The effect of perceived information sharing on trust and control in inter-firm
buyer-supplier relationships:
A mixed-method approach in the Dutch food industry
by
DENISE BAAS d.baas.1@student.rug.nl
S2227800
University of Groningen, the Netherlands Faculty of Economics and Business
Master of Business Administration Organizational and Management Control
January 2016
Supervisor: A. Rehman Abbasi Co-assessor: W. Kaufmann
Word count: 17 870
THE EFFECT OF PERCEIVED INFORMATION SHARING ON TRUST AND CONTROL IN INTER-FIRM BUYER-SUPPLIER
RELATIONSHIPS: A MIXED-METHOD APPROACH IN THE DUTCH FOOD INDUSTRY
Abstract
Purpose – This research aims to investigate in a principal-agent setting how perceived information sharing influences the levels of trust and control mechanisms in buyer-supplier relationships in the Dutch food industry.
Design/Methodology/Approach – A mixed method approach is undertaken, with both a qualitative and quantitative study. The qualitative part consists of eight semi-structured interviews with managers and entrepreneurs of buying firms in the food sector and the quantitative part incorporates a survey design with 56 managers of buying firms in the food sector.
Findings – The different analyses indicated that information sharing positively influences trust and informal control. Moreover, information sharing did not influence formal control.
Practical implications – Sharing information by suppliers is important for a collaboration between buyers and suppliers because trust and informal control consequently will be enhanced, and it does not have any effect on the extent of formal control. Since this paper recognizes the effects of sharing information by suppliers, they will definitely see the benefits to share information to come to closer and stronger collaborations and improve each other’s business.
Originality/Value – This paper is the first using a mixed method approach to indicate the effects of perceived information sharing on levels of trust and control mechanisms in buyer- supplier relationships in the food industry, which also investigates the underlying thoughts and reasons for the tested relationships besides empirically testing the relationships between the relationships.
Keywords
Information Sharing (strategic and operational), Control mechanisms, Trustworthiness, Inter- firm relationship (buyer-supplier relationship), Food Industry, Principal-Agent Theory
Research Field
Management Control & Trust in Inter-Organizational settings
Content
1. Introduction………...5
2. Theoretical Framework..………. 8
2.1 Theoretical foundation – an agency perspective……… 8
2.2 Trust……… 9
2.3 Control mechanisms……….. 11
2.4 Perceived information sharing………... 12
2. Hypotheses………..15
2.5 Hypothesis: Perceived information sharing and trust………15
2.5 Perceived information sharing and control mechanisms………16
3. Methodology ……….. 19
3.1 Research design………. 19
3.2 Data collection and sampling………. 20
3.3 Qualitative method ……… 21
3.3.1 Interviews and analysis………... 21
3.4 Quantitative method ……….. 22
3.4.1 Variables and measurement……… 22
3.4.2 Survey: Sampling and analysis………... 24
3.4.3 Sample statistics………. 25
3.5 Reliability……….. 26
3.6 Validity……….. 27
3.6.1 Content validity……….. 28
3.6.2 Construct validity……… 28
3.7 common method variance……….. 29
3.8 Multicollinearity……… 29
4. Results………. 30
4.1 Correlation and descriptive statistics……… 30
4.2 Hypothesis testing……… 31
4.2.1 Perceived information sharing and trust……… 32
4.2.2 Perceived information sharing and formal control……… 37
4.2.3 Perceived information sharing and informal control……….………… 42
5. Discussion……… 47
6. Conclusion………... 51
6.1 Conclusion and contributions……… 51
6.2 Limitations and suggestions for future research……… 52
7. Appendices………. 55
7.1 Appendix A: Item measurement……… 55
7.2: Appendix B: Interview questions……… 58
7.3 Appendix C: Conducted interviews……….. 59
7.4 Appendix D: Elaborated interviews………... 60
7.5 Appendix E: Factor analysis……….. 66
7.6 Appendix F: Test for non-response bias……… 67
7.7 Appendix G: Short overview results……….……… 68
8. References………69
1. INTRODUCTION
Increasingly, it has been recognised by practitioners and academics that developing closer and longer-term relationships with key suppliers of goods and services becomes more important (Van der Meer-Kooistra & Vosselman, 2000; Ring & van de Ven, 1992). Although there is a rich tradition of academic research focused on buyer-supplier relationships in business markets, work in this area has surged in the recent years. For instance due to the advent of information technology, which causes firms to have increasingly interconnected relationships due to the increased visibility of the supply chain (Ha & Tong, 2008). In addition, the
growing globalization increased the complexity of the supply chain dramatically (Majta, 2012).
As a result of these developments, Tomkins (2001) stated that sharing information between supplier and buyer becomes more important, since individual firms alone are no longer able to satisfy diverse needs of customers. The supplying firm should be willing to help the buying firm and should take initiatives in the interest of this firm (Van der Meer-Kooistra &
Vosselman, 2000). Also Fawcett & Magnan (2001) argued that information is the ‘life blood’
of an effective supply chain. Hence, information sharing increasingly is praised everywhere as the new prevailing way of conducting business since it is seen as a critical factor to realise the benefits of a collaboration (Malhotra et al., 2005; Zhou & Benton, 2007). Particularly in the food industry, transparency between buyer and supplier becomes more and more important (Bastian & Zentes, 2013). Food supply chains were criticised for the food scandals which has come to light in recent years. For instance the horse meat scandal which appeared in 2013, in which traders of the Netherlands were involved (Arumugam, 2013). In any case, these scandals indicated that food retailers were not satisfied with the transparency of their suppliers.
Despite the increased attention given to buyer-supplier relationships with regard to the value of information sharing (Sriram, 2005), previous research mainly investigated the factors which help enhance information exchange (Li et al., 2006) and to avoid prevailing information non-disclosure (Feldmann & Muller, 2003). However, the question remains largely unanswered of what the effects of information sharing would be. Importantly, Das &
Teng (2001) argued that monitoring mechanisms could be applied by buyers to keep track of
the supplier’s business, because in general buyers have less in-depth information about the
supplier. Moreover, providing information could be seen as an asset of value which implies
risk-taking. Consequently, issues of trust play a big role (Hopp et al., 2012). Despite this apparent importance of trust in supply chains, the development of trust perceptions is not a fully understood and established process (Dyer & Chu, 2000). In addition, the concept of information sharing in relationships often draws on agency theory (Ramaswami et al., 1997), but despite the fact that this theory could make apparent contributions to inter-firm
relationships (Stock, 1997), it is mostly researched in customer-to-customer (Jones &
Leonard, 2014) and business-to-customer relationships (Mishra et al., 1998).
