STRUCTURAL EQUATION MODELLING OF PRODUCTIVITY … · 2013-11-25 · Productivity Institute, 2006). theBy implementing such a program, followingsthe company can reduce waste • Increaseand

259Suranaree J. Sci. Technol. Vol. 17 No. 3; July - Sept 2010

STRUCTURAL EQUATION MODELLING OF PRODUCTIVITY ENHANCEMENT

Thanwadee Chinda*

Received: Jul 8, 2010; Revised: Aug 25, 2010; Accepted: Sept 1, 2010

Abstract Productivity has often been cited as a key factor in industrial performance, and actions to increase it are said to improve profitability and the wage earning capacity of employees. Improving productivity is seen as a key issue for survival and success in the long term. This paper focuses on examining key factors of productivity enhancement, and investigating the causal relationships among those key factors to better understand and plan for productivity improvement. The results prove 5 key productivity factors, including ‘leadership’, ‘strategic quality planning’, ‘people’, ‘data and information’, and ‘process management’, leading to a conceptual model. ‘Leadership’ is found to be the main driver to productivity enhancement. The strong commitment of a leader is thus crucial in improving productivity. ‘Leadership’ strongly influences ‘strategic quality planning’ and ‘process management’, and indirectly influences ‘people’ through ‘data and information’ and ‘strategic quality planning’. ‘People’ also plays a key role in the success of productivity enhancement. Additionally, there are direct and indirect relationships between ‘data and information’ and ‘people’ and between ‘data and information’ and ‘process management’, respectively.

Keywords: Exploratory factor analysis, productivity enhancement, structural equation modelling

Introduction

Productivity

Productivityisoneofthemostcommonmeasuresofanorganization’scompetitiveness.It has often been cited as a key factor inindustrialperformance,andactionstoincreaseit are said to improve profitability and thewageearningcapacityofemployees(CosmetatosandEilon,1983).Theconceptofproductivity,

generally defined as the relation betweenoutput and input, hasbeenavailable forover2 centuries and applied in many differentcircumstancesonvariouslevelsofaggregationin the economic system (Jorgenson andGriliches,1967).However,thereisnocommonagreementovertheunderstandinganddefinition

School of Management Technology, Sirindhorn International Institute of Technology, Thammasat University, 131 Tiwanont Road, Bangkadi, Muang, Pathumthani 12000, Thailand, Tel: 66-2-5013505 ext. 2111 Fax: 66-2-5013524, E-mail: [email protected] * Corresponding author

Suranaree J. Sci. Technol. 17(3)259-276

Structural Equation Modelling of Productivity Enhancement 260

of the productivity concept. Some ofproductivity’sdefinitionsareasfollows: • Productivity is the ratio of outputto input for a specific production situation.Risingproductivityimplieseithermoreoutputis produced with the same amount of inputs,or that fewer inputs are required to producethe same level of output (Rogers, 1998;RussellandTaylor,2009). • Productivity is the belief in humanprogress. It is a state of mind which aims atperpetualimprovement.Itisaceaselessefforttoapplynewtechnologyandnewmethodsforthe welfare and happiness of mankind. It isthetrainingofthemindsandthedevelopmentof attitudes of the people as a whole whichdetermines whether the nation will realizehigh productivity and an affluent life or lowproductivity and poverty (Asian ProductivityOrganization,2006). • Productivity is a measure of theamount of output per unit of input. Forexample, productivity in the auto industrymightbemeasuredbythenumberofhoursoflabourusedperautomobileproduced (HeizerandRender,2008;Investopedia,2008). • Productivityisaprocessofcontinuousimprovement in the production/supply ofquality output/service through efficient,effective use of inputs, with emphasis onteamwork for the betterment of all (TradeUnionists,2008). • Productivityisanattitudethatseeksthecontinuousimprovementofwhatexists.Itis a conviction that one can do better todaythan yesterday, and that tomorrow will bebetter than today. Furthermore, it requiresconstanteffortstoadapteconomicactivitiestoever-changingconditions,and theapplicationofnewtheoriesandnewmethods.Itisafirmbelief in theprogressofhumanity (EuropeanProductivityAgency,2008). • Productivity is the amount of work(in terms of goods produced or servicesrendered) produced during a period of time(usually per hour). Productivity analysestechnicalprocessesandengineeringrelationshipssuch as, how much of an output that canbe produced in a specified period of time

(Knowledgerush,2008). Fromtheabovedefinitions,thisresearchstudy defines productivity as “the ratio ofoutputs to inputs, and that, to improvethis ratio, continuous improvement must beperformed in the organizations”. It is arguedthatproductivity isoneof thebasicvariablesgoverning economic production activities,perhaps the most important one (Kilic andOkumus, 2005). Improving productivity isseenasakeyissueforsurvivalandsuccessinthe long term. To enhance productivity, theorganization may either consider reducinginputs while keeping outputs constant, orincreasing outputs while keeping inputsconstant (Kilic and Okumus, 2005). In aneconomicsense,inputsarelabour,capital,andmanagement, which are integrated into aproductionsystem(HeizerandRender,2008).Outputs, on the other hand, are goodsand services (Heizer and Render, 2008).Measurement of productivity is an excellentwaytoevaluateacountry’sabilitytoprovideanimprovingstandardoflivingforitspeople.Onlythroughincreasesinproductivitycanthestandardoflivingimprove.

Productivity in Thai Food Industry

Recently, productivity has receivedgrowing attention, both in the manufacturingand service industries. Herron and Braiden(2006),forexample,developedamodel,basedon a wide range of manufacturing efficiencyimprovements such as just-in-time and leanmanufacturing tools, to direct and generateproductivity improvement in a group ofmanufacturingcompanies.Kilic andOkumus(2005) investigated the factors influencingproductivityinhotelsinNorthernCyprusandfoundthatfactorssuchstaffrecruitment,stafftraining, meeting guest expectations, andservicequalityarethemainfactorsinimprovingproductivity.NPC(2003)attemptedtoestablishnetworkingandidentifybenchmarksandbestpractices throughdifferent industries, suchasthe textile industry, chemical industry, andfoodindustry,inordertoimproveproductivityand waste management. It was found thatfactors such as information technology,


