Impact of quality management practices on performance and competitive advantage, The

impact of quality management practices on performance and competitive advantage, The

Flynn, Barbara B


As decision makers become more involved in implementing Total Quality Management, questions are raised about which management practices should be emphasized. In this exploratory investigation of the relationship of specific quality management practices to quality performance, a framework was constructed. It focuses on both core quality management practices and on the infrastructure that creates an environment supportive of their use. In addition, it incorporates two measures of quality performance and their role in establishing and sustaining a competitive advantage.

Path analysis was used to test the proposed model, with multiple regression analysis determining the path coefficients, which were decomposed into their various effects. Weak linkages were eliminated. The trimmed model indicated that perceived quality market outcomes were primarily related to statistical control/feedback and the product design process, while the internal measure of percent that passed final inspection without requiring rework was strongly related to process flow management and to statistical control/feedback, to a lesser extent. Both measures of quality performance were related to competitive advantage. Important infrastructure components included top management support and workforce management. Supplier relationships and work attitudes were also related to some of the core quality practices and quality performance measures.

The results were interpreted in light of Hill’s concept of order winners and order qualifiers and Garvin’s eight dimensions of quality. They indicate that different core quality management practices lead to success in different dimensions of quality, and that those dimensions function differently as order winners and order qualifiers.

Subject Areas: Empirical Survey and Production/Operations Management.


As decision makers realize the importance of high quality production in establishing and maintaining a global competitive position, there has been a corresponding interest in research on quality management. Much of the literature to date has been descriptive; in particular, examining the practices used by Japanese companies (e.g., [2], [24], [35], [47], and [66]). The limited empirical literature focuses primarily on describing commonly used quality management practices, in both the U.S. and in Japan (see [6], [11], and [12]). There have been only a few attempts to empirically relate the use of certain quality management practices to quality performance and to the overall performance of the organization. For example, Garvin [12] studied quality practices and performance in the room air conditioner industry, and Roth, DeMeyer, and Amano [56] compared the relationship of various quality practices to quality performance in the U.S., Europe and Japan. Roth and Miller [57] found quality programs to be a strong predictor of manufacturing strength. However, the set of quality management practices included in these studies was quite limited. Quality infrastructure practices were not considered. In addition, these studies were conducted at the business unit (product-market) level, which may be problematic, since implementation of quality management practices can vary widely between plants within a common business unit.

In this exploratory study, we attempt to fill the gap by studying quality management at the plant level. We begin by suggesting a conceptual framework that synthesizes findings from the literature and discussions with managers about quality management practices and their expected relationship to quality performance. We take the position that the plants with the best quality performance will use a coordinated and integrated set of quality management practices, rather than focusing their efforts on a few popular quality management practices.


This paper takes the approach that quality management is an integrated, interfunctional means of achieving and sustaining competitive advantage. In examining the practices that comprise this approach, it groups management practices into core quality management practices, which are expected to lead directly to improved quality performance, and quality management infrastructure practices, which comprise the environment that supports effective use of the core quality management practices.

Figure 1 illustrates the proposed framework for quality management and its effect on performance. The five boxes at the left contain the quality management infrastructure components. At the center are the core quality management practices, indicated by rounded boxes. The product design process and statistical control/feedback are expected to have both a direct effect on quality performance and an indirect effect, through their impact on process flow management, which is also expected to be directly related to the measures of quality performance. The boxes at the right represent performance outcomes. The core quality management practices are expected to have an effect on both perceived quality market outcomes and the percent of items that pass final inspection without requiring rework, which are believed to make an important contribution to competitive advantage. Each of these components is described below.

Core Quality Management Practices

Process Flow Management. Effective process flow management is postulated to have a direct impact on quality performance through reduction of process variance. As process variance is reduced, the likelihood of defective parts is reduced accordingly. Process flow management practices include heavy reliance on preventive maintenance [10] [11], which emphasizes scheduled maintenance in order to avoid equipment breakdowns. Hayes notes that the cleanliness and organization of the workplace is also important, quoting a Japanese manager, “If you clean up the factory floor, you tend to clean up the thought processes of the people on it, too” [24, p. 59].

Foolproofing of a process reduces process flow variance by designing the process so that it is virtually impossible to perform it incorrectly [3] [18]. Also known by the Japanese term Poka-Yoke [61], foolproofing involves design of fixtures, use of electronic detection and signalling devices, product redesign and other practices that encourage or force proper performance of an operation, particularly assembly.

Another means of reducing process flow variance involves scheduling production to allow time for catching up at the end of the day. Knowing that there will be ample time to complete the day’s scheduled production encourages line workers to stop the production line whenever a process flow problem is detected. Also known by the Japanese term Jidoka, line-stop capability simplifies control of the process by directing attention to the stopped equipment and the worker who caused the stoppage [62]. This creates the “necessary pain” [23] to force attention to the solution of the process flow problem, preventing future line stoppages.

When the shop floor is laid out so that machines are in close proximity, rather than separated by in-process inventory, the visibility and subsequent solution of process flow problems is enhanced [18]. In cellular layouts, there may also be flexible workers and multiple machine tending, which facilitates group process problem solving, often without slowing down production. The frequent and visible presence of managers on the shop floor is also very helpful in solving process problems as they occur, keeping the process flowing [58].

Product Design Process. An effective product design process is believed to have a direct impact on quality performance through its effect on product reliability, product features and serviceability. High levels of reliability are achieved by considering the failure probabilities of each individual system and subsystem during the product design process [10]. All else being equal, the fewer the parts in a product, the lower its failure rate will be. Product features and serviceability may be enhanced by including customers on product design teams, incorporating the customers’ perspective in the design process. Product features improve quality by providing a product that better meets the needs of the customers, while serviceability influences the customers’ ease of use [22].

The product design process is also believed to have an indirect effect on quality performance through the impact of design for manufacturability on process flow management. The development of parts designs that are simple to manufacture and assemble leads to less process variance [21]. Key management practices include constant interaction between design engineers, the manufacturing function, customers and suppliers, extensive prototyping and trial production runs. The goal is to “get the bugs out” of a new product before production, rather than rushing to market with a questionable product [21].

