Met Police Exam cheat sheet

Met Police values framework (PEEL): Policing by Consent, Empathy, Integrity, Transparency - use these as the four lenses when ranking responses.

Ranking rule: first protect life and safety, then preserve evidence and legal integrity, then resolve the community issue, then manage paperwork and reputation.

A response that defers a decision to a supervisor is usually 'least effective' when the matter is within your own authority and time-critical.

'Most effective' answers act within role, record rationale, and escalate only where policy mandates - e.g. any use of force, any detention, any safeguarding referral.

National Decision Model (NDM): Code of Ethics → gather information → assess threat and risk → consider powers and policy → take action → review. Answers that skip 'review' are weaker.

Safeguarding triggers (child protection, domestic abuse, vulnerability, mental health crisis) require referral even when the victim declines - consent is not the gate.

Procedural justice markers expected in answers: voice (let people explain), neutrality (explain decisions), respect, and trustworthy motives.

Do not select answers that involve deception, unauthorised disclosure, or 'sorting it out informally' to avoid recording - these are scored as counterproductive.

Percentage change = (new − old) ÷ old × 100; percentage points compare two percentages directly and are not the same as percent change.

Reverse percentage: if a value is after a p% increase, original = value ÷ (1 + p/100); after a decrease, divide by (1 − p/100).

Rates per 1,000 population: rate = (count ÷ population) × 1,000; compare areas only after converting counts to rates.

Weighted average = Σ(value × weight) ÷ Σ(weights); never average percentages from unequal-sized groups unweighted.

Ratio scaling: to split a total T in ratio a : b : c, share = T × part ÷ (a + b + c).

Compound change over n periods: final = start × (1 + r)^n; for small r, approximate growth ≈ n × r.

Rounding trap: round only at the final step; interim rounding on multi-step percentage questions typically produces a wrong option.

Sanity check every result against the table's magnitude - an answer outside the range of the source data signals a misread unit (thousands vs millions).

True means the statement follows logically from the passage alone; False means it contradicts it; Cannot Say means the passage is silent or ambiguous.

'Cannot Say' is correct whenever the passage gives insufficient information - outside knowledge must never upgrade it to True.

Quantifier discipline: 'all', 'only', 'always', 'never' statements are False if the passage supports even one exception.

Distinguish a claim the author makes from a claim the author reports others making; only the former supports a True answer.

Beware scope shifts: a passage about one division, year, or category does not support statements about the whole organisation or other periods.

Cause versus correlation: a passage stating two things co-occur does not support a statement that one caused the other.

Work passage-by-passage and answer from the text, not from plausibility - plausible-but-unsupported options are the main distractor type.

Time rule: if a statement needs more than one re-read to justify, it is usually Cannot Say rather than True.

Chronological order with timestamps is the default structure for an incident report: who, what, when, where, how, and what action followed.

Use past tense, third person, and full names/roles at first mention; never write 'he/she' without an antecedent in the same paragraph.

Distinguish fact from opinion explicitly: 'I observed…', 'X stated…', 'In my opinion…' - opinion must be labelled, not blended into fact.

Record exact words in quotation marks for admissions, threats, or complaints; paraphrase only when labelled as a summary.

Every use of force, search, detention, or seizure needs the legal power relied on and the grounds for it stated in the report.

Avoid conclusory language ('aggressive', 'uncooperative') unless tied to observable behaviour: 'shouted, clenched fists, refused three requests to step back'.

The 5 Cs check: Clear, Concise, Complete, Correct, Chronological - a report missing any of these is returned for amendment.

Write for a reader who was not there and may read it months later in court: define abbreviations, spell out acronyms once, avoid jargon.

Syllogism rule: a conclusion is valid only if it must be true whenever all premises are true; one counterexample invalidates it.

Contrapositive is always valid: from 'if P then Q' infer 'if not Q then not P'; the inverse ('if not P then not Q') is invalid.

Necessary vs sufficient: P sufficient for Q means P→Q; P necessary for Q means Q→P. Confusing them is the most common trap.

Ordering puzzles: draw a linear or grid diagram and place the most constrained item first; re-check each clue once after placement.

For 'which assumption is required', negate the option - if the argument collapses without it, it is the necessary assumption.

Correlation-to-cause, hasty generalisation, and false dilemma are the recurring flaws; name the flaw before choosing the answer.

Set reasoning: 'some A are B' does not imply 'some B are not A'; 'no A are B' does imply 'no B are A'.

Work forwards from the data and backwards from the answer options; meeting in the middle is faster than full derivation.

Check character-level detail first: VRM (vehicle registration), dates, times in 24-hour format, and collar/shoulder numbers are the highest-error fields.

Confusable pairs to verify letter by letter: 0/O, 1/I/l, 5/S, 8/B, 2/Z; and in VRMs the format is two letters, two numbers, three letters (current series).

Date format standard is DD/MM/YYYY; '03/04' is ambiguous and must be written unambiguously in any record.

Time is recorded in 24-hour clock with four digits (e.g. 07:05, 23:40); midnight is 00:00, not 24:00.

Cross-field consistency: the time of the incident must precede the time of attendance, which must precede the time of report.

Transposition and omission are the most frequent data-entry errors - read numbers in groups (e.g. 4-3-2) rather than as a whole string.

When comparing two lists, compare line by line against a fixed reference rather than scanning for differences by eye.

Any figure that appears twice in a document (totals, counts, dates) must be reconciled; a mismatch invalidates the whole record.