A STATISTICAL METHOD FOR MEASURING “MARGINAL UTILITY” 181 
reach adjustment or equilibrium) will also have the same want- 
for-one-more unit of housing. 
This intercomparability is more truly applicable as between 
large groups of workingmen’s families, as revealed in the average 
family budgets, than as between two particular families. 
In fact, while it helped at the outset to picture three identical 
families, just as it helped to think of our measurements as in 
“pounds” or “square feet,” in practice—at least in the present 
state of our knowledge and statistics—we can scarcely expect 
to measure the wants of any individual family. Variability and 
chance enter in too much. To make any progress toward prac- 
tical measurements we must combine hundreds of families, and 
use only averages—albeit the “average family” is as mythical 
and non-existent, except as an average, as is the “economic man.” 
In this way we may hope to reach at least an approximate 
measure of man’s economic psychology in the mass. 
There is one field in which, without any guidance but common 
sense, we have expressed in figures the appraisal of mankind of 
the comparative value of money to people of different incomes. 
That field is taxation. Not only would it seem to all reasonable 
people unfair to assess the same number of dollars of taxes against 
the workingman as against the millionaire but to most people it 
would seem unfair to assess even the same rate per dollar of 
income. Even the philosophic doubter, if himself taxed unfairly, 
would be apt to know it! He would scarcely be satisfied if told 
that any comparison between his tax burden and others is mean- 
ingless because his mental phenomena and others’ are incom- 
mensurable. 
At any rate, whether justified or not, the method here set, forth 
does proceed on the assumption of commensurability, and my 
object in setting this forth is not so much to prove it correct as 
frankly to face it and point it out, as an assumption. 
Summary of Assumptions 
The following is a complete summary of the assumptions under- 
lying the second pair of equations, those on wantability. 
(a) Adjustment. The ‘budget groups used, like food, are 
assumed to be sufficiently subject to graduation in quantity and 
quality; and to be, in other ways, sufficiently adjustable, and 
adjusted, that the marginal dollar of an average or typical famaly