On this page
Try a free Reading mock
Every part in exam order on the official timer, with your answers checked at the end.
Choose the reading method from the task, identify the question's demand, find the evidence, then test every option against actor, condition, time and strength of claim.
That is the real CELPIP reading strategy. It is not a trick about absolute words or a reading-order rule you repeat for every part. It is the discipline of checking all four options against the text before you commit, because the wrong options in CELPIP are written to look right if you only check one feature and guess the rest.
The four Reading parts — correspondence, diagram, information, and viewpoints — each test a different reading job. Part 1 wants you to understand relationships and purpose. Part 2 wants you to cross-reference constraints. Part 3 wants you to distinguish stated facts from logical inferences. Part 4 wants you to map who believes what and why they believe it, including qualifications they add.
This guide shows you the task-specific method for each part, then walks through four original worked examples that show why every distractor fails.
Key takeaways
- The reading method depends on the task: match relationships in correspondence, check all constraints for diagrams, separate stated from unstated in information passages, and map person to position to reason in viewpoints.
- Wrong options fail on actor (who the statement is about), condition (a requirement the statement ignores), time (past, present, future), or strength (absolute claim when the text hedges, or vice versa).
- Review your errors as: locator (missed where the evidence was), paraphrase (synonym confused you), scope (chose a general claim when the text said something narrower), wrong speaker (attributed a view to the wrong person), missed condition (ignored "if" or "unless"), or unknown word.
- Do not rely on "always" and "never" as signals that an option is wrong. Official materials contain correct answers with absolute language when the passage makes an absolute statement.
The four-part strategy table
Each Reading part uses a different structure, and that structure tells you what to look for. Here is the method for each:
| Part | What it tests | How to read it | What makes an option wrong |
|---|---|---|---|
| Part 1: Correspondence | Understanding relationships, purpose, and tone | Read for who is writing to whom, why, and what they want or offer. The reply completes a second person's response. | Tone mismatch, wrong relationship, purpose stated differently than in the text, or reply content that contradicts the original. |
| Part 2: Apply a Diagram | Cross-referencing constraints | Read the diagram for names, times, prices, ages, and other requirements. Match each blank in the email to the row in the diagram that fits all constraints. | Fits one constraint but violates another: the right workshop on the wrong day, the right day at the wrong time, or a program with an age limit the writer's situation doesn't meet. |
| Part 3: Information | Matching statements to paragraphs based on exact information | Read each paragraph for what it explicitly states. A statement matches paragraph A only if A contains the exact information. Option E ("not given") is correct when no paragraph contains that information. | Common error patterns: logical inference (the passage suggests it, but never states it), scope error (the statement is more general or more specific than what the paragraph commits to), superlative escalation (passage says "third most popular," statement says "most popular"), invented mechanisms, or choosing a paragraph because it mentions the topic when it doesn't contain the exact information. |
| Part 4: Viewpoints | Mapping who believes what, why, and with what qualifications | Read for each person's name, their position, the reason they give, and any conditions or limits they add ("only if," "unless," "until"). Build a grid: person → view → reason → condition. | Attributing a position to the wrong person, omitting a condition the speaker includes, choosing a reason the passage does not connect to that person, or treating a qualified opinion as absolute. |
For all four parts: when a question asks "what must be true," test every option against the passage. Do not stop at the first one that looks good. The test knows which wrong option will look right if you only check one feature.
To run the full format with timing, start with the CELPIP Reading practice test. For drills by part, use the CELPIP Reading practice questions page.
Worked example 1: Correspondence (Part 1 style)
Original exercise. Read the following email, then answer the question below.
Subject: Community Garden Volunteer Orientation — Date Change
Dear Community Garden Volunteers,
The volunteer orientation originally scheduled for Monday, April 7, at 6:00 p.m. has been moved to Wednesday, April 9, at 6:00 p.m. at the garden shed. The session will cover plot assignments, watering schedules, and tool storage.
If you already registered for Monday's session, your registration automatically transfers to Wednesday — you do not need to register again. If you are a new volunteer and have not yet registered, please send your name and preferred plot size to gardencoord@city.ca by Tuesday, April 8.
We look forward to seeing you on Wednesday.
Regards,
Sanjay Mehta
Garden Coordinator
Question: What must an existing volunteer who registered for the Monday session do?
A. Attend the orientation on Wednesday, April 9
B. Re-register for the Wednesday session by Tuesday
C. Email their preferred plot size to the coordinator
D. Confirm their attendance by Monday, April 7
Answer: A
Why A is correct: The email explicitly states, "If you already registered for Monday's session, your registration automatically transfers to Wednesday — you do not need to register again." The existing volunteer's only obligation is to attend on the new date.
