A Progressive Policy Evaluation Framework for Construction Digitalization in China: Evidence from Wuhan

Abstract

Global digitalization drives policy-led transformation in the construction industry, yet its effectiveness hinges on localized implementation. However, research on China’s regional digital policies remains insufficient, particularly in systematic evaluation mechanisms. Focusing on Wuhan, this study proposes a progressive “3M” (macro–meso–micro) policy evaluation framework to analyze local policy efficacy under national strategies. Macro-level PESTEL analysis identifies weak legal frameworks as a critical gap. Meso-level PMC index modeling establishes a hierarchical optimization pathway prioritizing incentive measures, followed by policy timeliness, assessment mechanisms, policy focus, and policy nature. Micro-level Spearman’s correlation analysis further pinpoints five implementation drivers: pilot projects, long-term planning, detailed measures, talent cultivation, and regulatory reinforcement. The results indicate that Wuhan’s policies require targeted improvements: (1) synergizing pilot innovation with legal safeguards, (2) integrating green principles into long-term planning, (3) refining technical standards and policy alignment, (4) enhancing multidisciplinary talent development through industry–academia collaboration, and (5) establishing IoT-enabled dynamic monitoring platforms. This hierarchical evaluation system provides empirical evidence for optimizing China’s construction policies while offering a transferable governance framework for global cities navigating digital transitions. Future research should extend the temporal and spatial coverage while incorporating adaptive evaluation tools to address policy dynamism.

1. Introduction

With the rapid advancement of global digital technologies, production methods and management models across various industries are undergoing profound transformations, driving a fundamental shift in socio-economic paradigms [1]. As a vital pillar of national economies and a core component of modern industrial systems, the construction industry holds strategic significance in promoting sustainable development and enhancing international competitiveness through digital transformation [2]. Many countries worldwide are actively advancing the digital transformation of the construction industry (DTCI) through coordinated efforts among governments, enterprises, and industry associations. These efforts include formulating supportive policies and developing digital infrastructure to achieve industrial upgrading and bolster global competitiveness [3,4,5]. In alignment with global trends toward digitalization and intelligent construction, the Chinese government has adopted a progressive policy approach characterized by technological evolution. Since 2016, national policy has shifted from focusing on individual technological breakthroughs—such as Building Information Modeling (BIM)—to more integrated systems, culminating in the proposal of intelligent construction in 2020 and the promotion of full-domain data circulation in 2023. This series of policies not only reflects the progressive nature of technological development but also highlights a multi-tiered policy implementation framework. Specifically, it involves a top–down structure from national strategies (State Council) and sectoral plans (Ministry of Housing and Urban–Rural Development) to localized pilot projects at the municipal level (for details, see

Table 1. Key policies of China’s DTCI.

Policy Document	Department and Year	Content	Affect
2016–2020 Construction Industry Informatization Development Outline	Ministry of Housing and Urban–Rural Development, 2016	Promote the integrated application of technologies such as BIM, big data, cloud computing, and the Internet of Things to enhance the informatization level of the construction industry; drive digitalization, networking, and intelligent development; and establish internationally advanced IT application enterprises.	Establishes the strategic position of BIM technology and lays the foundation for construction digitalization.
Guiding Opinions on Promoting Coordinated Development of Intelligent Construction and Industrialization	Ministry of Housing and Urban–Rural Development and 13 other ministries, 2020	Drive digital and intelligent upgrades as core motivators to overcome key technologies and build a complete industrial system by 2025 and reach international advanced levels by 2035.	Constructs a collaborative development framework for intelligent construction and industrialization in the sector.
14th Five-Year Plan for the Development of the Digital Economy (2021–2025)	State Council, 2021	Designate the construction industry as a key sector for digital transformation; propose smart upgrading goals for the entire industry chain; and focus on BIM, intelligent construction equipment, and construction industry internet platforms.	Incorporated into the national digital economy strategy, strengthening policy support and resource allocation.
14th Five-Year Plan for Construction Industry Development	Ministry of Housing and Urban–Rural Development, 2022	Improve intelligent construction policies and industrial systems; promote digital collaborative design and prefabricated buildings; advance construction industry internet platform development; accelerate research and application of construction robots; and advocate for green construction methods.	Quantifies transformation objectives and establishes an assessable policy implementation system.
Overall Layout Plan for the Construction of Digital China	Central Committee of the CPC and State Council, 2023	Proposes the “2522” framework to consolidate digital infrastructure and data resource systems; advance deep integration of digital technology with economic, political, cultural, social, and ecological civilization construction; strengthen innovation systems and digital security capabilities; and optimize domestic and international digital development environment.	Advances national digital governance and builds a new development model driven by data elements.

; policy sources: the Central Government of China website, https://www.gov.cn/, accessed on 10 February 2025, and the Ministry of Housing and Urban–Rural Development website, www.mohurd.gov.cn, accessed on 10 February 2025).

Two strategic national-level policies have marked significant milestones in China’s digital transformation of the construction sector. First, the 14th Five-Year Plan for the Development of the Digital Economy (2021–2025), issued in 2021, identified the construction industry as a key domain for intelligent upgrading across the entire value chain. It emphasizes the integrated application of Building Information Modeling (BIM), industrial internet platforms, and intelligent construction equipment while promoting the in-depth integration of technologies such as 5G and artificial intelligence into the full lifecycle management of construction projects. The ultimate goal is to establish a comprehensive digital system that spans design, construction, and operation, thereby significantly improving industrial productivity. Second, the Overall Layout Plan for the Construction of Digital China, released in 2023, introduced the “2522” strategic framework, positioning data elements as a core driving force. The plan calls for the development of integrated databases that consolidate multiple indicators—including construction quality, energy consumption, and safety—and sets specific targets for smart construction pilot cities. As one of the leading cities in this initiative, Wuhan has taken the lead in promoting the numerical control of key construction processes and the widespread adoption of digital management systems. The release of these two policies not only reflects the Chinese government’s strong commitment to driving digital transformation within the construction industry but also provides actionable guidance and targets for local governments and enterprises. This policy direction is accelerating the sector’s shift toward a more intelligent and digitalized future [6]. However, despite preliminary progress achieved in some cities, the industry continues to face bottlenecks, such as the limited penetration of digital technologies and fragmented application scenarios [7].

The development of digital transformation in the construction industry (DTCI) is characterized by three major trends: technological integration, organizational transformation, and environmental sustainability. In terms of technological integration, the academic community has established a relatively systematic framework of digital technologies. Ullah, through bibliometric analysis, proposed the “Big9” technology system [8], which includes nine core technologies: unmanned aerial vehicles (UAVs), Internet of Things (IoT), 3D scanning, wearable devices, cloud computing, Software as a Service (SaaS), big data, virtual and augmented reality (VR/AR), artificial intelligence (AI), and robotics. Building on this foundation, Naji [9] further expanded the framework by adding digital twins, blockchain, geographic information systems (GISs), and cybersecurity, thereby forming a comprehensive architecture encompassing 12 core technologies.

Empirical studies have confirmed the multidimensional value of such technological integration. For instance, Mahajan [10] demonstrated the transformative effect of integrating UAVs and BIM for air–ground data fusion in construction supervision. Honcharenko [11] validated the energy efficiency improvements achieved through BIM-IoT integration. Kim [12] developed a BIM-based 3D laser scanning system that significantly enhanced the accuracy of schedule control through point cloud registration techniques. Awolusi’s research [13] confirmed the effectiveness of wearable devices in real-time safety monitoring at construction sites. Meanwhile, Oke [14] emphasized that cloud computing applications notably improve the sustainability and overall management efficiency of construction projects. In the domain of technological innovation, Mironova [15] achieved intelligent full-process decision-making by integrating big data, AI, and VR technologies. Afzal [16] constructed a digital twin and BIM framework that leverages real-time data to drive project performance. Li [17] proposed a hierarchical blockchain architecture (IaaS–BaaS–SaaS) to ensure the trustworthiness of data interactions. Moisa [18] utilized GIS-based spatial analysis to optimize project site selection. Alshammari [19] highlighted the critical role of cybersecurity in the integration of smart city systems.

In terms of organizational transformation, Samuelson [20] emphasized that adaptive adjustments to organizational structures and workflows form the foundation of successful digital transformation. Building on this, Osorio [21] argued that optimizing the organizational climate and information flow mechanisms is essential for unlocking the full potential of technological implementation. Zulu [22] identified key barriers at the leadership level, including resource constraints and insufficient risk awareness. Prebanić and Vukomanović [23] highlighted the importance of institutional restructuring in enhancing the synergistic application of digital technologies. Empirical research by Zhang [24] further revealed that top-level support and policy guidance are critical driving forces behind corporate digital transformation.

