mayer_et_al-2007.pdf

EDITOR’S FORUM

AN INTEGRATIVE MODEL OFORGANIZATIONAL TRUST: PAST, PRESENT,

AND FUTURE

F. DAVID SCHOORMANPurdue University

ROGER C. MAYERThe University of Akron

JAMES H. DAVISUniversity of Notre Dame

A considerable amount of research has examined trust since our 1995 publication. Werevisit some of the critical issues that we addressed and provide clarifications andextensions of the topics of levels of analysis, time, control systems, reciprocity, andmeasurement. We also recognize recent research in new areas of trust, such as affect,emotion, violation and repair, distrust, international and cross-cultural issues, andcontext-specific models, and we identify promising avenues for future research.

As we wrote our 1995 paper on trust (Mayer,Davis, & Schoorman, 1995), we were struck bythe relative scarcity of research in the main-stream management literature focusing directlyon trust. This led us to several bodies of litera-ture, including management, psychology, phi-losophy, and economics. We found that scholarsfrom diverse disciplines were presenting manyinsightful views and perspectives on trust butthat many of them seemed to talk past one an-other. Our goal was to integrate these perspec-tives into a single model.

This work came to fruition at about the sametime as several other works on trust. Papers ontrust by Hosmer (1995) and McAllister (1995) werealso published in Academy of Managementjournals that year, followed the next year by abook edited by Kramer and Tyler (1996). The con-fluence of these works, fueled by practical con-cerns raised by now infamous government andcorporate scandals over the next decade, pro-duced a groundswell of interest in understand-ing this basic and ubiquitous construct.

Since we were drawing perspectives frommultiple disciplines as inputs to the model, wewanted to provide a model that was generallyapplicable and would be used across multipledisciplines. We were gratified to find in a recent

search that our paper has been cited over 1,100times (according to Google Scholar). In additionto management and general business, it hasbeen cited in such diverse areas as marketing,accounting, finance, economics, informationsystems, industrial engineering, political sci-ence, communication, ethics, law, psychology,sociology, health care, and agribusiness. Wewould like to use this opportunity to revisit someof the issues raised by our 1995 paper and re-view how the field has dealt with them. We willalso discuss the new concerns and opportunitiesfor future research on trust.

CLARIFICATIONS AND EXTENSIONS OF THEMODEL OF TRUST

Trust As an Aspect of a Relationship

One of the difficult conceptual decisions thatwe faced as we developed our definition of trustwas to break with the widely accepted ap-proach, to that point, that trust was dispositionaland “trait-like” and to argue that trust was anaspect of relationships. That meant it variedwithin person and across relationships. Withsome exceptions (e.g., Driscoll, 1978; C. L. Scott,1980), the dominant conceptual and operational

� Academy of Management Review2007, Vol. 32, No. 2, 344–354.

Copyright of the Academy of Management, all rights reserved. Contents may notbe copied, emailed, posted to a listserv, or otherwise transmitted without thecopyright holder’s express written permission. Users may print, download, or emailarticles for individual use only.

344

definition of trust in the literature was Rotter’s(1967). We then went the next step and includedability as an antecedent of trust that allowed aparty’s trust to vary within a given trustee butacross domains. The dispositional aspects oftrust considered by Rotter are contained in theconstruct of propensity to trust in our model. Theliterature that has followed our model has notquestioned this decision and has accepted theview that trust is based in relationships.

Application Across Levels of Analysis

The importance of multilevel and cross-levelperspectives is gaining increasing attention inorganizational research. This has led to a callfor examining trust across levels of organization-al analysis (e.g., Rousseau, Sitkin, Burt, & Cam-erer, 1998). There is a need to understand trustboth within and between organizations becausemethodological difficulties can arise in the ab-sence of a clear multilevel conceptual model(Currall & Inkpen, 2002; Mossholder & Bedeian,1983; Rousseau, 1985).

We have heard from a number of scholars thatthe 1995 framework is fairly robust across levelsof analysis. A bit of history on the developmentof our theory may shed light on this issue. Earlydrafts of our paper developed our trust modelacross multiple levels of analysis. One of ourinitial goals was to develop a theory that wouldbe applicable across levels of analysis. We werecareful to develop constructs that would crosslevels of analysis, and we developed examplesof how cross-level applications of the modelwould work. Perhaps it was fortunate that earlyreviewers of our paper made the accurate obser-vation that the paper was very cumbersome(and long) because it developed the multilevelmodel. They recommended that we restrict ourpaper to a single level. The fact that our initialgoal was to develop a multilevel theory is prob-ably why the model works as well as it doesacross levels, but we do agree with those whoargue that one of the weaknesses in much of thecurrent trust research is that it is limited to re-lationships at a single level of analysis, consid-ering either dyadic trust relationships withinorganizations or trust between organizations.

Several authors have recognized differencesin trust for single referents at different hierar-chical levels within an organization (e.g., Cook& Wall, 1980; Driscoll, 1978; D. Scott, 1980). Recent

research points out that trust should be exam-ined at both the macro and micro levels withinan organization (McEvily, Perrone, & Zaheer,2003; Zaheer, McEvily, & Perrone, 1998). Whilethe need to understand trust has been noted inareas of study both within and between organi-zations, methodological difficulties can arise inthe absence of a clear multilevel conceptualmodel (Mossholder & Bedeian, 1983; Rousseau,1985).

