Emotional intelligence competencies that predict job performance beyond IQ (Goleman, 1995)
The construct of emotional intelligence (EI) emerged as a formal psychological concept in a seminal article by Peter Salovey and John D. Mayer (1990), who defined it as "the ability to monitor one's own and others' feelings and emotions, to discriminate among them, and to use this information to guide one's thinking and actions" (p. 189). This initial formulation positioned EI firmly within the intelligence tradition, conceptualizing it as a set of interrelated cognitive abilities rather than a personality trait or behavioral tendency. Salovey and Mayer drew on earlier work by Howard Gardner (1983) on multiple intelligences, particularly his constructs of interpersonal and intrapersonal intelligence, as well as on research in social cognition and emotion regulation.
Mayer and Salovey (1997) subsequently refined their model into a four-branch hierarchical framework that remains the most theoretically rigorous conceptualization of EI to date. The four branches are arranged in order of increasing psychological complexity:
Branch 1: Perceiving Emotions encompasses the ability to accurately identify emotions in oneself and others through facial expressions, vocal tone, body language, and other cues. This foundational ability includes the capacity to distinguish between genuine and feigned emotional expressions, a skill with demonstrated individual variation (Elfenbein & Ambady, 2002).
Branch 2: Using Emotions to Facilitate Thought refers to the ability to harness emotional states to enhance cognitive processes such as creativity, problem-solving, and decision-making. Research demonstrates that different emotional states facilitate different types of thinking: positive affect enhances creative problem-solving (Isen et al., 1987), while moderate negative affect can enhance analytical, detail-oriented processing (Forgas, 2007).
Branch 3: Understanding Emotions involves comprehension of the emotional lexicon, understanding how emotions blend and transition, and recognizing the causes and consequences of emotional states. This branch includes knowledge of emotion-eliciting situations, the ability to label complex emotional states, and understanding of emotional progression over time.
Branch 4: Managing Emotions represents the highest level of the hierarchy and encompasses the ability to regulate one's own emotions and influence others' emotional states. This includes the capacity to remain open to both pleasant and unpleasant feelings, to engage or detach from emotional states strategically, and to manage emotions in oneself and others to achieve personal and social goals.
Daniel Goleman's (1995) bestselling book Emotional Intelligence: Why It Can Matter More Than IQ brought the EI concept to mass public awareness, fundamentally shaping popular understanding of the construct. Goleman's model significantly expanded beyond Salovey and Mayer's ability framework, incorporating personality traits, motivational factors, and social competencies into a broader competency-based model. In his initial formulation, Goleman proposed five domains: self-awareness, self-regulation, motivation, empathy, and social skills.
Goleman (1998) subsequently refined this framework for organizational contexts in Working with Emotional Intelligence, reducing the model to four quadrants: Self-Awareness (emotional self-awareness, accurate self-assessment, self-confidence), Self-Management (emotional self-control, transparency, adaptability, achievement orientation, initiative, optimism), Social Awareness (empathy, organizational awareness, service orientation), and Relationship Management (inspirational leadership, influence, developing others, change catalyst, conflict management, building bonds, teamwork and collaboration).
While Goleman's work has been enormously influential in organizational development and leadership training, it has attracted significant scientific criticism for its expansive scope. Mayer, Salovey, and Caruso (2008) argued that Goleman's model conflates emotional intelligence with general personality and social competence, making it difficult to distinguish EI from established constructs already well-measured by existing personality instruments.
Reuven Bar-On (1997) developed a parallel conceptualization of emotional intelligence, introducing the term "emotional quotient" (EQ) and creating the Emotional Quotient Inventory (EQ-i) as a self-report measurement instrument. Bar-On's (2006) model defines emotional-social intelligence as "a cross-section of interrelated emotional and social competencies, skills and facilitators that determine how effectively we understand and express ourselves, understand others and relate with them, and cope with daily demands" (p. 14).
Bar-On's model encompasses five composite scales: Intrapersonal (self-regard, emotional self-awareness, assertiveness, independence, self-actualization), Interpersonal (empathy, social responsibility, interpersonal relationship), Stress Management (stress tolerance, impulse control), Adaptability (reality-testing, flexibility, problem-solving), and General Mood (optimism, happiness). The EQ-i has been validated across multiple cultural contexts and has accumulated a substantial normative database exceeding 100,000 respondents.
