5975
Abstract
We empirically examine a mediational model of TQM, in which TQM practices have a direct impact on customer
satisfaction and an indirect impact mediated through plant performance. We adopt a survey approach using the data from
339 manufacturing companies. We first establish convergent validity, discriminant validity, and reliability of the constructs.
We then examine the model using LISREL 8.10. The results suggest paradoxical relations among TQM practices, plant
performance, and customer satisfaction. TQM practices have a stronger impact on customer satisfaction than they do on
plant performance. Further, the plant performance, as described in the mediational model, fails to show a significant impact
on customer satisfaction. This observation is explained based on an institutional argument that states that loose coupling may
occur between TQM practices designed for customer demands and the activities on the plant floor designed for plant
performance. q 1998 Elsevier Science B.V. All rights reserved.
Keywords: Empirical research; Operation strategy; Interdisciplinary; Quality; Performance; Customer satisfaction; Institutional theory;
Supply chain
1. Introduction
US manufacturing firms have faced heightened
challenges from global competitors and a reduced
market share during the past two decades Meredith,
1992; Krajewski and Ritzman, 1996.. To renew their
competitiveness, many firms have adopted the practices of total quality management TQM., and their
experiences have been detailed in many books and
articles see, e.g., the cases presented in Dean and
Corresponding author. Present address. Arizona State University, College of Business, Department of Management and Department of Supply Chain Management, Tempe, AZ 85287-4006,
USA.
0272-6963r98r$ - see front matter q 1998 Elsevier Science B.V. All rights reserved.
PII: S 0 2 7 2 - 6 9 6 3 9 8 . 0 0 0 3 1 - X
60
2. Literature review
In the special issue of the Academy of Management Reiew, which was dedicated to Total Quality,
Dean and Bowen 1994., in their lead article, offered
an overarching perspective that the entire quality
management effort must be focused on achieving
customer satisfaction. Three articles that followed
echoed this perspective. Reeves and Bednar 1994.
suggested that the historical definitions of quality
espouse customer or consumer satisfaction as the
ultimate goal of quality management. Spencer 1994.
introduced three models of TQM mechanistic, organismic, and cultural. and argued that all three
models suggest quality as conforming to or satisfying internal or external customers. Anderson et al.
1994a,b. summarized Demings TQM principles and
delineated customer satisfaction as the final outcome
of quality management. Researchers e.g., Waldman,
1994; Madu et al., 1995. have acknowledged organization performance as another key construct of the
TQM model. McCune 1989. described performance
management as the key source of customer satisfaction. According to the model forwarded by Anderson
et al. 1994b., quality performance outcomes are
driven by quality management practices, which, in
turn, leads to customer satisfaction.
Therefore, these theoretical discussions offer general support for the significance of the constructs
shown in the mediational model of TQM in Fig. 1:
TQM practices, plant performance, and customer
satisfaction. In the remainder of Section 2, we review more carefully the literature related to these
three main constructs. We examine the extent to
which authors of the extant literature support or do
not support the relations among these three constructs. Finally, we summarize the literature into
specific propositions.
61
62
possible combinations of relationships, only 7 relationships were significant. Although marketing researchers have conducted studies on the relationship
between performance and customer satisfaction, they
focus mainly on the consumer satisfaction regarding
the performance of the purchased products Tse and
Wilton, 1988; Zeithaml, 1988; Anderson et al.,
1994a,b; Taylor and Baker, 1994.. Therefore, it may
be difficult to draw applicable conclusions for the
industrial purchasing context where the interaction
occurs between two organizations.
Overall, the positive relationships are suggested
among the three constructs in the TQM model e.g.,
Dean and Bowen, 1994; Capon et al., 1995r1994;
Ross, 1995 Black and Porter, 1996.. However, the
levels of support for these relationships vary. In
general, there is less dispute about the impact of
TQM practices on customer satisfaction than the
impact it has on plant performance results. Also, the
strength of the positive relationship between plant
performance, as affected by TQM practices, and
customer satisfaction, is still quite uncertain. In sum,
the level of impact may vary among the three constructs of Fig. 1. In this study, the level of impact is
operationalized by the level of correlation coefficient
according to Backstrom and Hursh-Cesar 1981.,
the level of correlations measures the level of impact
one construct has on the other, as the squared term of
the correlation coefficient represents explained variance that the associated two constructs have in common. Therefore, we offer the following propositions
to guide our study.