Just as perceptions about an individual’s abil-ity, benevolence, and integrity will have an im-pact on how much trust the individual can gar-ner, these perceptions also affect the extent towhich an organization will be trusted. We de-fined each of these trustworthiness dimensionsso that it could be applied to interpersonal, in-tergroup, or interorganizational levels of analy-sis.

At higher levels of analysis, such as betweenorganizations, viewing the trustee in terms ofability and integrity seems to be well accepted.At macro levels of analysis, however, benevo-lence has received little attention. We definedbenevolence as the extent to which a party isbelieved to want to do good for the trustingparty, aside from an egocentric profit motive.Does the other company hold the focal compa-ny’s best interests as highly important? Whilewe may be able to identify situations, such assole proprietorships, where the owners havestrong bonds that display significant benevo-lence toward one another, the more traditionalmode is probably one wherein each company ismotivated primarily by its own financial inter-ests. If this is indeed the norm, benevolence isnot likely to be the most important factor in thedevelopment of interorganizational trust. How-ever, acts of benevolence (e.g., allowing bench-marking) from a potential partner in a joint ven-ture would help to build trust.

We contend that all three factors of ability,benevolence, and integrity can contribute totrust in a group or organization. Consider, forinstance, a supplier-buyer relationship. Thebuyer may believe that a supplier is able toprovide a quality product in a timely fashion.However, this only assures that the suppliercould perform. This does not mean that it willperform, and, therefore, the supplier will notnecessarily be trusted. The perception that thesupplier has integrity suggests that it will fulfillagreements as promised. Yet even if there is an

2007 345Schoorman, Mayer, and Davis

agreement, if the supplier’s ability to deliver isquestionable, it will not be trusted. If the sup-plier is perceived as benevolent, it will have astrong desire to serve this particular buyer’sneeds. If the supplier’s integrity is suspect be-cause, for instance, its track record with otherfirms is inconsistent with its stated policies,trust will again be lacking. As the perception ofeach of these factors increases, we would expectan increase in willingness to take a risk in therelationship.

The trust of either the dominant coalition orthe management team is critical to understand-ing organizational trust, since it is this level oftrust that will govern the strategic actions of theorganization (Cyert & March, 1963; Simon, 1957).As with individuals, we propose that some or-ganizations develop greater propensities totrust than do others. For organizations, thesepropensities develop from geographic, industry,and economic histories. A series of previous in-teractions with other organizations that resultedin, for example, lawsuits or monetary losseswould lower an organization’s propensity totrust. Conversely, a series of such experiencesas mutual benchmarking with various organiza-tions that significantly improved the qualityprocesses for an organization would increase itstrust propensity.

In summary, groups and organizations canboth garner trust from other parties and trustother parties. Our model was designed to under-stand the major factors that explain trust fromnot only the individual level but from the groupand organizational perspectives as well.

The Time Dimension

One of the issues explicit in our theory wasthat “time” would play an important role in themeaningfulness of the variables in the model.We noted that propensity, as a dispositionalquality, would be an important factor at the verybeginning of the relationship. We also notedthat judgments of ability and integrity wouldform relatively quickly in the course of the rela-tionship and that benevolence judgments wouldtake more time. Proposition 3 states that “theeffect of integrity on trust will be most salientearly in the relationship prior to the develop-ment of meaningful benevolence data” (1995:722); Proposition 4 states that “the effect of per-ceived benevolence on trust will increase over

time as the relationship between the parties de-velops” (1995: 722).

Despite these assertions, in many empiricalstudies researchers have raised questions aboutthe high observed correlation between benevo-lence and integrity and have questioned the in-dependence of these variables. In a discussionof several empirical studies, Schoorman (2002)observed that the findings as a whole were com-pletely consistent with the model. Those studiesconducted in laboratory settings were morelikely to show a high correlation between be-nevolence and integrity because the relation-ships had not had time to develop any real dataabout benevolence. In field samples where theparties had longer relationships, benevolenceand integrity were more likely to be separablefactors. We continue to find this pattern to beconsistent in our research. We think it would beinteresting for future research to establish morespecifically the process and time frames inwhich each of the variables contributes to trust.

Trust, Risk, and Control Systems

In our model we argued that trust would leadto risk taking in a relationship (see Proposition5). Perceived risk moderates the relationship be-tween trust and risk taking in our model. Trust isthe “willingness to take risk,” and the level oftrust is an indication of the amount of risk thatone is willing to take. Clearly, control systemsare an alternate mechanism for dealing withrisk in relationships. Recently, several scholarshave speculated about the relationship betweentrust and control systems in dealing with risk(McEvily et al., 2003; Sitkin & George, 2005).

Our views on this issue are developed furtherin Davis, Schoorman, and Donaldson (1997), inwhich we argue that one of the major distinc-tions between agency theory and stewardshiptheory is the use of trust versus control systemsto manage risk. However, we do not see thesemechanisms as being mutually exclusive. Onthe contrary, when the risk in a situation isgreater than the trust (and, thus, the willingnessto take risk), a control system can bridge thedifference by lowering the perceived risk to alevel that can be managed by trust. For exam-ple, in an organization that has a culture of“open book management” and transparency innumbers (a control system), the levels of per-ceived risk may be lower. There is a greater

346 AprilAcademy of Management Review

opportunity to empower employees by trustingthem to manage larger budgets and for employ-ees to trust the supervisor that performance-based compensation is fair.

