1

Contextual Influences on Managerial Attitudes toward Supplier

Integration: A Cross-Cultural Study

ABSTRACT

Studies show the benefits of supplier integration, but behavioral constraints to

integration exist that can result from negative attitudes toward the practice. Research, however, is

lacking as to what influences such attitudes. Using the theory of reasoned action, our study

closes this research gap by investigating how various contextual drivers –– namely, a

collaborative organizational culture, time-based manufacturing practices (TBMP), and country

culture –– interact to affect managers’ attitudes toward supplier integration. A cross cultural

study is conducted by using secondary data collected from 224 US manufacturing managers and

117 Chinese manufacturing managers. We test the hypothesized model using a multi-group SEM

approach. The results show that collaborative organizational culture significantly increases the

positive attitudes towards supplier integration in both US and Chinese sample, while TBMP

significantly increases the positive attitudes towards supplier integration only in the Chinese

sample. In addition, we find the Chinese cultural context diminishes the effect that a

collaborative culture has on supplier integration due to the strong in-group collectivist belief

system. Our results show that overcoming negative attitudes will require more than simply

espousing the benefits of supplier integration; looking deeper into an organization’s situational

context is required.

Key Words: Supplier integration, the theory of reasoned action, cross-cultural study, multi-

group SEM, and measurement invariance.

2

INTRODUCTION

The positive impact of supplier integration for firms worldwide has been shown in supply

chain management literature (Petersen, Handfield, & Ragatz, 2005; Das, Narasimhan, & Talluri,

2006; Krause, Handfield, & Tyler, 2007; Paulraj & Chen, 2007; Terpend, Tyler, Krause, &

Handfield, 2008; Azadegan & Dooley, 2010; Lockström, Schadel, Harrison, Moser, & Malhotra,

2010; Lockstrom, Schadel, Moser, & Harrison, 2011). However, as noted by Fawcett, Fawcett,

Watson and Magnan (2012), different behavioral constraints that impede supplier collaboration

can emerge from different institutional, cultural, and organizational contexts. Recent supply

chain literature has particularly been interested in differences between the United States and

China (Zhao, Flynn, & Roth, 2007) . Relatively few companies have achieved high levels of

supply chain collaboration required to obtain breakthrough performance. In particular, behavioral

constraints, such as negative attitudes toward change and collaboration, make it difficult to

integrate customer and supplier resources for unique competitive advantage.

Adapted from Pagell (2004) and Das et al (2006), supplier integration is defined as a

process of interaction and collaboration in which the focal firm and its supplier(s) work to

synchronize supply processes in a cooperative manner for mutually acceptable outcomes.

Although both economic and behavioral reasons explain managerial attitudes toward supplier

integration (Kelley, Whatley, & Worthley, 1987; Landeros, Reck, & Plank, 1995; Kannan & Tan,

2004), the latter is largely overlooked in previous research. As such, the literature is unclear if

and how non-rational factors –– such as a-priori beliefs and tacit cultural norms –– affect

attitudes toward supplier integration. For instance, it is unclear if organizational experiences with

intra-organizational collaboration will lead toward a positive attitude toward inter-organizational

(i.e. supplier) collaboration. Likewise, it is not clear if tacit in-group collectivist cultural norms

3

(like those in China) will tend to view integration with external groups (i.e. suppliers) positively.

Thus, a closer study of the behavioral factors that could impact attitudes toward supplier

integration would aid understanding into why some firms resist integrating with suppliers and

some do not.

Using the theory of reasoned action (Fishbein and Ajzen 1975), our study closes this

research gap by investigating how various cultural and organizational contexts interact to affect

managers’ attitudes toward supplier integration. We conduct a cross-cultural study in the US and

China to examine how a collaborative organizational culture and an advanced manufacturing

practice –– i.e., time-based manufacturing practices (TBMP) –– can create beliefs that affect

managers’ attitudes toward supplier integration. Particularly, collaborative organizational culture

refers to a set of shared beliefs towards working collectively within an organization (Schein,

2004). TBMP refers to a set of manufacturing practices that allow quick and nimble production,

such as cellular manufacturing, quality improvement efforts, preventive maintenance, and pull

production (Nahm, Vonderembse, & Koufteros, 2003),essentially the synchronization of buyer

and supplier processes (Das et al., 2006). We hypothesize that both collaborative organizational

culture and TBMP increases managerial positive attitudes toward supplier integration. In

addition, we investigate differences between U.S. and Chinese firms to reveal if and how cultural

norms, particularly the in-group collectivist orientation in Chinese culture, moderate the

relationships.

Our results indicate that collaborative organizational culture increases managerial

positive attitudes toward supplier integration in both US and Chinese samples, while TBMP

increases these attitudes only in the Chinese sample. In addition, we find the support that the

Chinese cultural context diminishes the effect that a collaborative culture has on attitudes toward

4

supplier integration. The paper contributes to the current literature by examining how managerial

attitudes towards supplier integration are formed. In addition, we identified two drivers of

managerial attitudes towards supplier integration, namely collaborative organizational culture

and TBMP, and investigated the impacts under different cultural context.

BACKGROUND THEORY AND HYPOTHESIS DEVELOPMENT

Literature has made extensive use of Fishbein and Ajzen's theory of reasoned action for

understanding motivations for human behavior (Madden, Ellen, & Ajzen, 1992). At its core, the

theory describes processes for the formation of beliefs, attitudes, and intentions that collectively

lead toward predictions of behavior. For the purposes of our study on attitudes toward supplier

integration, we focus on the theory’s explanations of attitude formation within the context of

antecedent beliefs.

According to the theory of reasoned action, a belief is defined as “the subjective

probability of a relation between the object of the belief and some other object, value, concept, or

attribute” (Fishbein & Ajzen, 1975) (p.131). For instance, a manager may believe that arranging

production into an integrated work cell relates to faster production. Belief formation, then, is the

creation of links between objects that emerge either from direct observations of association

(called descriptive beliefs), from conjectured associations based on observations (called

inferential beliefs), or from outside sources claiming associations exist (called informational

beliefs). In a manufacturing context, employees continually update their beliefs as new

experiences, judgments, and information is presented.