Therefore, the aim of this study is to investigate how information sharing by suppliers (operational and strategic information) influence the different components of trust (goodwill and competence) and control (output, behavioural, and social), taken the principal-agent approach. In particular, the focus is on a specific buyer-supplier relationship in which the buyer’s perspective is addressed by applying a mixed method approach. This mixed method study is used to both examine the effect of information sharing on trust and control by means of a quantitative study and also investigate the underlying reasons of these effects by means of a qualitative study. The proposed relationships are suggested in the theoretical framework in figure 1.
From a practical perspective, this research contributes to the understanding of the effects of information sharing. Suppliers do not know whether truthfully sharing information is beneficial to them, since revealing private information on the one side could enhance a trustworthy image (Meijboom et al., 2006) and on the other side could be doubtful since probably private knowledge will be disclosed (Chen, 2003). When the supplier recognizes the consequences of information sharing, this might positively influence the extent of his
transparency which will lead to a successful relationship (Hoetker & Mellewigt, 2009). Also given the fact that parties in the food supply chain are under fire because of the recent scandals in the food industry, this research adds new insights into the current perceptions of food purchasers with regard to their suppliers.
The research question for this study will be ‘How does information sharing of the supplier influence trust and control in buyer-supplier inter-firm relationships in the food industry?’ To answer this research question, two sub questions are formed: ‘How do perceptions of
information sharing (strategic and operational information) influence the level of trust
(goodwill and competence) in buyer-supplier relationships?’ and ‘How do perceptions of
information sharing (strategic and operational information) influence the usage of control mechanisms (formal and informal control) in buyer-supplier relationships?’
The paper proceeds as follows. Chapter two provides an overview of the existing literature and the research framework will be given, and also a set of research hypotheses will be
formulated. Chapter three describes the research methodology. In chapter four the findings are
analysed and described and chapter five discusses the implications of this research for theory
and managerial practice. Finally, chapter six concludes the paper and ends by noting the
limitations of this study and providing directions for further research.
2. THEORETICAL FRAMEWORK
To get a deeper understanding of the different variables in this study, this chapter explores the relevant literature about those variables. First, a theoretical basis of the broader scope of the literature is given, which is drawn from agency theory. Despite the fact that there are several approaches and theoretical foundations in the context of buyer-supplier relationships, for instance social exchange theory, social psychology, agency theory and transaction cost theory (Lai et al., 2005), agency theory best assists the scope of this study. Furthermore, this chapter provides the relevant and established literature with regard to trust, control and perceived information sharing in buyer-supplier relationships. In addition, the presented relationships from figure 1 will be clarified after which the proposed hypotheses are presented.
2.1 Theoretical foundation for information sharing in the context of buyer-supplier relationships – an agency perspective
Agency theory has proved a highly influential lens for analysing the structure and outcomes of relationships between firms (Zenger, 1994). Ross (1973) and Eisenhardt (1989) argues that agency theory is relevant for situations wherein one party (the principal) delegates authority to another party (the agent). According to Steinle et al. (2014), buyer-supplier relationships could be interpreted as these principal-agent situations. The buyer asks the supplier to produce products or services on behalf of them which encompasses allocating some power and
privileges to the supplier. In this case, the supplier acquires compensation for his efforts, while the buyer benefits from the supplier’s efforts and competences (Jensen & Meckling, 1976).
This sounds like a win-win situation for both parties. However, by delegating decisions to the agent, the principal desires to take advantage of the agent’s information and specific
knowledge about his field of interest (Müller & Gaudig, 2011). The agent has, as far as his knowledge and capabilities are concerned, an informational advantage over the principal.
Thus, the principal-agent approach is based on an asymmetrical distribution of information
between buyer and supplier: the buyer does not have access to sufficient information on the
supplier’s behaviour as the supplier’s purposes and actions are not disclosed at all. These
arising problems are called agency problems. Consequently, the buyer tries to disclose some
information by implementing output, process and social controls according to the principal-
agent perspective (Stump & Heide, 1996).
As agency theory underlies these occurrences of individual preferences and the risk of pursuing the supplier’s own objectives, contrary to those asked by the buyer, the principal- agent perspective appears to be the most suitable theoretical lens to study relationships in the buyer-supplier context.
Concluding, the principal-agent perspective helps to explain the problem of existing information asymmetries in relationships in the supply chain and it analyses the challenges buyers have to deal with under circumstances of information asymmetry.
Below, the concepts of the theoretical framework, trust, control, and perceived information sharing are described and explained.
2.2 Trust
In the context of buyer-supplier relationships, trust is considered to be an important factor (Svensson, 2001). Empirical studies on trust are frequently investigated in the context of personal intra-firm relationships, which cannot instantaneously be translated to organizational inter-firm relationships (Mayer et al., 1995). Of course, the influence of globalization and the proliferation of impersonal technologies changed a lot with regard to long-term trusted relationships between firms. Therefore the concept of trust in buyer-supplier relationships has shifted a lot during the last decades. Consequently, this raises questions about the influence of this concept of trust in inter-firm relationships. Moreover, according to Free (2008), there has been little research which focuses on trust in asymmetrical power relationships, in spite of the fact that such relationships, for example buyer-supplier relationships, are widespread.
Many researchers state that trust and specifically inter-organizational trust is such a complex phenomenon that it cannot be generally defined to suit every theoretical purpose (Free, 2008).
Reasonably, definitions have to be expounded to fit specific research objectives, whereby different aspects of trust become important. This research endeavours to develop the analysis of trust in the context of inter-firm buyer-supplier relationships. The definition of Rousseau et al. (1998, p. 295) will be used to define the buyer’s trust in a supplier. In these terms, trust refers to the buyer’s willingness “to accept vulnerability based upon positive expectations of the intention or behaviour” of the supplier with an important replenishment of Mayer et al.
(1995, p. 712) that it is “irrespective of the ability to monitor or control the supplier”.
A critical note about this definition of inter-organizational trust needs to be added. The
concept of inter-organizational trust, as the term also explains, is mainly conceived as a
construct that occurs through interaction between parties. Therefore, inter-organizational trust is referred to two sets of actors, in which both are the trusting and trusted party’s (Blois, 1999). Due to this fact that trust is an interacting concept, it always depends on
trustworthiness. Trustworthiness is a characteristic of the trustee, which stems from the perception held by the trustor about the trustee. This concept of trustworthiness is specifically appointed in this study, compared to the previous and widely used concept of trust in inter- firm relationships (Stuart et al., 2011; Koh et al., 2012 ). Accordingly, the trust of the supplier in the perception of the buyer specifically will be researched in this study, which could also be mentioned as the trustworthiness of the supplier. So, the considerable difference between these two concepts of inter-organizational trust and trustworthiness has been explained clearly in order to prevent indistinctness about the scope of the study.
