CAST Permanent cheat sheet

True means the statement follows logically from the passage alone; False means it contradicts the passage; Cannot Say means the passage neither confirms nor denies it, even if outside knowledge suggests an answer.

Judge only against the passage text. Prior knowledge, plausibility, or general opinion is irrelevant and must not shift a Cannot Say to True or False.

A statement that overstates the passage (all, always, never, only) is usually False or Cannot Say; check whether the passage hedges with some, may, often, tends to.

Distinguish contradiction from silence: False requires the passage to state the opposite; if the passage is merely silent on the point, the answer is Cannot Say.

Watch quantifier and scope shifts: 'most' vs 'all', 'some' vs 'none', 'more likely' vs 'causes'. A changed quantifier usually changes the verdict.

Correlation in the passage does not license a causal statement; a statement claiming cause is Cannot Say unless the passage asserts causation.

Comparative statements need the same comparison set: a claim about 'the largest' is Cannot Say if the passage only compares two of several items.

Time and tense matter: a statement about the present or future is Cannot Say if the passage only describes the past.

Percentage change = (new - old) / old x 100; reverse a percentage increase by dividing by (1 + rate), e.g. remove 20% VAT by dividing by 1.20.

Percentage points vs percent: a move from 15% to 18% is +3 percentage points but a 20% relative increase (3/15).

Ratios scale by a common factor: if a:b = 3:5 and a + b = 64, one part = 64/8 = 8, so a = 24 and b = 40.

Currency and unit conversion: multiply by the exchange or conversion factor in the direction that cancels the unwanted unit; check the result's magnitude for plausibility.

Weighted averages: overall mean = sum(value x weight) / sum(weights); a simple mean of group means is wrong unless group sizes are equal.

Compound change: successive changes multiply, so +10% then -10% gives 1.10 x 0.90 = 0.99, a net -1%, not 0%.

Rate problems: distance = speed x time; work rate adds as 1/t1 + 1/t2 = 1/together; use consistent units before combining.

Estimate first by rounding to one significant figure to eliminate options, then compute exactly only if two options remain close.

Track each element separately across the series: shape type, count, colour/fill, position, rotation, and size often follow independent rules.

Rotation rules are stated in degrees and direction, e.g. +90 degrees clockwise per step; verify against two consecutive transitions before applying.

Alternating series split into odd and even positions when a single rule fails to fit all items.

Count arithmetic: element counts often follow +1, -1, doubling, or a repeating cycle; write the count sequence out numerically.

Distinguish rotation from reflection: reflection reverses orientation (a letter-like shape becomes mirrored), rotation preserves it.

Superposition rules in matrix puzzles: overlay cell 1 and cell 2; common conventions are keep-common, cancel-common (XOR), or add-without-overlap.

Position movement wraps or bounces at the grid edge; check whether an element that exits reappears on the opposite side.

Fill and shading rules are separate from shape rules; a distractor often keeps the shape rule but breaks the shading rule.

Answer as the ideal candidate in the stated role, not as yourself; the scoring key rewards the behaviour the employing organisation prescribes.

Rank by priority order typical of public-sector frameworks: safety and legality first, then delivery to the citizen/customer, then team and stakeholder communication, then efficiency.

Escalate when the issue exceeds your authority, is a safety or legal risk, or affects someone outside your remit; handle routine matters yourself first.

Prefer actions that gather facts before acting: check records, clarify the brief, or consult the relevant policy, rather than guessing or ignoring.

Communication rules: inform affected parties promptly, keep the line manager informed of material risks, and avoid bypassing the chain of command without cause.

Avoid extremes in response options: 'do nothing', 'ignore the rule', 'confront publicly', and 'wait for someone else' are usually ineffective.

In 'most/least effective' items, judge each option independently on effectiveness, not on how it compares with the other options.

Competency questions want a concrete STAR structure: Situation, Task, Action, Result, with the action owned by you and the result evidenced by a specific outcome.

