Читать книгу Statistics in Nutrition and Dietetics - Michael Nelson - Страница 21

Intervention trials are designed to prove that phenomena which are true in the population are also true in a representative sample drawn from that population.

Compare the relative impact of two iron preparations in the treatment of anaemia.

This may sound similar to the statement that was made under ‘Experimental Studies’. The two statements are different, however. In the intervention trial, we would try to ensure that the two groups in which we were comparing treatments were similar to each other and similar to the population from which they were drawn. In the experimental study, we chose a group of subjects, measured the exposure and outcome and other characteristics of the group, and assumed that if the outcome was true in that group, it would be true in the population with similar characteristics. These differences in approach and logic are subtle but important.

In practice, the aim of most studies is to find evidence which is generalizable to the population (or a clearly defined subgroup). The relationship between the type of logic used and the generalizability of the findings is discussed below. The limitations of inductive logic and their resolution are discussed lucidly by Popper [1, pp. 54–55].