As stated above that distinguishing between different aspects of trust is important, Das &
Teng (2001) distinguish trust into two components which are according to Sako (1992) very important in inter-firm relationships: goodwill and competence trust. In this study,
competence trust refers to perception of the buyer that the supplying party is capable of performing according to the expectations of the buyer, whereas goodwill trust refers to perception and expectation of the buyer that the supplying firm in the exchange relationship shows concern for the buying firm’s interests and even more important, the interests of the relationship, above or equal to his own interests (Das & Teng, 2001).
Competence trust. In this research, competence means that the supplier is able to constantly and consistently deliver specific products or services and meet the expectations of the buyer (Nooteboom, 1996). To have confidence in this, the buyer needs to have information of the supplier’s previous behaviour. This insight in the supplier’s information could possibly mean that there is a form of information disclosure of the supplier. A supplier who says what he does and reciprocally does what he says will more likely be trusted because he can be relied on to act in a predictable manner. In addition, in a buyer-supplier relationship competence also refers to skills of employees, such as having the right knowledge and communicating or negotiating with the buyer in the right way. These skills are specifically important for the specific employee in the supplying firm who orders and contacts with the buyer.
Goodwill trust. Goodwill means that the supplier is doing favours in tight situations. In other
words specific to this context, it is the willingness of the supplier to help the buying party in a
critical situation. This often leads to a deeper sense of goodwill trust by the buying firm and a
greater commitment to maintaining the relationship in the future (Sengun & Wasti, 2007).
Another indicator of goodwill trust is the perception of the buyer that the supplier assumes to protect the rights and interests of the buyer, even under extraordinary and unexpected
situations (Geyskens et al., 1996). Besides, the perception of sharing confidential information is another aspect of goodwill trust.
2.3 Control mechanisms
Research on control issues in the context of buyer-supplier arrangements have been explored for a few years (Jap & Ganesan, 2000; Heide & John, 1992). Otley (1994, p. 293) noted that
“there is increased monitoring and control between organizations along the supply chain”.
According to Dalley (2011), control is a very important concept in agency theory. Following this agency perspective in the research area of buyer-supplier relationships, control could be defined as the drawing up of clear agreements and the monitoring the buyer applies to its supplier as a procedure by which the buyer affects the activities of the supplier to achieve its objective (Flamholtz, 1996). These control mechanisms are mainly used to counteract the common opportunistic behaviour of the supplier (Wathne & Heide, 2000).
Many researchers will use a distinction to divide control mechanisms with regard to the manner in which a firm will be assessed by another firm (Das & Teng, 2001; Smith et al.
1995). In the literature about inter-firm relationships, the distinction between formal and informal control mechanisms is much disputed. Some researchers state that usage of formal controls is the common method to reduce uncertainty in the supplier, while others state that informal controls like relational norms are the best method to control the supplier’s behaviour (Li et al., 2010). Moreover, some researchers investigated that formal and informal control mechanisms are substitutes (Gulati, 1995; Dyer & Singh, 1998), while others argue that they are complementary mechanisms (Poppo & Zenger, 2002; Luo et al., 2001). Moreover, Das &
Teng (1998) indicated that the relationship between these controls is assumed to be more
complex than previously investigated. Due to these inconsistencies and conflicting views, the
usage of different control mechanisms will be divided into formal and informal control
mechanisms. Das & Teng (2001) and Ouchi (1979) differentiate formal control further into
output (or outcome) and behavioural control, and social control is an informal control
mechanism.
Output control. Output control is the assessment of the supplier’s performance and the accomplishment of certain performance goals. Buyers will check whether the supplier
conducts its business according to the expectations of the buyer. Inter alia, these expectations are about conditions for distribution and production, order accuracy and quality of products or services. Expectations are based upon verbally determined or formally, contractual
established goals and mutual clarified rules in the beginning of the relationship (Das & Teng, 1998).
Behavioural control. Behavioural controls are conducted to specify how the supplier is acting.
In addition, it monitors whether the actual behaviour of the supplier is in accordance with pre- arranged behaviour. Typical behavioural controls are procedures, rules and programs in which for instance the scheduled activities will be controlled and will contain checking devices with regard to product quality (Gulati & Singh, 1998). Das & Teng (1998) argue that behavioural control is important to make sure the supplier behaves desirable.
Social control. According to Dyer & Singh (1998), social control mechanisms are related to joint realization of promises. Often trust is also seen as a mode of social control in inter-firm relationships (Adler, 2001; Ring & Van de Ven, 1992) since social control focuses on the development of shared standards, beliefs, principles, and objectives among firms (Das &
Teng, 2001). Ouchi (1979) used the term of clan control, which is often used synonymously with social control. However, social control can exist when there is agreement on purposes or outcomes, without the required existence of shared belief systems. In other words, there does not need to be a ‘clan’, a collection of individuals within a unique organization. Social control is based on promises and reliance on the partner (Heide et al., 2007), which is also
acknowledged by Carson et al. (2006) who states that inter-organizational trust is a main foundation for the application of social control. Moreover, if a supplier acts opportunistically, the buyer who has been cheated can spread the world through his network, which may
damage the supplier’s reputation. Therefore, good reputation functions as social control by creating an implicit moral obligation for the supplier and increasing likelihood that he will behave in line with his reputation.
2.4 Perceived information sharing
As argued by Doney & Cannon (1997), information sharing between buyers and suppliers is
essential for developing long-term relationships. It is viewed as a major activity in inter-firm
relationships (Bensaou, 1997; Moberg et al., 2002; Krause et al., 2007). According to the principal-agent approach, this concept of perceived information sharing could be interpreted as the information exchange a principal (buyer) perceives to receive from the agent (supplying firm). It bears upon the distribution of information between the buying and supplying firm, incorporating the limited ability to detect opportunism by the supplying party (Van Elten, 2012). An important note to make here: it is not about the real information sharing between the buyer and seller but about the perceived information sharing in the belief of the buyer since the perceived information sharing has influences on the level of trust in the supplier, instead of the real information sharing. Information sharing involves the willingness of the supplier in the perception of the buyer to give useful, relevant and private information of its company (Mahama, 2006).
