Anda di halaman 1dari 20



ASAC,2008       MarylneGagn

Halifax,NovaScotia        ConcordiaUniversity

JacquesForest

UniversitduQubecMontral

MarieHlneGilbert

UniversitdeMontral

CarolineAub&EstelleMorin

EcoledesHautesEtudesCommerciales

AngelaMalorni(Student)

ConcordiaUniversity

THE MOTIVATION AT WORK SCALE: VALIDATION IN TWO LANGUAGES5




The Motivation at Work Scale was developed according to self-


determination theory. We examined the structure of the MAWS in
English and French. Results suggested that the structure of motivation at
work across languages is consistently organized in four different types:
intrinsic motivation, identification, introjection and external regulation.


5
Marylne Gagn, Department of Management, John Molson School of Business. Jacques Forest, School of Management.
Marie-Hlne Gilbert, Department of Psychology. Angela Malorni, Department of Psychology. We thank David Goodridge for
helping with data collection. Preparation of this article was facilitated through a grant from the Fonds Qubcois de la Recherche
sur la Socit et la Culture to the first author and a post-doctoral fellowship from the Institut de Recherche Robert-Sauv en Sant
et Scurit au Travail (IRSST) to the second author. Correspondence may be sent to Marylne Gagn, Department of
Management, John Molson School of Business, Concordia University, 1455 de Maisonneuve West, Montral, Qc, Canada, H3G
1M8, e-mail: mgagne@jmsb.concordia.ca.

157
The MAWS subscales were predictably associated with organizational
behavior constructs.


Despitethefactthatworkmotivationisoneofthemajortopicsinorganizationalbehavior,not
many work motivation surveys exist (exceptions count VandeWalle et al., and Amabile et al. ). We
developed and validated the Motivation at Work Scale based on the framework of selfdetermination
theory(Deci&Ryan,1985a,2000).SDToffersamultidimensionalconceptualizationofmotivationthat
allowstheassessmentoflevelofmotivationandtypeofmotivation.Combiningallfieldsofpsychology,
SDT has yielded over 400 empirical publications since the early 1980s. It is a dominant theory of
motivationinsocial,education,andsportpsychology.Inthesefields,validatedmeasuresofmotivation
alreadyexist(Grolnick&Ryan,1987;Guay,Vallerand,&Blanchard,2000;Pelletieretal.,1995;Ryan&
Connell,1989;Vallerand,Blais,Brire,&Pelletier,1989;Vallerandetal.,1992),butwedonothavea
reliable,valid,andpracticalmeasureofworkmotivationthatfollowsthetraditionofSDTinthefieldof
organizationalbehavior.Blais,Brire,Lachance,Riddle,andVallerand(1993)publishedthevalidationof
aFrenchmeasureofworkmotivationthatwasgroundedinSDTbuttherehavebeenreliabilityproblems
with some of the subscales as well as face validity problems with some of the items (Gagn, Brub,
Donia,Houlfort,&Koestner,2007;Gagn,Boies,Koestner,&Martens,2004;Gagn,Boies,Koestner,&
Martens, 2005). For example, an intrinsic motivation item in the Blais et al. scale is Because of the
pleasurableexperienceoflearningnewthingsatthisjob.Itisdifficulttoknowifthisshouldrepresent
identified or intrinsic motivation. In addition, it is a long 31item scale, and has not been validated in
English.FollowingtheSDTtradition(Grolnick&Ryan,1987;Guayetal.,2000;Pelletieretal.,1995;Ryan
&Connell,1989;Vallerandetal.,1989;Vallerandetal.,1992),wecreatedtheMotivationatWorkScale
in two different languages, assessed its structure using confirmatory factor analysis with invariance
tests,andexamineditslinkstoantecedentsandoutcomesthatarerelevanttoorganizationalbehavior
(Gagn&Deci,2005).

The Continuum of Motivation


 

SDTproposestwooverarchingtypesofmotivation,intrinsicandextrinsic.Intrinsicmotivationis
definedasdoingsomethingforitsownsake,becauseitisinterestingandenjoyable.Extrinsicmotivation
is defined as doing something for instrumental reasons. These instrumental reasons can differ,
dependingonhowinternalizedthemotivationis.Internalizationreferstotakinginaregulationthatwas
initially regulated by external factors, such as rewards or punishments, so that it becomes internally
regulated. Internalization can vary in terms of how well it is assimilated with a persons existing self
regulations,suchasvaluesandintereststhatthispersonalreadyholds.Thedegreetowhicharegulation
isinternalizedgivesrisetodifferenttypesofregulationsormotivations,sothatextrinsicmotivationcan
becompletelyexternallyregulated,orcanbepartiallyorfullyinternallyregulated.

158


At the low end lies external regulation, which refers to doing an activity in order to obtain
rewardsortoavoidpunishments.Behaviorsoregulatedisthereforecompletelynoninternalized.Next,
introjected regulation refers to the regulation of behavior through selfworth contingencies like ego
involvement and guilt.  It involves taking in a regulation so that it becomes internally pressuring, and
thus involves only partial internalization that remains controlling, not volitional.  Introjected people
engageinabehaviororcommittoanactivityoutofguiltorcompulsion,ortomaintaintheirselfworth
(Koestner&Losier,2002).Next,identifiedregulationreferstodoinganactivitybecauseoneidentifies
withitsvalueormeaning,andacceptsitasonesown,whichmeansthatitisautonomouslyregulated.
Identified peopleengage inabehaviororcommittoanactivity basedonits perceivedmeaningorits
relationtopersonalgoals(Koestner&Losier,2002).Finally,integratedregulationreferstoidentifying
withthevalueofanactivitytothepointthatitbecomespartofapersonshabitualfunctioningandpart
of the persons sense of self.  This is the form of extrinsic motivation that is most autonomous and
internalized.Thus,wewouldsaythatanurseisidentifiedifshefullyendorsesbathingapatientforthe
healthofthepatienteventhoughitmaynotbeanenjoyabletask(thusnotintrinsicallymotivated),but
the motivation would be integrated if the nurse goes further to say that her job is a vocation.
Integrationandidentificationdifferfromintrinsicmotivationinthattheactivityisdonenotsomuchfor
its own sake (because it is interesting and fun), but for the instrumental value it represents.
Identificationandintegrationaredrivenbyvaluesandgoals,whereasintrinsicmotivationisdrivenby
emotionsthatemergewhileengagingintheactivity.