In order to achieve the design of reliable and manufacturable products that meet the needs of the customers, interdisciplinary design teams are used, in conjunction with a formalized product design process. Rather than being limited to design engineers, design teams involve diverse constituencies, including direct laborers, manufacturing engineers, quality engineers, customers, suppliers and marketing representatives. They follow a prescribed process that is designed to obtain input from a number of perspectives at each stage, incorporate quality considerations into the design of new products and do so in an efficient manner [21].

Statistical Control/Feedback. The use of statistical control/feedback is proposed to have a direct effect on quality performance by detecting and feeding back information about defective parts to the operators and engineers. Detection of quality problems is achieved through the use of statistical process control (SPC), which establishes the limits of normal variability of the production process. Knowledge of the limits and the use of control charts empowers direct laborers by giving them knowledge of when it is appropriate to stop the production process in order to remedy a quality or process flow problem. This enables determination of the root cause of quality problems, in order to prevent their recurrence.

Feedback of information about the process often takes the form of various charts that portray SPC information, as well as other information, such as defect and breakdown rates, schedule compliance, plant productivity, etc. Feedback may also be verbally provided to individuals by their supervisors. Schonberger states that a key aspect of feedback is its timeliness: “When discovery of a mishap is delayed, the trail of evidence of causes grows cold and the number of combinations of causes quickly becomes astronomical” [59, p. 5].

The use of statistical control/feedback is also postulated to have an indirect effect on quality performance through its effect on process flow management. The use of statistical control/feedback facilitates stopping to remedy problems before the process drifts out of control, producing defectives [61].

Quality Management Infrastructure Practices

Customer Relationship. Establishment and maintenance of an open relationship with customers provides an input to the product design process by facilitating clarification of the customers’ needs and desires. Key to nurturing strong customer relationships is the establishment of communication links between the plant and customers [65]. Practices include frequent meetings with customers, visits to customer plants, customer visits to the plant [58] and encouragement of customer feedback on product and service quality.

The relationship with customers is hypothesized to have an indirect effect on quality performance in three ways. First, by improving initial design quality, a strong relationship with customers will improve quality performance by reducing the number of engineering change orders after the design has reached production, thereby reducing manufacturing process variability [58]. Second, the establishment of strong links with customers is useful in the development of manufacturable designs, allowing determination of which specifications and tolerances are critical from the customers’ perspective. Third, customer interaction is likely to lead to the design of new product features, which better meet the customers’ needs and satisfy customers.

Supplier Relationship. Suppliers can contribute to quality performance in a number of ways. Key to maximizing their contribution is the selection of a small number of suppliers and establishing a long-term relationship with them. Selection of suppliers based on quality considerations, rather than cost, encourages the provision of high quality parts; in contrast, suppliers selected through the competitive bidding process have no incentive to improve quality once a contract has been awarded [47]. Companies with a relationship with suppliers characterized by interdependence and cooperation establish links for systematically exchanging information. For example, supplier certification or qualification programs provide a means of conveying manufacturers’ quality expectations to suppliers, as well as providing assurance about the quality of incoming materials ahd parts [35]. Suppliers also contribute to the product design process through inclusion in product design teams, where they provide input about the capabilities of prospective materials and parts.

The supplier relationship is also expected to be directly related to process flow management, since purchased materials and parts are a dominant source of process variability [41].

Work Attitudes. Development of a work force with positive work attitudes, including loyalty to the organization, pride in work, a focus on common organizational goals and the ability to work with employees from other departments, facilitates teamwork and flexibility [29]. Effective communication between diverse parts of the organization, rather than rivalry, competitiveness or lack of effective means of communication, is vital. Knowledge of common organizational goals is essential in ensuring that teams will progress in a direction that is not inconsistent with the organization’s common goals [60]. Flexibility is particularly important in process flow management, permitting workers to shift assignments to cover bottlenecks and absences, smoothing out process variance [8] [31] [40]. It may also improve problemsolving ability.

When employees are loyal to the organization and have pride in being part of it, they will be more willing to take individual risks in order to better the organization. Thus, employees with positive work attitudes have the knowledge and confidence to attack problems, rather than defer to supervisors, and to be contributing team members.

Workforce Management. Managing the work force to enhance work attitudes requires a non-traditional approach to problem solving and compensation. This approach emphasizes the importance of employees’ ideas and their continuous growth and development [38] [44]. Interfunctional teams become the basis for problem solving, while compensation approaches include incentives for group performance, quality-based incentives and compensation based on breadth of skill [30] [31] [38] [39] [46].

In developing and encouraging team problem-solving approaches, it is important that supervisors take on a new role, functioning as coaches, rather than giving subordinates orders. A coaching supervisor is willing to let employees make their own mistakes, in order to learn how to be empowered and manage their own work; thus, supervisors function more as resources than as superiors in this type of environment.

Top Management Support. Top management support is expected to be integral to encouraging the practices and behaviors that lead to quality performance throughout the organization. It is expected to have an impact on all of the dimensions of the framework. For example, top management can encourage the development of strong customer relationships through the provision of resources for customer plant visits by employees, inviting customers to visit the plant, requiring the solicitation of detailed information about customer needs and specifications, and requiring the inclusion of customer representatives on product design teams. Strong relationships with suppliers can be nurtured by top management de-emphasizing price considerations in evaluating supplier selection and retention, providing the purchasing department with the tools needed to assess supplier quality levels, encouragement of long-term contracts with suppliers and requiring suppliers to be certified for quality [51]. Top management can encourage quality in the design process by sheltering the design function from pressures to rush new products to market before they have been thoroughly tested [21]. Quality in process flow management is encouraged when top management eliminates the use of short-term, output-based measures as the means of supervisor evaluation, and instead, provides rewards for process flow improvements. Top management influences work attitudes through the development and communication of a clear strategy that identifies the nature and direction of the organization as including quality performance, thus, encouraging goal congruence [63]. Statistical control/feedback is encouraged by top management’s insistence that key information about the process be recorded. More importantly, top management must act immediately on process feedback it receives, and encourage all managers and supervisors to do the same. Top management supports workforce management through its provision of resources to support training efforts, encouraging a thorough selection process and the development of compensation schemes which are related to quality goals.