innovation, and training help improveproductivity,andthusenhancethecompany’scompetitiveness. Attentionhasalsobeenpaidtoimprovingproductivity in thefood industry inThailand,as food is considered one of the importanteconomic sectors, constituting 14 percent ofthe country’s total exports, and generatingemployment for 20 million people (ThailandBoard of Investment, 2005). For example,Betagro Group, one of the biggest foodmanufacturingcompaniesinThailandadopteda number of quality management tools, suchas Six Sigma and Kaizen, in planning for aproductivity improvement program (ThailandProductivityInstitute,2006).Byimplementingsuch a program, the company can reducewaste and rework, minimize the work-in-process inventory, lessen transportationcosts,and eliminate idle time, thus increasingproductivity (Thailand Productivity Institute,2006). Theaboveillustrationsdemonstratethatproductivity is one of the most vital factorsaffecting a manufacturing company’s com-petitiveness. However, researchers argue thatproductivityisoftenregardedassecondrank,and neglected or ignored by those whoinfluence the production process (Tangen,2002).A major reason for this could be thatthere is no standard tool for assessing andmeasuringproductivity.Increasingproductivityrequires that attention be paid to using andmanipulating numerous factors, which isoften a challenging task (Poetscheke, 1995).Moreover,neithertheinteractionsamongkeyfactors influencing productivity, nor theconsequencesofproductivityinitiativesbeingundertaken is focused. This research study,therefore, aims at examining key factors ofproductivity enhancement. The StatisticalPackage for Social Sciences (SPSS) programversion15.0isusedtoensuredataconsistency,and to allow the results to be meaningfullyinterpreted.A number of data screening andpreliminaryanalyses, including thenormalitytest, the outliers test, and the reliability test,areperformed.Thescreeneddatawillthenbefurtheranalysedusingmorecomplexanalyses,

including the exploratory factor analysis andstructural equation modelling, to investigatethecausalrelationshipsamongkeyproductivityfactors,sothatthecompaniesareabletobetterunderstand as well as plan for productivityenhancement.

Productivity Attributes Productivityisconsideredasoneofthebasicvariables governing economic productionactivities (Tangen, 2002).At the same time,productivity is also seen as one of the mostvitalfactorsaffectingamanufacturingcompany’scompetitiveness.AccordingtoTangen(2002),theorganizationcaneitherconsideroneofthefollowingstoincreaseproductivitylevels: • Increase output and input, but theincrease in input is proportionally less thantheincreaseinoutput • Increaseoutputwhilekeeping inputconstant • Increaseoutputwhilereducinginput • Keepoutputconstantwhiledecreasinginput • Decrease output and input but thedecrease in input is proportionallymore thanthedecreaseinoutput Thereareanumberofattributesthatcanbe used to represent productivity. Theseattributesarecarefullyselected,withreferenceto the frequency of citations in recentmanufacturing-related, including the foodindustry,literature.Eachattributeisdescribedindetailbelow. 1. Financial incentive: Millea andFuess(2005)claimedthatmoneycanbeusedto motivate workers, which in turn, tends toincreaseproductivity. 2. Training: According to Black andLynch (1996) and Hoffman and Mehra(1999), a successful organization alwaysensures that its staff is properly trained, sothat they can carry out various activitieseffectively. 3. Work pressure:Highworkpressuresare caused by distress, unworkable scheduletimes, and workforce instability (highturnover)(Siuet al.,2004).These,inturn,lead


to decreased productivity (Black and Lynch,1996). 4. Personal recognition: Productivityimprovement requires highly committed andmotivated employees. Personal recognition,through trust support, acknowledgement,and empowerment, is the key to attainingcommitment (Allender, 1984).Workers withgood records should be recognized andrewarded(Marsidi,2009). 5. Workers’ attitude:Beck(2009)statedthatanimportantkeytohighproductivityandeffectiveleadershipistohaveandmaintainapositive attitude.A negative attitude leads toabsenteeism, high turnover, unnecessaryconflicts, and many other counter-productiveproblems(Marsidi,2009). 6. Teamwork: Very strong teamworkis a tool to improve management andorganizationalproductivity(AsianProductivityOrganization, 2007).The world-class organi-zations see the need to embrace teams andteamwork for competitive advantage andproductivityimprovement. 7. Knowledge background: Kilic andOkumus (2005) stated that the knowledgebackground of the employees affects theirworkingperformance.Employeeswithalackof knowledge show signs of problems inproductivity programs (Barocci and Wever,1982). 8. Workers’ experience: Lubbe (2000)claimedthatproductivitycanbeenhancedbyusingexperienced,professionalworkers.KilicandOkumus(2005)agreedthatworkerswithhigherexperienceperformbetterinthejob. 9. Workers’ recruitment:AsianProduc-tivity Organization (2007) stated that themanager’s role of hiring and offering aposition can affect the productivity andprofitabilityoftheentireenterprise. 10. Leader’s support: Job satisfaction,productivity and organizational commitmentare affected by leadership behaviours (Loke,2001). Lack of management support leads topoorproductivity(Savery,1998). 11. Job allocation: Bell (1988) statedthat people should be assigned to the tasksfor which they are best suited. Proper job

allocations in work schedules help anorganization reduce its total labour cost,enhance the flexibility of its operations, andimprove worker productivity (Peter et al.,2007). 12. Feedback:Pritchard(1995)concludedthat feedback plays a major role in creatingtheproductivityimprovement.Productivityinan organization, where feedback is given, ishigher than that in organizations where nofeedbackisgiven. 13. Safe workplace: According toThailand Productivity Institute (2008b) andSalesCreators(2009),goodhousekeepingandorderly plant operations create a pleasantworkingenvironment,which,inturn,leadstoincreasedproductivity. 14. Supervision: According to Bell(1988),poorsupervisionshouldbeavoidedinorder to reduce the possibility of decreasedworkers’ motivation, high turnover, andloweredproductivity. 15. Equipment effectiveness: Goodmaintenance programs help in reducing theidletime,increasingthemachineeffectiveness,and perhaps enhancing the productivity(ThailandProductivityInstitute,2008c). 16. Inventory management:Accordingto Thailand Productivity Institute (2008a),the storage cost can be reduced with goodinventory documentations (e.g. daily checksheets,storageplan,andproductsrecords). 17. Performance appraisal:Accordingto Dessler (2005) andThailand ProductivityInstitute (2008b), performance appraisalshouldbeemployedtomotivateworkers. 18. Job description:Clearjobdescriptionassists new employees in performing taskscorrectlyandeffectively(ThailandProductivityInstitute,2008b). 19. Operational audit:Batraet al.(2009)identifiedthatoperationalauditisaneffectiveway for enhancing the productivity andminimizing energy consumption. It has beenobserved and proven that production levelscanbeimprovedandenergyconsumptioncanbereducedthroughtheaudit. 20. Information technology investment:Swierczek and Shrestha (2003) noted that


productivity improvement can be achievedthrough information technology use, as itincreases outputs and decreases costs. It alsoincreases the organization’s competitivenessthrough differentiation and customer serviceimprovement, reduced costs, better riskavoidance, and maintaining the stability ofthecustomerbaseandmarketshare. 21. Service quality:An organizationstrives to achieve better service quality inorder to satisfy its customers (Haynes andDuvall,1992). 22.Customer satisfaction:Lubbe(2000)proposed that service productivity can beincreased by becoming more effective (byincreased customer satisfaction) or moreefficient(lessresourceusetodeliverservice). 23. Two-way communication:Relgolook(2009) stated that effective communicationhas become the order of the day for anyindividual or a business entity. Hoffmanand Mehra (1999) highlighted that 2-waycommunication should be encouraged in theorganization,aspoorcommunicationcanleadtoproductivityfailures. 24.Top management commitment:HoffmanandMehra(1999)statedthatlackoftopmanagementcommitmentandinvolvementis the major obstacle in successfullyimplementing a productivity improvementproject. 25. Advertisement and marketing:AccordingtoKilicandOkumus(2005),goodadvertising and sales support plan have apositiveinfluenceonproductivity. 26.Total quality management:The useoftotalqualitymanagementintheorganizationhelps in increasing productivity (ThailandProductivityInstitute,2008b). 27.Benchmarking system:According toHussain(2008),abenchmarkingsystemshouldbeencouragedinordertoimproveproductivity.

