A STATISTICAL METHOD FOR MEASURING “MARGINAL UTILITY” 175 
It is assumed that the same or similar grades exist in Oddland 
except that the corresponding step-up in cost is 33143% higher, 
r.e., is $1.3314 instead of $1.00. 
Since we are assuming that each grade costs Case 1 334% 
more than the same, or corresponding, grade costs Case 2, evi- 
dently Case 1, paying 3314% more than Case 2, may be said to 
be obtaining the same grade. A higher grade would be more than 
the 3314 % higher and a lower grade, less. 
Now if the two families are assumed to be so much alike in 
size and character, including tastes, education, and occupation, 
as well as in any other respect which might affect their want-for- 
one-more food unit or housing unit, as such units are above 
defined, the only essential difference between them being in the 
length of their purses, and the price levels in their respective 
markets, then, it seems reasonable to assume that their psycho- 
logical reactions to the same “physical” food rations, or to the 
same “physical” housing accommodation will be the same. That 
assumption is here made and, having made it, we need not be 
troubled by the fact that what has been called “physical” 
similarity cannot in the complexities of food and housing vari- 
ations be wholly disentangled from mental judgments. We may 
rest content with specifying that two food rations are sub- 
stantially equal, both physically and psychically, if the ratios 
th and ho are the same. That is, if these ratios are the same, 
we assume that either family, Case 1, or Case 2, would pronounce 
the two rations or bills of fare as practically the same, even 
though they were not absolutely identical. 
Food and Clothing Assumed Independent 
But I wish to call attention to an important assumption which 
is implied. This is that the want-for-one-more unit of food 
depends only on the food ration and not on the housing accom- 
modation, nor on any other circumstance differentiating Cases 1 
and 2; and, likewise, that the want-for-one-more unit of hous- 
ing depends only on the housing accommodation and not on the 
food ration, nor on any other circumstance differentiating Cases 
1 and 2. In other words: W; Fy, the want-for-one-more unit of 
food in Case 1, is assumed to be independent of all variables other 
than food itself; and so as to W, F.. Likewise W3 Rs, is assumed