CELPIP Reading Part 4 (Viewpoints): who thinks what

CELPIP Reading Part 4 is called Reading for Viewpoints. You read one article in which several named people give their opinions on a single issue, then answer 10 questions: multiple-choice questions about the article and its viewpoints, followed by a reader comment that reacts to the article and has drop-down blanks to complete. It is the last and heaviest read in the section, and it tests one skill: can you keep track of who believes what, and why, and apply that to a new piece of text.

The method that works: build a who-thinks-what map before you answer anything, use the multiple-choice questions to check that map, and fill the comment blanks by asking which viewpoint each sentence agrees or disagrees with.

The format: one article, several voices, two kinds of question

The article is built around a debate — a local policy, a workplace practice, a community decision — and presents the positions of several people, usually introduced by name and role. Typically one is for, one against, and one in between.

• The article: a few hundred words presenting multiple viewpoints on one issue. The named people are the key to everything that follows.
• Multiple-choice questions: what a particular person believes, why, what the writer implies, or where two viewpoints agree or clash.
• The reader comment: a short response from someone who has read the article. Several words or phrases are missing; each gap is a drop-down, and the right option must fit the grammar, the comment's logic, and the viewpoints in the article.

Officially, Part 4 has 10 questions, each worth the same as any other Reading question — and they come when fatigue and the clock both work against you.

Step one: build the viewpoint map

Do not start with question 1. Read the article once with a single goal: for each named person, know who they are, what side they are on, and the one reason or example they give.

1. Read the opening to get the issue. You cannot sort viewpoints until you know what question is on the table.
2. Each time a name appears, label it in five words or fewer: 'Okafor — for the change — cost.' 'Lindqvist — against — safety.' Jot the names on scratch paper with a plus, minus, or question mark.
3. Capture the reason, not just the side. Most questions do not ask whether someone supports the idea; they ask why, or what worries them.
4. Note the writer's own framing — a closing line often tilts the piece, and questions about what the writer implies depend on it.

By the end you should be able to say, without re-reading, what each person thinks and why. If you cannot, go back now — fixing the map once beats re-reading for every question.

Whose view is whose: the attribution trap

The most common Part 4 error is attaching the right opinion to the wrong person. The viewpoints overlap on purpose — two people may agree on the facts but disagree on what to do — and a distractor will often state a real viewpoint from the article, accurately, under the wrong name.

• When a question names a person, go back to that person's sentences specifically. Do not answer from your general sense of the debate.
• Check pronouns and reporting verbs. 'She argues' and 'he counters' mark a change of speaker; 'admits' or 'insists' tell you how strongly the view is held.
• Look for lines where one person responds to another. 'Unlike Okafor, Lindqvist believes…' gives you two positions and the relationship between them in one sentence.

An answer that is true according to the article can still be wrong — if the question asked about a different person.

Inference questions: what is implied, not stated

Some questions have no answer you can point to on the page: what a person would most likely think about a related situation, what the writer's purpose is, or what attitude a viewpoint reveals. Anchor every inference in something explicit — find the sentence the question is really about, then ask what it lets you conclude and no more. If someone calls the plan 'too expensive for a town this size', you can infer they would back a cheaper alternative; you cannot infer they oppose the idea in principle.

Answers one step from the text are usually right; answers three steps away, or that rely on your own opinion, are usually distractors. An inference never overturns a direct statement, so eliminate anything that contradicts a stated position. For questions about the writer, look at which viewpoint gets the most space and the last word.

The blank-completion technique for the reader comment

The comment is written by an imagined reader who agrees with one person, pushes back on another, or adds a personal experience. Each blank tests whether you know how the comment relates to the article — which is exactly what your map holds.

1. Read the whole comment once before choosing anything. Is the commenter persuaded, sceptical, or mixed? That stance rules out many options on its own.
2. For each blank, read the complete sentence, then find the viewpoint it refers to. 'I ____ with Lindqvist that safety comes first' can only take a word that matches what Lindqvist actually argued.
3. Decide whether the blank is logic or content. Connectors ('however', 'as a result') depend on the sentence before; content words must match a fact or opinion in the article.
4. Eliminate options that break the comment's own position. A commenter who opened by saying the article convinced them will not 'strongly reject' someone two sentences later without a 'but' or 'still'.
5. Re-read the filled sentence. All four options are usually grammatical; the right one keeps the comment consistent with the article and with itself.

If two options survive, prefer the precise match over the emphatic one. Distractors often overstate a view — 'demands' where the article had 'suggests'. Precision wins.

Timing: protect Part 4, do not rescue it

Reading runs on a single timer for all four parts. Part 4 comes last, holds 10 of the section's 38 questions, and needs the clearest head, yet most people reach it with the least time because Parts 1 and 2 felt easy and got overspent.

• Decide your Part 4 budget before you start: roughly a quarter of your total Reading time, in proportion to its share of the questions. Check the clock when Part 3 ends.
• Spend the first third of that budget on the article and your map. It feels slow; it is faster than re-reading.
• Answer the multiple-choice questions in order, returning to the named person's lines each time. If one drags, pick the best option and move on.
• Leave the comment for last. Its blanks are quick once the map is solid.
• Never leave a drop-down empty. CELPIP does not deduct marks for wrong answers, so a guess costs nothing and a blank guarantees zero.

The map-first approach is also the fastest, because it front-loads the thinking into one careful read. Practise Part 4 under a timer, first question to last blank, until the budget feels normal.

How Part 4 is scored

Part 4 has no separate score. Your answers feed into your overall Reading result on the CELPIP scale of 1 to 12, and each CELPIP level maps one-to-one to a Canadian Language Benchmark level — CELPIP 9 is CLB 9. Every Reading question carries the same weight; Part 4's ten are simply the ones most people leave to chance.

Read for the map, check each answer against the named person's own words, infer one step at a time, and fill the comment to stay consistent with the article. Do that until it is routine, and Part 4 becomes the part where you pull ahead.