This art icle was downloaded by: [ Tianj in Universit y] On: 02 Novem ber 2011, At : 01: 06 Publisher: Rout ledge I nform a Lt d Regist ered in England and Wales Regist ered Num ber: 1072954 Regist ered office: Mort im er House, 37- 41 Mort im er St reet , London W1T 3JH, UK Total Quality Management & Business Excellence Publicat ion det ails, including inst ruct ions f or aut hors and subscript ion inf ormat ion: ht t p: / / www. t andf online. com/ loi/ ct qm20 Validation of the theoretical model underlying the Baldrige criteria: Evidence from China Zhen He a , James Hill b , Ping Wang c & Gang Yue a a Depart ment of Indust rial Engineering, School of Management , Tianj in Universit y, Tianj in, PR China b Depart ment of Management Science, Fisher College of Business, The Ohio St at e Universit y, Columbus, OH, USA c Depart ment of Logist ics Engineering and SCM, School of Management , Tianj in Universit y, Tianj in, PR China Available online: 19 Feb 2011 To cite this article: Zhen He, James Hill, Ping Wang & Gang Yue (2011): Validat ion of t he t heoret ical model underlying t he Baldrige crit eria: Evidence f rom China, Tot al Qualit y Management & Business Excellence, 22: 2, 243-263 To link to this article: ht t p: / / dx. doi. org/ 10. 1080/ 14783363. 2010. 545562 PLEASE SCROLL DOWN FOR ARTI CLE Full t erm s and condit ions of use: ht t p: / / www.t andfonline.com / page/ t erm s- andcondit ions This art icle m ay be used for research, t eaching, and privat e st udy purposes. Any subst ant ial or syst em at ic reproduct ion, redist ribut ion, reselling, loan, sub- licensing, syst em at ic supply, or dist ribut ion in any form t o anyone is expressly forbidden. The publisher does not give any warrant y express or im plied or m ake any represent at ion t hat t he cont ent s will be com plet e or accurat e or up t o dat e. The accuracy of any inst ruct ions, form ulae, and drug doses should be independent ly verified wit h prim ary sources. The publisher shall not be liable for any loss, act ions, claim s, proceedings, dem and, or cost s or dam ages what soever or howsoever caused arising direct ly or indirect ly in connect ion wit h or arising out of t he use of t his m at erial. Total Quality Management Vol. 22, No. 2, February, 2011, 243 –263 Validation of the theoretical model underlying the Baldrige criteria: Evidence from China Zhen Hea, James Hillb, Ping Wangc∗ and Gang Yuea Downloaded by [Tianjin University] at 01:06 02 November 2011 a Department of Industrial Engineering, School of Management, Tianjin University, Tianjin, PR China; bDepartment of Management Science, Fisher College of Business, The Ohio State University, Columbus, OH, USA; cDepartment of Logistics Engineering and SCM, School of Management, Tianjin University, Tianjin, PR China The primary objectives of this study are (1) to develop a measurement model tied directly to the Baldrige criteria at both construct and dimension levels, and (2) to validate the theoretical model underlying the Baldrige framework based on evidence from China. Using both exploratory factor analysis and confirmatory factor analysis on data collected from 2302 manufacturing and service firms in China, authors find that the proposed theoretical model with 19 hypotheses are statistically supported at ,0.05 level. Contributions of this study are three-fold: (1) the newly developed measurement model provides a neat and reliable reference for both practitioners and researchers; (2) the theoretical model developed was statistically supported regarding the overall model fit indices and hypothesised linkages, demonstrating the robustness of the Baldrige framework; and (3) based on evidence from China, process management is the most important construct in the Baldrige framework, followed by leadership, which shows a slight departure from previous studies on industrialised countries. Keywords: Baldrige; quality management; China; process management 1. Introduction China has become the fourth largest manufacturing country in the world. The nation has experienced an annual GDP growth of approximately 9.75% between 1999 and 2008 (USChina Business Council), and their foreign trade growth is approximately 15% per year over the same period. However, the growth in manufacturing has brought many challenges. There have been numerous recalls in 2007, such as 450,000 Chinese-made light-truck tyres, 1.5 million toys manufactured with lead paint, 1.2 million space heaters with electrical cords that could overheat, and 97,000 halogen table lamps due to fire and burn hazards (Vardaman, 2007). These recalls have understandably become a concern for consumers worldwide and companies that outsource or manufacture products within China. The US-China Business Council (USCBC) has spoken with government officials in Washington DC and Beijing about the importance of moving quickly to maintain public confidence in the quality of products made in China. Quality problems may arise from a number of sources including: product design, quality programmes, policies and attitudes, vendor management practices, and production training and work force processes (Garvin, 1991). There is little evidence as to whether Chinese companies face these same sets of sources when addressing quality issues as do American firms. These are important issues as they have a direct bearing on the policies ∗ Corresponding author. Email: pwang607@tju.edu.cn ISSN 1478-3363 print/ISSN 1478-3371 online # 2011 Taylor & Francis DOI: 10.1080/14783363.2010.545562 http://www.informaworld.com Downloaded by [Tianjin University] at 01:06 02 November 2011 244 Z. He et al. companies should adopt when operating in China, and on the relevance of Chinese approaches to quality management. It has been argued in the literature that a necessary condition for successful quality performance is a management dedicated to improving quality (Sousa & Voss, 2002). Research on management’s role in improving product quality in China is sparse. The Chinese government recognised the importance of quality and started several national initiatives that focus on quality management. For instance, in August 2007, the Chinese government launched a nationwide campaign to improve product quality and safety in areas including agricultural products, food processing, catering, medicine, imports and exports, and other consumer goods involving human health and safety. Accompanied by a nationwide campaign on improving quality, the National Development and Reform Commission (NDRC) commissioned the China Association for Quality (CAQ) to conduct a nationwide survey on quality management practices in China from August 2007 to October 2007. The primary purpose of the survey was to identify critical factors that affect quality management practices in China, and to make necessary suggestions for improvement. Parts of this survey questionnaire were questions related to the criteria of the Malcolm Baldrige National Quality Award (MBNQA) which originated in the United States. The MBNQA has evolved from a means of recognising and promoting exemplary quality management practices to a