Top management should accept its responsibility for quality and provide active quality leadership through the types of actions listed above. It is important to have strong, visible leadership for quality in a well-developed, focused strategy that has quality as its central focus. A long-term orientation by top management is vital in order to prevent frustration if changes in quality performance progress more slowly than expected. In order to communicate this strategy to employees at all levels, it is necessary to create a managerial climate that focuses on quality performance [32], since employees behave as they perceive they are expected to by higher levels of management [67].

Performance Outcomes

Quality performance is a difficult concept to define precisely. Indeed, Garvin [13] lists eight critical dimensions of quality performance. Garvin’s list includes: performanceprimary operating characteristics of a product; features-characteristics that supplement the basic functioning of the product; reliability-the probability of the product malfunctioning or failing within a specified time period; conformance-the degree to which the product’s design and operating characteristics meet established standards; durability-the amount of use the customer gets from the product before replacement is preferable to continued repair; serviceability-the speed, courtesy, competence and ease of repair; aesthetics, which is based on individual preference for how the product looks, feels, sounds, tastes or smells; and perceived quality, which is based on image, brand name and advertising that makes inferences about quality. Maani and Sluti describe a conceptual model that combines quality dimensions into two constructs, stating that “the link between quality and business unit performance may be explained via two distinct paths, arising from two different definitions of quality: (1) manufacturing-based definition, or quality of conformance, and (2) productbased definition, or quality of design [45, p. 93].” While it is difficult to precisely measure the dimensions of the quality construct in an objective fashion, Figure 1 uses several proxies for quality performance.

Perceived Quality Market Outcomes. Perceived quality market outcomes focuses on management’s perception of the plant’s product quality and customer service, relative to its competition. As such, it is a multidimensional construct, implicitly including product characteristics such as conformance, reliability, performance and durability, as well as serviceability and perceptions of customer satisfaction, which could potentially include features and aesthetics.

Percent of Items That Pass Final Inspection Without Requiring Rework. The percent of items that passes final inspection without requiring rework is an internal measure of the plant’s ability to control its processes so that quality is designed and built into its products, rather than defects inspected out. This primarily measures Garvin’s conformance dimension. As conformance to specifications has an impact on performance, durability and reliability, the percent of items that pass final inspection without requiring rework is expected to be related to perceived quality market outcomes.

Competitive Advantage. A firm’s competitive advantage is the way in which it creates value for its customers. It does so by outperforming its competition on various dimensions, which allows it to establish and sustain a defensible position in its product market. Porter [53] [54] [55] describes two distinct competitive advantages: low cost and differentiation, which may include quality, features, delivery, follow-up service, ease of use and other non-cost means of differentiating a firm from its competitors. Hayes and Wheelwright [25] [68] suggest that there are five manufacturingbased competitive advantages: low cost, high quality, dependability, flexibility and innovativeness.

Traditional economic analysis suggests that these competitive advantages represent tradeoffs. More recently, Porter [53] [54] [55] suggests that every organization must make a choice about which competitive advantage to pursue in order to avoid being “stuck in the middle.” However, there is both theoretical [27] [34] [48] and empirical [19] [20] [69] support for the notion that simultaneous pursuit of several competitive advantages can lead to a stronger position in the market than focusing on a single competitive advantage. These researchers suggest that competing on several fronts simultaneously creates a position that is difficult for competitors to attack.

There have been several recent attempts to conceptualize competitive advantage and related constructs as a measure of multiple performance criteria. For example, Nemetz [49] suggests including measures of quality, material control, delivery, inventory, machine performance, flexibility, and cost, stating that measures included should meet the following criteria: (1) they should reflect manufacturing processes, (2) they should promote decisions congruent with long-term profitability, and (3) they should help to control operations. Wood, Ritzman and Sharma [70] empirically found four independent clusters for achieved performance. These were quality (including performance, durability, reliability and features), delivery (including both speed and dependability), price/cost and a second quality dimension that included performance, consistency and quality as perceived by the customer.


The relationships portrayed in Figure 1 give rise to a number of hypotheses. The first four deal with interrelationships among the quality infrastructure variables. They focus primarily on the foundation for quality management established by top management support.

H1: Customer relationship is directly related to top management support.

H2: Supplier relationship is directly related to top management support.

H3: Workforce management is directly related to top management support.

H4: Work attitudes are directly related to top management support and workforce management, as well as indirectly related to top management support through the mediating effect of workforce management.

The next three hypotheses focus on the determinants of the core quality management practices. They emphasize the supportive role of the quality infrastructure practices. H5: The product design process is directly related to customer relationship, top management support, supplier relationship and work attitudes. It is indirectly related to workforce management through the mediating effect of work attitudes; and to top management support through the mediating effect of supplier relationship, workforce management and work attitudes.

H6: Process flow management is directly related to top management support, supplier relationship, workforce management, work attitudes, product design process and statistical control/feedback. It is indirectly related to top management support, customer relationship and supplier relationship through the mediating effect of the product design process; and to top management support, workforce management and work attitudes through the mediating effect of statistical control/feedback.

H7: Statistical control/feedback is directly related to top management support, workforce management and work attitudes; and indirectly related to top management support and workforce management through the mediating effect of work attitudes.

The final set of hypotheses deals with the effects of the core quality management and infrastructure practices on the measures of plant performance. In the interest of parsimony, the mediators of the indirect effects are not listed.

H8: Perceived quality market outcomes are directly related to the product design process, statistical control/feedback and process flow management; and indirectly related to customer relationship, supplier relationship, workforce management, work attitudes and top management support.

H9: The percent of items that pass final inspection without requiring rework is directly related to product design process, statistical control/feedback and process flow management; and indirectly related to customer relationship, supplier relationship, workforce management, work attitudes and top management support.

H10: Competitive advantage is directly related to perceived quality market outcomes and the percent of items that pass final inspection without requiring rework. It is indirectly related to product design process, statistical control/feedback, process flow management, customer relationship, supplier relationship, workforce management, work attitudes and top management support.



