Karl Pearson and J. F. Tocher 
181 
The "Corrected" Deathrates in the table on p. 180 are based on the Age Groups 
of the total population for England and Wales in 1913. It will be seen that they 
modify entirely the order of the Crude Deathrates, which order appeared at first 
sight to be comparable with the cancer results, as showing a difference between the 
city and rural or urban areas — in this case indeed to the advantage of the former. 
Proceeding as in the case of cancer to compare significance of corrected deathrates 
when reduced to the general population of males (1913) as standard and when 
reduced to the standard population of maximum difference we find : 
Pair 
M' - M 
Q 
London and County Boroughs 
- 1-550 
•8898 
1-74 
4-10 
London and Urban Districts ... 
- 1-334 
-8794 
1-52 
3-16 
London and Rural Districts 
+ 0-151 
-8829 
0-17 
3-95 
County Boroughs and Urban Districts 
+ 0-216 
6727 
0-32 
2-34 
County Boroughs and Rural Districts 
+ 1-701 
-6998 
2-43 
5-76 
Urban Districts and Rural Districts . . . 
+ 1-485 
•6790 
2-19 
5-03 
Now with the single doubtful result for County Boroughs and Rural Districts 
none of the values of the ratio {M' — M)/aM'-M can be definitely considered as 
rendering M' — M significant. On the other hand all the differences of the corrected 
deathrates reduced to the standard population of maximum difference must, with 
one possible exception, i.e. County Boroughs and Urban Districts, be considered 
as markedly significant. This illustration is of great interest. For it is quite 
easy to select a standard population where for real age classes there is a significant 
difference between the " corrected" deathrates of London and the County Boroughs, 
but reduced to the general population (males) for 1913 there is no such difference. 
We see therefore that the reduction to an arbitrary standard population may be 
absolutely misleading as a means of testing whether two class deathrates are 
differentiated. Again the order of pairs for significance in the case of diabetes 
is for Q: 
Q 
Pair 
549 
County Boroughs and Rural Districts ... 
2-43 
503 
Urban Districts and Rural Districts 
2-19 
4-16 
London and County Boroughs ... 
1-74 
3-95 
London and Rural Districts 
0-17 
3-16 
London and Urban Districts 
1-52 
2-34 
County Boroughs and Urban Districts 
0-.32 
It will be seen that the significance of London and Rural Districts is considerably 
displaced by the general population as standard. Or, we conclude, as in the case 
of cancer, that relative significance as well as absolute is determined by the special 
standard population selected and we cannot accept the current view that the 
