Cognitive biases that shape the drivers and barriers to embracing green construction practices

Climate change began to show its noticeable environmental effects in the 1970s, which called for a united response from global organisations. This led the United Nations Environment Programme (UNEP) to form the International Panel on Climate Change (IPCC) in response to the growing concerns regarding the environmental impact caused by human activities and the resulting weather catastrophes that follow (MacGregor et al., 2018). In addition, through the initiative of the IPCC, two significant efforts were established to address climate change: the Kyoto Protocol in 1997 and the Paris Agreement in late 2015, with the primary collective goal of controlling greenhouse gas emissions that affect the overall global temperature. Many countries make serious efforts in this regard.

The concept of green building is the integration of green design with the primary focus on reducing a structure’s environmental impact while improving the occupants' quality of life during its whole lifecycle compared to a traditional “non-green” facility (Ali’and Al Nsairat, 2009). Aside from improving the environment and public health, Ali and Al Nsairat (2009) also indicated that green design lowers the structure’s lifecycle costs, enhances the marketability of buildings and organisations, boosts occupant efficiency, and aids in developing sustainable communities. In general, green building is associated with sustainable construction, but despite their frequent interchange, the terms “green” and “sustainable” are not synonymous in scholarly discourse (Doan et al., 2021). In contrast, Purvis et al. (2019) suggest that sustainable construction encompasses a holistic view of sustainable development, encompassing the economic, social, and environmental aspects of a building, which are the three pillars of sustainability. These three aspects have been long accepted for organisational sustainability reporting, and many building evaluation technologies have considered them to provide organised and comprehensible information on social, economic, and environmental solutions (Du Plessis, 2007). Ali and Al Nsairat (2009) identified these evaluation tools as a comprehensive system that measures the environmental impact of a building. Among the global green rating tools are BREEAM (Great Britain), GBTool (Canada), LEED (US), EcoProfile (Norway), and Environmental Status (Sweden) (Ali and Al Nsairat, 2009). New Zealand has the Green Star NZ as its primary green rating system (Doan et al., 2021). However, it was the least effective among the green rating tools described as limited to a single aspect of sustainability: the environment.

In the construction decision-making process, decisions can take many forms, from strategic to corrective, and are essential for all projects (Love et al., 2023). Decision-making is, therefore, necessary to ensure the success of projects. Love et al. (2023) also noted that prompt decision-making is crucial to ensure that no responsibilities are delayed and that the project’s timeline is met. Additionally, Kamranfar et al. (2022) suggested that it is vital to carefully evaluate the significance of each criterion, factor, and structure to enhance the quality of holistic decision-making. Literature is abundant regarding decision-making, addressing various construction perspectives and offering different approaches to identify factors that affect the construction industry. Conveying decision-making strategies in the construction sector assists in assessing and prioritising innovative drivers towards sustainable development (Van Nguyen, 2023). Kamranfar et al. (2022) related decision-making in uncovering the significance and connections among construction sustainability indicators. Moreover, Sun et al. (2022) employed a similar approach to reveal causal relationships and intensity within system elements, thereby facilitating the successful implementation of green supply chain management. Furthermore, decision-making is associated with understanding the causal effects of interrelationships among barriers in sustainable waste management (Negash et al., 2021).

From the construction industry perspective, heuristics are generally regarded as liabilities resulting from cognitive biases in decision-making, particularly when determining risks and unpredictable situations (Ika et al., 2022). Moreover, different factors affect the decision-making to achieve sustainable development, and these factors also influence human behaviour and judgement through cognitive biases (Korteling et al., 2023). Korteling et al. (2023) suggested that cognitive biases may lead to prompt, sensible, and fulfilling judgements in a natural and fundamental application. However, these decisions may be inefficient and potentially harmful in addressing various modern, complex, and long-term issues, such as mitigating climate change or preventing pandemics. One notable example is the influence of cognitive biases within the Lean environment. Purushothaman et al. (2023) implied that managers could stimulate cognitive biases to benefit from Lean and its applications. The study also contributed to the well-known biases and how they affect the ideas that now assess and lessen the challenges in the Lean environment. On the contrary, cognitive biases are particularly prominent in terms of drawbacks to decision-making. Collusive bidding (Peng et al., 2022) and contracting practices (Jennejohn et al., 2022) succumb to the adverse behavioural effects of cognitive biases.

This paper analysed different drivers and barriers associated with the global construction industry, which are further investigated through the association of cognitive biases. The research questions for this study are:

1. What drivers and barriers are prevalent in green and sustainable construction practices?

2. What are the associated cognitive biases that affect drivers and barriers to green and sustainable construction practices?