The theory of reasoned action further defines an attitude as a location on a bipolar

evaluative dimension with respect to some object, representing the overall favorableness toward

5

some object (Fishbein & Ajzen, 1975) – i.e., an expectancy-value model. For example, a

manager may locate the idea of fast production toward the high-end of favorableness for the

organization. In addition, as beliefs link and unlink objects to attributes, attitudes toward objects

shift as a function of attribute evaluations. This means, if fast production begins to associate with

quality failures then the favorableness of fast production decreases. Recent work finds that an

individual can have multiple attitudes toward a psychological object, and for this reason it is

crucial to understand the context within which the attitude forms and is studied (Ajzen, 2001).

As new contextual experiences occur, attitudes form through a cognitive interaction of

new beliefs (descriptive, informational, and inferential) that emerge from the new experiences, as

well as based on a priori beliefs that existed before the new experience (Fishbein & Ajzen, 1975).

That is, if a manager believed a priori that fast production was highly favorable, one observance

or inference that fast production created quality failures will not substantially alter the favorable

attitude toward fast production. It should be noted that an important aspect of the attitude process

involves internal consistency (cf. Fishbein and Ajzen 1975, p.144-145), which implies that if two

objects are associated with a positive attribute then both objects will tend to be positively favored.

Thus, if managers associate both work cell and pull production techniques with a positive

attribute like fast production, then similar attitudes will form toward both techniques. Studies

have proposed that affect (i.e., mood) also directly influences attitude, but empirical results are

inconclusive and instead suggest that only hedonic objects have an affect component – functional

objects (e.g., business practices) do not (Ajzen, 2001).

The object relevant to this study is the act of supplier integration. Our interest is in

explaining variations in attitudes toward supplier integration through the use of contextual

variables relating to managerial, organizational, and environmental (i.e., country) contexts. This

6

is important because regardless of a manager’s personal convictions, if contextual factors are

found to influence attitudes toward supplier integration then we will come closer to

understanding why negative attitudes may persist when they should not.

The Influence of Organizational and Managerial Contexts

Our first hypothesis is that the presence of a collaborative organizational culture will be

associated with positive attitudes toward supplier integration. According to Schein (2004), an

organizational culture is a set of shared assumptions, values, and beliefs about the world that

develops through challenges of internal integration and external adaptation. Thus, each

organization has embedded a priori beliefs about how work best gets accomplished. If, for

instance, an organization believes collaboration is an appropriate approach to accomplishing

goals, then because beliefs tend toward internal consistency (Fishbein & Ajzen, 1975), activities

with collaborative attributes will also be inferred as positive and appropriate.

Pagell (2004, p.460) defines the concept of integration as “a process of interaction and

collaboration in which (departments) work together in a cooperative manner to arrive at mutually

acceptable outcomes for their organization.” As can be seen, the concept of integration is

associated with the attribute of collaboration. Therefore, organizations that a priori view

collaboration favorably –– i.e., seeking employee involvement in decisions, working with others

and customers, avoiding excessive managerial control –– will also view integration activities

favorably. Inferential beliefs should carry over into attitudes toward supplier integration as well.

H1: A collaborative culture increases managerial positive attitudes toward supplier

integration.

7

Our second hypothesis is that when organizations implement TBMP, new informational

and descriptive beliefs form that will increase the positive attitude toward supplier integration.

Manufacturing plants implement practices in order to survive in an ever-growing, time-based

competitive environment (Stalk, 1991). Practices involved in TBMP –– such as pull production,

cellular manufacturing, quality improvement, and preventative maintenance –– seek to

coordinate supply with demand in a highly synchronized and robust way (Koufteros,

Vonderembse, & Doll, 1998). Research shows such practices are successful (Nahm et al., 2003).

As employees see and are told about the successes of such synchronized practices, descriptive

and informational beliefs will form that associate synchronization with positive outcomes.

Das, Narasimhan, & Talluri (2006) note that supplier integration is essentially the

synchronization of buyer and supplier processes. In other words, the act of supplier integration

should be highly associated with the attribute of synchronization. Similar to the reasoning in our

first hypothesis, because belief systems tend toward internal consistency (Fishbein and Ajzen

1975), synchronization-related activities will be inferred as positive in plants that have

successfully implemented TBMP.

H2: Time-based manufacturing practices increase positive managerial attitudes toward

supplier integration.

The Influence of Country Context

Our third hypothesis is that the Chinese cultural context will diminish the effect that the

organizational and managerial context has on positive attitudes toward supplier integration. The

effects of manufacturing practices and organizational beliefs do not take place in a country

8

cultural void, but instead have their influences observed within a country culture context (Ansari,

Fiss, & Zajac, 2010). This means that a priori beliefs can supersede or counteract new descriptive,

informational, and inferential beliefs. Because objects are multi-attribute entities (Fishbein and

Ajzen 1975), objects may be associated with some positive and some negative attributes.

Because country cultures have clear distinctions as to how various attributes are valued and

practiced (House, Hanges, Javidan, Dorfman, & Gupta, 2004), the influence that new attribute

associations have will vary from country to country.

The Chinese culture experiences a strong in-group collectivist orientation1 (House et al.,

2004), meaning that managers in China make clear distinctions between those who are within

their sphere of influence and interaction and those who are not. Strong loyalties exist for in-

group members, while strong discomfort exists for out-group members. This is important

because the act of supplier integration is an act associated with an out-group member. Therefore,

because a priori country contextual beliefs are negative toward out-group members, any positive

influences from collaborative organizational culture or TBMP will be diminished.

H3a: The Chinese cultural context will diminish the effect that a collaborative culture

has on attitudes toward supplier integration.

H3b: The Chinese cultural context will diminish the effect that TBMP has on attitudes

toward supplier integration.

1 The House et al. (2004) GLOBE study measured both cultural practices (what is) and cultural values (what should

be). We focus on cultural practices because that is what managers experience day-to-day.

9

************************************************

Insert Figure 1 Here

***********************************************

METHODOLOGY

To test the research hypotheses, secondary data collected by Nahm et al. (2004) and Li et

al.(2012) are used. The primary data were collected to investigate the impact of organizational

culture on time-based manufacturing and firm performances. The US sample was collected from

224 manufacturing managers and executives from four different industries. The Chinese sample

was collected from 167 manufacturing managers and executives enrolled in the MBA and

Executive MBA programs at Chongqing University during the 2010-2011 academic year.