 Research in different domains, such as education (Williams & Deci, 1996), sports (Vallerand &
Fortier,1998),work(Blais,Brire,Lachance,Riddle,&Vallerand,1993),andhealthcare(Williams,Grow,
Freedman,Ryan,&Deci,1996),hassupportedthatthetypesofmotivationformasimplexlikepattern
(Guttman,1954)thatrepresentsvariationintheunderlyingdegreeofinternalization,whichmeansthat
each subscale correlates most positively with the subscales closest to it (e.g., intrinsic and identified
regulation) and less positively or more negatively with subscales farther from it (e.g., intrinsic and
externalregulation).However,eveniftheyareononecontinuum,researchhasrepeatedlyshownthat
thereisaclearbreakintheconsequencesofeachtypeofmotivation.Ifwebreakthecontinuumdown
themiddle,wecanrecategorizethetypesofmotivationsuchthatexternalregulationandintrojection
represent controlled motivation, and identification, integration and intrinsic motivation represent
autonomousmotivation.Dependingontheresearchquestionofinterest,wecansometimesusethese
aggregates,andsometimesusethediscretetypesofmotivation.KoestnerandLosier(2003)hasshown
thatwecanfinddifferentbehavioralandattitudinaldifferencesbetweenintrojection,identificationand
intrinsic motivation in certain domains, like environmental behavior (Pelletier) and political behavior
(Koestneretal.).

159
MeasuringtheContinuumofMotivation


Intrinsicmotivationandeachtypeofextrinsicmotivationarereflectedindifferentreasonsfor
behaving, and these reasons provide a means for assessing the types of motivation (Ryan & Connell,
1989). Incumbents rate various reasons for doing a particular behavior or activity that reflect intrinsic
motivation or one of the types of extrinsic motivation in terms of how true they are for them.
Motivation can be assessed at different levels of analysis. Vallerand (1997) suggested that motivation
canbemeasuredatthelifeorgloballevel,atthedomainlevel(e.g.,work,education,leisure),andatthe
state level (e.g., specific task, specific period of time). The MAWS assesses work motivation at the
domainlevel,butotherscalesexisttomeasure,forexample,teachersworkmotivationforspecificjob
tasks(Fernet,Sencal,Guay,Marsh,&Dowson,inpress).

CreationoftheMotivationatWorkScale


Wecreateditemstomeasuredifferentworkrelatedbehavioralregulationsthatrepresentthe
range of the continuum of motivation to do a particular job. We chose to focus on specific types of
motivation that we thought were most useful to assess in the work domain. We did not include
amotivation (i.e., lack of any type of motivation) items because we wanted the MAWS to focus on
measuring the type of active motivation in organizational employees. We also did not include
integration items as it is typically very difficult to psychometrically distinguish integration from
identification(Vallerandetal.,1992).Wethereforecreated5itemsforeachofthefollowingsubscales:
external regulation, introjection, identification, and intrinsic motivation. Six of the items were taken
fromtheBlaisetal.(1993)scale,butsomeofthewordingwassimplified,andtheyweretranslatedinto
English.TherestwerecreatedsimultaneouslyinFrenchandEnglish.

TestingtheMAWS


We administered the MAWS along with other scales described below to various samples of
French and English speaking Canadian workers. We conducted confirmatory factor analyses to trim
downtheMAWSto12itemsandtestitsinvarianceacrossthetwolanguages.

To test the validity of the MAWS, we examined relations between the MAWS and known
antecedents and consequences of motivation. SDT suggests that intrinsic motivation and the
internalizationofextrinsicmotivationaredeterminedbythedegreetowhichpeoplecansatisfythree
basicpsychologicalneedsforautonomy,competenceandrelatednessintheenvironmentinwhichthe

160
activitytakesplace.Thesatisfactionoftheseneedscanbeaffectedbyenvironmentalpressures,suchas
deadlines, surveillance and contingent rewards (Amabile, DeJong, & Lepper, 1976; Deci, Koestner, &
Ryan, 1999; Fernet et al., in press; Lepper & Greene, 1975). It can also be affected by interpersonally
controlling or supportive behavior of an authority figure, such as a teacher or a manager (Deci et al.,
2001;Koestner,Ryan,Bernieri,&Holt,1984;Lepper&Greene,1975).Finally,itcanbeaffectedbythe
design of tasks and jobs (Gagn, Sncal, & Koestner, 1997). To test some of these premises, we
includedameasureofneedsatisfactioninoneofthesamples,expectingthatneedsatisfactionwould
be positively related to intrinsic motivation and identified regulation, and negatively related to
introjected and external regulation. We also used a highly used measure of perceived organizational
support (Eisenberger, Huntington, Hutchison, & Sowa, 1986), expecting that it would be positively
related to autonomous motivation and unrelated to controlled motivation. SDT also proposes that
certain individual differences or dispositions can influence the type of motivation adopted, such as
causalityorientations(Deci&Ryan,1985b).Wedidnottestcausalityorientations,becausethescalehas
low psychometric properties, but instead examined links between work motivation and optimism, a
stabledispositionthatwethinkcaninfluencethedegreeofinternalizationofanactivity;DeciandRyan
(1985b) found that an autonomous causality orientation, which is associated with more internalized
motivation towards activities, was positively associated with selfesteem and negatively with self
derogation.