This paper’s originality lies in its thorough examination of the diverse and interconnected factors that shape the construction practices associated with green buildings. It meticulously investigates the primary influences that encourage sustainable building methods, highlighting the motivating forces behind adopting eco-friendly practices. Additionally, the discussion addresses the various obstacles that impede the successful implementation of these practices, analysing the challenges faced by stakeholders in the industry. Furthermore, the paper examines the cognitive biases that can influence decision-making in this realm, illuminating how preconceived notions and biases may impact the choices made by builders, architects, and developers in their pursuit of sustainability.

To address the first research question in this study:

1. A systematic literature review (SLR) was employed to identify the drivers and barriers associated with green and sustainable construction practices.

2. The data analysis of the known drivers and barriers employs thematic analysis related to the well-established political, economic, social, technological, legal, and environmental (PESTLE) framework.

3. The results are presented in a system dynamic mapping, specifically a causal loop diagram, using the Vensim simulator.

To address the second research question in this study:

1. The identification of associated cognitive biases among different factors in the construction industry is through a systematic literature review (SLR) based on the initial results from the first question.

2. Data on the cognitive biases concerning the factors is analysed through narrative analysis.

This study employs an idealistic methodology that defines various experiences from various sources (Hughes and Sharrock, 2016; Ormston et al., 2014). The study focuses on the cognitive biases related to drivers and barriers and the decision-making process of green construction, suggesting a subjective approach because different construction practices would have distinct cognitive biases. Thus, idealism was selected as the ontological perspective of the study. Decision-makers constantly construct research solutions in idealistic situations (Bryman, 2016). This research adopts a constructivist epistemological position, focusing on known factors and cognitive biases. Following the constructivist approach, the study will adopt interpretivism as its theoretical position, as the associated factors and cognitive biases are interpreted through different analyses based on the individual scope of human behaviour, as implied by Cohen et al. (2002). Moreover, to discuss the comprehensive data analysis, the study will follow a qualitative narrative inquiry as its methodology to achieve a thorough comprehension (Murray, 2009; Savin-Baden and Niekerk, 2007). The summary of research methodology is viewed as a research onion, as shown in Figure 1.

This study employed the systematic literature review (SLR) method, a thoroughly constructed and methodological review of the literature that answers research questions by identifying, selecting, and critically assessing the findings from the studies included in the review process (Rother, 2007). Moreover, Pradana et al. (2023) stated that SLR provides a straightforward and impartial comprehensive assessment of different types of literature, uncovers research gaps, gathers and integrates evidence, and suggests further research possibilities. This method is similar to the methodology used by Purushothaman and Seadon (2023), where the literature retrieved and included in the research consists of English texts, primarily focusing on journal articles, conference proceedings, and books to ensure credibility and reliability. The papers were reviewed at least twice to ensure quality control of data and information, thereby enhancing the dependability of the study. In addition, careful analysis was practised by observing the theme of the topic, which focuses on architecture, engineering, management, sustainability, and construction.

The framework (shown in Figure 2) of the study began with Stage 1 Systematic Literature Review (SLR), which identified the various drivers and barriers associated with green buildings. The identified factors were thematically analysed following the concept of the PESTLE framework but using a more straightforward approach that consists of three sub-categories. After the factors were subcategorised, the interrelationships among the factors were discussed from the authors’ perspective. They were collated in pairwise analysis, which was transferred into a causal loop diagram to visualise the pattern of interrelationships. Finally, stage two of SLR concluded the research by identifying the cognitive biases corresponding to the sub-categories of the factors. These cognitive biases were added to the causal loop diagram to describe the interplay between cognitive biases and factors related to green building.

The research was conducted in two stages using three reliable academic databases and providers: Scopus, EBSCOhost, and ScienceDirect. The databases also employ similar methods to process and assess keyword searches by matching query details to the search fields, specifically title, abstract, and keywords (TIABKW), as described by Penning de Vries et al. (2020). Scopus and Science Direct have straightforward and similar search strategies. On the other hand, EBSCOhost is a database provider that offers numerous databases to categorise the search and make it easier for researchers. For this study, the databases available on EBSCOhost were selected to align with the research topic’s theme, specifically Art and Architecture Complete, Business Source Complete, and GreenFile.

The keyword search centres on the central theme of cognitive biases influencing the decision-making process in green building practice. Stage one of the SLR focused on identifying the factors, barriers, and/or drivers that affect green construction practices, while stage two focused on cognitive biases related to the identified factors.

