1 7G 0)1 Criteria for the Existence of Differential Deathrates 
Male Cancer Statistics, England and Wales, 1913. 
il^t/ \JTHJl*.IJ 
(h) County 
Boroughs 
(c) Urban 
Districts 
{d) Rural 
Districts 
0—15 
3-86 
1-73 
2-38 
2-74 
15—25 
6-26 
3-88 
3-02 
3-77 
25—35 
1405 
13-64 
11-89 
11-20 
35—45 
44-30 
54-30 
42-21 
37-28 
45—55 
220-94 
200-21 
165-43 
127-20 
OO U 'J 
^ 1 '^.97 
0 lo L i 
.Joy-u / 
65—75 
961-48 
786-96 
799-57 
755-43 
75—85 
1109-93 
1043-48 
953-53 
976-27 
85 and over 
1295-22 
645-45 
991-71 
800-00 
Corrected 
deathrates * 
Crude 
deathrates 
111-48 
109-21 
101-01 
92-67 
91-97 
88-27 
82-24 
100-02 
In the first place we take the "corrected deathrates" reduced to the male 
population of 1913. We have: 
Pair 
M'-M 
(M'~M)l<rj^^,,-M 
Q 
London and County Boroughs 
10-47 
2-666 
3-93 
6-96 
London and Urban Districts ... 
19-51 
2-522 
7-74 
8-95 
London and Rural Districts ... 
29-24 
2-536 
11-53 
13-09 
County Boroughs and LTrban Districts 
904 
1-848 
4-89 
7-39 
County Boroughs and Rural Districts 
18-77 
1-953 
9-61 
12-25 
Urban Districts and Rural Districts... 
9-73 
1-829 
5-32 
6-76 
To begin with it will be seen that there is a very considerable difEerence in the 
values of {M' — M)juM'-M and Q, or the reduction to the general population of 
males gives nothing like the same intensity of significance to the differences 
between the means as the reduction to the standard populations of maximum 
difference. We are unable to determine what is the standard population of real 
maximum difference in any case|, and this very fact seems to discredit the use 
of the deathrates corrected to an arbitrary standard population as a means of 
adequately testing differences in mortality. For, although in this case all the 
differences of the deathrates corrected to the general male population are 
significant, in the next case — for example diabetes — they may not be, while the 
differences of the deathrates corrected to the standard population of maximum 
difference may be — as in the case of diabetes practically, they are — of significance. 
* The "corrected deathrates " are here reduced to the assumed male population of 1913 of all England 
and Wales. They differ therefore somewhat from the Registrar-General's "corrected deathrates " — 1 10-0, 
98-9. 00-2, 80-6 respectively, which are deduced from the general population of England and Wales, 1901. 
t Q may be deduced from an unreal population of maximum difEerence, i.e. one with some age classes 
negative. 