Wealsoexaminedlinksbetweenthedifferenttypesofmotivationandsomeoftheoutcomes
thathavebeenstudiedinothervalidationsofsimilarSDTbasedscales,aswellassomeorganizationally
relevant outcomes: job satisfaction, organizational commitment, turnover intentions, wellbeing, and
psychological distress. Again, we expected that intrinsic and identified motivation would be positively
correlated to positive outcomes (i.e., job satisfaction, affective commitment, and wellbeing), and
negatively to negative outcomes (i.e., turnover intentions and psychological distress), and that the
opposite pattern would be found with external and introjected regulation. The magnitude of these
correlationsshouldbealignedwiththeirdegreeofinternalization(e.g.,positivecorrelationwithintrinsic
motivation should be slightly stronger than with identified motivation). Finally, we compared the
motivationscoresofemployeesindifferenttypesofjobstoseeifwecouldfindmotivationdifferences
basedonthetypeofworktheydo.Workersinjobsthattypicallyhavelowautonomyorlowdecision
makingpower,poorrelationshipsorlowcompetencewouldbeexpectedtoscorehigheroncontrolled
motivationandloweronautonomousmotivation.

Method


Participants


161
Data were collected from convenience samples of Canadian workers in different industries, in
twodifferentlanguages.

Englishversion.Wecontacted2795pilotsfromacommercialairlinecompanybyemailthrough
theiruniontocompleteawebsurveythatincludedtheMAWS,ashort8itemversionoftheperceived
organizational support scale (Eisenberger et al., 1986),  = .86, and the 5item satisfaction with work

scale(Gagnetal.,2007),=.71.Eighthundredeightyonepilotscompletedthesurvey(32%response
rate),outofwhich98%weremenwithanaverageageof52years.

Fiftyfivemiddlemanagersoutof66thatwerecontactedinfourdifferentgroundtransportation
companiesalsocompletedtheMAWS,69%ofwhichweremenwithanaverageageof43.91yearsand
3.5yearsoftenureinthecompany.Finally,130undergraduatecommercestudentswhoworkedpart
time(86%)orfulltimecompletedawebsurveythatcomprisedtheMAWSaswellasameasureofwork
satisfaction (Gagn et al., 2007) in exchange for extra credit. Fiftythree percent were men with an
averageageof22years.

French version. Twentythree advanced undergraduate students distributed paper surveys to
285 French Canadian workers who were categorized to work in one of five sectors: Education,
management/professional, sales, health, and secretarial work. The survey included the MAWS,
measuresofthesatisfactionoftheneedsforautonomy(=.87),competence(=.70),andrelatedness

( = .79), total  = .84 ((Morin, 2003), as well as a measure of optimism ( = .72; (Scheier, Carver, &
Bridges, 1994; Trottier, 1999). Sixtythree percent were men with an average age of 39 years and 7.5
years of tenure at their work. Twohundred fortynine correctional officers and conditional liberation
officersworkingatamaximumsecurityCanadianprison(outof398workers)completedapapersurvey
comprisingtheMAWS,affective(=.83),normative(=.73),andcontinuance(=.89)organizational
commitment(Allen&Meyer,1996),turnoverintentions(createdforthissurvey;Itishighlyprobable
thatIwillleavethisjob,Iwillverylikelylookforanewjobthisyear,r=.62,p<.001),andworkplace
wellbeing(23items,=.97),psychologicaldistress(25items,=.98),andselfreportedphysicalhealth

(5items,=.87)fromMassetal.(1998).Sixtythreepercentweremenwithanaverageageof42.43
yearsand11.5yearsoftenureintheorganization.

Statistical Analysis Strategy




162
Weconductedconfirmatoryfactoranalysisoneachofthesamplestoeliminateitemsthathada
poorloadingontheirownfactorandcrossloadings(basedonLagrangeMultipliertests).Weeliminated
2itemspersubscaletobringtheMAWSdownto3itemsperfactortomakeitparsimonious(from20to
12items)whilemaintainingitsreliability,stabilityandinterpretability(Tabachnik&Fidell,2006).Once
12 items were selected, we proceeded to test the factorial invariance of the MAWS across the two
languages. To test the factorial invariance of a measuring instrument implies testing different models
that becomes more stringent each time (Byrne, 2006; Cheung & Rensvold, 1999, 2002; Steenkamp &
Baumgartner, 1998; Vandenberg & Lance, 2000). We tested whether the item loadings, and then
whetherthefactorstructure(i.e.,correlationsbetweenlatentfactors)wereinvariantacrosslanguages.
Wedidnottestwhetherthelatentfactormeanswereinvariantacrosslanguagesbecauseweconsiderit
normalformeanstovarywithinandacrossgroups,asvariationwoulddependonjobtype,managerial
styles,andothersuchfactors.

To assess the fit of the models, goodnessoffit indices were used in combination with the
SatorraBentler 2 statistic. We used the comparative t index (CFI) (Bentler, 1990), the root mean
square error of approximation (RMSEA) (Steiger, 1990), and the RMSEA 90% confidence interval
(Cheung & Rensvold, 2002). Values between 0.90 and 0.94 for the CFI indicate adequate t, whereas
values of 0.95 and higher indicate excellent t. Values smaller than 0.10 for the RMSEA indicate
acceptablet,valuessmallerthan0.08indicatesgoodfit,andvalueslowerthan0.05indicateexcellent
t.RMSEA90%confidenceinterval(CI)wasalsousedtoassesshypothesesofverycloset(RMSEA<
0.05)andnot(RMSEA>0.10;(McCallum,Browne,&Sugawara,1996).

WethenexaminedthevalidityoftheMAWSbyexaminingcorrelationsbetweenthemotivation
subscales with antecedents and outcomes. We also tested for differences in motivation orientation
between groups of workers that would be expected to differ based on the type of work they do and
theirhierarchicallevel.