Proposition 1: TQM practices hae a strong impact on customer satisfaction.
Proposition 2: TQM practices hae a moderate
impact on plant performance.
Proposition 3: Plant performance has a weak
impact on customer satisfaction.
3. Data collection
3.1. Surey instrument
A list of TQM activities was compiled based on
the extant literature e.g., Feigenbaum, 1951; Juran,
1951; Crosby, 1979; Deming, 1986; Sashkin and
63
64
descriptions provided at the four-digit level, 55 classifications were selected that related to automotive
and electrical manufacturing.
Plant managers were determined as the most appropriate respondents, because they are most familiar
with their plants operating practices and performance outcomes. Of the 1679 mailed surveys, 67
were returned as undeliverable. Within two months
of mailing, 339 completed surveys were returned,
which was a 21% response rate.
According to the responses to the demographic
questions regarding principal products and plant size,
the sample was generally representative of the targeted industries, but responding plants were, in general, larger than the samples reported norms. The
target population had a mean sales of US$23.5 million and 183 employees, according to the Ohio Manufacturers List 1994.; however, the survey respondents reported a mean of US$40.8 million and 286
employees. It is possible that plant managers of
larger companies had more readily accessible data
and were able to take the time to complete and return
the surveys. It is also possible that some organizational growth had occurred, because there was a
two-year gap between the database figures 1993.
and survey results 1995.. According to the Ohio
Department of Development, Ohio showed the highest manufacturing growth among all 50 states in the
United States during this period.
Constructs
X1
X2
X3
X4
X5
X6
X7
X8
TQM practices
Plant performance
Customer satisfaction
0.68 0.05.
0.58 0.06.
0.59 0.05.
0.66 0.05.
0.70 0.06.
0.62 0.05.
0.76 0.05.
0.16 0.06.
0.55 0.05.
0.62 0.05.
0.68 0.06.
0.55 0.05.
0.29 0.06.
0.61 0.05.
0.37 0.06.
0.48 0.06.
0.36 0.06.
0.56 0.05.
0.37 0.06.
0.29 0.06.
X9
Range of t-values
0.68 0.05.
11.4214.70
5.2113.31
5.0413.13
Table 1
Convergent validity of constructs
65
66
Table 2
Discriminant validity of constructs
Unconstrained model
2
Constrained pair
Constrained model
x 2 differencea
Difference in d.f.
251
249
139
1
1
1
Significant at p - 0.05 after adjusting the a-level for the number of tests performed Anderson and Gerbing, 1988..
For instance, a o s 1 y 1 y a i . t , where a o is the significance level i.e., 0.05. and t is the number of tests, which in our case is 3.
a i is the adjusted significance level that should be used for each test, which in our case was computed to be 0.01 for a conservative
measure.
Table 3
Overall internal consistency of scales
Scale title
Number of respondents
Cronbachs alpha
TQM practices
Process quality
Human resource
Strategic quality planning
Information and analysis
310
279
304
250
0.69
0.78
0.92
0.71
5
7
8
7
1
2
0
1
Plant performance
Quality
Delivery
Cost
306
282
296
0.73
0.48
0.63
3
2
3
0
0
0
Customer satisfaction
Quality
Delivery
Cost
304
319
327
0.73
0.73
0.61
4
3
2
0
0
0
67
Table 4
Internal consistency by ownership and industry
Scale title
Plant ownership
US
Industry
Non-US
Auto
Electronics
Metal stamping
and coating
TQM practices
Process quality
Human resource
Strategic quality planning
Information and analysis
264
235
259
213
0.67
0.75
0.92
0.69
37
37
36
31
0.77
0.88
0.91
0.80
111
98
109
101
0.72
0.79
0.91
0.77
99
90
98
76
0.71
0.78
0.91
0.67
78
73
79
58
0.55
0.78
0.93
0.64
Plant performance
Quality
Cost
263
255
0.72
0.61
38
35
0.81
0.48
116
111
0.75
0.62
95
94
0.70
0.68
77
73
0.73
0.55
Customer satisfaction
Quality
Delivery
Cost
264
274
280
0.73
0.74
0.61
34
38
39
0.68
0.54
0.48
106
117
114
0.68
0.73
0.50
98
97
105
0.67
0.76
0.72
80
84
87
0.80
0.64
0.60
To further examine the scales we used, an additional analysis using Cronbachs alpha was completed through which we analyzed the scales internal consistency by ownership and industry of the
respondents see Table 4.. The industry types were
broken into three categories to reflect the identification provided by the plant managers when they
identified their core businesses. Metal stamping and
coating was an industry designation separated from
auto and electronics industries by the respondents.
