Research on the Influencing Factors of Trust Networks of Construction Project Participants

Abstract

The latest research on the governance of construction projects has centered on the cooperation efficiency among project participants and the enhancement of project performance. Trust networks can exert a remarkable role in this aspect, facilitating the cooperative behaviors of all project participants and achieving project value. Nevertheless, the facilitating factors of trust networks remain unclear, and the traditional classification factors of trust relationships between binary subjects have not been verified from the perspective of relationship networks. Hence, this study carried out a structural equation model analysis on 139 valid questionnaires to disclose the influence mechanism of trust types and trust networks, while taking into account the moderating effect produced by the objective situation of asymmetric dependence. The research indicates that cognition-based and affect-based trust positively affect the trust networks. Affect-based trust on trust networks is positively moderated by asymmetric dependence. Based on this, the trust networks research in construction project governance has established a distinct set of influence paths, and relevant governance measures will be systematically implemented.

1. Introduction

The mainstream research on construction project governance has shifted from a purely transaction cost economics perspective to a stage of integration between economics and sociology [1]. This is because a simple economic research paradigm is incapable of explaining the complex embeddedness of social relationships, and resolving such issues is crucial for significantly enhancing the cooperation efficiency of project participants and greatly improving project performance. Admittedly, previous studies related to contractual governance or informal governance have recognized the impact of the amount of social capital and acknowledged the significant role of the interaction relationships among project participants [2]. However, they are often confined to qualitative analyses and lack empirical research. Thus, which factors can facilitate the improvement of cooperation efficiency among project participants? What is the influence mechanism? How can such governance approaches be applied in the practical process of construction projects? The responses to these queries are precisely what this study intends to address.

The completion of a construction project is a lengthy process involving numerous participants. During the phases of engineering design, tendering, and construction, the main participants include the owner, the engineering design studio, independent enterprises engaged in cost consultation, construction, engineering supervision, etc., as well as building material suppliers and various administrative departments. All of them are key participants capable of determining the success and degree of value realization of the project. Previous studies have indicated that governing the contractual or informal relationships among the aforementioned project participants is beneficial for the improvement of project performance. However, the governance perspective should encompass both dyadic and network relationships in order to incorporate all project participants into a unified governance system and formulate a fair and reasonable management plan. In this respect, the trust network, as the core means of network relationship governance, has had its significant efficacy emphasized and verified by numerous previous studies. Among the project participants, there exists an objective relationship state of trust or distrust, and these relationship connections interweave to form a networked structure, namely the trust network structure. From the perspective of social network theory, the trust network of project participants objectively possesses quantitative indicators such as network density and centralization potential, as well as characteristics such as network stability and structural holes.

Accordingly, the embeddedness perspective emphasized by the social network theory thus exerts a distinctive role. That is to say, within the ambience of the trust network, the behavior or action of any participating party is inevitably influenced by the form of the network structure it is situated in. In actuality, numerous previous studies have disclosed the influence mechanism of the trust network on various behaviors such as opportunism, conflict, and cooperation, and thereby can impact the performance of engineering projects. Nevertheless, these studies have overlooked an important premise, namely, they have not yet clarified the source of the trust network of the project participants, or rather, which factors can affect the trust network. In this aspect, the influencing factors of the trust relationship between the two parties of the construction project’s owner and contractor have received considerable discussion, including but not limited to various key factors such as the environment, institution, emotion, regulation, and capability that affect the trust relationship. However, the trust relationships among various project participants from the perspective of trust relationship networking are not completely consistent, and thus the influencing factors of the trust relationship are also not identical. In order to obtain universal project governance strategies within the analytical framework of the social network theory, it is highly necessary to explore the common influencing factors that can affect the trust relationships among various project participants, that is, the intersection or the greatest common divisor of the sets of various influencing factors. This is precisely the key influencing factor of the trust network that this study strives to reveal.

2. Theoretical Background

2.1. Trust Network Structure Among Project Participants

Previous relevant research on inter-organizational trust has differentiated between the binary subject and multi-relationship perspectives. Traditional studies have emphasized the core status and crucial role of the principal–agent binary relationship [3]. Whereas, from the multi-relationship perspective, the latest mainstream research has focused on the trust network formed among the participants of construction projects. For example, some scholars have mainly focused on the social network structure in which individuals are located in real economic transaction activities and analyzed the interactive influence of this structure on trust relationships and individual behaviors. Some researchers in this aspect have introduced the social relationship distance among individuals in the real world into the trust game experiments based on uncertainty event preference induction. The researchers selected groups of people with common group characteristics and those with very close relationships in real life to participate in the preference-induced trust game experiments and obtained some results that were significantly different from those of pure stranger experiments, indicating that trust relationships can significantly affect the outcome of the game.

