Structural Equation Model (SEM)-Based Productivity Evaluation for Digitalization of Construction Supervision

Abstract

The construction industry continues to face declining productivity due to its heavy reliance on labor and the repetitive, non-value-adding nature of supervision tasks. This study provides an exploratory, practitioner-based evaluation of how selected digital technologies, PDF-based documentation systems, object recognition algorithms, and 3D vision technology may contribute to productivity improvements in construction supervision. A total of 82 valid responses from field engineers were collected to examine perceived task substitution effects across major construction work types and management functions. The findings indicate that higher work-adoption rates of digital technologies are generally associated with improved supervisory productivity, with the strongest perceived benefits observed for PDF-based documentation in reinforced concrete and formwork tasks. However, other expected relationships, particularly those involving work responsibility, did not appear consistently in the practitioner data, suggesting that such perceptions may be influenced by task context and adaptation burden. This study offers a practical and context-specific framework for understanding how digital tools may support productivity enhancement in supervision work. While the results reflect tendencies based on a limited sample, they provide field-grounded insights that can inform the phased and targeted application of digital technologies in construction supervision and guide future empirical model development.

1. Introduction

Rapid population aging is a global phenomenon driven by declining birth rates and increased life expectancy. The continuously rising proportion of people aged 65 and older has become a major factor contributing to the reduction in the labor force. In particular, in the construction industry, the proportion of workers aged 55 and over increased from 22.7% in 2013 to 37.2% in 2023 (an increase of 14.5%), indicating a faster aging rate compared to other industries such as manufacturing (an increase of 11.3%) and agriculture, forestry, and fisheries (an increase of 1.3%) [1]. Moreover, the construction industry is more heavily dependent on its workforce than other sectors, making productivity management increasingly crucial [2]. Therefore, there is an urgent need to devise measures to address the reduction in the construction labor force.

Over the past decade, the productivity of the construction industry has been relatively low compared to other industries. Specifically, supervisory tasks in construction projects have been characterized by repetitive and non-productive nature, often leading to greater labor demands and adversely affecting overall project performance, particularly in manpower intensive operations. Therefore, analyzing and responding to productivity decline in the construction industry is an urgent priority. Effective productivity management and enhanced efficiency in labor-intensive tasks are essential for the construction industry [3]. Improving labor productivity, in particular, can have a profoundly positive impact on the construction sector as a whole [4].

In recent years, there has been a noticeable shift away from traditional management methodologies towards the adoption of Construction Management (CM) techniques that leverage digital technologies to increase productivity. Indeed, numerous studies have identified increased productivity as one of the most commonly cited benefits of adopting technological advancements such as modern equipment, materials, and information technologies [5,6]. However, despite these efforts, the construction industry’s broad use of the productivity concept limits its practical application at construction sites, largely due to the qualitative and context dependent factors influencing productivity. To address these persistent challenges, new solutions such as technological innovation are necessary [7].

With the Fourth Industrial Revolution, various digital technologies are revolutionizing production systems and work methods across industries, including construction [8]. To keep up, the construction industry must adopt digital technologies in supervision. By applying these technologies on-site and considering practitioners’ insights, effective strategies can be developed to offer practical and realistic solutions.

Despite the growing interest in the adoption of digital technology in construction supervision, significant research gaps remain. Previous studies have primarily focused on the implementation of individual technologies or provided broad industry-level analysis, yet few have developed comprehensive evaluation models that integrate multiple digital technologies across various construction types and management functions. In particular, most existing research lacks empirical validation through field-based surveys of practicing professionals, thereby limiting their practical applicability. Furthermore, although international studies have explored digital transformation in construction, there is a lack of research specifically addressing the Korean construction supervision context, which features unique regulatory frameworks and industry practices.

Therefore, there is a need to further adopt digital technology into supervision work, and an effective strategy for their implementation must be formulated based on practitioner’s insights and field level applications.

This study seeks to conduct a productivity evaluation using a Structural Equation Model (SEM) approach to support the optimization and improvement of digitalization in construction supervision It aims to assess how the implementation of digital technologies can substitute for traditional labor tasks and enhance worker productivity. Framing the adoption of digital technologies as replacement for conventional labor tasks, this study establishes its scope and methodology to logically analyze and interpret the relationship between digital adoption and productivity improvement, rather than to statistically verify model validity. Therefore, the scope of this study is limited to evaluating the productivity of construction supervision through the adoption of digital technologies, emphasizing field-level applicability rather than theoretical development. By focusing on practitioner-based insights and on-site conditions, the study aims to provide a practical and context specific framework that addresses the actual needs of the construction supervision industry during its digital transition.

2. Literature Review

2.1. Review of Previous Studies

Existing studies, as summarized in

Table 1. Summary of Previous Studies Related to Technology Adoption in Construction Supervision.

Year	Title	Author(s)	Research Content
2024	Smart Construction Supervision Solution for the Digital Transition of Supervision Tasks	Kim T.H et al., [9].	Introduced digital supervision solutions to improve efficiency and reliability in analog-based supervision processes.
2024	Development of Digital-Based Construction Supervision and Automation Robot Technologies: Technology 4 Robotic Automation for Construction Supervision	Park J.B et al., [10].	Proposed Scan-to-BIM and robotic automation methods for optimized supervision and monitoring.
2011	A Study on the Development of BIM-Based Construction Supervision Process Models	Kim S.J [11]	Analyzed BIM-based supervision models and interrelationships among supervision process variables.
2021	From BIM to Digital Twins: A Systematic Review of the Evolution of Intelligent Building Representations in the AEC-FM Industry	Deng et al., [12].	Reviewed the evolution from BIM to digital twins and proposed a roadmap for future implementation.
2024	A Study on the Establishment of an Integrated Smart Platform for Construction Sites	Shin D.H et al., [13].	Identified limitations in current smart platforms and suggested integration strategies for on-site application.
2024	A Study on the Establishment of an Integrated Smart Platform for Construction Sites	Gharaibeh L et al., [14].	Reviewed BIM ROI evaluation methods and proposed a framework for quantifying BIM investment value.

, have primarily focused on developing smart supervision systems, automation platforms, or pilot-level applications of digital tools. While these efforts have contributed to technological advancement, most studies have not sufficiently reflected field-level applicability or practitioners’ perspectives. Moreover, the research emphasis has largely remained on technological feasibility rather than the empirical verification of productivity outcomes.

To address these limitations, the present study emphasizes a field-based productivity evaluation that integrates practitioner feedback and empirical analysis. This approach bridges the gap between theoretical and on-site research, ensuring that the proposed framework not only supports academic understanding but also provides actionable insights for real-world implementation.

