Research Trends and Patterns on Emotional Intelligence in Education: A Bibliometric and Knowledge Mapping During 2012–2021

2.1 Searching and Retrieving Data

On 2 July 2022, we undertook a comprehensive search of the Scopus database to assemble scholarly contributions addressing emotional intelligence in the field of education, spanning the years 2012 to 2021. Our search criteria targeted the title and abstract fields with the following terms: “emotional intelligence” AND (“teaching” OR “education”). The initial search yielded 2,786 results. In the second stage, the search was refined to encompass only those articles in the terminal stages of publication, and limited by publication type to journal articles, conference papers, and reviews. Additionally, we restricted our inquiry to works written in English. This search strategy yielded an initial pool of 1,810 entries. In the third stage, we cleaned the data removing incurrent and incomplete entries leaving us with 1,732 publication. In the final stage, we screened the titles and abstract for their relevance to the topic of investigation. Post-screening, a total of 1,696 articles met the requisites for inclusion and analysis in this review.

From this literature, we extracted bibliographic details, which encompassed: authorship, institutional affiliations, subject categorizations, journal designations, enlisted keywords, article titles, and publication chronology. Adhering to the stringent guidelines laid out in the PRISMA declaration (Preferred Reporting Items for Systematic reviews and Meta-Analyses, Moher, Liberati, Tetzlaff, & Altman, 2009), we meticulously sieved through the initial dataset. The procedural delineation of our selection regimen is visually represented in the PRISMA flowchart, depicted in Figure 1.

Figure 1

PRISMA flow diagram showing the flow of the search in the identification and screening of sources for the bibliometric analysis of research on emotional intelligence in education on emotional intelligence in education (PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses).

2.2 Data Analysis

Prior to analysis, the dataset extracted from the Scopus database, comprising 1,696 entries, was exported to a comma-separated values (.csv) format, priming it for further scrutiny. Our analysis relied on two pivotal software tools. First, VOSviewer (www.vosviewer.com) was utilized for knowledge mapping, specifically network analysis. This approach spotlighted the “link strength” amongst author keywords, countries, institutions, and researchers. Network analysis is indispensable for delineating a discipline’s contours and structural nuances and for pinpointing salient research clusters (Donthu, Kumar, Mukherjee, Pandey, & Lim, 2021). The visual representation of these bibliometric networks captures the nodes’ proximity, serving as an indicator of association intensity (Fabrikant, Montello, & Mark, 2010). Concurrently, Microsoft Excel facilitated the examination of articles and citations, with findings presented through intuitive charts and tables.

3.1 Trend in Publication and Citation

The inaugural objective of our research centred on discerning both the publication and citation patterns inherent in scholarship concerning emotional intelligence. Figure 2 offers a visual representation of the temporal trends in publication volume and citations spanning the years 2012 through 2021. This graphical elucidation highlights a consistent uptick in publications from 2012 onward. The years 2019 and 2020 mark pronounced inflection points with the most substantial rise, while 2021 witnessed a decelerated growth trajectory. While a myriad of bibliometric studies suggests an overarching geometric escalation in publication counts (Hallinger & Kovačević, 2019; Moosa & Shareefa, 2020), our observations pinpoint a more measured, linear growth specific to emotional intelligence scholarship. These findings underscore a sustained academic intrigue in the domain. Bibliometric data from other sources can be used to verify if this declined rate is a consequence of our limitation in data collection to a single database.

Figure 2

Pattern of publications and citations.

In our exploration of normalized citation patterns, we found the average number of citations per article, annually, to maintain a steady trend. Contrary to the progressive increase observed in publication volume, citation patterns remained relatively flat, averaging roughly two citations per article. The mean citation count identified in our study closely aligns with figures presented in preceding bibliometric research, wherein normalized citation counts typically range between 1 and 10 (Moosa, Khalid, & Mohamed, 2022; Moosa & Shareefa, 2020). The consistent and nearly horizontal trajectory evident in our normalized citation graph is striking. It underscores the persistent reference and relevance of literature published on the subject, indicating a sustained interest and engagement within the scholarly community.

