Energy-Saving Behavior in Workplaces: A Bibliometric and Visualization Analysis

3.1 Performance analysis

Performance analysis is descriptive analysis aiming to assess the major scientific players in contributing to a given research field's scientific productivity [42]. In this study, performance analysis is used to evaluate the top publication trend, top publication sources, top authors, top country, and top article of ESBW. Before conducting the analysis, this study depicted the preliminary information about the collected data for 194 publications related to ESBW (Table 1).

Table 1. The preliminary information on performance analysis

3.1.1 Publication trend

Bibliometric analysis reveals different publication trends of ESBW over the past two decades. The publication trend of ESBW is shown in Figure 2. It can be seen that the research topic of ESBW shows slow growth in the early years. There is only one publication in 2005 and 2006 and none in 2007, 2008, and 2010. However, the number of publications shows an overall increasing trend from 2009 to 2023, while there are some years that experience a decrease compared to the previous year, such as in 2010, 2013, 2015, 2018, and 2023. The highest number of publications occurred in 2022 with 27 articles. In particular, this research field has experienced significant growth starting from 2014 (10 articles) and a steady increase in output in the following years, especially from 2016 onwards, where the number of publications consistently exceeds 10 articles each year. The increase in publications may have been driven by external factors. For example, the increase in publications after 2014 coincided with the increasing global attention to Sustainable Development Goals (SDGs) and energy efficiency commitments. One of them was the Paris Agreement in 2015 [48]. These global and national policies may have motivated researchers to examine ESBW. In addition, the increase in publications is also in line with the increasing changes in energy efficiency technology [49].

2.png

Figure 2. Total publications and average citations per year

Despite the growing number of publications, the impact of ESBW publications fluctuated from 2005 to 2024. In this study, the impact of ESBW publications was seen based on the mean total citations (TC) per year. Figure 2 shows that the highest publication impact occurred in 2013, with the mean TC per year of 9.56. In fact, in 2013, only three articles on ESBW were published. In other words, there were articles in 2013 that had a big impact. Likewise, in 2017, the average TC was 6.76 with 17 articles, indicating that 2017 also produced very influential research. However, the last few years (2021-2024) have shown a decrease in the impact of the situation. There was a decrease in the average TC to 1.56 in 2023, even though the number of articles was high. This is reasonable because the ESBW publications from 2021 to 2024 have not been released for long. Based on the analysis, it is estimated that the number of citations may increase in the future because the citations tend to grow over time, especially considering the already large volume of publications between 2021 and 2024. Moreover, some considerations for ESBW research will increase the number of publications and citations in the coming year: (1) one of the main concerns of the SDGs' 2030 agenda is energy efficiency [50], (2) climate change will have an increasingly negative impact on human life globally if no efforts are made to address it [51-53], and (3) the use of energy-efficient devices is growing [54].

3.1.2 Top publication sources

Referring to previous research, top publication sources were analyzed by volume and citation [41, 42, 44]. The bibliometric analysis demonstrates the publication sources that have significantly contributed to the field of study of ESBW, both in terms of volume and citation impact (Table 2). Journal of Building and Environment stands out as the most prolific source, with 12 publications and a notable citation count of 735, reflecting its central influence on the subject. This journal is published by Elsevier and has been indexed by Scopus since 1976. This journal is the most prolific source because the scope of its discussion is very relevant to the theme of energy saving. Some of the subject areas discussed in this journal are renewable energy, sustainability and the environment, environmental engineering, and civil and structural engineering. In second place, Sustainability (Switzerland) published 11 articles, while Energy Research and Social Science contributed 9 articles. Other important sources include Energy and Buildings and Energies, each publishing 8 articles, and Applied Energy, with 7 articles. In terms of citations, Energy Policy and Applied Energy also demonstrated strong influence, with 345 and 297 citations, respectively. Although these journals, such as Journal of Building Engineering and Energy Efficiency, published fewer articles, they still contributed meaningfully to the discourse, with 129 and 110 citations, respectively. These findings demonstrate that the majority of journals with the highest volume also have the highest citation impact. However, it is not always consistent. For example, the Journal of Resources, Conservation, and Recycling has the second-highest number of citations (611), even though it is ranked 10^th in terms of total publications. Likewise, the Tourism Management Journal received 116 citations even though only 1 article was published. In other words, it demonstrated its significant influence in the research field of ESBW, even though it published fewer articles.