Measurement Model

Collaborative culture was measured as a second-order factor by five first-order factors,

namely beliefs on working with others, beliefs on customer orientation, beliefs in investing in

facilities and equipment, beliefs on making decisions that are global, and beliefs on management

control (reverse coded) (Nahm et al., 2004). Each first-order factor was measured by three to

four items, totaling seventeen items. TBMP was also measured as a second-order factor by four

first-order factors, namely cellular manufacturing, quality improvement efforts, preventive

maintenance, and pull production. Each was measured by three to five items, totaling fifteen

items. The construct of managerial attitudes toward supplier integration was measured by three

items (see Appendix for our measurement instrument).

To ensure accuracy of translation and the compatibility of the items across the two

samples (Rosenzweig, 1994), two individuals translated the original English version of the

survey into Chinese independently, and then corrected each other’s translations and reconciled

10

the differences through in-depth discussions. To further validate the translation, the items were

then back-translated by a blind independent translator, as recommended by McGorry (2000). The

back-translated English version was then checked against the original English version using two

measures of comparison, comparability and interpretability (Sperber, Devellis, & Boehlecke,

1994).

Fifteen doctoral students and faculty members in supply chain management participated

in evaluating the back-translated survey. Problematic items identified from the back-translation

are eliminated. In consequence, fourty items were used for further analysis. Although two group

SEM is needed for further analysis, we conducted an initial CFA using both samples to ensure

the coherence of the measurement model. The goodness-of-fit results show that the measurement

model exhibits good fit to the data: χ2(621) = 1229 (p < .001); S-Bχ2(621) = 1026 (p < .001);

RMSEA = 0.041 (90% CI 0.036; 0.045); CFI = 0. 942; and IFI = 0. 943. Discriminant validity

of the constructs was assessed by comparing measurement models where the correlation was

constrained to be 1. The initial CFA analysis also showed that sufficient level of reliability. The

cronbach's alpha values are reported in Appendix.

Multi-group SEM

We used structural equation modeling (SEM) with an EQS program to test the

hypothesized causal relationships between constructs. One of the key assumptions in conducting

multi-group SEM is that the set of items and number of underlying constructs is the same across

groups (Steenkamp & Baumgartner, 1998). Therefore, we performed two-group measurement

invariance tests across the two independent survey samples to establish whether the attitudes on

supplier integration vary between US managers and Chinese managers. Measurement invariance

refers to “whether or not, under different conditions of observing and studying phenomena,

11

measurement operations yield measures of the same attribute” (Horn & McArdle, 1992).

Followed the procedure suggested by Baumgartner and Steenkamp (1998) and Byrne (2006), the

following four steps were conducted between the two groups: 1) Equality of item means and

covariances test, 2) configural invariance test, 3) metric invariance test, and 4) scalar invariance

test. Note that because our hypothesized model consists of two second-order factors and one first

order factor, we conducted the invariance tests of first- and second- order factors sequentially.

Item covariance and mean variance

The first step is to test whether the item covariance matrices and the item means are equal

across groups. If the hypothesis of equality is rejected the invariance assessment continues.

Otherwise, between group invariance is verified and no further analyses is unnecessary. The two

group sample was tested for equality of covariance and means in SPSS. The null hypothesis of

equality of item covariance matrices (Σg) across groups was rejected based upon a Box’s M of

2627.9 (1326 df, p<.001). Additionally, the equality of means (µg) hypothesis was tested via a

between groups ANOVA. This hypothesis was also rejected (p<.05) for 32 of the 40 items

utilized in the general CFA. Therefore it was concluded that a lack of Σg and µ

g equality exists,

which allowed further investigation.

Configural invariance

The second step is the configural invariance test, which seeks to confirm model

invariance via constraining the salient (nonzero) and nonsalient (zero) loadings to identical

structures between groups. While allowing the parameters to be freely estimated, if all salient

loadings are significant, all factor correlations are significantly below unity, and adequate model

fit is achieved, then configural invariance is supported.

Since the model in the configural invariance test will be used as the baseline model for all

12

further test of model invariance, a good fit of the baseline model to the data for both groups are

needed (Yuan & Bentler, 2004). Therefore, we examined the measurement model for each group

respectively. Based on the fit statistics –– e.g., comparative fit index (CFI) and the root mean

square error of approximation (RMSEA) (Fan, Thompson, & Wang, 1999) –– and an item’s

error variance, the Lagrangian multiplier (LM) test (i.e., modification indices), and the residual

covariations (Anderson & Gerbing, 1988), five items were eliminated to improve model fit for

both groups. The items we used for analysis are provided in Appendix. The fit indices for both

US and Chinese samples indicate that the model provides good fit for the data. For the US

model, the goodness of fit indices were S-Bχ2(548) = 769 (p < .001); RMSEA = 0.042; CFI =

0.950; and IFI = 0.950. For the Chinese model, the goodness of fit indices were S-B χ2(548) =

646 (p < .001); RMSEA = 0.033; CFI = 0.939; and IFI = 0.941. The fit indices exceed the

recommended threshold values (Byrne, 2006), hence, we considered this baseline model to be

well specified.

To assess configural invariance, we performed a two-group model test consisting of the

baseline models of US and Chinese samples without imposing any equality constraints. As Table

1 shows, the model exhibits adequate fit to the data: χ2(1096) = 1647 (p < .001); S-Bχ

2(1096) =

1414 (p < .001); RMSEA = 0.039; CFI = 0.946; and IFI = 0.947. All first-order factor loadings

are salient in both groups. However, one the second-order factor loading –– the reverse-coded

managerial control construct loading on collaborative culture –– is not significant in Chinese

sample. This implies the concept of control is not opposed to collaboration in China but is in the

US (we explore this in the discussion section). In sum, all 35 first-order factor loadings and 8 of

9 second-order factor loadings are supported to be invariant. Therefore, partial configural

invariance was achieved. We relaxed the constraint of the noninvariance factor loading, i.e.,

13

managerial control construct on collaborative culture, in all later steps. Both first- and second

order factor loadings are provided in Appendix.

