THE OPSONIC INDEX-A MEDICO-STATISTICAL 
ENQUIRY*. 
By W. F. HARVEY, M.A., M.B., D.P.H., Captain I.M.S. and ANDERSON 
M^KENDRICK, M.B., Oh.B., Captain I.M.Sf. 
Much misunderstanding seems to exist as to the relationship which should 
hold between the statistician, anxious to scrutinise the validity of inferences 
drawn from observations, and the observer of the facts. The latter looks with 
suspicion on the former and is inclined to doubt whether the intrusion of his 
fellow scientist into his domain is for any more worthy purpose than simply to show 
that he is totally wrong in his conclusions. The mathematician is apt to be impa- 
tient at the want of knowledge of his technical terms displayed by the observer and 
at the often openly expressed contempt for some of his most cherished elementary 
principles. But the statistical mathematician must get his data from the doctor 
— for it is of medical science in particular that we speak here — and the doctor at 
present is dependent on or would be wise to obtain the assistance of the mathema- 
tician for the refining of his conclusions. The doctor may also obtain much help 
from the criticism of the mathematician regarding the numerical soundness of his 
control observations. The statistician takes very little account of the non-quanti- 
tative data which bulk so largely in support of the claims of medical science. He 
takes merely certain of the characters which lend themselves easily to analysis, 
such as case-mortality or its complement case-recovery, attack-rate and the like 
and on this basis alone examines the validity of the conclusions arrived at. The 
degrees of correlation between such characters and particular methods of treatment 
are, through the controversies to which they have given rise, becoming familiar to 
medical men. But it must be pointed out that these correlations measure only 
the numerical aspect of single pairs of phenomena, admittedly important though 
these be. In some cases they do not measure the actual facts under discus- 
* From the Biometric Laboratory, University College, London, and the Vaecino- Therapy Laboratory, 
St Mary's Hospital, London. 
t In explanation of the joint authorship of this paper I should say that an earlier unpublished paper 
on the same subject by Captain McKendrick has been embodied in this paper. Being responsible for the 
wording of the paper I have used the pronoun " I " in the text. W. F. H. 