Regarding environmental sustainability, digital technologies serve as enablers of resource optimization and energy efficiency, thus supporting green development goals. Shen and Wang [25] pointed out that policy incentives and green innovation represent essential pathways for ecological transformation. Song [26] found that digital transformation enhances environmental performance through mechanisms such as technological upgrading and process innovation. Qiao and Sun [27] confirmed that digital capabilities significantly improve resource utilization in prefabricated construction. Additionally, Yang’s empirical study [28] demonstrated that advancements in digital construction practices can effectively reduce the carbon intensity of construction enterprises. Collectively, these studies provide systematic insights into the mechanisms and practical approaches through which digital technologies contribute to the sustainable development of the construction industry. In summary, the advancement of DTCI relies not only on technological innovation but also on parallel organizational adjustments and the support of green development pathways, offering a robust theoretical foundation for future policy-oriented research.

However, current research on DTCI-related policies—particularly at the local government level—remains relatively limited and lacks systematic analysis. The existing literature predominantly focuses on technological pathways and industrial applications, while in-depth investigations into policy implementation effectiveness and evaluation mechanisms are scarce. As a result, there is insufficient guidance for policy refinement and urban governance practices. Accelerating the DTCI process requires robust policy support, which plays a vital role in guiding enterprises, offering incentives, bridging gaps in funding and technological application, fostering digital integration, and promoting the widespread adoption of digital tools, thereby expediting industry transformation [29,30]. The effectiveness of policy measures hinges on rigorous policy research to optimize the formulation and implementation processes, ensuring feasibility and rationality while advancing industry development toward its intended goals [31]. One of the major challenges in China’s DTCI efforts is regional disparity [32]. The efficiency of the construction sector across different regions largely depends on local governments’ ability to design digital development plans tailored to regional characteristics. A regional coordinated development strategy can help leverage regional advantages and enhance the overall efficiency of digital transformation efforts [33]. The lack of systematic research on local policy frameworks has resulted in fragmented and reactive approaches by local governments, limiting their ability to proactively drive the transformation agenda [34]. Against this backdrop, there is an urgent need to strengthen systematic research on local policies to support the efficient advancement of DTCI. Accordingly, this study aims to fill the research gap by focusing on local governments’ roles in DTCI, examining their policy responses and implementation mechanisms, evaluating policy outcomes, identifying key bottlenecks, and proposing feasible strategies for policy optimization.

This study selects Wuhan as the focal case for analysis, based on the following three considerations: First, Wuhan, as a core city in central China, holds significant representativeness and serves as a policy experimentation hub. According to data from the Hubei Provincial Department of Housing and Urban–Rural Development [35], the total output value of the construction industry in Hubei Province is projected to reach RMB 2.2 trillion in 2024, marking a year-on-year growth of 5.8%. Wuhan alone accounts for RMB 1.36 trillion—62% of the provincial total—with an annual growth rate of 11.3%. As a national central city and a strategic intersection of the Belt and Road Initiative and the Yangtze River Economic Belt, Wuhan has long undertaken national-level policy pilot tasks and has been designated as one of China’s demonstration cities for digital transformation in manufacturing. Its practices in DTCI not only reflect national strategic priorities but also embody local policy innovation, thereby offering rich empirical cases for analysis. Second, Wuhan’s strategic transportation position and industrial agglomeration effect provide a robust economic foundation and transformation impetus for this research. As a major transportation hub in China, Wuhan’s well-developed logistics network extends across the country, granting high transferability and regional applicability to research findings. Third, Wuhan’s local policy experimentation in DTCI demonstrates strong international comparability and adaptability. Several global cities—such as London (UK), Singapore, and Helsinki (Finland)—have also pursued locally driven approaches to DTCI. As a regional policy pilot city in central China, Wuhan has introduced innovations in financial support, digital infrastructure development, and public–private collaboration mechanisms. These experiences offer valuable insights from China for other large cities worldwide. Therefore, selecting Wuhan as a case study helps to deepen the understanding of DTCI policy evolution from the perspective of local governance practice while enhancing the global applicability and dissemination potential of the theoretical outcomes.

The theoretical foundation of this study is grounded in public policy theory, encompassing policy implementation, evaluation, and process research [36]. Policy implementation focuses on the transition from policy formulation to practical execution [37,38,39], while policy evaluation employs retrospective analysis to assess policy outputs and impacts [40,41]. Policy process research, in contrast, explores the full lifecycle and dynamic evolution of policy mechanisms [42,43]. With the advancement of the 14th Five-Year Plan for Digital Economy Development, the conditions are now favorable for evaluating policy implementation outcomes. This study adopts a policy evaluation perspective to systematically analyze the effectiveness of existing policies and identify areas for improvement, thereby offering scientific guidance for future policy refinement [44,45]. In the existing literature, policy evaluations often adopt multidimensional frameworks. For example, Li [46] carried out a quantitative assessment of carbon-emission regulations in the construction sector from the standpoint of policy instruments; Liu [47] evaluated building energy-conservation policies using a two-axis model based on policy objectives and instruments; and Zhang [48] proposed a three-dimensional evaluation system tailored to DTCI. These studies provide valuable insights into policy performance; however, most lack a coherent hierarchical logic chain capable of revealing causal relationships and dynamic interactions among policy components. This limitation constrains the precision of policy design and the effectiveness of implementation. To address this gap, recent studies have increasingly adopted a “macro–meso–micro” (3M) tiered framework to enable layered analysis and multidimensional integration. For example, Zhou [49] constructed a 3M model for digital healthcare services in aging populations, elucidating the interaction and behavioral mechanisms among policy elements across levels. Hedayatnezhad [50] employed the 3M framework to explore the multilevel driving forces behind urban regeneration in aging communities, demonstrating methodological systematization. Wang [51] applied the framework to evaluate urban green transformation policies, highlighting its broad applicability and analytical strength in complex policy assessment scenarios.

Building upon existing research, this study proposes an innovative 3M hierarchical evaluation framework tailored to local DTCI policy assessment. Unlike prior studies that primarily emphasize the structural delineation across macro, meso, and micro levels, this framework further highlights the transmission mechanisms among the three tiers. It emphasizes the coherence and synergy across strategic-level policy formulation (macro level), pathway design (meso level), and concrete implementation (micro level). Methodologically, the framework integrates multiple analytical tools: it adopts the PESTEL framework to conduct macro-level analysis of policy texts, applies the Policy Modeling Consistency (PMC) index model to evaluate the internal coherence of meso-level policy pathways, and utilizes Spearman’s rank correlation coefficient to identify critical variable linkages at the micro level. The result is a closed-loop, full-process evaluation system that balances analytical breadth and depth, facilitating a comprehensive understanding of the policy environment and enabling the precise identification of implementation bottlenecks. This significantly enhances the scientific rigor and policy relevance of optimization efforts.

Based on this framework, Wuhan is selected as the case city, and this research focuses on the following objectives:

(1)

To construct an innovative, progressive 3M policy evaluation system that clarifies the interaction mechanisms across hierarchical policy elements;

(2)

To assess the implementation effectiveness and contextual adaptability of local DTCI policies and identify key barriers and influencing factors within the policy transmission chain;

(3)

To propose targeted policy optimization strategies that support effective local DTCI advancement and offer replicable insights for other regional governments.

2. Methodology: Sample Sources, Regional Analysis, and Research Framework

This chapter introduces the sample sources, regional analysis, and research framework to clarify the basis for sample selection, analyze the regional coverage of policies, and construct a theoretical support framework.

2.1. Data Collection and Sources

Sample selection plays a pivotal role in ensuring the scientific validity and accuracy of research, as the degree of alignment between the selected samples and the research topic directly impacts the credibility of policy evaluation outcomes [52]. Adopting a bibliometric approach, this study first identifies the core dimensions of digital transformation within the construction industry. CiteSpace [53] was employed to conduct a thematic clustering analysis of 586 publications (spanning 2005–2024) retrieved from the Web of Science database under the topic “Digital transformation in the construction industry” (see Figure 1). Based on keyword frequency and node centrality, the top five keywords were extracted, and non-representative terms such as “construction”, “framework”, and “performance” were excluded. As a result, three fundamental characteristics of the industry’s digital transformation were distilled: industrialization, digitalization, and intellectualization. Specifically, industrialization is underpinned by key technologies such as BIM (frequency = 64) and Industry 4.0 (frequency = 43); digitalization is driven by process reengineering centered around technology (frequency = 71); and intellectualization is characterized by collaborative governance mechanisms represented by the keywords management (frequency = 67) and system (frequency = 27).

Based on the aforementioned keywords and their conceptual scope, the official website of the Wuhan Municipal Government [54] (https://www.wuhan.gov.cn, accessed on 10 April 2024) was selected as the authoritative data source, with the study period spanning from 2021 to 2023. A multi-tiered retrieval strategy was employed: the core search terms included “digitalization in the construction industry” and “intelligent construction”, while the extended keywords comprised cross-disciplinary terms such as “BIM”, “Industry 4.0”, and “prefabricated buildings”, closely related to industrialization and intelligent technologies. The research sample primarily consisted of policy documents, notices, regulations, and opinions publicly issued by the Wuhan municipal government, district and county governments, functional zones, and affiliated institutions. Through manual screening, documents irrelevant to digital transformation or with weak association to the construction sector were excluded to ensure the relevance and validity of the sample. Ultimately, 33 policy documents with substantive policy guidance were identified, including 20 municipal-level policies and 13 district, county, and functional zone policies (see Appendix A).