************************************************

Insert Table 1 Here

***********************************************

Metric invariance

The third step is the metric invariance test, which seeks to determine if factor loadings are

equal between groups. It shows whether respondents in different groups interpret and respond to

measurement items in an equivalent manner (Steenkamp and Baumgartner, 1998). This is

accomplished through constraining the factor loadings to be equal across groups and assessing

model fit and model invariance. Although chi-square difference tests are usually performed for

model invariance, recent studies (Cheung & Rensvold, 2002; Byrne, 2006) argued that this ∆χ2

value is as sensitive to sample size and non-normality as the χ2 statistics itself, thereby rendering

it an impractical and unrealistic criterion on which to base evidence of invariance. As a

consequence, there is an increasing tendency to argue for model invariance based on two

alternative criteria: (a) the multi-group model exhibits an adequate fit to the data, and (b) the

∆CFI values between models do not exceed 0.01 (Byrne, 2006). Therefore, we assess the model

invariance using overall model fit and the ∆CFI values. The ∆χ2 value is also reported.

First, we constrained all first-order factor loadings to be equal across the two groups. The

goodness-of-fit results reported in Table 1 show that the model exhibits good fit to the data:

χ2(1122) = 1708 (p < .001); S-Bχ2(1122) = 1464 (p < .001); RMSEA = 0.040; CFI = 0.942; and

IFI = 0.943. The model invariance is also achieved since ∆CFI<0.01. A review of the LM Test

statistics for the metric invariance test showed that only 3 out of 26 first-order factor loadings

were noninvariant across the two groups. In effect, the US and Chinese managers generally

14

interpret measurement items equivalently in most cases.

Second, we constrained all first- and second-order factor loadings to be equal across the

two groups except for managerial control on collaborative culture. The goodness-of-fit results

reported in Table 1 show that the model exhibits good fit to the data: χ2(1130) = 1723 (p < .001);

S-Bχ2(1130) = 1478 (p < .001); RMSEA = 0.040; CFI = 0. 941; and IFI = 0. 941. The model

invariance is also achieved since ∆CFI<0.01. A review of the LM Test statistics for the metric

invariance test showed that only one out of eight second-order factor loadings were noninvariant

across the two groups. Again, the results indicate that the US and Chinese managers generally

interpret organizational collaborative culture and TBMP equivalently in most cases. Therefore,

partial metric invariance is achieved between US and Chinese samples.

Scalar invariance

The fourth step is scalar invariance, which is required for determining factor mean

differences between groups. The scalar invariance test is accomplished by imposing intercept

constraints on the model of metric invariance. If scalar invariance is confirmed, group

differences in observed item means implies underlying differences in construct means. The

model invariance is assessed in the same way as in the third step.

We first conducted the scalar invariance test for the observed items by constraining

invariant factor loadings (both first- and second-order factor loadings) and all observed variable

intercepts regardless of whether the factor loading for a variable is fixed to 1.0 for model

identification or freely estimated due to its noninvariance across groups (Byrne, 2006). The

goodness-of-fit results of this test indicate that the model fitted data well: χ2(1164) = 2292 (p <

.001); S-Bχ2(1164) = 2110 (p < .001); RMSEA = 0.065; CFI = 0. 930; and IFI = 0. 931.

Although the fit is adequate, the model invariance is not achieved since ∆CFI>0.01. The LM

15

Test statistics show that 19 out of 35 observed item intercepts were noninvariant across groups.

Second, we conducted the scalar invariance test for the first-order factors by adding

constrains of all first-order factor intercepts regardless of whether the factor loading for a

variable is fixed to 1.0 for model identification or freely estimated due to its noninvariance

across groups (Byrne, 2006). The goodness-of-fit results of this test indicate that the model fitted

data well: χ2(1664) = 2111 (p < .001); S-Bχ

2(1664) = 2292 (p < .001); RMSEA = 0.065; CFI = 0.

930; and IFI = 0. 931. Although the fit is adequate, the model invariance is not achieved since

∆CFI>0.01. The LM Test statistics show that 5 out of 9 first-order factor intercepts were

noninvariant across groups. Although the noninvariance of item and first-order factor intercepts

won’t affect the path coefficients for hypotheses testing, the results imply that there may be

systematic differences between US and Chinese sample in organizational collaborative culture,

TBMP and supplier integration. Thus, we conducted factor mean differences test in the next step

to obtain more insights from the data.

Factor mean differences

This last step is achieved via initially constraining the factor means to be equivalent

between groups. Then a model modification test (e.g. Lagrangian multiplier test) is used to free

factor means where appropriate and allow for estimation.

We first conducted the first-order factor mean difference test by constraining invariant

factor loadings (both first- and second-order factor loadings) and setting all first-order factor

intercepts in the US group to be 0 while allowing those in the Chinese group to be estimated

freely (Byrne, 2006). In this case, a significant first-order factor intercept in the Chinese group

indicates factor mean difference between groups. The goodness-of-fit results of this test indicate

that the model fitted data well: χ2(1154) = 1980 (p < .001); S-Bχ

2(1154) = 1767 (p < .001);

16

RMSEA = 0.045; CFI = 0. 944; and IFI = 0. 945. From the unstandardized estimates, Customer

Orientation (t= 2.504), Quality Improvement Efforts (t= 10.578), Preventive Maintenance (t=

6.568) and Pull Production (t= 7.035) are significantly higher in the Chinese sample than those in

the US sample; while Beliefs on Management Control (reverse coded, t= -11.502) is significantly

lower in the Chinese sample than those in the US sample.

The second-order factor mean difference test was conducted using the approach

suggested by Byrne (2006). Usually, the model is under-identified when testing second-order

factor mean differences. To overcome the identification issue, three approaches were developed

from previous studies. Approach A constrains both first- and second-order factor intercepts to be

0 in group 1 (US), and allow those to be estimated in group 2 (Chinese). It also has to fix the

second-order factor intercepts (group 2) to be equal to one of the first-order intercepts that load

on it. Approach B fixed one of the factor loading to a constant (ML estimates from previous

runs) for each first-order factor. Estimate first-order factor intercepts for both groups but

constrain them to be equal across groups. Then, constrain the second-order factor intercept to be

0 for group 1 and freely estimate that for group 2. Approach C, constrains first-order factor

intercepts to be 0 for both groups, and then constrain the second-order factor intercept to be 0 for

group 1 and freely estimate that for group 2. All three approaches are valid, while approach A is

considered to be preference (Byrne, 2006). Therefore, we conducted the test using approach A.

The goodness-of-fit results of this test indicate that the model fitted data well: χ2(1152) =

1972 (p < .001); S-Bχ2(1152) = 1758 (p < .001); RMSEA = 0.052; CFI = 0. 945; and IFI = 0.