Questionnaire Survey and Data Collection

The above 27 productivity-relatedattributesareusedinthequestionnairesurveydevelopment. A written questionnaire isself-administered, and can be sent throughthe traditional mail system or by email. It is

importantthatamailsurveybeclearlywrittenand self-explanatory because no one will beavailable to answer questions regardingthe survey, once it has been mailed out.Questionnairesurveyshaveseveraladvantages,forexample,theygenerallyhavelesssamplingbias (a tendency for one group to be overre-presentedinasample)thanpersonalinterviews.Theyalsoallowtheresearchertocollectdataon more sensitive information. Participants,who may be unwilling to discuss personalinformation with someone face-to-face, maybe willing to answer such questions in awritten survey. This method is usually lessexpensiveandcancoveralargegeographicalarea. Further, the participants can take asmuch time as they need to answer thequestions without feeling the pressure ofsomeone waiting for the answers. However,themajorproblemsconcernthelowresponserate,andthemisinterpretationofthequestions(McBurney,1994). The questionnaire survey comprises 3parts. The first part is devoted to gatheringdemographical information about therespondentsandtheirrespectiveorganizations(such as position, name of the organization,years of experience, and job title) to ensurethat the respondents have the appropriatebackgrounds. It is useful in identifyingdiscrepancies in the received responses.The second part covers 27 statements tooperationallydefineproductivityenhancement.Each statement is designed to elicit therespondents’ opinions on the differentattributes in the context of productivityenhancement using a 5-point Likert scale,with point 1 representing ‘strongly disagree’andpoint5representing‘stronglyagree’.Thisapproach enables the evaluation of theorganization’sperceptionof,andcommitmenttowards, each construct to be carried out.Lastly,thethirdpartasksforsuggestions.AnexampleofaquestionnairesurveyisillustratedinAppendix1. The intention of this research was tostudy the food industry sector in Thailand.Medium to large organizations, with staff of100ormore, located inBangkokandnearby


variableisavariablewhosemeanisnotinthecentre of the distribution. Kurtosis, on theother hand, relates to the peakedness of adistribution;adistributioniseithertoopeaked(withshort,thicktails),ortooflat(withlong,thin tails). When a distribution is normal,the values of skewness and kurtosis are zero(Pallant,2005).Ifthereisapositiveskewness,there is a pileup of cases to the left, and therighttailistoolong;withnegativeskewness,the result is reversed. Kurtosis values abovezeroindicateadistributionthatistoopeaked,whilekurtosisvaluesbelowzeroarereversed.Non-normalkurtosisproducesanunderestimateof the variance of a variable.According toCurran et al. (1996), the values of skewness< ± 2 and kurtosis < ± 7 are consideredacceptable. The 355 data are performed with thenormalitytest,andtheresultsdemonstratethenormal distribution, with all the skewnessand kurtosis values in the limited ranges.This, thus increases confidence in the datacollected.

Outliers Test

Anoutlierisacasewithsuchanextremevalueononevariable(aunivariateoutlier),orsuch a strange combination of scores on twoor more variables (multivariate outlier), thatdistorts thestatistical results (TabachnickandFidell,2007).Therearemanywaystotestforoutliers, such as the use of the 5% trimmedmean,theuseofstandardizedscores(z-scores),and the use of boxplots (Pallant, 2005;Tabachnick and Fidell, 2007). In this study,the mean, the 5% trimmed mean, and thez-scoretestsareusedtodetectoutliers. The 5% trimmed mean is a meancalculatedfromthecasesinwhich5%ofthetop and thebottomof the cases are removed(Pallant, 2005).According to Pallant (2005),thebigdifference(>0.2)betweenameanandits5%trimmedmeanmayindicateaproblemwith an outlier. The results show thatthe mean differences (Δmean) of all 27 attributes are small, providing support fortheabsenceofoutliers. Tofurtherdetectoutliers,astandardized

provinces were selected for the sampling.Targeted respondents were both managementand working positions to gain diverseopinions on productivity improvement. Fourhundred and sixty eight questionnaires weredistributed,with372responsesrepresentingaresponse rate of 79.49%. From the returnedresponses, 17 questionnaires were deemedunusable, due to data incompleteness, andweresubsequentlydroppedfromthedataset.Asaresult,355usablequestionnairesprovidedatafortheanalyses. Among the respondents, 56.06% areworkers, while 43.94% are in managementpositions, including managers, head officers,and directors/owners.Apart from that, morethan50%oftherespondentshavemorethan5yearsworkingexperienceboth in thepresentorganization and in the Thai food industry.This indicates the reasonably high workexperiencerateoftherespondents.

Data Screening and Preliminary Analyses

After thedata iscollected,anumberofdata examination techniques, ranging fromthe simple process of visual inspection ofgraphical displays to statistical methods areconducted. Thus, statistical methods of thehandling of missing data, the normality test,theoutlierstest,andthereliabilitytestneededtobeperformedtoincreaseconfidenceinthedata. Each statistical method is described indetailbelow.

Test of Normality

The screening of continuous variablesfor normality is an important early step inalmosteverymultivariateanalysis(Tabachnickand Fidell, 2007).Although the normality ofthe variables is not always required foran analysis, the solution is usually moreappropriate if the variables are all normallydistributed. For this reason, the normality ofthevariablesisassessedbyeitherstatisticalorgraphicalmethods. Twoimportantcomponentsofnormalityare skewness and kurtosis (Tabachnickand Fidell, 2007). Skewness relates to thesymmetry of the distribution; a skewed


score(z-score)testisperformed.AccordingtoTabachnick and Fidell (2007), the cases withz-scores that exceed ± 3.29, at p < 0.01,two-tailed test, are the potential outliers.The results show that there are 63 z-scoresexceeding ± 3.29, in which most are fromquestionnairesurveynumber ‘98’, ‘198’,and‘248’. As a result, these three cases aredeletedfromthedatafile,leadingtoatotalof352dataretainedforfurtheranalyses.