comprehensive framework for world class performance. The use of the criteria has been regarded as a source of information on achieving business excellence (Bemowski & Stratton, 1995). Since its inception, the Baldrige award has established quite a legacy. There are several international awards that have been modelled on the Baldrige award including the European Quality Award, the Mexican National Quality Award, the Brazilian National Award, the Egyptian National Award, Japanese National Award, and the Chinese National Award, to name a few (Flynn & Saladin, 2001). The Chinese National Quality Award was primarily developed on the basis of the 1997 Baldrige criteria and was used to conduct nationwide surveys in 2001 and 2007. We found no previous studies that validated the theoretical model underlying the Baldrige framework with regard to evidence from China. The primary objectives of this study are both exploratory and confirmatory. For the exploratory purpose, we develop a comprehensive measurement model, with associated constructs and scales that accurately capture the content of the Baldrige criteria. For the confirmatory purpose, we validate whether the causational links derived from the data represent the theoretical model of the Baldridge framework. The results from this study will contribute to the emerging literature on quality management in China. To the best of our knowledge, this is the first attempt to develop a ‘measurement model’ tied directly to the Baldrige criteria at the dimension level, and to use evidence from China to validate the theoretical model of the Baldrige framework. The rest of the paper is organised as follows. The next section will provide a brief literature review on the validations of the Baldrige framework, and present the hypotheses to be tested. The third section includes a discussion of the research design and methodology. Results and discussion are provided in the fourth section. Finally, conclusions, limitations, and future research directions are discussed in the last section. 2. Literature review and hypotheses 2.1. Baldrige framework Flynn and Saladin (2001) reviewed the history of previous Baldrige frameworks (1988, 1992, and 1997). Our paper is based on the 2006 framework that has no significant Downloaded by [Tianjin University] at 01:06 02 November 2011 Total Quality Management 245 modifications from the 1997 framework. There are seven categories underlying the Baldrige criteria, as shown in Figure 1, including leadership, strategic planning, customer and market focus, measurement, analysis, and knowledge management (MAKM), human resource focus, process management, and results. MAKM was changed from its previous name, information and analysis in the 1997 Baldrige model, by adding a new dimension, measurement, analysis, and review of organisational performance, to the original construct. MAKM serves ‘as a foundation for the performance management system’, and is ‘critical to the effective management of an organisation’ (Criteria for Performance Excellence, 2006, p.6). Each category has two or three items, except for the results category that has six items. In total the Baldrige framework has 19 items, as shown in the Appendix. Throughout the rest of the paper, we adapt the terminology used in Wilson and Collier (2000) by referring to each category as a construct, and questions used to measure the items as scales. We use the term item and dimension interchangeably throughout this paper. The Baldrige ‘system’ is composed of six constructs located in the centre of Figure 1 that define the operations and results; the six constructs can be viewed as two triads, the leadership triad and the results triad (Criteria for Performance Excellence, 2006, p. 5). The leadership triad constitutes leadership, strategic planning, and customer and market focus, and the results triad includes human resource focus, process management, and results. 2.2. Validations of theoretical models underlying the Baldrige framework The Baldrige award provides a well-accepted framework for operationalising the constructs of quality management. Numerous studies have been conducted in different industries or in different countries or regions, in order to develop reliable measurement scales and to validate theoretical associations among constructs (Evans & Jack, 2003; Flynn, Schroeder, & Sakakibara, 1994; Flynn & Saladin, 2001; Handfield & Ghosh, 1995; Lau, Zhao, & Xiao, 2004; Meyer & Collier, 2001; Su, Li, & Su 2003; Wilson & Collier, 2000). Lau et al. (2004) is the first study to apply the Baldrige criteria to evaluate the quality practices and performances of firms in China. Their research is based on the Figure 1. 2006 Baldrige framework. Downloaded by [Tianjin University] at 01:06 02 November 2011 246 Z. He et al. Figure 2. Flynn and Saladin’s (2001) path model. 2001 national survey on quality management practices in China. In their paper they did not directly test the theoretical relationships underlying the Baldrige framework. Flynn and Saladin (2001) used path analysis to validate the causations in the Baldrige framework, on data selected from the World Class Manufacturing (WCM) database, Round II. They proposed that the theoretical relationships in the 1997 Baldrige framework could be expressed as Figure 2, with 15 causal links (hypotheses) to be tested. They showed that 12 relationships in the figure hold, except for three links, the link from strategic planning to customer and market focus, the link from information and analysis to business results, and the link from human resource (HR) development and management to business results. Su et al. (2003) developed 20 hypotheses in testing the theoretical causal relationships underlying the Taiwan National Quality Award (TNQA) as shown in Figure 3. TNQA was developed based on the 2000 Baldrige framework. They concluded that three hypotheses did not hold; (1) from customer and market development to process management; (2) from customer and market development to business results; and (3) from human resource and knowledge management to business results. Figure 3. Su et al.’s (2003) path model. Total Quality Management 247 2.3. Hypotheses 2.3.1 Hypotheses relevant to leadership Downloaded by [Tianjin University] at 01:06 02 November 2011 Based on previous literature, we reason that the theoretical model underlying the Baldrige criteria is recursive, containing no reciprocal causations or feedback loops. We agree with the viewpoint that leadership drives the system that creates results. Therefore, we propose that leadership has a positive impact on the five constructs within the ‘system’ that influences results. In addition, empirical evidence shows that leadership has a direct and positive impact on results (e.g. Flynn & Saladin, 2001), which leads to our sixth hypothesis. The six hypotheses are described below: H1: Leadership H2: Leadership H3: Leadership H4: Leadership H5: Leadership H6: Leadership 2.3.2. has a has a has a has a has a has a positive influence positive influence positive influence positive influence positive influence positive influence on strategic planning. on MAKM. on customer and market focus. on human resource focus. on process management. on results. Hypotheses relevant to measurement analysis and knowledge management In validating previous theoretical Baldrige models, many researchers have found that information and analysis is the second most important construct in the Baldrige framework (e.g. Flynn & Saladin, 2001; Su et al., 2003; Wilson & Collier, 2000). Both empirical and theoretical evidence support the propositions that information and analysis has a positive impact on the other five constructs in the Baldrige framework. As noted earlier, information and analysis has been changed to MAKM in the 2006 Baldrige model. The five hypotheses are given below: H7: MAKM has a positive influence on strategic planning. H8: MAKM has a positive influence on customer and market focus. H9: MAKM has a positive influence on human resource focus. H10: MAKM has a positive influence on process management. H11: MAKM has a positive influence on results. 