RELATION BETWEEN PAY AND STANDARD OF LIVING 29 
Further, the method of gross correlation does not show what part 
schooling of the operator and home maker and composition or 
make-up of the family play in the formation of desires and demands 
in connection with family living. Many social relationships influ- 
encing expenditures of the farm family fall outside the field of 
schooling. Reading at home, visiting and social contacts of all 
persons composing the families with persons of other families 
through neighborhood groups and other groups have not been con- 
sidered. Ways in which the family reacts psychologically to its 
social and economic situations are given no consideration. 
Finally, the method of gross correlation gives no Suggestion as 
to the adequacy of expenditures for one year as a measure of the 
standard of living. Expenditures for the year of study may be ex- 
tremely nontypical for many of the families represented. As here 
used, expenditures do not signify a cash outgo but the value of all 
economic goods used. The proportion of these goods furnished by 
the farm may be abnormally high or low with many families. 
Quantities of certain goods purchased during the years immediately 
preceding the study may cause other variations. Marketing facili- 
ties may influence the expenditure for family living as well as the 
returns from farming. 
Owing to the complexity of all factors involved the method of 
gross correlation is ineffective as a means of determining the part 
of the total variation of all factors which could be accounted for 
by those factors used in the analyses and of selecting from the 
factors used those exerting the greatest influence or having the 
closest relation to the standard of living. The nearest approach 
to the solution of these problems is the method of multiple cor- 
relation, one of the results of which is expressed in the coefficient 
of determination. Through the use of this method of analysis the 
percentage of the total relation accounted for by the independent 
variable factors selected and the proportional contribution of the 
two sets of these factors, those indicative of the ability to pay and 
those indicative of the desires and demands on family living, were 
determined. 
Coefficients of determination were ascertained for all the inde- 
pendent variables shown in Table 12 with each of the three dependent 
variables; that is, with expenditures in each of the three terms of 
measurement. With expenditure per family as the dependent vari- 
able, it was found that 48 per cent of the total variation of all fac- 
tors could be accounted for by the factors used in the analysis. 
Three-fourths of this 48 per cent of variation which could be ac- 
counted for, or 36 per cent of the total variation, is attributable to 
criteria of the ability to pay and one- fourth, or 12 per cent of the 
total variation, is credited to factors indicative of the desires or 
demands of the family on family living. 
With expenditures per cost-consumption unit as the dependent 
variable, 42 per cent of the total variation could be accounted for 
by the independent variable factors used. Less than three-fourths, 
08 per cent of this 42 per cent accounted for, or 28.6 per cent of the 
total variation, is attributable to criteria of the ability to pay and 
the remaining 32 per cent, or 13.4 per cent of the total variation, 
is due to factors indicative of the desires or demands of the family. 