However, there is an important caveat thatmust be noted. If there is a very strong system ofcontrols in an organization, it will inhibit thedevelopment of trust. Not only will there be fewsituations where there is any remaining per-ceived risk but trustworthy actions will be at-tributed to the existence of the control systemrather than to the trustee (cf. Strickland, 1958).Thus, a trustee’s actions that should be inter-preted as driven by benevolence or by integritymay be viewed simply as responses to the con-trol systems. The use of control systems is howagency theory proposes dealing with risk man-agement, and this does not foster the develop-ment of trust.

The Reciprocity of Trust

One of the limitations of our model that wenoted in the conclusion of our 1995 paper wasthat our conceptualization was unidirectional.We did not explore the reciprocity in trustingrelationships. This is a particularly salient issuein the area of leader-subordinate relationships,since the dominant view among leadership the-orists is that leader-member exchange (LMX) ismutual and reciprocal (Graen & Uhl-Bien, 1995;Liden, Wayne, & Stillwell, 1993). In an extensionof our model, Brower, Schoorman, and Tan (2000)argued that, unlike LMX, trust is not necessarilymutual and is not reciprocal. One of the impli-cations of this argument is that, in a relation-ship, A can trust B, but B may not trust A. This iscompletely consistent with the approach to trustand trust formation that we presented in ourmodel but is inconsistent with the views in theleadership literature. Empirical studies that ex-amine this reciprocal linkage of how one party’strust affects the other party’s trust in return,rather than assuming them to be equivalent, arerare (Serva, Fuller, & Mayer, 2005). We feel thatthis presents a fruitful area for future research.

Measurement of Trust As “Willingness to BeVulnerable”

While the theory of trust described in the 1995paper has been very influential in the develop-ment of trust research, the measurement of trust

has been a different story. We defined trust as awillingness to be vulnerable to another party.As such, suitable measurement of the constructnecessitates that questions be asked that assessthe extent to which a trustor is willing to volun-tarily take risks at the hands of the trustee.

We developed a short, four-item measure,with each of the items tapping into how willingthe trustor was to be vulnerable to the trustee.We (Schoorman, Mayer, & Davis, 1996a) foundthat veterinary doctors took bigger risks withthose employees they trusted more. The impactof trust went beyond that explained by the abil-ity, benevolence, and integrity of the trustee.Despite the measure’s brevity, we found its in-ternal consistency strong (Cronbach’s alpha �.82).

We (Davis, Schoorman, Mayer, & Tan, 2000)used the same measure in a restaurant settingto measure the trust that employees had in theirgeneral manager and found an alpha of .62. Thismeasure of trust in the leader significantly pre-dicted subsequent sales, profits, and employeeturnover in the restaurants (Davis, Mayer, &Schoorman, 1995; Davis et al., 2000). Based onthese results, we concluded that if trust in thegeneral manager could be developed and sus-tained, it would be a significant competitive ad-vantage to the firm, and the framework, includ-ing ability, benevolence, and integrity, meritedfurther consideration as an approach to build-ing trust in management.

Using this same measure, a quasi-experiment(Mayer & Davis, 1999) showed that trust in topmanagement was significantly improved byidentifying and replacing an invalid appraisalsystem. While the alpha in this study was lower(i.e., .59 and .60 in two waves of data), the test-retest reliability was quite strong, at .75 overfive months and .66 over nine months. Further-more, the quasi-experimental results were sig-nificant, even though the sizes of the groupsbeing compared were modest (i.e., twenty-twoand fifty-seven). In additional analyses wefound an average interitem correlation of r � .32.For each item we calculated a correlation be-tween it and a composite of the other threeitems. These correlations ranged from .23 to .49,with an average of .38.

These results fall well within Kline’s (1986)description of a measure of a complex constructthat has maximum validity. Kline noted thatsuch a measure would only have low internal


consistency reliability. Internal consistency es-timates suffer because of the brevity of this mea-sure. As Cortina says, “[Alpha] must be inter-preted with the number of items in mind” (1993:102). If the measure were truly unreliable, wewould not have been able to attain the signifi-cant results we have repeatedly found with itsuse (Nunnally, 1978). The results of these studiesall support the idea that while our four-itemmeasure of willingness to be vulnerable has attimes had lower than desired levels of Cron-bach’s alpha, it is a robust, stable, and validmeasure of the construct as we defined it.

More recent efforts to develop the measurehave yielded higher alpha levels. Mayer andGavin (2005) expanded the measure to ten items,yielding alphas of .82 and .76 for the plant man-ager and the top management team, respec-tively. Further analysis revealed two factors,however. One factor consisted of the originalmeasure with an additional item, whereas thesecond factor consisted of five items that gener-ally cited willingness to engage in specific be-haviors that would put a trustor at risk, such ascommunicating sensitive information to thetrustee. The five-item general willingness to bevulnerable scale yielded an alpha of .81 for thefocal plant manager and .72 for the top manage-ment team. Both of these were improvements of.06 and .07, respectively, over the four-item mea-sure. While results using the five-item scalewere reported in that study, the ten-item scaleyielded nearly identical results.

In what may be the most promising measureto date, Schoorman and Ballinger (2006) ex-panded the original measure to seven items,taking care to maintain the conceptual defini-tion but not create redundant items. It has pro-duced an alpha level of .84 in a sample of vet-erinary hospital employees. This measureappears in the Appendix.

Gillespie (2003) developed and validated ameasure of trust based on the willingness to bevulnerable definition. This ten-item BehavioralTrust Inventory has good psychometric proper-ties and shows promise for future researchbased on this conceptual definition.

