ScoreFluentHighlight Incorrect Words shows you a transcript of a recording while the recording plays. Some words in the transcript don't match what's actually said — you click every word that differs. Because you're given the text, it's tempting to relax into listening for overall meaning, the way you would for a normal comprehension task. That instinct is exactly what causes missed words.
What's actually scored.
Credit per correctly identified mismatched word, typically with a penalty for clicking words that are actually correct. This isn't a comprehension check — the passage is usually simple enough to understand easily. It's a word-for-word tracking task, and the errors inserted are often subtle: a plural swapped for a singular, a synonym substituted for the original word, a tense changed.
Read-along, don't listen-and-check
The highest-scoring approach is to silently mouth or eye-track each word in the transcript in exact sync with the audio, the way a karaoke bouncing ball works — not to listen to a phrase, then glance at the text to see if it "seems right." Errors are single-word substitutions, often words that don't change the sentence's overall meaning much ("a chair" vs. "the chair," "quickly" vs. "quietly") — meaning-level listening skates right past these, because the sentence still makes sense either way.
Why meaning-preserving errors are the hardest to catch.
The task designers deliberately choose substitutions that don't break comprehension — a changed word that made the sentence nonsensical would be too easy to spot. That's precisely why exact word-tracking, not meaning-tracking, is the skill this task rewards.
A practice method that transfers directly
Before the exam, practice with any transcript-and-audio pair (a podcast with published transcript works well): follow the text with your finger or cursor at the exact word being spoken, and deliberately try to notice if you ever feel a half-second of mismatch between what your eyes expect and what your ears hear — that feeling, sharpened through repetition, is the actual skill being tested. Most candidates have never trained exact synchronous tracking before encountering this task type, which is why it initially feels harder than it should.
Common mistake: waiting until the recording ends to review
Because you can click words at any point while the audio plays, waiting to review the whole transcript only after the recording finishes means relying on memory for word-level detail from 30+ seconds ago — memory that's reliable for meaning but not for exact wording. Click the moment you notice a mismatch, in real time; don't defer the decision.
The summary.
Track the transcript word-for-word in real time as the audio plays, rather than listening for overall meaning — the inserted errors are deliberately meaning-preserving, so meaning-level listening won't catch them. Click mismatches the instant you notice them rather than reviewing from memory afterward.
If you want to drill Highlight Incorrect Words with instant AI-assisted practice, ScoreFluent is opening early access on July 12.
— The ScoreFluent team