changes in the work environment for creativity during...

* Academy of Management Joumal1999, Vol. 42, No. 6, 630-640.

CHANGES IN THE WORK ENVIRONMENT FOR CREATIVITYDURING DOWNSIZING

TERESA M. AMABILEHarvard University

REGINA CONTIColgate University

This study examined the work environment for creativity at a large high-technologyfirm hefore, during, and after a major downsizing. Greativity and most creativity-supporting aspects of the perceived work environment declined significantly duringthe downsizing hut increased modestly later; the opposite pattern was ohserved forcreativity-undermining aspects. Stimulants and ohstacles to creativity in the workenvironment mediated the effects of downsizing. These results suggest ways in whichtheories of organizational creativity can he expanded and ways in which the negativeeffects of downsizing might he avoided or alleviated.

A downsizing such as this one is always difficult foremployees. But out of tough times can comestrength, creativity and teamwork. [We are] fortu-nate to retain committed, hard-working and creativeemployees, who will find substantial motivationand fulfillment as we all work together in 1995 toachieve [our] turnaround. (From the letter to share-holders in the 1994 annual report of a United Statessoftware company)

Our strategy is to reduce operating costs to a levelcommensurate with projected sales levels. To ac-complish this goal, we have implemented a numberof tough cost-cutting measures, which includedownsizing of the workforce, facilities and otheroperating costs. In addition to cutting costs, we arefocusing on product innovation through our highlyeffective Research and Development team. (Fromthe letter to shareholders in the 1996 annual reportof an international electronic games company)

American companies continued to downsizetheir workforces through most of the 1990s, but

The data for this study were collected when both au-thors were in the Department of Psychology at BrandeisUniversity. The data collection and analysis were sup-ported by a grant from the Genter for Innovation Manage-ment Studies at Lehigh University. We are most gratefulto the Genter and its corporate sponsors for their support.Preparation of this article was supported by the HarvardBusiness School Division of Research. We wish to ac-knowledge Jill Nemiro and Thomas Leahy for theirhighly competent work as interviewers. We also wish toexpress our appreciation to the managers and employeesof the company studied, especially certain individuals inthe organizational research department, for the crucialroles they played in the collection of the data reportedhere.

they also continued to emphasize the importance ofinnovation to long-term success. Do the two goalsgo hand-in-hand, as these letters to shareholderssuggest, or are they incompatible?

As noted by McKinley and his colleagues (Mc-Kinley, 1993; Mone, McKinley, & Barker, 1998],there is considerable controversy among theoreti-cians about the effects of downsizing on a variety oforganizational outcomes, including innovation.Most scholars have adopted a definition of down-sizing similar to that of Freeman and Cameron(1993), describing it as an intentional managementaction involving a reduction in force and designedto improve a company's competitive position.Some theorists have taken the viewpoint expressedin the first letter to shareholders quoted earlier,arguing that a reduction in force will have generallypositive effects on a company's efficiency, reducingwaste and leading to a more productive allocationof resources (e.g., Jensen, 1986). However, others—most notably Cascio (1993)—have suggested that in-novation will be likely to suffer during downsizing.

Innovation, generally defined as the implemen-tation or adoption of new, useful ideas by people inorganizations (Amabile, 1988, 1996; Van de Ven,1986), depends on creativity—the generation ofthose new and useful ideas (Amabile, 1983a, Am-abile, 1983b). This article focuses on creativity and,in particular, changes in the work environment forcreativity during downsizing.

If indeed creativity is dampened during down-sizing, the work environment may play a centralrole in this decline. In a theoretical model of theeffects of layoffs on survivors, Brockner (1988)highlighted the importance of the work environ-

630

1999 Amabile and Conti 631

ment, which he referred to as context. Context en-compasses all elements of the psychological cli-mate of both the formal organization (policies andprocedures) and the informal organization (values,norms, and interpersonal relationships). Brockner'sresearch has shown that context can be importantnot only in affecting survivors' reactions to layoffs,but also in determining the impact of those reac-tions on job performance and other outcomes (e.g.,Brockner, Davy, & Carter, 1985; Brockner, Green-berg, Brockner, Bortz, Davy, & Carter, 1986; Brock-ner, Grover, Reed, DeWitt, & O'Malley, 1987;Brockner, Wiesenfeld, Reed, Grover, & Martin,1993).

Previous research has uncovered some commonpatterns of change in organizational w ork environ-ments during downsizing. Organizations appear toundergo a deterioration of communication at manylevels (Cascio, 1993; Dougherty & Bowman, 1995;Noer, 1993) during such periods, even though com-munication appears to be particularly important atthese times (Rosenblatt, Rogers, & Nord, 1993).Similarly, downsizing organizations appear to suf-.fer a deterioration of trust (Buch & Aldridge, 1991;Cascio, 1993) and an increase in fear (Buch et al.,1991). Organizations in decline (which includemost organizations undergoing downsizing) aremarked by a resistance to change and a tendencytoward rigid behavior patterns (Cameron, Sutton, &Whetton, 1988). Finally, the work environments ofdownsizing organizations appear to be marked byhigh levels of uncertainty and chaos (Tombaugh &White, 1990).