946. From the unstandardized estimates, Collaborative Culture is significantly higher in the

Chinese sample than those in the US sample (t= 2.552). Both first- and second-order factor mean

different test results are provided in Table 2.

17

************************************************

Insert Table 2 Here

***********************************************

RESULTS

With the evidence of measurement invariance, we conducted the structural invariance.

We constrained the structural paths to be equal across groups and also retained all equality

constraints of factor loadings except the parameters found to be noninvariant in the metric

invariance test. The goodness-of-fit results of this more restricted model (Table 3) indicate that it

fitted data well: χ2(1128) = 1706 (p < .001); S-Bχ

2(1128) = 1465 (p < .001); RMSEA = 0.039;

CFI = 0. 943; and IFI = 0. 943. LM Test results show that the two structural paths are both

noninvariant across the two groups. Table 3 reports the standardized path coefficients and

invariance test results. Figure 2 and 3 show the standardized path coefficients and second-order

factor loadings of US and Chinese sample respectively.

************************************************

Insert Table 3 Here

***********************************************

As shown in Figs. 2 and 3, the majority of hypotheses are supported. In particular,

collaborative culture was hypothesized to be positively correlated with managerial attitudes

towards supplier integration (H1). The standardized path coefficient of US sample is 0.875

(t=7.997>1.96, significant at 0.05 level) and that of Chinese sample is 0.541 (t=3.924>1.96,

significant at 0.05 level). H1, therefore, was supported in both US and Chinese samples. TBMP

was also hypothesized to be positively correlated with managerial attitudes towards supplier

integration (H2). However, it was only supported by the Chinese sample (standardized path

18

coefficient=0.447, t=2.974), but not by the US sample (standardized path coefficient=-0.055, t=-

0.724). In addition, Chinese cultural context was hypothesized to diminish the effects that a

collaborative culture (H3a) and TBMP (H3b) have on attitudes toward supplier integration. The

former was supported (p=0.044) while the latter was not (p=0.054). The hypotheses results are

summarized in Table 4.

************************************************

Insert Figure 2 Here

***********************************************

************************************************

Insert Figure 3 Here

***********************************************

************************************************

Insert Table 4 Here

***********************************************

DISCUSSION

While the majority of hypotheses based on the theory of reasoned action are supported,

hypotheses regarding TBMP are problematic. In H2, TBMP positively and significantly affects

managerial attitudes towards supplier integration in the Chinese sample, but not in the US

sample. Perhaps this is because TBMP in the US is attributed more with efficiency than with

synchronization. US managers may view TBMP as“operationally focused”, used to avoid

operational problems and to examine ways to improve efficiency. This may be especially the

case in the US because the national culture is to be more assertive than China in seeking changes

and improvements in the workplace (House et al. 2004). Thus, US managers are likely to not link

the TBMP with synchronization. Alternatively, it may be that supplier integration is not as

19

attributable to synchronization in the US as it is in China. Assertive US cultural practices imply

that independence over one’s workplace is crucial in the US (House et al. 2004), and supplier

synchronization forces supplier dependence. US managers likely view supplier integration

mostly as a means to collaborate rather than a means to synchronize. On the other hand, Chinese

managers are more likely to consider TBMP as synchronization so that they will have the

internal consistency with their priori beliefs on collectivism. Therefore, more cultural reasons

may exist for why H2 was only supported by one group.

Additionally, another unexpected result from the invariance test is that the belief on

management control (reverse coded) did not load on the collaborative culture in the Chinese

sample as it did in the US sample. This means that the concept of management control is not

antithetical to a collaborative work environment in China. A possible explanation for the

difference between the samples is uncertainty avoidant practices, another national culture

dimension (House et al. 2004). In the process of collaboration, parties seek mutually beneficial

outcomes. In a culture with high uncertainty avoidance, as in China, working relationship

between managers and subordinates may require high degrees of specificity so that management

control is considered as potentially useful to collaboration. Therefore, the collaboration

dimension is less opposed to the idea of management control. We also ran some post-hoc

analysis, in which the construct of management control was eliminated in one model, while, in

the other model, the loadings of management control on collaborative culture are forced to be

equal across the two groups. The two post hoc models yield consistent results with our original

model.

Managerial Insights

20

As literature has noted, behavioral constraints to supplier integration are a key inhibitor to

achieving effective supply chain management (Fawcett et al. 2012). If behavioral intensions are

strongly driven by attitudes and beliefs, we have shown to managers culturally-contingent ways

to overcome the behavioral constraints caused by negative attitudes. Specifically, we show that

successfully implementing TBMP in a Chinese manufacturing plant can demonstrate to Chinese

managers the benefits to synchronization. Such a demonstration, we show, helps overcome latent

cultural practices against integrating with out-group members like suppliers. This process may

also work for other synchronization-type internal practices, like enterprise resource planning

systems. So long as the synchronizing practice is successfully implemented, a positive attitude

toward supplier integration in China is more likely to form. Note that the corollary is also likely,

an unsuccessful internal synchronization will lead to negative attitudes toward supplier

integration.

Yet the above TBMP effect is not observed in the US. Instead, positive attitudes toward

supplier integration are completely predicted by a plant’s collaborative culture. This means that

simply showing the benefits of synchronizing will not convince US managers that supplier

integration is good. Behavioral resistance to a supplier integration effort is much more likely in a

culturally non-collaborative US plant than in a similarly non-collaborative Chinese plant. US top

management, therefore, must be cognizant of their plant’s latent beliefs and assumptions before

engaging in certain supply chain management practices. Our results may be seen as implying that

a US plant culture should change before implementing supplier integration practices – this is a

dangerous interpretation. Organizational culture change is fraught with many difficulties and

negative consequences (Harris & Ogbonna, 2002). Rather, our results are best interpreted as

21

showing how the approach toward implementing supplier integration needs to be culturally

sensitive.

CONCLUSION

Although the benefits of supplier integration have been studied in the literature, the

formation of managerial attitudes is overlooked. It is the first paper to examine how managerial

attitudes towards supplier integration are formed. Applying the theory of reasoned action, we

identified two drivers, namely collaborative organizational culture and TBMP. In addition, we

investigated how the impacts differ within country cultures. The findings from the multi-group

SEM analysis indicate that collaborative culture positively impacts the attitudes towards supplier

integration in both US and Chinese sample, while TBMP has positive impact only in the Chinese

sample. In addition, the Chinese cultural emphasis on in-group collaboration diminishes the

effect of a collaborative culture on supplier integration, but not that of TBMP.