Table 2. Top publication source based on volume and citation

Source: Authors’ elaboration based on Scopus.

3.1.3 Top authors

Top authors were examined by volume and citation, with reference to previous studies [41, 42, 44]. The bibliometric analysis identifies top authors contributing to research on ESBW based on both publication output and citation impact (Table 3). Chen Chien-Fei is the most prolific author with six published articles. She is a professor at the Department of Sociology, Anthropology and Criminal Justice, Clemson University, USA.

Table 3. Top authors based on total publication volume and citation

Source: Authors’ elaboration based on Scopus.

In terms of citation impact, Wang Shanyong stands out as the most influential author, with a total of 87 citations, even though only published two publications. One of the articles written by Wang Shanyong is an article entitled “Application of the extended theory of planned behavior (TPB) to understand individuals’ energy saving behavior in workplaces”. This article has become the most cited article with a total number of citations of 379. This article is the result of Wang S.’s collaboration with his colleagues at the School of Management, University of Science and Technology of China, namely Gao Lan, Li Haidong, and Li Jun, placing the three of them in second place as the most influential authors. Furthermore, Zhang Yixiang, Wang Zhaohua and Zhou Guanghui also showed substantial influence by collecting 376, 369 and 355 citations respectively from their two collaborative publications and other publications. While Chen Chien-Fei, the most prolific author, has received 274 citations, authors such as Dong Bing (240 citations) and Xu Xiaojing (208 citations) have also played an important role in shaping the research landscape.

An interesting aspect of these results is that a higher number of publications by an author does not always correlate with a higher number of citations. For instance, while Chen Chien-Fei and Xu Xiaojing are the most prolific authors with six and five publications each, Wang Shanyong, with only two publications, leads in citations with a total of 387. Likewise, Gao Lan, Li Haidong, Li Jun, who only published one article in collaboration with Wang Shanyong has garnered a higher citation count. This suggests that the citation impact is not strictly tied to the number of publications but rather the significance and visibility of the individual studies.

In the context of this study, Wang Shanyong, Gao Lan, Li Haidong, and Li Jun emerged as the researchers with the highest citations. This dominance was mainly due to their article which had a very large number of citations, surpassing other researchers who may have higher publication volumes. This article is also the most influential article in the context of ESBW research due to its high relevance to the topic. The main contribution of this article lies in the development and extension of the TPB. TPB itself is one of the most established and widely used theoretical frameworks in understanding human intentions and behaviors, especially in environmental psychology and pro-environmental behavior. Furthermore, they empirically showed that the extended TPB model has better explanatory power than the original TPB model. The use of Structural Equation Modeling (SEM) in testing the energy-saving behavior model has also made this article an important reference for other researchers in further studies.

Based on a minimum of three publications, the collaborations among authors are illustrated in Figure 3. While the network displays the number of collaborations, the size of the nodes indicates the number of articles written by authors [44]. Figure 3 indicates that the collaboration network is predominantly composed of authors with names that appear to be of Chinese origin. However, upon further investigation, some of them turned out to be affiliated with institutions or organizations in the United States, such as Chen Chien-Fei, Xu Xiaojing, and Dong Bing. This indicates that the collaboration among authors that occurred was not only national in nature, but also involved international networks across institutions.

Chen Chien-Fei and Xu Xiaojing have the most extensive network-based research cooperation on the study of ESBW. Five of Chen Chien-Fei's six publications were co-authored with Xu Xiaojing. They have also collaborated with Dong Bing, Menassa Carol, Li Da, etc. Among the 24 authors with at least 3 publications, the results reveal 8 different collaborative groups of authors that were separate from each other. Similar article qualities are shared by authors who belong to the same clusters [55]. These collaborative groups are usually formed by the similarity of the authors' disciplinary backgrounds. If mapped, in general, there are 4 disciplines that highlight the topic of energy-saving behavior, namely Economics and Management, Psychology, Sociology and Engineering. Several authors who are included in the collaborative group of the Economics and Management group are Wang Shanyong, who collaborated with Gao, Lan, Li Jun, Li Haidong, then there is Zhang Yixiang, who collaborated with Wang Zhaohua and Zhou Guanghui and Kotsopoulos Dimosthenis, who collaborated with Bardaki Cleopatra. Furthermore, from the Psychology group, there is Endrejat Paul Constantin, who collaborated with Kauffeld Simone because they both have expertise in the field of Industrial/Organizational Psychology. In addition, there is also Xu Xiaojing, a social psychologist who collaborates with Chen Chien-Fei, an environmental sociologist. Likewise, Dong Bing, who has an engineering background, collaborates with Chen Chien-Fei and Xu Xiaojing. So, cross-disciplinary collaboration is needed in this research topic to enrich the perspective.

