[Plea for the application of statistical methods by computer to stomatological research]

In face of the double reflex of fascination and defiance which is right away created by statistical methods by computers, the authors have attempted by means of precise examples to demystify these methods by recalling: 1) That computer statistics can be defined as the association of a service and a machine, the combination should be considered as an instrument and only an instrument. 2) That the application of computer statistics to odonto-stomatological research of course evidently enables the resolution of problems which would be otherwise insoluble, but above all considerably enlarges the field of research by rendering visible phenomena which would otherwise remain hidden. In this way, statistical methods by computer can be compared with the microscope; a low magnification with the optical microscope justifies only an overall view of the section while changing to higher magnifications permits the demonstration of details which would remain unsuspected. Use of the electromicroscope opens up another world for research. 3) In the same way as the biologist who uses the microscope, if he knows the principle of it is not in so far a specialist in optics, neither does the research woker who uses statistics necessarily have to be a trained statistician, on the one condition that evidently he disposes of professional computer logistics. This is the case of the ACB (Nantes) rush technical units, who cosigned this work, where the engineers who pose the problems to be resolved and emit the working hypotheses from the results supplied by the statistical studies are not the same people who carry out these very studies. 4) An essential fact, the research worker's independance is complete, only he can pose the problems and emit the hypothesis from the correlation, i. e. point of fact concomitance between the various isolated parameters. It therefore does not seem that weighty statistical techniques presently occupy the place of choice which should be theirs in stomatological research and more particularly in research concerning growth and development of the face where their use seems to be electively implied in face of the number and complexity of the parameters looked at.