A review published in the journal of Annual Psychology found that higher emotional intelligence is positively correlated with (Mayer, 2008):
- Better social relations for children – Among children and teens, emotional intelligence positively correlates with good social interactions, relationships and negatively correlates with deviance from social norms, anti-social behaviour measured both in and out of school as reported by children themselves, their own family members as well as their teachers (Mayer, 2008).
- Better social relations for adults – High emotional intelligence among adults is correlated with better self-perception of social ability and more successful interpersonal relationships while less interpersonal aggression and problems (Mayer, 2008). Highly emotionally intelligent individuals are perceived more positively by others – other individuals perceive those with high EI to be more pleasant, socially skilled and empathic to be around (Mayer, 2008).
- Better family and intimate relationships – High EI is correlated with better relationships with the family and intimate partners on many aspects.
- Better academic achievement – Emotional intelligence is correlated with greater achievement in academics as reported by teachers but generally not higher grades once the factor of IQ is taken into account (Mayer, 2008).
- Better social relations during work performance and in negotiations – Higher emotional intelligence is correlated with better social dynamics at work as well as better negotiating ability (Mayer, 2008).
- Better psychological well-being – Emotional intelligence is positively correlated with higher life satisfaction, self-esteem and lower levels of insecurity or depression. It is also negatively correlated with poor health choices and behaviour (Mayer, 2008).
- Allows for self-compassion – Emotionally intelligent individuals are more likely to have a better understanding of themselves, and to make conscious decisions based on emotion and rationale combined. Overall, it leads a person to self-actualisation (Heffernan et al., 2010).
Criticisms of Theoretical Foundation
Cannot be recognised as form of intelligence:
Goleman’s early work has been criticised for assuming from the beginning that EI is a type of intelligence or cognitive ability.
(Eysenck, 2000, p.109) writes that Goleman’s description of EI contains unsubstantiated assumptions about intelligence in general, and that it even runs contrary to what researchers have come to expect when studying types of intelligence:
“[Goleman] exemplifies more clearly than most the fundamental absurdity of the tendency to class almost any type of behaviour as an ‘intelligence’… If these five ‘abilities’ define ’emotional intelligence’, we would expect some evidence that they are highly correlated; Goleman admits that they might be quite uncorrelated, and in any case if we cannot measure them, how do we know they are related? So the whole theory is built on quicksand: there is no sound scientific basis.”
Similarly, (Locke, 2005) claims that the concept of EI is in itself a misinterpretation of the intelligence construct, and he offers an alternative interpretation: it is not another form or type of intelligence, but intelligence – the ability to grasp abstractions – applied to a particular life domain: emotions. He suggests the concept should be re-labeled and referred to as a skill.
The essence of this criticism is that scientific inquiry depends on valid and consistent construct utilisation, and that before the introduction of the term EI, psychologists had established theoretical distinctions between factors such as abilities and achievements, skills and habits, attitudes and values, and personality traits and emotional states (Mattiuzzi, 2008).
Thus, some scholars believe that the term EI merges and conflates such accepted concepts and definitions.
Confusing skills with moral qualities:
Adam Grant warned of the common but mistaken perception of EI as a desirable moral quality rather than a skill, Grant asserting that a well-developed EI is not only an instrumental tool for accomplishing goals, but has a dark side as a weapon for manipulating others by robbing them of their capacity to reason (Grant, 2014).
Has little predictive value:
(Landy, 2005) claimed that the few incremental validity studies conducted on EI have shown that it adds little or nothing to the explanation or prediction of some common outcomes (most notably academic and work success). Landy suggested that the reason why some studies have found a small increase in predictive validity is a methodological fallacy, namely, that alternative explanations have not been completely considered:
“EI is compared and contrasted with a measure of abstract intelligence but not with a personality measure, or with a personality measure but not with a measure of academic intelligence” (Landy, 2005, p.419).
Similarly, other researchers have raised concerns about the extent to which self-report EI measures correlate with established personality dimensions.