2.3.3. Hypotheses relevant to strategic planning and customer and market focus Strategic planning and customer and market focus, the two constructs in the leadership triad, are presumed to influence each other in the 2006 Baldrige framework. To be consistent with assumptions and the logic of quality management in the literature, the bi-directional arrow between the two constructs is redrawn as unidirectional, pointing from strategic planning to customer and market focus (Flynn & Saladin, 2001; Pannirselvam & Ferguson, 2001; Wilson & Collier, 2000). Besides, these two constructs are presumed to influence human resource focus and process management in the results triad. The five proposed hypotheses are described as: H12: H13: H14: H15: H16: 2.3.4 Strategic planning has a positive influence on customer and market focus. Strategic planning has a positive influence on human resource focus. Strategic planning has a positive influence on process management. Customer and market focus has a positive influence on human resource focus. Customer and market focus has a positive influence on process management. Hypotheses relevant to human resource focus and process management In regards to human resource and process management we bring forth the same argument we made when discussing the causal relationship between strategic planning and customer 248 Z. He et al. and market focus. We propose that the relationship between human resource focus and process management are unidirectional. As outlined in the 2006 Baldrige framework we propose that both constructs have a positive impact on results. These three hypothesised causations are given below, and all 19 hypotheses are illustrated in a theoretical model based on the 2006 Baldrige framework, as shown in Figure 3. H17: Human resource focus has a positive influence on process management. H18: Human resource focus has a positive influence on results. H19: Process management has a positive influence on results. 3. Downloaded by [Tianjin University] at 01:06 02 November 2011 3.1 Research design and methodology Data collection and descriptive statistics 3.1.1 Survey development This research uses data collected by CAQ for the Third National Quality Management Practices Survey in 2007. The survey questionnaire was developed by experts from CAQ and scholars from Fudan University and The Chinese University of Hong Kong. The survey instrument has 148 questions, with 91 questions relevant to the Baldrige criteria. These questions of the Baldrige criteria were measured using six-point Likert scales (see the Appendix for survey questions used in this study). The basic establishment of the sampling frame was constructed using the results of the Second National Reconnaissance Survey in China carried out by the State Bureau of Statistics in 2001. According to this survey, 34 provinces and regions in China are divided into three groups: highly developed, moderately developed, and less developed. The sample frame covers highly developed regions (Beijing, Shanghai, Tianjin, Hong Kong, and Macau), moderately developed provinces and regions (Chongqing, Zhejiang, Jiangsu, Guangdong, Liaoning, Hubei, Fujian, and Shandong) and less developed provinces (Yunnan, Hebei, and Jiangxi). These 16 provinces and regions collectively represent about 70% of GDP and 75% of export from China (2005 Yearbook of the State Bureau of Statistics in China). Due to the budget constraint and time pressure, the sample size was set at 3000 firms. 3.1.2 Main sample response rate and test for bias Target firms in the Third National Quality Management Practices Survey were limited to those with annual revenues of $700,000 (five million Renminbi) or above. Within this sample frame there are about 100,000 manufacturing and service firms that meet this criterion, which are typically considered as medium to large companies in China. These 100,000 firms are classified into 48 stratifies from three aspects: sectors (manufacturing, service), size ($700, 000– 1.4 million, 1.4 – 7 million, 7 – 14 million, and 14 million and above), and ownership (state-owned, collective, private, conglomerate, incorporated, and foreign direct investment). The quota of 3000 firms was proportionally allocated in the 48 stratifies with regards to the number of firms in each stratify. Finally, the probability proportional to size sampling method is adopted to select firms in each stratify. Branches of CAQ in each selected province or region are in charge of questionnaire dissemination and collection. Printed copies of the questionnaire were sent to target firms whose contact information was maintained in the database of the State Bureau of Statistics. Thanks to the influence of, and efforts made by CAQ, 2673 questionnaires were returned within three months, at an initial response rate of 89%. Total Quality Management 249 Common method bias was assessed by using Harman’s single-factor test (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). We loaded all the variables in our study into an exploratory factor analysis and examined the unrotated factor solution. The results demonstrated that neither a single factor emerged nor did one general factor account for the majority of the covariance among the measures. Downloaded by [Tianjin University] at 01:06 02 November 2011 3.1.3 Description of the sample Prior to furthering our analysis, the data characteristics were screened regarding the 91 questions relevant to the Baldrige criteria. The initial data with 2673 responses demonstrated moderate departures from normality in that scores of some indicators were skewed to larger values (the number of six in our study), and revealed the existence of outliers. Replications (more than one response from the same organisation) and incomplete responses were deleted, resulting in the sample size being reduced to 2302 that represents a response rate of 77%. Re-examination and visual plots of the final data showed no outliers and no departures from normality. The Kaiser – Meyer –Olkin (KMO) measure of sampling was used to ensure that our data supports the use of factor analysis. The KMO indicates whether or not the variables are able to be grouped into a smaller set of underlying factors. The overall KMO score for the theoretical model was 0.992, supporting the use of factor analysis (Hill, Eckerd, Wilson, & Greer, 2008). Bartlett’s test of sphericity tests whether the correlation matrix is an identity matrix, which would indicate that the factor model is inappropriate. The result of the Bartlett test has a significance level of p , 0.0001, which means the correlation matrix is not an identity matrix (Hair, Anderson, Tatham, & Black, 1998). After screening the data, the profile of the respondents includes senior managers (547, 23.8%), mid-level managers (1202, 52.2%), and frontline employees (507, 22%), with 10, 8.6, and 4 years of experience in quality management on average, respectively. The final sample has 1474 manufacturing firms and 828 service firms, as shown in Table 1, and more than half of the responding firms have annual revenues less than $1.4 million (10 million Renminbi). 