In most instances, information non-disclosure favours the supplier, thereby leading to an advantage of the supplier (Nayyar, 1990). In line with agency theory, this information non- disclosure has been recognized as a common problem in buyer-supplier relationships in which suppliers usually possess more information than buyers (Mishra et al., 1998). Perceived information sharing could be defined as the buyer’s perception that the supplier attempts to share all relevant information about its business with the buyer (Ceric, 2010). Under
conditions of information concealment, buyers cannot easily distinguish among high and low quality suppliers because low-quality suppliers try to hide their true characteristics to extract unjustified profits, while high-quality suppliers find it difficult to convincingly convey their true quality. Even if buyers try to pre-contractually assess suppliers and product qualities specifically, a true inference can only be made after the purchase has been completed and fulfilled. Therefore, the refusal to share information could also be related to the inability to evaluate quality accurately prior to purchase. In addition, deliberately withholding
information in a relationship between firms is an inadequate behaviour that usually causes misunderstandings, inappropriate expectations, and struggles among buyer and supplier which increases coordination costs and decreases the quality of the relationship (Kwon, 2008;
Morris & Cadogan, 2001).
However, irrespective the potential high commitment in the relationship, information can
never be entirely transparently distributed towards the other parties in a transaction (Mishra et
al., 1998). Information asymmetries exist in all exchange relationships and are specifically
accentuated in buyer-supplier relationships (Yan & Child, 2004) as each supplying firm has
more information about his products and services compared to the buyer who frequently has no insight into how the products are made and what the real knowledge and skills of the employees are. In this situation in which one firm is dependent on the other firm, and if the parties are not equally familiar with the details of the situation, the flow of information is likely to be blocked. Therefore, in buyer-supplier relationships, the lack of sufficient information may make it difficult for the buyer to predict and respond to the supplier’s self- interest seeking action (Glaister et al., 2003). However, this information non-disclosure could also have another reason, suppliers are careful when sharing information about new products because of potential copying behaviour by for instance supermarkets. They are afraid
supermarkets are going to copy the new products under a private label and even launch it to the market at an earlier stage.
Importantly, researchers investigated that distinguishing different types of information needs will help any analysis of information sharing (Tomkins, 2001). Due to the fact that
information sharing could be defined as the extent to which the supplier shares information about its operations and strategies, among others Mentzer et al. (2000) and Moberg et al.
(2002) investigated that information could be identified as either strategic or operational.
They argue that suppliers in the relationship need to share both operational and strategic information in order to increase the success of the relationship and to maximise business performance of the buyer. Lee et al. (2010) defined operational and strategic information sharing as the extent to which the supplying firm intend to share respectively operational and strategic information.
Operational information. Operational information is information about short-term activities of the supplier which has the purpose to obtain operational efficiency. This type of information is related to operational activities including production, inventory and distribution of products and services (Lee et al., 2010). Operational information sharing is about programmed,
organized and routine activities. These activities are not that risky because the activity range is clearly defined, in contrast to strategic information, which will be explained below.
Strategic information. Strategic information is information about long-term activities of the
supplier which has the purpose to achieve strategic goals. These strategic objectives need to
have the intention to create value and profitability for the buyers. This information is among
others related to pricing strategy, marketing strategy, demand prediction, and new product and
technology development (Lee et al., 2010). Despite its benefits for buyers, the disclosure of
strategic information of the supplier may represent a cost for the supplying firm, because it could possibly reveal proprietary information. Because strategic information is about long- term activities, strategic information sharing with buyers is not likely to happen without support of the supplier’s top management. Therefore, information sharing with regard to the strategy requires a high level of top management support.
In the next subsection, the possible relationships between the variables control, trust and information sharing presented in the previous section are assumed. The relationship between perceived information sharing and levels of trust in the supplier is covered in hypothesis 1 and the relationship between information sharing and control mechanisms in hypothesis 2.
2.5 Perceived information sharing and trust
From prior studies it appears that information exchange and trust have been identified as increasingly important with regard to the formation of business relationships (Hofstede et al., 2004; Van Dijk et al., 2003). Scholars agree that information exchanges and trust as sources of competitive advantages should receive more attention (Van Dijk et al, 2003) .The question therefore arises whether and to what extent perceived information sharing of the supplier influences its trustworthiness.
According to Morgan & Hunt (1994) and Rousseau et al. (1998), for developing trust a certain level of transparency is required. Moreover, information sharing by suppliers enables them to signal they are behaving according to the expectations by demonstrating norms of equity and fair dealing (Ring & van de Ven, 1992). The extent to which suppliers disclose information to buyers provides a signal of ‘good faith’, they signal that their intentions could be confidential (Dyer, 1997). Dyer (1997) argued as well that it is through information sharing that suppliers will not violate the social contract by behaving opportunistically, they just signal their willingness to share information which demonstrates they want to be
perceived as an honestly firm. Agreeing upon this view, research on inter-firm relationships investigated that the sharing of information supports the social obligation between partner firms, which would alternately reduce unfair behaviour (Luo et al., 2001). The reduction of this unfair behaviour would signal the supplier’s readiness to behave trustworthy and fairly (Dyer, 1997).
Moreover, perceived information sharing implies that the buyer has the feeling the supplier
provides a good effort to disclose information to the buyer that may be relevant to his
business (Ceric, 2010). This is highly related to the concept of competence trust, which implies that the expectations of the buyer are in accordance with the supplier who has to be able to reliably and constantly deliver particular products or services (Das & Teng, 2001). In addition, for suppliers to acquire goodwill trust, the buyer should recognize that the supplier wants to maintain the relationship in the long term (Sengun & Wasti, 2007). As Doney &
Cannon (1997) argued, information sharing is essential for the development of these long- term trusting relationships.
Based on this previous research, it is proposed that perceived information sharing of the supplier will lead to an increase in the level of both goodwill and competence trust of the supplier. When the buyer firm perceives there is no fair information sharing by the supplying firm, the buying firm would act with restraint.
Hypothesis 1: In buyer-supplier relationships, perceived strategic and operational information sharing of the supplier will both increase goodwill trust and competence trust in the supplier.
2.6 Perceived information sharing & control mechanisms
Information asymmetry could be seen as a primary cause of opportunistic behaviour according to Williamson (1985). Opportunistic behaviour is highly probable when the supplier recognizes that his or her actions, procedures and qualities are not visible to a high extent (Wathne & Heide, 2000). In buyer-supplier relationships, monitoring is important since it enhances the buyer’s ability to detect whether the supplier shirks for instance its quality of products and services. Consequently, suppliers will share their information to a greater extent in order to prevent the perceived opportunism by buyers in their behaviour, which could possibly enhance their reputation. Therefore, monitoring makes it difficult for suppliers to behave opportunistically (Wathne & Heide, 2000). According to Lal (1990) and Wathne &
Heide (2000), the option to apply control mechanisms depends on whether there is evidence that requirements have not been satisfied. In this case, buyers are more or less forced to
control suppliers in order to counteract a potential unacceptable behaviour and to ensure better transparency.
