ijqrm an instrument for measuring tqm implementation for ... · 3/7/2015 · 17,7 730...
TRANSCRIPT
IJQRM17,7
730
International Journal of Quality &Reliability Management,Vol. 17 No. 7, 2000, pp. 730-755.# MCB University Press, 0265-671X
Received July 1999Revised December 1999
An instrument for measuringTQM implementation forChinese manufacturing
companiesZhihai Zhang, Ab Waszink and Jacob Wijngaard
Faculty of Management and Organisation, University of Groningen,Groningen, The Netherlands
Keywords TQM, Implementation, Measurement, Empirical study, China, Manufacturing
Abstract From an extensive review of the literature in the field of total quality management(TQM), 11 constructs of TQM implementation were identified. An instrument measuring theseconstructs was developed. The reliability and validity of the instrument were tested and validatedusing data from 212 Chinese manufacturing companies. Various methods were employed for thistest and validation. Comparisons between this instrument and the three other qualitymanagement instruments were made. It was concluded finally that the instrument presented inthis paper was reliable and valid. Researchers will be able to use this instrument for developingquality management theory. Industrial practitioners will be able to use this instrument to evaluatetheir TQM implementation so as to target improvement areas.
IntroductionThe level of awareness of total quality management (TQM) has increasedconsiderably over the past few years. There have been many discussions on thesubject of TQM. However, the contents of these papers differ to a considerabledegree. Some focus on the conceptual issues of TQM (e.g. Reeves and Bednar,1994; Anderson and Schroeder, 1994; Dean and Bowen, 1994; Waldman, 1994),while others focus on the practical and empirical issues of TQM (e.g. Forza andFilippini, 1998; Choi and Eboch, 1998; Anderson et al., 1995; Mann, 1992;Mann and Kehoe, 1994; Blauw, 1990). There is a consensus that TQM is away of managing an organisation to improve its overall effectiveness andperformance. There is less agreement as to what the primary constructs ofTQM are, or what the overall concept of TQM is. No uniform view of TQMexists today. So far, TQM has come to mean different things to different people(Hackman and Wageman, 1995).
Although many Chinese manufacturing companies began to implement totalquality control from 1978 onwards, China still lacks effective qualitymanagement systems and application at the enterprise level. Some basicquality principles and modern quality management methods have not beenwidely used by Chinese manufacturing enterprises (Zhao et al., 1995). Althoughgreat efforts have been made by the Chinese government to stimulatecompanies to implement TQM and improve product quality, there has not beensatisfactory progress. The country's product quality as a whole is still at arelatively low level (Zhang, 1998a). A number of quality management problems
The research register for this journal is available athttp://www.mcbup.com/research_registers/quality.asp
The current issue and full text archive of this journal is available athttp://www.emerald-library.com
An instrumentfor measuring
TQM
731
still remain unsolved. After reviewing the literature related to China's qualitymanagement, it became very clear that little empirical research has beenconducted in the area of TQM implementation in Chinese manufacturingcompanies. Therefore, the current situation of TQM implementation in Chinesemanufacturing companies still remains unclear. Owing to lack of empiricalstudies in the field of TQM, it is difficult for Chinese manufacturing companiesto obtain sufficient information to support their TQM implementationpractices. As a consequence, many Chinese manufacturing companies haveexperienced difficulties or failures in implementing TQM.
In order to bridge the gap and provide Chinese manufacturing companieswith practical assistance in the area of TQM implementation, this researchaimed at identifying TQM implementation constructs, developing aninstrument for measuring these constructs, and empirically validating theinstrument using data from Chinese manufacturing companies. Researcherswill be thus able to use this instrument for developing quality managementtheory related to Chinese manufacturing companies. Chinese industrialpractitioners will be able to use this instrument to evaluate the status of theirTQM implementation so as to target improvement areas.
Several similar studies dealing with empirically validated scales forintegrated quality management have already been conducted (e.g. Saraph et al.,1989; Flynn et al., 1994; Ahire et al., 1996; Anderson et al., 1995; Powell, 1995;Black and Porter, 1996). These instruments are different in terms of constructsand measurement items. Each one has its own strengths and weaknesses.According to the goal of this study and the authors' understanding of TQM,none of these instruments could have been directly employed in this study. Anew instrument for measuring TQM implementation for Chinesemanufacturing companies therefore had to be developed.
The rest of this paper is organised as follows. Section two presents thedevelopment of the 11 TQM implementation constructs. Section three involvesinstrument development and information concerning respondent companies.The procedures and the methods used for testing and validating the TQMimplementation instrument are provided in section four. Finally, section fiveconcludes this paper together with some discussions.
Developing TQM implementation constructsConstruct identificationOver the past few decades, writers such as Deming (1986), Crosby (1979), Juranand Gryna (1993), Feigenbaum (1991), Ishikawa (1985), and others havedeveloped certain propositions in the area of quality management. Theirinsights into quality management provide a good understanding of qualitymanagement principles. World-wide, there are several Quality Awards, such asthe Deming Prize (1992) in Japan, the European Quality Award in Europe(1994) and the Malcolm Baldrige National Quality Award (1997) in the USA.Each award is based on a perceived model of TQM. They do not focus solely onproduct, service perfection or traditional quality management methods, but
IJQRM17,7
732
consider a wide range of management activities, behaviour and processes thatinfluence the quality of the final offerings. These award models provide auseful audit or assessment framework against which organisations canevaluate their quality management methods, the deployment of these methods,and the end business results.
Based on the comprehensive review of the TQM literature, the following 11constructs were considered to be the TQM implementation constructs:
(1) leadership;
(2) supplier quality management;
(3) vision and plan statement;
(4) evaluation;
(5) process control and improvement;
(6) product design;
(7) quality system improvement;
(8) employee participation;
(9) recognition and reward;
(10) education and training; and
(11) customer focus.
Detailed explanations of the 11 theoretical constructs will be presented in thefollowing paragraphs.
LeadershipThe European Quality Award and the Malcolm Baldrige Quality Awardrecognise the crucial role of top management leadership in creating the goals,values and systems that guide the pursuit of continuous performanceimprovement. Recognition of the critical role of top management and itsresponsibility in pursuit of quality improvement echoes the arguments putforward by gurus of quality such as Deming et al. A predominant theme inTQM literature is that strong commitment from top management is vital. Thefoundation of an effective TQM effort is commitment. Lack of top managementcommitment is one of the reasons for the failure of TQM efforts (Brown et al.,1994). Garvin (1986) reports that high levels of quality performance havealways been accompanied by an organisational commitment to that goal; highproduct quality does not exist without strong top management commitment.Many such empirical studies have also found that top management support forquality is a key factor in quality improvement. If top managers are committedto quality, they should not only actively be involved in quality managementand improvement process, but also strongly encourage employee involvementin quality management and improvement process. In addition, they shouldlearn quality-related concepts and skills, and arrange adequate resources for
An instrumentfor measuring
TQM
733
employee education and training. Various quality-related issues should also beoften discussed in top management meetings. Top management should pursuelong-term business success and focus on product quality rather than yields.
Supplier quality managementSupplier quality management is an important aspect of TQM since materialsand purchased parts are often a major source of quality problems. The MalcolmBaldrige Quality Award (1997) also recognises the importance of supplierquality. Garvin (1983) finds that organisations that manufacture the highestquality products have purchasing departments that rank quality rather thancost minimisation as their major objective. Conversely, in organisations withthe lowest quality performance, he finds that the primary objective of thepurchasing department is to obtain the lowest price for technically acceptablecomponents. Poor quality of supplier products results in extra costs for thepurchaser; e.g. for one appliance manufacturer, 75 percent of all warrantyclaims were traced to purchased components for the appliances (Juran andGryna, 1993). If organisations pursue good supplier quality management, theyshould establish long-term co-operative relations with their suppliers, oftenparticipate in supplier quality activities, have detailed information concerningsupplier performance, give feedback on the performance of suppliers' products,regularly conduct supplier quality audits, and regard product quality as themost important factor for selecting suppliers.
