Project Delivery Systems and Project Performance: A Causal Relationship Analysis of Indonesian Stakeholders

1.1. Parameters of PDS

The organization of stakeholder interactions and project timeframes for effective facility development is facilitated by the PDS [12]. The Project Delivery Institute in Moore [13] asserts that establishing roles, responsibilities, and activity sequences is essential for success in a project. The implementation of PDS has a substantial influence on the pace, expenses, excellence, and administration of building projects [14-17]. Choosing the wrong PDS can negatively impact project performance [18]. According to Gajurel [19], the DBB approach is a conventional procurement procedure in which contractors submit bids based on complete blueprints. The DB approach allocates design and construction responsibilities to a single company [20]. The construction manager at risk (CMAR) approach encompasses the involvement of a development team across the whole project lifecycle, providing a high degree of flexibility [21]. The DBB methodology continues to be extensively employed on a global scale, including in Indonesia, for public projects [18, 19, 22-26]. A survey reported (Fig. 1) that 42% of construction in Indonesia relies on DBB, followed by DB and CMAR. Despite the integrated project delivery system (IPDS) being used (5%), a separate contract system is still used, which is unique in Indonesia, where BIM defines IPDS.

According to Morton and Thompson [27], the IPDS and standard PDS undergo stages from conceptualization to construction. However, IPDS distinguishes itself by focusing on early design decisions, contributing to improved efficiency and cost-effectiveness through collaborative endeavors. IPDS, DB, DBB, and CMAR are distinct contract types that differ in their stages and legal elements.

1.2. Variables in PDS Aspects

Previous studies have identified PDS aspects globally [12, 19, 20, 28-35], but these aspects have not been realigned with Indonesian ones to offer thorough iterations on the elements that contribute to and enhance project performance. Therefore, this study first confirms the presence of PDS aspects and identifies commonalities, despite Hansen [8] and Yasin [7] categorizing them by technical, legal, and administrative considerations in Indonesian construction contracts (Table 1). Afterward, it examines PDS variables as independent variables (x).

1.3. Project Performance Indicators

Project performance indicators have been developing over time. For instance, Zimmermann [32] identified safety as an essential prerequisite for the achievement of a PDS, whereas Barnes and Wearne [36] highlighted the importance of cost, quality, and time [37]. According to Freeman and Beale [38], project managers still rely on their instincts regarding these criteria despite attempts to change this behavior. This research examines the dependent variable (y) through a comprehensive literature analysis. The variables considered include time, cost, quality, safety (K3), user satisfaction, party satisfaction, function, environmental performance, and profits [39-49]. Ongoing research and development are being conducted to explore and create additional indicators, which may vary based on the project's size and complexity.

2.1. Data Collection using Semantic Differential Scale Questionnaire

Data was gathered from the owner, designer, QS/QE, construction management, contractor, subcontractor, and operations as individuals. The participants were not Indonesian committees or organization members. Individuals consented to be respondents, and the data was presented anonymously. Thus, explicit consent was not needed. As demonstrated above, quantitative variables from direct content literature were utilized to develop questions using a 5-point semantic difference scale to measure each variable's impact on project performance [52]. The snowballing sampling was employed to target architecture engineering construction participants involved in building construction projects across Indonesian cities [53, 54]. Despite not representing all construction members, the quantity of respondents strongly implies the initial preferences that will be examined using linear regression in the preliminary study [55]. This technique took several stages to decrease bias. Questionnaires were sent to numerous initial connections to get non-relative respondents with different viewpoints and experiences. According to Andrieux et al. [56], this approach reduces bias. Second, the respondents who provided linear replies were identified, assuming the questions in Table 3 are irrelevant. Finally, Cronbach's alpha reliability assessment was carried out to check all replies for linearity. However, this preliminary phase requires more research with a larger sample size to support and cross-validate the findings.

This study did not account for the duration of work experience, which could serve as a confounding variable for the subsequent study, and regression could be conducted using ANOVA. After identifying underlying latent variables that explain the observed correlations among multiple variables, factor analysis was carried out [57]. Multivariate regression analysis was then performed to examine the relationships between various independent variables and a dependent variable [58].

2.2. Analysis of Data Distribution

Data were gathered from 102 respondents between October 2022 and March 2023 across various cities in Indonesia. The distribution of respondents by city, ranked by percentage, is as follows: Jakarta 45%, Bandung 32%, Bandar Lampung 7%, with the remaining respondents from other cities, as shown in Fig. (2a). As stated before, this data is initial, and more extensive research is required to encompass several regions that have not been included. The predominant respondents were contractors, constituting the highest percentage, succeeded by D/E. The additional percentages of other stakeholders are depicted in Fig. (2b). Distribution analysis (fit y by x) was carried out among stakeholders regarding the use of PDS types to examine which type is most widely used based on the percentages given in Fig. (3).

