Enculturation Trajectories and Individual Attainment: An Interactional ...

This article was downloaded by: [171.67.216.23] On: 04 March 2017, At: 15:04Publisher: Institute for Operations Research and the Management Sciences (INFORMS)INFORMS is located in Maryland, USA

Management Science

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

Enculturation Trajectories: Language, CulturalAdaptation, and Individual Outcomes in OrganizationsSameer B. Srivastava, Amir Goldberg, V. Govind Manian, Christopher Potts

To cite this article:Sameer B. Srivastava, Amir Goldberg, V. Govind Manian, Christopher Potts (2017) Enculturation Trajectories: Language,Cultural Adaptation, and Individual Outcomes in Organizations. Management Science

Published online in Articles in Advance 02 Mar 2017

. http://dx.doi.org/10.1287/mnsc.2016.2671

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

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

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

Copyright © 2017, INFORMS

Please scroll down for article—it is on subsequent pages

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

http://pubsonline.informs.org

http://dx.doi.org/10.1287/mnsc.2016.2671

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

http://www.informs.org

MANAGEMENT SCIENCEArticles in Advance, pp. 1–17

http://pubsonline.informs.org/journal/mnsc/ ISSN 0025-1909 (print), ISSN 1526-5501 (online)

Enculturation Trajectories: Language, Cultural Adaptation, andIndividual Outcomes in OrganizationsSameer B. Srivastava,a Amir Goldberg,b V. Govind Manian,b Christopher Pottsc

aHaas School of Business, University of California, Berkeley, Berkeley, California 94720; b Stanford Graduate School of Business,Stanford, California 94305; cDepartment of Linguistics, Stanford University, Stanford, California 94305Contact: [email protected] (SBS); [email protected] (AG); [email protected] (VGM); [email protected] (CP)

Received: December 8, 2015Accepted: September 24, 2016Published Online in Articles in Advance:March 2, 2017

https://doi.org/10.1287/mnsc.2016.2671

Copyright: © 2017 INFORMS

Abstract. How do people adapt to organizational culture, and what are the consequencesfor their outcomes in the organization? These fundamental questions about culture havepreviously been examined using self-report measures, which are subject to reporting bias,rely on coarse cultural categories defined by researchers, and provide only static snap-shots of cultural fit. By contrast, we develop an interactional language use model thatovercomes these limitations and opens new avenues for theoretical development aboutthe dynamics of organizational culture. We trace the enculturation trajectories of employ-ees in a midsized technology firm based on analyses of 10.24 million internal emails.Our language-based model of changing cultural fit (1) predicts individual attainment;(2) reveals distinct patterns of adaptation for employees who exit voluntarily, exit involun-tarily, and remain employed; (3) demonstrates that rapid early cultural adaptation reducesthe risk of involuntary, but not voluntary, exit; and (4) finds that a decline in cultural fitfor individuals who had successfully enculturated portends voluntary departure.

History: Accepted by Olav Sorenson, organizations.Funding: This work has been supported by the National Science Foundation [Grant BCS 1456077],

the Stanford Data Science Initiative, the Stanford Graduate School of Business, and the GarwoodCenter for Corporate Innovation at the Haas School of Business, University of California, Berkeley.

Supplemental Material: The supplemental material is available at https://doi.org/10.1287/mnsc.2016.2671.

Keywords: organizational culture • enculturation • cultural fit • attainment • linguistic accommodation

IntroductionOrganizational scholars have long recognized theimportance of culture in shaping individual, group,and organizational success. For example, culture fea-tures prominently in research on the efficacy of new-comer socialization (e.g., Ashforth and Saks 1996), theproductivity of groups and teams (e.g., Chatman et al.1998), and organizational performance following themerger of two firms (e.g., Weber and Camerer 2003).Although the definitions of culture have varied some-what across these research streams, prior research hastended to treat organizational culture as a static con-struct and therefore emphasized the importance ofachieving cultural fit—an informal threshold that anorganizational member either ultimately succeeds, orfails, to cross (Van Maanen and Schein 1979, Ashfordand Nurmohamed 2012)—for various indicators ofperformance (O’Reilly et al. 1991, Rivera 2012). Yetorganizational enculturation is a dynamic and ongoingprocess. Cultural fit, therefore, is an elastic construct.In this paper, we examine the following question:How is the specific temporal pattern of a person’s cul-tural compatibility with colleagues in an organizationrelated to her career outcomes in that setting?

Although some prior work assumes that culturalfit can change over time, especially during early new-comer adjustment to an organization (Bauer et al.2007, Chatman 1991), compelling theoretical accountsof the dynamics and consequences of cultural fitremain largely absent from the literature (Shipp andJansen 2011). We trace this paucity of theoreticaldevelopment to a methodological source: the toolsthat have heretofore been used to measure culturewithin organizations—such as participant observation(Kunda 2006, Van Maanen 1991) or self-report sur-veys (e.g., O’Reilly et al. 1991, Jones 1986, Hofstedeet al. 2010, Van Maanen 1975)—are simply ill-suitedto detecting fine-grained, temporal variation in cul-tural fit. The absence of such a measurement toolhas constrained researchers to assuming that a per-son’s cultural compatibility with an organization isfixed or, at most, monotonically increasing. Accordingto this view, newcomers remain probationary mem-bers of an organization unless and until they crosssome threshold level of cultural fit. This conceptualiza-tion of cultural fit as threshold crossing, we contend,has impeded theoretical progress on the dynamics of

1

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

http://pubsonline.informs.org/journal/mnsc/

mailto:[email protected]






Srivastava et al.: Enculturation Trajectories2 Management Science, Articles in Advance, pp. 1–17, ©2017 INFORMS

enculturation and has concentrated research attentionon either person–organization matching (e.g., Kristof1996) or early organizational socialization tactics (e.g.,Klein and Weaver 2000, Allen and Meyer 1990).By contrast, we propose that people can exhibit

increases or decreases in cultural fit throughout theirtenures in an organization. We introduce the constructof enculturation trajectory, which represents an individ-ual’s temporal pattern of cultural fit, and argue thatthe rate and direction of cultural adjustment is conse-quential for individual attainment. Drawing on previ-ous work on organizational socialization, we proposethat understanding how cultural fit waxes and wanesat different stages of a person’s tenure can provide awindow into two core mechanisms that underpin cul-tural fit: (1) acceptance of a focal actor by her colleaguesand (2) the focal actor’s attachment to her colleaguesand the organization as a whole. Thus, we hypothesizethat different enculturation trajectories will be asso-ciated with different career outcomes—namely, reten-tion, voluntary departure, and involuntary departure.To evaluate these ideas, we propose a novel mea-

surement approach, which is based on the languagepeople use in communications with their colleaguesin an organization. Language, we contend, provides awindow into organizational culture that is less suscep-tible to reporting biases, less topically constrained, andmore granular and scalable than self-report measures.It allows us to observe cultural fit as it unfolds overtime, illuminating enculturation as a process rather thanan end state. We apply our measurement strategy to aunique data set, which includes the complete corpus of10.24 million emails exchanged over five years among601 full-time employees of a midsized U.S. for-profittechnology firm.Whereas prior studies using archived electronic

communications in organizations have relied oncontent-free metadata to infer positions in networkstructure (e.g., Kossinets and Watts 2006, Kleinbaumet al. 2013, Srivastava 2015, Aven 2015), we have accessnot only to metadata but also to the natural languageof email content. We use the tools of computationallinguistics to transform this natural language into time-varying measures of individual-level cultural fit withcolleagues in the organization. We then rely on per-sonnel data to explore the relationship between encul-turation trajectories and individual outcomes in theorganization.

To preview our results, we find that employees withslow enculturation rates in the early stage (i.e., withintheir first six months in the organization) are morelikely to exit involuntarily than those with rapid ini-tial enculturation rates and that positive enculturationcan offset the downsides of initial low cultural fit. Wealso find that cultural fit can decline for some employeeslater in their careers, and that, when it does, it portendstheir choice to exit voluntarily.

From Cultural Fit to Trajectories ofEnculturationCultural Fit as an End StateOrganizations exhibit remarkable cultural persistencedespite turnover, growth, and decline (Kotter andHeskett 1992, Harrison and Carroll 2006). How donewcomers become aligned with an organization’s cul-ture? Existing literature has generally highlighted twodistinct yet complementary mechanisms. One empha-sizes cultural matching that occurs at the hiring stage.This work typically assumes that matching operateson ostensibly fixed attributes relating to individuals’ingrained psychological characteristics (Kristof 1996,Kammeyer-Mueller and Wanberg 2003) or accumu-lated cultural capital (Rivera 2012). Thus, organizationsselect (and are concomitantly selected by) individualswhose dispositions fit with the organization’s climateor who are culturally congruent with those who havealready joined the organization.

The process of cultural alignment does not, how-ever, end once an individual joins an organization.A second body of work—commonly referred to asorganizational socialization theory—focuses on theenculturation that occurs post entry, when newcom-ers acquire organization-specific cultural knowledge(Wanous 1992).1 Both cultural matching and encultur-ation lead to cultural fit, the state of being culturallycompatible with one’s colleagues in an organization.Organizations differ substantially in the extent towhich they actively propagate specific desired culturalfeatures (Sørensen 2002) and in the relative emphasesthey put on cultural matching versus enculturation.Even in the absence of an intentional effort to developa strong corporate culture, matching and enculturationnaturally occur through a combination of homophilyand peer influence (Carley 1991, Harrison and Carroll2006), leading organizations to vary in the levels ofcultural homogeneity they exhibit. Some organizationsare strongly aligned with a purposefully cultivatedorganizational culture, whereas others are more frag-mented (Martin 1992, Chatman et al. 2014).

