LEED v4 Adoption Patterns and Regional Variations Across US-Based Projects

Abstract

Despite the widespread adoption of the Leadership in Energy and Environmental Design (LEED) rating system, there is limited empirical research examining how different sustainability categories are implemented in practice or how methodological patterns influence certification outcomes. This study contributes to this understanding by analysing LEED v4 Building Design + Construction certification patterns across 1252 newly constructed buildings in the United States to understand the methodological foundations and identify improvement opportunities for the LEED framework. Using credit achievement degree (CAD) analysis, regional variation assessment, and correlation analysis, we examined category adoption patterns across nine US climate regions, investigated relationships between LEED categories, and analysed certification level influences. The analysis reveals significant disparities in category adoption, with innovation (80.7%) and regional priority (66.6%) achieving high implementation rates while the category of material and resources (41.1%) consistently underperforms. Statistically significant regional variations exist across eight of nine categories (p < 0.05), with location and transportation showing the highest variability (CV = 20.1%). The category of energy and atmosphere demonstrates the strongest relationship with overall project performance (R² = 0.38), explaining 43% of total score variation and serving as the primary driver of higher certification levels. Most critically, inter-category correlations are weak (typically R² < 0.05), indicating that projects treat sustainability domains as separate challenges rather than integrated systems. Positive skewness across all certification levels (z-scores > 1.96) provides statistical evidence of strategic “point-chasing” behaviour, where teams target minimum thresholds rather than maximising comprehensive sustainability performance. These findings reveal fundamental methodological patterns that may limit LEED’s effectiveness in promoting holistic sustainability approaches. The compartmentalised implementation patterns and threshold-focused strategies suggest opportunities for structural refinements, including enhanced integration incentives, region-sensitive benchmarking, and certification frameworks that reward comprehensive rather than minimal compliance. This research contributes empirical evidence for evidence-based improvements to green building certification methodology and provides insights for more effective sustainability assessment tools.

1. Introduction

Over the past three decades, the awareness of the substantial environmental and energy impacts of buildings has steadily increased. The negative environmental consequences of conventional building practices have prompted the building sector to pursue efficiency in resource and energy use. Meanwhile, there has been increased emphasis on enhancing the social sustainability of buildings, focusing particularly on occupant comfort and well-being. Climate change has further intensified these challenges, requiring the development of building and infrastructure systems to enhance resilience to extreme weather events and changing environmental conditions [1]. These developments have catalysed a shift towards green buildings (GBs), where sustainable practices are integrated into the design, construction, and operation of buildings [2]. This approach aims to reduce environmental footprints while improving social and health outcomes. This transition requires countries to establish and deploy control mechanisms that govern and encourage sustainable practices within the built environment [3,4]. Consequently, multiple GB rating systems have emerged to improve energy efficiency, health outcomes, and comprehensive sustainability in buildings.

The Leadership in Energy and Environmental Design (LEED) rating system represents one of the most widely implemented GB frameworks. The U.S. Green Building Council launched LEED version 1.0 in 1998 [5]. By November 2024, LEED certification had been adopted across 186 countries globally [6]. Both researchers and practitioners have recognised the role of LEED in expanding beyond conventional energy efficiency metrics. For instance, the system incorporates social considerations including public health and community cohesion. This comprehensive approach has led to the creation of various implementation frameworks, that is, a multitude of LEED versions with a focus on different building types. This broader perspective of the LEED rating system has also increased its international reach [7,8].

Despite LEED’s widespread adoption and the extensive research on its outcomes, limited empirical research exists on its implementation patterns, particularly regarding LEED credit and category adoption. LEED categories are distinct sustainability domains that organise different aspects of GB performance, including energy and atmosphere (EA), indoor environmental quality (IEQ), location and transportation (LT), material and resources (MR), water efficiency (WE), sustainable sites (SS), innovation (IN), regional priority (RP), and integrative process (IP). Buildings are evaluated across all nine categories, with each category offering specific credits that projects can pursue to earn points toward their overall certification level. Whilst most studies focus on LEED’s effectiveness in promoting energy efficiency or compare it with other rating systems, limited attention has been given to examining how different LEED categories correlate and contribute to certification outcomes in practice. Whilst some studies have examined previous LEED versions (v2.0–v3), limited research has analysed large numbers of US projects certified under LEED v4.

Analysing the scores achieved by LEED-certified projects provides valuable insights into the practical application and effectiveness of different LEED categories and credits, revealing both strengths and areas for methodological improvement. LEED v4, with substantial adoption since its introduction, now provides sufficient data for the comprehensive empirical analysis of certification patterns and category relationships.

The inquiry into LEED adoption patterns and credit interrelationships offers practical benefits as well. First, project stakeholders often approach LEED certification with limited budgets and resources, requiring strategic decisions about which credits to pursue. Little is known about how these strategic choices affect the balance of credits across different categories. Second, regional and climatic differences across the US may significantly influence LEED category adoption patterns, but the systematic analysis of these variations remains limited. Third, understanding the relationships between different sustainability categories is crucial for promoting integrated design approaches [9], yet few studies have examined how categories correlate with each other and with overall project performance.

Additionally, the emergence of strategic certification behaviours—where project teams may prioritise meeting minimum thresholds rather than maximising comprehensive sustainability performance—requires empirical investigation to understand its implications for LEED’s effectiveness as a sustainability assessment tool. Such analysis can reveal not only which credits are most readily adopted by project teams, but also the patterns that may encourage compartmentalised rather than holistic sustainability approaches. This study addresses these research needs by analysing scorecards from 1252 LEED v4 Building Design + Construction (BD + C) projects across nine US climate regions, examining category adoption patterns, regional variations, and inter-category relationships to understand how the LEED methodology functions in practice and identify opportunities for system enhancement.

The aim of this study is to analyse LEED v4 category adoption patterns through the empirical examination of certified projects. The objectives of this study are as follows: 1. analyse category adoption rates and patterns across LEED v4-certified projects in the United States; 2. examine regional variations in category adoption across different US climate regions; 3. investigate correlations between LEED categories and their relationship to overall certification outcomes; 4. identify strategic certification behaviours and their implications for sustainability assessment; 5. provide evidence-based recommendations for enhancing LEED’s methodological effectiveness.

The paper is structured as follows: First, we review the relevant literature on the LEED rating system, its evolution, and previous studies on credit achievement. Next, we describe our empirical data collection from 1252 LEED v4-certified buildings across nine US climate regions. We then analyse LEED category adoption patterns, exploring regional variations and examining how different categories contribute to certification levels. Following this, we discuss the implications of our findings, particularly regarding the skewness of project scores and the “point-chasing mentality.” Finally, we offer recommendations for improving the LEED rating system and conclude with limitations and directions for future research.

2. Literature Review

2.1. Overview of the LEED Rating System

Version 1.0 of LEED was first launched by the U.S. Green Building Council in 1998 [5], with its first full version (LEED-NC v2.0) released in 2000. It has since undergone continuous evolution in scope and the refinement of its criteria [10]. The certification system evolved through multiple versions, beginning with LEED v1 and advancing through subsequent editions until reaching LEED v4, launched in 2013 [11]. The latest version, LEED v5, is slated for full implementation following the United States Green Building Council (USGBC) ballot process in 2025 [12]. Although LEED v5 represents the latest certification update, its changes are incremental rather than transformative. Limited adoption and project data restrict its potential for in-depth analysis. Therefore, this study focuses on the more established LEED v4, leveraging its widespread application and mature dataset.

Although LEED v4 was released in 2013 [11], it remains in use today. Currently, a transition is being made from LEED v4 to LEED v5. This largely mirrors the same credit categories as those in LEED v4 (see

Table 1. Comparison of LEED points across v3, v4, and v5.

Category	LEED v3 (2009) Scores and % Contribution	LEED v4 Scores and % Contribution	Number of Credits in Each Category (LEED v4)	LEED v5 Scores and % Contribution
Innovation (IN)	6 (5.5%)	6 (5.5%)	2	NA
Regional Priority (RP)	4 (3.6%)	4 (3.6%)	4	NA
Project priorities and innovation	NA	NA		10 (9%)
Integrative Process (IP)	NA	1 (0.9%)	1	1 (0.9%)
Energy and Atmosphere (EA)	35 (31.8%)	33 (30%)	7	33 (30%)
Water Efficiency (WE)	10 (9.1%)	11 (10%)	4	9 (8.2%)
Sustainable Sites (SS)	26 (23.6%)	10 (9.1%)	6	11 (10%)
Indoor Environmental Quality (IEQ)	15 (13.6%)	16 (14.5%)	9	13 (11.8%)
Location and Transportation (LT)	NA	16 (14.5%)	8	15 (13.6%)
Material and Resources (MR)	14 (12.7%)	13 (11.8%)	5	18 (16.4%)
Number * of US-Based Projects certified with BD + C New Construction (NC) Rating	8302	1648		0

* These numbers are based on the search of LEED project database “www.usgbc.org/projects (accessed on 15 June 2025)”.

), but includes seven new prerequisites and eight new credits [13]. LEED v5’s updated requirements aim to align the system with current market conditions, address the industry’s environmental requirements, and support ongoing decarbonization efforts. LEED v5 has placed increased emphasis on climate change, resilience, and equity [12]. Currently, few projects have achieved O + M: Existing Buildings (Operation and Maintenance) certification with LEED v5, and only a few have been certified under the BD + C category. As LEED v4 is being phased out in 2025 and will be replaced by v5, data collected for LEED v4 until the end of 2024 represents the sunset stage of v4, and therefore defines the scope of this study.

Under the LEED v4 rating system, buildings undergo evaluation according to their performance across nine key categories. Within each category there are specific credits that represent points buildings can acquire by satisfying sustainability criteria. The cumulative points earned from all categories determine the building’s overall rating, whereby higher total point values result in elevated certification levels.

LEED v4, BD + C awards 33 points to the EA category. These 33 points are divided among the four prerequisites and seven credits that comprise the EA category. For the remaining categories, as shown in Table 1, the point distribution allocates 16 points each to IEQ and LT, 13 points to MR, 11 points to water efficiency, 10 points to SS, 6 points to IN, 4 points to RP, and 1 point to IP [14].