Vision and plan statementA vision statement describes how a company wants to be seen in its chosenbusiness. As such, it describes standards, values, and beliefs. Above all, avision is an advertisement of the intention to change. It propels the companyforward and acts against complacency. A vision statement usually cascadesdown to mission statements that detail short-term site (plant) aims ordepartmental aims. In order to realise a vision statement, an organisation mustmake plan statements which support the realisation of the vision. These plansmay be a detailed business plan, a quality policy, a quality goal, and a qualityimprovement plan. These plans and statements should be well communicatedto the employees of organisations and in return employees will be encouragedin their commitment to quality. In order to make these plans and statements,employees from different levels should be involved.
EvaluationEvaluating the situation in organisations' quality management practicesprovides an important base for organisations to improve their qualitymanagement practices. Juran and Gryna (1993) state that a formal evaluation ofquality provides a starting point by providing an understanding of the size ofthe quality issue and the areas demanding attention. Benchmarking is apowerful tool to use as a continuous process of evaluating an organisation'sproducts, services, and processes against those of its toughest competitors or
IJQRM17,7
734
those of organisations renowned as world-class or industry leaders. Evaluationactivities consist of evaluation for various policies and strategies, quality audit,analysis of quality costs, department/function performance evaluation, andemployee performance evaluation. In order to conduct evaluation activities,companies should have various pieces of quality-related information, such asdefect rates and scraps. This information should be communicated toemployees in order to stimulate employees to make things better. Moreover, theaim of evaluation is for improvement rather for criticism.
Process control and improvementA key part of any total quality strategy is the management of processes (Porterand Parker, 1993). Process refers to some unique combinations of machines,tools, methods, materials, and people engaged in production (Juran and Gryna,1993). Process management focuses on managing the manufacturing process sothat it operates as expected, without breakdowns, missing materials, fixtures,tools, etc., and despite workforce variability. One important matter in processmanagement is to ensure that process capability can meet productionrequirements. One aspect of process management is equipment maintenance,which ensures that variation is kept within acceptable bounds, keeping themanufacturing process running smoothly. Good process management shouldinvolve precisely documenting various process procedures, with instructionsfor equipment operation in order to minimise the likelihood of operator errors.Some methods, such as PDCA cycle, seven QC tools, statistical process control(SPC), sampling and inspection are effective for process control and processimprovement.
Product designProduct design is an important dimension of quality management. For complexproducts, errors during product development cause about 50 percent of fitness-for-use problems (Juran and Gryna, 1993). Sound product design meets orexceeds the requirements and expectations of customers better than thecompetitors, leading to an increased market share. For improving productdesign, design engineers are required to have some shopfloor and marketingexperiences. Customer requirements and production cost should be thoroughlyconsidered during the process of product design. Different departments in anorganisation should participate in new product development. Beforeproduction, new product design should be thoroughly reviewed in order toavoid problems happening during production. Experimental design (Zhang,1998b) and quality function deployment (Daetz et al., 1995) are two importantand effective methods in product design.
Quality system improvementA documented quality system as part of a TQM strategy can contribute toTQM by managing the organisation's processes in a consistent manner. Aquality system is defined as the organisational structure, procedures, processes
An instrumentfor measuring
TQM
735
and resources needed to implement quality management (ISO 8402, 1994). In1987, ISO published the ISO 9000 standards series on quality management andquality assurance. When ISO 9000 is implemented, a quality manual, qualityprocedures, and work instructions are established. An organisation mayeventually apply to be registered as having an ISO 9001 (9002 or 9003) qualitycertificate.
Employee participationBy personally participating in quality improvement activities, employeesacquire new knowledge, see the benefits of the quality disciplines, and obtain asense of accomplishment by solving quality problems. The participation leadsto lasting changes in behaviour. Participation is decisive in inspiring action onquality improvement (Juran and Gryna, 1993). Participation may enable theemployees to improve their personal capabilities, increase their self-respect,commit themselves to the success of their organisations, and/or change certainpersonality traits. Participation may also change employees' negative attitudesand instil in the employees a better understanding of the importance of productquality. Participation may contribute to the establishment of a company-widequality culture. Employees in organisations should be encouraged to reporttheir work problems. Good employees' suggestions should be implementedafter being evaluated. Methods such as cross-functional teams, within-functional teams, QC circles, voluntary teams, and suggestion activities can beused for encouraging employee participation.
Recognition and rewardIt almost goes without saying that an important feature of any qualityimprovement program is showing due recognition for improved performanceby any individual, section, and department or division within the company(Dale and Plunkett, 1990). Both teams and individuals can be recognised andrewarded for their excellent performance. To effectively support organisations'quality efforts, they need to implement an employee compensation system thatstrongly links quality and customer satisfaction with pay (Brown et al., 1994).Recognition and reward activities should effectively stimulate employeecommitment to quality improvement. Otherwise, these activities are failures.Working condition improvements, salary promotions, position promotions,monetary or non-monetary rewards, financial awards for excellent suggestionsare good methods for recognition and reward.
Education and trainingDeming (1986) stresses the importance of education and training for continualupdating and improvement. Many research results reveal that education andtraining are one of the most important elements in a successful implementationof TQM (e.g. Mann, 1992). The research confirms what most organisationshave already realised, namely, that education and training are an integral andessential part of the TQM initiative. Investment in education and training is
IJQRM17,7
736
vitally important for TQM success. Employees should be regarded as valuable,long-term resources worthy of receiving education and training throughouttheir career. All management personnel, supervisors, and employees shouldaccept quality education and training such as quality awareness education andquality management methods education.
Customer focusTo achieve quality, it is essential to know what customers want and to provideproducts or services that meet their requirements (Ishikawa, 1985). Asuccessful organisation recognises the need to put the customer first in everydecision made. The key to quality management is maintaining a closerelationship with the customer in order to fully determine the customer's needs,as well as to receive feedback on the extent to which those needs are being met.The customer should be closely involved in the product design anddevelopment process, with inputs at every stage of the process so that there isless likelihood of quality problems once full production begins (Flynn et al.,1994). The ultimate measure of company performance is customer satisfaction,which may very well predict the future success or failure of an organisation(Kanji and Asher, 1993). In order to improve customer satisfaction, customercomplaints should therefore be treated with top priority. Warranty on soldproducts should also be provided. Methods that can be used for customer focusefforts include collections of customer complaint information, marketinvestigations, and customer satisfaction surveys.
Instrument comparisonA comparison between this instrument and other instruments was conductedin order to identify the characteristics of this instrument. Several TQMimplementation instruments have already been developed by differentresearchers (e.g. Saraph et al., 1989; Flynn et al., 1994; Anderson et al., 1995;Ahire et al., 1996; Powell, 1995; Black and Porter, 1996). In this paper, only threeinstruments were selected for this comparison. The three instruments weredeveloped by Saraph et al. (1989), Flynn et al. (1994), and Ahire et al. (1996),respectively. The instrument presented in this paper has two uniquecharacteristics.