How might such changes in the work environ-ment come about? Threat-rigidity theory (Staw,Sandelands, & Dutton, 1981) directly addressespossible changes in organizational environmentsunder negative circumstances such as those accom-panying downsizing. The theory is concerned withsituations of threat, defined as an external event orcircumstance in which individuals, groups, or or-ganizations perceive impending negative or harm-ful consequences for their vital interests. Clearly,this definition applies to most organizations thatundertake downsizing; it certainly applies to theorganization examined in the present study. Al-though threat is proposed to have conceptuallysimilar effects on individuals, groups, and organi-zations, it is threat-rigidity theory's organizationallevel of analysis that is most relevant to a consid-eration of changes in work environments duringdownsizing.

According to this theory, under threatening con-ditions, organizations undergo a "mechanisticshift" (Staw et al., 1981: 516). They centralize con-trol, conserve resources, restrict information flow.

and rely on familiar, well-practiced routines. Un-der some conditions, when the organizationalthreat arises from relatively small, incrementalchanges, such effects can be useful. However,threat-rigidity theory implies that, under the typesof radical changes that are likely to precede oraccompany downsizing, the resulting effects willbe dysfunctional. This position has received con-siderable empirical support, primarily from studiesof organizational crisis (e.g.. Billings, Milburn, &Shaalman, 1980; D'Aveni & MacMillan, 1990;Smart & Vertinsky, 1984; Tjosvold, 1984) and stud-ies of centralization, efficiency, and resistance tochange (e.g., Cameron, Whetten, & Kim, 1987;D'Aveni, 1989; Whetten, 1981).

Recent theories of organizational creativity stressthe role of an organization's work environment inaffecting creative behaviors of individuals andteams (Amabile, 1988; Woodman, Sawyer, & Grif-fin, 1993). According to the componential model ofcreativity and innovation in organizations (Am-abile, 1988, 1997; Amabile, Conti, Coon, Lazenby,& Herron, 1996), five environmental componentsaffect creativity: encouragement of creativity(which encompasses open information flow andsupport for new ideas at all levels of the organiza-tion, from top management, through immediate su-pervisors, to work groups); autonomy or freedom(autonomy in the day-to-day conduct of work; asense of individual ownership of and control overwork); resources (the materials, information, andgeneral resources available for work); pressures (in-cluding both positive challenge and negative work-load pressure); and organizational impediments tocreativity (including conservatism and internalstrife). As measured by the survey instrumentKEYS: Assessing the Climate for Creativity (Am-abile, 1996), the components fall into two generalcategories: stimulants to creativity (tapped byscales assessing organizational and supervisory en-couragement, work group support, sufficient re-sources, and challenging work), and obstacles tocreativity (tapped by scales assessing organiza-tional impediments and workload pressure). In astudy in a high-tech company (Amabile et al.,1996), R&D projects rated high in creativity hadsignificantly different work environments (as mea-sured by KEYS) from those projects rated low increativity. The high-creativity projects were gener-ally higher on work environment stimulants to cre-ativity and lower on work environment obstacles tocreativity. Thus, it appears that there is indeed arelationship between the work environment andthe level of creativity produced by individuals inteams.

The mechanistic shift described in threat-rigidity

632 Academy of Management Journal December

theory suggests consequential changes during down-sizing in the aspects of the work envirormient thathave been identified as important for creativity. Pre-vious research on change in organizations under-going downsizing, described above, also points tochanges in creativity-relevant aspects of the workenvironment. Moreover, there is some evidence thatcreativity itself may decline during downsizing.Cascio (1993) concluded that downsizing survivorsbecome narrow-minded and risk-averse. Other re-searchers have found evidence that members of orga-nizations in decline exhibit a tendency toward rigidbehavior patterns (Cameron et al., 1988). Given thatrisk-taking and flexibility in behavior are central tothe generation of creative ideas, this evidence sug-gests that downsizing survivors will indeed be lesscreative in their work. The present study examinedthree hypotheses relevant to these proposed changesin work environments, and in creativity, as a functionof downsizing.

Hypothesis 1. Creativity will decrease as afunction of downsizing.

Hypothesis 2. The relationship between cre-ativity and downsizing will be accounted for bydecreases in work environment stimulants tocreativity and increases in work environmentobstacles to creativity.

METHODS

Design Overview

This study examined the work environment forcreativity in a large high-technology firm before, dur-ing, and after a major downsizing. We assessed workenvironment and creativity by sampling independentgroups of svirviving employees at four points in time:baseline (predownsizing), wave 1 (middownsizing),wave 2 (official end of downsizing), and wave 3 (post-downsizing). Measures of downsizing were taken atthree points in time: wave 1, wave 2, and wave 3. Forexploratory purposes, we included a number of othermeasures and conducted semistructured interviewswith survivors.

Some individuals in this company experiencedconsiderable downsizing in their departments,some were spared, and a few actually experiencedgrowth in their departments during the periodstudied. Therefore, the design allowed for exami-nation of perceived work environment and variousother outcomes as a function of the degree and typeof downsizing change experienced. Downsizingwas operationally defined through three measuresof changes experienced by individuals in the orga-nization; each of these changes is a common ac-

companiment of a downsizing of one's departmentor work group: recently experienced downsizing inone's own department (downsizing experienced),anticipation of future downsizing in one's own de-partment (anticipated downsizing), and recent dis-ruption of the stability of one's own work group(work group stability).

Research Setting

This study was conducted in a Fortune 500 high-technology company with over 30,000 employeesthat provided diversified electronics products tointernational markets. In the course of conductingan earlier study, we had obtained work environ-ment assessments from a broad sample of the com-pany's employees in January 1993. Approximatelyfour months later, the chairman of the companyannounced that a major restructuring of the com-pany was necessitated by disappointing corporateprofitability and by a radical shifr in the company'smajor market. Employees were told that the restruc-turing would be accompanied by downsizing theworkforce by 12-15 percent. Additionally, theywere told that this reduction in force was to beaccomplished by successive waves of early retire-ment incentives, voluntary separation incentives,and involuntary layoffs. There was no announcedstandard procedure for determining layoffs.

