust, and the fifth model considers the exchange of infor-mation variable along with the distinction between trustin embedded others and trust in non-embedded others.The dependent variable in all five models is always customerreferrals. Table 1 presents the results for correlation analy-sis across all variables. Five different regressions yield theresults for the 5 models presented in Table 2; in each col-umn we reported standardized beta coefficients to enhancecomparability across different variables. QAP regressionwas performed using 10,000 permutations to improve theestimates of the standard error and increase significance(Borgatti, Everett, & Freeman, 2002).

As reported in Model 1 and as we expected, physical dis-tance, number of services and price homogeneity are signif-icantly related with customer referrals; hotels with similarprice and service or close to each other share referrals.Interestingly though, chain affiliation is not significantly re-lated to customer referrals (p = .134), nor is hotel category(p = .262).

The second model adds status and interpersonal trustvariables. As expected there is no referral toward lower sta-tus hotels or same status hotels (p is respectively equal to.314 and .11), while higher status hotels do benefit fromcustomer overflows (p < 0.014). Interpersonal trust isstrongly and positively related to our dependent variable(p < 0.000) thus supporting Hypothesis 3a.

In the third model we make the distinction betweendyadic and embedded trust. As expected, both trust in

embedded others and trust in non-embedded others arestrongly and significantly related to customer referrals. Aspredicted in Hypothesis 3b, we observe that the standard-ized coefficient for embedded relationships (b̂ ¼ :274558)is significantly larger than the corresponding coefficientfor dyadic relationships (b̂ ¼ :129437).

In the fourth model, we introduce the information-exchange variables. The most important variable to predictreferrals is the reciprocally acknowledged exchange ofinformation, thus supporting Hypothesis 2. To furtherstrengthen the robustness of this result, Model 4 also showsthat none of the non-reciprocal information sharing vari-ables are significantly related with cooperation.

In Model 5, the final model, we tested all the variablesof interest. With the introduction of the information-exchange variables, the significance of trust in non-embedded others and trust in embedded others remainsthe same even though the value of their standardizedcoefficients is lower when compared with those reportedin Model 3. However, in Model 5 the standardized betacoefficient for trust in embedded others remains higherthan the standardized beta coefficient for the reciprocalexchange of information. In this last model, along withhigher status (p < 0.007), only price difference(p < 0.018) and physical distance (p < 0.000) are still sig-nificantly related with interorganizational cooperation.Going from the first to the fifth model, the adjusted-R2

increased from 0.031 to 0.134.

344 M. Tortoriello et al.

Discussion

In this paper we have tried to contribute to the ongoingdebate about structural embeddedness and cooperation be-tween competing organizations. Overall, we have shownthat perception of status, reciprocal and confirmed ex-change of information, and interpersonal trust, influencecooperative behaviors between hotel managers, over andabove the prediction of conventional non-social rationales.

This study advances our understanding of network struc-tures and embedded relationships in that it explicitly fo-cuses on the difference between social actors and theposition they occupy in the social structure as a possiblesource of network heterogeneity. In fact, while there is ageneral agreement about the fact that network resourcesare unevenly distributed within social systems, much lessis known about the causes of this heterogeneity. Here wepropose status differentiation as an important source ofnetwork heterogeneity and we assess its impact on embed-ded economic behavior. There are three main implicationsof this study.

The first implication concerns the qualitative differenti-ation of actors within a network. The tendency to view net-work structures as primarily conduits channeling resourcesfrom one individual to another, have led us to believe thatindividuals within a system of relationships are ‘‘sociallyundifferentiated’’. We often think about networks as flat,rather than hierarchical, structures in which all the nodesare at the same social level. However, this may not neces-sarily correspond with what we observe in the context ofeconomic exchange. In fact, even without changing levelsof analysis (e.g. without looking at individual attributes suchas, for instance, personality traits to differentiate the nodesin the network), it is still possible to draw meaningful dis-tinctions between different nodes based on the pattern ofrelationships defined among them. Focusing strictly on therelationships defined between any two nodes in the networkdoes not reveal that those very relationships not only areimportant for the two nodes directly involved in the trans-action, but also have an informational value for all theother nodes in the network (Podolny, 2001). Network ties al-low others to make inferences about the quality of one orboth of the nodes involved in the transaction. This meansthat social networks allow for the qualitative differentiationof all the actors involved in the network.