Research along the aforementioned lines generally points out that the behavior of individuals within a social network will be significantly affected by the trust relationships they have with other individuals in the network [4]. This influence is mainly manifested in aspects such as the tendency to implement behaviors and behavioral decision-making of individuals. In fact, it is rather complex and difficult to explore trust relationships within the social network structure constituted by the various participants in economic transactions. Therefore, previous research in this aspect is relatively rare. The reason for this can be considered that trust relationships are difficult to be accurately described or quantitatively calculated, and the interactive influence mechanism within the social network structure is also rather complex. Thus, it requires researchers to delve into specific management contexts and conduct concrete or condensed research highly related to the trust relationships among the various participants in their specific management contexts [5].

It is notable that mature research outcomes regarding trust network measurement have emerged in recent years, enabling the measurement of the trust network structure among the construction participants of engineering projects within the framework of social network theory. Thus, the embeddedness issues and multi-party collaborative cooperation problems that previous studies on binary trust relationships were unable to solve can now be provided with novel governance solutions from the perspective of trust networks. In this aspect, existing research in the recent period has emphasized that the trust network among construction participants can impact cooperation efficiency, opportunistic behavior, and project performance. However, the source of trust networks remains in a black box state; that is, previous studies have tacitly assumed that trust networks are a form of relationship structure that objectively exists or spontaneously forms among the construction participants of engineering projects. Scant research has revealed their specific sources, thereby making it difficult to implement corresponding governance systems or management countermeasures in a targeted manner. This constitutes the key difficulty that this study intends to solve.

2.2. Multiple Types of Trust Relationships Among the Participants of Construction Projects

Admittedly, trust relationships are an inescapable topic in real social life and economic transaction activities, but their sources and the influence mechanism on relationship networks remain undefined [6]. In the research domain of construction project governance, many scholars recognize that the trust among project participants is an inter-organizational trust, and this concept represents the positive expectation that the participants believe that the other party will not exploit their weaknesses for personal gain [7]. Regarding its sources, it can be classified into three aspects: trust based on institutions, trust based on cognition, and trust based on affects, or trust based on integrity, trust based on capability, and trust based on institutions [8]. These are the rather mainstream classification approaches for the sources of binary trust, frequently observed between construction project owners and contractors or between owners and design enterprises. Put differently, among different participants in construction projects, there exist differentiated classifications of trust sources. The current available research achievements have not yielded a unified source of trust relationship applicable to all construction participants; they merely consider various types of trust relationship sources as a subset of the entire set of trust relationships.

However, as a complex construct, when the analytical perspective enters the trust relationships among the various participants in the social network, there is a dearth of robust evidence [9]. Previous studies have regarded the trust network structure as an objective existence among the participants in engineering construction, neglecting the common trust roots and individualized mutual trust factors of different participants. In fact, considering the trust network as an objective existence is to view the social network relationships among the participants from a macroscopic perspective, much like the ubiquitous cosmic microwave background radiation, which is a socialized trust ambience that does not require self-certification. Research indicates that this trust ambience is widespread but of low density and intensity [10]. In other words, numerous factors can influence the trust relationship among the participants in an engineering project; that is, the sources of trust exhibit diverse classification situations. Some of these factors might affect the trust relationship between the project owner and contractor, while several others might impact the contractor and subcontractor. The fundamental aim of this research is to sift out the crucial influencing factors from the set of these influencing factors that can exert an effect on the trust relationship of all participants. Or, to state it differently, to identify the intersection or the greatest common factor of the factors that can influence the trust networks from the set of numerous factors.

Previous studies have indicated that during the course of several years or even over a decade of work communication among the participants of construction projects, they are required to achieve an efficient cooperation state in order to realize the project value and reach the project performance goals [11]. However, a low-density and low-level trust ambience is hard to serve as the cornerstone of efficient cooperation. Hence, it is necessary to explore the specific sources of the networked structure of trust relationships and formulate corresponding governance strategies to realize the efficient cooperation among all the participants of the project.

2.3. Asymmetric Dependence Among the Participants of Construction Projects

Inter-organizational dependence constitutes a core construct within the research purview of the resource dependence theory (RDT), highlighting that participants in economic transaction activities are not isolated [12]. Instead, the disparity in resource demands should be taken into account throughout the entire transaction process. In specific empirical research, inter-organizational dependence is frequently decomposed into two forms, namely joint dependence and dependence asymmetry. In the domain of construction project governance, “asymmetric dependence” is more prevalent and is employed to describe the differences in resources possessed or demanded by each participant in economic transaction activities [13].

