getting the picture too late: handoffs and the

GETTING THE PICTURE TOO LATE: HANDOFFS AND THEEFFECTIVENESS OF IDEA IMPLEMENTATION IN

CREATIVE WORK

JUSTIN M. BERGStanford University

ALISA YUStanford University

Past research on idea implementation has focused on employees trying towin social sup-port for their own ideas, yet employees are often handed ideas to implement that weredeveloped by others. We propose and test hypotheses on such handoffs, focusing on howhanding employees relativelymature ideas to implementmay lead them to build less cre-ative finalproducts.Wetestedourhypothesesusing twostudies:anarchivalstudyof5,676moviesintheU.S.filmindustryandacomplementaryexperiment.Resultssuggestthat latehandoffs yielded less creative final products than no or early handoffs, meaning it wascostly to creativitywhenemployees implemented relativelymature ideaswithout drivingat least some of their prior development. However, “serialized late handoffs”—whereinimplementers were handed relatively mature ideas after an earlier handoff between twoother individuals—were lesscostly tocreativity than latehandoffs fromoneother individ-ual. Mediation results suggest that late handoffs reduced implementers’ creativity byrestrictingtheirsenseofpsychologicalownershipandthecoherenceoftheirfinalproducts.This researchadvances theoryonidea implementation,handoffs,andpsychologicalown-ership in creativework.

Scholars have long recognized that creativity andinnovation unfold in a multistage process (Amabile,1988; Anderson, Poto�cnik, & Zhou, 2014). This pro-cess begins with employees generating rough initialideas, thenelaborating initial ideas intomoredetailedplans, and eventually implementing the plans to pro-duce a finished product (Perry-Smith & Mannucci,2017). Thus, developing creative ideas is just thebeginning—ideasmustthenbeimplementedsuccess-fully to become creative final products. In organiza-tions, employees do not always implement theirown ideas. Rather, employees are often handed ideasto implement that were developed by someone else.Engineers may build products that they or othersdesigned (Hargadon & Sutton, 1997). Film directorsmay create films from screenplays that they or others

wrote (Baker & Faulkner, 1991). Marketers may exe-cute their ownor someoneelse’s idea for anewadver-tisingcampaign(Hirschman,1989).Organizationsareincreasingly crowdsourcing new ideas from employ-ees (Terwiesch & Ulrich, 2009) or from customersand other outsiders (Bayus, 2013), which employeesmay be asked to implement.

Despite the ubiquity of employees being handedideas to implement that they did not create, thissituation has been largely neglected in past research(Rouse,2013).Acommonassumptioninpriorresearchis that the same creators drive all stages of the creativeprocess, from idea generation through implementa-tion, overlooking the notion that ideasmay be handedoff between stages (Amabile & Pratt, 2016). As a result,past theory and research on idea implementation hasfocused on employees trying to implement their ownideas, as opposed to ideas developed by someone else(e.g., Baer, 2012; Fleming, Mingo, & Chen, 2007;Miron-Spektor, Erez, & Naveh, 2011; Perry-Smith &Mannucci, 2017; West, 2002). Furthermore, this priorworkhas focusedon thechallengesof employees gain-ing social support and recognition for their creativeideas.AsBaer (2012: 1102) explained, idea implemen-tation “is primarily a social–political process.” From

We thank Elizabeth Trinh for her pivotal help with datacollection. For research assistance, we thank the StanfordGraduate School of Business DARC group and BehavioralLab, especially Wonhee Lee, Mason Jiang, Daniel Choi, AlGourrier, Jr., and Anh Le. We are grateful to Bart de Jongand three anonymous reviewers for their very helpful feed-back in the review process. We are also grateful to BruceNash and OpusData for providing key data.

1191

Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder's expresswritten permission. Users may print, download, or email articles for individual use only.

rAcademy of Management Journal2021, Vol. 64, No. 4, 1191–1212.https://doi.org/10.5465/amj.2019.1330

https://doi.org/10.5465/amj.2019.1330

this view, idea implementation is about winningsocial acceptance for one’s creative ideas, which isoften challenging because creative ideasmay violateconventions or disrupt the status quo. Adopting thisperspective, scholars have built knowledge on thesocial–politicaldriversofeffectiveideaimplementa-tion, includingtheimpactofnetworkstructure(Baer,2012; Fleming et al., 2007; Perry-Smith &Mannucci,2017) and team dynamics (Miron-Spektor et al.,2011;West, 2002), in garnering support for one’s cre-ative ideas.

However, in addition towinning social support fortheir ideas, implementers must also build their ideasinto finished products (Perry-Smith & Mannucci,2017).Scholarshavepaidlessattentiontothisproduc-tion side of idea implementation, despite the recogni-tion that poor implementation may turn what wasonce a creative idea into a relatively uncreative finalproduct (King,1992;Paulus,2002).Tobedeemedcre-ative, ideas or final products must be judged as bothnovel and useful (Amabile, 1996). Creativity scholarstypicallymeasure creativity at the end of idea genera-tionorelaboration,beforenewideasare implementedinto full-fledged final products (Anderson et al.,2014). But turning new ideas into creative final prod-ucts is usually difficult, complex, and uncertainwork (Sternberg & Lubart, 1991). Following the logicof the planning fallacy (Buehler, Griffin, & Ross,1994),evenfairlyelaborateplansforcreativeproductswill have unforeseen holes and kinks that need to befilledinandfiguredoutas the finalproduct isactuallymade. When employees replicate final products thathavebeenbuiltbefore,theycanrelyoncompleteplansand existing exemplars. But when employees try tomake new ideas into creative final products, theyhave, at best, incomplete plans and cannot rely toomuch on existing exemplars without reducing thenovelty of their products (Smith, Ward, & Schu-macher,1993).Thus,successful idea implementationis not just about garnering social support for ideas—howeffectivelyimplementersbuildtheideasintotan-gible final productsmatters for creativity aswell.Yet,we know little about the conditions under whichimplementersareabletoeffectivelyconvertnewideasthat are not necessarily their own into creative finalproducts.

Drawing on theories of psychological ownership(Baer & Brown, 2012; Pierce, Kostova, & Dirks, 2001)and creative cognition (Ward, Smith, & Finke, 1999),wepropose that implementers’effectiveness inbuild-ingnew ideas into creative finalproductsdependsonwhenand fromwhomideas arehandedoff (if at all) tothe implementer. In particular, we introduce a

framework that differentiates handoff types based ontwo considerations: (1) how mature ideas are whenimplementers become the primary driver of the idea,and(2)whetherapriorhandoffbetweenotherindivid-uals took place before the implementer is handed theidea. We hypothesize that “late handoffs” are costlyto creativity, meaning that implementers build lesscreative final products when they are handed rela-tively mature ideas to implement without driving atleast some of their prior development. However, wealso hypothesize that “serialized late handoffs”—when implementers are handed relatively matureideas after an earlier handoff between two other indi-viduals—arelesscostlytocreativitythanlatehandoffsfrom one other individual. We tested our hypothesesusing two studies: an archival study of the U.S. filmindustry and a complementary experiment. Acrossthe two studies, results generally supported ourtheorizing.

This research advances theory on creativity andinnovationinorganizations inat least three importantways. First, we develop theory on the largely over-looked production side of idea implementation—that is, building ideas into tangible final products. Inso doing, our work complements the social–politicalfocus inprior research, providing amore comprehen-sive understanding of the drivers of successful ideaimplementation, including key interdependenciesbetween idea implementation and earlier stages ofthe creative process. Second, our work advances the-ory on handoffs in creative work by illuminatinghow andwhy handoffs to implementersmay becomeriskierwhentheyoccurlater(ratherthanearlier)inthecreativeprocess.Specifically,whenimplementersarehanded relatively mature ideas to implement, theylack the opportunity to develop a sense of psycholog-ical ownership or coherent vision for the emergingproduct, reducing the creativity of their final prod-ucts. Third, our research contributes to the small butgrowing literature on psychological ownership increativeworkbyelucidatinghowthestructuralcondi-tionsofhandoffsmayshapepsychologicalownershipand downstream creativity, with an emphasis on theimplementation stage of the creative process.

HANDOFFS, PSYCHOLOGICAL OWNERSHIP,AND PRODUCT COHERENCE

Whereaspastresearchhasfocusedonsocial–politicalaspectsofideaimplementation,wefocusonwhatPerry-SmithandMannucci(2017:59)calledthe“production”phase of idea implementation, in which “the idea isturned into something tangible—a finished product,

1192 Academy of Management Journal August

service,orprocess.” Individualswhodrive thisproduc-tionsideof implementation (i.e., implementers)maybehandedmature ideas tobuild that theydidnotdevelop.Receivingsuchhandoffsmaythreatentheirsenseofpsy-chological ownership over the ideas they are handed.Psychologicalownership is the“state inwhich individ-uals feel as though the target of ownership (material orimmaterial in nature) or a piece of it is ‘theirs’” (Pierceet al., 2001: 299). Past research has demonstrated thatfeelings of psychological ownership are associatedwith greater commitment, motivation, and persistencetowardthetargetofownership,suchasone’sjobororga-nization (Dawkins, Tian, Newman, & Martin, 2017;O’Driscoll, Pierce,& Coghlan, 2006;VanDyne&Pierce,2004). Given the association between creativity andmotivation/persistence (Amabile, 1996; Grant & Berry,2011; Lucas & Nordgren, 2015), psychological owner-ship is presumably a key driver of creativity.

Despite the seemingly important relationshipbetweenpsychological ownership and creativity, rel-ativelylittleresearchhas focusedonthis relationship.However, a fewstudieshave examinedpsychologicalownershipinthecontextofcreativework,providingahelpful foundation for the present research. In herqualitative study, Rouse (2013) took a relational per-spective, focusing on the dynamics between giversand receivers of handoffs in the creative process, andhow these dynamics shape their experiences of psy-chological ownership over the ideas that are handedoff (the impactofpsychologicalownershiponcreativ-itywasnot the focus,however).Thisworkhelpsmoti-vate our focus on handoffs as a potentially potentdriver of psychological ownership in creativework—employees are likely sensitive to ownershipover ideas following handoffs, and this is likely toshape their behavior. In a study at the team level,Gray, Knight, and Baer (2020) examined how leaderswho conceive an initial idea foster collective psycho-logical ownership among their followerswhodidnotgenerate the idea, finding that such feelings of collec-tive ownership predicted team performance. Thiswork focused on projects in which the leader drovethe whole process, from idea generation throughimplementation. We complement this past work byexamining psychological ownership and creativitywhen employees are the principal implementers ofideas that others have generated.

Otherpastworkhas lookedat the impactofpsycho-logicalownershiponthe typeof feedbackcreatorsuseor receive from others. In Baer and Brown’s (2012)work, creators with a sense of ownership over theirideas weremore likely to utilize feedback that addedto the core of their ideas, but less likely to utilize

feedback that subtracted from the core of their ideas.In Brown and Baer (2015), territorial marking ofone’s ideas (e.g.,“This ismyidea,notyours”) reducedthe creativity of others’ feedback when they had anindependentself-construal,butenhancedthecreativ-ityoftheir feedbackwhentheyhadaninterdependentself-construal. Although this past work hints that therelationship between psychological ownership andcreativity is meaningful, little is known about howpsychological ownership relates to downstreamcreativity.