2.2. Review of Digital Technologies in the Construction Industry

Digital technologies have evolved beyond mere tools and are now acknowledged as important drivers of added value [15]. A systematic literature review analyzing 175 journal articles published between 2001 and 2020 examined the application of digital technologies in the construction industry [16]. Their study identified 26 digital technologies and classified them according to their role within the construction process, such as data acquisition, analytics, visualization, communication, and design and construction automation. These digital technologies include, but are not limited to: artificial intelligence (AI), machine learning, or deep learning [17], augmented reality (AR), virtual reality (VR), or mixed reality (MR) [18], big data analytics [19], blockchain technology [20], building information modeling (BIM) [21], cloud-based technology [22], computer-aided design (CAD) [23], computer vision (CV) [24], design for manufacturing and assembly (DfMA) [25], drones/unmanned aerial vehicles (UAVs), global positioning systems (GPS) [26], the Internet of Things (IoT) [27], radio frequency identification (RFID) [28], robotic technology [29,30], sensor technology [31], and 3D printing or additive manufacturing (AM) [32]. Research into the advantages of digital technologies has consistently reported positive outcomes; however, the adoption of technology in the construction sector remains slower than in other industries [6]. Key barriers to adoption include industry reluctance to embrace change [33], concerns surrounding innovation risk [34], and challenges related to data and information sharing, regulatory compliance, and data ownership [35].

Digital technology can be applied to four categories of supervision work in construction: measurement, records management, document/data management, and site weather provision [36]. This study identified the three types of supervision work, measurement, record management, and document and data management, excluding site weather provision, which has less direct relevance to supervision work. In order to replace these works utilizing digital technology, it is essential to secure detailed technological elements. In particular, for implementing digital construction supervision work, the following technological elements can be utilized: mobile technology, drawing information technology, field information acquisition technology, and field information recognition and judgment (intelligent) technology and platform-linked reporting technology [37].

Measurement work involves measuring the length, angle, and horizontality of construction members. This can be implemented with object recognition algorithms and 3D vision technology, which are vision technologies that identify, inspect, and analyze objects in images. By developing these technologies, it is possible to reduce repetitive work and improve measurement reliability through automatic measurement of construction members. Records management and document/data management tasks involve creating and managing various document tasks performed in the field. These tasks can be implemented with PDF construction documentation system, which converts construction manuals into PDF format so that they can be accessed through tablet PCs or mobile devices, displays information in the documents, and shares them with stakeholders. This provides existing paper-based data on mobile and provides convenience by digitizing document creation and management, thereby facilitating the review of construction drawings and detailed evaluations of construction methods during supervision.

As a result, this study defined three technologies, object recognition algorithm, 3D vision technology, and PDF construction documentation system as feasible technologies for utilization. These three technologies have been mentioned in numerous studies as technologies fundamentally and universally applicable in supervision work. Therefore, it is considered significant to study the impact of the introducing these technologies on improving supervision work productivity.

2.3. Theoretical Framework for Technology Adoption

Examining models related to the adoption of technological advancements within organizations can help understanding how such advancements are adopted and what factors determine their uptake [38]. Three theories are widely regarded as influential in the field of digital technology adoption: diffusion of innovation (DOI), technology-organization-environment (TOE) framework, and technology acceptance model (TAM).

The DOI theory, initially developed in the field of communication and later extended in 2003 [39,40], explains how innovations, encompassing ideas, products or technologies gradually gain momentum and spread across social systems or specific populations. Meanwhile, the TOE framework postulates that a firm’s application and utilization of technological innovations depend technological, organizational, and environmental contexts. The technological context includes the firm’s existing tools, processes, and the extent of technology accessible externally, while organizational context encompasses factors such as the firm’s size, structure, degree of centralization, and human resources. Lastly, the environmental context refers to the competition, sector, and regulatory landscape of the firm’s operation [41]. Finally, the TAM framework, initially introduced [42], and enhanced in 1993, focuses on individuals’ technological acceptance by taking into account perceived usefulness and ease of use. The model progresses through steps involving information collection, technology adoption, and new technology implementation. These theories have been applied in recent research efforts with consistent relevance [43].

Although these theories have differences, they all share the common goal of assessing the extent to which technology is integrated into traditional work practices. This can explain that technology adoption is influenced by workforce and time, which in turn affects productivity and work responsibility. Hence, in this study, the adoption rate of digital technology will be comprehensively defined as the level to which digital technology is implemented into construction supervision work.

Collectively, the DOI, TOE, and TAM frameworks provide a comprehensive lens through which to interpret how digital technologies are adopted and utilized in construction supervision. Within this conceptual foundation, work adoption rate corresponds to the technological readiness and perceived usefulness dimension, work responsibility represents organizational and psychological factors affecting task execution, and productivity denotes the resulting performance outcome achieved through digital adoption.

These frameworks together explain how technological integration influences operational efficiency and support the establishment of a model that empirically examines the causal relationship between technology adoption and productivity improvement in construction supervision.

2.4. Initial Hypothesis Development and Variable Definition

Hypothesis Setup

The objective of this study is to conduct a productivity evaluation for the digitalization of construction supervision. To enhance the reliability and interpretive clarity of evaluation, a SEM-based approach was utilized as a supportive analytical framework to explore potential directional tendencies rather than to provide definitive statistical validation. Accordingly, before conducting the research, it was necessary to establish hypotheses to examine the potential relationship between digitalization and productivity. In this study, digitalization is defined as the adoption of digital technologies, and hypotheses were formulated to investigate whether such adoption may be associated with productivity. Research on the relationship between “productivity” and “work adoption rate” in construction supervision has been conducted from multiple perspectives. Technologies including materials and information technology have notably shown the potential to influence construction labor productivity in recent years [44]. Particularly, automation and system integration have been reported to contribute to substantial productivity improvements. This has prompted arguments for mechanization in various industries as a means of improving productivity [45].

Combining these research findings suggests that technology adoption may be linked to productivity improvement across industries. Increased task substitution through technology adoption has generally been observed to correlate with productivity gains. However, to explore the potential need for digital technology adoption based on this correlation, it remains important to examine whether a similar tendency may appear between increased task substitution rate identified in previous studies and productivity enhancement. Therefore, based on the reviewed literature, the following hypotheses were formulated to evaluate the relationship between technology adoption (work adoption) and productivity in construction supervision using relevant variables.

H0. 

Increasing work adoption rate does not affect productivity improvement.

H1. 

Increasing work adoption rate affects productivity improvement.