Formal control. In the case of information secrecy of the supplier, there is an increasing need
to monitor and coordinate the performance outcomes and the behaviour of the supplier in
order to alleviate the potential information deficit to a certain extent, and to gain more
knowledge about the operations of the firm (Ahimbisibwe, 2014). Consequently, to reach this higher state of knowledge about the outcomes and processes of the supplier, formal controls mechanisms need to be used in order to protect the relationship and to disclose information (Mayer & Argyres, 2004; Eisenhardt, 1985).
Therefore, we propose that perceived information sharing by supplying firms will negatively influence the usage of formal controls. When information is shared with the buyer, there will be no reason to use formal control mechanisms because there is no information deficit and monitoring becomes possibly unnecessary.
Hypothesis 2a: In buyer-supplier relationships, the higher the perceived strategic and operational information sharing of the supplier, the less will be the usage of output and behavioural controls of the buyer.
Informal control. As argued in the above sections, the supplier’s reputation could be damaged when a buyer recognizes he is cheated by the supplier due to the fact that the supplier acts opportunistically. Creating a kind of social obligation to the supplier could be seen as a social control mechanism, which decreases the probability that the supplier acts opportunistically (Das & Teng, 2001). According to Buckley & Casson (1988), this social obligation could be reached by developing shared values, which is caused by sharing information with the partner.
Moreover, Ring & Van de Ven (1994) argued that social control mechanisms stimulate close connections between partners and are among others exercised through fine-grained
information exchange and try to strengthen cooperation. In addition, also Van der Meer- Kooistra & Scapens (2008) argued that social control is important in encouraging information sharing and interaction among individuals. Moreover, to implement social control, firms must create circumstances that encourage social control (Turner & Makhija, 2006). According to Merchant (1985), this could include among others the provision of information.
Therefore, we propose that perceived information sharing will have a positive effect on the
usage of social control mechanisms by buyers, because social controls are related to solving
disagreements together and joint realization of goals, which will be better realized when the
supplier shares more information.
Hypothesis 2b: In buyer-supplier relationships, the higher the perceived strategic and operational information sharing of the supplier, the more will be the usage of social controls of the buyer.
Figure 1: Theoretical model
3. METHODOLOGY
In this section the methodology of this study is clearly formulated. First the research design is clarified, from which the implications of a mixed method study becomes clear. Then the data collection and sample selection are explained. Furthermore, due to the fact that a mixed- method is applied, both the qualitative research and quantitative research methods are explained. Finally, the reliability and validity of the study are discussed.
3.1 Research design
This study tries to understand further the phenomenon of perceived information sharing in buyer-supplier relationships. This phenomenon has been shown to occur when the buyer perceives the supplier discloses the information which could be relevant for the buyer. With the purpose to better understand this mentioned information sharing perception and its effects on trust and control, this study combines quantitative (dominated by the positivist paradigm) and qualitative (dominated by the constructivist or interpretivist paradigm) research methods.
This combination of research methods could be defined as mixed-method research (Mingers, 2001). Mixed method design refers to the usage of two or more research methods in a single study incorporating elements of both qualitative and quantitative research methods with the aim to deepen the understanding (Teddlie & Tashakkori, 2003). In contrast to multi-method studies, which includes different methods as well, but has one underlying research paradigm, mixed method studies have more than one research paradigm or worldview (Teddlie &
Tashakkori, 2003). Therefore, both interpretivist and positivist research is performed in the same study examining the same phenomena, ostensibly to realize more knowledge.
Though, some researchers argued that there are strong associations between paradigm (positivist or interpretive) and applied methodology, which would make combinations of methods logically impossible. However, Cavaye (1996) argues that the method chosen should depend on what one is trying to research instead of a commitment to a particular paradigm.
Therefore, this research focuses on the nature of the phenomena to be investigated, rather than the epistemology or ontology. Knowledge gained through interpretive (qualitative) research need not be nor cannot be verified by positivist (quantitative) methods. Among others Teddlie
& Tashakkori (2003) proposed that the best applicable epistemology for mixed methods
research will be pragmatism, which could be defined as the ‘philosophy of free choice’. Due
to this adoption of pragmatism, mixed methods research will be the third paradigm (Johnson
et al., 2007) in a trinity of otherwise incommensurable approaches.
The use of mixed methods has several important implications for this study. First, collecting different kinds of data by different methods provides a wider range that may result in a more complete picture of the unit under study than would have been achieved by using only one research method (Bonoma, 1985). Second, the use of mixed methods increases the robustness of results because findings can be strengthened through triangulation. According to
Armstrong et al. (1997), when triangulation of data is used, or in other words the use of mixed methods, then the usage of these several methodologies provides weight to the findings and the validity of the results become greater. Specifically, in this study it is about methodological cross-paradigmatic triangulation (Denzin; 1989; Patton, 1990), which could be referred to as the cross-validation realized when different types and sources of data converge and are found consistent (Kaplan & Duchon, 1988). Each of the combined methods is valuable because every approach provides a different perspective on the topic. Third, each method has its own limitations which can be compensated for by using another method (Brewer & Hunter, 2006).
According to Teddlie & Tashakorri (2012), researchers incorporating mixed research methods are able to use information at the same time (concurrently) or after one method is completed (sequentially). This study contains a simultaneous mixed-method research design, so the information of both methods is used at the same time. The results from the integration of a quantitative and qualitative research method will add up to more than the sum of the parts.
For example, the numerical quantitative evidence of the occurrence of trust, control and perceived information sharing in buyer-supplier relationships is strengthened and made more complete by detailed information on how specific buyers perceive these different aspects of trust, control and information sharing to occur in their relationship with suppliers in the food sector.
