Reading comprehension

In short

Reading comprehension (RC) is the ability to read a passage once, build an accurate mental map of what it argues, and then answer questions using only what the passage states or implies. It is the single largest verbal section in Indian placement and entrance tests — roughly 50-70% of CAT's VARC, a fixed block in IBPS/SBI, SSC, RRB and insurance exams, part of GATE's 15-mark General Aptitude, and a standard round in TCS NQT, Infosys, Wipro, Accenture, Capgemini and Cognizant tests. Universities examine it under AICTE/CBCS courses such as Communication Skills, Technical English or Business Communication, usually as an unseen passage with short answers, a précis and vocabulary items. The skill is not academic: analysts, consultants, testers, bankers and support engineers are paid to read requirement documents, tenders, SOPs and RBI/SEBI circulars correctly the first time. RC is also the most trainable verbal skill, because every answer is in the text and the wrong options fail in a small number of predictable ways.

Explained

What an RC set tests, and how Indian papers set it

• An RC set is a passage (200-900 words) plus 3-6 questions. Sources are usually journal essays, business and economics writing, science journalism, philosophy, history and social commentary — not technical manuals.
• The rule that governs every answer: the passage is the only authority. Outside knowledge, personal opinion and "what is generally true" are all irrelevant and are actively used to build wrong options.
• Passage types: (i) *argumentative* — the author defends a position; (ii) *expository/descriptive* — explains a phenomenon neutrally; (iii) *narrative* — recounts events; (iv) *comparative* — two views weighed against each other.
• Typical Indian patterns (verify against the current notification, these change):
  • CAT VARC — 4 passages, 4 questions each, in a 40-minute section with 8 non-RC questions; +3 / −1 for MCQs.
  • IBPS/SBI PO prelims English — about 30 questions in 20 minutes, 1 mark each, 0.25 penalty; RC is 8-12 of them.
  • GATE General Aptitude — 10 questions, 15 marks, verbal aptitude includes comprehension; negative marking 1/3 and 2/3 on MCQs, none on NAT/MSQ.
  • IT service companies — 1-2 short passages of 200-350 words with 3-5 questions, often inside a 20-25 minute English section.
• University exams add productive tasks the test papers do not: write a précis to one-third length, give a title, answer in your own words, explain a sentence in context.

The question types, and where the answer lives

1. Main idea / central theme / best title — the whole passage, especially the first and last paragraphs. The answer must cover the passage, not one paragraph.
2. Specific detail / "according to the passage" — one or two sentences; locate them, do not reason.
3. Inference / "it can be concluded" — not stated, but must follow necessarily from what is stated. One safe step beyond the text, never two.
4. Assumption / "the author presumes" — the unstated link the argument needs to stand. Test it by negation: if the option is false and the argument collapses, it is the assumption.
5. Strengthen / weaken — new information that makes the conclusion more or less likely; it may contradict your own beliefs and still be correct.
6. Vocabulary in context — the meaning the word carries *here*, which may not be its dictionary-first meaning; replace the word with each option and reread the sentence.
7. Function / purpose of a sentence, paragraph or example — ask "why did the author put this here?" Answers sound like *to illustrate, to qualify, to refute, to concede, to introduce*.
8. Tone / attitude of the author — the adjective describing the author's stance.
9. Structure / organisation — how the argument is built ("a claim is made, an objection is raised, the claim is restricted").
10. EXCEPT / NOT / LEAST questions — four options are supported, one is not; mark each option T/F before answering, or you will invert the task.
11. True / False / Cannot say (banking and SSC style) — "Cannot say" means the passage neither confirms nor denies it; it does not mean "probably false".

