Emotional Intelligence, Transformational Leadership, and Team Effectiveness: A Systematic Review and Correlational Meta-Analysis

Abstract

Emotional intelligence (EI) and transformational leadership (TL) have been identified as key factors in team effectiveness (TE); however, the empirical evidence remains fragmented and exhibits substantial conceptual and methodological heterogeneity, particularly in studies that simultaneously integrate these three variables. To address this gap, the present study examined the relationships among TL, EI, and TE by applying the PRISMA 2020 protocol and the PICO-S framework. A total of 728 studies published in Scopus, Web of Science, ScienceDirect, Emerald, ProQuest, and APA PsycNet were identified, of which 22 studies were included in the systematic review and 15 documents in the meta-analysis. The results revealed positive and statistically significant correlations between TL–TE (9 studies, 18 effects, N = 3480; r ≈ 0.45), EI–TE (8 studies, 15 effects, N = 3440; r ≈ 0.41), and EI–TL (4 studies, 6 effects, N = 1955; r ≈ 0.63), with effect sizes and levels of heterogeneity ranging from moderate to high. Additionally, variations in the strength of these relationships were observed according to sample size, year of publication, and methodological quality. In conclusion, EI emerges as a central resource that strengthens TL and, through psychological and relational mechanisms, consistently enhances TE in complex organizational contexts.

1. Introduction

Teamwork is a central component in organizations, as it facilitates the coordinated execution of projects and strategies through collaboration among members (Al et al., 2021). In this context, several studies have indicated that variables such as TL and EI are associated with different indicators of TE; however, the nature and magnitude of these relationships vary according to organizational context and the methodological designs employed. Empirical evidence suggests that EI contributes to TL and TE by fostering cohesion, motivation, and emotional management in complex organizational environments (Coronado-Maldonado & Benítez-Márquez, 2023; Park et al., 2024). Specifically, some studies have reported positive associations between EI and TE (Eslava Zapata et al., 2022; Karve et al., 2025), while others have examined the relationship between EI and TL, highlighting its influence on adaptability to change and the resolution of social challenges within teams (Abdallah & Mostafa, 2021; Waglay et al., 2020). Likewise, previous research has shown that TL is associated with higher levels of group performance, communication, and interpersonal coordination (Barczak et al., 2010; Sarkar & Oberoi, 2020). In this regard, Murmu and Neelam (2022) identified emotional intelligence as a mediating variable between personality traits and effectiveness in virtual teams, reinforcing the need to examine these variables in an integrated manner.

From a conceptual perspective, EI has been defined as the ability to perceive, understand, regulate, and use one’s own emotions and those of others, thereby promoting individual and collective well-being (Salovey & Mayer, 1990). This concept was later expanded by Goleman (1995) to include dimensions such as self-regulation, empathy, and social skills. The literature generally distinguishes two main approaches: the ability model, which conceptualizes EI as a cognitive capacity, and the mixed model, which integrates emotional skills with personality traits and individual attributes (Suleman et al., 2020). TL, in turn, is characterized by generating significant change through inspiration, empowerment, and the development of human capital, and is structured around four core dimensions: idealized influence, inspirational motivation, intellectual stimulation, and individualized consideration (Bass & Riggio, 2005). Meanwhile, TE is understood as the collective ability to achieve objectives, maintain functional relationships, ensure member satisfaction, and adapt to dynamic and changing environments (Endriulaitienė & Cirtautienė, 2021; Kim & Ko, 2021; Ladino, 2025).

In recent years, systematic and bibliographic reviews have examined these variables in a partial manner. For instance, Coronado-Maldonado and Benítez-Márquez (2023) reported that emotionally intelligent leadership can improve team performance and attitudes, while Karve et al. (2025) emphasized that the development of emotional intelligence at both individual and collective levels contributes to more cohesive teams. Similarly, Koutsioumpa (2023) and Madamang et al. (2023) documented the relationship between emotional intelligence and leadership, as well as its mediating effect on group cohesion across different organizational contexts. However, these reviews have primarily focused on pairs of variables and present methodological heterogeneity in their findings. Although it is widely acknowledged that TL and EI influence TE, uncertainty remains regarding how these three variables jointly interact within real organizational contexts. The existing literature reports heterogeneous findings and employs diverse methodological approaches, with relatively few studies applying robust quantitative techniques such as correlational meta-analysis; to date, only five studies have been identified that simultaneously examine TL, EI, and TE (Ahmed et al., 2014; Hur et al., 2011; Mysirlaki & Paraskeva, 2019, 2020; Polychroniou, 2009). This limited body of evidence reveals a significant gap in understanding how the interaction among these variables can enhance both organizations’ sustainable productivity and the comprehensive development of their members. Therefore, adopting an integrative approach that considers TL, EI, and TE is essential, not only to understand how teams achieve organizational objectives but also to assess how these dynamics strengthen team members by promoting a balance between performance, effectiveness, and workplace well-being.

Despite the existence of empirical studies that simultaneously address TL, EI, and TE, to the best of the authors’ knowledge, no systematic review has been identified that jointly integrates these three variables, nor a correlational meta-analysis that quantifies the magnitude of their relationships. To address this gap, the present study aimed to examine the relationship between TL, EI, and TE through a systematic review and a correlational meta-analysis. The specific objectives were: (SO1) to identify the types of correlations and methodological designs used; (SO2) to determine the control and additional variables considered in the relationship between EI, TL, and TE; (SO3) to analyze the measurement instruments employed and their reliability; (SO4) to evaluate the existence and magnitude of statistical relationships between EI, TL, and TE; and (SO5) to examine whether these relationships vary according to predominant sex, sample size, year of publication, geographical location, level of analysis, and methodological quality.

2. Materials and Methods

2.1. Methodological Design, Criteria, Questions, and Objectives

This systematic review and meta-analysis were based on the definitions of Siddaway et al. (2019) and Gurevitch et al. (2018), following the PRISMA 2020 protocol (Page et al., 2021) and methodological recommendations (Chapman, 2021; Clark et al., 2020; Shaheen et al., 2023). This approach is widely recommended when primary studies report associations between psychosocial and organizational constructs measured through observational or survey-based designs, in which experimental manipulation is not feasible (Borenstein et al., 2009). The present study was registered in PROSPERO (ID: 1062056) https://www.crd.york.ac.uk/PROSPERO/register last edited on 23 September 2025. The updates recorded in 2025 correspond to editorial and methodological refinements (clarification of search terms and wording), without modifications to the eligibility cri-teria, study variables, or outcomes, thereby ensuring temporal and methodological consistency with the 2024 editorial volume. The PRISM checklist, available in Supplementary Materials was generated through https://prisma.shinyapps.io/checklist/, accessed on 25 October 2025. The PICO-S (Population, Intervention, Comparison, Results, and Design) framework was used, omitting comparison because it was an observational study (Ansari & Iqbal, 2025), which facilitated the formulation of research questions (Abu-Ras et al., 2025; Balalle & Pannilage, 2025; Trică et al., 2024) and eligibility criteria (Hernández-Torrano et al., 2025). According to Van der Wal and Kok (2019), differentiated inclusion and exclusion criteria were applied: the systematic review included studies that addressed the study variables, while the meta-analysis required correlational data (

Table 1. Eligibility criteria based on the PICOS framework.