3.2 Geographic Distribution of Publications

To address our second research objective, we examined the global distribution of the 1,696 publications within our dataset. This geographic breakdown is visualized in Figure 3, generated using Microsoft Excel. For clarity and relevance, only countries contributing a minimum of three documents were incorporated into this analysis. Out of the 115 countries that published on the subject, 69 met this threshold and are thus represented in the map showcased in Figure 3.

Figure 3

Global contribution to publications on emotional intelligence.

Figure 3 and its accompanying table reveal that the United States (438 documents), Spain (219), and the United Kingdom (138) are the primary contributors to the body of literature on emotional intelligence. Each of these countries has produced at least 100 publications, with the US contributing over double the number of the second-highest contributor, Spain. Consistent with numerous bibliometric studies, Western nations are prominent in the volume of publications (Hallinger & Kulophas, 2020; Hallinger & Nguyen, 2020; Punnakitikashem & Hallinger, 2020).

However, a couple of noteworthy distinctions emerge. First, Spain’s significant contribution is remarkable. Although a European nation, its contribution typically does not parallel the dominance of leading contributors. Second, the presence of non-Western nations, including those from Asia and the Middle East, among the top ten is noteworthy. An extended analysis of the 69 countries, beyond the top ten contributors, further amplifies the substantial input from other non-Western nations.

3.3 Patterns of Collaboration Towards Publications on Emotional Intelligence

The third objective of our research aimed to uncover the collaborative dynamics among nations and authors that are actively publishing on the subject of emotional intelligence. To this end, we employed co-authorship analysis in VOSviewer, examining each of these units individually.

As illustrated in Figure 4, a network of collaboration is evident among the nations contributing to the emotional intelligence scholarship. For this analysis, we set a baseline threshold of a minimum of two documents per country. Using the full counting method, 74 countries met this criterion. Nonetheless, several countries, specifically Iraq, Romania, Sri Lanka, Oman, and Grenada, were found to have no collaborative links with others and are consequently omitted from Figure 4. Within the figure, the dimension of the nodes represents the magnitude of collaboration, with larger nodes signifying stronger collaborative ties.

Figure 4

Collaboration between countries towards publications on emotional intelligence. https://tinyurl.com/23kgal4w.

As delineated in Figure 4, the top five nations in terms of collaboration are the United States (with a collaboration strength of 96), the United Kingdom (89), Spain (51), Australia (42), and China (38). Though the sequence varies, there’s a discernible correlation between collaboration strength and the volume of publications for each country. However, this relationship becomes less defined when extended to the top-ten contributors. Notably, countries like the Netherlands, Ireland, Hong Kong, and Italy feature prominently among the top collaborators but are not ranked as top publishers. While the association between collaboration magnitude and publication volume remains debatable in bibliometric analyses, numerous studies have posited that collaboration potentially facilitates increased publication outputs (e.g. Bozkurt, Eryilmaz, & Boyraz, 2021). Our data suggest that while there might not be a direct proportional relationship between collaboration extent and publication volume, leading collaborators are typically among the top publishers in the domain of emotional intelligence.

Furthermore, apart from the extent, we delved into the character of these collaborations. This was discerned by examining the clusters in Figure 4. Specifically, we identified pronounced collaborative ties among Asian nations, countries in proximity to the Balkans, and those in Europe. Those with significant collaboration footprints were found to collaborate with a varied group of nations and were part of comparatively smaller clusters. For example, the United States, the foremost collaborator, is part of a five-country cluster, whereas the United Kingdom, the second most significant collaborator, is associated with just three countries. This could intimate a predilection for diverse collaborative networks over insular, limited groupings.

Progressing in our analysis, we extended our exploration to understand the collaborative dynamics among institutions affiliated with scholarship on SDG pertaining to education and climate. Figure 5 illuminates the collaboration network among authors who have contributed to the discourse on emotional intelligence. For this phase, we established a threshold of at least four documents per author, leveraging the full counting method. Fifty-one authors satisfied this criterion, and all are represented in Figure 5. Within the figure, node sizes are indicative of the number of affiliated documents.

Figure 5

Collaboration between authors towards publications on emotional intelligence. https://tinyurl.com/27sj87qb.

As depicted in Figure 5, collaborations amongst authors appear notably limited, predominantly constrained to considerably diminutive clusters. The most substantial cluster, highlighted in red, encompasses four authors, each exhibiting a collaborative strength ranging between 4 and 5. Intriguingly, their contributions to publications mirror their collaborative strengths, falling between 4 and 5 articles. A number of publications from these authors are concentrated on research on adolescents and those who interact with them.

Successively, groups with three authors form the second largest clusters. These include the clusters in green and gold. Unlike the red cluster, the collaboration strength of each author in these clusters is eight while the number of publications they contribute is four. Interestingly, the link strength between any two authors in this cluster is also four. This indicates that these authors have close collaboration within themselves but not with others outside their circle, particularly with respect to the publications associated with the analysis in Figure 5.

Notably, the principal cluster among those with three collaborators, denoted in blue, interlinks with another cluster, represented in purple. A discernible pattern emerges from this observation, where several authors, apart from being eminent collaborators, are also paramount contributors to the academic discourse on the subject. To illustrate, Fernandez-Berrocal, P., possessing a collaborative strength of 14, has contributed to a commendable 12 publications (e.g., Gutiérrez-Cobo et al., 2023). Closely following is Cabello, R, with a collaboration intensity of 14, and an academic contribution spanning nine publications (e.g. Vega, Cabello, Megías-Robles, Gómez-Leal, & Fernández-Berrocal, 2022). Authors in the blue cluster seem to be engaged in a wide range of studies including those on mental health, adolescents, teachers, as well as parents.

While these patterns offer compelling insights, it is imperative to recognize that the magnitude of collaboration does not invariably align with the volume of publications. A case in point is Dolev, N, who, despite a substantive contribution of seven publications, exhibits no discernible collaboration within the presented network. This incongruity extends to other scholars as well. Stoller J K, Zysberg L, Hen M, Petrides K V, and Taylor D C, each of whom has produced an impressive four publications, also do not manifest significant collaborative ties in the presented network.

3.4 The Intellectual Structure of Knowledgebase on Emotional Intelligence

To delineate the intellectual structure underpinning the knowledge base on emotional intelligence, we employed co-citation analysis. This technique measures the frequency with which two publications are jointly cited within a third publication, thereby identifying co-cited literature (Small, 1973; Zupic & Čater, 2015). Undertaking a co-citation analysis founded on cited references necessitated certain parameters. Specifically, a source had to possess a minimum of 10 citations. Furthermore, we applied a thesaurus file to address inconsistencies or variations in data entries. The resultant visualization, as showcased in Figure 6 and rendered using VOSviewer, comprises 40 pertinent documents meeting the established criteria. Within this visual representation, the size of each node corresponds to the cumulative link strength of a particular source.

Figure 6

Co-citation network of cites references association with publications on emotional intelligence. https://tinyurl.com/23s8gay8.

As revealed by Figure 6, the literature on emotional intelligence can be segmented into six discernible schools of thought. For the sake of succinctness, we will expound on the three most prominent ones.

Cluster 1 (in red) predominantly features co-cited documents such as “Development and validation of a measure of emotional intelligence” (Schutte et al., 1998), “Trait emotional intelligence: Psychometric investigation with reference to established trait taxonomies” (Petrides & Furnham, 2001), and “The location of trait emotional intelligence in personality factor space” (Petrides, Pita, & Kokkinaki, 2007). The central intellectual thrust of this cluster appears to be on EI quantification. Notably, Schutte et al. (1998) embarked on a seminal exploration to devise and validate an EI measure. Furthermore, Petrides and Furnham (2001) shed light on the nexus between trait emotional intelligence and existing trait taxonomies. This, in tandem with other empirical studies, underscores the evolution and validation of EI measures (Brackett & Mayer, 2003; Fernandez-Berrocal, Extremera, & Ramos, 2004).