2.2. Regional Disparities in Policy Coverage

Regional development inequities significantly shape policy implementation intensity, resulting in resource allocation imbalances that undermine spatial justice in economic growth patterns [55]. As depicted in Figure 2, the cartographic representation delineates Wuhan’s administrative divisions alongside construction sector digital policy penetration gradients. Geospatial analysis reveals distinct implementation clusters: blue zones denote robust regulatory frameworks, while red territories signal policy provision deficits (the district-level policies related to DTCD in this area are fewer than two).

Areas with weak policy support are primarily concentrated in the northern districts of Huangpi and Xinzhou. These regions are large but have demonstrated slower digital transformation advancement due to limited development efforts and delayed policy implementation. Additionally, the central and eastern districts, such as Hongshan and Qingshan, face similar challenges owing to insufficient policy support. Notably, the East Lake High-Tech Development Zone has achieved substantial progress in digital transformation through strengthened policy support, demonstrating the pivotal role of institutional arrangements in regional development. Uneven policy coverage may impede digital transformation progress in under-resourced regions, thereby constraining their capacity to harness emerging technological advantages and potentially retarding industry-wide advancement. Particularly in the construction sector of these areas, digital adoption persists at a competitive disadvantage, with limited prospects for bridging developmental gaps with economically prosperous regions.

2.3. Analytical Framework Design

The scientific selection of research methods and the construction of a systematic analytical logic are fundamental for evaluating complex policy systems [56]. Focusing on the digital transformation policy practices in Wuhan’s construction industry, this study proposes an integrated assessment framework that is hierarchical and progressive and combines both qualitative and quantitative approaches. The analytical structure spans three levels, macro, meso, and micro, with the overall process illustrated in Figure 3.

At the macro level, this study employs the PESTEL analytical model in conjunction with text mining techniques to analyze the frequency and visualization of key terms within policy documents. The PESTEL model encompasses six dimensions of the external environment—political, economic, social, technological, environmental, and legal—offering a holistic perspective for evaluating the focus areas and coverage of policies [57]. Additionally, a co-occurrence network analysis is conducted to reveal the connection patterns and centrality structures among keywords, thereby uncovering the underlying thematic structures and potential strategic directions of the policies [58]. This method not only provides strategic insights for policymakers but also serves as a foundational basis for constructing indicators at the meso level.

At the meso level, this study adopts the PMC (Policy Modeling Consistency) index model to quantitatively assess the policy samples. Based on the “Omnia Mobilis” hypothesis proposed by Ruiz Estrada, the model builds a multidimensional index system to evaluate policy effectiveness in a systematic manner [59]. Through the development of primary and secondary indicators, the model constructs a matrix for index computation to quantitatively analyze the structural completeness and internal consistency of policies. The PMC surface diagram is employed to visually present the strengths and weaknesses within the policy structure, allowing for targeted recommendations for policy improvement [60]. This layer of analysis provides a solid empirical foundation for micro-level mechanism exploration.

At the micro level, this study introduces the Spearman’s rank correlation coefficient to explore the relationships between policy indicators and their correlation with policy performance classifications. As a nonparametric statistical method suitable for identifying nonlinear relationships [61], Spearman’s analysis aids in identifying key variables embedded within the policy structure. By examining the significance and correlation levels of indicators, this study filters out policy factors requiring optimization, thereby offering empirical evidence for improving the meso-level framework. Simultaneously, it provides focused suggestions for policy implementation and operational refinement.

3. A 3M Progressive Policy Evaluation System

This chapter systematically outlines the analytical approach and technical methods used for policy evaluation. This research assesses the digital transformation policies in Wuhan’s construction industry from macro, meso, and micro perspectives, establishing an empirical and operable evaluation system through the integration of text mining, quantitative modeling, and statistical analysis tools.

3.1. Macro-Level Analysis: Policy Text Mining via PESTEL Framework

This section aims to identify the macro-level dimensions reflected in the policy documents by classifying high-frequency terms according to the PESTEL framework. This classification reveals the focus areas and possible gaps in policy attention across different external environmental factors. As a well-established tool for macro-environmental assessment, PESTEL has been extensively applied in strategic and policy-related studies. For instance, Ulubeyli [62] explored its applicability in the strategic analysis of the construction sector, Makvandi [63] utilized it to determine key external influences on the evolution of financial products, and De [64] applied it to evaluate the macro-level development conditions of Brazil’s new energy vehicle industry. These examples demonstrate the framework’s effectiveness in highlighting both the strengths and limitations of policies in adapting to external contexts.

3.1.1. Text Mining and Frequency Statistics

To enable a structured and quantitative review of policy content, this study employs ROSTCM6 (Vision 5.8) software for high-frequency term extraction and categorization, following a thorough review of the selected policy texts. The preprocessing phase involves standardizing the documents, eliminating redundant words, and unifying terminology to improve the accuracy of term frequency analysis. Subsequently, high-frequency terms indicative of policy orientation are identified through a combination of automated term statistics and manual selection. A co-occurrence network is then constructed using the NetDraw module, which visualizes the strength of connections and centrality among terms, thus exposing the underlying thematic structure of the policy discourse. Based on this network, keywords are classified into the six dimensions of the PESTEL framework, and the distribution of elements across each category is analyzed. The size of nodes and the density of links within the co-occurrence graph indicate the prominence and interrelation of terms, providing insights into policy emphasis and potential dimension imbalances. Existing studies have confirmed the practical value of ROSTCM6 in text mining and semantic network analysis, making it a valuable tool for enhancing policy structure and coherence [65,66].

3.1.2. Policy Indicator Derivation

Based on the weak links identified via high-frequency word analysis, this study proposes relevant indicators to support the development of a meso-level framework for assessing policies.

3.2. Meso-Level Analysis: Quantitative Policy Evaluation

This research adopts the PMC index approach to systematically analyze internal policy mechanisms and identify optimization pathways. Unlike the PESTEL framework, which assesses multidimensional external influences but lacks quantitative precision, the PMC method quantifies individual variables’ contributions to policy outcomes, providing a data-driven basis for improvement.

3.2.1. Indicator System Construction

The evaluation framework developed in this study integrates the textual analysis of policy documents with high-frequency word mining, drawing on Ruiz Estrada’s PMC index methodology [67] and prior research in policy assessment [68,69,70,71,72,73,74,75,76,77]. The constructed system comprises 10 tier-1 and 40 tier-2 metrics (

Table 2. Policy indicator evaluation system.

Primary Indicators	Secondary Indicators	Primary Indicators	Secondary Indicators
$X_1$: policy nature	$X_{1:1}$: planning	$X_6$: policy recipients	$X_{6:1}$: provincial level
	$X_{1:2}$: guidance		$X_{6:2}$: municipal level
	$X_{1:3}$: description		$X_{6:3}$: district and county level
	$X_{1:4}$: suggestions	$X_7$: policy assessment	$X_{7:1}$: clear positioning
	$X_{1:5}$: regulation		$X_{7:2}$: sufficient basis
			$X_{7:3}$: detailed measures
$X_2$: policy focus	$X_{2:1}$: talent cultivation	$X_8$: policy fields	$X_{8:1}$: political
	$X_{2:2}$: technological innovation		$X_{8:2}$: economic
	$X_{2:3}$: outcome transformation and integration		$X_{8:3}$: social
	$X_{2:4}$: market mechanism		$X_{8:4}$: technological
	$X_{2:5}$: service platforms		$X_{8:5}$: environment
	$X_{2:6}$: practical applications		$X_{8:6}$: legal
$X_3$: policy timeliness	$X_{3:1}$: long-term (five years or more)	$X_9$: incentive measures	$X_{9:1}$: fiscal subsidies
	$X_{3:2}$: mid-term (three to five years)		$X_{9:2}$: talent incentives
	$X_{3:3}$: short-term (within three years)		$X_{9:3}$: tax reductions
$X_4$: policy effectiveness	$X_{4:1}$: low-carbon ecology		$X_{9:4}$: pilot projects
	$X_{4:2}$: improving people’s livelihood		$X_{9:5}$: awareness and outreach initiatives
	$X_{4:3}$: cost reduction and efficiency improvement		$X_{9:6}$: financial support
	$X_{4:4}$: demonstration and promotion
$X_5$: policy targets	$X_{5:1}$: government departments	$X_{10}$: policy transparency	$X_{10:1}$: openness
	$X_{5:2}$: enterprises		$X_{10:2}$: non-openness

).

3.2.2. Policy Evaluation Matrix Design

Guided by multi-input–output table principles, this study constructed a standardized evaluation matrix (

Table 3. Policy evaluation matrix.