The first one is that this instrument covers a broader scope of TQM incomparison with their instruments. Ahire et al. (1996) strongly recommendedthat a combination of the three instruments should be undertaken for futureresearch on quality management. Following their suggestions, all of theconstructs in the three instruments were carefully examined. The authors triedto integrate their constructs into this instrument as much as possible. Table Ilists the comparison of these instruments. The two constructs, namely, productquality and supplier performance in the Ahire et al. instrument, were notincluded in this instrument since they represent TQM outcomes. Role of qualitydepartment in the Saraph et al. instrument was excluded in this instrumentsince every department in any organisation is involved in quality management.
An instrumentfor measuring
TQM
737
Benchmarking and internal quality information usage in the Ahireet al. instrument were integrated to form the construct of evaluation in thisinstrument. Process control and cleanliness and organisation in the Flynnet al. instrument were combined to formulate the construct of process controland improvement in this instrument. This instrument includes two moreconstructs, namely, quality system improvement and vision and planstatement, which are not found in their instruments. In addition, thisinstrument has 78 measurement items in total (the original instrument had 79items and one item was deleted after factor analysis), which is more than theiritems. This instrument therefore covers a broader scope of TQM.
The second one is that specific characteristics of Chinese manufacturingcompanies were taken into account in developing this instrument. Since the aimof this research was to develop an instrument for measuring TQMimplementation for Chinese manufacturing companies, the instrument had tobe suitable for use in China. Thus, specific characteristics of Chinesemanufacturing companies had to be taken into account when developing thisinstrument. For example, most Chinese companies are weak in terms of theirvisions and plans. Vision and plan statement is therefore one construct in thisinstrument. Most Chinese companies are trying to implement ISO 9000 in orderto improve their quality systems. This instrument thus includes the constructof quality system improvement. Some Chinese top managers prefer to pursueshort-term business success because of the nature of the country's institutionalsystem. Item 8 (top management pursues long-term business success) in scale 1
This instrumentSaraph et al.instrument
Flynn et al.instrument
Ahire et al.instrument
12
3
45
67
8
9
10
11
LeadershipSupplier qualitymanagement andquality policyVision and planstatementEvaluationProcess controland improvementProduct designQuality systemimprovementEmployeeparticipationproceduresRecognition andrewardEducation andtrainingCustomer focus
1
2
34
5
6
7
8
Role of divisionaltop managementRole of qualitydepartmentTrainingProduct/servicedesignSupplier qualitymanagementProcessmanagement/operatingQuality data andreportingEmployeerelations
1
2
345
6
7
8
910
11
QualityleadershipQualityimprovementrewardsProcess controlFeedbackCleanliness andorganisationNew productqualityInterfunctionaldesign processSelection forteamworkpotentialTeamworkSupplierrelationshipCustomerinvolvement
1
23
4
567
8
9
101112
Top managementcommitmentCustomer focusSupplier qualitymanagementDesign qualitymanagementBenchmarkingSPC usageInternal qualityinformation usageEmployeeempowermentEmployeeinvolvementEmployee trainingProduct qualitySupplierperformance
Table I.Construct comparison
IJQRM17,7
738
(leadership) is therefore a very important item for measuring leadership ofChinese manufacturing companies. This item cannot be found from theexisting instruments. Most Chinese companies implement reward and penaltymeasures for strengthening their management. An item `̀ employees' rewardsand penalties are clear'' was therefore developed for scale 9 (recognition andreward). For details, please refer to the Appendix, and note that it is beyond thescope of this paper to address other characteristics of Chinese manufacturingcompanies.
Research methodologyInstrument developmentThe aim of this study is to develop an instrument for measuring TQMimplementation for Chinese manufacturing companies. To reach such a goal, aset of items for measuring TQM implementation constructs had to be welldeveloped. This was realised on the basis of a thorough review of the TQMliterature, expert guidance, and input from colleagues.
The following describes some of the difficulties experienced in developingthis instrument. Most of the TQM literature reviewed in this study was inEnglish. The instrument was thus first developed in English. However, thisinstrument was actually used for collecting information in China. The Englishversion thus had to be translated into Chinese. This translation might havebiased the original design of the instrument. A few quality management terms,such as benchmarking, could not be precisely translated into Chinese terms.The instrument had to be easily understood by respondents so as to avoidconfusion. Otherwise, the research findings might have been biased.
In this study, various measures were taken in order to minimise thesepotential problems. The English version was translated into Chinese by aChinese researcher who was doing quality management research in a Westerncountry. A few years previously, he had worked in the field of qualitymanagement in China. Therefore, it was assumed that he had enoughknowledge of quality management in both English and Chinese. A few Englishterms were translated into Chinese by providing additional explanations sothat respondents could understand them better. After translation, the Chineseversion of the instrument was presented to three quality managers who workedin different Chinese manufacturing companies. They were asked whether:
. the items were stated in a shared vocabulary;
. the items were precise and unambiguous; and
. there were biased wordings.
Some alterations were made according to their suggestions. During theresearch visit to China, the Chinese version of the instrument was formally pre-tested on various people (i.e., governmental officials, consultants, researchers,practitioners, and quality managers). These people were asked to providefeedback on ease of comprehension, clarity of the specific items, suggestions for
An instrumentfor measuring
TQM
739
possible change, and suggestions for additional items, etc. Based on this, theChinese version of the instrument was further modified. After this step, theauthors were confident that the instrument could be used for the large-scalesurvey. The final Chinese version of the instrument consisted of 79 items.The Appendix lists the instrument that was retranslated back into Englishfrom the final Chinese version by the same translator. If readers requesta Chinese version of the instrument, the authors will provide it. Followingother similar studies (e.g. Flynn et al., 1994), a widely used five-point Likertscale was employed for scoring responses (1 = strongly disagree; 2 = disagree;3 = neutral; 4 = agree; 5 = strongly agree).
SampleA mail survey was judged to be the most appropriate method for collectingdata for empirically testing and validating the instrument. The survey wasadministered in Liaoning province in China between July and October 1998.The type of sample and the number of companies were determined on the basisof the information requirements. In this study, all of the sampled companieswere from Liaoning province, one of the most important industrial centres inChina. In the Liaoning region, there were more than 1,000 large and medium-sized manufacturing companies at the time of the survey. Almost all of thesecompanies implemented TQM, or at least part of TQM. Therefore, they hadsome knowledge of TQM implementation. In this study, only manufacturingfirms with annual sales greater than RMB 10 million (RMB is the Chinesecurrency; US$1 = RMB8.30) were randomly selected for investigations.
The information about the companies was obtained from the LiaoningProvincial Statistics Bureau. In Liaoning province, there were 2,929manufacturing firms with annual sales in excess of RMB10 million in the year1997. A sample of 900 manufacturing companies was randomly selected fromthe database. This was done with the help of a computer. The sample size wasdecided after considering the expected response rate, the requirements forperforming statistical analysis, and the survey cost. The 900 companies weredivided into four groups on the basis of industrial sectors, 301 for machinery,180 for chemicals, 97 for electronics, and 322 for other industrial sectors. TheLiaoning Provincial Machinery Bureau, the Liaoning Provincial ChemicalsBureau, the Liaoning Provincial Electronics Bureau, and the LiaoningProvincial Quality Management Association sent 301, 180, 97, and 322questionnaires, respectively, to quality managers in these companies. A total of212 questionnaires were eventually returned. As a whole, the response rate was23.56 percent. The above mentioned four organisations received 97, 44, 21, and50 questionnaires, respectively. Their respective response rates were 32.23,24.44, 21.65, and 15.53 percent.