Why each distractor fails:
- B fails on condition: The passage says existing registrants "do not need to register again." Option B invents a re-registration requirement the email explicitly removes.
- C fails on actor: The instruction to "send your name and preferred plot size" applies only to "new volunteer[s] who have not yet registered." Existing volunteers are not asked to send anything.
- D fails on time and action: Monday, April 7 is the cancelled date. The email gives no instruction to confirm attendance by that day; it only says the session has moved.
Evidence location: Second paragraph, first sentence. The transfer is automatic.
Worked example 2: Apply a Diagram (Part 2 style)
Original exercise. Use the workshop schedule below to complete the sentence that follows.
Community Centre Workshops — March 2026
| Workshop | Day | Time | Cost | Age requirement |
|---|---|---|---|---|
| Pottery Basics | Tuesday | 6:00–8:00 p.m. | $15 | 16+ |
| Teen Painting | Thursday | 6:00–8:00 p.m. | $10 | 12–17 |
| Watercolour Intro | Thursday | 2:00–4:00 p.m. | Free | 18+ |
Sentence to complete: Maya is 14 years old and is available only on Thursday evenings. The workshop that fits her schedule is ____________.
A. Pottery Basics
B. Teen Painting
C. Watercolour Intro
D. None of the listed workshops
Answer: B
Why B is correct: Teen Painting runs on Thursday at 6:00–8:00 p.m. and accepts ages 12–17. Maya is 14 (within the age range) and available Thursday evenings. All three constraints match.
Why each distractor fails:
- A fails on day: Pottery Basics runs on Tuesday, and Maya is available only on Thursday evenings.
- C fails on time and age: Watercolour Intro runs Thursday afternoon (2:00–4:00 p.m., not evening) and requires age 18+. Maya is 14, and she needs an evening session.
- D is wrong because B fits: When one option satisfies all constraints, "none" is incorrect.
Method for diagram questions: List every constraint the sentence gives (age 14, Thursday, evening). Check each workshop row. The moment one constraint fails, eliminate that row and move to the next.
Worked example 3: Information (Part 3 style)
Original exercise. Read the passage below, then decide whether each statement is supported by Paragraph A, Paragraph B, or is Not Given (E).
Passage: City Bike-Share Pilot Program
Paragraph A
The city launched a six-month bike-share pilot on June 1, 2026, with 200 bikes placed at ten docking stations across downtown. Riders use a mobile app to unlock a bike and return it to any station. The pilot will run until November 30, 2026.
Paragraph B
Each docking station includes a repair kiosk with a tire pump and basic tools. Riders who notice a damaged bike can report it through the app, and the bike is automatically locked out of the system until maintenance clears it. The city's transportation department reviews all damage reports weekly.
Statement 1: The city will make the bike-share program permanent after November 30, 2026.
Answer: E (Not Given)
Why E is correct: Paragraph A says the pilot "will run until November 30, 2026," but it says nothing about what happens after that date. The passage does not say whether the program will become permanent, end, or extend for another trial period.
Why this is not B: Paragraph B covers damage reporting and repair kiosks. It does not mention the program's future. Absence of information about permanence in every paragraph means the statement is Not Given.
Common trap: Readers infer that a successful pilot becomes permanent. Inference is not the same as a stated fact. If the passage does not say it, the answer is E.
Statement 2: Riders can return a bike to any docking station.
Answer: A
Why A is correct: Paragraph A explicitly states, "Riders use a mobile app to unlock a bike and return it to any station." The statement matches the exact information in the text.
Why this shows Part 3's matching rule: In Part 3, you match each statement to the paragraph that contains that exact information. The official CELPIP Reading Pro study pack (page 9, checked November 8, 2026) instructs you to "Read that part of the paragraph carefully to make sure the exact information in the statement appears in the text" and to "Choose 'E' if the information is NOT stated in the reading." Option E is used for unsupported claims — common error patterns include superlative escalation, invented mechanisms, or unstated consequences — not for paragraph-topic matches where the exact information differs.
Statement 3: The bike-share program includes 200 docking stations.
Answer: E (Not Given)
Why E is correct: Paragraph A says "200 bikes placed at ten docking stations." The passage tells you the number of bikes and the number of stations. It does not say "200 docking stations." Confusing the two numbers is a paraphrase trap.
Evidence mapping:
- Paragraph A: 200 bikes, 10 stations, pilot dates, mobile app, any-station return.
- Paragraph B: repair kiosk, damage reporting, maintenance lockout, weekly review.
