A biometric analysis of some insemination-labour and menstrual-labour curves in certain mammalia / by D. Berry Hart.

  • David Berry Hart
Date:
[1913?]
Catalogue details

Licence: In copyright

Credit: A biometric analysis of some insemination-labour and menstrual-labour curves in certain mammalia / by D. Berry Hart. 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.

    16/24 page 10
    With regard to “ smoothing ” observations, I am always against it, my view being that the observations treated (as here) by least squares will look after themselves better than any plan we can adopt. But if we want to save calculation and (as in the present case) reduce * the amount of numbers to be employed (say) by grouping in 48 hours or in 4 days, then the plan is (for 4 days) to take every adjacent 4 days possible. Thus Keid, p. 406, I should treat as below. The general theory of probability allows for no exceptional cases, i.e., it expects to involve every case, however bad we may think it. But if we know of any reason attached to a particular case which makes it obviously unreliable, then, of course, cut it out. At the same time Chauvenet (an American mathematician) has devised a criterion by which a test can be applied for cutting out doubtful observations. However, it practically amounts almost to a recalculation for each observation cut out, and is too laborious to be used in actual observations.—Yours sincerely. 1° plot all these 4-day points, then 2° use judgment in cutting out very astray ones, and calciTlate with the rest. But, after all, I prefer taking the actual observations themselves. / 260 6 and so on omittiny 4, 5, 8. J. D. Hamilton Dickson.
    digitised image 1
    digitised image 2
    digitised image 3