Respondent companiesOf these 212 manufacturing companies, 82 were large, 70 medium-sized, and 60small (firm sizes were categorized by the Chinese government according to
IJQRM17,7
740
their fixed assets and industrial sectors). The current number of employeesworking in the companies that responded ranged from 75 to 26,809. Onaverage, each firm employed 2,006 employees. Approximately half of thecompanies exported products to foreign countries. The annual sales of thesecompanies ranged from RMB10 million to 3,500 million. On average, the annualsales were RMB154 million. Of the companies that responded, 70 firms werelosing money, 18 firms were breaking even, and 124 firms were making aprofit. Of the 70 money-losing companies, a total of RMB195 million had beenlost. On average, each firm lost RMB2.80 million. Table II provides furtherinformation about these companies.
Empirical assessment of the instrumentIn this study, the instrument that was developed consisted of 11 scales (79items). These scales had to be empirically tested and validated. Many methodswere available for empirically assessing the reliability and validity of ameasurement scale. This section will detail how the reliability and validity ofthese scales were evaluated.
ReliabilityReliability relates to the extent to which an experiment, test, or any measuringprocedure yields the same results on repeated trials (Carmines and Zeller, 1979).
Table II.Information about therespondent companies
Characteristics Frequency Percentage
Industrial sectorsMachinery 97 45.8Chemical 44 20.8Electronics 21 9.9Building materials 11 5.2Textile 11 5.2Light industry 10 4.7Food industry 7 3.3Metallurgical industry 7 3.3Medicine industry 4 1.9
OwnershipState-owned company 137 64.6Collective company 19 9.0Township company 19 9.0Joint venture 13 6.1Others 24 11.3
Establishment yearBefore 1949 34 161950-1966 84 39.61967-1978 48 22.61979-1990 31 14.61991-1997 15 7.1
Note: 212 respondent companies in total
An instrumentfor measuring
TQM
741
It is a statistical measure of how reproducible the data of the survey instrumentare (Litwin, 1995). There are four methods which can be used for assessingreliability:
(1) the test-retest method;
(2) the alternate-form method;
(3) the split-halves method; and
(4) the internal consistency method (Nunnally, 1967).
Of these four methods, internal consistency reliability is the most commonlyused psychometric measure in assessing survey instruments and scales.Internal consistency is an indicator of how well the different items measure thesame concept. This is important since a group of items purporting to measureone variable should indeed be clearly focused on that variable. Internalconsistency is measured by calculating a statistic known as Cronbach'scoefficient alpha (Nunnally, 1967; Cronbach, 1951). Coefficient alpha measuresinternal consistency reliability among a group of items combined to form asingle scale. It is a statistic that reflects the homogeneity of the scale. Table IIIlists Cronbach's alpha for different scales. It should be noted that one itemwas deleted after factor analysis. Table III shows that the reliabilitycoefficients ranged from 0.8377 to 0.9245, indicating that some scaleswere more reliable than others. Generally, reliability coefficients of 0.70 ormore are considered good (Nunnally, 1967). Accordingly, the 11 scalesdeveloped for measuring TQM implementation constructs were considered tobe reliable.
Item analysisA method to evaluate the assignment of items to scales was developed byNunnally (1967). This method considers the correlation of each item with eachscale. Specifically, the item-score to scale-score correlations are used to
Table III.Internal consistency
analysis for the 11scales
Scales Number of items Deleted number Cronbach's alpha
1. Leadership 8 No 0.89222. Supplier quality management 6 No 0.83773. Vision and plan statement 8 No 0.91444. Evaluation 10 No 0.88955. Process control and
improvement8 1 0.8611
6. Product design 8 No 0.83937. Quality system improvement 5 No 0.92458. Employee participation 8 No 0.88299. Recognition and reward 6 No 0.8568
10. Education and training 6 No 0.884811. Customer focus 6 No 0.8747
IJQRM17,7
742
determine whether an item belongs to the scale as assigned, belongs to someother scales, or if it should be deleted. The scale-score is obtained by computingthe arithmetic average of the scores of the items that comprise that scale. If anitem does not correlate highly with any of the scales, it has to be eliminated.Saraph et al. (1989) used this method to evaluate the assignment of items toscales for developing their instrument. For this study, it was decided that itemanalysis should be performed in order to understand whether items had beenassigned appropriately.
Table IV lists the correlation matrix for the 11 scales and their measurementitems. The correlation matrix shows that the items correlated highly with thescales they intend to measure. All of the values in this table were greater than0.5. Item values lower than 0.5 do not share enough variance with the rest of theitems in that scale. It is therefore assumed that the items are not measuring thesame construct, and that it should be deleted from the scale (Kemp, 1999). As anexample item-score to scale-score correlation, item 1 in scale 1 had a highercorrelation with scale 1 than with the other scales (item correlations with otherscales were not given here due to text limitation). It was therefore concludedthat item 1 in scale 1 had been assigned appropriately to this scale. All otheritems were similarly examined and the results were satisfactory. It wastherefore concluded that all items had been appropriately assigned to scales.
ValidityValidity is defined as the extent to which any measuring instrument measureswhat it is intended to measure. There are three most popular methods toevaluate the validity of scales. These are content validity, criterion-relatedvalidity, and construct validity (Carmines and Zeller, 1979).
Content validity. Content validity depends on the extent to which anempirical measurement reflects a specific domain of content. It cannot be
Table IV.Item to scalecorrelation matrix(Pearson correlation)
Item NumberScales 1 2 3 4 5 6 7 8 9 10
1 0.784 0.810 0.709 0.702 0.851 0.624 0.751 0.816 ± ±2 0.741 0.787 0.816 0.690 0.683 0.753 ± ± ± ±3 0.844 0.861 0.553 0.861 0.859 0.863 0.652 0.823 ± ±4 0.741 0.738 0.768 0.703 0.656 0.772 0.769 0.627 0.653 0.6925 0.725 0.666 0.755 0.713 0.800 0.754 0.790 0.721 ± ±6 0.688 0.649 0.642 0.726 0.636 0.570 0.802 0.765 ± ±7 0.782 0.913 0.914 0.939 0.844 ± ± ± ± ±8 0.765 0.760 0.804 0.698 0.741 0.777 0.639 0.761 ± ±9 0.756 0.718 0.801 0.766 0.745 0.800 ± ± ± ±
10 0.828 0.863 0.846 0.680 0.703 0.854 ± ± ± ±11 0.813 0.737 0.838 0.849 0.649 0.813 ± ± ± ±Note: Item number in this Table is the same as the item number in the instrument.The symbol `̀ ±'' means not available.
An instrumentfor measuring
TQM
743
evaluated numerically ± it is a subjective measure of how appropriate the itemsseem to various reviewers who have some knowledge of the subject-matter.The evaluation of content validity typically involves an organised review of thesurvey's contents to ensure that it includes everything it should and does notinclude anything it should not. Strictly speaking, content validity is not ascientific measure of a survey instrument's accuracy. Nevertheless, it providesa solid foundation on which to build a methodologically rigorous assessment ofa survey instrument's validity. In this research, it was argued that the 11 scalesfor measuring TQM implementation constructs have content validity since thedevelopment of the measurement items was mainly based on an extensivereview of the literature and detailed evaluations by academicians andpractitioners.
The references list the literature that was reviewed by the authors during theperiod of conducting this study. The detailed process of developing theresearch instrument was already described more fully in the section onresearch methodology.
Criterion-related validity. Criterion-related validity, Nunnally (1978) notes:
is at issue when the purpose is to use an instrument to estimate some important form ofbehaviour that is external to the measuring instrument itself, the latter being referred to asthe criterion.