The upshot of this discussion is that measur-ing trust as we defined it involves asking ques-tions that measure a trustor’s willingness to bevulnerable. We are aware from numerous per-sonal communications from other researchersthat many of them are concerned with the lower

than desirable alpha levels of our original mea-sure. While we think that a combination of itsconceptual clarity, test-retest reliability, and re-lationship with other variables in the nomolog-ical net across a number of studies is just asimportant a consideration as the alpha level of afour-item measure, we recognize that develop-ment of the more recent longer measures withhigher alpha levels was warranted. The mea-sures we have developed more recently appearto have overcome concerns with Cronbach al-pha levels. We believe that as these measuresbecome more widely available, more research-ers will choose to measure trust as defined inour 1995 paper.

On a related note, work on trust would befacilitated by further development of measuresof propensity. Many researchers have found Rot-ter’s (1967) twenty-five-item measure too long toinclude as a variable in studies with many othervariables and are concerned about its multidi-mensionality. Our adaptation (Mayer & Davis,1999; Schoorman et al., 1996a), while muchshorter and unidimensional, has not consis-tently produced high Cronbach’s alphas (e.g., .55and .66 in Mayer & Davis, 1999). We are notaware of any brief, unidimensional publishedmeasure of propensity that produces consis-tently high alpha levels. Development of such ameasure might enable finding more relation-ships between propensity and other variables ofinterest, particularly early in the development ofa relationship.

NEW DIMENSIONS IN THE RESEARCH ONTRUST

Affect, Emotion, and the Impact on Trust

The basis of our model was to understand howparties process information about others,thereby deciding how much risk to take withthose others. Perceptions of others and percep-tions of risk inherent in the behaviors beingconsidered must be processed in order to cometo decisions about taking risks. For instance, asone evaluates a trustee’s ability in the domainof interest or considers relevant inputs about atrustee’s integrity, one is thinking. As such, ourmodel represents a cognitive approach to trust.More recent work has pointed to the fact thattrust also involves emotion. Williams (2001) haspointed out that affective responses influence


how people evaluate their level of trust in an-other party. Similarly, Jones and George (1998)have argued that emotions and moods providepeople with information on how they are expe-riencing trust. Dunn and Schweitzer (2005) foundthat emotional states—even unrelated to thetrustee or the situation—have an effect on trust.Weber, Malhotra, and Murnighan (2005) haveshown that emotional attachments can cause atrustor to take a sudden risk not warranted bythe available evidence.

Proponents of the strictly cognitive approachto decision making about trust would argue thatwhile emotions may create a temporary “irra-tionality” about the data on ability, benevo-lence, and integrity, after a period of time theperception would return to a rational perspec-tive. Nonetheless, it appears to be clear thatemotions do influence the perception of the an-tecedents of trust and, therefore, the trust in re-lationships. It is also likely that this emotiondoes dissipate over time after a violation oftrust. What is not clear is whether it ever com-pletely dissipates and returns to a nonemotionalevaluation. Alternatively, while emotions arebeing experienced, they may lead the trustor toupdate prior perceptions of the trustworthinessdimensions and trust such that even after theemotions dissipate, the effect on the cognitiveevaluations remains. We think the role of emo-tions is a very interesting area of research andwill add a new dimension to the model.

Violation and Trust Repair

Some exemplary work has been done to un-derstand trust violation and repair (e.g., Lewicki& Bunker, 1996; Morrison & Robinson, 1997; Rob-inson, 1996), but this topic warrants more re-search. As part of their contributions, these au-thors highlight the idea that violation of trust islikely to be an emotional event for the trustor. Atthe time of this writing, a special issue of theAcademy of Management Review (edited byDirks, Lewicki, and Zaheer) focused on trust re-pair is in preparation. This is a very appropriatetopic and promises to add valuable insight intothe process by which trust development canmove forward after trust has been damaged.

We believe that in order to repair trust, it iscritical to first understand how it was damagedin the first place, since different means of dam-aging trust are likely to require different repar-

ative responses. Dating back to the 1950s (e.g.,Kruglanski, 1970; Strickland, 1958), some workhas been done that begins to lay the ground-work for how attributions are made when trust isbroken. We suggest that which trustworthinessfactor is damaged and how it was damagedinfluence not only how repairable the damage islikely to be but how effective various repairstrategies are likely to be. Repair attemptsmight be looked at either as mediating pro-cesses within the feedback arrow in our modelor as influencing the trustworthiness factors di-rectly (e.g., receiving information about trust-worthiness factors outside the process of havingtaken a risk that went awry). Further theoreticalwork in this area to understand the conditionsunder which various repair strategies are effec-tive would help in developing prescriptive road-maps for repairing broken trust.

Forgiveness is another evolving area thatholds promise for understanding trust repair af-ter a violation. An insightful paper by Aquino,Grover, Goldman, and Folger (2003) provides animportant separation between resolving nega-tive emotions (i.e., forgiveness itself) and behav-iors that follow that restore the relationship.More work in this area is needed to determinesuch basic issues as the conditions under whichforgiveness enhances trust repair after a viola-tion, what conditions increase the likelihoodthat an offended party will forgive the violator,and the role that forgiveness plays in the trustrepair process.

The Concept of Distrust

There has been considerable discussion in theliterature about the concept of distrust, as wellas the relationship between trust and distrust.Lewicki, McAllister, and Bies (1998) framed thedebate in the organizational literature, arguingthat trust and distrust are separate dimensionsand not the opposite ends of a single continuum.They noted that two-factor models of satisfac-tion and dissatisfaction had been proposed be-fore (Herzberg, Mausner, & Snyderman, 1967),and that the contemporary evidence from stud-ies of positive and negative affectivity supportsthis view. Their main reason for suggesting suchan approach to the trust-distrust distinction wasbecause “relationships are multifaceted or mul-tiplex” (1998: 442), and we need a model that


allows for both trust and distrust to exist in thesame relationship.