One limitation of this paper is the use of secondary, country-level data instead of regional

cultural data. Research is emerging as to the difference between regions within a country

(Hofstede, V. Adriana, Garibaldi, Tanure, & Vinken, 2010) and this could be an interesting

avenue of future research as more data emerges on the subject. Another limitation is that the two

drivers we identified are not exhaustive, but rather inspiring. This means that while our model

explains variations in supplier integration attitudes fairly well, more social or technical reason

may be available to improve predictability – beliefs toward risk and uncertainty are just an

example. Also, we note that the attitudes towards supplier integration are likely to evolve

overtime. Therefore, a longitudinal study of attitudes towards supplier integration may yield

some insights not captured in this paper. Finally, more countries can be involved in future studies

22

to analyze and isolate how specific dimensions of country culture influence the attitude

formation process.

23

REFERENCES

Ajzen, I. (2001). Nature and operation of attitudes. Annual Review of Psychology, 52, 27-58.

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.

Ansari, S. M., Fiss, P. C., & Zajac, E. J. (2010). Made to fit: How practices vary as they diffuse. Academy

of Management Review, 35(1), 67-92.

Azadegan, A., & Dooley, K. J. (2010). Supplier innovativeness, organizational learning styles and

manufacturer performance: An empirical assessment. Journal of Operations Management, 28(6), 488-505.

Byrne, B. M. (2006). Structural Equation Modeling with EQS: Basic Concepts, Applications, and

Programming (second ed. ed.). New York: Lawrence Erlbaum Associates.

Cheung, G. W., & Rensvold, R. B. (2002). Evaluating Goodness-of-Fit Indexes for Testing Measurement

Invariance. Structural Equation Modeling: A Multidisciplinary Journal, 9(2), 233-255.

Das, A., Narasimhan, R., & Talluri, S. (2006). Supplier integration—Finding an optimal configuration.

Journal of Operations Management, 24(5), 563-582.

Fan, X., Thompson, B., & Wang, L. (1999). Effects of sample size, estimation methods, and model

specification on structural equation modeling fit indexes. Structural Equation Modeling, 6(1), 56-83.

Fawcett, S. E., Fawcett, A. M., Watson, B. J., & Magnan, G. M. (2012). PEEKING INSIDE THE

BLACK BOX: TOWARD AN UNDERSTANDING OF SUPPLY CHAIN COLLABORATION

DYNAMICS. Journal of Supply chain management, 48(1), 44-72.

Fishbein, M., & Ajzen, I. (1975). Belief, Attitude, Intention, and Behavior: An Introduction to Theory and

Research. Reading, MA: Addison-Wesley.

Harris, L. C., & Ogbonna, E. (2002). The unintended consequences of culture interventions: A study of

unexpected outcomes. British Journal of Management 13(1), 31-49.

Hofstede, G., V. Adriana, d. H., Garibaldi, S. M., Tanure, B., & Vinken, H. (2010). Comparing regional

cultures within a country: Lessons from Brazil. Journal of Cross-Cultural Psychology, 41, 336.

Horn, J. L., & McArdle, J. J. (1992). A Practical and Theoretical Guide to Measurement Invariance in

Aging Research. Experimental Aging Research, 18(Fall-Winter), 117-144.

24

House, R. J., Hanges, P. J., Javidan, M., Dorfman, P. W., & Gupta, V. (2004). Culture, leadership, and

organizations: The globe study of 62 societies. London: Sage.

Kannan, V. R., & Tan, K. C. (2004). Supplier alliances: differences in attitudes to supplier and quality

management of adopters and non-adopters. Supply Chain Management: An International Journal, 9(4),

279-286.

Kelley, L., Whatley, A., & Worthley, R. (1987). Assessing the Effects of Culture on Managerial Attitudes:

A Three-Culture Test. Journal of International Business Studies, 18(2), 17-31.

Koufteros, X. A., Vonderembse, M. A., & Doll, W. J. (1998). Developing measures of time-based

manufacturing. Journal of Operations Management, 16(1), 21-41.

Krause, D. R., Handfield, R. B., & Tyler, B. B. (2007). The relationships between supplier development,

commitment, social capital accumulation and performance improvement. Journal of Operations

Management, 25(2), 528-545.

Landeros, R., Reck, R., & Plank, R. E. (1995). Maintaining Buyer-Supplier Partnerships. Journal of

Supply chain management, 31(3), 2-12.

Li, B., Nahm, A. Y., Yang, Y., & Lo, B. W. N. (2012). Working paper: The impact of Chinese and U.S.

manufacturing managers’ beliefs upon time-based manufacturing practices: A comparative study.

Lockström, M., Schadel, J., Harrison, N., Moser, R., & Malhotra, M. K. (2010). Antecedents to supplier

integration in the automotive industry: A multiple-case study of foreign subsidiaries in China. Journal of

Operations Management, 28(3), 240-256.

Lockstrom, M., Schadel, J., Moser, R., & Harrison, N. (2011). Domestic Supplier Integration in the

Chinese Automotive Industry: The Buyer's Perspective. Journal of Supply chain management, 47(4), 44-

63.

Madden, T. J., Ellen, P. S., & Ajzen, I. (1992). A Comparison of the Theory of Planned Behavior and the

Theory of Reasoned Action. Personality and Social Psychology Bulletin, 18(1), 3-9.

McGorry, S. Y. (2000). Measurement in a cross-cultural environment: survey translation issues.

Qualitative Market Research: An International Journal, 3(2), 74-81.

Nahm, A. Y., Vonderembse, M. A., & Koufteros, X. A. (2003). The impact of organizational structure on

time-based manufacturing and plant performance. Journal of Operations Management, 21(3), 281-306.

25

Nahm, A. Y., Vonderembse, M. A., & Koufteros, X. A. (2004). The Impact of Organizational Culture on

Time-Based Manufacturing and Performance. Decision Sciences, 35(4), 579-607.

Pagell, M. (2004). Understanding the factors that enable and inhibit the integration of operations,

purchasing and logistics. Journal of Operations Management, 22(5), 459-487.

Paulraj, A., & Chen, I. J. (2007). Strategic Buyer–Supplier Relationships, Information Technology and

External Logistics Integration. Journal of Supply chain management, 43(2), 2-14.

