Sentence improvement questions give you a sentence — sometimes grammatically broken, sometimes not — and ask you to pick the version that is correct, or to identify that no improvement is needed. The trap the paper-setter exploits is your instinct. If you have read enough Hindi-medium classroom English or absorbed common spoken errors, your ear will validate the wrong answer. NDA paper-setters know this and they select sentences that feel right but are technically wrong.
Here is what sentence improvement is really testing: whether you can apply grammar rules systematically, not just "sound out" the sentence. Think of yourself as a copy-editor reviewing a military communiqué before it goes to the brigade commander. Every word must earn its place. The standard does not lower because the sentence sounds okay in conversation.
The analogy that works: imagine a parade drill. Each soldier must stand in the correct position. If the left-flanker steps one foot out of line, the drill is still imperfect even if the overall formation looks acceptable from a distance. Grammar works the same way — one misplaced preposition, one wrong verb form, one adjective where an adverb should stand, and the sentence fails inspection regardless of how smoothly it reads aloud.
In NDA's GAT section, sentence improvement questions typically require you to recognise errors in five recurring categories:
since vs for, and transitive verbs that take no preposition)Master these five, and you handle the vast majority of what NDA throws at you.
Two correlative pairs dominate NDA questions: either...or and neither...nor.
The proximity rule: When two subjects are joined by either...or or neither...nor, the verb agrees with the subject that is closest to it.
Look — this sounds mechanical, and it is. That is the point. You do not need to understand why; you need to apply the rule in 20 seconds.
Common trap: the paper-setter puts the plural subject first and the singular subject second, hoping you will default to a plural verb because you read the plural noun first.
When the reporting verb is in the past tense (said, suggested, told, instructed), the verb inside the reported clause must shift back one tense level. This is called backshift.
| Direct speech | Reported speech | |---|---| | may | might | | can | could | | will | would | | is/are | was/were | | have/has | had |
So "the report suggested that enemy forces may have crossed" is wrong — suggested is past, so may must shift to might. The corrected sentence: "the report suggested that enemy forces might have crossed."
This rule applies even when the fact is still true. In reported speech, backshift is grammatically required regardless of real-world logic.
since vs for:
Since marks a specific point in time: "since 2019", "since Monday", "since he joined."For marks a duration: "for five years", "for three months", "for a long time.""The regiment has been deployed since five years" — wrong. "Five years" is a duration, not a point. Correct: "for five years."
Transitive verbs that take no preposition:
Some verbs are transitive and take a direct object directly — no preposition in between. Emphasize is the most commonly tested.
Other verbs in this pattern: discuss (not "discuss about"), reach (not "reach to"), mention (not "mention about"), approach (not "approach to").
When two or more elements are coordinated (and, but, or), they must be in the same grammatical form.
to maintain (infinitive) and following (gerund) do not match.to shared across both.Here's the trick: scan the sentence for and, but, or. Then look at both sides of the conjunction and ask — same form? If not, find the option that makes them match.
Adjectives modify nouns. Adverbs modify verbs, adjectives, or other adverbs. When a word is modifying a verb, it must be an adverb.
The test: ask "what does this word describe?" If the answer is a noun or pronoun, use an adjective. If the answer is a verb or an adjective, use an adverb. This takes four seconds per question.
NDA consistently includes questions where the original sentence is already correct. This is not a trick — it is a test of overconfidence. Students who have seen too many errors start correcting things that do not need correcting. The discipline is to verify first, then decide.
Before selecting "No improvement required," run through all five categories above. If the sentence passes all five checks, trust the original.
With either...or / neither...nor, ignore the first subject completely. Lock your attention on the subject immediately before the verb. That subject alone determines singular or plural. Standard method: read both subjects, try to average them mentally — 30 seconds, often wrong. Proximity Lock: cover everything before or/nor, read the closer subject, decide — 8 seconds, always systematic.
Substitute the time expression with "Monday." If Monday fits naturally, use since. If it sounds absurd, use for. "The regiment has been deployed since Monday" — makes sense, so a specific point. "The regiment has been deployed since five years" — absurd, so it must be for. This substitution takes 5 seconds vs 20 seconds of trying to remember the rule abstractly.
Read the sentence, find the suspicious word, then ask: can I replace the verb it sits next to with "seems" or "is"? If yes, the word is a predicate adjective and stays as-is. If no, you need an adverb. "He handled it efficient" — can you say "he is efficient"? No, because the subject of efficiency here is the handling, not him. So you need "efficiently." This 2-step swap reduces adjective/adverb errors to a 6-second check vs reading multiple options and guessing.
Every time you see and, but, or or in a sentence improvement question, immediately look left and right of the conjunction. Write down (mentally) the grammatical form of what's on each side: infinitive, gerund, noun, clause. If they do not match, the answer must be the option that makes them match. If they do match, that is evidence for "No improvement required." This scan takes 10 seconds versus reading all four options first, which averages 35 seconds and invites distraction.