In our model we chose to take the opposite(and more traditional) view that trust and dis-trust are the opposite ends of the same contin-uum. This is consistent with dictionary defini-tions—for example, Webster’s defines distrustas “the lack or absence of trust” and RandomHouse as “to have no trust in.” In sociology,Ross, Mirowski, and Pribesh define “mistrust” asthe “absence of faith in other people” (2001: 568).Luhmann argues that distrust is a “functionalequivalent of trust” (1979: 71). Our definition oftrust—willingness to take risk (i.e., be vulnera-ble) in a relationship—means that at the lowestlevel of trust, one would take no risks at all. Wefelt that the complete lack of trust and distrustare the same thing.

In our model of trust, however, we argued thatability is an important antecedent of trust, alongwith benevolence and integrity. This was a de-viation from an emerging view that trust wasmore affective. This point is further clarified inSchoorman, Mayer, and Davis (1996b). The im-portant implication of the addition of ability tothe antecedents of trust is that it creates aframework of trust that is domain specific. Thisis noted in our original article (1995: 717) andfollows from the work of Zand (1972). Trust beingdomain specific allows for the multifaceted andmultiplex relationships about which Lewicki etal. (1998) raised concerns. For example, it may beappropriate to trust a colleague to do a good jobcollaborating on a research project but to nottrust him/her to do a good job teaching yourclass in your absence. The difference in the levelof trust within the same relationship is a func-tion of the different abilities across different do-mains. The skills required to present and inter-act effectively in class differ from thosenecessary to do research.

Lewicki et al. (1998) produced a chart with ahigh trust and high distrust condition in whichone would presumably “trust, but verify.” Wefeel this is not a reasonable argument withindomain. If you trust a partner, you do not need toverify. Doing so would be the clearest indicationthat you do not trust. We do agree that you mighttrust your colleague to produce a literature re-view but may need to verify his/her ability todeliver in the classroom by reviewing his/herlecture notes and presentation. Our model says

that you don’t need different constructs to ac-count for this.

McKnight and Chervany (2001) produced anexcellent summary of the literature on defini-tions of trust and distrust and the models thatdescribe each of the constructs. They reviewedthe literature on trust and distrust and devel-oped separate conceptual models (antecedentand contextual variables) for each construct.The resulting models are identical for both trustand distrust, which suggests to us that perhapswe do not need both models. In fact, these au-thors conclude that “most trust theorists agreethat trust and distrust are separate constructsthat are opposites of each other” (2001: 42). Wewould simply add that if they are opposites ofeach other, there is little added value to treatingthem as separate constructs.

A review of empirical work on the conceptual-ization of distrust in the literature produced sur-prising results. Some who argue that trust anddistrust are different dimensions only study oneof the constructs at a time, making it difficult todevelop data on the differences. It is particularlyinteresting that some researchers who study theconcept of distrust (e.g. McAllister, Pang, Tan, &Ruan, 2006) have used our measure of trust as awillingness to be vulnerable (Mayer & Davis,1999; Schoorman et al., 1996b) and reverse-scored it to represent their measure of distrust.In sum, we can find no credible evidence that aconcept of distrust that is conceptually differentfrom trust is theoretically or empirically viable.

International and Cross-Cultural Implicationsfor Trust

Over the same time frame in which interest intrust accelerated, there was a significant in-crease in interest in studying cross-national andcross-cultural differences. It is therefore not sur-prising that much of the explosion of interest intrust research has come from around the globe.Over 20 percent of the 1,100 studies listed inGoogle Scholar that cite our paper were writtenin a language other than English. The WorldEconomic Forum is made up of the world’s lead-ing politicians and business leaders. It meetsannually in Davos, Switzerland to discuss awide range of international issues. It has beenmonitoring public trust levels since 2003 througha biannual global public opinion poll conductedby GlobeScan Incorporated. The latest findings


from the poll show that trust in a range of insti-tutions has dropped significantly since January2004 to levels not seen since the months follow-ing the September 11 terrorist attacks in 2001(www.weforum.org). GlobeScan reports that theoverall trust level for global companies is thelowest since the tracking began. While we arenot sure what definition of trust was used in thispoll, these results present obvious reasons forconcern.

The recent GLOBE project by House and col-leagues (House, Hanges, Javidan, Dorfman, &Gupta, 2004; House, Javidan, Hanges, & Dorf-man, 2002) measured the cultural orientation ofsixty-two societies around the world basedlargely on the cultural dimensions identified byHofstede (1980). These scholars used nine di-mensions of culture, as well as twenty-one lead-ership dimensions. All of this research has led tothe inevitable question of how trust is differentacross cultures (Den Hartog, 2004; Wasti, Tan,Brower, & Onder, in press).

We believe that one of the ways in whichculture affects trust is through the propensityvariable. We have proposed that the anteced-ents of propensity include personality, experi-ences, and culture. There is evidence in the cul-ture literature that initial trust of strangersvaries across cultures. One of the dimensions ofculture that is most relevant to this issue is thetask versus relationship orientation of a culture.Task-oriented cultures seem to have a higherinitial trust of strangers and therefore a higherpropensity, while relationship-oriented culturesneed time to develop a relationship prior toworking on the task. The cultural variable ofuncertainty avoidance is well-established as apredictor of predispositions to take risk or berisk averse (Sully de Luque & Javidan, 2004). Iftrust is the willingness to take risk in a relation-ship, how does uncertainty avoidance as a dis-positional quality affect the development oftrust? We think there is considerable work thatneeds to be done in fine-tuning what we knowabout the influence of culture on the propensityto trust.