Ac.	Keywords	Inclusion	Exclusion
Population	Work teams	Studies that analyzed work teams in different organizational contexts, sizes, geographical locations, or types of organization were included	Studies that did not assess teamwork dynamics included participants with less than six months in the organization or involved teams without a clearly defined organizational structure.
Intervention	EI and TL	Research that evaluated the variables of TL and/or EI in TE was considered.	Studies that did not examine the association of EI and TL with TE.
Outcome	Team effectiveness (TE)	Studies were selected that reported correlation coefficients between TL and/or EI and TE	Studies without a representative sample (<30 participants) or that did not report correlation coefficients.
Design	Observational studies	Observational studies published in peer-reviewed scientific journals were included, without restriction of language or year of publication, and with access to the full text.	Qualitative studies, theoretical papers, systematic reviews, technical reports, unpublished theses, opinion articles, and experimental or qualitative studies.

).

2.2. Search Strategy: Databases, Terms, and Periods

Relevant studies were identified in six academic sources, including Scopus and Web of Science, which were considered the primary databases (Pranckutė, 2021), as well as ScienceDirect, Emerald, ProQuest, and APA PsycNet. Articles published up to February 2025 were included, without a time restriction. The search terms were defined based on recognized thesauri (Emtree, UNESCO, and ERIC), and the search strings were adapted to each database.

• Scopus: TITLE-ABS-KEY (“emotional intelligence) AND (“transformational leadership” OR “leadership styles”) AND (“team effectiveness”)
• Web of Science: TS = (“emotional intelligence” OR “emotional competence” OR “emotional quotient”) AND TS = (“transformational leadership” OR “leadership styles” OR “charismatic leadership”) AND TS = (“team effectiveness” OR “team performance” OR “work team efficiency”) AND TS = (“work teams” OR “group dynamics” OR “collaborative groups”)
• Science direct: (“emotional intelligence” OR “emotional competence” OR “emotional quotient”) AND (“transformational leadership” OR “leadership styles” OR “charismatic leadership”) AND (“team effectiveness” OR “team performance” OR “work team efficiency”) AND (“work teams” OR “group dynamics” OR “collaborative groups”)
• Others (Emerald, ProQuest, APA PsycNet): “emotional intelligence” AND “transformational leadership” OR “leadership styles” AND “team effectiveness” OR “work team efficiency.”

The search strategy was conducted in two phases. The first phase took place between 15 August and 30 September 2024, using advanced search strategies in the Scopus, Web of Science, and ScienceDirect databases. The second phase was conducted between 5 January and 28 February 2025, incorporating the “Others” section, which included the Emerald, ProQuest, and APA PsycNet databases. To ensure reproducibility, database-specific search strings were applied, using fields such as TITLE-ABS-KEY, Boolean operators (AND, OR), wildcards, and filters by language (English and Spanish), document type (peer-reviewed journal articles), and clearly defined time ranges for each search execution. During the second search phase, stricter inclusion criteria were applied. Only studies that explicitly addressed the variables of interest and analyzed correlational relationships among them were included. Documents without empirical analysis, those that did not report statistical relationships between variables, or those classified as reviews, editorials, book chapters, or grey literature were excluded. A total of 728 records were identified through the search process. Of these, 473 were retrieved from Scopus (469 in the first phase and 4 in the second), 48 from Web of Science (19 in the first phase and 29 in the second), 185 from ScienceDirect (150 in the first phase and 35 in the second), and 22 from other sources identified exclusively during the second phase. Record management and duplicate removal were performed using Mendeley Desktop (version 1.19.8) and Parsifal (version 2.2.0). Records were compared across fields such as title, authors, year of publication, and DOI. In cases of duplication, the most complete record with the highest level of indexation was retained.

2.3. Process of Selecting and Evaluating Studies

The articles selected according to the inclusion and exclusion criteria of the PICOS framework (Schardt et al., 2007) were exported in BibTeX format to the Parsifal platform (https://parsif.al/) to facilitate the systematic review process (Harrison et al., 2020; Khalil et al., 2022). The selection process was conducted in two phases: first, by screening titles and abstracts, and subsequently through full-text analysis (Van der Wal & Kok, 2019). Methodological quality was assessed using the checklist developed by Petticrew and Roberts (2006), later applied to social science research by Chan and Lee (2021); Liu et al. (2025), which consists of eleven items across five dimensions (

Table 2. Checklist Applied.

Dimensions	No.	Criteria
Research question	P01	Is the research question clear?
	P02	Has the appropriate research method been chosen to answer the research question?
Research sample	P03	Was enough research data collected to ensure the validity of the results?
	P04	Has the research context (participants) been clearly replicated?
Method	P05	Has the author provided a detailed description of the research methodology?
	P06	Has the author explained the reason for the choice of the research method?
	P07	Have other potentially influential variables been taken into account?
	P08	Did the author provide information on the validity and reliability of the research?
Data analysis	P09	Are the data analysed appropriate?
	P10	Are the results presented clearly?
Conclusion	P11	How clear are the links between data, results, and interpretation?

Note: Instrument constructed by Petticrew and Roberts (2006).

). The evaluation procedure and scoring scale used by these authors were replicated, and no psychometric validation, cultural adaptation, or methodological modification was performed, as the sole objective was to standardize the assessment of the methodological quality of the included studies rather than to measure constructs in a specific population. Quality assessment was conducted independently by three reviewers (JC, ZM, JH) under the supervision of MPS. A simple majority decision rule was applied, whereby the rating assigned by at least two of the three reviewers was considered the final score for each article, using a trichotomous ordinal scale of Good (1), Fair (0.5), and Poor (0). This metric allows for the capture of gradations in methodological rigor. The overall quality score was obtained by summing all items, without differential weighting by dimension, and only studies with a minimum score greater than 5.5, equivalent to 50% of the maximum possible score, were included. Inter-rater reliability was verified using Fleiss’ Kappa coefficient (Fleiss, 1971), which is appropriate for estimating agreement among three or more raters (M. Li et al., 2023; Kamso et al., 2023). Methodological quality was assessed using the checklist developed by Petticrew and Roberts (2006), later applied to social science research by Chan and Lee (2021) and Liu et al. (2025). The checklist comprises eleven items grouped into five dimensions (Table 2).

2.4. Data Extraction and Coding Process

Data extraction was carried out independently considering the following items: (1) Identification of the study (title, authors, year, journal and DOI); (2) Theoretical foundation and variables (mentioned theories, primary relationship, additional and control variables); (3) Methodological characteristics, (level of analysis (group or individual), composition of the teams, context, country and type of design); (4) Instruments, used to evaluate TE, TL and EI; (5) Sample characteristics, (type of sample, size, number of teams, inclusion and exclusion criteria, distribution by sex, average age, and other relevant data); (6) Statistical data, (Type of correlation, p value, main results by relationship, and correlations by dimension when a general correlation was not reported); (7) Analysis techniques (Applied statistical methods); and (8) Content synthesis, (summary of conclusions, implications, limitations, and suggestions).