Building on this prior work, we propose that psy-chologicalownership isvaluable forcreativityduringidea implementation, as implementers try to buildideas into creative final products. To turn new ideasinto creative final products, implementers likelyneedstrongconvictionforaboldvisionoftheproduct.Psychologicalownershipmaybeessentialtodevelop-ing such conviction. If implementers do not see theproduct as “theirs,” they may take shortcuts thatundermine their products in terms of novelty, useful-ness, or both—leading to a less creative final productoverall.

Along with psychological ownership, productcoherence may also be important during idea imple-mentation. Scholars have long noted the importanceof coherence in the evaluation of creative products(Besemer & Treffinger, 1981; Jackson & Messick,1965).Drawingonthispriorwork,wedefine“productcoherence”asthedegreetowhichthedifferentpartsorelementsoftheproductworkwell togetherasawhole.Beyond the general insight that coherence tends toenhance theperceivedcreativityof finalproducts, lit-tle research has addressed the role of product coher-ence in the creative process (for an exception, seeSeidel & O’Mahony, 2014). Extending priorwork,wepropose that product coherence plays an importantrole in driving howhandoffs to implementers impactthe creativity of their final products. Creativityinvolves bringing together previously disparate ele-ments of existing ideas into a novel anduseful combi-nation (Weick, 1979; Welch, 1946). For this reason,product coherence may be a principal challenge increative endeavors involving handoffs, as it may bedifficult for multiple individuals to integrate dispa-rate elements in a coherentway (Harvey, 2014; Seidel& O’Mahony, 2014), especially when the idea ishanded from person to person. Without a unifiedvision, handoffs may yield final products that lackcoherence,therebyunderminingcreativity,asobserv-ersareunlikely toseenoveltyandusefulness inahap-hazardcombinationofpartsorelements thatdonot fittogether in ameaningfulway.

2021 Berg and Yu 1193

Conversely, coherence can boost the creativity offinal products. Jackson andMessick (1965: 320) usedthe following analogy to explain the importance ofcoherence (or “condensation,” as they label it) forcreativity:

An assortment ofdebris gathered in a junkyard and theordered arrangement of the same material by an artistserves to illustrate the distinction being made … theordered arrangement, if it is worthy of artistic notice,containsmoremeaning than can be understood at firstglance.Thecolorandshapeof theobjects, their texture,their spatial location, and their original function allcombine to enhance their aesthetic appeal.

In short, product coherencemakes the creativity ofthewhole greater than the sum of its parts.

When implementers are handed relatively matureideas to build, it may be more difficult to develop asense of psychological ownership or coherent visionfor the product than if they are handed relativelyimmature ideas that have more room to be shapedprior to implementation. To distinguish betweentwo levels of maturity prior to implementation, weuse the following three stages of Perry-Smith andMannucci’s (2017) framework: (1) idea generation(generating a rough initial idea), (2) idea elaboration(fleshingout the initial idea intomoredetailedplans),and (3) idea implementation (building the final prod-uct). Distinguishing between idea generation andelaboration (prior to implementation) is useful fromboth a theoretical and practical standpoint. Theoreti-cally, thisdistinctionenablesanalysisofhowgenerat-ing the idea from inception versus elaboratingsomeoneelse’sideamayeachinfluencepsychologicalownership, product coherence, and ultimately crea-tivity. Practically, these three stages map onto legiti-mate milestones and deliverables in many creative

domains. For example, a musical artist may generatean initial idea for a new song, elaborate it into ademo recording, and then implement the song into acommercialized recording.1

Whereas Perry-Smith and Mannucci (2017: 55)“assume the creator remains the primary driver anddeveloper of his or her creative idea” throughout allstages, our framework focuses on whether and whenthe creator hands his or her idea off to the implemen-ter, meaning the individual who ultimately drivesimplementation.Continuingwiththemusicexample,the first one or two stages could be driven by a song-writer before the artist then becomes the principalimplementer of the song. Although our frameworkfocuses on the principal driver of each stage, thisdoes not necessarily mean only one individual isinvolvedineachstage. Inthemusicexample, theartistmay utilize input from their team of producers andsound engineers, but theywould still be theprincipaldriver of how the song is implemented (or, in somecases in the music industry, the principal driver ofimplementation is a producer, who also may or maynot drive the earlier stages). Thus, using these threestages allows our theorizing to address an importantpractical question: Are implementers more effectivewhen they drive all three stages, or can they takeover after generation or elaboration without sacrific-ing creativity?

Using these three stages, we propose four handofftypes (see Figure 1). First, “late handoffs” occur

FIGURE 1Handoff Types

Late Handoff

Generation Elaboration Implementation

Implementer

Implementer

Implementer

Implementer

Creator

Creator

Creator A Creator BSerialized Late Handoff

No Handoff

Early Handoff

1 For more examples of the three stages, see Table 2 inPerry-SmithandMannucci (2017:57).Theirmodel includesa fourth stage between idea elaboration and implementa-tion: ideachampioning (pitchingplans tokeystakeholders),whichweexcludedbecause thematurityoftheideadoesnotchange during this stage.


when the implementer is handed a mature idea aftersomeone else drives the generation and elaborationstages, meaning the implementer only drives theimplementation stage. Second, “no handoffs” occurwhen the implementer drives all three stages, fromgeneration through implementation. Third, “earlyhandoffs” occur when the implementer is handed aninitial idea after someone else drives the generationstage,meaningtheimplementerdrivesboththeelabo-rationand implementationstages.Fourth,“serializedlatehandoffs”occurwhenthe implementer ishandeda mature idea after the generation and elaborationstages are each driven by separate individuals(whereas late handoffs are from one individual whodrovebothgenerationandelaboration).Inthesectionsthat follow, we propose hypotheses on how andwhylatehandoffs lead implementers to build less creativefinal products than the other three handoff types.Wehypothesize that the negative effect of late handoffsoncreativity ismediatedbypsychological ownershipand product coherence (as depicted in Figure 2).

The Disadvantages of Late Handoffs versus No,Early, and Serialized Late Handoffs

Psychological ownership. When employeesreceive a late handoff, this means that they did notdrive the development of the mature ideas they arenow taskedwith implementing. Thismay put imple-menters who receive late handoffs at a disadvantagein terms of psychological ownership, compared toimplementerswhodrovebothgenerationandelabora-tion (no handoff) or just elaboration (early handoff)

prior to the implementation stage. Based on psycho-logical ownership theory (Pierce, Kostova, & Dirks,2003), individuals develop feelings of psychologicalownership through three routes: (1) controlling own-ership of the target, (2) intimately getting to know thetarget, and (3) investingtheself into the target.Drivingthegenerationandelaboration(orjust theelaboration)of ideasmay increase implementers’ senseof psycho-logical ownershipover the emergingproduct throughall three of these routes, as implementers are able toexert influence and shape the vision of the idea itselfduring these earlier stages (Baer & Brown, 2012).

Therefore, no and early handoffs should fostergreater psychological ownership before the imple-mentationstagethanlatehandoffs.Whenimplement-ers receive a late handoff, they are handed relativelydetailed plans for a product that they did not createbut now must build. At this point, it may be too latefor them to develop strong feelings of psychologicalownership over the emerging product, because theirabilitytoshapetheideahasbecomesubstantiallycon-strained. In contrast, when implementers receive noor early handoffs, they enter the process before initialideas become relatively detailed plans. This leavesplenty of opportunity for them to developpsycholog-ical ownership as they generate or elaborate initialideas to maturity. In sum, compared to no and earlyhandoffs, late handoffs may prevent implementersfrom developing psychological ownership over theemerging product prior to the implementation stage.

Late handoffsmay also lead to lower psychologicalownership thanserialized latehandoffs.Whereas latehandoffs occur when implementers are handed a

FIGURE 2Visual of Hypothesized Model

H2(–)Study 2

H3(–)Study 2

H1(–)Studies 1 & 2

(+)

Creativity ofFinal Product

(+)

ProductCoherence

PsychologicalOwnership

(a) No Handoff(b) Early Handoff(c) Serialized Late Handoff

Late Handoffvs.


mature idea fromone individualwhodrove both gen-erationandelaboration,serializedlatehandoffsoccurwhen implementers are handed a mature idea aftergeneration and elaborationwere eachdrivenby sepa-rate individuals. Although they are handed ideas ofequal maturity, recipients of serialized late handoffsmayfinditeasier todeveloppsychologicalownershiprelatively early in the implementation stage thanrecipients of late handoffs. Upon receiving handoffs,implementers are likely tocompare their opportunityforownershipto the individual(s)whodrovethepriordevelopment of the idea (Gray et al., 2020; Rouse,2013). When recipients of late handoffs comparethemselves to the individual who drove both genera-tionandelaboration, theymay feel that their opportu-nity to claim ownership over the idea is relativelyconstrained, as the individual already had ampleopportunity to control, get to know, and invest them-selves into the idea (Pierce et al., 2003).

Incontrast, recipientsof serialized latehandoffs arelikely to compare themselves to the two individualswho each drove only one prior stage (generation orelaboration). These two individuals each have asmaller claim to ownership than the referent individ-ual for latehandoffs,astheyeachhadlessopportunitytocontrol, get toknow, and invest themselves into theidea. Thus, ownership over the idea may seem more“up for grabs” to recipients of serialized late handoffsthan to recipients of late handoffs. This may make iteasier for recipients of serialized late handoffs todevelop psychological ownership—and reap the cor-respondingmotivationalbenefits—earlyintheimple-mentation stage. The key to serialized late handoffs isnotmerelytheinvolvementoftwoversusoneindivid-ual, but, rather, that two individuals eachdrivediffer-ent stages prior to the late handoff (if two individualsco-led the generation and elaboration stages together,their opportunity to control, get to know, and investthemselves into the idea would be more comparableto the referent individual for late handoffs from oneindividual).

Withoutastrongsenseofpsychologicalownership,recipients of late handoffs may lack the motivationand commitment that is needed to build new ideasinto creative finalproducts. Instead, theymaydefaulttothepathofleastcognitiveresistanceandditchnovelelements of the emerging product so they can rely onconventional exemplars and prototypes, undermin-ing the novelty of their final products (Kohn& Smith,2011;Smithetal.,1993;Ward,1994;Wardetal.,1999).Moreover,evenwithrelativelydetailedplans,unfore-seen gaps and problems are likely to arise with theproduct’sstructure,functionality,ortechnicalquality

as they build out the product (Buehler et al., 1994).Because they lack psychological ownership over theemerging product, recipients of late handoffs maytake shortcuts in addressing these unforeseen chal-lenges,underminingtheusefulnessoftheirfinalprod-ucts. Thus, as a result of their limited psychologicalownership over the mature ideas they are handed,recipientsoflatehandoffsmaybuildlesscreativefinalproducts than recipients of no, early, and serializedlate handoffs.