Exploratory Factor Analysis Inthisresearchstudy,2statisticalanalysesareperformed, including the exploratory factoranalysis (EFA) and the structural equationmodelling (SEM). The exploratory factoranalysis (EFA) method is often used inthe early stages of data analysis to gatherinformation about interrelationships amonga set of variables.According to Seo et al.(2004), the EFA is a precursor to the SEM.When conducting an EFA, three mainsteps are followed: 1) the assessment of thesuitabilityofthedata;2)thefactorextraction;and3)thefactorrotation(Pallant,2005).Thedetailsofeachsteparedescribedbelow.

Assessment of the Suitability of the Data for the Analysis

Two main issues facilitate the deter-mination of whether a particular data set issuitable for factor analysis.The first issue isthesamplesize.TabachnickandFidell(2007)notedthatitiscomfortingtohaveatleast300cases for factor analysis. However, theyconcededthatasmallersamplesize(e.g.150cases) should be sufficient, if the solutionshave several high loading marker variables(above 0.80). Pallant (2005), on the otherhand, recommended that five cases for eachitem are adequate in most cases. Coakes andSteed(2003)tookastanceinbetween,arguingthat a sample of 100 cases is acceptable,with a sample of 200 or more cases beingpreferable.Atotalof(usable)352cases(withcase numbers ‘98’, ‘198’, and ‘248’ deleted)are considered acceptable for the analysis inthisstudy.

The second issue concerns the strengthof the inter-correlations among the items.Bartlett’s test of sphericity and the Kaiser-Meyer-Olkin(KMO)testarenormallyappliedtoassessthefactorabilityofthedata(Pallant,2005). Bartlett’s test of sphericity should besignificant (p < 0.05) for factor analysis tobe considered appropriate, while the KMOindexshouldrangefromzerotoone,with0.6suggested as the minimum value for a goodfactoranalysis(TabachnickandFidell,2007).In this study, Bartlett’s test of sphericity issignificant (p = 0.00), with the KMO indexbeing0.87 (higher thanaminimumof0.60),thus indicating that the data is suitable forfactoranalysis.

Factor Extraction

Factor extraction, the second step inconductingtheEFA,involvesdeterminingthesmallestnumberoffactorsthatcanbeusedtobest represent the interrelations among thesetofvariables(TabachnickandFidell,2007).There are a variety of approaches that canbe used to extract (or identify) the numberof underlying factors. Some of the mostcommonly available extraction techniquesare principal components, principal axisfactoring, and general least square (Pallant,2005). According toCoakesandSteed (2003),the most frequently used techniques areprincipal components and principal axisfactoring. Principal components analysis is,however, normally used when the researcheris interested in discovering which variablesin the set form coherent subsets that arerelatively independent of one another.Variablesthatarecorrelatedwithoneanotherbut largely independent of other subsets ofvariablesarecombinedintofactors(Tabachnickand Fidell, 2007). For this reason, theprincipal components method is chosenfortheanalysis.

Factor Rotation and Interpretation

After the extraction, factor rotation isordinarily used to present the pattern ofloadingsinamannerthatiseasytointerpret.


Numerous methods of factor rotation areavailable, but the most commonly used is‘varimax’ (Coakes and Steed, 2003; Pallant,2005). Varimax is a variance maximizingprocedure.Thegoalofvarimaxrotation is tomaximize the variance of factor loadingsby making high loadings higher, and lowloadings lower, for each factor (Tabachnickand Fidell, 2007). The varimax method isusedforthefactoranalysisinthisstudy. In summary, the principal componentmethod, together with the varimax rotationmethod,areusedtoexaminethedimensionalityofthe27attributesofproductivityenhancement,and to achieve better interpretation of thefactor loadings. Due to the sample size (352data set), a cut-off factor loading of 0.40 isusedtoscreenouttheattributesthatareweakindicators of the constructs, as suggestedby Hair et al. (1998).The first run leads todroppingofthe‘feedback’attributeasitfailsto make the cut-off loading. Consequently,performing the next factor analysis extractsthe remaining 26 attributes into 5 factors,accounting for 52.04% of the total variance,asshowninTable1. Factor 1 is accounted for by 10 items,measuring mainly process management. Forexample, Thailand Productivity Institute(2008b) mentioned that good housekeepingandorderlyplantoperationscreateapleasantworking environment. They also suggestedthat the storage costs can be reduced withgood inventory documentations (e.g. dailycheck sheets, storage plan, and productsrecords). Pupat (2009), on the other hand,statedthataclearjobdescription,withaclearoutlineofdutiesandresponsibilities,helpsinreducing the reworkprocess, andmaking thescreening process as direct and focused aspossible. Peter et al. (2007) also concludedthatproper job allocations inwork scheduleshelp an organization reduce its total labourcost, enhance the flexibilityof itsoperations,andimproveworkerproductivity.Asaresult,

Factor 1 is called the ‘process management’ (PRO) factor. Factor 2 is explained by 5items, initially measuring leadership. Blazey(2008), for example, explained that leadersshould demonstrate a strong commitment toimproving productivity by focusing more oncustomer satisfaction. This can be achievedthrough, for instance, leaders developing aneffective system to understand customerrequirements, strengthening loyalty andcustomerrelationships,andresolvingcustomerproblemsimmediately.Thisfactoris,therefore,namedthe ‘leadership’ (LEA)factor. Factor 3 is accounted for by 3 items,measuringstrategicqualityplanning;asstatedin Thailand Productivity Institute (2008b)and Hussain (2008) an effective quality planand benchmarking system help increaseproductivity.Thisfactoris,therefore,labelledthe ‘strategic quality planning’ (STR) factor.Factor 4 is explained by 5 items, initiallymeasuring people, and, therefore, called the‘people’ (PEO) factor. This is confirmedby Allender (1984) who mentioned thatproductivity improvement requires highlycommitted and motivated employees.Trust,support, acknowledgement, andempowermentare thekeys toworkers’commitment.Lastly,Factor5isrepresentedby3items,measuringmainly workers’ data and information, andis hence called the ‘data and information’(DAT)factor. To increase confidence in the constructvalidity, the5 factors extracted are tested forinternal consistency (Cronbach’s alpha, α). Sharma (1996) stated that the alpha valuesranging from 0.6 to 0.7 are regarded assufficient, and the value higher than 0.7 isregarded as good in reliability testing. Theresults, shown inTable 2, have alpha valuesranging from 0.60 to 0.85, all of which areconsidered acceptable. This thus increasesconfidence in the contribution of the 26attributes to the measurement of theirrespectiveconstructs.


