What “answer engineering” means here
You don’t win CELPIP Reading by “feeling” the right option—you win by disqualifying wrong ones fast. This section gives you a mechanical system: detect typical wrong-answer archetypes, run a short set of kill-tests, and keep only the choice that survives text-based scrutiny.
The Four Horsemen of Wrong Answers
Most bad options are one (or more) of these. Name them as you see them; eliminate on sight.
-
Outside Scope
Mentions a topic in the passage but goes beyond what’s discussed (adds new reasons/claims, widens the group or time).
Tell: words like all, always, everywhere where the passage talks about a smaller group/time.
-
Over-Strong
Upgrades may/usually/some to will/always/most or must.
Tell: absolute verbs/adverbs that don’t match the passage’s hedged tone.
-
Misaligned Speaker
Attributes a claim to the wrong voice (author vs commenter vs other resident).
Tell: the option fits the text somewhere, but not the named speaker.
-
Time/Quantity Mismatch
Breaks before/after/by/until logic; flips at least ↔ exactly; misreads am/pm, zones, units, or footnotes.
Tell: numbers/dates look right at a glance but fail the digit-by-digit or footnote check.
The Five Kill-Tests (run them in order)
Use these on every item. Stop at the first fail.
-
Finger-on-Text Proof
Can you touch the line (or exact diagram cell) that supports the choice? If not, it’s suspect.
-
Strength Match
Compare modality words: must/should/may, will/might, definitely/likely. If strength differs, kill it.
-
Scope Match
People/time/place/quantity must match exactly: residents vs visitors, evenings vs weekends, ≥ / ≤, zone 2→3 not 1→3.
-
Polarity & Logic
Watch not, only, except, unless, before, after, by, until, from. One toggle flips the truth.
-
Speaker/Stance Alignment
Who would say this? If the option’s voice doesn’t match (or the stance intensity is wrong), out it goes.
If multiple options survive, keep the one that is closest to the text in wording/strength and least assumptive.
The “ACE” Filter (fast smell-test)
When you first read options, tag each as:
- A — Aligned: matches text’s idea, strength, scope. Candidate to verify.
- C — Contradicts: conflicts with a clear line/number/footnote. Eliminate.
- E — Extra: adds new claims/causes/conditions not in the text. Eliminate.
You should have 1–2 A’s after a quick pass; then run the Kill-Tests on those.
Numerical & Time Logic—Zero-Error Protocol
Numbers and time wording are a free way to eliminate options quickly.
- Digit-by-Digit: Read numbers digit by digit ($31 ≠ $13).
- Window math:
- by 6:00 = ≤ 6:00
- until 6:00 = up to 6:00, not after
- after 6:00 = > 6:00
- from 6:00 = includes 6:00
- Quantifiers: at least 3 = 3,4,…; more than 3 = 4,5,…; at most 3 = 3,2,1,0.
- Units: km vs m, $ vs ¢, per night vs total—match the label before you compute.
One mismatch → instant elimination.
Language Controls—Strength, Scope, Polarity (SSP)
When two options look good, the survivor is the one with perfect SSP match.
- Strength ladder: must > should > may.
- Scope ladder: all > most > some > few.
- Polarity toggles: only / except / unless / not until / before / after / by / until / from.
If the text says may and the option says will, it dies—no matter how elegant it sounds.
Final 20-Second Audit (before you click next)
- Any blanks? Fill them.
- Any chosen option with an absolute word that the text doesn’t justify? Downgrade it.
- For the comment/reply, does the sentence still read naturally after your drop-down choice? If tone or facts wobble, replace it.
If you can’t prove it, you shouldn’t choose it. Engineer the answer—don’t guess it.