Specifically, the participants involved in a construction project are numerous and have distinct professional skills and resource advantages, specifically manifested in aspects such as project supervision, engineering design, civil construction, mechanical and electrical equipment installation, cost accounting, organization and management of construction personnel, and supply of construction materials. In engineering practice, almost no single participant can master all the technical capabilities and resources of the construction project. Thus, objectively, it is necessary for numerous participants to collaborate efficiently to complete the construction project and achieve outstanding project performance. From the perspective of transaction cost theory, the specialized division of labor among the participants of an engineering project can reduce transaction costs, mainly learning costs. Therefore, professional skills, information, and resources are individually mastered by various enterprises, which can also be regarded as one of the theoretical roots of the asymmetric dependence in engineering practice. In reality, asymmetric dependence exists objectively among the participants in construction projects. Hence, incorporating it into the research domain is an inevitable tendency.

A considerable number of previous studies, proceeding from the perspective of “structural holes” in social network theory, have probed into how the participants in crucial and core positions within networked organizations affect the profits of other non-key participants, and the moderating effect of inter-organizational dependence has been effectively verified in the above-mentioned mechanism [14]. Other scholars, through relevant research on inter-enterprise cooperation risks and opportunism, have discovered that asymmetric dependence has a significantly negative moderating effect on the relationship strength and risk-sharing relationship among all parties. This indicates that maintaining an appropriate relationship strength between enterprises is conducive to achieving risk-sharing, and moderate trust and reciprocity will have an inhibitory effect on the moderating effect presented by asymmetric dependence. Therefore, when considering the networked structure of trust relationships among the participants in construction projects, the impact brought about by asymmetric dependence is inevitably involved [15]. Specialized analysis in this aspect will be conducive to enhancing the cooperation efficiency of the participants. In the practice of construction engineering, the party that holds resource superiority often imposes psychological pressure on the disadvantaged party, subsequently causing the resource-disadvantaged party to exhibit proactive defensive behavioral strategies such as suspicion and doubt. This objectively impedes the formation of trust relationships and undermines the implementation of cooperative actions. Whether considering such issues from the perspective of the asymmetric dependence effect of the resource dependence theory or the structural hole analysis of the social network theory, a relatively consistent real situation emerges; that is, because the resources held by the participating parties are objectively different, all parties need to collaborate to complete the engineering project. However, the relatively resource-advantaged and disadvantaged parties among them will have fluctuations in trust relationships. This study incorporates the objective situation of asymmetric dependence into the analytical framework and endeavors to explain its specific influence path on the network structure of trust relationships.

3. Model Development and Hypotheses

It can be verified through the theoretical background that the trust network among project participants originates from the trust relationship links among the participating parties. However, whether the trust relationships among each participant are influenced by the same or similar factors remains in an unknown state. Therefore, this study proposes the conceptual model as shown in Figure 1, with the aim of exploring the mechanism of the source of trust relationships on the trust network and simultaneously considering the objective influence of asymmetric dependence on the project participants.

3.1. Set of Factors Affecting the Trust Network

Previous studies have indicated that the trust relationship is a complex and multi-dimensional concept. In construction projects, there are numerous participants, and enterprise-level mutual trust relationships exist among them. In reality, the key factors influencing the trust relationships among them are not the same. The set of these factors comprises calculative, informational, identity-based, self-interested, social, and affective trust relationships. The trust between the project owner and contractor may be affected by several factors within the set, while the trust relationship between the contractor and subcontractor is influenced by other different factors. This research endeavors to analyze the common influential factors in the trust networks, that is, to identify the key intersection of factors from the set of influential factors that can exert an impact on the trust relationship links of all participating enterprises.

Different types of trust will impact the trust relationships among the participants in engineering projects. Such interwoven trust relationships, which may or may not exist and may be strong or weak, form a networked structure, thereby emerging trust network densities of varying degrees. Past studies have indicated that system-based trust, cognition-based trust, and affect-based trust can all positively influence trust relationships [16]. The microscopic influence mechanism herein requires specific elaboration. Namely, the types of participants in construction projects are diverse, and the trust relationships among them can all be categorized into the aforementioned three types. Otherwise, it would be challenging to conduct efficient questionnaire surveys or measure them within the same scale. In this respect, a considerable number of previous studies have pointed out that whether it is the owner, the contractor, the material supplier, or the design and engineering consulting enterprises, their trust relationships can all be classified as system-based trust, cognition-based trust, and affect-based trust. Thus, this study has the following hypotheses:

H1a: 

System-based trust exhibits a positive correlation with the trust network.

H1b: 

Cognition-based trust demonstrates a positive correlation with the trust network.

H1c: 

Affect-based trust shows a positive correlation with the trust network.

