Sampling and Representativeness

UCAT Decision Making questions sometimes ask you to evaluate whether a conclusion drawn from data is justified. A key factor is whether the sample (the group studied) is representative of the population (the group you want to draw conclusions about). This lesson teaches you to identify sampling bias, evaluate the strength of evidence, and recognise when data supports a conclusion — and when it does not.

---

Populations and Samples

Term	Definition	Example
Population	The entire group you want to draw conclusions about	All NHS patients in England
Sample	The subset of the population that was actually studied	500 patients at three London hospitals

The purpose of a sample is to learn something about the population without having to study every member. For this to work, the sample must be representative — it must reflect the characteristics of the population.

---

Types of Sampling Bias

Selection Bias

The sample is chosen in a way that systematically excludes certain groups.

Example: A survey of patient satisfaction conducted only during weekday mornings. This excludes working-age patients who attend evening or weekend appointments. The results may overestimate satisfaction (if older, retired patients are generally more satisfied) or underestimate it (depending on the population).

Self-Selection Bias (Volunteer Bias)

People who volunteer for a study may differ from those who do not.

Example: A survey sent to all staff asking about workplace stress. If only the most stressed staff respond, the results will overestimate the average stress level.

Survivorship Bias

Only the "survivors" (those who completed a process) are studied, ignoring those who dropped out.

Example: A medical school surveys its graduates and finds 95% are satisfied with their career. But this ignores those who left medical school before graduating — who may have been dissatisfied.

Recall Bias

Participants' memories of past events may be inaccurate or selective.

Example: Patients who developed cancer are asked about past dietary habits. They may overreport unhealthy behaviours because they are searching for explanations for their illness.

Healthy User Bias

People who engage in one healthy behaviour may also engage in others, confounding the results.

Example: People who take vitamin supplements may also exercise more, eat better, and avoid smoking. A study finding that supplement users are healthier may be measuring these other factors, not the supplements themselves.

---

Evaluating Whether a Conclusion Is Justified

When a UCAT question presents data and asks whether a conclusion follows, use this checklist:

Question	If "No" → conclusion may be unjustified
Is the sample large enough?	Very small samples may produce unreliable results
Is the sample representative of the population?	If the sample is biased, results may not generalise
Does the study measure what it claims to?	Proxy measures may not capture the true variable
Are there confounding factors?	Other variables may explain the observed pattern
Is the conclusion within the scope of the data?	Conclusions that go beyond the data are not supported

---

Sample Size

How Large Is "Large Enough"?

There is no fixed number that constitutes a "large enough" sample — it depends on the variability in the population and the precision required. However, some general principles apply: