Tell an ability measure from a self-report trait measure, and say what each one can and cannot tell you.
Two tests can both be labelled "emotional intelligence" and still measure different things, and when they do, neither test is broken. Before you trust any score, including one about yourself, find out which kind of test produced it.
Hui Min works in HR at a logistics company in Jurong East (an invented example). Her manager wants an EI test for the new team leads programme, and two vendors send brochures. To see what each one is like, she and a colleague each try both tests.
On the first test, Hui Min scores high and her colleague scores low. On the second test, the result is reversed. Nobody cheated and neither test is faulty. To see why, you need to know how researchers sort EI and what kind of test each approach uses.
Researchers usually sort EI into three families. Ability models treat EI as a set of mental abilities, something like a narrower kind of intelligence. Trait models treat it as part of personality: how you see yourself and your own feelings. Mixed models combine emotional abilities with personality, motivation and social skill.
Each family has its own kind of test. Most of the disagreement in the field starts here, because tests with the same label can ask completely different questions.
An ability test works like an exam. You have to produce answers, and you can't get credit just by saying you're good at something. One task might show you a photo of a face and ask which emotions it shows and how strongly. Another might describe a colleague who has just been passed over for promotion and ask how their feelings are likely to change over the next week. A third might ask which mood helps most when brainstorming and which helps when checking figures.
The best-known ability test is the MSCEIT (Mayer-Salovey-Caruso Emotional Intelligence Test), built by John Mayer, Peter Salovey and David Caruso. It is built around the four abilities you met in lesson 1.1.
Tests like this run into a scoring problem: there is no answer key for how a person in a story feels. The MSCEIT deals with this in one of two ways. Expert scoring compares your answers with those of a panel of emotion researchers. Consensus scoring compares them with a large sample of ordinary test takers. The closer your answers are to what the experts or most people chose, the higher your score.
An ability score shows how well someone did on those tasks, on that day, in a quiet room. It can't show whether they use the skill under real pressure, for example when a manager criticises them in front of the team at 6pm. Knowing that a reply will upset someone and choosing not to send it are separate things, and a test checks only the first.
Trait tests are self-report. They give you statements such as "I can usually tell how I am feeling" or "I find it hard to stay calm under pressure", and you rate how well each one describes you. The researcher most associated with trait EI is K. V. Petrides. His work describes trait EI as a set of self-perceptions that sit within personality.
Self-report questionnaires capture something real. Confidence in emotional situations affects what people attempt, and someone who believes they can handle a hard conversation is more likely to start one.
Self-report has two limits. First, people can only rate what they can see in themselves, and the habits that annoy others are often the ones people notice least. Second, it is easy to answer as the person you would like to be, especially when a promotion or a programme place depends on the result.
The best-known mixed model is Goleman's, from his 1995 book. Besides the emotional abilities, it includes motivation, self-confidence, social skill and qualities that read like a description of a good colleague. Many workplace questionnaires and 360-degree reviews draw on the mixed-model tradition.
The problem with mixed models is overlap. Ordinary personality tests already measure optimism, drive, conscientiousness and sociability. A high mixed-model score may mainly show that someone is calm, organised and easy to get along with, which a standard personality questionnaire would also show. That makes it hard to say what the EI part adds. Lesson 1.3, "What the research supports and what it oversells", picks up that question.
Back in Jurong East, the reversed results come from the two kinds of test. Hui Min's colleague is confident and generous with herself, so she did well on the self-report questionnaire. Hui Min rates herself modestly and did better on the ability tasks. Both scores are honest, and they answer different questions. Hui Min recommends neither vendor's test as a selection tool. She suggests using one only as a conversation starter in the programme.
You can sort out most tests quickly by asking three questions:
Did I perform tasks, or did I rate myself? Who decided what counted as a good answer? What does the test leave out entirely, such as how I behave when tired, rushed or provoked?
The first question tells you which family the test belongs to. The second tells you whose judgement the score reflects: experts, most people, or your own view of yourself. The third shows you the gaps, which help you decide how much weight the score deserves.
Most free online quizzes are short self-report questionnaires, and few of them publish any evidence for what they claim. They're a reasonable prompt for reflection and a poor basis for any decision about yourself or anyone else.
In the activity below you try one of these quizzes yourself. Read it the way Hui Min learned to read the brochures, as a test of something specific with gaps you can name.
Take one free online emotional intelligence quiz. Note whether it asks you to rate yourself or tests your performance, then write down two things about you it cannot measure.
Junxiong-WFG Organisation is an authorised representative of AIA Financial Advisers Private Limited (Reg. No. 201715016G).