Kagawa Entrance English common mistakes
Kagawa Entrance English · Understand the traps and how to avoid them.
Reading a passage once at normal speed and then answering detail questions from memory, choosing the option that 'feels right' rather than returning to the line that supports it.
Under time pressure candidates treat the passage as a gist task and trust a first impression; the distractors are written to match that impression, often reusing exact words from the passage in a false relationship.
What to do instead
For every detail or inference item, locate and mentally underline the one sentence that decides the answer, and reject any option you cannot tie to that sentence. Practice with a two-pass habit: skim for structure first, then answer with line-by-line verification.
Choosing an answer to a 'what does the author imply / what is the main idea' question that is true in real life but not stated or supported in the passage.
Candidates import background knowledge, especially on familiar topics like environment or technology, and mistake plausibility for textual support.
What to do instead
Before selecting, ask 'Which words in the passage would I point to?' If the answer requires information from outside the text, eliminate it. For main-idea items, check that the option covers the whole passage, not just one paragraph.
Answering grammar items by ear, picking the form that 'sounds natural' in Japanese-influenced English, such as 'I have been to Kyoto last year' or 'He suggested me to go'.
Classroom and textbook English plus first-language transfer create strong but wrong patterns; the ear recognises the translated structure, not the target rule.
What to do instead
Build a short personal error list of the five or six structures that keep appearing (tense with time markers, suggest/recommend + that-clause or gerund, articles with countable nouns). For each, memorise one model sentence and test new items against it explicitly.
Treating every blank in a cloze or grammar section as an isolated word choice, ignoring the tense, subject, and reference words in the surrounding sentences.
Items are read left to right and answered locally; candidates do not re-read the full sentence or the previous one to check agreement and sequence of tenses.
What to do instead
After choosing, re-read the complete sentence with the option inserted and check three things: subject-verb agreement, tense consistency with time expressions, and pronoun/antecedent match. Do this aloud in practice until it becomes automatic.
Learning vocabulary as Japanese-English word pairs only, so a word is recognised in isolation but not in collocations or with a changed part of speech, and confusable pairs (affect/effect, adapt/adopt, economic/economical) are mixed up.
Pair-list memorisation gives one meaning per word and no context; exam items test words inside phrases and in less common senses.
What to do instead
Record new words in a short phrase or sentence with a collocation (e.g. 'take measures against', 'a marked increase'), plus the part of speech and one antonym or confusable partner. Review by covering the word and recalling it from the phrase, not the reverse.
In sentence-ordering or error-identification items, building the sentence by translating a Japanese word order and stopping at the first arrangement that is grammatical, without checking whether it is the intended meaning.
Japanese puts the verb last and allows flexible scrambling; candidates accept any order that parses and do not test cohesion markers such as pronouns, connectors, and given-new information flow.
What to do instead
Identify the fixed anchors first (subject, main verb, any connective that must open a clause), then place the remaining pieces so that pronouns follow their referents and connectors sit at the start of their clause. Write out the final sentence in full and read it once for meaning.
Writing compositions that list opinions with no linking devices, repeat the same sentence pattern ('I think ... I think ...'), and exceed or fall short of the required length.
Candidates compose directly onto the answer sheet, translating from a Japanese plan, so paragraph logic and word count are never planned or checked.
What to do instead
Spend two minutes on a three-line outline (position, two reasons, conclusion) and assign one connector per paragraph (however, for example, therefore). Leave two minutes at the end to count words and replace repeated sentence openings with a participle or relative clause.
In listening sections, writing the answer while the next item is already being read, then losing the following question entirely.
Candidates try to capture every word rather than the answer, and they do not preview the question stems and options during the pause.
What to do instead
Use the pause before each section to underline the question word (who, when, why) and predict the answer type. During the audio, mark only a keyword or symbol; write the full answer after the item ends. Practise with single playbacks only, never replaying.
Missing negation, numbers, and contrast signals in listening and reading (not, hardly, unless, however, actually), so the answer is the opposite of what was said.
These words are unstressed or quickly spoken, and candidates listen for content nouns only; in writing they skim over the same markers.
What to do instead
Train specifically on minimal pairs and on sentences where a negation reverses meaning. In practice, transcribe short extracts and circle every negative and contrast word, then check how each one changes the answer.
Treating integrated tasks (summarising or responding to a passage plus a spoken or written source) as two separate answers, copying phrases from the source, or adding a personal opinion where only synthesis is asked.
Candidates do not identify the task's purpose before reading, and paraphrasing under time pressure feels riskier than lifting words.
What to do instead
Before reading, write the task verb and the required content in one line. Then note source points in note form and combine them with your own linking language, changing structure and word choice rather than copying. Check that the response answers the exact prompt and attributes each point to its source.