True means the statement follows logically from the passage alone; False means it contradicts the passage; Cannot Say means the passage neither confirms nor denies it, even if outside knowledge suggests an answer.

Judge only against the passage text. Prior knowledge, plausibility, or general opinion is irrelevant and must not shift a Cannot Say to True or False.

A statement that overstates the passage (all, always, never, only) is usually False or Cannot Say; check whether the passage hedges with some, may, often, tends to.

Distinguish contradiction from silence: False requires the passage to state the opposite; if the passage is merely silent on the point, the answer is Cannot Say.

Watch quantifier and scope shifts: 'most' vs 'all', 'some' vs 'none', 'more likely' vs 'causes'. A changed quantifier usually changes the verdict.

Correlation in the passage does not license a causal statement; a statement claiming cause is Cannot Say unless the passage asserts causation.

Comparative statements need the same comparison set: a claim about 'the largest' is Cannot Say if the passage only compares two of several items.

Time and tense matter: a statement about the present or future is Cannot Say if the passage only describes the past.

Percentage change = (new - old) / old x 100; reverse a percentage increase by dividing by (1 + rate), e.g. remove 20% VAT by dividing by 1.20.

Percentage points vs percent: a move from 15% to 18% is +3 percentage points but a 20% relative increase (3/15).

Ratios scale by a common factor: if a:b = 3:5 and a + b = 64, one part = 64/8 = 8, so a = 24 and b = 40.

Currency and unit conversion: multiply by the exchange or conversion factor in the direction that cancels the unwanted unit; check the result's magnitude for plausibility.

Weighted averages: overall mean = sum(value x weight) / sum(weights); a simple mean of group means is wrong unless group sizes are equal.

Compound change: successive changes multiply, so +10% then -10% gives 1.10 x 0.90 = 0.99, a net -1%, not 0%.

Rate problems: distance = speed x time; work rate adds as 1/t1 + 1/t2 = 1/together; use consistent units before combining.

Estimate first by rounding to one significant figure to eliminate options, then compute exactly only if two options remain close.

Track each element separately across the series: shape type, count, colour/fill, position, rotation, and size often follow independent rules.

Rotation rules are stated in degrees and direction, e.g. +90 degrees clockwise per step; verify against two consecutive transitions before applying.

Alternating series split into odd and even positions when a single rule fails to fit all items.

Count arithmetic: element counts often follow +1, -1, doubling, or a repeating cycle; write the count sequence out numerically.

Distinguish rotation from reflection: reflection reverses orientation (a letter-like shape becomes mirrored), rotation preserves it.

Superposition rules in matrix puzzles: overlay cell 1 and cell 2; common conventions are keep-common, cancel-common (XOR), or add-without-overlap.

Position movement wraps or bounces at the grid edge; check whether an element that exits reappears on the opposite side.

Fill and shading rules are separate from shape rules; a distractor often keeps the shape rule but breaks the shading rule.

Answer as the ideal candidate in the stated role, not as yourself; the scoring key rewards the behaviour the employing organisation prescribes.

Rank by priority order typical of public-sector frameworks: safety and legality first, then delivery to the citizen/customer, then team and stakeholder communication, then efficiency.

Escalate when the issue exceeds your authority, is a safety or legal risk, or affects someone outside your remit; handle routine matters yourself first.

Prefer actions that gather facts before acting: check records, clarify the brief, or consult the relevant policy, rather than guessing or ignoring.

Communication rules: inform affected parties promptly, keep the line manager informed of material risks, and avoid bypassing the chain of command without cause.

Avoid extremes in response options: 'do nothing', 'ignore the rule', 'confront publicly', and 'wait for someone else' are usually ineffective.

In 'most/least effective' items, judge each option independently on effectiveness, not on how it compares with the other options.

Competency questions want a concrete STAR structure: Situation, Task, Action, Result, with the action owned by you and the result evidenced by a specific outcome.