Culture can also affect the perception of abil-ity, benevolence, and integrity and the impor-tance given to each of these variables in themodel. More action-oriented, competitive, per-formance-oriented cultures—what Hofstede hascalled “masculine” cultures—tend to place ahigher value on the ability variable. More col-

laborative, being-oriented, “feminine” culturestend to put more emphasis on the benevolencevariable. While these are broad generalizationsof relationships between culture and trust, theyillustrate the potential value of future researchto develop these links more carefully.

Context-Specific Models of Trust

It was our intention in developing the modelto be as parsimonious as possible and to de-velop a model that would be generalizable tothe broadest number of contexts. In order toachieve this, we neglected many specific con-text variables that would be relevant to a morerestricted trust domain. We think it would beappropriate to specify contextual variables forthe model that are unique to studying trustwithin a particular context. For example, muchof the research on trust in organizations hasfocused on the relationship between supervisorsand subordinates. In this context, the hierarchi-cal power difference and the asymmetry of in-formation that exist between the two individu-als in the trusting relationship have someimportant implications for how trust might de-velop. If the supervisor has more access to infor-mation about the subordinate and can initiateopportunities to gather information about abil-ity, benevolence, and integrity, and if these op-portunities are not available to the subordinate,we would expect that the supervisor’s trust inthe subordinate would develop more quicklythan vice versa.

Additionally, since risk taking in the relation-ship is caused by an interaction between trustand risk, one’s perceptions of risk in the actionbeing contemplated (which our model separatesfrom the trustee in question) will affect risk-taking actions. Ceteris paribus, the party whohas more power in the relationship will likelyperceive— by virtue of that power—less riskand, thus, will engage in more risk-taking ac-tions. This would give the appearance that thisparty’s trust is higher still. In the context of trustbetween peers, there is likely to be a differentset of variables that predict the developmentand use of trust. We expect that studies in par-ticular contexts will develop additional vari-ables that help better explain the antecedentsand consequences of trust.


SUMMARY AND CONCLUSIONS

In this paper we have taken the opportunity toreflect on our thinking as we developed the 1995piece, “An Integrative Model of OrganizationalTrust,” and the knowledge context in which itoccurred. This paper has given us an opportu-nity to revisit some of the decisions that wemade and to evaluate them in light of the re-search that followed. A part of our charge inwriting this paper was to indicate how our ownresearch in this area has progressed and howwe view the growing body of literature on thetopic.

The levels-of-analysis issue continues to bean interesting topic of discussion, and more ex-plicit extensions to group and organizationallevels are warranted. We note that time issuesaddressed in our paper have not received theattention that we expected. Research attemptsto study trust in very short laboratory simula-tions have yielded mixed but not unexpectedresults. By including a consideration of time,studies of trust should lead to more predictableresults. The interplay of trust, risk, and controlsystems continues to be a much debated topic.We have attempted to clarify our thinking onthese dynamics. We see trust and control sys-tems as alternate and sometimes compatiblemeans for managing risk. We have extended ourthinking about the reciprocity of trust to explic-itly recognize the notion that, unlike relationalleadership constructs (e.g., LMX), trust is not mu-tual and not necessarily reciprocal. The mea-sures of trust that we had developed based onour definition of trust have produced mixed re-sults in terms of internal consistency reliabilityestimates. Here we have explored some of thedevelopments and advances in the measure-ment area and have noted the need for contin-ued improvement.

We also have reviewed some of the interest-ing new directions in the research on trust.Prominent among these is the inclusion of therole of affect and emotion, trust violations, andrepair. We believe these constructs will add newdimensions to the model of trust and provide forvaluable research in the future. Another areaseeing rapid growth in interest is the role thatinternational and cross-cultural dimensionsplay in the model of trust. We see the greatestopportunities in the development of the conceptof propensity across cultures, as well as for the

relative importance of ability, benevolence, andintegrity across cultures. Finally, we note thatwhile we deliberately focused on a model thatwas maximally generalizable and parsimoni-ous, it is appropriate to now examine context-specific variables that might add to the model oftrust. One such area that our research has takenus into is the specific relationship between su-pervisor and subordinate in the workplace. Is-sues of power and information asymmetry makethis relationship and the trust it produces some-what unique.

While a great deal of research has occurred inthe area of trust over the past decade, the newresearch only suggests that there is a lot more tobe done and many very promising avenues topursue. We hope that our comments in this pa-per can be a catalyst for some of this research.

APPENDIX

Trust Items from Schoorman and Ballinger(2006)

My supervisor keeps my interests in mind whenmaking decisions.I would be willing to let my supervisor have completecontrol over my future in this company.If my supervisor asked why a problem occurred, Iwould speak freely even if I were partly to blame.I feel comfortable being creative because mysupervisor understands that sometimes creativesolutions do not work.It is important for me to have a good way to keep aneye on my supervisor.Increasing my vulnerability to criticism by mysupervisor would be a mistake.If I had my way, I wouldn’t let my supervisor haveany influence over decisions that are important to me.