Primary Indicators	Secondary Indicators
$X_1$	$X_{1:1}$	$X_{1:2}$	$X_{1:3}$	$X_{1:4}$	$X_{1:5}$
$X_2$	$X_{2:1}$	$X_{2:2}$	$X_{2:3}$	$X_{2:4}$	$X_{2:5}$	$X_{2:6}$
$X_3$	$X_{3:1}$	$X_{3:2}$	$X_{3:3}$
$X_4$	$X_{4:1}$	$X_{4:2}$	$X_{4:3}$	$X_{4:4}$
$X_5$	$X_{5:1}$	$X_{5:2}$
$X_6$	$X_{6:1}$	$X_{6:2}$	$X_{6:3}$
$X_7$	$X_{7:1}$	$X_{7:2}$	$X_{7:3}$
$X_8$	$X_{8:1}$	$X_{8:2}$	$X_{8:3}$	$X_{8:4}$	$X_{8:5}$	$X_{8:6}$
$X_9$	$X_{9:1}$	$X_{9:2}$	$X_{9:3}$	$X_{9:4}$	$X_{9:5}$	$X_{9:6}$
$X_{10}$	$X_{10:1}$	$X_{10:2}$

) for systematic data processing and analysis. The framework consists of 10 primary indicators, each containing multiple secondary indicators, with equal weights allocated to all measures.

3.2.3. PMC Index Computation

The calculation process involves four key steps:

1.

Variable construction: The hierarchical policy variables are defined based on the analysis of 33 policy documents, as shown in Equation (1). These variables serve as the foundation for the subsequent steps in the calculation.

$$X\sim N(0,1)$$

(1)

2.

Multi-input–output table: A binary scoring system is used to evaluate each secondary indicator associated with the primary indicators. If a secondary indicator meets the specified criteria, it is assigned a score of 1; otherwise, it is assigned 0. This scoring system is captured in Equation (2).

$$X=\{XR:(0,1)\}$$

(2)

3.

Primary variable calculation: The primary indicators are calculated by aggregating the corresponding secondary indicators. Equation (3) demonstrates this calculation, where each primary indicator value is the weighted average of its associated secondary indicators.

$$X_t=\sum _{j=1}^n\frac{X_{tj}}{T\left(X_{tj}\right)}t=1,2,3,4,5,6,7,8,9,10$$

(3)

4.

Final PMC index calculation: The final PMC index is the sum of all primary indicator values. This is formally expressed in Equation (4), where the value of each primary indicator X_t is computed as the sum of the normalized values of its corresponding secondary indicators X_tj. The final PMC index is the aggregate of all primary indicator scores:

$$\begin{array}{c}PMC=X_1+X_2+X_3+\dots +X_{10}\\ =\left({\displaystyle \sum _{j=1}^n}\frac{X_{1}j}{T\left(X_{1}j\right)}\right)+\left({\displaystyle \sum _{j=1}^n}\frac{X_{2}j}{T\left(X_{2}j\right)}\right)+\left({\displaystyle \sum _{j=1}^n}\frac{X_{3}j}{T\left(X_{3}j\right)}\right)+\dots +\left({\displaystyle \sum _{j=1}^n}\frac{X_{9}j}{T\left(X_{9}j\right)}\right)+X_{10}\end{array}$$

(4)

where X_t represents the primary indicator for t = 1, 2, 3, 10, X_tj represents the secondary indicator for each primary variable, and T(X_tj) is the total number of secondary indicators for each primary indicator. This summation captures the overall effectiveness of the policy by evaluating the contribution of each primary indicator, normalized by the total number of associated secondary indicators.

3.2.4. Visualization of PMC Surface Plot

This study used Equation (5) to construct a PMC surface plot, providing a visual representation of the PMC index results for policies. While constructing the surface matrix, only the first nine primary indicators (X₁–X₉) were selected, as the policies analyzed were publicly available documents and did not include the primary indicator X₁₀. Owing to the equal number of row and column variables, the PMC matrix exhibited regular symmetry, facilitating a comparison between policies. The specific steps for constructing the matrix are as follows:

• The first nine primary variables (X₁, X₂, X₃… X₉) and their parameter values were selected.
• Using Equation (5), the parameter values of the three variables were grouped to construct a 3 × 3 matrix:

  $$PMC=\left(\begin{array}{c}{X}_1\hspace{1em}{X}_2\hspace{1em}{X}_3\\ X_4\hspace{1em}{X}_5\hspace{1em}{X}_6\\ X_7\hspace{1em}{X}_8\hspace{1em}{X}_9\end{array}\right)$$

  (5)

3.3. Micro-Level Analysis: Policy Effectiveness Correlation via Spearman’s Rank Test

To examine the relationships between policy indicators and evaluation levels, Spearman’s rank correlation coefficient was employed. This nonparametric method effectively captures monotonic dependencies, revealing how policy indicators influence evaluation outcomes. While the PMC index provides a comprehensive quantitative assessment framework and highlights both strengths and weaknesses through surface plot concavity analysis, it fails to identify key effectiveness determinants. Therefore, Spearman’s analysis was conducted to identify crucial indicators, offering evidence-based support for policy refinement.

3.3.1. Data Preprocessing and Validation

The data preparation and validation processes are as follows:

• Data preparation: based on the scores of nine primary policy indicators, final policy evaluation ratings (data from Section 4.2.1), and secondary policy indicator scores (data from Appendix B), IBM SPSS Statistics (Vision 29.0) was used for data preprocessing to construct the required research sample.
• Data transformation: normality tests and comparative analysis of correlation methods [78,79] confirmed the dataset’s non-normal distribution and absence of linear relationships, justifying the use of Spearman’s rank correlation for assessing monotonic dependencies.
• Correlation analysis: The coefficient was computed using Equation (6).

  $$r_s=1-\frac{6\sum d_i^2}{n\left(n^2-1\right)}$$

  (6)

  where “$r_s$” denotes Spearman’s coefficient (−1 to +1). $d_i$ reflects rank differences for the i-th sample. Values are interpreted as +1 (perfect positive), −1 (perfect negative), and 0 (no) correlation.

In addition, p-values were obtained through SPSS’s two-tailed significance test function to assess the statistical reliability of the observed correlations. A significance level of p < 0.05 was used to determine whether the relationships were statistically meaningful [61]. The Spearman’s coefficients were categorized as follows:

• 0.00–0.19 (negligible);
• 0.20–0.39 (weak);
• 0.40–0.59 (moderate);
• 0.60–0.79 (strong);
• 0.80–1.00 (high).

A correlation coefficient with a p-value lower than 0.05 is deemed statistically significant, suggesting a meaningful relationship between the indicators and the evaluation results. The stronger the correlation coefficient of a policy indicator, the larger the variation in the policy scores for that indicator, which indicates a more significant influence of the indicator on the policy outcomes.

3.3.2. Correlation Between Primary Indicators and Policy Effectiveness

The correlation coefficients between the scores of nine primary policy indicators and policy evaluation levels for all 33 policies were calculated using Equation (6). The results showed correlation differences among each primary indicator, and through p-value testing, the primary policy indicators requiring optimization were identified, providing feedback for comprehensive policy optimization path at the macro level.

3.3.3. Hierarchical Correlation Among Primary–Secondary Indicators

Employing Equation (6), we examined the relationships between core primary indicators and their corresponding secondary components across all 33 policies. This approach pinpointed the most influential secondary indicators per primary category, uncovering critical policy implementation gaps and informing targeted optimization strategies.

4. Results

This section synthesizes policy assessment findings through integrated macro-, meso-, and micro-level analyses, demonstrating their hierarchical interrelationships.

4.1. Macro-Level Policy Focus and Deficiencies

This section identifies key policy characteristics through term frequency and centrality analysis, highlighting both effective elements and limitations to inform meso-level indicator development.

4.1.1. PESTEL-Based Policy Theme Distribution

Based on the six domains of the PESTEL framework, 40 representative high-frequency terms extracted from policy samples were categorized (see

Table 4. Categorization of ley high-frequency terms.

Dimension	Keyword Frequency
Political	Construction (1970), Government (599), System (511), Planning (386), Demonstration (330), Reform (190), Pilot (160)
Economic	Enterprises (1396), Project (835), Management (463), Finance (346), Factors of Production (321), Market (291), Investment (241)
Social	Service (951), Talent (362), Cultivation (308), Livelihood (223), Integration (208)
Technological	Innovation (900), Technology (618), Platform (448), Design (403), Science and Technology (396), Application (367), Intelligent Construction (325), Industry 4.0 (276), BIM (223), Smart City (218), AI (156)
Environmental	Ecology (1175), Environment (644), Natural Resources (148), Green Building (129), Low Carbon (104)
Legal	Security (260), Regulation (220), Standard (200), Regulatory Framework (192), Oversight (91)

).

The frequency of high-frequency words reveals the focus areas and priorities of policies, whereas the centrality of the word co-occurrence network demonstrates interconnections among different fields (see Figure 4).