The coding refers to: Sex, according to the predominance in the sample (1 = most women, 2 = most men); Sample size, corresponds to the total number of participants reported by the study; Continent, indicates the geographical location (1 = Asia, 2 = America, 3 = Europe, 4 = Africa, 5 = Oceania, 6 = Virtual); year of publication, was coded as 1 = before 2018 and 2 = after 2018; Level of analysis, reflects whether the relationship was evaluated at the individual (1) or group level (2); and Methodological quality, was rated from 1 to 4 (1 = bad, 2 = regular, 3 = good, 4 = excellent).

2.5. Procedure for Data Analysis and Synthesis

A 2-phase synthesis was employed: a systematic review to provide an overview of empirical evidence, and a quantitative meta-analysis to estimate combined effect sizes (Curtis et al., 2025; Van der Wal & Kok, 2019). The statistical analysis used was the meta-correlation, also known as meta-analysis of the correlation coefficient r, which combines the correlation coefficients of different studies to estimate the general relationship between two continuous variables (Daraj et al., 2023).

All analyses were conducted in RStudio (version 4.3.1), using the packages sf, rnaturalearth, rnaturalearthdata, dplyr, ggplot2, and ggrepel for the systematic review, and metafor, clubSandwich, dplyr, ggplot2, and ggrepel for the meta-analysis. Both studies reporting overall correlations and those reporting correlations between dimensions of one instrument and the other variable were included, while correlations between dimensions of both variables were excluded. In the meta-analysis, reported correlation coefficients were transformed into Fisher’s Z values, which were used as the analytical metric to stabilize variance and improve the statistical properties of the estimates (Majumder & Ray, 2021). The results were then back-transformed to r coefficients to facilitate interpretation (Ansari & Iqbal, 2025; Ma et al., 2025; Ooi et al., 2025). To address dependence among multiple effect sizes derived from the same primary study, correlations were not averaged across dimensions. Instead, Robust Variance Estimation (RVE) was applied, allowing multiple correlated effect sizes to be included while appropriately weighting by sample size. An a priori random-effects model was specified, considering the expected conceptual, methodological, and contextual heterogeneity across studies, associated with differences in organizational contexts, measurement instruments, and sample characteristics. Heterogeneity was evaluated using τ², which estimates variability between studies in a manner robust to within-study dependencies, making it more appropriate than I² in this context. Results were displayed using forest plots (Daraj et al., 2023). Publication bias was assessed using funnel plots and Begg’s rank correlation test, chosen for its conservative nature and lower propensity to inflate Type I error in meta-analyses with a small number of studies. Additionally, influence analyses were conducted using DFBETAS to identify effect sizes disproportionately affecting the model intercept. Finally, Microsoft Excel was used as a complementary tool for data organization and figure preparation (Yang et al., 2024).

2.6. Importance and Procedure for Correlation Analysis

For each eligible study, correlation coefficients (r) were extracted as the primary effect size metric (Allen et al., 2017). In cases where primary studies reported multiple correlations within the same study (e.g., between different dimensions of EI or TL and TE), effect sizes were computed to avoid unit-of-analysis errors, following recommended aggregation procedures (Borenstein et al., 2009). Correlational meta-analysis constitutes a fundamental methodological tool, as it systematically integrates evidence from primary studies on associations between constructs and allows for more precise, robust, and generalizable estimates of the true effect size (Leongómez, 2023).

To stabilize variance and correct for non-normality, correlation coefficients were transformed using Fisher’s Z transformation prior to analysis. Meta-analytic aggregation was conducted using random-effects models, acknowledging the considerable conceptual and methodological heterogeneity expected across different organizational contexts, sectors, team compositions, and measurement instruments.

Heterogeneity was evaluated using the Q statistic and τ², which allowed for the identification of between-study variance not attributable to sampling error. When sufficient data were available, meta-regression analyses were conducted to explore potential moderating effects of variables such as sex predominance, level of analysis, geographic region, and quality assessment; however, the limited and heterogeneous reporting of moderator analyses constrained more detailed moderation analyses.

It is important to highlight that the correlational meta-analytic design aligns with the Input–Process–Output (IPO) and Input–Mediator–Output–Input (IMOI) theoretical frameworks, which conceptualize leadership and EI as inputs whose effects on TE are transmitted through team processes and emergent mediator states (Mathieu et al., 2008). Consequently, the use of correlational meta-analysis is theoretically justified and methodologically appropriate for examining complex, multilevel phenomena in team research. Previous meta-analyses in the domains of leadership, EI and TE have consistently employed this approach to estimate population-level associations and contribute to theoretical development and the integration of explanatory models (Miao et al., 2017; Coronado-Maldonado & Benítez-Márquez, 2023). In line with these precedents, the present study uses correlational effect sizes to provide a robust, integrative quantitative synthesis that complements the conceptual assumptions of the IPO and IMOI models.

3. Results

3.1. Selection of Studies

The initial search identified 728 items, which were selected in 3 stages (Figure 1): (1) Removal of duplicates: (n = 213). (2) Screening: The investigators independently reviewed the titles and abstracts, excluding (N = 497). (3) Inclusion and exclusion: Resulting in 22 studies for the systematic review, of which 15 documents served for the meta-analysis, subsequently completing the quality evaluation. Following the recommendations of Davis et al. (2014) and PRISMA (Page et al., 2021), the exclusions were recorded: (Dizgah & Keshavarz, 2015; Malik et al., 2022; Silva et al., 2020) did not report correlations between EI–TE and Sosik et al. (1997) did not report the correlation between LT–TE, (Othman et al., 2009; Mysirlaki & Paraskeva, 2019) were excluded because they reported only dimensional correlations between TL and TE.

Of the 493 studies excluded after the review, the majority did not meet the inclusion criteria defined according to the PICOS framework. First, 113 studies did not analyze the relationship between EI and/or TL on TE. For example, studies such as Strengthening the bond and enhancing team performance: Emotional intelligence as the social glue (Lu & Fan, 2017) and Group leader emotional intelligence and group performance (Y. Zhang et al., 2023) focused exclusively on EI and team performance without considering overall team effectiveness, or The effect of emotional intelligence, intellectual intelligence, and transformational leadership (Fareed et al., 2021).

Second, 247 studies were excluded because they did not address any of the study variables and instead focused on other aspects, such as personality traits, different types of leadership excluding TL, studies related to teams that did not measure effectiveness, or research conducted exclusively in the healthcare domain. For instance, Gender diversity and motivation in collaborative learning groups: The mediating role of group discussion quality (Curşeu et al., 2018) analyzed collaborative learning processes and group dynamics, as did studies conducted in inclusive leadership contexts (Egitim, 2022), without operationalizing TL or considering EI in relation to TE.

Additionally, 21 studies were excluded that, although conceptually relevant, were systematic reviews, narrative reviews, or previous meta-analyses, such as Team-Centric Leadership: An Integrative Review (Kozlowski et al., 2016) and The impact of project manager’s emotional intelligence on project performance: A meta-analysis (Q. Zhang et al., 2023), since the present work focused exclusively on primary empirical studies in accordance with methodological recommendations for correlational meta-analyses.

Another notable group of 12 exclusions consisted of research conducted in non-workplace or highly specific populations, such as university students, athletes, or educational and healthcare contexts, which limited the generalizability of results to organizational work teams. For example, Performance measurement in project management (Bosch-Rekveldt et al., 2023) was conducted in project management, while Relationship between Emotional Intelligence and Transformational Leadership in Physical Education Managers (Esfahani & Soflu, 2011) was restricted to a very specific professional sector and addressed only managers rather than the team as a whole.

