A cloze test is a passage — typically 150 to 250 words — with five to ten words removed and replaced by numbered blanks. Your job is to identify which of four options best fills each blank, keeping the passage grammatically correct and contextually coherent.
Here is the key distinction that most test-takers miss: a cloze test is not just a vocabulary exercise and it is not just a grammar exercise. It is both, simultaneously. Each blank has two filters. First, the option must be grammatically legal in that sentence slot. Second, it must make semantic sense given the paragraph's overall meaning. An option can pass the grammar filter and still be wrong because it contradicts the passage's tone or argument.
Think of it like this. Imagine you are reading a letter with ink blots covering certain words. You know what the letter is about because you can read the surrounding sentences. The blotted word must fit both the visible grammar of the sentence and the broader message of the letter. That dual constraint is exactly what the cloze test is testing.
For SSC CHSL specifically, the passages tend to be factual or mildly argumentative — topics like environment, society, human behaviour, or general science. The blanks are designed to test four recurring skill types: tense and voice, preposition collocation, modal verbs expressing degree of obligation or possibility, and vocabulary precision (noun vs verb vs adjective from the same root). You will rarely encounter a blank that requires an obscure word — the trap is usually a familiar word used in an unfamiliar grammatical slot, or a preposition that feels almost right but collocates incorrectly.
Read the whole passage before filling even the first blank. Thirty seconds of reading saves two minutes of backtracking.
These are the most mechanical blanks in the set. SSC CHSL setters love passive voice constructions because they look tricky but are rule-bound.
The rule for passive voice is auxiliary be (conjugated) + past participle. That means:
is/are + V3was/were + V3will be + V3should/can/must + be + V3Any option that pairs a modal directly with a past participle without be is grammatically wrong, full stop. For example, "will made" is wrong because it needs "will be made". "Can stopped" is wrong because it needs "can be stopped" or "should stop" (active). When you see these options, eliminate them without reading the passage — the grammar kills them.
For tense selection, look for temporal markers in the surrounding sentences. Words like "currently", "every day", "in general" push you towards present simple or present continuous. Words like "yesterday", "in 2001", "last year" push you towards past simple.
This is where most intermediate-level students lose marks. The correct preposition is determined by the verb or noun it collocates with, not by what sounds logical in isolation.
Key collocations tested repeatedly on SSC CHSL:
grown on trees (not grown from or grown of)believe in something (not believe on)made from/of a material (made from wood, made of gold)put one's mind to workovercome a challenge (verb takes a direct object, no preposition)When a preposition blank appears, do not try to construct a logical argument for why one preposition is correct. Recall the fixed collocation. If you cannot recall it, test each option by asking: "Have I seen this exact combination in standard English text?" Trust the collocation, not the logic.
Modals express degrees of possibility, permission, obligation, or recommendation:
must / have to — strong obligation or certaintyshould / ought to — recommendation, mild obligationcan / could — ability or possibilitymay / might — weak possibilityIn the context of a passage, the correct modal is determined by the passage's stance. If the passage is making a recommendation (e.g., "save nature"), should is more appropriate than must (which would mean absolute compulsion) or can (which would mean mere ability). Read the surrounding argument to judge whether the writer is commanding, recommending, or permitting.
These blanks present four words from the same root — for example, nature, natural, nurture, nurturing. The trap is that they all look similar and students pick the first one that "feels environmental". The grammar slot decides it: after a preposition like "harm to", you need a noun, not an adjective or a gerund. So "harm to nature" is correct; "harm to natural" is not.
Always identify the grammatical slot first:
After reading the full passage, build a one-sentence summary in your head: "This passage is about X and its argument is Y." For every blank, ask: "Does this option support that argument, or contradict it?"
When all four blanks in a multiple-answer question (like the "multiple thoughts" PYQ below) are filled by one option set, check coherence globally. The correct option set tells one consistent story. The wrong sets introduce jarring shifts in meaning — for example, switching from "multiple thoughts" to "challenging thoughts" changes the passage from being about cognitive science to being about difficulty, which contradicts the explanatory tone.