The PESTEL-based analysis of high-frequency policy terms reveals critical implementation factors and challenges:

• Political dimension: The term “Construction” appears most frequently and is centrally positioned, indicating that the goal of digital transformation is to drive the overall industry’s intelligent development, upgrading digital infrastructure. The high-frequency terms “Government” and “System” underscore governmental dominance in shaping digital policies and frameworks. “Planning” signals reliance on strategic long-term deployment, while “Demonstration” reveals a risk-averse approach through pilot programs. In contrast, the relatively low frequency of “Reform” and “Pilot” suggests insufficient efforts in promoting industry innovation and implementing pilot programs, which limit technological and managerial innovation, thus affecting transformation efficiency.
• Economic dimension: “Enterprise”, as a core term, underscores its key role in driving technological applications and market development in digital transformation. The term “Project” frequently appears and holds a high centrality, indicating its importance, particularly in the application of technology and platforms. The high frequency of “Management” emphasizes the higher demands of digital transformation in industry management models, covering smart construction, supply chain optimization, and data-driven decision-making, significantly improving efficiency. The frequent occurrence of “Finance” and “Market” highlights the critical role of policy and the market environment in digital investment. The term “Production Factors” shows the focus of a policy on resource optimization, but the lower frequency of “Investment” suggests that insufficient funding may constrain the advancement and depth of digital transformation.
• Social dimension: The term “Service” appears frequently with high centrality, highlighting the pivotal role of digital innovation in the service sector of the construction industry, which drives emerging businesses, such as smart construction and smart cities. The terms “Talent” and “Cultivation” point toward the increasing demand for skilled professionals, reflecting the focus of a policy on talent development. The term “Livelihood” appears infrequently with low centrality, indicating that the role of digitalization in the social security system is overlooked, particularly in areas such as smart housing and infrastructure construction. The low frequency of “Integration” reveals the insufficient integration of technology and collaborative application in digital transformation.
• Technological dimension: The term “Innovation” appears frequently with strong centrality, indicating that it is the core driving force behind digital transformation. The terms “Platform” and “Application” emphasize the crucial role of digital platforms in the transformation process, whereas “Design” highlights the importance of digital design throughout the project lifecycle. The frequent mention of “Intelligent Construction” hints at its growing importance in the direction of transformation that drives integration of technology, platforms, and innovation. “Industry 4.0” indicates that the construction industry is transitioning toward intelligent manufacturing models, promoting integration with other manufacturing sectors. The frequent occurrence of “Science and Technology” reflects the emphasis of the policy on technological advancement. The lower frequency of terms such as “BIM”, “AI”, and “Smart City” suggests insufficient attention to and depth of application in frontier technologies.
• Environmental dimension: The term “Ecology” demonstrates strong centrality, highlighting ecological sustainability as a priority in construction digitalization. Concurrently, “Environment” reflects policy-driven green development initiatives. However, the low frequency of terms such as “Natural Resources”, “Green Building”, and “Low Carbon” reveals deficiencies in policy support for promoting sustainable buildings and low-carbon ecological concepts. In particular, insufficient focus on the optimal use of natural resources, the promotion of green buildings, and support for low-carbon technology implementation suggests a lack of attention and resource allocation in these areas.
• Legal dimension: The term “Security” appears with low frequency and weak centrality, indicating insufficient attention to data and information security issues in the policy. The low-frequency terms “Regulation” and “Standards” indicate weaknesses in the legal framework and standardization, resulting in compliance deficiencies and implementation inconsistencies. The low frequency of terms such as “Regulatory Framework” and “Oversight” indicates insufficient attention to regulatory mechanisms, which may result in ineffective supervision, hindering the smooth progress of industry transformation.

4.1.2. Strategic Implications of Macro-Level Findings

The high-frequency term analysis identifies policy strengths and weaknesses in Wuhan’s construction sector digitization. The legal dimension shows the least attention, especially regarding frameworks, system standardization, and regulatory mechanisms, with inadequate data security focus. However, the analysis reveals several areas requiring improvement across policy dimensions. Politically, support for industrial innovation and pilot projects remains inadequate. Economically, digital transformation lacks sufficient financial backing. Socially, insufficient attention has been given to digitalization’s role in social security systems, coupled with poor technology integration. Technologically, the adoption rates of advanced solutions remain low. Environmentally, greater efforts are needed to promote green building and low-carbon concepts.

To address the identified gaps and enhance the PMC model’s effectiveness, key policy indicators are recommended:

• Political dimension: X_1:2 (guidance) and X_9:4 (pilot projects);
• Economic dimension: X_2:4 (market mechanism) and X_9:6 (financial support);
• Social dimension: X_4:2 (improving people’s livelihood) and X_2:3 (outcome transformation and integration);
• Technological dimension: X_2:2 (technological innovation) and X_2:6 (practical applications);
• Environmental dimension: X_4:1 (low-carbon ecology);
• Legal dimension: X_1:5 (regulation), X_7:2 (sufficient basis), and X_7:3 (detailed measures).

4.2. Meso-Level Policy Effectiveness and Optimization Pathways

This research employed the PMC approach to assess policy effectiveness. Macro-level term frequency analysis informed system optimization, with dimensional quantitative evaluations conducted. The results from the surface chart guided policy refinement, while index scores facilitated subsequent correlation analysis.

4.2.1. PMC Index Scores and Policy Grading

By analyzing the 33 policy samples, secondary indicators of the multi-input–output table were assigned values (see Appendix B). Based on the results and by applying Equations (1)–(4), the PMC index was calculated, yielding PMC scores for all 33 policies and their overall averages (see

Table 5. PMC index score table.

Serial Number	${\mathit{X}}_1$	${\mathit{X}}_2$	${\mathit{X}}_3$	${\mathit{X}}_4$	${\mathit{X}}_5$	${\mathit{X}}_6$	${\mathit{X}}_7$	${\mathit{X}}_8$	${\mathit{X}}_9$	${\mathit{X}}_{10}$	Total Score	Score Ranking
P1	0.6	1	0.333	1	1	0.667	0.333	1	0.167	1	7.1	29
P2	1	0.833	1	1	1	0.667	0.667	1	0.5	1	8.667	9
P3	1	1	0.667	1	1	0.667	1	1	0.5	1	8.834	6
P4	0.8	1	0.333	1	1	1	0.667	1	0.667	1	8.467	12
P5	1	1	0.333	0.75	1	1	0.667	1	0.667	1	8.417	13
P6	1	1	1	1	1	0.667	1	1	0.667	1	9.334	2
P7	0.8	0.667	0.667	1	1	0.667	1	1	0.333	1	8.134	18
P8	1	1	0.333	1	1	0.667	1	1	0.5	1	8.5	11
P9	0.8	0.333	0.333	1	1	0.667	1	1	0.667	1	7.8	23
P10	0.8	0.5	0.333	0	1	0.667	0	1	0.667	1	6.967	32
P11	0.8	1	0.333	0.75	1	0.667	0.667	0.833	0.833	1	7.883	21
P12	1	1	0.667	1	1	0.667	1	1	0.667	1	9.001	4
P13	1	0.667	0.667	1	1	0.667	0	0.833	0.333	1	7.167	28
P14	1	0.833	0.333	1	1	0.667	1	1	0.667	1	8.5	10
P15	0.8	0.833	0.333	1	1	0.667	1	1	0.667	1	8.3	16
P16	1	1	0.333	1	1	0.667	1	1	0.833	1	8.833	7
P17	1	0.333	0.667	0.75	1	0.667	0.667	1	0	1	7.084	30
P18	0.8	0.833	0.333	0.75	1	0.667	1	0.833	0.333	1	7.549	26
P19	1	0.667	0.333	1	1	0.667	1	1	0.167	1	7.834	22
P20	0.8	0.5	0.333	1	1	0.667	0.667	1	0.333	1	7.3	27
P21	0.6	1	1	1	1	0.667	0	1	0.333	1	7.6	25
P22	1	0.667	1	1	1	0.667	1	1	0.333	1	8.667	8
P23	0.8	1	0.333	1	1	0.333	1	1	0.667	1	8.133	19
P24	1	1	1	1	1	0.333	1	1	0.833	1	9.166	3
P25	1	1	0.667	1	1	0.333	0.667	1	0.667	1	8.334	14
P26	0.8	1	0.667	1	1	1	1	1	0.5	1	8.967	5
P27	1	1	0.333	0.75	1	0.333	1	1	0.333	1	7.749	24
P28	0.8	1	0.333	1	1	0.333	1	0.833	0.667	1	7.966	20
P29	0.6	0.333	1	1	1	0.333	0.333	1	0.333	1	6.932	33
P30	0.8	0.833	1	0.75	1	0.667	0.333	1	0.833	1	8.216	17
P31	1	0.833	0.333	1	0.5	0.333	0.333	1	0.667	1	6.999	31
P32	1	1	1	1	1	0.667	1	1	0.833	1	9.5	1
P33	1	0.833	0.333	1	1	0.333	1	1	0.833	1	8.332	15
Average	0.891	0.833	0.576	0.924	0.985	0.616	0.758	0.98	0.545	1	8.128	-

).