3.2 Data collection and sample selection
The unit of analysis for this research were buyer-supplier relationship. With each respondent
from a buyer’s perspective, a specific single buyer-supplier relationship was discussed. Data
will be collected from the buyer’s perspective because the buyer finally decides on whether to
buy from a certain supplier. Therefore, it is elected to seek information from the buyer’s
perspective. The sample frame consisted of purchasing managers of the buying firm or
purchasing managers of a specific product category in this firm. Each purchasing manager
was asked to report on only one relationship with a specific supplier. Although other
employees of the buying firm also could be in the situation to have knowledge about some
facets of a relationship with a supplier, their range of knowledge frequently is restricted. In contrast, logically it is the responsibility of the purchasing manager to be knowledgeable and capable to answer the questions about the supplier posed in the questionnaire and the
interview. The questions should all be responded concerning the supplier in the last purchasing decision with which the purchasing manager was involved. This procedure ensures that the managers will be directed to a specific relationship and they will indirectly focus on relationships with suppliers that had been more efficacious in gaining the business of the buying firm. This also prevents respondent bias to a great extent. However, these
directions also noticed that the respondent should not be concerned if this supplier is no ordinary and usual supplier. The strategy conducted for the collection of respondents is non- probability sampling (Blumberg et al., 2005)
3.3 Qualitative method 3.3.1 Interviews and analysis
For the qualitative method, semi-structured interviews with open-ended questions are
conducted with 8 respondents in buying food firms to explore the phenomenon of perceptions of information sharing and its effects on trust and control in buyer-supplier relationships. In table 3.1 the characteristics of these interviewees are included. The semi-structured interviews and its findings provided detailed information on the ‘how’ and ‘why’ of perceived
information sharing and its effects on trust and control. Therefore, the aim of the interviews is to provide additional insights and to analyse the perceptions of food firms with regard to their suppliers and they will discuss why they have certain levels of trust and control. The findings are not generalizable statistically, although interpretations of and comparisons between interviews contribute to the building of theories, which could be applicable in other
environments. 20 interview questions are included and these are mentioned in Appendix B.
The first 4 questions are about the characteristics of the firm and employee, the next 16 questions are about the variables of this study. The interview format is semi-structured because this allows the a priori concepts to be presented while allowing the respondents to discuss and elaborate freely on their experiences.
The responses are all written down by making transcripts. Delving into the data is reached by
reading the interview transcripts several times and gaining insight that emerged from the
interviews. This process continued until reasonable interpretations became apparent. A
directed approach is applied for content analysis, in which researchers begin by identifying
key concepts as initial coding categories (Potter & Levine-Donnerstein, 1999). The set of codes developed reflected the different themes of the model. The next step in the analysis is coding the passages of the transcripts applying the predetermined codes. Data that cannot be coded are analysed later to determine whether they represent a new category or maybe a subcategory of an existing code.
Then the contents of the transcripts with coded texts are elaborated into a spreadsheet from which discrete directions, distinct quotations and interpretations of the respondents become clear. This spreadsheet is attached in appendix D. From this spreadsheet it becomes clear which positions the respondents take and which opinions they have with regard to information sharing, trust and control. This makes it convenient to come to concrete numbers in the
qualitative results part.
In appendix C, background information and characteristics of the interviews are included. All interviewees were manager or entrepreneur at their company, which meant they were highly involved in interactions with suppliers. Quotations are included in the results section to describe experiences with more eloquence than could be achieved when written in the third person. Maybe 8 respondents does not seem much, however it became clear that the managers mentioned a lot of the same aspects and points. Therefore it could be said that the interviews reached a sufficient level of saturation (Mason, 2010).
3.4 Quantitative method
3.4.1 Variables and measurement
Individual scale items are attached in appendix A. The scale items were all taken from a variety of established scales that were obtained from the existing literature, to ensure sufficient evidence of reliability and validity. None of the items are self-generated. All variables are subjectively measured using a five-point Likert scale, with 1 being ‘strongly disagree’ and 5 being ‘strongly agree’ or 1 being ‘no monitoring’ and 5 being ‘extensive monitoring’. The actual items and key descriptive statistics are attached in appendix A.
Trust. For trust, this variable will be divided into competence and goodwill trust. Goodwill trust will be measured by 6 items, while competence trust will be measured by 3 items. These constructs are adopted from Sengun & Wasti (2007).
Control mechanisms. This factor will be divided into formal and informal control
mechanisms, in which formal control is subdivided into behavioural and output control. These
constructs describe the buyer’s usage of control mechanisms to control the supplier.
Therefore, the measures reflect the buyer’s perception of their usage. The constructs of output, behavioural and social control respectively have three, three and four items. All items for this variable are adopted from previous studies, which are mentioned in appendix A.
Perceived information sharing. The variable perceived information sharing is measured based on the work of Lee et al. (2010) which distinguishes strategic and operational information.
Both constructs include four items.
Control variables. Three control variables are included in the model, which may influence the key variables tested in the survey, as also indicated in table 1.
Length of the relationship. Research indicated that the length of the relationship could
influence the nature of the relationship (Lee et al., 2010). It is defined as the period for which the buyer and supplier form a business relationship. Bensaou (1997) argued that when there is an enduring business relationship, suppliers are more likely to exchange information required for business transactions. Therefore, the length of the relationship will have an influence on the proposed model. Moreover, previous experiences between the organizations can affect the effectiveness of control mechanisms (Fryxell et al., 2002).
Perceived interdependence. Perceived interdependence is included as control variable since high perceived interdependence will strengthen the need for maintaining the stability and the accurate effectiveness of the relationship between exchange partners (Sengun & Wasti, 2007).
Therefore, the relationship between the partners will be influenced when the inter-firm
dependency increases (Mentzer et al., 2000; Noorderhaven et al., 1998) and information of the supplier may possibly be more accurately and freely exchanged so that the preferences of the buyer may be explored (Pruitt, 1981), because both firms have a vested interest in a
continuing relationship (Dyer & Singh, 1998). The perceived interdependence between buyer and supplier is defined as the extent to which the buyer considers the relationship with the supplier as essential. Perceived interdependence by the buyer will be measured by 3 items.
This construct is adopted from Lee et al. (2010).
Relational risk. Das & Teng (2001) argued that relational risk is seen as a problem in inter-
firm relationships. It is the probability of not having satisfactory and effective cooperation and
it is about the likelihood of surreptitious activities. Given this fact, relational risk will have an
influence on the proposed model. Relational risk will be measured by 3 items. This construct
is adopted from Sengun & Wasti (2007).
3.4.2 Survey: sampling & analysis
The aim of the quantitative method is to indicate the responses of 56 respondents of buying food firms with regard to the constructs of the different variables by using a survey. The qualitative method provided findings which informed the choice of answers in the quantitative part. The survey is composed of and based on constructs from previous literature in order to ensure controllability, reliability and validity (Van Aken et al., 2012). Due to the usage of existent research construct items, they have already been tested in previous literature and there is no need to conduct a pilot test. The survey is conducted in the Netherlands in November 2015 utilising food firms in the Netherlands to fill in the questionnaires. The survey is about inter-organizational variables, this means that the behaviour of a second (supplying) organization needs to be measured, which is perceived by a respondent of the first (buying) organization. When the organization gives approval to contribute to the study, the email will be sent to the specific person who is the most appropriate to fill in the
questionnaire. The data is collected over a period of four weeks, after which a single follow- up was conducted in an attempt to increase the response rate. The response rate is 14.5 %, as 386 e-mails are sent to potential respondents. The questions that measure the same variable will be clustered. The data was processed in SPSS in which among others correlation and regression analyses are conducted in order to measure whether there is a significant relationship between the variables in the model.
