the kirton adaptation.pdf

7/27/2019 the kirton adaptation.pdf

1/8

Review of Public Personnel Adm1nistration spring 1999

Edi tor ia l ScaffPIIbIiaberLDouaWDobeonuse lnadrute olPublic AffirinEcIitor-la.cbWNkhoIu P.Lovrich,Jr.Wuhineron Srate UniveniryMaaaciDlEdftorSteVen W. HaysUniveniry of South Carolina

Seaior Editorial A'"'-Y BoardJ- S. Bowmaa FIorid8Scate UniwntcyDouWICIiaper FloridaIncemadon8IUntv.DmdH. Ro8nbIoom AmericanUnMmicyO. GIeaa SC8bIAr\tnaton.VA'I'I18IIJ. 'J'hoMp.oo SUNYIeAlban,

HighliglwEditorIal AdvI8ory Bo.rd

C8dyaBMa UntveniryoiPlcabulahE_a.- UntvmiryofCenll1llFloddaN.JoeepIa. ,. AriIonI~ UnIvenkyCMhyC- Pon~AuchorityDeDak D8Iey NCSU MRaIeIanAIIIIIoDeNW Rutam UniwnitvDeaai8 DNe.ac UniftnicyofWl.aninRobercEllioa Auburn Untwnityt.I8k ~ Univenicyof ConnecdcucGe..ldG.bria NonhemIllino8Un1wnicyC. BllueGnhala use acColumbia}oWy E. Guy FIodda Scare UnivenicyPII IridI"""" Syncu.e UnivenicyJ.I!dw8rdJCellouP UntvmiryofGeoqiaCamlLiaerSanM-. TX~Manouo TOWIOIISraeeUniYenItyJobaN.a-fim Uniwniryofx-s.n-~ UniftnicyolArltar..Feh A. Nipo Athena, OAIJo,d Nipo GeorgiaSraee UnivenicyJ- L Peny Indi8n8 UnivenicyGary Pokomy ElCcrrico,CANotma lUcc:ucd SUNYacAhmyGary RoI8t8 FairleiahDicldNoa Uniwni&yAbaSebeeia Calibn8SraeeUnivenityJ-ySWriu UntvmiryofPkabuJlhADIDIIIoSia.oe UnivenicyollDinoiaRoaaIcI D. s,t.Ia San.JoeeScareUnlYeniryr..- Selia \.JnMnityofM8louriJonMt- Weac UniwnlryolMiImIHoycN. Wheeler USCacColumbia

5 u.lna PerformanceMeasurementin HumanResourceManagementA SlII'Vey


2/8

The Kirton Adaptation--InnovationInventoryValidity Issues, Practical QuestionsMICHAEL BaBIC, EMMANUEl COllEGEERIC DAVIS, UNIVERSITYOF CALIFORNIA, RIvERSIDEROBERT CUNNINGHAM, UNIVERSITY OF TENNESSEE, KNOXVILLE

The Kirton AdaplallOl1./nnowUon Irwtntory (KAI) isdcsipcd co"",as"", proptnsi" coin""""".........pm" i" 10adapI, a pmonali" dimtnsian c/aimtd .ignificant.for wndmmding and buiIdint organitalianal,{fWMtvU, Thisani& p",mu G !niL! of ""lidiry chtdu for u.. KAI. r


3/8

controlled organization mayproduce goodsand serviceswhich are superior to alterna-tives, but clients are not yet ready to acceptthe new concepts. A cohesive dedslon-making team Including both adaptors andinnovators is more likelysensitive to envi-ronmental challenges.KAI:TIiEORYANDOPERATIONAUZATIONAssuming that personal ity II composed ofabilities. traits, and styles, the KAI measurescognitive style--()ne 's preferred or charac-teristic manner of processing information.Conceptually, KAI has three dimensions:

I . Rule/Group Conformi ty (R)2. Efficiency (E)3. Sufficiency vs Proliferation of Origi-nality (0) IRule/group conformiry (R) i s the degree

[0 which one works within the acceptedstructures or rejects cri tical elements ofsuchst ructures. Innovators may overlook or ig-nore pressures to conform to consensualviews on what is needed and how to get it;adaptors are more l ikely to abide bysystemdictates (IZ items). Examples ofconformityi tems inc lude "A person who l ikes the pro-tec tion of prec ise ins truc tions" (Q29), and"Aperson who prefers colleagues who neverrock the boa t" (Q32).Efficiency (E) exemplifies Weber's no-t ion of the legal -rat ional bureaucracy. Theefficient manager emphasizes precision, re-l iabi li ty , and efficiency. Innovat ion i s dis-cont inuous to the exist ing system, therebyto be rejected as hindering efficiency (7items). Sample items for efficiency include"A person who is thorough" (QI4), and "Apers


4/8

third factor, originality.Table 1presents the factor loadings and

the amount of variance expla ined by thosefacrors. All variables for a given factor loadat greater than .30, and in every case, thenext highest loading score isstatist ically sig-nif icant ly smaller than the score for the fac-tor chosen as its primary loading dimension.