Results


FactorialStructure

ThefactorialstructureoftheMAWSwasassessedthroughCFA.Foreachofthetwolanguages,
an initial model with four factors was postulated. These factors correspond to the four subscales and
weremadeupofthethreecorrespondingitemsforeachsubscale.Nocrossloadingswerehypothesized

163
and the covariance between the intrinsic motivation latent factor and the external regulation latent
factorwasnotincludedintheanalysessinceSDTpostulatesthroughthesimplexlikepatternthattheir
correlationshouldbeclosetozeroornonsignificant.Thefitofthisinitialmodelwaswithinacceptable
range,SB2(49)=450.43,p<.001,CFI=.920,RMSEA=.086and90%CI.079,.093inEnglishandSB2
(49)=166.56,p<.001,CFI=.963,RMSEA=.067and90%CI.056,.079inFrench(seeTable1).Wechose
thismodeltobethebaselinemodelforinvarianceanalyses(Byrne,2006).

Table1

FitStatisticsfortheMotivationatWorkScaleforeachLanguageSeparately,Languages
combined,andforInvarianceTests

Model RMSEA Model


N S-B2 df CFI RMSEA S-B2 df CFI
90% CI comparison

English only 1115 450.43 49 .920 .086 .079, .093

French only 529 166.56 49 .963 .067 .056, .079

Baseline model Two languages


1644 803.92 98 .910 .094 .088, .100
combined

Measurement model invariant Baseline vs.


1644 840.74 106 .906 .092 .086, .098 36.82 8 -.004
(factor loadings) Measurement

Structural model invariant (factor Measurement


1644 1032.35 111 .882 .101 .095, .106 191.61 5 -.024
covariances and factor loadings) vs. Structural

Structural model invariant (all


Structural vs.
except covariance between intrinsic 1644 916.35 110 .897 .094 .089, .100 -115.99 1 .015
Structural - 1
and identified regulation)

164
Invariance Analyses


When using data from both languages to test the baseline model, the model fit the data
acceptablywell,SB2(98)=803.92,p<.001,CFI=.910;RMSEA=.094;90%CI=.088,.100.Wetested
the factorial invariance of the MAWS using the procedure outlined in Byrne (2006). After testing a
baseline model, the second step is to verify if the measurement model (factor loadings) is group
invariant (i.e. languageinvariant) while the third step entails examining if the structural model
(correlationsbetweenlatentfactors)isalsolanguageinvariant(Byrne,2006).Byconstrainingthefactor
loadingstobeequalacrosstheEnglishandFrenchspeakingsamples,themodelstillfitthedatawell,S
B2 (106) = 840.74, p < .001, CFI = .906; RMSEA = .092; 90% CI = .086, .098, albeit with a negligible
deterioration6inmodelfit,SB2(8)=36.82,p<.001,CFI=.004.

Thefinalstepentailstheverificationofthestructuralmodelinvariancewhereallfactorloadings
aswellascorrelationsbetweenthelatentfactorsareconstrainedtobeequalacrosslanguages(apart
from the correlation between intrinsic motivation and external regulation, which was already
constrained at 0). This final invariant structural model had slightly lower fit indices than the previous
model,SB2(111)=1032.35,p<.001,CFI=.882;RMSEA=.101;90%CI=.095,.106,withasignificant
deteriorationinfit,=191.61,p<.001;CFI=.024.

Itisnotoftenthecasethatmodelssucceedingoingthroughallinvariancetestsandthatiswhy
Byrne(2006)suggeststestingpartialmeasurementinvariancewhereonlyspecificpartsofthemodelare
verified.Inthislogic,shesuggestsidentifyingcorrelationsbetweenlatentfactorsthatarenotinvariant
acrossgroups.Inthetwosamples,investigationofmodelmisspecificationwiththemaximumlikelihood
LagrangeMultiplier(LM)testforreleasingconstraintsrevealedthatoneconstraintdidnotbehavethe
same way in the two samples (i.e., intrinsic motivation with identified regulation). By releasing this
constraint,theoverallfitwasbetterandclosertothegenerallyrecognizedboundarywithSB2(110)=
916.35,p<.001,CFI=.897,RMSEA=.094,and90%CI.089,.100.Buttheimprovementinfitoverthe
previous fully constrained model was greater than the recommended cutoff, SB2 (1) = 115.99, p<
.001,CFI=.015.Forthisreasonwedecidedtokeepthismodelasourfinalone.

As can be seen in Figure 1 (which presents nonstandardized results, for simplicity), the
Motivation at Work Scale is best represented through four latent factors representing intrinsic
motivation, identified regulation, introjected regulation and external regulation. Standardized factor
loadingsrangedfrom.61to.95acrossthetwolanguages.ThisstructurewasinvariantacrossEnglishand
French versions. However, one correlation differed across the two samples. While the correlation
(standardized) between identified regulation and intrinsic motivation was .83 in English, it was .55 in

6
1A significant deterioration in fit would imply a CFI larger that .01 (Cheung & Rensvold, 2002).

165
French.Bothofthesecorrelationsarestillconsideredmoderatetohigh,andconcurwiththesimplex
likepatternpostulatedinthetheory.

Reliabilities and Simplex Pattern




UsingSPSS,overallCronbachalphaswerecalculatedforeachsubscaleusingtheentiresample
(EnglishandFrench)andvalueswere.75(External),.77(Introjected),.84(Identified)and.91(Intrinsic)
respectively.StandardizedcorrelationsbetweenthelatentfactorsinEnglishwere.14betweenexternal
andintrojectedregulation,.62betweenintrojectedandidentifiedregulation,.55betweenidentification
andintrinsic motivation,.21between externaland identifiedregulation,and .37betweenintrojection
andintrinsic motivation.CorrelationsinFrenchwere.12betweenexternalandintrojectedregulation,
.37 between introjected and identified regulation, .83 between identification and intrinsic motivation,
.15betweenexternalandidentifiedregulation,and.21betweenintrojectionandintrinsicmotivation.

