The Role of Statistics in the Methodology of the Life Sciences

Sir Karl Popper has oHered a demarcation between science and pseudoscience: a discipline is a science only if the theories it entertains are universal and falsifiable. Its laws are those theories which to date have survived attempts at falsification. Is this view applicable to the life sciences or only to the so-called 'hard' sciences? If it is applicable, what is the effect of the interaction between biologist and statistician on the formulation of theories and on their falsification? May we properly identify the rejection of statistical hypotheses with the falsification of universal theories? In the life sciences, theories are often expressed as probability statements. Are probability statements falsifiable? The solution to the problem of obtaining falsification rules for probability statements, proposed by Gillies (1971, BritishJoumal of Philosophy of Science 22, 231-261), is discussed and its conflict with the Neyman-Pearson theory of hypothesis testing examined. A modified form of Gillies' solution is proposed. The questions considered are illustrated by means of a critical evaluation of a study by Bansal and Gupta (1978, Biometrics 34, 653-658) of the probabilities of survival of irradiated cells.