2. The Core Method: A Disciplined Three-Step Approach

Most candidates who struggle with Sentence Improvement are not struggling with grammar — they are struggling with process. They read the sentence, feel a vague sense that something is off (or isn't), skim the four options looking for one that "sounds better," and pick it. This is exactly the approach the format is designed to punish, because at least two of the four wrong options are specifically engineered to "sound better" in a shallow read while being wrong in a way a careful reader would catch. The fix is to replace that instinctive skim with a fixed three-step sequence, applied to every single question, without exception, until it takes less than twenty seconds per question to execute.

Step 1 — Diagnose the underlined segment in isolation, before looking at the options

Cover the four options with your hand or your eye, and ask only one question: what, specifically, is wrong with the underlined segment on its own? Not "does the sentence sound odd" — that is too vague to be useful — but a precise diagnosis you could say out loud: "the preposition after 'insisted' is wrong," or "the tense doesn't match the time clue 'yesterday,'" or "the subject is singular but the verb is plural," or, critically, "nothing — this segment is already correct." Write this diagnosis down mentally (or, in a mock test, jot a two-word note in the margin) before you read a single option. This single habit — diagnosing first, reading options second — is the most important discipline in this entire guide, because it gives you an independent anchor to judge each option against, instead of judging options against each other.

If you cannot produce a specific diagnosis after a careful read, that is itself informative: it suggests the segment may already be correct, and you should lean toward "No improvement" once you confirm this in Step 3 — not abandon the method and start guessing.

Step 2 — Evaluate each option against your diagnosis, not against your ear

Now read options (A), (B), and (C) one at a time, and for each one ask two separate questions:

1. Does this option fix the exact problem I diagnosed in Step 1? If your diagnosis was "wrong preposition," an option that changes the verb tense but keeps the same wrong preposition has not fixed your problem — eliminate it, even if it looks superficially different from the original.
2. Does this option introduce a new problem while fixing the old one? This is the step most candidates skip, and it is where the format does most of its damage. An option can correctly swap "insisted to complete" for "insisted on completion" — the preposition is now right — but if "completion" turns the sentence's meaning subtly wrong, or breaks agreement elsewhere in the sentence, that option still fails. Read the entire replacement sentence with the option substituted in, start to finish, not just the underlined fragment in isolation. A trap option is often grammatically self-contained but wrong once you see it sitting inside the full sentence.

Eliminate every option that fails either test. In a well-constructed question, this leaves exactly one surviving option among (A), (B), and (C) — or none, which is your signal to examine (D) seriously.

Step 3 — Select "No improvement" only after positive confirmation, never by default

Do not treat "No improvement" as a fallback for when you are unsure — treat it as an option you must actively earn the right to select by confirming, specifically, that the original underlined segment contains no fixable error. Ask: "If I had to explain to someone why this sentence is correct as written, what would I say?" If you can produce a real answer — the idiom is standard, the tense is right for the time clue, the agreement is correct — then "No improvement" is a confident, earned choice. If you are choosing it only because none of (A), (B), or (C) "felt right" to you, stop and re-run Step 1: you likely mis-diagnosed the segment, or you are missing a subtler correct option. Section 5 of this guide exists entirely to sharpen this judgment, because it is the single place where guessing costs the most marks.

A Worked Illustration of the Full Method

Consider: "The selection board have decided to postpone the interview round due to unforeseen circumstances."

Step 1 — Diagnose: The underlined segment is "board have decided." "Board" is a collective noun referring to a single administrative body acting as one unit here, so it should pair with a singular verb. Diagnosis: subject-verb agreement error — "have" should be "has."

Step 2 — Evaluate options against the diagnosis. Suppose the four options are: (A) board has decided (B) boards have decided (C) board having decided (D) No improvement. Option (A) directly fixes the diagnosed agreement error and introduces nothing new — it survives. Option (B) changes "board" to "boards," which sidesteps the actual problem by changing the subject instead of the verb, and it also shifts meaning (now multiple boards, not one) — eliminated. Option (C) turns "decided" into "having decided," creating a dangling non-finite construction with no main verb for the clause — a new error — eliminated.

Step 3 — Confirm the pick. (A) is the only surviving option, and it directly answers the diagnosis from Step 1 with no side effects. Final answer: (A).

This is the exact rhythm — diagnose, evaluate against the diagnosis while hunting for new errors, confirm — that every worked example in Section 3 below will walk you through. Practise saying the diagnosis out loud (or writing it down) for the first fifty questions you attempt from the practice set in Section 8; once the habit is automatic, you can do it silently in under fifteen seconds per question.