The sample was comprised of data collected as part of a larger data collection effort, which measured management practices and performance characteristics of worldclass manufacturers in the U.S. It was constructed so that plant-level analysis could be conducted, because it is at the plant level that specific quality management practices and performance occur. In order to assure a broad range of practices and performance, a two-factor stratified sample design was used, with industry and plant type as the two factors. Data were solicited from a stratified random sample of 75 manufacturing plants with greater than 100 employees in the machinery, electronics and transportation components industries, defined at the 3-digit SIC code level. These industries were chosen because they are industries in transition, where a great deal of variability in performance and practices was expected to be present. Three plant types were chosen: Japanese-owned (operating in the U.S.), World-Class reputation (U.S.-owned) and traditional (U.S.-owned). These plant types were included because they were expected to provide information about a broad range of management practices and performance.

The plants, all located in the U.S., represented different parent corporations. A master list was developed for each substratum using Dun’s Industrial Guide: The Metalworking Directory [5] as the source for traditional plants, a Japanese-language source published by JETRO as the source for Japanese plants, and Schonberger’s [58] honor roll, and communication with industry leaders as the source of U.S.owned world-class plants. Within each master list, all plants were randomly selected.

Participation was solicited by telephone conversations with the plant managers. A total of 706 questionnaires were received, representing 60 percent (45 plants) of the 75 plants that were contacted. The 45 responding plants were not significantly different from the original universe of 75 plants in terms of plant size or location. Three plants were subsequently removed when it was discovered that they were inappropriate to the sample design (for example, fewer than 100 employees, or from the same parent corporation as another plant in the sample). Responses by plant type and industry are summarized in Table 1. Table 2 summarizes selected characteristics of the sample.


The measured variables correspond to the components of the framework. Top management support is an exogenous variable, not influenced by the other measured variables in the model, while all other variables are endogenous variables. The percent of items that passed final inspection without requiring rework was an objective measure, while the remaining ten variables were each operationalized as the mean value on a perceptual scale (complete scales are listed in the Appendix), measuring a specific component of the framework, where a value of one indicated the best performance and a value of five indicated the worst performance. The scales were developed by us to correspond to the dimensions of quality management practice that we believed would be important, based on the literature. The measure of competitive advantage focused on five competitive advantages (not including quality, in order to avoid tautological logic): low cost, fast delivery, volume flexibility, inventory turnover and cycle time. A low score on this scale indicates that plant management perceives that the plant has been relatively successful pursuing several of these competitive advantages simultaneously.

Table 3 contains the mean, standard deviation and variance of each variable and its zero-level correlation and covariance with all other variables. It indicates that there were no serious problems with unusually high standard deviations, nor unusual means. Table 3 does indicate the presence of multicollinearity, particularly between work attitudes, process flow management and several other variables. Although some multicollinearity between independent variables is not unusual, relatively high levels of multicollinearity between independent variables can lead to difficulties in drawing inferences on the basis of the regression estimates (due to higher standard errors of the estimated correlation coefficients) [2] [50]. However, Asher [1] indicates that even at r=0.8 the obtained results will be close to the true values, while they will be more disparate at higher levels of correlation. The correlation between workforce management and work attitudes (r=0.8108) was the only one that exceeded 0.80. Two tests were used to further assess the impact of the observed multicollinearity. First, Lewis-Beck [42] suggests that zero-order correlations fail to consider how each independent variable is related to all other independent variables simultaneously. He suggests that a stronger indicator of multicollinearity is a high R^sup 2^ value combined with statistically insignificant coefficients when each independent variable is regressed on all others. None of the independent variables was found to have high multicollinearity using this criterion; although several models had relatively high R^sup 2^ values, each had at least several statistically significant coefficients. Second, the Variance Inflation Factor (VIF), which measures the inflation in parameter estimates due to collinearities among the independent variables [50], was calculated for each model. This was done during the path analysis, and VIF values are listed in Table 6. All model variables were well within the VIF limit of 10, indicating that their multicollinearity did not have an undue influence on the least squares estimates. Thus, all variables in the model were retained for further analysis.

Because there were six plants that declined to provide information about the objective measure of the percent of items that passed final inspection without requiring rework, analysis of variance was used to test whether there was a systematic difference between the respondents and nonrespondents to that item, in terms of the other variables. Table 4 indicates that there were no statistically significant differences between respondents and nonrespondents, in terms of the percent of items that passed final inspection without requiring rework; thus, all 42 observations were retained for further analysis. The variables were standardized after descriptive statistics were calculated, but prior to further statistical analysis.

Table 5 contains a summary of the analysis of the reliability and validity of the scales [9]. Reliability was operationalized as internal consistency and was measured by Cronbach’s alpha [4]. Using item intercorrelation matrices as a guide, items that did not strongly contribute to alpha, and whose content was not critical, were eliminated. Table 5 shows that the final alpha values for all scales exceeded the minimum acceptable alpha value of 0.60 [33] [43] [52], and that most did so by a substantial margin, indicating that the scales were internally consistent.

Construct validity measures the extent to which the items in a scale all measure the same construct [14]. Within-scale factor analysis was used to test whether the items in a scale all loaded on a common factor. Table 5 shows that the eigenvalues for each of the scales exceeded the minimum eigenvalue of 1.00 [41], and the Appendix shows the factor loadings by item. Factor loadings of at least +0.40 are considered acceptable [18]; thus, all of the items contributed to their respective scales.

Instrument Development

Because the quality management scales were used to gather data as part of a larger study, they were assembled into seven questionnaires, along with other measures, each targeted at a different respondent. This approach was used in order to ensure the accuracy of the responses; questions and scales were targeted at the respondents expected to be most knowledgeable about their content. Thus, many of the scales were targeted at several managers only, or workers only. For example, while the process flow management scale was administered to a process engineer and three supervisors, the statistical control/feedback scale was administered to ten direct laborers and three supervisors. Altogether, eleven managers and ten workers in each plant received questionnaires. The instrument also collected objective data, including the measure of the percent of items that passed final inspection without requiring rework. All scale responses were eventually averaged into a single plant response per scale. For example, the responses of the ten workers were summarized by the average worker response on the work attitudes scale, or the sum of quality managers and plant accountant’s response was summarized by their average response on the statistical control/feedback scale.