Our study also has implications for understanding thesources of embeddedness. While the relevance of social sta-tus has been convincingly established in ‘‘markets’’ such asinvestment banking (Podolny, 1993), wineries (Benjamin &Podolny, 1999), and venture capitalists (Podolny, 2001),we still have a limited understanding of how, the samemechanisms, affect producers� behaviors in socially embed-ded contexts. The ongoing structure of social relationships,if compared to the market-place, should provide an evenstronger basis for the development of status differentiation.The opportunity for repeated and meaningful social interac-tions characteristic of socially embedded contexts (personalties, trust, ‘‘thick’’ information exchange) contribute tothe widespread diffusion of fine-grained information abouteach other, thus facilitating the definition of status orderingamong the different nodes in the network. An important

implication of this argument is that status differentiationcould be a primary source of influence on individual embed-ded behavior along with more traditional social mechanismslike friendship, exchange of information, interpersonaltrust, etc.

A third implication of our study concerns the value ofstatus differentiation as a useful way of complementingtraditional approaches to studying dyadic (as opposed tostructural) relationships. As Coleman (1990, p. 130)suggested in his discussion of status and social exchangesystems ‘‘the awarder of status may never even havemet the recipient.’’ While there might be no direct con-tact between the awarder of status and the recipient(Coleman, 1990), individuals� do have status perceptionsabout each others and do strive to be associated with pos-itively perceived others (Kilduff & Krackard, 1994). AsPodolny argued, if markets can be seen as socially con-structed structures, the affiliation with other, high status,producers provides competitive leverage (Podolny, 1993).Focusing on status differentiation is also important for an-other reason. To date, there has been a growing agree-ment on the fact that market actors primarily seekstatus affiliation with like-status others. Previous researchshowed that status differentiation would result in the ef-fort of high status producers to develop affiliation withother high status actors (thus avoiding lower status part-ners). At the same time, we noted that lower status pro-ducers should attempt to develop relationships with higherstatus producers in an effort to enhance their position inthe overall status ordering. In our study we focused onthis second kind of status dynamic; accordingly, weshowed how the perception of social status differentiationcreated status seeking behavior through the pattern ofcustomer referrals. Lower status producer (hotel manag-ers) would try to develop an affiliation with higher statusones through referrals to benefit from this association inthe eyes of some relevant public (tourists in our case).In this example, higher status producers enjoy the rentderived from their privileged position in the status orderin the form of receiving referrals from lower status hotels.At the same time though, lower status hotels, through thereferral process, signal to the tourists their own quality ina way to be able to compete for them in the future.

While this paper has outlined a tentative way to enlargeresearch perspectives on embeddedness and cooperativebehaviors among competitors, there are several limitationsthat should be acknowledged. A first relevant issue, quitecommon with cross-sectional research designs in networkstudies, is one of causality. We claim that different socialinteractions shape cooperative behaviors, but the cross-sectional design of our study doesn�t allow us to specificallytest for the direction of causality. While we have based ourarguments on existing theory and previous empiricalfindings, causality could conceivably go in the oppositedirection or may be reciprocal and we currently have noway to rule out these alternative hypotheses. In the future,a relational dataset measured on the same nodes at differ-ent points in time would enhance the possibility to developa stronger causal argument.

Also a second issue concerns the generalizability of ourfindings. Clearly, collecting data about network relation-

Cooperation among competitors as status-seeking behavior: Network ties and status differentiation 345

ships among hotels located in a given geographical area,presents limitations concerning the external validity ofour findings. For instance, relationships among hotels lo-cated in major cities and/or belonging to large hotelchains might be driven by different logics. While it is hardto collect network data across different sites, Ingram andRoberts� study of the Sydney Hotel District did provide evi-dence for the importance of social ties (friendship) be-tween managers of hotels in a context quite differentfrom ours (Ingram & Roberts, 2000). Thus, the mecha-nisms through which hotel managers establish forms ofcooperation (e.g. client referrals, sharing of information,etc.) appears to extend to different contexts.