3.png

Figure 3. Co-authorship network based on authors

3.1.4 Top countries

Based on previous studies, the top countries were identified by common practice in bibliometric analysis [41, 42, 44]. The bibliometric analysis reveals top countries contributing to research on ESBW based on publication output, citation impact, and corresponding authorship (Table 4). In general, the results indicate that the top countries are dominated by developed countries. It might occur because investment in research and development (R&D spending) in developed countries such as China, the UK, and the USA is very high compared with middle-income countries [56]. They also spend huge amounts of money on clean energy, including energy efficiency [57].

Table 4. Top countries based on total publication volume, citation and corresponding

China leads the field with 154 publications, making it the most prolific country in this area of research. It also dominates in terms of TC, accumulating 1,356 citations, highlighting the global influence of its research. In addition, China has the highest number of corresponding authors (39), further emphasizing its leadership role in the field. China leading top position in publications, citations, and corresponding authors can likely be attributed to its rapid growth in R&D spending, significant increase in patent numbers, and overall technological advancement in the new century [58]. Moreover, China has consistently been the world's largest investor in clean energy [57].

Following China, the UK and the USA are notable contributors, with 102 and 84 publications, respectively. However, while the UK ranks second in publication output, the USA surpasses it in citation impact, with 826 TC compared to the UK's 688. Both countries also rank highly in corresponding authorship, with the USA having 19 and the UK 16. Therefore, China, the UK and the USA are the top three, both in terms of publication output, citation impact, and corresponding authorship.

Other countries with notable contributions include Japan, which ranks fourth in both publication count (41) and corresponding authors (9), and Germany with 31 publications and 8 corresponding authors. Australia stands out with 26 publications and a relatively high citation count of 197, indicating a strong impact despite its smaller volume of research. Countries such as Greece, Italy, and Spain also contribute significantly to the research landscape, though their citation impact and corresponding author counts are more modest. Overall, the data indicate that while China, the UK, and the USA are the dominant forces in this area of research, a wide range of countries are contributing valuable insights into ESBW.

Based on a minimum of five publications, the collaborations among countries are depicted in Figure 4. It shows that international collaboration in ESBW research has occurred across multiple countries and even across three continents, namely Asia, America, and Europe. This indicates not only the global interest in the topic but also the breadth of collaboration across diverse geographic regions. More specifically, there are three key countries involved in the research collaborations on ESBW. They are the UK (13 networks), China (13 networks), and the USA (12 networks). Among the 19 countries with at least 3 publications, the results reveal 5 cluster of countries collaboration in this field: Cluster 1 (i.e., Germany, Italy, Netherland, New Zealand, Sweden, and Switzerland), Cluster 2 (i.e., Japan, Malaysia, Pakistan, and South Korea), Cluster 3 (i.e., Greece, Spain, and the UK), Cluster 4 (i.e., China, Hong Kong, and the USA), and Cluster 5 (i.e., Australia and Taiwan). An interesting aspect of these results is that the collaboration in research remains dominated by Western and Asian countries, with little to no involvement from African countries in these research partnerships.