In addition to the predownsizing baseline datacollected in January 1993, we collected data in July1994 (wave 1), when the downsizing was at about60 percent of its stated target; in December 1994(wave 2), just as the downsizing was reaching itspreviously announced completion; and in May1995 (wave 3).

Each wave of data collection targeted a differentrandom sample of employees. This between-subjects design was necessary because the com-pany required that all questionnaires be anony-mous, and because sample attrition due toemployees leaving the company (either voluntarilyor involuntarily) during the downsizing wouldhave made it impossible for us to follow one groupof people through all waves.

Respondents

A total of 754 employees participated in thestudy. There were 455 respondents to the baselinedata collection (a 45.5 percent response rate); 102 atwave 1 (34%); 87 at wave 2 (29%); and 110 at wave3 (37%). At waves 1, 2, and 3, respondents wereasked to volunteer for interviews, and 83 were sub-sequently interviewed: 34 at wave 1, 23 at wave 2,and 26 at wave 3.


These volunteers were interviewed in 20-minutesessions either in person or via telephone, by one offour interviewers. To avoid biasing responses, theinterviewers used the more neutral term "restruc-turing" rather than "downsizing." The questionsfocused on changes experienced during the restruc-turing, perceptions of various aspects of the workenvironment, general motivational and socialchanges experienced or observed, and suggestionsfor management about the restructuring.

Instruments

Respondents in all phases of data collectioncompleted the work environment survey KEYS:Assessing the Climate for Creativity (Amabile,1995; Amabile, Burnside, & Gryskiewicz, 1995;Amabile & Gryskiewicz, 1989). In addition, re-spondents for waves 1, 2, and 3 completed abackground questionnaire.

KEYS: Assessing the Climate for Creativity.The KEYS instrument is designed to assess factorsin the work environment that influence creativityand innovation, according to previous theory andresearch (see Amabile, 1988, 1996). KEYS includeseight work environment scales (six stimulants tocreativity and two obstacles to creativity) and twocriterion scales (creativity and productivity). The78-item inventory asks respondents to indicate theextent to which each statement describes theirwork environment. Thus, the instrument assessesrespondents' perceptions of specific stimulants andobstacles to creativity, as well as their perceptionsof the actual creativity and productivity of the workbeing done (the criterion scales). KEYS has beenshown to have satisfactory psychometric proper-ties, in terms of both reliability and validity(Amabile et al., 1996).

The six stimulants to creativity scales, listed herewith the major concepts they cover summarized inparentheses, are (1) organizational encouragement(15 items, Cronbach's alpha [a] = .92; an organiza-tional culture that encourages creativity through anactive flow of ideas, the fair and constructive judg-ment of ideas, reward and recognition for creativework, mechanisms for developing new ideas, and ashared vision of what the organization is trying todo); (2) sufficient resources (6 items, a = .82; accessto appropriate resources, including funds, materi-als, facilities, and information); (3) freedom (4items, a = .69; freedom in deciding what work todo or how to do it; a sense of control over one'swork); (4) challenging work (5 items, a = .80; asense of having to work hard on challenging tasksand important projects); (5) supervisory encourage-ment (11 items, a = .92; a supervisor who commu-

nicates effectively with the work group, serves as agood work model, sets goals appropriately, sup-ports the work group, values individual contribu-tions, and shows confidence in the work group);and (6) work group supports (8 items, a = .86; adiversely skilled work group in which people com-municate well, are open to new ideas, construc-tively challenge each other's ideas, trust and helpeach other, and feel committed to the work they aredoing).

The two KEYS obstacles to creativity scales(again, listed with major concepts) are: (1) organi-zational impediments (12 items, a = .84; an orga-nizational culture that impedes creativity throughinternal political problems, harsh criticism of newideas, destructive internal competition, an avoid-ance of risk, and an overemphasis on the statusquo) and (2) workload pressure (5 items, a = .77;extreme time pressures, unrealistic expectations forproductivity, and distractions from creative work).

The two KEYS criterion scales (listed with majorconcepts) are: (1) creativity (6 items, a = .84; acreative and innovative organization or unit, wherea great deal of creativity is called for and wherepeople believe they actually produce creativework) and (2) productivity (6 items, a = .88; anefficient, effective, and productive organization orunit).

Background questionnaire. This questionnaire,created for the present study, obtained measures ofchanges experienced and anticipated during thedownsizing, including (1) respondents' numericalestimates of the percent changes in the sizes of theirdepartments since the last data collection (down-sizing experienced), (2) the anticipated percentchanges in the sizes of their departments over thenext year (anticipated downsizing), and (3) changesin the membership of their w ork groups since thelast data collection (work group stability/disrup-tion). The questionnaire also included a number ofother exploratory measures, such as self-reportedrisk-taking and entrepreneurialism, perceived un-certainty and chaos, job satisfaction, morale, feel-ings, and likelihood of voluntary exit. The ques-tions on the background questionnaire werederived in part from the existing literature on theeffects of downsizing on employees (see Cascio,1993).