Another issue that deserves further attention is the inter-relation between physical proximity, social structure, andour findings. Indeed, the only control variable to remainstrongly and significantly associated with our dependentvariable across the five models presented is physical proxim-ity. This might be due to several reasons: hotels can be sim-ilar along some dimensions (Baum & Haveman, 1997);customers might express preferences about location or theymight not be willing to move far away from a particulararea; hotel managers might know more about hotels thatare located nearby, or physical proximity might just workas a short-cut in the managers� mind when they are facinga decision about referral. At the same time, social structureitself can be affected by physical proximity; for instance,managers whose hotels are located in the same area mayhave more opportunities for interactions than managers ofhotels located far apart and this in turn might shape theflows of information and the development of trust and/orcooperative relationships. While hard to disentangle thesemechanisms, it is important to notice that over and abovethe effects of hotel locations, the social network variablesconsidered in our study are strongly and significantly associ-ated with our dependent variable.

Lastly, while we characterized status differentiation interms of hotel managers� perception of other hotels stand-ings, it is also possible to consider other, more objectiveforms of differentiation among hotels which we could notconsider for lack of data. For instance the information pro-vided by guide books or independent third party sites, couldhelp shape individuals� perceptions of the differencesamong hotels independently of individual direct appraisal.However, as our theoretical focus was on the relationshipbetween status perceptions and cooperation among hotelmanagers, we deemed it appropriate to rely on hotel man-agers� judgment of other hotels.

Although there might be different approaches to thestudy of interorganizational cooperation among potentiallycompeting organizations, a social network approach to thestudy of status dynamics is a promising way to advanceour understanding of cooperative behaviors. In this paperwe have focused specifically on the development on clientreferrals among hotel managers as a form of cooperativebehaviors among competing organizations. We argued thatsince social status is a relevant resource for market ex-change (Podolny, 1993), individuals at competing organiza-tions engage in cooperative behaviors trading customersfor the benefits of being associated with higher statusorganizations. So far, the dominant view on status dynam-ics has highlighted the efforts undertaken by higher status

actors to maintain the status quo, without paying muchattention to the concomitant efforts of lower status actorsto change it. This study is a preliminary step in the direc-tion of understanding the asymmetric incentive systemsthat allows for cooperation between lower and higher sta-tus actors.

References

Baum, J. A. C., & Haveman, H. A. (1997). Love thy neighbor?Differentiation and agglomeration in the Manhattan hotelindustry. Administrative Science Quarterly, 42, 304–338.

Baum, J. A. C., & Ingram, P. (1998). Survival-enhancing learning inthe Manhattan hotel industry, 1898–1980. Management Science,44(7), 996–1016.

Baum, J. A. C., & Oliver, C. (1992). Institutional embeddedness andthe dynamics of organizational populations. American Journal ofSociology, 57(4), 540–559.

Benjamin, B. A., & Podolny, J. M. (1999). Status, quality, and socialorder in the California wine industry. Administrative ScienceQuarterly, 44, 563–589.

Bonacich, P. (1987). Power and centrality: A family of measures.American Journal of Sociology, 92(5), 1170–1182.

Borgatti, S. P., Everett, M. G., & Freeman, L. C. (2002). Ucinet forwindows: Software for social network analysis. Harvard: Ana-lytic Technologies.

Cialdini, R. B., Borden, R. J., Thorne, A., Walker, M. R., Freeman,S., & Sloan, L. R. (1976). Basking in reflected glory: Three(football) field studies. Journal of Personality and SocialPsychology, 34(3), 366–375.

Coleman, J. S. (1988). Social capital in the creation of humancapital. American Journal of Sociology, 94, 95–120.

Coleman, J. S. (1990). Foundations of social theory. Cambridge,Mass: Harvard University Press.

Gould, R. V. (2002). The origins of status hierarchies: A formaltheory and empirical test. American Journal of Sociology,107(5), 1143–1178.

Huber, L. J., & Shultz, J. (1976). Quadratic assignment as a generaldata analysis strategy. British Journal of Mathematical andStatistical Psychology, 29, 190–241.

Ingram, P., & Baum, J. A. C. (1997). Chain affiliation and the failureof Manhattan hotels, 1898–1980. Administrative Science Quar-terly, 42, 68–102.