While work on cultural fit and enculturation istoo vast to be comprehensively summarized here (forreviews, see, for example, Bauer et al. 2007, Kristof-Brown et al. 2005), we draw on two fundamentalassumptions that animate these literatures. The first isthat individual cultural fit is positively associated withindividual success in the organization. Although thereasons are multifaceted, two explanations for the linkbetween cultural fit and attainment are paramount.One is grounded in the psychological benefits of cul-tural fit. High cultural fit is thought to lead to greaterjob satisfaction, stronger identification and attachmentwith the organization, higher motivation, and reducedstress. As a result, people achieve higher levels of per-formance and a longer tenure with the organization

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Srivastava et al.: Enculturation TrajectoriesManagement Science, Articles in Advance, pp. 1–17, ©2017 INFORMS 3

(O’Reilly et al. 1991, Chatman 1991, Meglino et al.1989). The other is rooted in culture’s role as a solu-tion to the complexities and challenges of interper-sonal coordination under conditions of uncertainty.Colleagues who fit in culturally with each other areassumed to have more efficient and efficacious inter-actions with one another, resulting in better coordina-tion and higher productivity (Kreps 1990, Weber andCamerer 2003, Van den Steen 2010).A second common assumption in enculturation

research is that the process unfolds in distinct stages.Although they use different terminology and iden-tify slightly different break points, enculturation mod-els typically include three core stages (Bauer et al.1998): (1) anticipatory adjustment, which occurs priorto entry, (2) early adjustment, which occurs immedi-ately following entry, and (3) final adjustment, whennewcomers are fully accepted as insiders. It is oftenassumed that the second stage, when newcomers expe-rience high levels of uncertainty and stress as theylearn and update their expectations about the organi-zation and try to make sense of its normative order,is the most critical for subsequent attainment. This ispresumed to be the period of most consequential orga-nizational learning.

Enculturation as a ProcessTo summarize, the process of enculturation is oftenconceptualized, to use VanMaanen and Schein’s (1979)imagery, as a newcomer’s radial movement from out-side the organization’s formal boundary into its cul-tural core, as illustrated in panel (A) of Figure 1.Cultural matching occurs at the point of entry intothe organization during hiring, followed by a proba-tionary period of early cultural adjustment. Once thenewcomer passes an informal threshold of acceptance,he or she presumably becomes a full member of theorganization.

Figure 1. An Illustration of Cultural Fit as an End State (A) and of Enculturation as a Process (B)

Probationarymembership

Fullmembership

Culturalmatching

Peer acceptance

Attachment

(A) Cultural fit as an end state (B) Enculturation as a process

Notes. In the end-state framework (panel (A)), the newcomer needs to cross the formal organizational boundary (full line) and the informalacceptance boundary (dotted line) to attain cultural fit and become a full member. In the process framework (panel (B)), the cultural journey isongoing; the mechanisms of peer acceptance and attachment are consequential during different phases of this journey.

Although enculturation is often assumed to be anongoing process, empirical studies of socializationhave, in practice, tended to treat organizational cultureas fixed and monolithic and conceived of individual-level cultural fit as a static end state that people eitherachieve or fail to achieve through processes of selec-tion and posthire enculturation (e.g., Allen and Meyer1990). But maintaining cultural alignment requiresconstant investment. Moreover, culture is known to bean evolving, group-level adaptive response to inter-nal and external pressures (Schein 2010, Ravasi andSchultz 2006), represented by the jagged boundary inpanel (B) of Figure 1. Enculturation, as panel (B) illus-trates, is therefore better understood as a journey anorganizational member takes, rather than as a thresh-old he or she successfully traverses. This cultural jour-ney, we argue, is as important to understand as thedestination.

Thinking of enculturation as a process helps pointto two mechanisms through which cultural fit relatesto individual attainment: peer acceptance and individ-ual attachment. Cultural fit can lead to acceptance bycolleagues because it is interpreted as a signal thatan individual’s values, beliefs, and styles of work arecompatible with those of her coworkers. It also servesas a manifestation of her attachment to the firm sincethose who feel they belong in an organization are lesslikely to adopt countercultural behaviors to assert theirdivergent social identities. Although both peer accep-tance and attachment are consequential for attainmentin the organization—one cannot succeedwithout beingaccepted by others, and a lack of attachment damp-ens one’s impetus to be a productive organizationalmember—each pathway tends to bemore salient at dif-ferent stages of the enculturation process. We illustratethis in panel (B) of Figure 1.

Early enculturation relates to a newcomer’s ability togain acceptance by peers. It is during this probationary

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


phase that newcomers’ identities and behaviors aremost heavily scrutinized by their colleagues. And, asprevious research has shown, it is during this periodthat newcomers experience heightened anxiety anduncertainty and therefore a strong incentive to conformculturally (Jones 1986). Even if meticulously screenedon cultural matching and culturally trained during theorganization’s formal “onboarding” process, newcom-ers are still required to tune into and adopt the ineffa-ble aspects of the organizational code (March 1991), tolearn which behaviors are appropriate and which arefrowned upon, to assess what idiosyncratic rituals andsymbols signify, and to infer what implicit assumptionsand expectations are informing colleagues’ behaviors(Van Maanen 1991). As Schein (2010) points out, mak-ing sense of these cultural artifacts is rarely a straight-forward task. Yet whatever choices and actions thenewcomer takes, the success of this process is ulti-mately determined by his or her colleagues, whodecide whether to accept the newcomer as an insider(Wanous 1992).Once this implicit boundary is passed, however,

cultural alignment becomes less a matter of gainingacceptance by colleagues and more a challenge of self-maintenance of cultural compatibility. Although indi-viduals vary in the extent to which they buy intothe culture and in their ability to take on culturalfacades, diffusing the tension between front-stage nor-mative compliance and back-stage identity manage-ment requires significant and constant emotional work(Goffman 1959, Hochschild 1979, Cable et al. 2013,Kilduff and Day 1994, Grandey 2003). The employeesin Kunda’s (2006) ethnography of “Tech,” for exam-ple, constantly partook in exchanges of cynicism anddetachment as a means to reassert their authenticityand membership in the organization while resolvingthe inherent tension between both identities.

Organizational members need to put in work toremain normatively compliant even in the contextof a stable organizational culture. Their need to doso is amplified when the organizational landscapeitself changes. These changes need not be the resultof dramatic shocks or concerted cultural retooling.Rather, the cultural content in organizations constantlyand organically evolves as new symbols are intro-duced and existing ones reinterpreted, as recent eventsare mythicized and old stories are forgotten and asnew implicit agreements emerge to substitute for oldones (Danescu-Niculescu-Mizil et al. 2013, Rafaeli andPratt 2006, Ravasi and Schultz 2006). Because keepingpace with the organizational culture requires energyand attention, it also necessitates motivation. Unlessprompted by an unusual shift in the organization’sculture, a decline in the level of cultural alignmentby an organizational member who has already gained

peer acceptance should be reflective of that individ-ual’s declining attachment to the organization.

Conceptualizing cultural fit as a static end stateobscures these different processes. Whereas lowcultural fit might lead to negative evaluations by col-leagues, it could also be an indicator of low attach-ment to the organization. Just knowing that a personhas failed to achieve a high level of cultural fit tells usvery little about which of these twomechanisms—peeracceptance or attachment—might be operative. Rather,the timing and pattern of enculturation are likely to becrucial in disambiguating these underlying pathways.

We hypothesize that different outcomes in theorganization leave different enculturation signatures.Three such outcomes are particularly important: reten-tion, voluntary exit, and involuntary exit. We inter-pret retention to mean that a person has beenaccepted culturally by others and remains motivatedto stay culturally compliant with others’ expectations.We interpret voluntary departure as an indication oflow commitment to the firm in light of other outsideopportunities. Involuntary departures, on the otherhand, are imposed on the individual and therefore typ-ically indicate the inability to gain acceptance by one’scolleagues. Although these different exit types reflectdifferent underlying processes—attachment and peeracceptance—previous research on enculturation hasoften overlooked the distinction between them eitherby measuring turnover irrespective of exit type (e.g.,Kammeyer-Mueller and Wanberg 2003, Cable et al.2013) or by focusing only on voluntary exit (e.g., Allen2006, Chatman 1991). This inattention is reflective of atheoretical tendency to conflate the effects of culturalfit on attachment with its effects on evaluations by oth-ers and to treat cultural fit as a boundary that is eithercrossed or never traversed.

By contrast, we expect that these different pathwaysrelate to different patterns of enculturation and, corre-spondingly, different individual outcomes. Organiza-tional members who are successfully integrated intothe organization should exhibit a capacity to adjust cul-turally post entry—and concomitantly gain the accep-tance of their peers—aswell as continuedmotivation toincrease their cultural alignment after this inclusionaryboundary has been traversed. Their cultural fit shouldincrease steadily over time. By contrast, peoplewho failto adapt culturally in the early stages of their tenureare less likely to be accepted by their peers and there-fore face a greater hazard of experiencing involuntaryexit. Finally, people who succeed in adapting culturallyearly in their tenure and gain acceptance by colleaguesbut then—at a later stage in their tenure—experiencea decline in cultural fit are likely detaching from theorganization. We posit that such a pattern heralds theirvoluntary exit from the organization. In other words,we expect the following.

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Hypothesis 1. A secular increase in cultural fit is predic-tive of retention.

Hypothesis 2. Slow rates of enculturation early in a per-son’s tenure in an organization are predictive of involuntaryexit.

Hypothesis 3. A decline in cultural fit later in the tenureof a person who was previously enculturated into an organi-zation presages voluntary exit.

Enculturation and Language UseStudies of organizational culture have mostly es-chewed questions relating to enculturation trajectories,in large part because culture is a complex constructthat is difficult tomeasure consistently over time (Mohr1998, Goldberg 2011). Organizational scholars have,of course, studied cultural processes extensively, butmethodological limitations have precluded the sys-tematic analysis of enculturation patterns. Participantobservation provides rich insight into the workings ofenculturation (e.g., Kunda 2006, VanMaanen 1991), butgiven that a researcher can only be present in one set-ting at a given time, he or she cannot feasibly observeall organizational members on a consistent basis. Pre-vious work systematically examining individual vari-ability in enculturation has therefore mostly relied onself-reports to operationalize individual cultural fit.Self-reports suffer, however, from a variety of lim-

itations (Greenwald and Banaji 1995, Srivastava andBanaji 2011): they presuppose a small set of culturaldimensions, often overlooking organizationally spe-cific cultural manifestations; are subject to a varietyof social and cognitive reporting biases; and, by theirnature, sacrifice qualitative richness for observationalbreadth, leading to a focus on core cultural dimensionsthat are often most resistant to change. Most impor-tant, self-reports are inevitably limited in scope, giventhat individuals cannot be surveyed constantly andexhaustively.2 While they provide access to subjectivedispositions and perceptions, self-reports are limitedin their ability to systematically address fundamen-tal questions that relate to the evolution of individualenculturation over time.3

Language as a Signal of Cultural AlignmentHow one measures enculturation invariably relatesto how one defines culture. Although scholars haveoffered a variety of definitions, most would agreethat organizational culture comprises two fundamen-tal dimensions: a cognitive dimension, relating to orga-nizational members’ shared assumptions, beliefs, andvalues, and a behavioral dimension, relating to normsand expectations that emerge from these values andthat govern interaction in the organization (Schein2010, Hofstede et al. 2010, Ravasi and Schultz 2006,Chatman et al. 2014). Cultural fit can concomitantly

be thought of as an individual’s levels of cognitiveand behavioral alignment with her peers; namely, theextent to which she shares understandings with herpeers and is normatively compliant with their expec-tations. Cognitive cultural fit is rarely observed in theorganization; members have only limited and indirectaccess to their colleagues’ cognition. Instead, peopleinfer their own level of cultural fit, as well as assessthe cultural fit of their colleagues, by observing theirpeers’ behavior and comparing it to their own. Theyinterpret behavioral cultural alignment as an indica-tion of cognitive alignment. Such an interpretation isnot necessarily correct, as some individuals put onfacades that mask their true beliefs and values. Yetsuch incongruent “surface acting” is emotionally tax-ing and is often either resolved by readjusting one’sinner thoughts and feelings or by departing from theorganization (Hochschild 1979, Grandey 2003).