Cluster 2 (in green) is chiefly characterized by highly co-cited documents such as “Emotional intelligence” (Salovey & Mayer, 1990), “Human abilities: Emotional intelligence” (Mayer, Roberts, & Barsade, 2008), and “Trait emotional intelligence of Greek special education teachers in relation to burnout and job satisfaction” (Platsidou, 2010). Core themes of this cluster encapsulate the multifarious facets of EI and their ensuing ramifications. Salovey and Mayer (1990) delineate the seminal conception of emotional intelligence, framing it as an adeptness in perceiving, interpreting, and modulating emotions. The seminal works of Mayer et al. (2008) furnish an exhaustive analysis of human competencies, spotlighting EI. Moreover, studies accentuating educators unveil EI’s pivotal role in educators’ holistic well-being and pedagogical outcomes (Brackett, Palomera, Mojsa-Kaja, Reyes, & Salovy, 2010; Brackett, Rivers, & Salovey, 2011; Jennings & Greenberg, 2009).

Cluster 3, meanwhile, predominantly spotlights co-cited articles like “The Bar-On model of emotional-social intelligence” (Bar-On, 2006) and “The ability model of emotional intelligence: principles and updates” (Mayer et al., 2016). The canon within this cluster pivots around the foundational paradigms of emotional intelligence. Bar-On’s (2006) treatise introduces the Bar-On model of emotional-social intelligence (ESI), elucidating EI as a spectrum of interwoven emotional and social competencies that dictate holistic well-being and life success. Mayer et al. (2016) advance the ability model of emotional intelligence, accentuating aptitudes in discerning, employing, decoding, and navigating emotions adeptly. These frameworks enrich our comprehension of EI’s intricate architecture and its pervasive implications across life domains.

3.5 The Conceptual Structure of the Knowledgebase on Emotional Intelligence

Our final research objective focused on delineating the prevalent themes embedded within the keywords of publications on emotional intelligence. For this keyword analysis, we considered only author-provided keywords that manifested at least five times, employing a thesaurus file for consistency. Given that “emotional intelligence” was our primary search term and anticipated to have a higher frequency, we chose to omit this keyword from the analysis. To refine our analysis further, clusters comprising fewer than five keywords were consolidated into larger ones.

Figure 7 visualizes the network crafted by VOSviewer for keyword analysis. Evidently, the conceptual structure of the scholarship on emotional intelligence encompasses diverse themes. Given space constraints, we will limit our discussion to the five largest clusters. Additionally, Table 1 delineates the five most recurring keywords within each cluster.

Figure 7

Co-occurrence network of keywords. https://tinyurl.com/24fwucdx.

From Figure 7 and Table 1, it is apparent that in cluster 1 (coloured in red), the keyword “trait emotional intelligence” emerges most frequently (31 occurrences), trailed by ‘intelligence’ (16 times), and ‘emotion’ (15 times). Cumulatively, with other salient keywords, this cluster’s overarching theme resonates with facets of intelligence, emotion, and soft skills. Santos, Mustafa, and Gwi (2015) delved into the impact of trait emotional intelligence on the emotional labour strategies of customer service agents. Parallelly, Shrivastava, Martinez, Coletti, and Fornari (2022) underscored the synergy of emotional intelligence and soft skills in the realm of leadership development. Providing a broader perspective, Grobelny, Radke, and Maczka (2021) accentuated the cardinal role of trait emotional intelligence in gauging job performance across a plethora of professions. These empirical works jointly underscore the pivotal role of trait emotional intelligence in grasping emotions, bolstering skills, and fostering genuine emotional expressions in varied professional environments.

Cluster 2 (Green): The predominant keywords include “adolescence” (29 times), “mental health” (20), and “secondary education” (16). This cluster primarily underscores the mental health and socio-emotional regulation of adolescents. Notably, Yeager (2017) analysed the impact of SEL programs in secondary schools on adolescents, while Verzeletti, Zammuner, Galli, and Agnoli (2016) delved into emotion regulation strategies during adolescence. Sande et al. (2019) further accentuated the effectiveness of secondary school SEL programs in honing target competencies.