Four studies that were not open access, such as (Post, 2015), were also excluded. Finally, 96 studies were excluded because they addressed only a single variable and did not report correlation coefficients, which were a required inclusion criterion for the correlational meta-analysis (Kim & Ko, 2021).

3.2. Review of the Quality of the Analyzed Documents

Fifteen documents published between 2008 and 2024 were selected for the meta-analysis. The inter-rater reliability was good (κ = 0.79, SE = 0.15, z = 5.32, p < 0.001; 95% CI [0.50, 1.00]), which, according to Landis and Koch (1977), represents substantial agreement and is considered satisfactory in recent studies (Melguizo-Ibáñez et al., 2024; Valderrama-Diaz et al., 2025). No studies were classified as of poor quality, four were regular (30.8%) and nine were good (69.2%). The quality scores ranged from seven (Mntambo & Chan, 2024) to 10.83 (Dimas et al., 2018), with a median of 9.70.

3.3. Systematic Review

A total of 728 articles were identified, of which 213 were eliminated as duplicates and 493 did not meet the criteria, leaving 22 studies for the final analysis. Scopus contributed 473 articles (10 accepted; 45%), Science Direct 185 (0 accepted; 0%), Web of Science 48 (7 accepted; 32%), and other sources 22 (5 accepted; 23%), which shows that, although Scopus had greater volume, Web of Science and other bases presented greater relevance (Figure 2).

Likewise, Figure 3 shows the temporal evolution: between 1997, when the first study identified by Sosik et al. (1997), and 2010, only 4 articles were published (18.2%) between 2011 and 2017. The number increased to 7 (31.8%) between 2018 and 2024, reaching 11 publications (50%) in the last period, reflecting sustained growth and a notable boom (Figure 3).

Finally, Figure 4 presents the geographical distribution: Asia concentrates 50% of the studies, with South Korea (3), Pakistan (2), and Iran, China, Vietnam, Turkey, and Malaysia (1 each); America contributes 21.4%, with the United States (3), Brazil (1), and Peru (1); Europe 14.3%, with Portugal (2) and Greece (1); Africa 7.1% with South Africa (1); and Oceania 7.1% with Australia (1). There are two studies developed in virtual environments, but they were not included in the graph because they do not specify the location where they were conducted (Mysirlaki & Paraskeva, 2019, 2020).

3.4. Characteristics of the Selected Studies

It is essential to highlight the relevant findings identified in the systematic review that, had they not been carried out, would not have been revealed solely through meta-analysis. These data were considered valuable to offer a broader and more contextual view of the field of study (Van der Wal & Kok, 2019) (

Table 3. Description of items included for RS.

No.	Author(s)	Year	VE	Variables Extra	Variables of Control	NA	Context	Design	Country	Cont.
1	Ahmed et al. (2014)	1	3	NI	Education	1	Services (Private Sector)	Does not indicate	Pakistan	1
2	Dizgah and Keshavarz (2015)	2	3	NI	NI	1	Service (banks)	Does not indicate	Iran	1
3	Mysirlaki and Paraskeva (2019)	2	3	Team cohesion, cooperative group norms, social capital, and interdependence of tasks	NI	1	Video game forums, social networks (Facebook and Twitter), and universities.	Does not indicate	Virtual	NI
4	Silva et al. (2020)	2	1	Confidence, Psychological Capital, Collective Behavior, Intragroup Conflict, Team Psychological Safety, Team Cohesion	NI	2	Service (does not indicate category, but there is a 47% majority)	Does not indicate	Brazil	2
5	Hur et al. (2011)	1	3	Leader Effectiveness and Service Climate	Age, education, and size	2	Public Sector (Cases, Law Enforcement and Administration, Additional Law)	Transversal	South Korea	1
6	Polychroniou (2009)	1	3	NI	NI	1	Services, trade, financial services, and manufacturing	Does not indicate	Greece	3
7	Malik et al. (2022)	2	1	Service leadership, work stress, project success	NI	1	Service (Non-Governmental Organizations)	Transversal	Pakistan	1
8	Q. Zhang and Hao (2022)	2	2	Team cohesion, Duration of the team in the project	NI	1	(Construction Industry)	Does not indicate	China	1
9	Mysirlaki and Paraskeva (2020)	2	3	NI	Sex	1	Virtual teams (technology company), gaming forums, Facebook, and Twitter groups. University students	Does not indicate	Virtual	NI
10	Dunaway (2013)	1	2	NI	Team size and sex	2	Service (Education)	Does not indicate	United States	2
11	Paredes-Saavedra et al. (2024)	2	2	Work environment, organizational culture, and creative synergy, Team leadership	NI	1	Service (Education)	Transversal	Peru	2
12	Farh et al. (2012)	1	2	Work performance, Work context of managerial work demands MWD	Age, sex, supervisor status, seniority at work, personality traits, cognitive ability, contextual work factors (work complexity and emotional work demands)	1	Service (Education)	Does not indicate	United States	2
13	Choi et al. (2017)	2	1	Shared leadership	Sex, Age, Education level, Seniority in the organization	1	Services (financial and insurance)	Transversal	South Korea	1
14	Paolucci et al. (2018)	2	1	Affective Team Commitment	Team size	2	Services, industry, and research.	Transversal	Portugal	3
15	Tran and Vu (2021)	2	1	Teamwork orientation, Shared leadership	NI	1	Services (Information Technology, Real Estate, Financial, and Educational Products)	Transversal	Vietnam	1
16	Lee and Wong (2017)	2	2	Team process (task conflict and relationship conflict)	Team size, age diversity, and sex.	2	Banking, investments, healthcare, information technology, and pharmaceuticals.	Transversal	South Korea	1
17	Mitchell et al. (2014)	1	1	Interprofessional Motivation; Openness to Diversity; Negative Affective Tone	Professional Diversity; Average Age; Team Size	2	Service (Health).	Transversal	Australia	5
18	Özaralli (2003)	1	1	Empowerment	NI	1	Private companies: Textiles, advertising, communications, construction, aviation, energy, banking, and health.	Does not indicate	Turkey	1
19	Dimas et al. (2018)	2	1	NI	Team size	2	Services and industry	Transversal	Portugal	3
20	Mntambo and Chan (2024)	2	1	Transactional, managerial, and laissez-faire leadership	MTM experience and number of MTMs	1	Service (Mining)	Does not indicate	South Africa	4
21	Sosik et al. (1997)	1	1	Transactional leadership, group power, and level of anonymity	NI	2	Service (University)	Longitudinal	United States	2
22	Othman et al. (2009)	1	2	Work motivation	Age, sex, marital status, educational level, work experience, and position in the company	1	Services (accounting, banking, consulting, banks, hotels, call centers, messaging, hotel services, insurance, investment, legal, sales and services, telecommunications).	Does not indicate	Malaysia	1

Note. The coding was: 1 = Asia, 2 = America, 3 = Europe, 4 = Africa, 5 = Oceania and Virtual = 6; For type of relationship the coding was: 1 = LT–TE, 2 = EI–TE, 3 = EI–LT in TE and 4 = the 3, For year it was coded with 1 before 2015 and with 2 after 2015, Level of analysis (NA): 1 = members and 2 = Group.