For LEED v4, BD + C, the overall performance or the GB rating of a particular project is based on the total score it achieves from 110 available points [14]. A project’s certification level is based on its total score, which must fall within specific point ranges. Under LEED v4, certification levels are defined as follows: certified (40–49 points), silver (50–59 points), gold (60–79 points), and platinum (80 points or more) (Council, 2019). Projects that meet the minimum point requirement receive the certified rating, whilst those with the highest scores earn the prestigious platinum rating.

2.2. Prior Research on LEED Rating System

Academic interest in LEED emerged as early as the 1990s, with initial studies focusing on its market application [15]. Over time, the scope of LEED-related research has significantly expanded, aiming to refine the rating system and boost its adoption.

LEED research has evolved through distinct phases that reflect both the maturation of the certification system and increasing analytical sophistication. Early studies (2009–2012) focused on understanding basic adoption patterns and implementation barriers, exemplified by foundational work examining practitioner perceptions and regional variations [16,17,18]. The second phase (2016–2019) was characterised by large-scale comprehensive analyses and methodological innovations, including the introduction of statistical methods and association rule mining to understand credit interrelationships [9,19,20]. The current phase (2020–2024) has focused on LEED v4 analysis, though studies remain limited in scope compared to earlier comprehensive research on previous LEED versions.

Among the studies conducted on LEED, one of the key areas of inquiry concerns LEED credit adoption, which is also referred to as credit achievement degree (CAD) (see Table 2). This research stream has highlighted the varying difficulty levels associated with attaining particular LEED credits and categories. Research in this area provides comparative analysis of credits and categories, determining which ones are more commonly pursued by LEED-certified projects versus those that are of occasional interest to projects. Such analysis yields important insights into achievement patterns and provides practical recommendations for the adoption of underutilised credits [3,9,16,18,19,20,21,22,23].

Despite extensive research on the outcomes and external impacts of the LEED rating system, limited understanding exists regarding the interrelationships within LEED credits and the trends of credit adoption, particularly for LEED v4. Whilst some studies have addressed these areas for LEED v3 (BD + C) [9,20,22,24] and LEED NC v2.0, v2.1, v2.2 (BD + C) [16,18,19], comprehensive studies focusing on large numbers of LEED v4-certified US projects are currently missing. Research in this regard can have significant theoretical and practical implications, as LEED v4 introduced substantial changes to credit structure and point allocation compared to previous versions.

Table 2 indicates some patterns in the existing literature on LEED credit and category adoption. There is significant variation in geographic scope and sample sizes, ranging from regional studies with smaller samples (42 projects in Canada) to comprehensive worldwide analyses (5327 projects). The majority of studies have focused on earlier LEED versions (v2.0–v3), with limited attention paid to LEED v4.0 despite its widespread adoption since 2013. Among the few v4.0 studies, sample sizes remain relatively small (94–971 projects), indicating that LEED v4.0 research would benefit from larger-scale analysis. Some studies have also focused specifically on examining certain categories (particularly MR and EA). Most previous studies provide general credit achievement patterns without deep analysis of inter-category relationships or regional variations. Beyond credit adoption rates, prior research has explored associations with incremental costs, complexity levels, project ownership types, and energy performance, but few studies have systematically examined how different categories correlate with overall certification outcomes. This literature landscape shows that whilst credit achievement degree (CAD) analysis is well-established as a methodology, there remains a significant opportunity for comprehensive analysis of LEED v4.0 projects that combines CAD analysis with regional variation assessment and correlation analysis to understand the methodological foundations of the rating system.

The theoretical framework underlying LEED research focuses on credit achievement degree (CAD) as the primary quantitative methodology, whilst acknowledging that implementation is influenced by multiple factors including cost, complexity, regional conditions, and project characteristics. However, systematic correlation studies examining relationships between different LEED categories and their influence on overall certification outcomes are missing from the literature.

Table 2. Overview of previous studies related to LEED credit and category adoption.

Source	Region	Number of Certified LEED Projects Considered	LEED Certification and Version Considered	Study Objectives (Besides LEED Credit Adoption)	Specifically Focused Credit Categories
[19]	Worldwide	5327	LEED v2.2
[20]	Worldwide	3416	LEED BD + C: NC v3 (2009)
[3]	MENA region—Arab countries	452	LEED (NC, commercial interior, core and shell, existing buildings, neighbourhood development) (v1.0, v2.0, v2.1, v2.2, v2007, v2008, v2009)
[16]	USA	102	LEED BD + C: NC v2.0, v2.1, v2.2	Associations of credits with incremental cost and level of complexity
[22]	USA	95 *	LEED BD + C: NC v3 (2009)		MR; EA
[25]	USA	971	LEED BD + C: NC v4.0	Circular economy integration in LEED projects	MR
[9]	USA	1000	LEED BD + C: NC v3 (2009)	Associations among credits
[26]	USA	60	LEED BD + C: NC v4.0
[21]	Worldwide	222	LEED BD + C: NC v4.0	Associations among credits
[27]	USA	94	LEED O + M: EB v4.0	Associations of credits with project size
[24]	USA	1500	LEED BD + C: NC V3 (2009)	Associations of credits with project ownership type	EA
[18]	Arizona, USA	48	LEED BD + C: NC v2.0, v2.1, and v2.2	Associations of credits with energy use
[17]	Canada	42	LEED Canada BD + C: NC v1.0; LEED BD + C NC v2.2
[23]	Turkey	122	LEED BD + C: NC & C + S V3 (2009)

Note: * = Only fire station projects.

Table 2. Overview of previous studies related to LEED credit and category adoption.

Source	Region	Number of Certified LEED Projects Considered	LEED Certification and Version Considered	Study Objectives (Besides LEED Credit Adoption)	Specifically Focused Credit Categories
[19]	Worldwide	5327	LEED v2.2
[20]	Worldwide	3416	LEED BD + C: NC v3 (2009)
[3]	MENA region—Arab countries	452	LEED (NC, commercial interior, core and shell, existing buildings, neighbourhood development) (v1.0, v2.0, v2.1, v2.2, v2007, v2008, v2009)
[16]	USA	102	LEED BD + C: NC v2.0, v2.1, v2.2	Associations of credits with incremental cost and level of complexity
[22]	USA	95 *	LEED BD + C: NC v3 (2009)		MR; EA
[25]	USA	971	LEED BD + C: NC v4.0	Circular economy integration in LEED projects	MR
[9]	USA	1000	LEED BD + C: NC v3 (2009)	Associations among credits
[26]	USA	60	LEED BD + C: NC v4.0
[21]	Worldwide	222	LEED BD + C: NC v4.0	Associations among credits
[27]	USA	94	LEED O + M: EB v4.0	Associations of credits with project size
[24]	USA	1500	LEED BD + C: NC V3 (2009)	Associations of credits with project ownership type	EA
[18]	Arizona, USA	48	LEED BD + C: NC v2.0, v2.1, and v2.2	Associations of credits with energy use
[17]	Canada	42	LEED Canada BD + C: NC v1.0; LEED BD + C NC v2.2
[23]	Turkey	122	LEED BD + C: NC & C + S V3 (2009)

Note: * = Only fire station projects.

Correlation analysis provides valuable insights into how various sustainability dimensions interact and which categories most strongly influence certification outcomes. Examining correlations between different sustainability categories can reveal whether LEED projects address sustainability criteria holistically or through a compartmental approach. Strong correlations between categories would suggest integrated approaches, while weak correlations might indicate that project teams treat different LEED credits and categories as separate challenges rather than as interconnected dimensions of a unified system. This insight is particularly relevant to critiques of GB certification that question whether point-based systems encourage fragmented rather than holistic sustainability approaches [28]. Some researchers have begun exploring these relationships. Both Ma and Cheng [9] and Pham, Kim [21] employed association rule mining as their primary methodology to investigate LEED credit relationships, using measures of support, confidence, and lift to identify co-occurrence patterns and directional relationships between credits in the form A → B. Ma and Cheng [9] focused exclusively on association rule analysis with 1000 LEED-NC v3 projects, and discovered 442 strong association rules. Pham, Kim [21] used association rule mining as part of a broader statistical framework that included multiple methods (mean value ranking, Kolmogorov–Smirnov test, Mann–Whitney U-test, and Cliff’s delta) applied to 222 LEED-NC v4 projects, ultimately identifying 8 strong association rules. Apart from the above-mentioned studies, using association rule mining to determine credit interrelationships, correlation studies on the subject matter are missing.

The relationship between category achievement and overall certification levels represents an important but understudied aspect of LEED research. Understanding which categories most strongly correlate with overall project scores provides insights into how the LEED weighting system functions in practice, beyond theoretical point allocations. For instance, while EA constitutes 30% of available points in LEED v4, its actual influence on certification outcomes depends on how consistently projects achieve points in this category compared to others and how strongly these achievements correlate with overall project performance.

Therefore, this study extends the current understanding by (1) providing comprehensive analysis of a large sample of LEED v4 projects to establish robust baseline data to complement existing literature, (2) conducting systematic correlation analysis to understand inter-category relationships and their influence on certification outcomes, and (3) assessing regional variations to understand how LEED v4 performs across different geographic contexts. This approach builds upon the established CAD methodology while extending it through correlation analysis to provide deeper insights into how the LEED rating system functions in practice and to identify potential areas for refinement.

3. Methodology

Building upon the literature review, we employ a comprehensive four-pronged analytical approach, combining quantitative descriptive analysis, statistical testing, and correlation analysis to examine LEED v4 certification patterns across multiple dimensions (see Figure 1). The methodology is designed to address the complex, multi-faceted nature of GB certification systems and their implementation patterns across diverse geographical contexts within the US region. This study uses a quantitative empirical approach based on historical data from LEED v4-certified projects, following methodological precedents established in previous studies [17,19,20]. We analyse credit achievement patterns to identify trends and variations in sustainability implementation.