Generally, self-report EI measures and personality measures have been said to converge because they both purport to measure personality traits (Petrides, Pita and Kokkinaki, 2007).
Specifically, there appear to be two dimensions of the Big Five that stand out as most related to self-report EI – neuroticism and extraversion.
In particular, neuroticism has been said to relate to negative emotionality and anxiety. Intuitively, individuals scoring high on neuroticism are likely to score low on self-report EI measures.
The interpretations of the correlations between EI questionnaires and personality have been varied.
The prominent view in the scientific literature is the Trait EI view, which re-interprets EI as a collection of personality traits (Mikolajczak et al., 2007; Smith, CIarrochi and Heaven, 2008; Austin, 2008).
Criticisms of Measurement
Ability model: measures conformity, not ability:
One criticism of the works of Mayer and Salovey comes from a study by (Roberts, Zeidner and Matthews, 2001) which suggests that the EI, as measured by the MSCEIT, may only be measuring conformity.
This argument is rooted in the MSCEIT’s use of consensus-based assessment, and in the fact that scores on the MSCEIT are negatively distributed (meaning that its scores differentiate between people with low EI better than people with high EI).
Ability Model: measures knowledge, not ability:
Further criticism has been leveled by (Brody, 2004, p.234), who claimed that unlike tests of cognitive ability, the MSCEIT “tests knowledge of emotions but not necessarily the ability to perform tasks that are related to the knowledge that is assessed”.
The main argument is that even though someone knows how he or she should behave in an emotionally laden situation, it doesn’t necessarily follow that the person could actually carry out the reported behaviour.
Ability Model: measures personality and general intelligence:
New research is surfacing that suggests that ability EI measures might be measuring personality in addition to general intelligence.
These studies examined the multivariate effects of personality and intelligence on EI (Cavazotte, Moreno and Hickman, 2012) and also corrected estimates for measurement error (which is often not done in some validation studies).
For example, a study by (Schulte, Ree and Carretta, 2004), showed that general intelligence (measured with the Wonderlic Personnel Test), agreeableness (measured by the NEO-PI), as well as gender could reliably be used to predict the measure of EI ability.
They gave a multiple correlation (R) of .81 with the MSCEIT (perfect prediction would be 1).
This result has been replicated by Fiori and Antonakis (2011), they found a multiple R of .76 using Cattell’s “Culture Fair” intelligence test and the Big Five Inventory (BFI); significant covariates were intelligence (standardized beta = .39), agreeableness (standardized beta = .54), and openness (standardized beta = .46). (Antonakis and Dietz, 2011a), who investigated the Ability Emotional Intelligence Measure found similar results (Multiple R = .69), with significant predictors being intelligence, standardized beta = .69 (using the Swaps Test and a Wechsler scales subtest, the 40-item General Knowledge Task) and empathy, standardized beta = .26 (using the Questionnaire Measure of Empathic Tendency) – see also (Antonakis and Dietz, 2011b), who show how including or excluding important controls variables can fundamentally change results – thus, it is important to always include important controls like personality and intelligence when examining the predictive validity of ability and trait EI models.
Self-report measures susceptible to faking
More formally termed socially desirable responding (SDR), faking good is defined as a response pattern in which test-takers systematically represent themselves with an excessive positive bias (Paulhus, 2002).
This bias has long been known to contaminate responses on personality inventories (Holtgraves, 2004; McFarland and Ryan, 2000; Peebles and Moore, 1998; Nichols and Greene, 1997; Zerbe and Paulhus, 1987), acting as a mediator of the relationships between self-report measures (Nichols and Greene, 1997; Gangster et al., 1983).
It has been suggested that responding in a desirable way is a response set, which is a situational and temporary response pattern (Pauls and Crost, 2004; Paulhus, 1991).
This is contrasted with a response style, which is a more long-term trait like quality.
Considering the contexts some self-report EI inventories are used in (e.g. employment settings), the problems of response sets in high stakes scenarios become clear (Paulhus and Reid, 2001).
There are a few methods to prevent socially desirable responding on behaviour inventories.