3.2. Asymmetric Dependence as an Objectively Existing Situational Factor

As a crucial concept within the resource dependence theory (RDT), inter-organizational dependence is regarded as an expansion and extension of the resource dependence theory. The resource theory states that all organizations acquire power by managing resource dependence relationships, which is also its core assumption. In previous studies, numerous scholars have indicated that the asymmetric dependence relationship between the two transaction parties will significantly escalate the distrust and conflicts between them, which will cause substantial damage to their cooperation efficiency and satisfaction with the cooperative relationship. Or rather, the objective existence of asymmetric dependence reduces the buyer’s trust level in the supplier, concurrently increasing the possibility of inter-enterprise conflicts and ultimately lowering project performance. In construction projects, there are objectively existing superior and inferior participating parties. The manifestations of these resources can be in aspects such as funds, technology, information, or labor force, thereby forming an asymmetric dependence relationship among the participating parties of the engineering project.

Meanwhile, each participant in construction projects holds heterogeneous resources and capabilities. During the cooperative interaction process for completing the project, there objectively exists a situation of dependence asymmetry, that is, one party needs to rely on the resources or information of the other party to fulfill its own work tasks. In this case, although there are diverse types of trust relationships, the asymmetric dependence of one participant on other participants will influence the persistence or strength of its trust relationship. Research indicates that when there is asymmetric dependence between the two parties, the trust relationship between them will be weakened. Thus, it can be considered that asymmetric dependence negatively regulates the trust network structure. Thus, this study has the following hypotheses:

H2a: 

Asymmetric dependence negatively moderates the influence of system-based trust on the trust networks.

H2b: 

Asymmetric dependence negatively moderates the influence of cognition-based trust on the trust networks.

H2c: 

Asymmetric dependence negatively moderates the influence of affect-based trust on the trust networks.

4. Methods

4.1. Questionnaire Design

The questionnaire survey method was used to collect data, and the items of the survey scale were all from mature items of previous similar studies [17]. All the measurement items were measured on the 5-point Likert scale (1 = strongly disagree; 5 = strongly agree), and

Table 1. Items of measurement and reliability and validity analysis.

Construct	Variable	Item No.	Mmeasurement Items	Loading	Cronbach’s α	CR	AVE
Trust types	System-based	SB1	An organizational resource guarantee system that meets the needs of the participating parties is of great significance.	0.882	0.893	0.923	0.751
		SB2	The management system should be clearly specified for solving cost, time, risk, and safety issues.	0.900
		SB3	Establishing a stable system for project information exchange is of great significance to the project.	0.867
		SB4	A good communication system can prevent ambiguous situations and risky matters from occurring.	0.815
	Cognition-based	CB1	Frequent and proactive communication with other construction parties is conducive to enhancing mutual trust.	0.936	0.925	0.952	0.87
		CB2	Regular participation in exchange meetings related to engineering is conducive to better mutual understanding among the construction parties involved.	0.940
		CB3	Mastery of financial resources is an important factor in evaluating the credibility of the participating construction party.	0.922
	Affect- based	AB1	Showing concern and care for other construction parties at the appropriate time can enhance a sense of trust.	0.889	0.81	0.887	0.723
		AB2	Taking into account the needs of each party in the decision-making process will lead to compromise and satisfactory results.	0.832
		AB3	Spending appropriate time and energy to understand the business background of other construction parties can eliminate the sense of distrust in the work.	0.829
Trust networks		TN1	There is a general relationship of mutual trust between the various parties involved in the project.	0.836	0.859	0.904	0.703
		TN2	The participants have always tried to show that they are to be trusted.	0.869
		TN3	All parties involved agreed that the relationship of trust had a significant impact on the project.	0.820
		TN4	Project information is often shared efficiently, and trusting relationships emerge between the participants.	0.828
Asymmetric dependence		AD1	Key resources and information are always centrally controlled by individual participants.	0.802	0.881	0.918	0.737
		AD2	The inaction of one participant will profoundly affect the work of other participants.	0.875
		AD3	There are key participants in the project that can influence the working or relationship climate.	0.878
		AD4	Temporary replacement of any participants will have a negative impact on the project.	0.875

presents the items. In order to improve the questionnaire items, four experts were invited to discuss them. They represent the government, academia, engineering consulting, and design companies, respectively. All of them have a master’s degree or above and have participated in many construction projects with many years of work experience. According to the feedback from the experts, we revised the initial measurement to improve the readability and validity of the questionnaire [18].