3.1. Empirical Data Collection

Although we acknowledge that LEED v5 is the latest advancement in the certification system, our research focuses on LEED v4 for several reasons. As of June 2025, the public database of LEED projects contains no certified projects with LEED v5 BD + C NC. In contrast, LEED v4, which remains in use, has been applied in 1600+ US-based projects. Therefore, data access to a large number of projects is a key reason for the emphasis of this study on LEED v4. Another key reason to focus on LEED v4 is that we are close to its sunset stage as LEED v4 registrations for BD + C are set to close at the end of Q1 2026 [29]. Analysing LEED v4 credit achievement patterns near its sunset provides a complete understanding of how the system performed throughout its entire lifecycle, capturing both early challenges and mature implementation practices. Data collection near the sunset stage ensures sufficient sample size while maintaining the representation of mature certification practices. The resulting data reveals issues that should be addressed in LEED v5 design and helps develop effective transition strategies. The final performance patterns also establish definitive benchmarks for measuring LEED v5’s effectiveness and improvements over its predecessor (LEED v4).

Among the 21 rating systems available in LEED v4, the Building Design and Construction (BD + C): NC system applies to both new building constructions and major renovations [14]. For this study, we obtained detailed LEED credit data for projects certified with LEED v4.0 BD + C: NC. The BD + C category includes a variety of project types, including institutional buildings (schools, libraries, community centres), commercial buildings (offices, retail, mixed-use), high-rise residential buildings (nine storeys or more), light industrial buildings, and government facilities (police stations, civic centres). Focusing exclusively on BD + C: NC projects enables more targeted analysis and the development of actionable insights for stakeholders involved in new construction projects.

Using the publicly accessible scorecards available on the online USGBC LEED project database, we obtained detailed scores for the nine LEED categories from projects meeting three criteria: (1) based in the US, (2) certified with LEED v4.0, and (3) certified within the rating system of BD + C: NC. Among the total of 3098 projects certified worldwide by LEED v4.0 BD + C: NC, 1648 (53%) projects are based in the US. The initial dataset of 1648 US-based LEED v4 BD + C: NC-certified projects represents buildings that achieved certification between March 2014 and June 2025, spanning the full operational period of LEED v4 from its earliest implementations to near its sunset phase.

LEED credit and category adoption varies significantly from region to region owing to contextual factors [30]. For instance, regarding adoption rates of LEED credits, significant differences are noticeable across the USA [22], Turkey [23], and the MENA region-based projects [3], to provide a few examples. Accordingly, the analysis in this study is limited to US-based projects only. We collected scorecard data in June 2025. As LEED v4 is being phased out in early 2026 and will be replaced by v5, data collected for LEED v4 until mid-2025 approximately represents the sunset stage of LEED v4.

The use of LEED scorecards is an empirically reliable data collection technique practiced by previous studies investigating LEED credit adoption and categories across various regions [3,16,19]. Scorecards provide the standardised and consistent documentation of the credits achieved by a project, allowing for objective comparison and analysis across multiple projects. We initially shortlisted 1648 projects. Due to lacking scorecard information for some projects, we finally selected 1252 projects for detailed analysis. The excluded projects were randomly distributed across different US regions and certification years. These exclusions were due to missing category-level score breakdowns for some projects.

Among the final 1252 projects selected, the majority are in California (n = 203), Florida (n = 79), Texas (n = 75), Virginia (n = 68), Washington (n = 60), and Colorado (n = 57) (see Figure 2). We divided the final set of 1252 projects among nine US climate regions (see Figure 3) as defined by the National Centers for Environmental Information (NCEI), US, and proposed by Karl and Koss [31]. Among the selected LEED projects, 285 are certified, 492 are silver, 403 are gold, and 68 have platinum certification.

3.2. Analysis

Our methodological framework consists of four complementary analyses designed to work synergistically: CAD analysis identifies overall adoption patterns, regional analysis reveals geographical variations, correlation analysis uncovers relationships between credits and categories, and certification level analysis determines which factors drive higher performances. We conducted all analyses using SPSS version 29.0, with statistical significance set at α = 0.05.

Credit Achievement Degree (CAD) Analysis: Credit achievement degree (CAD) serves as the primary quantitative metric for assessing LEED category adoption patterns. The CAD calculation follows a standardised approach established in previous LEED research [19,20], with specific adaptations for LEED v4 category structure.

The CAD calculation methodology involves several systematic steps.

Step 1: Data Validation and Quality Control

Prior to CAD calculation, all scorecard data underwent rigorous validation procedures. We verified each project’s scorecard for completeness across all nine LEED categories. We excluded projects with missing category-level breakdowns from the analysis, resulting in the final dataset of 1252 projects from an initial pool of 1648 US-based LEED v4-certified projects.

Step 2: Category-Level CAD Calculation

For each LEED category, CAD is calculated using the formula provided in Equation (1).

CAD = (CO/TC) × 100%

(1)

In Equation (1), CO equals the credit/category score obtained and TC equals the total available points of the credit/category. For instance, in the project dataset available for this study, for the IN category, the average score (CO) is 4.8, the full score (TC) is 6, and the resulting CAD value is 80. This metric enables standardised comparisons across categories with different point allocations and has been extensively used in previous LEED research [19,20].

Step 3: Regional Variation Analysis

We determined variations in the credits and categories across nine US climate regions defined by the National Centers for Environmental Information. The use of these nine regions represents a methodological trade-off between spatial resolution and statistical power, ensuring adequate sample sizes for robust statistical testing while acknowledging that finer-scale variations may exist within regions.

Before determining the regional variations in credits and categories, we tested the assumption (H₀) regarding the normal distribution of the data using the Shapiro–Wilk test. Given the large sample size (n = 1252), the Shapiro–Wilk test provides robust normality assessment. The null hypothesis states that there is no significant difference between the data distribution and the normal distribution. With a p-value less than 0.05, the null hypothesis could be rejected, adopting the alternative hypothesis (H_a) implying that the data are not normally distributed. With a p value less than 0.05 for all LEED categories (see

Table 3. Test results related to Shapiro–Wilk test for normality.

	Shapiro–Wilk Test
LEED Categories	df	Sig.
EA	1252	0.000
IE	1252	0.000
IN	1252	0.000
IP	1252	0.000
LT	1252	0.000
MR	1252	0.000
RP	1252	0.000
SS	1252	0.000
WE	1252	0.000

), we rejected the null hypothesis and adopted the alternative hypothesis implying that the data is not normally distributed. Based on these findings, we deemed a non-parametric test called the Kruskal–Wallis one-way analysis of variance suitable for checking the variations in LEED credits and categories across US climate regions. This test, as also previously used for similar inquiries [19], the Kruskal–Wallis test is useful when comparing the data (credit and category-related) from two or more independent groups (climate regions) to determine whether there are statistically significant differences in their central tendencies (medians). To assess variability across regions, we calculate the coefficient of variation (CV = standard deviation/mean), which provides a standardised measure of dispersion that enables comparison across categories with different scales. We consider categories with CV values below 10% to have uniform adoption, while those above 20% indicate substantial regional variation.

Inter-Category and Category-Performance Correlation Analysis: We conducted Spearman correlation analysis to discover the precise strength and statistical significance of relationships between LEED credits, categories, and overall project scores. We chose Spearman correlation over Pearson correlation due to non-normal data distributions and the presence of ordinal relationships in LEED scoring. This analysis reveals systematic patterns in credit achievement, indicating which categories consistently perform together versus those that are systematically underutilised. The analysis includes both inter-category correlations (R²) to identify potential synergies or compartmentalization in sustainability approaches, and category-to-total score correlations to understand which categories most strongly influence certification outcomes. We squared correlation coefficients (r) to obtain R² values, representing the proportion of variance in one variable explained by another, and interpreted correlation strength as weak (R² < 0.10), moderate (0.10 ≤ R² < 0.30), or strong (R² ≥ 0.30), following established conventions in social science research.

Certification Level Analysis: We investigate the relationship between category scores and certification levels (certified, silver, gold, platinum) to identify which categories most strongly contribute to achieving higher certification levels. This includes examining the mean value changes between certification levels and analysing score distributions through skewness measures to identify strategic certification behaviours. We calculated skewness values with corresponding z-scores to test for statistical significance of distribution asymmetry.

We followed a clustering analysis protocol whereby we conducted K-means clustering analysis to identify distinct strategic approaches among certified projects. We determined the optimal number of clusters using the elbow method and silhouette analysis. We tested values of k ranging from 2 to 8, with k = 3 selected based on maximising both interpretability and statistical validity. We implemented K-means clustering using Lloyd’s algorithm due to its robust convergence properties and computational efficiency for large datasets. Cross-tabulation with certification levels provided interpretive context for identified clusters, revealing strategic patterns in LEED pursuit approaches.

Additional data preparation procedures included (1) the identification and documentation of zero scores across categories, (2) the classification of projects by climate region using geographic information from the National Centers for Environmental Information, and (3) the validation of certification level assignments. All statistical analyses account for the hierarchical nature of the data (projects nested within climate regions) and potential heteroscedasticity in variance across certification levels.

4. Results

Our analysis of 1252 LEED v4-certified BD + C projects across nine US climate regions reveals distinct patterns in category adoption and regional variations.

4.1. Category-Level Analysis

4.1.1. Category-Level Adoption Patterns

The CAD analysis reveals a clear hierarchy in category adoption rates. IN (80%) leads with the highest adoption rate, followed by RP (66%) and IP (57%). Since IP contains only one binary credit, the 57% CAD represents the percentage of projects that successfully achieved this credit across the dataset. EA (52%), WE (50%), and SS (48%) show moderate adoption levels. The categories with the lowest adoption rates are IEQ (45%), LT (45%), and MR (41%), with MR consistently trailing behind other categories (see

Table 4. Adoption of LEED categories across US climate regions.

		EA	IEQ	IN	IP	LT	MR	RP	SS	WE
	Overall (all regions)	52	45	80	57	45	41	66	47	50
US Climatic Region	Northeast	50	47	80	53	49	44	69	47	52
	Northern Rockies and Plains	59	50	81	60	27	42	86	48	51
	Northwest	49	44	83	52	56	42	66	54	48
	Ohio Valley	47	46	80	65	45	43	64	42	49
	South	50	46	79	56	36	44	58	45	50
	Southeast	47	43	80	52	40	39	66	45	53
	Southwest	60	43	78	64	37	37	62	49	45
	Upper Midwest	52	43	74	50	38	37	70	49	50
	West	62	45	84	62	50	39	69	49	50
	Kruskal–Wallis Asymp. Sig.	0.000	0.000	0.004	0.000	0.000	0.157	0.000	0.000	0.001
	CV	10.4	4.9	3.4	9.4	20.1	6.3	10.9	6.6	4.7

).