).

3.5. Theories That Address Studies

Of the total of 22 documents analyzed, 41% (n = 9) incorporated some theory into their conceptual framework, while 59% (n = 13) made no explicit reference to theories. Among the theories identified, the Input–Process–Output (IPO) Model (McGrath, 1964) is implemented by Mysirlaki and Paraskeva (2019) in conjunction with the IPO framework and the multiphase approach proposed by Marks et al. (2001). Similarly, Mntambo and Chan (2024) integrate the conceptual model of Hackman (2002) with the IPO, while Paredes-Saavedra et al. (2024) expand this approach using the Input-Mediator-Output-Input (IMOI) model and the foundational framework proposed by Hackman (1990). Q. Zhang and Hao (2022) also use the IPO model, along with the Theory of Effectiveness in Work Teams (Mathieu et al., 2008) and the Emotional Intelligence Skill Model (Wong & Law, 2002). In the field of leadership, Dimas et al. (2018) focus on the Transformational Leadership Theory, while Malik et al. (2022) combine the OSH Theory with the Transformational Theory. Dunaway (2013) adopts the Theoretical Framework of team effectiveness proposed by Tannenbaum et al. (1992). In relation to emotional intelligence and team functions, Lee and Wong (2017) employ the Role Theory to analyze the influence of emotional intelligence on team processes and outcomes, while Farh et al. (2012) utilize the Capacity-based Emotional Intelligence Model proposed by Mayer and Salovey (1997).

3.6. Types of Correlations and Methodological Designs Addressed in the Studies

The analysis of the 22 included studies shows that the vast majority (n = 21; 95%) employed a cross-sectional design, of which nine (41%) explicitly specified (Choi et al., 2017; Dimas et al., 2018; Hur et al., 2011; Lee & Wong, 2017; Malik et al., 2022; Mitchell et al., 2014; Paolucci et al., 2018; Paredes-Saavedra et al., 2024; Tran & Vu, 2021). Although 12 studies do not specify the design, due to the nature of their measurements and procedures, it is inferred that they are transversal (Ahmed et al., 2014; Dizgah & Keshavarz, 2015; Dunaway, 2013; Farh et al., 2012; Mntambo & Chan, 2024; Mysirlaki & Paraskeva, 2019, 2020; Othman et al., 2009; Özaralli, 2003; Polychroniou, 2009). Only the study by Sosik et al. (1997), which represents 5%, adopted a longitudinal factorial design in the laboratory. These results show a marked preference for measurements at a single time point, with little representation of longitudinal designs or experimental methods.

3.7. Control Variables and Additional Variables That Research Has Addressed

Based on the studies included in the systematic review, it is evident that research has consistently incorporated control and additional variables to more accurately understand how EI and leadership, particularly TL, relate to TE. In this context, control variables primarily focused on sociodemographic and structural factors, such as age, gender, education level, team size, organizational tenure, and position, aiming to isolate the specific effect of leadership and reduce potential biases in the estimates (Ahmed et al., 2014; Dunaway, 2013; Paolucci et al., 2018; Q. Zhang & Hao, 2022). Some studies expanded this approach by including more complex controls, such as personality traits, cognitive ability, professional diversity, job complexity, emotional demands, and prior experience in multidisciplinary teams, allowing for a more robust approximation of real organizational contexts (Farh et al., 2012; Mitchell et al., 2014; Mntambo & Chan, 2024).

The results of the systematic review show a clear convergence with the Input–Process–Output (IPO) and Input–Mediator–Output–Input (IMOI) theoretical frameworks, demonstrating that the effects of EI and TL, conceptualized as inputs, do not operate directly or in isolation on TE but are instead conditioned by a set of team processes and emergent states that act as key mediating variables (Hackman, 2002; Mathieu et al., 2008; McGrath, 1964). While the meta-analysis was limited to the three main study variables, the qualitative synthesis allowed the identification of a set of mediating and contextual variables that explain how TL and EI translate into improved collective outcomes.

Specifically, several studies included in the review incorporated variables conceptually aligned with psychological capital, such as trust, collective psychological capital, work motivation, and empowerment, which operate as positive psychological resources enhancing self-efficacy, hope, and optimism at both the individual and collective levels. These resources facilitate the manifestation of TL in proactive behaviors, greater affective commitment, and superior team performance (Othman et al., 2009; Özaralli, 2003; Paolucci et al., 2018; Silva et al., 2020).

Additionally, the systematic review identified variables closely related to psychological safety, both directly and indirectly. Team psychological safety was explicitly addressed by Silva et al. (2020), while other studies analyzed functional indicators such as team cohesion, cooperative norms, social capital, openness to diversity, and intragroup conflict processes. These emergent states create an interpersonal climate of trust that allows transformational leadership to promote participation, collective learning, and constructive conflict management (Lee & Wong, 2017; Mitchell et al., 2014; Mysirlaki & Paraskeva, 2019; Q. Zhang & Hao, 2022). Overall, the findings of the systematic review support an IMOI logic, in which TL, together with EI, influences TE indirectly through intermediate psychological and relational states that mediate and amplify its impact. In this way, the systematic review complements the results of the meta-analysis by providing a broader, process-oriented understanding of the mechanisms through which TL contributes to the TE.

3.8. Measurement Instruments

As shown in

Table 4. Instruments used by the authors.