Table 1. Five factors extracted from the 26 attributes

Attribute no. Attribute Factor extracted

1 2 3 4 5

18 Jobdescription 0.66 21 Servicequality 0.65 10 Leader’ssupport 0.64 20 Informationtechnologyinvestment 0.61 13 Safeworkplace 0.60 11 Joballocation 0.59 23 Two-waycommunication 0.59 16 Inventorydocumentations 0.59 15 Equipmenteffectiveness 0.55 6 Teamwork 0.45 24 Topmanagementcommitment 0.67 17 Performanceappraisal 0.63 22 Customersatisfaction 0.59 14 Supervisor 0.57 19 Operationalaudit 0.49 27 Benchmarkingsystem 0.79 26 Totalqualitymanagement 0.79 25 Advertisementandmarketing 0.58 2 Training 0.71 4 Personalrecognition 0.60 1 Financialincentive 0.54 3 Workpressure 0.53 5 Workers’attitude 0.52 8 Workers’experience 0.76

7 Workers’recruitment 0.64

9 Knowledgebackground 0.61

Extraction method: principal component analysis. Rotation method: varimax with Kaiser Normalization. Rotation converged in 14 iterations.


Productivity Enhancement Model Develop-ment

The EFA gives rise to a total of 26attributes grouped to explain5 latent factors,leading to the so-called conceptual modelof productivity enhancement, as shown inFigure1.Theovalsand rectangles, shown inthe baseline model, symbolise latent andobserved variables, respectively.The formerrepresent the 6 constructs of the baselinemodel, whereas the latter represent theirrespective attributes.The arrows connectingthe2setsofvariablesindicatethedirectionofthehypothesizedinfluence.Forexample,itishypothesized that ‘process management’is manifested by the achievement of its10attributes:Questions18,21,10,20,13,11,23,16,15,and6; thearrowsare thus shownoriginating from ‘process management’ toeachoneofthe10attributes. By investigating the conceptual model,it is found that its constructs match with the5 major categories of the Malcolm BaldrigeNationalQualityAward(MBNQA)Framework.TheMBNQAFrameworkisoneofthewidelyused quality models, and is used to promotequality awareness and thus improve thecompetitivenessofthecompanies(PannirselvamandFerguson,2001).Theframeworkconsistsof 7 categories (leadership, information andanalysis, strategic quality planning, humanresources development and management,management of process quality, quality andoperational results, and customer focus andsatisfaction) grouped into 4 basic elements

Table 2. Internal consistency of the five productivity factors

Factor Cronbach’s Alpha (α)

Processmanagement 0.85

Leadership 0.69

Strategicqualityplanning 0.81

People 0.67

Dataandinformation 0.60

Note:PRO=processesmanagement,LEA= leadership,STR = strategic quality planning, PEO = people,DAT=dataandinformation Figure 1. Conceptual model of productivity enhancement


(driver, system, measures of progress, andgoals). Lee and Quazi (2001), however,grouped the ‘measurement of progress’ and‘goals’ elements into 1 single group called‘results’, as it is expected that the betterimplement of the ‘driver’ and ‘system’elements helps achieve better ‘results’ (seeFigure2). The 5 productivity factors achievedfrom the EFA can be matched with the 5categories of the MBNQA Frameworkwithin the ‘driver’ and ‘system’ elementsi.e. the ‘leadership’ factor matches withthe ‘leadership’ category, the ‘data andinformation’ factor matches with the‘information and analysis’ category, the‘strategic quality planning’ factor matcheswith the ‘strategic planning’ category, the‘people’ factor matches with the ‘humanresources development and management’category,andthe‘processmanagement’factormatches with the ‘process management’category.Asstatedearlier,theimplementationof the ‘driver’ and ‘system’ elements helpsachievebetter‘results’,thus,thefocusofthisstudy is on the improvement of the 5 keyfactorsinordertoachievehigherproductivity.

Structural Equation Modelling Theoretically, SEM comprises 2 types of

models: measurement and structural models.The former is concerned with how well theobservedvariablesmeasurethelatentfactors,addressing their reliability and validity.The latter is concerned with modelling therelationships between the latent factors, bydescribing the amount of explained andunexplained variance, which is akin to thesystem of simultaneous regression models(WongandCheung,2005).

Measurement Model

Testingthestructuralmodelwouldhavebeen meaningless until it was established asa good measurement model. In this study,a confirmatory factor analysis is undertakento establish confidence in the measurementmodel. According to Byrne (2001), themeasurement model should be assessed by4 methods, including the feasibility ofparameter estimates, the appropriateness ofstandard errors, the statistical significance ofparameter estimates, and the model fit asawhole(usinggoodnessoffit,GOF,indices). The most widely used GOF indicesare he chi square to the degrees of freedom(χ2DF), the rootmeansquareerrorofapproximation (RMSEA), the comparative fit index(CFI),theincrementalfit index(IFI),andtheTucker-Lewis index (TLI) (Byrne, 2001;Garson, 2006;Tabachnick and Fidell, 2007).

Figure 2. Seven categories of the MBNQA Framework (Lee and Quazi, 2001)


Note:PRO=processesmanagement,LEA= leadership,STR = strategic quality planning, PEO = people,DAT=dataandinformation Figure 3. The best-fit measurement model

Normally, the value of χ2/DF less than 2represents the model as a good fit; however,thevalueoflessthan3isacceptable(Garson,2006). The value of RMSEA up to 0.05indicates a good model fit, with a value upto 0.10 indicating a reasonable error ofapproximation(TabachnickandFidell,2007).Kline(2005)suggestedthat theCFI,IFI,andTLIvaluesshouldbeat least0.90toindicateamodelfit.Garson(2006),however,usedthecut-offvalueof0.80asthecriterion. The basic productivity model is testedwithSEM,andtheGOFindicesrevealaneedto modify the model in order to improve themodel fit (seeTable 3).According to Kline(2005),therearepotentially2possibleoptionsinvolved in theprocess ofmodel refinement.The first option is to eliminate the links or‘paths’withverylowcorrelations.Thisoptionisfoundnottobeapplicabletotheconceptualmodelduetothehighcorrelationsofallpaths.Thesecondoption is to remove theobservedvariablesshownbythecomputedmodificationindices as having multicollinearity (e.g. theobserved variables/their error variancesthatcorrelated tomore than2otherobservedvariables/error variances).As a result, the‘financial incentive’ attribute shows signs ofhigh multicollinearity, and thus is removed,leading to the best-fit measurement modelwith the best GOF indices, as shown inTable 3 and Figure 3, respectively. Table 3highlights the significant improvement ofthe GOF indices’ values compared to thoseobtainedfortheconceptualmodel.

Table 3. The GOF indices of the basic and the best-fit measurement models

GOF index Recommended level Conceptual model Best-Fit measurement

Model

χ2/DF <2.00(Byrne,2001) 2.60 1.98

RMSEA ≤ 0.05 (Tabachnick and Fidell, 2007) 0.07 0.05

CFI >0.90(Kline,2005) 0.83 0.91

IFI >0.90(Garson,2006) 0.84 0.91

TLI >0.90(Kline,2005) 0.81 0.90 The Conceptual model is shown in Figure 1, while the best-fit measurement model is shown in Figure 3.