4.1.2. Regional Variations in Category Adoption

The application of Kruskal–Wallis one-way analysis of variance to category scores of LEED projects developed across the nine US climate regions shows that except for the material and resources category (p = 0.157), all other categories and credits exhibit statistically significant differences in the median score values. Except for one category, all other LEED categories show statistically significant variations in adoption rates across the nine US climate regions. This implies that regional factors substantially influence LEED certification patterns. These findings are largely consistent with the study by Wu, Mao [19], which found that, for LEED v2.2 projects, there are statistically significant variations in adoption trends of LEED credits and categories across the four global regions (Northern America, Eastern Asia, Western Asia, and Europe). This also applied within the 10 regions in the US.

To determine the credits and categories with greater spread across US climate regions, we used the coefficient of variance (CV). This standardised metric is calculated by dividing the standard deviation by the mean. CV can be considered a measure of the level of uniform achievement, meaning that, with low CV values, developers obtain similar scores in the LEED credits and categories [20]. Categories with the least spread are ‘IN’ (3.4), WE (4.7), and IEQ (4.9). This implies that for these LEED categories, developers across the nine US regions obtain similar scores. Categories with the highest spread are LT (20), ‘RP’ (10.9), EA (10.4), and ‘IP’ (9.4). This implies that for these categories, projects achieve significantly different scores across the US regions. We explore the reasons behind varying CV values of LEED categories in detail in Section 5.2.

4.1.3. Correlations Among LEED Categories

The correlation analysis of 1252 US-based LEED projects reveals predominantly weak relationships between different LEED categories, suggesting a high degree of independence in how these aspects are implemented. The R-squared values across most category pairs rarely exceed 0.10, indicating that achievement in one sustainability area explains minimal variation in another (see

Table A1. LEED category correlation analysis.

	R-Square Values
LEED Categories	Northeast	Northern Rockies and Plains	Northwest	Ohio Valley	South	Southeast	Southwest	Upper Midwest	West	All Climates
IEQ-EA	0.007	0.001	0.022	0.000	0.000	0.001	0.010	0.001	0.027	0.000
IN-EA	0.001	0.032	0.016	0.001	0.009	0.002	0.038	0.000	0.030	0.000
IP-EA	0.022	0.120	0.044	0.001	0.058	0.000	0.009	0.002	0.033	0.000
LT-EA	0.021	0.122	0.116	0.047	0.014	0.085	0.005	0.068	0.004	0.001
MR-EA	0.003	0.017	0.003	0.022	0.028	0.005	0.000	0.015	0.015	0.000
RP-EA	0.081	0.056	0.002	0.043	0.109	0.019	0.209	0.050	0.019	0.002
SS-EA	0.013	0.001	0.001	0.031	0.013	0.003	0.002	0.074	0.013	0.000
WE-EA	0.003	0.010	0.044	0.003	0.013	0.006	0.003	0.006	0.061	0.000
IN-IEQ	0.024	0.023	0.069	0.065	0.180	0.082	0.083	0.007	0.098	0.004
IP-IEQ	0.027	0.026	0.040	0.013	0.121	0.008	0.058	0.000	0.085	0.001
LT-IEQ	0.026	0.002	0.082	0.002	0.002	0.004	0.027	0.001	0.001	0.000
MR-IEQ	0.120	0.188	0.005	0.006	0.151	0.061	0.081	0.055	0.061	0.004
RP-IEQ	0.018	0.013	0.016	0.000	0.023	0.000	0.056	0.039	0.000	0.000
SS-IEQ	0.012	0.003	0.088	0.063	0.105	0.017	0.151	0.036	0.021	0.001
WE-IEQ	0.018	0.135	0.001	0.014	0.046	0.029	0.072	0.007	0.011	0.000
IP-IN	0.033	0.033	0.033	0.039	0.084	0.020	0.070	0.009	0.059	0.002
LT-IN	0.058	0.004	0.000	0.000	0.044	0.011	0.009	0.041	0.081	0.001
MR-IN	0.083	0.259	0.072	0.094	0.076	0.067	0.019	0.045	0.088	0.005
RP-IN	0.002	0.118	0.003	0.001	0.005	0.002	0.000	0.047	0.021	0.000
SS-IN	0.018	0.013	0.015	0.033	0.036	0.020	0.000	0.031	0.020	0.000
WE-IN	0.026	0.332	0.009	0.001	0.070	0.008	0.085	0.001	0.038	0.000
LT-IP	0.014	0.091	0.018	0.011	0.003	0.015	0.004	0.005	0.021	0.000
MR-IP	0.025	0.069	0.086	0.002	0.008	0.001	0.021	0.027	0.084	0.000
RP-IP	0.028	0.030	0.000	0.014	0.063	0.006	0.009	0.028	0.001	0.000
SS-IP	0.013	0.044	0.026	0.019	0.046	0.019	0.055	0.013	0.021	0.000
WE-IP	0.000	0.077	0.036	0.032	0.060	0.009	0.013	0.020	0.047	0.000
MR-LT	0.005	0.076	0.063	0.005	0.023	0.034	0.000	0.059	0.061	0.001
RP-LT	0.000	0.013	0.000	0.031	0.002	0.000	0.000	0.048	0.017	0.000
SS-LT	0.001	0.033	0.002	0.000	0.002	0.022	0.000	0.008	0.000	0.000
WE-LT	0.001	0.002	0.002	0.000	0.016	0.025	0.008	0.006	0.018	0.000
RP-MR	0.029	0.042	0.026	0.097	0.028	0.010	0.002	0.142	0.043	0.000
SS-MR	0.003	0.028	0.000	0.003	0.025	0.000	0.069	0.006	0.000	0.000
WE-MR	0.018	0.303	0.002	0.014	0.006	0.002	0.008	0.022	0.094	0.000
SS-RP	0.117	0.011	0.189	0.077	0.004	0.061	0.031	0.068	0.029	0.004
WE-RP	0.000	0.082	0.005	0.007	0.041	0.034	0.083	0.000	0.118	0.001
WE-SS	0.000	0.054	0.000	0.010	0.006	0.032	0.035	0.009	0.007	0.000

in Appendix A). Notable exceptions are shown in

Table 5. Good correlations between LEED categories.

Climate Region	LEED Category Correlations	R-Square Values
Southwest	RP-EA	0.209
Northern Rockies and Plains	MR-IN	0.259
Northern Rockies and Plains	WE-IN	0.332
Northern Rockies and Plains	WE-MR	0.303

, which indicates a moderate correlation between WE and IN in the Northern Rockies and Plains region (R² = 0.33), between WE and MR (R² = 0.30) and MR and IN in the same region (R² = 0.26), and between RP and EA in the Southwest (R² = 0.21). These findings suggest that in certain regions, building practices may foster stronger connections between specific categories.

The prevalence of weak correlations throughout the data is particularly telling in terms of how LEED projects are typically approached. The consistently low R-squared values (most below 0.05) between critical sustainability areas—such as EA with WE (R² = 0.02), LT with SS (R² = 0.00), and IEQ with EA (R² = 0.01)—indicate that project teams tend to address these domains as separate challenges rather than as interconnected systems. This compartmentalised approach suggests that integrated design thinking, despite being a core principle of GB, may not be widely implemented in practice. The weak correlations may also reflect the structure of the LEED rating system itself, which allows projects to prioritise certain categories to achieve certification while potentially neglecting others.

The correlation analysis between LEED category scores and overall project scores reveals varying degrees of influence across the nine sustainability categories (see

Table 6. Correlation of LEED category scores with overall project scores.

LEED Categories	R-Square Values	Total Achievable Scores (Percentage of Achievable Scores)
EA	0.38	33 (30%)
IN	0.18	6 (5.5%)
MR	0.15	13 (11.8%)
IEQ	0.14	16 (14.5%)
SS	0.13	10 (9.1%)
RP	0.16	4 (3.6%)
WE	0.09	11 (10%)
LT	0.12	16 (14.5%)
IP	0.09	1 (0.9%)

). EA demonstrates the strongest relationship with overall project performance, with an R-squared value of 0.43, indicating that this category explains approximately 43% of the variation in total project scores. However, this correlation may be partially inflated by EA’s disproportionate point weight (30% of available points), making it difficult to separate mathematical correlation from intrinsic categorical importance.

This substantial correlation highlights the central importance of energy performance in determining overall LEED certification levels. The remaining categories show more moderate correlations, with IN explaining approximately 18% (R² = 0.18), RP explaining 16% (R² = 0.16), MR explaining 15% (R² = 0.15), and IEQ explaining 14% (R² = 0.14) of the variation in total scores. SS follows with R-squared values of 0.13, LT with 0.12, while WE and IP each explain about 9% of score variation (R² = 0.09). These findings indicate that while EA serves as the primary driver of overall LEED performance, successful projects typically require attention across multiple categories.

4.2. Credit-Level Analysis

4.2.1. Credit-Level Adoption Patterns

In addition to the holistic view of LEED project practices, enabled by category-level analysis, a detailed view of practices, enabled by credit-level analysis, is also important. Credits with the highest adoption rates tend to cluster in IN, IEQ, SS, and WE categories (see Figure 4). LEED Accredited Professional credit (99%) and construction indoor air quality management plan (95%) have near-universal adoption. This is because both these credits are easy to achieve. Conversely, the least adopted credits are concentrated in LT, EA, and MR categories. Notably, some categories simultaneously include both highly and poorly adopted credits—IEQ contains both highly adopted credits (construction indoor air quality management plan—95%) and poorly adopted credits (acoustic performance—8% and daylight—13%).

LEED certification credits show significant regional variation across the United States (See

Table A2. Adoption of LEED credits across US climate regions.