Some researchers believe it is necessary to warn test takers not to fake good before taking a personality test (e.g., McFarland, 2003).
Some inventories use validity scales in order to determine the likelihood or consistency of the responses across all items.
Predictive power unsubstantiated
(Landy, 2005) distinguishes between the “commercial wing” and “the academic wing” of the EI movement, basing this distinction on the alleged predictive power of EI as seen by the two currents.
According to Landy, the former makes expansive claims on the applied value of EI, while the latter is trying to warn users against these claims. As an example, (Goleman, 1998) asserts that “the most effective leaders are alike in one crucial way: they all have a high degree of what has come to be known as emotional intelligence. …emotional intelligence is the sine qua non of leadership”.
In contrast, Mayer (1999) cautions “the popular literature’s implication – that highly emotionally intelligent people possess an unqualified advantage in life – appears overly enthusiastic at present and unsubstantiated by reasonable scientific standards.”
Landy further reinforces this argument by noting that the data upon which these claims are based are held in “proprietary databases”, which means they are unavailable to independent researchers for reanalysis, replication or verification (Landy, 2005).
Thus, the credibility of the findings cannot be substantiated in a scientific way, unless those datasets are made public and available for independent analysis.
In an academic exchange, Antonakis and Ashkanasy/Dasborough mostly agreed that researchers testing whether EI matters for leadership have not done so using robust research designs; therefore, currently there is no strong evidence showing that EI predicts leadership outcomes when accounting for personality and IQ (Antonakis, Ashkanasy and Dasborough, 2009). Antonakis argued that EI might not be needed for leadership effectiveness (he referred to this as the “curse of emotion” phenomenon, because leaders who are too sensitive to their and others’ emotional states might have difficulty making decisions that would result in emotional labour for the leader or followers).
A recently published meta-analysis seems to support the Antonakis position: In fact, Harms and Credé found that overall (and using data free from problems of common source and common methods), EI measures correlated only ρ = 0.11 (Spearman’s rank correlation coefficient: closer to 1 = more correlation) with measures of transformational leadership (Harms and Credé, 2010). Ability-measures of EI fared worst (i.e., ρ = 0.04); the WLEIS (Wong-Law measure) did a bit better (ρ = 0.08), and the Bar-On (Bar-On, 2006) measure better still (ρ = 0.18).
However, the validity of these estimates does not include the effects of IQ or the big five personality, which correlate both with EI measures and leadership (Antonakis, 2009). In a subsequent paper analyzing the impact of EI on both job performance and leadership, (Harms and Credé, 2010) found that the meta-analytic validity estimates for EI dropped to zero when Big Five traits and IQ were controlled for.
(Joseph and Newman, 2010) meta-analytically showed the same result for Ability EI.
However, it is important to note that self-reported and Trait EI measures retain a fair amount of predictive validity for job performance after controlling Big Five traits and IQ (Joseph and Newman, 2010). (Newman, Joseph and MacCann, 2010) contend that the greater predictive validity of Trait EI measures is due to their inclusion of content related to achievement motivation, self efficacy and self-rated performance.
Meta-analytic evidence confirms that self-reported emotional intelligence predicting job performance is due to mixed EI and trait EI measures’ tapping into self-efficacy and self-rated performance, in addition to the domains of Neuroticism, Extraversion, Conscientiousness, and IQ.
As such, the predictive ability of mixed EI to job performance drops to nil when controlling for these factors (Joseph et al., 2005).
NICHD pushes for consensus
The National Institute of Child Health and Human Development has recognised that because there are divisions about the topic of emotional intelligence, the mental health community needs to agree on some guidelines to describe good mental health and positive mental living conditions. In their section, “Positive Psychology and the Concept of Health”, they explain:
“Currently there are six competing models of positive health, which are based on concepts such as being above normal, character strengths and core virtues, developmental maturity, social-emotional intelligence, subjective well-being, and resilience. But these concepts define health in philosophical rather than empirical terms. Dr. [Lawrence] Becker suggested the need for a consensus on the concept of positive psychological health…” (Nitkin, 2006).