Structural Model

Having established confidence in themeasurement model, a structural equationmodelisdevelopedandtestedtoexaminethedirection of the relationships between the5constructs,andtoimprovetheoverallmodelfit.Wilson and Collier (2000) introduced thecasualrelationshipsof the5constructsof theMBNQA Framework, as shown in Figure 4.Anumberofmodelrunsare,then,carriedoutin this study to verify these relationships.Anylinkswithverylowcorrelations,oritemsshowing signs of multicollinearity, aredeleted. For each run, the GOF indices arecomputed and compared. According toClissold (2004), the model with the best fitwouldprovethedirectionalinfluences. The SEM results show that the causallink between ‘data and information’ and‘strategic quality planning’ has a negativecovariance.Moreover, the causal relationshipbetween ‘data and information’ and ‘processmanagement’ is very weak. These 2 linksare, therefore, removed, leading to the fittedstructuralmodel(showninFigure5),withthebestGOFindices(listedinTable4). The final productivity model is, then,achieved(Figure6).Itisclearthat‘leadership’isthemaindrivertoproductivityenhancement,and the strong commitment of the leader iscrucialinimprovingproductivity. As previously mentioned, the value ofSEM lies in its ability to depict both direct

and indirect effects between the variables.In light of this, the final productivity modelappears to indicate that ‘leadership’ stronglyinfluences ‘strategic quality planning’ (pathcoefficient=0.64)and‘processmanagement’(pathcoefficient=0.60).ThisissupportedbyDayan and Bedenetto (2008) who concludedthat comments given by and to leaders helpencouraging2-waycommunicationaswellascreating better teamwork. Leaders shouldalso motivate the search for new ideas

Figure 4. The causal relationships of the five constructs of the MBNQA Framework (Adapted from Wilson and Collier, 2000)

Table 4. The GOF indices of the best-fit structural model

GOF Index Recommended Level Best-Fit Measurement

Model Best-fit Structural

Model

χ2/DF <2.00(Byrne,2001) 1.98 1.87

RMSEA ≤ 0.10 (Tabachnick and Fidell, 2007) 0.05 0.05

CFI >0.90(Kline,2005) 0.91 0.92

IFI >0.80(Garson,2006) 0.91 0.92

TLI >0.90(Kline,2005) 0.90 0.90

The best-fit measurement model is shown in Figure 3, and the best-fit structural model is shown in Figure 5.


of productivity enhancement by means ofbenchmarking,aswellasseekingoutthebestpractices for productivity improvement(Hussain,2008). ‘Leadership’,however,bearsnostatisticallysignificant relationship with ‘people’ (pathcoefficient = 0.14). The relatively stronginfluences‘dataandinformation’and‘strategicquality planning’ have on ‘people’ (pathcoefficients of 0.36 and 0.22, respectively)suggest that ‘leadership’ indirectly influences‘people’throughthese2constructs.Swierczekand Shrestha (2003), for example, statedthat to improve productivity, leaders shouldprovide workers with adequate resources,including information technology resources.Atthesametime,theworkersshouldreceivepropertrainingonhowtousethoseinformationtechnology tools, so that they can carry out

variousactivitieseffectively(BlackandLynch,1996; Hoffman and Mehra, 1999). This, inturn, leads to positive attitude, resulting inlower turnover, lower absenteeism, andimprovedproductivity(Marsidi,2009). ‘People’playsakey role in thesuccessof productivity enhancement, as seen bythe direct link from ‘people’ to ‘processmanagement’. Productivity improvementrequires highly committed and motivatedemployees. Personal recognition, workers’perceptions of work pressure, workers’empowerment, and workers’ attitude towardsproductivity improvement are important keysin improving productivity (Allender, 1984;Siuet al.,2004;Beck,2009).Negativeattitudesmay lead to high errors, lack of team spirit,andlowproductivity(Umiker,1998). The results also show that ‘data andinformation’hasnorelationshipwith‘strategicquality planning’. Also, no statisticallysignificantrelationshipisfoundbetween‘dataand information’ and ‘process management’,indicating the absence of a direct effect.An indirect effect exists, though, through‘people’. It is worth noting that the variancefor‘processmanagement’is0.64,demonstratingthat64%of thevarianceassociatedwith thisparticularfactorisaccountedforby‘people’.Inotherwords,‘people’explains(orinfluences)64%ofthevariancein‘processmanagement’.For instance, Bell (1988) stated that workersshouldbeassignedtothetasksforwhichtheyare best suited, as it helps the organizationin, reducing lead time, cost, inventory,and improving labour productivity (RyurikManagement Solutions, 2009).A summaryof the direct and indirect path coefficients isshowninTable5.

Conclusions The basic productivity model, as a result ofthe exploratory factor analysis, is developedbasedontheMBNQAFramework.The5keyproductivity factors, including ‘leadership’‘strategic quality planning’, ‘people’, ‘dataand information’, and ‘processmanagement’,aretestedwiththestructuralequationmodeling

Note:PRO=processesmanagement,LEA= leadership,STR = strategic quality planning, PEO = people,DAT=dataandinformationFigure 5. The best-fit structural model


Table 5. The direct and indirect path coefficients between the five productivity factors

Factor Correlation Coefficient

Strategicqualityplanning (0.64*LEA)

People(0.14*LEA)+(0.22*STR)+(0.36*DAT)+(0.14*LEA*STR)+(0.16*LEA*DAT)

Dataandinformation (0.45*LEA)

Processmanagement (0.60*LEA)+(0.32*PEO)+(0.12*DAT*PEO)All path coefficients are significant at 0.05 probability level. PRO = processes management, LEA = leadership, DAT = data and information, STR = strategic quality planning, PEO = people

Note:Allpathcoefficientsaresignificantat0.05probabilitylevel.TheitalisednumbersshowvariancevaluesR2(R-square)ofthefactors.