Hypothesis 1. Late handoffs have a negative effect oncreativity, such that implementers who receive latehandoffs build less creative final products than imple-menterswhoreceive (a)nohandoffs, (b)earlyhandoffs,and (c) serialized late handoffs.

Hypothesis 2. The negative effect of late handoffs oncreativity is mediated by psychological ownership,such that implementers who receive late handoffshave lower psychological ownership of the emergingproduct—leading them to build less creative finalproducts—than implementers who receive (a) nohandoffs, (b) early handoffs, and (c) serialized latehandoffs.

Product coherence. In addition to psychologicalownership, recipients of late handoffs may also facea disadvantage in terms of product coherence. Com-pared to recipients of no or early handoffs, recipientsof late handoffs may have less opportunity to form aunifiedvisionfortheproductpriortotheimplementa-tion stage. The generation stage ends, meaning theelaborationstagebegins,witha relatively abstract ini-tialidea.Theoriesofcreativecognitionsuggest that,asemployees flesh out abstract initial ideas with moreconcrete details, they engage in an iterative, interpre-tive process in which they develop a view on whatthey see as the core concept of the idea (Finke, 1996;Ward et al., 1999). This process of defining the coreconceptofthe ideamayhelprecipientsofnoandearlyhandoffs develop a relatively unified vision for theproduct before the implementation stage.We borrowWest’s (1990: 310) definition of “vision” as “a valuedoutcome which represents a higher order goal andmotivating force at work.” Entering the implementa-tionstagewitha relativelyunifiedvision for theprod-uctmayhelprecipientsofnoandearlyhandoffsbuildmore coherent—and thus more creative—final prod-ucts than recipients of late handoffs.

In addition, recipients of serialized late handoffsmay pursue a vision that better coheres with theirpredecessor’s vision than would recipients of latehandoffs. Given that the creative cognition literaturefocuses on individuals developing their own ideas,


this assertion ismore speculative than our theorizingregardingtheimpactofserializedlatehandoffsonpsy-chologicalownership.Nonetheless,wespeculatethatknowingthattheirrole is thelast inasequenceofthreeindividuals, as opposed to simply taking over an ideafrom one other individual, may compel recipients ofserialized late handoffs to stay on the same trajectoryas their predecessor. Because the ideas have presum-ablybeenshapedandvettedbytwootherpeopleinde-pendently, recipients of serialized late handoffs maysee their role as carrying forth their predecessor’svision, rather than putting their own spin on the ideathey are handed. Anchoring one’s thinking to others’ideas is often harmful to creativity during the earlystages of idea generation and elaboration (Kohn &Smith, 2011; Larey & Paulus, 1999). But, for ideaimplementation, recipientsofserializedlatehandoffsmay benefit from building on their predecessor, asstaying true to the core concept that their predecessorhad in mind may help them build a more coherentfinalproduct.Moreover,seeingtheirroleastheimple-menter of their predecessor’s vision may increasethe coherence of implementers’ products withoutsubstantially undermining their sense of psychologi-calownership,astheyhaveauniquesetofresponsibil-ities in the sequence of roles that enables themto control, get to know, and invest themselves intothe product as much or more than their twopredecessors.

In sum, as compared to the other three handofftypes (no, early, and serialized late), recipients oflate handoffs may lack commitment to a unifiedvision for the emerging product that they areexpected to finalize. Without a unified vision toguide how different elements or parts should cometogether as a whole during implementation, recipi-ents of late handoffs may go in their own direction,completing the plans they have been handed withparts or elements that are most easily retrieved fromtheir base of knowledge and experience (Dane,2010;Ward et al., 1999). This is likely to underminethe coherence—and thus creativity—of their finalproducts. In thisway, althoughcombining individu-als’ differing perspectives may be beneficial to crea-tivity during idea generation or elaboration (Paulus&Yang,2000), itmaybackfireduringideaimplemen-tation for recipients of late handoffs.

Hypothesis 3. The negative effect of late handoffs oncreativity is mediated by product coherence, suchthat implementers who receive late handoffs buildless coherent—and thus less creative—final productsthan implementers who receive (a) no handoffs, (b)early handoffs, and (c) serialized late handoffs.

STUDY 1: METHODS

Data Collection and Procedures

ForStudy1,weassembledanarchivaldatasetofthefilm industry in the United States, focusing on 5,676films released from 1990 to 2016. The goal of Study 1was to test the impact of handoff type on creativity(Hypothesis1) in a real-life creative industry.Follow-ing guidelines for archival research (Barnes, Dang,Leavitt, Guarana, & Uhlmann, 2018), we focused onthevariables that couldbevalidlymeasuredona rela-tively largescale:handoff typeandcreativity. Inaddi-tion,wecomplementthisarchivalstudywithStudy2,anexperimentaltestofthewholemodel,includingthetwomediation hypotheses (Hypotheses 2 and 3).

The film industry is representative of the contextsthat are the focus of our theorizing. Films are discretecreative projects that unfold through the three stagesin our theorizing (Perry-Smith & Mannucci, 2017),providing variance in handoff type. Film directorsare the principal driver of the implementation stage,but theymayormaynotdrive thegenerationandelab-oration stages, as other individuals may generate theinitial idea and/or write the screenplay. This mirrorsmanyothercreativedomains, including theexamplesgiveninthefirstparagraphofthispaper.Also,creativ-ity is valued—and explicitly evaluated—in the filmindustry,providingmeaningfulincentivestoproducecreative films (Elsbach&Kramer, 2003).

Toidentifythe5,676filmsincludedinouranalyses,we startedwith the list ofmovies inOpusData, a pro-fessionally assembled data set capturing feature filmreleases in the United States since 1915, which isroughly when the industry originated (Thomson,2006). This preliminary list included 24,850 filmsreleased from 1915 to 2016. Our main source of datafor the independent variable (handoff type) and con-trol variables was IMDb (Internet Movie Database).Data were available on IMDb for 82% of films on thepreliminary list, narrowing the number of films inthedatasetto20,326.FilmsmissingfromIMDbtendedto be relatively obscure, and thus were likely notmeaningful comparisons to the more mainstreamfilms included in the data set.

The set of 20,326 films was used to create controlvariables. However, the observations in our analysesare anarrower subset of5,676 films thatwere selectedto ensure the sample was appropriate for testing theproposed theorizing. Our approach was to select thebroadest possible subset of films that would enable avalid and relativelyprecise test ofHypothesis1.First,becauseourdependentvariablewascriticscores fromRotten Tomatoes (a review-aggregation website),


films without such scores were omitted (n 5 6,088).Second, tomitigate biases thatmay come from criticsreviewing films with established reputations, weomitted films released prior to 1990 (n5 2,568). Forfilms older than this cutoff, most of the critic reviewsthat Rotten Tomatoes scored were published manyyears after the original release of the film, usuallywhen the filmwas released on DVD, Blu-ray, or backto theaters. Third, we omitted films that had morethanoneindividualcreditedforanyofthethreestagesin our theorizing (n 5 5,957). This ensured that onepersonwastheprincipaldriverofeachstage, facilitat-ing consistent and fair comparisons between the fourhandoff types.

Next, to categorize the final sample of 5,676 filmsinto the four handoff types, we used films’ writeranddirectorcreditstodeterminetheindividualscred-ited for each of the three stages. TheWriters Guild ofAmerica has well-established rules for determiningwriter credits (see Writers Guild of America West,2018). The rules require thatwriter credits for a givenfilm distinguish between the writers who created theinitial story/characters (generation stage) and thosewho wrote the screenplay (elaboration stage), whiledirector credits (implementation stage) are listed sep-arately.Thus,usingthewriteranddirectorcreditsforagiven film,wedelineated the individuals credited foreach of the three stages. See OnlineAppendix A for amore detailed description of this process.

Films were late handoff if another person droveboth generation and elaboration before the directorthen drove implementation (n 5 1,604; 28.26%).Films were no handoff if the director drove all threestages (n5 2,821; 49.70%). Filmswere early handoffif another persondrove generation before thedirectorthendrovebothelaborationand implementation (n5483, 8.51%). Lastly, films were serial late handoff ifone person drove generation, then a second persondrove elaboration, before the director finally droveimplementation (n5 768; 13.53%).

Measures

Creativity. Following past research on the creativ-ityof feature films,webasedourmeasureof creativityon the evaluations of professional film critics(Plucker, Holden, & Neustadter, 2008; Simonton,2002, 2004, 2005). Creativity is at the core of whatfilm critics evaluate in their reviews, as they areexpected to assess films’ novelty and usefulness inproviding the intended form of entertainment (Bau-mann, 2001; Simonton, 2011). We collected criticscores from Rotten Tomatoes, a website that

aggregates critics’ reviews into composite scores.The scoring procedures used by Rotten Tomatoesalign with the Consensual Assessment Technique(CAT) for measuring creativity (Amabile, 1996). Theessence of the CAT is having raters with relevantexpertise independently rate the creativity of a givenset of products, and then creativity is captured by theaverage of their ratings. Rotten Tomatoesworks simi-larly: criticspublish reviews to their respectivepubli-cations independently, and then the websiteaggregates them into 0–10 composite scores for eachfilm. Rotten Tomatoes also provides a 0–100 percent-age scorereflecting theproportionofcriticswhoratedthe film positively versus negatively. We primarilyreport results for the 0–10 mean ratings, as the meanscores are more consistent with the CAT and moregranular than the percentage scores (e.g., a film withallmediocre reviews could score 4.9/10but 0/100%).

To examine the validity of using Rotten Tomatoesscores to measure creativity, we had independentraters assess the extent to which creativity wasreflected in a subset of critic reviews scoredbyRottenTomatoes. In particular, we collected independentratingsof2,000criticreviewson200films (10reviewsper film).Werecruited2,000 independent raters,whoeach assessed 20 critic reviews. Results from theseindependent ratings support thevalidityofusingRot-tenTomatoesscores tomeasurecreativity.SeeOnlineAppendix B for details on the data collection proce-dures and full results, as well as excerpts from criticreviews that describe high or low levels of creativity.Lastly, to address the possibility that results could bebiased by idiosyncrasies in Rotten Tomatoes’ scoringsystem, we also collected all available critic scoresfrom MetaCritic (an aggregator of critic reviews likeRotten Tomatoes) and user/audience scores fromIMDb.Wereport resultsusingtheseadditionaldepen-dent variables, including the 0–100% Rotten Toma-toes scores, as supplementary analyses.