The figure shows the distribution of development projects throughout Indonesian cities, with Jakarta as the leading city (Fig. 2a).

The most prominent stakeholders in construction projects include contractors (38%), designers/engineers (28%), owners (16%), construction managers (14%), operations (3%), and quantity surveyors/quality control (1%) (Fig. 2b).

The highest usage of the DBB type was based on D/E experience, making up 16%, while the DB type was primarily derived from contractor experience, accounting for 15%. Based on the owner, the most frequently used PDS type was DBB, which was 8%. For IPDS, it was found that 2% and 3% of contractors and operations have had experience using this type of PDS. The results of this distribution test are very significant, with a Prob> Chisq value of <0.0001.

The most experienced DBB users on construction projects are designers/engineers (16%), contractors (14%), CMAR for construction management (8%), and IPDS for operations (3%) (Fig. 3).

The assessments of stakeholders of the lowest-performing stage of building projects were evaluated based on actual issues. The participants were queried regarding the phase that presented the most significant challenges impeding the project's progress. Fig. (4) shows the design/engineering (D/E), pre-construction, and construction phase results. These results indicated potential difficulties that require additional discussion and thorough scope clarification in future studies (Prob> ChiSq = 0.0048).

2.3. Data Reliability

To assess the reliability of the utilized component data, a Cronbach's α reliability test was conducted, employing a coefficient range interpretation as follows: 0.7 to <0.8 (good), 0.8 to <0.9 (very good), and 0.9 (excellent). This evaluation preceded the multivariate correlational analysis stage [58]. The reliability outcomes for the PDS aspect variable yielded an average Cronbach's α of 0.8484 (indicating very good reliability). At the same time, the project performance variable exhibited a Cronbach's α of 0.8982 (also denoting very good reliability).

2.4. Factor Analysis (FA), Principal Components Analysis (PCA), and Multivariate Regression Analysis

After data reliability was established, principal component analysis (PCA) and varimax rotation were carried out in factor analysis (FA) to identify latent variables from nine PDS aspects and project performance variables. Three factors, which exceeded a cumulative percentage of 75%, showed project performance, while four components represented PDS aspects. This approach maintained the independence of variables. Using JMP Pro, correlation and causation linkages were investigated between PDS aspect variables (χ) and project performance variables (y). This robust PCA effectively replaced missing values based on the patterns observed. In the meantime, utilizing JMP Pro's “explore missing values” showed zero missing results.

3.1. Regression of PDS Aspect Variables on Project Performance Variables

In the next step, the regression analysis was performed between variables χ1-4 and y1-3. The results revealed a regression relationship with an R-squared value of 0.26, a parameter β value of 0.4108, and a significance level (p-value) of <0.0001*, mainly observed in the management aspect of the quality aspect (Table 6).

This phenomenon elucidates that the management aspects, encompassing project management, organization/team, documentation of activities, and project completion, are pivotal in enhancing project performance, particularly regarding quality aspects, such as quality and K3. The subsequent significant cause-and-effect relationships (regression) were sequentially observed in the legal and technical implementation aspects toward the satisfaction aspect, the management aspect toward the satisfaction aspect, and the execution method aspect toward the quality aspect. This association indicated that these three criteria significantly affect Indonesian building project performance. The owner's legal and technical obligations are often unclear, resulting in extra work during the project. This suggests that the feasibility study and design are not mature enough to be the leading cause.

3.2. Regression of PDS Aspect Variables on Project Performance Variables Based on PDS Type

Regression analysis examined the link between project performance and PDS aspect variables for various PDS types. It sought to determine how the project development strategy elements, considering the particular PDS in use, predict project performance. According to Azhar et al. [23], this PDS type promotes efficiency and cooperation, which are critical for excellent project performance. The results demonstrate a cause-and-effect link in PDS-DB's legal and technical aspects (Table 7).

Moreover, in the PDS-DBB type, the management aspect has the highest regression. This aligns with the findings of theoretical studies, as discussed by Azhar et al. [23], which emphasized that the DBB type, characterized by multiple entities, exhibits a higher level of fragmentation and may not inherently promote teamwork. This underscores the significance of prioritizing elements, such as project management, organization/team, documentation of activities, and collaborative efforts in project completion to enhance project performance within the PDS-DBB context. This supports prior results that contractual system-isolated stakeholder connections are often the cause of problems in terms of diverse interests and objectives. Indonesian national, organizational, and professional cultural diversity affects stakeholder relationships. These differences can affect numerous parties’ communication, decision-making, and conflict resolution [59, 60].