Response Scale

1 2 3 4 5

stronglydisagree

somewhatdisagree

neitheragree

nordisagree

somewhatagree

stronglyagree

REFERENCES

Aquino, K., Grover, S. L., Goldman, B., & Folger, R. 2003. Whenpush doesn’t come to shove: Interpersonal forgivenessin workplace relationships. Journal of Management In-quiry, 12: 209–216.


Brower, H. H., Schoorman, F. D., & Tan, H. H. 2000. A model ofrelational leadership: The integration of trust and lead-er-member exchange. Leadership Quarterly, 11: 227–250.

Cook, J., & Wall, T. 1980. New work attitude measures of trust,organizational commitment and personal need nonful-fillment. Journal of Occupational Psychology, 53: 39–52.

Cortina, J. 1993. What is coefficient alpha? Journal of AppliedPsychology, 78: 98–104.

Currall, S. C., & Inkpen, A. C. 2002. A multilevel approach totrust in joint ventures. Journal of International BusinessStudies, 33: 479–495.

Cyert, R. M., & March, J. G. 1963. A behavioral theory of thefirm. Englewood Cliffs, NJ: Prentice-Hall.

Davis, J. H., Mayer, R. C., & Schoorman, F. D. 1995. The trustedgeneral manager and firm performance: A strategic ad-vantage. Paper presented at the Annual InternationalStrategic Management Society Conference, Mexico City.

Davis, J. H., Schoorman, F. D., & Donaldson, L. 1997. Towarda stewardship theory of management. Academy of Man-agement Review, 22: 20–47.

Davis, J. H., Schoorman, F. D., Mayer, R. C., & Tan, H. H. 2000.The trusted general manager and business unit perfor-mance: Empirical evidence of a competitive advantage.Strategic Management Journal, 21: 563–576.

Den Hartog, D. N. 2004. Assertiveness. In R. J. House, P. J.Hanges, M. Javidan, P. Dorfman, & V. Gupta (Eds.), Cul-ture, leadership and organizations: The GLOBE study of62 societies: 395–431. Thousand Oaks, CA: Sage.

Driscoll, J. W. 1978. Trust and participation in organizationaldecision making as predictors of satisfaction. Academyof Management Journal, 21: 44–56.

Dunn, J. R., & Schweitzer, M. E. 2005. Feeling and believing:The influence of emotion on trust. Journal of AppliedPsychology, 88: 736–748.

Gillespie, N. 2003. Measuring trust in work relationships: TheBehavioral Trust Inventory. Paper presented at the an-nual meeting of the Academy of Management, Seattle.

Graen, G. B., & Uhl-Bien, M. 1995. Relationship-based ap-proach to leadership: Development of leader-memberexchange (LMX) theory of leadership over 25 years: Ap-plying a multi-level multi-domain perspective. Leader-ship Quarterly, 6: 219–247.

Herzberg, F., Mausner, B., & Snyderman, B. B. 1967. The mo-tivation to work (2nd ed.). New York: Wiley.

Hofstede, G. 1980. Culture’s consequences. Beverly Hills, CA:Sage.

Hosmer, L. T. 1995. Trust: The connecting link between orga-nizational theory and philosophical ethics. Academy ofManagement Review, 20: 379–403.

House, R., Javidan, M., Hanges, P., & Dorfman, P. 2002. Un-derstanding cultures and implicit leadership theoriesacross the globe: An introduction to project GLOBE. Jour-nal of World Business, 37: 3–10.

House, R. J., Hanges, P. J., Javidan, M., Dorfman, P., & Gupta,V. 2004. Culture, leadership and organizations: TheGLOBE study of 62 societies. Thousand Oaks, CA: Sage.

Jones, G. R., & George, J. M. 1998. The experience and evolu-tion of trust: Implications for cooperation and teamwork.Academy of Management Review, 23: 531–546.

Kline, P. 1986. A handbook of test construction: Introductionto psychometric design. London: Methuen.

Kramer, R. M., & Tyler, T. R. (Eds.). 1996. Trust in organiza-tions: Frontiers of theory and research. Thousand Oaks,CA: Sage.

Kruglanski, A. W. 1970. Attributing trustworthiness in super-visor-worker relations. Journal of Experimental Psychol-ogy, 6: 214–232.

Lewicki, R. J., & Bunker, B. B. 1996. Developing and maintain-ing trust in work relationships. In R. M. Kramer & T. R.Tyler (Eds.), Trust in organizations: Frontiers of theoryand research: 114–139. Thousand Oaks, CA: Sage.

Lewicki, R. J., McAllister, D. J., & Bies, R. J. 1998. Trust anddistrust: New relationships and realities. Academy ofManagement Review, 23: 438–458.

Liden, R. C., Wayne, S. J., & Stillwell, D. 1993. A longitudinalstudy on the early development of leader-member ex-changes. Journal of Applied Psychology, 78: 662–674.

Luhmann, N. 1979. Trust and power. New York: Wiley.

Mayer, R. C., & Davis, J. H. 1999. The effect of the performanceappraisal system on trust for management: A field qua-si-experiment. Journal of Applied Psychology, 84: 123–136.

Mayer, R. C., Davis, J. H., & Schoorman, F. D. 1995. An inte-grative model of organizational trust. Academy of Man-agement Review, 20: 709–734.

Mayer, R. C., & Gavin, M. B. 2005. Trust in management andperformance: Who minds the shop while the employeeswatch the boss? Academy of Management Journal, 48:874–888.

McAllister, D. J. 1995. Affect- and cognition-based trust asfoundations for interpersonal cooperation in organiza-tions. Academy of Management Journal, 38: 24–59.