3.1.4 The measurement model The measurement model was purified by using a systematic, iterative process similar to the one proposed by Anderson and Gerbing (1988). Scale elimination was based on weak loadings, cross loadings, communalities, error residuals and theoretical determination (Prahinski & Benton, 2004). Of the 91 initial scales, 68 scales were retained after the measurement purification process.1 Scales retained were statistically significant indicators of their respective constructs and dimensions. The sub-dimension, measurement, analysis, and review of organisational performance under the fourth construct, was eliminated since all scales under this sub-dimension were eventually deleted. The fit indices in Table 2 indicated acceptable fit for the purified measurement model. 3.2 Reliability and validity of the model Content validity is based on the extent to which a measurement reflects the specific intended domain (Carmines & Zeller, 1991, p. 20). The content validity of the instrument was initially established before data collection, by grounding it in extant literature, refining the questionnaire used in 2001, discussing the instrument with industry experts, and 250 Z. He et al. Table 1. Description of the sample firms by industry. Downloaded by [Tianjin University] at 01:06 02 November 2011 (%) of firms in this study Manufacturing Agricultural and wooden products, tobacco Apparel, leather and textile products Chemicals, rubber and plastic Construction and SCGC Electrical products Equipments and instruments Food and kindred products Machinery Petroleum, coal, and other mining Primary metals and fabricated metal products Vehicles and parts Other manufacturing Total Service Business service (consulting, inspection, IT, sofeware, marketing, tech services) Financial services Hotel, tourist, food services, and entertainment Logistics Merchandise (auto dealer/repair, part sales, gas stations) Newspaper and allied products/printing/publishing Wholesaling and (international) trading Real estate Public services (government, health care, utility, communication, education) Retailing Other services Total Firm size Mean annual sales Median annual sales Range of annual sales 56 (2.43%) 176 (7.65%) 111 (4.82%) 106 (4.60%) 175 (7.60%) 96 (4.17%) 94 (4.08%) 136 (5.92%) 35 (1.52%) 123 (5.34%) 63 (2.74%) 303 (13.16%) 1474 (64.0%) 82 (3.56%) 47 (2.04%) 116 (5.04%) 48 (2.09%) 62 (2.68%) 27 (3.48%) 80 (3.48%) 38 (1.65%) 58 (2.52%) 84 (3.65%) 186 (8.08%) 828 (36%) RMB 82 million RMB 8 million RMB 5 –12,276 million Note: SCGC – stone, clay, glass and concrete products. conducting a panel discussion with experts from CAQ and scholars from Fudan University and The Chinese University of Hong Kong. Construct validity seeks agreement between the theoretical model and the survey instrument (Bagozzi, Youjae, & Philips, 1991). Unidimensionality was established using exploratory factor analysis (EFA) and confirmatory factor analysis (CFA). Un-rotated and rotated EFAs with the eigen-value-greater-than-one criterion were conducted on datasets with both 91 and 68 scales. Rotated EFA on the imputed dataset with 91 scales revealed that seven distinct factors account for 68.85% of the total variance in the data. Rotated EFA on the imputed dataset with 68 scales showed that the seven factors account for 70.08% of the total variance.This change corroborates the robustness of our measurement purification process. In order to overcome the bias associated with any single measure it is suggested that a variety of fit measures be evaluated (Shah & Goldstein, 2006). Cronbach’s alphas were calculated to determine the reliability of the constructs. As shown in the Appendix, all Total Quality Management 251 Table 2. Measures of model fit and statistical power. Desirable range 2 Downloaded by [Tianjin University] at 01:06 02 November 2011 x test statistic, p ¼ 0.000 x 2/d.f. GFI NFI RFI CFI RMSEA RMSEA 90% confidence interval Degree of freedom Effective number of parameters Hoetler’s N (at a ¼ 0.05, 0.01) Power at RMSEA0 ¼ 0.01 and alternate RMSEAa ¼ 0.05 ≤3.0 ≥0.90 ≥0.90 ≥0.90 ≥0.90 ≤0.08 Proposed model 7689 3.63 0.90 0.95 0.93 0.96 0.034 0.033, 0.035 2119 227 667, 681 .0.99 Cronbach’s alphas were above 0.70, exceeding thresholds for adequate reliability (Flynn, Sakakibara, Schroeder, Bates, & Flynn, 1990). Unidimensionality and construct validity were established by using the maximum likelihood estimation (MLE) method. All values for the model fit indices (GFI ¼ 0.90, NFI ¼ 0.95, and CFI ¼ 0.96) exceed a hurdle of 0.90 as recommended by Bentler and Bonett (1980), which indicates the model fit is acceptable. The value for normed x2 equals 3.63, with p ¼ 0.00, suggesting the rejection of a test of ‘exact fit’. However, it has been well discussed that when the sample size is large (.200), the conventional fit index of the significance of a x 2 test and the hypothesis test of ‘exact fit’ might not be appropriate (MacCaluum, Borwne, & Sugawara, 1996). We used the methodology recommended by MacCallum et al. (1996) to calculate the statistical power for assessing the 90% confidence interval (CI) of the test of ‘close fit’. When RMSEAa ¼ 0.08 the power for not rejecting the null hypothesis of close fit at the 0.05 level is 0.99. Thus, the result of the RMSEA ¼ 0.034 in our study indicates a good fit. Convergent validity was supported with all t-values greater than 2.0, as shown in Table 3, thus giving support that all indicators are significantly related to their underlying theoretical constructs (Bagozzi & Yi, 1988). Discriminant validity was conducted to confirm that constructs that theoretically should not be related to each other are, in fact, observed to not be related to each other. The 95% confidence intervals of the correlations between constructs did not include the value of 1.0 (Anderson & Gerbing, 1988). 