W.M.F. Petrie to K. Pearson, 13th Nov. 1894.- External URL
- Correspondence Details
-
Sent From (Definite): Sir William Matthew Flinders PetrieSent To (Definite): Karl PearsonDate: 13 Nov 1894
- Current Holder(s)
-
Holder (Definite): University College London: Special Collections
- No links match your filters. Clear Filters
-
Sent fromSir William Matthew Flinders Petrie
13 Nov 1894
Description:
[from ‘Naqadah, Upper Egypt’:]
‘My dear Sir,
Thanks for your letter & paper, from which I see what a huge mass of material you must need to get any real result out of such tests. For instance I should have certainly suspected a mixture of two large curved in Plate 4 by the lump & hollow on the right side, but I see you conclude it is practically one. Is it quite certain that such tests, however beautiful mathematically take count fully of the facts?
I am very sorry that I shall not have the advantage of having a pronouncement from you on some of the questions raised in the “Pyramids”; but I leave early tomorrow for Egypt. I will as you suggest keep all the skulls however broken, if their dates are known. Do you know that C.H. Read (Brit. Mus.) was digging up dozens of skeletons[?] in fine condition lately & reburied them all? There was a good chance for a long list of measurements.
Have you tried as a test of the numerical results of mathematical treatment dividing your materially casually into two halves & treating each apart; by working out the result... & tabulating how the resulting elements vary in the smaller, but more checkable groups you would have a fine insight into the extent of variation in the result caused by casual accidents in the figures.
I have found that is a very profitable view to show what uncertainty attends the numerical results. E.g. a group of 100 obs. Should have each 3 times the error of the results from 900 obs. Together: & then uncertainty of the 100 obs. Can be seen by the differences between the results of the groups. It would be nice to see how many of the 12 places of figures would remain the same in some of the results, pp. 98, 102, &c.
I hope you will list up the asymmetrical physical curves & consider what causes may be compounded in them, like the ripple & whirl in the barometer curve. My interest in the matter is brutally practical, & not at all in your aetherial mathematics.
Yours very truly,
W.M. Petrie.’
-
Sent toKarl Pearson
13 Nov 1894
Description:
[from ‘Naqadah, Upper Egypt’:]
‘My dear Sir,
Thanks for your letter & paper, from which I see what a huge mass of material you must need to get any real result out of such tests. For instance I should have certainly suspected a mixture of two large curved in Plate 4 by the lump & hollow on the right side, but I see you conclude it is practically one. Is it quite certain that such tests, however beautiful mathematically take count fully of the facts?
I am very sorry that I shall not have the advantage of having a pronouncement from you on some of the questions raised in the “Pyramids”; but I leave early tomorrow for Egypt. I will as you suggest keep all the skulls however broken, if their dates are known. Do you know that C.H. Read (Brit. Mus.) was digging up dozens of skeletons[?] in fine condition lately & reburied them all? There was a good chance for a long list of measurements.
Have you tried as a test of the numerical results of mathematical treatment dividing your materially casually into two halves & treating each apart; by working out the result... & tabulating how the resulting elements vary in the smaller, but more checkable groups you would have a fine insight into the extent of variation in the result caused by casual accidents in the figures.
I have found that is a very profitable view to show what uncertainty attends the numerical results. E.g. a group of 100 obs. Should have each 3 times the error of the results from 900 obs. Together: & then uncertainty of the 100 obs. Can be seen by the differences between the results of the groups. It would be nice to see how many of the 12 places of figures would remain the same in some of the results, pp. 98, 102, &c.
I hope you will list up the asymmetrical physical curves & consider what causes may be compounded in them, like the ripple & whirl in the barometer curve. My interest in the matter is brutally practical, & not at all in your aetherial mathematics.
Yours very truly,
W.M. Petrie.’
-