According to the policy ranking criteria set by Ruiz Estrada [67], the policy evaluation was refined into five levels, and the 33 policies were classified into the five levels. The evaluation results are as follows:

• Perfect (9–10 points): P₆, P₁₂, P₂₄, and P₃₂;
• Excellent (8–8.999 points): P₂, P₃, P₄, P₅, P₇, P₈, P₁₄, P₁₅, P₁₆, P₂₂, P₂₃, P₂₅, P₂₆, P₂₇, P₃₀, and P₃₃;
• Good (7–7.999 points): P₁, P₉, P₁₁, P₁₃, P₁₇, P₁₈, P₁₉, P₂₀, P₂₁, and P₂₈;
• Acceptable (5–6.999 points): P₁₀, P₂₉, and P₃₁;
• Poor (0–4.999 points): none.

The 33 policies achieved an average PMC score of 8.128, demonstrating overall “Excellent” effectiveness. However, 13 policies fell into “Good” or “Acceptable” ranges, revealing improvement opportunities. The implementation analysis identified varying strengths and weaknesses across policy tiers, warranting targeted optimization. The subsequent analysis utilized PMC surface plots to examine specific deficiencies.

4.2.2. PMC Surface Plots for Policy Weakness Diagnosis

Based on Equation (5), a 3 × 3 PMC index matrix was constructed, and PMC index surface plots were generated. The recessed areas in the surface plots reflect the inadequacy of the policy indicators. Figure 5 shows the PMC composite average index surface plot for all 33 policies. The overall surface is smooth but shows a trend of tilting toward both ends across four dimensions, namely, X₉ (incentive measures), X₃ (policy timeliness), X₆ (policy recipients), and X₇ (policy assessment).

In particular, several key deficiencies emerged in the policy evaluation. X₉ demonstrated insufficient development of incentive mechanisms, compromising its effectiveness. X₃ exhibited suboptimal timeliness due to the inadequate consideration of temporal spans and policy adaptability. X₆ faced limitations stemming from inter-policy disparities and insufficient regional coverage. Most notably, X₇ suffered from ambiguous positioning, weak justification, and flawed measure design, significantly undermining its implementation and continuous improvement capacity.

Policy text analysis enabled the categorization of all 33 surface plots. To clarify the direction for improvement, this study compared the first-level indicator scores of each policy with the overall average and identified shortcomings that fell below the average, indicating areas of policy implementation that required attention. The prioritization of optimization was based on the magnitude of score deviation, with indicators showing larger deviations given higher priority for improvement. For instance, X₂ should be prioritized for optimization if the score deviation of X₂ is greater than that of X₁. By comparing the surface plots of each policy with the comprehensive average surface plot, the strengths and weaknesses of all 33 policies were further analyzed. The findings are detailed below:

• Perfect-Level Policies: The surface plots show a smooth and balanced state, indicating that the indicators are well coordinated and the overall performance is perfect (Figure 6).

P₃₂, P₆, P₂₄, and P₁₂ were rated as “Perfect”, as these policies significantly outperformed the average, reflecting their high systematization and effectiveness. P₆ and P₁₂ provided comprehensive policy frameworks with strong macro guidance, whereas P₂₄ and P₃₂ demonstrated stronger regional specificity and execution effectiveness at the district and county levels. However, these policies continue to show deficiencies in tax reduction and talent incentives, and an optimization focus should be placed on enhancing X₉.

2.

Excellent-Level Policies: The surface plots exhibit some fluctuations, indicating that certain policy indicators perform well; however, there is room for improvement (Figure 7).

Policies P₂₆, P₂, and P₃₀ show deficiencies in X₉, X₁ (policy nature), and X₇, particularly in terms of supervision and performance evaluation. P₂₆ lacks an effective supervision mechanism, sufficient fiscal subsidies, tax reductions, and pilot projects; P₂’s support for digitalization is not sufficiently detailed, with unclear tax reductions and digital demonstration projects; and P₃₀ lacks industry research data, digital transformation demonstration projects, and clear supervision indicators. The recommended optimization paths are P₂₆, X₁–X₉; P₂, X₇–X₉; and P₃₀, X₇–X₄–X₁.

Policies P₃, P₂₂, and P₇ show deficiencies in X₉, X₂ (policy focus), and X₁, particularly in terms of fiscal subsidies and tax reductions. P₃’s tax reductions are insufficient and lack pilot projects during the promotion of industrial upgrading and innovation-driven initiatives. P₂₂ lacks a systematic approach to talent cultivation and recruitment in digital economy and industrial cluster construction, and its successful case promotion is also insufficient. P₇ fails to establish specialized talent development programs and digital service platforms for integrating advanced technologies. Additionally, the supervision standards and entities of P₇ are not defined, leading to an imperfect supervision mechanism. The recommended optimization paths are P₃, X₉; P₂₂, X₉–X₂; and P₇, X₉–X₂–X₁.

Policies P₁₆, P₁₄, and P₈ exhibit deficiencies in X₃, indicating that long- and medium-term planning is insufficient and that policy sustainability and operability need enhancement. P₁₆ lacks long-term mechanisms for intelligent construction; P₁₄ falls short in terms of promoting the marketization of data elements to empower digital economy; and P₈ lacks sufficient fiscal subsidies and tax reductions in X₉ and shows weakness while promoting digitalization pilot projects. The recommended optimization paths are P₁₆, X₃; P₁₄, X₃; and P₈, X₃–X₉.

Policies P₄, P₅, and P₁₅ demonstrate inadequate X₃ performance, reflecting insufficient multi-phase planning and requiring enhanced sustainability and operability. P₄ lacks detailed measures in X₇ during the promotion of full-process digital management and the “World Optics Valley” construction, and its X₁ lacks an effective supervision mechanism, affecting the progress of specific projects; in the context of intelligent manufacturing and industrial internet, P₅ lacks sufficient integration of green ecology and energy-saving measures in X₄ (policy effectiveness), and its support for the construction industry is limited, restraining digital transformation; P₁₅ lacks an effective supervision mechanism in X₁ during digital infrastructure construction and industrial digital upgrades, failing to ensure successful enterprise digital transformation. The recommended optimization paths are P₄, X₃–X₇–X₁; P₅, X₃–X₄–X₇; and P₁₅, X₃–X₁.

Policies P₂₅, P₃₃, and P₂₃ exhibit notable deficiencies in the X₆ dimension, primarily manifested in the limited scope of policy coverage, thereby constraining the overall effectiveness of policy coordination. P₂₅ lacks clear quantitative transformation goals and detailed technical application paths in X₇; P₃₃ lacks clear mid-term planning and short-term action guidelines in X₃, limiting policy continuity; and P₂₃ lacks clear long-term goals in X₃ and X₁ and an effective supervision mechanism, impacting policy execution. The recommended optimization paths are P₂₅, X₆–X₇; P₃₃, X₆–X₃; and P₂₃, X₆–X₃–X₁.

3.

Good-Level Policies: The surface plots show moderate fluctuations, indicating the uneven development of various indicators and the presence of multiple issues (Figure 8).

Policies P₂₈ and P₂₇ show deficiencies in X₆ and X₃, suggesting insufficient policy coverage and collaboration between municipal and district/county levels as well as unclear long-term and mid-term planning, which impacts policy sustainability and implementation. P₂₈ has insufficient coverage in the “Intelligent Valley” construction and in the integration of new technologies and industries and lacks a digital transformation legal framework and standards. The supervision entity of P₂₈ is also unclear; P₂₇, while promoting building enterprises to “Cloud and Chain”, shows inadequate coordination among policy recipients, incomplete mid-term planning, and insufficient X₉ such as tax reductions and lacks sufficient content related to environmental protection and low-carbon ecology in X₄. The recommended optimization paths are P₂₈, X₆–X₃–X₈–X₁, and P₂₇, X₆–X₃–X₉–X₄.

Policies P₁₁ and P₁₈ show deficiencies in X₃, X₄, X₈ (policy fields), and X₁, indicating unclear long- and short-term planning, inadequate supervision mechanisms, and insufficient support for legal frameworks and standards. In the innovation of software and information technology services, P₁₁ suffers from ambiguous technical guidance for smart manufacturing, inadequate public data openness, and the poor integration of ecological sustainability concepts in X₈ coupled with weak oversight mechanisms, collectively hindering policy execution. P₁₈ fails to establish a holistic strategy and sustained technical support for enabling key enterprises to advance construction sector digitization. In X₉, there are inadequate talent incentives and tax reductions, limited coverage of financial special funds, unclear green ecological goals, and incomplete relevant legal frameworks. The recommended optimization paths are P₁₁, X₃–X₈–X₄–X₇–X₁, and P₁₈, X₃–X₉–X₄–X₈–X₁.

Policies P₁₉ and P₁₇ show deficiencies in X₉ and X₂, indicating a lack of tax reductions, talent incentives, and financial support, as well as unclear optimization of talent cultivation and pilot demonstration. In enhancing review efficiency and establishing a result transformation platform, P₁₉ fails to develop coherent strategic objectives and operational plans in X₃ for improving review systems and transformation platforms. Meanwhile, P17 neglects to integrate sustainability metrics into X₄ when establishing urban management centers, while its X₇ lacks provisions for talent development, technology commercialization, and competitive market cultivation. The recommended optimization paths are P₁₉, X₉–X₃–X₂, and P₁₇, X₉–X₂–X₄–X₇.

