Master Critical Reasoning Inference

Inference and conclusion questions typically make up 20-25% of the Critical Reasoning component in CAT VARC. Given that CR usually has 12-16 questions total across the section, you can expect 3-4 inference questions per CAT.

These questions are identifiable by specific phrases: “must be true”, “can be properly inferred”, “properly concluded”, “follows logically”, or “best supported by the statements above”. Some ask what must be true, others ask what can be reasonably concluded. The distinction matters because “must” requires absolute certainty while “best supported” allows for strong probability.

The percentage makes these questions significant but not dominant. Your CR preparation should cover all question types—assumptions, strengthen/weaken, evaluate, bold-face—but inference questions deserve focused practice because they test pure logical reasoning without the complexity layers of other types.

Execute a rapid 5-step method: lock premises as facts (15 seconds), decode question type (5 seconds), combine information systematically (20 seconds), test options by trying to break them (20 seconds), prefer safe wording when choosing between candidates (10 seconds). This gives you 70 seconds total, leaving buffer for difficult premises.

The key is pre-work before reading options. Most test-takers jump straight to answer choices, trying to evaluate each against premises. This wastes time and invites confusion. Instead, spend 30-35 seconds understanding what the premises actually establish. Rephrase complex sentences into simple logical relations. Identify any chains, overlaps, or clear inferences you can draw immediately.

Then scan options quickly looking for safe language. Options with “some”, “may”, “often”, “tends to” should be evaluated before extreme options with “always”, “never”, “only”, “all”. Moderate language usually indicates correct answers in inference questions because premises are deliberately limited in scope.

First, verify they’re actually contradictory rather than just complex. CAT premises are designed to fit together logically, even if the connection isn’t immediately obvious. Apparent contradictions usually result from different scope or different contexts within the same argument.

Break down each premise individually. Identify the exact claim being made, including all qualifiers like “some”, “most”, “in certain conditions”, “typically”, “when X occurs”. These qualifiers prevent contradictions by limiting scope. Premise 1 might say “employees prefer remote work” while Premise 2 says “productivity increases in office settings”. These aren’t contradictory—they describe preference versus productivity, two different dimensions.

If premises genuinely contradict after careful reading, treat the question as potentially flawed and mark your best answer based on majority of premises. However, this is extremely rare in official CAT questions. More likely, you’re missing a qualifier or misreading a conditional relationship.

When confused, simplify using symbols or shorthand. Convert “Companies that invest in training see productivity gains” to “Training → Productivity↑”. Convert “Most employees prefer remote work” to “Employees_majority → prefer_remote”. This notation makes relationships clearer and helps spot how premises combine.

“Must be true” questions require absolute certainty. The correct answer holds in every possible scenario where premises are true, with zero exceptions allowed. These test deductive reasoning—if premises are facts, what necessarily follows? You should be able to defend the answer against any challenge by showing it’s forced by the premises.

“Best supported” or “most reasonable conclusion” questions allow strong probability rather than absolute necessity. The correct answer fits premises most tightly with fewest additional assumptions, but alternatives might exist. These test which inference is safest given limited information, not what must occur.

Testing differs for each type. For must-be-true, ask “Can I imagine any scenario where premises hold but this answer is false?” If yes, eliminate it. For best-supported, ask “How many unstated assumptions does this require?” Pick the option needing the fewest additional beliefs.

In practice, many test-takers treat both types identically and suffer for it. Must-be-true tolerates no speculation—safe, boring restatements of premise combinations often win. Best-supported allows modest extension beyond literal premises as long as the extension is strongly indicated rather than invented.

Trap answers work by triggering System 1 thinking—fast, intuitive, pattern-matching cognitive processes that don’t verify logical connections carefully. They feel right because they align with real-world knowledge, common sense, or general topic familiarity, even though premises don’t support them.

CAT constructs these traps deliberately by including options that: (1) discuss the topic intelligently, (2) use vocabulary from the premises, (3) make claims that are generally true in reality, (4) extend premises in ways that seem reasonable. All of these feel appropriate in casual reasoning. None make the option logically forced by premises.

Counter this by treating premises as the only reality that exists. Real-world facts, your personal knowledge, common sense—all irrelevant. Only what’s stated or necessarily implied counts. This requires conscious override of intuition.

Build the habit of asking “Where in the premises is this supported?” for every attractive option. If you can’t point to specific premise language that forces or strongly backs the claim, it’s probably a trap. Options that feel too obvious or too clever both deserve skepticism.

Target 85% accuracy through systematic elimination rather than positive identification. Your goal isn’t finding the perfect answer—it’s ruling out three obviously wrong options, then choosing between two plausible candidates based on tighter logical fit.

Step one: eliminate extreme language unless explicitly supported. “Always”, “never”, “only”, “all”, “none”—these demand total certainty that premises rarely provide. Cutting these immediately improves odds.

Step two: eliminate options requiring causal assumptions. If premises give correlation or coexistence, don’t accept conclusions about causation, motivation, or policy recommendations. “X happened after Y” doesn’t support “Y caused X” or “We should do Y to get X”.

Step three: eliminate options that bring in outside information. The correct answer might seem boring, obvious, or trivial—that’s fine. Sophisticated-sounding options that add context or broader analysis usually exceed premise scope.

After eliminating three options, you’re choosing between two that both seem reasonable. At this stage, compare assumption count. Which option requires believing fewer unstated things? Which sticks closer to actual premise wording? Which makes the smaller logical jump? Pick that one.

Rephrase selectively based on premise complexity. Simple, clear premises need no rephrasing—it wastes time without adding clarity. Complex premises with multiple clauses, conditionals, or quantifiers benefit enormously from translation into simpler logical notation.

The benefit isn’t the rephrasing itself but the verification that you understood correctly. When you rephrase “Companies that invest heavily in training see productivity increases within two years” as “Heavy training → Productivity↑ within 2 years”, you’re testing your comprehension. If you can’t simplify it, you probably misunderstood the original claim.

Use shorthand efficiently: arrows for causes/leads to, inequality symbols for comparisons (>, <), "all/some/none" for quantifiers, abbreviations for entities. "Most engineers prefer remote work under flexible schedules" becomes "Engineers_>50% → prefer_remote IF flex_schedule”. This takes 5-8 seconds but prevents misreading scope or conditions.

Don’t rephrase during option evaluation—that’s too late. Do it immediately after reading premises, before seeing answer choices. This pre-work prevents you from being swayed by how options frame the information. You’ve already locked in what the premises establish independently of how answers present themselves.