182 On Criteria for tlie Existence of Differential Deathrates 
"standard population" to which the deathrates are corrected is a matter merely 
of convenience. 
It is clear that there is nothing in the "corrected" deathrates to indicate in any 
group a substantial difference from the rate 11-883 per 100,000 for all England 
and Wales*. It will accordingly be of some interest in this case to test whether 
this result of non-significance which flows from the deathrates corrected to the 
general population is confirmed when more accurate methods are applied to test 
the distribution of deviations as well as mean rates. The following are the values 
of and xo^ '■ 
Pair 
London and County Boroughs ... 
London and LTrban Districts 
London and Rm-al Districts 
County Boroughs and Ui'ban Districts 
County Boroughs and Rural Districts 
Urban Districts and Rural Districts ... 
17-3188 
10-0126 
15-5939 
5-4693 
30-1470 
25-2912 
18-0702 
9-9831 
15-4501 
5-4236 
30-4870 
25-8263 
It will be seen as in previous cases that there is no substantial difference between 
Q" and xo^- 
The following table gives the deathrates from Diabetes per 100,000 exposed to 
risk in each age group of the four categories : 
Male Diabetes Statistics, England and Wales, 1913. 
Age 
Group 
(a) London 
(6) County 
Boroughs 
(c) Urban 
Districts 
(d) Rural 
Districts 
0—15 
1-23 
1-22 
1-11 
2-09 
15—25 
3-13 
3-48 
4-08 
5-30 
25—35 
5-51 
6-39 
6-78 
6-55 
35—45 
7-16 
8-22 
7-89 
8-87 
45—55 
22-94 
16-02 
17-38 
12-90 
55—65 
44-89 
47-70 
53-00 
38-29 
65—75 
68-99 
106-77 
90-06 
76-10 
75—85 
84-97 
108-99 
94-36 
76-39 
85 and over 
0 
112-28 
81-51 
10-96 
The problem before us is again whether the above distributions of deathrates 
as a whole are or are not significantly different, and not whether the differences 
in one of their statistical constants (i.e. the "corrected" deathrate) are or are 
not significant. We shall apply only three tests, i.e. (i) the test or Goodness 
of Fit Test, (ii) that of the difference of corrected deathrates for the standard 
* The deviations from the General Rate are (a) - -784, (b) + -7G0, (c) + -550 and {d) - -935, all 
of the order of the probable errors of the differences. 