The instrument was pretested at twelve plants, located throughout the United States. Each pretest included a site visit, structured interviews with managers and workers, and administration of the pilot instrument. Revisions were made to the items, based on comments of the pilot respondents.


Path analysis was used to analyze the model, with regression analysis determining the significance of the relationships between the independent and dependent variables [36] [64]. Because of the exploratory nature of this study, several different models were constructed and tested. A simpler initial model, previously tested using discriminant analysis, was refined and modified into the initial path model, shown in Figure 1. It was based on the findings of the initial study and inferences from the literature. This model was subsequently reduced for decomposition. All variables were standardized to conform to a standard normal distribution, following the requirements of path analysis [15] [26].

Prior to conducting the path analysis, the standard assumptions that underlie multiple regression analysis [17] were verified, using three preliminary multiple regression models, with internal quality, external quality and competitive advantage as the dependent variables. The assumptions of constant variance, no influential outliers and normality were verified using the following plots: residuals by predicted values, rankits plot of residuals, studentized residuals by case number, Cook’s distances by case number and Leverage values (hat matrix diagonal) by case number [50]. The Shapiro-Wilk statistic provided a further test for normality. Neither the plots nor the Shapiro-Wilk statistic indicated any potentially significant departures from the assumptions.

Path coefficients between each set of independent and dependent variables were represented by standardized regression coefficients [16]. In order to simplify the model prior to decomposition, all paths whose coefficients were not statistically significant at the 0.15 level or less were eliminated. The relatively high threshold level was chosen in the interest of being conservative in estimating direct and indirect effects; linkages that had even a slight effect remained in the model for decomposition. The correlations between all pairs of variables were then decomposed into the sum of their direct, indirect and spurious effects [1].

A direct effect between two variables is indicated by an arrow joining them in the path model, while an indirect effect is indicated by a pair or series of forwardpointing arrows; for example, top management support is hypothesized to have an indirect effect on the product design process through the mediating effect of customer relationship. A spurious effect exists between two variables because of a third variable or combination of variables that directly or indirectly affects both variables of interest. For example, customer relationship and supplier relationship may be spuriously related because of top management support, which is an antecedent to both. Such an effect, although mathematically part of the decomposition, does not represent a substantially meaningful effect Thus, the sum of the direct and indirect effects represents the total substantively meaningful effect of one variable on another [1].

The sum of all simple (direct) and compound (indirect and spurious) paths provides an indication of the adequacy with which the model was specified. If the model was specified correctly, the empirical correlation between any two variables should be numerically equivalent to the sum of the simple and compound paths linking the two variables, except for measurement error. There are no hard and fast rules for determining the magnitude of acceptable measurement error before concluding that the model specification was incomplete or inadequate; arbitrary decision rules, such as “differences greater than 0.10 suggest the need for model revision” [1, p. 24] are often applied.


Table 6 lists the results of the tests of path coefficient significance. Using the criterion of p

Relationship Between Infrastructure and Core Quality Practices

The determinants of the product design process were found to be top management support and supplier relationship. Top management support was more strongly related to the product design process, both directly and indirectly through the mediating effect of supplier relationship. Top management support encourages the development and use of interfunctional design teams to develop products that are manufacturable and meet the needs of the customers. Suppliers are a key component in these teams, providing information about potential components, as well as receiving information from them about how to better meet their needs.

The link between work attitudes and the product design process was eliminated in the reduced model. This appears counterintuitive, since teamwork is believed to be very important in the product design process. However, we found that the product design function is typically located at the corporate level, rather than at the plant. Thus, our measure of work attitudes (plant level) was not administered to many of those (corporate level) most likely to be key players in design teams. While plant respondents could describe the product design process, they were not a direct part of it. Future research needs to include an assessment of the work attitudes of all design team members.

A second link with the product design process that was eliminated was the link with customer relationship. The scale that measured customer relationship was relatively weak from a measurement perspective, and it contained only three items. Thus, we believe that the lack of significance of customer relationship was probably due to measurement problems, rather than to a lack of importance of the construct. This relationship should continue to be investigated in future research.

Statistical control/feedback was a function of workforce management, work attitudes and top management support. The effect of top management support was indirect, mediated by work attitudes and the combination of workforce management and work attitudes. The strongest total effect was the effect of workforce management on information management. Likewise, process flow management was also a function of top management support, workforce management and work attitudes, as well as statistical control/feedback. Work attitudes and workforce management had the greatest total effects on process flow management.

Top management support was directly related to work attitudes, providing an environment that encourages their development. Top management support was also directly related to workforce management, providing encouragement and rewards for use of practices that lead to the development of positive work attitudes. Positive work attitudes are important in the development of effective work teams, which work to detect and solve process problems, leading to reductions in process variance. Likewise, positive work attitudes are important in development of a work force that willingly and conscientiously collects the data necessary for the provision of timely and effective feedback about the manufacturing process.

Indicators of Quality Performance

The top indicators of perceived quality market outcomes were statistical control/ feedback, product design process, process flow management, top management support and the percent of items that passed final inspection without requiring rework. The effect of top management support was indirect through the mediating effects of the infrastructure and core quality practices. The relative importance of statistical control/ feedback and product design process is not surprising. Information about process quality plays an important role in the production of products that meet the customers’ expectations of high quality, which translates into market outcomes.

The effect of process flow management is in the direction contrary to expectations. The negative sign of the coefficient indicates that the use of poorer process flow management practices leads to better perceived quality market outcomes. This is counterintuitive and cannot be explained by the literature. It may, however, be related to the relatively high degree of multicollinearity exhibited by process flow management with a number of other variables.

Table 6 also indicates that the percent of items that passed final inspection without requiring rework played a significant role in determining perceived quality market outcomes. This is not surprising, indicating the importance of conformance quality to market outcomes. The negative sign of this coefficient is in the expected direction, indicating that plants producing a higher percent of products passed final inspection without requiring rework achieved better perceived quality market outcomes (scores closer to one on a one-to-five scale).

