Notes That Cost Him the Answer
Ben's Listening mock was 18/40. The audio is played once. His answer sheet shows a consistent shape: in each of the four sections the first two or three items are correct, then accuracy falls away sharply. The timestamps he annotated afterwards cluster his misses 4–6 seconds after signposting phrases - 'the thing is', 'however', 'what we found was'. He writes full sentences while listening: his notes for a five-minute section run to about 90 words. He is reliable on form completion (names, numbers, dates) and loses most marks on multiple-choice and matching. He has spent a month on 'just listen to more English' with no measurable change.
1. Diagnose Ben's problem. Is it vocabulary, memory, or attention allocation? Argue from the evidence and say what would falsify your diagnosis.
Read the worked answer
Attention allocation.
- The misses are positional, not lexical: they cluster after discourse markers - exactly the points where the content shifts and the answer to the next item begins. He is still writing the previous answer while the next one is spoken. That is a timing failure, not a comprehension failure.
- 90 words of notes for a five-minute talk is not note-taking, it is transcription. Transcription consumes working memory that should be holding the upcoming question.
- He is accurate on form completion, which is a local item type with predictable wording and a long window. He fails on multiple-choice and matching, which are distributed item types where options must be held while the speaker ranges across them. That split is the signature of overload, not of a small lexicon.
- A pure vocabulary problem would produce misses scattered at hard words, not clustered at signposts.
Falsifier: if he were given the transcript after a miss and could not identify the paraphrase that carried the answer, the problem would be lexical after all. Run that check on ten misses before committing to the plan.
2. Redesign his note-taking. Specify what he writes, what he stops writing, and how he keeps his place in a multiple-choice section.
Read the worked answer
Stop: full sentences, and anything that repeats the stem he has already read.
Start:
- Before the audio, underline the content word in each question and note the expected answer type (number, reason, contrast).
- During the audio, write symbols and single words only - arrows for cause, '≠' for contrast, initials for speakers. Target under 20 words per section, not 90.
- Treat signposting phrases as a look-up cue: when he hears 'however' or 'the thing is', his pen stops and his eyes move to the next unanswered question. He should train this as a reflex, not a decision.
- For multiple-choice, pre-mark each option with a one-word gist before the audio starts, so he is matching a heard paraphrase to a known label rather than re-reading four full options under time pressure.
Why this should work: it removes the writing lag that the timestamps implicate and converts the signpost from a distraction into a trigger.
3. Give a three-week success criterion that does not depend on his total score.
Read the worked answer
Process criteria, checked on two fresh mocks per week:
- Note volume in each section ≤ 25 words (baseline 90).
- First-attempt correctness on items 4+ of each section - the exact zone where he currently collapses. Target: at least 50% on these, up from roughly 10–15%.
- Count of answers written before the next signpost, taken from his own timestamps.
Why not total score: at 18/40 the paper mixes item types and sections, so a two-mark swing can come from an easier section rather than a changed habit. The positional metric isolates the behaviour being trained. If items 4+ do not move in three weeks while note volume does, the transcription theory is wrong and the transcript check from the first task should be run.