Apply four filters in sequence for any blank: Grammar slot (noun/verb/adj/adv?) → Voice check (active or passive structure needed?) → Collocation match (does this word pair with its neighbours?) → Passage purpose (does it align with the passage's argument?). Each filter eliminates at least one option. Standard approach: students read all four options and guess from meaning alone (60 seconds, ~50% accuracy). GVCP chain: 30 seconds, eliminates two options before even reaching meaning (boosting accuracy to ~85%).
Write this mental scale from left to right when you see a modal blank: can < may < should < must. The passage's tone places you on this scale. A factual-recommendation passage (common in SSC CHSL environment topics) sits at should. A warning or scientific fact sits at must. An ability or possibility context sits at can. Standard method: read all options and guess (45 seconds). With the scale: identify tone in 5 seconds, pick the matching modal in 5 seconds (10 seconds total).
Scan any option with a modal or auxiliary and count the verb forms. Legal passive = auxiliary + be + V3 (two words after the main auxiliary at most, or three in modals). Illegal forms: will made (missing be), can stopped (missing be), was make (wrong form, needs made). Each illegal option is eliminated in 3 seconds. No reading of context required. This alone eliminates 1–2 options per tense/voice blank, reducing the actual choice to 2 options in under 10 seconds versus the standard 40-second option-by-option evaluation.
For vocabulary precision blanks, cover the options and identify what part of speech the blank needs by reading the sentence frame. Then uncover options and cross out every option that is not that part of speech. Example: "harm to ______" needs a noun → cross out adjectives and gerunds immediately. You go from four options to one or two in under 8 seconds, versus standard meaning-based comparison which takes 30–40 seconds and is prone to confusables like nature vs nurture.
When a preposition blank appears, replace the blank with each option and say the resulting phrase aloud (mentally). Your ear will flag the unnatural collocation faster than logical analysis. "Grown on trees" — natural. "Grown of trees" — unnatural. "Grown out trees" — unnatural. "Grown for trees" — unnatural. This substitution method takes 10 seconds and exploits your existing passive knowledge of English collocations. Standard logical analysis ("which preposition means what?") takes 40+ seconds and is unreliable because prepositions are idiomatic, not logical.
In the exam hall, use this sequence:
Step 1 — Read the full passage (30 seconds). Do not fill any blank yet. Build a mental summary: topic + tone + argument direction.
Step 2 — Classify each blank (5 seconds each). Look at the sentence frame. Is this a grammar blank (tense, voice, modal) or a vocabulary blank (word form, collocation)?
Step 3 — Grammar blanks first. Apply the Passive Voice X-Ray and Modal Strength Scale. Eliminate illegal structures immediately. These blanks should take under 15 seconds each.
Step 4 — Vocabulary blanks. Apply Word-Form Slot Test to identify part of speech. Then use Collocation Anchor to test remaining options. These blanks take 15–25 seconds each.
Step 5 — Global coherence check (15 seconds). Re-read the passage with your chosen answers. Does the passage now read as one coherent argument? If a blank feels jarring, reconsider that one option — but do not second-guess answers that passed all four GVCP filters.
Never leave a blank empty. If genuinely stuck, the Collocation Anchor substitution method gives a better-than-random answer in 10 seconds.
Why this question: Tests passive voice form recognition — one of the highest-frequency grammar patterns in SSC CHSL cloze tests. Three of the four options are passive-voice-shaped but grammatically broken.
Solving path: Apply Passive Voice X-Ray immediately. The blank follows a plural subject ("Houses, chairs, and toys") and precedes "from wood", so passive voice is needed. was make — wrong auxiliary-V1 combination. had make — wrong V1 after a perfect auxiliary. will made — missing be. Only are made is the legal structure (present passive: are + V3). Time: 10 seconds.