Convergent and Discriminant Validity




Correlations with hypothesized antecedents and outcomes are presented in Table 2.
Autonomous types of motivation were related more strongly with reports of the satisfaction of the
needs for autonomy, competence and relatedness than controlled types of motivation. Autonomous
motivation was also more strongly related to perceived organizational support and optimism than
controlled motivation. Autonomous motivation was also more strongly and positively related to job
satisfaction,wellbeing,andselfreportedhealththancontrolledmotivation.Itwasalsomorestrongly
andnegativelyrelatedtoturnoverintentionsandpsychologicaldistress.InlinewithMeyer,Beckerand
Vandenberghe (2004) affective commitment was positively related to autonomous motivation, and
continuancecommitmentwaspositivelyrelatedtocontrolledmotivation.Normativecommitmentwas
positivelyrelatedtobothautonomousmotivationandintrojectedregulation.

166
Figure 1

Unstandardized Parameter Estimates from the Final Invariance Analysis

1 Intrins 1

1.021
Intrinsic Intrins 2
1.040

Intrins 3

1.205 / .362

1 Ident 1

1.087
.302
Identified Ident 2
1.092

Ident 3
.461

1 Intro 1

1.064
Introjected Intro 2
.156 .933

Intro 3
.129

Ext 1
1

.959
Extrinsic Ext 2
1.019

Note. Only one coefficient is reported when paths are constrained


Ext 3
to be equal across languages. For the correlation between latent
factor F1 (Intrinsic) and latent factor F2 (Identified), the English
coefficient to the left and the French coefficient to the right.

Mean Differences on the MAWS




We compared the 285 French workers from the 5 different sectors of employment on mean
subscale scores on the MAWS, similar to what Blais et al. (1993) did in their validation study. Based on
the assumption that some types of work may have lower propensity to satisfy basic psychological needs
(i.e., some jobs typically involve lower levels of autonomy or decision making power, are likely to
involve comparatively poorer relationships with managers and peers (e.g., competitive atmosphere in
sales) and require lower levels of education), we hypothesized that some types of jobs will breed lower
levels of autonomous motivation, and higher levels of controlled motivation. We dummy coded different

167
job types in the following ascending order of their hypothesized potential to foster autonomous
motivation: (1) sales, (2) secretarial work, (3) education, (4) health, and (5) management and professional
work. ANOVA tests were conducted on each subscale of the MAWS (see Table 3 for means). The
ANOVA was not significant for external regulation. It was marginally significant for introjection, F (4,
201) = 2.26, p < .10. Contrary to our hypothesis, Tukey tests revealed that health employees scored lower
on introjection than managers and professionals. The ANOVA was significant for identification, F (4,
205) = 13.67, p < .001. Supporting our hypothesis, Tukey tests revealed that sales workers scored lower
on identification than all other workers, and that secretarial workers scored lower than managers and
professionals. The ANOVA on intrinsic motivation was also significant, F (4, 208) = 5.09, p < .001.
Supporting our hypothesis, Tukey tests revealed that sales and secretarial workers scored lower on
intrinsic motivation than education workers, managers and professionals.


Table2

RelationswithHypothesizedAntecedentsandConsequencesofFormsofMotivation(Correlations)

External Introjected Identified Intrinsic


regulation regulation regulation m otivation
Need for autonom y .17** .36*** .60*** .55***
Need for com petence .09 .08 .27*** .25***
Need for relatedness .01 .07 .28*** .35***
Need satisfaction total (285
.12 .23*** .52*** .51***
varied workers)
Perceived organizational support
.04 .12** .21*** .20***
(881 pilots)
Optim ism (285 workers) .07 .06 .35*** .35***
Job satisfaction
(881 pilots) .13*** .27*** .53*** .58***
(130 students) .14 .42*** .60*** .60***
Turnover intentions (249 COs) -.03 -.12 -.27*** -.26***
Affective OC (249 COs) -.18** .38*** .64*** .59***
Norm ative OC (249 COs) -.24*** .36*** .47*** .41***
Continuance OC 249 COs) .26*** .15* -.04 -.11
W ell-being (249 COs) -.09 .14* .43*** .54***
Psychological distress (249 COs) .20** -.06 -.34*** -.48***
Self-reported physical health (249
.06 .10 .19** .26***
COs)
Note: CO = correctional officer, OC = organizational commitment.

* p < .05, ** p < .01, *** p < .001




168
Totestwhetherneedsatisfactioncouldexplainmotivationdifferencesacrossthesejobsectors,
wecodedtheemploymentsectors(sales=1,secretarial=2,education=3,health=4,management=5)
andenteredthisnewvariableinthefirststepofaregressiononeachtypeofmotivation.Weenteredan
aggregate of the three needs in a second step. This new job variable, although it did not account for
much variance in need satisfaction, was still a significant predictor,  = .17, p < .01, R2 = .03. The job
variablesignificantlypredictedintrinsicmotivation,=.19,p<.01,R2=.03,andthisrelationwentdown
slightly when adding need satisfaction as a mediator,  = .10, p < .10, and need satisfaction was also
significant,=.48,p<.001,R2=.22,Sobel=2.85,p<.01.Thejobvariablealsosignificantlypredicted
identifiedregulation,=.39,p<.001,R2=.15,andthisrelationwentdownslightlywhenaddingneed
satisfactionasamediator,=.31,p<.001,andneedsatisfactionwassignificant=.44,p<.001,R2=
.19,Sobel=2.84,p<.01.Thejobvariabledidnotpredicteitherintrojectedorexternalregulation.Thus
wefoundevidenceofmediationofneedsatisfactiontoaccountforrelationsbetweenjobsectorsand
autonomousmotivation.