Controls. We created control variables to help ruleout potential confounds (see Table 1 for descriptivestatistics). First, given that some films were scoredby many more critics than others, we collected thenumber of critics who reviewed each film on RottenTomatoes. Second, past research on creators’ careershighlights the importance of experience and quantityof products (e.g., Kozbelt, 2008; Simonton, 1997).Thus,we controlled for the cumulative count of filmsfor which the director andwriter(s) received directoror writer credits prior to the focal film. Third, giventhatpriorcollaborationsmay facilitate intersubjectiv-ity between writer and director (Seidel & O’Mahony,2014), we controlled for the mean number of films


https://osf.io/vgwme/?view_only&hx003D;d5711f0f492a4a168a14a9a8958cc883



onwhichthedirectorhadworkedwitheachwriteronthefilm.Fourth,tohelpcontrolforhistoryeffectssuchas cumulative advantage (Merton, 1968), we calcu-lated the cumulativemean critic score for thedirectorandwriter(s)priortothefocalfilm,andwealsocreateda dummy for whether the film was a sequel. When itwas a director’s first film or the film had no writersapart from the director (no handoff), the samplemean critic scorewas used. Fifth, in light of evidenceon the importance of genre in the film industry (Hsu,2006), we created dummies for each of 23 genresfrom IMDb represented in the sample. Sixth, to helpcontrol for possible changes in the film industry orRotten Tomatoes’ method over time, we created setsof dummies for the release year of the film(1990–2016), and directors’ career start year, whichwas the earliest year they received awriter or directorcredit (1957–2016).Lastly,giventhatstudioscanvarywidely in their goals and resources, we collected theprimary studio associated with each film—betweenmultiple sources, we identified a studio for 98% offilms (the remaining 2%were treated as if they werefrom the same studio in the analyses).

Results and Discussion

Themodels in Table 2 test the three components ofHypothesis1.Toaccountforthefactthatfilms(level1)were cross-nested in directors (level 2) and studios(level 3), we used random-interceptmodels (Rauden-bush & Bryk, 2002).Model 1 tests Hypothesis 1 with-out fixed controls, Model 2 includes controlsexcept for the three sets of dummies, and Model 3includes all controls. To facilitate interpretation, all

continuous controlswere standardized at the samplelevel, while the dependent variable (creativity mea-sured byRottenTomatoes score)was kept in its origi-nal 0–10 scale.

Across all models, films with directors whoreceived late handoffs scored significantly lower increativity than films with directors who received no(Hypothesis 1a), early (Hypothesis 1b), and serializedlate (Hypothesis 1c) handoffs. In particular, plannedcontrasts using the estimates from Model 1 showedthat late handoffs (M5 5.46,SE5 0.05) scored signif-icantly lowerthantheother threehandoff types:no(M55.98,SE50.04),p, .001,d5 .38;early(M56.01,SE50.07),p, .001,d5 .40;andserializedlate(M55.71,SE5 0.06),p, .001,d5 .18 (seeFigure3). In termsofabsolute differences, late handoffs scored 0.52 pointsbelow no handoffs, 0.55 points below early handoffs,and 0.25 points below serialized late handoffs. Tohelp put these differences in context, the elite groupof 87 films (out of 5,676) that were nominated for theAcademy Award for Best Picture scored 2.00 pointshigher on average than the rest of the sample (SE 50.13, p , .001).2 Using 2.00 points as a benchmarkfor a very large practical difference, the point differ-ences between late and the other three handoff typesseem large enough to carry practical significance.But, given the effect sizes, the disadvantage of latehandoffs is likely more meaningful in comparison tonoandearlyhandoffs than to serialized late handoffs.Although the effectswere generally largerwith fewer

TABLE 1Study 1: Correlations and Descriptive Statistics

Variable 1 2 3 4 5 6 7 8

1. Creativity (RT critic score)2. Number of critics .243. Prior film count for director 2.01 .364. Prior film count for writer(s) 2.06 .16 .165. Prior experience with writer(s) .02 .07 .15 .256. Prior RT critic score mean for director .32 .16 .06 2.01 .057. Prior RT critic score mean for writer(s) .19 .01 2.02 2.03 .09 .158. Sequel (yes vs. no) 2.10 .08 .05 .12 .11 2.09 2.09

Mean 5.81 59.49 2.89 1.27 1.06 6.08 6.03 0.03Standard deviation 1.38 60.26 4.77 3.85 0.38 0.81 0.52 0.17Minimum 0.00 5 0 0 1 1.40 2.30 0Maximum 9.20 400 46 54 9 9.20 8.80 1Median 5.90 34 1 0 1 6.07 6.07 0

Notes: Correlations greater than .03 and less than2.03were significant at p, .05. For brevity, the three lengthy sets of dummies are excluded(see Table C1 in Online Appendix C for complete lists of these dummies).

2 ThisdifferencewascalculatedusingModel 1 inTable2,with a dummy for award nomination.



controls (Models1and2), the results remainedsignif-icant with all controls included (Model 3).

Supplementary analyses.Weran additionalmod-elstotestkeyalternativeexplanationsandrobustness.In general, these supplementary results provide fur-ther support for Hypothesis 1, with some minor cav-eats noted. The full models are reported in OnlineAppendixC, butwe summarize somehigh-level take-aways here. First, supplementary analyses helpedruleout thepossibility that theadvantageofserializedlate(vs.late)handoffswasduetobenefitsofleveragingthe work of two writers rather than one (Table C2).

Second, the pattern of results was largely consistentusing the aforementioned supplementary dependentvariables (Table C3). Third, additional analyseshelpedruleout thepossibility that resultsweremean-ingfully biasedbyhighly prolific directors (Tables C4and C5), the exclusion of films released prior to 1990(Table C6), or films’ financial budgets (Table C7).

In sum, the results for Study 1 provide supportfor Hypothesis 1 in a real-world context. However,Study 1has a fewkey limitations,whichwedesignedStudy 2 to address. First, Study2was designed to testthe proposed mediators—psychological ownership

TABLE 2Study 1: Random Intercept Models Testing (Hypothesis 1)

Model 1No Controls

Model 2Some Controls(no dummies)

Model 3All Controls

(with dummies)

Handoff type (vs. Late handoff)No handoff (Hypothesis 1a) .52�� .42�� .38��

(.04) (.04) (.04)Early handoff (Hypothesis 1b) .56�� .34�� .29��

(.07) (.06) (.06)Serialized late handoff (Hypothesis 1c) .26�� .18�� .13��

(.06) (.05) (.05)Controls

Number of critics .46�� .66��

(.02) (.02)Prior film count for director 2.10�� 2.05

(.02) (.03)Prior film count for writer(s) 2.05�� 2.03�

(.02) (.01)Prior experience with writer(s) .03 .03�

(.02) (.02)Prior RT critic score mean for director .27�� .17��

(.02) (.02)Prior RT critic score mean for writer(s) .14�� .10��

(.02) (.01)Sequel (yes vs. no) 2.47�� 2.22�

(.09) (.09)Genre dummies Includeda

Release year dummies Includeda

Director career start year dummies Includeda

Fixed intercept 5.46�� 5.69�� 7.98��

(.05) (.04) (1.06)Random intercept: Level 2 (Director) .65�� .26�� .21��

(.06) (.06) (.06)Random intercept: Level 3 (Studio) .30�� .29�� .25��

(.04) (.04) (.03)Observations (Films) 5,676 5,676 5,676Directors 3,535 3,535 3,535Studios 996 996 996Log-likelihood 29,550.50 29,015.94 28,491.63

Notes: Standard errors in parentheses. All continuous control variables were standardized.a Results for these lengthy sets of dummies were excluded for brevity (see Table C1 in Online Appendix C for full results).� p , .05�� p , .01�� p , .001





(Hypothesis 2) and product coherence (Hypothesis3)—whichcouldnotbedirectlytestedinStudy1.Sec-ond,becauseStudy1wascross-sectional,Hypothesis1 could not be tested causally. Study 2 is an experi-ment designed to enable causal inferences. Third,the role of film director may be unique compared toimplementer roles in other industries, as the filmindustry has a relatively specific and strong culture.Thus, we designed Study 2 to simulate handoffs inmore typical implementerroles.Fourth,althoughcre-ativity is at the heart of what film critics evaluate(Simonton,2011),andtheindependentratingswecol-lected suggest that the Rotten Tomatoes scores reflectcreativity, this measure may still be relatively noisycompared to more direct measures of creativity. Weaddress this bymeasuring creativity more directly inStudy 2.

STUDY 2: METHODS

Participants and Procedures

This second study was an experiment to enablecausal tests of Hypotheses 1 to 3, complementing thereal-world archival data from Study 1. Participantsincluded 600 individuals in the United States, ofwhom 53.33% were female and their ages rangedfrom 18 to 71 (Mage 5 35.36, SDage 5 10.32). They

wererecruitedviaAmazonMTurk,andwerecompen-sated $6.00 for their participation.

Wedesignedtheexperimenttofosterbothmundaneandpsychological realism (Berkowitz&Donnerstein,1982), while following current recommendations forensuring high-quality data with online participants(Aguinis, Villamor, & Ramani, 2021; Porter, Outlaw,Gale,&Cho,2019).Tofostermundanerealism,thecre-ative task given to participants simulated the sort ofcreativetaskthatemployeesmayexperienceinorgan-izations. The task was developing an advertisementfor a new food product—a soft pretzel made withcricket flour. This task was divided into separatestages (each described in more detail below), mirror-inghowthecreativeprocessoftenunfoldsinorganiza-tions(Amabile,1988;Perry-Smith&Mannucci,2017).Along with these sources of mundane realism, psy-chological realism was fostered in three main ways.First, the creative task embedded participants in adomain that they were likely to care about and knowwell:consumerfoodproducts.Second, tohelpensuretask significance was high across conditions (Grant,2008; Hackman & Oldham, 1976), participants wereall given the same background information extollingthehealthandenvironmentalbenefitsofusingcricketflour in food products (see Online Appendix D).Third, in linewith research suggesting that creativity

FIGURE 3Study 1: Mean Creativity by Handoff Type

Cre

ativ

ity

(Z

-Sco

re o

f E

stim

ated

Mea

n)

0.5

0.4

0.3

0.2

0.1

–0.1

–0.2

–0.3

–0.4

–0.5

0.0

Late No Early Serialized Late

Notes: Z-scores are based on the estimatedmarginal means fromModel 1 in Table 2. Error bars are61 SE.



is enhanced by setting the explicit goal to producenovel anduseful content (Shalley, 1991),participantswere given the explicit goal to be creative,whichwasreinforcedwith a financial incentive. Theywere told:

The goal is for your presentation to be creative, mean-ing both novel and useful for convincing consumersto purchase the product. A large group of consumerswill rate the creativity of all the presentations we col-lect. Ifyourpresentationscoresinthetop10%ofpartic-ipants, you’ll receive a $10 bonus.

Whenstudiesarewelldesignedforonlineplatformssuch as MTurk, these platforms can provide high-qualitydatafromsamplesthataremorerepresentativeof the broader population than in-person lab research(Aguinisetal.,2021;Porteretal.,2019;Walter,Seibert,Goering, & O’Boyle, 2019). Our experiment wasdesigned to ensure that we collected high-qualitydata. Because the task was generating open-endedcontent, participants could not complete the taskwithout actually paying attention (their final adswere presentations consisting of three slides).3More-over, the manipulations were an inherent part of thetask,making ithighlyunlikely thatparticipantscouldoverlook their respective manipulations and stillcomplete the task. Additionally, the task did not fol-low an existing experimental paradigm, mitigatingthe risk thatparticipants’ experiences inprior studiescould bias the results.

