What is an average EQ score?

It depends on the test. On the MSCEIT — an ability-style emotional intelligence test — scores are standardized so the average is 100 with a standard deviation of 15, and most people land between roughly 85 and 115. On self-report EQ questionnaires there is no single published 'average', because each tool uses its own scale and its own sample, so a number only makes sense relative to the group that tool was normed on.

Why do EQ self-ratings come out higher than EQ test scores?

Because they measure different things. Research comparing self-report emotional intelligence questionnaires with ability-based tests like the MSCEIT has found close to zero correlation between them — people's beliefs about their emotional skills track their self-perception and personality, not their tested performance on emotion problems. Scoring yourself 'above average' on empathy is easy; solving an emotion-perception item under test conditions is another matter.

Is a high MSCEIT score proof of high emotional intelligence?

Not proof — the method is debated. The MSCEIT scores answers against expert consensus or a large sample, which means a high score reflects agreement with a popular answer key rather than an objectively correct one, and psychometricians have argued about that for years. It is the best-studied ability measure available, but treat any single instrument's number as evidence in conversation, not a verdict.

Does AssessWiki publish an average EQ score for its own assessments?

No — deliberately. Our EQ checks are self-report reflection tools delivered to an open, self-selected audience; we do not know the composition of that sample, so any average we printed would be a number wearing a lab coat. Each entry states what it measures and what it cannot tell you, and that is the honest limit.

Assessment norms & scores

Average EQ score: what "above average" emotional intelligence really means

Last reviewed: October 2026 · Reviewer: AssessWiki Editorial · 5 min read

The short answer: "average EQ" only exists inside a specific test. On the MSCEIT, an ability-style measure, the average is 100 by design. On the self-report questionnaires most people actually take online, there is no shared average at all — and the two kinds of tests barely agree with each other. Here is how to read an EQ number without fooling yourself.

Two families of EQ tests, two very different "averages"

Emotional intelligence measurement split into two camps early on, and the split decides what "average" means:

Ability tests — like the Mayer–Salovey–Caruso Emotional Intelligence Test (MSCEIT) — present emotion problems with better and worse answers. Its scores are standardized to a mean of 100 and a standard deviation of 15 (development and standardization reported in Mayer, Caruso, Salovey et al., 2025), so most test-takers fall between roughly 85 and 115, and "average" is a defined, published region of the scale.
Self-report questionnaires — most of what "EQ test" means online, and every EQ check in this library — ask you to rate your own empathy, regulation, and social awareness. Each tool uses its own items, its own scale, and its own convenience sample. There is no population-wide average to compare against, and a site claiming "the average score is 78" for a home-grown questionnaire is inventing a norm.

Why self-ratings run high — and it isn't cheating

When researchers correlate the two families, the result is stubbornly flat: a 2003 review (Fordke, Chandler, Mackie, Maultsby & Watson) found the two families — self-report EI questionnaires and ability-style tests — largely uncorrelated. People who rate themselves emotionally skilled often don't score that way on tested emotion problems, and vice versa. That doesn't make self-report worthless — it makes it a measure of self-perception, which is its own useful signal: how you see your emotional behavior shapes what you attempt. It just isn't the same number, and a self-rating where most respondents place themselves "above average" is describing the scale's floor, not the population's ceiling.

The honest limits of even the standardized score

Being centered at 100 doesn't settle the debate. The MSCEIT keys its items by expert consensus or large-sample agreement, and critics (e.g., Antonakis, 2003) have argued that "correct" answers agreed on by popular vote are not evidence of a mental ability the way, say, a solving-a-matrix score is. Defenders point to reliability, norm data, and meaningful correlations with life outcomes. The reasonable stance: treat any single EQ number — from any camp — as one informed voice in a conversation about your emotional skills, never as a verdict. The EQ vs IQ comparison covers what the construct itself can and cannot carry.

So what should you do with your result?

Skip the percentile hunt; use the shape of your own answers. Self-report EQ tools are at their best when they show asymmetry — a pattern where your self-awareness outruns your social awareness, or your regulation holds at work but not at home. That is actionable in a way "103 vs. average 100" never is.

If you want a comparison number anyway, demand three things:

A named norming sample (who, how many, when),
evidence the instrument was scored on ability rather than self-belief — or honesty that it wasn't,
the sample's representativeness stated, not footnoted.

Most consumer EQ pages fail all three, including, by design, ours — which is why ours print no average.

See your own pattern

Free, self-report EQ checks in this library — with the limits stated on the page, no norms invented:

Try the free workplace EQ check EQ test: how the tools differ More in this series: average personality scores

Sources

Mayer, J. D., Caruso, D. R., Salovey, P., Lin, I. Y., Hansma, B. J., Solomon, J., Sitarenios, G., & Romero Escobar, M. (2025). Measuring emotional intelligence with the MSCEIT 2: theory, rationale, and initial findings. Frontiers in Psychology, 16, 1539785. source ↗
Fordke, M., Chandler, J., Mackie, D., Maultsby, L., & Watson, M. (2003). Direct and mixed measures of emotional intelligence. Personality and Individual Differences, 35(5), 1177–1189.
Antonakis, J. (2003). On why “emotional intelligence” will not pass the g-sat test… Consulting Psychology Journal: Practice and Research, 55(4) — a critique of consensus-scoring methods.
Brackett, M. A., & Salovey, P. (2006). Assessing individual differences in emotional intelligence (MSCEIT scoring methods). Psicothema, 18(SUPPL.), 36–43. source ↗

AssessWiki is an independent assessment library. Statements about score distributions reflect published measurement research, not AssessWiki's own data; we cite that research above. We publish no numeric ratings, and placement is never for sale. This page is educational, not a clinical or hiring recommendation.