To test for a potential non-response bias, the early and late responses are compared to check whether there is a difference between these groups with regard to the answers given. This method is chosen because there is a lack of data from non-responding firms, so direct comparison of responding and non-responding firms was not possible (Zou et al., 1997).
Therefore, the wave analysis method is applied to compare the early and late responses (Armstrong & Overton, 1977). This wave analysis method assumes that the respondents who respond less readily are approaching non-respondents (Zou et al., 1997). In this method, an independent samples t-test is conducted to test whether there is a difference between the two groups of fast and slow responses. The t-test analysis is included in appendix F, which shows that there is no between-group difference at the 5% level for any of the variables in this study.
Therefore, it may be concluded that non-response bias is not influencing this research
(Skarmeas et al., 2002).
When only a qualitative approach is applied, it is normal to have multiple persons who are coding the same data. For instance, when only interviews are conducted multiple persons should work on the analysis. However, in this study there are multiple methods (quantitative and qualitative) to confirm the same phenomenon, therefore inter-rater reliability is not critical. The convergence between the methods improves our belief that the results are not a methodological artefact but are obtained in a valid way (Bouchard, 1976).
3.4.3 Sample statistics
In table 3.2 the statistics of the firms are included with regard to their (relative) firm size, relationship length and frequency of interaction with the supplier. The sample consisted of buying food firms that on average have 183 employees and suppliers that on average have 324 employees. Though 1.8 % of the buyer-supplier relationships existed for less than 1 year, 33.9 % were between 10-20 years and 48.2 % were more than 20 years. Furthermore, in 30.4
% of the cases the frequency of interaction was daily, 26.8 % was three times a week, 25.0 % was once a week and the rest was less than once a week. Moreover, table 3.2 B provides information with regard to the product category of the supplier. So taking table 3.2 into account, the resulting sample of buyer-supplier relationships reflects the diversity of the types of relationships in the food industry, except the relationship length which in 82.1 % of the cases existed for more than 10 years.
Table 3.2: Sample statistics (A) and distribution of product categories (B) (n=56) A: Sample statistics
N Mean ± s.d. Percentage
Employees buyer (#)
Less than 50 50-299 300 or more
22 24 10
182.91±310.32 39.3%
42.8%
17.9%
Employees supplier (#)
Less than 50 50-299 300 or more
15 23 18
324.09±494.86 26.8%
41.1%
32.1%
Relationship length (years)
<1 year 1-2 years 1-5 years
1 1 2
1.8%
1.8%
3.6%
6-10 years 10-20 years
>20 years
6 19 27
10.7%
33.9%
48.2%
Frequency of
interaction
Daily
3 times a week Once a week Every 2 weeks 3 times a month Once a month
< Once a month
17 15 14 6 2 2 0
30.4%
26.8%
25.0%
10.7%
3.6%
3.6%
0.0%
B: Distribution of product categories
Product category Percentage Number of respondents Dairy products
Drinks Frozen foods
Fruits, vegetables and potatoes Grocery products and non-food International food
Meat, poultry and fish Salads and meals Bread
Confectionary products & snacks, cookies, bars, baking & pastry Other
10.7 % 14.3 % 3.6 % 12.7 % 16.1 % 3.6 % 14.3 % 3.6 % 7.1 % 8.9 %
5.4 %
6 8 2 7 9 2 8 2 4 5
3
3.5 Reliability
Composite reliability and internal consistency was tested by the Cronbach’s alpha values for each construct. The Cronbach alpha of each construct has to exceed 0.7, this is considered the minimally acceptable level of reliability for instruments (Nunnally & Bernstein, 1994).
Previous research results showed significant reliability measures for the different constructs in
this study due to the fact that all constructs are from previous research. Therefore, the
Cronbach alpha’s are already known in advance and are included in the scale items section in appendix A.
The reliability of the constructs in the current study is also measured. Cronbach’s alpha’s of the constructs are included in appendix E. Due to the fact that this analysis indicated that all Cronbach alpha’s are above 0.7, there is high internal consistency of all subscales. Despite the fact that high Cronbach’s Alpha scores were obtained for all constructs, some constructs benefit from deleting a few items to reach an even higher Cronbach’s Alpha. Therefore, due to the elimination of certain items in some subscales, the value of the Cronbach alpha’s improved to a certain extent.
3.6 Validity
The validity relates to whether the research data accurately describes the investigated
phenomenon (Moisander & Valtonen, 2006) and is necessary to ensure that the findings could be confidential and credible (Merriam, 1995). The validity of the study is presumably high since it is likely that the managers of the buying firms know a lot about their business and have a clear thought about their relationships with suppliers. However, the validity will never be 100 percent, due to the fact that one can never guarantee that the respondents gave all the information they have, for instance since their private reasons or security. Moreover, the objectivity and the accuracy of interpretations of the researcher are key aspects of the validity (Moisander & Valtonen, 2006). Due to the fact that only one researcher is involved who is not checked by a second one, the interpretations of the researcher are not completely objective.
Influential factors like education, experiences and thoughts of the researcher play an important role.
3.6.1 Content validity
Content validity is ensured in this research, since the variables are taken from previous research and are thoroughly reviewed by researchers.
3.6.2 Construct validity
The construct validity is realized when a set of measured items correspond to the theoretical
latent construct they were intended to measure. In order to assess the construct validity of this
research two tests are performed, convergent and discriminant validity (Wong, 2013). To
confirm convergent validity, the Average Variance Extracted (AVE) is examined and a
Confirmatory Factor Analysis (CFA) is conducted for the constructs in the model. For the
constructs to reflect adequate validity, the AVE has to exceed 0.5 (Hair et al., 2007). The values of AVE for the constructs exceed 0.5 which indicates that these values are sufficient to confirm the convergent validity. The AVE values are included in appendix E.
Moreover, through performing a CFA, it is made sure that the items match the proposed constructs and therefore do not cross load across other constructs. Several items of different constructs therefore are removed to strengthen the constructs. For the CFA the factor loadings and the significance of the factor loadings are examined. The factor analysis is included in appendix E. To test whether the totality of correlations between items is low, the Bartlett’s test of sphericity is used. A significant value indicates that the correlation coefficients between the items significantly differ from zero. If this is the case, it can be concluded that the inter-correlations are not too low and a factor analysis is useful. The Bartlett’s test of sphericity resulted in a significant chi-square of 941,361 (df=210, p<0.001) which means that the model is significant. Furthermore, a Kaiser-Meyer-Olkin measure of sampling adequacy (KMO) is used to determine whether the data is appropriate for a factor analysis. This resulted in a KMO of 0.742. According to Field (2009), a value above 0.7 for the KMO test is
considered as good.
