The new bypass has not reduced congestion in the city centre. Traffic counts at the three main junctions are as high as they were before it opened, and the average commute across the centre has not shortened. Which of the following is the conclusion of the argument?

The new bypass has not reduced congestion in the city centre.

Traffic counts at the three main junctions are unchanged.

The new bypass has not reduced congestion in the city centre.

The average commute across the centre has not shortened.

The bypass should not have been built.

A publisher has decided to stop printing its weekly magazine and publish it only online, on the grounds that this will cut the magazine’s costs. Which of the following is an assumption on which the decision rests?

The cost of publishing the magazine online is lower than the cost of printing it.

The magazine’s readers would rather read it online than in print.

The cost of publishing the magazine online is lower than the cost of printing it.

Publishing online will increase the magazine’s readership.

The cost of paper and printing has risen in recent years.

A company found that employees who attended its optional wellness programme took 30% fewer sick days last year than employees who did not attend. It concluded that the programme improves employee health. Which of the following, if true, most weakens the conclusion?

Employees who were already fit and active were far more likely to sign up for the programme than those who were not.

The programme was offered free of charge to all employees.

Employees who were already fit and active were far more likely to sign up for the programme than those who were not.

The programme ran for six months rather than the full year.

Several attendees said they enjoyed the sessions and would attend again.

A city introduced a congestion charge in its central zone, and traffic entering that zone fell 18% over the following year. Officials concluded that the charge caused the fall. Which of the following, if true, most strengthens the conclusion?

Traffic in comparable zones of the same city, where no charge applies, was unchanged over the same year.

The introduction of the charge was reported widely in the press.

Traffic in comparable zones of the same city, where no charge applies, was unchanged over the same year.

The revenue from the charge was spent on improving public transport.

Many drivers complained that the charge was too high.

Every company in our survey that adopted the new scheduling software reported higher productivity within a year. Therefore any company that adopts the software will report higher productivity. The reasoning above is most vulnerable to which criticism?

It generalises to all companies from a group that chose both to adopt the software and to report on it, which need not be representative.

It assumes the very conclusion it sets out to establish.

It generalises to all companies from a group that chose both to adopt the software and to report on it, which need not be representative.

It treats a necessary condition as though it were sufficient.

It relies on the opinion of an authority rather than on evidence.

All the shortlisted candidates cleared the written test. Some of the shortlisted candidates also cleared the interview. Which of the following must be true?

Some candidates who cleared the written test also cleared the interview.

Everyone who cleared the written test was shortlisted.

Some candidates who cleared the written test also cleared the interview.

Everyone who cleared the interview was shortlisted.

Most of the shortlisted candidates cleared the interview.

A bank redesigned its mobile app to be far easier to use, and the number of customers banking through the app rose sharply. Over the same period, however, the number of calls to the bank’s telephone helpline also rose sharply. Which of the following best explains both facts?

The easier app brought in many first-time digital users, who then needed help with basic banking questions they had previously taken to a branch.

The bank promoted the new app heavily on television.

The easier app brought in many first-time digital users, who then needed help with basic banking questions they had previously taken to a branch.

The bank increased the number of staff answering the helpline.

Some customers continue to prefer visiting a branch in person.

Since the metro raised its ticket prices, ridership has fallen by 12%. The fare rise is therefore driving commuters away from the metro. Which of the following is an assumption of this argument, rather than a premise of it or an inference from it?

No other significant change affected commuters’ travel decisions over the same period.

Ridership on the metro has fallen by 12%.

No other significant change affected commuters’ travel decisions over the same period.

The metro will collect less revenue overall than it did before the fare rise.

Metro fares had not been raised for several years before this increase.

Critical Reasoning — Arguments, Assumptions and What Must Follow

An argument is a claim plus the evidence offered for it, and almost every question in this family is answered by first separating the two. Once you can point at the conclusion and at the premises, each of the six question types has one test. An assumption is something the argument needs but does not state — so negate the candidate answer, and if the argument collapses, that was the assumption. To weaken, attack the link between evidence and claim, not the evidence itself. To strengthen, close the gap the argument left open — most often by ruling out a rival explanation. A flaw is a named mistake, and four names cover most of them. An inference is what must be true, which is a far smaller set than what is likely to be true, and the commonest wrong answer in the whole family is a statement that is plausible, agreeable and not entailed. Read every option asking one question: does the argument need this, does it follow from it, or is it merely consistent with it? Those three are different, and the exam is built on the difference.