In this study, criterion-related validity was a measure of how well scalesrepresenting the various quality management practices are related to measuresof product quality performance (the criteria) (Flynn et al., 1994). Product qualityperformance was measured by asking respondents to rate (on a five-point scale,1 = worst in the industry; 2 = below average; 3 = average; 4 = above average;5 = best in the industry) seven indices compared with the other companieswithin the same industry in China (see the Appendix). These indices included,for example, performance, conformity rates, reliability, durability, defect ratesof their primary products, internal failure costs, and warranty claims costs as apercentage of their annual sales.
In this study, correlation analysis was employed for testing criterionvalidity. Accordingly, bivariate correlation (Pearson) was conducted to studythe interrelationships between the independent and dependent variable sets:the TQM implementation (predictor set) and the product quality performancemeasures (the criterion set). Their bivariate correlation coefficients are listed inTable V. Fortunately, the correlation within the 11 scales (predictor set), withinthe seven measures (criterion set), between the predictor set and criterion setwas significant at the 0.01 level. It can therefore be concluded that this set ofscales had good criterion-related validity.
Construct validity. Construct validity measures the extent to which the itemsin a scale all measure the same construct (Flynn et al., 1994). The constructvalidity can be evaluated by the use of factor analysis. Factor analysisaddresses the problem of analysing the interrelationships among a largenumber of variables and then explaining these variables in terms of their
IJQRM17,7
744
(a)
Wit
hin
pre
dic
tor
set
(TQ
Mim
ple
men
tati
onsc
ales
)S
cale
sM
ean
SD
12
34
56
78
910
11
1.L
ead
ersh
ip3.
900.
691.
000
2.S
up
pli
erq
ual
ity
man
agem
ent
3.68
0.67
0.74
91.
000
3.V
isio
nan
dp
lan
stat
emen
t3.
830.
740.
823
0.80
21.
000
4.E
val
uat
ion
3.68
0.65
0.75
70.
788
0.85
31.
000
5.P
roce
ssco
ntr
olan
dim
pro
vem
ent
3.57
0.67
0.71
70.
749
0.78
90.
822
1.00
06.
Pro
du
ctd
esig
n3.
640.
610.
716
0.64
40.
765
0.76
40.
775
1.00
07.
Qu
alit
ysy
stem
imp
rov
emen
t3.
730.
990.
655
0.70
30.
781
0.71
90.
745
0.66
01.
000
8.E
mp
loy
eep
arti
cip
atio
n3.
680.
650.
749
0.72
20.
809
0.81
30.
806
0.75
30.
717
1.00
09.
Rec
ogn
itio
nan
dre
war
d3.
700.
680.
650
0.70
40.
732
0.75
80.
739
0.66
50.
611
0.78
91.
000
10.
Ed
uca
tion
and
trai
nin
g3.
680.
700.
752
0.73
10.
810
0.78
00.
788
0.69
60.
739
0.82
40.
764
1.00
011
.C
ust
omer
focu
s4.
000.
640.
720
0.71
00.
789
0.75
50.
744
0.69
40.
710
0.80
30.
738
0.81
61.
000
(b)
Wit
hin
crit
erio
nse
t(p
rod
uct
qu
alit
ym
easu
res)
Pro
du
ctq
ual
ity
mea
sure
sM
ean
SD
12
34
56
7
1.T
he
per
form
ance
ofy
our
pri
mar
yp
rod
uct
s3.
970.
891.
000
2.T
he
con
form
ity
rate
sof
you
rp
rim
ary
pro
du
cts
3.99
0.88
0.83
81.
000
3.T
he
reli
abil
ity
ofy
our
pri
mar
yp
rod
uct
s4.
020.
800.
796
0.85
01.
000
4.T
he
du
rab
ilit
yof
you
rp
rim
ary
pro
du
cts
4.01
0.80
0.76
00.
793
0.85
61.
000
5.T
he
def
ect
rate
sof
you
rp
rim
ary
pro
du
cts
3.61
0.96
0.47
40.
575
0.51
00.
513
1.00
06.
Th
ein
tern
alfa
ilu
reco
sts
asa
per
cen
tag
eof
ann
ual
sale
s3.
570.
920.
548
0.60
30.
595
0.52
00.
791
1.00
0
7.T
he
war
ran
tycl
aim
sco
sts
asa
per
cen
tag
eof
ann
ual
sale
s3.
610.
940.
559
0.61
30.
599
0.57
60.
747
0.86
91.
000
(con
tinued
)
Table V.Bivariate correlationmatrices
An instrumentfor measuring
TQM
745(c
)B
etw
een
pre
dic
tor
set
and
crit
erio
nse
t
Pro
du
ctq
ual
ity
mea
sure
sA
ver
age
sev
enT
QM
imp
lem
enta
tion
scal
es1
23
45
67
mea
sure
s
1.L
ead
ersh
ip0.
631
0.61
10.
587
0.54
70.
390
0.40
90.
418
0.60
32.
Su
pp
lier
qu
alit
ym
anag
emen
t0.
524
0.56
10.
574
0.50
60.
346
0.41
90.
378
0.55
53.
Vis
ion
and
pla
nst
atem
ent
0.61
70.
645
0.63
40.
563
0.36
40.
460
0.42
90.
622
4.E
val
uat
ion
0.60
80.
628
0.61
20.
565
0.37
10.
432
0.39
30.
605
5.P
roce
ssco
ntr
olan
dim
pro
vem
ent
0.62
30.
646
0.63
00.
553
0.42
20.
523
0.46
70.
650
6.P
rod
uct
des
ign
0.53
60.
576
0.53
50.
500
0.37
10.
441
0.40
40.
566
7.Q
ual
ity
syst
emim
pro
vem
ent
0.54
40.
569
0.54
40.
508
0.30
10.
341
0.31
60.
522
8.E
mp
loy
eep
arti
cip
atio
n0.
547
0.60
50.
568
0.53
10.
351
0.43
40.
388
0.57
49.
Rec
ogn
itio
nan
dre
war
d0.
468
0.53
00.
538
0.46
30.
442
0.47
50.
401
0.56
010
.E
du
cati
onan
dtr
ain
ing
0.58
70.
598
0.57
50.
519
0.39
00.
456
0.40
90.
594
11.
Cu
stom
erfo
cus
0.57
20.
601
0.58
10.
509
0.38
40.
465
0.42
80.
595
Note
:P
ears
onco
rrel
atio
nis
sig
nif
ican
tat
the
0.01
lev
el;
SD
mea
ns
stan
dar
dd
evia
tion
;to
tal
nu
mb
eris
212
Table V.
IJQRM17,7
746
common underlying dimensions (factors). The general purpose of factoranalysis is to find a way of condensing or summarising the information intoa smaller set of new composite dimensions (factors) with a minimum lossof information (Hair et al., 1992). There are two forms of factor analysis,namely, exploratory factor analysis (EFA) and confirmatory factor analysis(CFA).
EFA is designed for a situation where links between the observed and latentvariables are unknown or uncertain. The analysis thus proceeds in anexploratory mode to determine how and to what extent the observed variablesare linked to their underlying factors (Byrne, 1998). Factor loadings are used topresent these relations. EFA helps to identify whether selected items cluster onone or more than one factor. The unidimensionality of factors is thus assessed.Usually, three or more items are selected for a latent variable or construct.However, the aim of CFA is to test or confirm a prespecified relationshipbetween indicators and latent variables. In respect of this study'scharacteristics, EFA would have to be employed for construct validation.Accordingly, principal component analysis was performed and each scale wasfactor analysed separately. The detailed results are listed in Table VI.