Language is central to these processes. It is amongthe most salient organizational indicators of an indi-vidual’s level of behavioral cultural alignment.4 At itsmost basic level, language is a set of conventions thatconnect symbols with meanings, providing a solutionto a complex coordination problem (Lewis 1969). Orga-nizations converge on distinct linguistic conventionsthat relate to their particular context and the oppor-tunities and challenges they face (Crémer et al. 2007).An individual’s level of compliance with these con-ventions is therefore essential for becoming a produc-tive member of the organization. But language is notmerely functional; these conventions also come to sig-nify social identities and roles, and their normative useis an indication of an organizational member’s degreeof assimilation. Whether conscious or not, people’slinguistic choices are crucial for establishing relation-ships with their interlocutors (Giles et al. 1991, Labov2001). For example, linguistic compatibility minimizesperceived social distance between interaction partners,whereas linguistic divergence strengthens symbolicboundaries between them (Gumperz 1982, Bernstein2003, Niederhoffer and Pennebaker 2002, Danescu-Niculescu-Mizil et al. 2012). This happens because anindividual’s tendency to accommodate others linguis-tically both affects others’ evaluations (e.g., Rickfordet al. 2015) and is a reflection of her self-perceived sim-ilarity with her interlocutors (e.g., Ireland et al. 2011).Thus, language use is intrinsically related to the pro-cesses by which individuals fit, or fail to fit, into theirsocial environments.

The language people use in their daily interactionscan also provide a window into their underlying cat-egories of thought and value systems (Pinker 2007).Take swearing as an example.5 Organizations—and thevarious subgroups they house—vary in the extent towhich they condone or reject the use of profanities.

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


A newcomer’s ability to comply with the norms con-cerning the degree and appropriate use of vulgar lan-guage serves as a strong signal of her ability to readthe organizational code and conform to it. But the useof swear words also taps deeper systems of mean-ings. In an ethnography of counterculture youth inEngland in the 1960s, for example, Willis (2014) findsthat while hippies use ornate forms of language to sig-nal their defiance of mainstream British society, bik-ers do the same through pervasive use of profanities.Willis links these different linguistic styles with thehippies’ middle-class and bikers’ working-class back-grounds, arguing that the latter’s use of vulgar lan-guage relates to their celebration of muscularity.As Pinker (2007) points out, swear words invoke

strong emotions and are a form of symbolic violenceand power display. By extension, an organizational cul-ture that is tolerant of the use of swear words mightindicate a shared value system that accepts aggres-sion and coercion as legitimate forms of interpersonalcoordination. Contrast such an organization with theBody Shop, where the expression of emotion is nor-matively encouraged.6 Such “bounded emotionality”(Martin et al. 1998) is reflective of an underlying beliefsystem that values personal well-being and commu-nity and that rejects the assumption that workplacestress enhances productivity. As these examples illus-trate, linguistic alignment between an individual andher peers can serve as an indication of that individual’slevel of cultural fit.

Linguistic Reference GroupAny investigation of cultural fit must contend withthe choice of reference group against which to com-pare a focal individual’s degree of fit. The extant lit-erature commonly distinguishes between two levels offit: person–organization (PO) and person–group (PG)fit, the latter normally conceived of as the individual’sfit with her department or functional unit. The dif-ference between these two constructs is not merely amatter of level of analysis. While PO fit relates to align-ment between the individual and the baseline beliefsand values shared across all members of the organiza-tion, PG fit is more attuned to the specific assumptionsand norms evolved in one’s particular organizationalunit. Research indicates that both types of fit tend tobe correlated and are generally predictive of individualattainment and positive group and organization out-comes (Kristof-Brown et al. 2005, Adkins and Caldwell2004, Elfenbein and O’Reilly 2007).Both constructs presume that organizational cul-

ture follows the contours of formal organizationalstructure. This assumption introduces problems forboth constructs. First, PO fit assumes a unitary cul-ture across the entire organization. Yet many organi-zations exhibit cultural differentiation across distinct

subcultures (Martin 1992). Under such circumstances,an organization-level culture is more an analytical fic-tion than an experienced organizational reality. PG fit,by contrast, allows for cultural variation within theorganization but assumes that such variation neces-sarily follows formal organizational boundaries. Yetresearch suggests that informal organizational rela-tionships chronically crisscross formal and semifor-mal boundaries (Biancani et al. 2014, Srivastava 2015).There is no a priori reason to believe that cultural vari-ation necessarily forms along formal rather than infor-mal fault lines (e.g., occupational ones; see VanMaanenand Barley 1984).

Rather than reifying formal organizational units asmeaningful cultural groups, we make two differentassumptions. First, we assume that the linguistic man-ifestations of cultural fit might vary across settingsand groups within an organization. Even organiza-tions with strong and uniformly shared beliefs andvalue systems might exhibit variation in the normativeexpressions of these shared understandings. Kitchenworkers in Fine’s (1996) ethnography of restaurantwork, for example, converge on a variety of linguis-tic conventions for expressing flavor (e.g., “cooked todeath” versus “soothing”), even if they have similarconceptualizations of what “good” taste constitutes.Second, we assume that in organizations with a suffi-ciently large number of employees (e.g., several hun-dred), people cannot feasibly interact with more thana subset of members. It is the set of colleagues withwhom an organizational member interacts on a fre-quent basis who form an impression of that individ-ual’s degree of cultural assimilation and who are mostconsequential for determining the individual’s cul-tural fit.

Consequently, we shift the focus from a formal toinformal structure and conceptualize cultural fit as thelinguistic alignment between an individual and herinteraction partners in the organization. In organiza-tions with a strong homogeneous culture, this opera-tionalization will be very consistent with PO and PGapproaches to measuring cultural fit. In more frag-mented (and arguably more typical) organizations,however, our approach has the advantage of beingrobust to cultural heterogeneity within the organiza-tion and to mismatches between the culture observedwithin formal organizational boundaries and that pre-vailing in informal patterns of interaction.

An Interactional Language-Use Model ofEnculturationWe define cultural fit as an individual’s level of linguis-tic compatibility with her interaction partners during agiven observation window and an enculturation trajec-tory as the temporal pattern of individual cultural fit.7

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Our measure of cultural fit is defined in termsof textual records of interactional language use (e.g.,email exchanges, text messages, phone call transcripts).We assume a method ϕ for mapping texts to linguisticunits in lexicon L (e.g., words, bigrams, noun phrases,emotional categories). To reduce the effects of domain-and task-specific vocabulary, we use the LinguisticInquiry and Word Count (LIWC) (Pennebaker et al.2007) lexicon as the mapping method ϕ to code eachemail message relative to a set of semantic categories.The LIWC is an established framework for measuringlinguistic style (e.g., as reflected in the use of pronouns,swearing, or negations) that allows us tomeasure inter-locutors’ normative, as opposed to substantive, linguis-tic congruence.8 Measuring alignment with respect tothese linguistic categories helps to ensure that ourmea-sure of cultural fit does not merely reflect functionalcoordination between two individuals. For example,two employees troubleshooting a customer problemmay be using the same terminology in email exchangesin which they diagnose the problem, but whereas oneinterlocutormay be using swearwords, the othermightnot. Such an interaction is culturally incongruent, evenif topically aligned.We segment the data into monthly observation win-

dows to study trajectories of enculturation. To mea-sure the cultural fit of individual i during period T,we tokenize each textual record—in our case, emailmessages—into LIWC category frequencies, and wecreate two probability distributions giving the normal-ized frequencies for linguistic units in i’s outgoing andincoming messages in T. Let −→m it be a message sent byperson i at time t, let ←−m it be a message received byperson i at time t, and let l ∈ L be the list of 64 LIWCcategories. Our procedure iterates over all messagesand for each produces −→m l

it , which counts the numberof terms relating to LIWC category l contained in mes-sage −→m it . It then aggregates over all messages −→m it sentby person i during period t ∈ T to produce the nor-malized probability of category l for person i duringperiod T, as follows:

O liT �

−→m liT∑

k∈L−→m k

iT

. (1)

The procedure similarly normalizes over all messagesreceived by person i during period T to produce thenormalized probability over LIWC categories in i’sincoming messages:

I liT �

←−m liT∑

k∈L←−m k

iT

. (2)

We define i’s cultural fit at time T as the negative log ofthe Jensen–Shannon (JS) divergence (Lin 1991) betweenthese two normalized distributions, stated formally:

CFT(i)�− log(JS(OiT ‖IiT)), (3)

where the JS divergence between the two probabilitydistributions is defined as

JS(O‖I)� 12KL(O‖M)+ 1

2KL(I‖M), (4)

and where M �12 (O + I) and KL(O‖M) is the Kullback-

Leibler (KL) divergence of M from O:

KL(O‖M)�∑l∈L

O(l) log2O(l)M(l) . (5)

JS divergence is a symmetric measure of dissimilar-ity between two probability distributions. It smoothsthe KL divergence values and ensures that they arealways finite. As we have defined it here in terms oflog2, its values always fall in the interval [0, 1]. Thisapproach builds on previous efforts to estimate linguis-tic accommodation using probabilistic language mod-els (Danescu-Niculescu-Mizil et al. 2012, Hughes et al.2012). We have found that the smoothing properties ofour measure are particularly well suited to the sparse,power-law distribution of words in natural languageuse (Zipf 1949, Baayen 2001, Piantadosi 2014).