Take the Argument Apart First

Every argument in this section has at most three parts, and finding them takes seconds once you know what to look for. The conclusion is the claim the author wants you to accept. The premises are the reasons offered. The gap is what the premises do not quite establish — and the gap is where every question in the family lives.

Signposts, and what usually follows them
Word	What follows it
therefore, so, hence, thus, it follows that, clearly	The conclusion. If the passage has one of these, the work is already done.
because, since, as, given that, for	A premise. The clause it introduces is evidence, never the claim.
although, however, but, while it is true that	A concession — the author is granting something before pressing on. The conclusion is usually on the other side of it.
No signpost at all	Ask which sentence the others are there to support. That one is the conclusion, whatever its position — it is often the first sentence, not the last.

⚠️ The conclusion is not the last sentence

A passage that opens "The new bypass has not reduced congestion" and then gives two traffic statistics has stated its conclusion first and its evidence after. Candidates who take the last sentence as the conclusion get this whole family wrong in a predictable way, and the papers know it.

Assumptions, and the Negation Test

An assumption is an unstated premise the argument cannot do without. That definition gives a mechanical test, and it is worth applying literally rather than by feel:

Negate the option. Assume the opposite of what it says.
Re-read the argument under the negation. Does the conclusion still follow from the premises?
If the argument collapses, the option was an assumption. If the argument survives — even weakened — it was not. An assumption is necessary, not merely helpful.

Worked example. A publisher will stop printing its weekly magazine and publish only online, on the grounds that this will cut its costs. Test the option "the cost of publishing online is lower than the cost of printing". Negate it: online costs the same or more. Then the plan cannot cut costs and the argument is dead — so it is an assumption. Now test "readers prefer reading online". Negate it: readers prefer print. The plan still cuts costs; it may lose readers, but the argument about costs stands. Not an assumption. That is the entire method, and it disposes of the "sounds relevant" options that make this question type feel subjective.

🧠 Assumption or inference?

They point in opposite directions. An assumption is something the argument stands on — it comes before the conclusion. An inference is something that stands on the argument — it comes after. If an option would have to be true for the author to reason as they did, it is an assumption; if it would be true given that the author is right, it is an inference. Options of both kinds are routinely offered under one question stem.

Weaken and Strengthen — Attack the Link, Not the Data

Most arguments in this section move from a correlation to a cause, and that move is the gap. Two facts occur together, and the author declares that one produced the other. Three rival explanations are always available, and every weakening answer is one of them: the causation runs the other way; a third factor produced both; or the pairing is a selection artefact — the group being compared chose itself.

What weakens, and what only looks as though it does
Option says	Effect
The group that showed the benefit had chosen to join it	Weakens hard. Self-selection: the outcome may have caused the participation rather than the other way round.
A comparable group with no intervention changed in the same way	Weakens hard. Something common to both groups explains the change; the intervention is not needed.
A comparable group with no intervention did not change	Strengthens hard. This is a control group, and it removes the rival "something else did it" explanation.
Some of those affected disliked the measure	Neither. Approval is not the claim. An unpopular measure can work perfectly.
The measure was expensive, or widely publicised, or ran for six months	Neither. These are true of the situation without touching whether the cause is the cause. Background detail is the commonest decoy in this family.

Named Flaws, and What Must Follow

When a question asks for the flaw, the answer is usually one of four, and recognising the shape is faster than reasoning from scratch:

Correlation taken for causation. Two things moved together, so one caused the other.
An unrepresentative sample. The evidence comes from a group that chose itself, or is too small, or was surveyed where the answer was likeliest. Note that a sample of people who volunteered a report is doubly selected.
Necessary confused with sufficient. "You cannot pass without practice" does not mean practice will get you through.
Circularity. The conclusion appears among the reasons for it, usually in different words.

For an inference question the bar is entailment: the answer has to be true whenever the passage is true, with no exception you can construct. Three habits catch almost every wrong option. Watch the quantifier — "some" does not license "most", and "most" does not license "all". Watch direction — "all shortlisted candidates cleared the test" says nothing about whether everyone who cleared the test was shortlisted, and swapping the two is the single commonest error here. And distrust anything that adds a cause, a motive or a recommendation, because a passage of facts entails none of the three. The borrowed analytical-reasoning chapter’s Venn method is the tool for the formal cases; this is the reading discipline for the rest.