It is interesting to note, however, that the percent of items that passed final inspection without requiring rework was not the only determinant of perceived quality market outcomes, nor was it the key determinant. This demonstrates that quality is a multidimensional construct in the market, including other dimensions, such as reliability, durability, customer service, features and aesthetics, as well as conformance to specifications. Process flow management was the primary determinant of the percent of items that passed final inspection without requiring rework, and its effect was in the expected direction. Its focus is on designing and maintaining a process flow that is predictable and in control, resulting in the production of items with minimal variability. The indirect effects in Table 7 show that process flow management is supported by several important elements. The presence of strong statistical control/feedback provides operators with process information, which allows them to determine when intervention is necessary, as well as providing them with immediate feedback about their own performance. Effective workforce management is important in selecting and developing members of production teams, which are key in solving process flow problems as they arise and creating process improvements. Process flow management, statistical control/feedback and workforce management were all supported by a strong foundation of top management support, which provides direction, examples and rewards for quality behaviors.

The link between the product design process and process flow management was not significantly different from zero. The lack of significance of this link is surprising, particularly in terms of the importance of design for manufacturability approaches in the literature. However, recall that the process flow management scale measures the use of practices that improve process control (variance minimization); it is not a measure of the degree of process control present. It is expected that product design process would be significantly related to a measure of process control; however, this remains to be tested.

In addition, the difference between the sum of paths and implied correlation for some of the variables contributing to the measures of quality performance is relatively high. This indicates that, while these models are statistically significant and have relatively good R2 values, there may be additional relevant dimensions of quality management that were not tested, or that the scope of these dimensions was incompletely measured.

Indicators of Competitive Advantage

The top contributors to competitive advantage were perceived quality market outcomes and the percent of items that passed final inspection without requiring rework. Percent without rework had less than half the direct effect on competitive advantage that perceived quality market outcomes had; a third of its total effect was the indirect effect of percent without rework on competitive advantage, which was mediated by its effect on perceived quality market outcomes. This may be that, in the case of many plants, the market is unaware of internal quality performance, because defects are sorted out before they ever reach the market. However, poor internal quality, when not coupled with sorting out defects at final inspection, could have a devastating effect on market performance. These results also suggest that conformance, while perhaps more critical in competitive performance a decade ago, may be evolving to “order qualifier” [28] status, where high conformance to standards is a prerequisite for even being in the market. The “order winner,” which leads to competitive position by delighting the customers, is related to dimensions of quality other than conformance. Order qualifiers are no less important than order winners; if a plant neglected to meet its internal quality standards, the impact on its competitive performance could be very serious. However, conformance quality performance only gains entrance into the market, while other dimensions of quality performance win orders and enhance competitive position.

Other fairly strong indicators of competitive advantage were top management support, statistical control/feedback, process flow management, product design process and workforce management. All had an indirect effect through their impact on the measures of quality performance. This illustrates the contribution of both the core and infrastructure quality management practices to competitive advantage.

In addition, the R2 value indicates that the combination of percent without rework and perceived quality market outcomes explained slightly over a third of the variance in competitive advantage. This indicates that competitive advantage is a multidimensional construct and that quality performance is only one of several contributors.


This exploratory study makes several important contributions to the literature. First, we developed and tested a framework that describes core and infrastructure quality management practices and their relationship to several dimensions of quality and plant performance. Through analysis of path coefficients and elimination of weak paths, we have refined the original model to a trimmed model that should be a useful departure point for future researchers interested in pursuing the relationship between quality management practices and performance. The initial framework proposed that customers, suppliers, top management and work force cooperate to form an infrastructure that is supportive of the use of the core quality management practices. Core practices include practices related to the product design process, process flow management and statistical control/feedback. While the product design process is important in determining market perceptions of quality, process flow management and statistical control/feedback contribute to the physical quality of the product.

We have also provided evidence that competitive advantage is a multifaceted construct. Although perceived quality market outcomes and the percent of items that passed final inspection without requiring rework both contributed significantly to its variance, roughly two-thirds of the variability of competitive advantage remains to be explained. This indicates that there are other factors that contribute to competitive advantage and suggests that focusing solely on quality improvement may not be a sufficient means for a plant to attain and sustain competitive position. For example, the use of JIT to achieve fast throughput and automation for cost reduction could be two additional factors important in explaining competitive advantage [7] [19].

The results suggest that different dimensions of quality performance function in different strategic ways. We believe that internal, conformance-related quality is an order qualifier, required for a plant’s product to even be considered as a part of the market. While conformance-related quality has been known as an order winner in the past, there is much anecdotal evidence to suggest that it is evolving into an order qualifier, as manufacturers have been forced to make substantial improvements to their conformance-related quality in order to remain in business [37]. Our results suggest that the order winner is quality characteristics that are related to features and aesthetics. As customers consider a number of alternative products, all of which meet conformance standards, it is the design features and aesthetics that win them over, creating or sustaining a competitive advantage for the firm. There is a significant opportunity for future research in expanding the measurement of Garvin’s eight dimensions of quality performance and assessing their relationship with various quality management practices.

Not surprisingly, top management support was found to be critical to both infrastructure and core quality management practices. This demonstrates the idea that quality management is a philosophy that pervades the entire organization, rather than the responsibility of a few isolated individuals or departments. Without strong top management support, the core and infrastructure practices would be ineffective.

It is important to view this study in the context of its limitations. Although the respondents varied by scale, the study relies heavily on the use of perceptual data. The measure of perceived quality market outcomes, in particular, is relatively weak, because it asks managers for their perception of market perceptions of the organization and the quality of its products and services. Measurement of quality in the eyes of the customers would be valuable in future studies. While the data collected for this project included a large amount of objective data, including financial measures, its potential use was severely compromised by a large number of missing values and biased data due to the large number of green field sites among the Japanese plants. Future studies should strive to include more objective data, particularly financial measures of performance. Although the sample was randomly selected within each substratum, it was not designed to be generalizable to all U.S. industry. The sample is limited to three industries and contains roughly two-thirds world-class plants. Thus, there may be some restriction of range.

Most of the previous research about quality management practices has been descriptive, while this study links quality management practices with quality performance at the plant level. We have established not only the link between quality management practices and quality performance, but have also described the link between quality performance and competitive advantage. Future research should continue to investigate and refine the linkages between various types of core and infrastructure quality management practices, dimensions of quality performance and competitive advantage. [Received: June 16, 1992. Accepted: September 6, 1995.]