Figure 6. The final productivity model

toempiricallyvalidatethecausalrelationshipsamong them. The results improve theunderstanding of the interrelationships anddetermine where the organization shouldfocusinordertoachievehigherproductivity. ‘Leadership’ is found to be the maindriver to productivity enhancement. Thestrong commitment of the leader is, thus,crucial in improving productivity. In thisstudy, it is found that ‘leadership’ stronglyinfluences ‘strategic quality planning’ and‘process management’. Leaders, therefore,

shouldhelptoencourage2-waycommunication,creategoodteamwork,motivatethesearchfornew ideas of productivity enhancement bymeansofbenchmarking,andseekoutthebestpractices for productivity improvement.‘Leadership’ also indirectly influences‘people’ through ‘data and information’ and‘strategic quality planning’. Leaders shouldthusprovideworkerswithadequateresources,including information technology resources.Atthesametime,theworkersshouldreceivepropertrainingonhowtousethoseinformation


technology tools, so that they can carry outvariousactivitieseffectively. ‘People’isalsofoundplayingakeyrolein the success of productivity enhancement.Personal recognition, workers’ perceptionsof work pressure, workers’ empowerment,and workers’ attitude towards productivityimprovementareimportantkeysinimprovingproductivity. Negative attitudes may leadto high errors, lack of team spirit, and lowproductivity.Additionally,‘dataandinformation’should ease the process implementationperformed by workers, as direct and indirectrelationships exist between ‘data andinformation’ and ‘people’ and between ‘dataand information’ and ‘process management’,respectively. The 5 key productivity factors, as wellastheircausalrelationships,canbeusedasaguideline for an organization in planningfor its productivity enhancement.The resultscan also be used as an empirical basis forcomparisoninfutureresearch.

References Allender,M.C.(1984).ProductivityEnhancement:

ANewTeamworkApproach. JohnWiley&Sons,NewYork,NY,U.S.A.

AsianProductivityOrganization.(2006).HandbookonGreenProductivity.Availablefrom:http://www.apo-tokyo.org/gp/e_publi/ebooks_gp/APO_Handbook_on_Productivity.pdf.Accesseddate:Oct16,2008.

AsianProductivityOrganization.(2007).Productivitymeasurement.APO News, Tokyo, Japan,37(5):1-8.

Barocci,T.A.,andWever,K.R.(1982).Productivity,industrial relations and human resourcemanagement: football on a soccer field.WorksPaper,SloanSchoolofManagement,Massachusetts Institute of Technology,Cambridge,Mass.,USA.

Batra, V.K., Kumar, K., Chhangani, P.N., andParihar, P.S. (2009). Plant Operations andProductivity Enhancement - a Case Study.Available from: http://www.holtecnet.com/web/content/references/TechnicalPapers/p_2004_5.pdf.Accesseddate:Jun4,2009.

Beck, M. (2009). Protect yourAttitude.Availablefrom: http://www.xleaders. com/articlesProtect%20Your%20Attitude. pdf.Accessed

date:Jun3,2009.Bell, C.R. (1988). Human Resources and Labour

Relations.AmericanManagementAssociation,ManagingSolutions.

Black, S.E., and Lynch, L.M. (1996). Human-capital investments and productivity. TheAm.Econ.Rev.,86(2):263-267.

Blazey,M.(2008).InsighttoPerformanceExcellence2008:An Inside Look at the 2008 BaldrigeAwardCriteria.ASQQualityPress,Milwaukee,WI,USA.

Byrne,B.M.(2001).StructuralEquationModellingwithAMOS: Basic Concepts,Applicationsand Programming. Lawrence ErlbaumAssociations,Inc.,Mahwah,NJ,USA.

Clissold,G.(2004).UnderstandingSafetyPerformanceusing Safety Climate and PsychologicalClimate.Working Paper 65/04, Departmentof Management, Monash University,Melbourne,Australia.

Coakes,S.J.andSteed,L.G.(2003).SPSSAnalysiswithoutAnguishVersion 11.0 forWindows.JohnWiley&Sons,Australia.

Cosmetatos, G.P. and Eilon, S. (1983). Effects ofproductivity definition and measurementon performance evaluation. Eur. J. Oper.Res.,14(1):31-35.

Dayan, M., and Bedenetto,A. (2008). Proceduraland Interactional Justice Perceptions andTeamwork Quality.Available from: http://w w w. e m e r a l d i n s i g h t . c o m / I n s i g h t /ViewContentServlet?contentType=Article&Filename=Published/Emerald FullTextArticle/Articles/0800230805.html.Accesseddate:Feb13,2010.

Dessler,G.(2005).HumanResourceManagement.10thed.PearsonEducationInc.,UpperSaddleRiver,NJ,USA.

European ProductivityAgency. (2008).Definitionof Productivity. Available from: http://www.npccmaur i t ius .com/def in i t ion/ .Accesseddate:Oct16,2008.

Garson,G.D.(2006).StructuralEquationModelling.Available from: http://www2.chass. ncsu.edu/garson/pa765/structur.htm.Accesseddate:Aug4,2006.

Hair, J.F.,Anderson,R.E.,Tatham,R.L.,andBlack,W.C. (1998). Multivariate DataAnalysis.5thed.Prentice-HallInternational,EnglewoodCliffs,NJ,USA.

Haynes, R.M. and Duvall, P.K. (1992). Servicequality management: a process-controlapproach.Int.J.Serv.Ind.Manag.,3(1):14-24.

Heizer, J. and Render, B. (2008). OperationsManagement. 9th ed. Prentice-Hall, Upper


SaddleRiver,NJ,USA.Herron,C.andBraiden,P.M.(2006).Amethodology

for developing sustainable quantifiableproductivity improvement in manufacturingcompanies.Int.J.Prod.Econ.,104:143-153.

Hoffman,J.M.andMehra,S.(1999).Managementleadership and productivity improvementprograms. Int. J.App. Qual. Manag., 2(2):221-232.

Hussain, M.R. (2008). Striving for OrganizationalExcellencethroughProductivityEnhancement.PSMB2008Conference,MalaysiaProductivityCorporation.Available from: http://www.hrdnet.com.my/imej%20conf%202008/5Striving.pdf.Accesseddate:Jun3,2009.

Investopedia.(2008).Productivity.Availablefrom:http://www.investopedia.com/terms/p/productivity.asp.Accessed date: Oct 17,2008.

Jorgenson, D.W. and Griliches, Z. (1967). Theexplanation of productivity change. Rev.Econ.Stud.,34(3):249-283.

Kilic,H.andOkumus,F.(2005).Factorsinfluencingproductivity in small islandhotels:evidencefromNorthernCyprus. International Journalof Contemporary Hospitality Management,17(4):315-331.

Kline, R.B. (2005). Principles and Practice ofStructural Equation Modelling. 2nd ed.GuilfordPress,NewYork,NY,USA.

Knowledgerush (2008). Productivity.Availablefrom: http://www.knowledgerush.com/ kr/encyclopedia/Productivity/.Accessed date:Oct20,2008.

Lee,P.M.andQuazi,H.A.(2001).Amethodologyfor developing a self-assessment tool tomeasurequalityperformanceinorganizations.Int.J.Qual.Rel.Manag.,18(2):118-141.

Loke, C.F. (2001). Leadership behaviours: effectson job satisfaction, productivity, andorganizational commitment. Journal ofNursingManagement,9(4):191-204.

Lubbe,B.(2000).TourismDistribution:Managingthe Travel Intermediary. Juta & CompanyLimited,CapeTown,SouthAfrica.

