Scoring
How is PCAT scored?
PCAT scores every section with AI, using criteria that put whether the message got through first. This page explains how each skill is scored and what the report shows.
Listening and Reading
How are Listening and Reading scored?
Listening and Reading are adaptive. Each question is chosen according to your earlier answers, so you spend less time on questions that are far too easy or far too hard. Your score is not the number of questions you got right: it is estimated from which questions you answered correctly and how difficult they were. Two people who answered different questions can therefore be placed on the same scale. Listening uses meetings with three or four speakers, calls and announcements, each played once; Reading uses work chat and short specifications.
Speaking and Writing
How are Speaking and Writing scored?
Each response is scored against two layers. Communication — task achievement, intelligibility, appropriateness, coherence and strategic competence — makes up 85% of the score; Language Quality — grammar, vocabulary and, for speaking, pronunciation — makes up 15%. Speaking responses are transcribed and assessed for pronunciation, then scored against the same criteria as writing. Errors are classified by their effect: an error that stops the listener or reader from understanding lowers the communication score, while an error the meaning survives only affects language quality.
Scoring principle
How much do grammar mistakes affect the score?
PCAT does not treat every error the same way. In speaking and writing, errors that interfere with meaning, understanding or getting the task done are separated from errors that are inaccurate but still let the message through. Communication — was the task done, was the message understood, did it fit the situation — makes up 85% of the score; Language Quality makes up 15%. Minor slips do not dominate the result, and accuracy still counts.
- Communication
- Task Achievement — did the task get done30%
- Intelligibility — can the listener understand22%
- Appropriateness — fit for the situation and audience18%
- Coherence & Cohesion — logical structure10%
- Strategic Competence — rephrasing and repair5%
- Language Quality
- Grammar · Vocabulary · Pronunciation — shown as bands15%
An error the other person cannot decode, or decodes wrongly. It hits Intelligibility directly and pulls the total score down.
An error the meaning survives. It only moves the Language Quality bands and barely affects the total.
The report
What does the score report show?
You receive a PCAT Score and an overall CEFR-aligned level, a score and level for each of the four skills, dimension-level scores with feedback for speaking (eight dimensions) and writing (six), and strengths and areas to work on. The PCAT Score runs from 80 to 230, with each CEFR band covering 20 points: below 100 is below A1, A1 starts at 100, A2 at 120, B1 at 140, B2 at 160, C1 at 180 and C2 at 200. Within a band, .1 marks the lower half and .2 the upper half — a PCAT sub-band, not a CEFR level of its own (B1.2 is the upper half of B1). The report also records the test conditions, such as whether a camera was used. Results arrive within 48 hours.
AI scoring
Who does the scoring?
AI does the scoring, with human review where it is needed. Each speaking and writing response is scored three times independently and the median is used, which evens out the variation of a single run. Responses where the three scores disagree, scores are unusually high or low, or speech recognition was uncertain are placed in a review queue; our staff check them and correct the score if needed, and a corrected report says so. Speaking transcription and pronunciation assessment use external AI services; the scoring criteria are PCAT’s own and are the ones described on this page. Your answers, recordings and video are not used to train general-purpose AI models. See the privacy policy for how data is stored and handled.
Take it once and see
From JPY 5,500 per credit (tax included). Credits are valid for one year.