Key takeaways

Separate the conclusion from the premises before reading a single option. The conclusion is often the first sentence, not the last.
For an assumption, negate the option: if the argument collapses it was an assumption, and if it merely weakens it was not. An assumption is necessary, not helpful.
An assumption comes before the conclusion and an inference after it. Both are offered under the same stem, so decide which direction the question is asking in.
Nearly every argument here goes from correlation to cause. Weaken it by reversing the causation, naming a third factor, or showing the group selected itself; strengthen it with a control group that did not change.
Cost, popularity, publicity and duration are background. They feel relevant and change nothing about whether the cause is the cause.
Four flaws cover most of the type: correlation for causation, an unrepresentative sample, necessary confused with sufficient, and circularity.
An inference must be true whenever the passage is. Check the quantifier, check the direction of an "all" statement, and reject any option that adds a cause, a motive or a recommendation.

Practice questions (8)

Attempt each one before opening the answer. Every explanation names the tempting wrong option as well as the right one, because that is where marks are lost.

The new bypass has not reduced congestion in the city centre. Traffic counts at the three main junctions are as high as they were before it opened, and the average commute across the centre has not shortened.

Which of the following is the conclusion of the argument?
1. Traffic counts at the three main junctions are unchanged.
2. The new bypass has not reduced congestion in the city centre.
3. The average commute across the centre has not shortened.
4. The bypass should not have been built.
Show answer
Answer: B — The new bypass has not reduced congestion in the city centre.
The two statistics are offered in support of the first sentence, which makes the first sentence the conclusion — a passage may state its claim before its evidence, and this one does. Options A and C are the premises; a premise is never the conclusion, however factual it sounds. Option D is a recommendation, and the argument makes none: it says the bypass did not work, not that building it was a mistake, and the difference matters because a road can fail at one purpose while serving another.
A publisher has decided to stop printing its weekly magazine and publish it only online, on the grounds that this will cut the magazine’s costs.

Which of the following is an assumption on which the decision rests?
1. The magazine’s readers would rather read it online than in print.
2. The cost of publishing the magazine online is lower than the cost of printing it.
3. Publishing online will increase the magazine’s readership.
4. The cost of paper and printing has risen in recent years.
Show answer
Answer: B — The cost of publishing the magazine online is lower than the cost of printing it.
Apply the negation test. Negate B: online publishing costs the same as printing, or more. Then the move cannot cut costs and the stated ground for the decision is gone — so B is an assumption. Negate A: readers prefer print. Costs still fall; the publisher may lose readers, but the cost argument survives, so A is not an assumption. C is a hoped-for benefit the argument never claims. D would explain the publisher’s motive, and a motive is not a premise: even if printing costs had been flat for a decade, the decision would still cut costs so long as B holds.
A company found that employees who attended its optional wellness programme took 30% fewer sick days last year than employees who did not attend. It concluded that the programme improves employee health.

Which of the following, if true, most weakens the conclusion?
1. The programme was offered free of charge to all employees.
2. Employees who were already fit and active were far more likely to sign up for the programme than those who were not.
3. The programme ran for six months rather than the full year.
4. Several attendees said they enjoyed the sessions and would attend again.
Show answer
Answer: B — Employees who were already fit and active were far more likely to sign up for the programme than those who were not.
The programme was optional, so the two groups were not assigned — they chose. If the already-healthy were the ones who signed up, then good health explains the attendance rather than attendance explaining the good health, and the 30% gap would have appeared with no programme at all. That is self-selection, and it attacks the causal link the conclusion depends on. A is background: free access says nothing about who took it up. C might mean the effect is understated rather than absent. D is about enjoyment, and a programme can be thoroughly enjoyable without improving anyone’s health.
A city introduced a congestion charge in its central zone, and traffic entering that zone fell 18% over the following year. Officials concluded that the charge caused the fall.

Which of the following, if true, most strengthens the conclusion?
1. The introduction of the charge was reported widely in the press.
2. Traffic in comparable zones of the same city, where no charge applies, was unchanged over the same year.
3. The revenue from the charge was spent on improving public transport.
4. Many drivers complained that the charge was too high.
Show answer
Answer: B — Traffic in comparable zones of the same city, where no charge applies, was unchanged over the same year.
B supplies a control group. The obvious rival explanation for the 18% fall is that something city-wide reduced driving everywhere — fuel prices, a downturn, a change in working patterns. If comparable zones without the charge did not change, that rival is ruled out and what is left is the charge. A and C are true of the policy without bearing on causation; C in fact introduces a second possible cause, so it slightly hurts rather than helps. D is about how the charge was received, and complaint is compatible with a charge that worked exactly as intended — arguably it is evidence that drivers noticed it.
Every company in our survey that adopted the new scheduling software reported higher productivity within a year. Therefore any company that adopts the software will report higher productivity.