3. Methodology

3.1. Research Method

The methodological flow of this study is summarized as follows. First, the research background, objectives, and scope were defined to establish a practical framework for evaluating productivity in construction supervision. Second, through a literature review, digital technologies applicable to supervision tasks were identified, and relevant theoretical foundations were reviewed to support the analytical approach. Third, variables and hypotheses were defined to analyze the relationship between digital technology adoption and productivity improvement. Instead of focusing on statistical validation, a Structural Equation Model (SEM) approach was applied as an evaluative framework to explore internal tendencies and interpret possible inter-variable patterns [46,47,48]. To complement the SEM analysis, supplementary correlation and regression analyses were also conducted to examine the directional tendencies of associations between digital adoption and productivity. Fourth, a structured questionnaire survey was designed and distributed to experienced field practitioners. The survey assessed expected productivity improvements associated with the adoption of three digital technologies (PDF documentation system, object recognition algorithm, and 3D vision technology) across five construction work types and seven management functions. Fifth, reliability and descriptive analyses were performed to confirm data suitability. Subsequently, SEM analysis was conducted to identify he potential pathways through which variables may be related, while the supplementary regression analysis provided additional perspective on the relationships observed. Finally, results were interpreted from both theoretical and practical perspectives to derive strategies for enhancing productivity through digital transformation in supervision tasks.

In summary, this study combines field-based empirical data with a mixed analytical approach integrating SEM with regression and correlation tests to establish a practically grounded and evaluation-oriented framework for evaluating productivity tendencies through digital technology adoption in construction supervision.

3.2. Data Collection

Based on the variables defined in the previously established hypotheses, a questionnaire survey was designed to evaluate the relationship between digital technology adoption and productivity. The subjects of this questionnaire consisted of construction project management executives employed at construction sites in South Korea. The questionnaire was administered over a 12-day period, from 7 March 2024, to 18 March 2024. The questionnaire was distributed through a combination of a non-face-to-face method, which involved requesting cooperation from each site and collecting emails, and a face-to-face method, where responses were collected in person after visiting the site and explaining the purpose of the questionnaire. A total of 106 responses were collected, and missing values in the sample were treated using a combination of imputation and casewise deletion methods. Of the collected questionnaires, 82 were used in the final analysis.

Frequency analysis was conducted on responses related to career characteristics in the construction industry and construction supervision field, which correspond to the general information of the sample, using SPSS 29, and the results are presented in

Table 2. General Information of Respondents.

Construction Industry Experience		Construction Supervision Experience
Less than 5 years	10	Less than 3 years	26
5~7 years	2	3~5 years	5
7~10 years	2	5~7 years	9
10~20 years	7	7~10 years	6
More than 20 years	61	More than 10 years	36
Total	82	Total	82

. In terms of the construction industry, 10 people (12.2%) with less than 5 years of experience, 2 people (2.4%) with 5 to 7 years of experience, 2 people (2.4%) with 7 to 10 years of experience, 7 people with more than 10 years and less than 20 years of experience, and 61 people (74.4%) with more than 20 years of experience responded to the questionnaire.

In terms of construction supervision field, 26 people (31.7%) with less than 3 years of experience, 5 people (6.1%) with 3 to 5 years of experience, 9 people (11.0%) with 5 to 7 years of experience, 6 people (7.3%) with 7 to 10 years of experience, 36 people (43.9%) for more than 10 years of experience responded to the questionnaire.

3.3. Composition of the Questionnaire and Measurement Tools

The study constructed the questionnaire by dividing the construction site supervision-related tasks into two perspectives: construction work type and management function. Construction work types were explained through five categories: structural work, designated and foundation work, earthwork, formwork, and reinforced concrete work. Management functions were delineated by seven categories: general management, design management, quality management, schedule management, contract management, safety management, and cost management. This categorization was based on the proportion of inspection items in the standard construction management system in South Korea, which reflects the distribution of task across work types and management functions. For each perspective, the questionnaire was structured into three sections: “Work Adoption Rate,” “Work Responsibility,” and “Productivity.” Concerning the work adoption rate, six questions were formulated regarding the expected reduction in time and workforce when introducing three selected digital technologies. In the case of work responsibility, two questions were raised about the work difficulty and increase in work responsibility when introducing digital technologies. Lastly, for productivity, two questions were designed regarding the anticipated efficiency and degree of efficiency improvement with the introduction of digital technologies. A total of 10 questions comprised 50 items from the perspective of construction work type and 70 items from the perspective of management function, for a total of 120 items. Responses to each questionnaire item were rated on a 5-point Likert scale, ranging from 1 “very low” to 5 “very high” [49]. The composition of the questionnaire is shown in

Table 3. Composition of Questionnaire.

Latent Variable	Construction Supervision Experience
Work Adoption Rate	Object recognition algorithm technology	Expected time reduction upon implementation of technology	5-Point Likert Scale
		Expected workforce reduction in upon implementation of technology
	3D vision technology	Expected time reduction upon implementation of technology
		Expected workforce reduction in upon implementation of technology
	PDF construction documentation system	Expected time reduction upon implementation of technology
		Expected workforce reduction in upon implementation of technology
	Level of difficulty in task execution before technology implementation
	Level of difficulty in task execution after technology implementation
	Productivity before technology adoption
	Productivity after technology adoption

.

3.4. Setting of Variables

In order to analyze the structural relationship between the application of digital technology and the productivity of construction supervision work, the following three variables were defined, and their meanings clarified.

3.4.1. Definition of Productivity in Construction Supervision

Construction productivity has been defined in various ways, but it generally refers to maximizing output while minimizing input. Construction Labor Productivity (CLP) is typically defined as the ratio of labor costs to output quantities. Alternatively, CLP can also be expressed in terms of earned hours, which requires establishing standard outputs, or “norms” for each unit operation and assigning a corresponding number of earned hours to completed units [50]. Nonetheless, the challenge lies in establishing reliable norms and standards, as well as in developing methodologies for measuring productivity that adequately account for multiple influencing factors [46,47,48].

$$Productivity={\displaystyle \frac{\mathrm{O}utput}{Labor+Equipment+Material}}$$

This approach is useful for assessing construction progress efficiency, understanding broader implications, and guiding data collection and analysis. Overseas research on construction productivity has often emphasized collecting and analyzing actual productivity data and investigating barriers to productivity through methods such as work sampling [46,50]. However, questionnaire-based importance analysis alone lacks sufficient empirical support. Therefore, this study seeks to strengthen existing approaches by empirically validating productivity factors, thereby enhancing the productivity of construction supervision through the quantitative identification of key factors and the development of effective digitalization strategies.

3.4.2. Definition of Work Responsibility in Construction Supervision

Examining definitions of work responsibility described it as conditions that consistently undermine workers’ physical or psychological well-being or pose threats within the work environment. Meanwhile, the National Institute for Occupational Safety and Health (NIOSH) defines work responsibility as the physical and emotional strain caused by mismatches between job demands and workers’ capabilities, resources, or needs.