Cluster 3 (Blue): Key terms here are “leadership” (65 instances), “nursing students” (39), and “resilience” (36). This theme mistakenly repeats the same focus as cluster 2; however, it should centre on leadership qualities and resilience, particularly among nursing students. Hughes, Cologer, Swoboda, and Rushton (2021) discussed resilience enhancement among nursing students, while Benson, Martin, Ploeg, and Wessel (2012) established correlations between EI, leadership, and caregiving in nursing undergraduates.

Cluster 4 (Olive Green): Central keywords comprise “students” (65 times), “academic performance” (39), and “adolescents” (36). This cluster zooms in on the influence of EI on the holistic development of adolescent students, covering cognitive, psychomotor, and affective domains. Carlson et al. (2008) revealed the positive impact of physical education on academic performance, whereas Kayani et al. (2021) showcased how self-concept mediates between physical activity and anxiety among university students.

Cluster 5 (Purple): The prominent terms are “well-being” (20 occurrences), “motivation” (17), and “self-awareness” (15). Here, the spotlight is on psychological nuances like self-awareness, self-esteem, and self-regulation in tertiary education students. Hofer, Busch, and Kärtner (2011) outlined the synergy between self-regulation, well-being, and identity motives in students. Similarly, Ardianto and Etlidawati (2022) emphasized the positive link between self-awareness and well-being in nursing student behaviour.

4.1 Implications

Beginning in 2012, there has been a marked escalation in the attention given to emotional intelligence, underlining the burgeoning recognition by scholars and educators of its profound implications within the educational arena. Such awareness extends beyond mere academic pursuits, encompassing holistic student outcomes. By harnessing the tenets of emotional intelligence – like heightened self-awareness and augmented empathy – students stand to benefit immensely in interpersonal facets, namely enhanced communication, efficacious teamwork, and adept conflict resolution. Consequently, educators are presented with a potent tool to amplify both scholastic achievement and holistic personal development by integrating emotional intelligence principles within their pedagogical approaches. Moreover, the diversified research interest spanning Western and non-Western nations underscores a global momentum to understand and leverage emotional intelligence in education, unbounded by cultural or geographical confines.

Our co-authorship examination accentuates the pivotal role of interdisciplinary synergies in crafting a nuanced, comprehensive perspective on emotional intelligence within educational settings. Notably, the data suggest that expansive collaborations often correlate with enriched publication output, fostering dissemination of knowledge that captures the richness of diverse cultural interpretations and applications of emotional intelligence in education. This extensive collaborative framework facilitated a more detailed grasp of the intellectual landscape, as well as thematic pillars underscoring emotional intelligence in pedagogical contexts. The derived knowledge themes and keyword analyses spotlighted the multifaceted significance of emotional intelligence, touching upon critical aspects such as mental well-being, socio-emotional modulation during adolescence, leadership acumen among nursing students, and the tangible impact on student achievements spanning cognitive to affective realms. Such findings unequivocally cement the centrality of emotional intelligence in fortifying overall wellness, academic accomplishments, and both personal and professional growth within the educational sector.

4.2 Limitations

The present investigation relied on data from a single bibliometric data source, specifically SCOPUS. Integration of data from other bibliometric data sources could enhance the vignette of the research landscape on the subject of emotional intelligence.

4.3 Future Research Directions

Based on the findings of the current bibliometric investigation, we proposed the following directions for future research.

With respect to geographical disparities in research output, contribution from Africa and some regions of South America needs to be increased. Collaboration with researchers from these regions is required to enhance knowledge of emotional intelligence at a global level.

As for potential themes of research, investigations on understanding the measurement of emotional intelligence with respect to specific behavioural characteristics, focus on students and youth, attention to particular disciplines such as nursing, teaching, hospitality, etc. seem to be worth further explorations. Moreover, investigations integrating emotional intelligence with various psychological factors such as well-being, self-esteem, and self-regulation are a potential domain for future research.

With regard to the limitations of bibliometric data employed in this study, we suggest future research on the topic engaging bibliometric methodology could use a wider range of data. Additionally, as bibliometric research is limited in generating in-depth knowledge on a specific theme, we also recommend that future research could engage techniques of systematic literature reviews including meta-analysis and meta-synthesis in order to delve into the details of specific domains of knowledge.