Lectures on state medicine delivered before the Society of Apothecaries at their hall in Blackfriars May & June 1875 / by F. S. B. François de Chaumont.

  • de Chaumont, F. S. B. François.
Date:
1875
Catalogue details

Licence: Public Domain Mark

Credit: Lectures on state medicine delivered before the Society of Apothecaries at their hall in Blackfriars May & June 1875 / by F. S. B. François de Chaumont. Source: Wellcome Collection.

Provider: This material has been provided by London School of Hygiene & Tropical Medicine Library & Archives Service. The original may be consulted at London School of Hygiene & Tropical Medicine Library & Archives Service.

    214/238 (page 178)
    correspond to the known mathematical probabilities. The proportions may be seen from the diagrams ^; if the successive data depart very much from such pro- portions, then we may conclude that the data are not trustworthy. In this way Quetelet had no difficulty in pronouncing certain returns concerning French rejected conscripts to be false, and in determining the degree and direction of the error. Another way is the deter- mination of the probable error, and from that the value of the series, thus: v=~^, or the relative value is inversely as the square of the probable error. Were some of these methods followed we should not be treated, as we sometimes are, to such extraordinary vagaries in the shape of statistics, like some of those put forward by the opponents of the Contagious Diseases Acts. Some people in Germany, too, have even tried to prove by statistics that vaccination increases the liability to attacks of smallpox. Surely the ‘ farce of figures can no further go.’ I cannot do better than cite Quetelet’s four chief rules, which ought to be hung up in every room or office where figures are dealt with :— 1. ISTever have preconceived ideas as to what the figures are to prove. 2. Never reject a number that seems contrary to what you might expect, merely because it departs a good deal from the apparent average. 3. Be careful to weigh and record all the possible causes of an event, and do not attribute to one what is really the result of the combination of several.