Twenty-five of the thirty. two items(78%) l oad strongest on t he predicted fac.t or, with t hree fact ors accounting for 33%of the variance. Such a resul t supports t he,'"nc"nli,'II, 1(11hI"\'('.dimt'nNiolllll NtrllCtllre.Since explained variance isa function of thedegree to which the i tems are measuring thesame thing, account ing for a hi gh propor-t ion of variance would suggest i tem redun-dancy. On the other hand, too little vari-ance accounted for indicates the absence ofa unique, clea rly defined dimension. Wi theach of the three factors accounting for arespecrable proport ion of the variance , thethree-fac tor solut ion i s a reasonable inter .pretat ion. The subgroup persona li ty mea.sures al so correlate moderately with eacho ther and with t he KAI overal l score, indi .eating that subscales are tapping a commonunderlying phenomenon; see Table 2.CONTENTVALIDITYConcl'T1lI!alidity checks whether al l s igni fi -cant aspect s of t he domain are i ncluded i nthe measure. This is difficult to detennine,for there are not clear ly defined behaviorsto match adaptive and innovative att itudes,except perhaps at the extremes. Dimensionsof Kirton's conceptual meaning wereoperationalized by choice-pairs of phrases tocrea te an a lternat ive measure (Al tkir t) ; seeAppendix. Respondents chose the ones tatement from a two.s ta tement pai r whichhener descri bes himself or herself. Wit h aPearson's r of .65, KAI and Al tki rt appearto he tapping the same dimension; see Table

3. Altkirt does not dispute a content valid-ity claim.CRITERION VALIDITYCriterionI!aliditycan be establishedbyshow-ing KAI scores appropriately relating toother measures, either as predictors or asmeasures of similar content. Here, KAIscores will be checked for (a) consistencywith a content analysisof leadership storiestold bymanagers, (b)stabilityover time (totest the assumption that KAIscoresare per-NonlllitymCIISllreswhich do not (hRnlle eM'ily), and (c) consistency when managerscarryout a task within a group fairlyhomo-geneous with respect to KAI scores.

Contentanalysisofleadershi pstories. Inthe 32.item KAI,Kirton sets the agenda bydefining the dimensions and constructingthe items. Respondents assess their com.fort levelwith Kirton's items. The instruc.tions read:"Howeasyor difficultdoyoufindit to present yourself, consistently, over along period of time as "

In 1985each middle manager was in.terviewed privatelyduring the managementdevelopment programand askedto telllead-ership stories about situations in which heor she wasinvolved. The interviews, whichlastedfrom20 minutes toover an hour,wereaudiotaped and transcribed. These storiesare assumed to represent typical ways thatthe manager presents self.Becausethe sto-ries which managers tell are self.generatedand personal, the respondent is obviouslydescribingamanagement stylewhichis com-fortable to himor her.The storiesbridgethecognitive stylelbehavior gap, forthe storiesare self.reports of a manager's behavior.

The manager's comfort level is re-flected by the customaty behavior presentin the story. The coder's challenge isto de-termine whether the manager's comfortlevel lies in adapting or in innovating.

H RI\ "IFW or r l JR! Ie rFRSnNNH. " "MINISTRAnON SPRING19Q9

"

TABLE 1.Kir ton Adaptat ion Inventory, Rotated Fac tor Mat rix (N-587)RULESR) EFFICIENCY(E) ORIOINALITY0)

Q24 0 .71 -.05 .05Q29 R .64 .17 .\2Q32 R .60 .00 .10Q27 R .53 .22 .08Q30 R .50 .21 .28QIO R .48 -.02 .25Q20 R .44 .14 .00Q8 R .42 .37 .03Q28 E .41 .24 .15QIJ 0 .40 .11 .1/1Q7 R .39 .27 .15Q2 R .35 .27 .21

Q14 E


5/8

TABLEZ. Pearson's r among Subscale. (N-S87)RULES

R0KAI .88---...-

ORIOINALm EmCIENCY.49 .44

.13

.6177

RULES ORIGINAUTYTABLE 3. Pearson's r: Altkirt with KAland Subscalel, Student Sample

KAIAlrkirtN=R4

.62 .50EFFICIENCY

.40 .65

TABLE 4. KAI Mean: Pre- and Post-Training by Year, State Managen

Kirton provides 14characteristics of adap-rors. and their corresponding innovatorcharacteristics (Kirton.1989.pp. 8-9). Re-spondenrs' stories were read and coded the-matically by Bobic and Davis separately.Each theme dimension was allocated onepoint. so each respondent's score rangedfrom I to 14. The coders then discussedand justified their assessments of eachrespondent's score on each ofthe 14dimen-sions" On the initial pass.80 percent agree-ment was achieved between the coders.Respondents who received different scoreslayclose to the midpoint. Consultation re-solved those disputes.