[1] Asher, H.B. Causal modeling. Beverly Hills, CA: Sage Publications, 1983. [2] Cohen, J., & Cohen, P. Applied multiple regression/correlation analysis for

the behavioral sciences. Hillsdale, NJ: Lawrence Erlbaum, Assoc., 1983. [3] Cole, R.E. The Japanese lesson in quality. Technology Review, 1981, 83, 29-40.

[41 Cronbach, L.J. Coefficient alpha and the internal structure of tests. Psychometrika, 1951, 16, 297-334.

[5] Dun’s industrial guide: The metalworking directory. New York: Dun & Bradstreet, 1986.

[6]Ebrahimpour, M., & Lee, S.M. Quality management practices of American and Japanese electronic firms in the United States. Production and Inventory Management Journal, 1988, 29(4), 28-31.

[71 Fawcett, S.E. The Japanese challenge: A note on the emergence of Japanese

competitiveness. Operations Management Review, 1989, 7, 46-52. [8]Feuer, D. Paying for knowledge. Training, 1987, 24, 57-66.

[9) Flynn, B.B., Schroeder, R.G., & Sakakibara, S. A framework for quality management research and an associated measurement instrument. Journal of Operations Management, 1994, 11, 339-366.

[10] Garvin, D.A. Quality on the line. Harvard Business Review, 1983, 61, 64-75. [11] Garvin, D.A. Japanese quality management. Columbia Journal of World Busi

ness, 1984, 19(3), 3-12.

[12] Garvin, D.A. Quality problems, policies and attitudes in the United States and Japan: An exploratory study. Academy of Management Journal, 1986, 29, 653-673.

[13] Garvin, D.A. Competing on the eight dimensions of quality. Harvard Business Review, November-December 1987, 65, 202-209.

[14] Ghiselli, E.E., Campbell, J.P., & Zedeck, S. Measurement theory for the be

havioral sciences. San Francisco: W.H. Freeman, 1981. [15] Graybill, F.A. An introduction to linear statistical models (Vol. 1). New York: McGraw Hill Book Company, 1961.

[16] Griffin, L.J. Causal modeling of psychological success in work organizations. Academy of Management Journal, 1977, 20, 6-33.

[17] Hair, J.F., Anderson, R.E., & Tatham, R.L. Multivariate data analysis. New York: MacMillan Publishing Company, 1987. [18] Hall, R. Zero inventories. New York: Dow-Jones, Irwin, 1983. [19] Hall, W.K. Survival strategies in a hostile environment. Harvard Business Review, September-October 1980, 75-85.

[20] Hambrick, D.C. High profit strategies in mature capital goods industries: A contingency approach. Academy of Management Journal, 1983, 26, 687-707. [21] Hartley, J.R. Concurrent engineering. Cambridge, MA: Productivity Press, 1992. [22] Hauser, J.R., & Clausing, D. The house of quality. Harvard Business Review, 1988, 66(3), 63-73.

[23] Hay, E.J. The just-in-time breakthrough: Implementing the new manufacturing basics. New York: John Wiley & Sons, 1988.

[24] Hayes, R.H. Why Japanese factories work. Harvard Business Review, 1981, 59, 56-66.

[25] Hayes, R., & Wheelwright, W.C. Restoring our competitive edge: Competing through manufacturing. New York: John Wiley, 1984.

[26] Heise, D. Problems in path analysis and causal inference. In E. Borgatta (Ed.),

Sociological methodology. San Francisco: Jossey-Bass, 1969, 38-73. [27] Hill, C.W.L. Differentiation versus low cost of differentiation and lost cost: A contingency framework. Academy of Management Review, 1988, 13, 401-412. [28] Hill, T. Manufacturing strategy. Homewood, IL: Richard D. Irwin, 1989.

[29] Huber, V.L., & Brown, K.A. Human resource issues in cellular manufacturing: A sociotechnical analysis. Journal of Operations Management, 1991,11,138-159. [30] Ingram, E. The advantages of knowledge-based pay. Personnel Journal, 1990, 69, 138-140.

[311 Jenkins, G.D., & Gupta, N. The payoffs of paying for knowledge. National Productivity Review, 1985, 4(2), 121-130.

[32) Joiner, B.L. The key role of statisticians in the transformation of North American industry. The American Statistician, 1985, 39, 224-234. [33] Jones, A.P., & James, L.R. Psychological climate: Dimensions and relationships of individual and aggregated work environment perceptions. Organizational Behavior and Human Performance, 1979, 23, 201-250. [34] Jones, G.R., & Butler, J.E. Costs, revenue and business-level strategy. Academy of Management Review, 1988, 13, 202-213.

[35] Juran, J.M. Japanese and Western quality–A contrast. Quality Progress, 1978,

II, 10-18.

[36] Kim, J.Q., & Mueller, C.W. Factor analysis: Statistical methods and practical issues. Newbury Park, CA: Sage Publications, 1978.

[37] Kim, J.S., & Miller, J. Building the value factory: A progress report for U.S. manufacturing. Boston University: Manufacturing Roundtable Research Report Series, 1992.

[38] Lawler, E.E. Pay and organizational development. Reading, MA: AddisonWesley, 1981.

[39] Lawler, E.E. Design of effective reward systems. In J. Lorsch (Ed.), Handbook of organizational behavior. Englewood Cliffs, NJ: Prentice-Hall, 1987, 255-272. [40] Lawler, E.E., & Ledford, G.E. Skill-based pay: A concept that’s catching on. Compensation and Benefits Review, 1985, 18, 54-61. [41] Leonard, FS., & Sasser, W.E. The incline of quality. Harvard Business Review, 1982, 60, 163-171.

[42] Lewis-Beck, M.S. Applied regression: An introduction. Beverly Hills, CA: Sage Publications, 1980.

[43] Lord, FM., & Novick, M.R. Statistical theories of mental test scores. Reading, MA: Addison-Wesley, 1968.

[44] Luthans, F., & Fox, M.L. Update on skill-based pay. Compensation and Benefits Review, 1989, 21, 62-67.