Specifically, in light of the revision suggestions proffered by experts, the definitions of trust network and asymmetric dependence were concretized and referred to in the questionnaire. Meanwhile, the definitions of the three types of trust were also elaborately depicted and expounded in the questionnaire. Among them, the trust network investigated in this study is the reticular structure formed by the trust relationship links among the participants in the engineering project. Thus, based on the theory of social networks, the network density of this structure, namely the degree of the number of the trust relationship links within a certain range, can be delineated [19]. Additionally, system-based trust indicates the trust arising from the various existing project management systems; cognition-based trust refers to the trust engendered by the mutual communication or exchange of information and resources among the participants; and affect-based trust primarily describes the trust generated in accordance with the trustworthy elements manifested by the participants themselves [20]. Concurrently, asymmetric dependence mainly implies that due to the disparity in the resources held by the participants, one party has to rely on the other party to provide resources for the completion of the project. This objectively existing state is a relationship context factor [12].

4.2. Sampling and Procedure

The snowball sampling method was used in this study. Previous studies have shown that the core participants of construction projects mainly include various government departments, engineering design and management consulting enterprises, financial enterprises, private investment, and construction enterprises. Based on this fact, we asked 11 engineering directors involved in major infrastructure or public facilities projects to send questionnaires to the participants of the projects they had worked on that they thought were competent. The researcher will explain to the respondents through the questionnaire that this study is only for academic purposes, and the information they fill in will be kept confidential. They will then be asked to continue to invite other participants who are qualified to fill out questionnaires for this research objective.

A total of 210 questionnaires were collected from July to December 2024. Respondents need to confirm that they have participated in the development, construction, or operation process of a construction project so as to ensure that they can understand the meaning of each item in the scale and give an exact answer. The forms of the questionnaire were divided into electronic and paper versions. After eliminating the questionnaires with incomplete answers, obvious arrangements of answers, or contradictory answers, 139 valid questionnaires were obtained, and the effective recovery rate was 66.19%.

Table 2. Characteristics of respondents.

Background Characteristics	Frequency	Percentage
Enterprise type
Representative of the host government	36	25.90%
Local suppliers of materials or equipment	38	27.34%
Local financing support agencies	19	13.67%
Local third-party consulting agency	24	17.27%
Local contractors or subcontractors	13	9.35%
Others	9	6.47%
The country of your enterprise
P.R. China	114	82.01%
Others	25	17.99%
The continent where your project is located
Asia	121	87.05%
Others	18	12.95%
Work experience
<5 years	71	51.08%
5–10 years	43	30.94%
>10 years	25	17.99%
Job position
Manager of the headquarters	28	20.14%
Project/department manager	43	30.94%
General management/technical personnel	68	48.92%

presents the characteristics of the respondents.

During the process of conducting the questionnaire survey, this study, in order to circumvent the drawbacks of the snowball sampling method, such as rapid sample saturation, a narrow respondent circle, and a lack of universality, has directly dispatched questionnaires to the employees of enterprises that are currently or have formerly participated in international projects through multiple differentiated channels, including international academic conferences or training courses related to international engineering. Moreover, they have been invited to provide contact information of the next suitable respondent for receiving the questionnaire, with the hope of enhancing the internationalization level of the respondents in this manner. Nevertheless, after the final screening, the respondents of the valid questionnaires obtained still predominantly consist of employees from Chinese enterprises, and the international projects covered in the questionnaires mainly focus on the Asian region, both accounting for more than 80%.

Consequently, the data analysis and conclusion summary of this study primarily target international engineering projects in the Asian region, and these projects often involve the participation of enterprises from China. Admittedly, this to a certain extent restricts the universality of the conclusions drawn from this study. However, the Asian region, including China, encompasses numerous developing countries, where there is an immense demand for construction projects. Moreover, domestic enterprises alone are unable to fulfill all the tasks required for major construction projects. Whether in terms of financing channels, construction techniques, or engineering design plans, extensive international cooperation is indispensable. Thus, although constrained by the issue of a single respondent country resulting from the snowball sampling method, the conclusions obtained from this study can effectively support the governance of international engineering projects in developing countries.

4.3. Construct Reliability and Validity

This research utilizes the method of structural equation modeling for model testing and hypothesis validation. In contrast to the traditional general linear equation model, structural equation modeling exhibits greater multiplicity and inclusiveness. Furthermore, there are two prevalent analysis techniques in structural equation modeling, namely covariance-based structural equation modeling (CB-SEM) and partial least squares-based structural equation modeling (PLS-SEM). This research selects the PLS-SEM approach primarily because it possesses superior statistical efficacy compared to CB-SEM and demonstrates preferable computational accuracy when confronted with small sample non-normal distribution data [21]. In this aspect, SmartPLS software can present more outstanding performance in processing PLS-SEM models with small sample data compared to other similar software. Therefore, this research undertook specific data analysis tasks via the SmartPLS software.