From this Table, it was very clear that all of the items had factor loadingsthat were greater than 0.50 on factor 1. Factor loadings greater than 0.30 areconsidered significant; loadings of 0.40 are considered more important; if theloadings are 0.50 or greater, they are considered very significant (Hair et al.,1992). In this study, a factor loading of 0.50 was used as the cut-off point. Whenthe items in a scale loaded on more than one factor, the rotated (varimax,quartimax if necessary) solution was examined. The factor analysis showedthat the items in nine of the 11 scales formed a single factor, except for scale 4(evaluation) and scale 5 (process control and improvement). In the cases of thescales for evaluation and process control and improvement, two factorsemerged according to the criterion of eigenvalues greater than 1. Since thelatent root criterion (eigenvalue) is the most commonly used technique forfactor extraction, it was therefore selected as criterion for factor extraction. Incomponent analysis, only the factors having eigenvalues greater than 1 areconsidered significant; all factors with eigenvalues less than 1 are consideredinsignificant and disregarded. Although percentage of variance and scree testcan also be used as criteria for factor extraction, the two criteria are, however,too subjective. For example, in the social sciences, where information is oftenless precise, it is not uncommon for the analyst to consider a solution thataccounts for 60 percent of the total variance (and in some instances even less)as a satisfactory solution (Hair et al., 1992).
Table VII lists the unrotated factor and rotated factor matrix for scale 4(evaluation). From Table VII, it was very clear that item 1, 2, 3, 4, 9, and 10constituted a factor that might be interpreted as the factor of audit. The otherfour items in scale 4 were in the other factor, which might form the factor of useof quality-related information.
An instrumentfor measuring
TQM
747
Table VI.Results of exploratoryfactor analysis for the
11 scales
Fac
tor
load
ing
s
Sca
les
Nu
mb
erof
fact
ors
Eig
env
alu
esIt
em1
Item
2It
em3
Item
4It
em5
Item
6It
em7
Item
8It
em9
Item
10%
ofv
aria
nce
Lea
der
ship
(sca
le1)
14.
620
0.78
40.
802
0.72
50.
677
0.83
90.
644
0.77
00.
816
57.7
50S
up
pli
erq
ual
ity
man
agem
ent
(sca
le2)
13.
394
0.74
80.
808
0.81
70.
696
0.66
50.
737
55.8
20V
isio
nan
dp
lan
stat
emen
t(s
cale
3)1
5.09
50.
848
0.86
10.
531
0.87
50.
873
0.87
30.
630
0.81
563
.692
Ev
alu
atio
n(s
cale
4)2
5.10
6(f
acto
r1)
0.73
70.
679
0.73
10.
751
0.70
10.
671
0.79
70.
785
0.64
70.
627
51.0
621.
009
(fac
tor
2)0.
223
0.17
00.
269
0.02
0±
0.09
4±
0.37
9±
0.29
4±
0.58
90.
273
0.44
410
.095
Pro
cess
con
trol
and
imp
rov
emen
t(s
cale
5)2
4.40
5(f
acto
r1)
0.72
70.
677
0.76
50.
725
0.79
30.
751
0.77
70.
714
55.0
601.
019
(fac
tor
2)0.
313
0.49
20.
395
0.13
8±
0.08
0±
0.40
8±
0.33
3±
0.46
912
.734
Pro
du
ctd
esig
n(s
cale
6)1
3.79
70.
709
0.66
50.
661
0.71
20.
637
0.57
20.
789
0.74
547
.469
Qu
alit
ysy
stem
imp
rov
emen
t(s
cale
7)1
3.87
80.
795
0.90
40.
902
0.94
00.
855
77.5
61E
mp
loy
eep
arti
cip
atio
n(s
cale
8)1
4.44
30.
754
0.74
00.
806
0.70
50.
754
0.78
10.
637
0.77
355
.539
Rec
ogn
itio
nan
dre
war
d(s
cale
9)1
3.51
40.
761
0.69
90.
759
0.77
60.
747
0.81
058
.572
Ed
uca
tion
and
trai
nin
g(s
cale
10)
13.
834
0.83
30.
858
0.84
40.
674
0.71
80.
848
63.8
95C
ust
omer
focu
s(s
cale
11)
13.
710
0.81
90.
748
0.82
50.
842
0.65
70.
813
61.8
83N
ote
:A
nei
gen
val
ue
gre
ater
than
1w
asu
sed
ascr
iter
ion
for
fact
orex
trac
tion
IJQRM17,7
748
Table VIII lists the unrotated factor and rotated (varimax and quartimax)factor matrix for scale 5 (process control and improvement). After orthogonaland oblique factor rotation was done, it was not easy to decide whether item 5belonged to factor 1 or factor 2. Item 5 loaded very significantly on both factor 1and factor 2. After the content of item 5 was examined, it was decided that item5 (our processes are designed to be `̀ fool proof'' in order to minimise the chanceof employee error) should be deleted from this scale. Thus, items 1, 2, 3, and 4formed a factor that might be interpreted as process control. Similarly, item 6,7, and 8 constituted a factor that might be interpreted as use of qualitymanagement methods.
According to Madu (1998), caution should be exercised in interpreting theresults obtained from statistical data analysis. There may be a temptation tooverstate the statistical findings. In fact, interpretations should be based on atotal view of research contents and sampling frame. In this study, two factorsemerged for scale 4 (evaluation) and scale 5 (process control and improvement)according to the statistical analysis. These results were obtained according tothe criterion of eigenvalues greater than 1. However, the two eigenvalues forfactor 2 were only 1.009 and 1.019, respectively. They were just slightly greater
Table VII.Factor matrix for scale4 (evaluation)
Unrotated factor Rotated (varimax) factorScale 4 1 2 1 2
Item 1 0.737 0.223 0.701 0.319Item 2 0.731 0.170 0.661 0.354Item 3 0.751 0.269 0.742 0.294Item 4 0.701 0.020 0.540 0.447Item 5 0.671 ±0.094 0.442 0.513Item 6 0.797 ±0.379 0.348 0.811Item 7 0.785 ±0.294 0.395 0.739Item 8 0.647 ±0.589 0.098 0.870Item 9 0.627 0.273 0.652 0.209Item 10 0.679 0.444 0.803 0.114
Unrotated factor Rotated factor(varimax)
Rotated factor(quartimax)
Scale 5 1 2 1 2 1 2
Item 1 0.727 0.313 0.295 0.734 0.753 0.245Item 2 0.677 0.492 0.134 0.826 0.833 0.077Item 3 0.765 0.395 0.264 0.820 0.836 0.208Item 4 0.725 0.138 0.417 0.609 0.636 0.375Item 5 0.793 ±0.080 0.619 0.503 0.544 0.583Item 6 0.751 ±0.408 0.820 0.240 0.295 0.802Item 7 0.777 ±0.333 0.786 0.311 0.364 0.763Item 8 0.714 ±0.469 0.837 0.171 0.228 0.823
Table VIII.Factor matrix for scale5 (process control andimprovement
An instrumentfor measuring
TQM
749
than 1. After the contents of items in scale 4 and 5 were carefully examined, itwas very clear that the aim of using quality-related information was to evaluatean organisation's quality management practices. Similarly, the aim of usingquality management methods was to control and improve processes. Therefore,scale 4 (evaluation) and scale 5 (process control and improvement) were notsplit into two scales, respectively. Note that item 5 in scale 5 was deletedand scale 5 consisted of 7 measurement items. Finally, it was concluded that the11 TQM implementation scales consisting of 78 items had good constructvalidity.