Policies P₉ and P₂₀ show deficiencies in X₂, X₃, and X₁, indicating unclear talent cultivation, pilot demonstrations, long- and mid-term planning, and supervision entities and standards, with a lack of effective supervision and feedback mechanisms. In the adoption of technologies like drones, blockchain, and BIM, P9 exhibits unclear X₂, insufficient transformation mechanisms and resource allocation in X₃, and a lack of proper supervisory management in X₁. During the construction industry’s industrialization, digitalization, and green transition, P₂₀ demonstrates incomplete X₉, inadequate talent cultivation, and restricted business expansion. Additionally, X₇ lacks clear application scenarios and has weak risk control and supervision mechanisms. The recommended optimization paths are P₉, X₂–X₃–X₁, and P₂₀, X₂–X₃–X₉–X₇–X₁.

Policies P₂₁, P₁₃, and P₁ show deficiencies in X₇, X₉, X₈, X₁, and X₃, indicating unclear development directions and insufficient incentive measures. P₂₁ fails to clearly define the construction sector’s role in regional economic growth, while X₁ has an inadequate supervision mechanism, and X₉ lacks sufficient fiscal subsidies and tax incentives; P₁₃, in X₇, does not clearly describe the direction of digital development for building an international consumption center city, X₉ for talent cultivation and introduction is lacking, and the legal framework for data security and intellectual property protection in X₈ is incomplete; P₁ lacks specific guidance for construction industry digitalization in X₇, clear digital guidelines in X₁, and long-term planning in X₃, failing to effectively enhance industry capability or improve the modern industrial system. The recommended optimization paths are P₂₁, X₇–X₁–X₉; P₁₃, X₇–X₉–X₂–X₈; and P₁, X₇–X₁–X₉–X₃.

4.

Acceptable-Level Policies: The surface plots show significant fluctuations, reflecting the instability of policy effectiveness and notable differences across various indicators (Figure 9).

P₃₁, P₁₀, and P₂₉ show multiple deficiencies in construction industry digitization. All three policies display X₇ weaknesses, indicating the need for clearer digital transformation positioning. P₁₀ and P₂₉ demonstrate inadequate X₂ execution and supervision, impairing policy implementation. Furthermore, P₃₁ and P₁₀ exhibit X₃ shortcomings, particularly in long-term goal refinement and implementation planning. Regarding intelligent technology investment, P₃₁ lacks precise X₅ (policy objects) guidance, shows poor X₆–municipal coordination, and fails to define clear industry digitization objectives. In green development contexts, P₁₀ demonstrates incomplete X₁ regulation and goal-setting, plus inadequate talent development support. For green buildings and smart cities, P₂₉ lacks defined X₁ digitization pathways, displays weak X₆–municipal alignment, and provides insufficient X₉ incentives for talent and taxation. The recommended optimization paths are P₃₁, X₅–X₇–X₆–X₃; P₁₀, X₇–X₂–X₃–X₁; and P₂₉, X₂–X₇–X₁–X₆–X₉.

4.2.3. Priority Ranking of Policy Optimization

The scoring deviations of key policy indicators from the overall policy average can clarify the priority order for policy optimization paths while evaluating policy effectiveness. The specific optimization path is as follows: X₉–X₃–X₆–X₇–X₂–X₁–X₄–X₈–X₅. This path provides a clear direction for identifying and strengthening weak links in policy implementation, thereby helping enhance policy effectiveness.

However, considering that each policy has specific contexts and implementation scenarios and differing objectives, an optimization path based on a comprehensive index may not be applicable to all policies. Therefore, the reasonableness of the optimization path should be further examined. To ensure that policy adjustments effectively improve the overall performance, this study delves deeper into the correlation between key policy indicators and policy performance levels, exploring more rational optimization paths. Additionally, this study analyzes the correlation between each key policy indicator and secondary policy indicators, revealing the specific impact of secondary indicators on policy outcomes at the micro level and providing a theoretical basis for precise policy optimization decision-making.

4.3. Micro-Level Drivers of Policy Effectiveness

This section examines the correlations between primary policy indicators, performance ratings, and secondary indicators using the PMC index results. Stronger correlations indicate a greater influence on the final scores, revealing policy volatility and potential deficiencies.

4.3.1. Key Indicators Influencing Policy Performance

Using Equation (6), the correlation coefficients between the scores of nine primary policy indicators and effectiveness ratings of all 33 policies were calculated. The results indicate that all primary policy indicators are positively correlated with policy effectiveness ratings, and the p-value is inversely associated with the strength of the correlation (see Figure 10).

X₇ demonstrated the strongest association (r = 0.540, p = 0.001), showing a moderate correlation with policy performance, highlighting the importance of regular evaluations and clear objectives. Other significant variables included X₂ (0.492, p = 0.004), X₉ (0.470, p = 0.006), and X₁ (0.433, p = 0.012), all statistically significant (p < 0.05). These results emphasize that policy focus definition, incentive implementation, and transparency are critical for construction industry digitization. While X₃ showed a weaker correlation (0.374, p = 0.032), its statistical significance suggests that policy timeliness and update frequency remain relevant considerations.

X₈, X₄, and X₅ (policy target) showed weaker correlations (0.335, p = 0.057; 0.325, p = 0.065; 0.299, p = 0.091), none reaching significance. Their mean PMC indices (0.98, 0.924, 0.985) indicate relatively strong policy performance, though with limited influence on overall ratings. Despite low significance, these indicators warrant optimization attention—particularly X₈ and X₄’s evaluation criteria, which may overlook environmental protection and cost reduction. X₅’s design lacked adequate applicability. X₆ demonstrated minimal correlation (0.188, p = 0.295), suggesting that the constrained indicator scope may limit cross-domain policy evaluation applicability.

4.3.2. Hierarchical Impact of Secondary Indicators

The five primary policy indicators with a significance level of less than 0.05 were selected—X₇, X₂, X₉, X₁, and X₃. Based on Equation (6), the correlation coefficients between these primary policy indicators and their secondary indicators were calculated, and the secondary indicators with the highest correlation with each primary indicator were identified, revealing specific deficiencies in policy implementation. The related data are shown in Figure 11.

The secondary indicators demonstrating the strongest correlations with primary policy indicators all showed statistically significant relationships (p < 0.001). The key findings include the following: X_7:3 (r = 0.886) reveals policy implementation inconsistencies; X_2:1 (talent cultivation, r = 0.849) indicates inadequate talent support; X_9:4 (r = 0.739) suggests pilot promotion deficiencies; X_1:5 (r = 0.87) highlights oversight shortcomings; and X_3:1 (long-term, r = 0.595) shows a moderate correlation with long-term goal neglect. These correlations pinpoint specific weaknesses in primary indicators, offering targeted optimization directions.

4.3.3. Data-Driven Policy Adjustment Strategies

Minor discrepancies emerged between the two evaluation approaches: policy performance assessment using the PMC index (Method 2) and indicator–performance correlation analysis via Spearman’s coefficient (Method 3). Method 2 determined the priority order for optimization based on the comprehensive mean deviation ranking of the nine policy indicators, resulting in an optimization path order (X₉–X₃–X₆–X₇–X₂–X₁–X₄–X₈–X₅). However, mean-based ranking may not fully reflect the context and impact of each policy. In contrast, Method 3 analyzed the correlation between primary policy indicators and policy performance grades, resulting in a different priority order (X₇–X₂–X₉–X₁–X₃–X₄–X₈–X₅–X₆). However, the correlation-based ranking does not imply causality. Although the significance of p-values helps identify policy indicators that can be excluded, a comprehensive judgment should still be made in consideration of the specific context. Therefore, the optimization path should combine the advantages and limitations of both methods, considering both the score deviation and correlation analysis, to ensure practical applicability and relevance.

Further analysis revealed that the significance levels of X₄, X₈, X₅, and X₆ exceeded 0.05. In Method 2, the optimization priorities of X₄, X₈, and X₅ were relatively low, which aligns with the rankings in Method 3. Given that these indicators had relatively high mean scores, it is reasonable to assume that their optimization priorities were lower. In contrast, X₆ had a low mean score in Method 2 (0.616) but ranked third in terms of optimization priority. Method 3 showed a weak correlation for X₆ (r = 0.188, p = 0.295), with p > 0.05 indicating non-significant results. Considering that this indicator relies heavily on government levels and that there are significant functional differences across these levels, placing it third in the optimization path is clearly not appropriate.

Thus, policy optimization should prioritize PMC index score rankings while incorporating Spearman’s correlation analysis to exclude highly significant but low-priority indicators. Indicator balance and adaptability require special consideration across policy contexts and levels. The final optimization path, based on the ranking in Method 2, excludes X₄, X₈, X₅, and X₆, with significance greater than 0.05, leading to the final optimization path of X₉–X₃–X₇–X₂–X₁. Based on the correlation between these primary policy indicators and their subordinate indicators, the specific secondary indicator optimization measures are X_9:4, X_3:1, X_7:3, X_2:1, and X_1:5.