As shown in Table 1, the square root of the average variance extracted (AVE) for each factor is larger than the maximum value of the correlation coefficients between that factor and the other factors; it can be considered that the scale has good discriminative validity. Then, factor analysis of KMO value and Bartlett spherical test index data was carried out by SPSS 24 software, and the results showed that KMO = 0.827, approximate chi-square value = 1509.964, df = 153, Sig = 0.000, and cumulative variance of interpretation = 76.05%. This indicates that the data obtained in this study are suitable for factor analysis [22]. SmartPLS 3.3.2 software was used to analyze the valid data collected by the scale [23]. The variables shown in Table 1 and

Table 3. Correlation matrix and the square root of AVE of factors.

Items	AB	AD	CB	SB	TN
AB	0.850
AD	0.210	0.858
CB	0.320	0.412	0.933
SB	0.001	0.500	0.414	0.867
TN	0.314	0.206	0.241	−0.108	0.838

Bold values on the diagonal represent the square root of AVE.

are validated. It can be considered that the scale used in this study has good reliability, and the obtained survey data have good validity [24].

4.4. Test of the Structural Model and Hypotheses

The global goodness-of-fit (GoF = $\sqrt{\overline{Communality}\times \overline{R^2}}$) is an important index to measure the goodness of fit of PLS-SEM [25]. After calculation, the GoF value in this study was 0.38, greater than the optimal standard value (GoF = 0.36), indicating that the model has a good goodness of fit standard. After the goodness-of-fit test, SmartPLS 3.3.2 software was used to carry out the analysis of the bootstrapping method [26]. The samples were 5000 times more recommended by previous research findings. The results of testing are shown in

Table 4. Results of hypotheses (H1) testing.

Path (Moderating)	Standard Path Coefficient	T-Statistics	p-Value	Confidence Intervals		Correlation	Hypothesis	Inference
				2.50%	97.50%
SB → TN	−0.234	1.769	0.077 N.S.	−0.494	0.078	/	H1a	/
CB → TN	0.223	2.245	0.025 *	0.062	0.439	Positive	H1b	Supported
AB → TN	0.169	1.971	0.049 *	0.037	0.380	Positive	H1c	Supported
SB (AD) → TN	−0.046	0.855	0.392 N.S.	−0.049	0.406	/	H2a	/
CB (AD) → TN	−0.035	0.361	0.718 N.S.	−0.309	0.045	/	H2b	/
AB (AD) → TN	0.132	3.012	0.003 *	0.123	0.438	Positive	H2c	Not Supported

Note: * means p < 0.05; N.S. means nonsignificant.

, where the p-value represents the level of significance. In general, the smaller the p-value, the better the significance. Therefore, it can be judged that some hypotheses are not supported by data [27].

Specifically, hypotheses H1b and H1c are supported by the results of data analysis. Cognition-based trust exerts a positive influence on the trust network, with a path coefficient of 0.223 (p-values = 0.025, and the confidence interval does not encompass 0). Concurrently, it can be posited that affect-based trust also has a positive effect on the trust network, with a path coefficient of 0.169 (p-values = 0.049, and the confidence interval does not contain 0). It can be argued that these two types of trust stem from the communication and interaction processes in practice and therefore can exert an impact on all the participating parties in the construction project. That is to say, the positive interactions among the participating parties of the construction project can facilitate the trust network, and these positive interactions mainly comprise cooperative behaviors, non-opportunistic behaviors, performance behaviors, etc. Corresponding project governance strategies can be initiated from these aspects and exert influence. For example, an information real-time exchange platform based on building information modeling (BIM) is established through network technological means, enabling numerous participating construction parties to intuitively observe the progress of the project. During this process, each participating construction party is required to communicate about the practical problems that occur during the work implementation, promoting the formation of a trust relationship based on cognition. Simultaneously, during the project implementation process, all participating construction parties can be regularly invited to jointly participate in informal interaction activities, mainly group activities such as sports and literary and artistic ones, to facilitate the formation of affect-based trust.

However, hypothesis H1a was not upheld by the data analysis outcomes; that is to say, system-based trust failed to exhibit its influence on the trust network in this research. A potential reason for this result is that the questionnaire survey of this study was directed towards construction engineering enterprises, which have been subject to relatively systematic organizational management constraints for a long time. Whether it is the administrative governance systems or the internal management systems of the projects, they have all become the common rules and endogenous and spontaneous behavioral habits that the project participants habitually abide by. Hence, the system-based trust derived from the administrative governance systems or the internal management systems of the projects is unable to significantly affect the trust behavior decisions of each participating party in construction and even the overall trust network structure in practice.