Category	Credit	Credit CAD (Overall)	Northeast	Northern Rockies and Plains	Northwest	Ohio Valley	South	Southeast	Southwest	Upper Midwest	West	CV	Asymp. Sig.
IN	LEED Aced Professional	99%	98%	100%	98%	99%	99%	99%	98%	100%	100%	1%	0.000
IEQ	Construction indoor air quality management plan	95%	97%	96%	96%	96%	96%	94%	94%	93%	96%	1%	0.000
LT	Sensitive land protection	88%	91%	88%	93%	89%	82%	87%	83%	80%	92%	5%	0.000
SS	Site assessment	83%	83%	88%	90%	82%	79%	85%	83%	80%	79%	4%	0.000
MR	Construction and demolition waste management	82%	86%	90%	88%	84%	71%	84%	68%	76%	83%	9%	0.000
IN	Innovation	77%	76%	77%	80%	76%	75%	76%	74%	69%	81%	4%	0.000
WE	Outdoor water use reduction	70%	81%	50%	63%	74%	64%	79%	52%	73%	61%	16%	0.000
IEQ	Low-emitting materials	69%	69%	69%	71%	73%	74%	70%	59%	57%	70%	8%	0.000
SS	Heat island reduction	69%	57%	68%	64%	64%	86%	73%	65%	59%	80%	13%	0.000
WE	Water metering	69%	73%	88%	72%	75%	67%	62%	62%	69%	65%	11%	0.000
IEQ	Enhanced indoor air quality strategies	67%	74%	88%	69%	71%	64%	63%	63%	68%	63%	11%	0.000
EA	Enhanced commissioning	62%	66%	69%	60%	66%	68%	57%	59%	64%	57%	7%	0.000
EA	Optimise Energy Performance	61%	58%	65%	58%	53%	56%	56%	73%	60%	75%	12%	0.000
SS	Open space	61%	64%	68%	72%	59%	57%	60%	69%	64%	52%	10%	0.000
SS	Light pollution reduction	60%	63%	76%	50%	51%	68%	46%	64%	69%	73%	16%	0.000
IP	Integrative Process	57%	53%	60%	51%	65%	57%	52%	64%	51%	62%	9%	0.000
LT	Reduced parking footprint	53%	61%	40%	53%	56%	38%	57%	45%	56%	52%	15%	0.000
MR	Building product disclosure and optimisation—material ingredients	53%	58%	40%	55%	56%	59%	52%	45%	46%	49%	12%	0.000
WE	Indoor water use reduction	52%	51%	58%	50%	48%	53%	54%	51%	51%	56%	6%	0.000
LT	Surrounding density and diverse uses	51%	57%	37%	60%	53%	44%	47%	37%	41%	59%	18%	0.000
MR	Building product disclosure and optimisation—environmental product disclosure	51%	52%	44%	53%	55%	53%	50%	49%	46%	49%	7%	0.000
LT	Green vehicles	48%	43%	48%	58%	41%	37%	45%	45%	54%	63%	16%	0.000
IEQ	Thermal comfort	47%	53%	60%	43%	55%	41%	48%	41%	52%	37%	15%	0.000
EA	Enhanced refrigerant management	44%	47%	96%	31%	49%	41%	49%	39%	64%	30%	38%	0.000
IEQ	Interior lighting	43%	43%	44%	35%	43%	39%	42%	47%	50%	45%	9%	0.000
IEQ	Quality views	42%	46%	60%	46%	43%	48%	41%	43%	33%	36%	17%	0.000
LT	Bicycle facilities	41%	43%	40%	50%	35%	23%	38%	53%	41%	46%	21%	0.000
EA	Advanced energy metering	36%	38%	40%	37%	36%	35%	30%	39%	31%	42%	10%	0.000
EA	Renewable energy production	33%	27%	37%	23%	21%	25%	27%	50%	29%	59%	37%	0.000
LT	Access to quality transit	33%	39%	2%	55%	34%	21%	25%	25%	24%	37%	47%	0.000
SS	Rainwater management	30%	35%	16%	44%	23%	15%	30%	33%	33%	31%	30%	0.000
IEQ	Indoor air quality assessment	30%	34%	30%	24%	28%	31%	28%	26%	34%	36%	13%	0.000
LT	High priority site	29%	36%	10%	32%	31%	28%	27%	20%	22%	33%	28%	0.000
EA	Green power and carbon offsets	27%	23%	40%	28%	16%	29%	22%	31%	15%	41%	33%	0.000
MR	Building product disclosure and optimisation—sourcing of raw materials	23%	25%	32%	15%	28%	37%	22%	17%	23%	22%	27%	0.000
MR	Building life cycle impact reduction	23%	25%	28%	24%	24%	26%	18%	25%	21%	21%	13%	0.000
SS	Site development—protect or restore habitat	21%	23%	32%	34%	17%	16%	14%	21%	30%	19%	31%	0.000
WE	Cooling tower water use	15%	16%	12%	14%	14%	18%	20%	10%	14%	14%	19%	0.000
IEQ	Daylight	13%	14%	13%	11%	12%	17%	10%	18%	14%	12%	18%	0.000
IEQ	Acoustic performance	8%	8%	16%	13%	9%	6%	6%	6%	3%	9%	44%	0.000
EA	Demand response	6%	6%	14%	3%	7%	10%	4%	4%	7%	6%	48%	0.000
LT	LEED for Neighbourhood Development location	0%	0%	0%	0%	0%	1%	0%	1%	0%	0%	142%	0.000

in Appendix A). The most consistently achieved credits nationwide include LEED Accredited Professional (99%), construction indoor air quality management plan (95%), and sensitive land protection (88%), with minimal regional variation in that CV values range from 1 to 5%. In contrast, credits with the highest variability (CV > 30%) include LEED for Neighbourhood Development location, which has an adoption of less than 1% and a CV of 142%; demand response, with an adoption of 6% and a CV of 48%; enhanced refrigerant management, with an adoption of 44% and a CV of 38%; and access to quality transit, with an adoption of 33% and a CV of 47%. These values reflect substantial regional differences in implementation. The West and Northern Rockies regions generally achieve higher credit rates in renewable energy and energy performance, while the South region shows strengths in heat island reduction (86%) but struggles with credits related to public transit and bicycle facilities. The data suggests that geographical, climatic, and infrastructure differences significantly influence which LEED credits are prioritised and achieved in different regions, with the highest overall attainment in professional credentials and basic environmental management practices.

4.2.2. Correlations Among LEED Credits

The analysis of correlations among LEED credits reveals significant insights into the relationships between LEED credits across building projects. The strongest correlation observed is between ‘surrounding density and diverse uses’ and ‘access to quality transit’ (R² = 0.44), indicating that approximately 44% of the variation in one credit can be explained by the other (see

Table 7. Top correlations among LEED credits.

Credit	Credit	R-Squared Value
Surrounding density and diverse uses	Access to quality transit	0.442
Renewable energy production	Optimise energy performance	0.280
Building product disclosure—environmental product disclosure	Building product disclosure—material ingredients	0.266
Building product disclosure—material ingredients	Low-emitting materials	0.138
Building product disclosure—material ingredients	Building product disclosure—sourcing of raw materials	0.116

). This strong relationship demonstrates how urban location significantly influences transportation options, creating natural synergies between site selection and transit accessibility. Another notable strong correlation exists between ‘renewable energy production’ and ‘optimise energy performance’ (R² = 0.28), highlighting how these complementary strategies are frequently implemented together as part of comprehensive energy management approaches. The materials category also displays significant internal correlations, with ‘building product disclosure and optimisation—environmental product disclosure’ strongly correlating with ‘building product disclosure and optimisation—material ingredients’ (R² = 0.27), suggesting that projects addressing one aspect of materials transparency often address others simultaneously. This strong correlation exists because of internal synergies among the material credits. All three LEED credits in the material category use the same worksheet. All these three credits are based on similar approaches and almost the same kind of Life Cycle Analysis. Therefore, these credits are inextricably linked, irrespective of their individual identification within the LEED framework.

Most correlations in the matrix are remarkably weak (R² < 0.05), indicating that the vast majority of LEED credits are implemented independently of each other. This compartmentalization suggests that project teams typically approach LEED credits as separate, disconnected strategies rather than as part of holistic sustainability systems. For example, credits that might logically be connected, such as “indoor water use reduction” and “outdoor water use reduction,” show minimal correlation (R² < 0.012), and “daylight” shows almost negligible correlation with “quality views” (R² < 0.08) despite their conceptual connection. Even within categories, many credits show little correlation with each other, suggesting that projects often cherry-pick credits based on ease of implementation or cost-effectiveness rather than pursuing comprehensive category-wide strategies. This pattern of weak correlations across most credits highlights a potential missed opportunity for more integrated approaches to GB design and points to the siloed nature of how sustainable building strategies are typically implemented in practice.

4.3. Certification Level Analysis

Certification levels are linked to changes in LEED category scores since some categories contribute more to changes in certification levels, and some categories contribute less. To monitor such phenomena, a measure that can be used is the ‘mean value of category scores’ obtained by projects belonging to different certification levels (i.e., platinum, gold, silver, and certified) (see

Table 8. Certification-wise comparison of LEED categories.

		EA	IEQ	IN	IP	LT	MR	RP	SS	WE
Platinum	Mean	28.55	10.22	5.93	0.94	9.86	8.32	3.71	7.03	8.29
	SD	3.37	2.19	0.31	0.23	3.80	2.00	0.59	2.13	1.76
	CV	0.12	0.21	0.05	0.25	0.39	0.24	0.16	0.30	0.21
Gold	Mean	20.46	7.77	5.33	0.69	8.75	5.89	3.00	5.34	5.85
	SD	5.39	2.30	0.97	0.46	3.90	2.22	0.97	2.18	1.98
	CV	0.26	0.30	0.18	0.67	0.45	0.38	0.32	0.41	0.34
Silver	Mean	15.64	7.14	4.79	0.56	6.60	5.11	2.55	4.57	5.30
	SD	4.84	2.20	1.24	0.50	3.88	2.04	1.06	2.01	1.76
	CV	0.31	0.31	0.26	0.89	0.59	0.40	0.42	0.44	0.33
Certified	Mean	12.48	5.74	3.91	0.35	5.29	4.14	2.09	3.65	4.75
	SD	4.93	2.38	1.51	0.48	3.53	2.19	1.04	1.89	1.62
	CV	0.40	0.41	0.39	1.37	0.67	0.53	0.50	0.52	0.34
Mean Credit Increase	Certified–Silver	3.16	1.41	0.88	0.21	1.30	0.97	0.46	0.92	0.56
	Silver–Gold	4.82	0.63	0.54	0.13	2.15	0.78	0.45	0.77	0.54
	Gold–Platinum	8.09	2.45	0.60	0.25	1.11	2.43	0.71	1.69	2.44

). As expected, these mean values change from certified to silver rating, and onwards. Based on the mean credit values, it can be observed that for projects moving from certified to silver levels, the three key categories which contribute to this are EA, IEQ, and LT. For projects moving from silver to gold levels, the two impactful categories are EA and LT. For projects moving from gold to platinum levels, the four impactful categories are EA, IEQ, MR, and WE. All these five categories which can be seen to move projects from low-tier certification to high-tier certification are also the categories which carry most (81%) of the achievable scores (89 out of 110). The category of energy and atmosphere (EA) outranks all other LEED categories in terms of achieving higher certifications (

Table 9. Skewness of LEED projects for various certification levels.