The experiment simulated the four handoff types:late, no, early, and serialized late. These four condi-tionswereyokedtooneanother,suchthatparticipantsin the late, early, and serialized late conditions werehanded ideas/plans created by participants in the nohandoff condition. Because the no handoff conditionhad to be run first to collect the ideas/plans thatwerehanded to participants in the other three conditions,the nohandoff conditionwas not randomly assigned.To address this, we included a separate no handoffcondition that was randomly assigned. Thus, theexperiment included a total of five conditions—fouryoked conditions and a separate no handoff condi-tion—eachwith 120 participants.

Thecreativetaskwasdividedintothethreestagesinour proposed theorizing (generation, elaboration,implementation). For each stage, the survey would

not advance until 10 minutes had passed, althoughparticipants could spend more time if they wanted(see Online Appendix F for analyses on time spentby stage/condition). As aforementioned, we ran theyoked no handoff condition first to collect contentthat was handed to participants in the other threeyoked conditions (late, early, and serialized late).Participants in the no handoff condition completedall three stages of the creative task. In the generationstage, participants were asked to generate an initialidea for their advertisement. They were told thattheir initial idea should capture the general themeor story for their ad that theywouldeventuallycreateas a slide presentation consisting of three slides.They were also told that their initial idea could berough and incomplete, but it should be understand-able to someone else reading it. To help prevent par-ticipants from converging on an obvious idea tooquickly (Ward et al., 1999), they were encouragedto spend time brainstorming different possibilitiesbefore they committed to one.

Next, in the elaboration stage, participants wereasked to flesh out their initial idea in a more detailedoutline of their ad. Their initial idea was carried for-ward to this survey page for their reference. Theywere given three blank fields to create their outline,one for each slide in their ad. They were told that,although theydidnotneed to cover everydetail, theiroutlineshouldbeafairlycompleteplanandshouldbeunderstandable tosomeoneelse reading it.Theywerealso told: “The initial idea is a starting point to buildon—you can deviate from it if you’d like.” Lastly, inthe implementation stage, participants created theirfinalad,whichconsistedofathree-slidepresentation.Their initial idea and outlinewere carried forward tothis survey page for their reference. They were told:“Feel free to add any images or whatever contentyou’d like to your presentation. The outline is a start-ing point to build on—you can deviate from it ifyou'd like.”

Oncewecollected responses from120participantsintheyokednohandoffcondition,theinitialideasandoutlinesthattheseparticipantscreatedwerethenusedin the other three yoked conditions (late, early, andserialized late).These threeconditionswere identicalto the no handoff condition, except for the followingdifferences. In the late condition, participants startedat the implementation stage, meaning they jumpedstraight to creating their actual slide presentation.They were “handed” an initial idea and outline fromone of the 120 no handoff participants, and told thattheir role was to pick up where this past participanthad left off. In theearlycondition,participantsstarted

3 A total of 39 individuals completed the survey but didnot submit an advertisement as instructed—they submitteda bogus file so the survey would allow them to advance.These 39 improper submissions were replaced. The fiveconditions did not significantly differ in terms of impropersubmissions or incomplete surveys.



at the elaboration stage, meaning their first stage wascreating an outline. They were “handed” an initialidea from one of the 120 no handoff participants andtold that their rolewas to pick upwhere this past par-ticipant had left off. In the serialized late condition,participants started at the implementation stage (cre-ating their actual slide presentation). Similar to thelate condition, they were “handed” an initial ideaand outline from one of the 120 no handoff partici-pants. However, unlike the late condition, they weretold that the initial idea and the outline had eachbeen created by a different past participant. This wasreinforced with headers that read “Initial Idea fromPast Participant #1” and “Outline from Past Partici-pant #2.” See Online Appendix D for examples of aninitial idea, outline, and two final advertisements—one advertisement that was rated high in creativityand another that was rated low in creativity (bothwere created by participants who were handed thesame initial idea and outline).

Alimitationofthisyokeddesignisthatthenohand-off conditionwasnot randomlyassigned,aswehadtofinishdatacollection for this condition first.Thus,weincluded a separate no handoff condition in whichparticipants went through the same procedures asthe yoked no handoff condition. Once we reached120 participants in the yoked no handoff condition,all subsequent participants were randomly assignedto one of the other four conditions (the three yokedconditions or the separate no handoff condition).The three yoked conditions each needed 120 partici-pants,asahandoffpartnerwasneededforeachpartic-ipant from the yoked no handoff condition. Forconsistency, we also targeted 120 participants in theseparate no handoff condition. This meant that wehad480slotstofill,120foreachofthefourconditions.The survey was programmed to randomly assign(without replacement)oneof these480 slots topartic-ipants, limiting each slot to one participant.

The same posting onMTurkwas used to recruit all600 participants across all five conditions. Once wereached 120 yoked no handoff participants, wechanged the survey link to lead to a revised surveythat randomly assigned participants to the remainingfourconditions.Allparticipantsweretoldthatthesur-vey should take them about 45 minutes, and that thesurvey would consist of two parts, the first of whichwas the creative task (the second was a post-survey).However, participants were not told how long eachofthetwosurveypartswouldtake.Thishelpedensurethatparticipantsexpected the samehourlypay rateasthey worked on the creative task, regardless of theircondition.

Measures

Creativity andproduct coherence.Tomeasure thecreativityandproductcoherenceofparticipants’ finalads, we used the CAT approach (Amabile, 1996). Toavoid common method biases (Podsakoff, MacKen-zie,Lee,&Podsakoff,2003), independentsetsofratersassessedcreativityandproductcoherence.Across thetwomeasures, a total of 600 consumers in the UnitedStates was recruited throughMTurk to serve as inde-pendent raters, 45.67% of whom were female andwith ages ranging from 19 to 74 (Mage 5 36.39, SDage

5 10.76). Each raterwas compensated $6.00.Given the large volume of content to be rated, we

randomly divided the ads into separate rater groups.Specifically, participants’ 600 final ads were ran-domlydivided into10groupsof60ads.Foreachmea-sure(creativityandproductcoherence),30raterswererandomly assigned to each of the 10 groups. Thus,each rater assessed 60 ads and each ad was rated by30 raters for creativity and 30 separate raters for prod-uct coherence (respondents who failed attentioncheckswere discarded and replaced until each grouphad 30 valid responses).4 Each group of 60 adsincluded 12 ads from each of the five conditions.Ads from the same yoked groupwere assigned to dif-ferent rater groups,whichprevented raters fromeval-uatingmultiple ads that derived from the same initialidea/outline. This grouping approach enabled us tocalculate interrater reliability and agreement acrossallratergroups,ICC(1,30),aswellasinterraterreliabil-itywithineachgroupof30raters, ICC(2,30) (LeBreton& Senter, 2008; Shrout & Fleiss, 1979). All measuresand rater groups met conventional standards forinterrater reliability and agreement. For creativity,ICC(1,30)5 .91 and ICC(2,30) ranged from .91 to .96.For product coherence, ICC(1,30)5 .79 and ICC(2,30)ranged from .81 to .91.

Consistent with how creativity was defined to par-ticipants in the experiment, creativitywas defined toraters as “Overall degree towhich the ad is bothnovelanduseful,”whichtheyevaluatedona7-pointscale(15 extremely low, 7 5 extremely high). Raters alsoassessed novelty and usefulness separately so thatwe could examine how each dimension compared tooverall creativity (e.g.,Berg,2014;Rietzschel,Nijstad,

4 We included two attention checks in each rater group.Pages asking the rater to fill in either all 2s or all 5swere dis-played,which otherwise looked like the surveypages raterscompleted.AcrossmeasuresinStudy2,32%ofrespondentsfailed one or both attention checks andwere thus discardedand replaced.



& Stroebe, 2010). Novelty was defined as “Degree towhich the ad is unique.” Usefulness was defined as“Degree towhichthead iseffective.”Meansbycondi-tionfornoveltyandusefulnessseparatelyarereportedin Online Appendix E. Product coherence wasdefined to raters as the “Degree towhich the differentparts or elements of the ad work well together as awhole,”which they rated using the same scale as thecreativity raters.

As expected, overall creativity was highly corre-latedwith bothnovelty (r5 .91,p, .001) and useful-ness (r 5 .85, p , .001). Also as expected, productcoherence was significantly correlated with overallcreativity (r5 .70,p, .001). Supporting the discrim-inant validity of product coherence versus creativity,product coherence was substantially less correlatedwith novelty (r5 .53, p, .001) than usefulness (r5.81, p , .001). This makes conceptual sense, as wewould expect product coherence to correlate withboth dimensions but to diverge more from noveltythanusefulness(whereasoverallcreativitywashighlyconvergent with both novelty and usefulness). SeeOnline Appendix E for confirmatory factor analysesthat further support thediscriminantvalidityofprod-uct coherence versus overall creativity.

Because the proposed theorizing focuses on theimplementation stage, it is important that the experi-mentaldesigncanisolate the impactof the implemen-tation stage on the final product. To this end,we usedsimilar procedures to collect independent ratings ofthe creativity and coherence of participants’ initialideas andoutlines (seeOnlineAppendixF for furtherdetail). These supplementary data suggest that theconditions did not differ in creativity or coherenceheading into the implementation stage, meaning thatany significant differences between conditions couldbe attributed to participants’ behavior during theimplementation stage.

Psychological ownership. We measured psycho-logicalownershipusing six itemsadapted frommeas-ures of psychological ownership in prior research(Baer & Brown, 2012; Pierce, O’Driscoll, & Coghlan,2004; Van Dyne & Pierce, 2004): “I feel a high degreeof personal ownership over this presentation”; “Isense that this presentation is mine”; “I feel like thisismypresentation”;“Idonotfeelasenseofownershipover thispresentation” (reverse-scored);“It ishardforme to think of this presentation as mine” (reverse-scored); and “This presentation belongs to me” (a 5.93). Participants rated these items directly after sub-mittingtheir finaladsusinga7-pointLikert-typescale(1 5 strongly disagree, 7 5 strongly agree). Asexpected, psychological ownershipwas significantly

correlatedwith creativity (r5 .27, p, .001). Psycho-logical ownership alsohad a significant but relativelysmall correlation with product coherence (r 5 .13,p5 .002).

Results and Discussion

To account for clustering of the four yoked condi-tions,weusedrandominterceptmodels(Raudenbush&Bryk,2002).Arandominterceptforyokedgroupwasincluded in each model. The 480 participants in thefour yoked conditions were clustered into 120 yokedgroups(coded1–120).The120participantsinthesep-aratenohandoffconditionwereeachassignedtotheirown yoked group (coded 121–240). This approachaccounted for the clustered nature of the data whileallowing all 600 participants to be included in thesame model(s). Results were virtually identicalwhentheseparatenohandoffconditionwasanalyzedindependently from the four yoked conditions, andthus we report models with all five conditionsincluded for conciseness.