Before reading the options in a tense/reported speech question, look at the reporting verb first — not the underlined portion. If the reporting verb is past (said, suggested, warned, told), the underlined modal must be the past-equivalent. The pairs: may → might, can → could, will → would, shall → should. Identify the reporting verb (2 seconds), identify the modal in the sentence (2 seconds), match to the pair (2 seconds). Total: 6 seconds. Standard approach of reading all options: 25 seconds minimum.
In the exam hall, follow this decision sequence for every sentence improvement question:
Step 1 — Identify the underlined or bracketed portion. Do not read the options yet.
Step 2 — Run the five-category scan in order:
either...or / neither...nor? Apply proximity rule.since/for or a transitive verb with an attached preposition? Check preposition legality.and/but/or? Check parallelism on both sides.Step 3 — If none of the five categories fire, the sentence is likely correct. Check "No improvement required" before moving to options.
Step 4 — Match your finding to the options. Only now look at the options. Eliminate wrong ones using your diagnosis.
Target: 45-60 seconds per question. If you cannot identify the error type within 20 seconds of scanning, skip and return — do not guess from options alone, that is how paper-setters catch you.
Why this question: Tests the proximity rule for neither...nor — the most commonly repeated agreement pattern in NDA English.
Solving path: Identify the correlative conjunction: neither...nor. The two subjects are "army personnel" (plural) and "equipment" (uncountable, treated as singular but closer to the verb). Apply proximity lock — "equipment" is closer to the verb. However, the explanation grounds on "equipment" being considered plural in context, making "were" appropriate. The correct answer is option B: "were damaged."
Why this question: Tests backshift in reported speech — a rule that feels unnatural because may sounds fine even in past-reported contexts in spoken Indian English.
Solving path: Spot the reporting verb first: suggested — past tense. The modal in the sentence is may. Apply backshift: may → might. The correct replacement is might have crossed. Option B is correct.
Why this question: Tests the same proximity rule as the first PYQ but with either...or instead of neither...nor, confirming the rule applies to both conjunctions.
Solving path: Correlative conjunction: either...or. Subjects: "captain" (singular, first) and "lieutenants" (plural, closer to verb). Proximity lock: "lieutenants" governs the verb. The verb must be plural: were. Option C is correct.
Why this question: Tests since vs for — one of the most common preposition traps in NDA, exploiting the fact that Indian classroom English often uses since indiscriminately.
Solving path: The time expression is "five years" — a duration, not a point in time. Apply the Duration vs Point Test: "since Monday" works; "since five years" does not. Duration requires for. Correct answer: for five years, option C.
Why this question: Tests adjective vs adverb — a category where students often overlook the error because the adjective form is more familiar.
Solving path: The word "efficient" appears after the verb "handle." Ask: what does "efficient" describe here? The action of handling — a verb. Verbs require adverbs, not adjectives. Replace "efficient" with "efficiently." Option C is correct.
Why this question: Tests overconfidence — the "No improvement required" option. The sentence is already parallel and correct; the trap is that one option offers a superficially appealing rearrangement.
Solving path: Run the parallelism scan. The structure is "to maintain discipline and follow the orders" — both infinitives, with to shared. This is valid parallel structure. No grammatical violation exists. Select "No improvement required," option C.
Why this question: Tests a transitive verb (emphasize) that is commonly — and incorrectly — used with a preposition in Indian English.
Solving path: The original uses "emphasized on the importance." Emphasize is transitive — it takes a direct object without a preposition. "Emphasized on" is wrong. The correct form: "emphasized the importance." Option A is correct.
Why this question: Tests whether you will "improve" a sentence that does not need improvement — a direct test of disciplined grammar checking over instinctive tinkering.
Solving path: Run all five checks. The verb saved with from + gerund (crashing) is standard and correct English. No agreement error, no tense inconsistency, no preposition violation, no parallelism break, no adjective/adverb confusion. Correct answer: "No improvement required," option D.
Treating neither...nor as automatically singular. Many students fix the verb to "was" regardless of which subject is closer. The proximity rule is non-negotiable — apply it every time, do not default to singular.
Leaving may unchanged in past reported speech. "May" sounds perfectly fine in sentences like "the report suggested that they may have..." — which is exactly why it appears. If the reporting verb is past, may must shift to might. No exceptions in formal written grammar.
Confusing since and for when the time expression has a number. "Since three days" is wrong regardless of how commonly you hear it. Numbers representing durations always take for.
Over-correcting parallel structures. When a sentence already has matching infinitives (e.g., "to maintain and follow"), students sometimes "improve" it to "to maintain and to follow" or "maintaining and following" — both are grammatically acceptable but represent unnecessary edits when "No improvement required" is available and the original is already correct.
Adding prepositions to transitive verbs out of reflex. Verbs like emphasize, discuss, mention, approach, and reach are complete without prepositions. The reflex to add "on," "about," or "to" comes from Hindi sentence structure (where postpositions are always required) — suppress it when writing formal English.
Reading options before diagnosing the sentence. If you read the options first, option A or D will anchor your thinking and you will evaluate the others relative to them rather than relative to the grammar rule. Always diagnose the error type in the original sentence before you open the options.