The intuition behind JS is fairly straightforward. Theterm log2(O(l)/M(l)) in Equation (5) equals 0 whenO(l) � M(l) (that is, when the probability of linguis-tic unit l is equal in both distributions). The prod-uct O(l) log2(O(l)/M(l)) increases as O(l) grows andO(l) � M(l). Thus, the summation in Equation (5)grows when high probability units in O have signifi-cantly lower probability in M.9 KL divergence can beinterpreted as the amount of information necessary totranslate one distribution into another; when it equalszero, the two distributions are identical. Because M isthe average between the two distributions O (i’s out-going messages) and I (i’s incoming messages), then ifthe two are identical, both perfectly predict their aver-age, leading to JS(O‖I) � 0. As O and I diverge, theiraveraged dissimilarity in Equation (4) increases, andtherefore, i’s cultural fit in Equation (3) decreases.10

Our language-based measurement approach over-comes the fundamental limitations of self-report mea-sures that are commonly used to measure culturalfit in organizations. First, because language use is abehavioral outcome, our method is not subject to self-report bias. Second, in relying on naturally occurringunstructured textual exchanges, it is not limited tocultural dimensions assumed by the researcher andcontained in a survey instrument. A language-basedapproach does not require the researcher to make theusual trade-off between the richness of ethnographicresearch and the reach of survey research. It insteadtaps into more subtle forms of cultural differenceamong people. Third, since language use is pervasive,we can measure cultural fit at scale and at high gran-ularity over time. Fourth, as noted above, measuring

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


cultural fit with respect to a person’s interaction part-ners allows for the possibility of cultural heterogene-ity across groups within the organization, as well asover time within individuals. Together, these featuresenable us to measure enculturation trajectories withhigh resolution and in a consistentmanner that enablescomparisons across individuals.

DataWe obtained access to the complete corpus of elec-tronic messages—including metadata and content—exchanged among the full-time employees at amidsized technology company between 2009 and 2014.To protect employee privacy and company confiden-tiality, we stored all data on secure research serversthat we purchased and installed at the firm, elim-inated messages exchanged with parties external tothe firm, excluded messages exchanged with any ofthe company’s attorneys, and deleted message con-tent and all identifying information about employeesafter applying our natural language processing algo-rithms. The resulting data set included 10,236,668 dis-tinct messages.In addition to email data, we obtained human

resource records that included employee age, gender,tenure, and, for employees who departed the company,whether this departure was voluntary or involuntary.We inferred departmental affiliations and promotionsfrom distribution lists and applied additional refine-ments to the data. The resulting data set includes 9,885person-month observations for 601 full-time employ-ees. These form the basis of the analyses reportedbelow.11

ResultsBefore testing our hypotheses related to encultura-tion trajectories, we sought to establish whether ourinteractional-language-use measure of cultural fit is

Figure 2. Cumulative Probability of (A) Being Promoted to a Managerial Position and (B) Exiting Involuntarily, as Estimatedby Separate Cox Proportional Hazard Models

6 12 18 24 30 36

Months employed

6 12 18 24 30 36

Months employed

0

0.1

0.2

0.3

0.4

0.5

Cum

ulat

ive

prob

abili

ty

0

0.1

0.2

0.3

0.4

0.5

Cum

ulat

ive

prob

abili

ty

(A) Promotion

Low cultural fit (5th %ile)

Median cultural fitHigh cultural fit (95th %ile)

(B) Involuntary exit

predictive of individual attainment.12 We reasonedthat, if our measure is reflective of cultural fit, it shouldbe positively associated with individual career suc-cess (O’Reilly et al. 1991). Consistent with this expecta-tion, our measure of cultural fit strongly predicts bothpositive and negative attainment in the organization.Figure 2 reports the cumulative probabilities of beingpromoted to a managerial position (positive attain-ment) and being asked to leave involuntarily (negativeattainment), as estimated by two separate Cox pro-portional hazard models (each including controls forsociodemographic and organizational attributes; seeTable 1 of the supplemental material for details). Rank-and-file employees with high cultural fit have a cumu-lative probability of 48% of being promoted to a man-agerial position by the end of their third year at thefirm (see Figure 2, panel (A)), which is 1.5 and 2.7 timesgreater than their counterparts who exhibit median fitand low cultural fit, respectively. The implications oflow cultural fit for involuntary exit are particularly dra-matic (see Figure 2, panel (B)): at 46%, the cumulativeprobability of involuntary exit after three years is fourtimes greater for an employee with low cultural fit thanit is for one with median cultural fit.

Consistent with our expectations, cultural fit is not,however, a static personal attribute. Rather, for theaverage employee, cultural fit follows an upward-sloping trend, as depicted in Figure 3. For ease ofinterpretation, cultural fit is standardized such thatzero cultural fit corresponds to the average employeeat the firm. As the figure illustrates, newly hiredemployees initially exhibit rapid linguistic accommo-dation, reaching the mean level in the firm by theend of their first year. The growth rate of their cul-tural fit gradually decreases thereafter. In other words,our method demonstrates that newcomers to the firmare, on average, culturally adaptable; they achieve cul-tural assimilation despite initially being culturally dis-tant from their colleagues. It is also consistent with

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Figure 3. Cultural Fit (Standardized) as a Function ofNumber of Months Employed

6 12 18 24 30 36

Months employed

–0.5

–0.4

–0.3

–0.2

–0.1

0

0.1

0.2

0.3

0.4

Cul

tura

l fit

6 12 24 30 36

–2

2Employed

Involuntarilydeparted

Voluntarilydeparted

Cul

tura

l fit

0

18

Notes. The main diagram plots a cubic linear fit with 95% confidenceinterval (black line) and mean observed values (gray dots). The insetplots enculturation trajectories of varying lengths of tenure at thefirm for 30 randomly sampled employees. Highlighted employeesvary in employment status.

previous work that assumes enculturation entails dis-tinct phases.Yet the general trend illustrated in Figure 3 masks

considerable heterogeneity. Employees vary signifi-cantly in their average and peak levels of linguisticaccommodation, as well as in their overall encultur-ation trajectories, as the inset of Figure 3 (plotting arandom sample of individuals) illustrates. While theaverage employee at the firm exhibits positive encul-turation throughout her career, some employees expe-rience a decline in cultural fit. Moreover, although thefirm in question puts a strong emphasis on hiring on

Figure 4. Marginal Effect of Tenure on Cultural Fit (Standardized), as Estimated by (A) Period Fixed-Effects Model,(B) Matched-Pair Fixed-Effects Model, and (C) Two Independent Individual Fixed-Effects Models

3 6 9 12 15 18 21 24 27 30 33 36

Tenure (months)

–0.9

–0.8

–0.7

–0.6

–0.5

–0.4

–0.3

–0.2

–0.1

0

0.1

0.2

0.3

0.4

Cul

tura

l fit

(A) Period fixed-effects model

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Tenure (standardized)

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Tenure (standardized)

–0.8

–0.7

–0.6

–0.5

–0.4

–0.3

–0.2

–0.1

0

0.1

0.2

0.3

0.4

Cul

tura

l fit

(B) Matched-pair fixed effects model

–0.6

–0.5

–0.4

–0.3

–0.2

–0.1

0

0.1

0.2

0.3

Cul

tura

l fit

(C) Individual fixed-effects models

NondepartedVoluntarily departedInvoluntarily departed

NondepartedVoluntarily departedInvoluntarily departed

Voluntarily departedInvoluntarily departed

Notes. Effects plotted by employment status. Shaded areas correspond to 95% confidence intervals.

cultural fit (as discerned from conversations we con-ductedwith its chief people officer), newcomers exhibitlarge variation in initial levels of fit. If cultural fit relatesto a person’s ability to integrate successfully with hercolleagues, as we hypothesized earlier, then we shouldfind that different enculturation trajectories explain dif-ferences in individual outcomes in the firm.

We test our hypotheses by differentiating amongthree types of employees: (1) those who remainedemployed, (2) those who left the firm involuntarily,and (3) those who left voluntarily. As noted above, weinterpret involuntary departure as indication of rejec-tion by colleagues and voluntary departure as an indi-cation of weakened attachment. Figure 4 reports themarginal effects of tenure on cultural fit for these differ-ent employee types as estimated by several fixed-effectsmodels.

The first model, reported in panel (A), estimatescultural fit as a function of months of employment.Tenure in months and its square term are interactedwith dummy variables for voluntary and involun-tary exit, such that nondeparted employees serve asthe omitted category. We include period (monthly)fixed effects to account for unobserved heterogene-ity that is time related—for instance, firm-level (e.g.,growth, contraction, or changes in hiring practices) andmarket-level (e.g., supply of job applicants) variationthat might systematically affect cultural fit, departureand entry rates, or individual outcomes. We constrainthe sample to three years of employment to enablea comparison between employee types.13 The resultsare consistent with our hypotheses; namely, retainedemployees exhibit an increase in cultural fit, involun-tarily departed employees do not exhibit a statisticallysignificant increase in cultural fit, and those departingvoluntarily follow an inverted U-shaped pattern ofenculturation.