No.	Authors	TE Instrument	TL Instrument	EI Instrument
1	Ahmed et al. (2014)	Adapted (De Dreu, 2007; Schaubroeck et al., 2007; Tsui et al., 1997)	Multifactor Leadership Questionnaire (MLQ) (Bass & Avolio, 1994)	Emotional intelligence WLEIS (Wong & Law, 2002)
2	(Dizgah & Keshavarz, 2015)	Team Effectiveness questionnaire (Bateman et al., 2002)
3	(Mysirlaki & Paraskeva, 2019)	It was based on the IPO Framework (Hackman, 1987)	Multifactor Leadership Questionnaire (Bass & Riggio, 2005)	Emotional intelligence WLEIS (Wong & Law, 2002)
4	(Mysirlaki & Paraskeva, 2020)	It measured team performance, team members’ satisfaction, and team viability (Hackman, 1987)	Multifactor Leadership Questionnaire MLQ (Bass & Riggio, 2005)	Emotional intelligence WLEIS (Wong & Law, 2002)
5	Hur et al. (2011)	Adapted (De Dreu, 2007; Schaubroeck et al., 2007; Tsui et al., 1997)	Multifactor Leadership Questionnaire (MLQ-Form 5X-Short) 20 Bass and Avolio items from 2000	Emotional intelligence WLEIS (Wong & Law, 2002)
6	Polychroniou (2009)	Does not indicate the name of the instrument	Multifactor Leadership Questionnaire (MLQ) (Bass, 1985)	Adapted from Emotional Quotient Index (EQI) by (Rahim et al., 2006, 2002)
7	(Malik et al., 2022)	Adopted from (Maynard et al., 2012)	N/A	Emotional intelligence WLEIS (Wong & Law, 2002)
8	(Q. Zhang & Hao, 2022)	Aplicó la escala de Y. Li (2013)	N/A	Emotional intelligence WLEIS (Wong & Law, 2002)
9	(Lee & Wong, 2017)	Measured team performance (Hackman, 1987), innovation (Anderson & West, 1998), and Cohesion (Seashore, 1954; Widmeyer et al., 1985)	N/A	WEIP-Short Version WEIP-S (Jordan & Lawrence, 2009)
10	Dunaway (2013)	Team learning beliefs & Behaviors—Questionnaire (Van den Bossche et al., 2006)	N/A	WEIP-Short Version WEIP-S (Jordan & Lawrence, 2009)
11	Farh et al. (2012)	Role-based performance scale (Welbourne et al., 1998)	N/A	Could not find
12	(Othman et al., 2009)	Team role assessment adapted from (Welbourne et al., 1998)	N/A	Emotional intelligence WLEIS (Wong & Law, 2002)
13	(Silva et al., 2020)	Team performance (Rousseau & Aubé, 2010), Quality of group experience (Aubé & Rousseau, 2005)	Developed by (Carless et al., 2000)	N/A
14	(Paredes-Saavedra et al., 2024)	Prepared by (Jaca García et al., 2011)	Scale Trait Meta-Mood Scale TMMS adopted by (Gurrutxaga Azurmendi, 2015)	N/A
15	Choi et al. (2017)	Scale developed by integrating output-and process-related items derived from previous research (Pearce & Sims, 2002)	Multifactor Leadership Questionnaire (MLQ) (Bass & Avolio, 1990)	N/A
16	Paolucci et al. (2018)	Measured team viability, quality of group experience (Aubé & Rousseau, 2005), and team process improvement (Rousseau & Aubé, 2010)	Scale developed (Carless et al., 2000) and adapted to Portuguese (van Beveren et al., 2017)	N/A
17	Tran and Vu (2021)	It measured team performance, quality of team experience, and team viability (Aubé & Rousseau, 2005)	Multifactor Leadership Questionnaire (MLQ) (Bass & Avolio, 1990)	N/A
18	Mitchell et al. (2014)	Based on (Mathieu et al., 2008)	Transformational leadership scale TLS (García-Morales et al., 2008)	N/A
19	Özaralli (2003)	Developed by the same author, measures: innovativeness, in-group communication, and performance	Multifactor Leadership Questionnaire (MLQ) (Bass & Avolio, 1990)	N/A
20	Dimas et al. (2018)	Team resilience (Stephens et al., 2013), quality of group experience, and team viability (Aubé & Rousseau, 2005)	Developed by (Carless et al., 2000)	N/A
21	(Mntambo & Chan, 2024)	Hackman’s team effectiveness survey (Cavanaugh et al., 2021)	Bass and Avolio from 2000	N/A
22	(Sosik et al., 1997)	Longitudinal	Longitudinal	N/A

, the analysis shows that, after 2014, the use of instruments increased in the three variables studied, evidencing a methodological evolution. In terms of TE, 71% of the studies used single scales, and 29% combined them, with a subsequent increase to 2014 (from 29% to 43% for single scales and from 10% to 19% for combined scales), reflecting a trend toward more comprehensive measurements. In TL, the MLQ of Bass and Avolio remains the dominant instrument (73%), although after 2014 the use of alternative scales increased (from 7% to 20%). In EI the WLEIS maintains leadership with 54% of the total, followed by the WEIP-S with 15%, and other instruments with 31%. After 2014, the WLEIS predominates (31%), with a lesser presence of alternative scales.

3.9. Determine the Statistical Relationship Between the Variables of Interest

For the correlation between TL and TE, Robust Variance Estimation (RVE) was employed based on 9 studies that contributed 18 effect sizes and included a total of 3480 team members. The analysis yielded Z = 0.493 (SE = 0.058, t(7.45) = 8.51, p < 0.001), equivalent to r ≈ 0.45, indicating a positive, moderate, and significant association (Figure 5). Heterogeneity was moderate and evaluated using variance components (τ²), with results showing variance both between studies (τ² ≈ 0.016) and within studies (τ² ≈ 0.018), supporting the use of a multilevel random-effects model and justifying the exploration of potential moderators. Publication bias was explored using Begg’s test, which indicated no funnel plot asymmetry (Kendall’s tau = 0.029, p = 0.92) (Figure 6). Influence analysis showed that only one effect had a moderate impact (DFBETAS = −0.84), suggesting a moderate influence on the model intercept; however, this did not compromise the overall stability of the results. Overall, the findings confirm a positive, moderate relationship between TL and TE.

Robust Variance Estimation (RVE) was employed based on 8 studies that contributed 15 effect sizes and included 3440 team members, showing a moderate, positive, and significant meta-correlation between EI and TE (Z = 0.436, SE = 0.113, t(6.98) = 3.87, p = 0.006), equivalent to r ≈ 0.41 (Figure 7). Heterogeneity was substantial at the between-study level (τ² = 0.093), while within-study variability was considerably lower (σ² = 0.010), indicating that most of the heterogeneity is explained by differences between studies rather than internal variations. Publication bias was explored using Begg’s test, with no evidence of funnel plot asymmetry or publication bias detected (Kendall’s tau = −0.286, p = 0.322) (Figure 8). Influence analysis identified the study by Paredes-Saavedra et al. (2024) as influential (DFBETAS = 1.06); its exclusion reduced the correlation magnitude from Z ≈ 0.44 to Z = 0.36, although the effect remained statistically significant (p = 0.005). Additionally, between-study heterogeneity decreased substantially (τ² from 0.093 to 0.037), indicating that this study contributed meaningfully to inter-study variability, while within-study variability remained practically unchanged. Overall, the results support the existence of a positive, moderate relationship between EI and TE, even after accounting for the influence of individual studies.

The results of Robust Variance Estimation (RVE), based on 4 studies contributing 6 effect sizes and including a total of 1955 team members, showed a high, positive, and statistically significant meta-correlation between EI and TL (Z = 0.74, SE = 0.123, t(2.97) = 6.04, p = 0.009), equivalent to r ≈ 0.63 (Figure 9). Heterogeneity was moderate at the between-study level (τ² = 0.044), while within-study variability was lower (σ² = 0.018). Publication bias was explored using Begg’s test, which yielded a coefficient of τ = 0.33, indicating a slight positive asymmetry in the funnel plot. However, this association was not statistically significant (p = 0.497), suggesting no evidence of publication bias (Figure 10). Influence analysis showed that most studies had minimal impact on the overall meta-correlation. Nevertheless, the study by Hur et al. (2011) exhibited a relatively high influence (DFBETAS = −0.72). Overall, the meta-analytic results demonstrate a high, positive, and statistically significant relationship between EI and TL, confirming the robustness of the overall effect.

3.10. Analysis by Subgroups and Meta Regression

Moderator analyses showed differentiated patterns depending on the type of meta-analyzed relationship. In TL–TE, studies published after 2018 (β = 0.349, p < 0.001) had a strong impact, while the level of analysis (β = −0.296, p = 0.028) and methodological quality (β = 0.415, p = 0.014) showed moderate and significant effects, indicating that more recent studies, with higher rigor and focused on team members, report stronger associations. In EI–TE, none of the moderators reached statistical significance (p > 0.05), suggesting a stable and robust relationship, independent of sample characteristics, geographic context, level of analysis, or study quality. Finally, in EI–TL, the European continent (β = 0.395) and virtual contexts (β = 0.453) showed high and significant effects (p < 0.001), whereas studies published after 2018 had a slight negative effect. These results should be interpreted with caution due to virtually no residual variance and the small number of studies. Overall, the findings indicate that while the EI–TE relationship is highly robust and minimally sensitive to study characteristics, the TL–TE and EI–TL relationships are more dependent on temporal, contextual, and methodological factors (

Table 5. Moderator effects in the meta-analytic relationships between TL, EI, and TE.