	Certified	Silver	Gold	Platinum
Total score	40–49	50–59	60–79	80+
Mean score	42.39	52.25	63.07	82.84
Skewness	0.86	0.90	1.49	1.20
Standard error of skewness	0.14	0.11	0.12	0.29
z-score	5.94	8.14	12.31	4.15

). The mean score for this category increases substantially between certified and silver (+3.16 points), and between silver and gold (+4.82 points), and rises most dramatically between gold and platinum (+8.09 points). In terms of mean score changes from certified to platinum levels, the highest degree of change is observed for the EA category. This implies that this category is most critical for moving the projects from lower ratings to higher ratings. This is not unexpected, since this category offers the highest score allocation (33/110) compared to other categories. For any project to perform reasonably well in LEED certification, owing to the large score value of the energy category and because of the incremental nature of the LEED rating system, performing well in this category is essential. RP, IP, and IN are the LEED categories with lower-level mean scores across the four levels of LEED certification. These findings are comparable to the analysis of LEED v3 (2009)-certified projects by Wu et al. (2017) [20], which found that all four levels of certification had relatively higher scores in SS and EA, medium scores for IEQ, and low scores for IN, MR, and WE.

To identify the categories with greater spread across each certification level, we calculated the coefficients of variation (CV), a standardised value, by dividing the standard deviation by the mean. The coefficient of variation decreases as certification level increases for most categories, indicating that higher-rated projects demonstrate more consistent sustainability practices across multiple dimensions. We can observe this trend in Figure 5 where except for the WE category, CV values of the mean scores for all LEED categories decrease from lower certification levels to higher certification levels. This indicates that when moving from certified level to platinum level, the uniformity in achieving the scores in LEED categories increases, implying that projects at the silver level obtain somewhat different scores for a certain category, and that projects at the platinum level obtain somewhat similar scores for a particular category.

4.3.1. Skewness of Project Scores

All four levels of LEED certification demonstrate significant positive skewness in their score distributions, as indicated by z-scores for all certification levels exceeding 1.96 (Table 9 and Figure 6). While minor deviations from the overall skewness trend can be observed across different score ranges within each certification level, the specific reasons for these variations would require additional qualitative data on project-specific strategies and stakeholder decision-making processes. The positive skewness values (ranging from 0.86 for certified to 1.49 for gold projects) indicate that score distributions are right-skewed, with projects clustering near the lower thresholds of certification levels. This pattern aligns with findings from studies on previous LEED versions [19,20] and suggests a strategic “point-chasing mentality”, where project teams target the minimum points needed for a desired certification level rather than maximising sustainability performance.

This behaviour appears rational within the current LEED framework, where achieving points beyond the certification threshold offers no additional certification benefits. The mean scores of certified, silver, gold, and platinum projects typically exceed their respective lower thresholds by only 2.25 to 3.07 points, indicating that teams likely implement a few contingent credits (typically 2 to 3 points) to ensure they meet their target certification level.

Further evidence of a strategic mindset in credit selection is seen in category avoidance (see

Table 10. Statistics related to LEED project dataset.

	EA	IEQ	IN	IP	LT	MR	RP	SS	WE
Maximum possible points in a category	33	16	6	1	16	13	4	10	11
Number of projects with 20 percent or less scores in a category	33	95	38	*	272	121	205	189	52
Number of projects with 0 scores in a category	1	0	2	533	13	13	21	17	1

Note: * For IP minimum possible score is ‘0’ and maximum possible score is ‘1’. Hence, identifying projects with 20% or less scores (i.e., 0.2 or less scores) is not possible for this category.

). Of the 1252 projects analysed, 533 (42.7%) received zero points in the IP category. While this might be expected given that this category offers only one point, numerous projects also received zero scores in RP (21 projects), SS (17 projects), LT (13 projects), and MR (13 projects) categories. Many projects that secure ‘0’ or very low scores (equal to or less than 20 per cent in a category) indicate that for LEED projects, performing at the bare minimum across a number of categories is not uncommon.

4.3.2. Strategic Clustering Analysis

To further investigate whether the observed positive skewness reflects simple threshold targeting or more complex strategic behaviours, we conducted k-means clustering analysis on category scores using jamovi 2.7.4. software. This analysis tested whether distinct strategic approaches exist beyond projects merely targeting minimum certification thresholds.

The k-means clustering analysis (using k = 3) revealed three distinct strategic approaches among the 1252 projects: a location-focused strategy (Cluster 1, n = 365) emphasising location and transportation scores (11.7 points average), an indoor quality-focused strategy (Cluster 2, n = 446) prioritising indoor environmental quality (7.5 points average), and energy-focused strategy (Cluster 3, n = 441) demonstrating strong emphasis on energy and atmosphere performance (24.4 points average compared to 12.8–13.7 in other clusters). These clusters demonstrate clear strategic trade-offs between categories rather than uniform performance scaling across all sustainability dimensions (see

Table 11. Centroids of clusters resulting from k-means clustering.

Cluster	WE	SS	RP	MR	LT	IP	IN	IEQ	EA
1	5.34	4.62	2.60	5.55	11.71	0.60	4.98	6.68	12.84
2	5.47	4.67	2.42	5.16	4.35	0.50	4.64	7.48	13.70
3	5.77	4.97	2.97	5.34	6.33	0.63	4.94	7.40	24.43

).

The cross-tabulation of clusters with certification levels reveals strategic complexity with important performance implications (see

Table 12. Cross-tabulation of clusters with LEED certification levels.

Cluster	Cluster Focus	Certified	Silver	Gold	Platinum	Total
1	LT	67	149	146	3	365
2	IEQ	187	232	26	1	446
3	EA	31	109	232	69	441
Total		285	490	404	73

). While multiple strategic approaches exist within each certification level, certain strategies prove more effective for achieving higher performance: 94.5% of platinum projects (69 out of 73) follow the energy-focused strategy, while IEQ-focused approaches rarely achieve gold or platinum levels (only 27 out of 446 projects). This suggests that while strategic diversity exists beyond simple “point-chasing,” energy emphasis provides the most viable pathway to top-tier certification, supporting our earlier finding about EA’s dominant influence on overall project performance.

5. Discussion

5.1. Comparison with Prior LEED Versions

When comparing the findings of this study (for LEED v4) with adoption rates across different LEED versions investigated by previous studies, both consistencies and notable shifts can be observed (see

Table 13. Comparison of adoption rates for LEED categories.

Category	Adoption Rate
IN	80.7	59.2		83		76	82.0	80.0	67.8
RP	66.6	77.6				88	96.0
IP	57.4	59.9
EA	52.2	56.9	46.2	48	40.2	44	47.0	43.4	40.3
WE	50.3	64.6	58.2	71	64.0	77	78.0	68.9	60.8
SS	47.6	49.4	54.9	53		76	66.0	57.3	61
IEQ	45.1	38.7	70.7	62	64.0	49	50.0	66.0	59.1
LT	44.9	49.9
MR	41.1	29.3	42.7	41	38	39	36.0	42.6	38.7
Source	This Study	[21]	[16]	[18]	[22]	[23]	[3]	[19]	[20]
Region	USA	Worldwide	USA	Arizona, USA	USA	Turkey	MENA region	Worldwide	Worldwide
No. of certified LEED projects considered	1252	222	102	48	95 *	122	452	5327	3416
LEED version	v4.0	v4.0	v2.0, v2.1, v2.2	v2.0, v2.1, v2.2	v3 (2009)	v3 (2009)	v1.0, v2.0, v2.1, v2.2, v2007, v2008, v2009	v2.2	v3 (2009)

Note: * = only fire station projects.

). LEED v4 (BD + C) has also been investigated by Pham, Kim [21]. However, unlike our study, Pham, Kim [21] considered a smaller sample of 222 projects spread worldwide. Some of our findings are in sharp contrast to those of Pham, Kim [21]. These variations may stem from their smaller sample size and worldwide scope compared to our US-focused analysis.

The consistent hierarchy of category adoption rates across different LEED versions suggests inherent patterns in the rating system design. IN achieves high adoption rates across all LEED versions. This indicates that IN credits may be more accessible compared to other categories. Conversely, the persistent low adoption of MR across LEED versions suggests a recurring challenge that also transcends regional contexts, since low performance in this category is noticed not only for US-based projects but also for projects located elsewhere. This pattern may reflect practical challenges in balancing standard construction practices with sustainable material considerations, as noted by Ma and Cheng [9]. The complexity and cost associated with sourcing sustainable materials, verifying supply chains, and implementing life cycle approaches may present significant challenges within the current LEED framework.

LEED professionals often select credits based on feasibility and simplicity, leading to higher adoption rates for those perceived as easy to implement [16]. For example, credits under the IN category are frequently adopted because they offer up to six points and often involve standardised templates, such as Green Education or Green Cleaning, which require minimal customization [32]. Similarly, the IP credit is commonly pursued since it only involves completing a simple worksheet and participating in design-phase discussions, making it a low-effort addition.

For consistently underperforming categories, we propose the following recommendations:

• We propose creating streamlined documentation processes for complex credits, thereby reducing the administrative burden that often deters teams from pursuing challenging credits. Digital transformation is particularly important for improving performance in the material and resources (MR) category, which has a low achievement rate of 41.1%. Complex paperwork, the tracking of materials information, and activities that consume excessive time make it less likely for project teams to pursue this credit. The introduction of digital tools can be instrumental in streamlining the processes related to the MR category.
• USGBC can focus on developing region-specific resources, case studies, and implementation guides focused on underutilised categories, highlighting cost-effective strategies and successful approaches.
• USGBC can partner with industry stakeholders and government agencies to create financial incentives or recognition programmes specifically targeting underutilised categories, offsetting the higher costs or complexity associated with these credits.