This distinction was, therefore, carried over into the
Table 5
Factor analysis by scale
Scale title
Factor loadings
Item a1 Item a2
Item a3
Item a4
Item a5
2.28
2.37
5.11
0.52
0.75
0.77
0.84
0.78
0.83
0.66
0.76
0.87
0.64
0.79
0.74
0.68
2.14
0.72
0.78
0.70
Plant performance
Quality
1
1.96
0.78
0.85
0.79
Customer satisfaction
Quality
1
Delivery
1
2.23
1.95
0.73
0.80
0.76
0.84
0.72
0.78
TQM practices
Process quality
Human resource
Strategic quality
planning
Information and
analysis
No. of
factors
Eigen value
1
1
1
0.78
0.87
Item a6
Item a7
Item a8
0.75
0.80
0.74
68
TQM
Performance results
Customer satisfaction
Quality
Quality
Process quality
Human resource
TQM practices
Process quality
Human resource
Strategic quality planning
Information and analysis
y
0.56)))
0.53)))
0.38)))
y
0.60)))
0.53)))
y
0.66)))
Plant performance
Quality
0.13)
0.13)
0.090
0.13)
Customer satisfaction
Quality
Delivery
0.35)))
0.29)))
0.40)))
0.31)))
0.43)))
0.27)))
0.41)))
0.34)))
0.080
0.083
) p- 0.05.
)) p- 0.01.
))) p- 0.001.
y
0.52)))
Delivery
Table 6
Correlation matrix for TQM, performance, and customer satisfaction
69
LISREL estimations were completed after 24 iterations. Based on maximum likelihood estimations, the
linkages between TQM practices and plant performance and between TQM practices and customer
satisfaction showed significant t values at 2.50 and
6.52, respectively. However, the linkage between
plant performance and customer satisfaction was not
significant with a t value at 0.13. The correlations
among the three constructs are shown in Fig. 2. The
present model showed a good fitthe comparative
fit index CFI. of Bentler 1990. was 0.96, and the
Tucker and Lewis 1973. index TLI. was 0.94. The
x 2 statistic was 38.31 d.f.s 12, p - 0.001., which
failed to achieve non-significance. However, this is
not necessarily an indicator for poor fit, because
models with good fit are known to be falsely rejected
by x 2 tests due to bias stemming from either nonnormality or sample size or both Podsakoff and
MacKenzie, 1994; Rich, 1997.. In this regard, the
CFI and TLI are found to be resistant to this bias
Marsh et al., 1988., and since CFI and TLI are well
above the recommended level of 0.90, the model is
considered to be a reasonable representation of the
data Hair et al., 1995..
5. Discussion
These results offer general support for our propositions. As shown in Fig. 2, all correlations among
the three constructs in the model are positive. TQM
practices are significantly correlated at the 0.05 level
70
Once TQM practices are implemented e.g., suggestion programs, quality circles, display of before
and after pictures of improvements., they are conducive to direct observations during plant visits by
customers. However, performance results such as
internal or external reject rates must be accumulated
over a period of time, and it may be that what is
more visually immediate to the customers has a more
direct impact on their level of satisfaction. Therefore,
as evidenced in our data i.e., plant managers perceived lack of relationship between plant performance and customer satisfaction., plant managers
may perceive customer satisfaction is driven less by
their performance results than by their conformance
to the customers requests. At the same time, the
plant managers know that if they did not have the
required performance they would jeopardize the
long-term relationships with the present customers.
This observation brings us to the last linkage that
was not statistically significantthe relationship between the plant performance and customer satisfaction. Given that there is a loose coupling between
what plants do to increase plant performance and
what they do to satisfy the customer, as argued
above, the plant manager would have difficulty seeing the immediate relationship between plant performance and customer satisfaction. After all, customers seem to take satisfaction from seeing the
plants conform to their requeststhe implementation of TQM practices. In other words, the part of
plant performance improvement that is gained
through the implementation of TQM practices is
thought to lead to little impact on customer satisfaction. Statistically speaking, the variance of plant
performance that is explained by the TQM practices
shows a statistically insignificant level of explained
variance for customer satisfaction.