How to read: strategies and the three passes

• Skimming — fast reading for gist and structure (headings, first/last sentence of each paragraph). Scanning — hunting a specific word, name, number or date. Intensive reading — slow, line-by-line, used only on the 2-3 sentences a question points to.
• Pass 1 (read): read the whole passage once at a comfortable speed, pausing at paragraph ends to fix one phrase of summary in your head — "Para 2: counter-example, costs ignored".
• Pass 2 (question): read the question *stem* fully before the options, decide the type, and predict the answer in your own words.
• Pass 3 (verify): return to the exact lines and confirm. An answer you cannot point to is a guess.
• Keyword mapping / paragraph tagging: mentally label each paragraph with 2-4 words. For 600+ word passages this is what makes detail questions fast.
• Transition words are the skeleton — *however, although, yet, on the contrary* signal a turn; *moreover, furthermore, indeed* continue; *therefore, hence, thus* mark the conclusion; *for instance, consider* mark evidence that is never itself the main idea.
• Passage-first vs question-first: passage-first suits long, argumentative CAT-style passages; question-first (read stems, then scan) suits short, fact-heavy banking passages where most questions are detail-based. Do not skim questions and passage simultaneously.
• SQ3R (Survey, Question, Read, Recite, Review) is the classic study method prescribed in Communication Skills syllabi; in a timed test it compresses to survey-read-verify.
• Subvocalisation (saying words in your head) caps speed near 200 wpm; reading in chunks of 3-4 words and avoiding regression (re-reading) is what raises it. Target 200-250 wpm with 80%+ comprehension, not raw speed.

Option elimination: the trap families

• Work by elimination, not attraction: strike three options with a reason each; the survivor is the answer even if it reads awkwardly.
• If two options survive, find the single word that differs — usually a quantifier (*some* vs *most*), a modal (*may* vs *will*) or a scope word.
• Correct inference options are typically cautiously worded; correct detail options are typically paraphrases, not verbatim lifts (a verbatim lift is often bait).

Fact, inference, assumption and tone

• Fact: verifiable and stated. Opinion: a judgement the author or a cited person holds. Inference: unstated but necessarily follows. Assumption: unstated but necessarily presupposed. Many students lose marks by treating a plausible continuation as an inference — plausible is not necessary.
• Quick test for an inference: *could the passage be entirely true and this option still false?* If yes, reject it.
• Tone vocabulary (learn to recognise, not to memorise lists):

• Tone answers are almost never at the extreme (*vitriolic, euphoric, contemptuous*) unless the passage is openly hostile; qualified tones win more often.
• A passage that presents two views and endorses neither is objective, even if one view is described at greater length.

Timing, selection and scoring strategy

• Select passages in sectional exams: skim the opening lines of all passages first, start with the one whose subject and sentence length you handle best, and leave the densest (abstract philosophy, dense economics) for last.
• Set a budget per passage and respect it; a passage where you are stuck after 8 minutes has already cost two easier questions elsewhere.
• Answer detail and vocabulary questions first within a set — they are fast and near-certain; inference and main-idea questions take longer.
• With negative marking, decide by the break-even probability, not by feel (see the formula box). Eliminating two of four options almost always makes an attempt worth it.
• Never leave a non-penalised question (GATE NAT/MSQ, many company tests with no negative marking) unattempted.

Comprehension at work and in interviews

• Technical and HR interviewers test the same skill verbally: *"Summarise this page in three sentences"*, *"What is this client actually asking for?"*, *"What has this email not said?"*
• At work it appears as reading a BRD/SRS before coding, a tender or RFP before bidding, an RBI/SEBI/GST circular before implementing, a contract clause before signing off, and a root-cause report before escalating.
• Précis writing (university exams, and real work as an executive summary): reduce to about one-third, keep only the main argument and its key support, use your own words, write in third person and the past/present tense of the original, and give it a title.

Key points

• Passage is the only authority - no outside knowledge, however true.
• Main idea - must cover the whole passage; a supporting example is never the main idea.
• Inference - one necessary step beyond the text; if it can be false while the passage is true, reject it.
• Assumption - the unstated premise; verify by negating it and watching the argument fall.
• Detail questions - locate the line; do not reason your way to them.
• Vocabulary in context - substitute each option into the sentence; context beats the dictionary.
• Function questions - answer "why is this sentence here", not "what does it say".
• Tone - prefer measured, qualified tones over extreme ones unless the text is openly hostile.
• Transition words (*however, therefore, moreover*) carry the argument's structure and flag the conclusion.
• Elimination over attraction - strike three options with a stated reason each.
• Trap families - extreme language, out of scope, half-right, distortion, too narrow, too broad, reversal.
• EXCEPT questions - mark every option true/false separately before choosing.
• Cannot say (banking format) means not determinable from the passage, not false.
• Reading speed - aim 200-250 wpm with 80%+ comprehension; effective rate = speed × comprehension.
• Break-even accuracy - attempt when your success probability exceeds penalty/(marks + penalty).
• Passage selection and a per-passage time budget usually add more marks than faster reading.
• Precis - about one-third length, own words, main argument only, with a title.
• Skim, scan, intensive read are three different speeds used for three different jobs.