Rel.	Moderator	Estimate (β)	SE	p-Value	Sig.	Impact
1	Intercept	−1.126	0.247	0.031	*	Moderate
1	Sex: Female	0.135	0.098	0.261	ns	Not significant
1	Continent: America	−0.167	0.05	0.079		Small
1	Continent: Europe	−0.132	0.118	0.382	ns	Not significant
1	Continent: Virtual	0.307	0.114	0.072		Small
1	Year: Post-2018	0.349	≈0	<0.001	***	High
1	Level: Members	−0.296	0.05	0.028	*	Moderate
1	Methodological quality	0.415	0.05	0.014	*	Moderate
2	Intercept	1.174	1.549	0.587	ns	Not significant
2	Sex: Female	−0.026	0.471	0.965	ns	Not significant
2	Continent: America	0.161	0.471	0.791	ns	Not significant
2	Continent: Virtual	−0.092	0.408	0.859	ns	Not significant
2	Year: Post-2018	0.274	0.407	0.623	ns	Not significant
2	Level: Members	0.261	0.471	0.678	ns	Not significant
2	Methodological quality	−0.336	0.409	0.563	ns	Not significant
3	Intercept	0.536	≈0	<0.001	***	Moderate
3	Continent: Europe	0.395	≈0	<0.001	***	High
3	Continent: Virtual	0.453	≈0	<0.001	***	High
3	Year: Post-2018	−0.039	≈0	<0.001	***	Small

Note. Relationships coded as follows: 1 = TL–TE; 2 = EI–TE; 3 = EI–TL. Asterisks indicate statistical significance levels: *** p < 0.001; * p < 0.05; ns = not significant.

).

3.11. Evaluation of the Certainty of the Evidence

To assess the robustness of the findings obtained in this systematic review with meta-analysis, an evaluation of the certainty of the evidence was conducted, following the fundamental principles of the GRADE framework (Guyatt et al., 2011). This assessment was complementary to the evaluation of the individual methodological quality of the primary studies, performed using the adapted checklist of Petticrew and Roberts (2006) and focused on the aggregate body of evidence for each key outcome.

The certainty was determined following the following criteria: risk of bias, referred to the methodological quality and transparency of the included studies, consistency, referred to the degree of homogeneity of the effects between the studies, precision, refers to the amplitude of the confidence intervals of the combined estimates, publication bias, this is examined by funnel graphs, applicability, referred to the relevance and generalizability of the findings for diverse organizational contexts.

The synthesized evidence confirms positive and statistically significant associations between TL, EI, and TE, with an overall moderate level of certainty. The TL–TE relationship shows a positive, moderate association, with heterogeneity distributed both between and within studies, supporting the use of multilevel models. The effect estimate is precise, reflected by a relatively narrow confidence interval, and the absence of evidence of publication bias supports its robustness and stability. For EI–TE, although high between-study heterogeneity is observed, the effect remains stable, positive, and significant, and sensitivity analyses confirm its robustness. Additionally, the confidence interval indicates acceptable precision. Finally, the EI–TL relationship shows a strong and consistent effect, although based on a smaller number of studies, which limits precision. Overall, the findings underscore the relevance of EI and TL as key factors for TE, justifying a moderate GRADE certainty and highlighting the need for future longitudinal and experimental studies (

Table 6. Frame GRADE.

Relation	Bias Risk	Consistency	Precision	Publication Bias
TL–TE	Moderate	Moderate (between-study τ2 ≈ 0.016; within-study τ2 ≈ 0.018)	[0.38, 0.61]	Begg’s test nonsignificant (τ = 0.03, p = 0.92)
EI–TE	Moderate	High (τ2 = 0.093 inter-study; σ2 = 0.010 intra-study)	[0.21, 0.66]	Begg’s test nonsignificant (τ = −0.29, p = 0.32)
EI–TL	Moderate	Moderate (τ2 = 0.044 inter-study; σ2 = 0.018 intra-study)	[0.50, 0.98]	Begg’s test indicates mild asymmetry but nonsignificant (τ = 0.33, p = 0.50)

).

4. Discussion

This study conducted an integrated analysis of the relationships between emotional intelligence (EI), transformational leadership (TL), and team effectiveness (TE) based on a systematic review and meta-analytic synthesis of previously reported empirical studies. The findings confirm and expand existing evidence regarding the positive and significant relationships among these constructs, aligning with prior research that highlights the mediating role of EI in the influence of leadership on organizational outcomes (Hur et al., 2011; Karve et al., 2025). However, this study adds value by quantitatively synthesizing these relationships and identifying methodological and contextual factors that modulate their strength and significance.

Regarding the results linked to SO1 and SO2, the analysis revealed that most included studies employed cross-sectional correlational designs and showed limited incorporation of control and mediating variables. As noted by Davis et al. (2014) and Hur et al. (2011), the scarce use of longitudinal designs (only 1 of 22 studies) restricts the ability to establish causal relationships, which aligns with previous critiques in the leadership and EI literature (Lee & Wong, 2017; Ooi et al., 2025). This methodological limitation contributes to the heterogeneity observed in effect sizes and underscores the need for more integrative analytical models. The repeated reference to the IPO model in the included studies (Mysirlaki & Paraskeva, 2019; Paredes-Saavedra et al., 2024) highlights the relevance of mediational processes, such as coordination, empathic communication, and stress management, in translating TL and EI into TE.

Regarding SO3, the results indicate that the reviewed studies employed well-established and reliable instruments for measuring EI, TL, and TE, supporting the consistency of findings in both the systematic review and meta-analysis. Furthermore, the diversity of measurement instruments—MSCEIT for EI, MLQ for TL, and various ad hoc scales for TE—reflects the multifaceted nature of these constructs, although it poses challenges for direct comparability of results, as acknowledged in previous meta-analyses (Liu et al., 2025; Yang et al., 2024). Similarly, the concentration of studies in Asia and the services sector limits the generalizability of findings to other cultural contexts and economic sectors, a gap previously identified as an area for improvement in the international literature on team effectiveness (Pranckutė, 2021; Valderrama-Diaz et al., 2025). Nonetheless, the diversity of measurement approaches may also explain part of the variability between studies.

Regarding results related to SO4, the meta-analysis revealed positive and moderately strong associations between EI–TE and TL–TE, as well as a high and significant relationship between EI and TL. These findings indicate that EI not only acts as an independent input fostering group cohesion, trust, and collective performance (Ratan Kumar & Sampath Kumar, 2024; Coronado-Maldonado & Benítez-Márquez, 2023; Mburu, 2020) but also serves as a facilitator of TL, enhancing emotional self-regulation, empathy, and individualized consideration, which strengthens team effectiveness (Munir et al., 2023; Vrabii, 2023). Additionally, reviewed studies indicate that TL increases motivation, reinforces trust, and promotes team performance, even in complex organizational contexts, consolidating its role as a critical input for coordination and collaborative processes (Mishra et al., 2019; Suwandana, 2019).