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


However, the estimates reported in panel (A) exhibita degradation in confidence intervals as time goesby because departed employees drop out of the sam-ple. Moreover, as Figure 3 illustrates, we observe thatemployees differ not only in their enculturation trajec-tories but also in their rates of enculturation: amongthose who enculturate, some do so quickly whereasothers take longer to meet their peers’ level of culturalfit. We assume that the consequences of enculturationare affected by these differences in individual tempo—that is, that cultural adaptation and its relationshipwith individual outcomes is related to an individual’slife cycle at the firm rather than to that person’s abso-lute number of months at the firm. Consequently, westandardized time by employees’ tenure at the firm,such that it ranges from 0 to 1. Let ei be the monthof entry for individual i, and let di be the month ofdeparture for that individual. We calculate standard-ized tenure as τi � (ti − ei)/(di − ei), where ti corre-sponds to the month individual i is observed at thefirm.While departed employees are observed throughout

their tenure in the firm, observations of nondepartedemployees are right censored: some may leave in thefuture. Because we do not observe their departure,we cannot standardize their tenure. To address theseproblems, we employ a matched-pairing approach.We randomly pair each departed employee with onenondeparted employee in the month of arrival to thefirm, and we model both employees’ cultural fit onlythroughout the departed employee’s tenure. We stan-dardize the nondeparted employee’s tenure by thedeparted employee’s. That is, for each departed indi-vidual i, we randomly matched a nondeparted indi-vidual i′ such that ei � ei′ and define τi′t � τit . Thus,we compare departed employees’ cultural fit to thatof their counterparts who had joined the firm atthe same time and have remained in the firm since.We model cultural fit as a function of standardizedtenure and, once again, use interaction terms to differ-entiate between exit types (with nondeparted employ-ees serving as the omitted category; see Section 3 ofthe supplemental material for more details). We alsoinclude matched-pair fixed effects. Our modeling strat-egy allows us to account for heterogeneity in indi-vidual tenure lengths among the departed as well asaddress unobserved time-related heterogeneity.Panel (B) of Figure 4 illustrates the marginal effects

of standardized tenure on cultural fit as estimated bythis matched-pairs model. The three employee typesexhibit distinct enculturation trajectories. Confirmingextant literature on the relationship between culturalfit and attainment, and consistent with Hypothesis 1,individuals who are retained by the organizationexhibit a gradual increase in their level of culturalalignment. Not only do these individuals seem to gain

their colleagues’ acceptance; we interpret their consis-tent positive enculturation as an indication of a strongattachment to the organization.

By contrast, and consistent with Hypothesis 2, thosewho eventually leave involuntarily fail to accommo-date their colleagues linguistically from the momentthey join the organization. The first third of their tenureis characterized by consistently low cultural fit, whichis then followed by a gradual decline that moves themculturally further apart from their nondeparted coun-terparts. This lack of cultural adaptability has manycauses, which vary across individuals and situations,and that may be related to individual motivation or tocapabilities (Weber and Camerer 2003, Harrison andCarroll 2006, Jones 1986); regardless, these individuals’inability to enculturate portends their failure to gaintheir peers’ acceptance and to integrate successfullyinto the firm.

Those departing voluntarily, on the other hand, fol-low a different trajectory. Initially, they are statisticallyindistinguishable from nondeparted colleagues whohad joined the firm at the same time. Both groupsfollow the same upward trajectory of enculturation.Once they peak in cultural fit, roughly at their half-life in the firm, those who depart voluntarily beginto exhibit a decline. Unlike those who end up leav-ing the firm involuntarily, those who exit voluntar-ily are clearly capable of adapting. It appears that atsome point in their tenure, they cease to accommodatetheir colleagues linguistically. Consistent withHypoth-esis 3, this late decline in cultural compatibility withcolleagues appears to foreshadow an intention to leavethe organization.14Although our modeling strategy allows us to com-

pare nondeparted employees to those departing vol-untarily or involuntarily within the same model, itprecludes usage of individual fixed effects (given thatexit type is fixed per person). We therefore cannot ruleout that the different patterns depicted in panel (B) ofFigure 4 are attributable to stable differences amongindividuals (such as those related to human capi-tal or to psychological capabilities that facilitate cul-tural assimilation). To address this limitation, wemodel cultural fit by standardized tenure using anindividual fixed-effects model estimated separatelyfor voluntarily and involuntarily departed individu-als (excluding nondeparted individuals). Individualfixed-effects models account for unobserved hetero-geneity across individuals and therefore mitigate con-cerns about omitted variable bias (Greene 2012). Theyallow us to isolate individual enculturation trajectoriesby examining the relationship of within-person tenurechange on cultural fit, net of an individual’s baselinecultural fit. The marginal effects estimated by thesemodels are illustrated in panel (C) of Figure 4.15 They

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


reproduce the trends illustrated in panel (B) of Figure 4,suggesting that the differences in enculturation trajec-tories by exit type cannot be explained merely by dif-ferences in individual baseline capacity for cultural fit.The different trajectories depicted in the three pan-

els of Figure 4 are striking. However, because we donot have access to the cognitive processes producingthese results, only to their behavioral manifestations,we cannot determine their causes. It is nevertheless evi-dent that, whereas the voluntarily departed are capableof enculturation, the involuntarily departed are eitherincapable of cultural adaptation or unwilling to adapt.Given that involuntary departure is imposed on theindividual, while voluntary departure is a choice, weinterpret these results as suggesting that lack of cul-tural adaptability relates to a negative reception by col-leagues, whereas a drop in cultural fit for previouslyencultured individuals is indicative of a decline in anindividual’s attachment to the organization.

The results in Figure 4 also point to the importanceof enculturation, relative to initial cultural fit. By thetime the departed leave the firm, the three employeetypes exhibit different levels of cultural fit: those stillemployed by the firm are significantly above average;those voluntarily exiting are significantly below aver-age; and those leaving involuntarily exhibit dramati-cally low levels of cultural fit, significantly lower thanthe average newcomer’s (−0.52 compared with −0.3;see Figure 3). This is not the case upon arrival atthe firm, however. Although the nondeparted exhibitrelatively high levels of cultural fit when they jointhe firm (panel (A)), because there is great variabilityin initial cultural fit and in tenure lengths, employeetypes are statistically indistinguishable when they areproperly matched (panel (B)). The different encultur-ation signatures depicted in Figure 4 strongly sug-gest that employees’ fates are not merely the resultof their prehire cultural fit but also their capacity forenculturation. As we hypothesized, initial encultur-ation seems particularly consequential for successfulintegration: those who do not adapt to their colleaguesearly on appear to be at high risk of being asked toleave.To explore this further, we calculated the encultura-

tion rate for each employee during her first six monthsat the firm, which, as our nonstandardized estimates(panel (A)) and previous evidence (Bauer et al. 1998)suggest, is the critical period during which early encul-turation unfolds. We do so by fitting a simple linearmodel, effectively measuring the slope of cultural fitduring a newcomer’s first six months. We estimatedtwo Cox proportional hazard models, estimating therisk of involuntary and voluntary departure as a func-tion of this slope and various control variables. Asthe results in Table 1 demonstrate, initial cultural fit

Table 1. Cox Proportional Hazard Models of Exit

Model 1: Model 2:Involuntary Voluntary

Enculturation rate 0.086∗∗ 0.281(−2.66) (−1.40)

Initial fit 0.575∗∗ 0.681(−2.97) (−1.57)

Age 1.119 0.967(1.00) (−0.25)

Age2 0.999 1.000(−0.49) (0.07)

Female 1.286 1.927∗(0.88) (2.10)

Manager 0.857 1.098(−0.31) (0.21)

Department controls Yes YesN 8,238 8,238χ2 37.370 20.386Log-likelihood −2,003.94 −1,656.18Number of exits 68 56

Note. Exponentiated coefficients; t-statistics in parentheses.∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001.

and early enculturation reduce the risk of involun-tary, but not voluntary, departure: a one-third standarddeviation increase in cultural fit per month (whichroughly corresponds to the 90th percentile of encul-turation rate) decreases the hazard ratio of involun-tary exit by 30%.16 In other words, failure to assimilateearly on appears to be related to a failure to receiveacceptance by others but not to one’s attachment to theorganization.

We report the results from Model 1 in Table 1as cumulative hazards in Figure 5. We distinguishbetween different levels of initial cultural fit and earlyadaptation rates, depicting hazard for newcomers withlow initial cultural fit and either high (black solid line)or low (black dashed line) enculturation rates, new-comers with high initial cultural fit and either high(gray solid line) or low (gray dashed line) encultura-tion rates, and newcomers at the median level of ini-tial fit and with median enculturation rate (light graysparsely dashed line). Although those entering the firmwith high cultural fit are at lower risk of being asked toleave (with a one-standard-deviation increase in initialfit reducing the overall risk of involuntary exit by morethan 40%; see Table 1), the rate of initial enculturationcan offset the consequences of initial cultural fit. New-comers with initially low cultural fit who are quick toadapt (solid black line) fare better than those enteringwith median fit and who adapt at a median rate (graysparsely dashed line), or even those entering with highcultural fit but who are culturally inadaptable (dashedgray line). It appears that one’s capacity to enculturateis at least as important as one’s initial level of fit.

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Figure 5. Hazard of Involuntary Exit as a Function ofInitial Cultural Fit and Rate of Enculturation During aNewcomer’s First Six Months, Estimated with a CoxProportional Hazard Model

Note. Cumulative probability is plotted for different levels ofemployee’s initial cultural fit (low at the 25th percentile, median, andhigh at the 75th percentile, shade coded) and rate of enculturation(low at the 5th percentile, median, and high at the 95th percentile,line styling).

Discussion and ConclusionThe past three decades have seen the proliferation ofa vast and multifaceted literature on cultural fit andenculturation in organizations. Across these studies,one theme appears to be pervasive: those who areable to fit culturally enjoy significant benefits, whetherin psychological well-being, increased performance,favorable perceptions by colleagues, or likelihood ofretention. Indeed, these benefits accrue not only tothe individual in question but also to the organi-zation as a whole; contemporary firms consequentlyinvest considerable resources in cultural matching andenculturation. Using a language-based method formeasuring cultural fit that is more scalable, more eas-ily generalized across settings, higher in resolution,and less susceptible to biases than existing self-reportmeasures, we were able to discern these difficult-to-observe effects as they unfolded over time. Lendingfurther support to the claim that cultural compati-bility leads to attainment, our findings are consistentwith a large body of work on cultural fit in orga-nizations. Building on this vast literature, we use alanguage-based measure to provide further evidencethat cultural alignment is consequential for individualsurvival and success in an organization.

Yet our results go beyond reaffirming that cultural fitmatters; importantly, they also shed light on the pro-cesses by which individuals adapt to their colleagueswithin the organization. Measuring cultural fit overtime enables us to theorize about and empirically testpropositions related to a novel construct in socializa-tion research: enculturation trajectories. We find thatpeople are, on average, highly capable of encultura-tion and that different outcomes in the organizationare associated with unique enculturation trajectories.In other words, how people enculturate, not merelywhether they enculturate, matters for their integra-tion into the firm. Previous literature has tended toconflate enculturation processes related to acceptanceby others and those related to intrinsic attachmentby treating turnover as a one-dimensional outcome.By contrast, we distinguish between voluntary andinvoluntary exits and identify their different encultur-ation signatures. Newcomers who do not rapidly con-form to cultural norms are rejected by their colleaguesand ultimately forced to exit, whereas those who hadsuccessfully enculturated earlier in their careers butsubsequently exhibited a decline in cultural fit appearto detach from the organization and subsequently exitvoluntarily.

