Almost everybody reviews, and almost everybody reviews in the wrong order. Reading through for spelling and grammar is what feels productive, so it happens first, and it consumes the minutes in which a real problem could still have been fixed. The expensive failures are structural — answering a different question, losing your position, a body paragraph that argues against your thesis — and they can only be repaired if you find them while there is time.
So the order runs from most to least expensive. First: reread the instruction and check you answered that one, not the one you expected. “To what extent do you agree” asks for a position; discussing both views without giving one is a different task. Second: is the position visible in the introduction, both bodies and the conclusion? Third: does each paragraph support it, or has one drifted? Only then, fourth: the words.
The other half of the skill is knowing what cannot be fixed. With five minutes left, restructuring produces a response that is half old and half new, and the seam is more visible than the problem. Adding a new idea produces a claim with no development behind it. What is available is a word swap, a deleted sentence, and a repaired position statement — and those are worth more than they sound.
Did you answer the question asked, or the one you expected? Most expensive failure, and it takes twenty seconds to check.
Visible in the introduction, both bodies and the conclusion? A drift here costs across the whole response.
Does every paragraph support the thesis? A drifted paragraph can often be fixed with one sentence.
Vague, overclaimed, informal, repeated. Last, because it is the cheapest to fix and the least costly to miss.