For the factor analysis, there are several possibilities with regard to the factor rotation. Given the kind of items an oblique rotation is chosen, from which is expected that the factors in which they are classified correlate. According to Costello & Osborne (2005), oblique rotation provides a theoretically more accurate result compared to an orthogonal rotation. Importantly, for oblique rotation it is not absolutely necessary for factors to correlate, as many researchers think. Moreover, all kind of oblique rotations produce similar results (Costello & Osborne, 2005). Therefore, randomly the Promax rotation is selected, in which the kappa is set to four because this is the best choice according to Costello & Osborne (2005). Changing this value makes interpreting these results much more complex.
In addition, discriminant validity is examined. For discriminant validity to be confirmed, the square root of AVE for each individual construct has to exceed the correlation between the latent constructs (Fornell & Larcker, 1981). This table is included in the results (table 4.1).
Due to the fact that for all constructs the square root of the AVE exceeds the inter-
correlations, this Fornell-Larcker criterium is acquired. Concluding, discriminant validity is
confirmed.
3.7 Common method variance
Due to the fact that the data is self-reported, there is an opportunity for the existence of common method variance (CMV). These variances are attributable to the measurement method, instead of the constructs the measures represent (Podsakoff et al., 2003). Common method biases are the major threat for the internal validity. To take this issue into account, several ex-ante remedies are used to prevent this variance. Especially, measurement items are carefully designed, in the sense of item wording, format scales and reverse coded items.
Moreover, the introduction of the questionnaire assured respondents that their supplier and the answers given with regard to this supplier are anonymous. Due to the fact that the answers needs to be based on one specific supplier, this prevents the possibility of giving socially desirable answers and being lenient in how they think the researcher wants them to respond (Tourangeau et al., 2000). Furthermore, to test for common method variance, the Harman’s Single Factor technique is applied as ex-post remedy (Podsakoff & Organ, 1986). With this test a factor analysis is conducted on all dependent and independent variables. An unrotated factor structure has been applied with only one factor to extract. This new common latent factor may not explain more than 50% of the variance. This factor explains 47.06%. So although that is a lot of variance to be explained by a single factor, it is not the majority. So common method bias is not present in this study.
3.8 Multicollinearity
A test for multicollinearity is performed to determine whether the constructs are truly different
from each other. The underlying reason to prevent multicollinearity is to make sure that the
variables are not overlapping in a significant way and thus act independently from each other. The
measured levels of variance inflation factors (VIF) can be found in the tables 4.1 until 4.5 in
the results section. The VIFs suggest that the independent constructs hardly overlap with each
other due to the fact that they are all below 3.0, meaning multicollinearity does not play a
significant role in the analysis (Farrar & Glauber, 1967).
4. RESULTS
This section presents results from the qualitative and quantitative analyses, yielding important insights into the context of perceived information sharing, trust, and control. The two research questions will be examined: ‘How do perceptions of information sharing (strategic and
operational information) influence the level of trust (goodwill and competence) in buyer- supplier relationships?’ and ‘How do perceptions of information sharing (strategic and operational information) influence the usage of control mechanisms (formal and informal control) in buyer-supplier relationships?’ For both research questions, a combination of quantitative and qualitative evidence is obtained in order to answer the questions. First, the descriptive statistics will be reported and second the hypotheses have been tested, including the control variables. An overview of the results is presented in appendix G.
4.1 Correlation and descriptive statistics
Descriptive statistics and the correlation matrix for the variables in this study are presented in table 4.1. From this correlation analysis it appears which variables are significantly associated with each other. To clearly indicate which variables are control variables, these are
distinguished by an interrupted line in the correlation matrix. As the Likert scale obtained in this research could be interpreted as interval data, the Pearson’s correlation coefficient is sufficient. Pearson’s correlation coefficient allows to indicate the strength of the linear relationship between the variables in this study (Hart, 1999).
Positive and significant correlations are generated between operational information sharing
and goodwill trust (r=.612, p<.01), competence trust (r=.636, p<.01) and social control
(r=.576, p<.01). Moreover, strategic information sharing correlated significantly with
goodwill trust (r=.517, p<.01), competence trust (r=.653, p<.01) and social control (r=.438,
p<.01). The relationships between both strategic and operational information and the formal
control variables all are close to zero so the Pearson correlations did not lead to an expected
relationship between strategic and operational information sharing and the formal control
variables (output and behavioural control). Consequently, regression analysis will be
conducted to ultimately accept or reject the hypotheses.
Table 4.1: Means, standard deviations, VIF, correlations and the square root of AVE for constructs
4.2 Hypotheses testing
A hierarchical multiple regression analysis is conducted in order to find if there is a causal relationship between the variables (Cohen & Cohen, 1983). Moreover, the interviews are content analysed in order to examine the directions of the interviews with regard to the variables under study and to investigate the underlying reasons for these directions.
For conducting the regression analysis, perceived strategic and operational information sharing are independent variables, while competence and goodwill trust and output,
behavioural and social control mechanisms are dependent variables. The regression analysis is conducted by including the three control variables in the first block and the independent variables in the second block, in which the control variables are included with the stepwise method. This allows to check whether the control variables influenced the model
significantly. Hierarchical multiple stepwise regression requires a minimum ratio of valid cases to independent variables of 5 to 1 (Schwab, 2007). Therefore, with 56 valid cases and Variable Mean (s.d.) OIS SIS GWT CT OC BC SC ID RR LR
OIS 3.67 (.85)
.800ªSIS 3.63 (.79)
.660** .860ªGWT 3.53 (.77)
.612** .517** .732ªCT 3.76 (.77)
.636** .653** .740** .780ªOC 4.06 (.89)
.024 -.099 -.164 -.181 .906ªBC 2.71 (1.20)
.085 -.087 .041 -.066 .412** .924ªSC 3.66 (.92)
.576** .438** .420** .401** .067 .130 .898ªID 3.26 (.82)
.223 .361** .011 .122 .031 .212 .127 --RR 2.41 (.82)
.75* .235 .111 .121 -.062 -.041 .105 .409 --LR 5.18 (1.08)
.126 .122 -.165 -.002 .133 -.044 .208 .118 .101 --** = Correlation is significant at the 0.01 level (2-tailed).
* = Correlation is significant at the 0.05 level (2-tailed).
ª = Square root of Average Variance Extracted (AVE) n = 56