Organizational scholars have theorized extensivelyabout the dynamics and consequences of encultura-tion in organizations (Van Maanen and Schein 1979,Harrison and Carroll 2006, Wanous 1992, Bauer et al.1998). Because individual enculturation is difficult tomeasure reliably and consistently, however, empiricalwork has often treated cultural fit as a static end state.Thus, cultural matching has typically been studied asa selection process whereby an individual either fitsor does not fit culturally with an organization, andenculturation has been viewed as an early postentryprocess whereby an individual either adapts success-fully or fails to do so. Our findings are not inconsistentwith the view that a priori cultural fit, or early encul-turation, is consequential for eventual integration intoa firm. Rather, we too find that initial cultural fit andearly enculturation predict longevity at the firm. Yetthe implications of cultural compatibility are not lim-ited to entry. Variation in cultural fit at different stagesin a person’s tenure in an organization can provide awindow into different underlying mechanisms. Earlyin an individual’s tenure, low cultural fit is likely to beassociated with the failure to gain social acceptance bycolleagues; later, it is likely to reflect low attachment.This suggests that researchers and practitioners alikeshould pay more attention to enculturation trajectoriesas signatures of acceptance and attachment and as dif-ferentiated predictors of integration and attainment.

Organizational leaders have not been blind to corpo-rate culture. To the contrary, some argue that the pre-vailing tendency to cultivate strong corporate cultures

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


constitutes a managerial fad (Abrahamson 1996). Pop-ular depictions of cultural management have tended tofocus on screening on cultural fit or on early culturaltraining (for a recent example, see Bouton 2015), but,as our findings show, enculturation is an ongoing pro-cess. It therefore requires continuous cultivation. In anorganization that consciously invests significant timeand effort to hire on cultural fit, it is striking that weobserve tremendous variability in initial cultural fit.This seems to suggest that the individual differencesin cultural compatibility observed in the literaturemay not be merely a function of person–organizationfit but also of variance in enculturability—an individ-ual’s capacity for and susceptibility to enculturation.It remains unclear whether enculturability is a fixedindividual trait that newcomers bring with them toany new organization, whether it varies by individualexperience (for example, if newcomers without previ-ous work experience are more amenable to culturaltransmission; see Battilana and Dorado 2010), whetherit is context dependent and therefore a property ofthe person–organization relationship, or whether itchanges during “sensitive periods” when people areespecially likely to be imprinted by their social environ-ments (Marquis and Tilcsik 2013). Although we cannotexplore these questions further in our data, our find-ings suggest that identifying antecedents to encultur-ability may be as effective as hiring on cultural fit orposthire cultural training.Questions naturally arise about the causal rela-

tionships among individual enculturation, linguisticaccommodation, and attainment. It is conceivable,for example, that unobserved attributes of individ-uals are associated with their tendency to encultur-ate and linguistically accommodate others, as wellas their likelihood of achieving success in the orga-nization. Although we cannot conclusively rule outthese possibilities, the individual fixed-effects mod-els reported in panel (C) of Figure 4—which accountfor time-invariant, unobserved heterogeneity amongindividuals—partially mitigate such concerns aboutspuriousness. At the same time, however, and in keep-ing with general findings in sociolinguistics that lan-guage use and social identity are inseparable (Rickfordand Eckert 2001), it is likely that anticipated attainmentoutcomes have reciprocal effects on enculturation andlinguistic accommodation. Language use is both anoutcome and a cause: it reflects self-perceptions aboutone’s social standing, and it acts as an identity signalthat affects others’ judgments. Our findings are consis-tent with such amutually constitutive interplay amonglanguage, identity, and social outcomes. We treat lan-guage use as the behavioral signature of the complexprocesses that underlie organizational integration.

Although one should take caution in generalizingfindings based on observational data from a single

setting, we suspect that these patterns are likely toextend to other for-profit and nonprofit organizations.Whether because of measurement difficulty or theoret-ical focus, economic research has tended to downplaythe effects of cultural fit and adaptation on organi-zational success. Our findings suggest, however, thatvariability in cultural adaptability is consequential forindividual outcomes and, as others have shown, influ-ences organizational effectiveness (Weber and Camerer2003, Harrison and Carroll 2006, Van den Steen 2010).

Although firms are particular types of social sys-tems, we expect that our results will also apply in other,less formal group settings (e.g., Fine 1987). For exam-ple, similar to culturally inadaptable employees, schoolchildren incapable of cultural adaptation are probablyat higher risk of being rejected by their classmates. Aninverted U-shaped trajectory of cultural adaptation, onthe other hand, would likely indicate a child’s transi-tion into a different social milieu at school, similar to anemployee’s imminent voluntary departure. Indeed, theinteractional language use model we have developedcan be readily adapted to analyzing not only schoolsocialization but also a wide range of other socialdynamics and their implications for productivity. Forexample, analyses of the communication patterns ofscientists could help research centers in selecting indi-viduals for, and constructing teams that engage in,interdisciplinary research projects.

Individual-level measures of cultural fit and adapta-tion can also be aggregated to higher levels of analysisand, in similar fashion, have the potential to pave newtheoretical pathways about culture change in groupsand organizations. For example, cultural fit can be cal-culated not between a focal actor and a reference groupof all active interlocutors but instead between all pairsof individuals that constitute the organization. Thisdyad-level measure could then be aggregated to thelevel of functions, departments, or teams. Group-levelmeasures of fit could be used to inform organizationdesign choices—for example, determining which sub-units would be most culturally compatible with oneanother if they were combined or which departmentsactually consist of multiple, culturally fragmented sub-groups. In a similar fashion, dyad-level measures ofcultural fit could be aggregated to the level of organi-zations as a whole. Such measures could, for example,yield useful diagnostic information about the relativeease or difficulty of merging two firms. Computationalsociolinguistic techniques will continue to provide uswith novel ways of understanding these cultural pro-cesses and their impact on organizational dynamics.

AcknowledgmentsThe first two authors listed are joint first authors; otherauthors are listed in alphabetical order. The authors thankSoo Min Cho for research assistance, as well as Jennifer

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Chatman, Mathĳs de Vaan, Andreea Gorbatai, Stine Grodal,Ming Leung, Jo-Ellen Pozner, Jesper Sørensen, Glenn Carroll,Shai Bernstein, and participants of the Strategic ManagementSeminar at the University of Toronto, the InterdisciplinaryCommittee on Organizational Studies (ICOS) Seminar Seriesat the University of Michigan, the Strategy and Policy Sem-inar at the National University of Singapore, the StrategySeminar at INSEAD, and the Organization and ManagementSeminar at Yale School of Management for valuable input.

Appendix A. Illustration of the Cultural Fit ModelTo illustrate how our cultural fit measure works, we pro-vide two email examples. To protect the company’s and itsemployees’ identities, we draw these emails from two pub-licly available data sources: the WikiLeaks Sony Archive(available at https://wikileaks.org/sony/emails, accessedAugust 25, 2016) and the Enron email archive (avail-able at http://www.ferc.gov/industries/electric/indus-act/wec/enron/info-release.asp, accessed August 25, 2016). Forcomparison, we include one email sent by Amy Pascal, whowas chairperson of the Motion Pictures Group at Sony Pic-tures Entertainment at the time, and another by Kenneth Lay,chairman and chief executive officer of Enron at the time.The email contents and their normalized frequencies overLIWC categories are illustrated in Figure A.1. The two emailsclearly differ in content, tone, and style, which translates intodifferent normalized frequencies.

Our procedure iterates over all emails sent and receivedby focal individual i during period T to create outgoing andincoming probability distributions, as described above. Weillustrate a probability distribution over LIWC categories fora hypothetical set of incomingmessages, as well as two distri-butions of outgoing messages for two hypothetical people in

Figure A.1. (Color online) Examples of Emails and Their Normalized Frequencies Over LIWC

0 0.02 0.04 0.06 0.08 0.10

Normalized frequency

0 0.02 0.04 0.06 0.08 0.10

Normalized frequency

YouWork

WeVerbsTime

TentativeSwearSpaceSocial

3rd personSexual

SeeSad

RelativityQuantifierPronounPresent

PrepositionPersonal pronoun

Positive emotionPerceptual

PastNumbers

Negative emotionNegation

MotionMoneyLeisure

Impersonal pronounInsight

InhibitionInclusive

IHumans

HearFuture

FriendsFillerFeel

ExclusiveDiscrepancyConjunction

Cognitive processCertainCause

BodyBiologicalAux verb

ArticleAngerAffect

AdverbsAchievement

Sony

YouWork

WeVerbsTime

TentativeSwearSpaceSocial

3rd personSexual

SeeSad

RelativityQuantifierPronounPresent

PrepositionPersonal pronoun

Positive emotionPerceptual

PastNumbers

Negative emotionNegation

MotionMoneyLeisure

Impersonal pronounInsight

InhibitionInclusive

IHumans

HearFuture

FriendsFillerFeel

ExclusiveDiscrepancyConjunction

Cognitive processCertainCause

BodyBiologicalAux verb

ArticleAngerAffect

AdverbsAchievement

EnronSony

Enron

Notes. For illustration, two words are highlighted and mapped to their corresponding LIWC categories. “Total function word” categoryomitted for ease of presentation.

Figure A.2. (Color online) Examples of NormalizedDistributions Over LIWC Categories During a GivenMonth for a Hypothetical Reference Group and TwoHypothetical Individuals

0

0.05

0.10

Reference group LIWC distribution

0

0.1

0.2

0.3

Person A LIWC distribution JS = 0.5197 CF = 0.6546

0

0.05

0.10

0.15

Person B LIWC distribution JS = 0.159 CF = 1.8386

Figure A.2 (for illustration purposes, we choose uneven dis-tributions). We also report the JS divergences between thesetwo outgoing distributions and the incoming distribution, aswell as their corresponding levels of cultural fit (CF). As iseasily visible, person B’s distribution is more congruent withthe incoming reference group’s and consequently has lowerJS and higher CF than person A’s.

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

https://wikileaks.org/sony/emails

http://www.ferc.gov/industries/electric/indus-act/wec/enron/info-release.asp

http://www.ferc.gov/industries/electric/indus-act/wec/enron/info-release.asp


The examples above are illustrative. To help validateour measure of cultural fit, we conducted a supplemen-tal analysis using another publicly available data set:the Corpus of Contemporary American English (COCA;http://corpus.byu.edu/coca/). We applied our cultural fitmodel to the entire 2010–2012 “spoken” subcorpus, whichconsists of speech fragments from television and radioshows. Our analysis, reported in full in Goldberg et al. (2016),demonstrates that TV networks exhibit distinctive cultures.