5. Discussion

In comparison with the existing literature, and compared with Liu’s study [47], which centers on building-energy policies and underscores the primacy of fiscal subsidies and economic incentives, our analysis shows that Wuhan’s construction sector has a far greater need for pilot and demonstration projects—an industry-specific preference that profoundly shapes the choice of policy instruments. Likewise, although Zhang [48] also confirms the critical role of incentives in digital transformation, their nationwide policy approach leans heavily on tax breaks and broad-based subsidies. In contrast, Wuhan’s acute talent shortage directs local authorities to prioritize human-capital incentives first, while the pressing need to strengthen legal frameworks in regional governance elevates rule-of-law safeguards to equal importance. These comparisons reveal that, even when pursuing identical policy goals, variations in regional industrial endowments (for example, workforce reserves) and institutional maturity can substantially redefine both the incentive architecture and its implementation path. Building on these insights and framed within the 3M research model, this paper therefore offers the following tailored policy recommendations to accelerate digital transformation in Wuhan’s construction industry.

Analyzing digital transformation policies across the macro, meso, and micro levels offers a systematic approach to advancing construction industry digitization. First, strategic policy directions should prioritize strengthening legal and regulatory frameworks to ensure that policies are effectively implemented within a legal framework. Building on this foundation, the policy optimization pathway should prioritize incentive measures, policy timeliness, policy assessment, policy focus, and policy nature, ensuring policy alignment across different levels, and facilitate a gradual, coordinated implementation. Specific implementation measures can be refined in the following five aspects:

• Introducing pilot projects to foster innovation: Designating specific regions or enterprises as pilot sites for digital transformation in the construction industry can provide innovative solutions, mitigate transformation risks, and stimulate innovation. Pilot project outcomes should align with societal needs and strengthen collaboration with governments, industry associations, and social organizations to ensure that the results of transformation are integrated with broader social development goals. This approach enhances technological innovation, social service integration, and mutual benefits for both industry and society.
• Balancing long-term goals with strategic planning: Policies should balance short- and long-term objectives while aligning themselves with industrial development trends. Eventually, technological innovation should drive industrial upgrades, integrate green building principles and low-carbon development with digital transformation strategies, and foster full-scale industry digitalization and intelligent development. In the short term, policies should focus on key areas, advance core technology applications, enhance policy support and financial investment, and ensure policy foresight and sustainability to facilitate a steady industrial transformation.
• Enhancing policy measures for comprehensive implementation: Based on empirical research and industry feedback, policies should be formulated to ensure practicality and effectiveness. Incentives should promote digital technology adoption and strengthen competitiveness. Policies should establish BIM, AI, and other technical standards to ensure platform compatibility and data security. The legal framework must be improved to enhance data protection and intellectual property rights management and ensure compliance and effective enforcement.
• Strengthening talent cultivation and investment: A comprehensive digital skills training system should be established to foster deep collaboration among enterprises, universities, and research institutions to nurture multidisciplinary talents. Targeted policies should be implemented to attract high-level digital technology professionals and ensure a steady supply of technical expertise. From an economic perspective, investment should be encouraged in digital technology research and talent development to ensure financial support and accelerate industry transformation, enhance innovation capacity, and reinforce a strong foundation for digital transformation.
• Enhancing supervision and governance: A modernized regulatory system should be developed to support digital transformation, leveraging big data and IoT technology for the real-time monitoring of policy implementation to ensure transparency and precision. The use of digital platforms should enhance management efficiency, and dedicated regulatory bodies should be established to conduct periodic policy evaluations and adjust relevant standards and regulations as required, strengthening industry supervision and governance mechanisms.

In summary, digital transformation policies must be implemented in a coordinated manner across the macro, meso, and micro levels, which will enable the integration of legal protection, strategic planning, detailed implementation measures, talent development, and regulatory mechanisms to comprehensively support digitalization in the construction industry. This process requires continuous policy optimization and active participation from all sectors of society to ensure sustainable and successful digital transformation.

6. Conclusions, Research Limitations, and Expectations

6.1. Conclusions

This study evaluates digitization policies in Wuhan’s construction sector within China’s national digital strategy framework. Utilizing the “3M” progressive policy evaluation system, this study examines policy strategies, optimization pathways, and implementation decisions from the macro, meso, and micro perspectives to enhance policy effectiveness and improve transformation efficiency.

At the macro level, a PESTEL-based frequency analysis of 33 policies revealed key policy attention patterns. Based on the identified deficiencies, indicator-setting recommendations were proposed to support the development of a meso-level policy evaluation system. Specifically, in the political dimension, the frequency of terms related to “Reform” and “Pilot Projects” was relatively low, leading to the establishment of “guidance” and “pilot projects” indicators to strengthen the guiding role of pilot policies. The economic dimension showed lower frequencies in “Market” and “Investment”, necessitating “market mechanism” and “financial support” indicators to reinforce market mechanisms and capital investment for transformation support. The social dimension revealed insufficient focus on “Livelihood” and “Integration”, prompting the adoption of “improving people’s livelihood” and “outcome transformation and integration” indicators to ensure social security and the integration of transformation outcomes with societal needs. The technological dimension showed limited focus on advanced technologies (e.g., BIM, AI), prompting added indicators for “technological innovation” and “practical applications”. Similarly, the environmental dimension’s sparse “Low Carbon” references necessitated a dedicated “low-carbon ecology” indicator to highlight its transformative role. The legal dimension had the lowest frequency and weakest centrality among all dimensions, prompting the introduction of “regulation”, “sufficient basis”, and “detailed measures” indicators to improve policy oversight, standardization, and the regulatory framework. The overall analysis highlighted that the legal dimension was the weakest in the policy system. Therefore, a macro-level policy strategy should focus on enhancing the legal and regulatory framework to strengthen legal support for digital transformation and ensure the structured and sustainable implementation of policies.

At the meso level, PMC index modeling identified local policy gaps, supporting a systematic evaluation framework. A systematic evaluation revealed that four policies were rated as “Perfect”, fifteen as “Excellent”, eleven as “Good”, and three as “Acceptable”. The average rating of “Excellent” indicates that most policies are reasonably structured and well formulated. However, a PMC surface analysis identified deficiencies in incentive measures, policy timeliness, policy recipients, and policy assessments. To address these issues, an optimization sequence was determined based on deviations in the composite average scores, which included incentive measures, policy timeliness, policy recipients, policy assessment, policy focus, policy nature, policy effectiveness, policy fields, and policy targets. These refinements enhanced policy precision and effectiveness. Moreover, PMC index scores from the meso level provided essential references for a micro-level correlation analysis.

Micro-level analysis employed Spearman’s coefficient to assess policy indicator–quality relationships. Correlation significance refined the meso-level optimization pathway, prioritizing incentive measures, policy timeliness, assessment, focus, and nature, ensuring scientific validity. Furthermore, the correlation between primary and secondary policy indicators was analyzed, identifying five key weaknesses—pilot projects, long-term planning, detailed measures, talent cultivation, and regulation. To address these deficiencies, targeted optimization measures have been proposed, including the introduction of pilot projects to foster innovation, clarifying long-term planning priorities, refining policy measures to ensure comprehensive implementation, increasing talent development and investment, and strengthening policy oversight and governance. These optimization strategies are designed to address critical weaknesses and enhance the systematic effectiveness of policy implementation.

This study developed a policy evaluation framework with targeted optimization strategies, offering clear guidance to improve construction industry digitization while ensuring policy effectiveness. The findings provide significant implications for Wuhan’s sector and valuable insights for other cities formulating similar digital policies.

6.2. Research Limitations and Expectations

The policy sample in this study covers only a brief period, limiting our ability to observe long-term effects. Future work should extend the temporal scope to capture the policies’ sustained and far-reaching impacts. Moreover, although 33 policies were analyzed, the sample remains modest in both size and diversity. Expanding the dataset would enhance the generalizability and robustness of the findings. Finally, our focus on Wuhan alone precludes any cross-regional comparisons; subsequent research could contrast multiple locales to examine regional coordination and policy adaptability.

From a methodological standpoint, PESTEL analysis at the macro level effectively maps the policy environment but lacks mechanisms to address environmental dynamism and uncertainty. Incorporating real-time or adaptive evaluation tools could strengthen responsiveness to changing conditions and refine assessments of policy efficacy. At the meso level, the PMC index model offers a quantitative framework for policy evaluation; however, the quantification of certain qualitative indicators remains challenging, as subjective judgments can affect the results. Accordingly, future research should refine the standards for converting qualitative criteria into quantitative measures to enhance the objectivity and reliability of policy assessments. At the micro level, Spearman’s rank correlation uncovers monotonic associations among policy variables but falls short in detecting complex non-monotonic patterns or inferring causal pathways. Combining SEM or other sophisticated multivariate techniques would help fill this gap, clarifying the causal relationships between PMC dimensions and policy outcomes and supplying more robust evidence to guide policy refinement and decision support.