Table3


MeanDifferencesbetweenWorkSectors

External Introjected Identified Intrinsic

regulation regulation regulation motivation

Sales (n = 74) 2.53a 1.92a 2.17a 2.83a

Secretarial (n = 66) 2.61a 2.10ab 2.68b 2.75a

Education (n = 36) 2.26a 2.09ab 2.95bc 3.31b

Health (n = 23) 2.72a 1.79ab 3.04bc 2.94ab

Management/professional (n = 2.69a 2.48b 3.29c 3.45b

14)

Note: Different superscripts within a column denote a significant difference in means.

Discussion


WedevelopedtheMotivationatWorkScaleandtesteditsfactorialstructureandvalidityintwo
languages, French and English. Confirmatory factor analysis showed a sound factorial structure of the

169
MAWS, similar to the one found with other scales using the selfdetermination theory continuum.
FurtherinvariancetestsshowedthattheMAWSitemsareinterpretedinverysimilarwaysinFrenchand
English.

The MAWS was related in expected directions with other constructs that are relevant to the
domainoforganizational behavior.Autonomousmotivationwasrelatedto hypothesized antecedents,
suchasthesatisfactionofthepsychologicalneedsforcompetence,relatednessandautonomyaswellas
with perceived organizational support and optimism. Controlled motivation was unrelated to these
antecedents.Autonomousmotivationwasalsopositivelyrelatedtohypothesizedoutcomes,suchasjob
satisfaction, well being, and affective commitment, and negatively related to turnover intentions and
psychologicaldistress.Althoughwehadexpectedcontrolledmotivationtobenegativelyrelatedtothe
outcomes,wefoundinsteadthatitwasunrelatedtothem.Theonlyoutcomethatcontrolledmotivation
wasrelatedto,asexpectedinMeyer,Becker,andVandenberghe(2004),wascontinuancecommitment.
This supports SDTs assertion that autonomous motivation yields more positive outcomes than
controlledmotivationandconcurswithanumberofstudiesthathaveshowntheadvantageofusinga
more differentiated assessment of motivation, as different forms of motivation yield different
consequences(Deci&Ryan,2000;Koestner&Losier,2002).

Meandifferencesbetweenscoresofworkersemployedindifferentsectorsdemonstratedhow
we can expect work motivation to vary as a function of how much work makes employees feel
autonomous,competentandrelatedtoothers.Regressionanalysesconcurredwiththepremisesofthe
theory, and also with the premises of the Job Characteristics Model ((Hackman & Oldham, 1975),
whereby jobs with different characteristics afford the satisfaction of basic psychological needs
differently, thereby influencing the type of motivation people will adopt at work. Overall, what the
present research shows is that we get much better outcomes with autonomous than with controlled
motivation. Controlled motivation is not necessarily bad, it just does not have much of an effect on
outcomesthatarevaluedbyorganizations.Westillneedresearchtoextendourfindingthatpromoting
autonomous motivation is the path to successful organizational outcomes. For example, Kacmar,
Andrews, Van Rooy, Steilberg, and Cerrone (2006) recently showed that turnover rates can influence
sales and profits. If autonomous motivation influences turnover intentions, it would provide a useful
mechanismtoinfluenceactualturnoverandorganizationalsuccess.

There are some limitations to the present research. First, all data were collected cross
sectionallyusingselfreports,whichcouldleadtocommonmethodvarianceissues.However,bylooking
atTable2,wecanseethatnotallcorrelationsaresignificant,whichindicatesthattherelationshipswe
found are less likely to be spurious (Spector, 2006). Further validation work should test the MAWS in
longitudinaldesignsandwithmultiplereports(e.g.,managersandcolleagues)ormorebehavioraland
objective measures (e.g., performance appraisals). We also used convenience samples of Canadian

170
workers,sofurtherworkwillneedtovalidatetheMAWSinotherculturesandlanguagesaswellaswith
other types of jobs and organizations. We did not test the social desirability of the MAWS, but other
similarscales,suchastheBlaisetal.(1993)scaleonwhichthepresentscaleispartlybased,foundvery
lowrelationshipsbetweenthemotivationsubscalesandtheMarloweCrownescale.

As proposed in Gagn and Deci (2005), work motivation is under the influence of both
dispositional and situational factors. Dispositional factors can include personality traits such as
optimism, as well as deeply ingrained causality orientations (Deci & Ryan, 1985b) that may influence
peoplesreactionstoworkrelevanteventsandcircumstances.Situationalfactorsmayincludetheway
theworkisdivided,organized,anddesigned,aswellasthequalityofrelationshipswithsuperiors,peers,
subordinates and clients. Reward and recognition systems could also influence work motivation. The
MAWS can serve as a useful tool to conduct research that examines how different types of work
motivationmaybeinfluencedbythesefactors.TheMAWScanalsobeusedtostudydifferentoutcomes
associated with different types of motivation. The different subscales of the MAWS can be used
separatelytoexaminetheirdiscreteeffects(Koestneretal),ortheycanbeaggregatedintoautonomous
and controlled types to simplify analyzes. These aggregates can also serve to test possible interaction
effects.Weadvisethistechniqueoverusingtheselfdeterminationindex(Ryan&Connell,1989),which
consistsofsubtractingcontrolledmotivationfromautonomousmotivation.Theuseofdifferencescores
hasbeenheavilycriticized(Zuckerman,Gagn,Nafshi,Knee,&Kieffer,2002)formaskingtheeffectsof
theirrespectivevariables.WehopetheMAWSwillhelptheproliferationoforganizationalresearchthat
usestheselfdeterminationtheoryframework,whichhasyieldedveryusefulresultsinotherfields.

Appendix


TheMotivationatWorkScale(MAWS).

Intrins1 BecauseIenjoythisworkverymuch.

Intrins2 BecauseIhavefundoingmyjob.

Intrins3 Forthemomentsofpleasurethatthisjobbringsme

Ident1 Ichosethisjobbecauseitallowsmetoreachmylifegoals.