The reasoning above is most vulnerable to which criticism?
1. It assumes the very conclusion it sets out to establish.
2. It generalises to all companies from a group that chose both to adopt the software and to report on it, which need not be representative.
3. It treats a necessary condition as though it were sufficient.
4. It relies on the opinion of an authority rather than on evidence.
Show answer
Answer: B — It generalises to all companies from a group that chose both to adopt the software and to report on it, which need not be representative.
The sample is selected twice over. Only companies that chose to adopt the software are in it — firms whose problems the software fits are likelier to buy it — and only those that chose to report appear in the survey, and a company that saw no gain has little reason to file a report. Neither group is a random draw from all companies, so nothing about "any company" follows. A is wrong: the conclusion is not among the premises, it is over-extended from them. C names a different error — there is no necessary-condition claim here at all. D is wrong because the argument cites survey data rather than an authority; that the data is badly selected is a separate objection from appealing to authority.
All the shortlisted candidates cleared the written test. Some of the shortlisted candidates also cleared the interview.

Which of the following must be true?
1. Everyone who cleared the written test was shortlisted.
2. Some candidates who cleared the written test also cleared the interview.
3. Everyone who cleared the interview was shortlisted.
4. Most of the shortlisted candidates cleared the interview.
Show answer
Answer: B — Some candidates who cleared the written test also cleared the interview.
Take the candidates the second sentence names: they are shortlisted, and the first sentence says every shortlisted candidate cleared the written test. So those people cleared both, which is exactly B. A reverses the first statement — "all shortlisted cleared the test" leaves the test-clearers who were not shortlisted entirely unaddressed. C makes the same reversal with the interview: the passage tells us nothing about anyone outside the shortlist, so there may or may not be interview-clearers who were never shortlisted. D upgrades "some" to "most", and "some" is satisfied by a single candidate.
A bank redesigned its mobile app to be far easier to use, and the number of customers banking through the app rose sharply. Over the same period, however, the number of calls to the bank’s telephone helpline also rose sharply.

Which of the following best explains both facts?
1. The bank promoted the new app heavily on television.
2. The easier app brought in many first-time digital users, who then needed help with basic banking questions they had previously taken to a branch.
3. The bank increased the number of staff answering the helpline.
4. Some customers continue to prefer visiting a branch in person.
Show answer
Answer: B — The easier app brought in many first-time digital users, who then needed help with basic banking questions they had previously taken to a branch.
A good resolution explains both halves of the discrepancy with one mechanism, and B does: the same improvement that raised app usage also changed who the users were, and newcomers to digital banking generate calls that experienced users do not. Note that the calls need not be about the app at all — the mechanism works even if every call is a basic banking question. A explains the rise in app usage and says nothing about calls. C explains how the bank could handle more calls, not why more were made — capacity is not demand. D is consistent with both facts and explains neither, which is the shape of the most tempting wrong answer in this question type.
Since the metro raised its ticket prices, ridership has fallen by 12%. The fare rise is therefore driving commuters away from the metro.

Which of the following is an assumption of this argument, rather than a premise of it or an inference from it?
1. Ridership on the metro has fallen by 12%.
2. No other significant change affected commuters’ travel decisions over the same period.
3. The metro will collect less revenue overall than it did before the fare rise.
4. Metro fares had not been raised for several years before this increase.
Show answer
Answer: B — No other significant change affected commuters’ travel decisions over the same period.
The argument moves from "fares rose and ridership fell" to "the fares caused the fall", and that move needs the field to be clear of other causes. Negate B — a new metro line opened elsewhere, or a fuel subsidy made driving cheap, or offices moved to remote working — and the 12% has another explanation, so the conclusion no longer follows. B is therefore the assumption. A is a premise, stated in the passage. C is an inference someone might draw from the conclusion, and a shaky one: fewer riders paying a higher fare can yield more revenue, not less. D is background about fare history; the argument works identically whether the last rise was last year or a decade ago.