Appendix B. Cultural ContentOur approach is generally agnostic to the cultural contentbeing exchanged, such that different individuals might bedetermined to have similar levels of cultural fit even if theyuse different words and communicate about distinct topics.What matters is the extent to which their communicationaligns with that of their interlocutors, as reflected in theirword distributions over LIWC categories.

Nevertheless, to further validate our approach, we imple-mented a backward selection analysis to identify the LIWCcategories thatmattermost for cultural fit in the specific orga-nization we study. The backward selection procedure ordersLIWC categories, weighted by their inverse frequency in theemail corpus, based on their relative contribution to vari-ance in individual cultural fit. Weighting term frequency byits inverse document frequency (conventionally referred toas term frequency-inverse document frequency, or TF-IDF inshort) is a common approach in linguistic analysis for identi-fying linguistic units with high information content.

Our iterative procedure removes LIWC categories one byone in order of their contribution to variance, until all cat-egories have been removed. In each step, the procedureidentifies the highest contributor to variance by runningmultiple regression analyses. Each regression estimates cul-tural fit with all remaining LIWC categories (weighted bytheir inverse document frequency), excluding a different cat-egory at a time. The LIWC category whose exclusion con-tributes to the greatest decline in R2 is then determined tobe the greatest contributor to variance for that step. At theend of the procedure, we arrayed the LIWC categories bythe magnitude of decline in R2 that their removal produced.The result of this procedure is reported in Table B.1.

Removing the “First person singular” category resultedin the biggest drop in R2. Removing “Sadness” resultedin the second biggest drop in R2, and so on. After cate-gory 18 (“Leisure”) the subsequent declines in R2 were nolonger statistically significant. Thus, we consider these 18 cat-egories as the most important for cultural fit. The fact thatcategories such as “Sadness,” “Friends,” “Death,” “Family,”“Social processes,” and “Negation” are on the list supportsthe view that our measure is measuring normative com-pliance, and potentially tapping more fundamental cogni-tive orientations. Organizational cultures (and subcultures)vary in the extents to which they implicitly allow or frownupon the expression of sadness or disagreement, discus-sions of death, and reference to friendship or other socialprocesses in email communication (as the illustrative exam-ples in Appendix A also demonstrate). Thus, it appearsour measure is indeed tapping into an important facet oforganizational culture and is indicative of individual-levelcultural fit.

Table B.1. Backward Selection Analysis

Column LIWC category

1 First person singular2 Sadness3 Friends4 Numbers5 Exclusive6 Article7 Future8 Relativity9 Family10 Causation11 Social processes12 Death13 Work14 Second person15 Quantifiers16 Fillers17 Negation18 Leisure

Endnotes1The term “socialization” is typically used to describe several dimen-sions of individual adjustment, which include role clarification, taskmastery, and cultural assimilation (Bauer et al. 2007). As Schneideret al. (2013) point out, the literatures on organizational culture andsocialization have grown increasingly apart in recent years. Workon socialization typically does not focus on cultural compatibility(Bauer and Erdogan 2014), whereas research on organizational cul-ture has tended to downplay processes of socialization. We use theterm “enculturation” because it specifically denotes the process ofcultural adjustment.2Longitudinal designs typically survey respondents in 4- to 12-month intervals, leaving much to be missed in between.3Self-report methods differ as to whether they elicit self-perceptionsof cultural fit (e.g., Chao et al. 1994) or use more indirect approaches(e.g., Chatman 1991). But because they invariably rely on data col-lected through surveys—as opposed to naturally occurring behav-ioral manifestations of cultural fit—they are all, to varying degrees,susceptible to measurement constraints. Scholars are naturally awareof these limitations (e.g., Bauer et al. 1998) and have devised inven-tive ways to overcome them. The Organizational Culture Profile(O’Reilly et al. 1991), for example, cleverly uses the Q-sort methodto elicit individual value orientations. This approach is neverthe-less resource intensive and therefore limited in granularity, relieson prominent informants to devise the parameters of organizationalculture, and is ultimately constrained by the dimensions containedin the survey.4Our distinction between behavioral and cognitive cultural fit is anal-ogous to de Saussure’s (1972) distinction between langue and parole.Interlocutors observe each other’s parole, the enactments of which aregoverned by a shared cognitive representation of langue. In that vein,spoken language can be thought of as the behavioral manifestationof cognitive cultural fit.5Swearing is the least common linguistic category in the email cor-pus used for this study. The use of swearwords, although rare, is thusan extremely strong linguistic indicator of cultural misalignment inthe organization used as our research site.6For a nuanced view of the seeming incompatibility between mas-culine and emotional cultures, see O’Neill and Rothbard (2015).7Our language-based measure taps into an important facet of behav-ioral cultural fit. It does not, however, encompass all aspects of

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

http://corpus.byu.edu/coca/


culture. For example, organizations have norms regarding dress andother nonverbal cues that are not captured by our measure.8For details about the LIWC lexicon, see Section 1 of the supplemen-tal material.9When O(l) → 0, the contribution to the summation nears zerobecause limx→0 x log2 x � 0.10We provide further illustration in Appendix A.11Descriptive statistics as well as additional details about the dataare provided in Section 2 of the supplemental material.12For ease of presentation, we report only results of interest through-out this section. For complete information about the models usedand the estimates they produce, see Section 3 of the supplementalmaterial.13Eight percent of employees are observed for longer than threeyears in our data, of which only three individuals departed, eithervoluntarily or involuntarily, after more than three years.14To rule out the possibility that this decline is caused by culturalchange at the organizational level, rather than at the individual level,we conducted an additional analysis with organizational culturalself-consistency as a control. We operationalize organizational cul-tural self-consistency as the organization’s cultural fit in the currentperiod relative to itself in the prior period. The estimates reported inFigure 4, panel (B) are unaffected by this specification.15We conducted an additional analysis to help rule out the possi-ble effects of changes in time-varying unobserved heterogeneity inindividuals’ capacity to enculturate. We proxy this capacity with ameasure of cultural self-consistency, which is operationalized as anindividual’s cultural fit in the current period relative to him- or her-self in the prior period. We find that adding this measure as a controlto the model reported in Figure 4, panel (B) does not substantiallyaffect the results.16When only departed employees are included inModel 1, the coeffi-cients for initial cultural fit and enculturation rate remain significant,suggesting that early fit and enculturation rate are significantly moreconsequential for involuntary exit than they are for voluntary exit.

ReferencesAbrahamson E (1996) Management fashion. Acad. Management Rev.

21(1):254–285.Adkins B, Caldwell D (2004) Firm or subgroup culture: Where does

fitting in matter most? J. Organ. Behav. 25(8):969–978.Allen DG (2006) Do organizational socialization tactics influ-

ence newcomer embeddedness and turnover? J. Management32(2):237–256.

Allen NJ, Meyer JP (1990) Organizational socialization tactics: A lon-gitudinal analysis of links to newcomers’ commitment and roleorientation. Acad. Management J. 33(4):847–858.

Ashford S, Nurmohamed S (2012) From past to present and intothe future: A hitchhiker’s guide to the socialization literature.Wanberg CR, ed. The Oxford Handbook of Organizational Socializa-tion (Oxford University Press, Oxford, UK), 8–24.

Ashforth BK, Saks AM (1996) Socialization tactics: Longitudinaleffects on newcomer adjustment. Acad. Management J. 39(1):149–178.

Aven BL (2015) The paradox of corrupt networks: An analysis oforganizational crime at Enron. Organ. Sci. 26(4):980–996.

Baayen RH (2001) Word Frequency Distributions (Kluwer AcademicPublishers, Dordrecht, The Netherlands).

Battilana J, Dorado S (2010) Building sustainable hybrid organiza-tions: The case of commercial microfinance organizations. Acad.Management J. 53(6):1419–1440.

Bauer TN, Erdogan B (2014) Delineating and reviewing the role ofnewcomer capital in organizational socialization. Annual Rev.Organ. Psych. Organ. Behav. 1(1):439–457.

Bauer TN, Morrison EW, Callister RR (1998) Socialization research:A review and directions for future research. Ferris GR, RowlandKM, eds. Research in Personnel and Human Resources Management,Vol. 16 (Emerald Group Publishing, Bingley, UK), 149–214.

Bauer TN, Bodner T, Erdogan B, Truxillo DM, Tucker JS (2007) New-comer adjustment during organizational socialization: A meta-analytic review of antecedents, outcomes, and methods. J. Appl.Psych. 92(3):707–721.

Bernstein BB (2003) Class, Codes and Control: Theoretical StudiesTowards a Sociology of Language (Routledge, New York).

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

Bouton K (2015) Recruiting for cultural fit.Harvard Bus. Rev. (July 17),https://hbr.org/2015/07/recruiting-for-cultural-fit.

Cable DM, Gino F, Staats BR (2013) Breaking them in or elicitingtheir best? Reframing socialization around newcomers authenticself-expression. Admin. Sci. Quart. 58(1):1–36.

Carley K (1991) A theory of group stability. Amer. Sociol. Rev.56(3):331–354.

Chao GT, O’Leary-Kelly AM, Wolf S, Klein HJ, Gardner PD (1994)Organizational socialization: Its content and consequences.J. Appl. Psych. 79(5):730–743.

Chatman JA (1991) Matching people and organizations: Selectionand socialization in public accounting firms. Admin. Sci. Quart.36(3):459–484.

Chatman JA, Caldwell DF, O’Reilly CA, Doerr B (2014) Parsing orga-nizational culture: How the norm for adaptability influencesthe relationship between culture consensus and financial perfor-mance in high-technology firms. J. Organ. Behav. 35(6):785–808.

Chatman JA, Polzer JT, Barsade SG, Neale MA (1998) Being differentyet feeling similar: The influence of demographic compositionand organizational culture on work processes and outcomes.Admin. Sci. Quart. 43(4):749–780.

Crémer J, Garicano L, Prat A (2007) Language and the theory of thefirm. Quart. J. Econom. 122(1):373–407.