Marsidi,A.(2009).ImprovingProductivitythroughPeople. Malaysian Institute of Management.Available from: http://mgv.mim.edu.my/MMR/8412/841203.htm.Accesseddate: Jun3,2009.

McBurney, D.H. (1994). Research Methods.Brooks/Cole Publishing Company, PacificGrove,California,U.S.A.

Millea,M.andFuess,S.M.(2005).Doespayaffectproductivity or react to it? Examination of

U.S.manufacturing. Q.Rev.Econ.Financ.,45:796-807.

NPC. (2003). International Best Practices StudyMissiontoTaiwan.TheNationalProductivityCorporation, New Delhi, India.Availablefrom: http://www.mpc.gov.my/webdoc/BP1a.pdf.Accesseddate:Jun3,2009.

Pallant, J. (2005). SPSS Survival Manual:A StepbyStepGuide toDataAnalysisusingSPSSforWindows (Version 12),Allen & Unwin,NewSouthWales,Australia.

Pannirselvam, G.P. and Ferguson, L.A. (2001).Astudy of the relationships between theBaldrige categories. Int. J. Qual. Rel.Manag.,18(1):14-34.

Peter, E., Matta, R., and Boe, W. (2007). Short-term work scheduling with job assignmentflexibility for amulti-fleet transport system.Eur.J.Oper.Res.,280(1):82-98.

Poetscheke, B. (1995). Key success factors forpublic/private-sector partnerships in islandtourism planning. In: Island Tourism:ManagementPrinciplesandPractice.Conlin,M., and Baum, T., (eds). John Wiley andSons,NewYork,NY,USA.

Pritchard, R.D. (1995). Productivity Measurementand Improvement: Organizational CaseStudies. Greenwood Publishing Group,Westport,CT,USA.

Pupat, N. (2009). The Impact of OrganizationIntervention on Service Process forCurriculum Development inAssumption:Case Study. Available from: http://gsbejournal .au.edu/Ph.DOD/1.THE%20IMPACT%20OF%20ORGANIZATION%20DEVELOPMENT%20INTERVENTION%20ON%20SERVICE%20PROCESS%20FOR%20CURRICULUM%20DEVELOPMENT%20IN%20ASSUM P T I O N % 2 0 U N I V E R S I T Y. p d f .Accesseddate:Dec20,2009.

Relgolook. (2009). Productivity EnhancementSolution for Effective Email Management.Available from: http://www.articlebase.com/management-ar t ic les / re lgolook-productivity-enhancement-solution-for-effective-email-management-800221.html.Accesseddate:Jun4,2009.

Rogers,M.(1998).TheDefinitionandMeasurementofProductivity.MelbourneInstituteWorkingPaper No. 9/98. Melbourne Institute ofAppliedEconomic andSocialResearch, theUniversityofMelbourne,Australia.

Russell,R.S.andTaylorIII,B.W.(2009).OperationsManagementalongtheSupplyChain.6thed.


JohnWileyandSons,Hoboken,NJ,USA.Ryurik Management Solutions. (2009). How to

Optimize JobsAllocation.Available from:http://bmstudies.com/allocations.html.Accesseddate:Sep6,2009.

SalesCreators.(2009).BetterWorkplace.Availablef rom:h t tp : / /www.Salescrea tors .com/Section4/betterWkPlace.html.Accesseddate:Jun3,2009.

Savery,L.K.(1998).Managementandproductivityincreases.JournalofManagementDevelopment,17(1):68-74.

Seo,D.C.,Torabi,M.R.,Blair,E.H.,andEllis,N.T.(2004).A cross-validation of safety climatescale using confirmatory factor analyticapproach.JSafetyRes,35:427-445.

Sharma,S.(1996).AppliedMultivariateTechniques.JohnWileyandSons,NewYork,NY,USA.

Siu, O., Phillips, D.R., and Leung, T. (2004).Safety climate and safety performanceamongconstructionworkers inHongKong:the role of psychological strains asmediators.AccidentAnal.Prev.,36:359-366.

Swierczek, F.W. and Shrestha, P.K. (2003).Information technology and productivity:a comparison of Japanese andAsia-Pacificbanks. The Journal of High TechnologyManagementResearch,14(2):269-288.

Tabachnick, B.G. and Fidell, L.S. (2007). UsingMultivariate Statistics. 5th ed. PearsonEducation,Inc.,NewYork,NY,U.S.A.

Tangen, S. (2002). Understanding the conceptof productivity. Proceedings of the 7thAsiaPacificIndustrialEngineeringandManagementSystems Conference; Dec. 18-20, 2002;Taipei,Taiwan,p.1-4.

ThailandBoardofInvestment.(2005).ProductivityManagementCockpit:ToolsforPerformanceExcellence. Seminar on Productivity and

Investment Climate Survey in Thailand.Available from: http://www2.ftpi.or.th/dwnld/achieve/pics/cockpit.htm.Accesseddate:Oct20,2008.

Thailand Productivity Institute. (2006).AnnualReport on Productivity.Available from:http://www.ftpi.or.th/Portals/0/download/prdarticle/pdf_0.pdf.Accessed date: Oct 20,2008.

ThailandProductivity Institute. (2008a). InventoryandHumanResourceManagement.Availablefrom: http://www.ftpi.or.th/LinkClick.aspx?fi leticket=capPRV81GyQ%3d&tabid=96&mid=656.Accesseddate:Jun3,2009.

Thailand Productivity Institute. (2008b). QualityControl System. Available from: http://www.ftpi.or.th/LinkClick.aspx?fileticket=AbV7AB00kAw%3d&tabid=96&mid=656.Accesseddate:Jun3,2009.

Thailand Productivity Institute. (2008c). SetupTime Reduction. Available from: http://www.ftpi.or.th/LinkClick.aspx?fileticket=qCRA1dzh0Xo%3d&tabid=96&mid=656.Accesseddate:Jun3,2009.

TradeUnionists.(2008).DefinitionofProductivitybyTrade Unionists.Available from: http://www.npccmauritius.com/tudef/.Accesseddate:Oct17,2008.

Umiker,W.O. (1998). Management Skills for theNewHealthcareSupervisor.3rded.JonesandBartlettPublishers,Sudbury,MA,U.S.A.

Wilson, D.D. and Collier, D.A. (2000). Anempirical investigation of the MalcolmBaldrige National Quality Award causalmodel.DecisionSci.,31(2):361-383.

Wong, P.S.P. and Cheung, S.O. (2005). Structuralequation model of trust and partneringsuccess.J.Manage.Eng.,April:70-80.

Documents

STRUCTURAL EQUATION MODELLING OF PRODUCTIVITY … · 2013-11-25 · Productivity Institute, 2006). theBy implementing such a program, followingsthe company can reduce waste • Increaseand