Discussion and conclusionsCompared to the other quality management instruments developed by Saraphet al. (1989), Flynn et al. (1994), and Ahire et al. (1996), the instrument presentedin this paper has the highest external validity for manufacturing industries ingeneral and for Chinese manufacturing companies in particular. The reason isthat the authors used data from 212 Chinese manufacturing companies in nineindustrial sectors for testing and validating this instrument. While its internalconsistency falls behind the Flynn et al. instrument and the Ahire et al.instrument, it is, however, better than the Saraph et al. instrument. Saraph et al.(1989) used data from 162 general and quality managers in 89 divisions of 20manufacturing and service companies. A main strength of Saraph et al.instrument is its highest level of external validity for manufacturing andservice industries (Ahire et al., 1996). Flynn et al. (1994) employed data from 716respondents at 42 plants in the transportation components, electronics andmachinery industries. The focus of their instrument is more on manufacturingindustries within limited sectors. Their instrument therefore has the secondhighest internal validity. Its external validity is, however, better than the Ahireet al. instrument. Ahire et al. (1996) utilised data from 371 manufacturingcompany in a single industry ± motor vehicle parts and accessories ± forvalidating their instrument. Their instrument therefore has the highest internalconsistency, but its external validity is the lowest. Ahire et al. (1996)summarised the differences and similarities of the three instruments in theirpaper.
In this study, data used for testing and validating this instrument only camefrom 212 manufacturing companies with annual sales of more than RMB10million in the Liaoning region. Strictly speaking, the generalisation is limited,although this study was the first one to aim at developing an instrument formeasuring TQM implementation for Chinese manufacturing companies. Inorder to improve external validity of the instrument, additional studies wouldbe needed with increased sample sizes, geographical diversity, company typediversity, and so on.
Although this instrument was empirically tested and validated using datafrom Chinese manufacturing companies, researchers and practitioners fromother countries will be able to use it. The reason is that this instrument was
IJQRM17,7
750
developed on the basis of an extensive literature review. However, it should benoted that this instrument is more valid for Chinese manufacturing companiesthan for companies in other countries.
In summary, the empirically validated TQM implementation instrumentconsisting of 11 scales (78 items) is reliable and valid. This validatedinstrument can be used directly in other studies for different populations.Industrial managers will be able to use this instrument to evaluate their TQMimplementation programs and identify problem areas that should be improved.Researchers will be able to use this instrument to develop quality managementtheory.
References
Ahire, S.L., Golhar, D.Y. and Waller, M.A. (1996), `̀ Development and validation of TQMimplementation constructs'', Decision Sciences, Vol. 27 No. 1, pp. 23-56.
Anderson, J.C. and Schroeder, R.G. (1994), `̀ A theory of quality management underlying theDeming management method'', Academy of Management Review, Vol. 19 No. 3, pp. 472-509.
Anderson, J.C., Rungtusanatham, M., Schroeder, R. and Devaraj, S. (1995), `̀ A path analytic modelof a theory of quality management underlying the Deming management method:preliminary empirical findings'', Decision Sciences, Vol. 26 No. 5, pp. 637-58.
Black, S.A. and Porter, L.J. (1996), `̀ Identification of the critical factors of TQM'', DecisionSciences, Vol. 27 No. 1, pp. 1-21.
Blauw, J.N. and During, W.E. (1990), `̀ Total quality control in Dutch industry'', Quality Progress,February, pp. 50-2.
Brown, M.G., Hitchcock, D.E. and Willard, M.L. (1994), Why TQM Fails and What to Do about It,Irwin, Burr Ridge, IL.
Byrne, B.M. (1998), Structural Equation Modeling with LISREL, PRELIS, and SIMPLIS, BasicConcepts, Applications, and Programming, Lawrence Erlbaum Associates, Inc.,Publishers, Mahwah, NJ.
Carmines, E.G. and Zeller, R.A. (1979), Reliability and Validity Assessment, Sage Publications,Beverly Hills, CA.
Choi, T.Y. and Eboch, K. (1998), `̀ The TQM paradox: relations among TQM practices, plantperformance, and customer satisfaction'', Journal of Operations Management, Vol. 17 No.1, pp. 59-75.
Cronbach, L.J. (1951), `̀ Coefficient alpha and the internal structure of tests'', Psychometrika, Vol.16 No. 3, pp. 297-334.
Crosby, P.B. (1979), Quality Is Free, McGraw-Hill, New York, NY.
Daetz, D., Barnard, B. and Norman, R. (1995), Customer Integration: The Quality FunctionDeployment (QFD) Leader's Guide for Decision Making, Wiley Corp., New York, NY.
Dale, B.G. and Plunkett, J.J. (1990), Managing Quality, Philip Allan, New York, NY.
Dean, J.W. Jr and Bowen, D.E. (1994), `̀ Management theory and total quality: improving researchand practice through theory development'', Academy of Management Review, Vol. 19 No. 3,pp. 392-418.
Deming, W.E. (1986), Out of Crisis, Massachusetts Institute of Technology, Center for AdvancedEngineering Study, Cambridge, MA.
Deming Prize (1992), Guide for Overseas Companies, Union of Japanese Scientists and Engineers,Tokyo, Japan.
An instrumentfor measuring
TQM
751
European Quality Award (1994), Self-Assessment Based on the European Model for Total QualityManagement: Guidelines for Identifying and Addressing Business Excellence Issues,European Foundation for Quality Management, Brussels, Belgium.
Feigenbaum, A.V. (1991), Total Quality Control, 3rd ed., McGraw-Hill, New York, NY.
Flynn, B.B., Schroeder, R.C. and Sakakibara, S. (1994), `̀ A framework for quality managementresearch and an associated measurement instrument'', Journal of Operations Management,Vol. 11, pp. 339-66.
Forza, C. and Filippini, R. (1998), `̀ TQM impact on quality conformance and customersatisfaction: a causal model'', International Journal of Production Economics, Vol. 55 No. 1,pp. 1-20.
Garvin, D.A. (1983), `̀ Quality on the line'', Harvard Business Review, Vol. 61, pp. 64-75.
Garvin, D.A. (1986), `̀ Quality problems, policies and attitudes in the United States and Japan: anexploratory study'', Academy of Management Journal, Vol. 29, pp. 653-73.
Hackman, J.R. and Wageman, R. (1995), `̀ Total quality management: empirical, conceptual, andpractical issues'', Administrative Science Quarterly, Vol. 40, June, pp. 309-42.
Hair, J.F. Jr, Anderson, R.E., Tatham, R.L. and Black, W.C. (1992), Multivariate Data Analysis withReadings, 3rd ed., Macmillan Publishing Company, New York, NY.
Ishikawa, K. (1985), What Is Total Quality Control? The Japanese Way, Prentice-Hall, London.
ISO 8402 (1994), Quality Management and Quality Assurance ± Vocabulary, InternationalOrganisation for Standardisation.
Juran, J.M. and Gryna, F.M. (1993), Quality Planning and Analysis, 3rd ed., McGraw-Hill, NewYork, NY.
Kanji, G.K. and Asher, M. (1993), Total Quality Management Process: A Systematic Approach,Advances in Total Quality Management Series, Carfax, Abingdon.
Kemp, R.G.M. (1999), Managing Interdependence for Joint Venture Success: An Empirical Studyof Dutch International Joint Ventures, PhD thesis, Faculty of Management andOrganisation, University of Groningen, The Netherlands.
Madu, C.N. (1998), `̀ An empirical assessment of quality: research consideration'', InternationalJournal of Quality Science, Vol. 3 No. 4, pp. 348-55.