RESULTS

Downsizing

Table 1 presents data on the reported downsizingchanges. (Recall that comparable data were not col-lected at the baseline point, although the wave 1


TABLE 1Downsizing Changes"

Variable Wave 1 Wave 2 Wave 3

Percent change in size of department since previous data collection''Anticipated percent change in department size in nextStability of work group''

Multivariate F,o. 520

-13.663(32.70)-17.03j30.45)

66.92^ (25.08)

-3.68b (35.25)-0.06b (39.27)76.57b (23.81)

4.40*

-6.11b (29.86)-2.90b (36.36)82.a8b (20.71)

° Values shovirn are means and standard deviations (in parentheses). Means vi ith different subscripts ("a" or "b") are significantlydifferent (p < .05) as revealed by contrasts.

^ The overall effect was significant at p < .05.• The overall effect was significant at p < .01.•* The overall effect was significant at p < .001.*** p < .001

questionnaire did ask for reports of changes expe-rienced since the baseline data collection.) Thesemeasures include the reported change in the size ofa respondent's department occurring since the pre-vious data collection (as the percent smaller orlarger), the reported anticipated change in depart-ment size in the upcoming year (as the percentsmaller or larger), and the reported stability in themembership of the respondent's work group (anaverage of three items, a = .74, asking about thepercentage of previous coworkers whom the re-spondent was still working with; a higher numbersignifies greater stability). These results confirmthat a large dow^nsizing was experienced by therespondents between the baseline and wave 1 datacollections and that stability increased thereafter.

Changes in Creativity

A multivariate analysis of variance (MANOVA)comparing the criterion scales from the baseline,wave 1, wave 2, and wave 3 showed a significantmultivariate effect and significant univariate ef-

fects for both the creativity and productivityscales (see Table 2). Univariate contrasts for cre-ativity showed that it was significantly belowbaseline levels at all three later time points! Pro-ductivity was significantly below baseline atwaves 1 and 2, but not at wave 3. Thus, Hypoth-esis 1 is supported: it appears that creativity diddecline during the downsizing. Notably, changesin employees' assessments of creativity weremuch greater and more enduring than werechanges in their assessments of productivity.

To obtain an additional and more objectivemeasure of creativity and innovation, we at-tempted to secure month-by-month informationon the number of invention disclosures andpatent applications at the company for the yearsbefore, during, and after the downsizing an-nouncement. Unfortunately, we were able to ob-tain only data aggregated by year for two of thoseyears (the year during which the downsizingstarted in August and the following year, whenthe downsizing ended in December); thus, thisinformation is only suggestive. The number of

TABLE 2Creativity and Productivity"

KEYS Criterion Scale*"

Creativity'Productivity''

Multivariate Fg jjjo

Baseline,January 1993

2.73 (0.65)2.91 (0.59)

6.62***

Wave 1,July 1994

2.35J0.63)2.68^ (0.63)

Wave 2,December 1994

2.49o (0.56)2.75e (0.54)

Wave 3,May 1995

2.54o (0.65)2.87 (0.60)

" Values shown are means and standard deviations (in parentheses). Means with the subscript "a" were significantly different from thebaseline measure (p < .05) in a contrast test. For the four data collections, n = 455, 102, 87, and 110, respectively.

'' The KEYS response scale ranges from 1, "never or almost never true of your current work environment," to 4, "always or almost alwaystrue of your current work environment."

" The overall effect was significant at p < .001.'' The overall effect was significant at p < .01.***p < .001


employees decreased by 15 percent from the yearduring which the downsizing started through theyear during which it ended, and the number ofpatent applications declined by about the sameamount (12%). These percentages indicate thatthe number of patent applications per employeeremained approximately constant during thedownsizing. However, because of the long pro-cess required in submitting patent applications,this indicator of creative activity would be ex-pected to show considerably lagged effects. Bycontrast, the number of invention disclosuressubmitted during a given period of time is a moreimmediate indicator of current technological in-vention. Importantly, invention disclosure sub-missions declined by a disproportionate 24 per-cent during the same period. In other words, percapita, people were producing fewer new ideas.These data provide some additional support forthe conclusion that, indeed, declines in creativityaccompanied the downsizing.

Although the MANOVA results demonstratedthat declines in creativity accompanied downsiz-ing, we also examined Hypothesis 1 more directlyby determining the extent to which creativity de-clined as a function of downsizing. To this end,using data across all postbaseline waves, we re-

gressed creativity on each of the three downsizingmeasures (downsizing experienced, anticipateddownsizing, and work group stability). Two of thethree measures (anticipated downsizing and workgroup stability) were significant predictors of cre-ativity, and the overall regression equation was sig-nificant 288 = 4.04, p < .01, R^ = .04).

Many of the interviewees' comments focused onhow the downsizing limited creativity; for exam-ple, one employee said the following: "Before, cre-ativity wasn't promoted; now it's promoted evenless. People feel stifled. They don't take risks be-cause of the ramifications. They're worried aboutgetting laid off." Another interviewee said this:"The quality of our work dropped, because thesense of pride declined. We were no longer really ateam. Instead, everyone was trying to protect theirjob—to justify themselves by looking good. They'renot focused on the work itself."