McAllister, D. J., Pang, K., Tan, H. H., & Ruan, Y. 2006. Socialdynamics of paranoia and distrust in teams. Paper pre-sented at the annual meeting of the Academy of Man-agement, Atlanta.

McEvily, B., Perrone, V., & Zaheer, A. 2003. Trust as an orga-nizing principle. Organization Science, 14: 91–103.

McKnight, D. H., & Chervany, N. L. 2001. Trust and distrustdefinitions: One bite at a time. In R. Falcone, M. Singh, &Y.-H. Tan (Eds.), Trust in cyber-societies: 27–54. Berlin &Heidelberg: Springer-Verlag.

Morrison, E. W., & Robinson, S. L. 1997. When employees feelbetrayed: A model of how psychological contract viola-tion develops. Academy of Management Review, 22:226–256.

Mossholder, K. W., & Bedeian, A. G. 1983. Cross-level infer-ence and organizational research: Perspectives on inter-pretation and application. Academy of Management Re-view, 8: 547–558.

Nunnally, J. C. 1978. Psychometric theory. New York:McGraw-Hill.


Robinson, S. L. 1996. Trust and breach of the psychologicalcontract. Administrative Science Quarterly, 41: 574–599.

Ross, C. E., Mirowsky, J., & Pribesh, S. 2001. Powerlessnessand the amplification of threat: Neighborhood disadvan-tage, disorder and mistrust. American Sociological Re-view, 66: 568–591.

Rotter, J. B. 1967. A new scale for the measurement of inter-personal trust. Journal of Personality, 35: 651–665.

Rousseau, D. M. 1985. Issues of level in organizational re-search: Multi-level and cross-level perspectives. Re-search in Organizational Behavior, 7: 1–37.

Rousseau, D. M., Sitkin, S. B., Burt, R. S., & Camerer, C. 1998.Not so different after all: A cross-discipline view of trust.Academy of Management Review, 23: 393–404.

Schoorman, F. D. 2002. Discussant at symposium titled Inte-grating trust perspectives: Foundations for a revised in-tegrative model of organizational trust. Annual meetingof the Academy of Management, Denver.

Schoorman, F. D., & Ballinger, G. A. 2006. Leadership, trustand client service in veterinary hospitals. Working pa-per, Purdue University, West Lafayette, IN.

Schoorman, F. D., Mayer, R. C., & Davis, J. H. 1996a. Empow-erment in veterinary clinics: The role of trust in delega-tion. Presented in a symposium on trust at the 11thAnnual Conference, Society for Industrial and Organi-zational Psychology (SIOP), San Diego.

Schoorman, F. D., Mayer, R. C., & Davis, J. H. 1996b. Organi-zational trust: Philosophical perspectives and concep-tual definitions. Academy of Management Review, 21:337–340.

Scott, C. L., III. 1980. Interpersonal trust: A comparison ofattitudinal and situational factors. Human Relations, 33:805–812.

Scott, D. 1980. The causal relationship between trust and theassessed value of management by objectives. Journal ofManagement, 6: 157–175.

Serva, M. A., Fuller, M. A., & Mayer, R. C. 2005. The reciprocalnature of trust: A longitudinal study of interactingteams. Journal of Organizational Behavior, 26: 625–648.

Simon, H. 1957. Administrative behavior. New York: Macmil-lan.

Sitkin, S. B., & George, E. 2005. Managerial trust-buildingthrough the use of legitimating formal and informalcontrol mechanisms. International Sociology, 20: 307–338.

Strickland, L. H. 1958. Surveillance and trust. Journal of Per-sonality, 24: 200–215.

Sully de Luque, M., & Javidan, M. 2004. Uncertainty avoid-ance. In R. J. House, P. J. Hanges, M. Javidan, P. Dorfman,& V. Gupta (Eds.), Culture, leadership and organizations:The GLOBE study of 62 societies: 603–644. ThousandOaks, CA: Sage.

Wasti, S. A., Tan, H. H., Brower, H. H., & Onder, C. In press.Cross-cultural measurement of supervisor trustworthi-ness: An assessment of measurement invariance acrossthree cultures. Leadership Quarterly.

Weber, J. M., Malhotra, D., & Murnighan, J. K. 2005. Normalacts of irrational trust: Motivated attributions and thetrust development process. Research in OrganizationalBehavior, 26: 75–101.

Weiner, B. 1986. An attributional model of motivation andemotion. New York: Springer-Verlag.

Williams, M. 2001. In whom we trust: Group membership asan affective context for trust development. Academy ofManagement Review, 26: 377–396.

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

Zand, D. E. 1972. Trust and managerial problem solving.Administrative Science Quarterly, 17: 229–239.

F. David Schoorman ([email protected]) is a professor of management in the Kran-nert Graduate School of Management at Purdue University. He earned his Ph.D. fromCarnegie Mellon University. He has published research in the areas of trust, leader-ship, decision making, and organizational effectiveness.

Roger C. Mayer ([email protected]) is professor of management at The Universityof Akron. He earned his Ph.D. from Purdue University. His research includes employeemotivation, decision making, organizational effectiveness, and social capital.

James H. Davis ([email protected]) is an associate professor, the Siegfried Director ofEntrepreneurship, and the John F. O’Shaughnessy Professor of Family Enterprise atthe University of Notre Dame. He earned his Ph.D. from the University of Iowa. Hisresearch interests include trust, corporate governance, family business, stewardshiptheory, and social capital.


mayer_et_al-2007.pdf

Documents