Causal links

Overview

These ideas are relevant to almost all the 'science in the making' issues that are an important part of the course. Statistical knowledge is not expected but an ability to check claims against a few simple qualitative criteria is. The ideas will be essential in evaluating many media stories about risk.

a Many of the complex risk issues in 9.2, health risks and 10.5, radiation, come into this category. So do the effects of medicines in 9.3, medical ethics.

b "Correlation does not equal causation" needs to be repeated and repeated at every opportunity. The web links give some examples that make the point.

c This links to an understanding of risk

d Students need to understand the concept of bias and how it might arise in selecting the sample. They need to recognise that the control and test groups can be selected by matched samples or by randomisation. The possible bias introduced in retrospective case control studies needs to be considered. This idea is particularly important in 9.2 health risks, 9.3 medical ethics in drug trials and 10.3 air quality.

e This will be relevant to 10.5 radiation and might also be introduced in 9.2 health risks, 9.6 genetic engineering and 10.3 air quality.

f Proving something is safe is often demanded by critics of a technology. Understanding this idea allows students to explain the need for a realistic evaluation of risk.

g Until there is a causative mechanism is found there is always the possibility that a correlation is due to a confounding variable. However there are times when one has to act on a correlation. This issue was behind the earlier debates on smoking and lung cancer in 9.2 health risks and current ones on risks from mobile phones or power lines in 10.5 radiation.