5.2. Addressing Regional Patterns

The statistically significant variations in category adoption across climate regions highlight an important consideration in the LEED methodology: the challenge of applying standardised criteria across diverse geographical contexts. Despite LEED v4 introducing regional priority credits and alternative compliance paths, our analysis demonstrates that regional disparities persist across multiple categories.

Regional patterns are particularly pronounced for certain categories. Some show minimal geographic variation while others demonstrate substantial differences across climate regions. These variations reflect complex interactions between local environmental conditions, infrastructure availability, regulatory frameworks, and market conditions.

Table 14. Reasons behind low and high CV values for LEED categories.

LEEED Category and Observed CV Across US Regions	Rationale
IN (3.4)	IN credits demonstrate considerable flexibility, enabling projects to either propose unique sustainability strategies or pursue established pathways such as green education, active design, or exemplary performance. Despite this flexibility, many projects implement similar innovation strategies—including occupant surveys and green cleaning protocols—irrespective of geographic location. The IN credit structure operates independently of regional variables, relying instead on comprehensive documentation and creative implementation. Since LEED guidelines explicitly delineate common innovation strategies, project teams nationwide tend to converge on similar approaches, resulting in limited variation across different regions.
WE (4.7)	Within the WE category, credits for indoor water use reduction are determined by the flow rates and specifications of fixtures such as faucets, shower heads, spray valves, and appliances (certified through ENERGY STAR, WaterSense labels, and similar programmes). This standardisation results in minimal variation across projects, as design teams typically select from commercially available products that meet budget constraints and functional requirements.
IEQ (4.9)	Regarding Indoor Environment Quality, projects typically adhere to standardised strategies that are uniformly defined or follow established protocols, resulting in consistent score distributions across projects. For instance, in the lighting credit category, most projects target a Colour Rendering Index (CRI) value of 90 or higher to secure one point for interior lighting. Similarly, thermal comfort credits are readily achievable since thermostats are standard equipment in virtually all projects, and HVAC system design universally follows ASHRAE standards. These credits are primarily dependent on building systems and material specifications rather than regional climatic conditions or geographic location.
IP (9.4)	The variation in this category stems from several factors that influence credit adoption. The implementation of this credit depends on the project team’s organisational culture, the client’s level of sustainability knowledge, and design team scheduling constraints, all of which vary significantly across regions and project types. The persistence of design–bid–build delivery methods in certain areas inherently limits opportunities for integrative approaches, thereby creating inconsistent credit pursuit patterns. Furthermore, project approaches demonstrate considerable variation in sustainability ambition as some projects aim only to meet minimum local compliance thresholds without pursuing enhanced sustainability objectives, while others exceed these baseline requirements, contributing to the observed high CV value.
EA (10.4)	Energy systems design strategies vary significantly across projects due to diverse climate zones, regional energy-saving incentives, and project typologies. The adoption of renewable energy systems, procurement of green energy certificates, implementation of advanced energy metering, and grid harmonisation strategies are primarily determined by economic considerations and owner preferences, resulting in considerable variation in project approaches and making standardised implementation challenging across different projects.
RP (10.9)	USGBC determines RP credits based on local environmental priorities, such as energy efficiency in the Northeast and water conservation in Arizona. This regionally differentiated approach to credit prioritisation reflects varying sustainability challenges across geographic contexts. Consequently, projects in different regions are incentivised to focus on distinct sustainability dimensions, resulting in variable scoring patterns across geographic areas.
LT (20.1)	These credits exhibit strong geographic dependency, with performance significantly influenced by location-specific factors such as proximity to transit systems, walkability indices, and urban density. Credit achievement is contingent upon the availability of transit access, mixed-use development patterns, and cycling and pedestrian infrastructure within the project’s immediate context.

presents a detailed rationale for the differences in coefficient of variation values for LEED categories across the US.

These variations reflect legitimate differences in regional priorities, infrastructure, and environmental conditions. The Western regions’ EA performance coincides with progressive energy regulations, and the Northwest’s LT performance reflects developed transit infrastructure. These patterns suggest that the current LEED methodology may benefit from further adaptation to regional contextual factors that influence sustainability priorities and feasibility.

Our analysis demonstrates significant regional variations in category adoption, reflecting differences in climate conditions, infrastructure, and regulatory contexts. Our evaluation based on climate zones shows that geographical context significantly affects how people adopt LEED credits. For example, projects in the West, Southwest, and Northern Rockies regions consistently achieve higher scores in the energy and atmosphere (EA) category. This is likely because they have greater solar potential and more progressive energy codes. The Northern Rockies and Plains region, on the other hand, has a very high adoption rate for regional priority (RP) credits (86%) and unique inter-category correlations (e.g., MR-IN: R² = 0.259), which suggests that strategies should be tailored to the region. The Northwest region receives 56% of its location and transportation (LT) credits, and the reason behind this strong score is that this region has extensive public transportation options. The South and Southeast, on the other hand, have lower adoption rates for some credits that depend on urban density (like access to quality transit), but they perform better at reducing heat islands and improving water efficiency, which shows that they are adapting to hotter, more humid climates. These regional differences show that a one-size-fits-all LEED framework is inadequate, and that region-sensitive benchmarking, alternative compliance paths, and weighted credit scoring based on local environmental conditions are all important.

Implementing climate-specific strategies could make LEED much more relevant and equitable across different regions. To address this, USGBC could explore a region-sensitive point allocation model, where region-specific credits are assigned region-adjusted point multipliers. Under this approach, the number of points awarded for selected credits would vary based on regional climate conditions and resource priorities. For instance, WE credits may carry greater point weight in arid regions, while thermal envelope performance could be emphasised in colder climates. However, such region-sensitive adjustments must be carefully designed to align with actual regional environmental priorities and resource constraints to avoid creating perverse incentives that reward inappropriate practices in unsuitable contexts. Additionally, USGBC could systematically develop alternative compliance paths (ACPs) for major credits that are sensitive to regional variations. This would help accommodate differences in building practices, available technologies, and infrastructure, thereby providing more equitable pathways to certification across diverse geographic contexts.

5.3. Addressing Point-Chasing Mentality

The positive skewness observed across all certification levels indicates strategic behaviour among project teams—what we term the “point-chasing mentality.” This approach is understandable within the current LEED framework, which focuses on threshold achievement, but may affect the system’s ability to promote comprehensive sustainability.

To promote more balanced sustainability approaches and address the strategic “point-chasing” behaviour, USGBC could consider introducing a greater number of minimum achievement thresholds or prerequisites. This would foster more holistic sustainability approaches while still allowing flexibility in emphasis. LEED has already introduced a number of credits as ‘prerequisites’ which are the mandatory baseline requirements that all projects must satisfy to be eligible for certification, serving as non-negotiable minimum standards that earn zero points but are essential for qualification. They ensure every certified project meets essential environmental and health standards, while regular credits reward performance above these baselines. LEED v4 has 13 prerequisites, and in the latest LEED version (v5), the number of prerequisites has been expanded to 16.

In LEED v4, system-level submetering was part of the advanced energy metering credit under the energy and atmosphere (EA) category. In LEED v5, the detailed submetering of major energy systems has been elevated to a prerequisite, reflecting a stronger emphasis on energy performance transparency and operational efficiency. This development indicates that when introducing new versions, LEED has gradually moved some point-based credits to the list of prerequisite credits. An increase in the number of prerequisites would help mitigate the negative implications of the point-chasing mentality, as achieving LEED certification is only possible if all prerequisites are met. The strategic selection of credits to be moved to the list of prerequisites can also help ensure integrated sustainability outcomes. This can be achieved if the credits being selected for prerequisites are those which reduce the effort involved in achieving other credits.

However, while the strategic use of prerequisites can promote more robust sustainability and discourage superficial compliance, over-reliance on this approach risks shifting LEED from a flexible performance-based system to a more prescriptive one. Therefore, a balanced application of prerequisites is essential to preserve the adaptability and context-sensitivity that have been central to LEED’s success.

Additionally, to address the positive skewness, for higher certification levels (such as gold or platinum), USGBC can mandate more balanced achievement across categories. For instance, requiring at least 60% point achievement within each category to be eligible for the targeted certification level.

LEED-certified buildings are more oriented toward design-based calculations and criteria. There are hardly any credits that focus on continuous monitoring and evaluation. Considering the use phase of the built environment, it is crucial to assess real-time performance. Buildings need to maintain ongoing performance. For example, in terms of lower energy consumption, reduced water use, better air quality, or higher occupant satisfaction. The recent updates to LEED and WELL standards—which now emphasise operational performance—combined with crosswalk documents that align these certification systems can help drive the industry toward greater focus on ongoing building performance monitoring rather than just design-phase compliance. A radical change suggested for LEED would be to transition from design-based to performance-based certification for existing buildings, requiring demonstrated performance improvements over time to maintain certification status.

Another radical step would be to complement or eventually replace the current four-level certification model (certified, silver, gold, platinum) with a continuous scoring scale that recognises every additional point earned. While LEED technically uses a continuous point system (ranging from 0 to 110), the current tier-based recognition framework creates threshold effects, encouraging teams to do just enough to reach the next level, rather than pursuing deeper sustainability improvements that fall short of the next threshold. In contrast, Japan’s CASBEE (Comprehensive Assessment System for Built Environment Efficiency) offers a useful reference. Its core evaluation metric—the Built Environment Efficiency (BEE) indicator—is a continuous ratio of Q (building environmental quality) to L (building environmental load). This performance is plotted on a graph, where a steeper gradient reflects a more sustainable building. Although CASBEE still assigns letter grades (C, B−, B+, A, S), the underlying continuous nature of the BEE score enables finer differentiation between projects, even within the same grade. Importantly, CASBEE’s grading is derived from the BEE ratio and not used as a performance target, which helps to minimise the behavioural distortions commonly seen in threshold-based systems like LEED. This model demonstrates how a continuous or hybrid performance scale can better capture nuanced sustainability outcomes, reward incremental improvements, and discourage the “point-chasing” mentality. Incorporating a similar approach in LEED, such as public disclosure of precise point scores or percentile rankings alongside certification levels, could preserve the clarity of tier labels while still promoting deeper performance and fairer comparisons across projects.