In this study, work responsibility is defined as the difficulties experienced during task performance, specifically conceptualized in terms of fatigue experienced during work. Fatigue encompasses both physical and mental aspects as identified in previous studies. Physical fatigue refers to the physical demands inherent in the workload, while mental fatigue relates to stress factors, including perceived burdens and environmental influences.

3.4.3. Definition of Work Adoption in Construction Supervision

Automation refers to the development and adoption of new technologies that allow capital to substitute for labor across a range of tasks. It also implies that adoption of new technologies reduces the labor share of value added [51]. In this study, technology adoption is defined as the capacity of technologies to replace or complement existing tasks. Therefore, the work adoption rate is defined as the proportion of existing task labor that technology can potentially replace or complement upon its introduction.

4. Research Model and Hypothesis Formulation

4.1. Research Model and Hypothesis Setting

Prior to analyzing and interpreting the survey results, a Structural Equation Model (SEM) approach was employed as an analytical framework to support the assessment and conceptual refinement of the productivity evaluation applied in this study.

To perform this assessment, specific research questions were established to analyze the possible structural tendencies between digital technology adoption and productivity, and a corresponding research model was developed. Rather than using SEM solely for statistical validation, it was utilized to optimize the evaluation structure and examine potential relational patterns among variables. Work adoption rate, work responsibility, and productivity were defined as the core variables. Among these, the work adoption rate was set as an independent variable, based on the understanding that the replacement rate of work when digital technology is adopted, as perceived by practitioners, affects the productivity of supervision work. Accordingly, productivity was set as a dependent variable, and a research model was configured with work responsibility as a parameter to study how technical work substitution rate affects productivity. The hypothetical research model presented in this study is shown in Figure 1 and Figure 2.

• [Research Problem 1] Work adoption rate of technology will affect productivity.
• [Research Problem 2] Work responsibility will affect productivity.
• [Research Problem 3] Work adoption rate of technology will influence work responsibility.

4.1.1. Relationship Between Productivity and Work Adoption Rate

[Research Problem 1] Work adoption rate of technology will affect productivity.

H1. Work adoption rate (expected time reduction and workforce resulting from the implementation of object recognition algorithm technology, 3D vision technology, and PDF construction documentation system) influences productivity, measured in terms of efficiency after technology introduction and the extent of efficiency improvement.

Various studies examined the relationship between “productivity” and “work adoption” in the construction industry. In recent years, construction labor productivity has been significantly influenced by materials and information technology [44]. Efficiency has improved and technological requirements have shifted through the automation and integration of tools, machinery, and information systems [45]. The importance of innovation for construction companies, especially large-scale firms, has grown steadily increasing due to globalization, technological advancement, and trends in public–private cooperation [52].

Significant improvements in labor productivity and partial factor productivity over the long term have been observed in activities experiencing substantial changes in material technology [5]. Machinery has also become more powerful and complex [6,53,54]. Therefore, it can be argued that information technology can innovate management information systems and provide accurate information to enable quick and accurate decisions in the field.

Productivity has also improved through mechanization such as the manufacture of structural steel, building systems, prefabricated components, and related techniques [55,56]. However, the construction industry needs further innovation to remain competitive with other sectors. These research findings suggest that work adoption can have both negative and positive effects on productivity in construction supervision. Therefore, efforts are needed to reduce work performance and improve productivity through efficient work distribution.

4.1.2. Relationship Between Productivity and Work Responsibility

[Research Problem 2] Work responsibility will affect productivity.

H2. 

Work responsibility (level of difficulty in task execution after technology adoption and potential changes in workload following technology adoption) influences productivity, measured by efficiency after technology adoption and the degree of improvement achieved.

Various studies have investigated the relationship between “productivity” and “work responsibility” in the construction industry. Workloads and job stress, or work responsibility, can have negative impacts on workers, which in turn adversely influence productivity [57,58]. Excessive workloads can lead to lowered morale and motivation, becoming a source of work fatigue for workers [49]. Additional studies have posited that work responsibility may also reduce work productivity, as factors such as physical, behavioral, personality, emotional, cognitive, and interpersonal aspects contribute to stress symptoms that can affect job performance [58,59].

Through these studies, a better understanding of the relationship between productivity and work responsibility in construction supervision can be gained, supporting the development of appropriate responses and improvement strategies.

4.1.3. Relationship Between Work Responsibility and Work Adoption Rate

[Research Problem 3] Work adoption rate of technology will influence work responsibility.

H3. 

Work adoption rate (expected time reduction and workforce resulting from the implementation of object recognition algorithm technology, 3D vision technology, and PDF construction documentation system) affects work responsibility, measured by task difficulty and potential changes in workload following technology adoption.

Research on the relationship between “work adoption” and “work responsibility” in construction supervision tasks has been examined from multiple perspectives. Studies have discovered that work adoption can alleviate work responsibilities and consequently improve work performance. Subjective well-being based on work adoption is positively related to satisfaction and well-being in various contexts. Adopting more tasks from one’s social support system has also been linked to higher levels of happiness [60]. Adoption of different levels of work has also been shown to enhance accountability [61]. Considering these relationships, the digital technology adoption in construction supervision tasks requires adjustments in task allocation and workload distribution.

5. Results

5.1. Data Analysis

The reason for adopting a Structural Equation Model (SEM)-based analytical approach in this study was to support the evaluation with a systematic and interpretive assessment framework rather than using SEM as a strict statistical validation tool.

Accordingly, data obtained from the questionnaire survey were analyzed following a structured procedure. The collected data were examined through a two-step analytical process [62]. In the first stage, the basic consistency and internal coherence of the measured variables were assessed through confirmatory factor analysis (CFA). In the second stage, the possible relational patterns and directional tendencies among variables within the evaluation framework were analyzed using SEM to contextualize and interpret the productivity-related relationships identified in the field data.

5.1.1. Variable Validation

The purpose of confirmatory factor analysis is to verify the relationship between items (observed variables) and their ability to accurately represent the underlying latent variables [63]. In this study, basic analysis was performed on questionnaire responses before conducting confirmatory factor analysis. As part of the basic analysis, reliability analysis (Cronbach’s alpha test) and descriptive statistical analysis were performed, to evaluate the reliability of the dataset.

Cronbach’s alpha value was employed to assess the internal consistency of questionnaire items. A Cronbach’s alpha value of 0.7 or higher indicates that respondents’ answers to the items were consistently provided. All responses to the observed variable items of the latent variables established in this study yielded Cronbach’s alpha values exceeding 0.7, demonstrating strong reliability. The results of Cronbach’s alpha values are presented in

Table 4. Cronbach’s Alpha Analysis Value.

Classification	Measurement Variable	Num. of Evaluation Items	Cronbach’s Alpha
By Construction Work Type	Work Adoption Rate	30	0.949
	Work Responsibility	10	0.956
	Productivity	10	0.901
By Management Function	Work Adoption Rate	42	0.969
	Work Responsibility	14	0.969
	Productivity	14	0.907

.

