U. S. DEPARTMENT OF AGRICULTURE 
lute income. It was understood that some members of the middle 
class had incomes larger than some members of the well-to-do class. 
In establishing the classification of the family the interviewer was 
guided by the housewife's statement of the occupation of the hus- 
band or wage earner of the family and by other observable factors, 
such as the residential area in which the home was located and the 
general characteristics and appearance of the home itself. The 
housewives interviewed were selected at random. 
In the negro or colored group only the poor and middle classes 
were interviewed. These classifications corresponded to those 
defined for the American white group. 
In the foreign group, division was made by nationality only. No 
attempt was made to set up income or standard of living groups. 
Effort was made, however, to obtain for each group questionnaires 
from families representative of the entire group. 
NUMBER OF QUESTIONNAIRES 
Four thousand four hundred and sixty-six housewives were inter- 
viewed in the 16 cities. The members of the households which these 
housewives represented totaled 20,744 persons. (Table 1.) 
The distribution of the questionnaires by city and by group and 
class is shown in Table 2. It will be noted in this table that for the 
most part the questionnaires for any one class totaled about 50 for 
each city, although there was some variation from this number. It 
has been found in previous surveys of somewhat similar character 
that a group of about 50 replies gathered at random was sufficiently 
representative of one population class in a single city. This basis of 
collecting questionnaires resulted in obtaining about 200 in each city 
when all the four classes of the American white group were inter- 
viewed. If any colored or foreign groups were surveyed, the total 
number of questionnaires for the city was increased correspondingly. 
Housewives in 2,912 American white households were interviewed. 
Of this total, 751 were of the poor class, 782 of the middle class, 796 
of the well-to-do class, and 583 of the wealthy class. In the colored 
group, 203 housewives of the poor class and 197 of the middle class 
were interviewed, a total of 400. In the foreign group, there were 
interviewed 100 English housewives, 45 Finnish, 50 French, 175 
German, 276 Italian, 113 Jewish, 195 Polish, 50 Russian, and 150 
Scandinavian, a total of 1,154 housewives for the entire group. 
Particular attention is called to the fact that all 16 cities were not 
represented in all population groups or group subdivisions. However, 
in each table in which are presented the results of this investigation, 
it should be understood that, unless otherwise stated in the heading 
or footnotes of the table, the cities included in the data were those 
listed in Table 2. 
The number of replies for several questions sometimes varied from 
that indicated in Table 2. These differences were partly accounted 
for by the fact that some of the questions were not answered because 
of inability of the interviewers to obtain satisfactory replies from all 
housewives for all questions. Other factors which contributed to 
the variation in the number of replies were the changes made in the 
questionnaire after the survey had been completed in Baltimore, 
Binghamton, New Haven, Philadelphia, and Pittsburgh, and partially 