Formulas and examples

Example 1 — reading speed and effective rate. A student reads a 750-word passage in 4 minutes 10 seconds and gets 3 of 5 questions right.

• t = 250 s = 250/60 = 4.167 min → R = 750 / 4.167 = 180 wpm.
• C = 3/5 = 0.60 → effective rate = 180 × 0.60 = 108 wpm.
• A second student reads at 150 wpm (5.0 min) but scores 5/5: effective rate = 150 × 1.00 = 150 wpm. The slower reader is the better reader — speed without comprehension is wasted.

Example 2 — building a time budget (CAT-style VARC). 40-minute section: 4 passages × 4 questions = 16 RC questions, plus 8 non-RC questions. Passages average 600 words.

• Reserve 12 min for the 8 non-RC questions → 90 s each. RC gets 40 − 12 = 28 min, i.e. 7 min per passage.
• Within 7 min: reading at 200 wpm takes 600/200 = 3 min, leaving 4 min for 4 questions = 60 s per question.
• If your natural speed is 150 wpm, reading takes 4 min and only 45 s per question remains — the fix is either faster reading or attempting 3 passages well instead of 4 badly.

Example 3 — attempts versus accuracy under negative marking (CAT, +3 / −1).

• Accuracy: A = 70%, B = 85.7%, C = 50%.
• Break-even p* = 1/(3+1) = 0.25. A is right to attempt widely because her hit rate (70%) is far above 25%; C is destroyed not by attempting more but by an accuracy below the useful range.

Example 4 — banking pattern, 0.25 penalty. IBPS PO prelims English: 30 questions, 1 mark each, 20 minutes, 0.25 penalty. A candidate attempts 24, of which 19 are correct.

• Net = 19 − (5 × 0.25) = 19 − 1.25 = 17.75.
• p* = 0.25 / (1 + 0.25) = 0.20, so any question where she can eliminate even one option (leaving 1-in-4 = 25% > 20%) is worth attempting. Blind guessing with 5 options (20%) is exactly break-even — pointless.

Example 5 — a solved passage. Read this (about 100 words), then the questions.

When India's banks began reporting UPI volumes, commentators read the growth
as proof that cash was dying. The data support a narrower claim. UPI has
replaced cash in small, frequent, urban payments - the chai, the auto fare,
the kirana bill - where the cost of handling coins was highest. Currency in
circulation, meanwhile, has continued to grow in absolute terms, because cash
still serves as a store of value and as the medium of the rural, the informal
and the unbanked. Digital payment is displacing cash as a medium of exchange
faster than it is displacing cash as a store of value.

• Q1 (main idea). Options of the type "UPI has destroyed cash" (reversal — the author argues the opposite), "Currency in circulation is growing" (too narrow — one supporting fact), "Rural India cannot use digital payments" (distortion and out of scope — the passage says cash *serves* that market, not that digital *cannot*). Correct: *digital payments are replacing cash in one of its functions, not in all of them.* Step: the second sentence ("a narrower claim") announces the thesis; the last sentence restates it.
• Q2 (inference). "It follows that cash in circulation will soon fall" — rejected: the passage says it has continued to grow and predicts nothing. "It follows that two different functions of money can change at different speeds" — accepted: it is a necessary restatement of the final sentence.
• Q3 (function). "the chai, the auto fare, the kirana bill" is there to illustrate the category "small, frequent, urban payments" — not to prove that UPI is cheap, and not to list India's main expenses.
• Q4 (vocabulary in context). *narrower* here means more limited in scope / more qualified, not "thinner" and not "biased". Substitute and reread: "The data support a more qualified claim" fits.