These recommendations aim to address the methodological patterns identified in our analysis while preserving the fundamental structure and recognition value of the LEED rating system. By implementing these changes, USGBC could enhance the system’s effectiveness in promoting comprehensive sustainability practices across diverse regional contexts.

5.4. Category Interrelationships and Integration

The fact that most LEED categories do not have a significant impact on each other creates a problematic pattern where buildings that excel in WE may not perform well in EA or IEQ. This lack of interdependence among LEED categories has led project teams to pursue credits based on what is feasible or what is crucial to project objectives rather than employing integrated sustainability strategies—an approach that undermines LEED’s foundational principle of holistic design. While there is no one-size-fits-all approach to LEED implementation, the predominant practice of pursuing categories separately represents a missed opportunity for truly integrated sustainable design that could serve as a model for broader adoption.

The correlation analysis between LEED categories reveals a critical insight into how sustainability is approached in practice. The consistently weak correlations between most category pairs (typically R² < 0.05) suggest a systematic compartmentalization of sustainability strategies. This pattern indicates that project teams generally treat different aspects of GB—such as energy efficiency, water conservation, and material selection—as separate technical challenges rather than interconnected dimensions of a holistic system.

This compartmentalised approach stands in contrast to LEED’s theoretical underpinnings, which emphasise integrated design and the interrelationships between different sustainability aspects. For instance, the weak correlation between EA and IEQ (R² = 0.01) suggests missed opportunities for strategies that simultaneously enhance energy efficiency and occupant well-being. Similarly, the minimal relationship between WE and SS (R² = 0.02) indicates limited integration between site design and water management strategies.

The weak correlations between categories indicate that project teams lack systematic procedures for integrating LEED categories and typically evaluate them sequentially rather than through integrated approaches [33]. The correlations between category scores and overall project scores provide further evidence of this dynamic. EA demonstrates by far the strongest relationship with overall project performance (R² = 0.38), which helps explain its outsized influence on certification levels. This strong correlation, combined with the category’s substantial point allocation (30% of available points), creates a certification landscape where energy performance can dominate at the expense of balanced sustainability approaches.

These findings suggest that the current LEED methodology, despite its intentions, has not prevented a siloed approach to sustainability. The credit structure, which allows certification through the accumulation of points across independent categories without requiring significant integration, may be contributing to this pattern. This insight provides an important consideration for future refinements of the LEED system, suggesting that promoting stronger connections between sustainability dimensions could enhance its effectiveness in fostering truly integrated GBs. To address the compartmentalised approach revealed by our correlation analysis, we recommend several structural changes to the LEED system:

• USGBC can introduce bonus points for achieving high scores across multiple categories, incentivising comprehensive excellence rather than minimum compliance.
• LEED can emphasise synergistic credits (i.e., credits which, when pursued, also help achieve other credits) by increasing their score values. This may also mean creating some credits specifically for rewarding connections and synergies between traditionally separate categories. For example, credits could be developed that simultaneously address energy efficiency, indoor air quality, and occupant comfort in an integrated manner. Additionally, for emphasising synergistic credits, point multipliers can be introduced that increase value when related credits across different categories are achieved together, incentivising teams to pursue complementary strategies rather than cherry-picking unrelated credits.
• Integration planning can be made a prerequisite for higher certification levels, requiring projects to document how strategies across categories work together as interconnected systems rather than independent components.
• The existing IP credit can be expanded and strengthened to include specific requirements for demonstrating how design decisions connect and optimise performance across multiple sustainability dimensions.

These changes would help align LEED’s methodological structure with its foundational principle that truly sustainable buildings require integrated approaches across all dimensions of sustainability, rather than excellence in isolated categories.

6. Limitations and Future Research

Our study provides valuable insights into the methodological foundations of LEED v4, but several limitations should be acknowledged. First, our analysis is limited to projects in the United States and specifically to BD + C: NC projects. Additionally, our use of coarse climate regions may mask finer-scale spatial variations that could reveal more localised certification practices. Future research could expand this analysis to include other countries and LEED rating systems other than v4 BD + C:NC, and examine adoption patterns at state or metropolitan levels to provide a more comprehensive understanding of adoption patterns.

LEED v5 somewhat incorporates a few of this study’s conclusions and suggestions, such as the addition of performance metrics, greater focus on equity, and decarbonization. The IEQ category focuses more on measuring and benchmarking performance rather than just the design phase. However, several structural problems are still barely addressed, including weak inter-category synergies, regional disparities in credit feasibility, and the incentive for threshold-focused scoring. For instance, dynamic region-sensitive benchmarking and integration-based bonus structures are not yet implemented in LEED v5, which still employs a categorical point-based framework with fixed thresholds. Therefore, this study provides a diagnostic assessment for LEED v4 as well as a foundation for evidence-based monitoring of LEED v5 effectiveness over time. Furthermore, our analysis revealed regional anomalies in category correlations (such as the unusually high MR-IN correlation in the Northern Rockies and Plains region) that warrant investigation into underlying project-specific practices, regional practitioner networks, and local material/innovation ecosystems that may drive unique sustainability strategy integration patterns across different US regions.

Although we compared the findings of our study (for LEED v4) with previous studies, this study employs a cross-sectional design approach. Longitudinal studies tracking changes in adoption patterns as projects transition from LEED v4 to LEED v5 would provide valuable insights into how evolving certification systems influence sustainability practices over time.

Second, our study focuses on credit achievement patterns without examining the underlying reasons for these patterns beyond regional variations. Future research could incorporate qualitative approaches such as interviews with project teams or case studies to better understand the decision-making processes that influence category adoption. While our quantitative analysis provides empirical evidence of LEED adoption patterns, future research incorporating qualitative methods such as interviews, focus groups, or case studies with LEED consultants, architects, and developers would provide valuable insights into the underlying motivations for strategic decisions, particularly regarding the persistent underutilization of material and resources credits and the decision-making processes that drive category selection and prioritisation.

Third, while we identified the “point-chasing mentality” through statistical analysis, we did not directly measure the motivations or strategic considerations of project teams. Future studies could employ surveys or interviews to directly assess how teams strategically approach LEED certification and make trade-offs between categories. Additionally, our study does not evaluate the actual environmental or social performance outcomes of the certified buildings. Future research could link credit achievement patterns to measurable sustainability outcomes, such as energy consumption, water usage, or occupant satisfaction, to assess how well LEED certification correlates with real-world performance.

Fourth, correlation analysis cannot account for potential confounding variables such as project budget, team experience, or local regulations. Additional statistical techniques for investigating category and credit interrelationships need to be tested.

Regarding relationships between LEED credits, the LEED Reference Guide includes “Related Credit Tips” that outline theoretical synergies between credits. However, our analysis focuses on empirical certification outcomes to assess whether these synergies are reflected in actual project implementation. The lack of strong statistical correlations between many theoretically related credits (such as EA and IEQ, or daylight and quality views) suggests that real-world implementation may not always follow theoretical guidance due to practical constraints. For example, a project might achieve quality views through well-designed windows and glazing but may avoid pursuing daylight credits due to glare issues for occupants or building use types that do not require daylighting (such as X-ray rooms or CT scan facilities). Future research could systematically compare the suggested credit pairings in the Reference Guide with their empirical co-adoption rates across projects, providing insights into which theoretical synergies are not being fully realised and identifying the underlying barriers.

Future studies may expand upon this study in multiple important avenues. First, as LEED v5 becomes more prevalent, studies that compare certification patterns between LEED v4 and v5 would be very helpful for figuring out how new requirements, credit weightings, and thematic priorities (like equity, resilience, and decarbonization) change sustainability strategies. Second, examining projects outside of the United States could show how regional governance, climate, and market maturity affect the global adoption of LEED. Third, comparing LEED to other international rating systems like Building Research Establishment Environmental Assessment Method (BREEAM), Deutsche Gesellschaft für Nachhaltiges Bauen (DGNB), Comprehensive Assessment System for Built Environment Efficiency (CASBEE), and Green Star can help identify areas where the methods are similar and where they differ. This comparison will help make green building certifications more consistent. For example, future research could investigate the performance of dual-certified projects across systems or evaluate the viability of integrated frameworks. Finally, investigating how new technologies like BIM-integrated LCA tools or real-time IoT-based building performance monitoring can help develop dynamic certification models would make sustainability assessments much more useful and accountable in real-world applications.

7. Conclusions

This research advances theoretical understanding of GB certification effectiveness through comprehensive empirical analysis of 1252 LEED v4 projects, contributing to three critical theoretical domains. First, this study provides robust empirical evidence of systematic threshold-focused optimisation behaviour in GB certification, where practitioners strategically target minimum certification requirements rather than maximising sustainability performance. The consistent positive skewness across all certification levels (z-scores >1.96) fundamentally challenges theoretical assumptions that point-based certification systems inherently drive comprehensive sustainability improvement. Second, our correlation analysis reveals unprecedented evidence of compartmentalised sustainability implementation, with weak inter-category correlations (typically R² < 0.05) contradicting foundational theories of integrated design promoted by GB certification. This systematic compartmentalization suggests that current certification methodologies may inadvertently reinforce siloed approaches to sustainability. Third, the statistically significant regional variations across eight of nine LEED categories demonstrate that standardised certification systems exhibit systematic adaptation to local contexts despite uniform requirements, contributing to sustainability governance theory by revealing inherent limitations of “one-size-fits-all” environmental certification approaches.

These theoretical insights support evidence-based recommendations including integration incentive mechanisms, regional sensitivity frameworks, and performance-based thresholds that address identified behavioural patterns. While geographically limited to US contexts, this research establishes an empirical foundation for understanding how certification systems function in practice versus theoretical intention. The identification of strategic clustering patterns and systematic point-chasing behaviour provides crucial insights for designing more effective sustainability assessment tools that better align certification processes with comprehensive environmental outcomes. This work moves GB research beyond assumption-based system evaluation toward empirically informed understanding of certification methodology effectiveness.