4. Results and discussion Structural equation modeling (SEM) was used to simultaneously assess the measurement model and test the hypothesised relationships as defined in Figure 4. The model parameters were calculated through the method of maximum likelihood estimation (MLE) using AMOS 7.0 (Joreskog & Sorbom, 1999). Path analysis is one form of SEM that allows empirical estimation of the strength of each causal relationship (Shah & Goldstein, 2006). Table 3 gives the results of the path analysis, revealing that each regression model is statistically significant at ,0.05, indicating all hypotheses are statistically significant. Figure 5 depicts the theoretical model with standardised path coefficients. The fundamental premise of the Baldrige framework is that leadership drives the “system” that creates results. However, how leadership drives the system at the construct Downloaded by [Tianjin University] at 01:06 02 November 2011 252 Z. He et al. Table 3. Regression results. R2 Independent variable Path coefficient S.E. Standardised path coefficient. t Probability Information and analysis Strategic planning 0.66 0.89 Customer and market focus 0.79 Human resource focus 0.88 Business process focus 0.90 Results 0.62 Leadership Leadership Information and analysis Leadership Information and analysis Strategic planning Leadership Information and analysis Strategic planning Customer and market focus Leadership Information and analysis Strategic planning Customer and market focus Human resource focus Leadership Information and analysis Business process focus Human resource focus 0.86 0.72 0.43 0.43 0.28 0.21 0.15 0.61 0.22 0.07 0.09 0.42 0.21 0.06 0.25 0.14 0.14 0.29 0.13 0.029 0.031 0.026 0.047 0.033 0.047 0.044 0.034 0.042 0.031 0.038 0.039 0.036 0.027 0.036 0.033 0.053 0.047 0.048 0.82 0.61 0.38 0.41 0.28 0.24 0.13 0.56 0.23 0.06 0.08 0.39 0.22 0.06 0.25 0.15 0.16 0.35 0.16 29.30 22.91 16.53 9.13 8.30 4.51 3.33 17.75 5.26 2.17 2.42 10.88 5.83 2.38 6.99 4.18 2.72 6.13 2.75 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.030 0.015 0.000 0.000 0.017 0.000 0.000 0.007 0.000 0.006 Dependent variable Downloaded by [Tianjin University] at 01:06 02 November 2011 Total Quality Management 253 Figure 4. Hypotheses based on the 2006 Baldrige framework. level has received minimum attention in the literature. In order to examine how leadership affects results, we decomposed the theoretical model into different ‘pathways’ from leadership to results. A pathway is defined as a set of connections that leads from leadership to results. Figure 5 reveals four such pathways: (1) pathway 1 connecting leadership directly to results; (2) pathway 2 connecting leadership triad and MAKM to results via HR focus; (3) pathway 3 connecting leadership to results via MAKM; and (4) pathway 4 connecting leadership triad, MAKM, and HR focus to results via process management focus. The four pathways are depicted in Figures 6 and 7, respectively. The impact of leadership on results was analysed by decomposing the total effect from leadership to results on the four pathways. The total effect from leadership to results (0.714) through pathway 1 is 0.147; through pathway 2 is 0.134; through pathway 3 is 0.130; and through pathway 4 is 0.303. This result shows that pathway 4 has twice the effects of the other three pathways, revealing the important role of process management focus in the result triad. Figure 5. Proposed 2006 Baldrige model with standardised path coefficients. Downloaded by [Tianjin University] at 01:06 02 November 2011 254 Z. He et al. Figure 6. Pathway 1 and pathway 2. Figure 7. Pathway 3 and pathway 4. 5. Conclusions This research, which uses data collected by CAQ, has two primary objectives: (1) to develop a reliable measurement model for the seven constructs and 19 dimensions regarding the 2006 Baldrige criteria; and (2) to validate the theory underlying the Baldrige framework based on evidence from China. To the best of our knowledge, this is the first study to develop a measurement model directly tied to the 2006 Baldrige criteria at both construct and dimension levels, and to validate the Baldrige framework based on evidence from China. The newly developed measurement model provides a neat and reliable reference for both practitioners and researchers. Practitioners can use this questionnaire to self-evaluate their companies’ quality management practices and find areas for improvement. Researchers can apply this measurement model into different contexts to validate the Baldrige framework. Downloaded by [Tianjin University] at 01:06 02 November 2011 Total Quality Management 255 The theoretical model developed was fully supported with regards to the overall model fit indices and hypothesised linkages. The importance of each construct in the Baldrige framework was analysed by examining the variance explained by each construct. Our results reveal that process management is the most important construct in the Baldrige framework, followed by the leadership construct. Previous studies on Baldrige frameworks have concluded that leadership is the most important construct, followed by information and analysis (Wilson & Collier, 2000). We conclude that quality management practices in China are different than those in Western countries. For companies doing business in China, our research suggests that improving process management is the most efficient and effective approach to pursuing business excellence. While this study provides some interesting findings, it also has some limitations regarding the sample. First, over 50% of the companies sampled have annual sales less than $1.2 million (RMB 8 millions). Firms at this size in industrialised countries are often classified as small companies. Wilson and Collier (2000) acknowledged that quality management practices in small and medium size companies are quite different from those in large companies. Another limitation is that the dimension of measurement, analyse and review of operational performance in the 2006 Baldrige framework is not included in our analysis. A new set of measurement scales for this dimension, which can be validated empirically, will be of interest to both practitioners and researchers. This study has led to some interesting ideas for future research. Chinese national culture is different from that of Western countries, and might explain the differences between quality management practices in different countries. Another potential research direction is to further investigate the roles of each category in the Baldrige system. Findings from such studies will contribute to the body of knowledge regarding quality management theories and practices. Acknowledgement This research was sponsored by the National Science Foundation, grant no. # 70931004. Note 1. In one iteration only one scale was deleted, which either had loading problems or had reliability issue. Scales D8.4.3 and D8.6.1 have low loadings, compared to other performance measures. Scales C4.1, C4.2, C4.3, C4.4, and C4.5 are cross loaded on constructs not intended to. Scales C6.2, C6.5, C6.6, C6.7, C6.8, C6.9, C6.11, C6.16, and C6.17 are cross loaded on more than one sub-dimensions with loadings bigger than 0.50. Scales C1.9, C2.9, C3.5, C5.1, and C5.9 emerged as a single-factor dimension under their original constructs. Scales C2.8 and C3.6 are loaded to