Danescu-Niculescu-Mizil C, Lee L, Pang B, Kleinberg J (2012) Echoesof power: Language effects and power differences in socialinteraction. Proc. 21st World Wide Web Conf. (ACM, New York),699–708.

Danescu-Niculescu-Mizil C, West R, Jurafsky D, Leskovec J, Potts C(2013) No country for old members: User lifecycle and linguisticchange in online communities. Proc. 22nd Internat. Conf. WorldWide Web (ACM, New York), 307–318.

de Saussure F (1972) Course in General Linguistics (Open Court Pub-lishing Company, Peru, IL).

Elfenbein HA, O’Reilly CA (2007) Fitting in: The effects of relationaldemography and person-culture fit on group process and per-formance. Group Organ. Management 32(1):109–142.

Fine GA (1987) With the Boys: Little League Baseball and PreadolescentCulture (University of Chicago Press, Chicago).

Fine GA (1996) Kitchens: The Culture of Restaurant Work (University ofCalifornia Press, Oakland).

Giles H, Coupland N, Coupland J (1991) Accommodation theory:Communication, context, and consequence. Giles H, CouplandN, Coupland J, eds. Contexts of Accommodation (Cambridge Uni-versity Press, Cambridge, UK), 1–68.

Goffman E (1959) The Presentation of Self in Everyday Life (AnchorBooks, New York).

Goldberg A (2011) Mapping shared understandings using relationalclass analysis: The case of the cultural omnivore reexamined.Amer. J. Sociol. 116(5):1397–1436.

Goldberg A, Srivastava SB, Manian VG,MonroeW, Potts C (2016) Fit-ting in or standing out? The tradeoffs of structural and culturalembeddedness. Amer. Sociol. Rev. 81(6):1190–1222.

Grandey AA (2003) When “the show must go on”: Surface actingand deep acting as determinants of emotional exhaustion andpeer-rated service delivery. Acad. Management J. 46(1):86–96.

GreeneWH (2012) Econometric Analysis, 7th ed. (Prentice Hall, UpperSaddle River, NJ).

Greenwald AG, Banaji MR (1995) Implicit social cognition: Attitudes,self-esteem, and stereotypes. Psych. Rev. 102(1):4–27.

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

https://hbr.org/2015/07/recruiting-for-cultural-fit


Gumperz JJ (1982) Discourse Strategies (Cambridge University Press,Cambridge, UK).

Harrison JR, Carroll GR (2006) Culture and Demography in Organiza-tions (Princeton University Press, Princeton, NJ).

Hochschild AR (1979) Emotion work, feeling rules, and social struc-ture. Amer. J. Sociol. 85(3):551–575.

Hofstede G, Hofstede GJ, Minkov M (2010) Cultures and Organiza-tions: Software of the Mind, 3rd ed. (McGraw-Hill, New York).

Hughes JM, Fotia NJ, Krakauer DC, Rockmore DN (2012) Quantita-tive patterns of stylistic influence in the evolution of literature.Proc. Natl. Acad. Sci. USA 109(20):7682–7686.

Ireland ME, Slatcher RB, Eastwick PW, Scissors LE, Finkel EJ,Pennebaker JW (2011) Language style matching predicts rela-tionship initiation and stability. Psych. Sci. 22(1):39–44.

Jones GR (1986) Socialization tactics, self-efficacy, and newcom-ers’ adjustments to organizations. Acad. Management J. 29(2):262–279.

Kammeyer-Mueller JD, Wanberg CR (2003) Unwrapping the organi-zational entry process: Disentangling multiple antecedents andtheir pathways to adjustment. J. Appl. Psych. 88(5):779–794.

Kilduff M, Day DV (1994) Do chameleons get ahead? The effectsof self-monitoring on managerial careers. Acad. Management J.37(4):1047–1060.

Klein HJ, Weaver NA (2000) The effectiveness of an organizational-level orientation training program in the socialization of newhires. Personnel Psych. 53(1):47–66.

Kleinbaum AM, Stuart TE, Tushman ML (2013) Discretion withinconstraint: Homophily and structure in a formal organization.Organ. Sci. 24(5):1316–1336.

Kossinets G, Watts DJ (2006) Empirical analysis of an evolving socialnetwork. Science 311(5757):88–90.

Kotter JP, Heskett JL (1992) Corporate Culture and Performance (FreePress, New York).

Kreps DM (1990) Corporate culture and economic theory. AltJE, Shepsle KA, eds. Perspectives on Positive Political Economy(Cambridge University Press, Cambridge, UK), 90–143.

Kristof AL (1996) Person-organization fit: An integrative review ofits conceptualizations, measurement, and implications. Person-nel Psych. 49(1):1–49.

Kristof-Brown AL, Zimmerman RD, Johnson EC (2005) Con-sequences of individuals’ fit at work: A meta-analysis ofperson–job, person–organization, person–group, and person–supervisor fit. Personnel Psych. 58(2):281–342.

Kunda G (2006) Engineering Culture: Control and Commitment in aHigh-Tech Corporation (Temple University Press, Philadelphia).

Labov W (2001) Principles of Linguistic Change, Social Factors (Black-well Publishers, Oxford, UK).

Lewis D (1969) Convention: A Philosophical Study (Harvard UniversityPress, Cambridge, MA).

Lin J (1991) Divergence measures based on the Shannon entropy.IEEE Trans. Inform. Theory 37(1):145–151.

March JG (1991) Exploration and exploitation in organizationallearning. Organ. Sci. 2(1):71–87.

Marquis C, Tilcsik A (2013) Imprinting: Toward a multilevel theory.Acad. Management Ann. 7(1):195–245.

Martin J (1992) Cultures in Organizations: Three Perspectives (OxfordUniversity Press, Oxford, UK).

Martin J, KnopoffK, BeckmanC (1998) An alternative to bureaucraticimpersonality and emotional labor: Bounded emotionality at theBody Shop. Admin. Sci. Quart. 43(2):429–469.

Meglino BM, Ravlin EC, Adkins CL (1989) A work values approachto corporate culture: A field test of the value congruence pro-cess and its relationship to individual outcomes. J. Appl. Psych.74(3):424–432.

Mohr JW (1998) Measuring meaning structures. Annual Rev. Sociol.24:345–370.

Niederhoffer KG, Pennebaker JW (2002) Linguistic style matching insocial interaction. J. Language Soc. Psych. 21(4):337–360.

O’Neill O, Rothbard N (2015) Is love all you need? The effectsof emotional culture, suppression, and work-family conflicton firefighter risk taking and health. Acad. Management J.,ePub ahead of print December 14, http://dx.doi.org/10.5465/amj.2014.0952.

O’Reilly CA III, Chatman J, Caldwell DF (1991) People and orga-nizational culture: A profile comparison approach to assessingperson-organization fit. Acad. Management J. 34(3):487–516.

Pennebaker JW, Chung CK, Ireland M, Gonzalez A, Booth RJ (2007)The development and psychometric properties of LIWC2007.Technical report, LIWC.net, Austin, TX.

Piantadosi ST (2014) Zipf’s law in natural language: A critical reviewand future directions. Psychonomic Bull. Rev. 21(5):1112–1130.

Pinker S (2007) The Stuff of Thought: Language as a Window into HumanNature (Viking, New York).

Rafaeli A, Pratt MG (2006) Artifacts and Organizations: Beyond MereSymbolism (Lawrence Elbaum Associates, Mahwah, NJ).

Ravasi D, Schultz M (2006) Responding to organizational identitythreats: Exploring the role of organizational culture. Acad. Man-agement J. 49(3):433–458.

Rickford JR, Eckert P (2001) Style and Sociolinguistics Variation(Cambridge University Press, Cambridge, UK).

Rickford JR, Duncan GJ, Gennetian LA, Gou RY, Greene R, KatzLF, Kessler RC, et al. (2015) Neighborhood effects on use ofAfrican-American vernacular English. Proc. Natl. Acad. Sci. USA112(38):11817–11822.

Rivera LA (2012) Hiring as cultural matching: The case of elite pro-fessional service firms. Amer. Sociol. Rev. 77(6):999–1022.

Schein EH (2010) Organizational Culture and Leadership, 4th ed.(Jossey-Bass, San Francisco).

Schneider B, Ehrhart MG, Macey WH (2013) Organizational climateand culture. Annual Rev. Psych. 64(1):361–388.

Shipp AJ, Jansen KJ (2011) Reinterpreting time in fit theory: Craftingand recrafting narratives of fit in medias res. Acad. ManagementRev. 36(1):76–101.

Sørensen JB (2002) The strength of corporate culture and the reliabil-ity of firm performance. Admin. Sci. Quart. 47(1):70–91.

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

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

Van den Steen E (2010) Culture clash: The costs and benefits of homo-geneity. Management Sci. 56(10):1718–1738.

VanMaanen J (1975) Police socialization: A longitudinal examinationof job attitudes in an urban police department.Admin. Sci. Quart.20(2):207–228.

VanMaanen J (1991) The smile factory. Frost PJ, Moore LF, LouisMR,Lundberg CC, Martin J, eds. Reframing Organizational Culture(Sage Publications, , Newbury Park, CA), 58–76.

Van Maanen J, Barley SR (1984) Occupational communities: Cultureand control in organizations. Staw BM, ed. Research in Organiza-tional Behavior Vol. 6 (JAI Press, Greenwich, CT), 287–365.

Van Maanen J, Schein EH (1979) Toward of theory of organizationalsocialization. Staw BM, ed. Research in Organizational BehaviorVol. 1 (JAI Press, Greenwich, CT), 209–264.

Wanous JP (1992) Organizational Entry: Recruitment, Selection, Orien-tation, and Socialization of Newcomers, 2nd ed. (Addison-Wesley,Boston).

Weber RA, Camerer CF (2003) Cultural conflict and merger failure:An experimental approach. Management Sci. 49(4):400–415.

Willis PE (2014) Profane Culture (Princeton University Press, Prince-ton, NJ).

Zipf GK (1949) Human Behavior and the Principle of Least Effort(Addison-Wesley, Cambridge, MA).

Dow

nloa

ded

from

info

rms.

org

by [

171.

67.2

16.2

3] o

n 04

Mar

ch 2

017,

at 1

5:04

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

http://dx.doi.org/10.5465/amj.2014.0952

http://dx.doi.org/10.5465/amj.2014.0952

Enculturation Trajectories and Individual Attainment: An Interactional ...

Documents

Transcript of Enculturation Trajectories and Individual Attainment: An Interactional ...