A high and statistically significant relationship between EI and TL was also observed. This evidence aligns with Vrabii (2023), who reported a direct relationship between leaders’ EI, perceived TL, and TE. Similarly, Munir et al. (2023) found that individuals with higher levels of EI are more likely to display transformational leadership behaviors. These results suggest that EI enhances transformational leadership capabilities by facilitating emotional regulation, empathy, and individualized attention, thereby increasing trust and willingness to collaborate. In this sense, EI functions not only as an independent input but also as a key facilitator of TL, reinforcing its impact on TE.

The moderate-to-high heterogeneity observed in the relationships among EI, TL, and TE indicates that these associations depend on contextual, methodological, and sample-related factors. Variables such as sample size, methodological quality, and construct operationalization influence the stability of effects, whereas sample gender composition showed inconsistent patterns due to its limited treatment as a moderator in primary studies (Ahmed et al., 2014; Hur et al., 2011; Munir et al., 2023; Vrabii, 2023). Additionally, there was a tendency toward larger effect sizes in more recent studies, suggesting a progressive evolution in the measurement and operationalization of EI, TL, and TE (Ratan Kumar & Sampath Kumar, 2024; Coronado-Maldonado & Benítez-Márquez, 2023). Variations based on geographic location and level of analysis further underscore the importance of cultural context and team-focused approaches for a proper understanding of collective effectiveness (Mburu, 2020; Vrabii, 2023). Collectively, these findings support the IPO and IMOI models, demonstrating that contextual and methodological variables act as moderators influencing the impact of emotional intelligence and transformational leadership on team effectiveness.

Overall, these findings reinforce the theoretical validity of the analyzed constructs and the relevance of the IPO and IMOI frameworks, showing that the effects of EI and TL on TE are not solely direct but dynamic, context-dependent, and mediated by team processes. This underscores the need for integrative approaches that consider both individual inputs and emergent team processes within organizational contexts to fully understand collective effectiveness.

4.1. Limitations

Despite the methodological rigor applied, this review has limitations that must be considered. Publication bias is a concern, as an exhaustive search was conducted across multiple databases, and a funnel graph was used to evaluate bias. However, it is possible that unpublished studies with null or non-significant results were not included. This could inflate estimates of the effect size (Page et al., 2021; Van der Wal & Kok, 2019). Methodological heterogeneity, characterized by high variability in study designs, measurement instruments, and contexts, introduces heterogeneity that, although addressed with random effects models and subgroup analysis, complicates the unified interpretation of the results (Higgins et al., 2003; Leongómez, 2023). Exclusion of qualitative and gray studies was achieved by limiting inclusion to quantitative observational studies published in refereed journals, theses, and technical reports, while excluding qualitative studies that could have provided contextual depth and understanding of the mechanisms underlying the relationships studied (Chapman, 2021; Khalil et al., 2022). Dependence on cross-sectional correlations means that the predominance of cross-sectional designs (in 95% of studies) precludes the establishment of causal relationships. Future research should incorporate more longitudinal and experimental designs to strengthen causal inference (Davis et al., 2014; Siddaway et al., 2019). Limitations in the search strategy, although multiple databases and Boolean terms were used, some relevant studies may not have been captured due to language restrictions (only studies accessible in full text were included) or indexing in non-consulted databases (Pranckutė, 2021).

4.2. Practical Implications

The results show that companies can maximize their capabilities through programs designed to enhance and develop their capabilities. A leader demonstrates a solid capacity to manage their emotions; in addition, they must develop their emotional intelligence and strengthen the transformational leadership approach they apply, generating a positive impact on their work team (Madamang et al., 2023; Kulat & Shah, 2025). It is also suggested that leaders incorporate the practice of mindfulness, as it contributes to the strengthening of emotional intelligence, improves social skills, and fosters a supportive environment within companies (Coronado-Maldonado & Benítez-Márquez, 2023).

The findings demonstrate that emotional intelligence (EI) and transformational leadership (TL) are essential elements to maximize team effectiveness. Therefore, organizations should promote workshops that combine the development of emotional competencies, such as self-awareness, empathy, and self-regulation, with training in transformational leadership. This approach would not only optimize collective performance but also strengthen commitment and encourage innovation within the organization.

4.3. Implications for Future Research

The findings of this review, along with the limitations identified, suggest several promising lines for future research:

Longitudinal and experimental designs: to overcome the predominance of cross-sectional studies and establish causal relationships, the implementation of longitudinal and quasi-experimental designs is recommended to examine how the relationships between EI, TL, and TE evolve, as well as the impact of specific interventions on the development of emotional and leadership competencies. Expanding the Geographic and Sectoral Context: Given the overrepresentation of studies in Asia and the services sector, future research should explore these relationships in diverse cultural contexts (especially in Latin America and Africa) and in other economic sectors such as manufacturing, technology, or the public sector, to evaluate the generalizability of the findings.

Investigation of mediating and moderating mechanisms: It is crucial to deepen the study of mediating variables (for example, trust, psychological safety, learning climate, and cohesion) and moderators (for example, organizational culture, modality of face-to-face vs. virtual work, and type of task) that explain how and under what conditions EI and TL influence TE. A multilevel structural equation modeling approach would be particularly valuable. Integration of mixed methods: It is recommended to complement quantitative approaches with qualitative methodologies (such as interviews, case studies, and ethnographic observations) to gain a richer and more contextualized understanding of emotional and leadership dynamics within teams.

Focus on collective EI and shared leadership: While the literature has predominantly focused on EI and TL at the individual leader level, future studies should further explore the role of group emotional intelligence and shared leadership practices as critical predictors of team effectiveness. Intersection with technology and artificial intelligence (AI): Given the increasing digitization of work, there is a need to investigate how EI and TL manifest and are enhanced (or challenged) in virtual, hybrid environments, mediated by emerging technologies such as AI. This exploration could examine the role of leadership in managing emotions in digital spaces.

5. Conclusions

This systematic review and correlational meta-analysis were conducted with the aim of clarifying the relationship between emotional intelligence (EI), transformational leadership (TL), and team effectiveness (TE), integrating scattered and methodologically heterogeneous empirical evidence. Consistent with the stated objectives, the synthesis of studies revealed a predominance of correlational and cross-sectional designs, as well as the recurrent use of the r coefficient as a measure of effect size, supporting the methodological relevance of the adopted approach. Additionally, the review allowed for the identification of the main control and additional variables considered in the literature, including sociodemographic factors, psychological resources, team processes, and organizational contexts, as well as the most commonly used measurement instruments and their adequate levels of reliability, providing a structured view of how this relationship has been investigated across different contexts and levels of analysis.

The meta-analysis results confirmed the presence of positive and statistically significant relationships among EI, TL, and TE, with effect sizes ranging from moderate to high. However, moderation analyses suggest that the strength of these relationships varies depending on sample, contextual, and methodological characteristics, highlighting the importance of considering organizational context and study quality when interpreting the results. Overall, this work not only consolidates existing knowledge but also provides a solid foundation for designing leadership development programs and guiding future research on team effectiveness.