Furthermore, to verify the moderating effect of asymmetric dependence in this process, drawing on similar practices from previous studies, a structural equation model was constructed and fitting tests were carried out. The measured sample data of this study were imported into the model for analysis, and the analysis results are presented in Table 4. The analysis results suggest that among the three paths where trust factors might have an influence, asymmetric dependence only produced one kind of moderating effect, namely, it exerted a positive moderating role in the influence path from affect-based trust to the trust networks. Considering that the aforementioned content of this study has already indicated the positive effect of affect-based trust on the trust networks, that is to say, the objective existence of asymmetric dependence can facilitate and enhance the positive effect of affect-based trust on the trust networks. Nevertheless, the data analysis results in this regard are not in line with hypotheses H2a to H2c of this study; that is, none of these three hypotheses are supported by the data analysis results.

In reality, the analytical outcomes of this study concerning hypothesis H2c present a contradiction to the previous theoretical analysis. According to the resource dependence theory, the party in a resource-disadvantaged position has no option but to rely on the advantaged party; otherwise, the disadvantaged party would be unable to accomplish its engineering construction tasks. This passive reliance on the advantaged party constitutes asymmetric dependence, and such a situation typically impedes the establishment or persistence of a trust relationship. Nevertheless, the data analysis results of this study indicate that affect-based trust within the context of the trust networks is positively modulated by asymmetric dependence. The possible reason for this particular outcome might be that affect-based trust mainly represents emotional or psychological trust relationships. In the objective circumstance of asymmetric dependence, the influence of subjective factors such as emotions and psychology is instead accentuated. That is, the objective disparity would enhance the effect of subjective emotions, thereby facilitating the formation of a more stable trust relationship. Conversely, system-based and cognition-based trust were not modulated by asymmetric dependence, meaning that hypotheses H2a and H2b were not supported by the data analysis results. This could perhaps be because the sources of these two types of trust mainly rely on objective administrative systems or internal management systems of the construction project. Therefore, this objectivity determines that regardless of the presence or absence of an asymmetric dependence situation, the advantaged party should provide the necessary resources to the disadvantaged party as stipulated by the system, or in other words, the resource-advantaged party should comply with the system and refrain from oppressing the disadvantaged party, thereby significantly attenuating the modulating effect of asymmetric dependence.

5. Results Discussion and Implications

5.1. Cognition-Based and Affect-Based Trust Positively Affects the Trust Networks

The analytical outcomes of this research present an intriguing result, namely that the impact of system-based trust on the trust networks is not affirmed by the data analysis results. It can also be perceived that within the purview of this study, the influence of system-based trust on the trust networks is not remarkable. This thus contradicts the previous relevant studies from the binary perspective of the contractor and the owner. A large number of previous studies contend that system-based trust is the core source of the fundamental trust relationship during the project governance stage of engineering projects. Nevertheless, when the analytical perspective enters the networked structure of trust relationships composed of numerous participating parties, the analytical results of this study point out the ineffectiveness and impotence of system-based trust. The underlying reason is that among the participating parties of construction projects, there is no explicit institutional design or market mechanism enabling mutual trust among all parties. Each participating party merely converges incidentally to complete the construction project, without an inevitable foundation of mutual trust or the initial trust resulting from institutional arrangements.

Specifically, the trust networks structure emphasized in this research is constituted by the trust relationships linking the participants in the engineering project. Previous studies have highlighted the significant influence of system-based trust on the trust relationship between the project owner and contractor but have failed to confirm its degree of influence on the trust relationships among other participants. The data analysis results of this research indicate that system-based trust does not significantly and positively affect the trust network structure. It can be considered that the trust relationships among participants other than the project owner and contractor have not been affected by it. This implies that the existing management systems in engineering practice are difficult to govern all the trust relationships among the project participants. Hence, it is necessary to refine the institutional arrangements in this aspect to ensure that they can exert an effect on the numerous participants in the engineering project, such as implementing detailed project master control plans or participant management outlines.

However, the analytical results of this study indicate that cognition-based and affect-based trust positively affect the trust networks. These two types of trust stem from the interactions and communications during the project implementation process. Therefore, as the work tasks of each participating party advance, a certain degree of mutual trust will be generated during the process of communication or information exchange among all parties. Such mutual trust deepens as the project progresses and ultimately affects the trust network formed by all participating parties of the project. Hence, establishing an information interaction platform based on artificial intelligence technology or building information modeling technology can effectively enhance the information exchange and resource interaction among all participating parties, promote communication to a certain extent, and thereby sustain the stability of the trust network among all participating parties of the project, which in turn can enhance cooperation efficiency and achieve project performance.

