Writing: Proofreading Strategy for the Last Few Minutes

The last few minutes of a timed Writing task are usually spent one of two ways: re-reading the whole essay from the top in the hope of catching whatever's wrong, or not proofreading at all because there's a vague sense that it "probably won't help much" with so little time left. Both are weaker strategies than a third option — a targeted pass that checks specifically for a small, known set of error types rather than attempting a general quality re-read. The mistake pattern this addresses is treating proofreading as an unfocused, general activity ("read it again and see if anything looks wrong") when in the two or three minutes realistically available, an unfocused re-read catches far less than a deliberately narrow, prioritised check for the specific errors most likely to be present and most costly if left in.

The reason a targeted pass outperforms a general re-read comes down to how reading under time pressure actually works: when you re-read your own writing quickly, you tend to read for meaning and fluency, which your brain processes automatically and which therefore glosses over small mechanical errors your eye slides past because the sentence still makes sense despite them. A missing article, a subject-verb mismatch, or a wrong preposition rarely breaks comprehension enough to interrupt a fast read-through, which is exactly why a general re-read misses these errors so reliably — you're reading too fast, and too focused on sense, to notice them. A targeted pass works differently: instead of reading for meaning, you scan specifically for one category of error at a time, which is a fundamentally different, slower, and more effective kind of attention.

The highest-priority, fastest checks to run in this order are: article usage (a/an/the) on your most important nouns, subject-verb agreement in your longer or more complex sentences, and your actual word count against the minimum requirement. Article errors are the single most common recurring error category in submissions at Band 5-6.5 and are genuinely fast to check once you know what you're looking for — scan noun phrases in your topic sentences and any sentence introducing a new idea, since these carry the most weight if wrong. Subject-verb agreement errors cluster specifically in longer sentences with intervening clauses, where the true subject and the verb have drifted apart — a sentence like "The number of students who attend university despite financial difficulties have increased" contains an agreement error because "have" should agree with "number" (singular), not with the nearer noun "difficulties," and this kind of long-distance mismatch is exactly what a fast general read tends to miss but a targeted grammar-specific scan catches quickly. Word count is worth a final quick check specifically because falling under the minimum is directly penalised regardless of how strong the content is, and a rough visual estimate of length is less reliable than you'd expect.

Pronoun reference is a distinct check worth its own pass: scan for every use of "it," "this," "that," or "these" and confirm each one clearly points back to a specific, identifiable noun rather than to a vague, unstated idea from a few sentences earlier. A sentence like "Governments have increased spending on public transport in recent years. This has led to significant improvements" is fine if "this" clearly refers to the increased spending — but a sentence like "Governments have increased spending on public transport, reduced fuel subsidies, and introduced new regulations. This has led to significant improvements" leaves "this" genuinely ambiguous between three possible referents, which is a Coherence and Cohesion issue that's easy to miss while writing at normal speed (where you know exactly what you meant) but that a targeted reference-check, reading specifically for what each pronoun points to rather than for overall meaning, catches reliably.