In Stage One, the main keywords used in the search were “factors,” “affect,” and “green construction.” The keywords were assessed utilising synonymous ideas, as shown in Table 1. The final keyword search was formatted using the Boolean search strategy for the different databases. Moreover, proximity search was utilised on all platforms to refine the search queries, and its implementation varies across different databases, as shown in Table 2. The stage one keyword search yielded 2,279 publications, of which 58 relevant articles were rigorously and systematically analysed using bibliometric and qualitative approaches. The keywords from articles retrieved from all databases were imported into VOSviewer to determine the co-occurrences of the keywords based on the bibliographic data from the initial systematic literature review (SLR), as shown in Figure 3. The most common keywords across all publications are construction industry, sustainable development, sustainability, green building, and project management, all of which are relevant to the research topic.

In Stage Two, the keyword search centres on cognitive bias and the factors associated with green and/or sustainable construction based on themes identified through a PESTLE analysis. However, instead of focusing on the individual PESTLE category, the factors were divided into three main sub-categories: environmental and health-related factors, industry and economy-related factors, and policy and awareness factors, as shown in Table 3. Different studies inspired the idea to form a more understandable system from complex information (Al Harazi et al., 2023; Zulu et al., 2023). Table 4 shows the Stage 2 keyword search, which generated 274 articles that were analysed, resulting in 24 relevant papers.

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) diagram (Page et al., 2021) was used to keep a systematic journal that highlights the specific steps conducted to analyse the publications that yielded the final 58, as shown in Figure 4. The stage one PRISMA diagram initial screening is based on the following criteria:

1. Duplicated Records—The study was optimised by automating the detection of duplicated records from different databases offered by the reference manager EndNote. In stage one, 627 articles were identified as duplicates.

2. Records within the selected timeline—The literature in stage one included in the study were within the 2019 to 2023 timeline to maintain the recency in the current digital era, relevance and dependability of the journal articles. This screening criteria removed 791 papers that were not within the selected timeline.

3. Non-English texts—This study focuses solely on the English language as its medium and found only one article that is not written in English.

4. Title screening—This concept opted for the direct approach in selecting the articles, which were then thoroughly screened based on their titles. The primary objective of this screening process was to select literature that included factors, drivers, and/or barriers in its titles, thereby maintaining coherence with the study’s intended focus. In total, 576 articles were excluded solely based on this criterion.

5. Retrieved full-text papers—The study focused on articles with open access to avoid inconvenience in searching for relevant information, as securing data requires additional time and cost. This screening process ensured that all papers included for the second screening were available online to the public, yielding 270 retrieved documents.

The 270 retrieved papers were then screened for the second time following the criteria to maintain relevance according to the research questions:

1. Reasons 1 and 2: Methodology of the articles—This study employs SLR as its primary method, thereby excluding papers that use a similar process to highlight the valuable quantitative methods of other articles. This paper also excluded papers that utilised “modelling,” which included articles that used variables based on a data matrix, as these papers are very comprehensive and would require extensive time to analyse. Fifty-two papers were excluded from this screening due to these reasons.

2. Reason 3: Articles out of scope – This study focuses on the construction industry and excludes papers that involve technology integration, such as artificial intelligence, blockchain, and cloud computing, as these research areas fall outside the scope. In total, 117 articles were removed following this reason.

3. Reason 4: Papers that are within scope but not the topic—This paper focuses on the construction industry, excluding other elements such as waste management, supply chain, and geotechnical data. These papers tend to focus more on specific information instead of the adoption factors in green construction, which is one of the main ideas of this study. This reason excluded a total of 11 papers.

4. Reason 5: No relevance—This study excluded 32 articles that were not directly relevant to the information on factors influencing the adoption of green construction.

Figure 5 presents a summary of stage two of the SLR, which follows the same concept and format as stage one of the SLR. The initial stage of the screening in stage two is very similar to stage one but with only a difference in the selected timeline of 2020–2023 in stage two to analyse more recent information and exclude non-academic articles instead of the title exclusion for stage two. The main difference was in the second screening, which stage two had the following criteria:

1. Reason 1: Review papers – This section removed two SLR papers, similar to stage one.

2. Reason 2: Medical articles—Cognitive bias focuses on the medical aspect of literature, which was excluded in this research. A total of 31 medical-focused articles were removed from the selection.

3. Reason 3: Papers that do not have specific information—Four papers were excluded from this study because they addressed the general idea of cognitive biases. The study omitted them since this paper needed to identify specific cognitive biases.

4. Reason 4: Articles not in the construction decision-making process—This research focuses on cognitive biases regarding decision-making in the green construction industry. In this regard, 15 papers were excluded since they focused on decision-making in different sectors.

This section provides detailed information on the SLR results, a discussion of the various factors, their interrelationships, and the associated cognitive biases that influence the decision-making process in green construction.