Ident2 Becausethisjobfulfillsmycareerplans.

Ident3 Becausethisjobfitsmypersonalvalues.

171
Intro1 BecauseIhavetobethebestinmyjob,Ihavetobeawinner.

Intro2 BecausemyworkismylifeandIdontwanttofail.

Intro3 Becausemyreputationdependsonit.

Ext1 Becausethisjobaffordsmeacertainstandardofliving.

Ext2 Becauseitallowsmetomakealotofmoney.

Ext3 Idothisjobforthepaycheck.

Note.ThestemisUsingthescalebelow,pleaseindicateforeachofthefollowingstatementsto
whatdegreetheypresentlycorrespondtooneofthereasonsforwhichyouaredoingthisspecificjob
andisaccompaniedbythescale:1=notatall;2=verylittle;3=alittle;4=moderately;5=strongly;6=
verystrongly;7=exactly.TheFrenchversionofthescaleisavailableuponrequestfromthefirstauthor.

References


Allen, N. J., & Meyer, J. P. (1996). Affective, continuance, and normative commitment to the
organization:Anexaminationofconstructvalidity.JournalofVocationalBehavior,49,252276.

Amabile, T. M., DeJong, W., & Lepper, M. R. (1976). Effects of externally imposed deadlines on
subsequentintrinsicmotivation.JournalofPersonalityandSocialPsychology,34(9298).

Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 117, 497 
529.

Blais, M. R., Brire, N. M., Lachance, L., Riddle, A. S., & Vallerand, R. J. (1993). L'inventaire des
motivations au travail de blais. [blais' work motivation inventory]. Revue Qubcoise de
Psychologie.,14,185215.

Blais, M. R., Brire, N. M., Lachance, L., Riddle, A. S., & Vallerand, R. J. (1993). L'inventaire des
motivations au travail de blais. [blais's work motivation inventory]. Revue Qubcoise de
Psychologie,14,185215.

Byrne, B. M. (2006). Structural equation modeling with eqs: Basic concepts, applications, and
programming.Mahwah,NJ:LawrenceErlbaum.

Cheung, G. W., & Rensvold, R. B. (1999). Testing factorial invariance across groups: A
reconceptualizationandproposednewmodel.JournalofManagement,25,127.

172
Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodnessoffit indexes for testing measurement
invariance.StructuralEquationModeling,9,233255.

Deci,E.L.,Gagn,M.,Ryan,R.M.,Leone,D.L.,Usunov,J.,&Kornazheva,B.P.(2001).Needsatisfaction,
motivationandwellbeingintheworkorganizationsofaformereasternbloccountry:Across
culturalstudyofselfdetermination.PersonalityandSocialPsychologyBulletin,27(8),930942.

Deci, E. L., Koestner, R., & Ryan, R. M. (1999). A metaanalytic review of experiments examining the
effectsofextrinsicrewardsonintrinsicmotivation.PsychologicalBulletin(125),627668.

Deci,E.L.,&Ryan,R.M.(1985a).Intrinsicmotivationandselfdeterminationinhumanbehavior.New
York:Plenum.

Deci, E. L., & Ryan, R. M. (1985b). The general causality orientations scale: Selfdetermination in
personality.JournalofResearchinPersonality,19,109134.

Deci, E. L., & Ryan, R. M. (2000). The 'what' and 'why' of goal pursuits: Human needs and the self
determinationofbehavior.PsychologicalInquiry,11,227268.

Eisenberger, R., Huntington, R., Hutchison, S., & Sowa, D. (1986). Perceived organizational support.
JournalofAppliedPsychology,71,500507.

Fernet,C.,Sencal,C.,Guay,F.,Marsh,H.,&Dowson,M.(inpress).Theworktasksmotivationscalefor
teachers(wtmst).JournalofCareerAssessment.

Gagn,M.,Brub,N.,Donia,M.,Houlfort,N.,&Koestner,R.(2007).Validationofthesatisfactionwith
workscale.PaperpresentedattheAdministrativeSciencesAssociationofCanada,Ottawa,Ont.

Gagn, M., Boies, K., Koestner, R., & Martens, M. (2004). How work motivation is related to
organizational commitment: A series of organizational studies.Unpublished manuscript,
Montreal,ConcordiaUniversity.

Gagn,M.,Boies,K.,Koestner,R.,&Martens,M.L.(2005).Whyiworkdeterminesmyattachmentto
my employer: Using selfdetermination theory to understand the relationship between work
motivation and organizational commitment., Paper presented at the Annual Meeting of the
AdministrativeSciencesAssociationofCanada.Toronto,Canada.

Gagn, M., & Deci, E. L. (2005). Selfdetermination theory and work motivation. Journal of
OrganizationalBehavior,26,331362.

Gagn, M., Sncal, C., & Koestner, R. (1997). Proximal job characteristics, feelings of empowerment,
and intrinsic motivation: A multidimensional model. Journal of Applied Social Psychology, 27,
12221240.

Grolnick,W.S.,&Ryan,R.M.(1987).Autonomyinchildren'slearning:Anexperimentalandindividual
differencesinvestigation.JournalofPersonalityandSocialPsychology,52,890898.

173
Guay,F.,Vallerand,R.J.,&Blanchard,C.(2000).Ontheassessmentofsituationalintrinsicandextrinsic
motivation:Thesituationalmotivationscale(sims).MotivationandEmotion,24(3),175213.

Guttman, L. (1954). A new approach to factor analysis: The radex. In P. Lazarfeld (Ed.), Mathematical
thinkinginthesocialsciences(pp.258348).NewYork,NY:FreePressofGlencoe.

Hackman,J.R.,&Oldham,G.R.(1975).Developmentofthejobdiagnosticssurvey.JournalofApplied
Psychology,60,159170.