[45] Maani, I.E., & Sluti, D.G. A conformance-performance model: Linking quality strategies to business units’ performance. In J.E. Ettlie, M.C. Burstein & A. Fiegenbaum (Eds.), Manufacturing strategy. Norwell, MA: Kluwer Academic Publishers, 1990.

[46] Majchrzak, A. The human side of factory automation. San Francisco: JosseyBass Publishing, 1988.

[47] Manoochehri, G.H. Building quality into the product. Business, 1985, September, 47-50.

[48] Murray, A.I. A contingency view of Porter’s `generic strategies.’ Academy of

Management Review, 1988, 13, 390400.

[49] Nemetz, PL. Bridging the strategic outcome measurement gap in manufacturing organizations. In J.E. Ettlie, M.C. Burstein & A. Fiegenbaum (Eds.), Manufacturing strategy. Norwell, MA: Kluwer Academic Publishers, 1990.

[50] Neter, J., Wasserman, & Kutner, L Applied linear statistical models. Homewood, IL: Irwin, 1990.

[51] Note on quality: The views of Deming, Juran and Crosby. Harvard Business School Note #9-687-011, 1986.

[52] Nunnally, J.C. Psychometric theory. New York: McGraw-Hill Book Company, 1978.

[53] Porter, M.E. Competitive strategy: Techniques for analyzing industries and competitors. New York: The Free Press, 1980.

[54] Porter, M.E. Competitive advantage: Creating and sustaining superior performance. New York: The Free Press, 1985.

[55] Porter, M.E. The competitive advantage of nations. New York: The Free Press, 1990.

[56] Roth, A.V., Demeyer, A., & Amano, A. Manufacturing strategy, manufacturing strength, managerial success and economic outcomes. In J.E. Ettlie, M.C. Burstein & A. Fiegenbaum (Eds.), Manufacturing strategy. Norwell, MA: Kluwer Academic Publishers, 1990.

[57] Roth, A.V., & Miller, J. International manufacturing strategies: A comparative analysis. In E. Ferdows (Ed.), Managing international manufacturing. Amsterdam: Elsevier Science Publishers B.V., 1989.

[58] Schonberger, R. World class manufacturing. New York: The Free Press, 1985.

[59] Schonberger, R. Building a chain of customers. New York: The Free Press, 1990. [60] Shalley, C. Oldham, G., & Porac, J. Effects of goal difficulty, goal setting method and expected external evaluation on intrinsic motivation. Academy of Management Journal, 1987, 30, 553-566.

[61] Shingo, S. Non-stock production: The Shingo system for continuous improve

ment. Cambridge, MA: Productivity Press, 1988.

[62] Sugimori, Y, Kusunoki, K., Cho, F., & Uchikawa, S. Toyota production system and kanban system: Materialization of just-in-time and respect for human system. International Journal of Production Research, 1977, IS(6), 553-564. [63] Suzawa, S. How the Japanese achieve excellence. Training and Development Journal, 1985, May, 110-117.

[64] Swamidass, PM., & Newell, W.T Manufacturing strategy, environmental uncertainty and performance: A path analytic model. Management Science, 1987, 33(4), 509-524.

[65] Tillery, K.R. An exploratory study of the quality function with five manufacturing organizations. Unpublished doctoral dissertation, Georgia State University, 1985.

[66] Weiss, A. Simple truths of Japanese manufacturing. Harvard Business Review, 1984, 62, 119-125.

[67] Wheelwright, S.C. Japan-Where operations are really strategic. Harvard Business Review, 1981, 59, 57-64.

[68] Wheelwright, S.C., & Hayes, R.H. Competing through manufacturing. Harvard Business Review, 1985, 63, 99-109.

[69] White, R.E. Generic business strategies, organizational context and performance: An empirical investigation. Strategic Management Journal, 1986, 7, 217-231.

[70] Wood, C.H., Ritzman, L.P., & Sharma, D. Intended and achieved competitive priorities: Measures, frequencies and financial impact. In J.E. Ettlie, M.C. Burstein & A. Fiegenbaum (Eds.), Manufacturing strategy. Norwell, MA: Kluwer Academic Publishers, 1990.

Barbara B. Flynn is an associate professor of operations management at Iowa State University. She holds an A.B. in psychology from Ripon College, an M.B.A. from Marquette University and a D.B.A. in operations management from Indiana University. Her previous publications have appeared in Decision Sciences, Journal of Operations Management, International Journal of Production Research, and other journals. Her current research interests include quality, JIT and operations strategy. Professor Flynn’s research has been funded by the National Science Foundation, the Center for Innovation Management Studies, the Japan-U.S. Friendship Commission and the U.S. Department of Education. She has held leadership positions in the Decision Sciences Institute, the Academy of Management and The Institute of Management Science.

Roger G. Schroeder is a professor of operations management and codirector of the Quality Leadership Center at the Curtis L. Carlson School of Management at the University of Minnesota. He received a B.S. and M.S.I.E. degree in industrial engineering from the University of Minnesota and his Ph.D. in operations/ research from Northwestern University. He is on the faculty of the Minnesota Executive Program and has consulted with many public and private organizations. Professor Schroeder is the author of the textbook Operations Management, published by McGraw-Hill. His research has been funded by the Ford Foundation, the Exxon Education Foundation, the American Production and Inventory Control Society, the Japan-U.S. Friendship Commission and the National Science Foundation. His current research interests include strategy in operations management, quality improvement and management of technology.

Sadao Sakakibara is a visiting assistant professor of operations management at Keio and Gakushuin Universities, Tokyo, Japan. He holds a B.S. in mechanical engineering from Aichi Institute of Technology, Japan, an M.S. in operations research, an M.A. in statistics, a professional degree in industrial engineering and a doctorate in management systems, all from Columbia University. His research interests are comparative evaluation of manufacturing practices between Japanese and U.S. firms in topics related to JIT, quality management, manufacturing strategy and technology. Professor Sakakibara has previously published articles in Journal of Operations Management, Production and Operations Management, Business Horizons, and other journals. His research has been funded by the Japan-U.S. Friendship Commission and the National Science Foundation.

Copyright American Institute for Decision Sciences Sep/Oct 1995

Provided by ProQuest Information and Learning Company. All rights Reserved