Читать книгу Sociology - Anthony Giddens - Страница 85

An example of how difficult it can be to establish causal relations involved in a correlation is given by the long history of studies of smoking and lung cancer. Research has consistently demonstrated a strong correlation between these two. Smokers are more likely to contract lung cancer than non-smokers, and very heavy smokers are more likely to do so than light smokers. The correlation can also be expressed the other way around. A high proportion of those who have lung cancer are smokers or have smoked for long periods in the past. There have been so many studies confirming these correlations that today it is generally accepted that a causal link is involved, but the exact causal mechanism is thus far largely unknown.

Researchers may want to know why some young Muslims in the UK wear headscarves while others do not. However, it can be difficult to establish a causal relationship between the various factors involved.

However much correlational work is done on any issue, there always remains some doubt about the possible causal relationship. Other interpretations of the correlation are at least theoretically possible. It has been proposed, for instance, that people who are predisposed to lung cancer are also predisposed to smoke. On this view, what causes lung cancer is not smoking per se but, rather, some in-built biological disposition to both smoking and cancer. Identifying causal relationships is normally guided by previous research into the subject at hand. If we do not have some reasonable idea beforehand of the causal mechanisms involved in a correlation, we would probably find it very difficult to discover what the real causal connections are. In short, we would not know what to test for.