Changeooertime. Kirton (1989, chap.i) insists that KAIis a measure ofpersonal-ity.therefore stable over time and not easilychanged. Van der Molen (1989) and Clapp(1993) offer supporting evidence. Oold-

smith (1989)reports that training programsappear to have no impact on KAI scores.KAIstability has important Implicationsfororganizations which Invest In training pro-grams to teach people to be innovative.

If training programscan teach innova-tive behavior.then not only doesKAItheoryerr in assuming the adaptation-Innovationdimension as an Integral aspectof personal-Ity.but organizations are spending moneywiselyas they train people to innovate, as-suming that their organization lacksInnovativeness. However,ifKirtonandoth-ers are correct on this issue.organizationsseeking to change their culture shouldcon-sider recruitment, replacement, or buildingdiverseteamsrather than training toachievethe needed balanceof managerialstylesOverthewholecognitivetylerange. .State manager training sought to en-

14 Rf\'lfW Of M.'III.I( 'PERSONNel AOMINISTRATION SPRING1999

.-1

i.~

courage empowerment, risk-taking, and in-novation. This intensive residential pro-gram took state managers away from theirworksltel Intoa campus environment wheretheyworkedwithpractitioners and academ-Ics In an experiential program incorporat-Ing both Individual and group projects. Ifthe program Increased propensity to Inno-vate. participants should have gravitatedtoward the Innovative pole of the KAI Inthe three-month Interval betWeenorienta-tion and program conclusion. Enthusiasticevaluation of the program by participantssuggeststhat their KAI scores should havemoved toward the innovative end of thespectrum; yet aggregateKAI scores did notbudge. see Table 4. The slight changes arerandom.

This evidence supports the argumentthat training cannot easily change KAIscores. Perhaps the messageof the trainingprogram takes longer to soak in. In thatcase, scoresshould rise inone to three yearsafter the management development experi-ence. Some participants filled In KAI pro-tocols one to three years after the programwas concluded. The results are found InTable 5. Ai8l'egate scores remain stableacrosstime, whichsupportsthe originalfind-Ing that training programs do not changeKAI scores.

However, racial differences emerge.Table 6 contrasts the mean KAI score be-fore and after training. Becauseof the .mallnumber of black respondents for whom in-formation from three time points is avail-able, the data are pooledforall respondentsfrom 1985 through 1987. Black managersbefore training are lessinnovative than theirwhite peen. At the end ofthe training pro-gram,KAIscores forblackmanagersremainunchanged. However, after one to threeyears, innovation propensity among blackrespondents hu Increased to the point that

they are as innovative as their white col-leagues, who remain unchanged.This isa small data set (7blackman-agers and 43 white managers) from whichto extract a firmconclusion. but the find-Ingsoffer an Idea to explore. Perhaps mi-norities have a narrower range of behav-Iors acceptable to majorities. Successfulminority managers are sensitive to the pa-rameters of group norms, and stay withinthem. The training program encouragedmanagersto gobeyond the boundaries. Thereinforcement of superiors, peers. and in-structors that innovative behavior15accept-able. even desirable. may have fallen on afertile field among minorities, who brokebarriers to achieve these management posi-tions and who may have constrained theirnatural Innovative inclinations inorder notto risk damaging their careers.

Workenvironment mayInfluenceone'.Innovative or adaptive attitude. The Influ-ence of work environment Is reported sig-nificant In Holland. Bowskil l & Bailey(1991), whofounda regression to the mean.where new employees' KAIscores differedfrom the mean of the organization, but nochange where new employees' KAI scoresdid not differ fromthe mean.