Descriptive statistical analysis was conducted to check the mean, standard deviation, and normality of key variables. The main observed variables were analyzed, and if skewness < 3.0 and kurtosis < 8.0 were satisfied, the observed variables were considered to follow a normal distribution. Since the responses to the main variables in this study satisfy all relevant conditions, the reliability of the questionnaire data was confirmed. The results of descriptive statistical analysis are shown in

Table 5. Results of Technical Statistical Analysis by Construction Work Type.

Observed Variables	Mean	Std. Dev.	Skewness		Kurtosis
	Statistic	Statistic	Statistic	Error	Statistic	Error
Structural work: Expected time reduction upon adoption of technology 1	3.39	0.846	−0.454	0.239	0.138	0.474
Structural work: Expected workforce reduction in upon adoption of technology 1	3.01	0.928	0.056	0.239	−0.332	0.474
Structural work: Expected time reduction upon adoption of technology 2	3.41	1.047	−0.580	0.239	−0.144	0.474
Structural work: Expected workforce reduction in upon adoption of technology 2	3.20	0.975	−0.275	0.239	−0.287	0.474
Structural work: Expected time reduction upon adoption of technology 3	3.28	1.047	−0.280	0.239	−0.323	0.474
Structural work: Expected workforce reduction in upon adoption of technology 3	2.94	0.953	0.189	0.239	0.098	0.474
Structural work: level of difficulty in task execution after technology adoption	2.89	0.922	0.063	0.239	−0.484	0.474
Structural work: increase in workload after technology adoption	2.81	1.022	0.100	0.239	−0.443	0.474
Structural work: productivity after technology adoption	3.63	0.832	−0.258	0.239	−0.410	0.474
︙	︙	︙	︙	︙	︙	︙
Reinforced concrete work: level of difficulty in task execution after technology adoption	3.00	0.975	0.131	0.239	−0.579	0.474
Reinforced concrete work: increase in workload after technology adoption	2.57	1.148	0.269	0.239	−0.925	0.474
Reinforced concrete work: productivity after technology adoption	4.07	0.836	−1.065	0.239	1.605	0.474
Reinforced concrete work: increase in productivity after technology adoption	4.00	0.901	−0.828	0.239	0.511	0.474

and

Table 6. Results of Technical Statistical Analysis by Management Function.

Observed Variables	Mean	Std. Dev.	Skewness		Kurtosis
	Statistic	Statistic	Statistic	Error	Statistic	Error
General Management: Expected time reduction upon adoption of technology 1	3.35	0.935	−0.491	0.266	0.432	0.526
General Management: Expected workforce reduction in upon adoption of technology 1	3.12	0.986	−0.171	0.266	0.192	0.526
General Management: Expected time reduction upon adoption of technology 2	3.54	0.996	−0.641	0.266	0.365	0.526
General Management: Expected workforce reduction in upon adoption of technology 2	3.29	1.024	−0.408	0.266	0.127	0.526
General Management: Expected time reduction upon adoption of technology 3	3.52	0.820	−0.424	0.266	0.291	0.526
General Management: Expected workforce reduction upon adoption of technology 3	3.32	0.941	−0.135	0.266	0.124	0.526
General Management: level of difficulty in task execution after technology adoption	3.11	0.832	0.186	0.266	−0.077	0.526
General Management: increase in workload after technology adoption	3.04	1.059	0.053	0.266	−0.395	0.526
General Management: productivity after technology adoption	3.67	0.754	−0.251	0.266	−0.127	0.526
︙	︙	︙	︙	︙	︙	︙
Cost Management: level of difficulty in task execution after technology adoption	3.12	1.011	−0.030	0.266	−0.444	0.526
Cost Management: increase in workload after technology adoption	2.93	1.152	−0.053	0.266	−0.712	0.526
Cost Management: productivity after technology adoption	3.84	0.853	−0.298	0.266	−0.534	0.526
Cost Management: increase in productivity after technology adoption	3.89	0.832	−0.054	0.266	−1.005	0.526

.

5.1.2. Supplementary Correlation and Regression Analysis

Given the relatively small sample size (n = 82), supplementary analyses were performed to reinforce the reliability of the SEM-based productivity evaluation. To verify the consistency of the results, correlation and simple regression analyses were conducted using the aggregated mean data from both perspectives (construction work types and management functions). The correlation analysis indicated strong positive associations between the degree of digital task substitution (time and labor reduction) and productivity after digital technology adoption. Specifically, for construction work types, the correlations were r = 0.96 (time reduction) and r = 0.94 (labor reduction). For management functions, similarly high correlations were observed (r = 0.88 and r = 0.85, respectively). Regression analyses further confirmed these relationships, yielding coefficients of β = 0.91 (R² = 0.93) and β = 0.87 (R² = 0.89) for construction work types, and β = 0.81 (R² = 0.77) and β = 0.76 (R² = 0.72) for management functions.

These consistent and statistically significant results reinforce the robustness of the SEM findings, confirming that higher levels of digital technology adoption led to greater productivity improvements across both analytical perspectives.

5.2. Analysis of Structural Model

5.2.1. Confirmatory Factor Analysis Results of the Initial Structural Model

To verify whether the observed variables (items) for the main variables that have been validated for reliability were appropriately classified into latent variables, an initial model was established (Figure 3) and confirmatory factor analysis (CFA) was conducted. In the case of factor analysis, exploratory factor analysis and confirmatory factor analysis are generally performed to verify the validity of variables. However, when validity is tested using variables identified from previous research, CFA is considered more appropriate than EFA [63,64]. Therefore, in this study, only CFA was conducted to verify the variables. First, a model analysis was conducted to verify each evaluation index for 3 latent variables and 50 (by construction work type) and 70 (by management function) observed variables. This is to verify the appropriateness of the hypothetical initial research model proposed to explain the structural relationship that the application of digital technology has on the productivity of construction supervision work. Among the model suitability indices, x² (CMIN), TLI, CFI, and RMSEA, were examined and the results are presented in

Table 7. Summary of Criteria and Analysis Results.

Factor	Construction Work Type	Management Function
$x^2$ (CMIN)	0.000	0.000
TLI (Turker-Lewis index)	0.324	0.324
CFI (Comparative Fit Index)	0.353	0.345
RMSEA (Root Mean Square of Approximation)	0.202	0.183

.

Upon reviewing the values in Table 7, it was observed that the factor values of the research model did not reach a satisfactory level. This indicates that the model may not be fully suitable for representing the population data.