constructs not intended to, and without theoretical support. References Anderson, J.C., & Gerbing, D.W. (1988). Structural equation modeling in practice: a review and recommended two-step approach. Psychological Bulletin, 103(3), 411–423. Bagozzi, R.P., & Yi, Y. (1988). On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16(1), 74–95. Bagozzi, R.P., Youjae, Y., & Philips, L.W. (1991). Assessing construct validity in organizational research. Administrative Science Quarterly, 36(3), 421–458. Bemowski, K., & Stratton, B. (1995). How do people use the Baldrige award criteria? Quality Progress, 28(5), 43–47. Bentler, P.M., & Bonett, D.G. (1980). Significance tests and goodness of fit in the analysis of covariance structures. Psychological Bulletin, 88(4), 588–606. Downloaded by [Tianjin University] at 01:06 02 November 2011 256 Z. He et al. Carmines, E.G., & Zeller, R.A. (1979). Reliability and validity assessment. Beverly Hills, CA: Sage Publications. Criteria for Performance Excellence, 2006. Malcolm Baldrige National Quality Award. United States Department of Commerce, Technology Administration, National Institute of Standards and Technology, Gaithersburg, MD. Evans, J.R., & Jack, E.P. (2003). Validating key results linkages in the Baldrige performance excellence model. Quality Management Journal, 10(2), 7–24. Flynn, B.B., & Saladin, B. (2001). Further evidence on the validity of the theoretical models underlying the Baldrige criteria. Journal of Operations Management, 19(6), 617–652. Flynn, B.B., Sakakibara, S., Schroeder, R.G., Bates, K.A., & Flynn, E.J. (1990). Empirical research methods in operations management. Journal of Operations Management, 9(2), 250–284. Flynn, B.B., Schroeder, R.G., & Sakakibara, S. (1994). A framework for quality management research and an associated measurement management. Journal of Operations Management, 11(4), 339–366. Garvin, D.A. (1991). How the Baldrige award really works. Harvard Business Review, 69(6), 80–93. Hair, J.F., Anderson, R.E., Tatham, R.L., & Black, W.C. (1998). Multivariate data analysis. Upper Saddle River, NJ: Prentice Hall. Handfield, R.B., & Ghosh, S. (1995). An empirical test of linkages between the Baldrige criteria and financial performance. Proceedings of the Decision Sciences Institute, 3 (pp. 1713–1715). Hill, J.A., Eckerd, S., Wilson, D., & Greer, B. (2008). The effect of unethical behavior on trust in a buyer-supplier relationship: The mediating role of psychological contract violation. Journal of Operations Management, 27(4), 281–293. Joreskog, K.G., & Sorbom, D. (1999). LISREL 8: User’s Reference Guide. Chicago, IL: Scientific Software International Inc. Lau, R.S.M., Zhao, X., & Xiao, M. (2004). Assessing quality management in China with MBNQA criteria. International Journal of Quality & Reliability Management, 21(7), 699–713. MacCaluum, R.C., Borwne, M.W., & Sugawara, H.M. (1996). Power analysis and determination of sample size for covariance structure modeling. Psychological Methods, 1(1), 130–149. Meyer, S.M., & Collier, D.A. (2001). An empirical test of the causal relationships in the Baldrige Health Care Pilot Criteria. Journal of Operations Management, 19(4), 403–425. Pannirselvam, G.P., & Ferguson, L.A. (2001). A study of the relationships between the Baldrige categories. International Journal of Quality and Reliability Management, 18(1), 14–34. Podsakoff, P.M., MacKenzie, S.B., Lee, J.Y., & Podsakoff, N.P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5), 879–904. Prahinski, C., & Benton, W.C. (2004). Supplier evaluations: Communication strategies to improve supplier performance. Journal of Operations Management, 22(1), 39–63. Shah, R., & Goldstein, S.M. (2006). Use of structural equation modeling in operations management research: Looking back and forward. Journal of Operations Management, 24(2), 148–169. Sousa, R., & Voss, C.A. (2002). Quality management re-visited: A reflective review and agenda for future research. Journal of Operations Management, 20(1), 91–109. Su, C.T., Li, S.C., & Su, C.H. (2003). An empirical study of the Taiwan National Quality Award causal model. TQM and Business Excellence, 14(8), 875–893. Vardaman, E.J. (2007). Total recalls. Circuits Assembly, 18(9), 26. Wilson, D.D., & Collier, D.A. (2000). An empirical investigation of the Malcolm Baldrige National Quality Award causal model. Decision Sciences, 31(2), 361–390. Total Quality Management 257 Appendix Measurement instrument: selected from the Third National Quality Management Practice Survey questionnaire Please refer to the following self-evaluation guidance regarding each process description. Anchor the status of your process management to the most appropriate number by check “3” or “highlight”. Guidance of self-evaluating the status of process management: l ¼ Extremely disagree; 2 ¼ Disagree; 3 ¼ Somewhat disagree; 4 ¼ Somewhat agree; 5 ¼ Agree: 6¼ Extremely agree. Downloaded by [Tianjin University] at 01:06 02 November 2011 Process description Internal Factor loadings loadings Mean 1. Leadership (Cronbach’s a ¼ 0.926, initial eigen value ¼ 1.88) 1.1 Senior leadership (Cronbach’s a ¼ 0.82) a. Vision and values C1.1 Senior executives create and communicate a 0.83 vision and values focused on quality management to all employees. C1.2 Senior executives create and communicate a 0.81 vision and values focused on quality management to suppliers and partners. b. Legal and ethical requirements Note: C1.3 is highly correlated with C1.4, C1 5, and C1.6. 1.2 Governance and social responsibility (Cronbach’s a ¼ 0.87) a. Organisational governance (Cronbach’s a ¼ 0.84) C1.3 Senior executives create an organisational 0.63 environment that follows the legal and ethical requirements. C1.4 Senior executives create an organisational 0.51 environment of empowerment, innovation, and learning. C1.5 Senior executives periodically evaluate 0.70 organisational performance and the progress of business objectives, and transform evaluations into action plans. C1.6 Our company’s governance mechanism warrants 0.81 management’s behaviors to the interests of the company, shareholders, and other stakeholders. b. Communication and organisational performance (Cronbach’s a ¼ 0.79) C1.7 Our top management participates and 0.79 encourages employee involvement in quality improvement activities. C1.8 Senior executives predict and take actions to 0.74 reduce possible impacts on the public and environment, due to questionable products, services, and operations. C1.10 Senior executives proactively participate in 0.88 community services. medical cares, education and environment protections. c. Support of key communities C1.9 Our company emphasises on building long-term partnerships with suppliers. Note: C1.10 is highly correlated with C1.7 and C1.8. S.D. 0.69 3.84 1.34 0.66 3.92 1.31 0.65 4.39 1.3 0.63 3.95 1.34 0.61 4.02 1.41 0.85 4.08 1.34 0.55 4.06 1.38 0.55 3.98 1.38 0.41 3.69 1.37 4.41 1.3 (Continued) 258 Z. He et al. Appendix 1. Continued. Downloaded by [Tianjin University] at 01:06 02 November 2011 Process description 2. Strategic planning (Cronbach’s a ¼ 0.95, initial eigen value ¼ l.43) 2.1 Strategy development (Cronbach’s a ¼ 0.90) a. Strategy development process (Cronbach’s a ¼ 0.85) C2.1 Our company develops corporate strategic plans based upon analysing key operational factors and relative data. C2.2 Our company creates clear strategic plans, objectives and timetables for product/service quality improvement. b. Strategic objectives (Cronbach’s a ¼ 0.83) C2.3 Employees can feedback their opinions to strategic plans and business objectives. C2.4 Strategic objectives can face challenges and balance all stakeholders’ requirements. 