4.png

Figure 4. Co-authorship network based on countries

3.1.5 Top cited articles

The bibliometric analysis highlights the most-cited articles that have significantly contributed to the research of ESBW. The top cited article is the research work by Gao et al. [12] with 379 citations and 47.38 TC per year. It can be a top cited article for ESBW because the number of citations has continued to increase since its publication in 2017 until 2022. The article of Gao et al. [12] was published in Resources, Conservation and Recycling. It applied the TPB to examine the determinants of energy-saving behavior among individuals in workplaces. They found that attitudes, perceived behavioral control, descriptive norms, and personal moral norms play crucial roles in shaping energy-saving intentions, while subjective norms were found to have little effect. The paper by Gao et al. [12] was published in Resources, Conservation and Recycling, a journal covering subject areas in economics, econometrics, finance, and environmental science, with a focus on environmental management. As for citations, the Gao et al. [12] paper was cited by several subject areas, including environmental science, social science, energy, business management and accounting, and engineering. Conversely, the majority of authors citing this paper originated from China, aligning with the country of origin of Gao et al. [12] themselves. Aside from China, other prominent countries of origin for authors citing Gao et al.'s paper [12] include Iran, the United States, Vietnam, and Malaysia. It can be seen that ESBW is relevant to issues in developed or developing countries. Despite being published in 2017, the paper of Gao et al. [12] continues to demonstrate significant impact, as evidenced by the high number of citations it received in more recent years, particularly in 2022 and 2024.

The second most cited article is the study of Zhang et al. [5] with 257 citations published in Energy Policy. It explored the antecedents of employee electricity-saving behavior using the norm activation model (NAM), revealing that personal norms strongly influence behavior, and organizational climate can moderate this relationship. Awareness of consequences and ascription of responsibility were also found to be significant in shaping personal norms. This paper was cited by several subject areas, including accounting, environmental science, social science, energy, and engineering. The distribution of countries of origin of authors who cite Zhang et al. [5] is from the United States, South Korea, Malaysia, and Australia. An analysis of citations by publication year reveals a concentration of citations in recent years, particularly in 2024, 2023, 2022, and 2020. In other words, this suggests the paper of Zhang et al. [5] maintains its relevance within the field, despite having been published in 2013.

Other notable contributions include Hwang et al. [59], which has 157 citations and focused on occupant behavior and thermal comfort in hot, humid climates, and Mustapa et al. [60], which studied thermal comfort and adaptive behavior in university buildings, accumulating 133 citations. Lopes et al. [61] further enriched the research landscape by integrating the TPB, NAM, and human reliability theories to assess industrial workers' energy-saving behaviors, achieving 118 citations.  Additionally, newer studies like Amasyali and El-Gohary [62] used machine learning to predict cooling energy consumption in office buildings, showcasing the impact of occupant behavior on energy use. Further, Li et al. [6] proposed an integrated Motivation-Opportunity-Ability (MOA) framework to investigate energy-saving behaviors in the American workplace. Meanwhile, the more detailed information about the top articles related to top citing subject areas, countries, and years is presented in Table 5.

These results reveal several key findings. First, the top ten most-cited articles predominantly focused on developing and testing models related to ESBW. Several of these studies investigated the determinants factors of ESBW, e.g., [5, 6, 12, 61], while others, such as Hwang et al. [59] and Mustapa et al. [60], focused on thermal comfort and adaptive behavior in workplace settings. Second, all the top ten most-cited articles utilized quantitative methods, except Li et al. [63]. Moreover, the most used quantitative approach is the questionnaire survey, e.g., [5, 6, 12, 61]. Third, the TPB emerges as the most widely applied theoretical framework across these studies, e.g., [6, 12, 61, 64], highlighting its significant contribution to research on ESBW. Finally, all ten of the most-cited articles come from a single publisher, namely Elsevier, except Dong and Lam [64], published by Taylor & Francis Online. In other words, Elsevier has a concentration of high-impact publications in the research area of ESBW.

Table 5. Top articles based on TC

3.2 Science mapping

Science mapping is a descriptive analysis to assess the comprehensive structural connections based on science elements to comprehend important advancements in a specific field of research [42]. Referring to previous studies, such as AlNemer [47], Farooq [66] and Farooq [67], this study conducted science mapping on keyword co-occurrence analysis and thematic mapping. Further, keyword co-occurrence analysis is employed to determine which researchers are the most representative of a given field [44]. In this study, it is used to determine the most keywords discussed in the research of ESBW. On the other side, a thematic map is conducted to identify distinct thematic clusters in the research of ESBW based on the level of development (density) and level of relevance (centrality) [47]. The results of keyword co-occurrence analysis and the thematic map are explained below.