5.2.2. Model Revision and Configuration of Final Structural Model

In the case of the initial structural models, the indices of model appropriateness did not reach an acceptable level and were therefore considered limited in suitability for this study. However, the primary objective of this research was not to achieve strict statistical validation but rather to conduct a field-based survey that reflects practical construction site conditions. This approach provides meaningful research value by incorporating real-world industry perspectives. Therefore, even if the initial model did not achieve strong statistical performance, it was still possible to derive valuable findings by refining the model. Accordingly, this study undertook a systematic process to derive a more practically interpretable model, analyzed the results, and interpreted the evaluation outcomes within the context of this study.

To support this objective, this study used a variable elimination method to enhance the measurement model’s coherence and improve the model. This approach, provided by AMOS, involves removing less influential variables that hinder explanatory power. Afterwards, structural equation modeling was selected by reconstructing the structural model excluding the deleted variables. The hypothesis set was verified through path analysis for the selected model, and the results were interpreted. This study iteratively removed variables identified as weak contributors to the interpretive consistency of the study results according to the criteria in

Table 8. Methods for Improving Model appropriateness (Removing Variables).

Category	Exclusion Criteria	Relationship with Model Appropriateness
p-value of Regression Weights	Over 0.05	If the model appropriateness is good but the p-value is above 0.05, the variable is deleted.
Standardized Regression Weights	Under 0.5	If the model appropriateness is good, variables with regression weights below 0.5 are not deleted.
Variances	Negative (“-“)	If the model appropriateness is good but the regression weight is negative, the variable is deleted.
Squared Multiple Correlations	Under 0.4	If the model appropriateness is good, variables with regression weights below 0.4 are not deleted.

, to derive a refined model with greater interpretive alignment.

First, variables with p-values of regression weights of 0.05 or higher were deleted. Subsequently, variables with Standardized Regression Weights estimate values below 0.5 were removed in descending order of data. Variables with negative variances were then removed, followed by removing variables with Squared Multiple Correlations of less than 0.4. These steps resulted in a revised model with improved structural clarity. The variables corresponding to these criteria were removed to enhance the interpretability and stability of the final model. As a result, a total of 26 variables by construction work type and 25 variables by management function were removed. The model appropriateness improvement values of the initial research model and the revised research model are shown in

Table 9. Comparison of the Initial Model and the Revised Model.

Factor	Construction Work Type		Management Function
	Initial	Revised	Initial	Revised
$x^2$ (CMIN)	<0.0001	<0.0001	<0.0001	<0.0001
TLI (Turker-Lewis index)	0.324	0.584	0.324	0.536
CFI (Comparative Fit Index)	0.353	0.625	0.345	0.560
RMSEA (Root Mean Square of Approximation)	0.202	0.216	0.183	0.172

.

The ultimate significance of this study lies in conducting a practitioner-oriented evaluation of productivity improvement through digital technology adoption. As this study incorporates practical construction site conditions, it represents a novel approach within the domestic research landscape. The proposed final model provides a practically grounded framework for digital technology adoption in supervision tasks, as it is derived from real-world industry insights rather than theoretical constructs. In particular, this study combined a field-based evaluation with an SEM-supported analytical review to enhance the reliability of the assessment. Furthermore, by constructing the structural model and refining it through a logical process, the study provides a foundation for uture research to strengthen statistical rigor as larger datasets become available. Therefore, this study indicates the potential for digital transformation in supervision tasks to yield practical benefits. Consequently, the proposed model is positioned to offer results that are practically informative compared to earlier studies relying solely on conceptual or document-based analysis.

5.3. Hypothesis Testing Through Final Structural Model

This study aims to elucidate the structural relationship between digital technology adoption and productivity in construction supervision and, by verifying the support or rejection of the established hypotheses, proposes meaningful interpretations of productivity evaluation.

A structural equation model was established to analyze the impact of “Work Adoption Rate” and “Work Responsibility” on “Productivity,” and the results of hypothesis testing are presented in

Table 10. Results of Hypothesis Testing through Structural Equation Model.

Factor	Value		C.R (t)		p-Value		Result
	Work Type	Management Function	Work Type	Management Function	Work Type	Management Function	Work Type	Management Function
The work adoption rate of technology is expected to affect productivity.	0.294	0.235	5.037	1.972	***	0.049 (*)	Accept	Accept
Work responsibility is expected to influence productivity.	0.052	−0.152	1.249	−1.829	0.212	0.067	Reject	Reject
The adoption rate of technology is expected to influence work responsibility.	−0.27	0.139	−0.303	0.871	0.762	0.384	Reject	Reject

(* p < 0.05, *** p < 0.001).

. The significance of each hypothesis was determined based on the p-values (* p < 0.05, ** p < 0.01, *** p < 0.001), which determined their acceptance or rejection decision. Among the three hypotheses tested from both the perspectives of construction type and management function, two hypotheses were rejected while one was accepted.

Specifically, the hypothesis related to Research Problem 1, “Work Adoption Rate will affect Productivity,” was supported, while those related to Research Problem 2, “Work Responsibility will affect Productivity,” and Research Problem 3, “Work Adoption Rate will affect Work Responsibility,” were not supported. The structural paths excluded during model refinement were also taken into account in deriving the final model.

Resultantly, the structural patterns observed in the final model can be interpreted as follows. Only the pathway from work adoption rate to productivity emerged as a meaningful relationship, while the other two pathways, (1) work responsibility → productivity and (2) work adoption rate → work responsibility, did not appear to have noticeable effects. This outcome reflects the characteristics of the practitioner-based data used in this study. Many respondents were senior field engineers, and the concept of work responsibility may have been interpreted more as a short-term psychological or cognitive burden associated with unfamiliar digital tools than as an objective workload factor. As a result, perceived responsibility did not consistently align with productivity or technology-adoption tendencies in the model. These findings indicate that work responsibility may behave differently depending on task context and personal perception, suggesting the need for clearer construct definitions and broader datasets in future studies.

6. Analysis of Evaluation Findings

6.1. Theoretical Analysis of Hypothesis Testing Results

6.1.1. Construction Work Type Perspective

Firstly, the hypothesis (H1) suggesting that the work adoption rate influences productivity was confirmed, with a path coefficient of 0.294 (p < 0.001). This result reveals that an increase in the rate at which digital technologies replace traditional tasks has a positive effect on the productivity of supervision tasks. Thus, successful implementation of digital technologies in supervision tasks can enhance productivity by substituting conventional practices.

Secondly, the hypothesis (H2) suggesting that work responsibility affects productivity was not confirmed, with a path coefficient of 0.052 (p = 0.212), indicating no significant relationship between work responsibility and productivity. Lastly, the hypothesis (H3) suggesting that the work adoption rate influences work responsibility was also not confirmed, with a path coefficient of −0.027 (p = 0.762), indicating no significant relationship between the two variables.