2.2 Strategy deployment (Cronbach’s a ¼ 0.93) a. Action plan developments and deployments (Cronbach’s a ¼ 0.90) C2.5 Our company creates clear strategic plans, objectives and timetables for operational processes. C2.6 Our company’s strategic action plans and human resource plans support company’s key strategic plans. C2.7 Our company creates clear strategic plans, objectives and timetables for organisational reforms. C2.8 Wien selecting our suppliers, their capabilities to meet our quality requirements is the primary consideration. C2.9 Our company can trace the deployment of action plans and make adjustments quickly when necessary. b. Performance prediction (Cronbach’s a ¼ 0.87) C2.10 Our company uses key performance indicators (KPIs) to trace the deployment of strategic objectives, and compares our KPIs with that of competitors or benchmarks. C2.ll Our company invests sufficient resources in order to achieve strategic objectives. C2.12 Our company systematically communicates strategic plans and objectives in a “top-down” fashion. 3. Customer and market focus (Cronbach’s a ¼0.919, Initial eigen value¼ 1.00) 3.1 Customer and market knowledge (Cronbach’s a ¼ 0.88) a. Customer and market knowledge C3.1 Our company segments customers so that we can better define and understand customers’ needs. C3.2 Our company systematically listens and understands the requirements and preferences from different customer and market. Internal Factor loadings loadings Mean S.D. 0.77 0.51 4.03 1.39 0.65 0.53 3.95 1.4 0.83 0.53 3.62 1.32 0.61 0.55 3.74 1.36 0.66 0.57 3.96 1.38 0.67 0.57 3.82 1.35 0.77 0.60 3.76 1.44 4.41 1.32 3.9 1.38 0.75 0.45 3.76 1.43 0.75 0.68 3.94 1.35 0.68 0.47 3.77 1.38 0.84 0.62 4.11 1.34 0.79 0.82 4.03 1.31 (Continued) Total Quality Management 259 Appendix 1. Continued. Downloaded by [Tianjin University] at 01:06 02 November 2011 Process description Internal Factor loadings loadings Mean 3.2 Customer relationship and satisfaction (Cronbach’s a ¼ 0.919) a. Customer relationship building (Cronbach’s a ¼ 0.84) C3.3 Our company defines product-service 0.71 0.64 4.13 characteristics in accordance with customers’ voice. C3.4 Our company continuously improves customer 0.83 0.55 4.1 service processes, in order to provide convenience for customer enquiries, transactions, and complaints. C3.5 Our company systematically reviews customers’ 3.97 loyalty and satisfaction in order to improve product quality and service processes. C3.6 Our company plans well on NPD/NSD 3.91 processes. b. Customer satisfaction determination (Cronbach’s a ¼ 0.83) C3.7 Our company proactively initiates and builds 0.77 0.64 4.34 partnerships with customers. C3.8 Senior executives periodically meet with 0.88 0.57 4.05 customers. 4. Measurement, analysis, and knowledge management (Cronbach’s a ¼ 0.92. Initial eigen ¼1.12) 4.1 Measurement, analysis, and review of organisational performance a. Performance measurement C4.1 Our company systematically collects data and 3.92 information, in order to trace, review and improve organisational performance. C4.2 Our company communicates with partners 4.03 frequently regarding design changes and key factors affecting product/service quality. b. Performance analysis and review C4.3 Our company does well in integrating 3.81 performance information with innovation. C4.4 Senior executives in our company analyze data 4.02 by themselves for strategic planning and decision making. C4.5 Our company provides the results of 3.97 performance data analysis to business units or departments. 4.2 Information and knowledge management (Cronbach’s a ¼ 0.92) a. Data and information availability (Cronbach’s a ¼ 0.83) C4.6 Employees in our company can easily acquire 0.83 0.56 3.47 and use corporate information and data. C4.7 Our suppliers, partners and customers can share 0.81 0.73 3.31 our company’s data and information. b. Organisational knowledge management (Cronbach’s a ¼ 0.86) C4.8 Our company asks suppliers to participate in our 0.83 0.63 3.71 quality improvement projects. S.D. 1.32 1.36 1.4 1.37 1.32 1.41 value 1.36 1.34 1.36 1.36 1.37 1.36 1.37 1.43 (Continued) 260 Z. He et al. Appendix 1. Continued. Process description Downloaded by [Tianjin University] at 01:06 02 November 2011 C4.9 Internal Factor loadings loadings Mean Our company acquires data and information 0.68 0.54 from employees, customers, suppliers and partners, and shares the data and information inside our company. C4.11 We often ask suppliers for suggestions 0.66 0.53 regarding product/service designs. c. Data, information, and knowledge quality (Cronbach’s a ¼ 0.78) C4.10 Our company’s data and information are 0.63 0.43 complete, consistent, and accurate. C4.12 Employees can use quality management tools 0.84 0.49 to analyze data and information, and look for quality improvement opportunities. 5. Human resource focus (Cronbach’s a ¼ 0.95, Initial eigen value ¼ l. 15) 5.1 Work systems (Cronbach’s a ¼ 0.87) a. Organisation and management of work C5.1 We have effective communication and techniques sharing within our company. b. Employee performance management system C5.2 Our company uses salary and other incentive 0.74 0.51 systems to strengthen productivity and customer orientation. c. Hiring and career progression C5.3 We have an effective recruiting process to hire 0.79 0.47 employees with required techniques. C5.4 We take a number of approaches to explore 0.64 0.55 employees’ potential, and help employees to achieve their career goals. 5.2 Employee learning and motivation (Cronbach’s a ¼ 0.90) a. Employee education, training, and development C5.5 We systematically evaluate the training 0.8 0.53 requirements of employees at different levels, and develop teaching and training plans C5.6 Our company systematically assesses training 0.81 0.50 requirements of employees at different levels, and helps them achieve their career objectives. b. Motivation and career development Note: C5.4 is highly correlated with Dimension C5.1. 5.3 Employee well-being and satisfaction (Cronbach’s a ¼ 0.90) a. Work environment C5.7 We stress team-work spirit, and emphasise 0.75 0.59 collaboration. C5.8 Our company keeps improving employees’ 0.81 0.50 working environment b. Employee support and satisfaction C5.9 Our organisational culture is good for promoting authorisation and innovation. C5.10 Our company provides many services, welfare, 0.78 0.64 and policies, in order to ensure employees’ interests and to improve employees’ S.D. 3.75 1.33 3.78 1.38 3.89 1.4 3.54 1.41 3.9 1.34 3.87 1.35 3.82 1.32 3.73 1.37 3.82 1.35 3.8 1.36 4.26 1.33 4.12 1.37 3.8 1.42 3.97 1.33 (Continued) Total Quality Management 261 Appendix 1. Continued. Downloaded by [Tianjin University] at 01:06 02 November 2011 Process description Internal Factor loadings loadings Mean C5.11 We determinine key areas of improvement for 0.59 our working environment that supports employees on the basis of the relationship between employee satisfaction and business results. 