How this test is scored
Most online typology tests derive their scoring weights empirically: they collect self-reports from people who already identify with a type, then weight each question by how those populations answered. That approach bakes in whatever stereotypes and self-typing errors exist in the sample. This test takes the opposite route: every weight is derived directly from Socionics function theory, and the full scoring pipeline is deterministic and inspectable.
What is measured
The test measures 8 bipolar dimensions: the four classic Socionics dichotomies (extraversion/introversion, intuition/sensing, logic/ethics, rationality/irrationality) plus four secondary axes (strategic/tactical, independent/collective, constructive/restructuring, abstract/concrete). On top of these, each of the 8 information elements (Ne, Ni, Se, Si, Te, Ti, Fe, Fi) is measured directly by its own dedicated question cluster.
Question design
Each of the 250 items is written to isolate a cognitive processing preference, not a behavior, a mood, or a value statement. Introversion is not conflated with low energy. Sensing is not conflated with anxiety or practicality stereotypes. Rationality (judgment-first processing) is kept distinct from strategic planning scope, which many tests mix together.
From answers to a type
Answers on the 5-point scale are centered around neutral internally. Dimension scores are weighted averages over all items touching that dimension. Function scores blend two signals: direct answers to that function's own items, and alignment between the function's theoretical profile and your dichotomy scores. A type score is dominated by its leading function, supported by its creative function, with small dichotomy adjustments separating near-ties. The asymmetric weighting is what allows, say, strong Te and moderate Ti to appear side by side instead of being forced into mirror images.
Transparency and limits
Your result page shows every dimension score, the full function ranking, all 16 type scores, and which axes each individual answer influenced. You can export everything as JSON and re-import it later; imported answers are re-scored by the same engine, so results are always reproducible. The test is in public beta: weights are refined based on feedback, and like every self-report instrument it measures how you see yourself, not an objective ground truth.