Changes in the Work Environment

Table 3 presents work environment means oneach of the KEYS work environment scales for thebaseline and for waves 1, 2, and 3. Overall, thepattern is quite clear and consistent. All six KEYSenvironmental stimulants to creativity declined no-

TABLE 3Work Environment"

KEYS Scale"*

Stimulants to creativityFreedom'Challenge''Sufficient resources"Supervisory encouragementWork group supports'*Organizational encouragement''

Multivariate F^g 2.11a

Obstacles to creativityOrganizational impediments"Workload pressures"

Multivariate F^ ,5,0

Baseline,January 1993

2.93 (0.62)2.95 (0.65)2.78 (0.58)2.83 (0.73)3.07 (0.54)2.58 (0.60)

4.87***

2.25 (0.59)2.59 (0.61)

3.73**

Wave 1,July 1994

2.76^ (0.58)2.64^ (0.61)2.58^ (0.57)2.70b (0-60)2.71, (0.59)2.24^ (0.54)

2.44e (0.56)2.61 (0.63)

Wave 2,December 1994

2.80b (0-63)2.72, (0.64)2.68 (0.61)2.74 (0.68)2.95b (0-55)2.35^ (0.52)

2.27 (0.54)2.55 (0.56)

Wave 3,May 1995

2.97 (0.58)2.74, (0.63)2.89b (0.53)2.85 (0.71)2.92^ (0.58)2.47b (0.62)

2.18 (0.51)2.37,, (0.57)

" Values shown are means and standard deviations (in parentheses). Means with the subscript "a" were significantly different (p < .05)from the baseline measure in a contrast test. The subscript "b" indicates marginal difference [p < .10) from the baseline. For tbe four datacollections, Ji = 455, 102, 87, and 110, respectively.

'' Tbe KEYS response scale ranges from 1, "never or almost never true of your current work environment," to 4, "always or almost alwaystrue of your current work environment." The reliabilities of the BCEYS scales in this study were comparable to tbose published in theliterature (Amabile et al., 1996). The median alphas were .88 for tbe baseline, .86 for wave 1, .85 for wave 2, and .85 for wave 3.

" The overall effect was significant at p < .05.'' The overall effect was significant at p < .001." Tbe overall effect was significant at p < .01.

**p < .01***p < .001


tably from the baseline to wave 1. In contrast, theKEYS obstacle measure, organizational impedi-ments, increased. These results support the Hy-pothesis 2 prediction that, during downsizing,work environment stimulants to creativity will de-crease and work environment obstacles willincrease. The trend seems to be slowly revers-ing, however, as one moves from wave 1 to waves2 and 3; the creativity stimulants are increasing,and organizational impediments are declining.MANOVAs followed by univariate analyses of vari-ance and simple contrasts confirmed these patterns(see Table 3). Interestingly, although two aspects ofthe work environment showed a significant im-provement at wave 3 (there was a higher level ofsufficient resources and a lower level of workloadpressure), two other aspects (challenge and workgroup supports) showed a continuing significantdepression of the work environment for creativityeven five months after the downsizing ended.

The interviewees' comments largely reinforcedthe view that the work environment deterioratedduring the downsizing. For example, several com-mented on supervisor support; one intervieweestated this: "Supervisory support? None, zero—hasgone from bad to worse. . . . They're in limbo too."Many described structural changes relating to un-certainty and chaos: "During and after the restruc-turing, there was frustration at the chaos of whomoved into where, and what the new structurelooked like. I no longer knew where to go to dowhat I needed to do. Many people had no ideawhere they themselves were in the new organiza-tion." About an equal number of respondents men-tioned social changes; often, these changes inter-fered with unit functioning: "The impact has beennegative, in all respects. Communication is not asopen and honest. Cooperation is politically based,not oriented toward the customer's benefit. Trustwithin the group has deteriorated, because peopleare worried about their jobs." Effects on personalfunctioning were rarely reported. But when theydid occur, the circumstances were quite severe:"I'm very stressed and angry. My ulcer started act-ing up again recently, after 13 years. I thrive ongood stress, but it's all bad now" and "Our place isbeing shut down . . . we've had two suicides here."

Hypothesis 2 suggests that the relationship be-tween downsizing and creativity is mediated (seeJudd & Kenny, 1981) by such changes in the workenvironment. As noted earlier, two of the three down-sizing measures (anticipated downsizing and workgroup stability) significantly predicted creativity. Ad-ditional regression analyses demonstrated that thesesame downsizing measures predicted both environ-mental stimulants (F3 288 - 8.44, p < .001) and en-

vironmental obstacles (F3 288 ~ 4.33, p < .01) tocreativity. Furthermore, with the effects of both envi-ronmental stimulants ()3 = .84, t = 18.10, p < .001)and environmental obstacles 0 = .22, t = 4.79, p <.001) on creativity controlled, these downsizing mea-sures no longer significantly predicted creativity(Afl = .01, p > .25). Thus, the effects of downsizingon creativity appear to be fully mediated by the workenvironment stimulants and obstacles to creativity.

In all of the regressions involving the downsizingmeasures, anticipated downsizing and work groupstability played a more important role than down-sizing experienced. The relative importance of thedownsizing measures is highlighted by correla-tional patterns between those measures and severalexploratory measures from the background ques-tionnaire. Experienced departmental downsizingreported by respondents correlated significantly(p < .05) with only four measures: reports of per-ceived job security (r = .19), morale (self, r = .14,and coworkers, r = .12), and core capability of theunit (r = .21). By contrast, anticipated downsizingcorrelated significantly with several additionalself-reported measures. Respondents who expectedgreater downsizing in their departments in thecoming year reported lower levels of job satisfac-tion (r = .17), job security (r = .37), morale (self, r =:23, and coworkers, r = .35), and core capability ofthe unit (r = .24); they reported higher probabilitiesof voluntarily taking a job in another companyin the same industry (r = .20) and higher probabil-ities of voluntarily leaving the industry altogether(r = .26).