Ingram, P., & Inman, C. (1996). Institutions, intergroup rivarly, andthe evolution of hotel populations around Niagara Falls. Admin-istrative Science Quarterly, 41, 629–658.

Ingram, P., & Roberts, P. W. (2000). Friendships among competitorsin the Sydney hotel industry. American Journal of Sociology,106(2), 387–423.

Kilduff, M., & Krackhardt, D. (1994). Bringing the individual backin: A structural analysis of the internal market for reputationin organizations. Academy of Management Journal, 37(1),87–108.

Krackhardt, D. (1987). QAP partialling as a test of spuriousness.Social Networks, 9, 171–186.

Krackhardt, D. (1988). Predicting with networks: Nonparametricmultiple regression analysis of dyadic data. Social Networks, 10,359–381.

Latour, B. (1987). Science in action. Cambridge, MA: HarvardUniversity Press.

Mayer, R. C., Davis, J. H., & Schoorman, D. F. (1995). An integrationmodel of organizational trust. Academy of Management Review,20(3), 709–727.

McEvily, B., & Tortoriello, M. (2011). Measuring trust in organiza-tional research: Review and recommendations. Journal of TrustResearch.

346 M. Tortoriello et al.

McEvily, B., Perrone, V., & Zaheer, A. (2003). Trust as an organizingprinciple. Organization Science, 14(1), 91–103.

Podolny, J. M. (1993). A status-based model of market competition.American Journal of Sociology, 98(4), 829–872.

Podolny, J. M. (1994). Market uncertainty and the social characterof economic exchange. Administrative Science Quarterly, 39,458–483.

Podolny, J. M. (2001). Networks as the pipes and prisms of themarket. American Journal of Sociology, 107(1), 33–60.

Roberts, P. W., & Khaire, M. (2009). Getting known by the companyyou keep: Publicizing the qualifications and former associationsof skilled employees. Industrial and Corporate Change, 18(1),77–106.

Rousseau, D. M., Sitkin, S. B., Burt, R. S., & Camerer, C. (1998). Notso different after all: A cross-discipline view of trust. Academyof Management Review, 23, 393–404.

Uzzi, B. (1997). Social structure and competition in interfirmnetworks: The paradox of embeddedness. Administrative Sci-ence Quarterly, 42, 35–67.

Zaheer, A., McEvily, B., & Perrone, V. (1998). Does trust matter?Exploring the effects of interorganizational and interpersonaltrust on performance. Organization Science, 9, 141–159.

MARCO TORTORIELLO is an assistant pro-fessor of strategic management at IESEBusiness School. He received his Ph.D. inorganizational behavior and theory from theTepper School of Business, Carnegie MellonUniversity. His research focuses on themechanisms through which social structureaffects the distribution and utilization ofknowledge within and across organizationsand how that relates to individual innovativecapabilities and performance. His research

has been published in Academy of Management Journal, Organi-

zation Science, and Journal of Trust Research. He is part of the

editorial board of the Academy of Management Review and Orga-nization Science.

VINCENZO PERRONE is a full professor oforganization theory at Bocconi Universityand director of the Institute of Organizationand Management Information Systems(IOSI). He has served on the editorial boardof Organization Science. From 1992 to 1994he taught and conducted research at theCarlson School of Management of the Uni-versity of Minnesota (USA), as a visitingprofessor. His research has been publishedin Organization Science, California Man-

agement Review, and International Journal of Purchasing and

Materials Management.

BILL MCEVILY is an associate professor ofstrategic management at the Rotman Schoolof Management, University of Toronto,where he teaches courses on strategyimplementation, social networks, and orga-nizational theory. His research exploressocial capital as an organizational and stra-tegic resource. Professor McEvily�s researchhas been published in Administrative Sci-ence Quarterly, Management Science,Organization Science, Organizational

Behavior and Human Decision Processes, and the Strategic Man-

agement Journal. He is a Senior Editor at Organization Science andhas served as guest editor for special issues of Management Scienceand Organization Science. Professor McEvily holds a B.S. fromClarkson University and a Ph.D. from the University of Minnesota.