Negative effect of late handoffs on creativity(Hypothesis 1). Model 1 in Table 3 tests Hypothesis1,whichpositedthatthelatehandoffconditionwouldfinish lower in creativity than the other four condi-tions. InsupportofHypothesis1,latehandofffinishedsignificantly lower in creativity thaneachof theotherfourconditions (seemeansinTable4):no (yoked),b5.32,p5 .002,d5 .37; early,b5 .28,p5 .006,d5 .32;serialized late, b5 .30,p5 .003,d5 .34; and no (sep-arate),b5 .24,p5 .02,d5 .27.Noother comparisonsbetweenconditionswere significant.As illustrated inFigure 4, late handoff finishedwell below the samplemean in creativity, whereas the other four conditionsfinishedslightlyabovethesamplemean.Theseresultsprovide additional support for the three componentsof Hypothesis 1.

Psychological ownership (Hypothesis 2) andproduct coherence (Hypothesis 3) as mediators.Models2to4inTable3testwhetherthenegativeeffectoflatehandoffsoncreativitywasmediatedbypsycho-logical ownership (Hypothesis 2) and product coher-ence (Hypothesis 3). (See Table 4 and Figure 4 formeans by condition.) Consistent with Hypothesis 2andHypothesis3,latehandoffwassignificantlylowerin psychological ownership and product coherencethaneachof theother conditions,withoneexception.Although late handoff was numerically lower in psy-chological ownership than serialized late handoff,thisdifferencewasnot significant (p5 .10).Asshownin Figure 4, late handoff was well below the samplemean in relation to both psychological ownership





andproduct coherence,whereas theother conditionswere slightly above the sample mean (except serial-ized late handoff, which was slightly below themean in psychological ownership).

To calculate 95%confidence intervals (CIs) testingthe indirect effects proposed in Hypotheses 2 and 3,we used the Monte Carlo method recommended formultilevel mediation (Bauer, Preacher, & Gil, 2006).Forpsychologicalownership(Hypothesis2), theindi-recteffectsweresignificantforthreeoffourconditions(vs. late): no (yoked), 95% CI [.04, .13]; early, 95% CI[.02, .10];andno(separate),95%CI[.03, .13].Theindi-rect effect was not significant for serialized late, 95%

CI [2.007, .08]. For product coherence (Hypothesis3), the indirect effectsweresignificant for all fourcon-ditions (vs. late): no (yoked), 95% CI [.17, .42]; early,95% CI [.14, .40]; serialized late, 95% CI [.16, .42];and no (separate), 95% CI [.12, .42]. Overall, theseresults support Hypothesis 2 and Hypothesis 3—exceptthattheresultsforHypothesis2werenotsignif-icant for serialized late versus late handoff.

Supplementary analyses on changes made dur-ing the implementation stage. Results from Study 2suggest that the late handoff condition fell behindthe other conditions in coherence and creativity dur-ing the implementation stage.5 To unpack how andwhy this occurred, we collected supplementary rat-ings of the changes that participantsmade during theimplementation stage (fromtheir outline to their finalad).SeeOnlineAppendixFforfurtherdetailsonthese

TABLE 3Study 2: Random Intercept Models Testing (Hypotheses 1–3)

Dependent Variable:

Model 1(H1)

Model 2(H2)

Model 3(H3)

Model 4(H2 and H3)

Creativity Psych. Own. Prod. Coh. Creativity

Condition (vs. Late handoff):No handoff .32�� .57�� .31�� 2.05

(.10) (.15) (.07) (.07)Early handoff .28�� .41�� .29�� 2.04

(.10) (.15) (.07) (.07)Serialized late handoff .30�� .24 .31�� 2.02

(.10) (.15) (.07) (.07)No handoff (separate) .24� .56�� .28�� 2.10

(.11) (.15) (.08) (.08)Psychological ownership .14��

(.02)Product coherence .94��

(.04)Fixed intercept 3.54�� 5.21�� 4.44�� 21.35��

(.08) (.10) (.06) (.20)Random intercept (yoked group) .04 .00 .06�� .04��

(.00) (.00) (.02) (.01)Log-likelihood 2736.82 2929.31 2539.96 2516.59

Note: All three continuous variables are in their original 7-point scale.� p , .05�� p , .01�� p , .001

TABLE 4Study 2: Means by Condition

Condition CreativityPsychologicalOwnership

ProductCoherence

Late 3.54 (.71) 5.21 (1.25) 4.44 (.76)No 3.86 (.87) 5.78 (1.05) 4.76 (.52)Early 3.82 (.81) 5.62 (1.16) 4.73 (.51)Serialized late 3.84 (.75) 5.45 (1.15) 4.75 (.59)No (separate) 3.78 (.98) 5.78 (1.10) 4.72 (.61)All conditions 3.77 (.84) 5.57 (1.16) 4.68 (.62)

Notes: Standard deviations are in parentheses. For eachcondition, n 5 120 participants (N 5 600 participants total).

5 The generation and elaboration stages also mattered, asthe creativity of the output during these first two stages sig-nificantly predicted final creativity (see correlations inTable F1 in Online Appendix F). This is consistent withpast studies that have tracked ideas over time, and haveshown that early quality significantly predicts final quality,but plenty of variance in final quality is also explained bylater stages (Berg, 2014; Kornish & Ulrich, 2014).




ratings and the analyses summarized here. Resultssuggest that, while a moderate degree of change wasoptimal, late handoff participants tended to changethe content of their emerging products too much. Inthe four conditions other than late handoff, partici-pants tended tomake refinements and enhancementsto thecoreconceptwithwhich theystarted the imple-mentation stage. In contrast, latehandoff participantstendedtomakemoredrasticchangesthatstrayedfromthe core concept they were handed—presumablybecausetheywerenotverycommittedtothiscorecon-cept. In turn, their relatively drastic changes reducedthe coherence and creativity of their final products.These results support the theorizing for Hypothesis3, and also fit past research showing that those higherinpsychological ownership tend tostaycommitted tothe core concept of their ideas, while those lower inpsychological ownership are less committed to doingso (Baer & Brown, 2012).

GENERAL DISCUSSION

Across two studies—an archival study of the U.S.film industry and a complementary experiment—wetested hypotheses on the impact of when and fromwhomideas arehandedoff (if at all) to the individualswhoultimately implement them. Consistentwith theproposed hypotheses, our results suggest that latehandoffs led implementers to build less creative finalproducts than the other three handoff types in ourframework. This means it was costly to creativitywhen individuals implemented mature ideas thatwere developed by someone else (late handoffs), asopposedtotakingchargeofthecreativeprocessbeforeideas were mature (no and early handoffs) or after anearlier handoff between twoother individuals (serial-ized latehandoffs).Mediation results suggest that latehandoffsreducedimplementers’creativitybyrestrict-ing their sense of psychological ownership and the

FIGURE 4Study 2: Means by Condition

0.5

0.4

0.3

0.2

Mea

n Z

-Sco

re

0.1

–0.1

–0.2

–0.3

–0.4

–0.5Late No Early Serialized Late No (Separate)

Creativity Psychological Ownership Product Coherence

0.0

Note: Error bars are61 SEM.


coherence of their final products, as compared to theother three handoff types.

Theoretical Implications

Idea implementation.Past researchhasuncoveredimportant knowledge on the social–political driversofideaimplementation,oftenfocusingonthenetworkstructures or team dynamics that help implementerswin social support for their creative ideas (e.g., Baer,2012; Fleming et al., 2007; Miron-Spektor et al.,2011; Perry-Smith & Mannucci, 2017; West, 2002).Ourresearchcomplements thispriorworkbyadvanc-ing theoryontheproductionsideof idea implementa-tion—building ideas into tangible final products—which has been largely overlooked. Because plansfor new ideas are rarely complete when they arehanded off, implementers have degrees of freedomas they build out their emerging products that maymake or break the creativity of their final products.In particular, our research suggests that idea imple-mentation may fail not just due to social resistance,butalsobecause implementersmightlackthepsycho-logicalownershipandcoherentvisionit takestobuildnew ideas into creative final products. In Study 2, allconditions entered the implementation stage withequivalent ideas, yet recipients of late handoffsmadelesscreativefinaladsthanrecipientsoftheotherhandofftypes.Inthisway,ourfocusontheproductionside of idea implementation provides a complemen-tary perspective to the social–political focus inprior work.

Inaddition, our researchsheds light onkeydynam-ics between idea implementation and the earlierstages of idea generation and elaboration. In studiesthat have looked at the whole creative process, fromidea generation through implementation, the samecreatorsdriveallthreestages(e.g.,Baer,2012;Flemingetal.,2007;Kornish&Ulrich,2014).Asaresult,littleisknown about the impact of handoffs between stages,despite the notion that such handoffs are common inorganizations. Our results suggest that driving ideageneration and elaboration (or just elaboration) posi-tively relates to successful idea implementation, asrecipientsofnoandearlyhandoffsbuiltmorecreativeproducts thanrecipientsof latehandoffswhoskippedover generation and elaboration. Furthermore, ourresearch helps explain the mechanisms driving thislink between idea generation/elaboration and imple-mentation: psychological ownership and productcoherence.Developing ideasbefore theybecomerela-tively detailed plans may give implementers anopportunity to form psychological ownership and a

unified vision for the emerging product, facilitatingthe coherence and thus creativity of their finalproducts.

Moreover, the results on serialized late handoffssuggest that, if implementers skip idea generationand elaboration, knowledge of an earlier handoffbetween these two stages may facilitate their creativ-ity. Implementers’ awareness of an earlier handoffwas the only difference between the serialized lateandlateconditionsinStudy2.Althoughtheserializedlate conditionhad a clear advantageover the late con-dition in final creativity, mediation results were notsignificant forpsychologicalownershipbut relativelystrong for product coherence. This hints that productcoherence may be a better explanation than psycho-logical ownership for the advantage of serialized lateover late handoffs, althoughmore research is neededto test this comparison more directly. Broadly, byadvancing theory on the relationship between ideageneration/elaboration and implementation, thiswork helps answer calls for a more complete under-standingofdynamicsbetweenkeypartsofthecreativeprocess (Amabile & Pratt, 2016).

Handoffs increativework.Handoffsbetweenindi-viduals have been extensively studied in themedicalcontext, focusing on issues that can arise when apatient’s care is handed fromone caregiver to another(LeBaron, Christianson, Garrett, & Ilan, 2016; Patter-son &Wears, 2010). Handoffs have been largely over-looked in organizational research broadly (Pentland,Recker, & Wyner, 2017), including in the context ofcreativework.AkeyexceptionisRouse’s(2013)afore-mentioned research on handoffs in creative work,focusingonhowrelationaldynamicsbetweenthegiv-ers and receivers of handoffsmay impact psychologi-cal ownership. We build on this work and advancetheory on handoffs in creative work by highlightingthe important role of idea maturity in shaping howrecipients respond to handoffs. By examining themediatingrolesofpsychologicalownershipandprod-uct coherence, our research also elucidates themech-anisms throughwhichhandoffs to implementersmayimpact the creativity of their final products.