- Statement 3's claim: in neither paragraph.
Method for information questions: Do not match a statement to a paragraph because the topic feels related. Match it only if the paragraph's words commit to that claim. When no paragraph does, choose E.
Worked example 4: Viewpoints (Part 4 style)
Original exercise. Read the opinion article and the reader comment below, then choose the correct option for each blank.
Opinion: Should the City Library Extend Sunday Hours?
The city library currently opens 10:00 a.m. to 4:00 p.m. on Sundays. A proposal would extend Sunday hours to 10:00 a.m. to 8:00 p.m. year-round, adding four hours and increasing the annual cost by $80,000.
Marion Zhao, Head Librarian: "I support a trial extension of Sunday hours from September to December this year. Circulation data shows Sunday afternoon demand, and a trial would let us measure evening use before committing to a permanent change. However, I would not support a permanent extension if it reduces weekday staffing."
Carlos Rivera, Downtown Resident: "I support extended Sunday hours, but only if the library maintains current weekday hours and staffing. My family uses the library Tuesday and Thursday evenings, and cutting those hours to fund Sundays would hurt more people than it helps."
Priya Kapoor, Budget Officer: "Extending Sunday hours year-round would require hiring two additional staff members at an annual cost of $80,000. I cannot recommend that investment until we have attendance data showing consistent evening demand. A trial could provide that data."
Reader Comment (with blanks to complete)
"I read the article about extending library hours on Sundays. Marion Zhao's position is that [BLANK 1], and Priya Kapoor opposes a permanent extension because [BLANK 2]."
Blank 1 options:
A. the library should extend Sunday hours permanently without a trial
B. a trial extension from September to December would let the library measure evening demand
C. the library should reduce weekday hours to fund the Sunday extension
D. extending Sunday hours is only worthwhile if circulation data already shows evening demand
Answer for Blank 1: B
Why B is correct: Marion Zhao says, "I support a trial extension of Sunday hours from September to December this year." She then adds that the trial would "let us measure evening use before committing to a permanent change." Option B accurately paraphrases both her position (trial) and her reason (measure demand).
Why each distractor fails:
- A fails on condition: Zhao explicitly supports a trial, not a permanent extension without one. She wants data first.
- C is the opposite of her condition: Zhao says, "I would not support a permanent extension if it reduces weekday staffing." She opposes cutting weekday hours, not supports it.
- D misattributes Priya Kapoor's reasoning to Marion Zhao: Kapoor is the one who says she cannot recommend the investment "until we have attendance data." Zhao supports a trial to gather that data; she does not say the data must already exist.
Blank 2 options:
A. the $80,000 annual cost is too high for any library program
B. the library has not shown consistent evening demand on Sundays
C. extending hours would require reducing weekday staffing
D. Marion Zhao has already recommended against a permanent change
Answer for Blank 2: B
Why B is correct: Priya Kapoor says, "I cannot recommend that investment until we have attendance data showing consistent evening demand." She opposes the permanent extension because that data does not yet exist. Option B captures her reasoning.
Why each distractor fails:
- A overstates her position: Kapoor objects to the investment "until we have attendance data." She is not saying $80,000 is inherently too high; she is saying she needs evidence first. The scope is too broad.
- C confuses Carlos Rivera's condition with Priya Kapoor's reasoning: Rivera says he supports extended hours "only if the library maintains current weekday hours." Kapoor's concern is attendance data and cost, not weekday staffing.
- D misattributes an actor: Marion Zhao's position is mentioned in Blank 1. Priya Kapoor's reasoning for opposing permanence is the demand for data, not Zhao's recommendation.
Viewpoint map for this passage:
| Person | Position | Reason | Condition / Qualification |
|---|---|---|---|
| Marion Zhao | Supports a trial Sept–Dec | Measure evening use before committing | Would oppose permanent extension if it reduces weekday staffing |
| Carlos Rivera | Supports extended Sunday hours | (Not explicitly stated in this excerpt) | Only if weekday hours and staffing stay the same |
| Priya Kapoor | Opposes permanent extension now | No attendance data yet showing evening demand | Open to reconsidering after a trial provides data |
When you map the viewpoints this way, you see that every person includes a condition. Wrong options often strip those conditions or assign one person's condition to another person.
Review your errors with a diagnostic log
Most people review a Reading mock by counting correct answers, then feeling either good or bad about the total. That tells you nothing about which pattern to fix.