Petersen, K. J., Handfield, R. B., & Ragatz, G. L. (2005). Supplier integration into new product

development: coordinating product, process and supply chain design. Journal of Operations Management,

23(3–4), 371-388.

Rosenzweig, P. M. (1994). When can management science research be generalized internationally? .

Management Science, 40(1), 28-39.

Schein, E. H. (2004). Organizational Culture and Leadership.

Sperber, A. D., Devellis, R. F., & Boehlecke, B. (1994). Cross-Cultural Translation : Methodology and

Validation. Journal of Cross-Cultural Psychology, 25, 501-524.

Stalk, G. (1991). The strategic value of time. In J. D. Blackburn (Ed.), Time-based competition: The next

battleground in american manufacturing. Homewood, IL: Business One Irwin, 67-101.

Steenkamp, J.-B. E. M., & Baumgartner, H. (1998). Assessing Measurement Invariance in Cross-National

Consumer Research. Journal of Consumer Research, 25, 78-90.

Terpend, R., Tyler, B. B., Krause, D. R., & Handfield, R. B. (2008). BUYER–SUPPLIER

RELATIONSHIPS: DERIVED VALUE OVER TWO DECADES. Journal of Supply chain management,

44(2), 28-55.

Yuan, K.-H., & Bentler, P. M. (2004). On Chi-Square Difference and z Tests in Mean and Covariance

Structure Analysis when the Base Model is Misspecified. Educational and Psychological Measurement,

64(5), 737-757.

Zhao, X., Flynn, B. B., & Roth, A. V. (2007). Decision Sciences Research in China: Current Status,

Opportunities, and Propositions for Research in Supply Chain Management, Logistics, and Quality

Managemen. Decision Sciences, 38(1), 39-80.

26

APPENDIX: CONSTRUCT MEASURES WITH RELIABILITY, FACTOR LOADING, AND T-

VALUE FOR US AND CHINESE SAMPLES

Measurement Items US Chinese

Factor

loadings

t-valuesa Factor

loadings

t-values 关于顾客关于顾客关于顾客关于顾客

Customer Orientation (α=0.882; 0.716)b

管理者应该关注于寻求满足顾客需求的方法。

We believe that managers should focus on finding ways to satisfy our

customers.

0.907 0.706

管理者应该关注于为顾客提供价值。

We believe that managers should focus on providing value to customers.

0.897 17.46 0.719 7.765

我们应该努力接近顾客。

We believe that we should strive to get closer to our customers.

0.755 12.593 0.621 5.107 关于设施和设备投资的观念关于设施和设备投资的观念关于设施和设备投资的观念关于设施和设备投资的观念

Beliefs on Investing in Facilities and Equipment (α=0.922; 0.799)

我们想通过设施和设备的投资, 鼓励我们的员工用创新的方式来工作。

Through investments in facilities and equipment, we want to encourage our

workers to work in innovative ways.

0.873 0.604

我们想通过设施和设备的投资, 增加我们员工的智力工作。

Through investments in facilities and equipment, we want to increase

intellectual work among our workers.

0.856 17.214 0.698 5.239

我们想通过设施和设备的投资, 增加我们员工的创造性。

Through investments in facilities and equipment, we want to increase

creativity among our workers.

0.897 19.728 0.875 5.572

我们想通过设施和设备的投资, 促进我们员工为改进产品而努力。

Through investments in facilities and equipment, we want to support product

improvement efforts among our workers.

0.832 19.602 0.677 4.686 关于与他人一起工作的观念关于与他人一起工作的观念关于与他人一起工作的观念关于与他人一起工作的观念

Beliefs on Working with Others (α=0.887; 0.674)

各职能部门应该作为一个团队共同努力工作。

We believe that functional departments should work together as a team.

0.892 0.574

来自一个部门的员工应该与来自其他部门的员工一起共同工作。

We believe that employees from one department should work with

employees from other departments.

0.831 14.306 0.532 5.07

所有员工应该像一个团队一样共同工作。

We believe that employees should work together as a team.

0.851 16.985 0.63 5.713

员工应该清楚其他部门的工作性质。

We believe that workers should understand the nature of work in other

departments.

0.729 13.945 0.623 5.544 关于制定全局性决策的观念关于制定全局性决策的观念关于制定全局性决策的观念关于制定全局性决策的观念

Beliefs on Making Decisions that Are Global (α=0.876; 0.736)

在制定决策的时候，应该考虑该决策的整体效益。

We believe that when making decisions, the overall effects of a decision

should be considered.

0.891 0.776

27

决策应该基于公司的总目标。

We believe that decisions should be based on overall company objectives.

0.872 13.401 0.682 8.37

制定决策应该根据多方面的输入来进行。

We believe that decision making should include input from many areas.

0.776 16.016 0.66 7.025 关于管理控制的观念关于管理控制的观念关于管理控制的观念关于管理控制的观念

Beliefs on Management Control (α=0.839; 0.752)

管理者应该严格控制他的下属。

We believe that managers should take tight control upon their subordinates.

0.812 0.586

命令和控制是最佳的管理方法。

We believe that command and control is the best way to manage.

0.908 11.078 0.973 5.356

员工应该简单地服从管理者给予他的指示。

We believe that workers should simply follow the directions given by their

managers.

0.676 11.047 0.598 6.389 关于整合供应商的观念关于整合供应商的观念关于整合供应商的观念关于整合供应商的观念

Beliefs on Integrating with Suppliers (α=0.820; 0.665)

我们的供应商是我们增强竞争能力的战略伙伴。

We believe that our suppliers are strategic partners in building up our

competitive capabilities.

0.833 0.671

最好的供应商是那些让我们能够为顾客提供价值的供应商。

We believe that the best suppliers are the ones who enable us to provide

value to customers.

0.905 14.844 0.692 9.147

供应商应该参与到产品设计的决策过程中。

We believe that suppliers should be involved in decision making about

product design.

0.642 10.988 0.561 6.335 单元式制造单元式制造单元式制造单元式制造

Cellular Manufacturing (α=0.880; 0.810)

拥有相似的设计或工序要求的产品应该归为一类产品。

Products that share similar design or processing requirements are grouped

into families of products.

0.836 0.787

产品应该按照相似的工序要求进行分类。

Products are classified into groups with similar processing requirements.

0.945 17.818 0.736 8.706

产品应该按照相似的工艺路线要求进行分类。

Products are classified into groups with similar routing requirements.