We tested the influence of work envi-ronment on KAI scores. The governor inofficefromJanuary 1979untilJanuary 1987qot only encouraged training to Improvemanagement, he also preached to cabinetmembers and other top level officials thephilosophyof empowerment and devolutionof authority. Cabinet members were ex-pected toworkwith their own lineand staffto create more effective state government.The commissionermade the decisionsin thedepartment; the governor's staff memberswere subordinate. and stayedout ofdepart-ment matters unless Invited In. Informalconversations withmiddle managersduring

KIRKTONAOAmON.INNOVATION... CIINNINGHAMETAt. n

1985 1986 1987 1988 1989 1990N=24 N=23 N=23 N=27 N=25 N"'27rr".rraininl: 100.8 104.8 105.7 98.6 97.6 99.8P",r-rraining 99.9 104.9 108.2 96.7 96.2 101.6


6/8

TABLES. Kirton Scores: Pre-training, Post-training and Follow-up.1987N=16

106.56108.31106.81

"Indu,te, onlv tho", ".10 manaRenwhohave ocarelforallthr.. observations._d___.

TABLE6. Mean Kirton Scoresand Standard Error Over Time by.Race. StateManagers Only

the summertrainingprogramsindicated thatparticipants understood the governor'smes-sa~eand appreciated the opportunities forinnovation offered by this managerial phi-losophy.The next governor adopted a contrarymanagerial philosophy. Policy decisionswere made by the governor's personal sraff,and cabinet members were subordinate tostaff on policymatters. Deviation from ex-isting policy had to be cleared bycommis-sioners through the governor's staff. Com-missioner and middle manager innovationwasnot encouraged. The managerial envi-ronment emphasizedadaptation rather thaninnovation. Several managers expressedtheir discomfortat havingdecisionsimposedupon them bythe governor'sstaff.One cabi-net member stated privately that he some-times learned from the newspaper aboutdecisions affecting his department. Thismanagement style differs sharply from theprevious governor.Note fromTable4 that KAI pre-train-

ing scores from 1983-1987 (firstgovernor)are at or above 100. Pre-training scoresfor1988-1990 (second governor) do not reach100. The differences betWeenthe tWogov-ernors are consistent over the data set, andsupport a speculation that while trainingmay have no effect on KAI score, the cul-ture of the workplace appears to have animpact. An innovative workplace culturefostershigher KAIscores;an adaptive work-place culture encourages adaptation. Em-phasizingthe influence ofthe environmentis compatible with the Deming argument(Bowman, 1994) that the workplace envi-ronment and task structure are significantsources fororganization effectiveness. Theinference from this finding is that to effectchange, the organization executive shouldrestructure theenvironment orredesignthejob rather than flre employees. Sendingpeople to training programsmay teach spe-cmc work behaviors, but does not appear togenerate an innovative orientation. How-ever, a suppressed innovative attitude can

I t> RI' \' IFW t)F r ll l\UC rERS0NNEt AOMINIS11lAllON srRINO 1m

,IB


7/8

TABLE7. KAIGroup Score and Innovative Group Behavior, State ManagenINNlW A'I\)RS

GROUP

Mean KAI score*Departure from

OI:mJard practice

I2Z7

MllmLESGROUP

ADAPTORSGROUP

975

793

.Composj.e expen score(I =srandard practice. 10=sharp departure).

middle group as offering the best projectplan.

These three tests of criterion validitysupport the KAI asa valid measure ofdeci-sion-making style, corroborating previousevidence forconstruct andcontent validity.Scoresremain stableovertime, andKirton'shypothesis that balanced teams outperformteams nfhomogeneous adaptors or innova-tOrsisconfirmed.POLICY IMPLICATIONSOrganizations face the paradoxical chal-lenge of maintaining the satisfaction leveloftheir current stakeholder basewhileseek-ing new opportunities and more effectivewaysofdeveloping, producing, and deliver-ingthe product or service.Becauseindividu-als differin their propensity to welcome orseek change, the work group which incor-porates and celebrates a wide spectrum ofemployeeorientations to change willbepre-pared !>othto hold onto the bestof the oldand to introduce the new.Byassessingindi-vi,llIals "n Ihe a,IAptlltinn.innovlltion di.mensi"", KAI can assist management inbuilding effective work teams. However,one's KAI score may also be influenced byone's work environment or minority socialstatus. Individuals take cues from manage-ment and peers . KAIcan be useful inen-hancing innovative problem solutions, butan executive's demonstrated behaviors re-

main a factor influe~cing a subordinate'spropensity to innovate.SUMMARYBasedupon multiple data sets and multiplemethods, this research findsthe KAIto be avalidmeasureof the adaptation-innovationdimension ofmanagerial decisionstyle.Fac-tor analysis supports tri.dimenslonality forthe overall KAI,and confirms the Internalstructure of subscales.Triangulation byem-ploying an alternative method of testingadaptation-innovation (Altklrt), a contentanalysis of manager stories, and expertevaluation of three groups' project propos-alsarguethat KAImeasuresthe innovation-adaptation continuum.