Respondents whose work groups were less stablewere significantly (p < .05) more likely to reporthigher levels of depression (r = -.17), worry (r =-.13), and guilt (r = -.20), lower levels of happi-ness (r = .13), and a greater likelihood of leavingthe company voluntarily (r = —.19). Moreover, theyreported significantly, (p < .05) lower levels of en-trepreneurial behavior on their teams (r = .17), lessrisk-taking (r = .17), lower job satisfaction (r = .13)and morale (r = .18), lower core capability in theirunits (r = .16), and less focus in their own work(r = .14).

DISCUSSION

This study suggests that the event of organiza-tional downsizing is accompanied by negativechanges in the work environment for creativityalong several specific dimensions that have, in pre-vious research (Amabile et al., 1996), been shownto play a particularly important role in project teamcreativity. As expected, this study also uncoveredevidence that creativity itself diminishes during a


downsizing. Most importantly, however, the medi-ational analysis revealed that the relationship be-tween downsizing and creativity can he largelyaccounted for hy negative changes in an organiza-tion's work environment.

For the first time, a study has tracked the timecourse of work environment changes that accom-pany downsizing and has identified varyingchanges in different aspects of the work environ-ment. Although Noer (1993) suggested that survi-vors may not recover from the negative effects ofdownsizing, our data suggest that the perceivedwork environment does improve to some extent.Perhaps, hy wave 3, some people were beginning toaccept the fact of ongoing change within this com-pany, as Noer suggested people must do.

Unfortunately, although productivity and thework environment improved on many dimen-sions, creativity appeared to have still suffered,considerably, even several months after thedownsizing ended. If, indeed, the creativity ofindividuals and teams continued to be depressedwell beyond the end of the downsizing, and wellbeyond an apparent improvement in several as-pects of the work environment, the product inno-vation future of this high-tech company could bein serious danger.

It is particularly interesting that actually experi-enced downsizing was a much weaker predictor ofperceived work environment than was work groupstability or anticipated downsizing; future researchshould focus particular attention on these aspectsof downsizing. The work group stability results arelargely consistent with the theory of the need tobelong (Baumeister & Leary, 1995), which suggeststhat ongoing relational human bonds are a power-ful, fundamental, and pervasive motive, havingstrong positive effects on emotional patterns andcognitive processes. The anticipated downsizingresults suggest that, even if an employee's workunit has been decimated, the certainty of knowingthat the process is over leads to a generally morepositive work environment than the expectation offuture downsizing in a currently intact unit. Con-sistent with affective forecasting theory (Gilbert,Pinel, Wilson, & Blumbferg, 1997), the anticipationof the negative event may be less tolerable than theactual experience.

The results of this study suggest the possibility ofadding a dynamic element to the componentialmodel of organizational creativity (Amabile, 1988,1996). In its current articulation, the componentialmodel is static. It specifies relationships betweenthe perceived work environment and creative be-havior at any one point in time. It does not addressthe dynamics of change in the work environment. It

does not specify how events and patterns of eventsoccurring within organizations might lead indi-viduals to perceive their work environments ascreativity-supporting or creativity-undermining.What sorts of events give rise to such environ-ments, and what sorts of events lead to change inthose environments? If causality can be firmly es-tablished in future research, the relationships iden-tified in the present study will begin to point to-ward an elaboration of the componential model bysuggesting one class of events that may prove tohave a particularly powerful effect on the workenvironment for creativity.

One promising avenue of research could beinvestigation of the mechanisms by which down-sizing events, particularly anticipated downsiz-ing and work group instability, might lead todegraded work environments. Threat-rigiditytheory (Staw et al., 1981) could be especiallyhelpful in guiding such investigations, because itdirectly addresses possible changes in organiza-tional environments under negative circum-stances such as those accompanying downsizing.Specifically, the following effects could be pre-dicted: (1) a centralization of control would leadto perceptions of lower autonomy/freedom, (2) aconservation of resources would lead to percep-tions of less sufficient resources, (3) restriction ofinformation flow would lead to perceptions ofless encouragement of creativity from the organi-zation overall (organizational encouragement),from one's own supervisor (supervisory encour-agement), and from one's work group (workgroup supports), and (4) reliance on familiar rou-tines would lead to perceptions of more organi-zational impediments to creativity, through agenerally greater conservatism. Future researchtesting these predictions derived from threat-ri-gidity theory could profitably expand the compo-nential model of creativity by providing evidenceof specific antecedents to the various work envi-ronment factors. Thus, researchers may begin tounderstand how downsizing and other organiza-tional events bring about change in the perceivedwork environment for creativity.

Clearly, the conclusions that can be drawn fromthis study are limited. First, and perhaps most im-portantly, this study identified relationships, notcauses. It is possible that something else was goingon, at the same time as the downsizing, that led tosome or all of the changes noted. For example,there may have been changes in the employee pop-ulation from which we drew our four random sam-ples, owing to the layoffs and resignations. Respon-dents for the four waves of data collection mayhave used different frames of reference to complete

638 Academy ofManagement Joumal December

the surveys. However, given that our results areconsistent with previous theory and research onboth creativity and downsizing, and given that thedownsizing was the most notable change occurringin this organization during this time, our confi-dence in attributing these results to the downsizingis bolstered. Moreover, our correlational patternsand, in particular, our regression results suggestthat it was indeed anticipated downsizing anddownsizing-related changes in work group stabilitythat accounted for many of the work environmentchanges.

