EJU Japan and the World common mistakes
EJU Japan and the World · Understand the traps and how to avoid them.
Reading the Japanese passage word-by-word and translating each sentence into your first language before moving on, so you run out of time on the 読解 section and leave the final long passage blank.
School foreign-language habits train literal translation, but EJU 読解 rewards skimming for the writer's claim and paragraph function; the passage volume is calibrated so that a translating reader cannot finish.
What to do instead
Practise timed passes where you first read only the question stems, then hunt for the paragraph containing the answer using discourse markers (しかし, つまり, したがって, たとえば). Train yourself to identify the topic sentence and the conclusion sentence of each paragraph without translating the middle.
Treating every 読解 question as a search-and-copy task and choosing an answer option that repeats wording from the passage verbatim.
Distractor options are built by copying surface vocabulary from the text while reversing or overgeneralising the logic; candidates who match on keywords rather than on propositional meaning get trapped.
What to do instead
For each option, underline the exact clause in the passage it would need to be true, and check the logical connector and the scope (all/some, always/sometimes). Eliminate options that add absolutes (必ず, すべて, 決して) that the passage does not license.
Answering 聴解 questions from the first number, name, or date you hear, then stopping listening while the speaker corrects or revises it.
EJU listening items routinely embed a self-correction (いや、やっぱり, あ、違う) or a conditional (〜なら, 〜の場合) after the initial mention; candidates who write the first value down miss the revision.
What to do instead
Take abbreviated notes in a fixed column layout (person, action, time, place) and keep writing until the item ends. Mark provisional values with a symbol and overwrite them when a correction marker appears. Drill with audio played once only, as in the real test.
Ignoring 聴読解 diagram and chart conventions, such as which axis is time, which arrow means increase, or which symbol marks the speaker's own choice.
Candidates focus on the audio and glance at the figure only after it is too late; the figure carries information (legend, units, annotations) that the audio assumes you have already absorbed during the pre-reading pause.
What to do instead
In the instruction pause, systematically scan every figure: title, axis labels, units, legend, and any handwritten-style annotations. Predict which categories the audio will compare before it starts, so you only need to tick boxes as you listen.
Writing the 記述 essay as a personal opinion piece with anecdotes and no engagement with the prompt's stated positions.
Candidates transfer first-language essay habits (narrative, emotion, rhetorical questions) that do not satisfy the EJU rubric, which rewards accurate comprehension of the prompt, a clear stated position, and reasoned support.
What to do instead
Use a fixed structure: state your position in the first sentence, summarise the competing view fairly in one or two sentences, give two concrete reasons with examples, acknowledge a counterargument, and restate your conclusion. Keep the register formal (だ・である体) and avoid first-person anecdotes.
Losing points on 記述 for mechanical issues: running over or under the word limit, using です・ます inside a だ・である essay, or writing kanji you cannot actually produce.
Under exam pressure candidates default to polite forms learned in conversation classes and pad or trim content at the last minute, and they attempt low-frequency kanji that come out illegible or wrong.
What to do instead
Before the exam, fix your register and practise counting characters to hit the target band with a margin. In the exam, write only kanji you can render cleanly; use simple, correct vocabulary over ambitious vocabulary. Reserve two minutes to check register consistency and character count.
Memorising vocabulary as isolated word lists with single first-language glosses, then misreading words whose meaning shifts with context or collocation.
EJU 語彙 items and reading passages test words in context (e.g. かける, とる, あがる with multiple senses); a one-to-one gloss produces wrong readings of idioms and compound verbs.
What to do instead
Learn vocabulary in collocations and short sentences, noting the particle and the register. Keep a log of multi-sense verbs and 慣用句 with one example sentence each, and review by covering the Japanese and producing it from the situation, not from the gloss.
Skipping or rushing the 数学 section's reading of the question stem and solving a problem you assumed rather than the one asked.
EJU mathematics stems are short but dense with conditions (整数, 実数, ただし, のとき); candidates pattern-match to a familiar problem type from their home curriculum and answer a different question.
What to do instead
Underline every condition and constraint before starting. After solving, re-read the stem and verify that your answer satisfies each stated condition (domain, sign, integer requirement). Practise problems where an unassuming condition (e.g. x is an integer) changes the answer.
Assuming the 理科 or 総合科目 syllabus matches your home country's curriculum and skipping topics your school never covered.
EJU 総合科目 in particular blends economics, politics, geography, history, and contemporary society with a Japan-relevant framing, and 理科 subject choices have specific scope statements; candidates discover gaps only in the exam.
What to do instead
Obtain the official syllabus scope for your chosen subjects and audit it topic by topic against your notes, marking gaps explicitly. Build a checklist and schedule the missing topics first, since unfamiliar content costs more marks than polishing known content.
Managing the clock uniformly across sections instead of budgeting by item difficulty, so that a single hard 読解 passage or 数学 problem consumes time needed for easy marks elsewhere.
Candidates are used to finishing every question in school tests and feel uncomfortable leaving a blank; perfectionism on one item cascades into unanswered easy items later in the same section.
What to do instead
Set a per-item time cap in practice (e.g. roughly one minute per 読解 question) and enforce a skip rule: if an item exceeds the cap, mark it and move on. Do two full mock runs under real timing to calibrate the cap, and always answer every item since there is no penalty for guessing.