Inferences
8 questions · Recommended 5-6 minutes
What it tests. Judging whether a conclusion is genuinely supported by the text or is a logical leap.
Worked example. Given a passage where a carbon tax was introduced and five automakers then relocated to untaxed jurisdictions, 'the tax was the primary catalyst for relocation' is Probably True (strong chronological and destination correlation, but motives are not explicitly stated).
Common traps. Injecting external knowledge - e.g. knowing from the news the automakers actually left over semiconductor shortages.
How to handle it. Watch qualifiers like 'all', 'some', 'solely' and 'primarily'; a jump from a subset (manufacturing) to a macro category (all sectors) is usually a trap.
Recognizing Assumptions
8 questions · Recommended 5 minutes
What it tests. Identifying the unstated premises someone must take for granted to make an assertion.
Worked example. For 'to win premium M&A mandates we must upgrade to an AI document-review platform', the assumption 'clients value or require that capability' is Made; 'the upgrade will cut trainee headcount' is Not Made.
Common traps. Agreeing with an assumption because it seems realistic or sensible in a professional setting.
How to handle it. Apply the Negative Rejection Test: invert the assumption; if that destroys the original statement, the assumption is Made.
Deductions
8 questions · Recommended 6 minutes
What it tests. Syllogistic, strict logical reasoning where a conclusion must unavoidably follow.
Worked example. From 'all senior partners are corporate-law experts; some experts are arbitrators', the conclusion 'some senior partners are arbitrators' Does Not Follow - the arbitrator subset may not overlap the partner subset.
Common traps. Using everyday language norms - in conversation 'some' implies 'not all', but in strict logic 'some' means at least one and potentially all.
How to handle it. Sketch quick Venn diagrams or symbolic strings on scratch paper to map intersections and exclusions.
Interpretation
8 questions · Recommended 6 minutes
What it tests. Weighing evidence and judging a conclusion solely on the provided data, without extrapolation.
Worked example. Given resolution rates of 75% for IP disputes versus 25% for infrastructure disputes within six months, 'IP cases resolved faster on average' Follows; 'evidence volume is the reason infrastructure takes longer' Does Not Follow (the data shows no causal proof).
Common traps. Treating an author's or study's subjective conclusion in the text as an objective, data-verified fact.
How to handle it. Be extremely literal: if a statement requires inferring a causal mechanism and the text only gives correlation, it Does Not Follow.
Evaluation of Arguments
8 questions · Recommended 5-6 minutes
What it tests. Distinguishing strong from weak arguments by relevance and materiality.
Worked example. On mandating tech expertise on FTSE 100 boards, an argument citing a 40% rise in cyber-attacks and severe shareholder-value destruction is Strong; one based on board members' personal discomfort with technology is Weak.
Common traps. Confusing an argument you personally agree with for a strong argument.
How to handle it. A strong argument must be both highly relevant to the exact question and highly material (economic loss, safety, legal non-compliance, systemic risk); true-but-trivial is automatically Weak.