What 2,417 people got on the Socionics quiz

The quiz on this site has been recording aggregate results since June. Not who took it, not what anyone answered, just which of the sixteen types each completed run landed on. Here is the full picture, and the honest reading of what it means.

Socionics quiz results by type, 2,417 completed runs Horizontal bar chart of the sixteen Socionics types, ordered by how often each came up as a quiz result. ILI is the most common at 14.7 per cent and ESE the rarest at 1.1 per cent. The same sixteen figures are given as a table below the chart. ILI 14.7% IEI 14.0% LII 10.8% SLI 9.5% LSI 9.2% EII 8.9% SEI 7.2% ESI 6.3% ILE 4.6% IEE 3.7% SLE 2.5% EIE 2.3% LIE 2.2% LSE 1.7% SEE 1.3% ESE 1.1%
Socionics Insight quiz results, 17 June 2026 to 16 October 2026. n = 2,417.
Socionics quiz results by type, 17 June 2026 to 16 October 2026, n = 2,417.
TypeResultsShare
ILI35614.7%
IEI33914.0%
LII26010.8%
SLI2309.5%
LSI2239.2%
EII2158.9%
SEI1757.2%
ESI1526.3%
ILE1114.6%
IEE903.7%
SLE602.5%
EIE552.3%
LIE522.2%
LSE401.7%
SEE321.3%
ESE271.1%

The obvious finding, and why it is not what it looks like

Introverted types account for 80.7% of every result. The top two, ILI and IEI, take 28.8% between them on their own. ESE comes last with 27.

That is a real pattern in who reaches this page and finishes twelve questions. It is not a finding about how common these types are in the population, and I would not want anyone citing it that way. Online typology audiences have skewed introverted and intuitive in every dataset I have seen, and this one does nothing to break that.

Nearly every result is a close call

93.6% of results came out tied or one point apart on at least one dichotomy. That is 2,262 of 2,416 runs where the flag was recorded.

This is what a twelve question forced choice instrument does. Three questions per dichotomy means the only possible outcomes are 3-0 and 2-1, and 2-1 is a single answer away from a coin flip. The quiz was built as a starting point rather than a verdict and this is the number that proves it. If your result felt borderline, it almost certainly was.

The quadras come out level

Split the same data by quadra and the skew largely disappears.

Socionics quiz results by quadra, 2,417 completed runs Horizontal bar chart of the same 2,417 results grouped into the four quadras. The same four figures are given as a table below the chart. Alpha 23.7% Beta 28.0% Gamma 24.5% Delta 23.8%
The same 2,417 results, grouped by quadra.
Socionics quiz results by quadra, n = 2,417.
QuadraResultsShare
Alpha57323.7%
Beta67728.0%
Gamma59224.5%
Delta57523.8%

Alpha, Beta, Gamma and Delta land within 5 points of each other despite the type level spread being 13 to one from top to bottom.

I do not have a confident explanation for that. It may be that the dichotomy questions push hard on the individual axes while leaving the underlying value clusters roughly balanced. It is the finding I would most like more data on.

What this data is

Aggregate event counts from Umami, the privacy-first analytics the site runs. No accounts, no cookies, no personal data, nothing that identifies anyone. The only thing recorded on completion is the resulting type code, whether the run arrived via the quiz or via a shared result link, and whether the result was borderline. 2,394 of the 2,417 were genuine completions rather than shared links.

Snapshot taken 17 October 2026. I will refresh it as the sample grows.

Nine in ten quiz results are borderline

A written analysis works from your full responses and rules out the close alternatives, with the reasoning shown.

From £40 · written report in 5 working days →