Emotional intelligence (EI) enjoyed its 30th birthday last year and is fast becoming middle-aged, that is, if we take the writings of Peter Salovey & John (Jack) Mayer (1990) as the starting point for modern research into EI.  This anniversary gives us much to reflect on – certainly for me personally, as a researcher who has spent over a decade investigating EI (a third of its life) – but also for those of us interested in psychology more broadly and the study of individual differences. 

I began my EI adventure in 2008.  At that time EI was everywhere, and it had become a catch-all buzzword for improving society’s ills – take, for example, David Cameron’s infamous speech about emotion and compassion being at the core of understanding the causes of youth crime, which became known as the ‘hug a hoodie’ campaign. Self-help books were springing up regularly, telling us how to be more emotionally smart and how this was the ticket to solving all of our interpersonal problems.   This explosion of interest hadn’t escaped the attention of the UK government, who invested significant resources in school-based social and emotional learning programmes  in an effort to train EI in young people.  But we actually had very little scientific evidence about what EI was, whether training this was possible, and whether this made a tangible difference to young people in terms of academic success and mental wellbeing. 

In the academic world, EI was viewed as either representing a collection of specialist skills for perceiving, understanding, and managing emotions in ourselves and others, or as a collection of emotion-related personal qualities and competencies. This seemed to split opinion amongst researchers; some suggested EI was too ill-defined to be meaningful, others argued it was nothing new, and we were just measuring things we already knew a good deal about, like personality or intelligence, and cautioned against falling into the “jingle” or “jangle” fallacy trap.  Yet there were others still, who wholeheartedly embraced this as a new form of intelligence that could predict success in all areas of life – here is a detailed summary of the main arguments for those that want to delve more deeply.

As a newly minted PhD researcher, all of this debate and uncertainty meant there was plenty for me to explore.  I set about examining the ‘what’, ‘when’, ‘why’, and ‘how’ of EI, starting with finding out whether EI was actually useful for young people who were experiencing a range of different (everyday) stressors, but I found that this wasn’t clear-cut – if we measured EI through self-reports (as emotional self-efficacy), higher levels were useful for maintaining lower levels of symptoms in young people experiencing family dysfunction.  But if we measured EI as an ability (or skill set through IQ style testing), very high levels were counter-productive for those young people experiencing socio-economic adversity. 

With colleagues, we then went on to find other interesting (and potentially counter-intuitive) things.  We confirmed that the way that you measure EI matters because this predicts different ways of coping and engaging with emotion (here and here), but that universally ‘high’ levels of EI may not always be beneficial in all contexts – here is an open access review we wrote on this.  


So emotional skills and self-efficacy are different sides of the EI coin, but doubt remains over whether we are actually capturing an ‘intelligence’ and whether this is always a good thing.

Fast forward to 2020, and some of our most recent work has, in a sense, gone forward into the past to try to answer some of these questions.   One of the core tenets of EI theory is that emotionally intelligent individuals should be adept at recognising emotional expressions in others to facilitate effective social interactions.

Photo by Simon Schwyter on Unsplash

We tested 92 people in the lab, to see whether those with high scores on measures of EI could recognise dynamic emotional expressions at an earlier, more subtle stage of presentation than those with lower scores.  Using the Emotion Recognition Task, we showed people a set of video clips of faces morphing from neutral expressions into different emotions, from subtle through to intense levels, and found that: 1) emotional self-efficacy (self-reported EI) was not associated with recognition performance, 2) higher skill in emotional understanding did relate to better performance of emotion cues (subtle through to intense and particularly negative emotion) but that, 3) fluid cognitive ability was a better predictor of performance. In a separate study with young people, we found EI skills did not predict accurate recognition of emotion in voices.

So, this continues to raise questions about whether existing measures of EI are sensitive enough to represent individual differences in socially relevant aspects of emotion recognition, and also the validity of EI as a discrete, and meaningful, individual differences factor.  

What can we say about all of this?  

Three decades on, and it seems we are still grappling with key questions from the past.  Many researchers picked EI up and ‘ran with it’, without us really knowing what makes someone emotionally intelligent, or us being able to measure EI ‘in action’.  I don’t doubt that for most researchers, this urgency came from a good place, and a desire to make a difference to people’s lives through intervention.  However, a critical eye is needed. 

I believe there is room for further high-quality scientific work which could still lead us to make progress with EI measurement. But we will only succeed if we better integrate with the work going on in Emotion sciences more broadly.  For instance, the constructed theory of emotion tells us that rather than being ‘triggered’ in situations, our emotional experience arises from brain-based predictions, created from prior experiences that are specific to our culture and upbringing.  So, developing rich and informative (early) experiences seems to be pivotal to train our system to learn about, and learn from, socio-emotional interactions.  Similarly, this suggests that our current methods of testing EI are way off the mark (e.g., identifying posed emotional expressions; multiple choice questions about effective emotion management strategies, etc).  We may find the future lies in technological solutions (such as virtual reality) and we could look to allied fields of study for clues e.g., recent developments in the gamification of cognitive testing. Until we are better able to capture authentic EI ‘in action’ I fear we will continue to chase our tails.

The future

This is far from being just another ‘academic’ debate – emotional intelligence is being readily embraced by companies interested in developing artificial intelligence systems for reading emotions in humans, which has far-reaching implications for us all…just how reliable are facial movements for judging how someone is actually feeling? What are the implications of using emotion AI for decision-making (e.g., screening for criminal intent; an angry student at school; suitability of a job applicant)? As we surge further ahead into the technological epoch, this it is something we should all care about.


Dr Sarah Davies

Dr Sarah Davis

Sarah is a Senior Lecturer in Psychology & the Interpersonal Relationships & Wellbeing Research Group Lead.  She is currently writing a new critical primer on EI for Oxford University Press with Dr Bérénice Mahoney.