The 58 relevant SLRs, as shown in Table 5, have identified 48 barriers and 47 drivers, totalling 95 factors that influence the adoption of green construction. Table 6 shows the individual factors identified and organised according to the sub-category, as classified using thematic analysis in Table 7. Moreover, a stacked bar chart in Figure 6 was created to visualise the total number of studies concerning the countries where the articles were conducted. By geography, Nigeria has the highest number of related articles, with a total of seven references between 2019 and 2023. Malaysia follows it with five papers, and then South Africa with four. These results indicate that most factors, particularly barriers, are prevalent in developing nations, as illustrated in Table 5.

The qualitative thematic analysis revolved around the PESTLE analysis, which is a management tool to identify associated factors, drivers, and/or barriers concerning the political, economic, social, technological, legal, and environmental aspects that are prone to influence complex processes (Capobianco et al., 2021). They further implied that the PESTLE framework assists decision-makers in contemplating possibilities that may affect the efficacy of their decisions. Moreover, the analysis is also inspired by the three pillars of sustainability: economic, social, and environmental factors. This study identified three sub-categories in identifying the drivers and barriers to adopting green construction, as shown in Table 7.

Table 8 and Table 9 show the rankings of the drivers and barriers in green buildings. The rankings are according to the number of occurrences of each factor from the 58 screened academic journals. The top ten drivers and barriers were listed to highlight the prevalent factors in adopting green building construction. A total of thirteen drivers and ten unique barriers were identified. PARD-04 was the prevailing driver with 28 occurrences. At the same time, PARB-11 had the most occurrences for the barriers, with 45 references stating that lack of awareness about the benefits of sustainable and green construction is the predominant reason for not adopting green building construction.

The absence of a quantitative study in the SLR necessitated a dynamic diagram of the system to analyse the collected information thoroughly. Thus, a causal loop diagram (CLD) or a system dynamic map is created to identify the interlinks between the factors and the cognitive biases associated with them (Purushothaman and Seadon, 2023). The diagram illustrates the visual representation of the links and polarities among the elements in this study, which include barriers (B), drivers (D), and cognitive biases (CB). Blair et al. (2021) discussed how the CLD begins with a primary objective and proceeds backward, adding either positive (+) or negative (−) factors until it reaches viable interrelations.

The system dynamic mapping was created using the information taken from Table 10. However, due to the large number of factors (95), the top 10 drivers and barriers from Tables 8 and 9 were considered. This method simplifies the understanding of the interrelationship between the most common factors that were identified in the tables. There were 10 total drivers and 13 barriers; all were listed and interconnected using the Vensim application. On the other hand, the data in Table 10 are based on the pairwise comparison between the individual factors in Table 6. Nordstokke and Stelnicki (2014) referred to pairwise comparison as a statistical process in evaluating connections and links between two or more data. Moreover, Nordstokke and Stelnicki (2014) further implied that pairwise comparison would comprehensively analyse the relationship between the dependent variables. The factors from Table 6  are listed in the rows and columns where the SLR articles determine the intersecting relationship among each variable. The relationship may be either positive, where factor A amplifies or supports factor B, or negative, where the connection diminishes the impact between the factors. In Table 10 , the positive and negative relationships are denoted by P and N, respectively. The digits beside P and N represent the number of occurrences of each relationship from different references.

This research uncovered 433 positive interactions or relationships between drivers and barriers in green construction adoption. These relationships exhibit varying frequencies of occurrence: 397 instances where each unique relationship was mentioned only once (1P), 22 cases where a relationship was cited twice (2P), 10 occurrences where relationships were referenced three times (3P), two instances where a relationship appeared four times (4P), and five occurrences where distinct relationships were reiterated twice (5P). On the other hand, a total of 315 negative relationships were identified. The frequencies are as follows: 306 were found once (1N), eight relationships were repeated twice (2N), and a single negative relationship was reiterated thrice (3N).

The 24 articles included in this research for Stage Two have identified a total of 71 cognitive biases based on the sub-categories identified in Table 7. Table 11 presents the number of cognitive biases identified by each author, along with the countries in which the studies were conducted. Moreover, the top 8 cognitive biases were ranked according to the number of occurrences in the articles, as shown in Table 12. On the other hand, Table 13 presents the cognitive biases identified from the 24 articles found in Stage 2, categorised by the sub-factors from Stage 1. The top bias is the status quo bias, with eight references, which is also part of the related barrier from stage one SLR, the fifth common barrier (PARB-15), with a total of 34 occurrences. The articles cover various aspects of the construction industry, including quality of living, policymaking, collusive bidding, lean construction, and risk management.