Weakening by Identifying Confounding Factors

Many LNAT arguments draw causal conclusions from correlational evidence — claiming that because two things occur together, one must cause the other. One of the most effective ways to weaken such arguments is to identify a confounding factor: a third variable that explains the observed relationship without the claimed causal link. This is one of the most frequently tested skills in LNAT Section A.

---

What Is a Confounding Factor?

A confounding factor (or confounding variable) is a variable that is related to both the supposed cause and the supposed effect, creating the appearance of a causal relationship where none may exist — or where the relationship is more complex than claimed.

Visual representation

Claimed relationship:     A ──causes──> B

With confounding factor:  C ──causes──> A
                          C ──causes──> B
                          (A and B appear related, but C drives both)

Classic Example

Claimed relationship: "Ice cream sales cause drowning deaths" (both increase together in summer).

Confounding factor: Hot weather increases both ice cream sales and swimming, and more swimming leads to more drowning.

Ice cream does not cause drowning. The confounding factor (hot weather) drives both variables.

---

Correlation vs Causation

The distinction between correlation and causation is fundamental to critical thinking and to the LNAT.

Concept	Definition	Example
Correlation	Two variables change together (both increase, both decrease, or one increases as the other decreases)	Countries with more hospitals also have higher death rates
Causation	One variable directly produces a change in the other	Smoking causes lung cancer
Confounded correlation	Two variables appear related because a third factor drives both	Countries with more hospitals have higher death rates because larger populations need more hospitals and produce more deaths — population size is the confounder

Key Principle: Correlation does not establish causation. Whenever an LNAT passage moves from "X and Y occur together" to "X causes Y", look for a confounding factor that could explain the relationship.

---

Common Confounding Factors in LNAT-Style Arguments

1. Socioeconomic status

Socioeconomic status (income, education, occupation) confounds a vast number of observed relationships.

Argument: "Children who are read to at bedtime perform better in school. Parents should read to their children to improve academic outcomes."

Confounder: Parents who read to their children tend to have higher education levels, higher incomes, and more time for parental involvement — all factors independently associated with children's academic performance.

The reading itself may contribute, but the correlation may also reflect broader socioeconomic advantages.

2. Self-selection

When people choose to participate in an activity, those who choose may differ systematically from those who do not.

Argument: "People who attend fitness classes are healthier than those who do not. Fitness classes improve health."

Confounder: People who choose to attend fitness classes are likely already more health-conscious, more physically active, and healthier than average. The classes may not cause better health — they may attract people who are already healthier.

3. Pre-existing differences

Groups being compared may differ in ways other than the variable of interest.

Argument: "Students at selective universities earn more after graduation. Selective universities provide a better education."

Confounder: Students at selective universities were already high-achieving before they enrolled. Their subsequent earnings may reflect their pre-existing ability, motivation, and family networks rather than the quality of education they received.

4. Time and historical change

Over time, many things change simultaneously, making it difficult to attribute outcomes to any single factor.

Argument: "Since the introduction of CCTV in city centres, crime has fallen. CCTV reduces crime."

Confounder: Over the same period, policing strategies changed, economic conditions improved, and demographic shifts occurred. The crime reduction may be due to any combination of these factors.

---

Worked Example 1

"A study found that employees who work flexible hours report higher job satisfaction and are 20% more productive than those who work fixed hours. Companies should therefore introduce flexible working to boost productivity."

Question: Which of the following, if true, would most weaken this argument?

A. Flexible working arrangements are expensive to administer. B. The employees who work flexible hours in the study were predominantly senior staff in creative roles who were already high performers — suggesting that flexible working may attract or be offered to productive employees rather than making employees more productive. C. Not all jobs can be done flexibly. D. Some employees prefer the structure of fixed hours.