Given evidence that anchoring effects and pathdependenciesearlyinthecreativeprocessplayanout-size role in shaping the creativity of the final product(Berg, 2014; Kornish & Ulrich, 2014; Ward et al.,1999), onemight assume that handoffs late in the cre-ative process would be relatively inconsequential.However, our research suggests that late handoffs toimplementers are more consequential for creativitythan earlier handoffs. Recipients of late handoffs arehanded relatively mature ideas, preventing them


fromdevelopingpsychological ownershipor a coher-ent vision for theproductprior to the implementationstage. In contrast, recipients of no and early handoffsscored equally high in creativity (Studies 1 and 2), aswell as psychological ownership and product coher-ence(Study2).Thissuggests thatdriving just theelab-oration stage influences implementers’ ultimatecreativity toapproximately the sameextentasdrivingbothgenerationandelaboration.Moreover,theresultson serialized latehandoffs suggest that an early hand-off between twoother individualsmayhelp avoid thenegative effects of handing off mature ideas to imple-menters, assuming implementers areawareof theear-lierhandoff between the two individuals. In thisway,handoffs to implementers may become riskier as theoriginal creator of the idea drives the creative processfor longer—once the idea becomes fairly mature, itmay be too late for the original creator to hand it offto the implementerwithout sacrificing creativity.

Psychological ownership and creativity. Ourresearch contributes to the small but growing litera-ture on psychological ownership in creative work.Past researchhashighlighted thatpsychologicalown-ership is important in the context of creativework, asmultiple individuals often have a claim to ownershipin creative projects (Gray et al., 2020; Rouse, 2013).However, little is known about the structural condi-tions that enable or constrain psychological owner-ship and how this relates to downstream creativity.Ourworksheds lightonhowthestructureofhandoffsto implementersshapes theircapacity todeveloppsy-chological ownership and ultimately the creativity oftheirfinalproducts.Inparticular,ourresearchempha-sizes the importanceof the stageatwhich implement-ers become the primary driver of the creativeprocess:whencreativeprojectsare structuredsuchthat imple-menters receive late handoffs, their psychologicalownership and downstream creativitymay suffer.

Furthermore, our work hints that psychologicalownershipmay be especially conducive to creativityduring the implementation stage of the creative pro-cess. Our findings suggest that it may be importantfor implementers to develop psychological owner-shipbeforeideasarematureandthustheimplementa-tionstagebegins,aspsychologicalownershipmaybeakey driver of creativity during implementation.Moreresearch is needed to fullyunderstandwhenandhowpsychological ownership is related to creativity ineach stage of the creative process. However, it seemsplausible that psychological ownership could be lessconducivetocreativityintheearlierstagesofideagen-erationandelaboration, asbeingprotectiveoverone’sideasmaypreventdivergenceandexploration(Baer&

Brown, 2012;Ward et al., 1999). In contrast, ourworksuggests that, during the implementation stage, psy-chologicalownershipmaygiveimplementersthecon-viction that is required to turnnew ideas into creativefinal products.

Limitations and Future Directions

Our studieshave important limitations thatmaybeaddressed in future research. First, the generalizabil-ity of our resultsmay be limited by the uniqueness ofthe film industry in Study 1 and the creative task inStudy 2. For instance, the final products in both stud-ies haveaesthetics as a key aspect,whichmaybe rele-vant to interpreting the results. Future research couldexploresimilardynamics inothercontexts to test gen-eralizability and identify boundary conditions. Sec-ond, we followed Perry-Smith and Mannucci’s(2017) simplifying assumption that one individualcan be identified as the principal driver/leader ofeach stage. Although this is often the case in practice(including for the films in Study 1),multiple individ-uals may co-lead stages or parts of stages. We see ourindividual-level focus as a startingpoint, andencour-age future researchon the effects of sharedownershipof stages. Given that collective psychological owner-ship grows from the same roots as individual psycho-logical ownership (Pierce & Jussila, 2010), we wouldexpect similar results if equally sized groups simplyreplaced individuals in our handoff framework. Thedynamics may become more complex (and poten-tially more interesting) when the groups vary in sizeand stability—and handoffs vary in frequency—such that individuals may enter, exit, and re-enterthe process in a less structured way than in ourframework.

Third, in Study 2, psychological ownership wasmeasuredafterparticipants finalized their ads,whichwas a conservative approach, given that recipients oflatehandoffsprobablydevelopedsomedegreeofpsy-chological ownership after working on their ads. Ifparticipants rated psychological ownership beforethe implementation stage, the late handoff conditionmay have scored even lower in psychological owner-ship (but this would likely introduce confounds,whichiswhywedidnotusethisapproach).Similarly,product coherence was based on participants’ finalproducts, and therefore Study 2 could not addressthe cognitive ormotivational processes that ledprod-uct coherence to be lower in the late handoff condi-tion. Future research could explore other ways ofmeasuring psychological ownership, product coher-ence, and related processes to complement these


limitations in our approach. Fourth, our theorizingfocused on late handoffs compared to the other threehandoff types, but the results hint that more researchmay be needed with regard to how the other threehandofftypescomparetooneanother.InStudy1,seri-alized late handoffs scored lower than no and earlyhandoffs, but, in Study 2, serialized late handoffs didnot differ from no and early handoffs. This is likelydue to idiosyncrasies in the two contexts (i.e., filmindustry vs. experiment task). Future research couldexplore more comparisons between handoff types,complementing our focus on late handoffs versus theother three.

Fifth,ourresearchfocusedonhandoffreceivers,butfuture research could address dynamics between giv-ers and receivers. For instance, in the film industry,directors know the writers’ identities and it varieswhether writers expect to hand their ideas off (butthis could not be measured). In Study 2, participantsdid not know the identities of the handoff givers orexpect to hand their ideas off. Future research couldexplore the role of prior relationships between giversand receivers and of givers anticipating handoffs.Lastly, in practice, a common rationale for handoffsis leveraging individuals’ unique skillsets. Per Study1, many film directors can and do write screenplays,and thus specialized skill presumably played amini-mal role in the results, but this could not be tested orruled out. In Study 2, random assignment held skillconstant across conditions, which is how skill isimplicitly treated in our theorizing. Future researchcould unpack the intersection of handoffs andskill—late handoffs may be beneficial when theyhelp leverage individuals’ unique skillsets.

Practical Implications and Conclusion

These studiesmayoffer useful insights for employ-ees inimplementerrolesandthosewhomanagethem.The main takeaway is simple: try to avoid late hand-offs. Our results highlight potential risks of specializ-ing in idea implementation. It may be vital foremployees tasked with implementing ideas to startdriving the process before ideas become too mature.Once the idea is developed into relatively detailedplans, it may be too late for implementers to developa sufficient sense of psychological ownership orcoherent vision, undermining the creativity of theirfinal products. As such, this research suggests thatorganizations ought to be careful about outsourcingidea implementation. The employee who developeda given idea may be best positioned to implementthe idea effectively, assuming the employee has the

requisite skillset. Despite the costliness of late hand-offs, in Study 1, they were used for 28.26% of films,whichwasmore frequent thanearly (8.51%)and seri-alized late (13.53%) handoffs combined (22.04%).This hints that many managers may not be aware ofthe costs associated with late handoffs. If handoffsneed to occur late in the creative process to leverageemployees’ specialized skillets, the results on serial-ized latehandoffs suggest that addinganearlierhand-off—and highlighting this to implementers—mayfacilitate their creativity. Overall, implementers maywant to err on the side of jumping in early than riskbeing too late.

REFERENCES

Aguinis, H., Villamor, I., & Ramani, R. S. 2021. MTurkresearch: Review and recommendations. Journal ofManagement, 47: 823–837.

Amabile,T.M.1988.Amodelofcreativityandinnovationinorganizations. In B. M. Staw & L. L. Cummings (Eds.),Research in organizational behavior, vol. 10: 123–167. Greenwich, CT: JAI Press.

Amabile, T. M. 1996. Creativity in context. Boulder, CO:Westview Press.

Amabile,T.M.,&Pratt,M.G.2016.Thedynamiccomponen-tialmodelofcreativityandinnovationinorganizations:Making progress, making meaning. Research in Orga-nizational Behavior, 36: 157–183.

Anderson,N., Poto�cnik, K., & Zhou, J. 2014. Innovation andcreativity in organizations: A state-of-the-sciencereview, prospective commentary, and guiding frame-work. Journal of Management, 40: 1297–1333.

Baer, M. 2012. Putting creativity to work: The implementa-tion of creative ideas in organizations. Academy ofManagement Journal, 55: 1102–1119.

Baer, M., & Brown, G. 2012. Blind in one eye: How psycho-logical ownership of ideas affects the types of sugges-tions people adopt. Organizational Behavior andHumanDecision Processes, 118: 60–71.

Baker, W. E., & Faulkner, R. R. 1991. Role as resource in theHollywood film industry.AmericanJournal ofSociol-ogy, 97: 279–309.

Barnes, C.M., Dang, C. T., Leavitt, K., Guarana, C. L., & Uhl-mann,E.L. 2018.Archivaldata inmicro-organizationalresearch: A toolkit formoving to a broader set of topics.Journal ofManagement, 44: 1453–1478.

Bauer, D. J., Preacher, K. J., & Gil, K.M. 2006. Conceptualiz-ing and testing random indirect effects and moderatedmediation in multilevel models: New proceduresand recommendations. Psychological Methods, 11:142–163.


Baumann, S. 2001. Intellectualization and art world devel-opment:FilmintheUnitedStates.AmericanSociolog-ical Review, 66: 404–426.

Bayus, B. L. 2013. Crowdsourcing new product ideas overtime: An analysis of the Dell IdeaStorm community.Management Science, 59: 226–244.

Berg, J.M.2014.Theprimalmark:Howthebeginningshapestheendinthedevelopmentofcreative ideas.Organiza-tional Behavior andHumanDecisionProcesses, 125:1–17.

Berkowitz, L., & Donnerstein, E. 1982. External validity ismore than skin deep: Some answers to criticisms oflaboratory experiments. American Psychologist, 37:245–257.

Besemer, S. P., & Treffinger, D. J. 1981. Analysis of creativeproducts: Review and synthesis. Journal of CreativeBehavior, 15: 158–178.

Brown,G., & Baer,M. 2015. Protecting the turf: The effect ofterritorial marking on others’ creativity. Journal ofApplied Psychology, 100: 1785–1797.

Buehler, R., Griffin, D., & Ross, M. 1994. Exploring the“planning fallacy”: Why people underestimate theirtask completion times. Journal of Personality andSocial Psychology, 67: 366–381.

Dane, E. 2010. Reconsidering the trade-off between exper-tise and flexibility: A cognitive entrenchment perspec-tive.Academy ofManagement Review, 35: 579–603.

Dawkins, S., Tian, A. W., Newman, A., & Martin, A. 2017.Psychological ownership: A review and researchagenda. Journal of Organizational Behavior, 38:163–183.

Elsbach, K.D., & Kramer, R.M. 2003.Assessing creativity inHollywoodpitchmeetings:Evidence for adual-processmodel of creativity judgments. Academy of Manage-ment Journal, 46: 283–301.