Malcolm Baldrige National Quality Award (1997), Criteria for Performance Excellence, NationalInstitute of Standards and Technology, United States Department of Commerce,Gaithersburg, MA.
Mann, R.S. (1992), The Development of a Framework to Assist in the Implementation of TQM,PhD thesis, Department of Industrial Studies, University of Liverpool, UK.
Mann, R. and Kehoe, D. (1994), `̀ An evaluation of the effects of quality improvement activities onbusiness performance'', International Journal of Quality & Reliability Management, Vol. 11No. 4, pp. 29-44.
Nunnally, J. (1967), Psychometric Theory, McGraw-Hill, New York, NY.
Nunnally, J. (1978), Psychometric Theory, McGraw-Hill, New York, NY.
Powell, T.C. (1995), `̀ Total quality management as competitive advantage: a review andempirical study'', Strategic Management Journal, Vol. 16, pp. 15-37.
Reeves, V.A. and Bednar, D.A. (1994), `̀ Defining quality: alternatives and implications'', Academyof Management Review, Vol. 19 No. 3, pp. 419-45.
Saraph, J.V., Benson, G.P. and Schroeder, R.G. (1989), `̀ An instrument for measuring the criticalfactors of quality management'', Decision Sciences, Vol. 20, pp. 810-29.
Waldman, D.A. (1994), `̀ The contribution of total quality management to a theory of workperformance'', Academy of Management Review, Vol. 19 No. 3, pp. 510-36.
IJQRM17,7
752
Zhang, Z.H. (1998a), `̀ State supervision and inspection of product quality in China'', QualityProgress, Vol. 31 No. 12, pp. 53-7.
Zhang, Z.H. (1998b), `̀ Application of experimental design in new product development'', TQMMagazine, Vol. 10 No. 6, pp. 432-7.
Zhao, X.D., Young, S. and Zhang, J.C. (1995), `̀ A survey of quality issues among Chineseexecutives and workers'', Production and Inventory Management, Vol. 36 No. 1, pp. 44-8.
AppendixTQM implementation instrumentThe five-point likert scale (1 = strongly disagree; 2 = disagree; 3 = neutral; 4 = agree; 5 =strongly agree) was used to ask the respondents to state how much they agree with thesestatements.
Scale 1: leadership
(1) Top management actively participates in quality management and improvementprocess.
(2) Top management learns quality-related concepts and skills.
(3) Top management strongly encourages employee involvement in quality managementand improvement activities.
(4) Top management empowers employees to solve quality problems.
(5) Top management arranges adequate resources for employee education and training.
(6) Top management discusses many quality-related issues in top management meetings.
(7) Top management focuses on product quality rather than yields.
(8) Top management pursues long-term business success.
Scale 2: supplier quality management
(1) Our company has established long-term co-operative relations with suppliers.
(2) Our company regards product quality as the most important factor for selectingsuppliers.
(3) Our company always participates in supplier activities related to quality.
(4) Our company always gives feedback on the performance of suppliers' products.
(5) Our company has detailed information about supplier performance.
(6) Our company regularly conducts supplier quality audit.
Scale 3: vision and plan statement
(1) Our company has a clear long-term vision statement.
(2) The vision effectively encourages employees' commitment to quality improvement.
(3) Our company has a clear short-term business plan.
(4) Our company has a clear quality policy.
(5) Our company has a detailed quality goal.
(6) Our company has an effective quality improvement plan.
(7) Various policies and plans are well communicated to the employees.
(8) Employees from different levels are involved in making policies and plans.
An instrumentfor measuring
TQM
753
Scale 4: evaluation
(1) Our company regularly audits various policies and strategies.
(2) Our company regularly conducts quality audits.
(3) Benchmarking is used extensively in our company.
(4) Our company uses quality-related costs extensively.
(5) Our company has detailed quality-related data such as defect rates and scrapsa.
(6) Quality-related data are used to evaluate the management of our companya.
(7) Quality-related data are used to evaluate the performance of all departmentsa.
(8) Quality-related data are used to evaluate the performance of employeesa.
(9) Quality-related information is displayed on the shopfloor.
(10) The aim of evaluation is for improvement, not for criticism.
Scale 5: process control and improvement
(1) Our company is kept neat and clean at all times.
(2) Process capability can meet production requirements.
(3) Production equipment is maintained well according to maintenance plan.
(4) Our company implements various inspections effectively (e.g., incoming, process, finalproducts).
(5) Our processes are designed to be `̀ fool proof'' in order to minimise the chance ofemployee errorb.
(6) Our company uses the Seven QC tools extensively for process control andimprovementa.
(7) Our company uses SPC extensively for process control and improvementa.
(8) Our company uses PDCA cycle extensively for process control and improvementa.
Scale 6: product design
(1) The design engineers are required to have some shopfloor experiences.
(2) The design engineers are required to have some marketing experiences.
(3) The customer requirements are thoroughly considered in new product design.
(4) Various departments participate in new product development.
(5) New product designs are thoroughly reviewed before production.
(6) Cost is emphasised in the product design process.
(7) Experimental design is used extensively in product design.
(8) Quality function deployment (QFD) is used extensively in product design.
Scale 7: quality system improvement
(1) The quality system in our company is continuously improved.
(2) Our company uses ISO 9000 as a guideline for establishing our quality system.
(3) Our company has a clear quality manual.
(4) Our company has clear procedure documents.
(5) Our company has clear working instructions.
IJQRM17,7
754
Scale 8: employee participation
(1) Our company has cross-functional teams.
(2) Our company has several QC circles (within one function).
(3) Employees are actively involved in quality-related activities.
(4) Our company implements suggestion activities extensively.
(5) Most employees' suggestions are implemented after an evaluation.
(6) Employees are very committed to the success of our company.
(7) Employees are encouraged to fix problems they find.
(8) Reporting work problems is encouraged in our company.
Scale 9: recognition and reward
(1) Our company improves working conditions in order to recognise employee qualityimprovement efforts.
(2) Our company has a salary promotion scheme for encouraging employee participation inquality improvement.
(3) Position promotions are based on work quality in our company.
(4) Excellent suggestions are financially rewarded.
(5) Employees' rewards and penalties are clear.
(6) Recognition and reward activities effectively stimulate employee commitment to qualityimprovement.
Scale 10: education and training
(1) Employees are encouraged to accept education and training in our company.
(2) Resources are available for employee education and training in our company.
(3) Most employees in our company are trained on how to use quality management methods(tools).
(4) Quality awareness education is given to employees.
(5) Specific work-skills training is given to all employees.
(6) Employees are regarded as valuable, long-term resources worthy of receiving educationand training throughout their career.
Scale 11: customer focus
(1) Our company collects extensive complaint information from customers.
(2) Quality-related customer complaints are treated with top priority.
(3) Our company conducts a customer satisfaction survey every year.
(4) Our company always conducts market research in order to collect suggestions forimproving our products.
(5) Our company provides warranty on our sold products to customers.
(6) Our company has been customer focused for a long time.
Note: a means that these items in that scale formed a factor and the other items in that scaleformed another factor.b means that this item was deleted after factor analysis.
An instrumentfor measuring
TQM
755
Measurement for product qualityCompared with the other companies within the same industry in China, please state the situationof your primary products (1 = worst in the industry; 2 = below average; 3 = average; 4 = aboveaverage; 5 = best in the industry):
(1) The performance of your primary products
(2) The conformity rates of your primary products
(3) The reliability of your primary products
(4) The durability of your primary products
(5) The defect rates of your primary products
(6) The internal failure costs as a percentage of annual sales
(7) The warranty claims costs as a percentage of annual sales