Where it is used

• IT services placement tests — TCS NQT, Infosys, Wipro Elite NLTH, Accenture, Capgemini and Cognizant all run an English section with short RC passages; clearing the sectional cut-off is mandatory regardless of coding scores.
• MBA and government entrances — CAT, XAT, SNAP, NMAT, CMAT (VARC), IBPS/SBI PO and clerk, RBI Grade B, SSC CGL/CHSL, RRB NTPC, LIC AAO and the CSAT paper of UPSC Prelims.
• GATE and university exams — GATE General Aptitude (15 marks across all papers); AICTE/CBCS courses in Technical English, Communication Skills and Business Communication, with unseen passages, précis and comprehension questions.
• Business analyst, consulting and product roles — reading BRDs, SRSs, RFPs, tender documents and client mail trails and extracting the actual requirement; the ambiguity you fail to notice becomes a defect later.
• Banking, insurance, legal and compliance work — RBI master directions, SEBI and IRDAI circulars, GST notifications, loan and policy documents, where a misread clause has a rupee cost.
• Testing, support and technical writing — reading specifications to write test cases, reading logs and vendor documentation, condensing release notes and root-cause analyses into summaries management will read.

Interview questions

• Q: What is the difference between an inference and an assumption? A: An inference is a conclusion that follows from what is stated; an assumption is an unstated premise the argument needs in order to hold. Inferences come after the argument, assumptions sit underneath it.
• Q: How do you find the main idea of a passage quickly? A: Read the first and last paragraphs closely, tag each middle paragraph with two or three words, and look for the sentence after a turn word such as "however" — the thesis usually sits there.
• Q: An option is factually true in the real world but not mentioned in the passage. Is it correct? A: No. RC answers must come from the passage alone; outside truth is one of the commonest traps.
• Q: What is your reading speed and how would you measure it? A: Words in the passage divided by minutes taken; around 200-250 wpm with 80 per cent comprehension is the useful target, and effective rate is speed multiplied by comprehension.
• Q: How do you handle an "all of the following are true EXCEPT" question? A: Mark each option true or false against the text separately, then pick the one marked false — attempting it by intuition inverts the task.
• Q: Two options look equally good. What next? A: Find the single differing word — usually a quantifier, a modal or a scope term — and check which one the passage actually supports.
• Q: Why are extreme words like "always" or "only" usually wrong? A: Academic and journalistic writing is hedged, so a universal claim normally overstates what the author committed to; it is right only if the passage itself is absolute.
• Q: Would you read the questions before the passage? A: For short, fact-based passages, yes — most questions are detail-based and scanning works. For long argumentative passages it wastes time, because inference and tone questions need the whole structure anyway.
• Q: In a test with 0.25 negative marking and four options, is guessing worth it? A: Break-even probability is 0.25/1.25 = 20 per cent; blind guessing at 25 per cent is marginally positive, and eliminating even one option makes it clearly worth attempting.
• Q: How do you determine the author's tone? A: From evaluative adjectives, concessions and rhetorical questions rather than the subject matter; most correct tone answers are measured, such as critical, cautious or analytical.
• Q: You get a 30-page requirements document and two hours. How do you read it? A: Skim the contents and section headings first, scan for scope, assumptions and exclusions, read those sections intensively, list the ambiguities, and send them as questions before starting work.
• Q: Summarise an article you read recently in three sentences. A: State the author's central claim, the main evidence, and the limitation or counter-view — that structure shows comprehension rather than recall.

Common mistakes

• Answering from general knowledge or personal opinion instead of the passage, especially on familiar topics such as technology, climate or education.
• Treating a plausible continuation as an inference; correct inferences must be necessary, not merely likely.
• Picking a true supporting detail as the main idea, or picking a sweeping statement the passage only partly supports.
• Reading only half the option; half-right options are written specifically to be chosen by fast readers.
• Mis-marking EXCEPT, NOT and "Cannot say" questions by reversing the required judgement.
• Chasing speed and losing accuracy — a student at 300 wpm with 40% accuracy scores less than one at 180 wpm with 85%.
• Spending 12 minutes on one hard passage and leaving an easy one unread; and in interviews, retelling the passage line by line when asked for a summary instead of giving the claim, evidence and limitation.

Related topics

Para jumbles and sentence completion · Cloze tests and sentence improvement · Vocabulary, synonyms and antonyms · Critical reasoning: statement, assumption and conclusion · Précis, report and email writing · Group discussion and verbal communication.