Finke, R. A. 1996. Imagery, creativity, and emergent struc-ture.Consciousness and Cognition, 5: 381–393.

Fleming, L., Mingo, S., & Chen, D. 2007. Collaborative bro-kerage, generative creativity, and creative success.Administrative Science Quarterly, 52: 443–475.

Grant, A.M. 2008. The significance of task significance: Jobperformance effects, relational mechanisms, andboundary conditions. Journal of Applied Psychology,93: 108–124.

Grant, A. M., & Berry, J. 2011. The necessity of others is themother of invention: Intrinsic and prosocial motiva-tions, perspective-taking, and creativity. Academy ofManagement Journal, 54: 73–96.

Gray,S.M.,Knight,A.P.,&Baer,M.2020.Ontheemergenceof collective psychological ownership in new creativeteams.Organization Science, 31: 141–164.

Hackman, J. R., & Oldham, G. R. 1976. Motivation throughthe design of work: Test of a theory. OrganizationalBehavior andHuman Performance, 16: 250–279.

Hargadon, A., & Sutton, R. I. 1997. Technology brokeringand innovation inaproductdevelopment firm.Admin-istrative Science Quarterly, 42: 716–749.

Harvey,S.2014.Creativesynthesis:Exploringtheprocessofextraordinary group creativity. Academy of Manage-ment Review, 39: 324–343.

Hirschman, E. C. 1989. Role-based models of advertisingcreation and production. Journal of Advertising, 18:42–53.

Hsu, G. 2006. Jacks of all trades andmasters of none: Audi-ences’ reactions to spanning genres in feature filmproduction. Administrative Science Quarterly, 51:420–450.

Jackson, P. W., & Messick, S. 1965. The person, the prod-uct, and the response: Conceptual problems in theassessment of creativity. Journal of Personality, 33:309–329.

King,N. 1992.Modelling the innovationprocess:Anempir-ical comparison of approaches. Journal of Occupa-tional and Organizational Psychology, 65: 89–100.

Kohn, N. W., & Smith, S. M. 2011. Collaborative fixation:Effects of others’ ideas onbrainstorming.AppliedCog-nitive Psychology, 25: 359–371.

Kornish,L. J.,&Ulrich,K.T.2014.Theimportanceof therawidea in innovation: Testing the sow’s ear hypothesis.JMR, Journal ofMarketing Research, 51: 14–26.

Kozbelt, A. 2008. Longitudinal hit ratios of classicalcomposers: Reconciling “Darwinian” and expertiseacquisition perspectives on lifespan creativity. Psy-chology of Aesthetics, Creativity, and the Arts, 2:221–235.

Larey, T. S.,& Paulus, P. B. 1999.Grouppreference andcon-vergent tendencies in small groups: A content analysisof group brainstorming performance. CreativityResearch Journal, 12: 175–184.

LeBaron, C., Christianson,M. K., Garrett, L., & Ilan, R. 2016.Coordinating flexible performance during everydaywork: An ethnomethodological study of handoff rou-tines.Organization Science, 27: 514–534.

LeBreton, J.M., & Senter, J. L. 2008.Answers to 20questionsabout interrater reliability and interrater agreement.Organizational ResearchMethods, 11: 815–852.

Lucas, B. J., & Nordgren, L. F. 2015. People underestimatethe value of persistence for creative performance.Journal of Personality and Social Psychology, 109:232–243.

Merton, R. K. 1968. TheMattheweffect in science.Science,159: 56–63.


Miron-Spektor, E., Erez, M., & Naveh, E. 2011. The effect ofconformist and attentive-to-detail members on teaminnovation:Reconcilingthe innovationparadox.Acad-emy ofManagement Journal, 54: 740–760.

O’Driscoll, M. P., Pierce, J. L., & Coghlan, A.-M. 2006. Thepsychology of ownership: Work environment struc-ture, organizational commitment, and citizenshipbehaviors. Group & Organization Management, 31:388–416.

Patterson, E. S., &Wears, R. L. 2010. Patient handoffs: Stan-dardized and reliable measurement tools remain elu-sive. Joint Commission Journal on Quality andPatient Safety, 36: 52–61.

Paulus, P. B. 2002. Different ponds for different fish: A con-trasting perspective on team innovation.Applied Psy-chology, 51: 394–399.

Paulus, P. B., & Yang, H.-C. 2000. Idea generation in groups:A basis for creativity in organizations.OrganizationalBehavior and Human Decision Processes, 82: 76–87.

Pentland, B. T., Recker, J., &Wyner, G. 2017. Rediscoveringhandoffs. Academy of Management Discoveries, 3:284–301.

Perry-Smith, J.E.,&Mannucci,P.V.2017.Fromcreativity toinnovation: The social network drivers of the fourphases of the idea journey. Academy of ManagementReview, 42: 53–79.

Pierce, J. L., & Jussila, I. 2010.Collectivepsychological own-ership within the work and organizational context:Construct introduction and elaboration. Journal ofOrganizational Behavior, 31: 810–834.

Pierce, J. L.,Kostova,T.,&Dirks,K.T.2001.Towarda theoryofpsychologicalownershipinorganizations.AcademyofManagement Review, 26: 298–310.

Pierce, J. L., Kostova, T., & Dirks, K. T. 2003. The state ofpsychological ownership: Integrating and extending acentury of research. Review of General Psychology,7: 84–107.

Pierce, J. L., O’Driscoll, M. P., & Coghlan, A.-M. 2004.Workenvironment structure and psychological ownership:Themediating effects of control. Journal ofSocial Psy-chology, 144: 507–534.

Plucker, J.A.,Holden, J.,&Neustadter,D.2008.Thecriterionproblemandcreativity infilm:Psychometriccharacter-istics of various measures. Psychology of Aesthetics,Creativity, and the Arts, 2: 190–196.

Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff,N. P. 2003. Common method biases in behavioralresearch: A critical review of the literature and recom-mended remedies. Journal of Applied Psychology,88: 879–903.

Porter,C.O.,Outlaw,R.,Gale, J.P.,&Cho,T.S.2019.Theuseof onlinepanel data inmanagement research:A review

and recommendations. Journal of Management, 45:319–344.

Raudenbush, S.W., & Bryk,A. S. 2002.Hierarchical linearmodels: Applications and data analysis methods.Thousand Oaks, CA: SAGE.

Rietzschel, E. F., Nijstad, B. A., & Stroebe, W. 2010. Theselection of creative ideas after individual idea genera-tion: Choosing between creativity and impact. BritishJournal of Psychology, 101: 47–68.

Rouse, E. 2013. Kill your darlings? Experiencing, maintain-ing, and changing psychological ownership in creativework [Doctoral dissertation, BostonCollege]. Retrievedfrom http://hdl.handle.net/2345/3239

Seidel, V. P., & O’Mahony, S. 2014. Managing the reper-toire: Stories, metaphors, prototypes, and conceptcoherence in product innovation. Organization Sci-ence, 25: 691–712.

Shalley, C. E. 1991. Effects of productivity goals, creativitygoals, and personal discretion on individual creativity.Journal of Applied Psychology, 76: 179–185.

Shrout,P.E.,&Fleiss, J.L.1979. Intraclasscorrelations:Usesin assessing rater reliability. Psychological Bulletin,86: 420–428.

Simonton, D. K. 1997. Creative productivity: A predictiveand explanatory model of career trajectories and land-marks. Psychological Review, 104: 66–89.

Simonton,D.K. 2002.Collaborative aesthetics in the featurefilm: Cinematic components predicting the differentialimpact of 2,323 Oscar-nominated movies. EmpiricalStudies of the Arts, 20: 115–125.

Simonton, D. K. 2004. Group artistic creativity: Creativeclusters andcinematic success in feature films. Journalof Applied Social Psychology, 34: 1494–1520.

Simonton,D.K.2005.Filmasartversusfilmasbusiness:Dif-ferential correlates of screenplay characteristics.Empirical Studies of the Arts, 23: 93–117.

Simonton,D.K.2011.Greatflicks:Scientificstudiesofcin-ematic creativity and aesthetics. New York, NY:Oxford University Press.

Smith, S. M., Ward, T. B., & Schumacher, J. S. 1993. Con-straining effects of examples in a creative generationtask.Memory & Cognition, 21: 837–845.

Sternberg, R. J., & Lubart, T. I. 1991.An investment theoryofcreativity and its development.HumanDevelopment,34: 1–31.

Terwiesch, C., & Ulrich, K. T. 2009. Innovation tourna-ments:Creatingandselectingexceptionalopportuni-ties. Boston, MA: Harvard Business School Press.

Thomson,D.2006.Thewholeequation:AhistoryofHolly-wood. New York, NY: Vintage.

Van Dyne, L., & Pierce, J. L. 2004. Psychological ownershipand feelings of possession: Three field studies


http://hdl.handle.net/2345/3239

predicting employee attitudes and organizational citi-zenship behavior. Journal of Organizational Behav-ior, 25: 439–459.

Walter,S.L.,Seibert,S.E.,Goering,D.,&O’Boyle,E.H.2019.A tale of two sample sources: Do results from onlinepanel data and conventional data converge? Journalof Business and Psychology, 34: 425–452.

Ward, T. B. 1994. Structured imagination: The role of cate-gory structure in exemplar generation. Cognitive Psy-chology, 27: 1–40.

Ward, T. B., Smith, S. M., & Finke, R. A. 1999. Creative cog-nition. In R. J. Sternberg (Ed.),Handbook of creativity:189–212. New York, NY: Cambridge University Press.

Weick, K. E. 1979. The social psychology of organizing(2nd ed.). Reading, MA: Addison-Wesley.

Welch,L.1946.Recombinationof ideas increativethinking.Journal of Applied Psychology, 30: 638–643.

West, M. A. 1990. The social psychology of innovation ingroups. In M. A. West & J. L. Farr (Eds.), Innovationand creativity at work: 309–333. Chichester, U.K.:Wiley.

West, M. A. 2002. Sparkling fountains or stagnant ponds: Anintegrativemodelofcreativityandinnovationimplemen-tation inworkgroups.AppliedPsychology, 51:355–387.

WritersGuildofAmericaWest.2018.Screencreditsmanual.Retrieved from https://www.wga.org/uploadedfiles/credits/manuals/screenscredits_manual18.pdf

Justin M. Berg ([email protected]) is an assistantprofessor of organizational behavior at the StanfordGraduate School of Business. His research focuses oncreativity and innovation. He studies how to successfullydevelop, evaluate, and implement creative ideas in andoutside organizations. He received his PhD in managementfrom theWharton School of theUniversity of Pennsylvania.

Alisa Yu ([email protected]) received her PhD inorganizational behavior at the Stanford Graduate Schoolof Business. Her research focuses on emotions andinterpersonal dynamics in organizations.


https://www.wga.org/uploadedfiles/credits/manuals/screenscredits_manual18.pdf

https://www.wga.org/uploadedfiles/credits/manuals/screenscredits_manual18.pdf

mailto:[email protected]

mailto:[email protected]

getting the picture too late: handoffs and the

Documents