A statistical study of oral temperatures in school children with special reference to parental, environmental and class differences / by M.H. Williams and Karl Pearson.

  • Williams, Mary Hamilton.
Date:
1914
Catalogue details

Licence: In copyright

Credit: A statistical study of oral temperatures in school children with special reference to parental, environmental and class differences / by M.H. Williams and Karl Pearson. Source: Wellcome Collection.

Provider: This material has been provided by The Royal College of Surgeons of England. The original may be consulted at The Royal College of Surgeons of England.

    13/124 (page 13)
    affected by temporary short period variations and that the correlation we find between first and second temperatures is due either to physiological individuality, to permanent environmental conditions or to chronic pathological states. The accompanying Diagram I indicates graphically how little significance the interval between tempera- tures has either on their difference or on the variability of their difference. For the boys no standard deviation of an array, and no mean exceeds ± 2'5 times the probable error of a random sample. For the girls two standard deviations—those for 35 and 5'5 months’ interval—differ in opposite direction by more than 2‘5 times the probable error, and one mean that for 8'5 montlis. We should anticipate that a mean would fall outside these limits twice in our twenty trials *. The actual data are given in Tables XXI\ and XXV. We have looked at the matter from another standpoint, namely, by taking the differences of temperature vrithout regard to sign for each interval we obtained. Table D. Arithmetic Differences of Temperatures and Intervals. 1 Boys i Girls j Mean Arithmetic Mean Arithmetic | Differences of Temperatures Differences of Temperatures | 0-5 •42 + -03 — — 1-5 •29 + -03 2-25 •33 + -03 2-5 •33 + -02 3-5 ■39 + 04 3-5 •44 + -03 4-.5 •40+ -03 4-5 •41 + -03 55 •40 + -03 5-5 •31 + -02 6-5 •40 + -03 6-5 •37 -H -02 7-75 •35 + -03 7-5 •36 + -03 1 8-5 •38 + -03 10-4 [ -37 + -04 10-1 •27 + -04 1 11-75 •37 + 03 14-5 •34 + -04 15-9 •32 + 03 All Intervals •36 + -01 All Intervals •35 + -01 1 Having regard to the probable errors, there are practically no significant differences here. It looks as if the boys had the highest differences of temperatures for inter- mediate intervals, but this is only partially supported by the girls, and we conclude from this, as from the previous methods of examining the data, that it makes no substantial difference whether the temperature be taken a second time at an interval of two or of twelve months. Short period high or low persistent temperatures appear * If cr be the S.D. of the total population the probable errors have been found from •67449o-/\/« and •67449o-/x/2 n for mean and standard deviation res{)ectively, where n is the number of population in any array. The real values are somewhat different from these because o- is of course liable to error and as the array is part of the whole population, the more general formula, Biometrika, Vol. v. pp. 181—3, ought strictly to have been used, but the present method is sufficiently approximate here. By a slip of the draughts- man the last dot of Diagram I, Standard Deviations, Girls, has been placed at 14'5 instead of 15-9 months.