6. Process management (Cronbach’s a ¼ 0.96. Initial eigen value ¼ 3.58) 6.1 Value creation processes (Cronbach’s a ¼ 0.88) a. Value creation processes C6.1 Key production/service processes have 0.72 performance indices that are measurable and articulated. C6.2 Our company uses information from supplies, customers, and partners to design processes. C6.3 Our company standardises and documents all 0.81 production/service operational procedures. C6.4 Our company integrates new technology and 0.68 organisation knowledge with designs of production/service processes. C6.5 Our company maintains clean and near working environment to improve productivity. C6.6 Our company uses statistical methods to control production/service process fluctuations. C6.7 Our company will consider effectiveness and efficiencies (such as cycle tintes, productivities, and cost) in designing production/service processes. C6.8 Our company encourages frontline employees to participate in production/service procedure improvement projects. C6.9 Departments and business units in our company often share production/service improvement benefits with other departments and units. 6.2 Support processes and operational planning (Cronbach’s a ¼ 0.95) a. Support processes (Cronbach’s a ¼ 0.91) C6.18 We have systematic project management 0.58 processes to support new product/service development. C6.19 Employees can adjust unnecessary procedures 0.68 in production/service processes when needed. C6.20 Our company places emphasis on equipment 0.76 modifications and evaluates existing techniques in timely fashion. C6.21 Our company continuously improves 0.75 organisational technical and innovative capabilities. C6.22 Our company starts quality control at the 0.72 product/service design stage, in stead of frequently testing and inspecting in the production b. Operational planning (Cronbach’s a ¼ 0.91) C6.10 Our company implements continuous 0.68 improvement projects for production/service processes through project teams. S.D. 0.52 3.68 1.35 0.54 3.94 1.35 3.74 1.36 0.63 4.06 1.39 0.59 3.94 1.37 4.31 1.29 3.78 1.43 3.96 1.35 4.01 1.34 3.71 1.38 0.58 3.63 1.43 0.49 3.63 1.34 0.60 3.94 1.38 0.52 4.04 1.34 0.50 3.84 1.43 0.53 3.76 1.42 (Continued) 262 Z. He et al. Appendix 1. Continued. Downloaded by [Tianjin University] at 01:06 02 November 2011 Process description C6.11 We invite our suppliers to participate into our new product/service designs. C6.12 Our company uses information from suppliers, customers, and partners to improve production/ service processes. C6.13 Senior executives supervise the implementation of critical production/service process improvement projects. C6.14 Our company uses information technology as a means to dramatically improve production/ service processes. C6.15 Our company reasonably allocates resources to ensure the implementation of action plans for process improvement and the continuity of operations under emergency. C6.16 Our company encourages a working philosophy of “jumping out of the frames”. C6.17 Our company systematically collects quality data and information to adjust production/ seivice processes. Internal Factor loadings loadings Mean S.D. 3.41 1.41 0.78 0.56 3.75 1.32 0.56 0.57 4.06 1.33 0.67 0.62 3.78 1.39 0.58 0.62 3.89 1.33 3.63 1.42 3.86 1.39 Please refer to the following self-evaluation guidance regarding each performance description. Compare your performance with that of your competitors or benchmarkings, please anchor your performance level with the most appropriate number by check “3” or “highlight”. Guidance of self-evaluating the status of business results: l ¼ Our performance is worst. 2 ¼ Our performance is poor; 3 ¼ Our performance is fair; 4¼ Our performance is good; 5 ¼ Our performance is very good; 6 ¼ Our performance is excellent. 7. Results (Cronbach’s a ¼ 0.97, Initial eigen value ¼ 37.51) 7.1 Product and service outcomes (Cronbach’s a ¼ 0.91) D8.1.1 Key performance indices (KPIs) relevant to 0.73 0.74 4.06 1.12 product and service quality, such as reliability, safety, and stability D8 1.2 KPIs relevant to product and service cost, such 0.72 0.76 3.99 1.1 as price and value D8.1.3 KPIs relevant to the delivery and support of 0.68 0.74 4.04 1.13 product and service, such as order cycle, payment, customer service, and technical 7.2 Customer-focused outcomes (Cronbach’s a ¼ 0.88) D8.2.1 KPIs relevant to customer satisfaction. 0.70 0.76 4.07 1.1 D8.2.2 KPIs are relevant to the values perceived by 0.71 0.75 4.15 1.14 customer, such as customer loyalty, customer retention rate, “word of mouth” by customer, and establishment of customer relationships 7.3 Financial and market outcomes (Cronbach’s a ¼ 0.89) D8.3.1 KPIs relevant to financial returns and economic values, such as ROI, profit, turnover rate of capital, ratio of asset to liability, valueadded per employee, etc. D8.3.2 KPIs relevant to market, such as market share, business growth, or new market entry. 7.4 Human resource outcomes (Cronbach’s a ¼ 0.89) 0.72 0.78 3.91 1.13 0.68 0.78 3.96 1.14 (Continued) Total Quality Management 263 Appendix 1. Continued. Downloaded by [Tianjin University] at 01:06 02 November 2011 Process description D8.4.1 KPIs relevant to the effectiveness and efficiency of working environment, such as position simplification, position changeover, employee retention, inside promotion rate, etc. D8.4.2 KPIs relevant to the learning and development, such as innovation rate, number of reasonable suggestions, and performance improvement etc. D8.4.3 KPIs relevant to employee welfare and employee satisfaction, such as numbers of emergencies, employee, absence etc. 7.5 Organisational effectiveness outcomes (Cronbach’s a ¼ 0.90) D8.5.1 KPIs relevant to value-creation processes, such as productivity, cycle time, performance of suppliers and partners, as well as other measurements of effectiveness and efficiency. D8.5.2 KPIs relevant to supportive processes. 7.6 Leadership and social responsibility outcomes (Cronbach’s a ¼ 0.90) DS.6.1 KPIs relevant to financial responsibilities inside and outside, such as the independence of auditors or the auditing department. D8.6.2 KPIs relevant to ethical and behavioral issues in the corporate governance, such as the rate of independent directors. D8.6.3 KPIs relevant to legal and regulations, such as environment protection, energy consumption, resources recycling and reuse, etc. D8.6.4 KPIs revelant to community support, such as the number of public activities. Note: Italic items are removed after the data purification process. Internal Factor loadings loadings Mean S.D. 0.73 0.75 3.88 1.14 0.72 0.73 3.88 1.16 3.86 1.16 0.63 0.77 3.92 1.11 0.67 0.76 3.94 1.12 3.99 1.17 0.70 0.73 4.02 1.17 0.77 0.69 4.18 1.2 0.72 0.70 4.02 1.23