How Rorschach Scoring Works
A Rorschach result is not an interpretation of your bat or your butterfly. It is the product of coding every answer on half a dozen dimensions, adding the codes up across all ten cards, and comparing the totals with reference ranges. This is how that happens under the Exner Comprehensive System and R-PAS — and what Kyōgai does to approximate it.
First, the two phases
Nothing is scored until the whole session is recorded. In the Response phase you say what each card might be. In the Inquiry phase the examiner revisits each answer with neutral questions — "Where do you see that?", "What made it look like that?" — because a percept cannot be coded without knowing which part of the blot it used and which features drove it. See what is the Rorschach test for the full procedure.
Coding each response
Every single answer gets a string of codes. The main categories:
1. Location
Which part of the blot was used: the whole blot (W), a common detail area (D), an unusual detail (Dd), and whether the white space was incorporated (S). The balance of W, D and Dd across a protocol says something about how a person approaches complex information — taking it all in at once, breaking it into sensible chunks, or focusing on small, odd pieces.
2. Developmental quality
How well-organised the percept is: a synthesised response that relates distinct objects to each other (DQ+), a single ordinary object (DQo), or something vague with no real form ("clouds", "smoke" — DQv).
3. Determinants
What feature of the blot produced the answer. This is the richest category:
- Form (F) — shape alone.
- Movement — human (M), animal (FM) or inanimate (m). "Two people dancing" is M.
- Chromatic colour — C, CF or FC depending on whether colour or form dominates.
- Achromatic colour (C′) — black, white or grey used as colour.
- Shading — texture (T, "furry"), diffuse shading (Y, "misty"), or depth/vista (V).
- Form dimension (FD) — perspective implied by shape rather than shading.
- Reflections and pairs — using the blot's symmetry as a mirror image, or seeing "two of" something.
Several determinants in one answer make a blend, which is itself a sign of complexity.
4. Form quality
How well the answer fits the actual contours of the blot: ordinary (o), unusual but defensible (u), or a poor fit (−). The published systems use large tables listing which answers, in which areas of which card, count as which. The proportion of good-fit answers is one of the best-supported Rorschach variables; it relates to perceptual accuracy and reality testing.
5. Content
What the thing is: whole human (H), part human (Hd), fictional human ((H)), animal (A), anatomy (An), botany, clothing, blood, fire, landscape, art, and so on. Human content and how humans are seen (whole, realistic, interacting) feed the interpersonal part of the summary.
6. Popular
Whether the answer is one of the handful of responses given by a large share of the reference population for that exact card area. Most adults give four to seven. Our guide to the ten cards lists the common ones.
7. Special scores
Flags for unusual reasoning or themes. Cognitive special scores mark odd language, implausible combinations ("a bat with human hands"), fused percepts or strained logic. Thematic scores mark aggressive movement (AG), cooperative movement (COP), morbid or damaged content (MOR), and personal justification (PER). Cognitive special scores are among the most researched variables because they index disordered thinking.
The structural summary
The codes are tallied and turned into ratios. A few of the headline ones, in plain terms:
| Variable | Roughly what it reflects | Typical adult range |
|---|---|---|
| R | Total number of responses — engagement and productivity | 17–27 |
| Lambda | Share of pure-form answers; higher = simpler, more economical processing | 0.3–1.0 |
| EB | Human movement vs. weighted colour: an "ideational" vs. "expressive" style | — |
| EA | Available psychological resources (M + weighted colour) | 6–10 |
| es | Experienced stimulation — demands currently being felt | — |
| D score / Adj D | Resources minus demands, standardised; a stress-tolerance gauge | −1 to +1, mode 0 |
| Affective ratio | Responses to the three colour cards ÷ the first seven | 0.5–0.75 |
| Egocentricity index | Reflections and pairs relative to R — self-focus | 0.30–0.44 |
| Populars | Conventionality of perception | 4–7 |
| Isolation index | Nature/landscape content relative to R | — |
R-PAS keeps the better-supported of these, drops the weaker ones, adds a few (notably a complexity score), and reports everything as standard scores against international norms, which makes it easier to read and harder to over-interpret.
From numbers to an interpretation
A clinician reads the summary in clusters — processing, mediation, ideation, affect, self-perception, interpersonal perception, controls and stress — and only draws a conclusion where several variables point the same way, and then only alongside an interview and other tests. A single out-of-range value means little on its own.
How Kyōgai scores your test
Kyōgai records your Response and Inquiry phases exactly as above, then sends the full transcript to a large language model with the Exner/R-PAS coding rules spelled out. The model codes every response on all seven dimensions with a written rationale, computes the structural summary, compares it with the approximate reference ranges above, and writes two things: a plain-language reflection with every piece of jargon removed, and an honest limitations paragraph.
Three limitations are worth stating plainly:
- The exact form-quality, popular and D-score conversion tables are copyrighted and not reproduced; the model applies its trained knowledge of them and says so.
- A self-administered, typed or spoken protocol is not the same as one taken in a room with an examiner.
- The output is a reflection for a curious adult, not a clinical assessment or diagnosis.
The full response-by-response coding and all the ratios are available on the results page under "technical details" if you want to see the workings.