Making Valid Inferences

Inference questions are among the most challenging in UCAT Verbal Reasoning. They ask you to go beyond what is explicitly stated in the passage — but only slightly. The critical skill is knowing where the boundary lies between a valid inference (something that logically follows from the passage) and an over-inference (something that might be true but is not supported by the passage). Getting this right is what separates high scorers from average ones.

---

What Is a Valid Inference?

A valid inference is a conclusion that must be true (or is very strongly supported) based on the information in the passage, even though it is not directly stated.

Category	Description	Example
Explicitly stated	The passage says it directly	"The study found a 15% reduction" → The study found a reduction
Valid inference	It logically follows from what the passage says	"All participants were aged 18–25" → No participants were over 25
Over-inference	It might be true but is not guaranteed by the passage	"The study was conducted at a university" → The participants were university students
Speculation	It is a guess with no basis in the passage	"The researchers were probably funded by the government"

The UCAT Standard: A valid inference is something that must logically follow from the passage. If it merely "could" be true or is "likely" true, it is not a valid inference — it is speculation, and the answer is "Can't Tell."

---

The Inference Boundary

This concept is fundamental. Think of the passage as establishing a set of facts. A valid inference stays within the logical boundaries of those facts. An over-inference steps outside them.

Example

Passage: "The company reported a 30% increase in revenue in 2022 compared to 2021. However, operating costs rose by 45% over the same period."

Statement	Category	Reasoning
"Revenue increased in 2022"	Explicitly stated	Directly said
"Operating costs grew faster than revenue in 2022"	Valid inference	45% > 30%, so costs grew faster
"The company's profits decreased in 2022"	Over-inference	Revenue up 30% but costs up 45% — profits probably decreased, but we do not know the relationship between revenue, costs, and profit without knowing the base amounts
"The company will report a loss in 2023"	Speculation	No information about 2023

The third statement is particularly instructive. Many candidates would select "True" because it seems mathematically obvious. But consider: if the company's revenue was £10 million and operating costs were £2 million, then a 30% increase in revenue (£3 million) would far exceed a 45% increase in costs (£0.9 million). The percentage increase is higher for costs, but the absolute increase could be higher for revenue. Without knowing the base figures, we cannot determine whether profits decreased.

Key Rule: A higher percentage increase does not necessarily mean a larger absolute increase. Be extremely cautious with mathematical inferences unless the passage provides enough data to make the calculation definitive.

---

Types of Valid Inference

1. Logical Entailment

If the passage states A, and A necessarily implies B, then B is a valid inference.

Passage: "All medical students at the university are required to complete a placement in general practice during their fourth year."

Valid inference: "A fourth-year medical student at the university who has not completed a GP placement has not fulfilled all requirements." — This follows logically from "all... are required."

Over-inference: "Medical students enjoy their GP placements." — We know nothing about their feelings.

2. Numerical Inference

When the passage provides numbers, certain mathematical conclusions are valid.

Passage: "The survey received 1,200 responses. Of these, 720 respondents agreed with the proposed change."

Valid inference: "More than half of respondents agreed with the proposed change." — 720/1,200 = 60%, which is more than half.

Over-inference: "The majority of the population supports the change." — The survey represents respondents, not the entire population.

3. Comparative Inference

When the passage compares two or more things, you can often infer the relative position of those things.

Passage: "Hospital A had a readmission rate of 12%. Hospital B had a readmission rate of 8%. Hospital C had a readmission rate of 15%."

Valid inferences:

• Hospital B had the lowest readmission rate of the three
• Hospital C had a higher readmission rate than Hospital A
• Hospital A's readmission rate was between those of Hospital B and Hospital C

Over-inference: "Hospital B provides the best care." — Readmission rate is one metric; "best care" is a broader claim.

4. Definitional Inference

If the passage defines a term or concept, you can infer things that follow from that definition.

Passage: "A renewable energy source is one that is naturally replenished on a human timescale, such as solar, wind, and hydroelectric power."