The regression analysis of the impact of PDS aspects on quality aspects demonstrated the strongest causal relationships, particularly between management aspects related to the quality aspects of PDS-DB type (β 0.51 and a p-value of 0.0025*) and DBB type (β 0.43 and 0.0121*) (Table 8). Previous studies have stressed the importance of management, supervision, and collaboration in attaining high-quality project outputs, especially when the contractor's management engagement begins during design [23]. In the Indonesian context, project control techniques like internal control systems are important for quality management. Indonesian higher education institutions have found that internal control systems promote communication, cooperation, and leadership, which are necessary for quality improvement [61].

The inverse regression coefficient (β -1.52, p-value 0.0138) of IPDS exhibits value-based commonalities rather than individual pursuits in older PDS, thus supporting the reward aspect of the theoretical framework by AIA [33]. In typical PDS rewards, this standard of transparency is lacking, with only increasing penalties for lower labor quality (the opposite of PDS rewards) being standard offerings. However, IPDS incentivizes parties to share their experiences with fines, prizes, and compensation. This statistic, which accounts for only 5% of respondents, highlights the impact contract compensation and rewards have on project quality. Indonesia's construction industry is booming and growing in workload. The challenges and long working hours faced by construction workers in Indonesia can lead to work-family issues, so improving motivation concerning incentives is important [62]. However, IPDS's small sample size (N=5) makes the regression unreliable, necessitating further causality research.

PDS-DB type has substantial causal correlations between legal and technical implementation (β 0.45, p-value 0.012*) and management aspects (β 0.33, p-value 0.043*), which affect project success and cost. The reward aspect correlates with a parameter β value of 0.42 and a significant p-value of 0.018* (Table 9). This suggests that rewards and sanctions improve project performance, especially cost, in the Indonesian DBB type.

These findings are significant in the context of Indonesia, as Dewi et al. [63] reported comparable results in a specialized construction firm, along with two additional studies conducted in Indonesian general firms and in developing nations [64-66]. The result corroborates the traditional premise that incentives can enhance performance, but they must be meticulously designed to prevent negative consequences, supporting the theory of motivation [67, 68]. DBBs struggle with unclear stakeholder compensation and rewards. The legal component of the contract turns this ambiguity into cost-driven, individualized endeavors to avoid late punishments or fines, aligning with the findings of a study by Yasin [7]. The subsequent regression model illustrates the causal relationship between PDS aspects and performance aspects (Fig. 5).

Fig. (5). PCA derives four latent variables (x) representing PDS aspects and three latent variables (y) representing project performance aspects, with the highest regression on appreciation towards quality β -1.5178 with statistically significant results.

The regression model of execution method aspects (parameter=0.6757) is IPDS-specific. However, the regression's p-value is not significant, and the number of respondents (n = 5) (5% of 100%) for the IPDS type might be a confounding factor in determining significant causality from other aspects. However, this IPDS factor is still necessary as a comparative variable for causal relationships with conventional PDS. Obtaining causal facts in further research may be challenging, as this IPDS is still rarely utilized as a construction contract in Indonesia. With a larger IPDS sample, like DBB and DB, the p-value may become substantial, corroborating research on collaborative project management as a risk-reduction method [33]. On the other hand, this compensation and reward phenomenon can be a long-term risk for DBB if it is not addressed in a legal contract, for which it is obvious that IPDS is the mitigation solution. According to regression research, PDS quality has the highest R-squared value and contributes most to project performance. This indicates the extent to which management influences the current condition of Indonesian construction and project quality.

3.3. Regression of PDS Aspect Variables on Project Performance Variables Based on Stakeholders

Regression analysis of the impact of PDS aspect variables on project performance variables by stakeholders examines the relationship between these variables. This analysis examines how the PDS affects and predicts project performance, taking into account stakeholders' perspectives and roles. This regression analysis, focused on stakeholders, identifies aspects that significantly impact project performance from their perspectives. The regression model analysis identified significant connections between distinct stakeholder groups, which are as follows (Fig. 6):

3.3.2. D/E

a. A significant relationship was demonstrated between legal and technical implementation aspects and satisfaction and cost aspects.

b. A significant relationship was found between management aspects and quality aspects.