There are other limitations. Because this studywas conducted in only one company, conceptualreplications in other organizations will be requiredbefore firm conclusions can be drawn. Moreover,the low response rate for wave 2 is of concern, eventhough the wave 2 results do generally fit withinthe trajectory from wave 1 through wave 3. In ad-dition, results on the levels of downsizing andwork group stability experienced must be inter-preted cautiously, because they were obtained fromself-reported measures and not from objective de-partmental data. Similarly, because the creativitymeasure was also based on self-reports, it would bedesirable in future research to have additional ex-ternal measures of creativity, such as the expertratings used in earlier research linking the workenvironment to creativity (Amabile et al., 1996).

This study supports previous findings in severalways, but it also goes well beyond those findings. Inthe major previous empirical study of the possibleimpact of downsizing on innovation, Dougherty andBowman (1995] examined changes at the organiza-tional level—problem solving in product conceptual-ization, functional linkages, and strategic linkagesacross the organization they studied. In the presentstudy, we examined changes through the experiencesand perceptions of individuals working in projectteams. We found disruption in innovation activitiesat this level, just as Dougherty and Bowman (1995)found them at the strategic level. In another empiricalstudy, Tombaugh and White (1990) obtained findingssimilar to ours on work group disruptions. We havebeen able to go beyond these previous studies in boththe scope of changes examined and the level of detailuncovered.

Importantly, our methodology embedded severaladvances recommended by previous downsizingresearchers (e.g., Brockner, 1988; Tombaugh &White, 1990): the use of uncontaminated predown-sizing baseline data, examining both "high-down-sizing" environments and "low-downsizing" envi-ronments within a large company, examining thetime course of changes during a downsizing, andcollecting data on group-level variables, such as

work group stability. Moreover, by using surveysassessing the current work environment at variouspoints in time, we were able to avoid at least someof the retrospective biases that might have enteredinto many previous field studies of downsizing.

Given that there is considerable consistencywithin our results and between these results andprevious research, it is possible to identify severalimplications for management practice. This adviceis congruent with much of that given by other re-searchers and theorists (e.g., Caplan & Teese, 1997;Dougherty & Bowman, 1995; O'Neill & Lenn, 1995).First, and most obviously, it is important to do itright. Because the effects of downsizing on thework environment for creativity appear to be soconsistently negative, managers must be sure thatdownsizing is a truly necessary course of action inthe first place. Second, when it is not possible tomaintain team stability, it may be helpful to under-take team-building efforts as soon as new groupsare formed—especially if these are groups fromwhich high levels of creativity are desired. Finally,organizational creativity will be less likely to sufferin a downsizing if the process is concluded in atimely manner and if a downsizing moratorium canbe identified for some meaningful period of timeafterward, thus minimizing the anticipation ofdownsizing.

After the mid-1990s, there was some suggestionin the business press that the downsizing trend wasreversing. However, it is likely that, over time, eco-nomic circumstances will repeatedly trigger a se-ries of corporate workforce contractions. Our study,along with the literature examining long-term eco-nomic outcomes (Cascio, 1993), suggests that cor-porate decision makers of the future should ap-proach downsizing with great caution. The long-term negative effects of such actions on creativityand innovation may only retrigger the corporatewoes that started the cycle in the first place.

REFERENCES

Amabile, T. M. 1983a. The social psychology of creativ-ity. New York: Springer-Verlag.

Amabile, T. M. 1983b. Social psychology of creativity: Acomponential conceptualization. Journal of Person-ality and Social Psychology, 45: 357-377.

Amabile, T. M. 1988. A model of creativity and innova-tion in organizations. In B. M. Staw & L. L. Cum-mings (Eds.), Research in organizational behavior,vol. 10: 123-167. Greenwich, CT: lAI.

Amabile, T. M. 1995. KEYS: Assessing the Climate forCreativity. Greensboro, NC: Center for CreativeLeadership.


Amahile, T. M. 1996. Creativity in context: Update tothe social psychology of creativity. Boulder, CO:Westview.

Amabile, T. M. 1997. Motivating creativity in organiza-tions: On doing what you love and loving what youdo. California Management Review, 40(1): 39-58.

Amabile, T. M., Burnside, R. M., & Gryskiewicz, S. S.1995. User's guide for KEYS: Assessing the Climatefor Creativity. Greensboro, NC: Center for CreativeLeadership.

Amabile, T. M., Conti, R., Coon, H., Lazenby, J., & Herron,M. 1996. Assessing the work environment for cre-ativity. Academy of Management Journal, 39:1154-1184.

Amabile, T. M., & Gryskiewicz, N. 1989. The creativeenvironment scales: The work environment inven-tory. Creativity Research Journal, 2: 231-254.

Baumeister, R. F., & Leary, M. R. 1995. The need tobelong: Desire for interpersonal attachments as afundamental human motivation. Psychological Bul-letin, 117: 497-529.

Billings, R. S., Milburn, T. W., & Shaalman, M. L. 1980. Amodel of crisis perception: A theoretical and empir-ical analysis. Administrative Science Quarterly, 25:300-316.

Brockner, J. 1988. The effects of work layoffs on survi-vors: Research, theory, and practice. In B. M. Staw &L. L. Cummings (Eds.), Research in organizationalbehavior, vol. 10: 213-255. Greenwich, CT: JAIPress.

Brockner, J., Davy, J., & Carter, C. 1985. Layoffs, self-esteem, and survivor guilt: Motivational, affective,and attitudinal consequences. Organizational Be-havior and Human Decision Processes, 36: 229-244.