6.1.2. Management Function Perspective

Firstly, hypothesis 1 (H1) suggesting that the work adoption rate influences productivity was confirmed, with a path coefficient of 0.235 (p < 0.05). This indicates a positive impact of respondents’ work adoption rate on productivity, implying that as the rate at which digital technologies replace tasks increases, it positively affects the productivity of supervision tasks. Therefore, when digital technology introduced in supervision work successfully replaces existing (traditional) work, work productivity can be expected to improve.

Secondly, hypothesis 2 (H2) proposing that work responsibility affects productivity was not confirmed, with a path coefficient of −0.152 (p = 0.067), indicating no significant relationship between work responsibility and productivity. Lastly, hypothesis 3 (H3) suggesting that the work adoption rate influences work responsibility was also not confirmed, with a path coefficient of 0.139 (p = 0.384), indicating no significant relationship between the two variables.

6.1.3. Variable Appropriateness and Interpretation Analysis

The variable of work responsibility initially assumed to directly or indirectly influence productivity did not show statistically significant relationships with either work adoption rate or productivity in the structural equation model. However, this finding does not necessarily indicate that work responsibility has no effect. Rather, the rejection of hypotheses 2 and 3 may be explained by respondents conflating short-term and long-term perspectives of work responsibility when answering the questionnaire.

Based on a detailed review of the survey responses and demographic characteristics, several interpretive factors can be identified. Many respondents were senior practitioners with extensive field experience and thus may have associated work responsibility with a short-term psychological burden or adaptation pressure caused by unfamiliar digital tools. This perception could have led to higher self-reported fatigue or stress levels immediately following digital technology adoption, even if long-term efficiency improvements were anticipated. Consequently, the concept of work responsibility might have been perceived not as an objective workload factor but as a subjective psychological state, reducing its statistical association with productivity.

This interpretation aligns with findings from prior studies in human technology interaction, which suggest that perceived responsibility and cognitive load can function as moderating variables that influence how individuals respond to technological change [65,66]. In this context, it is plausible that when supervision tasks involve greater decision pressure or accountability, the initial benefits of digital adoption may be temporarily mitigated by cognitive and emotional strain. Conversely, in tasks with lower perceived responsibility, digital technologies may enhance productivity more immediately.

Therefore, while work responsibility did not show direct or mediating effects in this study, its role as a potential context-dependent moderator warrants further investigation. Future research should clearly define the concept of responsibility distinguishing between physical workload, psychological burden, and accountability and test its moderating effects using larger datasets and hierarchical or multi-group SEM approaches. Establishing such conceptual clarity will help reduce interpretation bias and improve the accuracy of productivity related analyses in digital construction supervision.

6.1.4. Theoretical Interpretation and Comparative Discussion

While the statistical validation confirmed the proposed relationships, further interpretation is required to contextualize these findings within established theoretical and empirical frameworks. To ensure that the quantitative outcomes are not interpreted in isolation, the following discussion links the results to relevant theories and previous studies to explain the underlying mechanisms. These findings can be further interpreted through theoretical perspectives that describe how technology adoption influences organizational performance. The Diffusion of Innovation (DOI) theory posits that innovations perceived as useful and compatible with existing workflows are more readily adopted, leading to measurable efficiency gains. Similarly, the Technology–Organization–Environment (TOE) framework highlights that technological readiness and contextual factors jointly shape innovation performance. In the construction domain, previous research has reported that adopting digital tools such as BIM-based coordination, automated documentation, and computer-vision inspection improves information accuracy, reduces manual workload, and enhances overall productivity [67,68].

The empirical evidence of this study reinforces these theoretical assumptions. The positive influence of the work-adoption rate on productivity indicates that digital task substitution enhances supervision efficiency by minimizing repetitive processes and facilitating information accessibility. However, the extent of this improvement varies depending on construction work type and management function, suggesting that the productivity impact of digital adoption is context-dependent. This observation extends previous studies—many of which were conceptual or simulation-based—by providing quantitative field evidence derived from practitioner evaluations. Consequently, the present study not only substantiates existing theoretical frameworks but also advances them toward a practical, evidence-driven understanding of how digital technologies translate into real productivity gains within construction supervision.

6.2. Practical Analysis of Hypothesis Testing Results

A practical interpretation was derived based on the theoretical analysis of the previously analyzed hypothesis testing results. Following the structural equation model, this study verified the hypothesis that ‘work adoption rate’ influences productivity, as hypothesis 1 was supported. This implies that the higher the work adoption rate, the higher productivity improves. By identifying variables that have a significant influence on the work adoption rate and productivity, it is possible to enhance the productivity of construction supervision. Therefore, this study provides indicators to guide the development of effective strategies for the digitalization of supervision by analyzing variables that critically affect work adoption rate and productivity. Ultimately, through practical interpretation, this study seeks to support decision-making for digitalization strategies by identifying digital technologies, types of work, and management functions most urgently required in construction supervision.

6.2.1. Construction Work Type Perspective

Variables that have a key impact on work adoption rate and productivity were analyzed for each type of work, and the results are shown in

Table 11. Analysis of Key Variable by Construction Work Type.

Latent Variable	Measurement Variable	Construction Work	Factor Loading
Work Adoption Rate	Expected time reduction upon adoption of technology	Formwork	0.997
		Reinforced concrete work	0.996
		Designated and foundation work	0.931
		Earthwork	0.872
	Expected workforce reduction in upon adoption of technology	Reinforced concrete work	1.266
		Formwork	1.214
		Designated and foundation work	1.188
		Earthwork	1.116
		Structural work	1.000
Productivity	Productivity after technology adoption	Designated and foundation work	1.431
		Earthwork	1.430
		Reinforced concrete work	1.300
		Formwork	1.103

. In the case of work adoption rate, variables (items) related to object recognition algorithm technology and 3D vision technology (technologies 1 and 2, respectively) were excluded because they were not statistically significant. As a result, only the pdf construction document system technology (technology 3) was statistically significant.

For technology 3, work adoption rate can be interpreted primarily in terms of amount of time required and the amount of workforce invested. It has been observed that when technology replaces work, the reduction in amount of workforce invested plays a more significant role in reducing amount of time required. Comparing the effects by construction type, it can be predicted that in terms of amount of workforce invested, the efficiency is in the following order, from high to low: reinforced concrete work > formwork > designation and foundation work > earthwork. In terms of time reduction, the effect is in the following order, from high to low: formwork > reinforced concrete work > designation and foundation work > earthwork. This implies that introducing digital technologies into reinforced concrete work and formwork can replace traditional tasks more effectively and improve productivity.

6.2.2. Management Function Perspective

Variables that have a key impact on work adoption rate and productivity were analyzed for each management function, and the results are shown in

Table 12. Analysis of Key Variables by Management Function.