A better review asks: why did the wrong option look right? Here are the six recurring error types and the drill that fixes each:
| Error type | What it looks like | How to diagnose it | Next drill |
|---|---|---|---|
| Locator | You chose an option because you remembered seeing related words, but the evidence for the correct answer was two paragraphs earlier or later. | Check where in the passage the correct answer's evidence actually appears. Did you look there? | Practice scanning for keywords in the question before reading the whole passage. Then locate the section that addresses that keyword. |
| Paraphrase | The correct answer said "postponed" and the passage said "moved to a later date," and you missed the match because you were looking for the word "postponed." | Write out the exact phrase in the passage that supports the correct answer. Does it use different words? | Drill synonym recognition: read an official explanation, underline the passage phrase and the answer option that matches it, and note the substitution (e.g., "moved to a later date" = "postponed"). |
| Scope | The passage said "some employees" and the wrong option said "all employees," or the passage gave a specific time frame and the option removed it. | Compare the correct answer and the wrong option you chose. Is one more general, more absolute, or wider in time than what the passage actually claims? | Practice writing down the exact scope of each option (who, when, how many, under what condition) and comparing it to the passage's scope. |
| Wrong speaker (Part 4 only) | You chose an option that accurately describes someone's view, but the question asked about a different person. | Check whose name is in the question. Then check whose name is attached to the position you selected. | Build a person-position-reason table for every Part 4 passage before you answer any questions. |
| Missed condition | The passage said "if X, then Y" and the option said "Y" without mentioning X. Or the passage said "unless," "only if," "provided that," or "as long as," and the option ignored it. | Underline every conditional word in the passage ("if," "unless," "only if," "provided that," "when," "until"). Check whether the option you chose includes that condition. | Mark conditional words while you read. When an option reuses a phrase from the passage but drops "if" or "unless," it is probably wrong. |
| Unknown word | A key word in the passage or in an option was unfamiliar, and you guessed. | Write down the word. Look it up. Check whether knowing it would have changed your answer. | Keep a vocabulary log of words that cost you marks. Review it before your next practice session. If the same word family appears in multiple tests (e.g., "subsidize," "subsidy," "subsidized"), prioritize learning it. |
Example log entry:
| Question | Part | Error type | Evidence in passage | Why I chose the wrong option | What to drill next |
|---|---|---|---|---|---|
| Q. 14 | 3 | Scope | Para. B says "some residents reported issues"; correct answer says "some residents." I chose "most residents" because I thought "reported issues" implied a majority. | The passage never quantifies beyond "some." I added an inference. | Practice reading every quantifier literally. "Some" does not mean "most." |
| Q. 22 | 4 | Wrong speaker | The question asks what Dr. Chen believes. I chose the option stating Dr. Martinez's position because both doctors discussed the same topic and I confused their names. | I matched the topic but not the speaker. | Write each speaker's name next to their position in my notes. Then match the question's name to my notes before answering. |
| Q. 8 | 2 | Missed condition | Workshop B is for ages 12–17. I chose it for an 18-year-old because I saw "Thursday evening" matched. | I checked day and time but skipped the age row in the diagram. | Check every row in a diagram. If the question gives three constraints, verify all three. |
Copy this table into your notes after each practice session. Fill in one row for each error. Then count how many rows have the same error type. That tells you which drill to prioritize.
Seven-day practice plan
Here is a sample schedule that moves from isolated parts to mixed practice. The plan assumes you have already completed at least one CELPIP Reading practice test as a baseline.
| Day | Focus | What to do | Time |
|---|---|---|---|
| Day 1 | Baseline review | Review your baseline test. Fill in the diagnostic log for every error. Count how many of each error type. | 30–45 min |
| Day 2 | Correspondence (Part 1) | Do 10 Part 1 practice questions. Focus on identifying purpose, tone, and logical reply content. Log errors. | 20 min |
| Day 3 | Diagram (Part 2) | Do 8 Part 2 practice questions. Build a checklist of constraints for each question (age, day, time, cost). Verify every constraint before choosing. | 20 min |
| Day 4 | Information (Part 3) | Do 9 Part 3 practice questions. Practice writing "stated" or "not stated" next to each statement before selecting A, B, C, D, or E. | 20 min |
| Day 5 | Viewpoints (Part 4) | Do 10 Part 4 practice questions. Build the person → position → reason → condition table before answering any questions. | 25 min |
| Day 6 | Mixed review | Do one question from each part without timing. Focus on applying the correct method for each part (correspondence: purpose; diagram: constraints; information: stated vs. inferred; viewpoints: map speakers). | 30 min |
| Day 7 | Fresh timed baseline | Complete a new full Reading practice test with timing. Compare the error count and error types to Day 1. | 50–60 min |
This plan practices each part in isolation before mixing them, because the method is task-specific. If Day 7's error log shows the same error type recurring, spend the next week drilling that error type's fix (see the diagnostic log table above).