Brockner, J., Greenberg, J., Brockner, A., Bortz, J., Davy,J., & Carter, C. 1986. Layoffs, equity theory, and workperformance: Further evidence of the impact of sur-vivor guilt. Academy of Management Joumal, 29:373-384.

Brockner, J., Grover, S., Reed, T., DeWitt, R., & O'Malley,M. 1987. Survivors' reactions to layoffs: We get bywith a little help for our friends. AdministrativeScience Quarterly, 32: 526-541.

Brockner, J., Wiesenfeld, B. M., Reed, T., Grover, S., &Martin, C. 1993. Interactive effect of job content andcontext on the reactions of layoff survivors. Joumalof Personality and Social Psychology, 64: 187-197.

Buch, K., & Aldridge, J. 1991. O.D. under conditions oforganizational decline. Organization DevelopmentJoumal, 9(1): 1-5.

Cameron, K., Whetten, D., & Kim, M. 1987. Organiza-tional dysfunctions of decline. Academy of Man-agement Journal, 30: 126-138.

Cameron, K. S., Sutton, R. I., & Whetton, D. A. (Eds.).1988. Readings in organizational decline: Frame-

works, research and prescriptions. Cambridge, MA:Ballinger.

Caplan, G., & Teese, M. 1997. Survivors: How to keepyour best people on board after downsizing. PaloAlto, CA: Davies-Black.

Cascio, W. 1993. Downsizing: What do we know? Whathave we learned? Academy of Management Exec-utive, 7(1): 95-104.

D'Aveni, R. A. 1989. The aftermath of organizationaldecline: A longitudinal study of the strategic andmanagerial characteristics of declining firms. Acad-emy of Management Journal, 32: 577-605.

D'Aveni, R. A., & MacMillan, I. C. 1990. Crisis and thecontent of managerial communications: A study ofthe focus of attention of top managers in survivingand failing firms. Administrative Science Quar-terly, 25: 634-657.

Dougherty, D., & Bowman, E. 1995. The effects of orga-nizational downsizing on product innovation. Cali-fornia Management Review, 37(4): 28-44.

Freeman, S. J., & Cameron, K. S. 1993. Organizationaldownsizing: A convergence and reorientation frame-work. Organization Science, 4(1): 10-29.

Gilbert, D. T., Pinel, E. C, Wilson, T. D., & Blumberg, S. J.1997. Affective forecasting and the durability bias:The problem of the invisible shield. Paper pre-sented at the Society for Experimental Social Psy-chology, Toronto.

Jensen, M. 1986. Agency costs of free cash flow, corpo-rate finance, and takeovers. American EconomicReview, 76: 323-329.

Judd, C. M., & Kenny, D. A. 1981. Process analysis: Esti-mating mediation in treatment evaluations. Evalua-tion Review, 5: 602-619.

McKinley, W. 1993. Organizational decline and adapta-tion: Theoretical controversies. Organization Sci-ence, 4: 1-9.

Mone, M. A., McKinley, W., & Barker, V. L. 1998. Orga-nizational decline and innovation: A contingencyframework. Academy of Management Review, 23:115-132.

Noer, D. M. 1993. Healing the wounds: Overcoming thetrauma of layoffs and revitalizing downsized or-ganizations. San Francisco: Jossey-Bass.

O'Neill, H. M., & Lenn, D. J. 1995. Voices of survivors:Words that downsizing CEOs should hear. Academyof Management Executive, 9(4): 23-33.

Rosenblatt, Z., Rogers, K. S., & Nord, W. R. 1993. Towarda political framework for flexible management ofdecline. Organization Science, 4: 76-91.

Smart, C, & Vertinsky, I. 1984. Strategy and the environ-ment: A study of corporate responses to crises. Stra-tegic Management Journal, 5: 199-213.

Staw, B. M., Sandelands, L. E., & Dutton, J. E. 1981.Threat-rigidity effects in organizational behavior: A

640 Academy of Management foumal December

multilevel analysis. Administrative Science Quar-terly, 26: 501-524.

Tjosvold, D. 1984. Effects of crisis orientation on manag-ers' approach to controversy in decision making.Academy of Management Journal, 27: 130-138.

Tombaugh, J. R., & White, L. P. 1990. Downsizing: Anempirical assessment of survivors' perceptions in apostlayoff environment. Organization Develop-ment Journal, 8(2): 32-43.

Van de Ven, A. H. 1986. Central problems in the man-agement of innovation. Management Science, 32:590-607.

Whetten, D. A. 1981. Organizational responses to scar-city: Exploring the obstacles to innovative ap-proaches to retrenchment in education. EducationalAdministration Quarterly, 17(3): 80-97.

Woodman, R. W., Sawyer, J. E., & Griffin, R. W. 1993.Toward a theory of organizational creativity. Acad-emy of Management Review, 18: 293-321.

Teresa M. Amabile is the Class of 1954 Professor ofBusiness Administration and Senior Associate Dean forResearch at Harvard Business School. She received herPh.D. in psychology from Stanford University. Her currentresearch interests focus on the impact of specific eventson project team work environments, motivation, and cre-ativity.

Regina Conti earned her Ph.D. in psychology at BrandeisUniversity and is an assistant professor of psychology atColgate University. Her current research investigates self-regulation processes, including the maintenance of in-trinsic motivation, the prevention of procrastination, andthe strengthening of willpower.

changes in the work environment for creativity during...

Documents