Factor	Technology	Measurement Variable	Management Function	Factor Loading
Work adoption rate	Object recognition algorithm	Expected time reduction upon adoption of technology	Contract Mgmt.	1.025
			Design Mgmt.	1.019
			General Mgmt.	1.000
			Schedule Mgmt.	0.956
			Safety Mgmt.	0.842
	3D vision technology		Contract Mgmt.	1.208
			General Mgmt.	1.119
			Cost Mgmt.	1.086
			Design Mgmt.	1.002
			Schedule Mgmt.	0.948
	Object recognition algorithm	Expected workforce reduction in upon adoption of technology	General Mgmt.	1.408
			Contract Mgmt.	1.334
			Cost Mgmt.	1.332
			Design Mgmt.	1.315
			Schedule Mgmt.	1.258
			Safety Mgmt.	1.088
			Quality Mgmt.	1.055
	3D vision technology		Contract Mgmt.	1.408
			General Mgmt.	1.389
			Cost Mgmt.	1.336
			Schedule Mgmt.	1.327
			Design Mgmt.	1.274
			Safety Mgmt.	1.132
			Quality Mgmt.	1.090
Productivity	Productivity after technology adoption		Safety Mgmt.	1.127
			Cost Mgmt.	1.107
			Schedule Mgmt.	1.076
			Quality Mgmt.	1.011

. In the case of work adoption rate, the variables (items) related to the pdf construction document system technology (technology 3) were excluded because they were not statistically significant. Consequently, it can be inferred that the object recognition algorithm technology and 3D vision technology (technologies 1 and 2, respectively) were identified as statistically significant.

Similarly to the construction work type, the work adoption rate was interpreted in two perspectives, amount of time required, and the amount of workforce invested. It can be inferred that, on average, the reduction in workforce plays a more critical role in tasks where technology substitutes for them than the reduction in time required. When comparing the effects by management function, it can be implied that technology 1 will be effective in the order of general management > contract management > cost management > design management > schedule management > safety management > quality management in terms of reducing workforce. For Technology 2, it can be predicted that it is effective in the order of contract management > general management > cost management > schedule management > design management > safety management > quality management in terms of reducing workforce. In terms of time reduction, it can be implied that technology 2 will be effective in the order of contract management > design management > general management > schedule management > safety management. For technology 2, it can be predicted that it is effective in the order of contract management > general management > cost management > design management > schedule management. Overall, introducing digital technology in contract management and general management aspects would be more effective in substituting existing tasks and significantly improving productivity.

Productivity was also analyzed for each type of construction, and how the expected productivity of management functions when introducing digital technology would affect the entire construction project. As a result, it can be determined that the productivity of the management function in the order of safety management > cost management > schedule management > quality management would exert a critical influence on the productivity of the entire construction project. The results in terms of management function also showed different results compared to the work adoption rate and productivity aspects. However, the reflection of the potential impact of these management functions on the success of the overall construction project is believed to be consistent across all aspects.

7. Conclusions

7.1. Empirical Findings and Key Results

Over the past 10 years, the productivity of the construction industry has remained relatively low compared to other sectors, largely due to its heavy reliance on labor. Consequently, productivity management is becoming increasingly important within the field. Therefore, this study focused on evaluating productivity improvements through digitalization, particularly in light of the industry’s strong dependence on manpower. To address these issues, this study conducted an exploratory, field-based productivity evaluation focusing on the adoption of digital technologies in construction supervision.

In this study, three selected digital technologies (object recognition algorithm technology, 3D vision technology, and PDF construction documentation system) were examined to analyze their effects on supervision productivity. Based on these, 3 hypotheses were formulated to test the relationships among work adoption rate, work responsibility and productivity, three key variables related to construction supervision productivity.

A total of 106 questionnaires were collected from field experts, of which 82 were valid and used for analysis. Structural Equation Model (SEM) was applied in an exploratory manner to review internal consistency and directional tendencies among variables rather than to generalize the model statistically.

Subsequently, three hypotheses were established to examine the relationships among the three variables: “Work adoption rate of technology will affect productivity (H1),” “Work responsibility will affect productivity (H2),” and “Work adoption rate of technology will influence work responsibility (H3).”

As a result, the analysis indicated that H1 was the only pathway showing a meaningful directional tendency, while H2 and H3 were rejected. H1 showed different levels of significance when divided from the perspectives of construction work types and management functions. However, with path coefficient values of 0.294 and 0.235, respectively, it was found that respondents’ “work adoption rate” tended to positively influences “productivity” to a similar extent. Among the evaluated technologies, the PDF-based construction documentation system exhibited the strongest observed productivity gains, particularly in reinforced concrete and formwork tasks. Object recognition and 3D vision technologies also showed positive tendencies in general management and contract management functions.

These empirical findings provide practitioner-based insights into the positive association between digital technology adoption and supervision productivity. They further suggest that the magnitude of improvement differs by technology type and management function, offering insight into how specific technologies can be effectively integrated into supervision workflows.

7.2. Discussion and Implications

The results of this study should be interpreted as exploratory rather than conclusive due to methodological and data-related constraints. Although the SEM analysis reviewed the internal consistency of relationships among variables, the model shows limited analytical appropriateness because of the relatively small sample size (n = 82) and perception-based survey responses. Therefore, the identified relationships should be understood as indicative tendencies indicative trends reflecting practitioners’ field experience rather than universal causal effects.

Within these limitations, the study contributes to the field by proposing a practical and context-specific evaluation framework that reflects real-world supervision conditions and industry needs. This framework enables practitioners and organizations to assess which technologies and supervision functions can yield measurable productivity improvements. In particular, the observed effectiveness of the PDF-based documentation system in reinforced concrete and formwork supervision suggests a potential focus area for future digital transformation initiatives. However, such implications should be regarded as suggestive guidance derived from exploratory data, not as direct policy recommendations. Moreover, the study’s methodological approach integrating field-based practitioner insights with quantitative analysis provides a novel empirical foundation distinct from prior academic or simulation-driven studies. By linking quantitative results with contextual observations, this research offers evidence-based perspectives that can inform subsequent model refinements and support practical decision-making for supervision digitalization.

Several limitations remain to be addressed. First, the modest sample size restricts broad generalization and the robustness of the SEM outcomes. Second, the data relied on respondents’ perceptions rather than hands on experience with digital tools, which may introduce bias. Third, potential ambiguity in the interpretation of “work responsibility” could have influenced responses, suggesting the need for a more precise construct definition.

Future research should expand the dataset to enhance statistical validity, employ pilot demonstrations or experimental studies to observe actual technology implementation, and investigate whether work responsibility operates as a moderating variable in the relationship between digital adoption and productivity. Such studies will strengthen the empirical foundation for evidence-based digital transformation strategies in construction supervision and help establish a more rigorous and scalable productivity evaluation model for the industry.