The Mayer-Salovey-Caruso Emotional Intelligence Test (MSCEIT; Mayer et al., 2002) represents the primary ability-based measure of EI and operationalizes the four-branch model through performance-based tasks rather than self-report. The MSCEIT presents respondents with stimuli requiring them to perceive emotions in faces and landscapes (Branch 1), judge which emotions would facilitate specific cognitive tasks (Branch 2), understand emotional vocabulary and emotional transitions (Branch 3), and evaluate strategies for managing emotions in self and others (Branch 4).
Scoring utilizes both consensus and expert scoring methods, with research demonstrating acceptable convergence between the two approaches (Mayer et al., 2003). Test-retest reliability has been established at approximately r = .86, and the instrument demonstrates discriminant validity from both personality measures and traditional cognitive intelligence (Brackett & Mayer, 2003). However, the MSCEIT has been critiqued for its consensus-based scoring methodology, which some argue confounds emotional knowledge with social conformity (Roberts et al., 2001).
The predictive validity of EI for important life outcomes has been examined through several comprehensive meta-analyses, providing a nuanced picture of the construct's utility.
O'Boyle, Humphrey, Pollack, Hawver, and Story (2011) conducted a landmark meta-analysis examining the relationship between EI and job performance across 43 studies (N = 5,795). Their findings indicated that all three streams of EI (ability-based, self-report, and mixed models) demonstrated incremental validity in predicting job performance beyond cognitive ability and the Big Five personality traits. The corrected correlation between EI and job performance was rho = .28 across all measures, with ability-based EI showing rho = .24 and mixed-model EI showing rho = .30. Critically, this meta-analysis demonstrated that EI contributes unique predictive variance even after controlling for established predictors, addressing a major critique of the construct.
Joseph and Newman (2010) performed a meta-analysis that examined the cascading model of EI, proposing that the relationship between EI and job performance follows a sequential path: emotion perception leads to emotion understanding, which leads to emotion regulation, which in turn predicts job performance. Their findings supported this cascading structure and further demonstrated that EI predicted job performance most strongly in jobs requiring high emotional labor (rho = .45) compared to those with low emotional labor demands (rho = .12).
Mattingly and Kraiger (2019) meta-analyzed EI training interventions (N = 58 studies), finding that EI-related competencies are trainable, with an overall effect size of d = 0.46. Training effects were stronger for ability-based programs than for knowledge-only interventions, and effects persisted at follow-up assessments, suggesting durable improvement in emotional competencies through structured intervention.
Emotional intelligence has been subject to vigorous academic debate. Antonakis (2009) presented a particularly forceful critique, arguing that EI research suffers from several methodological shortcomings: inadequate control for confounding variables (particularly general mental ability and personality), weak research designs that cannot establish causal relationships, and construct proliferation that adds little to existing well-validated constructs. Antonakis questioned whether EI represents a genuinely distinct construct or merely relabels aspects of personality and cognitive ability.
Locke (2005) offered a conceptual critique, arguing that the term "emotional intelligence" is an oxymoron because emotions and intelligence operate through fundamentally different psychological mechanisms. Locke contended that what EI researchers study is better understood as a combination of introspective skill, emotional knowledge, and social skill, none of which constitute "intelligence" in the traditional psychometric sense.
Landy (2005) raised additional concerns about the commercial exploitation of EI, arguing that the gap between popular claims and scientific evidence is substantial, and that organizations invest heavily in EI-based interventions without adequate empirical justification for their effectiveness.
The relevance of emotional intelligence to remote and distributed work has become a significant area of applied research. In virtual environments, several EI-related challenges intensify: reduced availability of non-verbal emotional cues (Daft & Lengel, 1986), increased potential for miscommunication in text-based channels, emotional isolation from colleagues, and difficulty in building trust without face-to-face interaction (Jarvenpaa & Leidner, 1999).
Research suggests that emotionally intelligent leaders are more effective in virtual contexts, demonstrating better ability to detect team member distress through limited communication channels, more effective conflict resolution in asynchronous environments, and stronger virtual team cohesion (Avolio et al., 2014). Workers with higher emotion regulation capacity show better adaptation to remote work demands, including more effective boundary management between work and personal life (Allen et al., 2015) and lower susceptibility to technostress and digital burnout (Salanova et al., 2013).
The practical implications for remote work include the importance of explicit emotional communication (compensating for reduced non-verbal cues), deliberate relationship maintenance behaviors, strategic use of video communication for emotionally complex interactions, and development of digital emotional literacy, the ability to accurately interpret and convey emotion through digital media.