The times are task time, not including review. Add 10 minutes after each day's drill to log errors.
Frequently asked questions
Should I read the questions first or the passage first?
It depends on the part. For Part 1 (Correspondence) and Part 4 (Viewpoints), read the passage first so you understand the situation or the structure of who says what. For Part 2 (Diagram) and Part 3 (Information), you can read the question first to know what you are locating, then find it in the diagram or passage. There is no universal rule; the task decides the method.
The official CELPIP-General Overview study pack gives task-specific tips for each part (checked November 8, 2026). It does not prescribe a single reading order for all four parts.
What should I do when I do not know a word in the passage?
Try to understand the sentence without it. Ask: does the sentence still make sense if I skip that word? If the unknown word is central to a question, look at the options. Eliminate the ones you know are wrong based on other evidence. Test the remaining options against what you do understand.
If the unknown word blocks you completely, mark the question and move forward. Spending three minutes on one vocabulary gap can cost you time on five questions you could answer. Come back if time remains.
Keep a log of words that cost you marks, and prioritize learning word families that appear multiple times (e.g., "allocate," "allocation," "reallocate").
How do I handle Part 3 (Information) when I am not sure if something is stated or just implied?
Part 3 asks whether the exact information in the statement appears in the text. According to the official CELPIP Reading Pro study pack (page 9, checked November 8, 2026), the Part 3 matching loop requires you to "Read that part of the paragraph carefully to make sure the exact information in the statement appears in the text" and to "Choose 'E' if the information is NOT stated in the reading."
Common error patterns that lead to E include superlative escalation (passage says "third most popular," statement says "most popular"), invented mechanisms (passage describes an animal's behavior, statement claims scientists understand why), or unstated causal claims (passage mentions a policy, statement claims a law was passed). If the passage does not explicitly state the information, the answer is E, even if the passage discusses a related topic.
Example: The passage says, "The library will trial extended Sunday hours from September to December." You read the statement, "The library will evaluate the trial results before making a permanent decision." That sounds logical, but unless the passage explicitly uses words that mean "evaluate the results" or "before a permanent decision," the answer is E — the information is not stated.
When stuck, ask: can I point to the words in the passage that state this exact information? If no, choose E.
In Part 4 (Viewpoints), how do I keep track of who believes what when there are multiple people?
Build a grid before you answer any questions. Draw three columns: Person, Position, Reason/Condition. Fill in one row per speaker as you read. When a question asks, "What is X's view?" or "Why does Y support Z?" check your grid instead of re-reading the passage.
For harder passages, there is often a fourth column: qualifications or conditions ("only if," "unless," "provided that"). Map those too, because the wrong options often strip a condition from a speaker's position.
If you skip this step and rely on memory, you will confuse who said what. Part 4's difficulty is not vocabulary; it is attribution.
Do absolute words like "always" and "never" mean the option is automatically wrong?
No. That is a common myth.
An option with "always" or "never" is correct if the passage makes an absolute claim. An option with "always" or "never" is wrong if the passage hedges with "some," "often," "may," or another qualifier.
Example: If the passage says, "All participants must register by Friday," then "Participants must always register by Friday" is a correct paraphrase for that situation. If the passage says, "Participants are encouraged to register by Friday," then "Participants must always register" is wrong because the passage does not require it in all cases.
Test every option against the passage. Do not use "always" or "never" as an automatic filter.
The CELPIP-General Overview study pack discusses absolute language and scope in its Reading tips (page 4, checked November 8, 2026).
What to do next
- Complete one CELPIP Reading practice test with timing if you have not done so already. That gives you a baseline error log.
- Fill in the diagnostic log (above) for every error. Count how many of each error type you made.
- Drill the part and the error type that costs you the most marks. Use the CELPIP Reading practice questions by part to isolate that part.
- After one week of part-specific drills, take a fresh timed baseline and compare the error counts.
The full CELPIP Reading format, including the score chart that maps raw scores to CELPIP levels, is covered in the CELPIP Reading practice test with answers guide. For Reading format and timing within the whole test structure, see CELPIP test format.
CELPIP Reading is not a guessing game. It is a discipline of locating evidence, testing all four options, and choosing the one that matches the passage's exact claim — including the actor, condition, time, and scope. The four worked examples above show that discipline in action.
Put it into practice
Try a free full Reading mock
Exam 1 is free in Speaking, Writing, Listening and Reading. Practice with the timers, then review your work. A free account includes one welcome AI feedback credit for Speaking or Writing; test access is separate.