3.2.1 Keyword co-occurrence analysis

The keywords co-occurrence network was built by VOSviewer using a frequency limit of four. It is called keyword co-occurrence when two or more keywords appear in the same article simultaneously [36].  Figure 5 depicts the visualization map result of the co-occurrence of author keywords, which shows: (1) the connection between the nodes symbolizes the co-occurrence between keywords (i.e., keywords that co-occur or occur together); (2) the node's size suggests the keyword's occurrence (i.e., the frequency with which the keyword appears); (3) the link's thickness indicates the occurrence of co-occurrences between keywords (i.e., how many times that the keywords co-occur or occur together); and (4) each color represents a thematic cluster, wherein the keywords are part of the same cluster based on the same color [35]. Therefore, the bigger the node’s size, the bigger the keyword occurrence, while the thicker the connection between nodes, the larger the occurrence of the co-occurrences between keywords [35].

5.png

Figure 5. Visualization map of keywords co-occurrence analysis

Keywords co-occurrence analysis offers details on the keywords that are most frequently discussed [44]. The most appeared keywords happened to be energy saving with 48 occurrences, followed by energy efficiency (26), occupant behavior (18), workplace (9) and energy conservation (10). Further, the results identified four clusters of co-occurrence analysis (Table 6). The first cluster (marked red) is the most prominent cluster that focuses on energy-saving behavior within commercial and office buildings. Cluster 1 highlights the importance of behavioral interventions (e.g., TPB) to foster energy-efficient practices. Several previous studies have examined this topic, including Gao et al. [12], Lopes et al. [61], and Obaidellah et al. [68]. Thermal comfort also emerges as a key factor in energy-saving behavior, suggesting that balancing comfort and efficiency is central to behavior modification in workplaces.

Table 6. The theme clusters of keywords co-occurrence analysis

Cluster 2 (marked green) suggests a focus on motivational and social approaches to energy-saving behavior. The use of feedback mechanisms and techniques like motivational interviewing seems central to driving behavior change. Social factors such as social norms and sustainability consciousness play a crucial role in promoting pro-environmental behaviors. In addition, cluster 3 (marked blue) connects the role of employee behavior and workplace habits in influencing overall energy conservation. The focus here is on the habitual nature of energy use at work, implying that long-term behavioral changes are required to reduce energy consumption. Cluster 4 (marked yellow) looks at energy consumption in relation to occupant behavior, which emphasizes the role of individuals within a building in driving or reducing energy use. This final cluster focuses on energy efficiency implies a more technical or architectural perspective in tandem with behavior.

Furthermore, there are at least 96 keywords that occurred fewer than five times but have moderate to low, or even high, total link strength (TLS). Several factors may explain this pattern: the keywords might represent niche or emerging concepts, redundant or variant terms, non-standard or context-specific expressions, or semantic outliers [69-71]. For instance, keywords like “gamification”, “motivational interviewing”, and “norm activation model” are well-defined but narrowly used, suggesting they may be emerging concepts that have not yet gained broader traction. Some low-frequency yet high-TLS keywords might also be the result of redundant or variant terms, such as differences in spelling (e.g., “behavior” vs. “behaviour”), that fragment their representation. Non-standard or context-specific terms, like “building energy management”, may also appear with limited connections to the dominant themes due to their specificity. Lastly, some keywords may function as semantic outliers, indicating highly specialized or novel tools, such as “spatiotemporal analysis”.

3.2.2 Thematic map

The occurrence of keywords was mapped out based on density (y axis) and centrality (x axis) using RStudio (Figure 6). The measures of density and centrality allow the visualization of themes and their relationships in a two-dimensional map to analyze the importance and magnitude of research on the theme of investigation [47]. The centrality measures the importance of the selected theme based on the degree of correlation amongst different topics, while density measures the level of research that has already been undertaken on the chosen theme [47]. Based on the centrality and density measures, the research keywords have been plotted into four quadrants, namely (1) motor themes, (2) basic themes, (3) niche themes, and (4) emerging or disappearing themes [66, 67].

6.png

Figure 6. Thematic map based on density and centrality