5.2. Affect-Based Trust on the Trust Networks Is Positively Moderated by Asymmetric Dependence

Inter-organizational dependence is an objectively existing condition among the participants of construction projects. Whether it is joint dependence or asymmetric dependence, both are frequently occurring interaction states among the participants. Previous studies have emphasized the influence of inter-organizational dependence, particularly asymmetric dependence, on the cooperative behavior of project participants and project performance. However, the perspective of this study focuses on the trust network of project participants. The data analysis results obtained indicate that the moderating effect of asymmetric dependence is not obvious. This is because the mutual trust relationship among the participants mainly arises in the process of communication and information interaction. Therefore, regardless of whether there is a dependence relationship among the participants, a trust relationship can be formed through communication and frequent interaction. Thus, when attempting to establish the trust network of project participants in construction projects in practice, there is no need for excessive intervention in the asymmetric dependence relationship between organizations. However, when the information interaction is sufficient and the communication channels are well-developed, the trust networks will gradually form spontaneously.

It should be noted that the data of the project participants collected in this research are sourced from regions with nearly identical cultural backgrounds, thereby resulting in the conclusion that the moderating effect of asymmetric dependence is insignificant. Nevertheless, when the cultural backgrounds of the project participants vary, perhaps different conclusions will be reached. Since existing studies have demonstrated that participants with dissimilar cultural backgrounds hold distinct viewpoints on trust, risk, resource disadvantages, opportunism, or cooperative behavior and make disparate behavioral decisions. Then, when a construction project involves participants from different cultural backgrounds, the impact of cultural factors on the trust relationship among the participants still requires consideration, or the possibility of a significant asymmetric dependence situation due to the differences in cultural background should be accounted for.

5.3. The Research on Trust Networks in the Governance of Construction Projects Has Established a Distinct Line of Thought

Traditionally, studies on the governance of construction projects have primarily concentrated on the binary perspective of the principal–agent relationship, mainly exploring the routes to enhance cooperation efficiency and project performance from the viewpoints of contractual governance and relational governance. Nevertheless, apart from the owner and main contractor, other participants in construction projects can equally exert a profound influence on project value and success. To compensate for the deficiencies of traditional research, studies based on social network theory that investigate the cooperative behavior of various project participants and project performance have emerged. Regarding this, existing research has indicated the influence of trust networks on the opportunistic behavior of construction project participants, which subsequently affects cooperation efficiency. Subsequent related studies have followed this theoretical thread to explore the impact path of trust networks on project performance, encompassing both empirical research and research outcomes using modeling and simulation, all of which have provided beneficial research conclusions [28]. Moreover, previous related studies have explored the influence of trust networks among construction project participants on inter-organizational conflicts and discovered that they can suppress conflicts and improve project performance.

Furthermore, some scholars have carried out research on the measurement of the existing network and formulated mature scales and questionnaires. Based on the aforementioned research findings, this study has identified the fundamental influence of the heterogeneity of trust types on the trust networks. Hence, a clear influence path can be obtained, which constitutes an important theoretical contribution of this study, as depicted in Figure 2. Subsequently, based on the discoveries of this study and in conjunction with previous related research achievements, a series of project management systems can be established to facilitate the formation of trust networks among the participants of construction projects.

6. Conclusions

This study uncovers the primary sources of the trust network among the participants in construction projects, representing a significant milestone. Previous studies have indicated that establishing a trust network in the domain of project governance can enhance project performance from multiple perspectives, such as cooperation efficiency, opportunistic behavior, and inter-organizational conflicts. Nevertheless, they have failed to clearly determine the sources of the trust network, thereby creating obstacles for the application of governance measures. The analytical results of this study demonstrate that cognition-based and affect-based trust can significantly facilitate the trust network, and the project governance scheme based on this can effectively address the challenge of improving project performance.

During the implementation of this study, certain limitations were identified. For instance, due to the constraints imposed by regional sampling and static trust measurement, the generalizability and universality of the conclusions drawn from this study are somewhat restricted. Additionally, given the use of the snowball sampling method, there is a potential risk of self-reporting bias in the questionnaire responses.

Moreover, construction project management modes exhibit considerable diversity, with significant variations in participants’ relationships across different management modes. For example, in the Design–Bid–Build (DBB) construction management mode, there is a direct contractual relationship between the owner and other participating parties, but there is no direct contractual relationship constraint among other participating parties. Thus, there are no systematic initial conditions for the formation of trust among them, and their trust network structure is gradually formed during the project implementation process. In contrast, previous studies have indicated that in the Public–Private Partnership (PPP) model, the prerequisite for the establishment of cooperative relationships among all participants is mutual trust [5]. Therefore, this type of project has the initial state of the trust networks. Moreover, the trust network structure of the participating parties in the Engineering–Procurement–Construction (EPC) model is also rather unique. In this model, both the project owner and the general contractor are in the position of structural holes in the network structure. Hence, the trust networks in this mode often present a more obvious centralization tendency. In conducting this study’s questionnaire survey, scale data originating from various project management modes were not rigorously differentiated. Consequently, it is possible that distinct research conclusions may emerge under varying management modes, leaving this area open for further exploration in subsequent studies.