In summary, the institutional interpretation of the
results offer insights into the paradoxical dynamics
of TQM practices in plants. Powerful institutional
actors such as industrial customers promote TQM
practices, and their actions are sanctioned and reinforced by the professional societies, universities, and
governmental units. Consequently, plant managers
succumb to this pressure and call for compliance.
Plants duplicate the TQM practices, what institutional theorists call legitimated elements, into their
structures. This act of conformance leads to customer
71
72
cater to those customers. However, firms are typically reluctant to disclose exact figures of their performance data, but plant managers are well acquainted with the performance data and could give
us an accurate assessment. Also, it would have been
a daunting task to track down and receive feedback
from the customers of hundreds of plants. For any
study, there is a trade-off between the sample size
and obtainment of objective performance results and
customer satisfaction data e.g., the larger the sample, the more difficult it would be to receive customer satisfaction data from the matching customer
firms.. We balanced this trade-off by using plant
managers as respondents because they would be the
most familiar with both performance data and customer satisfaction.
The potentially damaging weakness is the issue of
the common method variance Avolio et al., 1991.
in our case, data for all three constructs came from
one respondent. Contending with the common methods variance was a conscious choice on our part,
because we had to consider the trade-off between the
number of plants we surveyed versus getting multiple respondents from each plant. Given the constraint, we felt that plant managers would be the
correct people to inform us about all three constructs
TQM practices, plant performance, and customer
satisfaction. However, when we received the data,
we had to examine the effects of common method
variance. The confirmatory analysis gave us a strong
result validating the independence of the three constructs. Therefore, although we did not solve the
problem of the common method variance, we were
reasonably confident that all three constructs were
independent of one another. From the onset, our
intent was to explore the topic in an uncharted area
and to examine the puzzling relationships among
TQM practices, plant performance, and customer
satisfaction. Even with the stated weaknesses, we
believed that we gained a glimpse of the dynamics
that are described confusingly in the literature.
We have only scratched the surface of the complex dynamics of institutionally driven quality initiatives. As discussed in Section 6, our results have
implications for other externally driven programs,
such as ISO 9000. Research conclusions, consistent
with our findings, are beginning to appear in publications. For instance, Terziovski et al. 1997. failed to
find a significant effect of ISO 9000 on organizational performance. Then, what should the management do? How could the management internalize this
type of externally driven changes to reap the potential performance benefits? Or do we just have to bite
the bullet and concede that the benefits of these
externally driven programs end at the institutional
level and have little to no consequences on plant
performance?
73
74
References
Adam, E.E., 1994. Alternative quality improvement practices and
organization performance. J. Operations Manage. 12, 2744.
Ahire, S.L., Golhar, D.Y., Waller, M.A., 1996. Development and
validation of TQM implementation constructs. Decision Sci. J.
27, 2356.
Anderson, E.W., Fornell, C., Lehmann, D.R., 1994a. Customer
satisfaction, market share, and profitability: findings from
Sweden. J. Marketing 58 3., 5366.
Anderson, J.C., Gerbing, D.W., 1988. Structural Equation Modeling in Practice: A Review and Recommended Two-step Approach, 103, pp. 411423.
Anderson, J.C., Rungtusanatham, M., Schroeder, R.G., 1994b. A
theory of quality management underlying the Deming management method. Acad. Manage. Rev. 19, 472509.
Avolio, B.J., Yammarino, F.J., Bass, B.M., 1991. Identifying
common methods variance with data collected from a single
source: an unresolved sticky issue. J. Manage. 17 3., 571587.
Babich, P., 1992. Customer satisfaction: how good is good enough?
Quality Prog. 25 12., 6567.
Backstrom, C.H., Hursh-Cesar, G., 1981. Survey Research. Wiley,
New York.
Bagozzi, R.P., Yi, Y., Phillips, L.W., 1991. Assessing construct
validity in organizational research. Admin. Sci. Q. 36, 421
458.
Bamford, J., 1994. Driving America to tiers. Financial World,
163, pp. 2427.
Baron, R.M., Kenny, D.A., 1986. The moderator-mediator variable distinction in social psychological research: conceptual,
75