0.856 12.609 0.73 9.813

设备应该按照相似的产品系列进行组织。

Equipment is grouped to produce families of products.

0.594 9.887 0.624 6.998 质量改进努力质量改进努力质量改进努力质量改进努力

Quality Improvement Efforts (α=0.863; 0.802)

我们使用鱼刺图来查找质量问题的根源。

We use fishbone type diagrams to identify causes of quality problems.

0.699 0.657

我们使用试验设计法（如：田口方法）

We use design of experiments (i.e., Taguchi methods).

0.711 12.537 0.654 6.087

我们使用质量控制图（如 SPC图）

Our employees use quality control charts (e.g., SPC charts).

0.82 11.59 0.765 6.768

我们进行过程能力分析。

We conduct process capability studies.

0.84 11.258 0.777 6.682 预防性维修预防性维修预防性维修预防性维修

Preventive Maintenance (α=0.904; 0.872)

28

我们强调很好的预防性维修。

We emphasize good preventive maintenance.

0.889 0.753

我们进行预防性维修。

We do preventive maintenance.

0.895 20.431 0.902 10.481

我们在非生产时间做预防性维修。

We do preventive maintenance during non-productive time.

0.7 11.977 0.796 8.477

我们定期维修我们的设备。

We maintain our equipment regularly.

0.884 21.336 0.725 8.785 拉动式生产拉动式生产拉动式生产拉动式生产

Pull Production (α=0.881; 0.835)

生产由成品的出货拉动。

Production is “pulled” by the shipment of finished goods.

0.76 0.663

每一站的生产由下一站的当期需求拉动。

Production at stations is “pulled” by the current demand of the next stations.

0.865 12.451 0.849 7.098

我们使用一个“拉动式”生产系统。

We use a “pull” production system.

0.915 12.446 0.88 7.88

Second-Order Factor: Collaborative Cultures

Customer Orientation 0.805 12.84 0.814 6.896

Beliefs on Investing in Facilities and Equipment 0.598 8.708 0.499 4.051

Beliefs on Working with Others 0.816 11.372 0.97 5.648

Beliefs on Making Decisions that Are Global 0.868 10.334 0.837 6.785

Beliefs on Management Control 0.444 5.268 -0.158 -1.495

Second-Order Factor: Managerial Context

Cellular Manufacturing 0.441 5.381 0.774 9.436

Quality Improvement Efforts 0.777 8.643 0.716 5.614

Preventive Maintenance 0.663 9.426 0.731 8.268

Pull Production 0.545 6.142 0.76 5.636

a. p-value is based on ML Robust estimation method.

b. US and Chinese α values, respectively.

29

Table 1: Invariance test results (model fit indices).

Model X2 SB-X

2 df p-value

Rmsea

(90% conf) CAIC CFI IFI ∆ X

2 ∆ CFI

1a. Baseline_US_v1 768 548 <.001 .042

(.035 - .049)

-2746 0.950 0.950

1b. Baseline_Chinese_v1 646 548 0.002 .033

(.021 - .043)

-2700 0.939 0.941

2.Confi gural Invariance 1647 1414 1096 <.001 .039

(.032 - .044)

-6218 0.946 0.947

3.1. Partial Metric

Invariance (1st order

factor loadings)

1708 1464 1122 <.001 .040

(.034 - .045)

-6349 0.942 0.943 61 0.004

3.2. Partial Metric (1st

and 2nd

order factor

loadings)

1723 1478 1130 <.001 .040

(.034 - .045)

-6391 0.941 0.941 76 0.005

4.1. Partial Scalar

Invariance (observed

item intercepts)

2292 2110 1164 <.001 0.065

(.060 - .069)

-5995 0.930 0.931 645 0.016

4.2. Partial Scalar

Invariance (observed

item intercepts and 1st

order factor Intercepts)

2292 2111 1664 <.001 0.065

(.060 - .069)

-5995 0.930 0.931 645 0.016

5.1. Partial Mean

Difference Test(1st order

factor means )

1980 1767 1154 <.001 0.045

(.039 - .050)

-6269 0.944 0.945 333 0.002

5.2. Partial Mean

Difference Test(1st and

2nd

order factor means )

1972 1758 1152 <.001 0.052

(.047 - .057)

-6264 0.945 0.946 325 0.001

6. Partial Structural

Invariance

1706 1465 1128 <.001 0.039

(.033 - .045)

-6389 0.943 0.943 59 0.003

30

Table 2: First- and Second-Order Factor Mean Differences (from model 5.1. and 5.2.)

Factors Mean differences

(Chinese-US) t-statistics

Customer Orientation 0.22* 2.50

Beliefs on Investing in Facilities and Equipment 0.12 1.32

Beliefs on Working with Others -0.09 -1.20

Beliefs on Making Decisions that Are Global 0.09 0.08

Beliefs on Management Control -0.95* -11.50

Beliefs on Supplier Integration 0.08 1.04

Cellular Manufacturing 0.14 1.68

Quality Improvement Efforts 0.98* 10.58

Preventive Maintenance 0.62* 6.57

Pull Production 0.64* 7.04

Collaborative culture 0.11* 2.55

Managerial Context 0.07 1.68

* significant at .05 level.

31

Table 3: Structural Paths Coefficients and Invariant Test Results.

Structural Path US Chinese p-value of structural

invarance test Coefficients t-value Coefficients t-value

Collaborative cultures-->

Attitudes on supplier integration 0.875* 7.997 0.541* 3.924 0.044*

TBMP-->

Attitudes on supplier integration -0.055 -0.724 0.447* 2.974 0.054†

* significant at .05 level.

† significant at .10 level.

32

Table 4: Summary of hypotheses.

Number Hypothesized relationship US model Chinese model

H1 Collaborative cultures--> Attitudes on supplier integration Supported Supported

H2 TBMP--> Attitudes on supplier integration Not Supported Supported

H3a The Chinese cultural context will diminish the effect that a collaborative culture

has on supplier integration. Supported

H3b The Chinese cultural context will diminish the effect that TBMP has on supplier

integration. Not Supported

Figure 1: Hypothesized Model

33

Hypothesized Model

Figure 2: Structural Model of US Sample

* significant at .05 level.

34

Structural Model of US Sample

Figure 3: Structural Model of Chinese Sample

* significant at .05 level.

35

Structural Model of Chinese Sample