Kacmar,C.M.,Andrews,M.C.,VanRooy,D.L.,Steilberg,R.C.,&Cerrone,S.(2006).Sureeveryonecan
be replaced.But at what cost? Turnover as a predictor of unitlevel performance. Academy of
ManagementJournal,49,133144.

Koestner,R.,&Losier,G.F.(2002).Distinguishingthreewaysofbeinginternallymotivated:Acloserlook
at introjection, identification, and intrinsic motivation. In E. L. Deci & R. M. Ryan (Eds.),
Handbookofselfdeterminationresearch(pp.101121).Rochester,NY:UniversityofRochester
Press.

Koestner, R., Ryan, R. M., Bernieri, F., & Holt, K. (1984). Setting limits on children's behavior: The
differential effects of controlling vs informational styles on intrinsic motivation and creativity.
JournalofPersonality,52,231248.

Lepper, M. R., & Greene, D. (1975). Turning play into work: Effects of adult surveillance and extrinsic
rewardsonchildren'sintrinsicmotivation.JournalofPersonalityandSocialPsychology,31,479
486.

Mass, R., Poulin, C., Dassa, C., Lambert, J., Blair, S., & Battaglini, A. (1998). The structure of mental
health: Higher order confirmatory factor analyses of psychological distress and wellbeing
measures.SocialIndicatorsResearch,45,475504.

McCallum,R.C.,Browne,M.W.,&Sugawara,H.M.(1996).Poweranalysisanddeterminationofsample
sizeforcovariancestructuralmodelling.PsychologicalMethods,1,130149.

Meyer, J. P., Becker, T. E., & Vandenberghe, C. (2004). Employee commitment and motivation: A
conceptualanalysisandintegrativemodel.JournalofAppliedPsychology,89,9911007.

Morin,E.M.(2003).Stimulerlasantparletravail:Estcepossible?InR.Foucher,A.Savoie&L.Brunet
(Eds.),Concilierperformanceorganisationnelleetsantpsychologiqueautravail.Montreal,QC:
EditionsNouvellesAMS.

Pelletier,L.G.,Fortier,M.S.,Vallerand,R.J.,Tuson,K.M.,Brire,N.M.,&Blais,M.R.(1995).Towarda
newmeasureofintrinsicmotivation,extrinsicmotivation,andamotivationinsports:Thesport
motivationscale(sms).JournalofSport&ExercisePsychology,17,3553.

Ryan,R.M.,&Connell,J.P.(1989).Perceivedlocusofcausalityandinternalization:Examiningreasons
foractingintwodomains.JournalofPersonalityandSocialPsychology,57,749761.

174
Scheier, M. F., Carver, C. S., & Bridges, M. W. (1994). Distinguishing optimism from neuroticism (and
traitanxiety,selfmastery,andselfesteem):Areevaluationofthelifeorientationtest.Journal
ofPersonalityandSocialPsychology,67,10631078.

Spector,P.E.(2006).Methodvarianceinorganizationalresearch.OrganizationalResearchMethods,9,
221232.

Steenkamp,J.B.E.M.,&Baumgartner,H.(1998).Assessingmeasurementinvarianceincrossnational
consumerresearch.JournalofConsumerResearch,25,7890.

Steiger, J. H. (1990). Structural model evaluation and modification: An interval estimation approach.
MultivariateBehavioralResearch,25,173180.

Tabachnik,B.G.,&Fidell,L.S.(2006).Usingmultivariatestatistics(6thed.).Boston,MA:Allyn&Bacon.

Trottier,C.(1999).quivalenceetvalidationtransculturellesdequestionnairespsychologiquesmesurant
l'optimisme auprs d'une population canadiennefranaise. [equivalence and crosscultural
validation of psychological questionnaires to measure optimism in a french canadian
population].UnpublishedmastersthesisfromtheUniversitdeMontral.

Vallerand,R.J.(1997).Towardahierarchicalmodelofintrinsicandextrinsicmotivation.InM.P.Zanna
(Ed.), Advances in experimental social psychology (Vol. 29, pp. 271360). San Diego, CA:
AcademicPress.

Vallerand,R.J.,Blais,M.R.,Brire,N.M.,&Pelletier,L.G.(1989).Constructionetvalidationdel'chelle
demotivationenducation(eme)[constructionandvalidationoftheacademicmotivationscale
infrench(eme)].CanadianJournalofBehavioralScience,21,323349.

Vallerand, R. J., & Fortier, M. S. (1998). Measures of intrinsic and extrinsic motivation in sport and
physical activity: A review and critique. In J. Duda (Ed.), Advancements in sport and exercise
psychologymeasurement(pp.83100).Morgantown,WV:FitnessInformationTechnology.

Vallerand, R. J., Pelletier, L. G., Blais, M. R., Brire, N. M., Sncal, C., & Vallires, E. F. (1992). The
academic motivation scale: A measure of intrinsic, extrinsic, and amotivation in education.
EducationalandPsychologicalMeasurement,52,10031017.

Vandenberg, R. J., & Lance, C. E. (2000). A review and synthesis of the measurement invariance
literature: Suggestions, practices, and recommendations for organizational research.
OrganizationalResearchMethods.

Williams,G.C.,&Deci,E.L.(1996).Internalizationofbiopsychosocialvaluesbymedicalstudents:Atest
ofselfdeterminationtheory.JournalofPersonalityandSocialPsychology,70,767779.

Williams,G.C.,Grow,V.M.,Freedman,Z.R.,Ryan,R.M.,&Deci,E.L.(1996).Motivationalpredictorsof
weightlossandweightlossmaintenance.JournalofPersonalityandSocialPsychology,70,115
126.

175
Zuckerman,M.,Gagn,M.,Nafshi,I.,Knee,C.R.,&Kieffer,S.C.(2002).Testingdiscrepancyeffects:A
critique, a suggestion, and an illustration. Behavior Research Methods, Instruments, and
Computers,34,291303.

176

Anda mungkin juga menyukai