Why this question: Tests preposition collocation — specifically the phrase "grown on trees", which is idiomatic and cannot be derived logically.
Solving path: Apply Collocation Anchor. Say each option in the phrase: "grown on trees" / "grown out trees" / "grown of trees" / "grown for trees". Only "grown on" is a natural English collocation — fruits and vegetables are said to grow on trees. out, of, and for produce unnatural phrases. Time: 12 seconds.
Why this question: Tests modal verb selection in a recommendation context — a frequent trap where can, may, and should all seem plausible.
Solving path: Apply Passive Voice X-Ray first: may stopped and can stopped are missing be (to be correct passives, they would need may be stopped / can be stopped). These are eliminated in 5 seconds. have stop is wrong — after have, the V3 form stopped is needed. Only should stop is both grammatically complete (modal + base verb, active voice) and contextually appropriate — the passage is making a recommendation about conservation, not stating an absolute command (must) or a mere ability (can). Time: 15 seconds.
Why this question: Tests vocabulary precision — distinguishing between near-identical words from the same root (nature vs nurture vs natural vs nurturing).
Solving path: Apply Word-Form Slot Test. The sentence frame is "stop causing harm to ______". After the preposition to, a noun is needed. nurturing is a gerund/adjective, natural is an adjective — both eliminated by the slot test. Now between nurture (noun/verb meaning to care for and develop) and nature (noun meaning the natural world): the passage is entirely about trees and environmental conservation. "Harm to nature" fits that topic directly. "Harm to nurture" is possible abstractly but out of scope for this passage. Time: 18 seconds.
Why this question: This is a multi-blank question where all four blanks are answered by one option set — requiring global coherence evaluation, not just sentence-level grammar.
Solving path: The passage explains the cognitive phenomenon of seemingly simultaneous thoughts. Check Option C: "multiple thoughts" (plausible — describes many thoughts), "myriad" (means very many — consistent with the idea of numerous alternating thoughts), "alternating" (key word — thoughts switching rapidly, which explains why they seem simultaneous), "succession" (quick succession — natural English phrase meaning one after another in rapid sequence). Now check the others: Option A introduces "ramification" (a consequence/branching), which makes no sense in "quick ramification". Option B uses "unchanging", which contradicts the whole point (thoughts must be changing/alternating for the phenomenon to work). Option D uses "interruption" which changes the meaning to disruption. Only Option C is internally consistent and explanatorily accurate. Time: 35 seconds for the whole set.
Stopping at the single sentence. Many test-takers fill each blank using only the sentence containing the blank, ignoring adjacent sentences. This is the single biggest source of wrong answers in vocabulary-type blanks, where the correct word is determined by the passage's overall subject matter, not just the immediate grammatical context.
Accepting the first grammatically legal option. Two or three options often pass basic grammar inspection. "Can stop" and "should stop" are both grammatically legal active-voice constructions — you must go to the passage's tone to distinguish them. Students who stop at grammar step alone lose these marks.
Confusing word-family members. nature and nurture look alike. distinguishing and distinguished both exist and are both adjectives in certain contexts. The distinction is meaning-based: "distinguishing individuals" means individuals who are characterised by specific traits; "distinguished individuals" means already-famous, eminent individuals. The passage context determines which is correct.
Trusting logical preposition derivation. You cannot reason your way to the correct preposition. "Made from" and "made of" are both correct English in different contexts (material transformation vs material composition) — you cannot pick the right one without knowing the specific collocation convention. Trust ear-test substitution, not logic.
Ignoring illegal passive forms quickly enough. Options like will made, can stopped, was make should be killed in 3 seconds each. Students who spend time "considering" these options are wasting time the exam design intended them to spend on harder vocabulary decisions.
Second-guessing answers that passed all filters. After filling all blanks, students re-read the passage and feel uncertain about answers that are actually correct. The GVCP filter chain is reliable. If an answer passed all four filters — grammar slot, voice check, collocation, passage purpose — do not change it based on a vague feeling.