A STATISTICAL METHOD FOR MEASURING ‘MARGINAL UTILITY’ 165 
market prices and index numbers of these. Let us suppose Fj, 
the food price level in Oddland, to be a third greater than F. 
(=1), the food price level in Evenland, or F;—$1.3314 per 
pound. 
That 1s, we divide W;F;=1 by F;—1.3314 and obtain 
W,=.75 of a “wantab.” 
We have calculated W,, the want-for-one-more dollar of the 
family called Case 1. This calculation has been made on the 
basis of data relating to food alone; but, in accordance with 
well known economic theory, we assume that the want-for-one- 
more dollar of a given family is the same as the want-for-one- 
more dollar’s worth of food, clothing, shelter or any other item 
of expenditure. 
The above process, or chain of calculations, by which W; is 
found from W, may be tabulated as follows: 
Given W; = 1 wantab 
Given F; = $1.00 
Multiplying, we get WF: = 1 wanta. 
This is same as WF, = 1 wantat 
Given F, = $1.333 
Dividing, we get W; = .75 of a wantab 
Want-for-one-more 
dollar in Case 2. 
Price Index of Food, 
Case 2. 
Want-for-one-more 
pound of food, Case 2. 
Want-for-one-more 
pound of food, Case 1. 
: Price Index of Food, 
Case 1. 
Want-for-one-more 
dollar, Case 1. 
We have now found W; from W.. We can next find Wj from 
W. analogously. Briefly: 
Given W. = 1 want~h 
PR- pie 
W.-R, 
= Want-for-one-more 
dollar, Case 2. 
Price Index of Rent, 
Case 2. 
Want-for-one-more 
square foot of housing, 
Case 2. 
Want-for-one-more 
square foot of housing, 
Case 3. 
Price Index of Rent, 
Case 3. 
= Want-for-one-more 
dollar, Case 3. 
1 wants 
Same as 
W3R; = 1 wantab 
Given 
Dividing. 
Wi: = .331 of a wantab 
wd 
ta 
We have now calculated W3, the want-for-one-more dollar of 
the family called Case 3. This calculation was made from house 
rent data, but of course represents the want-for-one-more dollar 
expended for anything else.