KAIappears to be tapping a stable per-sonality dimension. A significant trainingexperience did not significantlyelevate (orlower) innovation scores. While the dataindicate that one's KAIscore remains stableover time, minority group status or the or-ganization culture communicated by topmanaKement likelynudge. scores to comein linewith the executive's predisposition.In each test , Kir ton's theory Iscon-firmed. Innone ofthe tests does the KAIorKirton's theory fail. Seeking organizationaleffectiveness through work-groupdiversityalongthe adaptation-innovation dimensionreceivessupport and iscompatiblewith priorresearch showing that organizational effec-

ltI RF\'IE\\' ("IfrlJRue: PERSONNELADMINISTRATION SPRING1m

tiveness isproduced bythe requisite varietyprinciple (Miller, 1993;Watson, Kuman &Michaelsen, 1993;Weick, 1979).The prac-tical managementstrategy fororganizationaleffectiveness is to recruit or build diversedecision teams.

Ifthe adaptation.lnnovation dimensionIssignificantforsupplyingneeded variety foran organization, and i( the theory of requi-site variety applies to organizations, then(other (acton being equal) organizationswhich use managers' KAI scores for build.ingdiversework teams willbemoresuccess-ful than organizations which ignore suchinformation about their employees whenbuilding teams.

':,:

Notes'More .han II0reoeRrch art icinmakl ll lluoeo(the KAIare reponed In the P8ycINFO 1996. lmdatabue.I TI le I I.. of I tems comp rill "8 e ac h l ubo ea le c an b efound In Klnon (1976).'Oulon (1980) and Mel li ck (1988) I ll te t ha t val id it y

I Imore a uni ta ry t han a l ri nl tl ry concept . T Il e cat el /Oryoeparatlon8 are.uaed here for orlfllnlzR.fonal PUrpmel.. Conceptua ll y, t he orl ll nR ll ty . e ff ic iency and rul e

con(ormlt y l uboeal e. a re orrho lOnal (uncorrel lt ed) Inmul ti dimens iona l Ipace . and t he overa ll KAl il l inear Intwo-dimensional space. Inrwo.dlmen.lonallpace the Itemlform a .1"8le me..ure; 10delplte the orthotronalltruc.ureof the three component (acton. .he"" .ub.scorel wil lhavemodes. pooluve In.ercorrelatlon8. Bydellnltlon, orthogonalfacton are uncorrelated. Howeve. . when lpeciAc Item. are. ummed tocrea te l ubocal el . t he (ac .o r l oadl Rndeffect lo f non luboeal e i tems are removed , . herebv Rll owln ll (orthe I'OIIIbility of either pool.lve or nelflltlve correl"lon.all lOtl l lubocalel . In thl l Cloe. al expected, the lu, ,"cole.lhawed modest pooltlve IntercorreIA.ionl. Thll mean. thaIthe .uboeales are Independent . ~ . related to the umbrellAdimension of adap'ation.lnnovotlon.

, The d lf fi irence berween the " '0 l arBeIt l oadl nl ll I ..09. (the smalln. difference Inour do.a) and 18"""Il IcAntat ,,-.0202 (t- 2.304).

. I n . h. l. own penona l KAI le nr e, n oh ic I nJ DRv l.leore approximately one . .anJArJ d.vlRllon from.he meanIn oppot ite dlrectlonl. Thl l dlvenlry be!Ween the codenreduce. the likelihood that a .Igniflcant theme IIoverl ooked. and the con venat ion. ( li te ra ll y vehemenlargumenu) between the codell after their InlUalIndependent jud(ll1len18lncreRaed the likelihood o(a validInt et pf tt at ion of t hemel f rom the sro ti e. . The d if fe rentp enpe cti Ve i of Bobk: a nd Dav ll I nd .h at Impa cl o n l hel eorl "8 point s up t he Impor tl nc . o f seekl ll ll d lver ll ty Inoel ec tl ng code. .. Val id it y I I enhRnced by a t horoughdlacuaion amotll coden with dUI'eri", perspective.. Scorillll

done bycoden wl.h . imil .r ocore. onthe dlmen.lons .el ledmAYachieve reliability, vel .acriAce validity.

, Only . he pr" ll 'am coord lnAton knew of t hl . 1110111 '"'AIiAcation.

ReferencesAlhby,w. R. (1956).In.mJ..crion10c,btmtric


8/8

anJ ",,'a""c. Journal..f AppheJ rs~,h"I"I

the kirton adaptation.pdf

Documents